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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
e0b05ae0-4f39-4f40-8ac2-aa183541b4f2 | intrinsic-bayesian-optimisation-on-complex | 2301.12581 | null | https://arxiv.org/abs/2301.12581v1 | https://arxiv.org/pdf/2301.12581v1.pdf | Intrinsic Bayesian Optimisation on Complex Constrained Domain | Motivated by the success of Bayesian optimisation algorithms in the Euclidean space, we propose a novel approach to construct Intrinsic Bayesian optimisation (In-BO) on manifolds with a primary focus on complex constrained domains or irregular-shaped spaces arising as submanifolds of R2, R3 and beyond. Data may be collected in a spatial domain but restricted to a complex or intricately structured region corresponding to a geographic feature, such as lakes. Traditional Bayesian Optimisation (Tra-BO) defined with a Radial basis function (RBF) kernel cannot accommodate these complex constrained conditions. The In-BO uses the Sparse Intrinsic Gaussian Processes (SIn-GP) surrogate model to take into account the geometric structure of the manifold. SInGPs are constructed using the heat kernel of the manifold which is estimated as the transition density of the Brownian Motion on manifolds. The efficiency of In-BO is demonstrated through simulation studies on a U-shaped domain, a Bitten torus, and a real dataset from the Aral sea. Its performance is compared to that of traditional BO, which is defined in Euclidean space. | ['Claire Miller', 'Mu Niu', 'YuAn Liu'] | 2023-01-29 | null | null | null | null | ['bayesian-optimisation'] | ['methodology'] | [-9.64317843e-02 2.47310132e-01 5.79298377e-01 5.84819913e-02
-3.86415124e-01 -2.75895655e-01 8.44860613e-01 -3.66588444e-01
-5.41086137e-01 5.73071063e-01 7.78291523e-02 -1.26934335e-01
-7.58782268e-01 -7.15411603e-01 -5.40562212e-01 -1.19225943e+00
-4.23213780e-01 6.32526159e-01 3.10514029e-02 2.23992631e-01
3.37006629e-01 6.93859696e-01 -1.12746286e+00 -9.90898728e-01
7.38854170e-01 5.12822390e-01 3.26105356e-01 6.29372895e-01
1.84401646e-01 -1.84924498e-01 -3.36735487e-01 -1.25108734e-01
2.66321331e-01 -1.35168219e-02 -4.48388010e-01 4.44784284e-01
-2.42944956e-01 2.49192029e-01 -2.93662786e-01 1.06302381e+00
3.09904635e-01 6.40233219e-01 1.27641785e+00 -7.17772007e-01
-2.34427348e-01 -1.73434671e-02 -4.59845275e-01 1.15855604e-01
-1.60259902e-01 8.08370952e-03 4.21656013e-01 -1.08686578e+00
2.22327411e-01 1.62884581e+00 4.96751577e-01 2.80008554e-01
-1.56135929e+00 1.08072609e-01 -6.85588568e-02 -2.44795263e-01
-1.71101213e+00 -1.98919967e-01 7.09590197e-01 -9.06263590e-01
3.89706671e-01 1.75291702e-01 3.88129741e-01 7.41909862e-01
4.33076143e-01 2.59916425e-01 8.97912204e-01 -2.28324607e-01
7.26330757e-01 -1.04771167e-01 8.71139392e-02 3.33561450e-01
2.99219936e-01 1.69066399e-01 -2.15592295e-01 -4.88589942e-01
8.58745813e-01 -1.72057971e-01 -3.07998240e-01 -6.35203838e-01
-1.42080593e+00 1.06183326e+00 1.86761394e-02 8.77683237e-02
-5.00277579e-01 -1.15417294e-01 -2.07027718e-01 -2.29516149e-01
6.10658169e-01 3.14776361e-01 -4.02477711e-01 -2.63463974e-01
-8.04966390e-01 8.05812627e-02 1.09997141e+00 9.31021392e-01
7.15266466e-01 1.83887079e-01 3.78876150e-01 6.24921620e-01
1.04322946e+00 8.73165011e-01 1.69973925e-01 -1.02340341e+00
2.40368277e-01 2.04995543e-01 2.95133173e-01 -8.21445882e-01
-4.76915061e-01 -1.93533972e-01 -1.12275040e+00 1.45428047e-01
6.09823108e-01 -3.21284354e-01 -6.86274350e-01 1.53903425e+00
8.77604961e-01 3.32859159e-01 2.92995811e-01 8.54450047e-01
2.89325684e-01 8.16068172e-01 -3.02545846e-01 -1.15728244e-01
1.31993055e+00 -2.26099432e-01 -4.72498983e-01 7.09855855e-02
5.73914289e-01 -4.70213801e-01 5.75518966e-01 4.37881261e-01
-7.30706871e-01 -2.50866473e-01 -7.57233799e-01 5.23600221e-01
-2.93885499e-01 1.89598352e-02 6.53976202e-02 9.47906792e-01
-9.35654044e-01 6.71246171e-01 -1.17739105e+00 -3.40272397e-01
1.31716162e-01 3.12676996e-01 -3.97002250e-01 2.31692374e-01
-8.92422557e-01 8.29610348e-01 1.70214504e-01 5.83031714e-01
-8.52490127e-01 -5.90447187e-01 -8.88113439e-01 -2.89084524e-01
7.50905229e-03 -6.02181911e-01 4.45572346e-01 -1.98623925e-01
-1.86624753e+00 3.64993036e-01 9.51722357e-03 -6.74272850e-02
4.87616956e-01 -2.01957732e-01 -1.58560544e-01 6.55880347e-02
-4.50437963e-02 2.31864512e-01 1.21338844e+00 -8.93740654e-01
-1.48383537e-02 -6.97028399e-01 -1.65631801e-01 1.80948868e-01
1.85287088e-01 -1.50535358e-02 2.30475850e-02 -7.18390942e-01
5.40497839e-01 -1.27720118e+00 -5.28667986e-01 -3.78308982e-01
-4.31032807e-01 -2.23330542e-01 8.47662926e-01 -7.19276547e-01
8.13688755e-01 -2.35352206e+00 8.59224081e-01 4.33109194e-01
-1.07172698e-01 -8.38757902e-02 9.57399160e-02 2.76562750e-01
8.99625272e-02 1.92075416e-01 -7.59372115e-01 -2.42045462e-01
7.73694292e-02 3.11425298e-01 1.55073637e-02 1.24310076e+00
4.61104363e-01 1.66076466e-01 -8.71456742e-01 -3.19284052e-01
2.72632688e-01 6.39129758e-01 -2.76376635e-01 1.07144028e-01
-3.90478708e-02 7.58619308e-01 -7.79851675e-01 2.45251253e-01
8.19356382e-01 9.20895189e-02 -3.19656521e-01 2.07421795e-01
-1.66004062e-01 -1.72061637e-01 -1.67645526e+00 1.54235423e+00
-1.61842182e-01 4.27898854e-01 5.53783417e-01 -9.52926695e-01
1.33010006e+00 2.57939994e-01 5.19622564e-01 2.08151862e-01
-4.21751626e-02 6.45814091e-02 1.59160152e-01 -3.35478246e-01
3.52754563e-01 -2.37150520e-01 1.57379970e-01 2.85263628e-01
-6.72184750e-02 -3.95289540e-01 -2.81194597e-01 2.80778725e-02
1.00767219e+00 5.17802596e-01 -1.41160972e-02 -9.75363553e-01
7.05244303e-01 -3.08173537e-01 5.63534319e-01 3.73237044e-01
-6.56185951e-03 6.29283488e-01 3.71904075e-01 -2.48780042e-01
-1.00804484e+00 -1.35555935e+00 -9.61427808e-01 3.79580617e-01
7.61725977e-02 -1.49225786e-01 -7.33750045e-01 -3.28394622e-01
-9.16492268e-02 6.92196727e-01 -6.95904791e-01 -4.84299250e-02
-4.40731138e-01 -1.38227785e+00 2.03400910e-01 -1.89720154e-01
3.64729404e-01 -6.43943727e-01 -6.19978130e-01 3.14491153e-01
8.51816982e-02 -8.93860877e-01 -3.37197512e-01 3.56167033e-02
-1.24467480e+00 -9.59908068e-01 -7.72073090e-01 -2.42220283e-01
7.60276556e-01 -7.29335472e-02 6.59131169e-01 -9.12675619e-01
-2.72027433e-01 9.73138273e-01 -1.43050715e-01 -3.77509028e-01
-3.14145356e-01 -3.45923036e-01 4.09238398e-01 5.62047184e-01
4.29632545e-01 -5.28055012e-01 -4.47605968e-01 8.99954498e-01
-8.84726107e-01 -5.49229443e-01 2.06604719e-01 9.04789507e-01
3.53698790e-01 2.72143185e-01 4.43886012e-01 -8.89623910e-02
3.20025057e-01 -9.65450823e-01 -9.55502987e-01 -1.63750604e-01
-1.74313545e-01 2.96288788e-01 -1.95097715e-01 -6.71613753e-01
-9.25235569e-01 9.16710198e-02 4.27093834e-01 -3.03379744e-01
-2.28604555e-01 6.11170888e-01 -3.24115336e-01 1.14244223e-01
5.13248086e-01 1.61022767e-01 3.44731688e-01 -7.02556849e-01
3.03269356e-01 8.07948351e-01 1.81663319e-01 -9.06369567e-01
8.27473640e-01 7.69389510e-01 6.51081562e-01 -1.58102572e+00
-2.30909586e-01 -4.98610198e-01 -1.12873185e+00 -2.84373045e-01
1.03539300e+00 -6.70903265e-01 -4.46420848e-01 4.21908766e-01
-1.12987006e+00 -3.94553959e-01 -2.90688604e-01 1.02236676e+00
-8.01473081e-01 4.27938044e-01 -1.34155288e-01 -1.28997660e+00
1.00892514e-01 -1.10870194e+00 1.08879566e+00 -1.25007033e-01
-8.16321075e-02 -1.52831149e+00 4.46455359e-01 7.60100409e-02
2.57391751e-01 4.71348166e-01 6.45379126e-01 -3.73121291e-01
-5.42496383e-01 -9.23076645e-02 2.24813864e-01 5.42258918e-01
2.56282181e-01 2.09313884e-01 -7.60319889e-01 -2.66005844e-01
8.08045268e-01 3.64560217e-01 3.84262025e-01 7.57091165e-01
4.86440867e-01 -7.18710795e-02 -2.77573258e-01 4.98142153e-01
1.12101221e+00 7.71909207e-02 5.79919815e-01 3.27372998e-01
4.72362012e-01 1.03159380e+00 6.57465935e-01 5.76589167e-01
2.95550436e-01 6.70168459e-01 5.37245512e-01 3.67531598e-01
4.80703831e-01 3.57502848e-01 5.59950292e-01 8.52636397e-01
-2.14555949e-01 3.51625025e-01 -1.00918388e+00 6.90280318e-01
-1.79838240e+00 -6.44217491e-01 -4.47671920e-01 2.69446230e+00
4.48427051e-01 -2.56130815e-01 2.75993999e-02 -1.02730833e-01
1.09099412e+00 -1.40818015e-01 -3.49318087e-01 -1.23575494e-01
-1.31485656e-01 -1.87238216e-01 6.78445399e-01 7.30082989e-01
-1.11111200e+00 2.32896447e-01 5.87736130e+00 7.45861232e-01
-6.91435635e-01 6.22627586e-02 3.86377424e-01 1.95591375e-01
1.33254919e-02 2.64054965e-02 -8.60071719e-01 6.79537952e-01
1.27503479e+00 1.98713705e-01 4.15686369e-01 6.30639493e-01
5.82188964e-01 -2.58954883e-01 -8.83489668e-01 8.42941761e-01
-2.01921448e-01 -6.69356823e-01 -3.92679006e-01 8.64171445e-01
5.72604179e-01 2.26487160e-01 3.47634740e-02 -8.40410963e-02
2.99405038e-01 -8.46272528e-01 4.60938632e-01 1.02808261e+00
4.81368512e-01 -6.31974936e-01 5.89307606e-01 5.68707585e-01
-8.68257582e-01 3.88767362e-01 -6.46189630e-01 2.64088541e-01
2.01230243e-01 6.99100494e-01 -8.74309778e-01 6.15823746e-01
7.90232778e-01 6.65907919e-01 -2.35012844e-01 1.15647411e+00
2.08668515e-01 8.04845691e-01 -8.91739249e-01 8.50536302e-03
4.04097795e-01 -1.22547066e+00 1.47427714e+00 1.00462723e+00
7.19491959e-01 6.10914007e-02 -2.80433774e-01 1.02146256e+00
5.56328535e-01 -1.16769791e-01 -9.21815574e-01 5.33628762e-02
1.05194792e-01 1.31940031e+00 -5.81240892e-01 1.71552047e-01
-3.13997120e-01 3.47943068e-01 -2.55832464e-01 5.18375933e-01
-5.26926219e-01 -2.67182797e-01 7.06288397e-01 1.31277323e-01
5.50221026e-01 -6.14669025e-01 7.32518733e-02 -1.01390445e+00
-1.07016832e-01 -2.13001579e-01 3.35092992e-02 -5.16934216e-01
-1.38568270e+00 4.43359941e-01 4.03360039e-01 -1.23891807e+00
-4.42969687e-02 -8.21370482e-01 -4.92068738e-01 1.30086863e+00
-9.29980040e-01 -7.91748047e-01 8.81969556e-02 5.43657005e-01
1.66586250e-01 -4.23212051e-01 7.27055848e-01 -1.43758401e-01
-5.02662301e-01 -4.29784685e-01 8.46273482e-01 -4.80999649e-02
1.36595055e-01 -1.59387600e+00 2.88230330e-01 7.48079062e-01
-2.09292024e-01 6.51791036e-01 7.91995823e-01 -6.62886143e-01
-1.63901091e+00 -1.12142408e+00 1.42484158e-01 -7.59184182e-01
1.02586329e+00 -4.48226899e-01 -9.65130985e-01 5.52066624e-01
-3.29887182e-01 5.59742302e-02 7.65398622e-01 -9.54642221e-02
5.37786074e-02 1.21368498e-01 -1.32834947e+00 3.97105694e-01
4.34669763e-01 -4.84705456e-02 -8.39026511e-01 3.63350034e-01
4.09119278e-01 3.01344786e-02 -1.30667210e+00 4.54647541e-01
2.08536565e-01 -3.66159737e-01 9.27227080e-01 -5.64035356e-01
-3.33092958e-01 -6.09137058e-01 -4.51348215e-01 -1.46403873e+00
-2.16699660e-01 -1.16241276e+00 -3.72287408e-02 1.04192817e+00
4.25542742e-02 -8.78312051e-01 7.46341467e-01 5.90997517e-01
1.35810031e-02 -3.89336944e-01 -1.69806528e+00 -1.04679477e+00
4.11444038e-01 -3.48781049e-01 3.70840341e-01 6.72365367e-01
-1.44595131e-01 1.93601459e-01 -1.42485365e-01 6.50636435e-01
1.22261155e+00 -5.46409488e-01 8.84069681e-01 -1.91765237e+00
-2.23850816e-01 -1.71486557e-01 -7.67889857e-01 -9.68927324e-01
2.07748443e-01 -4.39464331e-01 3.17530960e-01 -1.03492153e+00
-3.47209871e-01 -6.78769171e-01 2.59211034e-01 -5.59643745e-01
-2.57773418e-02 -3.34282637e-01 -2.61398584e-01 3.61375451e-01
-6.46224841e-02 1.08864844e+00 1.10413826e+00 -3.05772237e-02
-2.88086712e-01 3.36973161e-01 9.43319779e-03 8.77223611e-01
6.68067038e-01 -5.85239410e-01 -1.91851199e-01 -2.46624183e-02
1.89665370e-02 4.10209037e-02 3.46655011e-01 -8.54204535e-01
5.69322985e-03 -1.36968344e-01 -1.88596159e-01 -6.47930503e-01
6.44263506e-01 -1.08197439e+00 4.50257540e-01 1.21806838e-01
3.02205205e-01 -6.95042536e-02 -2.24846527e-02 1.19409573e+00
-1.30701333e-01 -5.28510690e-01 9.14756298e-01 -4.68040667e-02
-6.22937875e-03 1.20087951e-01 -4.17628169e-01 -1.35959238e-01
1.10132992e+00 -2.77420402e-01 -4.23809476e-02 -1.38987809e-01
-1.11671174e+00 1.29255757e-01 3.99169266e-01 5.38900793e-02
7.45956182e-01 -1.13851511e+00 -9.90796208e-01 1.45575568e-01
-1.80419639e-01 5.35208642e-01 3.65336910e-02 1.09587944e+00
-5.38519561e-01 3.99899065e-01 2.62314796e-01 -1.01742196e+00
-8.01174700e-01 4.87488568e-01 5.45481920e-01 1.93230584e-01
-6.79372132e-01 3.20111543e-01 4.10483003e-01 -8.12210441e-01
-5.84182562e-03 -2.37993464e-01 -2.04682395e-01 -1.40360981e-01
2.63339758e-01 7.77724087e-01 -1.91625923e-01 -1.10773313e+00
-1.37192622e-01 6.02123022e-01 7.29281604e-01 -5.05673766e-01
1.38633192e+00 -4.37574685e-01 -5.42690277e-01 7.29006708e-01
7.80663431e-01 -9.80090946e-02 -1.60678256e+00 -3.27263087e-01
2.95179486e-01 -3.30183595e-01 2.72122771e-01 5.49277812e-02
-5.80919981e-01 8.53694916e-01 4.47633445e-01 3.41642767e-01
4.67273980e-01 4.02822718e-02 -2.44636655e-01 3.38136584e-01
4.61439818e-01 -8.32769096e-01 -1.72189087e-01 4.99280810e-01
1.06857049e+00 -8.33343446e-01 -1.21249259e-01 -3.36193234e-01
-2.92221636e-01 1.22642219e+00 -2.24917121e-02 -3.05891067e-01
1.36212409e+00 -1.50144979e-01 -3.52728844e-01 -1.17636636e-01
-3.85543048e-01 -1.93996076e-02 3.06691170e-01 8.41470242e-01
-8.31767321e-02 1.38301656e-01 -1.03326086e-02 1.41160578e-01
-6.55177087e-02 -5.60765266e-01 7.64147520e-01 8.04946125e-01
-4.55540359e-01 -7.41473913e-01 -1.15638506e+00 3.19490701e-01
-3.58759522e-01 6.05216697e-02 2.11148262e-01 7.49089301e-01
-1.75350368e-01 1.10824537e+00 3.36533546e-01 6.82693049e-02
3.60650122e-02 -4.01750170e-02 1.47983536e-01 -5.64852536e-01
7.21696541e-02 4.71640229e-01 -6.97311834e-02 -1.02350317e-01
-5.26326239e-01 -1.28405392e+00 -7.13912070e-01 -2.68668961e-03
-4.55918372e-01 6.49544597e-01 1.26476908e+00 9.20423150e-01
2.44418979e-01 6.75074458e-02 6.68674469e-01 -1.27233207e+00
-7.35092163e-01 -1.52651346e+00 -1.16536009e+00 -2.12619841e-01
5.19217491e-01 -1.03127122e+00 -9.72391605e-01 1.20577164e-01] | [6.830724716186523, 3.9220240116119385] |
0aeddb47-4efc-4aff-a553-172eb2a9a63a | persona-guided-planning-for-controlling-the-1 | null | null | https://aclanthology.org/2022.naacl-main.245 | https://aclanthology.org/2022.naacl-main.245.pdf | Persona-Guided Planning for Controlling the Protagonist’s Persona in Story Generation | Endowing the protagonist with a specific personality is essential for writing an engaging story. In this paper, we aim to control the protagonist’s persona in story generation, i.e., generating a story from a leading context and a persona description, where the protagonist should exhibit the specified personality through a coherent event sequence. Considering that personas are usually embodied implicitly and sparsely in stories, we propose a planning-based generation model named ConPer to explicitly model the relationship between personas and events. ConPer first plans events of the protagonist’s behavior which are motivated by the specified persona through predicting one target sentence, then plans the plot as a sequence of keywords with the guidance of the predicted persona-related events and commonsense knowledge, and finally generates the whole story. Both automatic and manual evaluation results demonstrate that ConPer outperforms state-of-the-art baselines for generating more coherent and persona-controllable stories. Our code is available at https://github.com/thu-coai/ConPer. | ['Minlie Huang', 'Jian Guan', 'Jiaxin Wen', 'Zhexin Zhang'] | null | null | null | null | naacl-2022-7 | ['story-generation'] | ['natural-language-processing'] | [ 1.51100799e-01 4.62872237e-01 -1.12471886e-01 -5.04337490e-01
-6.03874147e-01 -6.28239334e-01 1.25205886e+00 -2.90525388e-02
3.09468620e-02 6.57314718e-01 1.13921094e+00 3.79682958e-01
2.07615614e-01 -8.57475698e-01 -6.95196986e-01 -1.61224484e-01
4.34830457e-01 6.91980302e-01 -2.81421930e-01 -2.84976661e-01
3.37492943e-01 2.08535213e-02 -1.27481782e+00 5.80346048e-01
7.86471426e-01 3.90111983e-01 3.77237767e-01 5.91664851e-01
5.37014604e-02 1.34097552e+00 -8.45849037e-01 -8.85079324e-01
-2.40903243e-01 -1.10869920e+00 -9.58198071e-01 5.28282523e-01
4.22480702e-02 -2.18304083e-01 -4.32073921e-01 6.85211837e-01
3.12351167e-01 2.98822343e-01 9.93118942e-01 -1.10033619e+00
-7.25361884e-01 1.64645088e+00 -2.64943331e-01 -1.94226488e-01
8.65533710e-01 2.91301638e-01 1.11262882e+00 -7.36182094e-01
1.10724127e+00 1.39983475e+00 3.30173969e-01 5.94186723e-01
-1.08737123e+00 -3.49696785e-01 1.85582653e-01 1.77137688e-01
-1.28402960e+00 -4.41781759e-01 1.21030617e+00 -5.21825671e-01
6.82909429e-01 2.97638178e-01 1.01612008e+00 1.76156509e+00
-3.89554016e-02 9.36555982e-01 6.50674939e-01 -3.02320778e-01
5.95230535e-02 1.10748999e-01 3.59388851e-02 3.94372940e-01
-1.63502768e-01 -2.12167338e-01 -9.87311900e-01 -6.25084564e-02
7.21225739e-01 -1.55945942e-01 -1.15766287e-01 1.49167433e-01
-1.81114960e+00 7.06974089e-01 2.15984192e-02 3.45345914e-01
-8.29639733e-01 2.47899026e-01 4.43966657e-01 -2.23263383e-01
1.72156811e-01 9.06815469e-01 1.63176894e-01 -4.15011555e-01
-9.04997945e-01 1.02999783e+00 1.01471508e+00 1.25935197e+00
2.04565257e-01 -1.06418170e-01 -8.94587636e-01 6.78389668e-01
-8.79452005e-02 3.06489289e-01 3.03968668e-01 -9.93420601e-01
4.67287540e-01 6.69822276e-01 5.09297550e-01 -1.11537385e+00
-1.53793260e-01 -3.34116966e-01 -5.70694506e-01 -3.91134351e-01
2.91381598e-01 -1.89961299e-01 -4.40280765e-01 1.78326035e+00
2.01666355e-01 4.01408188e-02 1.22936249e-01 8.76312673e-01
9.51283157e-01 1.02093291e+00 4.86674383e-02 -4.16159898e-01
1.62170792e+00 -1.01246750e+00 -8.36562455e-01 -3.23572546e-01
2.46676773e-01 -6.03667080e-01 1.45808053e+00 2.37300366e-01
-1.32122314e+00 -4.82376903e-01 -8.83179665e-01 -2.18468666e-01
2.48982564e-01 5.34894407e-01 4.25857395e-01 -9.84545797e-02
-2.99062490e-01 5.01801610e-01 -5.33361614e-01 -6.52163684e-01
3.24711233e-01 -4.60990280e-01 1.55470982e-01 1.11412473e-01
-1.26208806e+00 5.87167621e-01 6.50218010e-01 -3.91749531e-01
-1.21012437e+00 -7.20012128e-01 -9.01007116e-01 1.00778742e-02
6.04827464e-01 -8.23102176e-01 1.42332995e+00 -6.83164954e-01
-1.46265626e+00 1.02314007e+00 -3.58663738e-01 -6.24203146e-01
8.63403320e-01 -4.04006690e-01 -3.85885328e-01 -1.99048687e-02
3.95952672e-01 6.79004252e-01 5.68202972e-01 -1.14590108e+00
-5.24666190e-01 2.39753455e-01 1.84467450e-01 5.26919246e-01
-9.93765444e-02 4.27889675e-01 -4.87228602e-01 -7.85529792e-01
-2.71423668e-01 -6.16603434e-01 -7.44261295e-02 -6.62271500e-01
-1.11214745e+00 -5.09140193e-01 3.57324064e-01 -6.56846881e-01
1.42390418e+00 -2.03532004e+00 3.63261700e-01 -2.62524188e-01
9.56556275e-02 -4.52189714e-01 1.35308757e-01 9.51740086e-01
2.16619998e-01 5.44062667e-02 -3.40360887e-02 -4.46327418e-01
3.05694789e-01 -3.30206454e-01 -5.96452236e-01 -4.44318401e-03
5.82894273e-02 1.03507113e+00 -1.09257233e+00 -6.55699611e-01
-1.93257090e-02 3.74641955e-01 -3.17421675e-01 4.56755608e-01
-7.58433521e-01 3.93029690e-01 -4.38935727e-01 3.31868798e-01
-1.68653116e-01 -3.86765540e-01 1.98251709e-01 -2.08250836e-01
-1.62036493e-01 7.95891285e-01 -7.96344757e-01 1.66781080e+00
-8.76026899e-02 5.28834999e-01 -7.16899037e-01 -2.40374729e-01
1.14068985e+00 4.05246198e-01 -1.53977145e-02 -2.21601408e-02
1.18449219e-01 -2.53089488e-01 -1.50591567e-01 -5.48687756e-01
7.40728676e-01 -2.60440975e-01 -8.45982671e-01 8.03925812e-01
-2.52821110e-02 -6.61050260e-01 5.19201338e-01 5.90337753e-01
8.07130039e-01 4.18368876e-01 7.45929420e-01 -3.32618244e-02
1.71242073e-01 2.68788278e-01 6.12175941e-01 9.47748244e-01
1.47292659e-01 6.34296238e-01 8.44943225e-01 -4.41892326e-01
-1.30054832e+00 -8.76485407e-01 3.91150951e-01 7.32101798e-01
8.51427466e-02 -8.60495150e-01 -8.40127647e-01 -5.14379203e-01
-4.43592072e-01 1.73257363e+00 -8.07390153e-01 4.50889282e-02
-6.48570538e-01 -1.59586802e-01 4.97572213e-01 2.66718268e-01
5.75599015e-01 -1.57074261e+00 -7.70656049e-01 4.09887791e-01
-7.35476911e-01 -1.10783589e+00 -8.96453977e-01 -4.30419534e-01
-4.90955412e-02 -6.53312802e-01 -4.14702415e-01 -5.06152272e-01
5.50857723e-01 -1.08033940e-01 1.21717966e+00 -2.29565829e-01
2.12563127e-01 3.81249972e-02 -4.77025658e-01 -6.54477537e-01
-6.40452504e-01 2.07754206e-02 -6.79056942e-02 1.09466381e-01
3.20459217e-01 -7.33322144e-01 -2.73050994e-01 5.31470999e-02
-5.28024137e-01 1.24591827e+00 1.03893220e-01 5.74760616e-01
5.72657645e-01 2.11372450e-01 3.20088327e-01 -1.10275471e+00
8.52519870e-01 -4.82321203e-01 2.00424660e-02 1.34608075e-01
-1.15865849e-01 -2.65443325e-01 6.75717473e-01 -6.71568453e-01
-1.26802278e+00 9.60344896e-02 3.49566758e-01 -1.21869884e-01
-3.17026228e-01 5.88443875e-01 -4.52095836e-01 1.22264111e+00
9.54296172e-01 4.21379507e-01 -6.92677319e-01 -1.44764949e-02
4.22469467e-01 2.84729451e-01 9.00919914e-01 -7.96160161e-01
8.65677178e-01 2.46337786e-01 -4.38522816e-01 -2.77044833e-01
-1.30173099e+00 -7.67553449e-02 -3.15029562e-01 -8.58747602e-01
8.78510356e-01 -1.02771258e+00 -4.87574339e-01 4.62873787e-01
-1.57081330e+00 -5.10647655e-01 -5.73335588e-01 2.78532654e-01
-1.03692043e+00 -3.65061015e-01 -4.77989256e-01 -7.33579040e-01
-2.89038181e-01 -6.21618927e-01 7.73481905e-01 4.85967457e-01
-1.16779792e+00 -6.38061941e-01 -9.12738219e-02 4.54414159e-01
-2.56688654e-01 7.04095960e-01 7.22575188e-01 -7.62965858e-01
-3.78030747e-01 -3.06336582e-01 1.19387560e-01 -2.86255181e-01
1.18445002e-01 1.02359019e-01 -5.86564720e-01 4.86874104e-01
-1.19621074e-02 -5.12635231e-01 3.67368281e-01 1.42555222e-01
7.94695318e-01 -9.44843590e-01 -3.35368693e-01 3.14760834e-01
8.29599321e-01 1.69711307e-01 6.60448432e-01 1.90134317e-01
7.02090621e-01 7.79173493e-01 8.29711616e-01 8.45202744e-01
7.89943993e-01 7.11517036e-01 -7.47400969e-02 3.99444163e-01
-1.66198656e-01 -1.22482109e+00 5.96413672e-01 3.73319626e-01
-2.21194416e-01 -5.99195600e-01 -8.55915189e-01 8.48194957e-01
-2.11721277e+00 -1.57969534e+00 -1.65166065e-01 1.49910164e+00
1.20469546e+00 2.81887949e-01 3.41111809e-01 -1.04345210e-01
8.54209781e-01 5.16199410e-01 -5.20570636e-01 -1.80573463e-01
-3.00659806e-01 -6.33031011e-01 -1.84161142e-01 4.00776356e-01
-7.60818124e-01 1.27514768e+00 5.26921844e+00 8.02895129e-01
-6.47198558e-01 7.07293600e-02 6.97380602e-01 -5.02903044e-01
-6.06375635e-01 1.00215115e-01 -9.82619941e-01 4.74862248e-01
6.93301499e-01 -1.02703953e+00 5.58233321e-01 7.81161189e-01
8.32873464e-01 1.70707598e-01 -1.27977693e+00 7.12793887e-01
4.33186799e-01 -1.57750642e+00 3.31420153e-01 -1.85994163e-01
8.75994980e-01 -7.98822284e-01 -3.17458540e-01 3.37633193e-01
6.39176726e-01 -9.14356351e-01 1.52598178e+00 8.37770164e-01
6.18235469e-01 -8.23606670e-01 2.30703339e-01 5.35570323e-01
-8.95918489e-01 1.81057185e-01 8.19062591e-02 -4.20447409e-01
7.52081215e-01 4.81189817e-01 -1.05098224e+00 2.47338191e-01
2.88173676e-01 9.74081337e-01 -1.46559134e-01 4.36729908e-01
-1.06957591e+00 9.39617276e-01 1.37253657e-01 -3.96109909e-01
5.56078292e-02 -7.19619989e-02 1.07429659e+00 1.30734384e+00
4.91018355e-01 6.78884685e-01 2.95832425e-01 1.56978357e+00
-3.01786691e-01 1.43080443e-01 -4.44831938e-01 -5.50443947e-01
8.39132726e-01 1.24697590e+00 -5.85419238e-01 -5.56455195e-01
2.91430745e-02 1.21337640e+00 3.69043142e-01 2.60457784e-01
-9.86342907e-01 -1.69337034e-01 2.38072276e-01 5.71520269e-01
7.97285065e-02 1.41537741e-01 -5.49989998e-01 -1.04988515e+00
-1.88377693e-01 -9.31276023e-01 2.23842636e-01 -1.61888134e+00
-1.07150972e+00 6.39253139e-01 2.14015588e-01 -1.05826879e+00
-5.59409440e-01 2.92410821e-01 -1.20622933e+00 7.78500021e-01
-4.64233845e-01 -1.60448778e+00 -3.20521861e-01 1.99456856e-01
9.22699094e-01 -4.81331907e-02 6.07260704e-01 -5.34115732e-01
-4.91766661e-01 3.30283195e-01 -8.83321404e-01 1.61598757e-01
5.94329655e-01 -1.14710784e+00 7.52796829e-01 1.05478847e+00
2.13184908e-01 5.53412795e-01 1.42167103e+00 -1.15059757e+00
-8.46463740e-01 -1.14466965e+00 1.60109353e+00 -5.58122754e-01
8.04311872e-01 -5.10059297e-01 -5.65396965e-01 8.37858737e-01
5.27226150e-01 -9.31596756e-01 5.60670197e-01 1.02292420e-02
-3.55399817e-01 3.87945801e-01 -9.41051900e-01 1.23763537e+00
1.38042104e+00 -3.93729359e-01 -8.85333240e-01 6.18336797e-01
9.80460763e-01 -5.92639029e-01 -4.75929976e-01 -1.67255446e-01
2.67602861e-01 -5.97537935e-01 5.98775327e-01 -5.91880798e-01
1.42219043e+00 -3.73363823e-01 1.04127824e-01 -1.31383431e+00
-4.85749245e-01 -1.00508535e+00 -1.96241453e-01 1.78415406e+00
5.66396534e-01 2.19195962e-01 3.47564310e-01 7.48779118e-01
-2.37007990e-01 -3.82518917e-01 -1.85049251e-01 -5.50612390e-01
-4.66180474e-01 -4.18760478e-01 8.29585671e-01 1.00964439e+00
4.67610627e-01 7.04609811e-01 -9.34337199e-01 -5.28906584e-02
4.89645511e-01 2.72915691e-01 9.93453562e-01 -8.17545414e-01
-4.90248621e-01 -3.77727926e-01 5.49061038e-02 -7.98735619e-01
2.02593997e-01 -7.83322334e-01 2.84035474e-01 -1.82333267e+00
7.66184270e-01 3.46314281e-01 2.78287709e-01 4.72644150e-01
-3.88597786e-01 -1.73322275e-01 2.94209212e-01 3.04864734e-01
-6.52606785e-01 4.92895126e-01 1.44587398e+00 -8.98285732e-02
-5.97675204e-01 -7.48572722e-02 -1.09309042e+00 8.95197392e-01
8.65923107e-01 -4.96623099e-01 -4.33550268e-01 -6.14550486e-02
4.41226423e-01 4.19952124e-01 5.00869691e-01 -9.12638009e-01
2.12447360e-01 -7.86322415e-01 2.87630409e-01 -6.32910132e-01
4.38214570e-01 -7.62231201e-02 5.76436162e-01 1.11227401e-01
-1.00802600e+00 -1.91383317e-01 -1.97882965e-01 2.20598474e-01
-1.47247940e-01 -1.28240749e-01 4.83158261e-01 -3.53017271e-01
-4.09682959e-01 6.40216395e-02 -4.24997807e-01 7.16190562e-02
1.12870872e+00 -1.12372376e-01 -2.78774232e-01 -1.03710783e+00
-5.55207431e-01 1.95777312e-01 7.19924450e-01 3.71180385e-01
5.88392138e-01 -1.52907562e+00 -1.35103667e+00 -4.49448049e-01
3.72916043e-01 -4.07411791e-02 3.40780109e-01 2.85057813e-01
-1.01537056e-01 1.43873677e-01 -1.13162108e-01 9.97871161e-03
-1.19642854e+00 4.97130066e-01 1.11733064e-01 -3.37491989e-01
-7.99054027e-01 9.23770368e-01 3.31359148e-01 -3.43674161e-02
-1.52243236e-02 1.10781968e-01 -4.29494321e-01 5.48516475e-02
9.19054568e-01 1.83185469e-02 -7.80048668e-01 -7.20423043e-01
-4.92739603e-02 -1.47267312e-01 -2.11487189e-01 -6.91307604e-01
1.39723670e+00 -9.45580378e-02 1.46524906e-02 9.36828256e-01
3.30973446e-01 2.67796189e-01 -1.41891015e+00 -1.64801627e-01
-1.71188131e-01 -2.25510210e-01 -4.00390357e-01 -1.20815790e+00
-3.43755990e-01 2.71389872e-01 -6.62483394e-01 9.44271162e-02
8.27765405e-01 4.83298987e-01 7.36139357e-01 9.18182284e-02
4.27202702e-01 -9.42604661e-01 3.56224626e-01 5.98929286e-01
1.68710589e+00 -6.60322547e-01 1.40796071e-02 -4.47309077e-01
-1.52470052e+00 7.97698081e-01 7.42555618e-01 -1.04800291e-01
-1.89008154e-02 3.68335135e-02 -1.44413203e-01 -3.40265065e-01
-1.18606341e+00 8.38742107e-02 3.57928008e-01 4.38506126e-01
5.29198647e-01 5.43950438e-01 -5.52711368e-01 1.40935922e+00
-1.00407958e+00 1.08846359e-01 8.97008240e-01 4.61521834e-01
-2.52456844e-01 -6.33658350e-01 -2.32308045e-01 3.75053644e-01
-9.05374289e-02 -3.82907130e-02 -9.99525607e-01 4.69212383e-01
1.57548800e-01 8.49565923e-01 -4.36155982e-02 -4.76854801e-01
4.80516881e-01 1.48201570e-01 2.88124651e-01 -8.78313065e-01
-7.46734679e-01 9.53468867e-03 6.72476470e-01 -4.28115994e-01
-2.39316702e-01 -1.30734313e+00 -1.42977798e+00 -5.50681472e-01
2.72463292e-01 1.32690772e-01 2.21114129e-01 1.04746640e+00
1.46071509e-01 5.51607132e-01 4.78620052e-01 -6.96697533e-01
-1.31716756e-02 -1.17533278e+00 -3.54323238e-01 7.18659043e-01
-3.75379831e-01 -2.78887004e-01 -1.21429391e-01 6.13045692e-01] | [11.782376289367676, 8.821413040161133] |
2f6dd64d-0e26-4fea-a84b-4c59fc89bf15 | memorization-capacity-of-multi-head-attention | 2306.02010 | null | https://arxiv.org/abs/2306.02010v1 | https://arxiv.org/pdf/2306.02010v1.pdf | Memorization Capacity of Multi-Head Attention in Transformers | In this paper, we investigate the memorization capabilities of multi-head attention in Transformers, motivated by the central role attention plays in these models. Under a mild linear independence assumption on the input data, we present a theoretical analysis demonstrating that an $H$-head attention layer with a context size $n$, dimension $d$, and $O(Hd^2)$ parameters can memorize $O(Hn)$ examples. We conduct experiments that verify our assumptions on the image classification task using Vision Transformer. To validate our theoretical findings, we perform synthetic experiments and show a linear relationship between memorization capacity and the number of attention heads. | ['Christos Thrampoulidis', 'Renjie Liao', 'Sadegh Mahdavi'] | 2023-06-03 | null | null | null | null | ['memorization'] | ['natural-language-processing'] | [-6.72449693e-02 1.78833038e-01 8.09973404e-02 -1.73297629e-01
-4.57914650e-01 -1.12409284e-02 4.87864278e-02 -5.59698679e-02
-5.76012552e-01 6.06793284e-01 1.11729046e-03 -5.15432000e-01
-1.08284481e-01 -6.91727459e-01 -8.30162644e-01 -6.64243400e-01
-1.56891674e-01 1.56004488e-01 1.34990320e-01 -2.05239281e-02
3.99788529e-01 2.88459390e-01 -1.65320373e+00 -2.41449624e-01
7.30574787e-01 1.27077639e+00 6.19861662e-01 8.37679803e-01
1.76771984e-01 1.14715433e+00 -5.93678355e-01 -6.89841151e-01
1.84580013e-01 -3.64572436e-01 -1.16490793e+00 1.89472303e-01
5.33160448e-01 -1.39033824e-01 -5.52682519e-01 1.15857673e+00
6.31755173e-01 2.28398517e-01 5.30509889e-01 -1.00295341e+00
-1.28492761e+00 6.83627903e-01 -5.90406239e-01 1.08877409e+00
1.49927959e-01 2.65095979e-01 1.27369356e+00 -1.07412875e+00
2.00121954e-01 1.15470517e+00 3.99663866e-01 5.46458423e-01
-1.02543223e+00 -7.71850348e-01 2.42111549e-01 6.14427209e-01
-1.41200149e+00 -5.36296189e-01 6.07757211e-01 -3.27210426e-01
1.36934972e+00 1.66235700e-01 6.97883010e-01 4.67838198e-01
4.37438458e-01 8.00015748e-01 1.02999866e+00 -7.53019273e-01
1.18377887e-01 2.60278046e-01 5.08220792e-01 9.94920075e-01
4.09983128e-01 -2.12496445e-01 -6.52748585e-01 2.53750920e-01
8.57597649e-01 -2.15868130e-01 -4.25699055e-01 9.44535900e-03
-4.65596914e-01 9.41957057e-01 7.44119465e-01 2.96306938e-01
-5.77787459e-01 5.02421796e-01 1.45520335e-02 3.08570474e-01
1.52497381e-01 4.04783040e-01 6.88646212e-02 1.97639063e-01
-4.94695455e-01 -4.01615143e-01 4.33036566e-01 1.37140083e+00
7.53366590e-01 2.94591069e-01 -8.59959573e-02 6.51397824e-01
-9.26537812e-03 8.16487610e-01 6.54653311e-01 -9.47650313e-01
3.94283921e-01 2.12472305e-01 -1.23316184e-01 -4.85322654e-01
-1.57334059e-01 -6.27204061e-01 -8.68327439e-01 -2.62519717e-01
1.73161447e-03 4.01511826e-02 -8.59482586e-01 2.03262019e+00
-5.36600590e-01 -4.58450615e-01 -1.81494117e-01 6.74910367e-01
3.42547804e-01 2.07665607e-01 3.64636540e-01 -5.08231461e-01
1.35066330e+00 -7.13262558e-01 -7.21140265e-01 -4.91099417e-01
2.52376556e-01 -5.66384435e-01 1.40114641e+00 9.48137343e-02
-1.74156642e+00 -6.83927238e-01 -8.35800648e-01 -1.99947640e-01
-2.01505870e-01 -1.43102884e-01 7.57962525e-01 5.94674110e-01
-1.48172677e+00 -4.70621623e-02 -2.93866277e-01 -4.48977888e-01
4.53642935e-01 6.46056890e-01 2.38284213e-03 -2.99784154e-01
-1.16916883e+00 1.00567997e+00 1.34523036e-02 -3.35663408e-02
-1.09701610e+00 -2.57668108e-01 -7.93538094e-01 4.54322845e-01
8.49895030e-02 -9.73932505e-01 1.35697293e+00 -6.05261981e-01
-8.27572823e-01 8.36595893e-01 -5.25296688e-01 -7.86539435e-01
-1.16293266e-01 -3.37191343e-01 4.81960550e-02 1.69785067e-01
1.10313132e-01 5.92606485e-01 9.83772337e-01 -9.47782576e-01
-7.35806167e-01 -4.42071110e-01 3.18637133e-01 -7.35916058e-03
-5.45263171e-01 -2.90792882e-02 -4.04329002e-01 -3.97491574e-01
-1.07249752e-01 -8.33333671e-01 -1.15491122e-01 -3.78519893e-01
-3.25584352e-01 -2.10689172e-01 2.62913764e-01 -3.54065329e-01
1.41583574e+00 -2.26937580e+00 2.18413353e-01 1.85460731e-01
5.53160608e-01 -8.76254141e-02 1.18248135e-01 -6.05107425e-03
2.15256531e-02 4.90121618e-02 -1.39513776e-01 -5.45009136e-01
3.30173224e-02 1.66996002e-01 -4.34016347e-01 4.04354811e-01
-4.55942787e-02 7.41648078e-01 -5.22056520e-01 -5.41281223e-01
-9.01345164e-02 3.66905451e-01 -7.09380805e-01 4.79556769e-01
2.51761168e-01 -1.88873559e-01 -2.12250680e-01 4.52336550e-01
3.68422896e-01 -8.30808282e-01 1.70201614e-01 -1.83705632e-02
-6.28835633e-02 1.37260944e-01 -5.60492277e-01 1.01947916e+00
-6.74107134e-01 8.35344613e-01 1.70591071e-01 -9.52479541e-01
5.51257372e-01 1.75916031e-01 1.61904842e-02 -1.02368522e+00
4.99241292e-01 -1.05503999e-01 3.88639867e-01 -3.90955001e-01
5.49180627e-01 -3.83971870e-01 -3.86848375e-02 5.93404293e-01
3.24731797e-01 1.43877536e-01 4.37668450e-02 3.19301218e-01
1.03731382e+00 -1.04856706e+00 3.90726447e-01 -6.31827176e-01
4.35860842e-01 -4.94106084e-01 2.75055677e-01 1.13475692e+00
-4.66985494e-01 1.43410385e-01 3.63535136e-01 -2.56223559e-01
-1.04239488e+00 -9.04142737e-01 -1.00550704e-01 1.64687693e+00
1.43962607e-01 -1.41122743e-01 -9.71014380e-01 -2.45949939e-01
-1.90040976e-01 7.11670339e-01 -9.74396884e-01 -2.77394205e-01
-5.55188417e-01 -6.94994330e-01 2.89908111e-01 1.14911938e+00
8.78767431e-01 -1.26982379e+00 -9.38842118e-01 -1.36422262e-01
-5.01887463e-02 -1.00778532e+00 -6.31194115e-01 5.83785057e-01
-5.89123547e-01 -1.05186331e+00 -8.73882830e-01 -1.02643168e+00
8.61399472e-01 4.09029007e-01 1.19415486e+00 1.25637338e-01
-3.77062321e-01 8.69493067e-01 2.23663822e-01 -5.24038076e-01
1.94640562e-01 3.19453448e-01 1.24012455e-01 -5.34787416e-01
3.69321257e-01 -5.35239995e-01 -8.31195235e-01 1.60851821e-01
-7.41171002e-01 -1.82418570e-01 9.65145230e-01 7.74649143e-01
3.26522768e-01 -4.94279228e-02 3.98320079e-01 -5.88601589e-01
8.43238533e-01 -3.04066867e-01 -5.26842415e-01 2.81677246e-01
-7.44102418e-01 3.05483699e-01 6.47500753e-01 -2.48910919e-01
-7.37103820e-01 -2.63713986e-01 3.04878354e-02 -5.98509133e-01
2.94425517e-01 3.97530854e-01 -2.30238453e-01 -1.60732016e-01
4.64707613e-01 4.98187184e-01 -2.63773531e-01 -2.39097938e-01
3.82227063e-01 2.70355284e-01 4.49235588e-01 -5.76665103e-01
4.34925556e-01 2.48077363e-01 -1.98219791e-01 -8.39701474e-01
-7.74160862e-01 -3.41747329e-02 -3.23307931e-01 -4.91118245e-02
7.93284476e-01 -8.00726295e-01 -1.16702223e+00 2.06010863e-01
-8.58136296e-01 -3.78068566e-01 -3.16112101e-01 3.80626500e-01
-7.58623421e-01 6.64019734e-02 -8.86115193e-01 -1.05759728e+00
-5.36490381e-01 -1.15895021e+00 3.84483337e-01 2.15038866e-01
-5.22536039e-02 -8.40643525e-01 -3.75348449e-01 9.71373543e-02
7.72122681e-01 -7.93990374e-01 1.39315701e+00 -3.05790663e-01
-9.78091002e-01 1.05847910e-01 -4.36105490e-01 2.39079401e-01
-1.46159038e-01 -5.80495596e-01 -1.14036071e+00 -4.16469485e-01
4.31206107e-01 -3.23893487e-01 1.00315070e+00 6.74732804e-01
1.49999917e+00 -4.38963264e-01 -2.39227787e-01 2.98064679e-01
1.32403517e+00 2.77638674e-01 6.68880820e-01 1.23130724e-01
4.64458078e-01 1.20488979e-01 2.40943104e-01 5.02141178e-01
4.97127920e-01 9.71439183e-02 4.41154480e-01 5.84763959e-02
-1.15609042e-01 7.11996332e-02 -3.82683873e-02 1.20388341e+00
-3.87869328e-01 -4.15352434e-01 -8.88866127e-01 9.54098821e-01
-1.37124598e+00 -8.40320647e-01 6.11263216e-01 2.25951147e+00
7.29272604e-01 5.18684208e-01 1.40873641e-01 1.83432639e-01
8.99784863e-01 -3.89105864e-02 -6.04064167e-01 -5.52231252e-01
-1.58559024e-01 7.93377399e-01 7.17374742e-01 7.45773554e-01
-6.07326388e-01 8.60157192e-01 7.91654539e+00 5.57557225e-01
-1.07270479e+00 1.46734774e-01 7.70492315e-01 -3.15547019e-01
-1.88100606e-01 -3.71685356e-01 -8.41312945e-01 3.92147332e-01
8.98409188e-01 -5.64726830e-01 4.53342766e-01 9.46403027e-01
-4.82436508e-01 -1.66819811e-01 -1.25927591e+00 1.06207395e+00
2.62845874e-01 -1.03543985e+00 2.42072538e-01 2.42668957e-01
2.34944731e-01 -2.61636376e-01 8.15339327e-01 3.72531027e-01
3.93500358e-01 -1.08714950e+00 6.00874722e-01 3.77287596e-01
8.51155698e-01 -7.20152318e-01 5.98028421e-01 2.58692294e-01
-1.25798893e+00 -6.12886786e-01 -5.30609131e-01 -2.05540031e-01
6.45591272e-03 2.56313056e-01 -6.23937845e-01 -1.28397137e-01
9.01809335e-01 1.36253774e-01 -6.54684186e-01 7.80817926e-01
-2.06611931e-01 4.42244291e-01 -1.85642287e-01 -2.12801993e-02
4.20412943e-02 4.23695177e-01 4.19897735e-02 1.12287652e+00
3.30193251e-01 7.76096642e-01 -4.11689311e-01 6.44558132e-01
-4.19299543e-01 -7.48103932e-02 -4.07448143e-01 2.16042474e-01
5.65534472e-01 6.65052474e-01 -5.75219393e-01 -4.93620276e-01
-3.23632300e-01 7.82260656e-01 7.26883471e-01 2.30392396e-01
-8.99192274e-01 -4.55144286e-01 5.51528275e-01 2.85707831e-01
6.67333543e-01 -3.18088800e-01 -3.10795516e-01 -8.96205723e-01
-9.09254029e-02 -3.51246268e-01 3.53762746e-01 -8.63742590e-01
-1.12335646e+00 8.48864615e-01 -2.35292226e-01 -5.50419211e-01
-2.90570850e-03 -5.64744771e-01 -6.12279475e-01 8.95531476e-01
-1.44660068e+00 -6.68831527e-01 -1.28677607e-01 8.37164402e-01
5.58548212e-01 4.02950915e-03 7.28869081e-01 2.47294977e-01
-7.34249830e-01 1.13170516e+00 -4.01366264e-01 1.19032942e-01
2.43628189e-01 -1.25033832e+00 3.16320956e-01 7.51963496e-01
-1.69493347e-01 1.11747742e+00 7.98367381e-01 -1.80183902e-01
-1.46068537e+00 -8.15028071e-01 9.17923689e-01 -2.89815694e-01
5.07660151e-01 -2.98879683e-01 -8.02343369e-01 1.13565218e+00
6.59529924e-01 -1.01095431e-01 1.03294039e+00 2.94775456e-01
-3.74572188e-01 -7.15565234e-02 -1.06230569e+00 4.17488813e-01
1.22433329e+00 -8.85978758e-01 -7.70815670e-01 -1.08505853e-01
6.38810515e-01 -3.40803824e-02 -8.38398635e-01 1.40551269e-01
4.29323882e-01 -9.92184997e-01 8.97514164e-01 -4.85231608e-01
2.95264393e-01 3.44271451e-01 -4.59285915e-01 -1.02031398e+00
-8.20939600e-01 -2.53049165e-01 -1.43631071e-01 6.79312766e-01
3.95890146e-01 -6.47899866e-01 4.57162857e-01 9.20479655e-01
2.89398921e-03 -4.96235162e-01 -9.98676956e-01 -5.63253760e-01
2.10079968e-01 -3.32092613e-01 4.25452501e-01 5.90729415e-01
3.79027635e-01 7.43463337e-01 -2.90847152e-01 1.13073766e-01
7.15188384e-01 1.14100933e-01 2.52020359e-01 -9.86274421e-01
-3.44294101e-01 -4.62688476e-01 -2.69743621e-01 -1.29725933e+00
1.68881178e-01 -3.24441373e-01 7.97739923e-02 -1.11121428e+00
6.79997623e-01 -3.72507781e-01 -7.34979868e-01 3.59962225e-01
-3.19102496e-01 8.83942246e-02 3.08753431e-01 2.74630427e-01
-9.54774439e-01 5.47299862e-01 9.49923515e-01 -7.43592754e-02
2.75504559e-01 -2.96307117e-01 -1.02792799e+00 6.18286908e-01
6.47359967e-01 -1.16760813e-01 -6.08096004e-01 -8.75132263e-01
1.05844975e-01 8.16481560e-02 1.04127683e-01 -1.12759149e+00
4.92090672e-01 1.09313063e-01 3.48114163e-01 -3.26400071e-01
5.49927533e-01 -8.19840252e-01 -4.29688692e-01 4.61211890e-01
-5.76918662e-01 4.29021120e-01 3.65953147e-01 3.62245530e-01
6.81613162e-02 -7.16820266e-03 8.74858201e-01 -3.85006875e-01
-5.93534827e-01 3.00562739e-01 -5.91396511e-01 1.25734389e-01
7.33265638e-01 4.90448140e-02 -4.65794623e-01 -5.73291421e-01
-1.06996942e+00 1.44753680e-01 3.07419270e-01 -1.09195232e-01
7.30851114e-01 -1.12711263e+00 -1.74029440e-01 4.43939835e-01
4.29660678e-02 -2.68264592e-01 2.44907126e-01 6.88437879e-01
4.45480123e-02 8.63401711e-01 -2.93378800e-01 -4.15884316e-01
-1.07219088e+00 1.07404113e+00 3.27414006e-01 -4.82378267e-02
-1.20987386e-01 1.23849916e+00 5.97309291e-01 6.49610341e-01
3.35878074e-01 -4.36839074e-01 -1.13834757e-02 -2.47494176e-01
6.55733287e-01 1.96464673e-01 -1.59975260e-01 -4.25278455e-01
-3.17309886e-01 4.75196451e-01 -5.03846943e-01 -2.18137130e-01
9.57566202e-01 -2.96153337e-01 1.48040280e-01 2.75343806e-01
9.50137377e-01 -1.76193520e-01 -9.29349542e-01 -3.71168226e-01
-2.52661556e-01 -3.69662255e-01 4.23158379e-03 -2.97517091e-01
-1.21649969e+00 1.18845892e+00 6.38952792e-01 6.73281670e-01
1.30868053e+00 4.26102608e-01 5.11620939e-01 4.72301990e-01
4.69489843e-01 -9.43585396e-01 2.05838516e-01 6.54309452e-01
8.47862065e-01 -8.46558750e-01 -2.97639310e-01 -2.82630473e-01
-5.01125574e-01 4.73986715e-01 1.02596331e+00 -4.59554568e-02
7.70019293e-01 3.04249644e-01 -4.18212533e-01 -4.35502082e-01
-9.81957674e-01 -5.14733911e-01 -1.88328162e-01 4.30759251e-01
4.94879872e-01 -9.70152095e-02 5.64561263e-02 6.04194045e-01
-4.83713478e-01 1.06192874e-02 3.67821872e-01 1.00370598e+00
-9.71960664e-01 -4.78227615e-01 -2.26113632e-01 3.13108265e-01
-6.39998674e-01 -5.71033657e-01 -1.46688595e-01 7.40276694e-01
-1.10515706e-01 7.74422646e-01 4.36238766e-01 -3.02278280e-01
3.15428436e-01 1.53892323e-01 8.15721214e-01 -4.53026325e-01
-3.54933023e-01 -1.87368929e-01 -4.61812466e-01 -2.44115129e-01
-2.87988305e-01 -3.35229814e-01 -9.75014806e-01 -8.20814669e-01
-3.88419479e-01 3.94872010e-01 1.89357817e-01 8.90720427e-01
3.36840451e-01 6.40010476e-01 5.93817592e-01 -4.37411278e-01
-3.86995643e-01 -1.22670293e+00 -6.35315835e-01 1.95620254e-01
3.57900143e-01 -7.34239221e-01 -5.81786394e-01 1.46308988e-01] | [9.522573471069336, 2.554103136062622] |
f399d3db-7ab4-48bd-8cba-e6c8cdcf3ea8 | generalizing-discrete-convolutions-for | 1904.02375 | null | https://arxiv.org/abs/1904.02375v5 | https://arxiv.org/pdf/1904.02375v5.pdf | ConvPoint: Continuous Convolutions for Point Cloud Processing | Point clouds are unstructured and unordered data, as opposed to images. Thus, most machine learning approach developed for image cannot be directly transferred to point clouds. In this paper, we propose a generalization of discrete convolutional neural networks (CNNs) in order to deal with point clouds by replacing discrete kernels by continuous ones. This formulation is simple, allows arbitrary point cloud sizes and can easily be used for designing neural networks similarly to 2D CNNs. We present experimental results with various architectures, highlighting the flexibility of the proposed approach. We obtain competitive results compared to the state-of-the-art on shape classification, part segmentation and semantic segmentation for large-scale point clouds. | ['Alexandre Boulch'] | 2019-04-04 | null | null | null | null | ['3d-part-segmentation', 'lidar-semantic-segmentation'] | ['computer-vision', 'computer-vision'] | [ 6.97711185e-02 -2.18013953e-02 -1.43516948e-02 -3.36672425e-01
-2.69385129e-01 -5.29881299e-01 4.48091209e-01 3.41702998e-01
-5.57837844e-01 2.74232149e-01 -7.30936587e-01 -4.30485189e-01
-7.46408477e-02 -1.17588079e+00 -1.08881867e+00 -2.90221959e-01
-1.46445841e-01 9.89780188e-01 5.34841239e-01 -8.28816369e-02
1.52003795e-01 1.51017845e+00 -1.61348212e+00 2.88337499e-01
6.75554216e-01 1.17483044e+00 1.69401243e-01 3.92333865e-01
-7.33321249e-01 1.11623473e-01 -3.76391381e-01 -3.57012719e-01
5.25571048e-01 4.43252295e-01 -8.13180089e-01 3.16634685e-01
8.82092535e-01 -3.06880951e-01 -7.48448372e-02 1.03042006e+00
1.40062019e-01 2.68020062e-03 7.69376218e-01 -1.28870070e+00
-1.01665974e+00 8.89772177e-02 -3.04954797e-01 -1.61125496e-01
-2.37786949e-01 -2.56793380e-01 7.08151519e-01 -1.02290809e+00
4.11251694e-01 1.36103988e+00 1.15457046e+00 4.01252240e-01
-1.01640332e+00 -4.06535178e-01 1.06762178e-01 -2.24018488e-02
-1.33189857e+00 1.41238451e-01 8.65367174e-01 -4.41776901e-01
1.14933455e+00 2.71665812e-01 9.70175862e-01 4.73231077e-01
-2.46447042e-01 9.48217809e-01 6.66094065e-01 -4.36426759e-01
4.23746258e-01 -1.22498438e-01 2.26076812e-01 5.95170379e-01
3.31423730e-01 -6.89688995e-02 2.02684090e-01 -3.52546573e-01
1.37809503e+00 4.69889075e-01 5.47250248e-02 -8.35858285e-01
-1.07651412e+00 9.27306235e-01 9.48833585e-01 3.08675349e-01
-4.27595705e-01 3.43758941e-01 3.11008006e-01 1.90997496e-01
7.99859166e-01 1.37970328e-01 -6.00713670e-01 2.35735312e-01
-1.13623524e+00 4.26915497e-01 7.04401255e-01 1.27740014e+00
8.45875978e-01 8.67483318e-02 2.53202736e-01 8.53464603e-01
2.43021026e-01 3.27373415e-01 2.25673184e-01 -8.83771479e-01
1.93247691e-01 9.15320158e-01 6.59350008e-02 -8.11038315e-01
-5.29935062e-01 -2.01733202e-01 -8.93789649e-01 6.51316822e-01
1.28907949e-01 3.03076357e-01 -1.24703848e+00 7.06390142e-01
2.78481811e-01 4.13909882e-01 -1.13693409e-01 8.40582073e-01
1.06685209e+00 6.76387906e-01 -9.38487351e-02 4.59945410e-01
1.16627693e+00 -8.33983064e-01 -1.81494698e-01 -3.76560912e-02
2.57308275e-01 -4.90968704e-01 8.04559469e-01 4.03229743e-01
-1.15001881e+00 -7.78524935e-01 -9.97005284e-01 -2.80469507e-01
-7.14846432e-01 2.12582588e-01 7.61885285e-01 4.76693928e-01
-1.40185559e+00 1.15537274e+00 -1.15596890e+00 -3.75040323e-01
8.37800145e-01 8.13182831e-01 -2.16693908e-01 1.72335580e-01
-6.55189455e-01 5.35482943e-01 5.94710588e-01 2.32506126e-01
-2.37074852e-01 -6.85480893e-01 -6.65755332e-01 1.36228904e-01
-3.61862965e-02 -8.04709792e-01 1.40058386e+00 -9.30925310e-01
-1.49696159e+00 1.06665289e+00 7.16022030e-02 -6.97320521e-01
5.09719372e-01 -1.11597672e-01 1.27246439e-01 3.11007291e-01
-1.14555225e-01 1.13645029e+00 8.37377369e-01 -1.56559563e+00
-5.85117877e-01 -4.35105056e-01 1.34388521e-01 4.63945307e-02
-1.81066141e-01 -1.41466588e-01 -5.42712927e-01 -3.66439193e-01
4.17523921e-01 -1.03266072e+00 -4.36788768e-01 4.66860294e-01
-2.60822415e-01 -4.91268069e-01 1.20393860e+00 -1.39681354e-01
2.63106734e-01 -2.01525116e+00 -5.98134287e-02 1.73741266e-01
2.22919926e-01 5.30202925e-01 -9.89628062e-02 2.45916098e-01
-2.50791907e-01 2.80911803e-01 -5.63672602e-01 -5.27715445e-01
1.38572499e-01 5.61983347e-01 -2.21845835e-01 5.22295654e-01
5.17034292e-01 1.22893465e+00 -5.38760304e-01 -3.42997611e-01
6.28080070e-01 6.22564375e-01 -4.04823929e-01 9.52275284e-03
-4.51015025e-01 2.55354106e-01 -5.67512333e-01 7.29346395e-01
1.19536138e+00 -2.95644283e-01 -3.90374362e-01 -1.25182420e-01
-3.07996958e-01 -2.17225537e-01 -1.00904751e+00 1.65797770e+00
-5.01215279e-01 5.35530448e-01 2.11502612e-01 -1.05198288e+00
1.19169128e+00 3.02211046e-01 7.99136996e-01 -1.81003690e-01
9.05667171e-02 3.96120161e-01 -2.73435265e-01 -1.23286076e-01
4.23940778e-01 -1.68707997e-01 2.55564570e-01 -5.61534315e-02
1.33869976e-01 -7.51546144e-01 -1.12271093e-01 -4.30030227e-01
5.95598996e-01 9.82717127e-02 9.68434736e-02 -2.19111890e-01
4.53202575e-01 3.36859584e-01 3.45561914e-02 6.64394736e-01
1.44900545e-01 8.99389982e-01 2.04989433e-01 -9.50182080e-01
-1.33983910e+00 -1.16823554e+00 -3.98109436e-01 7.58111894e-01
1.69725493e-01 1.29575029e-01 -8.13183248e-01 -4.04435158e-01
3.89338851e-01 3.71323079e-01 -4.47297305e-01 4.24360693e-01
-7.74407387e-01 -2.99109340e-01 5.83569050e-01 9.39791262e-01
4.91966963e-01 -1.36299181e+00 -5.03364027e-01 1.37347341e-01
4.36233789e-01 -1.30236387e+00 1.17256694e-01 2.28477195e-01
-1.58253157e+00 -9.66432452e-01 -7.41746068e-01 -1.13757908e+00
6.36705816e-01 3.77801985e-01 1.46204329e+00 2.56912112e-01
-2.75262706e-02 4.48388636e-01 -4.03687507e-01 -8.76509070e-01
-8.79805908e-02 2.73506761e-01 -2.10976213e-01 -1.59052342e-01
4.99226153e-01 -6.97616994e-01 -4.61833715e-01 1.19399559e-02
-1.36573124e+00 -2.06461921e-01 6.36945128e-01 6.39640868e-01
9.13246632e-01 -8.72129723e-02 1.27420157e-01 -9.31227088e-01
4.69390541e-01 -2.00183317e-01 -9.11959767e-01 -3.91952358e-02
-1.33786917e-01 -3.46703142e-01 7.06979573e-01 -1.96346670e-01
-4.85239506e-01 5.26206851e-01 -6.54262602e-01 -1.09116018e+00
-9.10415232e-01 1.37580141e-01 7.30778873e-02 -6.50904298e-01
3.18184555e-01 1.43523991e-01 2.10135002e-02 -7.79576480e-01
5.98951459e-01 7.33782530e-01 4.00999844e-01 -5.91479182e-01
8.28571737e-01 9.96147573e-01 1.45433187e-01 -1.23556495e+00
-3.15986753e-01 -7.39464223e-01 -1.47365439e+00 7.53651559e-02
9.78949785e-01 -7.08285987e-01 -7.38071442e-01 5.84944427e-01
-1.60114801e+00 -2.56054521e-01 -6.12116337e-01 1.14908569e-01
-8.67243469e-01 3.29124600e-01 -7.84497917e-01 -6.30893290e-01
-4.48622644e-01 -1.09326243e+00 1.55626345e+00 1.06058881e-01
3.80478621e-01 -1.18193364e+00 -1.60573438e-01 -4.31753090e-03
2.50996292e-01 3.44294548e-01 9.99158621e-01 -8.17291021e-01
-8.08647871e-01 -4.55033779e-01 -5.37223339e-01 5.34058928e-01
-1.41673520e-01 1.53204456e-01 -8.79302680e-01 -1.97504446e-01
-9.74850059e-02 -2.62926131e-01 7.83194721e-01 6.22492790e-01
1.70370924e+00 -1.81105677e-02 -4.28205162e-01 7.93083906e-01
1.66673887e+00 1.45145189e-02 7.98338771e-01 4.32204545e-01
9.33421493e-01 3.09881181e-01 1.97803721e-01 1.38457105e-01
1.17255591e-01 5.52037776e-01 8.50394726e-01 -5.66842914e-01
1.72289580e-01 9.27210727e-04 -4.66325343e-01 7.20503092e-01
-2.95195758e-01 -8.95849243e-02 -1.15017712e+00 6.11925006e-01
-1.77168751e+00 -5.29254913e-01 -6.26279175e-01 1.85423601e+00
2.08964989e-01 1.08622208e-01 6.79151863e-02 9.83413085e-02
7.49640048e-01 -4.21952717e-02 -3.60644877e-01 -4.52797741e-01
1.06561966e-01 8.55811775e-01 7.96054125e-01 1.25862181e-01
-1.49385691e+00 1.06407380e+00 6.86245155e+00 9.21782851e-01
-1.26088381e+00 1.63008392e-01 2.92017758e-01 3.24200690e-01
-1.04558520e-01 -3.74165386e-01 -4.96286988e-01 1.52344495e-01
6.07350588e-01 3.50762010e-01 4.63870317e-02 1.18576133e+00
-5.96924536e-02 5.56454182e-01 -1.12709880e+00 1.05299246e+00
-3.11872393e-01 -1.63301778e+00 2.33296558e-01 -1.95620190e-02
7.02058196e-01 5.58311105e-01 6.21936396e-02 1.23403691e-01
1.53671265e-01 -1.15937114e+00 7.84260869e-01 2.63274729e-01
6.96090341e-01 -7.69051075e-01 7.21075177e-01 4.64957863e-01
-1.32501876e+00 1.51205972e-01 -9.91668940e-01 -6.68263584e-02
-2.65517794e-02 5.52638710e-01 -8.83360505e-01 5.72068453e-01
8.34824264e-01 7.21464753e-01 -4.95738506e-01 1.40827298e+00
1.54727384e-01 2.99891800e-01 -6.67449117e-01 -8.59861001e-02
6.87779307e-01 -5.01817942e-01 3.40455502e-01 1.25342977e+00
2.54324913e-01 2.45350953e-02 3.95590514e-01 1.16751051e+00
-1.87853724e-01 8.25022459e-02 -9.21680212e-01 1.29793897e-01
2.88801700e-01 1.26496923e+00 -1.20733976e+00 -3.57248545e-01
-6.39737964e-01 9.84417558e-01 4.82323855e-01 3.02114427e-01
-4.74229127e-01 -3.16115260e-01 6.28187776e-01 1.70381114e-01
7.50403941e-01 -6.87057853e-01 -4.12407935e-01 -8.78136396e-01
-2.55593155e-02 -1.87354535e-01 -1.42866239e-01 -8.17686558e-01
-1.71488678e+00 7.79049039e-01 1.32578611e-01 -1.51933086e+00
1.56937957e-01 -1.17604971e+00 -6.65513217e-01 7.58525431e-01
-1.74339724e+00 -1.48150992e+00 -2.82772601e-01 5.71715236e-01
6.04666471e-01 1.25231475e-01 8.06012690e-01 2.32318848e-01
7.75847211e-02 9.53345373e-02 3.81519824e-01 4.91764933e-01
-1.41298935e-01 -1.54011369e+00 1.06842351e+00 3.56316686e-01
3.19291025e-01 4.29962873e-01 1.17776275e-01 -4.71539140e-01
-1.30164111e+00 -1.41870618e+00 3.26921105e-01 -3.41759056e-01
4.67758298e-01 -5.14317036e-01 -1.21089697e+00 5.98213136e-01
1.79835241e-02 4.87684429e-01 4.78189558e-01 -6.30985126e-02
-1.98943794e-01 2.39888847e-01 -1.31705213e+00 2.63364136e-01
1.00585008e+00 -3.16529334e-01 -6.32438302e-01 5.30624270e-01
9.37353849e-01 -6.79066479e-01 -1.04140246e+00 6.22121274e-01
1.70839205e-01 -9.34188724e-01 1.48783898e+00 -7.24784374e-01
2.40254030e-01 -1.15088388e-01 1.90660562e-02 -1.19959390e+00
-5.07193267e-01 1.29773974e-01 -9.72360559e-03 6.09939814e-01
7.39553869e-02 -5.03155231e-01 1.26361680e+00 3.02863300e-01
-5.05884826e-01 -8.72124195e-01 -1.12281084e+00 -9.46548581e-01
6.43685102e-01 -5.94141245e-01 8.80632818e-01 8.38946044e-01
-8.15134645e-01 -2.41884887e-01 2.17277721e-01 4.08335328e-01
6.84064806e-01 4.52848941e-01 5.27816772e-01 -1.86503351e+00
1.95246309e-01 -6.28621995e-01 -8.76072764e-01 -1.09491098e+00
4.29815024e-01 -9.11202788e-01 -4.15000431e-02 -1.82354391e+00
-4.12586242e-01 -7.52771556e-01 -2.56287381e-02 5.49960077e-01
3.81737411e-01 6.40385330e-01 3.41975182e-01 5.10431349e-01
-3.34197104e-01 4.87444788e-01 1.36233270e+00 -4.00081903e-01
-1.88375145e-01 3.44913036e-01 1.26328636e-02 9.69972491e-01
8.69459867e-01 -3.97013932e-01 -1.13799781e-01 -6.76086545e-01
-8.41025189e-02 -1.89796314e-01 7.46587694e-01 -1.18697858e+00
2.50823051e-01 5.49554043e-02 5.73429704e-01 -1.15807509e+00
4.99217004e-01 -1.23549819e+00 9.32788402e-02 2.51774609e-01
3.81780453e-02 4.70591299e-02 5.28891802e-01 4.88019854e-01
-3.67153615e-01 -5.19789219e-01 6.83395743e-01 -6.84819698e-01
-6.36996984e-01 7.30597675e-01 -9.92889982e-03 -4.06126052e-01
9.22194242e-01 -6.34624302e-01 9.23412666e-02 4.00557145e-02
-9.79411304e-01 -8.05121139e-02 7.75853038e-01 4.45161700e-01
1.08263624e+00 -1.53682959e+00 -5.80352187e-01 2.79751182e-01
-3.40287723e-02 7.10616648e-01 1.54172495e-01 4.28435445e-01
-1.28208470e+00 7.13344753e-01 -3.75648320e-01 -1.06329644e+00
-9.49732006e-01 7.78326333e-01 4.54228997e-01 1.87482819e-01
-9.36210930e-01 7.84565568e-01 9.89123806e-03 -9.97363448e-01
1.43569887e-01 -8.31665814e-01 -1.93450511e-01 -1.76593035e-01
-6.75963834e-02 2.29749233e-01 6.25111818e-01 -6.12551153e-01
-2.47782066e-01 8.35358918e-01 2.26840287e-01 2.75837928e-01
1.54830933e+00 3.98800969e-01 -4.27480280e-01 4.97617513e-01
1.45663893e+00 -5.29026031e-01 -1.10091424e+00 -1.03862122e-01
-7.22726956e-02 -4.85194415e-01 1.96140204e-02 -1.97861269e-01
-1.16618955e+00 1.18033683e+00 7.04671919e-01 6.41140342e-01
7.71376729e-01 1.21205449e-01 8.37696552e-01 5.95493495e-01
6.72196805e-01 -7.18028486e-01 -3.45201492e-01 7.30345130e-01
8.62636149e-01 -1.15985811e+00 -1.13036007e-01 -7.79610336e-01
-1.72329441e-01 1.62153018e+00 4.60859716e-01 -7.70102084e-01
9.73996222e-01 2.37818971e-01 -3.01025454e-02 -4.69884604e-01
-2.54245460e-01 -4.35772538e-01 4.08128381e-01 9.23384070e-01
2.80428529e-01 2.17977270e-01 1.90075979e-01 7.09351301e-02
-1.21484213e-01 1.68767393e-01 3.75808835e-01 1.08292747e+00
-4.83040065e-01 -1.10943973e+00 -6.62289023e-01 6.23701751e-01
-1.18166789e-01 -9.50965192e-03 -5.39453745e-01 9.32898879e-01
3.22174549e-01 3.57147485e-01 6.88720167e-01 -1.20734796e-01
5.46426713e-01 -3.43431979e-02 4.74473059e-01 -8.81482542e-01
-6.76491976e-01 -8.09561908e-02 -5.13321579e-01 -3.80547434e-01
-6.22964382e-01 -5.61173737e-01 -1.32558608e+00 -2.55459756e-01
-2.97739506e-01 1.38284238e-02 9.06155646e-01 6.95963144e-01
3.11713994e-01 4.64260012e-01 3.60604405e-01 -1.59608114e+00
-3.29370290e-01 -8.14240277e-01 -6.44464731e-01 3.86756450e-01
2.65164614e-01 -7.25715220e-01 1.97711624e-02 -1.11458458e-01] | [7.99898099899292, -3.578993797302246] |
cfd281e1-fd5e-4036-bfeb-fef85b570ba0 | few-shot-3d-shape-generation | 2305.11664 | null | https://arxiv.org/abs/2305.11664v1 | https://arxiv.org/pdf/2305.11664v1.pdf | Few-shot 3D Shape Generation | Realistic and diverse 3D shape generation is helpful for a wide variety of applications such as virtual reality, gaming, and animation. Modern generative models, such as GANs and diffusion models, learn from large-scale datasets and generate new samples following similar data distributions. However, when training data is limited, deep neural generative networks overfit and tend to replicate training samples. Prior works focus on few-shot image generation to produce high-quality and diverse results using a few target images. Unfortunately, abundant 3D shape data is typically hard to obtain as well. In this work, we make the first attempt to realize few-shot 3D shape generation by adapting generative models pre-trained on large source domains to target domains using limited data. To relieve overfitting and keep considerable diversity, we propose to maintain the probability distributions of the pairwise relative distances between adapted samples at feature-level and shape-level during domain adaptation. Our approach only needs the silhouettes of few-shot target samples as training data to learn target geometry distributions and achieve generated shapes with diverse topology and textures. Moreover, we introduce several metrics to evaluate the quality and diversity of few-shot 3D shape generation. The effectiveness of our approach is demonstrated qualitatively and quantitatively under a series of few-shot 3D shape adaptation setups. | ['Jian Yuan', 'Jiansheng Chen', 'Huimin Ma', 'Jingyuan Zhu'] | 2023-05-19 | null | null | null | null | ['3d-shape-generation'] | ['computer-vision'] | [ 3.85984895e-03 -5.68599394e-03 1.56953990e-01 -1.31334037e-01
-6.68854892e-01 -6.17920458e-01 7.38755941e-01 -3.09082419e-01
1.83010563e-01 8.42979848e-01 6.36416152e-02 3.18162411e-01
2.76620209e-01 -1.18697011e+00 -8.25545132e-01 -6.30096436e-01
3.87559503e-01 7.91169286e-01 1.39918000e-01 -4.26224649e-01
1.46710649e-01 8.66497695e-01 -1.47985542e+00 -1.27857924e-01
1.06082749e+00 7.91288018e-01 2.20307410e-01 5.95033586e-01
-2.99712956e-01 1.01458065e-01 -7.22037673e-01 -4.84992266e-01
5.25861204e-01 -7.45603681e-01 -1.92510918e-01 5.18888414e-01
2.53108978e-01 -3.79005760e-01 -2.27689311e-01 9.30576265e-01
9.24885392e-01 2.81464159e-01 1.10260475e+00 -9.72363651e-01
-1.06691563e+00 1.18323406e-02 -6.54874265e-01 -2.39571199e-01
2.36208320e-01 5.67025185e-01 4.04973418e-01 -1.01562750e+00
1.06837010e+00 1.11875474e+00 5.70851982e-01 8.43482912e-01
-1.49736023e+00 -5.28885663e-01 -1.09282739e-01 -3.20872188e-01
-1.33984601e+00 -3.97243917e-01 1.19832158e+00 -3.14861447e-01
4.90113825e-01 -7.09349737e-02 9.48096693e-01 1.37237477e+00
6.02899380e-02 6.98240578e-01 8.12008619e-01 -2.21684709e-01
5.71763158e-01 1.00544751e-01 -8.71318161e-01 4.55487370e-01
2.76955783e-01 1.45935059e-01 -3.60267192e-01 -1.51948124e-01
1.52191103e+00 7.84533098e-02 -9.30797458e-02 -9.24890459e-01
-1.12239492e+00 9.14127052e-01 5.22654295e-01 2.51165241e-01
-5.23112833e-01 5.68078943e-02 5.21862805e-02 1.36365503e-01
7.15330362e-01 4.38534200e-01 -3.98834683e-02 -1.36956990e-01
-7.80729830e-01 4.50919122e-01 7.01414764e-01 1.29991961e+00
7.83645213e-01 6.06195688e-01 -2.30581656e-01 1.13064444e+00
6.57460885e-03 9.57828403e-01 4.48780239e-01 -8.82250428e-01
1.50025964e-01 3.93121183e-01 2.02369720e-01 -1.03150415e+00
1.45632058e-01 -3.55050445e-01 -1.20071328e+00 3.00123096e-01
2.76107430e-01 -3.00413996e-01 -1.25536072e+00 1.66428483e+00
5.94212294e-01 1.30048171e-01 -8.41504559e-02 9.31456268e-01
7.64746249e-01 7.19124973e-01 -2.78955400e-01 -1.41557768e-01
5.53814352e-01 -6.80448651e-01 -3.15275431e-01 -7.29159638e-02
2.17512965e-01 -8.29096437e-01 1.25353169e+00 -1.58599332e-01
-1.46510112e+00 -6.42290592e-01 -8.18242669e-01 2.54683286e-01
-4.91659604e-02 -2.80134439e-01 3.54478240e-01 5.60774863e-01
-8.53473067e-01 5.68017423e-01 -6.41947925e-01 -2.47308478e-01
8.34654152e-01 -9.40007903e-03 -1.37997448e-01 -2.68686622e-01
-8.47527206e-01 5.56977272e-01 2.57390477e-02 -4.34814692e-01
-1.01530325e+00 -8.10204268e-01 -8.84706616e-01 -5.67588545e-02
1.69640288e-01 -1.10986745e+00 1.06819093e+00 -6.92402661e-01
-1.93467999e+00 7.80827403e-01 -7.05496594e-03 4.18129973e-02
5.59898913e-01 2.48496249e-01 -1.02776624e-02 -1.05381228e-01
9.38972086e-02 8.48591566e-01 1.08027589e+00 -1.63161504e+00
-1.08743340e-01 -2.29556262e-01 -2.14484856e-01 3.14424932e-01
-1.38674766e-01 -6.44116282e-01 -3.20726693e-01 -9.40216780e-01
-1.55835282e-02 -7.52704859e-01 -5.14575422e-01 1.98126629e-01
-1.51769638e-01 2.26680726e-01 7.54785419e-01 -9.08158198e-02
5.35610914e-01 -2.13450980e+00 3.74960899e-01 1.74358800e-01
2.02950746e-01 2.97802895e-01 -4.02465284e-01 3.95280868e-01
3.02834988e-01 1.11505367e-01 -3.75289619e-01 -2.57506549e-01
-3.18445405e-03 2.01684788e-01 -2.89160222e-01 1.62886441e-01
3.06597501e-01 1.25933635e+00 -9.91279185e-01 -2.73438185e-01
3.67269307e-01 6.54874325e-01 -7.11871147e-01 3.63382548e-01
-4.58999425e-01 7.89450645e-01 -5.61357379e-01 5.42040229e-01
7.24014521e-01 -3.34668517e-01 -2.00968400e-01 -8.23472533e-03
3.14713031e-01 -2.80217648e-01 -9.86462057e-01 1.98792481e+00
-6.37200296e-01 2.87391692e-01 -2.12205112e-01 -7.77249753e-01
1.51521635e+00 1.07420787e-01 4.95857894e-01 -7.17860222e-01
2.28065297e-01 3.41396093e-01 -2.04662070e-01 -2.14169368e-01
4.49014932e-01 -5.65338850e-01 -1.85504407e-01 6.29275262e-01
7.93878660e-02 -9.87554371e-01 -1.75425671e-02 9.55717713e-02
6.06478095e-01 2.26525053e-01 2.42690817e-01 9.75684673e-02
1.12438157e-01 -7.05919489e-02 4.08266008e-01 5.35854101e-01
1.58851698e-01 1.17079127e+00 1.66499376e-01 -4.30936992e-01
-1.80185807e+00 -1.40076709e+00 2.24800348e-01 4.33035105e-01
2.74714947e-01 4.13557924e-02 -6.84154153e-01 -4.73929614e-01
1.75810885e-02 8.30777228e-01 -5.61964691e-01 -4.13592190e-01
-5.65760076e-01 -5.61591804e-01 1.85914978e-01 4.67402548e-01
3.89335901e-01 -1.22136629e+00 -3.78886938e-01 2.86888391e-01
2.15404898e-01 -8.56676221e-01 -6.62631512e-01 -3.26271415e-01
-9.91006017e-01 -7.20816314e-01 -1.43808854e+00 -8.08553159e-01
9.42982435e-01 2.31260881e-01 1.12757170e+00 -1.34064093e-01
-3.84874105e-01 2.84231007e-01 -2.91313440e-01 -3.52892637e-01
-7.02336431e-01 4.87672873e-02 -2.24526953e-02 -1.60490256e-02
-1.28350377e-01 -1.07170582e+00 -6.32264733e-01 3.73800844e-01
-9.55892503e-01 1.71570525e-01 6.58654153e-01 1.02948463e+00
9.69017506e-01 -2.26427957e-01 7.53213167e-01 -7.11742997e-01
8.18444908e-01 -3.96498024e-01 -3.90083849e-01 8.02870542e-02
-2.41708755e-01 -5.32771973e-03 8.72795463e-01 -8.51609170e-01
-1.10131669e+00 -7.35139921e-02 -1.42412826e-01 -1.01071572e+00
-3.40981781e-01 4.14482169e-02 -1.90297201e-01 -6.21937066e-02
9.51929390e-01 6.19578660e-01 2.94850141e-01 -2.88193285e-01
6.07598186e-01 3.06943178e-01 3.21428239e-01 -6.77055120e-01
1.06941342e+00 5.04952312e-01 5.05667878e-03 -8.35113883e-01
-5.61073244e-01 1.35764778e-01 -6.89037621e-01 -2.20236033e-01
4.80759382e-01 -6.30469024e-01 -1.54401630e-01 6.55835509e-01
-1.00564110e+00 -6.40852988e-01 -9.64447021e-01 1.19885735e-01
-9.65287328e-01 1.33247212e-01 -2.49734223e-01 -6.25604331e-01
-4.87300515e-01 -9.45285261e-01 1.20063996e+00 3.87500256e-01
-9.42402259e-02 -9.89274740e-01 2.50699610e-01 -2.37960145e-01
7.42578566e-01 6.27704978e-01 8.88866127e-01 -1.10697903e-01
-5.72254598e-01 -2.37490714e-01 -5.58804981e-02 9.03757587e-02
5.97293854e-01 -1.07023738e-01 -5.20413101e-01 -2.76505500e-01
-1.61776245e-01 -4.45523351e-01 4.47704375e-01 6.18652999e-01
1.09755456e+00 -9.82479751e-02 -1.80081159e-01 7.43485272e-01
1.27182949e+00 2.75897205e-01 6.87404752e-01 -2.78531671e-01
7.55169511e-01 3.48895669e-01 3.64881217e-01 5.92845023e-01
6.60505146e-02 6.44621551e-01 1.01639070e-01 -6.88842610e-02
-4.50491488e-01 -6.92603290e-01 -3.21005546e-02 7.78036833e-01
-2.10646108e-01 -3.33151579e-01 -6.43099368e-01 6.59665048e-01
-1.53049433e+00 -1.01165068e+00 4.82567817e-01 2.24193835e+00
8.33663642e-01 1.23824120e-01 2.51241803e-01 -2.91076094e-01
8.15940678e-01 7.14500099e-02 -9.79689479e-01 -7.71500319e-02
-1.95197642e-01 3.20059866e-01 3.24782878e-02 2.48145744e-01
-4.86812681e-01 1.15459096e+00 6.06463623e+00 1.06609201e+00
-1.33921278e+00 -2.66607311e-02 8.48195255e-01 -4.03666377e-01
-7.84738064e-01 -3.07954848e-01 -5.51362455e-01 5.09464979e-01
3.83680671e-01 -4.85006630e-01 4.53744739e-01 8.81767333e-01
1.27020359e-01 1.88895479e-01 -7.27212429e-01 1.25691438e+00
1.03012174e-01 -1.71026981e+00 3.16704690e-01 6.72757700e-02
1.48167539e+00 -2.16432869e-01 2.69228965e-01 9.34389010e-02
5.47721505e-01 -1.03735507e+00 7.32148468e-01 6.79686904e-01
1.32749188e+00 -8.96215081e-01 3.05380821e-01 4.01395470e-01
-1.06650066e+00 3.80890518e-01 -6.38750672e-01 2.13336870e-01
5.51120520e-01 6.47155523e-01 -8.35224271e-01 4.23399597e-01
3.39983165e-01 8.34288836e-01 -1.58560947e-01 1.08797991e+00
-1.61625445e-01 1.42392814e-01 -3.75321984e-01 -2.59501129e-01
7.39889517e-02 -3.58012915e-01 7.92078197e-01 5.29104769e-01
8.07167828e-01 1.52201340e-01 6.87610433e-02 1.34659660e+00
-2.03481004e-01 9.15281475e-02 -1.07118297e+00 -1.65887907e-01
7.27300167e-01 9.25185382e-01 -7.62667358e-01 -3.83688092e-01
9.35206190e-03 1.06650746e+00 1.78184822e-01 3.77663910e-01
-6.30707920e-01 -2.63881952e-01 5.71939945e-01 3.80146027e-01
5.12363672e-01 -5.32164574e-01 -5.05452812e-01 -1.15614629e+00
-2.05154911e-01 -7.60746181e-01 -1.85047403e-01 -8.72525752e-01
-1.65393806e+00 6.74646080e-01 -5.01809791e-02 -1.48599005e+00
-5.70500433e-01 -1.02613203e-01 -7.54424572e-01 8.26776683e-01
-1.01030207e+00 -1.22772765e+00 -5.19162953e-01 5.33357203e-01
7.84921288e-01 -4.90048945e-01 7.36309111e-01 -4.18065041e-02
-1.41377404e-01 6.21623635e-01 2.09684864e-01 -8.19667205e-02
5.97737849e-01 -8.98308933e-01 9.62686062e-01 5.02555728e-01
1.40749618e-01 3.15526783e-01 5.73135078e-01 -8.73469591e-01
-1.21184051e+00 -1.12198961e+00 1.39317274e-01 -1.86313972e-01
1.28906563e-01 -3.01457673e-01 -8.89400065e-01 1.67919070e-01
-8.06723014e-02 1.47724688e-01 3.05311382e-01 -3.61930609e-01
2.64267661e-02 1.91474542e-01 -1.35773945e+00 8.69500101e-01
1.46140075e+00 -1.52715951e-01 -2.16138780e-01 9.34425294e-02
5.44447780e-01 -6.34840608e-01 -9.92404938e-01 2.86265969e-01
4.14180160e-01 -9.40612793e-01 1.08792520e+00 -4.99734044e-01
5.48843682e-01 -9.13780108e-02 -1.41108006e-01 -1.88531542e+00
-3.17189455e-01 -7.71559417e-01 -1.78679779e-01 9.97036755e-01
2.59311020e-01 -4.63709772e-01 1.11907053e+00 3.85650992e-01
-1.45580158e-01 -8.26119184e-01 -5.75181603e-01 -9.38986838e-01
3.05036098e-01 -5.69742993e-02 1.09791911e+00 7.40873039e-01
-6.67129576e-01 2.51922727e-01 -4.81277823e-01 -3.40338528e-01
7.04035580e-01 5.02738416e-01 1.27748513e+00 -1.11104476e+00
-2.05514193e-01 -4.78168190e-01 -4.00747895e-01 -1.23533762e+00
-4.35026810e-02 -7.50822723e-01 -4.35429020e-03 -1.58169734e+00
9.25278198e-03 -7.05114543e-01 3.64258796e-01 4.14843559e-02
-4.15106043e-02 4.18104500e-01 1.83347002e-01 2.35015899e-01
-1.01892412e-01 1.31193745e+00 2.04960704e+00 -1.04110211e-01
-4.74047184e-01 9.13260058e-02 -6.71873093e-01 4.82242078e-01
7.09490001e-01 -1.85854197e-01 -7.09893644e-01 -4.22997504e-01
-3.74975689e-02 1.09942444e-01 3.22616458e-01 -8.66165936e-01
-1.20630994e-01 -5.03904998e-01 7.49613702e-01 -4.78628308e-01
4.86130327e-01 -4.87622052e-01 4.03644681e-01 2.58776248e-01
-3.33307162e-02 -3.48742753e-01 6.38338998e-02 6.10833228e-01
-2.08406849e-03 -2.04973994e-03 1.15140152e+00 -3.71401370e-01
-5.06917179e-01 8.58748376e-01 8.82254764e-02 4.16346967e-01
1.18093121e+00 -6.25742912e-01 8.49852115e-02 -6.79170012e-01
-7.03668714e-01 -1.48696408e-01 1.10096753e+00 4.33195442e-01
1.02366149e+00 -1.85577822e+00 -7.91523635e-01 5.64376771e-01
5.50214248e-03 5.66874921e-01 4.90091830e-01 2.78393716e-01
-4.64391410e-01 -5.23024164e-02 -5.54577172e-01 -6.82990849e-01
-6.44874275e-01 5.32310367e-01 3.65101337e-01 9.35100690e-02
-8.26827705e-01 8.23021233e-01 3.79568070e-01 -5.94114244e-01
-2.30683535e-01 5.74257821e-02 2.13655710e-01 -3.18341553e-01
3.02728146e-01 7.05393478e-02 -3.13969135e-01 -4.33521122e-01
1.64235204e-01 8.75045717e-01 1.26904756e-01 -1.03328764e-01
1.57486892e+00 -7.52284080e-02 4.49617684e-01 2.81419247e-01
1.02153265e+00 -1.25441039e-02 -1.80650771e+00 -2.09547058e-01
-7.24426329e-01 -8.83654773e-01 -2.98630625e-01 -4.26096976e-01
-1.37686026e+00 8.81711602e-01 3.38301212e-01 7.73471547e-04
8.35848510e-01 1.56937450e-01 1.15359282e+00 -2.09951424e-03
6.07785761e-01 -9.06526089e-01 6.44393861e-01 3.35396737e-01
1.08899474e+00 -1.03321052e+00 -2.31349602e-01 -2.84943432e-01
-8.01537931e-01 8.10643196e-01 7.18755364e-01 -4.21996772e-01
5.97136974e-01 2.36707866e-01 -1.73090398e-01 -1.93394512e-01
-5.64985693e-01 -2.60198005e-02 3.03762794e-01 1.16372526e+00
9.46665034e-02 -4.70355004e-02 1.14512220e-01 2.57571369e-01
-3.33147854e-01 9.12922155e-03 4.50645208e-01 6.83924675e-01
-3.52202535e-01 -1.14043522e+00 -2.56274819e-01 4.35226440e-01
1.66066214e-01 2.05226973e-01 -3.76985282e-01 6.24352515e-01
-1.36616468e-01 2.78176457e-01 1.84026733e-01 -2.77096093e-01
3.29225093e-01 -2.03978673e-01 7.90583789e-01 -7.36913621e-01
1.00433759e-01 1.82775512e-01 -3.65164220e-01 -1.76172540e-01
-2.11363003e-01 -5.53355992e-01 -1.04786110e+00 -4.92614508e-01
-6.18230458e-03 -1.22895241e-01 4.18698758e-01 6.03855133e-01
7.03466833e-01 3.34902197e-01 9.23960030e-01 -1.38047850e+00
-5.34647644e-01 -8.36682439e-01 -6.46712482e-01 6.76799893e-01
-1.45807326e-01 -8.17531228e-01 -1.11565910e-01 1.20433718e-01] | [9.040556907653809, -3.549912691116333] |
e1e4590c-26e1-417a-9801-64825b55e281 | the-sjtu-system-for-dcase2021-challenge-task | null | null | https://dcase.community/documents/challenge2021/technical_reports/DCASE2021_Xu_119_t6.pdf | https://dcase.community/documents/challenge2021/technical_reports/DCASE2021_Xu_119_t6.pdf | THE SJTU SYSTEM FOR DCASE2021 CHALLENGE TASK 6: AUDIO CAPTIONING BASED ON ENCODER PRE-TRAINING AND REINFORCEMENT LEARNING | This report proposes an audio captioning system for the Detection
and Classification of Acoustic Scenes and Events (DCASE) 2021
challenge task Task 6. Our audio captioning system consists of a
10-layer convolution neural network (CNN) encoder and a tempo-
ral attentional single layer gated recurrent unit (GRU) decoder. In
this challenge, there is no restriction on the usage of external data
and pre-trained models. To better model the concepts in an audio
clip, we pre-train the CNN encoder with audio tagging on AudioSet.
After standard cross entropy based training, we further fine-tune the
model with reinforcement learning to directly optimize the evalua-
tion metric. Experiments show that our proposed system achieves a
SPIDEr of 28.6 on the public evaluation split without ensemble1. | ['Kai Yu', 'Mengyue Wu', 'Zeyu Xie', 'Xuenan Xu'] | 2021-07-06 | null | null | null | dcase-challenge-2021-7 | ['audio-tagging', 'audio-captioning'] | ['audio', 'audio'] | [ 3.99308592e-01 1.00306623e-01 3.04811418e-01 -5.76910079e-01
-1.52554345e+00 -4.25910532e-01 -2.50715315e-02 -2.70247217e-02
-5.04160225e-01 4.72287536e-01 4.97094870e-01 1.92442853e-02
4.73500639e-01 -2.88302809e-01 -8.60778689e-01 -3.01015884e-01
-4.30480480e-01 1.78914532e-01 4.68309000e-02 4.39337455e-02
-2.90618569e-01 -1.24312572e-01 -1.42777967e+00 8.06080699e-01
1.01710476e-01 1.52019644e+00 2.06314370e-01 1.40833378e+00
4.98423517e-01 9.67499554e-01 -7.04887629e-01 -3.51842642e-01
-1.67167172e-01 -4.88043606e-01 -8.78953636e-01 -3.56435180e-01
2.62244701e-01 -2.87998676e-01 -4.44811702e-01 7.29499102e-01
8.30398262e-01 3.16423893e-01 3.30863893e-01 -1.09201956e+00
-3.60137403e-01 1.09789169e+00 -1.54901762e-02 5.20174503e-01
3.79582316e-01 3.92466187e-02 1.47468579e+00 -8.76899302e-01
1.14641890e-01 9.65970218e-01 7.26010859e-01 6.73126340e-01
-8.67096305e-01 -1.08315802e+00 -1.10300221e-01 4.72683161e-01
-1.51634145e+00 -6.41855657e-01 8.38431239e-01 -3.24759930e-01
1.15710282e+00 3.22346479e-01 3.82076979e-01 1.44357240e+00
-1.25010222e-01 7.39662528e-01 2.60414630e-01 -2.81815767e-01
3.13078851e-01 -1.85782954e-01 -6.37684837e-02 6.95208460e-02
-8.22971106e-01 1.26584535e-02 -6.56028152e-01 -8.85257870e-02
2.09267706e-01 -7.78611064e-01 -1.13627538e-01 3.09089988e-01
-8.27653229e-01 6.11883283e-01 3.68807882e-01 1.15436696e-01
-3.58356655e-01 6.42487168e-01 9.06738102e-01 1.81262091e-01
5.07667422e-01 6.23422086e-01 -6.07861876e-01 -5.02397239e-01
-9.89697397e-01 4.54308130e-02 4.67348039e-01 5.84858596e-01
1.49340197e-01 3.31229746e-01 -5.59387207e-01 1.11828458e+00
2.16428906e-01 4.02262658e-01 7.06361830e-01 -9.05336738e-01
7.27894545e-01 -2.74567753e-01 -2.19131753e-01 -4.06418085e-01
-2.27754772e-01 -6.23505771e-01 -6.55874908e-01 -4.15516436e-01
-2.16846585e-01 -5.43569088e-01 -8.91962826e-01 1.86623085e+00
-1.81671992e-01 7.32220471e-01 1.33739606e-01 8.75161469e-01
9.69505370e-01 1.14812613e+00 4.56579924e-01 -6.92514097e-03
1.41409826e+00 -9.33421135e-01 -8.10161710e-01 -1.47882015e-01
4.04867798e-01 -7.69011855e-01 9.00422096e-01 2.15559959e-01
-1.14096868e+00 -8.14001799e-01 -1.10036278e+00 -6.87915608e-02
-1.32082611e-01 3.81959707e-01 4.24592383e-02 3.96493495e-01
-8.51281881e-01 3.26587260e-01 -7.44539440e-01 7.83995315e-02
2.80500770e-01 4.19415504e-01 -5.25837839e-02 5.46238542e-01
-1.61570394e+00 3.19526494e-01 9.07130480e-01 2.43650172e-02
-1.53704762e+00 -8.63028944e-01 -9.39350784e-01 4.35502529e-01
4.51933630e-02 -2.59385616e-01 1.84623551e+00 -9.82239902e-01
-1.80843568e+00 6.93956375e-01 1.89151406e-01 -1.14688909e+00
1.00124337e-01 -5.14221370e-01 -8.38483930e-01 2.37564847e-01
-1.70309097e-01 1.02668703e+00 8.06029677e-01 -6.86519146e-01
-6.53969526e-01 5.23571253e-01 -9.70103666e-02 4.51525420e-01
-3.88864934e-01 5.93188286e-01 -4.12686378e-01 -7.91135013e-01
-5.32471061e-01 -8.69602740e-01 1.30215451e-01 -7.02577889e-01
-5.05716980e-01 -2.41781697e-01 8.01322579e-01 -9.94919479e-01
1.40114176e+00 -2.71857953e+00 -2.43002713e-01 -6.34737164e-02
-2.26437867e-01 3.99206787e-01 -5.26120067e-01 1.43979684e-01
-4.61614996e-01 -4.99687120e-02 -1.73371360e-01 -7.01541662e-01
1.23594582e-01 -1.16741106e-01 -8.11132967e-01 -4.31296527e-02
5.12982547e-01 6.43655241e-01 -9.51918125e-01 -3.24066669e-01
9.47618037e-02 7.78506577e-01 -9.06154275e-01 6.86348319e-01
-3.40917557e-01 2.19855189e-01 -8.76104012e-02 2.09109262e-01
2.61549085e-01 -4.44543734e-02 -1.00393757e-01 -2.56379824e-02
1.45263121e-01 1.06896770e+00 -1.04003406e+00 1.78654265e+00
-7.24709272e-01 8.71010721e-01 -1.59699783e-01 -8.55433881e-01
7.62572885e-01 9.93548751e-01 4.81936872e-01 -4.84395742e-01
1.36626437e-01 1.88573245e-02 -1.49703056e-01 -3.87394130e-01
4.91357923e-01 -6.16992153e-02 -3.69806945e-01 3.25483620e-01
6.08541191e-01 -2.02172622e-02 -1.96830839e-01 -1.97145529e-02
1.13048840e+00 -8.35201517e-02 -1.08205050e-01 4.16915268e-01
5.02731800e-01 -5.43477237e-01 5.34116864e-01 4.77261096e-01
-6.36275038e-02 1.22298110e+00 1.92976981e-01 -1.76613003e-01
-1.07333171e+00 -9.35044169e-01 -8.31238627e-02 1.62466168e+00
-5.21495223e-01 -8.27031374e-01 -8.64194334e-01 -5.52853823e-01
-4.80984986e-01 7.58183062e-01 -7.19620109e-01 -3.69346559e-01
-5.96360981e-01 -6.11181617e-01 1.30002773e+00 6.81785464e-01
1.95419908e-01 -1.48562324e+00 -5.31460941e-01 4.77259487e-01
-6.80118561e-01 -1.34642661e+00 -7.33107626e-01 5.89162111e-01
-3.84987593e-01 -5.11423528e-01 -5.04464030e-01 -8.57032359e-01
-1.18936196e-01 -4.47144181e-01 1.19820654e+00 -4.83378857e-01
6.17049336e-02 2.11075947e-01 -4.63659167e-01 -5.47872484e-01
-5.88036418e-01 5.20065606e-01 -6.10089973e-02 2.60111660e-01
4.05920744e-01 -5.94566107e-01 -4.73104149e-01 -3.57805863e-02
-5.22198856e-01 -2.42060255e-02 3.32923472e-01 7.62976706e-01
7.55871356e-01 -2.68662781e-01 8.40979517e-01 -1.26879558e-01
7.06702828e-01 -4.66678977e-01 -3.84873629e-01 1.17479274e-02
5.96929051e-04 -1.74318463e-01 5.35237074e-01 -5.21243989e-01
-8.25358152e-01 3.96696985e-01 -5.83068430e-01 -7.62533605e-01
-2.88277954e-01 3.81786376e-01 -1.58922538e-01 6.18672907e-01
5.14362335e-01 6.96330816e-02 -7.00713754e-01 -7.33725488e-01
2.29927823e-01 1.17403519e+00 9.91765738e-01 -2.58622527e-01
4.45761442e-01 -2.62187123e-01 -6.50317550e-01 -5.88045835e-01
-1.28110099e+00 -4.29462552e-01 -2.42290169e-01 -2.50227004e-01
1.39066911e+00 -1.39111030e+00 -7.26177692e-01 1.31825926e-02
-1.25779724e+00 -3.10078561e-01 -4.90586847e-01 7.81108141e-01
-7.50017166e-01 -2.60402203e-01 -7.86976337e-01 -9.79883969e-01
-6.67574584e-01 -8.75629187e-01 1.20421600e+00 4.13921662e-02
-4.19443965e-01 -4.95097131e-01 5.07607162e-01 4.53895658e-01
5.00445604e-01 8.10989738e-02 3.69049788e-01 -1.06858706e+00
-2.01020017e-01 -1.92400530e-01 8.14133286e-02 8.12871516e-01
-2.73766875e-01 -1.72668278e-01 -1.89158487e+00 -1.26981840e-01
-3.20190191e-01 -7.17953146e-01 9.17858362e-01 4.05412465e-01
1.77184939e+00 -2.57779032e-01 2.00520054e-01 6.55539989e-01
9.98061299e-01 4.55121160e-01 5.86636186e-01 7.77010098e-02
4.72323775e-01 2.72901297e-01 2.72025973e-01 4.94405717e-01
1.94507688e-01 8.16904783e-01 4.53567415e-01 8.79821703e-02
-3.09637755e-01 -5.60650527e-01 7.01063573e-01 1.27438939e+00
4.36197892e-02 -2.75675982e-01 -8.85184169e-01 7.09637165e-01
-1.40522325e+00 -1.11395550e+00 3.60814452e-01 1.96178782e+00
1.18373609e+00 1.84951618e-01 1.31037980e-01 3.67946595e-01
7.92361081e-01 4.00244724e-03 -1.87496737e-01 -6.22855484e-01
1.94282848e-02 5.09827971e-01 6.26015216e-02 5.12704909e-01
-1.57744086e+00 7.04984844e-01 6.60582685e+00 7.85714209e-01
-1.16436219e+00 3.21032077e-01 7.37905145e-01 -5.02638817e-01
2.15747386e-01 -4.26387578e-01 -7.05608308e-01 6.22628272e-01
1.87324286e+00 -1.11925513e-01 4.40903932e-01 6.93555355e-01
8.49061236e-02 5.59057117e-01 -1.16768599e+00 1.06926048e+00
4.50185761e-02 -9.45087135e-01 -1.76173910e-01 -2.02014565e-01
5.04476130e-01 4.45776701e-01 3.00106764e-01 8.17745030e-01
1.41996607e-01 -1.06229699e+00 8.78296733e-01 3.20837587e-01
9.91615355e-01 -8.89829278e-01 9.04942572e-01 -2.45478183e-01
-1.30647004e+00 -1.12115987e-01 -1.82930365e-01 7.18551725e-02
3.31681877e-01 3.70268673e-01 -1.32780159e+00 2.83294499e-01
9.89640117e-01 4.69521672e-01 -3.56242448e-01 1.27257144e+00
-1.25247583e-01 1.42032027e+00 -3.75832528e-01 2.38144696e-01
1.54809415e-01 5.28734982e-01 6.73393965e-01 1.72053969e+00
3.08907837e-01 -1.38826787e-01 -1.21778183e-01 4.66914058e-01
-6.33891106e-01 1.33968294e-01 -1.01670697e-01 -1.01457015e-01
4.69374418e-01 9.69039142e-01 -1.65266812e-01 -3.93470675e-01
-1.05878346e-01 7.60742784e-01 1.30150989e-01 2.86905646e-01
-1.16448390e+00 -7.01938152e-01 7.43900359e-01 -3.04642022e-01
6.85484111e-01 2.15804338e-01 7.82463923e-02 -1.00010478e+00
-1.98111087e-01 -7.28924572e-01 6.03863180e-01 -1.08606660e+00
-1.00302744e+00 9.88336205e-01 -1.63037807e-01 -1.23422945e+00
-6.89239919e-01 -1.38561413e-01 -5.87658346e-01 7.38286614e-01
-1.36204290e+00 -8.55066538e-01 8.14611167e-02 4.80583817e-01
6.24090970e-01 -2.73075551e-01 1.12805712e+00 6.44993484e-01
-5.33657312e-01 8.94608796e-01 -1.59879267e-01 5.14513016e-01
8.40243161e-01 -1.31031132e+00 5.01436770e-01 6.02140248e-01
4.27252263e-01 -1.03573024e-01 6.92547321e-01 -2.47646734e-01
-5.38957059e-01 -1.53178334e+00 1.13526154e+00 -3.95690948e-01
7.86022127e-01 -8.47983897e-01 -1.00189114e+00 7.08288431e-01
4.74628419e-01 5.92712834e-02 1.03806770e+00 2.41930708e-01
-3.31625521e-01 -2.22693190e-01 -4.85206753e-01 3.25170718e-02
4.85264182e-01 -1.21462405e+00 -7.31644332e-01 2.23378673e-01
1.18091106e+00 -6.19116724e-01 -9.66349840e-01 4.99805242e-01
5.41853845e-01 -2.80550361e-01 1.05122340e+00 -7.86307096e-01
4.54098612e-01 -1.69681817e-01 -4.49437678e-01 -1.23578572e+00
-8.29083249e-02 -7.57959247e-01 -2.52885491e-01 1.38643694e+00
6.96073949e-01 2.49426514e-01 4.46282089e-01 5.73085062e-02
-5.06474018e-01 -4.17373508e-01 -1.10235071e+00 -7.31981754e-01
-2.34082803e-01 -1.19045627e+00 6.13762617e-01 7.29870498e-01
4.05417383e-03 6.35476351e-01 -6.31414711e-01 4.42683011e-01
1.63172379e-01 -3.67314965e-01 2.96458900e-01 -1.04529548e+00
-5.64364374e-01 -1.50403157e-01 -3.69838208e-01 -6.37977242e-01
1.44166857e-01 -8.43744993e-01 4.43795204e-01 -9.51891959e-01
-2.04539225e-02 -9.49297473e-02 -1.11225736e+00 6.67872071e-01
-2.40051914e-02 6.18766963e-01 2.63619810e-01 -1.93563342e-01
-1.07986331e+00 8.11874092e-01 5.51020384e-01 -3.23882908e-01
-3.41654390e-01 1.24897741e-01 -2.88663417e-01 3.94845039e-01
1.00222385e+00 -7.88582742e-01 -2.27613658e-01 -3.69686902e-01
2.79870838e-01 8.64044949e-02 3.08163077e-01 -1.44179940e+00
3.96647975e-02 3.51053745e-01 2.17582420e-01 -6.42031372e-01
6.54517114e-01 -4.80611563e-01 4.19002660e-02 1.37821570e-01
-1.07874286e+00 -2.43213937e-01 4.89105552e-01 3.92890424e-01
-6.89470947e-01 -1.51100367e-01 6.36530817e-01 3.83703917e-01
-2.40734458e-01 1.62182868e-01 -5.10374308e-01 2.50948697e-01
5.94761968e-01 5.67110360e-01 3.88795622e-02 -6.72779977e-01
-1.18101025e+00 1.41567320e-01 -4.05336469e-01 8.10451746e-01
4.53589290e-01 -1.65596187e+00 -1.11260736e+00 1.19333185e-01
1.83451995e-01 -2.54262179e-01 3.80916417e-01 2.83671319e-01
-1.02533959e-01 5.88806510e-01 -7.77920410e-02 -4.50264961e-01
-1.12052155e+00 2.32573941e-01 6.30703211e-01 -2.93880761e-01
-3.33929956e-01 1.14804077e+00 3.38212907e-01 -1.31126493e-01
7.89412260e-01 -4.91258383e-01 -3.76816422e-01 5.63055836e-02
7.62171030e-01 8.86625871e-02 3.16204578e-01 -7.21432388e-01
-3.74721825e-01 1.64698847e-02 3.82415913e-02 -4.49678272e-01
1.37445319e+00 8.15639794e-02 4.85992253e-01 6.02035224e-01
1.50579095e+00 -2.93385476e-01 -1.20949864e+00 -7.99038187e-02
-2.05934778e-01 3.33115131e-01 3.71807516e-01 -8.43044043e-01
-9.82276320e-01 1.26905715e+00 9.32369232e-01 3.27520609e-01
1.32651103e+00 2.71291677e-02 9.52599764e-01 2.83303738e-01
-3.79803717e-01 -1.27794087e+00 2.11074591e-01 8.80340099e-01
1.05260932e+00 -9.66581166e-01 -6.54561579e-01 1.29362077e-01
-6.95312023e-01 9.89224255e-01 4.98766214e-01 -1.86939567e-01
7.38883138e-01 3.14909428e-01 2.45037326e-03 9.85262468e-02
-1.24117792e+00 -2.33691692e-01 6.93324089e-01 4.65036392e-01
5.34922838e-01 2.08178118e-01 3.91014010e-01 1.12304676e+00
-7.65051067e-01 -2.72161420e-02 3.02833170e-01 4.70916420e-01
-3.01621050e-01 -7.54337847e-01 -2.10996658e-01 5.06381169e-02
-1.02722561e+00 -3.05519521e-01 -1.95173234e-01 3.90528440e-02
1.06355347e-01 9.27904963e-01 5.12131572e-01 -7.68799007e-01
3.03478301e-01 4.78277355e-01 -7.38049969e-02 -6.73590362e-01
-9.10917461e-01 2.52235502e-01 3.52867424e-01 -4.40183520e-01
-1.14902116e-01 -5.71184039e-01 -1.00475943e+00 5.47699451e-01
-3.35816741e-01 5.96886098e-01 8.74161065e-01 7.63641477e-01
4.39535290e-01 1.11735845e+00 8.61017525e-01 -5.98139584e-01
-3.51539999e-01 -1.33936191e+00 -2.32893944e-01 1.72473058e-01
6.29154205e-01 -1.81493849e-01 -2.99689800e-01 5.78464568e-01] | [15.234484672546387, 5.02278470993042] |
c9376b6f-61aa-4d99-babc-e6e869468436 | design-considerations-of-a-coordinative | 2212.08535 | null | https://arxiv.org/abs/2212.08535v2 | https://arxiv.org/pdf/2212.08535v2.pdf | Design Considerations of a Coordinative Demand Charge Mitigation Strategy | This paper presents a coordinative demand charge mitigation (DCM) strategy for reducing electricity consumption during system peak periods. Available DCM resources include batteries, diesel generators, controllable loads, and conservation voltage reduction. All resources are directly controlled by load serving entities. A mixed integer linear programming based energy management algorithm is developed to optimally coordinate of DCM resources considering the load payback effect. To better capture system peak periods, two different kinds of load forecast are used: the day-ahead load forecast and the peak-hour probability forecast. Five DCM strategies are compared for reconciling the discrepancy between the two forecasting results. The DCM strategies are tested using actual utility data. Simulation results show that the proposed algorithm can effectively mitigate the demand charge while preventing the system peak from being shifted to the payback hours. We also identify the diminishing return effect, which can help load serving entities optimize the size of their DCM resources. | ['PJ Rehm', 'Di wu', 'Ning Lu', 'Hanpyo Lee', 'Hyeonjin Kim', 'Kai Ye', 'Rongxing Hu'] | 2022-12-16 | null | null | null | null | ['energy-management'] | ['time-series'] | [-4.56420571e-01 -3.24264541e-02 -5.19298792e-01 1.76306382e-01
-2.62226969e-01 -8.44146192e-01 2.88304597e-01 1.96983740e-01
3.20354372e-01 1.09645724e+00 1.13188446e-01 -1.77159801e-01
-4.90978509e-01 -1.15714264e+00 -8.65221471e-02 -9.78299260e-01
-1.05206259e-01 6.06848598e-01 -1.04438424e-01 -1.28011152e-01
2.29254231e-01 4.56214309e-01 -1.20621121e+00 -2.47499704e-01
1.62630069e+00 8.45103800e-01 7.55246997e-01 1.53691128e-01
2.82348454e-01 1.39164314e-01 -6.49198115e-01 4.19444144e-01
3.79363388e-01 -2.15790942e-01 1.86024323e-01 -3.76189917e-01
-1.32890105e+00 -5.73362470e-01 9.35874730e-02 1.05516982e+00
9.09247518e-01 2.96591729e-01 4.50691909e-01 -2.22870040e+00
-4.38847214e-01 7.76564777e-01 -4.98395443e-01 4.02102053e-01
9.31370945e-04 5.27250767e-02 6.45466387e-01 -5.45793176e-01
-3.24606597e-02 1.06436706e+00 1.01120465e-01 2.28331864e-01
-8.53148580e-01 -5.47375977e-01 1.45832226e-01 5.50450206e-01
-1.33515012e+00 1.69316396e-01 1.06942618e+00 -2.41693154e-01
1.35908961e+00 6.38218760e-01 1.09571159e+00 -3.77722263e-01
4.40518945e-01 8.35427999e-01 8.41735184e-01 -4.07342523e-01
3.81310791e-01 2.26332322e-01 -2.52228677e-01 -5.33805490e-01
5.09118557e-01 -2.17638522e-01 -3.00824437e-02 -2.34007269e-01
-1.52385220e-01 -5.39944433e-02 -5.48071802e-01 -1.49520099e-01
-3.51005584e-01 3.27903062e-01 9.82129648e-02 1.17284477e-01
-6.79082036e-01 -3.37894738e-01 1.55088320e-01 -3.62398773e-01
5.00921726e-01 1.12437168e-02 -5.81678391e-01 4.91770729e-02
-6.79035246e-01 -8.23018402e-02 7.24361479e-01 1.11899495e+00
2.15540871e-01 6.82495475e-01 -4.41933274e-01 4.33928430e-01
2.45829985e-01 1.10863578e+00 4.20623600e-01 -8.36763263e-01
3.91776770e-01 6.60321116e-01 8.29995394e-01 -5.99563062e-01
-3.05960238e-01 -4.03501004e-01 -7.31689632e-01 8.66907313e-02
-1.39528826e-01 -5.68864226e-01 4.48388513e-03 1.47019041e+00
6.31871000e-02 -3.93716656e-02 5.29762059e-02 6.45063877e-01
-2.52761960e-01 1.23204565e+00 -8.66958965e-03 -1.11890972e+00
1.10135341e+00 -5.68840861e-01 -1.06616819e+00 2.10618466e-01
3.99875969e-01 -3.93613517e-01 1.06521353e-01 -8.04469436e-02
-1.70073986e+00 3.20179947e-02 -8.78514767e-01 4.97673720e-01
-5.23891509e-01 3.70645016e-01 -1.82258353e-01 2.43956029e-01
-9.01727140e-01 4.59381878e-01 -6.07802093e-01 2.06435889e-01
2.62683295e-02 6.01909943e-02 7.67597258e-01 4.43407089e-01
-1.53754151e+00 1.11252999e+00 6.22062802e-01 5.39818347e-01
-4.80631262e-01 -1.05897593e+00 -5.01196980e-01 7.23723829e-01
3.22721303e-01 -3.38257253e-01 1.19819748e+00 -3.40471178e-01
-1.02453864e+00 -2.00840477e-02 -1.44407317e-01 -5.35195887e-01
5.27302384e-01 3.10286254e-01 -4.97845203e-01 -9.73089561e-02
-7.07376450e-02 -1.97611889e-03 7.08455965e-02 -1.09861112e+00
-9.34100032e-01 -1.99844316e-01 -5.45892358e-01 8.32627535e-01
-4.39724386e-01 -5.41113377e-01 2.43961185e-01 -3.67269814e-01
-7.22641587e-01 -7.55329728e-01 -1.65021285e-01 -7.85058320e-01
-6.17913723e-01 -8.75023723e-01 1.07507896e+00 -8.19430709e-01
1.47035182e+00 -1.72842002e+00 -2.12883297e-02 4.47123289e-01
-4.48268592e-01 8.33133385e-02 3.55917603e-01 9.93026912e-01
-2.19543949e-01 9.75695923e-02 1.34672716e-01 1.53263122e-01
5.77562988e-01 4.83496338e-01 -4.07216042e-01 9.10347179e-02
-3.68956238e-01 6.75824165e-01 -8.84245574e-01 3.08450669e-01
5.81805110e-01 1.15461156e-01 -7.61650801e-02 1.83734387e-01
-3.04864854e-01 -2.25349694e-01 -4.10519242e-01 6.05871856e-01
1.26274383e+00 7.38726109e-02 6.24852955e-01 -3.48609209e-01
-5.45410991e-01 -1.56041086e-01 -1.14098179e+00 5.90858042e-01
-5.13928711e-01 6.23085976e-01 3.92352521e-01 -1.16632807e+00
5.05081892e-01 8.30050558e-02 9.76467073e-01 -1.10849011e+00
-1.90758258e-01 1.38454691e-01 -2.95855142e-02 -2.84727186e-01
3.94521415e-01 2.00820476e-01 2.40230948e-01 5.06877422e-01
-7.47242212e-01 6.03532568e-02 6.06750131e-01 -2.69300509e-02
2.16980577e-01 -3.64788026e-01 2.81022489e-01 -1.04907978e+00
6.72956884e-01 -4.48816717e-02 1.18076396e+00 -2.77999163e-01
-1.49585128e-01 -5.30213594e-01 6.25657737e-01 2.83075929e-01
-9.43249822e-01 -9.13196743e-01 -4.03287821e-02 6.93642735e-01
6.02785468e-01 1.41694278e-01 -5.69046974e-01 2.91340034e-02
4.39819813e-01 1.83078074e+00 -1.10663204e-02 -2.18871772e-01
-5.86722553e-01 -1.22800088e+00 -5.39418340e-01 6.31992161e-01
2.82553613e-01 -4.95015115e-01 -8.72698128e-01 2.58394569e-01
-7.56216347e-02 -5.87136030e-01 -6.46743715e-01 1.78660154e-01
-2.51985878e-01 -1.01685774e+00 -8.92164767e-01 -4.40444916e-01
1.11546314e+00 4.16456819e-01 9.99363363e-01 2.12204918e-01
-7.21012875e-02 4.47727829e-01 3.85839492e-01 -9.61178362e-01
-2.27693394e-02 -3.40368181e-01 3.72896492e-01 -7.09932387e-01
-7.81485753e-04 -4.46958423e-01 -9.24163461e-01 2.29949698e-01
-5.74631095e-01 2.93661833e-01 4.57286313e-02 6.24093890e-01
7.45242119e-01 9.35009181e-01 1.46579373e+00 -2.45179817e-01
1.10266662e+00 -9.36759293e-01 -1.13713801e+00 7.03659177e-01
-1.45871282e+00 -1.56840920e-01 9.07339990e-01 -3.44286636e-02
-1.22662902e+00 -2.12344810e-01 4.56595242e-01 -2.51737952e-01
5.96783459e-01 2.21438020e-01 -7.90137589e-01 4.77177352e-01
-6.66195214e-01 4.32560593e-01 -2.53239214e-01 -3.69470865e-01
5.73010668e-02 4.48696434e-01 3.88092369e-01 -3.79466474e-01
9.26901221e-01 -6.91217333e-02 3.25930864e-01 8.28204583e-03
-4.38430905e-02 -2.07630545e-01 -1.67996898e-01 -4.96845067e-01
5.01930475e-01 -1.12064934e+00 -1.48155177e+00 5.48592806e-01
-7.54916906e-01 -2.70912707e-01 -1.22726738e-01 1.91202104e-01
-3.36804658e-01 2.36702219e-01 -1.37574464e-01 -1.28449380e+00
-6.12482548e-01 -8.76513600e-01 1.73827201e-01 8.84777486e-01
5.59687376e-01 -1.13037741e+00 4.71058860e-02 -2.32732207e-01
4.10260171e-01 3.58775407e-01 1.20712519e+00 -2.39594251e-01
-7.48475671e-01 2.40510002e-01 5.18684871e-02 4.17709678e-01
2.86803514e-01 8.38928744e-02 -4.29867595e-01 -9.00161207e-01
-3.13862830e-01 6.39400542e-01 -1.24158032e-01 5.14441848e-01
1.21468604e+00 -8.07791829e-01 -8.26825798e-01 9.55171138e-02
2.02374268e+00 1.16656280e+00 5.37474513e-01 3.09063524e-01
2.45300591e-01 4.50226605e-01 8.52876425e-01 1.10129011e+00
1.07734358e+00 6.12991929e-01 8.19894552e-01 4.91285436e-02
5.65728486e-01 -2.56055724e-02 2.18766525e-01 7.92379677e-01
2.83330493e-02 -8.41338456e-01 -7.40264237e-01 7.63970792e-01
-1.88267446e+00 -1.00492644e+00 -2.83190310e-01 2.29195762e+00
5.23776770e-01 1.68814678e-02 -7.37931654e-02 2.39990488e-01
8.67589653e-01 -2.28634343e-01 -9.97216642e-01 -5.92163682e-01
-5.05568385e-01 -6.48148537e-01 8.73872519e-01 3.29652816e-01
-1.33865252e-01 -2.41891652e-01 5.53330183e+00 7.82518744e-01
-7.62314677e-01 -1.48072854e-01 8.69455993e-01 -2.85807818e-01
-7.83092558e-01 -1.22019619e-01 -6.61640346e-01 1.52767813e+00
1.19609439e+00 -1.75295246e+00 7.97018290e-01 8.25728178e-01
1.15705633e+00 -6.55631065e-01 -7.24576294e-01 6.42732501e-01
-2.80589938e-01 -1.26781273e+00 -4.14911091e-01 1.83487281e-01
1.05741060e+00 -2.62675166e-01 -4.64116514e-01 2.07597733e-01
3.43773142e-02 -3.93775046e-01 6.90268517e-01 9.45636153e-01
1.52261272e-01 -1.57072580e+00 7.79478967e-01 7.76359200e-01
-1.55671203e+00 -7.08867967e-01 -1.44861668e-01 -9.50322077e-02
9.28181946e-01 9.19366121e-01 -6.58607304e-01 7.15724349e-01
5.72899163e-01 2.76985168e-01 -1.51114479e-01 9.99718487e-01
-9.33390483e-02 1.93822950e-01 -4.40523773e-01 -3.65146101e-02
-4.29677933e-01 -6.79772794e-01 4.23432231e-01 5.35020769e-01
5.54932654e-01 4.26055908e-01 3.52746457e-01 6.54355466e-01
5.56356385e-02 -3.41615468e-01 -3.44407350e-01 5.16371131e-02
1.38184488e+00 1.34232783e+00 -6.43085480e-01 -4.46634203e-01
-1.92383319e-01 5.36524177e-01 -5.37106216e-01 4.10772324e-01
-8.35409641e-01 -3.72993439e-01 5.89996099e-01 -7.85712674e-02
-1.82141498e-01 3.06613185e-02 -8.21086466e-01 -7.11473346e-01
1.71238586e-01 2.15958610e-01 3.42188150e-01 -1.05609095e+00
-1.30229115e+00 -3.88769478e-01 4.05266315e-01 -1.14279175e+00
-8.35522950e-01 -1.26155242e-01 -1.39685309e+00 1.35064340e+00
-2.11558819e+00 -3.86848241e-01 -1.09197214e-01 1.83652982e-01
7.07555592e-01 2.85551157e-02 4.30799872e-01 7.20100939e-01
-1.35011494e+00 5.42992316e-02 9.91717219e-01 -4.08765793e-01
-1.16641112e-01 -1.53665531e+00 -4.52446252e-01 1.01247454e+00
-1.23434234e+00 -1.45905256e-01 6.72210038e-01 -4.85948145e-01
-1.82113528e+00 -8.34856093e-01 6.99844718e-01 3.75434786e-01
4.85540271e-01 -1.05191238e-01 -9.23474193e-01 1.43232495e-01
1.10829282e+00 -5.76259375e-01 3.95699322e-01 -1.08054674e+00
7.28089631e-01 -3.87800097e-01 -1.36814523e+00 3.89216781e-01
3.08618009e-01 -1.30928442e-01 -1.79964118e-02 4.23868567e-01
1.11223817e-01 -2.08334267e-01 -9.48179364e-01 4.24091339e-01
1.88688233e-01 -1.79991126e-01 6.90014243e-01 1.21892199e-01
-3.90231162e-01 -3.88281047e-01 -6.48425817e-02 -2.24842596e+00
-2.43499756e-01 -8.09136987e-01 -5.67062616e-01 1.64721787e+00
2.13747382e-01 -1.04501987e+00 1.87630311e-01 1.12149441e+00
-2.73470432e-01 -7.59974062e-01 -1.19277346e+00 -8.17049861e-01
2.12268412e-01 1.92700654e-01 1.13004613e+00 1.02100503e+00
4.72831815e-01 -5.08974716e-02 3.13350894e-02 4.85347837e-01
1.05432439e+00 4.24421906e-01 -4.56566066e-02 -7.89947569e-01
4.71326441e-01 -5.99497557e-01 5.21587074e-01 -1.99685305e-01
3.16228002e-01 -7.27612555e-01 -7.76055083e-02 -2.21937728e+00
2.62463987e-01 -2.25090578e-01 -4.52567428e-01 6.67055011e-01
-5.55569455e-02 -4.53446567e-01 5.10793328e-01 3.13679039e-01
1.09780431e-01 1.04976153e+00 8.18958342e-01 -9.43095759e-02
-3.04888695e-01 1.63778976e-01 -3.56732249e-01 5.91086268e-01
1.21182048e+00 -1.62823536e-02 -1.02293801e+00 -1.92122266e-01
2.79083818e-01 4.60387915e-01 -3.50877009e-02 -6.23759508e-01
3.83040994e-01 -9.33105826e-01 2.54487991e-01 -1.46299064e+00
-2.42684454e-01 -1.22378063e+00 8.41236174e-01 8.44531417e-01
1.60914913e-01 3.58014286e-01 1.96691126e-01 4.95332420e-01
3.33230346e-01 8.70275497e-02 5.26589274e-01 2.76115596e-01
-8.31655025e-01 -9.94630828e-02 -5.78024805e-01 -2.75008976e-01
1.85599124e+00 2.56912899e-03 -9.54380035e-01 -4.06207561e-01
-5.26293337e-01 1.58987987e+00 2.01861799e-01 5.25590956e-01
1.10358998e-01 -1.40211236e+00 -4.27490205e-01 -3.12935501e-01
-6.15874410e-01 -4.32823867e-01 6.36278450e-01 6.98827267e-01
-5.22381142e-02 8.01072717e-01 -2.26567432e-01 -2.55731642e-01
-6.74800873e-01 5.26637852e-01 6.44803882e-01 -2.71998197e-01
-8.86846557e-02 -1.71892226e-01 -7.23309889e-02 1.42826527e-01
-1.35116696e-01 -2.52447665e-01 -2.08503217e-01 6.49326146e-01
3.69384170e-01 1.41891813e+00 4.20706160e-02 -1.48643941e-01
-6.22555554e-01 -1.34177050e-02 8.25647414e-01 4.79592294e-01
1.34750938e+00 -8.10518503e-01 -1.58329710e-01 1.61417633e-01
8.15286100e-01 -2.23033689e-02 -1.27591431e+00 4.88043219e-01
-1.52317509e-01 -2.01102555e-01 1.72303185e-01 -1.19843698e+00
-1.41263938e+00 2.01646551e-01 4.86687034e-01 1.08251166e+00
1.86256886e+00 -6.21585011e-01 7.65776515e-01 -9.65174362e-02
3.92482787e-01 -1.76982725e+00 -6.78968072e-01 1.59429297e-01
8.39220703e-01 -5.69834828e-01 2.71600783e-01 -2.34490633e-01
-5.47971010e-01 9.04311121e-01 5.56600273e-01 -2.62828860e-02
7.49074996e-01 4.08834577e-01 -8.16216052e-01 4.96283650e-01
-1.06119680e+00 8.76989067e-02 -1.07998125e-01 4.56119120e-01
-2.09523156e-01 6.46213472e-01 -9.14353430e-01 8.88493001e-01
2.54936904e-01 -1.55716976e-02 1.01784968e+00 9.67059135e-01
-2.89709210e-01 -5.99062979e-01 -3.68423045e-01 4.56364453e-01
-2.04567969e-01 2.14894861e-01 2.97232777e-01 6.07436299e-01
3.38805974e-01 8.79882991e-01 6.00455999e-01 4.16783303e-01
6.98224664e-01 -1.23189993e-01 -3.21208358e-01 1.40168723e-02
4.24394645e-02 2.97303677e-01 -3.63729801e-03 -6.72601461e-02
-6.33383617e-02 -6.50891304e-01 -1.89619756e+00 -4.83619958e-01
-5.68678975e-01 6.78528368e-01 8.26475203e-01 9.24301505e-01
6.28070176e-01 6.09973073e-01 1.53378463e+00 -9.53117549e-01
-6.23314857e-01 -5.32634199e-01 -9.81971085e-01 -3.87273371e-01
-1.29213795e-01 -6.50860608e-01 -7.61777401e-01 -2.19936460e-01] | [5.643326282501221, 2.4878182411193848] |
c8c41e6c-90b7-4453-b2a8-79744da0865f | elastichash-semantic-image-similarity-search | 2305.04710 | null | https://arxiv.org/abs/2305.04710v1 | https://arxiv.org/pdf/2305.04710v1.pdf | ElasticHash: Semantic Image Similarity Search by Deep Hashing with Elasticsearch | We present ElasticHash, a novel approach for high-quality, efficient, and large-scale semantic image similarity search. It is based on a deep hashing model to learn hash codes for fine-grained image similarity search in natural images and a two-stage method for efficiently searching binary hash codes using Elasticsearch (ES). In the first stage, a coarse search based on short hash codes is performed using multi-index hashing and ES terms lookup of neighboring hash codes. In the second stage, the list of results is re-ranked by computing the Hamming distance on long hash codes. We evaluate the retrieval performance of \textit{ElasticHash} for more than 120,000 query images on about 6.9 million database images of the OpenImages data set. The results show that our approach achieves high-quality retrieval results and low search latencies. | ['Bernd Freisleben', 'Markus Mühling', 'Nikolaus Korfhage'] | 2023-05-08 | null | null | null | null | ['image-similarity-search'] | ['computer-vision'] | [-7.39616528e-02 -7.00947344e-01 -4.52475160e-01 -3.99562985e-01
-1.24271107e+00 -5.41131914e-01 3.69676471e-01 7.20286489e-01
-6.76787198e-01 1.12383083e-01 2.81928092e-01 1.16740711e-01
-2.24059969e-01 -8.70500326e-01 -6.02781713e-01 -6.04154468e-01
-3.70704651e-01 7.46016383e-01 8.65062237e-01 5.39302127e-03
6.27989054e-01 4.85587150e-01 -2.06862879e+00 4.62098867e-01
3.73721272e-02 1.38545823e+00 4.18807238e-01 7.27570772e-01
1.69237062e-01 4.93143559e-01 -2.85114884e-01 -4.45336491e-01
4.07279044e-01 -5.46732694e-02 -1.01001024e+00 -5.69394946e-01
6.81147218e-01 -8.66017640e-01 -9.92153943e-01 9.72135425e-01
9.58685100e-01 1.71301186e-01 3.44440430e-01 -1.22175884e+00
-7.45392203e-01 3.79969180e-01 -4.65030611e-01 6.32463872e-01
6.75603747e-01 4.60290872e-02 1.39164066e+00 -1.13841569e+00
7.59977520e-01 1.28355670e+00 7.60297477e-01 7.67179728e-02
-1.01217985e+00 -7.19026148e-01 -1.01209760e+00 6.18835688e-01
-2.22671556e+00 -4.55399662e-01 -1.38818800e-01 -6.54707029e-02
1.25249887e+00 2.63652027e-01 5.73958039e-01 2.71417379e-01
3.20453614e-01 5.38136840e-01 7.86117315e-01 -2.70944953e-01
3.15856844e-01 -3.71227801e-01 2.13491514e-01 6.57792687e-01
6.36146665e-02 3.67949992e-01 -7.26128221e-01 -6.78782105e-01
5.47802150e-01 1.14131622e-01 8.65494609e-02 -2.09914133e-01
-1.27447915e+00 1.05572546e+00 8.94114614e-01 1.58064201e-01
-6.28792226e-01 4.79960710e-01 7.31498003e-01 4.12550479e-01
-3.09076458e-01 9.03669447e-02 1.36386126e-01 4.45286706e-02
-1.25321686e+00 4.21927154e-01 7.36663222e-01 8.97663772e-01
1.00471950e+00 -9.29029703e-01 -4.33688313e-01 8.49316478e-01
1.83459207e-01 9.44645643e-01 7.70053029e-01 -1.17820084e+00
-6.26811385e-02 4.27664332e-02 -1.74810812e-01 -1.41758168e+00
9.21098292e-02 2.77857959e-01 -5.70087194e-01 -2.38041475e-01
-1.68498203e-01 9.41266239e-01 -8.28882515e-01 1.13935602e+00
3.83440912e-01 1.27624840e-01 -1.59865811e-01 9.54583645e-01
9.83538032e-01 8.58906627e-01 4.38580150e-03 1.52402103e-01
1.67740273e+00 -1.12180865e+00 -4.06454653e-01 2.51394212e-01
4.46739763e-01 -1.14381135e+00 8.35565507e-01 -6.10254519e-02
-1.34046662e+00 -7.72359550e-01 -9.22417700e-01 -5.00922859e-01
-5.04392982e-01 -5.62248886e-01 1.97640553e-01 1.63387030e-01
-1.65512490e+00 6.14336371e-01 -4.93000001e-01 -4.95608121e-01
1.55797735e-01 4.92052257e-01 -4.29125816e-01 -4.14271295e-01
-1.34330547e+00 5.14853060e-01 6.00462794e-01 -7.60711432e-01
-7.80750394e-01 -4.70964342e-01 -7.67063081e-01 3.35127920e-01
-2.78956354e-01 -4.07927871e-01 1.29850733e+00 -4.87973988e-02
-7.13409781e-01 1.43597186e+00 -2.05758095e-01 -6.87716484e-01
-2.70938933e-01 -1.50032667e-02 -3.60776216e-01 9.43336904e-01
4.30064082e-01 1.15143299e+00 9.05972123e-01 -6.21655822e-01
-6.67120814e-01 -3.99234444e-01 -3.57469052e-01 1.74142256e-01
-4.86956179e-01 3.55714560e-01 -9.13889706e-01 -6.55939043e-01
7.11210668e-02 -1.04689014e+00 1.56663302e-02 4.00592476e-01
4.22350951e-02 -4.06986177e-01 7.66156852e-01 -4.22148526e-01
1.55005264e+00 -2.50255656e+00 -1.28836349e-01 6.03537619e-01
2.37915561e-01 1.93226978e-01 -4.34964329e-01 6.82310820e-01
2.47668877e-01 -1.34388864e-01 2.60601163e-01 -5.97667024e-02
2.47092739e-01 1.79774284e-01 -4.82542723e-01 3.79874438e-01
-5.54495037e-01 1.22939074e+00 -7.68704593e-01 -1.06044590e+00
4.82567539e-03 4.19582486e-01 -5.38275421e-01 3.21545094e-01
4.07898247e-01 -6.23807192e-01 -2.97228426e-01 6.98909402e-01
6.11316025e-01 -7.99210012e-01 -3.52590889e-01 -5.20051599e-01
4.51894552e-02 2.43516326e-01 -9.22333837e-01 1.92221081e+00
-3.69613171e-02 4.47042942e-01 -2.72333682e-01 -4.39859867e-01
7.04101264e-01 5.20332120e-02 5.24723291e-01 -1.33371794e+00
-1.74056709e-01 5.66271126e-01 -7.82995939e-01 -1.38035253e-01
8.04304361e-01 4.35907632e-01 -3.30632746e-01 8.33689153e-01
-1.53583720e-01 -2.36070901e-01 1.30254179e-01 3.43769789e-01
1.30194545e+00 -7.32195437e-01 3.16373765e-01 -2.38411754e-01
5.54009676e-01 1.63569570e-01 -1.38599887e-01 1.04729760e+00
-2.46500313e-01 6.10292792e-01 -5.49753308e-01 -8.95224154e-01
-1.28782451e+00 -1.36925709e+00 -3.02247345e-01 1.29878151e+00
7.75208056e-01 -9.03253376e-01 -7.18982458e-01 -1.40473455e-01
4.18530226e-01 -2.50951231e-01 -2.32024714e-01 -4.35547054e-01
-4.05485392e-01 -2.47742653e-01 8.15983653e-01 2.58043349e-01
8.74510169e-01 -1.34721327e+00 -1.00538754e+00 1.09164901e-02
-3.64818126e-01 -1.12405515e+00 -1.04002094e+00 3.14588994e-02
-5.73057473e-01 -8.45998645e-01 -9.23086107e-01 -1.32552505e+00
1.76739752e-01 7.25358963e-01 1.35082686e+00 6.18770480e-01
-9.43043888e-01 2.44184956e-01 -4.88409340e-01 4.97801214e-01
-2.51428396e-01 -5.37683209e-03 -7.03181699e-02 -5.19390762e-01
7.13858843e-01 -2.50047594e-01 -1.23064578e+00 6.98463082e-01
-1.04076493e+00 -6.72313392e-01 6.75130427e-01 8.51915956e-01
1.11156225e+00 1.91192672e-01 -1.52838379e-01 1.38893098e-01
4.12088007e-01 -3.11775386e-01 -6.31157160e-01 3.51178557e-01
-9.10696149e-01 2.04463333e-01 9.99840870e-02 -4.20341104e-01
-5.90033494e-02 8.24165419e-02 -1.66821748e-01 -5.49085557e-01
5.49202785e-02 2.91812688e-01 6.44661307e-01 -5.15676022e-01
7.02034593e-01 6.90560460e-01 5.87858148e-02 -2.38935500e-01
3.26539576e-01 9.36693728e-01 9.63519275e-01 -3.10762852e-01
7.93383062e-01 4.35941041e-01 -1.57716706e-01 -5.15845835e-01
-3.69693071e-01 -1.05622065e+00 -3.38522106e-01 2.09489226e-01
9.59705651e-01 -1.12388337e+00 -8.66524458e-01 4.21642333e-01
-8.79420877e-01 -2.01022863e-01 -2.60904998e-01 2.65386581e-01
-6.37684822e-01 6.40074968e-01 -1.19507539e+00 -1.29868668e-02
-8.90398979e-01 -1.15260744e+00 1.79598236e+00 1.77518874e-01
-2.09661216e-01 -3.05990428e-01 4.09170628e-01 3.26355934e-01
5.46293199e-01 -4.20299679e-01 8.18297923e-01 -6.40570521e-01
-9.67145979e-01 -2.68148005e-01 -6.57734096e-01 -2.39067167e-01
-2.95905709e-01 -5.08102834e-01 -5.94600022e-01 -7.29269445e-01
-5.98067403e-01 -1.00233185e+00 7.06690907e-01 4.76183631e-02
1.30928588e+00 -2.29794070e-01 -4.52703327e-01 6.60592854e-01
1.67435217e+00 1.41155124e-01 1.02516615e+00 5.62786758e-01
9.68580619e-02 2.40418110e-02 8.82531047e-01 7.89961994e-01
6.00579917e-01 9.87225413e-01 1.61087364e-01 -8.40531662e-03
-2.81587571e-01 -3.25996637e-01 -6.40162975e-02 7.64779389e-01
6.75505698e-01 -7.32995421e-02 -8.33468795e-01 6.47243977e-01
-1.54680908e+00 -1.23152125e+00 2.83603579e-01 2.38096380e+00
9.82453942e-01 -9.07759666e-02 1.29354745e-01 1.57147437e-01
1.03026843e+00 1.55333966e-01 -5.09352148e-01 -2.24546462e-01
-8.85345340e-02 5.53336978e-01 6.29296899e-01 2.40403578e-01
-9.91378784e-01 1.00813603e+00 7.21838999e+00 1.23013067e+00
-9.47584569e-01 -2.11819876e-02 3.71741980e-01 1.23832069e-01
-1.16187736e-01 -2.09539875e-01 -9.26516473e-01 5.67979395e-01
9.84075904e-01 -2.72453427e-01 6.83553815e-01 8.67064059e-01
-5.51805317e-01 -2.13060781e-01 -8.52426052e-01 1.64670169e+00
1.62168890e-01 -1.58535612e+00 3.85311455e-01 1.99771032e-01
5.40934443e-01 3.76306146e-01 2.26526096e-01 5.16680032e-02
3.06695700e-02 -8.82350445e-01 6.55718207e-01 8.76692981e-02
1.15252995e+00 -6.74196601e-01 5.79460859e-01 -4.19675149e-02
-1.63296390e+00 -1.33566245e-01 -5.72948456e-01 5.25223315e-01
-2.36445349e-02 3.12596351e-01 -4.23184901e-01 -9.91023704e-02
1.48163736e+00 4.91956860e-01 -6.21717334e-01 1.27249002e+00
5.75140715e-01 -1.94689244e-01 -4.84007120e-01 1.01678535e-01
3.31323057e-01 5.46033740e-01 -3.92285362e-03 1.28413093e+00
5.24543643e-01 3.49383265e-01 1.87199637e-01 3.46656263e-01
-1.86407834e-01 2.93376118e-01 -7.24315286e-01 -4.59290966e-02
1.10386872e+00 1.04902303e+00 -9.28783834e-01 -6.05250478e-01
-1.88003719e-01 1.56212819e+00 4.96298857e-02 -2.00978026e-01
-7.18802512e-01 -8.21181536e-01 5.74138284e-01 1.59517210e-02
7.35094190e-01 -6.30134940e-02 4.96483147e-01 -6.65323198e-01
-7.84774497e-02 -9.94078398e-01 9.98814881e-01 -9.82882500e-01
-1.18144059e+00 7.07165420e-01 -1.54841349e-01 -9.81658936e-01
-4.33051229e-01 -4.86953445e-02 -4.94300015e-02 5.98874271e-01
-1.59587348e+00 -6.91559196e-01 -6.21234000e-01 1.23216748e+00
4.05660182e-01 7.57362519e-04 1.13124812e+00 7.06904054e-01
2.37755865e-01 9.59540665e-01 3.05049926e-01 -3.58698927e-02
9.56305623e-01 -7.89780617e-01 8.45511377e-01 1.45078883e-01
2.24704966e-01 6.43131495e-01 3.37466985e-01 -5.14129639e-01
-1.53307378e+00 -6.19680762e-01 1.42855906e+00 2.74532642e-02
6.13057911e-01 -6.36401549e-02 -9.81049299e-01 -1.35638919e-02
7.79129714e-02 7.09658206e-01 8.07286024e-01 -7.45035291e-01
-9.39586163e-01 -2.97873527e-01 -1.32582200e+00 2.76488811e-01
9.70081627e-01 -1.23340714e+00 -6.00021303e-01 3.10675085e-01
8.36047530e-01 -2.85422355e-01 -1.35494244e+00 3.97772491e-01
9.67837095e-01 -1.04485595e+00 1.68809676e+00 9.37530845e-02
1.39815256e-01 -2.49618828e-01 -6.98771954e-01 -4.21557933e-01
-6.67207778e-01 -6.03859961e-01 -7.60212690e-02 4.23701882e-01
-9.30520073e-02 -2.50121564e-01 7.33259916e-01 2.27478936e-01
4.89075869e-01 -6.37647688e-01 -8.55925143e-01 -8.86306167e-01
-4.28149104e-01 3.19692940e-01 8.66478384e-01 5.38105905e-01
-1.27266258e-01 -6.34002611e-02 -8.28832984e-02 1.66390806e-01
9.69497740e-01 7.76771665e-01 4.95197415e-01 -1.15614104e+00
-3.19292217e-01 -2.32598200e-01 -8.50840390e-01 -1.19409788e+00
1.76148955e-03 -8.38274419e-01 1.98309675e-01 -1.00990343e+00
7.07156241e-01 -6.00482166e-01 -4.12500620e-01 4.25483793e-01
1.64023846e-01 1.11233270e+00 1.10274397e-01 8.68291199e-01
-1.29266739e+00 3.47166032e-01 7.17101097e-01 -2.17024013e-01
4.40757364e-01 -6.39314234e-01 -4.76085785e-04 3.30658220e-02
3.78011256e-01 -6.33137524e-01 -3.31905484e-01 -4.05554712e-01
1.24569930e-01 1.76263928e-01 5.74938774e-01 -1.17531300e+00
8.35466802e-01 3.07538420e-01 1.84349626e-01 -9.37812150e-01
2.82172948e-01 -7.77162015e-01 2.25011423e-01 9.29863095e-01
-6.27729595e-01 8.31644356e-01 -5.91511354e-02 3.42997074e-01
-5.32552898e-01 -1.42849132e-01 8.87343764e-01 -1.06194109e-01
-1.14831519e+00 5.57087421e-01 -6.97973296e-02 1.32641137e-01
8.54999721e-01 -3.07841152e-01 -1.60596624e-01 -4.06875879e-01
-4.17846322e-01 1.91914022e-01 7.82402873e-01 4.12083834e-01
1.07388973e+00 -1.73055875e+00 -4.04771119e-01 3.56883883e-01
6.31217360e-01 -4.10456449e-01 3.30973119e-02 1.83458641e-01
-7.55327225e-01 5.54449379e-01 -3.83882225e-01 -7.24897265e-01
-1.57759452e+00 8.08373332e-01 -1.16468504e-01 -1.23404577e-01
-6.73520088e-01 8.43962550e-01 -1.30395815e-01 2.03840926e-01
5.12548149e-01 3.39366138e-01 2.98552305e-01 -1.46601528e-01
1.11146331e+00 3.57332140e-01 1.50755703e-01 -8.65017891e-01
-4.53368872e-01 1.10994947e+00 -2.07708210e-01 -2.16782138e-01
9.08005714e-01 -5.34294188e-01 -4.49104875e-01 -1.81044906e-01
2.00855875e+00 -5.49523473e-01 -5.00609457e-01 -5.65467358e-01
1.51224196e-01 -7.64283776e-01 1.72419623e-01 -3.24707240e-01
-9.86651182e-01 4.10780549e-01 1.03520262e+00 -6.85825795e-02
1.36917007e+00 3.89261037e-01 1.80579567e+00 8.37195456e-01
7.97393978e-01 -1.10181201e+00 4.89840448e-01 3.65990341e-01
6.03088319e-01 -1.25615752e+00 -1.33987129e-01 -2.58846339e-02
-5.16342111e-02 1.05243826e+00 2.42024586e-01 -1.49550810e-01
7.88349271e-01 2.62978762e-01 -1.47243720e-02 -5.12319088e-01
-6.61446512e-01 -1.31072327e-01 1.78548828e-01 1.74962819e-01
-1.59756988e-01 -3.15039635e-01 -1.32536396e-01 -3.54448080e-01
-1.79869279e-01 4.06192020e-02 -3.90881866e-01 1.19818842e+00
-7.53426671e-01 -1.00332344e+00 -4.90091473e-01 1.51839122e-01
-3.92607987e-01 -5.16422987e-01 -6.20629899e-02 2.57370740e-01
-2.73603946e-01 6.54625416e-01 3.69137675e-01 -6.59784317e-01
6.54472411e-02 -3.27001095e-01 3.49905074e-01 -1.62892833e-01
-6.14007235e-01 -9.96610597e-02 -6.06781781e-01 -1.31242454e+00
-1.06973104e-01 -3.97697181e-01 -1.48225415e+00 -6.13025308e-01
-8.78816247e-02 4.86979246e-01 6.96018696e-01 2.44730145e-01
7.33733058e-01 -6.61873519e-01 8.76502156e-01 -7.38779843e-01
-8.58943999e-01 -1.05205625e-01 -4.89527285e-01 7.17615724e-01
4.72880334e-01 -2.84326643e-01 -4.45341796e-01 3.13375331e-02] | [11.20203971862793, 0.9536350965499878] |
e5a86f46-e51d-4f84-8f09-0160f1a8898f | instance-segmentation-of-multiple-myeloma | null | null | https://openreview.net/forum?id=T1ZK_GYtdbn | https://openreview.net/pdf?id=T1ZK_GYtdbn | Instance Segmentation of Multiple Myeloma Cells via Hybrid Task Cascade | Multiple Myeloma (MM) is a blood cancer that develops when plasma cells expand abnormally in the bone marrow. Early detection of MM is beneficial for accurate treatment in time and draws increasing recognition. There are several endeavors to construct computer-assisted automatic diagnostic tools for myeloma cell detection. Of late, deep learning based methods have been expected to apply for the detection and segmentation of cells of interest from microscopic images. In this paper, an ensemble framework is designed for the detection and segmentation of myeloma cells. Extensive experimental results on the SegPC-2021 Challenge dataset demonstrate that the proposed method achieves promising performance. | ['Anonymous'] | 2021-07-20 | null | null | null | miccai-workshop-compay-2021-9 | ['cell-detection'] | ['computer-vision'] | [-1.28884450e-01 -2.64685541e-01 2.42335275e-02 -2.37972751e-01
-1.04642868e+00 -3.82813020e-03 4.37443256e-01 5.05091906e-01
-3.99270803e-01 6.53588474e-01 -1.62469432e-01 4.89067361e-02
5.49449503e-01 -7.32562959e-01 -5.62929036e-03 -1.18801677e+00
2.27360949e-01 1.40408671e+00 2.06452444e-01 1.69465810e-01
2.74303854e-01 1.12621975e+00 -6.80740297e-01 5.70705116e-01
8.14657211e-01 7.42182493e-01 1.26573056e-01 1.12937033e+00
-5.52697897e-01 1.19402826e+00 -6.05214834e-01 -1.79255903e-01
-1.89691618e-01 -4.98047531e-01 -8.57928753e-01 4.59628910e-01
1.09420218e-01 -1.62013292e-01 -3.12015206e-01 7.93017030e-01
7.20128834e-01 -3.53253573e-01 1.10663342e+00 -1.04976726e+00
-1.05288759e-01 4.01060581e-01 -9.61355388e-01 7.93319523e-01
1.47733480e-01 4.51385617e-01 3.95228177e-01 -7.50523031e-01
5.17806649e-01 1.01180840e+00 5.67002833e-01 6.54210269e-01
-1.05541217e+00 -5.02767682e-01 -6.69596374e-01 3.86353910e-01
-1.42186809e+00 -2.54991859e-01 1.31065339e-01 -8.17782223e-01
9.70092893e-01 3.86901289e-01 5.13558865e-01 7.32242763e-01
5.77260077e-01 1.24426365e+00 1.21237087e+00 -5.87350190e-01
3.30907524e-01 -2.66279906e-01 5.32643676e-01 8.10207009e-01
7.17283487e-02 -6.18375391e-02 -5.10969639e-01 -1.13932006e-01
8.65758121e-01 2.30337039e-01 -1.65887937e-01 1.20720305e-02
-9.13438976e-01 9.58688438e-01 3.60941947e-01 6.54641151e-01
-4.07514811e-01 1.94611758e-01 2.24720359e-01 -4.52485383e-02
3.44125599e-01 -5.53526916e-02 1.91880956e-01 2.76500732e-02
-1.14483261e+00 2.79793411e-01 5.99855244e-01 4.05841678e-01
1.74873188e-01 -3.66599917e-01 -3.59981567e-01 7.70775914e-01
3.78145933e-01 4.69955236e-01 6.16999686e-01 -5.24845481e-01
-2.27852046e-01 1.21575809e+00 -1.06337681e-01 -4.76469487e-01
-9.70791399e-01 -4.00956005e-01 -1.14886522e+00 2.92654693e-01
8.76529396e-01 1.79670721e-01 -1.29689443e+00 8.53452384e-01
4.09950882e-01 4.31819052e-01 -2.10762739e-01 9.81244564e-01
8.64075422e-01 2.49023795e-01 2.27314025e-01 -7.97303915e-02
1.46135032e+00 -7.68359780e-01 -4.63219732e-01 6.30013347e-02
8.42196941e-01 -6.58302486e-01 4.92367476e-01 1.57627389e-01
-8.12032700e-01 1.19479217e-01 -6.74552977e-01 1.97920561e-01
-2.27272063e-01 -2.74842978e-01 8.44410360e-01 7.29565740e-01
-1.06263995e+00 2.59636432e-01 -1.24975443e+00 -7.56622553e-01
9.05039966e-01 7.69283533e-01 -1.92091852e-01 -2.75544971e-02
-4.89752114e-01 9.13171768e-01 2.63744146e-01 -3.57374698e-02
-5.70979714e-01 -7.01844454e-01 -2.72770643e-01 -2.79670417e-01
-4.01646882e-01 -1.30967009e+00 1.32584262e+00 -3.73893768e-01
-1.21846557e+00 1.45878363e+00 -4.19445574e-01 -7.65787899e-01
6.26173079e-01 3.80888999e-01 -3.47334087e-01 4.57618237e-01
-7.56652653e-02 5.52507699e-01 3.17205876e-01 -9.88945544e-01
-1.41327965e+00 -8.64013016e-01 -6.71181023e-01 -5.48807904e-02
2.55921274e-01 3.43770266e-01 -4.65868145e-01 -4.66242582e-01
3.44219118e-01 -9.38950062e-01 -5.63908994e-01 -3.41875553e-01
-5.65550089e-01 -4.56527263e-01 7.54329920e-01 -5.89910507e-01
5.55531919e-01 -1.43281591e+00 2.51066536e-01 2.55177379e-01
7.16726422e-01 -9.24046431e-03 5.22297621e-01 -1.49700031e-01
2.68095523e-01 -1.46907777e-01 4.95120846e-02 -4.73571986e-01
-4.55470160e-02 7.36607835e-02 1.49781480e-01 7.09511697e-01
-1.69486821e-01 1.31159294e+00 -3.37574571e-01 -7.61724710e-01
3.56123537e-01 4.94639128e-01 -1.76212460e-01 3.73492986e-01
-2.69618839e-01 8.64150524e-01 -3.62247616e-01 1.32218659e+00
5.61818779e-01 -8.53483558e-01 -1.66089803e-01 1.92042485e-01
3.42925578e-01 -5.31346977e-01 -7.07170069e-01 1.31079662e+00
1.34951547e-01 4.99995679e-01 1.90213859e-01 -7.80182898e-01
3.96717012e-01 1.87223077e-01 8.26901197e-01 -7.60628104e-01
6.77602172e-01 2.72991300e-01 9.03720781e-03 -4.58186328e-01
-1.22176828e-02 -7.15322375e-01 2.43617460e-01 1.92867234e-01
-1.94374859e-01 -1.60977785e-02 4.54461187e-01 2.27578104e-01
1.54549706e+00 -8.30208361e-01 1.10407218e-01 -1.20919287e-01
9.43144798e-01 2.54902810e-01 4.71152604e-01 5.08823574e-01
-6.05798006e-01 7.12314010e-01 -8.32016766e-03 -4.70956177e-01
-6.47611499e-01 -1.38214958e+00 -3.80131543e-01 6.72104299e-01
1.45627558e-01 3.67285639e-01 -5.95765471e-01 -6.03127837e-01
4.42774922e-01 5.20603836e-01 -5.97541928e-01 4.55056190e-01
-7.65593469e-01 -1.35982561e+00 4.96664584e-01 5.94493091e-01
3.04148465e-01 -1.06014144e+00 -2.71310955e-01 4.51985925e-01
-1.14021070e-01 -9.83411312e-01 6.65787160e-02 3.01146835e-01
-1.05450821e+00 -1.42017341e+00 -1.00382423e+00 -9.85266089e-01
7.77753651e-01 -5.95229082e-02 1.08477533e+00 3.98953080e-01
-9.97284770e-01 3.10420811e-01 -1.77289802e-03 -4.73486185e-01
-6.00122631e-01 -1.41705915e-01 -2.40826771e-01 -3.87034938e-02
1.14905894e+00 -2.87442893e-01 -7.58171439e-01 1.01550736e-01
-5.49046993e-01 2.21617430e-01 8.45356047e-01 6.14001274e-01
1.05491698e+00 -2.48691812e-01 4.82148319e-01 -1.09983110e+00
3.02223176e-01 -4.51609910e-01 -3.37815344e-01 1.83735326e-01
-4.89429444e-01 -4.58749354e-01 1.91237539e-01 1.54057696e-01
-7.02971160e-01 1.28131166e-01 -4.27957594e-01 -1.03450269e-01
-6.74593568e-01 1.12604514e-01 2.31066361e-01 -4.18272436e-01
3.14733744e-01 4.81981188e-01 -3.18664499e-02 -3.04222405e-01
-9.94399637e-02 7.91202724e-01 7.90547848e-01 -2.32571974e-01
2.33057201e-01 9.80603576e-01 4.24268574e-01 -6.79911554e-01
-5.35164833e-01 -1.03752470e+00 -6.20601475e-01 -5.03033817e-01
9.23731267e-01 -7.48563707e-01 -7.73603141e-01 1.00784290e+00
-8.12311709e-01 -1.85576469e-01 4.12638523e-02 2.02433273e-01
-6.20055139e-01 3.20742846e-01 -1.56149626e+00 -3.58084947e-01
-8.59155238e-01 -1.27198970e+00 1.10673785e+00 9.02217329e-01
-3.95047009e-01 -1.14510119e+00 3.98482502e-01 9.29485142e-01
5.26379764e-01 3.30375493e-01 8.79008055e-01 -1.01208127e+00
-8.27442467e-01 -5.35474718e-01 -2.39367396e-01 -3.49649131e-01
3.84061784e-02 -2.80137453e-02 -8.52609873e-01 -4.10887301e-01
-2.00283233e-04 -1.69624060e-01 1.05368698e+00 7.67553806e-01
9.66791689e-01 5.20314217e-01 -1.04821944e+00 8.43486130e-01
1.57581425e+00 5.20125508e-01 5.10237217e-01 6.35345697e-01
4.66550559e-01 2.13458434e-01 1.87405810e-01 4.65764612e-01
2.11166903e-01 2.67788649e-01 2.70379156e-01 -3.25091720e-01
-2.10736290e-01 4.99605864e-01 -3.35787237e-01 5.27177393e-01
2.91714579e-01 -3.23528320e-01 -1.56748605e+00 5.13873696e-01
-1.64761007e+00 -7.23202586e-01 -6.52292192e-01 1.49217594e+00
6.77039981e-01 5.19731194e-02 1.21807419e-01 -1.79503530e-01
9.18059707e-01 -7.17839479e-01 -7.32326150e-01 1.44879401e-01
-4.39327836e-01 4.75110650e-01 3.03703576e-01 4.54263002e-01
-1.03623819e+00 5.39437294e-01 7.03003740e+00 8.05647612e-01
-1.22631288e+00 1.17854655e-01 1.16223872e+00 -2.81691104e-01
2.18375280e-01 -3.15194964e-01 -1.12879336e+00 6.47772133e-01
6.70062721e-01 -2.65277147e-01 -3.03757697e-01 3.19366962e-01
1.06943965e-01 -6.54184103e-01 -1.28552020e+00 1.16682470e+00
-1.28113836e-01 -1.81070662e+00 -1.04566365e-01 6.58297464e-02
6.02882445e-01 1.21156774e-01 4.96998206e-02 2.96193421e-01
4.75635439e-01 -1.22921550e+00 -1.29086584e-01 7.98153818e-01
2.62145549e-01 -1.04058802e+00 1.22910893e+00 5.28636873e-01
-7.93994963e-01 2.72380620e-01 -3.83681387e-01 1.96617112e-01
2.11318940e-01 8.62253964e-01 -1.43135202e+00 1.51711598e-01
3.49591017e-01 2.43024617e-01 -8.92230272e-01 1.55418193e+00
2.42946908e-01 6.39786184e-01 -5.03527343e-01 3.94390039e-02
-1.95673198e-01 -2.01726705e-01 4.11206841e-01 1.51306236e+00
1.44697860e-01 2.48554692e-01 1.75880492e-01 7.75599599e-01
-8.57827961e-02 2.87713408e-01 1.13590330e-01 -7.22891325e-03
-1.60037670e-02 1.49478292e+00 -1.40768337e+00 -2.68376976e-01
-3.68795812e-01 7.88837671e-01 3.59365672e-01 4.96722339e-03
-6.02431953e-01 2.58712411e-01 2.69043565e-01 1.52376771e-01
-1.42370135e-01 2.37426907e-01 -8.68229508e-01 -8.79271269e-01
-7.59827852e-01 -4.07930613e-01 1.03885388e+00 -3.98524195e-01
-1.56512368e+00 3.27792317e-01 -9.06495988e-01 -6.46986425e-01
-2.65169263e-01 -5.43370306e-01 -9.41181421e-01 8.22655797e-01
-1.17275131e+00 -1.39917743e+00 -4.41620320e-01 5.40887237e-01
4.42014158e-01 -3.83134633e-01 7.89678991e-01 7.77062774e-02
-8.92639637e-01 5.60700715e-01 3.17416370e-01 1.93059623e-01
5.21051407e-01 -1.71799564e+00 -1.47025660e-01 6.09548330e-01
-1.72618926e-01 1.73618749e-01 9.25178111e-01 -6.61889970e-01
-1.35251188e+00 -9.69878137e-01 5.79584897e-01 -5.92497170e-01
6.78981900e-01 8.37626979e-02 -8.10739636e-01 6.18693590e-01
4.55287904e-01 -1.67618901e-01 1.11383760e+00 -2.80584425e-01
3.35595995e-01 2.55280763e-01 -1.49574256e+00 4.78278071e-01
1.07427262e-01 -2.15216298e-02 -4.10306543e-01 8.00919831e-01
6.48605675e-02 -7.89528310e-01 -1.09879684e+00 4.84891057e-01
5.83816767e-02 -1.12762737e+00 8.20212901e-01 -3.48695487e-01
-2.03762986e-02 -3.51442993e-01 -3.53773274e-02 -9.88668203e-01
-2.51167804e-01 -1.26665950e-01 -5.26285589e-01 1.13781142e+00
1.30317360e-01 -2.39287645e-01 1.90319097e+00 6.13178790e-01
-6.13889657e-03 -1.02466881e+00 -1.06194818e+00 -3.76110911e-01
5.85594893e-01 1.33808842e-02 2.28398606e-01 7.72893250e-01
1.42810509e-01 3.17291260e-01 4.84501034e-01 1.81532264e-01
1.05598009e+00 5.03764212e-01 7.13217497e-01 -1.25987852e+00
-1.89728498e-01 -9.39633787e-01 -6.73063278e-01 -8.16663444e-01
1.29144285e-02 -1.05462146e+00 -4.18996423e-01 -1.78613806e+00
7.80871034e-01 -2.40321711e-01 -6.11755490e-01 -2.58530527e-02
-2.94781059e-01 2.15638071e-01 -6.46225959e-02 4.42512721e-01
-8.07283998e-01 1.05454393e-01 1.07781112e+00 -3.93286794e-01
2.04200834e-01 3.49665821e-01 -3.59253764e-01 8.13913465e-01
8.46962273e-01 -3.27888161e-01 3.33170950e-01 1.86886057e-01
-3.86179537e-01 4.45437729e-01 1.88098416e-01 -1.07273781e+00
7.76312709e-01 -2.46145532e-01 8.93865585e-01 -1.34823751e+00
1.69903263e-01 -4.92600322e-01 2.64047742e-01 9.27071154e-01
-6.38122931e-02 -2.71420419e-01 -3.92372580e-03 5.26603520e-01
-2.39927709e-01 -1.05831832e-01 1.59218216e+00 -2.90323406e-01
-7.13993490e-01 5.02995968e-01 -7.08147883e-01 -1.69547513e-01
1.42957044e+00 -3.58501345e-01 -2.94689447e-01 2.96267085e-02
-1.03219211e+00 7.20459580e-01 7.60389149e-01 -1.36300594e-01
6.55491292e-01 -1.14549661e+00 -1.00564575e+00 -1.03785709e-01
1.13277137e-01 1.91349953e-01 3.42860699e-01 1.19548726e+00
-1.30290830e+00 6.89131975e-01 -1.15812272e-01 -1.20506072e+00
-1.46440160e+00 2.93153703e-01 8.34300578e-01 -8.48631024e-01
-7.98609018e-01 1.54482234e+00 2.35203698e-01 -1.86917260e-02
2.50292361e-01 -4.46681120e-02 -2.27302015e-01 -1.86653733e-01
8.35770786e-01 3.86245281e-01 2.28057012e-01 -7.53964961e-01
-3.63159120e-01 1.61599904e-01 -6.89066350e-01 4.49199468e-01
1.36305475e+00 -1.40421450e-01 -5.02474189e-01 2.33511865e-01
7.57760942e-01 -3.38500291e-01 -6.55637503e-01 -2.28357270e-01
4.79564130e-01 -3.23610961e-01 1.05090767e-01 -8.79769325e-01
-1.32155931e+00 7.08821416e-01 9.55951214e-01 3.10802869e-02
1.12427592e+00 2.94264734e-01 1.15477967e+00 -2.83105709e-02
5.49015880e-01 -8.29093516e-01 -1.22813858e-01 2.10028514e-01
3.02374363e-01 -1.35463345e+00 -1.42833859e-01 -4.37918961e-01
-1.53806046e-01 1.35923791e+00 8.93283665e-01 -1.16355941e-01
4.71388429e-01 8.78725052e-01 3.63817543e-01 -4.80005294e-01
-8.68315578e-01 -2.20752686e-01 -2.13550106e-01 7.21874714e-01
5.84506452e-01 2.37349585e-01 -2.18089342e-01 5.80678344e-01
8.06122646e-02 9.74400342e-02 4.53259885e-01 9.49538052e-01
-1.07959628e+00 -9.67563033e-01 -7.17873573e-01 8.02556753e-01
-7.85935521e-01 3.51844728e-01 -7.13515162e-01 6.95243657e-01
-1.05673917e-01 8.39043617e-01 1.65862545e-01 1.70387268e-01
2.21115202e-02 1.40612543e-01 7.00211883e-01 -2.86214799e-01
-5.55735588e-01 2.66744345e-01 -4.99880880e-01 -2.25749880e-01
-2.30043381e-01 -7.08221912e-01 -2.07144189e+00 -6.50563538e-01
-2.17669591e-01 1.79034676e-02 2.72553205e-01 1.17911124e+00
-1.55814201e-01 5.34509301e-01 4.53425318e-01 -4.30682153e-01
-1.82679459e-01 -8.83797586e-01 -1.23017967e+00 6.38957679e-01
9.95832458e-02 -4.41137850e-01 -3.01634103e-01 9.43906233e-02] | [15.046609878540039, -3.0533251762390137] |
c1c46f12-c58d-4f73-b104-4f3c98a548b0 | drawing-attention-to-detail-pose-alignment | 2302.04800 | null | https://arxiv.org/abs/2302.04800v1 | https://arxiv.org/pdf/2302.04800v1.pdf | Drawing Attention to Detail: Pose Alignment through Self-Attention for Fine-Grained Object Classification | Intra-class variations in the open world lead to various challenges in classification tasks. To overcome these challenges, fine-grained classification was introduced, and many approaches were proposed. Some rely on locating and using distinguishable local parts within images to achieve invariance to viewpoint changes, intra-class differences, and local part deformations. Our approach, which is inspired by P2P-Net, offers an end-to-end trainable attention-based parts alignment module, where we replace the graph-matching component used in it with a self-attention mechanism. The attention module is able to learn the optimal arrangement of parts while attending to each other, before contributing to the global loss. | ['Jameel Hassan', 'Mohamed El Amine Boudjoghra', 'Salwa Al Khatib'] | 2023-02-09 | null | null | null | null | ['graph-matching'] | ['graphs'] | [ 4.17718776e-02 2.42134154e-01 -1.97376683e-02 -3.88600349e-01
-4.30157334e-01 -3.94108891e-01 4.58927304e-01 3.29447836e-02
-2.28275284e-01 2.95512468e-01 3.00220251e-01 5.47029912e-01
-1.31313160e-01 -7.73494959e-01 -8.71822715e-01 -5.40676773e-01
1.82708576e-01 3.28387439e-01 6.61412418e-01 -2.41826072e-01
3.36110026e-01 7.57699966e-01 -1.45077360e+00 3.34537297e-01
8.39967310e-01 1.11153769e+00 1.91303864e-01 2.50559807e-01
-3.62757176e-01 4.76749182e-01 -5.27230740e-01 -6.58517957e-01
6.43638432e-01 -2.49001935e-01 -8.99565876e-01 3.93799931e-01
7.67343581e-01 8.73055309e-02 -5.68508983e-01 1.19491756e+00
4.94662881e-01 1.89308837e-01 5.44908941e-01 -1.23944390e+00
-8.43385637e-01 4.20541883e-01 -6.23428404e-01 1.91444024e-01
3.14966530e-01 3.23571377e-02 1.03831375e+00 -6.29485965e-01
4.59597528e-01 1.39789689e+00 6.45908415e-01 5.71697176e-01
-1.12703574e+00 -4.12448198e-01 4.42696810e-01 4.57370967e-01
-1.42250264e+00 -1.51466519e-01 1.23606122e+00 -3.37528974e-01
9.60971713e-01 1.39840230e-01 7.74873614e-01 8.84618580e-01
4.78480875e-01 7.64817834e-01 7.77370393e-01 -4.58499938e-01
-5.02118170e-02 -1.49809629e-01 4.65353578e-02 6.05663002e-01
9.81767774e-02 -1.56665847e-01 -2.33270153e-01 3.35200354e-02
7.99885213e-01 3.89363945e-01 -4.59404975e-01 -8.64156723e-01
-1.02229178e+00 4.87658203e-01 1.20869815e+00 3.71033579e-01
-3.95039201e-01 1.14882633e-01 2.05509096e-01 1.80425659e-01
4.12850648e-01 3.60893548e-01 -7.17459798e-01 3.38337302e-01
-5.54845810e-01 3.64014655e-02 4.92978722e-01 8.39978755e-01
8.84058356e-01 -4.51785833e-01 -4.57376331e-01 9.38278973e-01
4.09517944e-01 1.36326682e-02 7.74812281e-01 -5.74293017e-01
5.20402372e-01 1.00811315e+00 -3.13172847e-01 -1.15059316e+00
-4.40957189e-01 -5.52918851e-01 -8.89679551e-01 1.75648078e-01
1.23596832e-01 3.38006467e-01 -1.25782764e+00 1.70420921e+00
5.37206411e-01 2.48913877e-02 -4.16412532e-01 7.67904639e-01
6.97121263e-01 1.39852136e-01 -4.12586667e-02 3.39337498e-01
1.28499901e+00 -1.45771027e+00 -3.60466361e-01 -2.61512578e-01
8.46831724e-02 -7.99862564e-01 8.89130592e-01 -3.52547094e-02
-9.58603680e-01 -9.07391727e-01 -9.64856863e-01 -2.92566121e-01
-5.90956092e-01 -3.72993439e-01 7.26521462e-02 3.57005447e-01
-1.06815588e+00 7.68878460e-01 -6.59629226e-01 -3.68375391e-01
8.82938921e-01 4.85700220e-01 -7.24661529e-01 -1.38500273e-01
-6.60581172e-01 7.83798099e-01 3.43794167e-01 3.38092893e-01
-4.91723269e-01 -3.41379642e-01 -9.53498125e-01 2.44802549e-01
2.49565065e-01 -9.13984001e-01 7.92293549e-01 -1.53182995e+00
-1.57435453e+00 9.05719280e-01 1.62777528e-01 -2.19438344e-01
5.96571803e-01 -1.56124398e-01 -1.37713104e-01 -5.36939688e-02
1.05841711e-01 7.72288263e-01 9.25597906e-01 -1.15969884e+00
-2.61621803e-01 -7.21200109e-01 1.09454356e-01 3.23000371e-01
-2.19128922e-01 2.30561104e-02 -7.03220546e-01 -8.41233790e-01
4.19286788e-01 -9.64701653e-01 -3.45521748e-01 2.51336247e-01
-3.93453330e-01 -2.47499779e-01 8.80594611e-01 -6.01064742e-01
5.57519734e-01 -2.03283906e+00 3.34481061e-01 1.82822436e-01
3.18340987e-01 3.31454813e-01 -5.59111178e-01 3.43560338e-01
-6.33943081e-02 -1.03545912e-01 -2.25339755e-01 -4.23727393e-01
-7.32139498e-02 1.33215576e-01 2.91347485e-02 6.06315196e-01
5.08118808e-01 1.15580201e+00 -6.25512481e-01 -3.80785197e-01
3.02931130e-01 6.38195693e-01 -4.56760317e-01 3.20284784e-01
-1.29373550e-01 3.86496037e-01 -3.92570108e-01 6.20728612e-01
9.49762583e-01 -1.27012268e-01 -1.59828752e-01 -2.71226674e-01
2.38904461e-01 1.35587215e-01 -1.10410035e+00 1.87022460e+00
-3.11369240e-01 3.43184501e-01 1.61952153e-01 -1.13797402e+00
8.50676417e-01 4.11555916e-02 3.38917017e-01 -8.09018910e-01
2.95471072e-01 -3.94729190e-02 1.32772937e-01 -3.25073451e-01
1.36805445e-01 1.64333239e-01 1.03478439e-01 -6.16389699e-02
2.59763211e-01 4.79513332e-02 -9.94887948e-02 -1.05708502e-02
9.95169222e-01 3.83212507e-01 3.22351366e-01 -2.25737348e-01
7.24378586e-01 -4.14018840e-01 7.59990096e-01 2.64740497e-01
-3.47826451e-01 1.09646976e+00 1.78453401e-01 -9.05431092e-01
-6.32192791e-01 -7.67922521e-01 -4.11398187e-02 8.86695325e-01
6.00987196e-01 -2.10637808e-01 -7.59640276e-01 -1.09749603e+00
1.06160678e-02 -9.28056091e-02 -8.68765950e-01 -2.31843784e-01
-6.45514190e-01 -3.20293903e-01 -1.15529206e-02 7.40996063e-01
8.14695477e-01 -1.22295272e+00 -4.45714146e-01 4.07554477e-01
-2.49554738e-01 -1.15303183e+00 -8.17914724e-01 2.88903773e-01
-8.25450480e-01 -1.37371850e+00 -9.41520751e-01 -8.81593823e-01
9.44719613e-01 4.41958636e-01 1.23549438e+00 2.38783598e-01
-4.65025634e-01 2.27517977e-01 -4.64027494e-01 -4.12085027e-01
5.89087531e-02 8.96826014e-02 -3.53162199e-01 3.80105406e-01
1.01122633e-01 -8.25917900e-01 -8.19804847e-01 6.42028928e-01
-6.97926879e-01 -1.93159491e-01 8.76855314e-01 7.47524500e-01
9.78575408e-01 -2.71889836e-01 3.37755904e-02 -6.63883805e-01
1.73924863e-01 -6.20956756e-02 -4.96914119e-01 4.99083757e-01
-2.33665660e-01 1.83106184e-01 7.18414724e-01 -1.81124896e-01
-5.78391790e-01 2.44584799e-01 -2.88918763e-01 -6.17109954e-01
-2.38825470e-01 -1.05972581e-01 -7.54815280e-01 -6.21243358e-01
2.97778010e-01 2.98677206e-01 -1.03516258e-01 -6.28644705e-01
3.35100979e-01 4.09637153e-01 3.92482162e-01 -2.91550994e-01
8.63370717e-01 4.16611999e-01 -3.73964459e-02 -6.02639854e-01
-8.94586802e-01 -6.18577600e-01 -1.21538734e+00 8.53294283e-02
9.98305857e-01 -9.04842734e-01 -2.64232069e-01 8.51347029e-01
-1.16637897e+00 -3.69093627e-01 -5.76491892e-01 1.18044361e-01
-5.30232191e-01 5.35019159e-01 -3.28179419e-01 -2.45249391e-01
-5.71622074e-01 -1.07405066e+00 1.25966537e+00 3.85425925e-01
2.79928952e-01 -8.47268045e-01 3.16777110e-01 3.82852048e-01
4.07186002e-01 2.68819660e-01 6.74038768e-01 -6.03574216e-01
-7.39908814e-01 -3.46885890e-01 -3.80612433e-01 5.35320103e-01
3.56366545e-01 -1.11520268e-01 -8.82435620e-01 -3.95835072e-01
-1.64385751e-01 -1.87588990e-01 7.81416714e-01 2.72427648e-01
1.55318284e+00 -4.69701767e-01 -4.95385051e-01 8.94479573e-01
1.38571501e+00 -1.91088885e-01 8.42799604e-01 3.06631982e-01
1.17964113e+00 5.80302179e-01 2.35087708e-01 -9.64100100e-03
5.57015300e-01 1.05043471e+00 7.46985316e-01 -2.23512635e-01
-4.97649431e-01 -1.31128430e-01 9.84648541e-02 5.16929030e-01
3.98885757e-02 -2.70409256e-01 -5.08139193e-01 4.19128239e-01
-1.76307821e+00 -8.82932901e-01 1.33644655e-01 2.08484817e+00
3.70984375e-01 2.03179136e-01 7.63701424e-02 -3.84480394e-02
7.45896935e-01 3.14358860e-01 -4.20490801e-01 -2.67986000e-01
-8.35809112e-02 1.41074002e-01 3.93096715e-01 2.83355504e-01
-1.13644850e+00 9.08857524e-01 6.12710381e+00 7.15996861e-01
-1.20825613e+00 -1.98640898e-02 6.15134597e-01 3.15937728e-01
-1.84624612e-01 -1.14472441e-01 -5.54632843e-01 5.03972471e-01
2.32831717e-01 3.04873556e-01 2.47371763e-01 7.62560904e-01
-2.77917027e-01 2.94271767e-01 -9.59011018e-01 8.80264878e-01
3.91701579e-01 -9.20969188e-01 1.77033752e-01 1.18957214e-01
8.35330188e-01 2.38548443e-01 -3.04763824e-01 2.22307876e-01
4.42474941e-03 -8.27418983e-01 7.18171239e-01 4.55533624e-01
2.57389277e-01 -7.38507986e-01 8.80877733e-01 2.72832930e-01
-1.50107920e+00 -6.74166828e-02 -5.95861912e-01 -1.79486908e-02
-1.04621440e-01 3.47651809e-01 -4.02958721e-01 7.81050503e-01
7.33279049e-01 6.27673388e-01 -7.71234989e-01 1.34800506e+00
-2.46928424e-01 -7.07502440e-02 -9.53642428e-02 2.65929848e-01
1.15378201e-01 -1.77270994e-01 3.96453470e-01 9.33241904e-01
4.97071892e-02 -1.46596387e-01 5.32156467e-01 6.21468365e-01
-3.62968802e-01 1.53677091e-01 -3.08617294e-01 4.76843506e-01
-5.49546592e-02 1.47938776e+00 -9.47430491e-01 -1.16927207e-01
-5.23568034e-01 1.56688356e+00 6.73079312e-01 -3.50753218e-02
-6.76637053e-01 -5.46715617e-01 8.17228675e-01 2.70912498e-01
6.97601080e-01 -6.48138300e-03 -1.04341775e-01 -1.15587068e+00
3.45218331e-01 -8.03863108e-01 4.82277632e-01 -5.86324096e-01
-1.44720006e+00 8.55923951e-01 -3.06532025e-01 -1.26794422e+00
3.19900453e-01 -6.24363840e-01 -8.88432443e-01 8.29210103e-01
-1.75897598e+00 -1.44836068e+00 -6.86516762e-01 7.69237339e-01
6.68365717e-01 9.11482200e-02 6.27936482e-01 2.50918955e-01
-6.16056681e-01 9.20781136e-01 -1.91044524e-01 4.42582101e-01
5.97541332e-01 -1.25800323e+00 7.08619595e-01 9.35059488e-01
5.26891530e-01 2.30212167e-01 3.05919856e-01 -3.48159015e-01
-1.09070420e+00 -1.28656507e+00 8.14038217e-01 -4.29514319e-01
2.59032130e-01 -3.88588339e-01 -1.06151247e+00 6.17222130e-01
2.51366496e-01 8.18526924e-01 3.85561764e-01 -1.50958270e-01
-7.28218377e-01 -3.61686319e-01 -1.28778815e+00 3.19332361e-01
1.31898153e+00 -4.35633838e-01 -6.83450878e-01 1.98220730e-01
6.19468391e-01 -5.30147314e-01 -6.31312668e-01 4.55653906e-01
5.06561697e-01 -9.39067960e-01 9.84759152e-01 -5.32748520e-01
2.60043293e-01 -3.23462456e-01 6.97293924e-03 -1.42483902e+00
-8.69571626e-01 -4.29730952e-01 -8.07978306e-03 1.16582012e+00
7.05123171e-02 -6.60434484e-01 7.86520183e-01 5.92099011e-01
-3.65482330e-01 -8.31994474e-01 -1.05957103e+00 -7.44382143e-01
-4.40859683e-02 1.41207531e-01 7.21758783e-01 5.49990356e-01
-3.94805312e-01 1.75349966e-01 -8.71900916e-02 2.28460550e-01
4.11024541e-01 3.76910716e-01 8.59102070e-01 -1.40320420e+00
-4.89030004e-01 -7.34755814e-01 -1.07207501e+00 -1.16381693e+00
1.32611126e-01 -9.16291356e-01 1.73877388e-01 -1.62662506e+00
3.92796844e-01 -2.46632621e-01 -6.25642955e-01 5.38167238e-01
-3.72300923e-01 6.34401739e-01 3.49453032e-01 2.09323779e-01
-6.82479739e-01 8.68333459e-01 1.46376634e+00 -4.23016369e-01
-1.17221273e-01 1.22596167e-01 -7.15670943e-01 6.75774872e-01
5.52203774e-01 -4.62118715e-01 -5.94874956e-02 -6.34670258e-01
-3.43938440e-01 -4.58328873e-01 3.83567989e-01 -1.32008994e+00
4.97173741e-02 3.57796922e-02 6.22657359e-01 -5.04114866e-01
2.95733333e-01 -1.13268554e+00 -6.81396574e-02 5.34794688e-01
-9.90089104e-02 2.93009847e-01 1.08734615e-01 7.68919051e-01
-4.50983077e-01 2.68042441e-02 9.76280928e-01 -2.58201420e-01
-7.10421503e-01 7.90431917e-01 1.79745331e-01 7.06900954e-02
1.16717076e+00 -3.98712367e-01 -9.70356390e-02 9.29177105e-02
-5.59173405e-01 1.80599183e-01 6.31164670e-01 5.88541508e-01
4.32613015e-01 -1.36714625e+00 -6.39065266e-01 3.38997900e-01
3.49461555e-01 1.99641451e-01 4.10525233e-01 6.96674228e-01
-3.99767548e-01 1.89669952e-01 -6.01992667e-01 -5.34461856e-01
-1.32329953e+00 7.74503291e-01 7.49453664e-01 -3.65353644e-01
-6.01420403e-01 1.24455202e+00 3.61007810e-01 -4.72503215e-01
3.23696434e-01 -3.67536932e-01 -2.48877391e-01 -6.26447648e-02
3.10379416e-01 1.50180697e-01 2.84971595e-01 -8.66130054e-01
-5.14467061e-01 1.20328045e+00 -9.87649187e-02 3.69006664e-01
1.24931991e+00 -1.53854504e-01 -8.94802660e-02 2.41610687e-03
1.44185722e+00 5.34918010e-02 -1.46847761e+00 -3.78418177e-01
-2.56170958e-01 -6.25028074e-01 -7.95415789e-02 -8.09249341e-01
-1.45236468e+00 8.40215504e-01 6.78729296e-01 1.87959433e-01
1.16749525e+00 -6.64616153e-02 8.79202724e-01 1.11463949e-01
5.01914501e-01 -8.94159019e-01 2.44675010e-01 4.44861412e-01
1.09632683e+00 -1.36654568e+00 -1.77582249e-01 -5.89869797e-01
-3.53496671e-01 1.09014690e+00 9.52299893e-01 -4.38399911e-01
8.35647404e-01 -1.90536141e-01 1.50526479e-01 -2.08912969e-01
-3.10202360e-01 -3.56041163e-01 7.87299693e-01 6.82529926e-01
2.54906535e-01 -1.12450123e-01 -1.72052279e-01 2.36653179e-01
1.15145378e-01 -3.70183676e-01 -6.70414791e-02 8.13215017e-01
-2.44704500e-01 -1.41519248e+00 -2.46160403e-01 1.94268510e-01
-4.40068990e-01 2.51630157e-01 -6.42059684e-01 6.28112078e-01
3.69706929e-01 5.09875238e-01 3.01546514e-01 -4.31962609e-01
6.73553705e-01 -1.54580548e-02 6.72894061e-01 -6.69052362e-01
-7.08243012e-01 -6.92621469e-02 -5.01278102e-01 -9.66047943e-01
-4.32242632e-01 -4.32659954e-01 -8.41270208e-01 1.18823074e-01
-4.52132970e-01 -1.41042560e-01 3.58769864e-01 9.10835683e-01
6.85844004e-01 6.42504334e-01 9.54570234e-01 -1.27307296e+00
-4.99633968e-01 -8.75364423e-01 -4.02062058e-01 7.56711066e-01
3.01012009e-01 -5.48957884e-01 -1.16380543e-01 -3.55086923e-01] | [9.592514991760254, 1.9480019807815552] |
f711c082-d8e4-4699-8c5c-0fdec0134654 | intrusion-detection-in-iot-using-artificial | null | null | https://jwcn-eurasipjournals.springeropen.com/articles/10.1186/s13638-021-01893-8 | https://jwcn-eurasipjournals.springeropen.com/articles/10.1186/s13638-021-01893-8 | Intrusion detection in IoT using artificial neural networks on UNSW-15 dataset | Internet of Things (IoT) devices are well-connected; they generate and consume data which involves transmission of data back and forth among various devices. Ensuring security of the data is a critical challenge as far as IoT is concerned. Since IoT devices are inherently low-power and do not require a lot of compute power, a Network Intrusion Detection System is typically employed to detect and remove malicious packets from entering the network. In the same context, we propose feature clusters in terms of Flow, Message Queuing Telemetry Transport (MQTT) and Transmission Control Protocol (TCP) by using features in UNSW-NB15 data-set. We eliminate problems like over-fitting, curse of dimensionality and imbalance in the data-set. We apply supervised Machine Learning (ML) algorithms, i.e., Random Forest (RF), Support Vector Machine and Artificial Neural Networks on the clusters. Using RF, we, respectively, achieve 98.67% and 97.37% of accuracy in binary and multi-class classification. In clusters based techniques, we achieved 96.96%, 91.4% and 97.54% of classification accuracy by using RF on Flow & MQTT features, TCP features and top features from both clusters. Moreover, we show that the proposed feature clusters provide higher accuracy and requires lesser training time as compared to other state-of-the-art supervised ML-based approaches. | ['Syed Ali Haider & Muhammad Safeer Khan', 'Hasan Tahir', 'Muhammad Zeeshan', 'Qaiser Riaz', 'Muhammad Ahmad'] | 2021-01-21 | null | null | null | eurasip-journal-on-wireless-communications | ['network-intrusion-detection'] | ['miscellaneous'] | [ 1.06315300e-01 -3.50979984e-01 -3.61954659e-01 -5.61877251e-01
-1.21845543e-01 -5.33070207e-01 2.57462472e-01 3.07201415e-01
-4.63883251e-01 7.76873708e-01 -4.35803354e-01 -6.77932560e-01
-6.40930951e-01 -1.13269150e+00 -1.28958583e-01 -8.40552688e-01
-2.08067037e-02 4.94947225e-01 4.65850890e-01 8.56433958e-02
2.78827667e-01 6.68626547e-01 -1.56928670e+00 3.05626959e-01
6.21426582e-01 1.62590361e+00 -5.95403492e-01 6.89243793e-01
-2.12020472e-01 5.79544783e-01 -7.75324523e-01 -3.44574690e-01
3.05765599e-01 1.05184257e-01 -7.26373911e-01 -2.69344002e-01
-4.08583432e-01 8.23002085e-02 -1.01225995e-01 8.06519449e-01
3.40720147e-01 -1.85190201e-01 6.89637899e-01 -1.97958255e+00
-9.58113968e-02 6.04684412e-01 -5.63268304e-01 2.44977221e-01
2.42554069e-01 -2.54287481e-01 4.82542932e-01 -1.45342290e-01
1.38804033e-01 9.07708824e-01 6.77819967e-01 4.29590493e-01
-1.03529215e+00 -1.16999590e+00 -3.28008294e-01 4.25137132e-01
-1.34419382e+00 -2.72134632e-01 6.35884345e-01 -2.27841705e-01
9.22803223e-01 7.09693670e-01 3.83986473e-01 7.66384244e-01
5.09629011e-01 1.57753766e-01 9.33004081e-01 -2.33494356e-01
4.12758857e-01 4.13557678e-01 4.35879380e-01 3.21509391e-01
5.93090355e-01 -9.57617238e-02 1.32936209e-01 -4.04955417e-01
7.42143393e-02 5.42999029e-01 2.12854773e-01 1.61070004e-01
-1.00384629e+00 8.67292643e-01 2.46996716e-01 4.73399401e-01
-4.34841692e-01 -1.80737749e-01 5.01216948e-01 4.15053576e-01
-6.29410669e-02 -2.07590863e-01 -1.02917480e+00 -3.07369493e-02
-6.99593246e-01 -5.25561690e-01 1.07393527e+00 1.11863422e+00
6.99493885e-01 -8.91973302e-02 3.80493015e-01 3.68101746e-01
4.89945114e-01 8.23787272e-01 5.94202578e-01 -5.29992521e-01
4.34051037e-01 8.78930509e-01 -2.26010218e-01 -8.80658984e-01
-9.28524792e-01 -1.32805258e-01 -1.40191221e+00 -7.83811230e-03
-9.21862051e-02 -8.74094442e-02 -6.73494995e-01 1.13603747e+00
5.59140384e-01 2.97995478e-01 2.85564005e-01 8.54828358e-02
6.52404904e-01 6.09585345e-01 3.24021995e-01 -5.22244573e-01
1.33702254e+00 -3.56014818e-01 -6.02633715e-01 5.11109054e-01
5.65767527e-01 -9.36928928e-01 3.92233044e-01 5.36912739e-01
-5.78824937e-01 -6.22879863e-01 -9.04483199e-01 5.82166433e-01
-7.10310519e-01 -3.39647979e-01 7.41412938e-01 1.28368175e+00
-6.09904051e-01 5.34002960e-01 -7.16084957e-01 -5.04006565e-01
6.87042058e-01 1.11195397e+00 -1.93922445e-01 6.39102086e-02
-7.62745559e-01 2.65879661e-01 4.25146967e-01 -3.20986956e-01
-7.56365880e-02 -4.60056543e-01 -4.04166013e-01 -1.63672194e-01
-1.19475082e-01 -5.35756707e-01 6.67112231e-01 -4.36663985e-01
-1.26243925e+00 2.18705162e-01 2.82821786e-02 -3.75511944e-01
7.04015270e-02 2.68589795e-01 -1.15959990e+00 1.86892405e-01
-1.24412872e-01 3.22228760e-01 8.41235578e-01 -8.91634703e-01
-1.04226160e+00 -4.73646134e-01 -4.10726309e-01 -8.06720316e-01
-4.48581308e-01 8.31161067e-02 2.98235297e-01 -4.01646309e-02
3.79795790e-01 -1.04961443e+00 -1.23100035e-01 -3.92439187e-01
-7.39334822e-01 -5.80774069e-01 1.59067297e+00 -1.94544762e-01
1.10921299e+00 -2.02596426e+00 -7.24513054e-01 7.92131603e-01
1.90041050e-01 6.05439067e-01 3.22641104e-01 3.09834272e-01
9.25178900e-02 6.17093563e-01 1.45604134e-01 -2.69082114e-02
-5.20036696e-03 3.77451450e-01 3.76001894e-02 5.27127862e-01
-2.43559983e-02 5.06702662e-01 -5.15150428e-01 -5.87925375e-01
8.36083829e-01 5.93823195e-01 -3.69712144e-01 -9.34648961e-02
2.51334727e-01 6.43900037e-01 -8.04490566e-01 7.13802457e-01
9.75144684e-01 -2.15854675e-01 3.80750671e-02 -6.14791393e-01
-1.76703736e-01 2.71540821e-01 -1.21725821e+00 9.03592885e-01
-4.88042504e-01 1.59718454e-01 -3.08799237e-01 -1.19565606e+00
1.11902893e+00 4.91606116e-01 1.01402175e+00 -8.03074360e-01
6.96224689e-01 2.56920159e-01 8.20112899e-02 -7.35348582e-01
-1.70958415e-01 1.07317768e-01 -2.23956719e-01 6.42814279e-01
-2.29427993e-01 3.80409986e-01 1.26044616e-01 -1.41522586e-02
1.42920852e+00 -6.61746979e-01 3.63976061e-01 -2.42174909e-01
8.47396374e-01 -1.99916855e-01 4.02334034e-01 6.62132204e-01
-3.63126665e-01 1.30245816e-02 6.43006116e-02 -5.80617785e-01
-8.67152572e-01 -9.40555632e-01 -4.04923469e-01 6.98064387e-01
4.21816669e-02 -1.73003197e-01 -3.64127815e-01 -8.49097133e-01
9.84727070e-02 5.12507141e-01 -2.33587682e-01 -1.69866845e-01
-5.21541655e-01 -9.82499838e-01 5.63141227e-01 1.69086814e-01
6.41767442e-01 -1.00483608e+00 -6.32828951e-01 2.57008910e-01
1.87869668e-02 -1.50442815e+00 3.69116992e-01 2.59694338e-01
-9.59302902e-01 -1.38529599e+00 4.77500558e-01 -6.74649298e-01
4.35652703e-01 2.72536278e-01 9.53714013e-01 2.22180665e-01
-4.07500297e-01 1.35637343e-01 -5.69778740e-01 -5.37748635e-01
-3.72776836e-01 1.49168581e-01 3.04405242e-01 1.89222082e-01
7.59655118e-01 -1.04819107e+00 -6.26821220e-01 6.87557936e-01
-7.93948650e-01 -5.68986297e-01 5.84858596e-01 2.40488708e-01
3.71256143e-01 7.62430131e-01 8.58343363e-01 -6.93128407e-01
6.68880045e-02 -9.14778233e-01 -5.68465173e-01 -6.00519739e-02
-6.53686762e-01 -1.85598820e-01 1.02933824e+00 -4.88432318e-01
-4.48868215e-01 1.15456574e-01 -3.28946829e-01 -1.60133820e-02
-5.78574479e-01 -4.07420009e-01 -2.82630742e-01 -2.65055478e-01
2.93634415e-01 -7.05605596e-02 -3.87271792e-01 -4.11347955e-01
-1.99202150e-01 1.50959158e+00 3.94595563e-02 -3.88637334e-02
9.99623656e-01 8.48277390e-01 2.92970002e-01 -9.65148270e-01
-5.14657617e-01 -6.19417667e-01 -6.46255851e-01 9.57581773e-02
8.15402746e-01 -4.78971779e-01 -1.52221620e+00 2.46133968e-01
-8.30608189e-01 3.62355351e-01 -9.51271206e-02 6.39258564e-01
-3.81427933e-03 2.07752645e-01 -3.68834585e-01 -1.08800983e+00
-8.00609589e-01 -1.00123966e+00 7.42627859e-01 2.70806432e-01
-3.01622033e-01 -7.73281455e-01 -4.76246327e-01 3.54023635e-01
6.13073885e-01 3.90959233e-01 1.02146578e+00 -1.12135446e+00
-3.30017686e-01 -4.89706397e-01 -3.92648071e-01 2.75703669e-01
4.87875313e-01 1.47990376e-01 -9.99640048e-01 -2.80976146e-01
2.72564411e-01 1.53685093e-01 3.43006641e-01 1.71965659e-01
1.26222038e+00 -4.56494927e-01 -5.88890791e-01 5.99971354e-01
1.47120094e+00 7.46590316e-01 7.21253753e-01 2.41251618e-01
5.91090381e-01 5.22506952e-01 3.36537182e-01 7.38149881e-01
3.72724980e-01 3.82294953e-01 5.15021086e-01 9.00713354e-02
3.23160648e-01 2.07424775e-01 -2.30468102e-02 7.47792721e-01
1.47135034e-01 -3.42879832e-01 -6.72709703e-01 1.82830598e-02
-1.62214100e+00 -9.45387661e-01 -7.20385790e-01 2.20416331e+00
1.08889915e-01 4.58917469e-01 4.68057692e-01 1.28006494e+00
8.65481853e-01 -5.30322313e-01 -4.26736087e-01 -3.94187957e-01
2.39621982e-01 4.69947129e-01 9.39959407e-01 1.66330244e-02
-1.20160723e+00 2.85761654e-01 4.83697510e+00 6.97719932e-01
-1.21075773e+00 2.43851870e-01 5.44336617e-01 2.79756486e-01
1.85822204e-01 -3.44491214e-01 -7.91608095e-01 8.82493496e-01
1.47638202e+00 3.26948196e-01 3.66048545e-01 6.94885790e-01
4.51522827e-01 3.01492233e-02 -7.69039214e-01 1.18081355e+00
-4.18804735e-01 -8.74969780e-01 -2.68310487e-01 1.59183085e-01
3.78211379e-01 1.54905766e-01 -2.90168196e-01 9.05027762e-02
2.65764207e-01 -7.58391976e-01 2.53210634e-01 1.45286024e-01
5.51250458e-01 -9.79330122e-01 1.06961095e+00 2.18324810e-01
-1.35783052e+00 -5.51557899e-01 -3.46396416e-01 -2.16408423e-03
-4.90293419e-03 9.97413099e-01 -6.28072977e-01 6.38340592e-01
1.06826818e+00 4.37750638e-01 -3.28389615e-01 9.43335116e-01
1.93184868e-01 7.41197824e-01 -7.87391603e-01 -4.33535516e-01
-1.63549662e-01 4.69959453e-02 1.70394659e-01 1.04617453e+00
2.49974549e-01 1.31470472e-01 2.47295961e-01 1.82474867e-01
1.51398480e-01 5.24228700e-02 -3.14129144e-01 3.46646160e-01
8.01833391e-01 1.48483896e+00 -1.02837741e+00 -3.96161050e-01
-4.64166760e-01 5.44510961e-01 -4.63674903e-01 -4.99313660e-02
-8.46183956e-01 -6.99919879e-01 6.49300933e-01 6.77709430e-02
3.89697522e-01 -5.57077006e-02 -5.56109786e-01 -8.10410798e-01
-1.64019793e-01 -3.63622516e-01 4.93475646e-01 -2.38802806e-01
-1.45033634e+00 7.85374403e-01 -2.93438107e-01 -1.45877957e+00
7.86024239e-03 -5.14631271e-01 -5.57308018e-01 2.07341656e-01
-1.56204724e+00 -9.53090727e-01 -5.20013034e-01 1.24834585e+00
1.00479454e-01 -2.67581511e-02 9.25238609e-01 7.85450101e-01
-6.07184231e-01 5.43335915e-01 1.69520855e-01 9.91654918e-02
1.10191390e-01 -6.75254345e-01 1.36882693e-01 3.03748727e-01
-1.83887869e-01 4.66209054e-01 2.94524401e-01 -4.28564727e-01
-1.51290822e+00 -1.31788182e+00 8.08050275e-01 -1.56004742e-01
6.55556858e-01 -4.81213629e-01 -3.74158621e-01 4.53967839e-01
-2.08171725e-01 3.15456241e-01 1.05646956e+00 -3.34010988e-01
-2.27783397e-01 -6.30831718e-01 -2.03100896e+00 -4.30575944e-02
7.07930863e-01 -2.27777883e-02 3.98019701e-03 2.57507950e-01
6.17535949e-01 4.70754355e-01 -1.21302283e+00 4.26902831e-01
6.78250253e-01 -1.15409803e+00 1.04195356e+00 -4.49171573e-01
-5.59375048e-01 -3.76023680e-01 -4.33291525e-01 -5.07975638e-01
-1.94697246e-01 -7.77982473e-01 -6.59047216e-02 1.69070566e+00
4.15463567e-01 -1.06882739e+00 8.05925846e-01 4.46256429e-01
2.25417867e-01 -3.67712945e-01 -1.05747807e+00 -7.60363221e-01
-3.33917797e-01 -7.74819434e-01 1.01763427e+00 9.20117676e-01
-8.80222544e-02 5.92989028e-01 -1.46918371e-01 3.20746154e-01
8.62501562e-01 -7.06199184e-02 6.39995277e-01 -1.99103451e+00
2.25100949e-01 -1.87856518e-02 -8.76942277e-01 -1.63389742e-01
-1.93593144e-01 -6.11639857e-01 -5.61188936e-01 -1.34893179e+00
-1.77053154e-01 -1.18055296e+00 -6.36071324e-01 6.22705877e-01
5.78820944e-01 6.15032911e-01 3.18203457e-02 3.06443810e-01
-6.77776754e-01 -7.76912197e-02 5.80295742e-01 -7.93636031e-03
-2.68617064e-01 7.80155599e-01 -4.97672677e-01 6.93833411e-01
1.67497540e+00 -9.28505182e-01 -3.08918536e-01 1.10284559e-01
-5.88078648e-02 -2.12791979e-01 1.71459049e-01 -1.25839877e+00
1.60962731e-01 -1.37680443e-02 6.23694301e-01 -7.24161625e-01
2.23495990e-01 -1.62365961e+00 2.84516633e-01 9.72609997e-01
3.23634326e-01 2.75455654e-01 -1.14699483e-01 4.13344413e-01
3.68757427e-01 -4.15416472e-02 8.41028631e-01 1.71432689e-01
-1.73022822e-01 3.64403814e-01 -4.85446513e-01 -3.51713777e-01
1.28190470e+00 -3.61689806e-01 -4.95011836e-01 -1.01448007e-01
-4.78010416e-01 -1.09103814e-01 1.36695728e-01 2.88159013e-01
3.87965560e-01 -1.11868143e+00 -2.72503912e-01 5.30610681e-01
-1.22824609e-01 -2.91644633e-01 4.73714024e-02 9.80757833e-01
-2.44407028e-01 5.17440557e-01 -1.44564465e-01 -7.67700136e-01
-1.40762234e+00 7.61200130e-01 -2.81509787e-01 -2.06889629e-01
-3.04168850e-01 3.13994288e-01 -7.76369095e-01 -4.34380263e-01
1.92000106e-01 -2.90312201e-01 -3.35582554e-01 -8.23407546e-02
6.17398381e-01 1.01750028e+00 4.46964532e-01 -6.43452466e-01
-7.81095684e-01 9.19213474e-01 1.96745902e-01 4.26783592e-01
1.35122967e+00 -2.98445493e-01 -3.23655963e-01 1.43481058e-03
1.64325917e+00 -4.11387868e-02 -3.28618228e-01 2.01528162e-01
2.16688022e-01 -2.67523408e-01 -1.42588556e-01 -6.17138207e-01
-1.20230436e+00 7.23599732e-01 1.02600944e+00 1.17779386e+00
1.26692724e+00 -4.68309000e-02 1.12923348e+00 4.16695356e-01
6.22554302e-01 -7.31127501e-01 -3.38098019e-01 1.31896198e-01
-3.00264865e-01 -1.21375358e+00 -9.17370021e-02 -7.79631078e-01
-2.04021577e-02 1.24375808e+00 2.40385130e-01 -1.02013700e-01
1.27370250e+00 5.99977195e-01 -7.08711818e-02 3.38036381e-02
-5.29116631e-01 -8.71398449e-02 -2.94309914e-01 9.98171985e-01
-1.09220140e-01 2.40036771e-01 -1.86242148e-01 4.55934495e-01
-5.45325518e-01 7.68314004e-02 4.08293992e-01 1.19709826e+00
-6.12802148e-01 -1.25674367e+00 -3.88723910e-01 8.62807631e-01
-8.73272061e-01 1.04321763e-01 8.41649398e-02 7.00703204e-01
4.86003757e-01 1.86319494e+00 2.55742192e-01 -1.05590796e+00
1.73000202e-01 -1.82852328e-01 1.00273170e-01 -9.00216401e-02
-6.59709930e-01 -2.35697076e-01 -1.04077868e-01 -5.13084769e-01
-5.43239176e-01 -3.93083781e-01 -1.49069381e+00 -8.67487311e-01
-4.57743943e-01 3.62326413e-01 1.21151900e+00 1.09531593e+00
4.76485819e-01 5.48189759e-01 1.20429206e+00 -4.68314558e-01
-1.38561174e-01 -8.42327178e-01 -4.85943675e-01 3.37834775e-01
3.07893455e-01 -4.90398943e-01 -6.19996905e-01 -7.78387338e-02] | [5.185660362243652, 7.146861553192139] |
bcb59317-d8a5-44d8-89fd-6ad1ec549b22 | weighted-automata-extraction-and-explanation | 2306.14040 | null | https://arxiv.org/abs/2306.14040v1 | https://arxiv.org/pdf/2306.14040v1.pdf | Weighted Automata Extraction and Explanation of Recurrent Neural Networks for Natural Language Tasks | Recurrent Neural Networks (RNNs) have achieved tremendous success in processing sequential data, yet understanding and analyzing their behaviours remains a significant challenge. To this end, many efforts have been made to extract finite automata from RNNs, which are more amenable for analysis and explanation. However, existing approaches like exact learning and compositional approaches for model extraction have limitations in either scalability or precision. In this paper, we propose a novel framework of Weighted Finite Automata (WFA) extraction and explanation to tackle the limitations for natural language tasks. First, to address the transition sparsity and context loss problems we identified in WFA extraction for natural language tasks, we propose an empirical method to complement missing rules in the transition diagram, and adjust transition matrices to enhance the context-awareness of the WFA. We also propose two data augmentation tactics to track more dynamic behaviours of RNN, which further allows us to improve the extraction precision. Based on the extracted model, we propose an explanation method for RNNs including a word embedding method -- Transition Matrix Embeddings (TME) and TME-based task oriented explanation for the target RNN. Our evaluation demonstrates the advantage of our method in extraction precision than existing approaches, and the effectiveness of TME-based explanation method in applications to pretraining and adversarial example generation. | ['Meng Sun', 'Yihao Zhang', 'Xiyue Zhang', 'Zeming Wei'] | 2023-06-24 | null | null | null | null | ['model-extraction', 'model-extraction'] | ['adversarial', 'methodology'] | [ 6.39596760e-01 3.30989778e-01 -2.34387457e-01 -8.59473124e-02
-4.62807715e-01 -4.05822277e-01 6.86244786e-01 -1.74784064e-01
-1.24861494e-01 4.60051596e-01 5.17932475e-01 -7.95941651e-01
-1.18950278e-01 -7.02397525e-01 -5.93689084e-01 -4.10095394e-01
2.67749447e-02 1.15237966e-01 1.80204943e-01 -2.66811311e-01
1.56387836e-01 5.22817075e-01 -1.57570231e+00 2.16410980e-01
7.40976095e-01 6.56774700e-01 2.14007542e-01 8.05585384e-01
-5.54916739e-01 1.00660515e+00 -2.88456410e-01 -2.83653736e-01
5.28441109e-02 -5.26658714e-01 -8.88837099e-01 -1.86816547e-02
-1.15628168e-02 -2.11491957e-01 -7.25412130e-01 8.88702750e-01
2.32825398e-01 3.70737880e-01 7.31763422e-01 -1.46588111e+00
-5.79398215e-01 1.00981677e+00 -6.39935257e-03 1.07416190e-01
1.10159650e-01 1.32388547e-01 1.15899551e+00 -8.98847222e-01
2.97950894e-01 1.23188436e+00 6.47230744e-01 1.10370958e+00
-1.03397274e+00 -7.01287508e-01 4.42336321e-01 2.32987165e-01
-9.76583123e-01 -5.69248617e-01 8.62322271e-01 -1.91426530e-01
1.45183992e+00 1.56617433e-01 6.03085518e-01 1.32495415e+00
8.45883340e-02 9.94850993e-01 9.08745944e-01 -6.16536379e-01
1.15232810e-01 -1.70674369e-01 4.79126781e-01 1.06539488e+00
7.90342242e-02 1.38783172e-01 -3.34822029e-01 -8.47407579e-02
8.06826413e-01 3.10154378e-01 3.39087769e-02 -1.18770160e-01
-1.09805274e+00 8.16788733e-01 3.79052013e-02 3.09755504e-01
-1.87943190e-01 3.39556575e-01 5.45971036e-01 1.16538070e-01
2.44565964e-01 4.30806369e-01 -5.45770407e-01 -4.67650026e-01
-6.66200161e-01 2.58894712e-01 7.02391624e-01 8.45984161e-01
5.58577359e-01 4.62939739e-01 -1.82413876e-01 7.17351139e-01
3.08191270e-01 5.06154120e-01 8.91893804e-01 -5.61418533e-01
6.00543439e-01 9.75428998e-01 -2.57910699e-01 -1.02358711e+00
-4.39187557e-01 -2.24566519e-01 -1.00133598e+00 -2.15741992e-01
1.88342169e-01 4.93992791e-02 -1.14005101e+00 2.00657272e+00
1.52568206e-01 4.09254014e-01 3.67917538e-01 5.69272459e-01
5.71290851e-01 8.20407152e-01 -1.31049931e-01 -9.00772363e-02
1.32943511e+00 -1.13540518e+00 -9.23689306e-01 -4.29833293e-01
9.12555158e-01 -2.66611844e-01 1.26226199e+00 1.31801113e-01
-7.04846799e-01 -4.53614205e-01 -1.13614559e+00 4.26942296e-02
-3.69410247e-01 6.40444756e-02 6.91689253e-01 4.83532071e-01
-6.27853334e-01 5.17587841e-01 -1.12306702e+00 -1.85717996e-02
2.83794761e-01 6.16001070e-01 -2.44758546e-01 1.07401706e-01
-1.49259901e+00 7.92085946e-01 5.47041893e-01 1.96983784e-01
-7.06823885e-01 -3.51339310e-01 -1.28633904e+00 1.78737968e-01
5.59880376e-01 -3.63290280e-01 1.35834897e+00 -7.84842551e-01
-1.63529098e+00 7.89468959e-02 -4.30235147e-01 -8.20699036e-01
-2.80390650e-01 -5.65710068e-02 -6.08687639e-01 -2.66408443e-01
-2.78346837e-01 4.92067397e-01 8.86261702e-01 -1.01893079e+00
-5.15063584e-01 -2.18621001e-01 8.19930136e-02 -7.15105906e-02
-6.94401503e-01 -1.57722592e-01 -2.72786707e-01 -8.38561416e-01
1.06836013e-01 -1.11343920e+00 -5.18847764e-01 -7.21051514e-01
-5.51035225e-01 -3.42291564e-01 8.84958267e-01 -6.42159700e-01
1.74206316e+00 -1.94874477e+00 1.69102699e-01 3.96848768e-01
4.16765362e-01 7.05289662e-01 -2.55175948e-01 4.10482466e-01
2.34398749e-02 2.60833621e-01 -5.22714794e-01 -3.97569537e-01
1.70215279e-01 6.77506387e-01 -7.66225219e-01 -2.42423043e-01
5.18617630e-01 1.04646242e+00 -8.14098179e-01 -3.89979869e-01
7.15632215e-02 3.72295290e-01 -4.50412929e-01 3.26242834e-01
-3.28781217e-01 7.01848343e-02 -4.40927804e-01 3.77117336e-01
1.65458560e-01 -5.98338246e-02 2.02814788e-01 2.64247265e-02
6.77525997e-02 5.62817395e-01 -1.30729246e+00 1.23507869e+00
-7.50266790e-01 5.94923019e-01 -6.03418648e-01 -9.57903981e-01
1.01183987e+00 3.46270353e-01 -7.12134293e-04 -4.66628432e-01
1.69600621e-01 1.68694869e-01 2.20170319e-01 -5.37480772e-01
6.56130910e-01 -2.68818550e-02 -1.49181753e-01 6.95354283e-01
2.08306871e-02 1.67892545e-01 1.13497637e-01 2.47896805e-01
1.32663035e+00 9.08406302e-02 3.92423302e-01 1.05609179e-01
6.47249341e-01 -7.39925206e-02 7.32366800e-01 6.30734086e-01
-9.25720856e-02 2.87205726e-01 4.70538974e-01 -6.87843382e-01
-1.16186225e+00 -5.99352121e-01 4.08741295e-01 7.57012963e-01
-1.41743287e-01 -6.83260024e-01 -8.75431478e-01 -8.71264637e-01
-3.63741368e-01 7.37857580e-01 -5.52679181e-01 -5.04377365e-01
-8.47058713e-01 -3.88347507e-01 9.41069365e-01 1.01162887e+00
4.68822777e-01 -1.37790287e+00 -4.50667679e-01 3.07628483e-01
-3.39678556e-01 -1.39158177e+00 -6.94377363e-01 2.31624395e-01
-1.10570347e+00 -8.39369416e-01 -1.35536164e-01 -6.49197519e-01
7.77698398e-01 1.33284763e-01 5.81330478e-01 1.53037563e-01
-6.81281015e-02 1.81618825e-01 -3.79012078e-01 -3.90929520e-01
-7.11446047e-01 3.18535954e-01 4.91438061e-01 4.18912023e-02
3.48818004e-01 -5.28225124e-01 -1.99370697e-01 2.80501008e-01
-1.17352223e+00 3.15328717e-01 8.11571062e-01 9.63445902e-01
5.90771496e-01 1.23841856e-02 5.48980951e-01 -1.05461121e+00
9.40558434e-01 -2.15890944e-01 -3.79572421e-01 1.64813116e-01
-8.69098902e-01 7.02684224e-01 1.08385599e+00 -8.30181718e-01
-9.46375251e-01 2.79114127e-01 -3.24887156e-01 -5.99205554e-01
-6.14341386e-02 5.99610090e-01 -2.19889358e-01 2.35088632e-01
4.65432882e-01 5.77653170e-01 1.71193361e-01 -2.09733099e-01
4.39572364e-01 6.51755333e-01 4.17953044e-01 -4.14448470e-01
8.63823354e-01 7.44251087e-02 3.17423977e-02 -7.43177831e-01
-8.01906884e-01 -1.69837236e-01 -4.28537369e-01 6.78299591e-02
5.96824884e-01 -4.17100996e-01 -5.32709122e-01 2.28189185e-01
-1.34756589e+00 -3.64540428e-01 -2.86494911e-01 5.29493809e-01
-4.42323655e-01 4.39723551e-01 -5.95688403e-01 -1.13823879e+00
-6.49681568e-01 -1.27012420e+00 6.87562823e-01 5.14990576e-02
-4.61471975e-01 -1.00456285e+00 2.10748285e-01 1.64195731e-01
4.50825721e-01 -4.63839732e-02 1.15618503e+00 -1.12383294e+00
-4.63637233e-01 -2.93428987e-01 -1.14277929e-01 4.39745963e-01
1.21514797e-01 -4.66859639e-02 -8.79263878e-01 -1.02785025e-02
-1.16988406e-01 -1.31778777e-01 6.98926628e-01 2.09036976e-01
1.15021384e+00 -7.29226828e-01 -3.85985285e-01 3.15902174e-01
1.05213487e+00 2.86987036e-01 6.05579793e-01 2.81028867e-01
9.53895092e-01 4.12230343e-01 4.79241580e-01 2.54227668e-01
3.83379996e-01 5.77866137e-01 4.17159379e-01 1.80144176e-01
-9.14285854e-02 -5.65530539e-01 7.43495405e-01 1.38938975e+00
-1.50286004e-01 -1.50722474e-01 -9.80719924e-01 6.96276844e-01
-2.07538605e+00 -8.86392117e-01 1.01242559e-02 1.92856193e+00
6.87308788e-01 4.33304906e-01 4.46018800e-02 4.80733454e-01
5.65419555e-01 2.07705304e-01 -5.05721331e-01 -8.29424798e-01
2.11168513e-01 2.64594495e-01 2.53074586e-01 5.61178267e-01
-7.39903212e-01 1.11175835e+00 5.76062250e+00 1.02111769e+00
-8.30962539e-01 -3.01806740e-02 1.67091921e-01 3.41419578e-01
-5.31094134e-01 1.47123694e-01 -1.14149249e+00 2.84430087e-01
1.32163513e+00 1.50191203e-01 6.12088978e-01 7.29346275e-01
1.21133350e-01 6.46776795e-01 -1.01145673e+00 7.82558739e-01
-2.42573153e-02 -1.33745372e+00 3.74394149e-01 9.01827291e-02
4.76726830e-01 -3.18092644e-01 -3.57187726e-02 6.65989816e-01
2.78771430e-01 -1.03901398e+00 4.81501013e-01 5.38814306e-01
5.65314651e-01 -9.34208214e-01 8.28022182e-01 5.69456339e-01
-1.47394276e+00 -3.12066019e-01 -2.92484224e-01 -3.02062303e-01
-4.55994047e-02 3.93778980e-01 -9.67486918e-01 3.86094242e-01
1.55540049e-01 4.58495677e-01 -3.78322423e-01 3.14420909e-01
-4.30614293e-01 9.09241915e-01 -1.66677460e-01 -4.60695535e-01
3.85368437e-01 -7.83689767e-02 5.32258451e-01 1.27787399e+00
2.28721425e-01 2.46957481e-01 -7.77349472e-02 9.27169561e-01
-1.38845265e-01 2.57865861e-02 -9.32153940e-01 -5.16302824e-01
6.19680226e-01 1.22084272e+00 -6.80835664e-01 -1.42197281e-01
-3.55603099e-01 8.71915519e-01 5.37115335e-01 1.30923316e-01
-9.49403524e-01 -5.88506341e-01 6.03315055e-01 -1.28535733e-01
1.93285018e-01 -4.97974277e-01 -1.28360242e-01 -1.23149204e+00
3.33878212e-02 -1.25819826e+00 1.71671420e-01 -3.76092702e-01
-7.27654874e-01 9.23577785e-01 -1.26959579e-02 -1.11480916e+00
-5.57233810e-01 -6.43868923e-01 -7.43445754e-01 4.60011810e-01
-1.10852897e+00 -1.30665350e+00 1.86872631e-01 4.65452701e-01
9.42854941e-01 -3.87515128e-01 9.13006246e-01 4.80129234e-02
-8.85704875e-01 8.17532361e-01 -2.80982316e-01 2.87954301e-01
-6.31497726e-02 -1.07967532e+00 8.71502578e-01 1.13074327e+00
6.07475638e-01 9.67980385e-01 5.76646447e-01 -6.31148815e-01
-1.65578544e+00 -1.29728138e+00 1.10153925e+00 -4.67771679e-01
7.03366995e-01 -5.56322157e-01 -9.90795910e-01 8.86954904e-01
-1.08046778e-01 -2.18997613e-01 5.38257182e-01 2.27467403e-01
-3.09622675e-01 1.65710509e-01 -6.71173275e-01 1.12370956e+00
1.15547132e+00 -7.40087211e-01 -8.54126573e-01 -9.73543674e-02
1.08428276e+00 -2.87092775e-01 -5.58866739e-01 4.98271406e-01
7.09643424e-01 -3.85009348e-01 7.75579154e-01 -9.82869685e-01
6.06758356e-01 -1.49344578e-01 -1.29750371e-01 -1.25566804e+00
-1.37219504e-01 -7.54056454e-01 -7.71474421e-01 1.40931940e+00
6.83247447e-01 -6.39585733e-01 7.94304371e-01 6.95321441e-01
-2.23248959e-01 -1.13683712e+00 -7.43331313e-01 -7.82992125e-01
-2.63434410e-01 -1.03771603e+00 7.77679682e-01 6.75896764e-01
1.72636479e-01 6.04847014e-01 -6.41461790e-01 2.28288263e-01
2.36027360e-01 3.92665863e-02 7.76224315e-01 -9.33658421e-01
-2.99647391e-01 -3.27564895e-01 -4.03379947e-01 -1.13178122e+00
5.99379420e-01 -9.64575887e-01 1.51076049e-01 -1.47382534e+00
1.67658612e-01 -1.85930401e-01 -3.71550977e-01 8.64096642e-01
-1.40887335e-01 -4.64616083e-02 1.12771273e-01 1.71975121e-01
-4.74820554e-01 6.37909174e-01 9.34628427e-01 -1.21282235e-01
-3.77441227e-01 1.32890761e-01 -5.26636779e-01 7.70881236e-01
9.25955415e-01 -6.46658778e-01 -6.29412949e-01 -3.40089858e-01
1.94367185e-01 1.57611407e-02 2.10649282e-01 -8.50914001e-01
2.35771298e-01 -1.04343943e-01 -2.86093503e-01 -4.44600791e-01
1.97237954e-01 -8.99925470e-01 -1.25457674e-01 5.74938953e-01
-5.71371734e-01 2.13465348e-01 3.36721361e-01 8.24395657e-01
-1.26884133e-01 -3.70372295e-01 2.41457939e-01 2.59529740e-01
-6.68787241e-01 4.34013128e-01 -7.92388439e-01 3.66468541e-02
6.00507915e-01 -2.11077288e-01 -6.00828230e-02 -3.99245143e-01
-6.35194898e-01 1.37732431e-01 -6.63817078e-02 5.44578195e-01
1.06918585e+00 -1.53997660e+00 -2.39266187e-01 5.01209497e-01
8.35867226e-02 -2.10766658e-01 8.03772584e-02 7.80377448e-01
-8.91191959e-02 6.15000248e-01 -1.10244215e-01 -5.63036464e-02
-1.32817471e+00 6.32974982e-01 1.26774564e-01 -7.61316299e-01
-6.97706640e-01 4.98569071e-01 -7.62807131e-02 -7.54043818e-01
2.99648970e-01 -7.95960188e-01 -3.72674108e-01 -3.90576810e-01
5.32213271e-01 3.41906607e-01 -6.74097333e-03 -4.14984405e-01
-2.21053198e-01 2.84699082e-01 -3.16011518e-01 -1.61939874e-01
1.19306028e+00 3.18040773e-02 1.03340045e-01 5.00465512e-01
8.75096977e-01 -9.44429263e-02 -9.81719017e-01 -3.89002711e-01
2.87127495e-01 2.93534938e-02 -2.56061852e-02 -3.20527911e-01
-7.98397899e-01 1.13594234e+00 3.99053931e-01 4.56427485e-01
8.96151364e-01 -2.35564008e-01 1.34624708e+00 6.81305647e-01
6.32264540e-02 -8.77448201e-01 1.21164136e-02 9.34828520e-01
5.88801682e-01 -9.31228936e-01 -4.44476873e-01 -1.27590284e-01
-7.17961371e-01 1.31611025e+00 5.45430362e-01 8.60251933e-02
4.21342283e-01 4.53550845e-01 -2.35523567e-01 -7.59375095e-02
-9.34854627e-01 -2.55179405e-01 3.79865915e-01 4.04013932e-01
1.73435196e-01 -6.25257790e-02 -9.51392502e-02 1.08825386e+00
-1.08673282e-01 -1.57076582e-01 4.84484673e-01 8.49250913e-01
-4.46866244e-01 -1.33602870e+00 -1.22228920e-01 6.11854792e-01
-3.23919564e-01 -3.94087851e-01 -5.21757305e-01 7.07500279e-01
-1.71951085e-01 7.69934118e-01 -2.64091998e-01 -9.93969262e-01
4.48670387e-01 5.18212080e-01 2.77488351e-01 -6.61393166e-01
-3.87344062e-01 -2.87252754e-01 1.53240532e-01 -4.16216254e-01
-2.96115786e-01 -4.29682285e-01 -1.73078465e+00 -5.24701700e-02
-6.69797719e-01 1.26991272e-01 5.50622940e-01 1.42491150e+00
3.45244944e-01 6.89820766e-01 5.87634265e-01 -6.02598310e-01
-6.87612951e-01 -8.79221797e-01 -3.20682645e-01 2.83258021e-01
2.62558967e-01 -5.33924341e-01 -4.38860387e-01 -1.92717202e-02] | [10.752954483032227, 6.924848556518555] |
512690d1-81bf-480a-a346-0cceb3afed64 | learning-with-partial-labels-from-semi | 2211.13655 | null | https://arxiv.org/abs/2211.13655v2 | https://arxiv.org/pdf/2211.13655v2.pdf | Learning with Partial Labels from Semi-supervised Perspective | Partial Label (PL) learning refers to the task of learning from the partially labeled data, where each training instance is ambiguously equipped with a set of candidate labels but only one is valid. Advances in the recent deep PL learning literature have shown that the deep learning paradigms, e.g., self-training, contrastive learning, or class activate values, can achieve promising performance. Inspired by the impressive success of deep Semi-Supervised (SS) learning, we transform the PL learning problem into the SS learning problem, and propose a novel PL learning method, namely Partial Label learning with Semi-supervised Perspective (PLSP). Specifically, we first form the pseudo-labeled dataset by selecting a small number of reliable pseudo-labeled instances with high-confidence prediction scores and treating the remaining instances as pseudo-unlabeled ones. Then we design a SS learning objective, consisting of a supervised loss for pseudo-labeled instances and a semantic consistency regularization for pseudo-unlabeled instances. We further introduce a complementary regularization for those non-candidate labels to constrain the model predictions on them to be as small as possible. Empirical results demonstrate that PLSP significantly outperforms the existing PL baseline methods, especially on high ambiguity levels. Code available: https://github.com/changchunli/PLSP. | ['Jihong Ouyang', 'Yiyuan Wang', 'Changchun Li', 'Yuanzhi Jiang', 'Ximing Li'] | 2022-11-24 | null | null | null | null | ['partial-label-learning'] | ['methodology'] | [ 2.52396375e-01 5.79061151e-01 -7.61815190e-01 -9.23543215e-01
-1.25297880e+00 -5.81971705e-01 4.26144332e-01 1.36139372e-03
-2.29218200e-01 1.05025315e+00 -6.23586662e-02 2.11094785e-02
-5.64073250e-02 -3.49049777e-01 -8.66418362e-01 -6.97968721e-01
4.34300780e-01 8.99048150e-01 -1.53211460e-01 4.49604064e-01
-3.63166369e-02 1.92110285e-01 -1.49172187e+00 2.98576713e-01
1.11796653e+00 1.29203689e+00 6.75760955e-02 -1.04098313e-01
-1.58108398e-01 8.01048219e-01 -3.15937430e-01 -3.99215341e-01
3.67342234e-01 -3.99778992e-01 -9.57589328e-01 4.77658689e-01
5.61891973e-01 -8.69540945e-02 1.15656406e-01 1.24285030e+00
2.85231084e-01 1.04107469e-01 6.76067352e-01 -1.52238178e+00
-7.87505984e-01 7.04219282e-01 -6.04709685e-01 -3.11634153e-01
-3.78052518e-02 -1.21250510e-01 1.34244096e+00 -1.32732224e+00
5.58860838e-01 1.22362983e+00 7.22002089e-01 6.89345717e-01
-1.41387331e+00 -8.64788890e-01 1.67695194e-01 1.24759555e-01
-1.38087308e+00 -3.51446897e-01 9.57285941e-01 -3.06164891e-01
2.79625684e-01 3.85679305e-02 -9.36204218e-04 1.19796407e+00
-4.07748729e-01 9.91832137e-01 1.32199585e+00 -4.69033867e-01
3.50717962e-01 3.85769337e-01 4.28823233e-01 7.25759268e-01
2.49425054e-01 1.42795011e-01 -4.16191161e-01 -1.29042387e-01
3.68023604e-01 -5.04563674e-02 -1.42259732e-01 -6.25926852e-01
-1.06364572e+00 8.01659048e-01 4.10778701e-01 2.39033182e-03
-2.13227853e-01 -2.34922558e-01 2.41232947e-01 4.04995345e-02
6.59661174e-01 4.17352527e-01 -9.47995842e-01 4.72324967e-01
-8.90104771e-01 1.84111192e-03 5.43412864e-01 1.14430499e+00
1.15694940e+00 -7.39887357e-02 -3.63022923e-01 1.19873750e+00
5.81269145e-01 3.98532718e-01 5.07158756e-01 -1.17595315e+00
5.03388405e-01 7.53049433e-01 2.80917943e-01 -4.46759671e-01
-3.86869937e-01 -5.69347560e-01 -6.56292617e-01 -5.62572479e-02
2.34333694e-01 -2.58127630e-01 -1.04857373e+00 2.09091711e+00
4.44107682e-01 3.13125014e-01 2.14226723e-01 9.22670722e-01
8.21821034e-01 6.67037666e-01 3.86616170e-01 -5.22260308e-01
9.71253872e-01 -1.39828014e+00 -6.26080930e-01 -5.86628556e-01
6.40974700e-01 -5.82694471e-01 1.26471484e+00 3.32659662e-01
-8.22666645e-01 -6.59167588e-01 -9.75871861e-01 -5.95535524e-02
-2.17176363e-01 7.61145055e-01 3.16765696e-01 2.39438757e-01
-8.43885601e-01 5.94612777e-01 -5.68481088e-01 3.75290290e-02
5.82079530e-01 4.24980044e-01 -4.33584481e-01 -2.56154180e-01
-1.21515107e+00 5.65358043e-01 7.84126878e-01 1.31181404e-01
-8.96346569e-01 -4.97559220e-01 -1.05426288e+00 6.43866807e-02
7.11780131e-01 -1.76285863e-01 1.31501961e+00 -1.46702266e+00
-1.21193326e+00 1.30757523e+00 -1.67028710e-01 -1.66464776e-01
5.10888755e-01 -1.94945440e-01 -3.93574417e-01 -1.84595138e-02
6.45982802e-01 9.84577417e-01 8.49655390e-01 -1.89690793e+00
-6.55951798e-01 -2.46391237e-01 -2.07548574e-01 3.94514680e-01
-2.55246341e-01 -2.91872650e-01 -3.57954890e-01 -5.04189312e-01
5.62745035e-01 -1.21271968e+00 -9.92856324e-02 -1.70859426e-01
-7.42205918e-01 -5.50029635e-01 6.10001445e-01 -4.31110203e-01
7.80766487e-01 -2.12617922e+00 -1.47666000e-02 7.01956600e-02
1.40476331e-01 4.72505629e-01 -2.47332036e-01 4.14811261e-02
-2.72760004e-01 1.36906549e-01 -5.42307317e-01 -7.04169512e-01
1.44693583e-01 3.95219594e-01 -3.95228028e-01 1.68942317e-01
4.02686387e-01 8.92344832e-01 -1.09138370e+00 -5.26867986e-01
4.65488322e-02 1.78136721e-01 -2.17756748e-01 4.57593799e-01
-4.49053943e-01 6.49126172e-01 -4.37396854e-01 8.27513278e-01
7.93280602e-01 -8.08755875e-01 3.65185499e-01 -2.06320390e-01
1.33696854e-01 2.01332867e-01 -1.10339165e+00 1.51827323e+00
-2.82144070e-01 6.84519336e-02 6.30918890e-02 -1.16063190e+00
1.01574004e+00 2.67549872e-01 3.61272633e-01 -3.49275231e-01
-7.93450177e-02 4.66830790e-01 -4.62373763e-01 -3.75972956e-01
1.95146322e-01 -2.54203707e-01 6.71966076e-02 4.13019478e-01
3.84059012e-01 2.44902112e-02 1.79907069e-01 1.42486040e-02
5.37555695e-01 4.37070489e-01 2.37558931e-01 -1.76630482e-01
6.30908072e-01 -2.94433295e-04 1.11233568e+00 5.97251892e-01
-4.41014647e-01 7.02119410e-01 4.53449517e-01 -2.55839705e-01
-9.17201042e-01 -1.19548285e+00 -3.71411800e-01 1.16785085e+00
4.18016970e-01 -1.70459718e-01 -5.46858013e-01 -1.30958188e+00
9.65635404e-02 7.42173612e-01 -5.08596063e-01 -2.86055386e-01
-3.78078431e-01 -7.77945518e-01 8.78195837e-02 5.22912741e-01
6.49935663e-01 -1.28158200e+00 2.52128839e-01 2.45256890e-02
-2.07299560e-01 -1.02253819e+00 -3.81824523e-01 6.01000249e-01
-7.95236468e-01 -1.08286524e+00 -4.42294985e-01 -1.27631223e+00
1.04795623e+00 3.10510602e-02 1.17069721e+00 -1.02083877e-01
3.53236198e-01 8.15070644e-02 -3.93042564e-01 -4.39333022e-02
-3.81359309e-01 -1.17303440e-02 1.82222813e-01 2.02097714e-01
4.73572820e-01 -2.79163331e-01 -3.59826148e-01 3.67571294e-01
-8.13358128e-01 1.08783416e-01 6.96564794e-01 1.31097734e+00
1.17795229e+00 -7.71064609e-02 1.04237103e+00 -1.42814243e+00
3.27349752e-01 -6.67620957e-01 -4.59962845e-01 5.38341463e-01
-1.06139350e+00 1.73825234e-01 6.82411253e-01 -5.13672531e-01
-1.21316886e+00 3.75173658e-01 1.12889009e-02 -6.36399686e-01
-2.71009833e-01 5.27340472e-01 -3.56956065e-01 9.26870629e-02
5.26526749e-01 1.27816409e-01 -2.14129522e-01 -4.66528028e-01
3.60208631e-01 6.27871990e-01 3.86467785e-01 -7.12418973e-01
6.61821365e-01 3.21440876e-01 -1.89486042e-01 -1.95510760e-01
-1.88276184e+00 -3.49371850e-01 -7.99466133e-01 -8.11403990e-02
5.62364459e-01 -9.86874521e-01 -2.24748343e-01 2.47400612e-01
-8.33059907e-01 -5.46418488e-01 -4.76042628e-01 4.41616923e-01
-5.05559325e-01 4.05996710e-01 -6.05210304e-01 -5.84592402e-01
-1.70788422e-01 -1.08087790e+00 1.16790724e+00 3.53697807e-01
-1.06180616e-01 -8.62987459e-01 -1.43806608e-02 6.98509753e-01
-8.56350586e-02 -7.10476860e-02 7.70783961e-01 -1.15511024e+00
-2.91561276e-01 -1.39223725e-01 -3.85738283e-01 8.04633677e-01
1.24597460e-01 -3.02927136e-01 -1.07452095e+00 -3.80448371e-01
-5.92238037e-03 -1.10698926e+00 8.50109994e-01 3.19412231e-01
1.33261931e+00 -3.32985103e-01 -3.18247378e-01 5.17404556e-01
1.28455365e+00 4.25529368e-02 1.68652803e-01 1.07226335e-01
8.22171569e-01 6.62307858e-01 1.03734362e+00 2.89947450e-01
3.81945759e-01 4.09675419e-01 4.36123967e-01 -1.92796122e-02
-7.93401599e-02 -5.83219707e-01 -1.88117083e-02 6.93316519e-01
5.52654207e-01 -1.98632136e-01 -9.03126121e-01 2.67233610e-01
-2.12478471e+00 -4.85900819e-01 1.18465059e-01 2.22974300e+00
1.08794951e+00 1.71249554e-01 -3.88558358e-01 1.32353026e-02
1.08790028e+00 2.16734353e-02 -9.37339902e-01 2.32111648e-01
-2.66349912e-01 -4.72017303e-02 2.64615953e-01 4.74877954e-01
-1.57711339e+00 1.18636191e+00 5.53295660e+00 9.34487820e-01
-8.90499115e-01 2.47023955e-01 1.15722466e+00 1.52032688e-01
-3.73248547e-01 2.23045498e-02 -9.70277965e-01 6.03728175e-01
7.35805929e-01 1.84853822e-01 1.96180433e-01 1.13769174e+00
-1.51977250e-02 7.86069483e-02 -1.20163059e+00 1.02302742e+00
1.65102720e-01 -9.95122075e-01 6.83612451e-02 -3.25045645e-01
1.17603886e+00 -8.21198449e-02 2.40258396e-01 6.81123376e-01
3.60975027e-01 -8.39412570e-01 7.71764755e-01 2.09893644e-01
9.20062721e-01 -5.51764607e-01 8.42043340e-01 5.05139053e-01
-8.67420137e-01 -2.63333440e-01 -4.26027358e-01 1.63651586e-01
2.51844525e-02 6.37514114e-01 -5.95075369e-01 3.61720741e-01
4.78993833e-01 9.56432343e-01 -5.30034065e-01 7.49373615e-01
-7.30486751e-01 8.56296182e-01 -5.14646918e-02 3.57677698e-01
3.01423877e-01 -3.89874309e-01 8.57020393e-02 8.19911122e-01
1.16479084e-01 3.17099392e-02 5.69011033e-01 1.13505304e+00
-5.10488093e-01 1.09954432e-01 -2.69384444e-01 1.13564149e-01
8.75049531e-01 1.39705706e+00 -5.89433074e-01 -5.17489731e-01
-4.29571152e-01 9.03578520e-01 8.42954040e-01 4.50883538e-01
-7.04967797e-01 1.41965419e-01 1.03735916e-01 -2.13088185e-01
-6.07683510e-02 2.52994299e-01 -4.33036625e-01 -1.16509056e+00
1.53435156e-01 -6.33877575e-01 5.59496820e-01 -9.39201415e-01
-1.73843217e+00 5.38157344e-01 -2.16252893e-01 -1.38249409e+00
-8.59385803e-02 -4.41896051e-01 -2.50627488e-01 6.75768495e-01
-1.72823822e+00 -1.32911658e+00 -2.18943447e-01 3.27786595e-01
6.06755376e-01 -1.73418060e-01 6.94515288e-01 2.16624305e-01
-7.19170511e-01 6.01844132e-01 2.87260056e-01 4.83264253e-02
1.00666237e+00 -1.43670678e+00 -1.33511141e-01 6.29573047e-01
2.09599584e-01 3.73791188e-01 3.67734104e-01 -6.77358508e-01
-6.73801303e-01 -1.43411911e+00 1.12463844e+00 -2.98123151e-01
3.99267823e-01 -2.78401315e-01 -1.08742774e+00 9.66034472e-01
-1.65805489e-01 4.53850031e-01 7.23638773e-01 -4.32905108e-02
-4.96068060e-01 -1.40205681e-01 -1.24022174e+00 4.02669072e-01
9.05825794e-01 -4.25429463e-01 -6.48271382e-01 8.09781969e-01
8.91380787e-01 -3.42269123e-01 -5.21111429e-01 8.22439551e-01
3.74071933e-02 -7.43105769e-01 8.24863672e-01 -5.57881117e-01
4.28254932e-01 -2.91595459e-01 2.91260257e-02 -1.27613330e+00
-3.24248463e-01 -1.39825761e-01 -8.06184262e-02 1.35513699e+00
6.68622077e-01 -5.44249296e-01 1.19961548e+00 8.71138334e-01
-4.56588000e-01 -9.13767159e-01 -7.98215330e-01 -8.14903319e-01
2.80317869e-02 -1.61682460e-02 2.55286634e-01 1.42831862e+00
-3.54795679e-02 4.64844644e-01 -4.97207642e-01 2.61991262e-01
8.21730256e-01 4.45075274e-01 1.92782477e-01 -1.41731942e+00
-2.56203860e-01 -4.92966454e-03 2.99767137e-01 -9.75494623e-01
8.21599245e-01 -1.16367197e+00 3.55166554e-01 -1.34752584e+00
4.40073222e-01 -9.15711045e-01 -6.02710545e-01 9.50418711e-01
-4.47605312e-01 2.98338294e-01 -2.16137484e-01 6.46516740e-01
-9.78395164e-01 6.41097546e-01 1.23753262e+00 -1.09797731e-01
-2.17018589e-01 1.48288444e-01 -6.54276669e-01 7.37045407e-01
9.37304676e-01 -6.23334110e-01 -5.09668291e-01 -2.39611790e-01
1.21984586e-01 -7.35550374e-02 5.36534265e-02 -7.26857662e-01
-6.58929721e-02 -2.02941850e-01 3.46041262e-01 -5.26220858e-01
2.28943542e-01 -7.85552382e-01 -1.18649423e-01 2.21937522e-01
-9.81767356e-01 -4.73793030e-01 -1.62270084e-01 5.31653464e-01
-3.66549104e-01 -4.86919731e-01 9.23396826e-01 -9.29386243e-02
-8.50736499e-01 4.15886849e-01 2.24672019e-01 3.35885435e-01
9.72656488e-01 1.62175328e-01 -3.29345047e-01 -5.75482473e-02
-1.07998312e+00 6.20949209e-01 2.80031562e-01 3.07765365e-01
4.29932058e-01 -1.57090569e+00 -5.70150852e-01 1.38132155e-01
3.16096008e-01 3.16075206e-01 1.91126660e-01 4.35216099e-01
-9.52807739e-02 4.46518362e-01 -6.74837530e-02 -6.31605089e-01
-9.77394879e-01 7.65967190e-01 1.48730323e-01 -2.43696824e-01
-2.10484251e-01 1.03743386e+00 3.77214640e-01 -1.03704035e+00
3.94340575e-01 2.98558995e-02 -2.20726997e-01 -1.03497906e-02
1.10647112e-01 1.62933782e-01 -9.34254471e-03 -6.91546857e-01
-1.86236635e-01 3.97325784e-01 -1.67295650e-01 1.36155665e-01
1.12047565e+00 -1.87223852e-01 -7.60735050e-02 6.12760544e-01
1.26206684e+00 -2.23807767e-01 -1.69536161e+00 -6.83003843e-01
3.83710474e-01 -1.80138290e-01 -1.28651649e-01 -1.17159593e+00
-1.05786479e+00 6.77930832e-01 4.61795032e-01 -1.70103595e-01
9.40848351e-01 3.42539430e-01 4.57747430e-01 4.27623272e-01
2.61128664e-01 -1.25791454e+00 3.58703196e-01 5.60181379e-01
6.49061620e-01 -1.96473563e+00 -2.01186150e-01 -5.61231971e-01
-9.21061218e-01 7.72010267e-01 1.03780055e+00 2.27019284e-02
5.58970571e-01 -6.18756637e-02 2.40815759e-01 -3.29883657e-02
-6.99998915e-01 -7.32573867e-02 3.99568975e-01 3.12251776e-01
4.54751760e-01 2.38359362e-01 -4.00250107e-01 9.34780002e-01
2.22897679e-01 -1.34235442e-01 2.19910830e-01 7.31438339e-01
-4.44088548e-01 -1.19191599e+00 -1.59860224e-01 6.02305412e-01
-2.06894159e-01 -7.53925592e-02 -4.48100150e-01 5.10462105e-01
4.31033760e-01 7.74108350e-01 -1.67431399e-01 -3.24941218e-01
-4.41162400e-02 4.47897106e-01 1.21301688e-01 -1.10237205e+00
-2.47062102e-01 1.09304860e-01 1.50548480e-02 -3.22606653e-01
-4.91324961e-01 -5.64466357e-01 -1.40438986e+00 3.70624155e-01
-4.24394816e-01 3.26497495e-01 2.90986091e-01 9.84691799e-01
2.42135167e-01 1.26720652e-01 7.78559387e-01 -7.37798154e-01
-8.65501106e-01 -1.01073027e+00 -8.27944040e-01 7.39880800e-01
1.31698579e-01 -8.82349133e-01 -6.49182558e-01 8.07654019e-03] | [9.506559371948242, 3.8527393341064453] |
6c9c00d1-e628-4870-822b-a4ba7ded4303 | graphwoz-dialogue-management-with | 2211.12852 | null | https://arxiv.org/abs/2211.12852v1 | https://arxiv.org/pdf/2211.12852v1.pdf | GraphWOZ: Dialogue Management with Conversational Knowledge Graphs | We present a new approach to dialogue management using conversational knowledge graphs as core representation of the dialogue state. To this end, we introduce a new dataset, GraphWOZ, which comprises Wizard-of-Oz dialogues in which human participants interact with a robot acting as a receptionist. In contrast to most existing work on dialogue management, GraphWOZ relies on a dialogue state explicitly represented as a dynamic knowledge graph instead of a fixed set of slots. This graph is composed of a varying number of entities (such as individuals, places, events, utterances and mentions) and relations between them (such as persons being part of a group or attending an event). The graph is then regularly updated on the basis of new observations and system actions. GraphWOZ is released along with detailed manual annotations related to the user intents, system responses, and reference relations occurring in both user and system turns. Based on GraphWOZ, we present experimental results for two dialogue management tasks, namely conversational entity linking and response ranking. For conversational entity linking, we show how to connect utterance mentions to their corresponding entity in the knowledge graph with a neural model relying on a combination of both string and graph-based features. Response ranking is then performed by summarizing the relevant content of the graph into a text, which is concatenated with the dialogue history and employed as input to score possible responses to a given dialogue state. | ['Pierre Lison', 'Stefan Ultes', 'Nicholas Thomas Walker'] | 2022-11-23 | null | null | null | null | ['dialogue-management'] | ['natural-language-processing'] | [ 1.89543828e-01 8.89239788e-01 -4.99354452e-02 -5.42149961e-01
-4.21134323e-01 -7.45747685e-01 9.68349993e-01 1.01144671e+00
-3.21708173e-01 8.26255798e-01 7.72699535e-01 1.47246197e-01
-8.21139216e-02 -9.34547424e-01 -1.34450555e-01 -2.62100194e-02
-2.30940863e-01 1.10342264e+00 4.95532960e-01 -9.01102424e-01
2.56806999e-01 3.36716734e-02 -1.37404132e+00 3.19050074e-01
1.27771124e-01 6.33152962e-01 1.27833560e-01 9.34583068e-01
-3.62830162e-01 1.07737112e+00 -1.00406790e+00 -3.56743425e-01
-2.86950976e-01 -5.53817570e-01 -1.66621089e+00 1.77913696e-01
-1.73194081e-01 -3.13905180e-01 -4.47449356e-01 6.65196300e-01
4.86989945e-01 8.18972945e-01 4.06922281e-01 -1.25215638e+00
3.31142843e-02 9.18653309e-01 2.60786682e-01 -5.23859188e-02
1.39231360e+00 -4.39469740e-02 1.17136633e+00 -4.66529489e-01
1.07244658e+00 1.61587453e+00 4.13979352e-01 6.10105813e-01
-8.78920078e-01 1.81641906e-01 2.36955225e-01 2.15889052e-01
-8.21141660e-01 -6.41638160e-01 4.33780909e-01 -4.27354574e-01
1.47463250e+00 4.34057087e-01 4.16370451e-01 7.55249679e-01
-3.39904577e-01 5.17828524e-01 3.25633943e-01 -7.21542954e-01
2.11044878e-01 2.31220156e-01 7.10223794e-01 8.84654343e-01
-5.84248304e-01 -6.20126784e-01 -7.89985657e-01 -6.51352704e-01
4.36874241e-01 -5.47332466e-01 -2.79297054e-01 -3.75051737e-01
-1.27522874e+00 8.09403956e-01 1.60585284e-01 2.55399436e-01
-3.91225129e-01 -3.18468183e-01 8.64234090e-01 4.33389068e-01
3.45016837e-01 5.33564210e-01 -6.27792239e-01 -4.43221152e-01
3.13888758e-01 4.12304938e-01 1.71046841e+00 8.60534012e-01
9.47913945e-01 -5.61688125e-01 -3.85128438e-01 1.19912958e+00
1.62971511e-01 2.99570039e-02 4.19740468e-01 -1.05919611e+00
6.43243730e-01 1.20664191e+00 4.74475622e-01 -1.16195607e+00
-8.29195976e-01 6.64923012e-01 -3.96545678e-01 -4.08725053e-01
4.95599627e-01 -2.93965429e-01 -2.23871186e-01 1.82952285e+00
6.84709251e-01 -1.06150404e-01 5.04283488e-01 5.78776598e-01
1.44188678e+00 6.02879405e-01 4.53154966e-02 -3.69408339e-01
1.66918409e+00 -7.92820096e-01 -9.80567873e-01 -4.12407182e-02
1.04886210e+00 -4.45307702e-01 5.87267518e-01 -1.72082812e-01
-9.49197173e-01 -2.69311160e-01 -5.41028798e-01 -1.08642466e-01
-6.55694366e-01 -3.03923965e-01 4.86953884e-01 1.72483921e-01
-1.13573456e+00 3.72511804e-01 -4.46781546e-01 -1.03573108e+00
-7.27874517e-01 4.34805989e-01 -5.53237200e-01 3.88407230e-01
-1.68234289e+00 1.34410489e+00 7.25631893e-01 4.75910399e-03
-3.49362820e-01 2.30889350e-01 -1.51480579e+00 -1.40653672e-02
6.44634306e-01 -5.67940950e-01 1.88233232e+00 -4.66383457e-01
-2.10023022e+00 8.28602016e-01 -1.37218729e-01 -3.95105034e-01
1.18413851e-01 3.82238887e-02 -3.38681608e-01 3.93678024e-02
3.49132232e-02 3.69499594e-01 1.18890367e-01 -1.06596875e+00
-9.62287307e-01 -3.14805269e-01 8.47291529e-01 7.59531081e-01
1.37317404e-01 3.28189999e-01 -7.91546345e-01 3.25062156e-01
3.12114935e-02 -9.34213698e-01 -2.47080445e-01 -9.06309307e-01
-7.63576925e-01 -8.97278666e-01 6.46134317e-01 -7.16475666e-01
1.42884624e+00 -1.66035700e+00 4.48300153e-01 2.37775028e-01
3.13961536e-01 1.40330911e-01 -1.23363614e-01 1.13318527e+00
1.46435574e-01 -2.15490341e-01 6.76264912e-02 -3.27526331e-01
3.24719369e-01 2.19277397e-01 3.48890498e-02 9.16258097e-02
-1.15003340e-01 6.18615568e-01 -1.34595072e+00 -3.30745667e-01
4.22161132e-01 4.78775203e-02 -1.43232837e-01 6.77266836e-01
-4.41039026e-01 4.04814869e-01 -4.80464280e-01 1.01022713e-01
-2.23464429e-01 -1.31981805e-01 8.30776393e-01 -1.08722292e-01
-3.35396640e-02 9.48136926e-01 -9.65558410e-01 1.56045389e+00
-4.57204789e-01 5.54673553e-01 7.07274973e-02 -7.21523643e-01
9.63073313e-01 7.54271746e-01 1.87451348e-01 -2.89625823e-01
1.52857095e-01 -1.64254650e-01 -2.54013926e-01 -8.25586319e-01
1.00863826e+00 4.71416086e-01 -7.63744473e-01 7.66438544e-01
3.64210784e-01 -8.64741504e-02 4.71988916e-01 7.00831711e-01
1.39665568e+00 -9.61323380e-02 8.92467022e-01 3.79984558e-01
6.86774254e-01 6.07990921e-02 1.73572764e-01 6.97001159e-01
-9.43197533e-02 -1.07124195e-01 8.80234122e-01 -3.83127391e-01
-7.92163074e-01 -5.90467691e-01 4.08503950e-01 1.54349589e+00
2.72952169e-01 -7.46650696e-01 -8.55078638e-01 -7.94943333e-01
-2.54868299e-01 6.35181069e-01 -5.69408536e-01 -2.27899686e-01
-6.56597197e-01 -2.23616496e-01 4.12072748e-01 5.03076762e-02
4.51846451e-01 -1.56407022e+00 -6.05082095e-01 4.23675865e-01
-7.60321677e-01 -1.22378087e+00 -4.04004872e-01 1.40407756e-01
-2.61118978e-01 -1.35174060e+00 -2.71659903e-02 -9.45750475e-01
4.60710466e-01 -1.30431382e-02 1.35461700e+00 1.47284225e-01
1.14854805e-01 9.60661232e-01 -6.42222524e-01 -6.54392987e-02
-9.81693923e-01 9.56079811e-02 3.60783301e-02 -4.12172154e-02
2.71477729e-01 -1.43095404e-01 -2.11414218e-01 3.77429008e-01
-4.68737811e-01 6.75264969e-02 -2.12741733e-01 8.53465915e-01
-1.74066603e-01 -1.44653663e-01 6.11187100e-01 -1.23659837e+00
1.14646924e+00 -6.79021001e-01 -1.94387451e-01 5.81592500e-01
1.00042149e-02 -4.59385775e-02 2.44959757e-01 -3.12794328e-01
-1.04079533e+00 8.96192119e-02 5.83494566e-02 3.38217586e-01
-3.35190713e-01 9.05535460e-01 -2.20201373e-01 3.31145227e-01
6.55928731e-01 -5.07508852e-02 6.76184893e-02 -1.47553846e-01
6.54999495e-01 9.65175092e-01 7.76055694e-01 -6.44498050e-01
1.75442979e-01 -1.65721387e-01 -5.41269243e-01 -9.83368635e-01
-6.51891351e-01 -9.43760455e-01 -8.09415698e-01 -6.00192130e-01
7.78627098e-01 -5.24460077e-01 -1.47144234e+00 4.42531884e-01
-1.52546239e+00 -7.14994788e-01 -9.82542634e-02 2.61816949e-01
-6.88279331e-01 5.35607338e-01 -9.03048754e-01 -1.11509955e+00
-1.94296077e-01 -7.13882923e-01 9.33519125e-01 1.55478016e-01
-7.48798609e-01 -1.33115947e+00 2.68313140e-01 3.49690646e-01
5.17365411e-02 2.82496601e-01 9.54076707e-01 -1.24414206e+00
-2.93657519e-02 -5.03484190e-01 -5.09492774e-03 -1.65235519e-01
3.15733254e-01 -1.90274075e-01 -6.31011188e-01 -1.51459947e-01
-2.48601303e-01 -5.20456195e-01 1.56450391e-01 -1.83297589e-01
2.28424385e-01 -6.11120999e-01 -4.56571668e-01 -3.96256298e-01
6.86685145e-01 5.31510830e-01 4.98749644e-01 3.74641120e-01
5.14938414e-01 1.24902093e+00 5.97276866e-01 6.68124080e-01
1.09006345e+00 1.21943402e+00 2.44524702e-01 8.38642865e-02
1.22935720e-01 -1.68515503e-01 4.10499811e-01 8.98951292e-01
-1.20481728e-02 -6.18588984e-01 -8.69496107e-01 5.08953452e-01
-2.37497711e+00 -9.85441923e-01 -1.75443850e-02 2.23324251e+00
9.56328273e-01 -1.79563791e-01 3.49200964e-01 -2.03445062e-01
1.13790858e+00 1.17778741e-01 -4.30695862e-01 -5.13923228e-01
1.78758040e-01 -2.98849016e-01 7.13678300e-02 1.04900682e+00
-1.05806458e+00 1.19333220e+00 5.99374819e+00 1.26460776e-01
-6.03675783e-01 -9.96624213e-03 1.86369613e-01 5.54761350e-01
1.60491168e-01 -2.47674510e-02 -8.12827229e-01 7.26572052e-02
1.03755128e+00 -2.99441516e-01 5.30713916e-01 4.09219265e-01
3.64968665e-02 -3.97653133e-01 -1.31407011e+00 7.00384974e-01
1.35262787e-01 -1.16763973e+00 -1.59041673e-01 -3.10695797e-01
8.69896114e-02 -2.24749014e-01 -6.55076146e-01 6.22175038e-01
9.95145261e-01 -5.31619191e-01 4.24127758e-01 4.54193532e-01
6.25681162e-01 -4.05506372e-01 9.05250490e-01 4.28874224e-01
-1.20156550e+00 1.24797486e-01 6.38390630e-02 -1.76464513e-01
3.04520458e-01 -2.79605091e-01 -1.72737658e+00 7.34449089e-01
3.00652921e-01 3.78772438e-01 -3.97572517e-02 5.47709703e-01
-3.06444615e-01 7.24071264e-02 -2.28888378e-01 -2.84226358e-01
-3.93626839e-02 1.16144177e-02 7.52559543e-01 1.35424078e+00
-1.55410245e-01 5.46070933e-01 6.34227991e-01 1.77393571e-01
-1.44448787e-01 2.33741626e-01 -7.80556858e-01 1.00749388e-01
8.12249601e-01 1.34989524e+00 -5.14919162e-01 -6.13119006e-01
-3.54546517e-01 9.35975492e-01 5.03231168e-01 3.60886216e-01
-2.45062053e-01 -4.95125592e-01 6.07692897e-01 -2.27232724e-01
-2.39338040e-01 -2.64390521e-02 6.21541202e-01 -9.55218911e-01
-3.53133142e-01 -7.86442876e-01 5.95314503e-01 -7.70056486e-01
-1.06305075e+00 8.97878945e-01 1.60531387e-01 -8.43701422e-01
-9.63799238e-01 -2.25745976e-01 -5.38591206e-01 7.75344670e-01
-9.08494711e-01 -8.16286743e-01 -3.81471634e-01 6.62430584e-01
5.86012006e-01 -2.70173937e-01 1.59139454e+00 -2.24947616e-01
-6.64947927e-01 3.33822519e-01 -3.66976470e-01 5.08134842e-01
7.56173730e-01 -1.37144649e+00 5.63978851e-01 4.41770181e-02
-7.57127181e-02 6.08493209e-01 8.02291334e-01 -7.86470532e-01
-1.09474742e+00 -7.29903519e-01 1.34316170e+00 -7.54599452e-01
7.08472311e-01 -3.89074385e-01 -9.39578831e-01 9.35461581e-01
3.90717268e-01 -5.45327008e-01 6.52571380e-01 5.13325095e-01
-1.67069018e-01 4.93432432e-01 -1.04362392e+00 5.45969665e-01
8.51040661e-01 -7.36225605e-01 -1.01606870e+00 5.59067547e-01
9.35872555e-01 -1.05456233e+00 -9.20470536e-01 1.76374540e-01
1.33622572e-01 -6.38952613e-01 6.82983518e-01 -8.80108654e-01
1.86776686e-02 -3.55924517e-02 -3.73006538e-02 -1.45952177e+00
4.62850295e-02 -1.00299573e+00 -1.21845774e-01 1.25629199e+00
3.83458197e-01 -5.28079212e-01 4.70749319e-01 9.32165086e-01
-2.67314196e-01 -2.90358812e-01 -8.09581459e-01 -1.53325394e-01
-8.28114152e-01 -2.17304826e-02 5.82810044e-01 1.13664174e+00
1.10085475e+00 1.14269614e+00 -3.95645499e-01 1.87105939e-01
-2.47918628e-02 -1.52703583e-01 1.14327145e+00 -1.47526717e+00
-6.15344569e-02 -2.82379836e-01 -4.89942908e-01 -1.06479132e+00
3.62102360e-01 -6.71289325e-01 5.36521554e-01 -1.84747291e+00
-2.75888509e-04 -3.30156356e-01 2.81246692e-01 7.83952713e-01
-9.77087766e-02 -4.09034014e-01 8.29390958e-02 2.07405478e-01
-1.16986191e+00 3.02822948e-01 6.47942662e-01 -1.14141189e-01
-7.63462126e-01 1.09136365e-01 -1.88681230e-01 7.52512634e-01
7.07268178e-01 5.07371454e-03 -2.44747832e-01 3.60871218e-02
2.23840162e-01 9.12859559e-01 -5.04493788e-02 -4.89200830e-01
5.93141377e-01 -8.41217339e-02 -4.58385617e-01 -2.41076425e-01
6.62946224e-01 -4.68652487e-01 1.19577058e-01 1.22925818e-01
-7.68681347e-01 -5.31013384e-02 4.89776880e-02 6.06652975e-01
-3.80080104e-01 -2.96466827e-01 1.81175500e-01 -2.55515814e-01
-9.34248328e-01 -1.76045224e-01 -6.24500453e-01 -1.87307850e-01
8.68327320e-01 -8.88873488e-02 -5.95382333e-01 -9.70835805e-01
-1.12980115e+00 6.13975644e-01 2.00092331e-01 5.90988755e-01
4.17312473e-01 -9.54290390e-01 -5.51471114e-01 -3.27880979e-01
3.75649691e-01 -5.84021118e-03 2.49614730e-01 3.50012422e-01
-1.76460728e-01 3.92737359e-01 -4.68242988e-02 -2.51415312e-01
-1.59844375e+00 2.01947108e-01 3.40395659e-01 -5.63507497e-01
-4.73999888e-01 6.07571721e-01 -1.96627947e-03 -1.06521380e+00
5.52412271e-01 6.38358714e-03 -1.00245845e+00 3.96125555e-01
5.76348066e-01 2.18233570e-01 7.38580227e-02 -1.01607311e+00
-2.56568402e-01 3.50750121e-03 8.53665266e-03 -2.93254346e-01
1.14099503e+00 -5.58306038e-01 -5.36693871e-01 9.27101672e-01
8.18603218e-01 -7.66444057e-02 -8.24941397e-01 -7.33832240e-01
4.30123866e-01 -1.68345906e-02 -6.55718982e-01 -8.65638733e-01
-3.36282104e-01 1.75054431e-01 8.39020833e-02 9.74018395e-01
4.95437682e-01 3.97576272e-01 6.63000941e-01 9.16101336e-01
7.40046740e-01 -1.12805033e+00 2.79943854e-01 1.13942611e+00
1.04132843e+00 -1.22033584e+00 -3.03989470e-01 -5.45379281e-01
-9.33507204e-01 1.17153788e+00 6.59231544e-01 4.96110231e-01
2.40580589e-01 -2.86029398e-01 3.38740706e-01 -5.74862361e-01
-1.07278264e+00 -4.52475846e-01 -3.65986899e-02 6.69296086e-01
4.18595582e-01 2.13764608e-01 -1.64364591e-01 4.14402753e-01
-3.23113918e-01 -4.46694255e-01 6.61615908e-01 9.05179024e-01
-5.88694692e-01 -1.26173830e+00 -9.03830230e-02 3.11041743e-01
5.51721007e-02 2.54177243e-01 -8.41961026e-01 6.26764476e-01
-4.02707130e-01 1.47489929e+00 9.66834873e-02 -6.52221441e-01
9.21412885e-01 3.82534325e-01 1.43109053e-01 -1.26092601e+00
-1.12585354e+00 -4.13807988e-01 9.99947667e-01 -5.08832395e-01
-4.28053200e-01 -6.26230299e-01 -1.54450345e+00 -2.27167636e-01
-4.94053394e-01 5.38836122e-01 5.05324066e-01 9.57865119e-01
1.00604333e-01 4.50150222e-01 6.56796336e-01 -7.90772080e-01
-2.03243375e-01 -1.22950900e+00 -3.51289928e-01 6.91075325e-01
1.27554521e-01 -5.64769983e-01 -1.32238269e-01 5.12176305e-02] | [12.609683990478516, 7.99213171005249] |
2f072b09-35f0-42fe-8ab3-51708ec9cd07 | scaling-to-many-languages-with-a-triaged | 2104.02125 | null | https://arxiv.org/abs/2104.02125v3 | https://arxiv.org/pdf/2104.02125v3.pdf | SpeakerStew: Scaling to Many Languages with a Triaged Multilingual Text-Dependent and Text-Independent Speaker Verification System | In this paper, we describe SpeakerStew - a hybrid system to perform speaker verification on 46 languages. Two core ideas were explored in this system: (1) Pooling training data of different languages together for multilingual generalization and reducing development cycles; (2) A novel triage mechanism between text-dependent and text-independent models to reduce runtime cost and expected latency. To the best of our knowledge, this is the first study of speaker verification systems at the scale of 46 languages. The problem is framed from the perspective of using a smart speaker device with interactions consisting of a wake-up keyword (text-dependent) followed by a speech query (text-independent). Experimental evidence suggests that training on multiple languages can generalize to unseen varieties while maintaining performance on seen varieties. We also found that it can reduce computational requirements for training models by an order of magnitude. Furthermore, during model inference on English data, we observe that leveraging a triage framework can reduce the number of calls to the more computationally expensive text-independent system by 73% (and reduce latency by 59%) while maintaining an EER no worse than the text-independent setup. | ['Ignacio Lopez Moreno', 'Quan Wang', 'Jason Pelecanos', 'Roza Chojnacka'] | 2021-04-05 | null | null | null | null | ['text-independent-speaker-verification'] | ['speech'] | [-7.27638183e-03 -1.82102658e-02 -6.21755868e-02 -8.09590340e-01
-1.47165048e+00 -8.24357033e-01 4.85499859e-01 -1.59399062e-01
-4.38446552e-01 2.90558428e-01 7.41996095e-02 -9.65127885e-01
2.44329333e-01 -6.76425546e-02 -6.80810153e-01 -3.95258665e-01
-6.16179369e-02 6.22993112e-01 1.52497679e-01 -2.32154801e-01
-1.09115340e-01 5.07088602e-01 -1.75385153e+00 1.72981337e-01
6.17160380e-01 6.93923652e-01 8.03210959e-02 8.53631735e-01
2.04181876e-02 3.13758403e-01 -8.80506039e-01 -5.01145005e-01
2.80260950e-01 -1.01989239e-01 -8.39878321e-01 -1.85983449e-01
8.94881368e-01 -1.69977292e-01 -7.35473037e-02 7.48378932e-01
8.64907086e-01 -3.74757983e-02 8.08231533e-02 -1.37569928e+00
-2.38053054e-01 1.06503904e+00 -2.08882079e-01 1.61395669e-01
5.39914608e-01 8.58745500e-02 7.45630264e-01 -8.30891311e-01
3.93421501e-01 1.34789002e+00 6.16479993e-01 9.09061491e-01
-1.37321901e+00 -1.00796545e+00 2.96509951e-01 -2.15311870e-01
-1.82273567e+00 -1.43369818e+00 2.32496962e-01 -2.13287964e-01
1.31075442e+00 5.24377882e-01 1.17802851e-01 1.03038156e+00
-3.12524915e-01 5.90554178e-01 9.25094664e-01 -6.32608056e-01
2.54575551e-01 7.47062564e-01 2.28572801e-01 7.45232999e-01
8.29744190e-02 -9.37674344e-02 -1.20205784e+00 -3.02395433e-01
-1.17275722e-01 -6.46225154e-01 -1.82664707e-01 1.12453133e-01
-1.04103875e+00 5.17507195e-01 -3.44760805e-01 2.11377665e-01
8.46830159e-02 -2.38108501e-01 5.32535672e-01 7.38989055e-01
3.97358447e-01 2.38382712e-01 -7.26534784e-01 -3.68557751e-01
-1.25779903e+00 7.43546486e-02 1.25881696e+00 1.24600101e+00
6.57384098e-01 2.39836216e-01 3.55285197e-01 6.75013661e-01
6.05041385e-01 8.53219986e-01 7.39806235e-01 -3.81981999e-01
8.03187490e-01 2.89444089e-01 -2.24407658e-01 2.44430290e-03
-3.78626347e-01 -2.28260964e-01 -1.03172414e-01 3.58153954e-02
4.63798434e-01 -4.98834431e-01 -7.91496694e-01 1.97631574e+00
3.89098078e-01 1.05881497e-01 2.00126141e-01 4.96195555e-01
6.63799703e-01 4.85661477e-01 6.60336167e-02 -2.87083033e-02
1.42574263e+00 -7.74619758e-01 -4.88115251e-01 -3.25279087e-01
1.10162771e+00 -1.08117771e+00 1.06107283e+00 3.87956470e-01
-1.03675652e+00 -4.15201455e-01 -1.36940753e+00 -1.18234791e-01
-5.15507698e-01 1.00851268e-01 5.69013894e-01 1.56421244e+00
-1.69305468e+00 -7.70949498e-02 -1.03898537e+00 -6.80908859e-01
-1.98263705e-01 6.05350077e-01 -4.63214487e-01 1.67926595e-01
-9.01776910e-01 8.80905092e-01 5.63738570e-02 -1.56548396e-01
-8.75466466e-01 -6.96535945e-01 -9.48678672e-01 -5.11002876e-02
1.90622792e-01 -2.56924927e-01 1.46966398e+00 -6.76408887e-01
-1.84244072e+00 8.75754237e-01 -7.63444364e-01 -4.98214483e-01
3.34118396e-01 -1.50016487e-01 -1.00255024e+00 -2.73131400e-01
-8.85179639e-02 3.82028639e-01 6.12461448e-01 -8.89730334e-01
-6.65817976e-01 -5.47770798e-01 -2.68878728e-01 1.91245005e-01
-5.69935799e-01 4.59360749e-01 -6.32501602e-01 -1.41499326e-01
1.35020941e-01 -1.16795504e+00 3.33660245e-01 -3.47746581e-01
-4.56865638e-01 -1.19222537e-01 9.87248421e-01 -1.12507665e+00
1.30983651e+00 -2.34666991e+00 -1.25620797e-01 2.74155080e-01
-1.39518529e-01 2.59435058e-01 -5.74079230e-02 4.02319342e-01
6.79850951e-02 2.36396074e-01 -5.90001903e-02 -9.38503861e-01
1.09834105e-01 -3.93340960e-02 -3.87284398e-01 4.63697642e-01
-1.40018746e-01 5.61761200e-01 -2.06410989e-01 -5.17587125e-01
-1.29319817e-01 5.14833450e-01 -5.06354809e-01 -3.28732580e-02
9.77668911e-02 5.72688393e-02 -3.69506255e-02 7.83704758e-01
6.66633010e-01 2.37853959e-01 3.20846260e-01 3.14351737e-01
-2.84488291e-01 9.48304772e-01 -1.41969705e+00 1.86014950e+00
-9.02975619e-01 7.68584430e-01 7.20331430e-01 -4.86462027e-01
8.39730859e-01 6.62442684e-01 -2.33220562e-01 -4.72247839e-01
-2.49392837e-01 4.37797755e-01 -3.78902480e-02 -2.96010703e-01
5.30518949e-01 2.24298425e-03 -2.41118416e-01 6.89639986e-01
2.00801089e-01 -1.59303471e-01 -1.22768078e-02 1.71087056e-01
7.93427467e-01 -1.95529938e-01 -1.64472163e-01 -3.85916084e-01
5.89141250e-01 -3.76549065e-01 4.73068684e-01 7.00829208e-01
-2.36535490e-01 -1.23511534e-02 2.08812151e-02 -4.23056595e-02
-6.70514226e-01 -1.07717621e+00 -1.97652489e-01 1.55089843e+00
-2.68490404e-01 -5.74955821e-01 -9.52048481e-01 -5.01444936e-01
1.11245677e-01 9.49007452e-01 -3.76316160e-02 1.59019575e-01
-5.55428147e-01 -4.32431310e-01 1.30388832e+00 4.05136436e-01
2.65801162e-01 -3.56097162e-01 -1.65318951e-01 8.03019945e-03
-3.55085135e-02 -1.31104708e+00 -7.71576345e-01 3.71662945e-01
-7.23345935e-01 -3.28431398e-01 -3.46207470e-01 -9.69393730e-01
3.91908228e-01 5.47523191e-03 9.23332691e-01 -4.97614630e-02
-1.67244710e-02 3.02741557e-01 7.39184842e-02 -5.88928819e-01
-7.86206007e-01 4.23932672e-01 5.23056567e-01 -1.37431517e-01
6.08958662e-01 -4.00511652e-01 -9.99023858e-03 5.45398057e-01
-4.78154540e-01 -1.09564938e-01 4.12094891e-01 4.65534031e-01
-3.10043097e-02 -2.18498528e-01 6.20985508e-01 -4.90323901e-01
4.31658745e-01 -7.87380934e-02 -8.00694346e-01 5.44004977e-01
-7.53394306e-01 2.09642634e-01 3.58908266e-01 -5.09107769e-01
-1.21412396e+00 2.62357354e-01 -2.75896698e-01 8.97956714e-02
-1.44604936e-01 1.88521653e-01 -4.72900957e-01 -2.21545547e-01
6.64987683e-01 4.90544736e-01 8.84435698e-03 -6.03934407e-01
2.92274833e-01 1.34371388e+00 4.53725100e-01 -4.94759023e-01
6.64731741e-01 2.69777894e-01 -4.95459944e-01 -1.19660020e+00
-1.66098893e-01 -6.36893988e-01 -4.70705271e-01 3.23002636e-02
5.68407774e-01 -1.35109925e+00 -8.91590953e-01 6.58521473e-01
-8.31467628e-01 -5.10415375e-01 1.57122865e-01 5.95699012e-01
-2.11531550e-01 2.40618825e-01 -6.61467731e-01 -1.04088056e+00
-4.37939972e-01 -1.26914465e+00 1.32957590e+00 1.19431712e-01
-2.40627363e-01 -8.05939138e-01 -7.22645745e-02 6.18880570e-01
6.65587008e-01 -6.38167024e-01 4.64028984e-01 -1.10918200e+00
-3.37642908e-01 -4.25693035e-01 2.18289956e-01 2.82750428e-01
-6.37439936e-02 -1.32350802e-01 -1.52397311e+00 -5.93602598e-01
2.79052779e-02 -2.69309998e-01 3.55631143e-01 1.73317641e-02
5.81117809e-01 -1.47415757e-01 -5.04957974e-01 5.60027063e-01
9.37192738e-01 1.62293047e-01 6.81266114e-02 5.24271792e-03
4.30642188e-01 4.66470003e-01 7.56704248e-03 4.74224500e-02
6.70770347e-01 1.01258910e+00 -3.18022609e-01 -5.38400523e-02
-1.70634851e-01 -2.81790167e-01 8.03771853e-01 1.02793491e+00
4.29630876e-01 -2.59821832e-01 -1.25901639e+00 5.38291872e-01
-1.19136751e+00 -7.45779753e-01 2.29196519e-01 2.55937600e+00
8.94705892e-01 2.07317069e-01 3.25457156e-01 1.32464483e-01
7.65364289e-01 -1.90083891e-01 -3.97200167e-01 -6.41893089e-01
-6.77110702e-02 1.73719198e-01 5.54611802e-01 1.02369392e+00
-7.67949164e-01 9.98314023e-01 6.47005177e+00 4.48290288e-01
-1.61024272e+00 3.45952660e-01 4.28923637e-01 -3.30567181e-01
-8.40165988e-02 4.18729782e-02 -1.49029148e+00 2.48694256e-01
1.70449531e+00 -4.11216348e-01 5.88247597e-01 9.34888303e-01
-6.35601953e-02 8.62369910e-02 -1.42686510e+00 9.10728037e-01
4.89715606e-01 -7.60065913e-01 -1.45646304e-01 2.54665762e-01
3.62023592e-01 3.99428874e-01 3.48481834e-02 4.37226504e-01
3.06368887e-01 -7.30546653e-01 1.07583022e+00 7.15973899e-02
8.91154408e-01 -5.13309538e-01 4.56809968e-01 4.62743402e-01
-1.26094937e+00 8.01029727e-02 1.68258920e-01 2.07094133e-01
2.85935223e-01 2.33064368e-01 -1.20476103e+00 4.32385057e-01
6.29221678e-01 1.75768416e-03 -7.05312788e-01 6.96332037e-01
4.16039824e-02 9.30378199e-01 -6.90886021e-01 -7.82737881e-03
-1.84619159e-01 4.30478930e-01 5.26761174e-01 1.46425629e+00
4.34125930e-01 -4.06819701e-01 4.44718935e-02 3.37571114e-01
8.27635266e-03 1.70135215e-01 -3.84258032e-01 -7.87288323e-03
8.21330905e-01 9.43987787e-01 -3.79519522e-01 -5.13133883e-01
-3.74755591e-01 1.01012063e+00 1.73328854e-02 3.91555101e-01
-5.86774588e-01 -4.02984351e-01 6.57116115e-01 3.29130664e-02
1.25578523e-01 -3.11986446e-01 -2.95234770e-01 -1.15235126e+00
2.65103281e-01 -1.09773636e+00 3.70367646e-01 -3.31390321e-01
-7.77136385e-01 8.01087439e-01 -1.31646767e-01 -6.85722172e-01
-5.26434481e-01 -3.59456450e-01 -3.50325108e-01 1.37669182e+00
-1.19771111e+00 -1.10359251e+00 7.69119561e-02 5.66954613e-01
5.86757541e-01 -4.20115799e-01 1.22627866e+00 4.94383067e-01
-6.18227303e-01 1.25052083e+00 -1.81696620e-02 2.64051966e-02
9.09108520e-01 -1.16296053e+00 9.48720396e-01 1.07070911e+00
3.66120756e-01 1.13148189e+00 6.39244854e-01 -4.66275275e-01
-1.77575505e+00 -6.36957288e-01 1.43755996e+00 -5.58806121e-01
4.74728316e-01 -1.06498659e+00 -7.81960905e-01 9.23486054e-01
2.06027001e-01 -3.75498563e-01 8.46047938e-01 6.07632995e-01
-7.47646511e-01 -3.54577154e-01 -1.22885406e+00 5.22261024e-01
8.46555769e-01 -1.21909738e+00 -2.36751720e-01 2.99986362e-01
7.87230909e-01 -5.18800676e-01 -7.13151753e-01 -3.14630494e-02
7.26681352e-01 -6.47600114e-01 6.35179877e-01 -2.96228677e-01
-5.73443890e-01 -2.31627017e-01 -4.21506673e-01 -9.97233331e-01
3.23993891e-01 -1.05823708e+00 5.68194129e-02 1.66955745e+00
1.05418885e+00 -1.04516041e+00 6.21248305e-01 8.40170145e-01
-1.73529327e-01 -1.37229994e-01 -1.33459163e+00 -1.02790201e+00
-3.47710475e-02 -6.95811689e-01 8.44316721e-01 8.12183380e-01
1.90724209e-01 3.37561011e-01 -6.77235425e-02 7.06435919e-01
3.17956984e-01 -2.88920432e-01 9.19601440e-01 -8.65077674e-01
-4.43849683e-01 -3.17258984e-01 -4.34292495e-01 -1.05662596e+00
2.65341967e-01 -1.02236903e+00 1.88319355e-01 -8.37714612e-01
-9.67655852e-02 -4.48764026e-01 1.05494848e-02 6.05000854e-01
1.32443890e-01 -6.62901029e-02 8.22154954e-02 -1.20003372e-01
-2.75074035e-01 -5.64815402e-02 1.92243800e-01 -1.44720733e-01
-3.80797863e-01 1.73477873e-01 -7.60042906e-01 3.94201577e-01
7.04172969e-01 -3.79875541e-01 -4.82168168e-01 -5.88458776e-01
-5.83992787e-02 4.87899268e-03 -6.60909787e-02 -1.12048101e+00
5.67651629e-01 4.25287127e-01 -5.46403974e-02 -3.89225096e-01
5.19453824e-01 -5.29933929e-01 7.25069270e-02 3.51906955e-01
-4.15302813e-01 3.37360680e-01 7.63809979e-01 1.20049283e-01
-6.54513240e-02 6.72307909e-02 4.52528805e-01 3.93277228e-01
-5.34233153e-01 -2.64219344e-01 -5.90853751e-01 -9.49610397e-02
7.32675016e-01 6.18017949e-02 -2.87193984e-01 -4.27178502e-01
-5.99890351e-01 1.01607367e-01 4.30220515e-01 6.07777715e-01
4.59928550e-02 -7.89983630e-01 -7.40201235e-01 6.17037356e-01
1.43362582e-01 -5.48474014e-01 1.55133560e-01 5.45743644e-01
-3.55998725e-01 7.65329301e-01 4.04478103e-01 -8.40725899e-01
-1.93966007e+00 2.51744509e-01 5.40150404e-01 1.03169091e-01
-2.13117868e-01 1.29338837e+00 -9.78054628e-02 -9.19293821e-01
6.08551741e-01 -2.62682915e-01 5.31422079e-01 -4.21384461e-02
7.94570684e-01 2.06005022e-01 6.94874227e-01 -7.98797786e-01
-7.74401605e-01 2.72482425e-01 -3.86313289e-01 -8.01042974e-01
1.02489030e+00 -2.09585875e-01 1.19147182e-01 6.05917335e-01
1.11038351e+00 5.28712571e-01 -7.56555200e-01 -4.03056443e-01
1.47582322e-01 -7.09179342e-02 2.12426528e-01 -1.00430453e+00
-6.64841950e-01 6.83823287e-01 8.60315204e-01 2.66640276e-01
9.32091355e-01 1.74393162e-01 6.31232858e-01 5.15239298e-01
4.84316260e-01 -1.18647099e+00 -6.71353340e-01 2.26341426e-01
5.50632417e-01 -1.09685838e+00 -2.50562847e-01 -3.60712081e-01
-5.31376481e-01 7.76031435e-01 4.75044966e-01 8.75784874e-01
5.99194586e-01 7.89224327e-01 5.11645913e-01 1.82367861e-02
-9.27495241e-01 1.45604074e-01 2.26960495e-01 4.21998322e-01
6.74463868e-01 4.09030795e-01 2.07398117e-01 3.43149185e-01
-6.25478268e-01 -3.91150624e-01 2.38757491e-01 1.00775373e+00
-1.71558812e-01 -1.24201179e+00 -5.40487051e-01 1.25931323e-01
-4.32369769e-01 -3.24375689e-01 -3.89859736e-01 7.50399590e-01
-1.07326537e-01 1.52580571e+00 -6.98262900e-02 -4.63832974e-01
2.50739634e-01 8.60232353e-01 1.55128464e-01 -6.30857944e-01
-8.70441496e-01 -2.12462451e-02 4.40181643e-01 -4.18436974e-01
5.69838583e-02 -1.07104623e+00 -8.85486901e-01 -3.86561185e-01
-6.41064167e-01 1.90637454e-01 1.41378701e+00 8.66819203e-01
5.98279834e-01 1.80415809e-01 5.78259408e-01 -5.28251886e-01
-7.87676573e-01 -1.01723361e+00 -4.88965780e-01 -1.66222572e-01
3.92149359e-01 -5.84597886e-02 -5.57561338e-01 1.47187799e-01] | [14.30822467803955, 6.342647075653076] |
35940089-a02b-4b7f-9ade-f35cadcebf8a | natural-image-stitching-with-the-global | null | null | https://www.cmlab.csie.ntu.edu.tw/project/stitching-wGSP/ | https://www.cmlab.csie.ntu.edu.tw/project/stitching-wGSP/ECCV-2016-NISwGSP.pdf | Natural Image Stitching with the Global Similarity Prior | This paper proposes a method for stitching multiple images together so that the stitched image looks as natural as possible. Our method adopts the local warp model and guides the warping of each image with a grid mesh. An objective function is designed for specifying the desired characteristics of the warps. In addition to good alignment and minimal local distortion, we add a global similarity prior in the objective function. This prior constrains the warp of each image so that it resembles a similarity transformation as a whole. The selection of the similarity transformation is crucial to the naturalness of the results. We propose methods for selecting the proper scale and rotation for each image. The warps of all images are solved together for minimizing the distortion globally. A comprehensive evaluation shows that the proposed method consistently outperforms several state-of-the-art methods, including AutoStitch, APAP, SPHP and ANNAP. | ['Yu-Sheng Chen; Yung-Yu Chuang'] | 2016-10-01 | null | null | null | european-conference-on-computer-vision-2016 | ['image-stitching'] | ['computer-vision'] | [ 4.35848296e-01 -3.09014022e-01 -1.12001635e-01 1.27728758e-02
-3.63680124e-01 -6.41367376e-01 7.42886662e-01 -3.77453625e-01
-1.51004657e-01 2.87618279e-01 2.49845728e-01 2.16233581e-01
-3.72158848e-02 -4.90121216e-01 -4.98216569e-01 -9.94957983e-01
2.80429214e-01 2.59786069e-01 6.85684264e-01 -4.21228349e-01
6.87563121e-01 5.70234776e-01 -1.14917541e+00 -9.11467299e-02
9.52481329e-01 7.74636745e-01 3.45420033e-01 5.41158736e-01
2.96946973e-01 1.18302926e-01 -6.00439429e-01 -3.62882793e-01
6.80175126e-01 -5.42069376e-01 -8.37789595e-01 5.53435922e-01
5.84935129e-01 -2.16528922e-01 -1.47149652e-01 1.10838628e+00
3.55834007e-01 4.08481658e-01 5.91574907e-01 -1.22159851e+00
-3.53004575e-01 9.52482447e-02 -1.11403084e+00 1.00670028e-02
3.82240504e-01 1.25853762e-01 7.21983612e-01 -5.83962679e-01
9.18427527e-01 1.10591865e+00 5.77522337e-01 8.13098401e-02
-1.33092356e+00 -6.04543924e-01 -3.33850801e-01 -1.36874184e-01
-1.41351092e+00 -3.22385252e-01 1.10914087e+00 -2.73919255e-01
4.58423436e-01 4.38974082e-01 6.24189734e-01 8.62402678e-01
9.38014984e-01 2.56851494e-01 9.65314209e-01 -6.56844974e-01
4.25367467e-02 -1.72854304e-01 -2.07102016e-01 4.59279031e-01
1.15535326e-01 2.03197688e-01 -4.92551982e-01 -3.20931554e-01
1.19765282e+00 -3.28084826e-02 -2.15079606e-01 -5.12481749e-01
-1.56268191e+00 6.28339231e-01 3.92743409e-01 4.88539457e-01
-1.64034426e-01 1.98759571e-01 1.09703042e-01 1.09333016e-01
3.77304316e-01 6.32416487e-01 1.22889780e-01 1.10390037e-01
-1.22852385e+00 2.72442371e-01 3.59363198e-01 9.23902512e-01
7.42300510e-01 1.15282029e-01 1.64855435e-01 7.75596499e-01
1.42852396e-01 5.46961010e-01 3.98426622e-01 -1.15297127e+00
3.76638949e-01 1.93250135e-01 1.70831829e-01 -1.51230204e+00
5.03576174e-02 -7.32349828e-02 -6.97841525e-01 4.33134824e-01
7.70048574e-02 6.11794321e-03 -9.96447921e-01 1.26832867e+00
4.02677149e-01 4.57976550e-01 -2.63350815e-01 9.34995353e-01
3.05847198e-01 9.68156457e-01 -4.80132729e-01 -3.20172668e-01
1.26093209e+00 -9.93766308e-01 -9.80277061e-01 -4.34063047e-01
-4.23400104e-02 -1.51099324e+00 8.03480566e-01 3.10997993e-01
-1.10673082e+00 -6.40208781e-01 -1.38092268e+00 1.26692057e-01
1.47233009e-01 -3.16933393e-01 9.60531682e-02 3.35766613e-01
-9.73095179e-01 7.81797171e-01 -7.79242992e-01 -3.89710546e-01
-3.48785967e-01 3.14946264e-01 -3.53374422e-01 2.70001620e-01
-8.86491776e-01 1.06902444e+00 4.35647249e-01 -1.66977242e-01
-5.53714216e-01 -3.32026154e-01 -6.67556047e-01 -2.03377917e-01
1.65458262e-01 -5.20331860e-01 9.95467246e-01 -1.16191721e+00
-1.78721929e+00 8.78132105e-01 -1.29292812e-02 -1.60043329e-01
5.18965781e-01 -6.82822466e-02 -2.98955590e-01 2.01916754e-01
-1.11208610e-01 6.73716366e-01 1.54320097e+00 -1.75209200e+00
-4.90143538e-01 6.19528368e-02 -5.25592804e-01 4.63321447e-01
-1.40613094e-01 2.45425954e-01 -7.27067709e-01 -1.17326558e+00
2.64523983e-01 -1.15702796e+00 -2.79777020e-01 -2.32544631e-01
-4.23440844e-01 4.75387573e-01 1.41002429e+00 -7.16788113e-01
1.31323206e+00 -2.31582785e+00 3.98405910e-01 4.91992980e-01
4.47604917e-02 -1.83128510e-02 -2.64975518e-01 6.90730989e-01
-1.06167004e-01 -1.49855748e-01 -2.75466532e-01 -4.69694406e-01
-4.44561392e-01 4.03801858e-01 -5.86211145e-01 7.74574578e-01
-1.14950448e-01 5.27667224e-01 -7.49852359e-01 -6.79196477e-01
2.96971411e-01 5.42415142e-01 -3.69188786e-01 4.02874798e-01
2.12207049e-01 4.57381129e-01 -2.67109543e-01 2.04509452e-01
9.75322723e-01 2.92520989e-02 1.21009415e-02 -4.15976524e-01
-4.83028233e-01 -2.25812405e-01 -1.18980980e+00 1.53825521e+00
-4.36293259e-02 6.21200919e-01 -5.97706111e-03 -5.70127428e-01
1.32746792e+00 3.21749747e-01 7.51220644e-01 -2.30066538e-01
4.40628231e-01 2.88691103e-01 -1.44775555e-01 -3.46170872e-01
7.63098121e-01 -2.54610777e-01 1.03671633e-01 4.80430186e-01
-8.95721465e-02 -8.87949824e-01 6.69426396e-02 -7.65607413e-03
4.64779824e-01 -1.27201655e-03 4.36970860e-01 -6.90137088e-01
4.22661483e-01 3.95417772e-02 5.24239302e-01 3.53026807e-01
-3.18526290e-02 1.08554971e+00 1.00265123e-01 -5.46011865e-01
-1.45995963e+00 -8.30193043e-01 -8.26354325e-02 5.17793298e-01
8.27320516e-01 -1.81408376e-01 -1.02219629e+00 -3.32020104e-01
-3.06962699e-01 4.75216508e-01 -6.46989584e-01 -5.65071926e-02
-9.64334011e-01 -2.27036804e-01 2.39296854e-01 7.32120275e-02
6.88816309e-01 -8.33949924e-01 -6.74553394e-01 3.85606214e-02
-3.37670505e-01 -1.15765941e+00 -1.42502379e+00 -2.70525187e-01
-1.06503415e+00 -9.16536808e-01 -8.50928903e-01 -1.18021548e+00
9.92295206e-01 7.33473659e-01 5.96552849e-01 5.21801859e-02
-1.49305044e-02 -5.78944106e-03 -4.50740129e-01 1.64865747e-01
-5.42242289e-01 -2.61361182e-01 1.10270731e-01 2.28256449e-01
-4.18480337e-01 -6.33643389e-01 -6.56299472e-01 8.78593743e-01
-1.23156869e+00 2.38078356e-01 2.20877588e-01 7.71365762e-01
6.46634758e-01 3.38441283e-01 -4.46677625e-01 -2.61222601e-01
7.01504588e-01 1.44070297e-01 -6.30019248e-01 2.19482869e-01
-4.47897583e-01 1.43588632e-01 5.85380733e-01 -9.15760279e-01
-1.06487143e+00 1.90118961e-02 3.30028415e-01 -5.66204906e-01
2.19552800e-01 8.93815085e-02 1.65135279e-01 -6.41732156e-01
6.43568873e-01 2.96351761e-01 3.70352983e-01 -2.99033582e-01
2.70006478e-01 3.30265820e-01 8.77497971e-01 -5.13996542e-01
1.31964099e+00 7.20923185e-01 6.36191219e-02 -9.97467220e-01
-2.32136369e-01 -3.52753520e-01 -8.21419537e-01 -3.40396434e-01
8.05672288e-01 -2.15166122e-01 -4.27171528e-01 7.19091654e-01
-1.18040526e+00 -9.14676767e-03 -1.82738841e-01 3.87113869e-01
-7.18769610e-01 8.20090294e-01 -3.54700655e-01 -4.87286061e-01
-4.13070172e-01 -1.54751790e+00 9.02120709e-01 2.86867738e-01
-4.42061543e-01 -1.06101012e+00 6.10397696e-01 1.60111114e-01
4.51628387e-01 5.82706034e-01 4.41743761e-01 -1.35934904e-01
-3.57693702e-01 -1.42735451e-01 1.29034325e-01 2.22821042e-01
5.10129511e-01 6.54491544e-01 -4.83760029e-01 -6.05898857e-01
2.99141705e-01 1.21004857e-01 4.84526306e-01 6.09622717e-01
4.69442487e-01 -6.26744807e-01 -4.31153238e-01 9.54733670e-01
1.49172711e+00 6.52360260e-01 9.21219289e-01 6.11256421e-01
4.36988056e-01 5.13577461e-01 8.25815737e-01 3.14823300e-01
-2.14535981e-01 1.08143640e+00 4.25510228e-01 -2.56767780e-01
-2.95914244e-02 -2.37182334e-01 3.24059635e-01 7.54213333e-01
-2.69922137e-01 -1.85782418e-01 -8.02367508e-01 3.73148501e-01
-1.77753985e+00 -9.97000515e-01 2.51270115e-01 2.46864891e+00
7.69882917e-01 -9.62069333e-02 -2.84709781e-02 1.77772790e-01
1.06595242e+00 7.01425076e-01 -4.69627604e-02 -4.98207629e-01
-7.86468387e-02 -1.24761425e-02 6.66952431e-01 1.01841629e+00
-1.01864552e+00 1.00277936e+00 7.31688404e+00 9.19447303e-01
-1.30024469e+00 -4.03664917e-01 3.49832267e-01 3.14352334e-01
-4.00334954e-01 3.64831120e-01 -5.62309325e-01 5.29042661e-01
2.81030029e-01 -1.81701407e-01 5.54252148e-01 3.77065301e-01
3.44298780e-01 -6.77968115e-02 -5.15604615e-01 8.75079274e-01
8.66503567e-02 -1.28772998e+00 1.06229506e-01 -4.62610386e-02
1.07433319e+00 -3.56378883e-01 2.66260356e-01 -7.13640094e-01
2.58011490e-01 -7.49437153e-01 8.77896667e-01 3.40339333e-01
7.64882207e-01 -9.30598497e-01 5.63615322e-01 -1.42686022e-02
-1.23733675e+00 2.58490384e-01 -1.64068669e-01 4.11312044e-01
4.94504869e-01 3.21739554e-01 -5.87078810e-01 3.90023470e-01
4.58302796e-01 4.65452313e-01 -1.62982300e-01 9.70797420e-01
-1.97860882e-01 1.63557127e-01 -4.31503057e-01 4.13055837e-01
3.62222791e-01 -6.72429681e-01 9.29281771e-01 1.12946236e+00
6.63953960e-01 1.76203176e-01 2.61750132e-01 5.49663961e-01
3.10732365e-01 4.87740450e-02 -5.77555120e-01 8.32732841e-02
6.30310833e-01 1.04332411e+00 -8.58053327e-01 -9.17314515e-02
-4.99051102e-02 1.23619473e+00 -3.80316615e-01 4.01059031e-01
-7.33225703e-01 -4.71279234e-01 6.64670587e-01 2.01625764e-01
2.56255627e-01 -4.66273695e-01 -3.35069060e-01 -9.61683929e-01
-1.34434015e-01 -1.07619774e+00 2.43158698e-01 -8.32286417e-01
-9.15178776e-01 7.34237075e-01 1.77558139e-01 -1.82689750e+00
-1.95815757e-01 -1.74403816e-01 -9.60600078e-01 7.26480186e-01
-1.09893024e+00 -1.13609469e+00 -4.43926215e-01 6.72879934e-01
5.49224913e-01 2.47895960e-02 2.66930044e-01 -3.58337909e-01
-3.89429778e-01 2.46025190e-01 1.15585916e-01 -1.55733660e-01
1.17078698e+00 -6.65113688e-01 6.52110398e-01 1.29646242e+00
-2.10759670e-01 7.15050638e-01 1.35388970e+00 -8.64563167e-01
-1.18180943e+00 -4.86654907e-01 6.92424059e-01 5.84078655e-02
7.26159513e-01 1.22856759e-01 -9.34613526e-01 5.78732729e-01
7.89810121e-01 -9.06913728e-02 1.17558807e-01 -6.87754512e-01
-3.35936248e-01 -1.66441262e-01 -1.28502357e+00 7.52187371e-01
5.60569823e-01 -1.28455147e-01 -6.23261929e-01 -3.26846689e-02
7.27665067e-01 -7.19801307e-01 -8.47160399e-01 1.36892676e-01
5.73291540e-01 -1.01187694e+00 1.17176306e+00 1.66575722e-02
5.68812788e-01 -5.53978026e-01 1.44124731e-01 -1.52042139e+00
-7.25290477e-01 -1.32641542e+00 1.44229949e-01 1.01421773e+00
-6.69970214e-02 -7.00341463e-01 5.23582339e-01 4.09746975e-01
3.47984731e-02 -3.36047202e-01 -7.72445679e-01 -9.39200878e-01
-2.34112799e-01 3.06617647e-01 5.57618201e-01 1.13482368e+00
1.54365018e-01 -4.43924367e-02 -8.43298674e-01 3.62852514e-01
9.66987312e-01 8.30377862e-02 7.86388516e-01 -8.06240439e-01
-1.89204842e-01 -4.85815436e-01 -4.64565128e-01 -8.54450107e-01
-2.73242861e-01 -1.44657701e-01 1.57475367e-01 -1.04150116e+00
1.56896770e-01 -1.62457705e-01 2.57552803e-01 2.90169269e-01
-1.75252095e-01 2.91745722e-01 3.82864952e-01 6.11585975e-01
2.29838833e-01 3.69993865e-01 1.78386068e+00 6.60638139e-02
-4.64172453e-01 -2.24889576e-01 -2.53584027e-01 5.39019585e-01
8.23655307e-01 -4.25124288e-01 -3.22629273e-01 -5.42839766e-01
-1.91505522e-01 2.11156651e-01 -5.67324348e-02 -8.17142963e-01
3.93468946e-01 -7.66192853e-01 -1.09890318e-02 -5.17168939e-01
4.50426996e-01 -1.05311942e+00 5.57672560e-01 7.12632895e-01
-1.65419817e-01 3.03952307e-01 -6.80486858e-02 3.38556394e-02
-4.60184008e-01 -2.72244126e-01 1.44342804e+00 2.90865839e-01
-3.26594263e-01 2.79943317e-01 -1.41470190e-02 -2.79552281e-01
1.14367247e+00 -6.10591352e-01 -2.59296715e-01 -5.51706731e-01
-8.99924934e-02 -1.44985942e-02 1.05699754e+00 4.13090765e-01
7.40452588e-01 -1.45509958e+00 -5.90447903e-01 3.90422434e-01
-1.98903963e-01 -9.14940313e-02 4.53179739e-02 6.01966798e-01
-9.81345952e-01 3.86961699e-02 -6.63947344e-01 -4.74675775e-01
-1.64679265e+00 6.22457802e-01 1.67180166e-01 -2.38214523e-01
-5.90967238e-01 5.69217384e-01 2.10258603e-01 6.19175285e-02
-3.85131314e-02 -7.91216120e-02 -1.04799993e-01 -2.08982974e-01
5.89647174e-01 4.77835745e-01 -3.23365241e-01 -9.46511567e-01
-2.70057052e-01 1.33025384e+00 -1.44918844e-01 -5.45441806e-01
1.30282307e+00 -2.21099541e-01 -2.75115013e-01 -1.59686118e-01
1.21322882e+00 3.01875114e-01 -1.62113345e+00 -1.95329472e-01
-4.48255062e-01 -1.01210940e+00 2.47496087e-02 -1.72764853e-01
-1.26209176e+00 4.15848285e-01 2.44873956e-01 9.68701318e-02
1.22433567e+00 -4.59451556e-01 9.40024197e-01 -2.25284621e-01
2.79467225e-01 -1.01510036e+00 5.17023504e-01 5.08843362e-01
1.16643870e+00 -7.04151690e-01 4.54666764e-01 -4.81113911e-01
-8.00574541e-01 1.30794168e+00 2.74478614e-01 -6.12419188e-01
2.95373440e-01 6.41150713e-01 2.11310849e-01 8.24241638e-02
-2.50256568e-01 4.92823243e-01 5.11645436e-01 5.31208277e-01
1.23131156e-01 -2.98925191e-01 -4.63438988e-01 -3.40599686e-01
-3.42172951e-01 -2.68605679e-01 4.73708928e-01 9.51663792e-01
-7.33809531e-01 -1.28874743e+00 -1.13444614e+00 -4.21664596e-01
-1.62190467e-01 1.99262977e-01 -4.58827913e-01 6.91565692e-01
-1.95049345e-01 7.53609002e-01 7.89472759e-02 -6.29222393e-01
2.50674516e-01 -3.48073751e-01 4.76151347e-01 6.86220243e-04
-7.52975047e-01 5.71225166e-01 -1.53983057e-01 -5.86254716e-01
-2.87827700e-01 -4.57107306e-01 -1.08983278e+00 -6.89949572e-01
-3.75279605e-01 2.33582426e-02 4.83901560e-01 7.08876312e-01
1.21417873e-01 -1.53727941e-02 1.11345446e+00 -1.28733623e+00
-3.77272666e-01 -5.04318833e-01 -5.74179828e-01 5.02456307e-01
4.32986498e-01 -4.50252146e-01 -4.76632118e-01 4.19777870e-01] | [9.409327507019043, -2.3505759239196777] |
d1673a4a-8cc9-4d21-a529-c1c927abba7a | spatial-and-spectral-deep-attention-fusion | 2002.01626 | null | https://arxiv.org/abs/2002.01626v1 | https://arxiv.org/pdf/2002.01626v1.pdf | Spatial and spectral deep attention fusion for multi-channel speech separation using deep embedding features | Multi-channel deep clustering (MDC) has acquired a good performance for speech separation. However, MDC only applies the spatial features as the additional information. So it is difficult to learn mutual relationship between spatial and spectral features. Besides, the training objective of MDC is defined at embedding vectors, rather than real separated sources, which may damage the separation performance. In this work, we propose a deep attention fusion method to dynamically control the weights of the spectral and spatial features and combine them deeply. In addition, to solve the training objective problem of MDC, the real separated sources are used as the training objectives. Specifically, we apply the deep clustering network to extract deep embedding features. Instead of using the unsupervised K-means clustering to estimate binary masks, another supervised network is utilized to learn soft masks from these deep embedding features. Our experiments are conducted on a spatialized reverberant version of WSJ0-2mix dataset. Experimental results show that the proposed method outperforms MDC baseline and even better than the oracle ideal binary mask (IBM). | ['Jian-Hua Tao', 'Bin Liu', 'Zhengqi Wen', 'Jiangyan Yi', 'Cunhang Fan'] | 2020-02-05 | null | null | null | null | ['deep-attention', 'deep-attention'] | ['computer-vision', 'natural-language-processing'] | [-5.22569381e-02 -4.87935036e-01 2.72440106e-01 -3.46800834e-01
-1.13539422e+00 -3.98057491e-01 2.76232600e-01 -1.02641255e-01
-5.03570855e-01 2.15604961e-01 2.34637767e-01 -2.11796060e-01
-3.73882681e-01 -1.91939279e-01 -4.31225508e-01 -1.26725245e+00
-1.59337267e-01 -5.76820970e-02 -7.23990798e-02 6.85050935e-02
-7.91363046e-02 1.61976039e-01 -1.28831339e+00 2.50042617e-01
1.01624346e+00 1.07430279e+00 5.88007450e-01 6.56988263e-01
8.99006873e-02 4.76287425e-01 -7.93615937e-01 7.77834877e-02
2.13138297e-01 -6.33043528e-01 -2.85586685e-01 4.42656167e-02
-4.15739836e-03 -4.61554900e-02 -4.40822124e-01 1.39426923e+00
9.84562933e-01 4.00186568e-01 6.71137750e-01 -1.18083203e+00
-6.41407549e-01 7.81114221e-01 -7.08604872e-01 3.49588990e-01
-2.05089673e-01 1.56540200e-01 9.90876555e-01 -9.11860824e-01
-3.69662702e-01 1.13667786e+00 4.60692585e-01 2.68280864e-01
-1.13935220e+00 -7.74338484e-01 1.75966457e-01 5.16747534e-01
-1.59265196e+00 -8.79684806e-01 1.10691571e+00 -2.90814787e-01
7.80054510e-01 3.89465362e-01 2.93555945e-01 8.79578948e-01
-5.65036535e-01 8.03333521e-01 1.07518804e+00 -6.29071832e-01
1.66435793e-01 1.73133701e-01 1.49771899e-01 4.00685608e-01
-1.63585022e-01 1.30079433e-01 -3.15808952e-01 6.64966181e-02
3.37280422e-01 -7.24413469e-02 -7.55020618e-01 2.51913220e-01
-1.21061480e+00 6.23278201e-01 4.84858513e-01 4.72969770e-01
-1.48164645e-01 1.37312889e-01 1.46046355e-01 2.13187918e-01
3.54227632e-01 3.44977289e-01 -4.39844877e-01 -8.12727734e-02
-9.19050694e-01 -2.21052825e-01 3.85897547e-01 6.02616370e-01
7.65973091e-01 3.19175661e-01 -2.09548891e-01 1.32002473e+00
5.94259620e-01 7.39792705e-01 8.30012262e-01 -9.65870857e-01
7.54453838e-01 -4.56412463e-03 -1.01039462e-01 -1.32038689e+00
-2.53288090e-01 -7.78977156e-01 -1.07647872e+00 -1.11357562e-01
8.04626867e-02 -3.22745204e-01 -8.98756802e-01 1.65406013e+00
9.26541686e-02 4.55993265e-01 2.12906718e-01 1.36756968e+00
6.86142087e-01 1.03728878e+00 -5.11165500e-01 -2.73047179e-01
1.04741347e+00 -1.10647786e+00 -1.10195112e+00 -2.49337241e-01
2.24630058e-01 -9.82718766e-01 9.19946492e-01 4.12751824e-01
-8.07699978e-01 -7.96598852e-01 -1.32401812e+00 1.90048605e-01
-3.10871780e-01 5.07816672e-01 2.54028648e-01 6.73768222e-01
-7.75235355e-01 3.97806495e-01 -9.74510133e-01 3.22958320e-01
1.71549648e-01 3.15980643e-01 -6.40003234e-02 5.03160805e-03
-1.31599057e+00 4.80192989e-01 4.55748856e-01 5.58669031e-01
-9.13166344e-01 -2.34393165e-01 -7.98898876e-01 2.67932147e-01
1.69719905e-01 1.89255133e-01 8.22319031e-01 -9.73645270e-01
-1.73153830e+00 4.10141021e-01 -9.25686806e-02 -3.45151126e-01
1.13772243e-01 -2.07547247e-01 -7.79887676e-01 3.59862715e-01
8.42720792e-02 4.14273351e-01 1.08299077e+00 -1.46905637e+00
-5.02166450e-01 -1.01638295e-01 -1.76210165e-01 3.41919839e-01
-5.63152850e-01 -7.54425749e-02 -6.46067023e-01 -9.31209743e-01
2.98726439e-01 -5.81039548e-01 -8.89053717e-02 -4.67304617e-01
-6.44868374e-01 3.41908559e-02 7.82964885e-01 -1.01319349e+00
1.29188120e+00 -2.79099536e+00 2.20172033e-01 3.39979082e-01
1.72232732e-01 3.22196305e-01 -3.68966967e-01 -5.09765670e-02
-1.99893802e-01 -3.72304693e-02 -5.15347302e-01 -5.24297893e-01
2.00356886e-01 6.18109619e-03 -2.08894238e-01 6.30748868e-01
1.79753244e-01 4.17170048e-01 -6.44020855e-01 -5.78269601e-01
3.51871312e-01 6.00191474e-01 -5.51437378e-01 1.75149903e-01
2.63493657e-01 2.66263664e-01 -3.45402732e-02 2.22296953e-01
1.10516536e+00 -1.14772528e-01 7.33490586e-02 -3.33912998e-01
-5.39178997e-02 3.16970229e-01 -1.26702666e+00 1.77884507e+00
-6.38302088e-01 7.73220301e-01 4.27102894e-01 -1.53179049e+00
7.19239950e-01 5.55189252e-01 4.10735935e-01 -5.76524019e-01
1.11228406e-01 2.60618091e-01 3.64194810e-01 -4.76398289e-01
-9.86142680e-02 -8.27142596e-02 1.76105678e-01 2.16709957e-01
2.33250812e-01 -2.35553235e-02 -3.30133200e-01 -9.26120393e-03
7.95788467e-01 -4.56761003e-01 -5.29874377e-02 -9.78466049e-02
6.77506268e-01 -3.94971699e-01 9.29766119e-01 4.12923396e-01
-2.61801392e-01 9.45052326e-01 2.60353118e-01 2.68220037e-01
-4.63155985e-01 -1.14293051e+00 -1.42531902e-01 6.66661441e-01
3.75039428e-01 -3.53870600e-01 -9.00117695e-01 -4.98410493e-01
-2.88093537e-01 6.09227180e-01 -2.00612545e-01 -4.02219892e-01
-6.14178717e-01 -1.09995234e+00 6.09932661e-01 3.61677259e-01
6.43657207e-01 -6.53340995e-01 -9.35393199e-02 3.24210256e-01
-6.40875995e-01 -1.03994453e+00 -6.93714321e-01 5.04858255e-01
-2.00267270e-01 -7.67552793e-01 -6.83880270e-01 -9.05870199e-01
4.34657395e-01 6.50053620e-01 3.46706420e-01 -2.59733875e-03
-1.23085342e-01 8.12969208e-02 -5.07970870e-01 -2.08949968e-01
2.45928764e-02 -5.54581322e-02 1.17229722e-01 5.52479565e-01
2.82012105e-01 -7.88990617e-01 -7.00638413e-01 3.11976075e-01
-7.43514895e-01 -1.27517998e-01 6.04100287e-01 9.20397043e-01
3.82556766e-01 8.71798337e-01 5.74513793e-01 -1.97712928e-01
5.43644249e-01 -4.14690763e-01 -4.80633616e-01 6.04655333e-02
-4.07242030e-01 7.70651549e-02 8.18444610e-01 -6.15133762e-01
-1.00349927e+00 -8.70692637e-03 -1.77009434e-01 -6.67210579e-01
-4.58708890e-02 4.87564653e-01 -7.51574814e-01 3.08305413e-01
2.62167990e-01 3.75038087e-01 -8.80271718e-02 -7.62416422e-01
1.39519215e-01 1.16015136e+00 5.14790118e-01 -4.11005110e-01
9.26635325e-01 3.14113975e-01 -5.46164513e-01 -9.87114012e-01
-7.71865964e-01 -5.97683370e-01 -4.93276358e-01 4.15469445e-02
1.10259950e+00 -1.05416131e+00 -5.05240440e-01 6.42664552e-01
-1.19929707e+00 -3.62795502e-01 1.57757595e-01 9.62537348e-01
-1.36243790e-01 2.82620609e-01 -6.15114868e-01 -8.10688198e-01
6.25252649e-02 -1.43429995e+00 9.81403947e-01 1.04564264e-01
2.98455358e-01 -9.06025827e-01 -1.76740274e-01 4.30515647e-01
2.94300765e-01 -5.77474684e-02 8.42394769e-01 -5.99840820e-01
-4.12712038e-01 -6.30899668e-02 -2.45380014e-01 8.35996807e-01
4.94343489e-01 -1.38491765e-01 -1.39032614e+00 -3.40983033e-01
5.01425862e-01 6.59246519e-02 1.10094190e+00 4.89329457e-01
1.58038783e+00 -3.37456524e-01 -2.06380323e-01 9.26078022e-01
1.21956825e+00 4.68323290e-01 4.74743456e-01 -4.39789407e-02
9.15415466e-01 4.19137686e-01 2.52894074e-01 2.01187342e-01
2.18943581e-01 6.69656992e-01 2.07743496e-01 -3.12514275e-01
-2.82503843e-01 -2.25753691e-02 4.43791121e-01 1.68159533e+00
8.84068012e-02 -2.24009365e-01 -7.47011006e-01 4.13137317e-01
-1.66235209e+00 -9.08755720e-01 1.95713621e-02 2.16278076e+00
9.87519026e-01 8.95615965e-02 -3.24195713e-01 6.12158358e-01
1.02348864e+00 3.31712931e-01 -5.08609235e-01 1.57629848e-01
-3.85941386e-01 1.98508605e-01 2.63020307e-01 8.08057964e-01
-1.39480793e+00 5.54068208e-01 5.10873032e+00 1.28064752e+00
-1.36863685e+00 3.96643549e-01 5.42784095e-01 -3.47011805e-01
-1.69802949e-01 -1.61239952e-01 -3.72090846e-01 8.36761177e-01
7.74337053e-01 3.08211207e-01 9.45869505e-01 3.48963439e-01
2.24883661e-01 2.16950610e-01 -7.93695450e-01 1.41984510e+00
1.69103175e-01 -7.41350532e-01 -3.72238308e-01 1.38755649e-01
5.44001937e-01 -1.04350783e-01 3.82094473e-01 3.59965235e-01
2.52948344e-01 -9.90883231e-01 8.21706831e-01 3.52298439e-01
6.39605343e-01 -8.40482175e-01 5.30344725e-01 3.76516402e-01
-1.20527494e+00 -2.25848705e-01 -4.40130562e-01 2.39967376e-01
-8.64469260e-02 9.09065127e-01 -4.05209184e-01 5.28318763e-01
7.24654377e-01 7.68402398e-01 -3.53260785e-01 8.23427558e-01
-2.68389583e-01 9.52063441e-01 -2.64489502e-01 3.35575759e-01
2.13442057e-01 -2.64116853e-01 4.40680355e-01 1.26033437e+00
3.06809783e-01 5.99130541e-02 1.43531263e-01 7.04327881e-01
-2.95386929e-02 -1.07220069e-01 -1.19272873e-01 -5.74288554e-02
6.54899538e-01 1.19021356e+00 -5.59327543e-01 -1.65414661e-01
-4.52408671e-01 1.26429045e+00 1.13080978e-01 7.84460485e-01
-1.14250958e+00 -8.66808534e-01 8.44317019e-01 -3.11386347e-01
4.43320990e-01 -2.82180190e-01 -1.23182207e-01 -1.21053398e+00
8.57763812e-02 -1.03200817e+00 -6.43440858e-02 -5.82217693e-01
-1.23573494e+00 5.75970650e-01 -3.35934758e-01 -1.41689062e+00
2.40903839e-01 -5.69448948e-01 -8.05484533e-01 1.15561879e+00
-1.53419077e+00 -6.51004314e-01 -7.61000887e-02 6.20587707e-01
4.17584807e-01 -3.53000671e-01 6.04112566e-01 6.92077756e-01
-9.82164443e-01 7.47743666e-01 6.47551894e-01 4.53623444e-01
8.24834943e-01 -1.40397859e+00 1.04928026e-02 1.10750186e+00
2.73816288e-01 5.44274926e-01 3.61522704e-01 -1.17387608e-01
-9.75574076e-01 -1.12658417e+00 3.04120272e-01 1.00405194e-01
4.87914562e-01 -5.57037294e-01 -9.37946916e-01 2.83697098e-01
4.31231499e-01 1.34481177e-01 8.16321015e-01 -1.20190091e-01
-3.17106873e-01 -6.36168242e-01 -6.60014331e-01 2.25084543e-01
6.97680950e-01 -8.60997856e-01 -4.78049636e-01 3.16748738e-01
1.10841882e+00 -1.62700683e-01 -4.52121258e-01 2.47840598e-01
1.49940073e-01 -8.34586740e-01 1.26296544e+00 -1.46040678e-01
1.85218781e-01 -7.25374162e-01 -6.49393678e-01 -1.75418806e+00
-3.41943264e-01 -5.21194160e-01 -7.46703893e-02 1.62003231e+00
5.95688045e-01 -7.59896100e-01 9.81730670e-02 -5.98493107e-02
-4.35456038e-01 -3.91663939e-01 -8.97336423e-01 -8.20528805e-01
-7.61151686e-02 -5.17194748e-01 8.08562577e-01 1.21345937e+00
-1.98844016e-01 3.89805436e-01 -3.13612401e-01 8.27095807e-01
6.87296510e-01 1.33913413e-01 3.17267865e-01 -9.10750389e-01
-6.79126441e-01 -5.20381451e-01 -1.58687919e-01 -1.09511328e+00
3.40974599e-01 -1.06102633e+00 4.33364093e-01 -1.32197905e+00
-1.85352176e-01 -5.87464035e-01 -8.60118389e-01 2.09868833e-01
-5.12122869e-01 -2.99794134e-02 1.52803227e-01 2.01416403e-01
-4.66528505e-01 8.34292769e-01 1.04176819e+00 -5.84251404e-01
-1.90702096e-01 1.10906452e-01 -7.15924263e-01 6.60767555e-01
9.60629165e-01 -4.67495292e-01 -4.90874678e-01 -8.46702218e-01
-1.70053050e-01 1.52697461e-02 4.09538567e-01 -1.23947012e+00
3.68410349e-01 -5.37065731e-04 4.29484457e-01 -5.85922718e-01
6.28631175e-01 -8.42863858e-01 -2.90334374e-01 2.32293665e-01
-3.12665969e-01 -4.59928483e-01 -3.67474146e-02 5.46137869e-01
-6.74006939e-01 -2.07557604e-01 8.76100540e-01 7.31088072e-02
-2.45074987e-01 8.90075862e-02 -3.40841860e-01 9.39757079e-02
5.90337336e-01 1.26121968e-01 -1.54507861e-01 -2.78039604e-01
-6.71635449e-01 3.95748824e-01 -2.83595733e-02 2.94664294e-01
7.17493653e-01 -1.60529304e+00 -6.20696187e-01 3.64225715e-01
-1.08320616e-01 1.66475147e-01 3.93906623e-01 9.65071082e-01
-3.02858084e-01 1.93630099e-01 3.55485350e-01 -5.92194080e-01
-9.11151707e-01 7.24393189e-01 6.43677950e-01 2.43001461e-01
-3.87097359e-01 9.11772132e-01 3.28120500e-01 -5.58586836e-01
5.26658058e-01 -3.74769270e-01 -1.32768869e-01 8.91122222e-02
5.86130857e-01 3.48768800e-01 5.03187999e-02 -5.61663091e-01
-5.12048006e-01 5.30610979e-01 2.27204368e-01 -4.77146745e-01
1.35324550e+00 -3.98115098e-01 -2.49561414e-01 5.26325524e-01
1.74607182e+00 3.36115956e-01 -1.20210826e+00 -4.11761910e-01
-2.69076407e-01 -3.71306211e-01 3.94490451e-01 -6.72460377e-01
-1.40020490e+00 1.52257872e+00 8.56500387e-01 3.83864671e-01
1.41664791e+00 -2.23379076e-01 8.80909443e-01 1.11047909e-01
-2.96064109e-01 -1.23921180e+00 2.13400722e-01 3.28128219e-01
7.66775846e-01 -1.29882181e+00 -4.18566346e-01 -1.31091177e-01
-4.74598676e-01 8.42614651e-01 5.70829690e-01 8.20225775e-02
1.08006442e+00 4.22969609e-01 4.01655883e-01 1.18569210e-02
-4.34818298e-01 -3.68705869e-01 2.70149708e-01 6.19086742e-01
2.27762654e-01 3.06404322e-01 1.47186339e-01 8.52134943e-01
-2.31254473e-01 -9.16653454e-01 5.29334955e-02 2.60921240e-01
-4.16645020e-01 -8.43738139e-01 -7.37262309e-01 2.95747876e-01
-4.42801654e-01 -3.75634819e-01 -8.61491784e-02 2.56407082e-01
2.95745879e-01 1.41059899e+00 4.41789888e-02 -6.50306344e-01
1.05829865e-01 -8.84349868e-02 9.27241817e-02 -4.09942746e-01
-1.45215794e-01 5.83904505e-01 -3.28713924e-01 -1.61357969e-01
-4.36051071e-01 -5.76545119e-01 -1.32442844e+00 7.91464150e-02
-5.50169706e-01 5.09486020e-01 6.38646781e-01 8.64077508e-01
1.07350908e-01 8.89354944e-01 1.21291304e+00 -7.79201448e-01
-4.40697193e-01 -1.12847352e+00 -6.72256470e-01 2.74519891e-01
8.56402099e-01 -5.99116266e-01 -7.89918482e-01 1.03048362e-01] | [14.958309173583984, 5.864485740661621] |
49cda9bf-66e0-440c-8177-1a7c0eacab92 | classifying-fonts-and-calligraphy-styles | 1407.2649 | null | http://arxiv.org/abs/1407.2649v1 | http://arxiv.org/pdf/1407.2649v1.pdf | Classifying Fonts and Calligraphy Styles Using Complex Wavelet Transform | Recognizing fonts has become an important task in document analysis, due to
the increasing number of available digital documents in different fonts and
emphases. A generic font-recognition system independent of language, script and
content is desirable for processing various types of documents. At the same
time, categorizing calligraphy styles in handwritten manuscripts is important
for palaeographic analysis, but has not been studied sufficiently in the
literature. We address the font-recognition problem as analysis and
categorization of textures. We extract features using complex wavelet transform
and use support vector machines for classification. Extensive experimental
evaluations on different datasets in four languages and comparisons with
state-of-the-art studies show that our proposed method achieves higher
recognition accuracy while being computationally simpler. Furthermore, on a new
dataset generated from Ottoman manuscripts, we show that the proposed method
can also be used for categorizing Ottoman calligraphy with high accuracy. | ['Alican Bozkurt', 'Pinar Duygulu', 'A. Enis Cetin'] | 2014-07-09 | null | null | null | null | ['font-recognition'] | ['computer-vision'] | [ 2.49542773e-01 -7.76583612e-01 2.52486207e-02 -3.62642944e-01
-2.08215415e-01 -7.63650417e-01 9.15398836e-01 2.02540830e-01
-2.72643924e-01 4.70755041e-01 -5.18340953e-02 -4.24029917e-01
-2.05165371e-01 -8.25537443e-01 -1.30820632e-01 -7.15017974e-01
1.52743205e-01 4.40015584e-01 3.87070566e-01 -3.02064925e-01
7.88589418e-01 1.09666908e+00 -1.71253133e+00 4.85229731e-01
6.97386205e-01 1.02989161e+00 3.31831276e-01 7.29155183e-01
-5.22187173e-01 4.32188779e-01 -8.61190379e-01 -5.79902291e-01
1.69801936e-01 -3.26910913e-01 -4.91981804e-01 7.86525249e-01
5.57614148e-01 1.74760401e-01 -2.05036432e-01 1.01685214e+00
2.66781777e-01 1.34795271e-02 1.14616859e+00 -4.06405896e-01
-9.12417829e-01 3.61864090e-01 -6.97593868e-01 1.81882545e-01
-2.55950335e-02 -4.18485314e-01 7.53768742e-01 -9.63751853e-01
5.79657018e-01 1.15816188e+00 2.75817037e-01 1.32776082e-01
-1.20895398e+00 -4.24804032e-01 -8.28266591e-02 5.09037554e-01
-1.40592027e+00 -1.75996825e-01 9.48705733e-01 -5.29649615e-01
4.05144185e-01 7.33923078e-01 4.90682662e-01 8.02819431e-01
2.21877888e-01 6.61802649e-01 1.60015559e+00 -9.05745149e-01
1.94492772e-01 3.10884088e-01 4.18463081e-01 7.14318216e-01
2.68207937e-01 -6.86434329e-01 -3.25312525e-01 1.19708702e-01
7.42932439e-01 -8.11836347e-02 -1.02096647e-01 -1.62021145e-02
-1.09172583e+00 6.91210389e-01 -2.92903125e-01 6.30779803e-01
-2.86094416e-02 -4.81587678e-01 4.31001157e-01 4.54015195e-01
3.58397186e-01 3.04209024e-01 -5.37984399e-03 -2.29346529e-01
-1.18178320e+00 7.07305670e-02 8.31453919e-01 9.25830722e-01
2.15464741e-01 2.59250164e-01 3.07463080e-01 1.43092668e+00
-1.96718387e-02 9.10408497e-01 6.69491053e-01 -3.23666632e-01
3.84470582e-01 5.60440063e-01 -7.78856203e-02 -1.21290183e+00
-2.43981555e-01 3.35366242e-02 -1.02171648e+00 5.08043766e-01
6.50393188e-01 6.38884008e-01 -8.73547256e-01 5.20843863e-01
-5.87502494e-02 -7.23068476e-01 7.80171454e-02 8.52624655e-01
4.04941589e-01 8.19944978e-01 -3.45718294e-01 -5.89059815e-02
1.70659018e+00 -5.99708021e-01 -7.06445456e-01 1.74292073e-01
1.47760278e-02 -1.52962255e+00 1.24831462e+00 1.15238416e+00
-4.86019850e-01 -7.01091230e-01 -1.16857171e+00 1.57696322e-01
-4.66648251e-01 1.01011086e+00 5.06170034e-01 1.21178937e+00
-4.33105797e-01 3.10145646e-01 -7.75994241e-01 -6.59767509e-01
1.08533934e-01 4.57800888e-02 -2.08778217e-01 2.34156057e-01
-4.30016339e-01 8.40054870e-01 3.01348239e-01 1.42416850e-01
-1.85215309e-01 6.36219010e-02 -2.82207936e-01 -8.10535774e-02
1.60920426e-01 5.90267599e-01 8.16674471e-01 -8.54919434e-01
-1.75123262e+00 8.78035784e-01 2.59762332e-02 -1.42826959e-01
6.74698234e-01 5.71392365e-02 -1.21197844e+00 2.23199084e-01
-3.37626249e-01 -5.63724376e-02 1.31517303e+00 -9.47990477e-01
-5.95706522e-01 -4.52593654e-01 -6.06108725e-01 -2.68583775e-01
-9.50769782e-01 2.95103163e-01 -4.39201355e-01 -1.15982795e+00
5.71665645e-01 -9.52708483e-01 4.47923809e-01 2.16723755e-01
-2.63224751e-01 -1.50016636e-01 1.14048612e+00 -1.01423240e+00
1.10071552e+00 -2.09751415e+00 9.01895482e-03 7.23834634e-01
-1.87163293e-01 3.35794330e-01 1.93249464e-01 4.21618044e-01
2.40165263e-01 -2.25406170e-01 -1.76330402e-01 8.96758437e-02
-2.15709865e-01 2.76838422e-01 -5.35967410e-01 5.38750410e-01
1.40976399e-01 8.77685547e-02 -1.81273341e-01 -7.53016651e-01
4.18646157e-01 4.49968249e-01 1.79507777e-01 -3.53343666e-01
-9.98026971e-03 -7.23576173e-02 -5.43074369e-01 9.46981907e-01
7.89758801e-01 -3.34870741e-02 4.08652753e-01 -8.23238939e-02
-4.59021688e-01 -3.82984519e-01 -1.46592057e+00 1.01843202e+00
-3.93347830e-01 1.22866488e+00 -1.93739027e-01 -9.74936187e-01
1.73857391e+00 -8.03633705e-02 -1.80745035e-01 -7.12442696e-01
1.08330987e-01 5.95654786e-01 -5.38008623e-02 -4.67736155e-01
9.62795436e-01 1.86996743e-01 1.24767005e-01 3.70259643e-01
-3.25989068e-01 -1.65727720e-01 7.74087608e-01 -1.71800345e-01
3.73337686e-01 -5.27559482e-02 1.03655696e-01 -5.89270592e-01
1.05956888e+00 1.76325947e-01 5.79150654e-02 4.19685662e-01
4.02786076e-01 5.82932532e-01 3.03093523e-01 -6.05268300e-01
-1.31557167e+00 -8.93462300e-01 -5.95080793e-01 8.03253353e-01
-1.82645008e-01 -2.33149186e-01 -6.56081975e-01 -1.45847067e-01
5.93053810e-02 3.31160992e-01 -4.98653352e-01 5.04383624e-01
-8.73753548e-01 -7.96960592e-01 6.09274745e-01 5.01691580e-01
5.40392220e-01 -9.11737800e-01 -7.07664192e-01 1.56987965e-01
2.19515294e-01 -1.08633065e+00 -2.41772413e-01 -8.07207897e-02
-9.68693554e-01 -1.04407823e+00 -1.10893476e+00 -9.61585641e-01
6.87834203e-01 3.28434676e-01 6.45855904e-01 2.34248340e-02
-4.42393959e-01 1.90221101e-01 -5.94962478e-01 -4.56216544e-01
-7.07009792e-01 1.81601495e-01 -8.33505541e-02 5.26223660e-01
3.28278720e-01 -2.52729446e-01 -1.37244657e-01 5.12968004e-01
-9.23231125e-01 -1.55509770e-01 5.82776129e-01 6.26985729e-01
4.07501459e-01 2.65717447e-01 -3.95913534e-02 -7.32033074e-01
7.51403213e-01 1.10200591e-01 -1.04749453e+00 6.03677392e-01
-6.18987322e-01 1.58191293e-01 1.15117812e+00 -5.57128966e-01
-1.13219965e+00 -1.75618276e-01 1.80926099e-01 2.36283436e-01
-1.35888606e-01 2.82492340e-01 -1.33518815e-01 -3.61732692e-01
6.70783997e-01 7.94659972e-01 1.78571954e-01 -1.01935983e+00
-3.14114653e-02 1.31681454e+00 7.68495858e-01 -8.01216066e-01
6.85831845e-01 6.95667148e-01 2.26433381e-01 -1.73506653e+00
-3.55984457e-03 -3.07541758e-01 -8.29622865e-01 -5.60824573e-01
6.83102667e-01 -3.40311438e-01 -6.33181989e-01 7.24963725e-01
-1.00870657e+00 3.23770404e-01 1.37972698e-01 5.05518377e-01
-2.77227938e-01 7.85758376e-01 -5.02733529e-01 -8.83957565e-01
-1.81687370e-01 -9.58364129e-01 7.88082540e-01 1.90498352e-01
-9.10574123e-02 -8.89332056e-01 -1.73525855e-01 1.67911187e-01
2.07435027e-01 -3.84828374e-02 1.15544760e+00 -4.11286056e-01
-3.26088160e-01 -5.62744081e-01 -3.14478368e-01 2.86408633e-01
3.09861124e-01 4.96190637e-01 -8.22604954e-01 -5.68713993e-02
-2.10412592e-01 2.63584908e-02 8.26126993e-01 -1.89046741e-01
1.29408896e+00 3.38753201e-02 9.40140784e-02 2.82431483e-01
1.33091879e+00 4.01898205e-01 5.79906881e-01 7.41101086e-01
4.90124226e-01 6.79786801e-01 4.90581036e-01 7.61031389e-01
-2.59004951e-01 6.25515640e-01 -2.48220682e-01 2.50986576e-01
-2.36609966e-01 1.89747080e-01 1.15328565e-01 1.10834420e+00
-5.66465080e-01 -2.27534726e-01 -1.06071877e+00 1.77658230e-01
-1.45365727e+00 -9.10945892e-01 -4.38795686e-01 2.11655593e+00
6.02689505e-01 9.75490063e-02 2.83186167e-01 8.57428610e-01
8.38877678e-01 1.96014717e-01 -7.92596415e-02 -7.82635808e-01
-6.34232342e-01 2.84039587e-01 5.10234177e-01 1.79127052e-01
-9.73583579e-01 7.33897746e-01 6.11784267e+00 1.02082241e+00
-1.57770109e+00 -3.34014267e-01 4.62817222e-01 3.30984294e-01
9.93393362e-02 -3.03449184e-01 -8.63989592e-01 5.32624841e-01
5.09737849e-01 -4.73928340e-02 2.89766192e-01 6.77102983e-01
5.44421487e-02 -1.51025951e-01 -7.03781545e-01 1.12838995e+00
4.65581864e-01 -1.31419325e+00 2.85331398e-01 8.76194090e-02
5.34021258e-01 -7.12423980e-01 1.66157320e-01 -2.80535012e-01
-2.16219783e-01 -8.21239710e-01 1.10469294e+00 5.96117616e-01
6.16437137e-01 -6.87915206e-01 3.55329990e-01 5.36628328e-02
-1.13856959e+00 -6.90860152e-02 -7.77088284e-01 1.24025401e-02
-9.90286097e-02 6.04192436e-01 -5.09278893e-01 3.83887708e-01
6.00010633e-01 4.85373199e-01 -9.49529052e-01 9.49559629e-01
8.38654190e-02 6.83493257e-01 -2.98773885e-01 -7.08411038e-01
2.09914982e-01 -5.76202989e-01 2.84281194e-01 1.49641776e+00
4.65268761e-01 -1.89492017e-01 -1.39965713e-01 4.57952887e-01
3.08853924e-01 7.31129408e-01 -9.08133015e-03 -3.34361166e-01
2.37457827e-01 1.12281060e+00 -1.47588289e+00 -4.25235718e-01
-4.96636808e-01 1.04429865e+00 -5.43678440e-02 7.27397799e-02
-3.95237595e-01 -6.64463162e-01 3.86795104e-02 1.24279514e-01
4.69069690e-01 -7.67790735e-01 -5.57796240e-01 -1.01263714e+00
3.71563643e-01 -9.20927286e-01 2.95711607e-01 -5.29111445e-01
-1.20076239e+00 6.75705850e-01 -3.49138714e-02 -1.61667430e+00
2.19757661e-01 -1.40608895e+00 -3.91376764e-01 8.15861285e-01
-9.25849020e-01 -1.12312722e+00 -3.09974164e-01 3.97974879e-01
8.90863538e-01 -8.31951380e-01 8.19921315e-01 2.88806856e-02
-4.57937777e-01 4.16085571e-01 9.30868864e-01 2.05380008e-01
7.57058561e-01 -1.22516978e+00 2.83648908e-01 8.30474198e-01
4.79753077e-01 4.64396864e-01 6.71111405e-01 -5.02751589e-01
-1.52363610e+00 -6.68200374e-01 8.91332269e-01 -1.07811965e-01
8.47828448e-01 -4.64577466e-01 -9.34560895e-01 3.03447619e-02
1.10510603e-01 -2.90163040e-01 6.82496130e-01 -8.31862614e-02
-5.84355056e-01 -3.46618831e-01 -8.62683654e-01 5.89264274e-01
4.96128649e-01 -4.17725861e-01 -5.33852875e-01 3.64374936e-01
-3.21388423e-01 -3.09694987e-02 -9.69223678e-01 -5.09820223e-01
1.06433308e+00 -7.64144480e-01 7.41403759e-01 -1.84407875e-01
4.36751664e-01 -3.10298473e-01 -3.61143768e-01 -9.22767758e-01
-2.60010660e-01 -4.15829629e-01 3.04079622e-01 1.27420998e+00
2.35855192e-01 -6.01991117e-01 8.43395650e-01 1.91309154e-01
1.90073431e-01 -1.61724210e-01 -7.13775992e-01 -1.28450048e+00
5.06831752e-03 -6.35302588e-02 5.62834561e-01 9.00077939e-01
-5.76944090e-02 -2.89305616e-02 -4.94207025e-01 -1.36742517e-01
8.00916433e-01 5.36050737e-01 6.39192760e-01 -1.61136377e+00
-2.78810352e-01 -6.94816470e-01 -8.85091603e-01 -5.76062322e-01
-1.80294458e-02 -8.44659448e-01 -3.25990587e-01 -1.08428466e+00
-6.77542537e-02 -2.95720428e-01 1.17990315e-01 1.57615677e-01
3.08874935e-01 4.71937776e-01 4.49169666e-01 4.73915815e-01
1.22161970e-01 2.12295964e-01 1.05293882e+00 -4.68502134e-01
-5.29990010e-02 -7.59920105e-02 -1.41905919e-01 6.43184006e-01
7.85393953e-01 -2.08692476e-01 -6.72024563e-02 -3.36758226e-01
-7.14076310e-02 -3.26590002e-01 -5.71879223e-02 -1.30030346e+00
4.11877893e-02 -1.84802637e-01 8.06474209e-01 -6.61894739e-01
2.54856467e-01 -8.73102784e-01 -6.08222894e-02 4.49461371e-01
-1.42803982e-01 4.05029476e-01 1.34013325e-01 4.68964934e-01
-1.35985121e-01 -4.52230245e-01 8.16671491e-01 1.58418894e-01
-9.88977313e-01 -2.46944413e-01 -9.06161904e-01 -3.78134936e-01
7.65700221e-01 -7.16857493e-01 -4.03212994e-01 1.53730199e-01
-4.42932963e-01 -4.67096776e-01 5.66813052e-01 5.56267381e-01
6.18897557e-01 -1.14147747e+00 -8.45024049e-01 4.25859392e-01
3.92980635e-01 -8.64042401e-01 9.56622697e-03 4.19887275e-01
-1.34658730e+00 3.96909505e-01 -7.38756657e-01 -6.40079916e-01
-1.69432831e+00 2.35900879e-01 -3.15211207e-01 1.72328591e-01
-6.98546112e-01 1.98747292e-01 -5.37797809e-01 -5.48658930e-02
1.24910727e-01 -2.55059123e-01 -4.67629194e-01 2.27665797e-01
6.38183653e-01 7.33460605e-01 4.96795833e-01 -8.34422588e-01
-1.77775964e-01 9.38795567e-01 -3.15319210e-01 -3.03213358e-01
1.16305411e+00 2.69411147e-01 -1.98949561e-01 6.49880886e-01
9.69423175e-01 3.39155138e-01 -6.14537716e-01 -6.07150309e-02
3.79046708e-01 -8.26166093e-01 -2.44508356e-01 -5.90447068e-01
-8.11963081e-01 1.02597415e+00 8.43128622e-01 3.29275221e-01
1.13783896e+00 -5.01524270e-01 3.13629925e-01 8.50363791e-01
3.67099494e-01 -1.61843908e+00 -1.48674503e-01 2.94476360e-01
1.02922904e+00 -8.94776046e-01 3.59370261e-01 -5.47408342e-01
-6.52807713e-01 1.83696377e+00 2.66407013e-01 -4.49966639e-02
5.96397281e-01 2.66832829e-01 2.76235312e-01 2.85111934e-01
-1.83280498e-01 1.86638296e-01 3.49753559e-01 4.67834502e-01
6.93599582e-01 3.40408981e-01 -9.48249698e-01 3.34314704e-01
-4.16103542e-01 -3.54582548e-01 9.25793886e-01 1.18078732e+00
-6.10367060e-01 -1.53614616e+00 -1.16541481e+00 7.33723462e-01
-3.93243998e-01 1.73640117e-01 -6.24428689e-01 8.63691688e-01
-3.53894860e-01 7.57830143e-01 1.90846175e-01 -3.52192402e-01
2.14076027e-01 8.18213001e-02 7.03005731e-01 -2.11324058e-02
-4.47699219e-01 3.26070726e-01 8.58715549e-02 2.16175139e-01
-2.72519320e-01 -9.67522025e-01 -8.62345457e-01 -4.97638404e-01
-1.39475390e-01 2.90566921e-01 1.12142885e+00 5.50110936e-01
-2.07192704e-01 1.65014252e-01 6.83148861e-01 -5.00429869e-01
-7.54681468e-01 -9.03002441e-01 -1.13015056e+00 3.75415355e-01
-1.29932761e-01 -5.70097625e-01 7.88367093e-02 5.31620979e-01] | [11.881258964538574, 2.5684380531311035] |
78e5c05e-3ed5-4498-94b3-12c0331e8bd3 | stock-price-prediction-using-generative | null | null | https://thescipub.com/abstract/jcssp.2021.188.196 | https://thescipub.com/pdf/jcssp.2021.188.196.pdf | Stock price prediction using Generative Adversarial Networks | Deep learning is an exciting topic. It has been utilized in many areas owing to its strong potential. For example, it has been widely used in the financial area which is vital to the society, such as high-frequency trading, portfolio optimization, fraud detection and risk management. Stock market prediction is one of the most popular and valuable areas in finance. In this paper, it proposes a stock prediction model using Generative Adversarial Network (GAN) with Gated Recurrent Units (GRU) used as a generator that inputs historical stock price and generates future stock price and Convolutional Neural Network (CNN) as a discriminator to discriminate between the real stock price and generated stock price. Different from the traditional methods, which limited the forecasting on one-step-ahead only, by contrast, using the deep learning algorithm is possible to conduct the multi-step ahead prediction more accurately. In this study, it chose the Apple Inc. stock closing price as the target price, with features such as S&P 500 index, NASDAQ Composite index, U.S. Dollar index, etc. In addition, FinBert has been utilized to generate a news sentiment index for Apple Inc. as an additional predicting feature. Finally, this paper compares the proposed GAN model results with the baseline model. | ['Amir Jafari', 'Gaofeng Huang', 'Chen Chen', 'HungChun Lin'] | 2021-04-02 | null | null | null | journal-of-computer-science-2021-4 | ['stock-market-prediction', 'portfolio-optimization', 'stock-price-prediction', 'stock-prediction'] | ['time-series', 'time-series', 'time-series', 'time-series'] | [-4.54503566e-01 -2.38080308e-01 1.70888565e-02 -1.68857262e-01
-3.05504024e-01 -4.76692855e-01 6.14453197e-01 -3.01861078e-01
-2.66042829e-01 1.17080164e+00 2.48734891e-01 -3.10742646e-01
2.99526632e-01 -1.49972129e+00 -4.26727533e-01 -6.13330781e-01
1.73330426e-01 1.93146095e-01 5.67184165e-02 -5.50642133e-01
5.75886071e-01 5.15863538e-01 -1.19247639e+00 1.45431608e-03
5.61066508e-01 1.44396913e+00 -1.06731437e-01 1.35289341e-01
-4.21367258e-01 1.21712387e+00 -9.37489629e-01 -7.20184565e-01
7.15761781e-01 -8.00954342e-01 -1.23273440e-01 -3.88000488e-01
-5.24339676e-01 -6.82741642e-01 -3.24562371e-01 1.13643456e+00
3.20589423e-01 -6.21808358e-02 4.78616059e-01 -1.25491941e+00
-7.77361929e-01 8.02158237e-01 -7.84762383e-01 5.60482979e-01
-3.47678602e-01 1.26652181e-01 7.63805211e-01 -5.75926125e-01
2.64019310e-01 8.58590424e-01 5.08351445e-01 2.76313007e-01
-5.91938615e-01 -1.28691053e+00 -1.32970035e-01 7.78691918e-02
-6.24958038e-01 2.71673203e-01 1.10264838e+00 -3.35105121e-01
8.21607232e-01 1.27887443e-01 1.14716923e+00 8.22129369e-01
8.48173201e-01 6.77537680e-01 1.26961529e+00 -1.64662153e-02
2.08390713e-01 1.22824639e-01 -2.39377379e-01 6.03429303e-02
3.91036153e-01 4.54159319e-01 -1.04285799e-01 -2.87953429e-02
1.10810518e+00 5.41519642e-01 -5.93850575e-02 4.93471950e-01
-1.08370757e+00 1.18256044e+00 6.28941894e-01 4.82900023e-01
-6.17770016e-01 9.45173129e-02 4.92080778e-01 6.83067501e-01
6.46507919e-01 3.42107475e-01 -4.14185613e-01 -4.06540453e-01
-1.23215556e+00 5.71821213e-01 7.73673415e-01 4.51567292e-01
2.97703564e-01 1.00743771e+00 8.55536461e-02 3.19333494e-01
2.28159741e-01 4.54813480e-01 1.41749871e+00 -3.14667851e-01
4.33331609e-01 5.59561014e-01 1.28580332e-01 -1.09232962e+00
-2.64085978e-01 -9.74292874e-01 -1.09728992e+00 7.58060455e-01
3.11884768e-02 -5.08275509e-01 -7.52425611e-01 1.23109174e+00
-1.42600581e-01 5.25368035e-01 3.83303285e-01 6.03419185e-01
4.20215726e-01 1.05478716e+00 -2.29040027e-01 -3.75513852e-01
1.07223725e+00 -7.63341427e-01 -5.96557438e-01 9.64841247e-02
-7.01194629e-02 -8.82801175e-01 5.00097096e-01 4.68830734e-01
-9.86210108e-01 -6.18903518e-01 -1.14897251e+00 4.61859256e-01
-6.20720923e-01 -7.32982308e-02 6.00512743e-01 4.89095807e-01
-6.73261166e-01 8.08528721e-01 -6.59121931e-01 3.48150551e-01
1.79657638e-01 1.38368249e-01 1.58634692e-01 7.99812078e-01
-1.67811751e+00 9.85824466e-01 7.04564333e-01 1.27530754e-01
-4.01658416e-01 -3.75016868e-01 -4.66269791e-01 3.10046852e-01
-1.82863533e-01 -3.62708181e-01 1.19627213e+00 -1.19134080e+00
-1.72474897e+00 3.70780796e-01 6.22605085e-01 -1.16925859e+00
8.28203022e-01 -1.94100775e-02 -7.62323081e-01 -4.26573217e-01
-7.92969242e-02 1.09976441e-01 6.46428704e-01 -2.98485398e-01
-7.20371902e-01 -2.85411477e-01 -2.28598148e-01 1.88737273e-01
-4.23041731e-03 -3.94740924e-02 3.94604146e-01 -1.57070994e+00
4.74621262e-03 -5.12437224e-01 -3.47526185e-02 -6.97785199e-01
-2.40823478e-01 3.08581777e-02 8.39760721e-01 -1.05966687e+00
1.18950641e+00 -1.86288524e+00 -3.97958517e-01 5.29952645e-01
-1.87397063e-01 1.05105266e-01 4.40354824e-01 4.66061115e-01
-3.88445467e-01 2.14961201e-01 -1.94183394e-01 2.25164011e-01
-1.30193196e-02 -1.85586035e-01 -8.43648612e-01 5.53393401e-02
2.72503883e-01 1.17565608e+00 -3.98402572e-01 1.51479766e-01
7.39519671e-02 1.14717305e-01 -1.21189980e-03 2.58543700e-01
-3.20604384e-01 2.87797540e-01 -4.57182527e-01 5.66102922e-01
6.06346011e-01 -1.24355435e-01 -4.33611840e-01 2.72212531e-02
-2.75643647e-01 3.16108353e-02 -1.14251947e+00 7.08070219e-01
-3.91169876e-01 4.81373310e-01 -5.74468970e-01 -9.24582005e-01
1.70251870e+00 3.92303407e-01 3.19616258e-01 -7.81017184e-01
2.82479227e-01 7.42865562e-01 1.17373079e-01 2.73470357e-02
4.83807743e-01 -6.79887295e-01 8.31280742e-03 6.47386014e-01
-3.65390748e-01 -5.39279953e-02 3.67783308e-02 -3.54008377e-01
6.49302602e-01 1.39911935e-01 4.52452242e-01 1.24844916e-01
4.38327789e-01 -1.38864629e-02 7.01642454e-01 -1.05928347e-01
1.57605916e-01 5.31129718e-01 5.64073563e-01 -6.78468227e-01
-1.15029752e+00 -6.90430045e-01 -5.55982664e-02 1.33750737e-01
-2.31194556e-01 4.65937078e-01 -4.99367326e-01 -3.82441401e-01
1.79852456e-01 9.30930138e-01 -7.20359802e-01 -1.46154419e-01
-4.14107710e-01 -1.05209208e+00 4.73880440e-01 7.19179511e-01
1.29881644e+00 -1.68270004e+00 -6.40359640e-01 6.28145397e-01
4.50432032e-01 -3.64790916e-01 -1.19342811e-01 3.66343483e-02
-9.68194664e-01 -9.29831684e-01 -1.35197461e+00 -4.90520000e-01
9.52562839e-02 -4.14428562e-01 1.01280093e+00 -1.51372761e-01
1.01434171e-01 -3.87100875e-01 -4.03147519e-01 -9.23578858e-01
-3.84390622e-01 7.02119023e-02 -2.98064917e-01 1.58155695e-01
3.01413566e-01 -4.21720505e-01 -6.21811032e-01 -2.90717594e-02
-1.00964200e+00 -5.89839965e-02 7.65455186e-01 1.09180164e+00
4.67682779e-01 5.26385844e-01 1.26258147e+00 -1.11170363e+00
8.81397128e-01 -8.49336743e-01 -9.87444222e-01 5.36647774e-02
-9.17737365e-01 -1.77510142e-01 8.35593224e-01 -1.53848290e-01
-1.28157055e+00 -5.39802134e-01 -3.18347156e-01 -2.45906517e-01
4.06137228e-01 8.38293731e-01 8.19579065e-02 2.82084405e-01
6.65764958e-02 7.05432415e-01 8.56356844e-02 -4.83130932e-01
-3.45403850e-01 4.48985696e-01 4.03075576e-01 1.95756286e-01
9.27774549e-01 -1.07525215e-01 -4.07734327e-03 -2.08414748e-01
-4.04337883e-01 1.25098303e-01 -9.32382271e-02 -2.57884152e-02
7.24734724e-01 -8.42738748e-01 -5.15173376e-01 1.08488035e+00
-9.63070571e-01 7.95683563e-02 -2.89281726e-01 6.19918764e-01
-4.73084986e-01 -2.70871520e-01 -8.70069385e-01 -9.86695230e-01
-8.03000927e-01 -8.65073085e-01 2.78703392e-01 6.51406407e-01
1.00305922e-01 -1.09123802e+00 1.91090077e-01 -1.41176566e-01
7.12629914e-01 9.49582994e-01 6.68327332e-01 -1.13648462e+00
-5.73459625e-01 -6.02084160e-01 -6.97667524e-02 8.67474079e-01
2.68620431e-01 -6.57202005e-02 -5.52371144e-01 -1.30050823e-01
5.68518043e-01 -1.47302806e-01 7.49591947e-01 2.90313631e-01
8.06743681e-01 -5.49425602e-01 4.04876649e-01 7.12469935e-01
1.59628284e+00 1.03824949e+00 9.67980623e-01 9.06585455e-01
3.37338775e-01 1.41062170e-01 6.40529335e-01 5.61288238e-01
-1.21943198e-01 9.83418990e-03 3.53164971e-01 -2.66422331e-02
5.15655041e-01 -4.04187381e-01 5.16976357e-01 7.61697590e-01
-2.10021168e-01 -5.14463261e-02 -6.81904078e-01 -4.82728258e-02
-1.59146321e+00 -1.41067934e+00 3.15838635e-01 1.93102980e+00
7.05218196e-01 5.87123692e-01 9.86429602e-02 4.44137931e-01
6.32646859e-01 3.27729523e-01 -8.98846030e-01 -2.09925905e-01
-2.91120589e-01 5.60288548e-01 6.39699101e-01 -6.67684674e-02
-8.84194434e-01 5.73111236e-01 4.68814564e+00 7.98825026e-01
-1.66495740e+00 -1.89315349e-01 1.28284025e+00 1.59998208e-01
-4.83584732e-01 -3.65502894e-01 -7.96372771e-01 1.22219980e+00
7.66822457e-01 -6.39876246e-01 2.16267064e-01 1.07563329e+00
2.25503236e-01 2.71835595e-01 -4.41869020e-01 8.10585737e-01
-3.25346708e-01 -1.52102351e+00 2.24161506e-01 1.18197232e-01
7.88120687e-01 -2.41377130e-01 4.55907106e-01 3.90038878e-01
1.17437162e-01 -9.98336792e-01 6.54214263e-01 9.06801701e-01
4.44677591e-01 -1.46062779e+00 1.54031265e+00 3.81688744e-01
-1.01469409e+00 -6.61715940e-02 -2.96609908e-01 -9.66663659e-02
1.28165424e-01 5.49436092e-01 -5.74990630e-01 6.21147871e-01
5.72003365e-01 7.29347825e-01 -3.38176072e-01 9.12335038e-01
-3.09555866e-02 6.22373164e-01 -1.38064951e-01 -3.56423318e-01
4.45964694e-01 -6.62766933e-01 8.73693675e-02 5.92112899e-01
8.52321565e-01 1.58602804e-01 -2.10771292e-01 8.38830352e-01
-2.20170915e-01 2.94452310e-01 -5.54348111e-01 -1.53947771e-01
2.53702760e-01 7.64688909e-01 -7.17451334e-01 -5.25548100e-01
-2.79782712e-01 8.80700052e-01 -3.35737944e-01 2.23891020e-01
-8.69619131e-01 -7.25650489e-01 2.09680662e-01 2.24752039e-01
2.43154109e-01 7.10320547e-02 -3.81211698e-01 -1.31816983e+00
5.96170127e-02 -7.51084089e-01 1.73945278e-01 -6.71020985e-01
-1.49541068e+00 6.92161322e-01 -3.07960898e-01 -1.73855305e+00
-7.71772265e-01 -5.67284524e-01 -1.30488729e+00 1.53870249e+00
-1.74970615e+00 -8.98050010e-01 -2.12007575e-02 4.35268700e-01
5.50513148e-01 -1.06588697e+00 5.78404605e-01 4.90815975e-02
-4.97723579e-01 4.18785393e-01 5.46244621e-01 5.98387003e-01
1.91163808e-01 -1.22406149e+00 8.69982362e-01 8.38103294e-01
5.28713502e-02 3.82797360e-01 3.96183789e-01 -9.81731296e-01
-5.83107114e-01 -1.22194827e+00 5.13883233e-01 3.56727779e-01
8.86637211e-01 3.97989333e-01 -9.80427325e-01 7.50711501e-01
3.84988189e-01 -2.56872594e-01 4.57604825e-01 -8.32297266e-01
1.05671600e-01 -4.11462009e-01 -1.43125248e+00 3.00227404e-01
-1.22105494e-01 7.20732212e-02 -5.83893895e-01 -1.19110897e-01
6.20332658e-01 -4.94069278e-01 -8.39288235e-01 1.94015682e-01
4.98988301e-01 -1.34859991e+00 5.15393376e-01 -3.10147941e-01
5.74138403e-01 -1.08211495e-01 2.28627026e-01 -1.46481717e+00
-2.27769107e-01 -3.24491084e-01 -1.48910627e-01 1.40184879e+00
4.77362633e-01 -1.32300365e+00 1.02632439e+00 5.13702750e-01
1.17650211e-01 -1.01933575e+00 -8.09588671e-01 -6.31543219e-01
1.70266405e-01 7.14695528e-02 1.23688400e+00 8.92367661e-01
-4.79403734e-01 -2.68294550e-02 -5.10321558e-01 -3.55829656e-01
4.08974200e-01 2.85415441e-01 5.34611583e-01 -1.10346508e+00
-5.85025072e-01 -7.57579386e-01 -4.37078774e-01 -5.68330050e-01
-9.90587100e-03 -5.33863783e-01 -6.61320090e-01 -1.22358990e+00
-4.80792820e-01 -2.92029321e-01 -8.13093185e-01 7.68449530e-02
6.25484437e-02 7.95504749e-02 3.20521414e-01 4.23301458e-01
6.32146299e-01 5.10048270e-01 1.29160881e+00 -1.56613216e-01
-1.21484406e-01 6.89154983e-01 -4.71571475e-01 6.09855413e-01
1.41887128e+00 -2.15959195e-02 -2.78384149e-01 1.60088867e-01
3.12926710e-01 3.06434631e-01 9.04019251e-02 -8.58422637e-01
-3.84821296e-02 -2.01665461e-01 9.88981485e-01 -8.50957930e-01
1.26398340e-01 -7.07493782e-01 5.04631281e-01 9.42927599e-01
-1.19426921e-01 7.78751850e-01 2.86933710e-03 3.82237583e-01
-8.87531519e-01 -4.13798630e-01 6.38861299e-01 -5.77842712e-01
-7.21709847e-01 4.76339966e-01 4.47882749e-02 -1.21551767e-01
1.29378796e+00 -1.62947208e-01 -1.58975214e-01 -5.58520615e-01
-3.71382862e-01 1.03482015e-01 2.49024242e-01 3.80477965e-01
8.47130239e-01 -1.48479342e+00 -8.98183703e-01 2.95998573e-01
-3.88622969e-01 -7.21399933e-02 1.04018807e-01 2.58510828e-01
-1.06257010e+00 2.09825873e-01 -5.95365107e-01 3.48234773e-01
-5.43995738e-01 3.18350196e-01 4.35667634e-01 -6.85845196e-01
-4.97050285e-01 5.08808434e-01 -2.53975213e-01 2.98621148e-01
-8.70619565e-02 -4.55260694e-01 -5.89516521e-01 3.81988466e-01
5.94404817e-01 4.94878888e-01 3.23235653e-02 -3.30533832e-01
3.09623539e-01 3.76780987e-01 -7.76220672e-03 -2.86776483e-01
1.58099174e+00 3.91453832e-01 -1.69960514e-01 7.61903286e-01
1.04284167e+00 -9.46182534e-02 -1.03540015e+00 6.92878664e-02
2.20914871e-01 -1.79474443e-01 -2.08194051e-02 -8.91582847e-01
-1.58191788e+00 8.00213814e-01 5.10445654e-01 5.92591405e-01
1.17127228e+00 -7.91439772e-01 1.54148078e+00 7.96792656e-02
5.17282665e-01 -9.47589040e-01 -1.97412729e-01 3.05025935e-01
1.05898213e+00 -1.17051029e+00 -2.44147614e-01 3.32878411e-01
-7.25310504e-01 1.45401001e+00 4.25442606e-01 -6.91294074e-01
9.15971935e-01 3.70046318e-01 2.63169646e-01 2.51531065e-01
-5.48127353e-01 2.99367458e-01 -1.08242556e-01 8.51319283e-02
5.50745308e-01 2.45030597e-02 -3.31691325e-01 8.29245210e-01
-8.52411628e-01 2.13959038e-01 6.93925679e-01 8.49728167e-01
-3.44376802e-01 -1.05075622e+00 -4.36162263e-01 1.05769801e+00
-8.98679733e-01 -1.77763909e-01 4.70537655e-02 8.93582582e-01
-9.11323875e-02 2.17369467e-01 5.20226836e-01 -4.70915973e-01
2.47209340e-01 1.22643992e-01 -1.83090255e-01 -4.75688040e-01
-9.69232678e-01 -3.19840461e-02 -3.93230855e-01 4.41614091e-02
-2.20544085e-01 -6.49104714e-01 -1.23557091e+00 -3.79543722e-01
3.95682603e-02 2.63835013e-01 5.31228125e-01 6.49207771e-01
-1.21666595e-01 5.44613242e-01 1.21341062e+00 -5.63109636e-01
-1.05192447e+00 -1.06621265e+00 -1.07912707e+00 2.86323607e-01
2.17510909e-01 -4.05138373e-01 -3.39405388e-01 -1.05403244e-01] | [4.469584941864014, 4.218649387359619] |
8dc14aa6-54b0-46ea-9d51-6700e9946080 | textmi-textualize-multimodal-information-for | 2303.15430 | null | https://arxiv.org/abs/2303.15430v2 | https://arxiv.org/pdf/2303.15430v2.pdf | TextMI: Textualize Multimodal Information for Integrating Non-verbal Cues in Pre-trained Language Models | Pre-trained large language models have recently achieved ground-breaking performance in a wide variety of language understanding tasks. However, the same model can not be applied to multimodal behavior understanding tasks (e.g., video sentiment/humor detection) unless non-verbal features (e.g., acoustic and visual) can be integrated with language. Jointly modeling multiple modalities significantly increases the model complexity, and makes the training process data-hungry. While an enormous amount of text data is available via the web, collecting large-scale multimodal behavioral video datasets is extremely expensive, both in terms of time and money. In this paper, we investigate whether large language models alone can successfully incorporate non-verbal information when they are presented in textual form. We present a way to convert the acoustic and visual information into corresponding textual descriptions and concatenate them with the spoken text. We feed this augmented input to a pre-trained BERT model and fine-tune it on three downstream multimodal tasks: sentiment, humor, and sarcasm detection. Our approach, TextMI, significantly reduces model complexity, adds interpretability to the model's decision, and can be applied for a diverse set of tasks while achieving superior (multimodal sarcasm detection) or near SOTA (multimodal sentiment analysis and multimodal humor detection) performance. We propose TextMI as a general, competitive baseline for multimodal behavioral analysis tasks, particularly in a low-resource setting. | ['Ehsan Hoque', 'Mohammed Ibrahim Khan', 'Iftekhar Naim', 'Wasifur Rahman', 'Sangwu Lee', 'Md Saiful Islam', 'Md Kamrul Hasan'] | 2023-03-27 | null | null | null | null | ['multimodal-sentiment-analysis', 'sarcasm-detection', 'humor-detection', 'multimodal-sentiment-analysis'] | ['computer-vision', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 2.90176868e-01 -1.70953438e-01 -1.47651598e-01 -3.29104632e-01
-1.06725502e+00 -7.23900378e-01 5.78000844e-01 1.90070674e-01
-8.19063246e-01 2.36632988e-01 5.14927030e-01 -2.07719386e-01
5.02422988e-01 -1.78308383e-01 -5.94957054e-01 -4.39510763e-01
4.59281653e-01 4.04016256e-01 4.25205156e-02 -2.85973489e-01
-8.69245753e-02 9.81004611e-02 -1.53146517e+00 7.29822814e-01
4.97794688e-01 8.98579895e-01 1.01585455e-01 9.38532174e-01
-5.53840548e-02 1.23935843e+00 -4.00817037e-01 -7.26772785e-01
-3.61037135e-01 -5.48695266e-01 -6.61482334e-01 2.52042174e-01
4.32943463e-01 -5.77811658e-01 -3.07880342e-01 7.98629940e-01
5.29246449e-01 3.90498489e-02 6.44220471e-01 -1.28470159e+00
-2.77200907e-01 5.59501946e-01 -7.20908165e-01 -1.10635146e-01
6.31597221e-01 1.90864116e-01 9.53144908e-01 -9.44558978e-01
5.79806030e-01 1.53941035e+00 5.16808867e-01 7.45409012e-01
-1.13889122e+00 -4.68111783e-01 -6.33554310e-02 2.03796089e-01
-9.41559672e-01 -6.64441943e-01 7.50816345e-01 -5.68481803e-01
9.91447151e-01 3.27984929e-01 3.68331075e-01 1.63270760e+00
-2.96784729e-01 1.24377847e+00 7.92625189e-01 -1.71392024e-01
1.92211363e-02 3.69792879e-01 1.30798504e-01 7.83292055e-01
-4.17775065e-01 -4.34189856e-01 -9.13012803e-01 -9.14722383e-02
1.89837158e-01 -9.59421024e-02 -9.78954360e-02 -4.58494835e-02
-1.13551986e+00 9.83427048e-01 1.24284863e-01 8.81099999e-02
-2.43516177e-01 2.26793155e-01 8.92515182e-01 3.88051927e-01
3.51108521e-01 2.48448789e-01 -1.74015075e-01 -5.22890091e-01
-8.57128680e-01 1.67668507e-01 6.92003667e-01 5.03872275e-01
3.92472982e-01 -2.17676070e-02 -4.31054346e-02 1.22115242e+00
2.82473415e-01 7.95905411e-01 5.51960349e-01 -9.58421946e-01
7.77194738e-01 5.72951019e-01 -1.23420037e-01 -1.08494973e+00
-8.73162150e-01 1.48735285e-01 -6.14522874e-01 -5.64187951e-02
5.88952184e-01 -2.50623584e-01 -6.28564835e-01 1.89390624e+00
8.74341205e-02 -2.67569035e-01 1.54599667e-01 1.12143183e+00
1.10891485e+00 8.43863189e-01 2.56462395e-01 -2.17719495e-01
1.60254800e+00 -9.77057755e-01 -5.93045652e-01 -6.22473836e-01
7.64966428e-01 -7.75482833e-01 1.36942315e+00 4.47537005e-01
-1.22818696e+00 -2.84667760e-01 -6.59249067e-01 -4.64142174e-01
-2.18385279e-01 2.76701689e-01 3.75719786e-01 5.09832084e-01
-7.92257965e-01 -6.71822801e-02 -8.08057666e-01 -7.02570558e-01
1.90979227e-01 3.14146101e-01 -5.22496879e-01 1.70629714e-02
-8.83566856e-01 7.78551400e-01 -2.25441642e-02 -2.33208407e-02
-8.38251352e-01 -1.84117928e-01 -1.21988583e+00 -4.15627658e-02
5.71399391e-01 -5.53095520e-01 1.45457649e+00 -1.27353597e+00
-1.48850691e+00 9.82147217e-01 -4.04741794e-01 -1.88381732e-01
4.19932693e-01 -8.97220224e-02 -2.43789360e-01 5.04547715e-01
-2.12300792e-01 8.88706744e-01 1.06112921e+00 -1.11491442e+00
-4.09502149e-01 -4.13527787e-01 2.16810077e-01 3.57164025e-01
-8.70489180e-01 5.16347647e-01 -7.50507593e-01 -4.64185357e-01
-3.68198067e-01 -1.09436500e+00 5.64231277e-02 -1.97836667e-01
-3.79214406e-01 5.42174187e-03 8.01403821e-01 -8.37441146e-01
1.23500717e+00 -2.31928349e+00 5.41621923e-01 -7.72080794e-02
2.53786087e-01 1.76667005e-01 -4.64051306e-01 4.77220863e-01
2.82096237e-01 8.39635134e-02 -2.84571320e-01 -8.56234431e-01
1.47844493e-01 4.30254415e-02 -9.86570790e-02 2.85576612e-01
2.43011773e-01 1.07155633e+00 -6.72961414e-01 -5.37319183e-01
2.82633573e-01 5.24668574e-01 -8.59979391e-01 2.90188611e-01
-1.76754579e-01 3.85462016e-01 -2.13116229e-01 6.92570448e-01
1.17404766e-01 -3.52972776e-01 1.95798695e-01 -1.88975811e-01
1.80748805e-01 1.30158082e-01 -6.68288946e-01 1.55964351e+00
-7.70647824e-01 9.36602712e-01 1.64825559e-01 -8.23245347e-01
4.68116343e-01 3.17340225e-01 3.89821023e-01 -7.03381896e-01
3.91553313e-01 -4.47659865e-02 2.82188994e-03 -9.73410368e-01
4.69200552e-01 -3.40958565e-01 -5.67710161e-01 4.75875646e-01
1.32745132e-01 -1.31317889e-02 2.90756494e-01 3.23820114e-01
1.04552674e+00 -1.67629883e-01 1.93118468e-01 4.42508698e-01
5.73121488e-01 3.08790002e-02 3.44803743e-02 5.82226038e-01
-9.12482291e-02 6.21787965e-01 8.43956649e-01 4.42913510e-02
-1.04211116e+00 -7.72988081e-01 2.56041199e-01 1.75922811e+00
-1.02672957e-01 -6.54320121e-01 -8.39762926e-01 -6.16091311e-01
-1.38343439e-01 4.03582007e-01 -4.47805613e-01 -2.30495825e-01
-4.45836395e-01 -7.28217125e-01 6.62041962e-01 5.84508717e-01
4.05266881e-02 -1.17636347e+00 -6.28427804e-01 5.82714751e-02
-5.15455484e-01 -1.57199538e+00 -3.68662477e-01 1.38470754e-02
-3.92290086e-01 -7.01161325e-01 -5.32307386e-01 -5.90948820e-01
3.44239116e-01 2.04568475e-01 9.20891106e-01 -1.93802863e-02
-2.28025332e-01 6.16416574e-01 -4.07934606e-01 -2.90048122e-01
-6.72995925e-01 -1.11848950e-01 2.88111493e-02 1.85005978e-01
4.12334800e-01 -9.64573547e-02 -2.48164654e-01 2.37603083e-01
-1.00148726e+00 2.52638519e-01 4.62442517e-01 8.95971954e-01
1.05663858e-01 -4.87027049e-01 5.75586736e-01 -6.23714745e-01
6.75921261e-01 -3.97560865e-01 -1.55133873e-01 -3.19009759e-02
1.60553426e-01 -1.95430011e-01 7.40491867e-01 -7.16426194e-01
-8.60825360e-01 1.52429044e-01 -3.18226218e-01 -5.09635687e-01
-2.96595693e-01 9.09651458e-01 -3.99372391e-02 1.69436574e-01
3.33845198e-01 -2.53414828e-02 1.66774198e-01 -5.22651851e-01
5.41218519e-01 8.46876383e-01 7.28448272e-01 -2.09171534e-01
5.45305252e-01 3.94568443e-01 -1.51303604e-01 -1.15902126e+00
-8.38933527e-01 -6.39625430e-01 -5.34151256e-01 -3.55470121e-01
1.15419304e+00 -1.06506968e+00 -9.31757510e-01 5.08674383e-01
-1.08983886e+00 -3.26359510e-01 2.50852048e-01 4.70930725e-01
-6.92058325e-01 5.89882731e-01 -7.53066838e-01 -9.47816432e-01
-1.64643124e-01 -1.21864974e+00 1.28865659e+00 -5.38177080e-02
-6.01346195e-01 -8.72719109e-01 -8.59224126e-02 1.02985859e+00
1.82414744e-02 -5.03148809e-02 9.60681379e-01 -7.51894832e-01
2.68873235e-04 -4.64413226e-01 -3.52586865e-01 3.18940133e-01
-3.75785589e-01 -8.71942490e-02 -1.23469734e+00 -1.32198632e-01
-1.64832711e-01 -1.18929577e+00 1.03787136e+00 2.24610910e-01
9.87591445e-01 -2.52664357e-01 7.30580017e-02 2.78289050e-01
8.30903172e-01 -5.73044755e-02 2.96535015e-01 6.07801862e-02
9.47942495e-01 8.67779911e-01 4.83333260e-01 5.83649993e-01
7.24170268e-01 9.10019040e-01 4.30217564e-01 -2.93276608e-01
-1.23538814e-01 -2.36016691e-01 9.57728267e-01 9.44029510e-01
1.29744694e-01 -6.14744484e-01 -9.99219477e-01 3.10255110e-01
-2.07860303e+00 -1.21170282e+00 -5.72245382e-02 1.87524462e+00
7.07199812e-01 -4.97485511e-02 5.91427386e-01 -3.90721820e-02
2.83259243e-01 1.55521154e-01 -4.81561869e-01 -4.68673825e-01
-2.47415707e-01 -4.05759692e-01 2.21826732e-02 5.55079699e-01
-1.15948164e+00 1.02717793e+00 5.68177557e+00 6.56902492e-01
-1.33045328e+00 3.12196404e-01 6.38219953e-01 -5.20066082e-01
-1.86177939e-01 -4.62717414e-01 -5.41823447e-01 1.81948185e-01
1.32080841e+00 2.11782396e-01 4.54954952e-01 5.95751226e-01
5.86987019e-01 -2.13519990e-01 -1.36029911e+00 1.50240457e+00
6.80084229e-01 -9.42427874e-01 5.60203195e-02 -1.03054605e-01
2.57184982e-01 9.84695703e-02 2.10982263e-01 5.31393170e-01
-3.29743326e-01 -1.25655603e+00 8.74501348e-01 1.90786734e-01
6.99726224e-01 -6.56541288e-01 7.08913803e-01 4.61786896e-01
-9.45540309e-01 -3.07821333e-01 -3.48907337e-02 -5.31220287e-02
4.14407670e-01 -1.15738839e-01 -6.59804821e-01 3.78908291e-02
4.29886967e-01 7.45405436e-01 -7.46595263e-01 5.76670706e-01
-5.17444536e-02 6.69218600e-01 -3.26621026e-01 -2.54299998e-01
3.57066005e-01 6.31664097e-02 4.94241178e-01 1.47832906e+00
5.41146211e-02 3.26610655e-02 4.03366834e-01 5.78041911e-01
-1.83847010e-01 4.21130747e-01 -6.20153725e-01 -5.42590380e-01
-8.63947868e-02 1.38027978e+00 -4.59336817e-01 -4.29971963e-01
-8.04934263e-01 1.09238052e+00 3.34845275e-01 3.92661601e-01
-9.51078176e-01 -2.81308163e-02 5.14752686e-01 -1.35129288e-01
-5.18801771e-02 -1.55224085e-01 -2.04555929e-01 -1.49693239e+00
-1.83065794e-02 -1.17274487e+00 5.02085268e-01 -9.36827421e-01
-1.14797258e+00 5.43130696e-01 -3.67594622e-02 -9.25145030e-01
-7.08280325e-01 -8.24146569e-01 -4.03099686e-01 3.58945727e-01
-1.12696779e+00 -1.47433519e+00 -3.00923765e-01 7.13952124e-01
8.45085740e-01 -7.90491402e-02 6.17113352e-01 4.29020047e-01
-7.84190059e-01 6.29151344e-01 -1.70936987e-01 2.26945087e-01
8.80586863e-01 -9.31389928e-01 -1.68400496e-01 5.91101170e-01
2.12341234e-01 3.05887550e-01 8.47285926e-01 -2.66596675e-01
-1.80254316e+00 -7.91403532e-01 7.90521204e-01 -6.57224000e-01
1.07539046e+00 -7.76233912e-01 -8.36479604e-01 5.16996622e-01
1.84774876e-01 -4.68491703e-01 8.50057781e-01 2.02802524e-01
-5.08681238e-01 2.11594746e-01 -7.07043231e-01 7.87156343e-01
6.23584211e-01 -9.45234060e-01 -3.73055309e-01 2.73532480e-01
5.38222134e-01 -1.25190869e-01 -7.06654370e-01 3.32317084e-01
7.03395724e-01 -7.58569121e-01 9.69361842e-01 -9.05234873e-01
8.72437596e-01 5.02539538e-02 -2.97866017e-01 -9.87367392e-01
1.84208661e-01 -4.82679367e-01 -1.17534861e-01 1.21103084e+00
6.77013576e-01 -1.07794181e-01 5.26133358e-01 6.34208560e-01
8.08374733e-02 -5.62353313e-01 -6.98358774e-01 -3.43653709e-01
-1.32920891e-01 -8.09327960e-01 -1.27413303e-01 7.23070502e-01
7.04994857e-01 8.96643758e-01 -8.69126320e-01 -2.19085533e-02
2.90078998e-01 6.89477026e-02 1.00704432e+00 -8.52829814e-01
-4.12515581e-01 -6.14838183e-01 -3.76108021e-01 -1.08270001e+00
5.28302312e-01 -8.76116991e-01 -3.14549357e-02 -1.25796866e+00
7.75770426e-01 2.89948553e-01 -6.03774842e-03 8.02258730e-01
-8.30319673e-02 6.54397964e-01 5.09950638e-01 3.95541377e-02
-8.57703924e-01 5.10037661e-01 8.49534690e-01 -1.97298497e-01
-2.56990701e-01 -2.48449594e-01 -5.34280658e-01 1.15008903e+00
5.35321295e-01 -1.67088360e-01 -3.55436891e-01 -3.87629718e-01
4.38170522e-01 4.65956062e-01 6.03969276e-01 -7.06876099e-01
6.39485642e-02 -1.09696835e-01 1.22023851e-01 -4.43143845e-01
9.41357374e-01 -6.14923298e-01 -3.99633050e-01 1.16408385e-01
-6.19150043e-01 3.25501873e-03 2.88722456e-01 4.08689499e-01
-3.33102494e-01 -1.93303898e-01 7.33680904e-01 7.51560852e-02
-6.75109386e-01 -6.21829294e-02 -7.59590209e-01 1.33945548e-03
6.91835463e-01 -8.33516940e-03 -5.24328411e-01 -9.40853536e-01
-8.87101829e-01 2.40538016e-01 5.40548980e-01 6.69567943e-01
6.69957697e-01 -1.02402961e+00 -6.33906007e-01 2.41454281e-02
3.88328731e-01 -6.15647376e-01 3.99972916e-01 1.34275818e+00
-9.06415656e-02 3.68987888e-01 -6.04926497e-02 -8.02135587e-01
-1.75562978e+00 4.88508075e-01 1.67483166e-02 -2.84514506e-04
-4.42488968e-01 5.95236659e-01 6.12046361e-01 -3.15085948e-01
4.40864742e-01 -1.68872818e-01 -3.38388681e-01 4.36897695e-01
7.83183873e-01 6.70007020e-02 -2.07887128e-01 -1.06814325e+00
-3.64631593e-01 6.24633670e-01 9.87191945e-02 -4.61316109e-01
1.25029457e+00 -4.32337612e-01 -4.20167227e-04 8.06477547e-01
1.46083021e+00 9.50321369e-03 -9.45226014e-01 -4.16068584e-02
-1.10417038e-01 -1.83653325e-01 -4.75843757e-04 -8.30644250e-01
-8.24129820e-01 1.24985218e+00 2.65557289e-01 7.80692026e-02
1.19325316e+00 4.34810281e-01 8.60566020e-01 6.91448808e-01
8.24938645e-04 -1.05660748e+00 6.25645638e-01 6.03958845e-01
8.73248935e-01 -1.62853265e+00 -3.01264137e-01 -1.44674227e-01
-1.30830359e+00 1.14411223e+00 5.72217643e-01 4.23502713e-01
8.98671001e-02 3.06516379e-01 2.46390134e-01 -3.03101599e-01
-9.95160222e-01 -2.91161060e-01 5.36292195e-01 3.28106552e-01
5.41254342e-01 -9.98595431e-02 -2.31374688e-02 8.13635290e-01
-1.68028086e-01 -2.21205220e-01 7.01524675e-01 4.82356012e-01
-5.36543846e-01 -5.93594313e-01 -3.76577675e-01 2.62644023e-01
-5.14889181e-01 -1.69217259e-01 -7.50779390e-01 5.50873995e-01
-2.00824380e-01 1.25495577e+00 3.30913323e-03 -4.91783977e-01
2.60508269e-01 2.45354071e-01 3.12909573e-01 -4.92909074e-01
-5.84160089e-01 5.87156594e-01 5.48965693e-01 -6.74227178e-01
-4.99921590e-01 -6.72549546e-01 -1.21419942e+00 -2.92109698e-01
1.47669822e-01 -3.33072573e-01 7.17226863e-01 1.14181435e+00
7.00307637e-02 2.05111191e-01 2.99078703e-01 -9.73431706e-01
-3.80912662e-01 -8.07079077e-01 -1.31960362e-01 7.58572578e-01
5.51634848e-01 -4.21268672e-01 -2.23180935e-01 3.02082330e-01] | [13.104141235351562, 5.134629726409912] |
f6e04dca-1a7e-42ac-87e7-57fca1dd05c6 | conditional-score-based-reconstructions-for | 2303.14795 | null | https://arxiv.org/abs/2303.14795v2 | https://arxiv.org/pdf/2303.14795v2.pdf | MRI Reconstruction with Side Information using Diffusion Models | Magnetic resonance imaging (MRI) exam protocols consist of multiple contrast-weighted images of the same anatomy to emphasize different tissue properties. Due to the long acquisition times required to collect fully sampled k-space measurements, it is common to only collect a fraction of k-space for each scan and subsequently solve independent inverse problems for each image contrast. Recently, there has been a push to further accelerate MRI exams using data-driven priors, and generative models in particular, to regularize the ill-posed inverse problem of image reconstruction. These methods have shown promising improvements over classical methods. However, many of the approaches neglect the additional information present in a clinical MRI exam like the multi-contrast nature of the data and treat each scan as an independent reconstruction. In this work we show that by learning a joint Bayesian prior over multi-contrast data with a score-based generative model we are able to leverage the underlying structure between random variables related to a given imaging problem. This leads to an improvement in image reconstruction fidelity over generative models that rely only on a marginal prior over the image contrast of interest. | ['Jonathan I. Tamir', 'Kannan Ramchandran', 'Ajil Jalal', 'Brett Levac'] | 2023-03-26 | null | null | null | null | ['mri-reconstruction', 'anatomy'] | ['computer-vision', 'miscellaneous'] | [ 5.73018491e-01 4.45687212e-02 1.35170162e-01 -5.96839786e-01
-1.15415263e+00 -4.36183244e-01 5.26377261e-01 9.41464528e-02
-6.09760880e-01 7.53229022e-01 3.39898318e-01 -1.98578194e-01
-6.91654146e-01 -4.84034985e-01 -5.28475642e-01 -1.01677096e+00
-6.41401932e-02 7.75139153e-01 2.88756728e-01 2.12415859e-01
1.80398256e-01 4.65749711e-01 -9.75341916e-01 1.63514897e-01
8.21214259e-01 6.18929982e-01 5.93106091e-01 8.38928342e-01
2.02075332e-01 9.06435847e-01 -9.06069949e-02 1.07203722e-01
2.16982476e-02 -7.01894701e-01 -1.07454503e+00 2.59782583e-01
1.76459923e-01 -4.11099672e-01 -3.25016707e-01 1.04946053e+00
6.71240985e-01 2.45962337e-01 7.72468328e-01 -7.09965885e-01
-2.10082635e-01 4.84586984e-01 -6.66508496e-01 6.19939089e-01
2.26864088e-02 5.02422228e-02 4.51920152e-01 -4.63263422e-01
7.60814428e-01 6.14110649e-01 5.37591398e-01 2.94853806e-01
-1.70727205e+00 -2.50001103e-01 -2.49717325e-01 2.95078427e-01
-1.03829741e+00 -2.04937935e-01 7.31575489e-01 -4.87591922e-01
5.44668615e-01 2.07702234e-01 6.70462549e-01 8.26963246e-01
5.39241791e-01 5.46801567e-01 1.72349286e+00 -4.57127571e-01
2.12948218e-01 -3.60136256e-02 1.38616815e-01 5.02844274e-01
2.75495797e-01 8.78629833e-02 -2.31961593e-01 -1.72960207e-01
8.59474301e-01 7.48281926e-02 -5.08712828e-01 -7.06578910e-01
-1.22646773e+00 8.01045477e-01 3.68930757e-01 5.31098604e-01
-7.30134428e-01 2.67434400e-02 1.16424941e-01 -1.23232022e-01
1.71405926e-01 4.30719733e-01 1.09614238e-01 -5.73682822e-02
-1.38713169e+00 1.53167620e-01 6.10731661e-01 3.92000616e-01
5.58930933e-01 -2.86025077e-01 -1.60727501e-01 5.59233129e-01
2.80565232e-01 4.29411232e-01 6.36520863e-01 -1.11183000e+00
-9.52764526e-02 -8.28848630e-02 -1.15910240e-01 -5.88903606e-01
-3.55454534e-01 -6.19325638e-01 -7.62429476e-01 2.60202706e-01
5.74519873e-01 8.76789540e-02 -1.11969888e+00 1.72066736e+00
4.51439291e-01 2.84562111e-01 -3.75371903e-01 1.05550432e+00
3.64790648e-01 1.75601378e-01 7.62474686e-02 -3.51779759e-01
1.23590767e+00 -5.53632677e-01 -8.52138042e-01 -3.43492664e-02
2.98716217e-01 -8.22235405e-01 7.19534934e-01 5.45040965e-01
-1.50723195e+00 -1.70690730e-01 -1.16230619e+00 1.88689038e-01
2.17090771e-01 -3.30076069e-01 5.43813050e-01 7.67164111e-01
-1.05154419e+00 7.63428152e-01 -1.33534110e+00 1.68747291e-01
4.35926676e-01 4.04096425e-01 -3.82149398e-01 -6.36064708e-01
-9.10742342e-01 1.19615984e+00 1.01718388e-01 2.26324070e-02
-1.01334918e+00 -1.12734401e+00 -6.16850674e-01 -3.90372545e-01
3.90599489e-01 -6.66908205e-01 1.04406691e+00 -5.95022738e-01
-1.35290110e+00 8.20496559e-01 -7.11432546e-02 -2.05354944e-01
7.00579107e-01 1.95943147e-01 -1.20212458e-01 6.99648499e-01
1.85115203e-01 3.46170783e-01 7.78613806e-01 -1.27584434e+00
7.25564584e-02 -3.91371667e-01 -1.30280375e-01 1.27676278e-01
3.34374070e-01 -1.57684028e-01 -3.28899503e-01 -3.05366635e-01
3.90600324e-01 -1.25113523e+00 -5.84313810e-01 -2.70044327e-01
-2.91306078e-01 6.95757747e-01 3.24625522e-01 -9.34866726e-01
8.95507395e-01 -1.93137693e+00 3.77654374e-01 3.09952140e-01
5.45896232e-01 -8.39863420e-02 4.20998409e-02 -1.11172557e-01
-4.02452290e-01 -2.15321273e-01 -5.93688011e-01 -1.53848827e-01
-4.31118041e-01 4.14405555e-01 1.53835580e-01 9.25622404e-01
-2.47426406e-02 8.90186727e-01 -1.07935750e+00 -5.38073480e-01
4.18577075e-01 7.47384250e-01 -6.62796378e-01 1.80605128e-01
3.63195896e-01 1.27642417e+00 -5.74677944e-01 -3.05295759e-03
9.62648094e-01 -6.35819256e-01 5.24467051e-01 -3.69323969e-01
-1.12442620e-01 1.92155465e-01 -1.14405692e+00 2.04692292e+00
-5.49945593e-01 1.46207869e-01 3.02958757e-01 -1.22699046e+00
1.49917170e-01 6.03628099e-01 1.04428852e+00 -6.08021379e-01
8.77054483e-02 4.11723763e-01 4.44601268e-01 -5.31636298e-01
1.57286257e-01 -8.60807955e-01 3.54821742e-01 5.16955018e-01
2.05486715e-01 -6.65923297e-01 1.86257884e-02 3.22859347e-01
1.31374407e+00 2.82886568e-02 5.99644482e-02 -4.98659104e-01
3.94780278e-01 -1.30967334e-01 3.36120188e-01 9.44460094e-01
-1.77089810e-01 1.01662028e+00 2.25596964e-01 -2.34304875e-01
-1.21141291e+00 -1.30655003e+00 -6.71155393e-01 1.69648781e-01
-7.58882836e-02 3.16488072e-02 -6.36094868e-01 -4.01454896e-01
-3.54421198e-01 4.95194256e-01 -7.89504588e-01 -8.76838490e-02
-8.18160474e-01 -1.43718565e+00 2.79902428e-01 2.30901182e-01
7.60245323e-02 -6.72771037e-01 -8.15342724e-01 5.31501889e-01
-4.84790742e-01 -1.15972137e+00 -2.80225307e-01 5.39442897e-01
-1.16068900e+00 -9.45448935e-01 -1.17451084e+00 -1.08171761e-01
7.51956999e-01 1.62070826e-01 1.26353061e+00 -1.38024896e-01
-7.55855620e-01 7.04581916e-01 -1.47908881e-01 -5.05219474e-02
-4.96258885e-01 -2.45384872e-01 -2.47692198e-01 -8.23997781e-02
-3.08735192e-01 -8.79123211e-01 -8.71695340e-01 5.00879958e-02
-1.22052896e+00 4.77666371e-02 6.20308220e-01 1.04578006e+00
9.03989494e-01 -8.84955600e-02 8.31288174e-02 -1.02547014e+00
2.99107939e-01 -5.19164324e-01 -3.68847996e-01 2.89396912e-01
-5.39331973e-01 4.99802053e-01 6.43393621e-02 -5.16936183e-01
-1.07188141e+00 -1.53202891e-01 -1.15319364e-01 -2.73222417e-01
-7.70136639e-02 6.02169991e-01 2.68251866e-01 -4.77758884e-01
4.79538023e-01 3.09679836e-01 3.70440513e-01 -2.13941246e-01
2.15859916e-02 1.63125116e-02 4.58151340e-01 -6.82969630e-01
3.99666160e-01 8.60581100e-01 7.00349391e-01 -7.69784451e-01
-8.89944375e-01 -4.63599533e-01 -7.30862677e-01 -3.59670907e-01
1.14671004e+00 -5.42271852e-01 -5.25899887e-01 2.12663934e-01
-7.38236725e-01 -2.75134325e-01 -5.17847955e-01 1.20085692e+00
-7.62748659e-01 5.69771290e-01 -6.94633722e-01 -5.25717556e-01
-6.80816732e-03 -1.84882343e+00 7.59084880e-01 8.21600705e-02
1.42327799e-02 -1.18722856e+00 3.60365093e-01 3.99637938e-01
8.57796848e-01 3.28096211e-01 8.66470695e-01 -1.73857719e-01
-7.66015291e-01 -1.12713173e-01 -8.87465030e-02 3.88695568e-01
9.44946334e-02 -5.57309628e-01 -7.18825221e-01 -2.08879828e-01
7.16503561e-01 -2.84145385e-01 6.62703693e-01 1.16660798e+00
1.24942482e+00 5.08336961e-01 -4.52460796e-02 7.55573690e-01
1.58650744e+00 -2.32472457e-02 7.96255589e-01 5.19424863e-02
4.07007724e-01 5.34165740e-01 2.01570749e-01 1.69937372e-01
2.59796530e-01 6.72412038e-01 2.47499481e-01 -3.17737252e-01
-3.81641835e-01 8.64081085e-02 -2.65286624e-01 8.07706952e-01
-3.70691180e-01 1.55253574e-01 -9.38694179e-01 5.84731281e-01
-1.33708656e+00 -9.51024950e-01 -4.22550082e-01 2.33349252e+00
8.78593326e-01 -1.76131546e-01 -1.73446476e-01 8.41095373e-02
5.09352267e-01 -8.01160466e-04 -5.12467742e-01 3.08949441e-01
-1.63382217e-02 6.05668902e-01 8.87259781e-01 8.19352865e-01
-7.42406130e-01 9.84558985e-02 7.33506107e+00 6.67104900e-01
-1.26666772e+00 5.24392128e-01 6.48202300e-01 -1.12047352e-01
-5.23033082e-01 2.18059123e-01 -3.46574396e-01 4.33356613e-01
1.03228319e+00 5.25557511e-02 5.64038515e-01 1.10293016e-01
2.07866997e-01 -7.82057941e-01 -7.82910585e-01 7.35590816e-01
-2.05922499e-02 -1.10627377e+00 -2.88073927e-01 4.96149689e-01
7.42037773e-01 2.93236792e-01 2.23233476e-01 -2.42657408e-01
3.42155457e-01 -1.08511198e+00 2.78352052e-01 8.18327367e-01
7.84694493e-01 -4.08987164e-01 5.70167303e-01 2.55042285e-01
-4.91092622e-01 5.00265837e-01 8.05022847e-03 3.78956348e-01
4.45822090e-01 1.00668025e+00 -7.57414579e-01 7.12566972e-01
4.20565099e-01 3.53103161e-01 -2.61534870e-01 1.12293363e+00
-9.84496474e-02 6.07822776e-01 -2.86289603e-01 6.43767595e-01
2.63601542e-01 -4.18363035e-01 6.10752583e-01 8.55993211e-01
3.37208509e-01 4.45462793e-01 1.31775588e-01 8.18670452e-01
4.74986732e-01 -1.47071123e-01 -1.68143496e-01 2.88811982e-01
-2.33302757e-01 1.33456373e+00 -9.61192310e-01 -3.84875238e-01
-3.84714216e-01 8.14362228e-01 4.65460904e-02 3.11644614e-01
-7.08589792e-01 2.37677619e-01 -9.76247042e-02 5.47592640e-01
1.42015383e-01 -4.01576906e-01 -1.73853368e-01 -1.14064646e+00
-1.05031669e-01 -6.97670162e-01 2.96232343e-01 -7.82787323e-01
-1.08279026e+00 4.92548078e-01 4.33982253e-01 -9.08003211e-01
-4.51654226e-01 -3.16257298e-01 -3.21582407e-01 1.28194642e+00
-1.83869123e+00 -7.46068954e-01 -1.49669826e-01 5.89504421e-01
-6.85028881e-02 4.90929812e-01 7.27994323e-01 4.21406865e-01
8.57151970e-02 -5.76807745e-02 2.50977516e-01 -2.19763741e-01
7.43727863e-01 -1.59896922e+00 -5.11723578e-01 6.78145587e-01
-2.46666253e-01 7.61343777e-01 8.22745681e-01 -6.76091909e-01
-1.32588208e+00 -3.32231104e-01 3.76222342e-01 -3.74625772e-01
6.60124779e-01 1.24945499e-01 -1.02230370e+00 5.93577623e-01
1.54306069e-01 6.98122919e-01 8.83656383e-01 -1.86792493e-01
2.34416276e-01 2.21481755e-01 -1.39839733e+00 1.01829711e-02
5.41066289e-01 -3.69989485e-01 -4.21196610e-01 1.07179701e-01
7.14181960e-02 -6.94199920e-01 -1.20969176e+00 3.25811505e-01
4.19170231e-01 -9.97178435e-01 1.17740858e+00 -2.62796491e-01
5.41984439e-01 -1.81129977e-01 -2.29825824e-02 -1.35859227e+00
-3.85458112e-01 -2.70211518e-01 2.17656106e-01 4.55644906e-01
7.45299533e-02 -4.91085529e-01 6.98387384e-01 8.90246689e-01
-1.83621764e-01 -3.26068342e-01 -1.16199565e+00 -4.75390851e-01
2.38319710e-01 -3.87282431e-01 5.10277525e-02 9.75166976e-01
-2.53424227e-01 -4.54736426e-02 -2.48434365e-01 1.90738618e-01
1.21663928e+00 7.12231621e-02 1.24639936e-01 -1.09406197e+00
-8.52789879e-01 -1.14144385e-01 -2.80856639e-01 -7.98503458e-01
-1.64373770e-01 -1.14517677e+00 6.57321140e-02 -1.28817856e+00
7.09086120e-01 -7.27203846e-01 -4.21780497e-01 -9.20088682e-03
-3.02025825e-01 3.27663332e-01 -1.02827579e-01 2.07155287e-01
-3.36796165e-01 1.33985594e-01 1.91664898e+00 1.12285212e-01
2.11288586e-01 -1.40621200e-01 -4.66276616e-01 4.55111593e-01
2.75444001e-01 -7.69490182e-01 -4.67760444e-01 -3.47715557e-01
1.26502618e-01 7.42493808e-01 5.95502675e-01 -9.61142361e-01
2.83099264e-01 -2.00992394e-02 5.38031876e-01 -3.83054286e-01
3.47305506e-01 -7.44101942e-01 5.03451049e-01 6.44046605e-01
-2.94844329e-01 -2.55850792e-01 5.32188676e-02 3.42221975e-01
-3.47030908e-01 -7.05207705e-01 1.14867723e+00 -4.92511839e-01
-3.16974558e-02 3.34828287e-01 -3.62068564e-01 1.86029375e-01
6.21400893e-01 6.67110533e-02 4.28235471e-01 -3.26469064e-01
-1.26835871e+00 -2.65223473e-01 3.56807739e-01 -2.36366063e-01
4.29928213e-01 -9.66694891e-01 -8.03418875e-01 1.29979342e-01
-3.42875153e-01 4.49313037e-03 9.31586027e-01 1.60719001e+00
-4.38811153e-01 3.48294079e-01 -4.51267719e-01 -8.79702985e-01
-9.64727044e-01 3.70927572e-01 6.50595725e-01 -9.70759749e-01
-7.59586573e-01 5.93418598e-01 2.25786701e-01 -2.57457703e-01
-4.97307777e-01 -8.28087181e-02 7.11133331e-02 -2.45624140e-01
5.86492956e-01 1.44120827e-01 4.52613443e-01 -6.27496004e-01
-2.58061320e-01 5.82446992e-01 -3.06707472e-01 -5.89928210e-01
1.55907786e+00 -1.70979172e-01 -6.59105647e-03 4.12232280e-01
1.07199275e+00 8.42824280e-02 -1.43648744e+00 -2.54412025e-01
-3.02893281e-01 -4.84223872e-01 7.78833508e-01 -8.98264766e-01
-1.07002079e+00 9.48833883e-01 6.48326457e-01 5.52874953e-02
1.01691651e+00 -1.00996308e-01 3.81317496e-01 -4.16588426e-01
4.90651041e-01 -7.87394404e-01 -1.68365166e-02 1.20586030e-01
6.18386209e-01 -1.21626914e+00 4.09673095e-01 -3.64619970e-01
-5.36248028e-01 9.16211545e-01 -2.70617843e-01 -3.49403650e-01
7.30793834e-01 6.09393895e-01 -1.41217381e-01 -3.27006429e-01
-2.70871550e-01 2.89504379e-02 3.18932742e-01 5.40240347e-01
6.00812495e-01 2.44306363e-02 -3.89419705e-01 8.55605453e-02
2.17761859e-01 3.05794835e-01 6.51344061e-01 1.13608170e+00
1.32565061e-02 -1.50115502e+00 -5.41470587e-01 3.75162542e-01
-8.76360476e-01 -1.57522231e-01 4.55892056e-01 5.53308845e-01
-2.65023828e-01 5.97622275e-01 -1.65876821e-01 4.41987783e-01
-4.23511602e-02 -7.37101212e-02 1.27735829e+00 -6.10642612e-01
-3.88007998e-01 3.68777007e-01 -4.21648651e-01 -5.87208450e-01
-7.36204684e-01 -1.10172141e+00 -1.28621054e+00 -5.15622497e-02
-1.56151325e-01 2.30062306e-01 1.10183871e+00 1.00140011e+00
-4.34012823e-02 9.41900015e-01 4.88512397e-01 -9.75888133e-01
-5.30833542e-01 -7.40356147e-01 -8.50904465e-01 5.15529633e-01
3.16369504e-01 -6.88108385e-01 -3.34708691e-01 -4.66148704e-02] | [13.531808853149414, -2.396821975708008] |
657acfc7-8de8-421f-ba85-53cb86fa0bc1 | learning-harmonic-molecular-representations | 2303.15520 | null | https://arxiv.org/abs/2303.15520v1 | https://arxiv.org/pdf/2303.15520v1.pdf | Learning Harmonic Molecular Representations on Riemannian Manifold | Molecular representation learning plays a crucial role in AI-assisted drug discovery research. Encoding 3D molecular structures through Euclidean neural networks has become the prevailing method in the geometric deep learning community. However, the equivariance constraints and message passing in Euclidean space may limit the network expressive power. In this work, we propose a Harmonic Molecular Representation learning (HMR) framework, which represents a molecule using the Laplace-Beltrami eigenfunctions of its molecular surface. HMR offers a multi-resolution representation of molecular geometric and chemical features on 2D Riemannian manifold. We also introduce a harmonic message passing method to realize efficient spectral message passing over the surface manifold for better molecular encoding. Our proposed method shows comparable predictive power to current models in small molecule property prediction, and outperforms the state-of-the-art deep learning models for ligand-binding protein pocket classification and the rigid protein docking challenge, demonstrating its versatility in molecular representation learning. | ['Hao Zhou', 'Fei Ye', 'Lihao Wang', 'Shi Chen', 'Yuning Shen', 'Yiqun Wang'] | 2023-03-27 | null | null | null | null | ['drug-discovery'] | ['medical'] | [ 3.06323707e-01 -4.24022637e-02 -4.12030250e-01 -2.77029812e-01
-9.27925110e-01 -4.26864952e-01 3.26122612e-01 3.68745595e-01
-2.71041870e-01 7.91608453e-01 6.34719953e-02 -6.96067810e-01
-3.96145433e-01 -7.78083324e-01 -9.49338078e-01 -9.22727466e-01
-7.78790414e-01 3.88600439e-01 -5.66170990e-01 -3.42486560e-01
3.77461940e-01 1.06142604e+00 -9.54883814e-01 3.04272562e-01
8.60955656e-01 5.52732289e-01 1.14092097e-01 6.42533183e-01
7.85981715e-02 8.56095433e-01 -2.35810623e-01 -3.61852795e-01
5.16164815e-03 -2.60299206e-01 -9.30871189e-01 -4.41813767e-01
5.11924028e-01 2.27541979e-02 -8.60814452e-01 7.69067466e-01
7.84163117e-01 5.37796855e-01 1.13892531e+00 -5.07885456e-01
-1.06075561e+00 1.72579482e-01 -5.17977476e-01 1.53712288e-01
5.97119749e-01 -9.56351534e-02 1.18643653e+00 -1.25051844e+00
8.69881988e-01 1.30149829e+00 6.97229207e-01 6.23229682e-01
-1.56020117e+00 -5.49560964e-01 -9.68643874e-02 4.53987598e-01
-1.81155205e+00 -1.33537352e-01 6.10095501e-01 -5.52500963e-01
1.28459525e+00 2.56391436e-01 3.86002511e-01 9.32888627e-01
7.95428991e-01 5.12277007e-01 4.53910172e-01 7.48424307e-02
2.72768050e-01 -5.70752025e-01 -2.05145791e-01 1.11627162e+00
4.40383516e-02 -1.68884069e-01 -5.07226110e-01 -5.50586522e-01
8.69235814e-01 5.39860010e-01 -3.94008726e-01 -6.60669506e-01
-1.09589660e+00 1.34629405e+00 8.85131001e-01 1.07181206e-01
-4.76716131e-01 3.87878060e-01 1.04673669e-01 1.84144929e-01
3.78037632e-01 8.17967534e-01 -1.69128418e-01 3.40272784e-02
-5.83233297e-01 1.63242966e-01 5.92144430e-01 5.52477241e-01
6.41298473e-01 -1.12751862e-02 1.75388992e-01 6.34844661e-01
5.09505749e-01 1.35295093e-01 1.98771253e-01 -6.85526431e-01
3.09975803e-01 5.34955025e-01 -2.49770224e-01 -1.27861702e+00
-6.33979499e-01 -3.86816949e-01 -1.19931066e+00 9.03369039e-02
1.79042947e-02 3.11190963e-01 -5.23005307e-01 1.50837708e+00
2.97892958e-01 2.01727152e-01 3.34601253e-01 9.02475059e-01
9.00300920e-01 9.55521107e-01 1.17994532e-01 -5.84898591e-02
1.01428747e+00 -4.64544982e-01 -3.09287876e-01 6.10585988e-01
1.14026010e+00 -5.72263122e-01 5.41758895e-01 2.55394548e-01
-8.75679493e-01 -2.26058856e-01 -1.33350694e+00 -2.83272833e-01
-5.06877422e-01 -3.10391169e-02 1.27311945e+00 5.07143736e-01
-6.40654683e-01 1.35586333e+00 -9.14926231e-01 2.22948715e-02
7.18134701e-01 9.51686442e-01 -8.64448845e-01 -3.04529190e-01
-1.07758653e+00 8.74005914e-01 2.23715350e-01 -1.36450119e-02
-9.53485548e-01 -1.07951546e+00 -1.05379164e+00 -2.44915903e-01
-3.56429487e-01 -6.54483557e-01 6.50056660e-01 -3.30243587e-01
-1.81808770e+00 6.50423408e-01 -2.56792530e-02 -4.66563851e-01
5.14441654e-02 1.56067625e-01 -2.52546996e-01 1.61881998e-01
-5.09731293e-01 6.36287987e-01 5.44989705e-01 -5.45070648e-01
1.29191980e-01 -5.94449282e-01 2.60416210e-01 2.97273338e-01
2.32395027e-02 -4.48187232e-01 3.52018893e-01 -5.17124474e-01
2.29793668e-01 -8.59516203e-01 -6.11705840e-01 -9.74829271e-02
-2.07742080e-01 -3.35054696e-01 4.09555644e-01 -3.98555338e-01
9.72145140e-01 -1.80757821e+00 9.12856102e-01 3.84317100e-01
4.85013068e-01 1.55421481e-01 -3.71584713e-01 6.11475170e-01
-5.37549257e-01 1.14233971e-01 -1.54998094e-01 -1.09645091e-01
-7.63450265e-02 8.69126525e-03 -2.78779536e-01 1.14598715e+00
2.30466068e-01 9.68845308e-01 -9.69588161e-01 1.10798539e-03
1.12645999e-01 1.23625231e+00 -9.16060746e-01 1.03038669e-01
-1.48598179e-01 5.40838838e-01 -6.90324903e-01 5.75013816e-01
7.57319987e-01 -4.31075305e-01 2.32680023e-01 -5.01300573e-01
6.04305267e-02 5.35759389e-01 -8.73728096e-01 2.36105776e+00
-3.28327209e-01 3.85900855e-01 -4.15569395e-01 -1.00088513e+00
8.52187157e-01 2.14700490e-01 7.61507094e-01 -6.24927819e-01
-1.21098623e-01 2.43999541e-01 2.39189699e-01 -1.71407819e-01
3.62510592e-01 -1.37179360e-01 3.99284601e-01 1.19888134e-01
1.27945051e-01 5.39318956e-02 -2.05440894e-01 1.29498258e-01
8.83231699e-01 2.91208684e-01 1.76346064e-01 -3.06211263e-01
6.36218667e-01 -2.34655544e-01 2.76159883e-01 3.36652786e-01
3.38935018e-01 6.08313680e-01 2.91958213e-01 -8.33595574e-01
-9.72788692e-01 -1.00369656e+00 -5.42697966e-01 1.16657698e+00
-6.49806932e-02 -7.58243859e-01 -5.44807434e-01 -4.76882458e-01
7.54830390e-02 2.30331808e-01 -6.98540390e-01 -5.97896397e-01
-5.60921133e-01 -1.27432406e+00 5.69946468e-01 1.36097148e-01
-4.18886393e-02 -5.97867548e-01 -6.30662665e-02 4.97537762e-01
2.38600880e-01 -4.25863922e-01 -5.73567331e-01 3.27612668e-01
-1.03694212e+00 -1.22860503e+00 -8.60184669e-01 -7.09990323e-01
6.99184239e-01 2.87126839e-01 7.82428145e-01 -3.83364022e-01
-9.23031569e-01 1.52034149e-01 -2.51422934e-02 2.03739274e-02
-3.58362883e-01 2.76070505e-01 3.05470169e-01 -1.72190517e-02
3.68347973e-01 -6.22409105e-01 -1.04169142e+00 1.72467723e-01
-8.03079665e-01 1.42753907e-02 4.53871369e-01 7.75933802e-01
1.03393543e+00 -2.91677624e-01 4.68000054e-01 -6.76629424e-01
5.72735429e-01 -5.50437450e-01 -4.36309725e-01 8.66630450e-02
-7.50504211e-02 5.58117509e-01 6.84305012e-01 -3.39219034e-01
-3.19425851e-01 2.47990161e-01 -2.95647770e-01 -2.72972494e-01
2.29043230e-01 6.79621696e-01 -2.96725649e-02 -8.50240171e-01
8.49280179e-01 1.93826094e-01 1.34145971e-02 -5.65255284e-01
5.31789184e-01 4.59122807e-01 9.35080945e-02 -7.87455320e-01
3.89795303e-01 3.55807036e-01 8.14716101e-01 -1.03738427e+00
-4.98880595e-01 -3.67656022e-01 -8.15115571e-01 4.30196404e-01
9.82278228e-01 -7.97800899e-01 -1.31837451e+00 -1.07846096e-01
-1.29366529e+00 2.14063581e-02 1.17839172e-01 8.01457822e-01
-6.87557936e-01 5.27773619e-01 -6.58046722e-01 -3.33926827e-01
-5.27941585e-01 -1.52782667e+00 1.20602477e+00 -2.62263477e-01
-8.25606510e-02 -1.15129685e+00 4.21910465e-01 1.65996194e-01
2.05125034e-01 6.28724098e-01 1.29232430e+00 -4.58195865e-01
-7.27912188e-01 -3.01368743e-01 2.04984229e-02 -7.81640485e-02
1.58934787e-01 -2.87724704e-01 -7.09283471e-01 -5.95431328e-01
-5.19162357e-01 -2.80079573e-01 8.86593580e-01 2.74007827e-01
1.49706900e+00 -4.02432054e-01 -2.42673352e-01 1.14638913e+00
1.26396680e+00 3.00892472e-01 7.39319980e-01 7.42444694e-02
1.12498260e+00 2.71485120e-01 5.25712371e-02 4.26989645e-01
2.39906803e-01 7.45774150e-01 5.79172432e-01 -1.62470862e-01
2.19652012e-01 -2.93365449e-01 4.26607907e-01 8.60199451e-01
-5.75739324e-01 6.75824061e-02 -7.16407597e-01 -3.03831369e-01
-1.75249541e+00 -1.02143693e+00 1.80300981e-01 2.38338876e+00
8.21733832e-01 -4.60082650e-01 1.59422513e-02 -2.13128492e-01
3.26582313e-01 1.87373444e-01 -7.51800716e-01 -4.93422002e-01
-2.75939465e-01 5.40257692e-01 6.06544733e-01 7.00573802e-01
-1.09478295e+00 8.54522228e-01 6.08898687e+00 9.63238418e-01
-1.11712837e+00 -1.23873547e-01 6.60185933e-01 1.62487328e-01
-3.08937341e-01 -4.74435031e-01 -6.50902867e-01 -5.17903902e-02
1.08163643e+00 5.26254205e-03 6.24188662e-01 6.92431450e-01
3.43011580e-02 6.27674699e-01 -1.54711008e+00 1.45589399e+00
1.49649814e-01 -2.19470763e+00 8.23503137e-01 5.51291347e-01
6.17264569e-01 1.25299573e-01 5.25023341e-01 -1.11737356e-01
-2.52261877e-01 -1.95230877e+00 -1.03561237e-01 6.04316235e-01
1.08075571e+00 -1.28908968e+00 4.13108736e-01 -1.32177800e-01
-1.34362125e+00 3.79552633e-01 -9.89098907e-01 2.20626876e-01
-3.69284689e-01 -9.80065092e-02 -7.67736554e-01 6.46559894e-01
-1.94784608e-02 1.19870043e+00 -4.21205282e-01 8.51314485e-01
4.20748532e-01 4.27034460e-02 -2.73955753e-03 -7.59875253e-02
5.15230775e-01 -8.45412254e-01 3.86924326e-01 1.26122713e+00
1.12396568e-01 4.25518453e-01 1.32738173e-01 1.03705454e+00
-5.45437276e-01 5.41086555e-01 -9.63574052e-01 -4.01593804e-01
2.13751793e-02 1.03039718e+00 -2.80738294e-01 1.84419140e-01
-3.00416172e-01 1.08423984e+00 4.31116760e-01 3.95321697e-01
-7.32423484e-01 -7.55841672e-01 1.31472254e+00 -1.96342338e-02
5.56994900e-02 -5.58567762e-01 3.54063004e-01 -1.24092162e+00
-3.82104993e-01 -9.37120140e-01 2.06964701e-01 -2.96805799e-01
-1.04871058e+00 4.05756682e-01 -5.37053287e-01 -1.02575767e+00
1.82473585e-01 -1.29159975e+00 -3.10513526e-01 9.22727883e-01
-1.48289490e+00 -8.71952355e-01 3.85110110e-01 6.53480172e-01
1.75939679e-01 -5.15940428e-01 1.42550933e+00 4.49611425e-01
-5.81196189e-01 7.69315541e-01 6.18962407e-01 -1.45584002e-01
2.96859980e-01 -1.19529116e+00 4.60319281e-01 -7.28781521e-02
2.80591220e-01 1.17956710e+00 3.62806350e-01 -2.73042113e-01
-2.38867712e+00 -1.35868371e+00 3.86331260e-01 -5.31189144e-01
5.63004494e-01 -2.46922821e-01 -1.00913048e+00 3.66422355e-01
-3.48224878e-01 -1.25529496e-02 1.39547861e+00 -1.62024260e-03
-7.29126275e-01 1.48619115e-01 -1.09711945e+00 7.29892552e-01
1.08540487e+00 -9.93835628e-01 -1.70170724e-01 9.62141335e-01
7.10450351e-01 -4.27762628e-01 -1.48344564e+00 3.80194366e-01
4.66301113e-01 -3.08386207e-01 1.55577087e+00 -1.30516779e+00
2.41222903e-01 -4.27941948e-01 -4.78569478e-01 -1.18410039e+00
-6.10810637e-01 -1.01960850e+00 -3.40230584e-01 2.59805650e-01
6.18194520e-01 -4.20655787e-01 7.93886065e-01 1.83710054e-01
-1.39869973e-01 -1.15318358e+00 -1.23040199e+00 -4.84276712e-01
5.26797891e-01 -1.24727003e-01 4.65686053e-01 9.54555273e-01
3.04769367e-01 5.01517296e-01 -2.60434985e-01 1.51676193e-01
6.81511402e-01 9.74872187e-02 4.14374262e-01 -1.31032169e+00
-3.06948453e-01 -5.03019035e-01 -9.98536289e-01 -1.21632051e+00
4.68150347e-01 -1.71996665e+00 -7.61241496e-01 -1.40311849e+00
3.34720254e-01 -1.79616436e-01 -4.49874401e-01 9.50515047e-02
3.53456289e-01 2.96634346e-01 -1.74478158e-01 1.39533699e-01
-8.76381338e-01 9.09537971e-01 1.38735187e+00 -5.18590748e-01
-2.44746134e-01 -1.81926653e-01 -4.31828707e-01 4.89346027e-01
7.09318995e-01 -2.50178993e-01 -1.99201509e-01 -2.42770031e-01
4.65499878e-01 -1.01112917e-01 1.11701742e-01 -6.78812623e-01
3.89848985e-02 -8.52915347e-02 6.43967092e-01 -1.53312027e-01
6.32138073e-01 -4.96866912e-01 2.57828236e-01 5.58096766e-01
-3.97723168e-01 1.01674192e-01 6.93472698e-02 7.57314980e-01
5.00250794e-02 2.05772385e-01 1.09499478e+00 -6.42415881e-02
-1.52756393e-01 1.07918477e+00 -3.75690877e-01 -2.50082254e-01
5.81633568e-01 -2.13413283e-01 -3.93306911e-02 8.72505158e-02
-7.62599051e-01 -2.79979229e-01 4.39504743e-01 2.93684065e-01
1.12844503e+00 -1.47730851e+00 -6.79586649e-01 2.61318982e-01
3.07504356e-01 -2.09366575e-01 1.31154418e-01 6.21228218e-01
-9.82760489e-01 7.63916016e-01 2.85877250e-02 -4.32608992e-01
-1.13946450e+00 4.21329409e-01 9.27616179e-01 9.92951468e-02
-6.44928455e-01 7.43281722e-01 4.43234295e-01 -4.30474639e-01
3.31774622e-01 -1.12030990e-01 -2.43304595e-01 -1.26534387e-01
7.44159758e-01 4.35899884e-01 2.13798031e-01 -9.61933613e-01
-4.64931339e-01 9.16892529e-01 -3.94671172e-01 4.95496392e-01
1.51190531e+00 4.85084862e-01 -3.27024907e-01 1.92552492e-01
1.93213129e+00 -5.00726640e-01 -1.03186095e+00 -1.29970893e-01
-4.29846114e-03 -2.26977065e-01 5.25507368e-02 -2.82686830e-01
-5.33135712e-01 1.17595029e+00 7.01052070e-01 -6.04740441e-01
2.30116501e-01 -1.33161262e-01 6.09938622e-01 1.22208881e+00
5.52912951e-01 -6.58633590e-01 1.23610631e-01 6.09719515e-01
1.11749995e+00 -1.18982303e+00 1.95794314e-01 -1.79042980e-01
-1.55816764e-01 1.48686123e+00 8.87971148e-02 -2.05626562e-01
5.40996850e-01 -5.87780297e-01 -5.00589550e-01 -5.66813111e-01
-4.74748164e-01 2.53196239e-01 7.98356593e-01 6.30970895e-01
1.12915659e+00 1.89171299e-01 -4.98053618e-02 2.03478307e-01
-1.47367110e-02 -5.10949790e-01 2.52906322e-01 6.55003667e-01
-4.69990015e-01 -1.26119041e+00 -1.29424974e-01 6.46877587e-02
-6.29943788e-01 -3.54785770e-01 -4.79989082e-01 3.63560408e-01
-1.91000506e-01 5.85122287e-01 -2.75221616e-01 -3.07200938e-01
2.78843671e-01 -1.13057300e-01 1.05565274e+00 -6.54707372e-01
-1.57300711e-01 -2.76486605e-01 -4.29778397e-01 -7.62710810e-01
-2.65377909e-01 -1.12023860e-01 -1.50073647e+00 -5.45401335e-01
-1.35662451e-01 5.44262648e-01 7.87742615e-01 5.12033582e-01
9.50473368e-01 4.77204829e-01 7.40533531e-01 -1.21712923e+00
-5.43353915e-01 -5.21657765e-01 -6.51629210e-01 2.82321751e-01
5.89189470e-01 -6.50955260e-01 -8.36161226e-02 -2.12123498e-01] | [5.08051061630249, 5.810705184936523] |
1d354288-29e9-41ce-b11d-b5078a6d35bb | improving-generalization-ability-of | 2305.10940 | null | https://arxiv.org/abs/2305.10940v1 | https://arxiv.org/pdf/2305.10940v1.pdf | Improving Generalization Ability of Countermeasures for New Mismatch Scenario by Combining Multiple Advanced Regularization Terms | The ability of countermeasure models to generalize from seen speech synthesis methods to unseen ones has been investigated in the ASVspoof challenge. However, a new mismatch scenario in which fake audio may be generated from real audio with unseen genres has not been studied thoroughly. To this end, we first use five different vocoders to create a new dataset called CN-Spoof based on the CN-Celeb1\&2 datasets. Then, we design two auxiliary objectives for regularization via meta-optimization and a genre alignment module, respectively, and combine them with the main anti-spoofing objective using learnable weights for multiple loss terms. The results on our cross-genre evaluation dataset for anti-spoofing show that the proposed method significantly improved the generalization ability of the countermeasures compared with the baseline system in the genre mismatch scenario. | ['Junichi Yamagishi', 'Erica Cooper', 'Xiaoxiao Miao', 'Xin Wang', 'Chang Zeng'] | 2023-05-18 | null | null | null | null | ['speech-synthesis'] | ['speech'] | [ 4.42951322e-01 -1.24473222e-01 -1.36592925e-01 -7.18020499e-02
-1.12044036e+00 -5.41764677e-01 4.40444648e-01 -2.51565963e-01
-1.15585171e-01 6.73098326e-01 3.97011012e-01 -1.54470459e-01
1.19652845e-01 -2.95538127e-01 -8.97984982e-01 -7.85336733e-01
1.10316806e-01 1.46993799e-02 1.08433820e-01 -3.91425073e-01
-1.15851238e-01 6.66401386e-02 -1.52577460e+00 6.55532598e-01
7.58329809e-01 9.27703857e-01 -9.90763232e-02 5.72183073e-01
5.51890731e-01 6.38772964e-01 -1.22044051e+00 -6.20995820e-01
2.46519074e-01 -5.08624792e-01 -7.31256664e-01 3.10457218e-02
8.09210479e-01 -2.33752638e-01 -7.15623200e-01 1.23847437e+00
8.90746713e-01 2.03734078e-02 4.29803908e-01 -1.34089625e+00
-5.49158156e-01 1.04041612e+00 -1.68645337e-01 3.53709251e-01
3.23290348e-01 9.87935811e-02 1.05836606e+00 -5.34031451e-01
4.36304808e-01 1.61864603e+00 6.90692067e-01 8.06328773e-01
-1.31968832e+00 -1.25060713e+00 2.35007685e-02 4.87171799e-01
-1.23488939e+00 -7.88631320e-01 1.15263546e+00 -8.32444876e-02
7.81888783e-01 4.45172876e-01 2.57794887e-01 2.34535646e+00
-3.59812379e-02 8.87551486e-01 1.05004299e+00 -2.02539340e-01
-1.12304553e-01 3.46992195e-01 -1.45759195e-01 2.40069136e-01
1.88646078e-01 5.68291426e-01 -9.63015497e-01 -1.90117896e-01
-1.55958816e-01 -7.53829956e-01 -7.66240954e-01 6.12524413e-02
-1.12312579e+00 8.93662214e-01 9.85854045e-02 3.28816444e-01
2.98112445e-02 3.03816468e-01 7.62771070e-01 8.10478806e-01
7.10099161e-01 8.27897370e-01 -5.22180140e-01 6.97131529e-02
-9.23543096e-01 4.06042308e-01 6.83980763e-01 6.77750170e-01
6.45789951e-02 6.58179283e-01 -2.32695043e-01 1.04438198e+00
2.98481937e-02 8.21159184e-01 6.01520956e-01 -6.35810733e-01
1.00983083e+00 -4.31961894e-01 5.86693846e-02 -1.34381127e+00
-2.33696446e-01 -1.01490009e+00 -7.53082097e-01 -4.65021640e-01
9.23883319e-02 -1.76674679e-01 -5.87202132e-01 2.06866169e+00
6.86105564e-02 7.73914993e-01 1.23154879e-01 9.04126585e-01
6.75846756e-01 5.87708235e-01 -3.71101916e-01 -3.72305691e-01
1.07065892e+00 -1.11610198e+00 -1.19867921e+00 -2.64231473e-01
6.18299782e-01 -1.01843452e+00 1.06800580e+00 8.70734930e-01
-6.39508367e-01 -6.38217747e-01 -1.41671729e+00 2.34847650e-01
-3.51079196e-01 6.10431954e-02 3.55983466e-01 1.19653320e+00
-5.47769547e-01 5.80357969e-01 -4.20024157e-01 1.73491821e-01
2.65992790e-01 -1.39616042e-01 -2.59119183e-01 2.03101173e-01
-1.74708843e+00 6.46416485e-01 6.95917785e-01 -8.48957151e-02
-1.43627977e+00 -6.08052909e-01 -6.81176424e-01 4.09447476e-02
4.58775103e-01 -5.15993834e-01 1.07227075e+00 -9.18003023e-01
-1.51382029e+00 7.64670551e-01 3.15202028e-01 -8.23505819e-01
5.12575984e-01 -4.42162782e-01 -1.13982856e+00 1.40890956e-01
-6.26318157e-02 2.45915353e-01 1.55816996e+00 -1.37451267e+00
-4.42988515e-01 -2.23854370e-02 -1.56344727e-01 -1.68834269e-01
-6.33660853e-01 7.49072731e-02 9.52115357e-02 -1.48563397e+00
-1.10237733e-01 -8.71851265e-01 4.49763983e-01 -7.15205729e-01
-8.01463425e-01 1.36973456e-01 1.22609806e+00 -9.45916474e-01
1.58753526e+00 -2.36302233e+00 2.65529186e-01 -1.09503657e-01
-1.46539748e-01 5.93047023e-01 -4.28105056e-01 3.68337929e-01
-4.73973155e-01 2.50379235e-01 -2.34824438e-02 -6.49602652e-01
8.75305682e-02 -1.09799542e-01 -9.88707066e-01 5.59022844e-01
2.51245558e-01 2.46008247e-01 -9.60614979e-01 -2.40200460e-01
-1.37119055e-01 4.52770203e-01 -9.41690207e-01 2.35333517e-01
-2.33977705e-01 4.52320188e-01 -1.83432698e-02 6.22935474e-01
7.14024782e-01 2.20755517e-01 -5.14369719e-02 -2.37833813e-01
3.62765133e-01 6.83129847e-01 -9.67890084e-01 1.55889857e+00
-5.06823242e-01 7.04082012e-01 3.08071971e-01 -1.29330599e+00
8.03431094e-01 7.10309386e-01 3.42501640e-01 -4.08101141e-01
2.72010863e-01 4.18197900e-01 -1.73749421e-02 -7.19322443e-01
3.85611773e-01 -3.55937809e-01 -1.58822998e-01 1.65410310e-01
5.28383315e-01 -3.88443381e-01 -1.53198600e-01 -1.30534902e-01
1.08017492e+00 -2.06612140e-01 -4.01708215e-01 1.63695320e-01
6.86024189e-01 -3.84476274e-01 7.31778622e-01 1.02274907e+00
-3.23910117e-01 6.70951009e-01 3.62807125e-01 6.10523447e-02
-7.58491576e-01 -9.56788599e-01 -2.89447337e-01 8.91803086e-01
5.12398593e-02 -6.48992181e-01 -7.28478611e-01 -9.48983967e-01
4.97897863e-02 9.26210046e-01 -1.96549684e-01 -6.41289294e-01
-7.07804918e-01 -7.70366251e-01 1.19164312e+00 -4.06197365e-03
5.15291214e-01 -8.70544076e-01 1.10986248e-01 4.03720737e-01
-8.85537326e-01 -1.58293903e+00 -5.70642829e-01 1.20381311e-01
-4.76371497e-01 -7.73865283e-01 -3.96649420e-01 -7.55408108e-01
-2.46894136e-01 2.76971012e-01 8.49111438e-01 -1.20619193e-01
1.19613316e-02 6.04071394e-02 -5.54371178e-01 -6.28634512e-01
-8.44386995e-01 1.32965818e-01 5.07323384e-01 4.35807139e-01
-4.29945700e-02 -6.52472138e-01 -1.37325883e-01 5.26947081e-01
-9.15229380e-01 -3.35258901e-01 2.10298181e-01 1.23734164e+00
2.05198199e-01 4.16542768e-01 8.54925752e-01 -6.28961325e-01
8.75849426e-01 -5.37814856e-01 -3.73684406e-01 -5.23313768e-02
-3.98640960e-01 4.37409617e-02 6.67869508e-01 -9.86863315e-01
-7.55537629e-01 -4.67792124e-01 -3.64493310e-01 -7.54335523e-01
1.45490199e-01 2.44839922e-01 -5.41598082e-01 9.22224671e-02
6.64168060e-01 1.56307444e-01 -3.46799552e-01 -7.46694446e-01
1.87856436e-01 1.16818917e+00 7.37203956e-01 -5.89341104e-01
1.05607092e+00 3.94349366e-01 -3.25957447e-01 -1.00578690e+00
-9.43184674e-01 -4.06023234e-01 2.81511787e-02 1.69606894e-01
6.86061561e-01 -8.97006094e-01 -4.74990755e-01 9.12282050e-01
-1.55417395e+00 -1.32401753e-02 2.99168322e-02 7.58046985e-01
-7.20858574e-01 6.57582462e-01 -6.13596976e-01 -7.15934396e-01
-2.61831880e-01 -1.35487413e+00 1.11206293e+00 -6.44351840e-01
-1.41278598e-02 -6.66333139e-01 1.84638515e-01 6.69313014e-01
2.98355132e-01 1.43620864e-01 8.52881610e-01 -1.04028428e+00
-1.87325358e-01 -2.56497651e-01 1.40024349e-01 1.13329601e+00
1.04384288e-01 -3.82631749e-01 -1.53806067e+00 -6.51016831e-01
4.57409173e-01 -6.85051560e-01 1.16232944e+00 3.64582129e-02
1.37669528e+00 -6.54351056e-01 -2.23138720e-01 9.29680645e-01
8.82941663e-01 1.48064923e-02 4.96583432e-01 2.03284562e-01
5.67933083e-01 4.82187659e-01 5.50680995e-01 2.98033148e-01
-2.08644509e-01 1.10644794e+00 5.48620164e-01 2.59200096e-01
-2.49614105e-01 -4.71496671e-01 7.48595297e-01 1.36015284e+00
2.96142608e-01 -8.35248649e-01 -4.57746804e-01 4.71015930e-01
-1.48091769e+00 -1.18333280e+00 2.15223357e-01 2.22360134e+00
7.48407841e-01 2.84634054e-01 9.03174654e-03 7.94677913e-01
8.74304235e-01 6.56549931e-01 -1.02132842e-01 -4.42300111e-01
-5.71106195e-01 2.32525945e-01 3.78189623e-01 4.35270846e-01
-1.55969226e+00 1.06309640e+00 5.90935946e+00 1.43117273e+00
-1.36317956e+00 5.40092051e-01 2.93757200e-01 -1.43169910e-01
-1.94925174e-01 -8.40602592e-02 -7.11433589e-01 7.66696274e-01
1.05114782e+00 1.27928585e-01 7.63951778e-01 4.14996803e-01
2.77997583e-01 8.32279444e-01 -8.96008313e-01 1.14899826e+00
3.60859305e-01 -1.06129539e+00 5.88297360e-02 -2.03888137e-02
6.19950294e-01 -1.28926560e-01 3.40200007e-01 5.11763632e-01
-1.84074312e-01 -1.02103186e+00 1.10516620e+00 -2.30077766e-02
8.07567894e-01 -6.89944208e-01 7.05143392e-01 3.75008464e-01
-7.58434951e-01 -3.70695531e-01 -2.56768882e-01 4.43015397e-01
2.46812135e-01 3.90790045e-01 -9.14070547e-01 6.70044720e-01
6.07321799e-01 5.05606115e-01 -2.43383139e-01 7.67172456e-01
-3.20826739e-01 1.11383224e+00 -1.59019068e-01 2.85969645e-01
4.49987054e-02 4.47317004e-01 1.22248387e+00 1.19553006e+00
2.72767574e-01 -5.74617624e-01 -5.15745766e-02 7.18733191e-01
-3.37378979e-01 -1.13034599e-01 -6.19971514e-01 -2.90250033e-01
6.13954306e-01 6.76554084e-01 1.40983865e-01 -3.86605039e-03
-6.96603656e-02 9.83058691e-01 2.84711495e-02 4.08468843e-01
-1.27032924e+00 -6.65916085e-01 7.85732210e-01 -3.54179665e-02
4.13244039e-01 9.90659446e-02 2.61797905e-01 -1.49847639e+00
4.02191766e-02 -1.72326243e+00 3.07964951e-01 -6.16297603e-01
-1.29698467e+00 8.69917095e-01 -1.20076433e-01 -1.38609493e+00
-1.07521169e-01 -5.71518242e-01 -3.00530851e-01 4.76149112e-01
-1.40149486e+00 -1.02286065e+00 1.53265730e-01 5.47274292e-01
6.21118903e-01 -4.67878759e-01 7.97591090e-01 6.16825342e-01
-6.72051191e-01 1.18760490e+00 6.32318705e-02 5.02360016e-02
1.06542301e+00 -8.21897507e-01 2.96277016e-01 7.84056664e-01
3.38621914e-01 3.96873921e-01 1.05104876e+00 -5.00780106e-01
-9.80672181e-01 -1.12066007e+00 8.92034233e-01 -3.32390100e-01
9.05751348e-01 -7.36998379e-01 -9.36440647e-01 3.97448719e-01
2.76744366e-01 -1.07680708e-02 4.46886480e-01 -2.33442634e-01
-8.62604141e-01 -2.20464006e-01 -1.15234494e+00 3.16052973e-01
1.19523060e+00 -9.19878721e-01 -5.99629998e-01 4.89239216e-01
1.31691921e+00 -3.32083672e-01 -5.78398645e-01 5.58924854e-01
4.80929464e-01 -8.22950840e-01 1.34327316e+00 -1.06848991e+00
1.81760117e-01 1.09101720e-02 -5.44298768e-01 -1.64470804e+00
-4.33812886e-02 -1.22161782e+00 -1.82173252e-01 1.52470911e+00
4.12369728e-01 -6.70683086e-01 3.32088619e-01 -7.29043603e-01
-3.17052066e-01 -3.63206327e-01 -1.20192814e+00 -1.33322334e+00
1.11900419e-01 -8.07140052e-01 7.89002597e-01 1.18879128e+00
-2.30099887e-01 2.73792237e-01 -1.07895744e+00 5.63383520e-01
6.17931306e-01 -3.94638598e-01 7.65629292e-01 -8.29758763e-01
-8.16694736e-01 -2.30548635e-01 -4.23719913e-01 -1.06764817e+00
6.48750722e-01 -9.74421680e-01 -1.91121876e-01 -3.16673338e-01
-3.52741539e-01 -2.64667213e-01 -3.90089154e-01 -2.87704244e-02
-1.05967902e-01 1.86365411e-01 2.69442588e-01 1.74412802e-02
-1.76013842e-01 8.73007596e-01 1.05108202e+00 -7.33339787e-01
2.03306843e-02 2.63596743e-01 -3.94881427e-01 6.37460232e-01
6.94765389e-01 -1.00638008e+00 -3.45596433e-01 -4.15916145e-01
-3.94245274e-02 3.42364728e-01 3.16778332e-01 -1.21034408e+00
-3.85890484e-01 3.64248604e-02 -3.06540549e-01 -3.55964661e-01
7.03950822e-01 -7.49841273e-01 -1.03905439e-01 4.13008332e-01
-5.01797140e-01 -8.10171515e-02 1.36677802e-01 9.78800118e-01
-5.30273855e-01 -1.55555591e-01 7.65770853e-01 2.37355724e-01
4.24413849e-03 1.61286164e-02 -2.12814718e-01 4.07800227e-01
4.39095944e-01 1.76527157e-01 -5.93625188e-01 -5.33103526e-01
-6.56883836e-01 -1.51204303e-01 1.67943959e-04 8.52444470e-01
6.01425886e-01 -1.41793346e+00 -9.18663085e-01 2.71255553e-01
1.58028439e-01 -6.44213200e-01 1.55543804e-01 7.93437541e-01
-1.96093634e-01 4.98053402e-01 1.17886938e-01 -3.96564573e-01
-1.30486941e+00 8.91752422e-01 3.35557371e-01 -4.36250567e-01
-2.17475370e-01 1.05384421e+00 6.60187304e-02 -4.40732718e-01
6.45139039e-01 -2.29755133e-01 8.37852359e-02 3.15586068e-02
2.82281756e-01 5.96180737e-01 3.28646988e-01 -8.83597732e-01
-3.71382087e-01 2.21499845e-01 6.84659779e-02 -2.38801733e-01
1.22751558e+00 -9.64129195e-02 3.18805665e-01 2.40184039e-01
1.58908534e+00 4.76014882e-01 -7.52514541e-01 -2.30826229e-01
4.38578241e-02 -6.00871027e-01 1.93702221e-01 -6.89781725e-01
-1.14008689e+00 1.04240501e+00 6.67689204e-01 3.84123653e-01
1.12571931e+00 -2.59908825e-01 1.14726996e+00 1.86717600e-01
2.59487659e-01 -1.02021384e+00 5.42538106e-01 4.18842882e-01
1.25996375e+00 -1.01522708e+00 -4.09089714e-01 -5.69678366e-01
-5.03924668e-01 8.32817852e-01 3.65959823e-01 1.21455424e-01
5.23808300e-01 8.59066397e-02 1.38479128e-01 1.40232682e-01
-7.88290083e-01 1.44944072e-01 3.42483908e-01 7.11062610e-01
1.89783778e-02 5.08790798e-02 -2.61221349e-01 7.84383476e-01
-5.57747126e-01 -5.46146214e-01 4.05018151e-01 4.82860863e-01
-6.06774352e-02 -1.13525903e+00 -6.42961204e-01 2.62396447e-02
-9.28073406e-01 -9.16897431e-02 -5.91448367e-01 4.80618179e-01
3.00216883e-01 1.52390635e+00 -3.52199733e-01 -1.05669761e+00
3.40554476e-01 5.17117716e-02 3.68195206e-01 -4.54255551e-01
-7.28084087e-01 2.22828299e-01 4.68775392e-01 -4.39852357e-01
-2.85093307e-01 -3.40741158e-01 -3.29449236e-01 -1.35208681e-01
-1.01178038e+00 1.02446958e-01 6.20852530e-01 7.19674766e-01
1.92408964e-01 6.05630040e-01 1.05064094e+00 -7.04374909e-01
-1.24315035e+00 -1.11536622e+00 -5.33958256e-01 7.62892127e-01
8.48624766e-01 -8.24669242e-01 -1.07508838e+00 -2.37908587e-01] | [14.094097137451172, 5.825326442718506] |
cc01d60b-834d-4e3d-ba85-c12ea4fbde1c | ego2hands-a-dataset-for-egocentric-two-hand | 2011.07252 | null | https://arxiv.org/abs/2011.07252v3 | https://arxiv.org/pdf/2011.07252v3.pdf | Ego2Hands: A Dataset for Egocentric Two-hand Segmentation and Detection | Hand segmentation and detection in truly unconstrained RGB-based settings is important for many applications. However, existing datasets are far from sufficient both in terms of size and variety due to the infeasibility of manual annotation of large amounts of segmentation and detection data. As a result, current methods are limited by many underlying assumptions such as constrained environment, consistent skin color and lighting. In this work, we present a large-scale RGB-based egocentric hand segmentation/detection dataset Ego2Hands that is automatically annotated and a color-invariant compositing-based data generation technique capable of creating unlimited training data with variety. For quantitative analysis, we manually annotated an evaluation set that significantly exceeds existing benchmarks in quantity, diversity and annotation accuracy. We provide cross-dataset evaluation as well as thorough analysis on the performance of state-of-the-art models on Ego2Hands to show that our dataset and data generation technique can produce models that generalize to unseen environments without domain adaptation. | ['Tony Martinez', 'Fanqing Lin'] | 2020-11-14 | null | null | null | null | ['hand-segmentation'] | ['computer-vision'] | [ 2.05527753e-01 -2.15662867e-01 1.41918585e-01 -4.58769590e-01
-8.78258944e-01 -1.02580643e+00 2.43341297e-01 -3.15887272e-01
-3.23848605e-01 6.26986146e-01 -2.50046644e-02 1.35749439e-02
1.84416369e-01 -4.45006847e-01 -5.26911378e-01 -5.40360332e-01
2.45634556e-01 7.70315111e-01 4.83534694e-01 -1.39071822e-01
-6.48341700e-02 5.56210220e-01 -1.60286903e+00 6.92464039e-02
5.30263364e-01 8.73736203e-01 6.51177242e-02 9.60336447e-01
1.30890131e-01 3.28680366e-01 -4.74283129e-01 -2.30264544e-01
6.50737524e-01 -4.17427927e-01 -9.53923702e-01 3.81157786e-01
9.50737119e-01 -7.83433795e-01 -4.00685489e-01 7.69153655e-01
1.00302029e+00 2.52642244e-01 5.81125200e-01 -1.42574966e+00
-5.33478677e-01 1.72210038e-01 -6.51034057e-01 -1.35335520e-01
5.73259592e-01 6.80945873e-01 8.79009366e-01 -5.51466286e-01
1.13000405e+00 1.00340986e+00 5.56855917e-01 1.00922906e+00
-1.41982627e+00 -4.55988526e-01 3.49291950e-01 -2.11068049e-01
-1.35383999e+00 -3.55276287e-01 8.75908077e-01 -5.28579652e-01
9.00264680e-01 3.34648162e-01 1.08115101e+00 1.51864970e+00
-5.02289116e-01 1.25641477e+00 1.41119909e+00 -4.76142406e-01
4.22313541e-01 1.02263957e-01 -3.15229632e-02 7.40272105e-01
9.53411758e-02 -1.51611879e-01 -5.44944406e-01 1.41796544e-01
1.14392543e+00 -3.92767787e-01 -5.80789149e-02 -7.91708231e-01
-1.25009823e+00 4.08935130e-01 5.81943154e-01 -7.10470006e-02
-2.82364696e-01 3.85186285e-01 3.66301984e-01 -2.40473479e-01
2.90384352e-01 3.70093018e-01 -6.95061564e-01 -2.41885468e-01
-1.10692358e+00 5.25145650e-01 6.81077182e-01 1.47133613e+00
4.62334722e-01 3.68389413e-02 -1.38789982e-01 5.71490943e-01
2.27480024e-01 5.29757440e-01 2.72716135e-01 -1.21965885e+00
4.74622786e-01 5.63377023e-01 2.03299195e-01 -4.76759523e-01
-5.57818174e-01 -1.44740045e-01 -7.97756314e-02 4.76837516e-01
9.68884408e-01 -2.59898603e-01 -1.39428151e+00 1.68871248e+00
5.52441061e-01 -4.06980515e-01 -2.79102117e-01 1.19508207e+00
8.14825535e-01 -7.72464797e-02 2.92858273e-01 2.79263288e-01
1.21580899e+00 -9.15730059e-01 -4.37089711e-01 -2.86874771e-01
3.96638155e-01 -6.48835242e-01 1.65337241e+00 5.13512135e-01
-9.61636901e-01 -3.79968077e-01 -6.70710921e-01 -2.12272406e-01
-5.37002206e-01 2.86214165e-02 1.02506733e+00 1.15642214e+00
-9.57178712e-01 3.36309612e-01 -8.58705640e-01 -8.61404240e-01
7.15469003e-01 1.99743733e-01 -4.69879687e-01 -1.97602659e-02
-5.78704834e-01 5.21961868e-01 5.37501693e-01 2.51118261e-02
-8.24320674e-01 -4.24985379e-01 -6.45901740e-01 -5.28104782e-01
4.18415010e-01 -5.52553117e-01 1.41262960e+00 -7.40089536e-01
-1.29999804e+00 1.06107152e+00 4.07459736e-02 1.62027702e-02
1.09274030e+00 -1.56297952e-01 1.14500569e-02 2.52746016e-01
1.15409926e-01 9.85958040e-01 6.88552141e-01 -1.51408732e+00
-5.13165832e-01 -7.28439867e-01 1.24701329e-01 2.71215141e-01
-1.91853493e-01 -9.63863432e-02 -7.64124811e-01 -6.20833635e-01
6.14005737e-02 -1.11929178e+00 -1.74562752e-01 7.36858770e-02
-7.99367964e-01 5.44793159e-03 8.40432882e-01 -7.67809093e-01
7.26546943e-01 -1.89518046e+00 1.17610544e-02 2.21427679e-01
1.60051420e-01 9.28316489e-02 -2.08672494e-01 9.10489038e-02
1.70904130e-01 1.93262771e-01 -1.43208697e-01 -4.86795574e-01
1.77316740e-01 4.46050130e-02 -6.22676965e-03 4.57660109e-01
-1.63294934e-02 9.93620098e-01 -1.05164623e+00 -1.01952720e+00
5.12004495e-01 5.28572202e-01 -5.43630481e-01 2.52415359e-01
-5.99032581e-01 6.62673950e-01 -4.83578473e-01 1.23592472e+00
2.92324722e-01 -7.20567033e-02 -3.63733396e-02 -2.78624475e-01
1.79934755e-01 -1.36581719e-01 -1.31106365e+00 2.32276821e+00
-2.35340908e-01 6.21389985e-01 -3.97935584e-02 -3.21730003e-02
3.34106803e-01 2.59388536e-01 6.57513201e-01 -4.01805133e-01
4.57946688e-01 7.24470541e-02 -1.88956395e-01 -4.52866346e-01
5.47380447e-01 2.72637725e-01 1.98833994e-03 6.91149473e-01
-4.16555367e-02 -6.45444512e-01 2.87936211e-01 2.53868103e-01
8.77790093e-01 7.75863826e-01 -1.08335555e-01 1.28403351e-01
-1.98117390e-01 4.84441638e-01 1.55103907e-01 6.39025211e-01
-7.16525495e-01 1.11379337e+00 1.40167579e-01 -2.04495072e-01
-1.35650265e+00 -1.20465326e+00 -1.80234373e-01 1.19241345e+00
1.48138896e-01 -1.67971030e-01 -1.36482108e+00 -7.70207107e-01
-2.96430495e-02 5.56756556e-01 -7.19097137e-01 3.88371080e-01
-2.52837330e-01 -6.44295573e-01 6.59572005e-01 1.17886829e+00
6.17536426e-01 -1.18927169e+00 -9.68564391e-01 5.83372004e-02
-9.12750587e-02 -1.34983587e+00 -5.11847079e-01 -4.25195843e-02
-7.20839441e-01 -1.40567529e+00 -8.10773969e-01 -6.35717034e-01
5.99375248e-01 1.86306059e-01 1.03519571e+00 -3.11698109e-01
-9.89089727e-01 9.20522332e-01 -4.33858812e-01 -4.29610819e-01
-7.43116364e-02 2.21456632e-01 -2.00755261e-02 -5.23413897e-01
5.87904036e-01 -2.38875493e-01 -9.34276819e-01 3.06404084e-01
-6.06746614e-01 -7.17964396e-02 3.80157083e-01 4.66386616e-01
7.09687054e-01 -3.64057899e-01 3.12643111e-01 -7.08400846e-01
5.24051845e-01 1.20407909e-01 -4.13899094e-01 3.77023697e-01
-5.33722281e-01 -2.06379458e-01 7.20542744e-02 -3.42287004e-01
-1.24141979e+00 7.22826421e-01 9.83829573e-02 -2.92266548e-01
-6.07444823e-01 -3.22909325e-01 -1.63364291e-01 -1.87518999e-01
1.07505834e+00 9.18437913e-02 -2.87810355e-01 -4.48766261e-01
9.36694741e-01 8.86711299e-01 7.10969925e-01 -8.05394113e-01
5.53522229e-01 5.99509358e-01 -2.44377226e-01 -7.70702720e-01
-4.51230437e-01 -4.85966951e-01 -1.24924242e+00 -4.26886946e-01
1.05515838e+00 -8.00117671e-01 -6.05366349e-01 8.06419730e-01
-9.32654202e-01 -8.00378740e-01 -4.33312476e-01 2.28593543e-01
-8.70021224e-01 3.32731366e-01 -4.99334037e-01 -1.07152474e+00
-3.28914732e-01 -1.09714925e+00 1.32493865e+00 1.27733350e-01
-4.87000376e-01 -7.88208842e-01 -1.07827261e-01 6.36876523e-01
1.24719650e-01 7.25439787e-01 3.66623998e-01 -4.02701586e-01
-6.69330716e-01 -5.30787230e-01 -4.45636451e-01 2.24930927e-01
2.03697801e-01 8.45135152e-02 -1.39250720e+00 -2.55316615e-01
-7.56644487e-01 -7.55614698e-01 4.29469496e-01 2.60947227e-01
1.28824723e+00 1.68094039e-01 -2.88217217e-01 5.42373776e-01
1.26640260e+00 -7.21471310e-02 4.04379517e-01 3.69166344e-01
1.11215174e+00 6.12528086e-01 4.76733834e-01 4.07467186e-01
3.07873458e-01 5.58621466e-01 3.59741688e-01 -1.60740480e-01
-5.51877558e-01 -2.78547853e-01 -3.14200163e-01 -8.13770220e-02
-5.47049582e-01 -2.78847158e-01 -9.28754807e-01 7.55520344e-01
-1.61904705e+00 -7.13297188e-01 -9.69981104e-02 2.03843689e+00
9.94809508e-01 -1.28840894e-01 8.63902032e-01 1.97544888e-01
6.18235648e-01 -1.33730695e-01 -8.94168377e-01 -8.54955241e-02
-5.96113540e-02 2.03276947e-01 6.22248173e-01 1.22392178e-01
-1.24058819e+00 1.28576648e+00 7.08582306e+00 4.45270449e-01
-7.19297588e-01 1.54200569e-01 4.23115224e-01 -3.32092285e-01
9.49070379e-02 -4.51117128e-01 -6.00154877e-01 1.43021211e-01
1.09245725e-01 5.91312468e-01 8.66342306e-01 1.31173015e+00
-3.28707248e-02 -8.96893293e-02 -1.27297568e+00 1.20409358e+00
1.02948457e-01 -7.22368121e-01 -2.66320229e-01 -7.21810432e-03
9.20548201e-01 1.81939248e-02 -2.52884608e-02 -1.88330516e-01
5.20369470e-01 -8.90432060e-01 1.12769294e+00 1.01095453e-01
1.06589627e+00 -3.24092954e-01 3.34784329e-01 -1.52891427e-01
-1.05706668e+00 1.95183903e-01 1.61497712e-01 1.98546454e-01
1.08510934e-01 -1.08619362e-01 -8.95416558e-01 -1.85216982e-02
9.31994498e-01 2.28143021e-01 -7.15149701e-01 7.60506570e-01
-4.78504002e-01 3.18336755e-01 -6.13646448e-01 -1.36630341e-01
1.36893243e-01 1.10724680e-01 2.96992779e-01 1.31040514e+00
-1.72066465e-01 2.87255906e-02 2.57499933e-01 9.81308222e-01
-4.75515164e-02 9.96886119e-02 -4.89999384e-01 -6.22622818e-02
5.65607667e-01 1.24775910e+00 -1.14301896e+00 -2.08179846e-01
-2.48801470e-01 1.41124439e+00 2.06817120e-01 5.94869912e-01
-6.42482102e-01 -3.36399436e-01 5.69287658e-01 2.33773485e-01
4.02937904e-02 -5.29945314e-01 -6.87699974e-01 -1.02370250e+00
1.46794513e-01 -6.97707534e-01 3.53921622e-01 -9.31182683e-01
-1.33026385e+00 5.79891741e-01 4.11411524e-02 -8.72011006e-01
-3.36235076e-01 -9.79287386e-01 -1.04955547e-01 7.19669342e-01
-1.09166467e+00 -1.81307662e+00 -9.74000514e-01 8.14251304e-01
6.38996482e-01 -7.82956034e-02 8.54786992e-01 6.58139363e-02
-6.40760064e-01 7.29831576e-01 -1.39897093e-01 4.65150088e-01
7.51197457e-01 -1.80226827e+00 5.73706210e-01 7.87031054e-01
1.45177886e-01 6.89044476e-01 6.13966405e-01 -6.21040463e-01
-1.48487377e+00 -8.98263395e-01 -4.73979898e-02 -1.06420457e+00
4.25333291e-01 -5.65893471e-01 -3.26899529e-01 1.01098335e+00
-4.69586626e-02 1.69387043e-01 5.73098242e-01 1.57191664e-01
-5.75403988e-01 1.41673326e-01 -1.52573824e+00 7.53464520e-01
1.50031877e+00 -5.66272378e-01 -3.01528305e-01 6.02301300e-01
3.00776035e-01 -7.60240257e-01 -8.36782694e-01 -2.40529049e-02
1.11195362e+00 -9.48005497e-01 1.05834162e+00 -8.55225980e-01
1.56624421e-01 -2.81909972e-01 -4.70546246e-01 -9.74470556e-01
1.41424481e-02 -6.07935309e-01 -1.12292049e-02 1.30979276e+00
2.23135665e-01 -3.33365768e-01 1.23974621e+00 1.34531832e+00
1.30485728e-01 -3.34381193e-01 -6.77907944e-01 -9.53753531e-01
-5.83129153e-02 -7.55457759e-01 6.28367126e-01 6.50767624e-01
-1.19899146e-01 -2.10667737e-02 5.80934882e-02 -1.29470363e-01
1.00197399e+00 5.01667410e-02 1.23782182e+00 -9.81474578e-01
-1.47668689e-01 -4.84563768e-01 -5.24970055e-01 -7.20710099e-01
-1.62527964e-01 -4.42351490e-01 3.55183750e-01 -1.80110252e+00
2.81822056e-01 -7.77136147e-01 1.80296600e-02 6.55287683e-01
-1.02770172e-01 8.79983068e-01 2.29803041e-01 2.36072093e-01
-4.83711421e-01 7.60383308e-02 1.32003582e+00 -1.40581802e-01
-4.82565790e-01 -3.33874345e-01 -4.51891363e-01 7.65078723e-01
8.44149828e-01 -3.24558537e-03 -4.85742003e-01 -4.36828256e-01
-3.32792550e-02 -6.11184239e-01 6.82741880e-01 -9.55964684e-01
-1.65374175e-01 -2.31172770e-01 7.02197671e-01 -4.20288235e-01
5.03846169e-01 -7.84086108e-01 -1.71740621e-01 1.47696137e-01
-2.08664864e-01 -3.28214765e-01 1.32642090e-01 4.93729562e-01
4.39978898e-01 9.56982225e-02 6.13723457e-01 -4.73566324e-01
-1.15609908e+00 3.73720199e-01 8.74321163e-02 2.84074426e-01
1.19786084e+00 -6.65718257e-01 -2.64562041e-01 -2.53258795e-01
-7.50363410e-01 1.00640720e-03 9.78806138e-01 4.88371134e-01
3.48914444e-01 -9.85030293e-01 -4.23740059e-01 3.52280475e-02
4.41064775e-01 3.80713016e-01 4.58121225e-02 4.33513284e-01
-8.52490783e-01 2.31916055e-01 -3.93360287e-01 -6.95882201e-01
-1.25265694e+00 4.87900257e-01 2.53471941e-01 4.02603507e-01
-6.39705539e-01 9.66325939e-01 -1.26962408e-01 -5.65725982e-01
4.79633301e-01 -3.86572063e-01 5.13960540e-01 -1.31720185e-01
3.33818227e-01 6.54735982e-01 -9.57861841e-02 -6.45450234e-01
-4.84826177e-01 5.99663317e-01 2.27088585e-01 -3.57650876e-01
1.01026154e+00 3.58995460e-02 3.37880641e-01 1.54327765e-01
8.20030391e-01 -2.11004913e-01 -1.66078210e+00 1.37865037e-01
-4.53084737e-01 -7.26206183e-01 -1.49377715e-02 -1.25216353e+00
-1.05208659e+00 9.90149200e-01 9.65119481e-01 8.60853773e-03
9.97169316e-01 2.94135869e-01 9.08044338e-01 3.75686646e-01
6.12178802e-01 -1.63791764e+00 3.46744597e-01 6.24811836e-03
8.12416673e-01 -1.39924967e+00 -3.69458634e-04 -5.92000127e-01
-7.70720601e-01 8.71453285e-01 9.57063138e-01 1.96006238e-01
2.06539169e-01 2.26277366e-01 3.39611679e-01 -2.97643214e-01
3.03454995e-01 -5.03341258e-01 1.18279658e-01 1.14533782e+00
3.64519387e-01 2.42360115e-01 1.58169851e-01 2.71744221e-01
-3.71531695e-01 5.44112250e-02 1.39382303e-01 1.22600579e+00
-8.28425586e-02 -8.53071213e-01 -4.71727699e-01 3.74361336e-01
-1.62712604e-01 2.70205736e-01 -8.22568476e-01 1.15554750e+00
1.50748476e-01 9.33488846e-01 -1.56699926e-01 -4.36529160e-01
4.08830583e-01 1.53934106e-01 1.17009628e+00 -6.58439696e-01
-3.11198324e-01 7.56848976e-02 -2.45436151e-02 -6.21214747e-01
-3.85547310e-01 -8.90133917e-01 -1.32458496e+00 -1.91582382e-01
-5.41364193e-01 -6.13492191e-01 9.87609506e-01 8.92234087e-01
1.56347811e-01 5.12611449e-01 5.79291806e-02 -1.14432478e+00
-4.39772248e-01 -1.00573730e+00 -8.09154630e-01 7.56288469e-01
-5.90598397e-02 -6.55566216e-01 7.80956028e-03 3.28439713e-01] | [6.699012756347656, -0.6924402713775635] |
c38c1830-f34e-4793-8d7d-f413d38f1c68 | textbf-p-2-a-a-dataset-and-benchmark-for | 2207.12730 | null | https://arxiv.org/abs/2207.12730v1 | https://arxiv.org/pdf/2207.12730v1.pdf | $\textbf{P$^2$A}$: A Dataset and Benchmark for Dense Action Detection from Table Tennis Match Broadcasting Videos | While deep learning has been widely used for video analytics, such as video classification and action detection, dense action detection with fast-moving subjects from sports videos is still challenging. In this work, we release yet another sports video dataset $\textbf{P$^2$A}$ for $\underline{P}$ing $\underline{P}$ong-$\underline{A}$ction detection, which consists of 2,721 video clips collected from the broadcasting videos of professional table tennis matches in World Table Tennis Championships and Olympiads. We work with a crew of table tennis professionals and referees to obtain fine-grained action labels (in 14 classes) for every ping-pong action that appeared in the dataset and formulate two sets of action detection problems - action localization and action recognition. We evaluate a number of commonly-seen action recognition (e.g., TSM, TSN, Video SwinTransformer, and Slowfast) and action localization models (e.g., BSN, BSN++, BMN, TCANet), using $\textbf{P$^2$A}$ for both problems, under various settings. These models can only achieve 48% area under the AR-AN curve for localization and 82% top-one accuracy for recognition since the ping-pong actions are dense with fast-moving subjects but broadcasting videos are with only 25 FPS. The results confirm that $\textbf{P$^2$A}$ is still a challenging task and can be used as a benchmark for action detection from videos. | ['Dejing Dou', 'Feixiang Lu', 'Jun Zhao', 'Jun Cheng', 'Xuhong LI', 'Chen Liu', 'Jun Huang', 'Haoyi Xiong', 'Qingzhong Wang', 'Jiang Bian'] | 2022-07-26 | null | null | null | null | ['video-classification', 'action-localization'] | ['computer-vision', 'computer-vision'] | [ 1.91716328e-01 -3.21780145e-01 -3.87085319e-01 -5.14572263e-02
-7.84699261e-01 -3.59813660e-01 1.07008614e-01 -3.12593013e-01
-7.75543272e-01 5.12355089e-01 -6.58288747e-02 7.83870071e-02
-1.58557624e-01 -5.21629930e-01 -1.01933849e+00 -6.98796451e-01
-6.66802347e-01 4.93420176e-02 7.50561476e-01 -2.85507470e-01
1.17235884e-01 4.38309729e-01 -1.97453821e+00 6.71110570e-01
2.48874217e-01 1.44488609e+00 -1.10359259e-01 9.20485377e-01
3.28000605e-01 1.26444757e+00 -7.49178588e-01 -5.31400025e-01
6.18523777e-01 -5.80334187e-01 -5.92978835e-01 5.77779189e-02
7.51534700e-01 -5.43942034e-01 -9.16390479e-01 8.53775620e-01
3.89922470e-01 4.50867593e-01 2.12253302e-01 -1.50974238e+00
2.46601701e-01 3.63326728e-01 -7.24813581e-01 8.69158626e-01
6.39280200e-01 3.29191417e-01 7.54686654e-01 -6.94935322e-01
6.23233140e-01 1.14299965e+00 6.65527761e-01 3.44228208e-01
-6.01296604e-01 -1.02439392e+00 1.99944302e-01 7.38963008e-01
-1.41846454e+00 -3.16987991e-01 3.50341707e-01 -4.31739360e-01
8.85552943e-01 4.27506477e-01 1.03583813e+00 1.08961117e+00
2.61994600e-01 1.19028747e+00 6.41088009e-01 -1.11932896e-01
1.36575282e-01 -4.74060714e-01 -1.79702967e-01 6.79435790e-01
-1.87870935e-01 9.90399122e-02 -1.12024069e+00 3.14915568e-01
1.01109111e+00 1.20492399e-01 -1.02701217e-01 1.55963764e-01
-1.35450375e+00 6.29792154e-01 2.95350224e-01 1.18974566e-01
-3.55520934e-01 4.90059108e-01 6.15736604e-01 3.65319401e-01
1.67169333e-01 5.94373569e-02 -2.41673067e-01 -1.13399148e+00
-8.89511466e-01 4.43708897e-01 4.16957557e-01 1.09760857e+00
5.05510092e-01 2.83348173e-01 -2.87809730e-01 7.24768221e-01
-3.29137295e-01 6.70309484e-01 2.81907022e-01 -1.40487134e+00
6.64122999e-01 5.57694197e-01 2.48902529e-01 -1.18256247e+00
-5.07610500e-01 9.12305200e-04 -5.74773848e-01 6.79471269e-02
7.32915759e-01 -2.62789458e-01 -7.09080219e-01 1.32707834e+00
1.78256512e-01 3.87817413e-01 -3.61930490e-01 1.25839961e+00
8.70513022e-01 4.90161926e-01 1.69921547e-01 -1.32577956e-01
1.52788508e+00 -9.95313704e-01 -3.69854301e-01 -3.41851950e-01
7.25288153e-01 -5.27837396e-01 1.00113511e+00 7.42249310e-01
-1.09732604e+00 -8.49733829e-01 -7.65657306e-01 1.87831163e-01
4.06331979e-02 2.57502049e-01 6.30159497e-01 6.29626632e-01
-7.34990060e-01 6.00251496e-01 -1.00913048e+00 -3.17690074e-01
5.68353117e-01 4.01797473e-01 -6.26350164e-01 -2.84641504e-01
-1.06932795e+00 4.72914219e-01 1.78275824e-01 2.21004382e-01
-1.35868168e+00 -5.15357316e-01 -7.78747559e-01 -1.66366547e-01
9.44293797e-01 1.70366645e-01 1.03837883e+00 -1.11075199e+00
-1.24500144e+00 9.34246421e-01 2.88397163e-01 -5.76069951e-01
6.45608962e-01 -5.41677535e-01 -6.61516488e-01 5.75609446e-01
2.15904012e-01 6.02252424e-01 6.63020492e-01 -4.66989338e-01
-1.13107002e+00 -3.09010476e-01 4.83041942e-01 2.54014283e-01
-1.99451685e-01 6.03781402e-01 -6.92559898e-01 -6.86126590e-01
2.51267135e-01 -1.02395797e+00 1.22724630e-01 1.08656272e-01
-3.08748960e-01 -1.39545828e-01 6.61607444e-01 -7.45459378e-01
1.20755601e+00 -2.31666541e+00 6.55186772e-02 2.46724188e-02
1.38978928e-01 2.73887664e-01 -1.06593326e-01 3.17987800e-01
-4.11115680e-03 -2.47367755e-01 3.10765445e-01 7.09499493e-02
-2.57656090e-02 2.60697365e-01 2.17099324e-01 6.99514627e-01
-3.65562260e-01 6.28903031e-01 -7.59230971e-01 -5.44703960e-01
2.43512735e-01 1.97946250e-01 -4.28391308e-01 5.72196208e-04
5.01595102e-02 3.09139013e-01 -4.59938169e-01 1.01444876e+00
3.01556677e-01 1.63251311e-01 -3.57376481e-03 -2.15591699e-01
-2.65833080e-01 -7.49322399e-02 -1.61929607e+00 1.73178184e+00
1.43186957e-01 7.44128525e-01 3.84252042e-01 -1.30221343e+00
6.29075050e-01 1.62500918e-01 1.05104303e+00 -9.04875755e-01
2.80582637e-01 1.84284393e-02 9.10525098e-02 -8.15181792e-01
5.00315130e-01 -1.06322549e-01 -2.52466470e-01 3.01579058e-01
1.40938699e-01 3.63609523e-01 8.04446936e-01 2.17802167e-01
1.61823058e+00 3.26954305e-01 -8.77422839e-02 -5.95783927e-02
3.38088572e-01 3.14838886e-02 8.41261148e-01 1.13501477e+00
-6.16828322e-01 5.74342072e-01 8.37124109e-01 -7.59496331e-01
-6.92358971e-01 -8.40640128e-01 2.60285974e-01 1.64054751e+00
3.36004615e-01 -5.87398171e-01 -7.51278520e-01 -7.32086241e-01
-1.03940755e-01 8.31222814e-03 -5.76083541e-01 -2.05084532e-01
-1.01343644e+00 -4.63838726e-01 8.39686871e-01 9.60969806e-01
8.47438097e-01 -1.32902229e+00 -9.52730656e-01 1.16102405e-01
-3.48924637e-01 -1.43768013e+00 -4.74098682e-01 -5.54033257e-02
-4.43606019e-01 -1.39496779e+00 -7.68751264e-01 -5.65937757e-01
2.29346946e-01 3.49559665e-01 1.11150730e+00 -2.52620317e-02
-6.05899036e-01 5.86035550e-01 -8.34677279e-01 -2.91938543e-01
5.56266308e-02 -3.71403486e-01 2.80231178e-01 2.62043625e-01
5.90024173e-01 -3.21654886e-01 -8.41298044e-01 8.67777348e-01
-7.52784073e-01 -2.59910494e-01 6.01547182e-01 4.67130929e-01
6.42378569e-01 7.83511922e-02 -8.87371451e-02 -3.33533496e-01
-2.35660776e-01 -9.85866562e-02 -4.00260270e-01 -5.42040123e-03
2.54895568e-01 -7.05447793e-01 3.53150666e-01 -5.77455759e-01
-4.56620783e-01 2.68497281e-02 -2.60877311e-01 -8.34903002e-01
-2.48301104e-01 2.16076359e-01 -9.60391760e-02 -6.26170710e-02
7.78527975e-01 4.46704477e-01 -3.11443117e-03 -3.78539771e-01
-1.22842669e-01 3.82304460e-01 6.58255577e-01 -3.98337215e-01
4.00667965e-01 6.29977107e-01 -5.42314397e-03 -9.17924643e-01
-6.92721844e-01 -5.78562915e-01 -5.02575099e-01 -8.65723729e-01
1.08471036e+00 -1.19534910e+00 -1.25935018e+00 6.33694053e-01
-5.46732247e-01 -3.30374032e-01 -4.05295312e-01 1.02356517e+00
-7.32736349e-01 5.23137331e-01 -6.97863996e-01 -6.87790275e-01
2.52977833e-02 -1.11470807e+00 1.02457964e+00 1.43841341e-01
-7.43476227e-02 -1.90148845e-01 -2.96162486e-01 8.12404275e-01
1.80136748e-02 3.12831402e-01 1.03624865e-01 -4.09942985e-01
-5.41120470e-01 -6.04153574e-01 -1.24435179e-01 5.86499453e-01
-2.52681255e-01 -1.64693490e-01 -5.62912583e-01 -3.40930879e-01
-1.87562481e-01 -4.67776239e-01 7.81803429e-01 6.58492088e-01
1.35059643e+00 -8.92698690e-02 -1.79744110e-01 6.41611457e-01
8.14380288e-01 3.35251838e-01 9.94013548e-01 4.35385734e-01
5.63097060e-01 2.61715770e-01 1.00545239e+00 7.78302670e-01
2.53208168e-02 9.28331494e-01 5.87468684e-01 9.22918469e-02
-5.07994853e-02 4.02734429e-02 7.33296275e-01 1.62191108e-01
-8.72495949e-01 -1.33318782e-01 -7.78276920e-01 2.57793218e-01
-1.82686436e+00 -1.31298912e+00 -1.26229286e-01 2.10575676e+00
3.86592656e-01 3.99787456e-01 7.72923052e-01 3.03115189e-01
6.03896677e-01 2.98077464e-01 -4.32897031e-01 5.20216450e-02
-1.48790032e-01 2.78488517e-01 5.69446921e-01 -1.83083460e-01
-1.43981218e+00 7.85606325e-01 4.92457199e+00 1.15660644e+00
-9.85106170e-01 1.08731039e-01 4.28388000e-01 -7.40883291e-01
7.46925950e-01 -2.57335275e-01 -8.07000220e-01 7.52962351e-01
8.24374437e-01 2.57511973e-01 1.15307331e-01 1.10769999e+00
4.18928802e-01 -4.99156028e-01 -9.97727752e-01 1.45274436e+00
2.87571043e-01 -1.30563676e+00 -4.16704953e-01 -1.43132135e-01
3.12492162e-01 -9.99701768e-02 -2.16128394e-01 7.69104540e-01
-4.63863350e-02 -9.49518204e-01 9.22518969e-01 2.83914596e-01
9.11992192e-01 -5.82546234e-01 7.09412575e-01 3.08747143e-01
-1.42023039e+00 -5.13319612e-01 -2.12719157e-01 -4.43590701e-01
3.05410773e-01 1.82011157e-01 -1.22478316e-02 4.55972761e-01
1.33333409e+00 1.05880010e+00 -3.71725291e-01 1.00019634e+00
-1.22000039e-01 6.10235631e-01 -3.96881461e-01 9.23175178e-03
2.89463073e-01 -4.44307923e-03 4.15381014e-01 1.10927951e+00
5.02692521e-01 5.29948294e-01 6.72934115e-01 1.33804306e-01
-5.34850545e-02 7.09188133e-02 -1.96860224e-01 2.31605880e-02
-2.08203215e-02 1.04938602e+00 -1.04183161e+00 -4.07034397e-01
-4.39456880e-01 8.15915823e-01 -1.50589690e-01 1.77791402e-01
-1.34090841e+00 -3.11956942e-01 9.28752720e-01 6.11829638e-01
4.15123612e-01 -2.21016034e-01 3.72904807e-01 -1.28285921e+00
2.25157470e-01 -1.16556346e+00 6.57336354e-01 -7.26350784e-01
-6.56054735e-01 3.28222483e-01 2.17693567e-01 -1.79287279e+00
1.66853890e-02 -8.56063247e-01 -3.69306087e-01 1.29460618e-01
-6.33654892e-01 -8.25957179e-01 -6.21197045e-01 9.04416978e-01
7.45496750e-01 -2.74719924e-01 3.19613189e-01 7.86058545e-01
-7.74850786e-01 6.25661910e-01 -1.55949414e-01 6.12376213e-01
4.13674563e-01 -7.68845916e-01 -6.66919202e-02 8.70967507e-01
3.09130579e-01 -7.43031204e-02 7.22230375e-01 -5.21590054e-01
-1.65670228e+00 -7.49417424e-01 2.35596776e-01 -3.92890960e-01
5.61383724e-01 -4.35337722e-01 -3.90951723e-01 7.32948363e-01
-2.65885770e-01 3.02170098e-01 4.90241081e-01 -1.72165111e-01
5.09583317e-02 -3.78119439e-01 -8.27565134e-01 3.09333593e-01
1.46786618e+00 -1.97950259e-01 -7.25895762e-02 7.24463940e-01
9.63572934e-02 -6.71253979e-01 -9.06444728e-01 3.92011136e-01
7.87751734e-01 -1.28316200e+00 1.11933386e+00 -6.56525731e-01
4.78795111e-01 -9.48658958e-02 -2.27494404e-01 -5.81494510e-01
-3.97977941e-02 -5.77894092e-01 -2.18164288e-02 7.23378181e-01
-3.93697880e-02 -4.34024818e-02 1.26118922e+00 5.88557720e-01
-4.03726757e-01 -5.80444753e-01 -1.50351536e+00 -1.02413130e+00
-2.85962433e-01 -7.87534833e-01 -1.01322755e-01 6.27961338e-01
4.03255038e-02 -3.57610345e-01 -8.24368477e-01 -3.29971015e-01
4.40790981e-01 -1.57002762e-01 1.00267124e+00 -7.45751381e-01
-4.19750810e-01 -2.83880770e-01 -1.13934970e+00 -1.40477335e+00
-2.58055985e-01 -3.37750524e-01 2.71029249e-02 -1.10556436e+00
1.75864249e-01 -1.83761239e-01 -3.62106532e-01 6.33888483e-01
2.23945215e-01 7.40625143e-01 3.20900619e-01 2.02694967e-01
-1.07096398e+00 1.76062271e-01 1.02449083e+00 -1.09165400e-01
9.17587802e-02 2.24959880e-01 -1.12179182e-02 9.38089371e-01
3.32281083e-01 -4.51815069e-01 -7.10946470e-02 -2.26781234e-01
2.19563574e-01 4.93545324e-01 6.20163143e-01 -1.51914072e+00
1.28988281e-01 -3.45002860e-01 2.92985976e-01 -4.87905949e-01
8.34982872e-01 -3.94351095e-01 -1.08665824e-02 6.57432854e-01
-1.95447907e-01 3.17235515e-02 9.95117649e-02 5.21218002e-01
-4.09524709e-01 1.06694974e-01 7.08434105e-01 -4.21873212e-01
-1.34323061e+00 3.34355533e-01 -5.18650711e-01 2.41712332e-01
1.37049246e+00 -6.54956222e-01 -3.66215259e-01 -4.55170482e-01
-1.04038525e+00 2.15105295e-01 -1.77453812e-02 5.35491705e-01
6.12842441e-01 -1.15396416e+00 -8.62983584e-01 1.98350981e-01
1.97835639e-01 -4.19911087e-01 9.30007875e-01 1.39213967e+00
-9.04934525e-01 1.75665513e-01 -5.16585827e-01 -6.97397172e-01
-1.41790843e+00 2.66114444e-01 2.16560498e-01 8.15320411e-04
-8.59755933e-01 1.32565069e+00 -7.59410486e-02 2.26862013e-01
4.64783430e-01 -2.92442858e-01 -1.67136714e-01 1.37469292e-01
7.42015004e-01 8.49607229e-01 -9.03286412e-02 -8.56664896e-01
-5.81893504e-01 4.30610776e-01 1.23955846e-01 2.26624563e-01
1.17364049e+00 1.94627151e-01 2.19705120e-01 2.13142619e-01
9.67624128e-01 -3.42963457e-01 -1.52547562e+00 1.58706918e-01
-3.43186706e-01 -8.75815749e-01 -3.48769486e-01 -5.01873970e-01
-1.37378943e+00 8.72837007e-01 8.80322039e-01 1.25782683e-01
1.05696213e+00 1.03483476e-01 7.96529293e-01 3.37420851e-01
6.63400948e-01 -1.54857934e+00 4.96626794e-01 2.85657257e-01
6.73896134e-01 -1.20266986e+00 1.62768781e-01 -2.29155362e-01
-8.60636830e-01 8.80864382e-01 9.00448442e-01 -2.67586857e-01
4.75030810e-01 1.15188733e-01 1.35160610e-03 -3.00419658e-01
-3.96113515e-01 -2.00919911e-01 2.72138715e-02 3.85554343e-01
9.34295207e-02 -2.37163082e-02 -3.34277630e-01 5.72753966e-01
-1.18645228e-01 1.19104460e-01 6.26887619e-01 1.22773719e+00
-5.37748814e-01 -5.93626082e-01 -3.03849548e-01 6.75452232e-01
-5.69162369e-01 2.45417729e-01 3.22708003e-02 1.10507321e+00
4.31655526e-01 9.71256435e-01 9.17169750e-02 -7.68845558e-01
6.66482329e-01 -1.04095861e-01 3.54427993e-01 -3.42562586e-01
-7.66407371e-01 7.19286501e-02 3.43640029e-01 -1.21343517e+00
-5.75893283e-01 -8.98634195e-01 -1.06107402e+00 -5.36953509e-01
1.53118074e-01 8.05248320e-02 2.67001867e-01 8.45039964e-01
9.16740373e-02 3.39316279e-01 2.40446746e-01 -1.06405210e+00
-2.69608825e-01 -9.74928021e-01 -1.18745041e+00 6.54010296e-01
-1.79715842e-01 -1.02044713e+00 -2.66839266e-01 1.55999750e-01] | [7.874356269836426, 0.21046769618988037] |
14ca3b19-2341-4125-857a-b34c12ce3762 | deep-defocus-map-estimation-using-domain | null | null | http://openaccess.thecvf.com/content_CVPR_2019/html/Lee_Deep_Defocus_Map_Estimation_Using_Domain_Adaptation_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Lee_Deep_Defocus_Map_Estimation_Using_Domain_Adaptation_CVPR_2019_paper.pdf | Deep Defocus Map Estimation Using Domain Adaptation | In this paper, we propose the first end-to-end convolutional neural network (CNN) architecture, Defocus Map Estimation Network (DMENet), for spatially varying defocus map estimation. To train the network, we produce a novel depth-of-field (DOF) dataset, SYNDOF, where each image is synthetically blurred with a ground-truth depth map. Due to the synthetic nature of SYNDOF, the feature characteristics of images in SYNDOF can differ from those of real defocused photos. To address this gap, we use domain adaptation that transfers the features of real defocused photos into those of synthetically blurred ones. Our DMENet consists of four subnetworks: blur estimation, domain adaptation, content preservation, and sharpness calibration networks. The subnetworks are connected to each other and jointly trained with their corresponding supervisions in an end-to-end manner. Our method is evaluated on publicly available blur detection and blur estimation datasets and the results show the state-of-the-art performance.In this paper, we propose the first end-to-end convolutional neural network (CNN) architecture, Defocus Map Estimation Network (DMENet), for spatially varying defocus map estimation. To train the network, we produce a novel depth-of-field (DOF) dataset, SYNDOF, where each image is synthetically blurred with a ground-truth depth map. Due to the synthetic nature of SYNDOF, the feature characteristics of images in SYNDOF can differ from those of real defocused photos. To address this gap, we use domain adaptation that transfers the features of real defocused photos into those of synthetically blurred ones. Our DMENet consists of four subnetworks: blur estimation, domain adaptation, content preservation, and sharpness calibration networks. The subnetworks are connected to each other and jointly trained with their corresponding supervisions in an end-to-end manner. Our method is evaluated on publicly available blur detection and blur estimation datasets and the results show the state-of-the-art performance.
| [' Seungyong Lee', ' Sunghyun Cho', ' Sungkil Lee', 'Junyong Lee'] | 2019-06-01 | null | null | null | cvpr-2019-6 | ['defocus-estimation'] | ['computer-vision'] | [ 8.15141946e-02 -2.92849123e-01 4.42006439e-01 -5.44738412e-01
-1.84803177e-02 -5.11658788e-01 4.17184442e-01 -6.84311688e-01
-4.42085326e-01 1.02612555e+00 1.83864057e-01 9.92616173e-03
-3.42940800e-02 -5.17147720e-01 -1.03770030e+00 -7.97490954e-01
1.44877443e-02 5.16489439e-04 3.19762647e-01 7.33898431e-02
1.25742793e-01 4.38002259e-01 -1.52664065e+00 1.31076157e-01
1.21227407e+00 1.02340615e+00 6.29680514e-01 9.40132380e-01
3.67860734e-01 9.37559009e-01 -5.23846447e-01 -5.00600552e-03
2.14717522e-01 -1.99234292e-01 -5.81427813e-01 1.41404361e-01
1.00079155e+00 -1.10336769e+00 -9.83990669e-01 1.33839512e+00
7.66895235e-01 -1.25219762e-01 6.75422549e-01 -1.12216485e+00
-8.98857892e-01 1.62956312e-01 -4.32229370e-01 3.23340327e-01
-9.03396532e-02 5.55742025e-01 1.54634044e-01 -8.38007629e-01
6.12928450e-01 1.17647874e+00 6.75142169e-01 4.67828810e-01
-1.03239393e+00 -5.85491419e-01 -1.10106289e-01 2.23450854e-01
-1.07075727e+00 -3.86173189e-01 7.31652379e-01 -5.85354865e-01
7.51684368e-01 -3.34641308e-01 4.25485790e-01 9.52646673e-01
6.21981978e-01 5.93667388e-01 1.01115048e+00 -1.03368543e-01
3.70614529e-01 -2.07036540e-01 -2.22217202e-01 6.60413146e-01
1.47297636e-01 6.35430753e-01 -3.62475783e-01 4.70620811e-01
1.34902620e+00 3.89615372e-02 -1.02318323e+00 -6.60249949e-01
-1.48071039e+00 3.89426321e-01 9.58969712e-01 1.39556870e-01
-1.99388266e-01 3.41949970e-01 -3.16023603e-02 1.42531276e-01
4.20176238e-01 4.97561246e-01 -4.18568581e-01 -1.96827367e-01
-7.97773957e-01 2.04859644e-01 4.94046748e-01 1.25462341e+00
1.11367297e+00 -4.49589379e-02 -4.75018561e-01 9.23278928e-01
-1.26808612e-02 7.38632083e-01 5.74141800e-01 -1.34237170e+00
2.47150272e-01 2.34940693e-01 6.32997930e-01 -5.99842191e-01
-4.52836305e-01 -5.34708738e-01 -1.15557480e+00 4.23359215e-01
3.68869275e-01 -5.08961499e-01 -1.09832990e+00 1.88227379e+00
2.34771818e-02 4.47852790e-01 1.73379645e-01 1.47893572e+00
8.54280174e-01 4.50360149e-01 -6.76371038e-01 1.20220527e-01
1.26536858e+00 -1.21531308e+00 -5.97354770e-01 -3.66212130e-01
1.73832878e-01 -9.98927116e-01 8.57128620e-01 1.05417393e-01
-1.18622315e+00 -8.58494043e-01 -1.12061822e+00 -2.19136521e-01
-4.41050343e-03 2.59475678e-01 2.98603833e-01 2.02989459e-01
-1.37833679e+00 6.32850647e-01 -6.67503476e-01 -1.10425137e-01
5.71863770e-01 1.35418445e-01 -3.61110657e-01 -4.51924026e-01
-1.25245345e+00 7.44172990e-01 3.45930129e-01 1.47978902e-01
-1.24892855e+00 -1.02445722e+00 -1.08250320e+00 1.33047253e-01
1.79996472e-02 -1.11196613e+00 1.51981401e+00 -8.17273438e-01
-1.27322447e+00 5.18055797e-01 3.83973941e-02 -4.67581183e-01
7.55129635e-01 -1.52667746e-01 -1.81179374e-01 2.11090356e-01
1.84155732e-01 8.03214014e-01 1.11591184e+00 -1.29892635e+00
-6.88632667e-01 -1.05883786e-02 1.25883505e-01 1.67584538e-01
8.60855505e-02 -2.12327421e-01 -3.83980870e-01 -5.38335204e-01
-2.46289268e-01 -7.37070739e-01 2.24294990e-01 3.56193781e-01
-3.68656844e-01 3.83875996e-01 9.61427569e-01 -2.76501417e-01
9.24277544e-01 -2.48352718e+00 1.22867279e-01 -8.09591353e-01
7.03216910e-01 3.09451163e-01 -2.74673849e-01 8.43173862e-02
-2.07615793e-01 -4.34422225e-01 -3.74985039e-01 -5.20309567e-01
-1.96259946e-01 1.77218929e-01 -1.68771341e-01 6.03475809e-01
3.81217957e-01 7.14164317e-01 -1.45896745e+00 1.75482798e-02
6.12420440e-01 7.38788843e-01 -4.86334056e-01 8.20859492e-01
-2.85309523e-01 6.00512505e-01 1.28403995e-02 1.86931849e-01
1.49530387e+00 -1.73064902e-01 -5.14808238e-01 -8.22437763e-01
-5.12950957e-01 -1.06554799e-01 -8.10838521e-01 1.92826986e+00
-5.79943240e-01 1.09932423e+00 1.60459369e-01 -3.17777604e-01
5.90251505e-01 1.20138638e-01 2.33627468e-01 -3.95336866e-01
1.27230361e-01 4.14760292e-01 5.69953993e-02 -6.03533208e-01
2.87551075e-01 -1.54789299e-01 2.26412147e-01 3.24481636e-01
3.17530364e-01 -7.14107811e-01 6.54625893e-02 1.54963657e-01
1.28306556e+00 2.41778474e-02 -1.77521646e-01 -1.69508517e-01
3.98772985e-01 -4.46975291e-01 3.57443959e-01 9.25195277e-01
-4.98764932e-01 1.27222693e+00 3.89469892e-01 -6.00210130e-01
-1.24862349e+00 -1.42790425e+00 -4.11994129e-01 3.22670907e-01
5.45166314e-01 1.71101540e-01 -1.01036131e+00 -6.38881445e-01
1.63129225e-01 4.63260442e-01 -6.06907189e-01 -2.53827542e-01
-3.60832274e-01 -2.63950408e-01 2.76503563e-01 3.47661436e-01
1.40170920e+00 -9.86341953e-01 -8.61655354e-01 2.79801518e-01
-2.75822908e-01 -1.55819368e+00 -1.07294929e+00 1.19464099e-01
-3.70517373e-01 -1.12636578e+00 -1.22867835e+00 -6.33840024e-01
6.47931099e-01 7.51394153e-01 1.10818803e+00 -3.04595560e-01
-1.81938529e-01 1.13375053e-01 1.17486604e-02 -2.67497450e-01
-3.29454899e-01 -2.10975066e-01 1.80041250e-02 2.43505761e-01
-8.93666372e-02 -7.64507055e-01 -1.23247600e+00 5.42150974e-01
-1.22446513e+00 3.38879496e-01 5.66061854e-01 9.98179615e-01
1.75573632e-01 1.14265390e-01 8.61504599e-02 -4.37228799e-01
4.30130839e-01 -2.49066293e-01 -8.59319985e-01 -8.05913284e-02
-3.12917888e-01 1.25351101e-01 6.39208972e-01 -4.46995676e-01
-1.24051952e+00 -1.58557177e-01 2.77719259e-01 -1.06759715e+00
-2.87468970e-01 7.02372864e-02 2.37395186e-02 -1.03475191e-01
8.74553025e-01 2.24360093e-01 9.70845148e-02 -3.59898448e-01
2.56940663e-01 1.01701772e+00 1.24285007e+00 -1.07082918e-01
7.37872124e-01 4.16195631e-01 -1.18410373e-02 -4.52536196e-01
-1.05588996e+00 -2.70346642e-01 -4.77303118e-01 -3.58081460e-01
9.04045522e-01 -1.33058107e+00 -5.89627564e-01 1.42235959e+00
-1.37846351e+00 -6.47417963e-01 3.19511965e-02 7.36604214e-01
-7.29148865e-01 2.97321498e-01 -6.82973921e-01 -2.57512361e-01
-1.70016021e-01 -1.45230019e+00 1.12642097e+00 6.13996327e-01
3.48222882e-01 -9.48249698e-01 -1.61830887e-01 2.69154847e-01
7.87104785e-01 2.08966017e-01 6.18759453e-01 2.62327373e-01
-9.57511902e-01 3.44083250e-01 -9.94651496e-01 7.60180414e-01
6.38944745e-01 -1.76042736e-01 -1.17725742e+00 -5.72432339e-01
1.23832807e-01 -2.00219169e-01 9.87357974e-01 7.46417046e-01
1.16228056e+00 -9.74310189e-02 -1.05260745e-01 1.23907804e+00
1.49725223e+00 4.71437015e-02 7.82605827e-01 1.85215071e-01
6.40157759e-01 1.62916347e-01 3.63122553e-01 2.91779220e-01
4.19484854e-01 7.40780711e-01 8.31254900e-01 -4.08355206e-01
-9.09151196e-01 -1.64213166e-01 1.60754055e-01 -4.63206694e-03
2.35137790e-01 -3.65369767e-01 -5.99863887e-01 7.68209219e-01
-1.71996176e+00 -7.27611125e-01 -5.93235269e-02 2.00950074e+00
1.04905713e+00 -1.05616927e-01 -1.83501363e-01 -5.02595782e-01
8.77550960e-01 2.92616099e-01 -9.58635569e-01 -1.96796343e-01
-2.82481581e-01 -1.50929049e-01 4.57110941e-01 6.74917698e-01
-1.12158942e+00 7.23767817e-01 4.22462749e+00 4.83000129e-01
-1.40224791e+00 -6.96941838e-02 5.31827569e-01 -1.53236136e-01
-5.82834631e-02 -2.23969519e-01 -4.22734022e-01 1.10815132e+00
5.15065312e-01 1.48326933e-01 1.01852965e+00 5.39979517e-01
3.03260982e-01 -4.94904459e-01 -1.27544141e+00 1.38947201e+00
-4.37819175e-02 -1.29442000e+00 -4.27908987e-01 -2.83510685e-01
1.08333945e+00 4.28421080e-01 7.41365328e-02 -2.01157946e-02
3.63618165e-01 -1.02148426e+00 7.25944161e-01 6.98011875e-01
1.25228512e+00 -4.33342278e-01 9.93661463e-01 2.73013920e-01
-8.19330037e-01 2.52217520e-02 -5.02432227e-01 6.42740354e-02
1.98141605e-01 1.00330830e+00 -8.72792065e-01 5.28388143e-01
1.01710069e+00 9.86559808e-01 -3.84719193e-01 1.37485659e+00
-3.13987076e-01 1.75792530e-01 6.75763935e-02 3.37686777e-01
1.88487068e-01 -4.51290645e-02 5.78550220e-01 1.26672709e+00
5.10040164e-01 -3.44118059e-01 -3.60416651e-01 1.22661269e+00
-4.43629920e-01 -1.11309147e+00 -4.48773235e-01 4.32573706e-01
4.75396782e-01 1.06591094e+00 1.08463271e-02 -2.42528304e-01
-2.15010986e-01 1.21139729e+00 3.76822323e-01 7.07559764e-01
-9.98558402e-01 -8.53374362e-01 1.11270475e+00 1.30330682e-01
4.93398190e-01 2.04737723e-01 6.56025708e-02 -1.45523489e+00
-3.67006958e-02 -5.51114082e-01 -2.12893248e-01 -1.71214759e+00
-1.42212403e+00 7.44706452e-01 1.69017255e-01 -1.29547894e+00
-1.68712027e-02 -5.70597112e-01 -6.81472659e-01 1.46363473e+00
-2.21310425e+00 -8.14151347e-01 -1.22549176e+00 6.36675775e-01
4.41786855e-01 2.88048554e-02 2.64033318e-01 2.05186009e-01
-2.87030101e-01 2.08023712e-01 3.54169220e-01 2.25610122e-01
1.27019250e+00 -1.43649483e+00 4.28651273e-01 8.88213277e-01
-7.59697437e-01 2.04475939e-01 6.61064446e-01 -3.41214508e-01
-1.04710174e+00 -1.37201226e+00 5.44942677e-01 2.79552191e-02
5.21270633e-01 -3.57677251e-01 -9.61812377e-01 4.26717520e-01
4.97560531e-01 4.25962567e-01 -3.58717710e-01 -6.04490399e-01
-2.04865977e-01 -4.70708996e-01 -1.19642687e+00 4.09164578e-01
1.08151746e+00 -6.14087403e-01 -3.54651600e-01 3.49159122e-01
8.91306341e-01 -1.08783734e+00 -5.04856825e-01 6.61808789e-01
3.43072265e-01 -1.59361279e+00 9.50396538e-01 1.98025942e-01
9.36744988e-01 -7.74113357e-01 2.24207595e-01 -2.11179590e+00
-4.10101920e-01 -7.41237760e-01 -1.87214389e-01 8.59549463e-01
-1.20629139e-01 -9.34155881e-01 4.16107148e-01 2.27917358e-01
-5.87075770e-01 -7.59312272e-01 -7.73697555e-01 -7.71449625e-01
8.84832069e-03 1.06731199e-01 8.06057572e-01 6.87038660e-01
-6.10854268e-01 2.95077056e-01 -3.80288035e-01 2.69201517e-01
8.16014588e-01 1.56823277e-01 7.78770566e-01 -7.97702074e-01
-2.39109188e-01 -2.95064211e-01 -3.90397698e-01 -1.67144227e+00
-9.73999724e-02 -2.00980738e-01 4.28619325e-01 -1.46840620e+00
2.26285443e-01 -2.42496490e-01 -1.41629636e-01 1.15230598e-01
-2.42920980e-01 8.83529428e-03 1.01723887e-01 2.18311980e-01
-1.00798391e-01 4.62958872e-01 1.96624637e+00 -3.17627013e-01
-3.94194536e-02 -4.53384966e-02 -4.43106681e-01 4.65344578e-01
4.37364817e-01 -1.82068631e-01 -6.80055916e-01 -9.76770937e-01
-2.62215912e-01 3.51625055e-01 7.58867085e-01 -1.23979902e+00
3.19050133e-01 -6.15785131e-03 7.53756404e-01 -5.06466925e-01
1.98553756e-01 -6.14330053e-01 9.20457542e-02 3.13657731e-01
-1.55591853e-02 -3.75130981e-01 2.86085725e-01 3.73108119e-01
-3.05577606e-01 -8.00103843e-02 1.32154071e+00 2.97095813e-02
-4.48494345e-01 4.62842464e-01 -9.35469121e-02 1.35136977e-01
7.40999341e-01 -1.31721213e-01 -9.74040747e-01 -6.70314789e-01
-2.32920066e-01 1.10200502e-01 8.35393250e-01 2.47417957e-01
6.86261058e-01 -1.23750103e+00 -7.59626985e-01 7.73079753e-01
1.32698163e-01 7.08712041e-01 5.57542861e-01 8.43395710e-01
-8.18960607e-01 3.99862677e-01 -3.46890688e-01 -8.06767166e-01
-7.25271344e-01 6.03885770e-01 1.05106521e+00 2.29907781e-01
-1.98891908e-01 1.06114602e+00 7.87995815e-01 -1.10364497e-01
1.49359643e-01 -1.00169027e+00 2.23048329e-01 -7.35412300e-01
5.43055892e-01 2.82949507e-02 6.55434001e-03 -2.94617146e-01
-1.38679042e-01 6.28542066e-01 -2.28591636e-02 3.09881330e-01
1.33447182e+00 -2.08259046e-01 -1.60921097e-01 -1.33576274e-01
1.38462293e+00 -9.47420895e-02 -2.36296463e+00 -2.74367899e-01
-9.60186124e-01 -9.88175035e-01 4.13656473e-01 -1.18011379e+00
-1.28259706e+00 9.51742053e-01 9.17776883e-01 -1.96209028e-01
1.59962094e+00 -8.69759470e-02 7.17781723e-01 -4.92722802e-02
2.03926280e-01 -5.81360936e-01 1.85683668e-01 6.98874593e-01
9.83105481e-01 -1.27931821e+00 -3.69502306e-01 -3.63680422e-01
-3.96012783e-01 1.18481624e+00 8.13991308e-01 -2.06782177e-01
5.45650542e-01 2.35135019e-01 1.35907233e-01 7.15872645e-02
-5.54979503e-01 6.87788054e-02 3.23648334e-01 9.20230985e-01
1.58479571e-01 -2.50244766e-01 4.49599028e-01 8.15751478e-02
-1.37881115e-01 2.87432611e-01 7.58304954e-01 5.90530217e-01
-3.86345387e-01 -6.56722724e-01 -2.90820986e-01 5.52580804e-02
2.46965513e-01 -2.31585160e-01 -4.96570878e-02 5.14362931e-01
5.03393829e-01 9.33406055e-01 3.43720764e-01 -2.31292620e-01
4.73849922e-01 -5.46879411e-01 7.86837518e-01 -3.55604500e-01
-2.94576824e-01 -9.86671820e-02 -1.02285728e-01 -4.02207792e-01
-1.84664607e-01 -3.57470602e-01 -9.63796139e-01 -4.72894430e-01
-6.29122630e-02 -2.00234756e-01 4.87554222e-01 7.30881572e-01
6.51939809e-01 6.65479362e-01 9.81239617e-01 -1.08685696e+00
-4.89531249e-01 -1.21070921e+00 -6.41397655e-01 4.49987441e-01
1.18608677e+00 -8.20990145e-01 -6.61938965e-01 6.85215816e-02] | [11.309432029724121, -2.736599922180176] |
cc0201cc-5fba-492a-a0d8-0d0a988532fa | overview-of-the-sv-ident-2022-shared-task-on | 2209.09062 | null | https://arxiv.org/abs/2209.09062v1 | https://arxiv.org/pdf/2209.09062v1.pdf | Overview of the SV-Ident 2022 Shared Task on Survey Variable Identification in Social Science Publications | In this paper, we provide an overview of the SV-Ident shared task as part of the 3rd Workshop on Scholarly Document Processing (SDP) at COLING 2022. In the shared task, participants were provided with a sentence and a vocabulary of variables, and asked to identify which variables, if any, are mentioned in individual sentences from scholarly documents in full text. Two teams made a total of 9 submissions to the shared task leaderboard. While none of the teams improve on the baseline systems, we still draw insights from their submissions. Furthermore, we provide a detailed evaluation. Data and baselines for our shared task are freely available at https://github.com/vadis-project/sv-ident | ['Philipp Mayr', 'Kai Eckert', 'Andrea Zielinski', 'Simone Paolo Ponzetto', 'Yavuz Selim Kartal', 'Tornike Tsereteli'] | 2022-09-19 | null | https://aclanthology.org/2022.sdp-1.29 | https://aclanthology.org/2022.sdp-1.29.pdf | sdp-coling-2022-10 | ['variable-disambiguation', 'variable-detection'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.56094939e-01 7.81146856e-03 -3.55572104e-01 -4.92960125e-01
-1.35612178e+00 -1.02796316e+00 8.29427302e-01 2.31055886e-01
-2.71038532e-01 9.63922858e-01 3.61208886e-01 -2.33899236e-01
-8.37953761e-02 -2.53706425e-01 -6.03000700e-01 3.29334773e-02
2.35408485e-01 5.88807523e-01 -1.24606274e-01 -1.16926283e-01
6.97420716e-01 7.31349215e-02 -1.07335269e+00 7.93030083e-01
6.50113642e-01 5.96265912e-01 3.42701524e-01 1.10233366e+00
-5.46840549e-01 7.91676044e-01 -9.39741969e-01 -6.03497744e-01
-4.57589136e-04 -2.47151643e-01 -1.35759974e+00 -4.37660873e-01
9.65756178e-01 2.43085280e-01 -2.67231971e-01 9.41856027e-01
3.87181729e-01 3.62060457e-01 6.81154609e-01 -1.41832030e+00
-7.54236937e-01 1.40574110e+00 -3.75224680e-01 6.00447536e-01
6.70348525e-01 -1.43621385e-01 1.66472960e+00 -1.28270674e+00
1.12151575e+00 1.35141706e+00 6.23152912e-01 4.29930538e-01
-1.18650115e+00 -7.66743600e-01 4.19112355e-01 3.99465024e-01
-1.33364725e+00 -8.64000678e-01 5.55554032e-01 -6.32920563e-01
1.22007525e+00 6.71821475e-01 2.47437015e-01 1.44366896e+00
-9.08966642e-03 1.09461212e+00 7.27740943e-01 -4.34137672e-01
-6.81833252e-02 1.50747418e-01 8.21963966e-01 4.98498857e-01
2.64426377e-02 -7.55673707e-01 -8.47468972e-01 -2.20885083e-01
1.57358572e-01 -5.92551708e-01 -3.17088574e-01 4.16438222e-01
-1.50696695e+00 5.70461094e-01 -3.17354292e-01 4.33490515e-01
-5.57084978e-02 1.25547543e-01 6.33025050e-01 3.69576633e-01
7.08753109e-01 8.93090367e-01 -6.83318019e-01 -6.71556473e-01
-1.13017583e+00 6.73879564e-01 1.22074723e+00 1.29912496e+00
3.79832536e-01 -3.50727677e-01 -7.23485410e-01 9.02298152e-01
1.52696326e-01 4.72455084e-01 1.94570929e-01 -1.33387232e+00
8.30824316e-01 2.77137220e-01 -1.54778868e-01 -9.71581638e-01
-1.18163191e-01 -1.25824004e-01 -4.23388660e-01 -4.87645775e-01
2.84234077e-01 -3.06987226e-01 -7.22733557e-01 1.36201739e+00
-4.30726081e-01 -1.75880164e-01 -2.11042643e-01 5.05065978e-01
1.29480195e+00 7.11891949e-01 -1.41351357e-01 -1.38021901e-01
1.21424901e+00 -1.11747754e+00 -1.22049046e+00 -2.91617304e-01
6.42761648e-01 -1.17491436e+00 8.89463842e-01 5.46733379e-01
-1.25993943e+00 -3.81194860e-01 -9.07163858e-01 -7.07249761e-01
-5.03371418e-01 1.24040917e-01 4.27814305e-01 2.15204433e-01
-1.03324950e+00 7.24589527e-01 -7.83827722e-01 -3.28672081e-01
4.72533494e-01 -4.98768389e-02 -1.33351639e-01 2.09398150e-01
-1.20131636e+00 7.38741517e-01 -1.69747904e-01 -2.30668306e-01
-6.16937995e-01 -9.11625504e-01 -5.83344877e-01 -7.92575255e-02
5.28619289e-01 -5.83721101e-01 1.86002922e+00 -4.71725464e-01
-1.30724394e+00 1.06263769e+00 -7.59363234e-01 -2.14366391e-01
6.17560267e-01 -6.70185566e-01 -2.94490844e-01 -7.26487488e-02
5.05435765e-01 -2.46727373e-02 4.00651664e-01 -1.01475489e+00
-5.25448263e-01 -3.03406209e-01 -2.53326207e-01 -1.13037884e-01
-1.12911172e-01 7.22662330e-01 -8.42878878e-01 -7.60588288e-01
-2.79947221e-01 -7.55981922e-01 6.28582090e-02 -2.57795990e-01
-9.24039543e-01 -5.85943341e-01 6.85915172e-01 -9.47416842e-01
1.28398395e+00 -2.10279083e+00 4.65827882e-01 1.08859122e-01
4.78979230e-01 -1.25349924e-01 -1.33060575e-01 6.67764127e-01
-7.34518701e-03 7.57650852e-01 -2.38842577e-01 -1.10299575e+00
2.51249343e-01 -1.17555410e-01 -6.70945406e-01 3.47357750e-01
-6.48522228e-02 7.83563793e-01 -9.84022856e-01 -7.01760650e-01
-4.98808399e-02 1.77368209e-01 1.87491030e-01 -6.64871186e-03
-5.18675745e-01 2.21994340e-01 -6.86239183e-01 5.05938172e-01
4.30682003e-01 -3.06126863e-01 -3.93046029e-02 5.38317002e-02
-4.91075367e-01 8.54613245e-01 -1.06722128e+00 1.83941746e+00
-2.76299894e-01 1.26323426e+00 2.63627350e-01 -6.38602138e-01
8.52461398e-01 3.17187846e-01 2.40847662e-01 -4.14562821e-01
2.71425676e-03 2.73111194e-01 -3.87303472e-01 -3.99090916e-01
1.02434742e+00 5.05916119e-01 -3.26224595e-01 5.53581297e-01
9.88437906e-02 -4.07448739e-01 1.07176518e+00 8.82610738e-01
1.30082309e+00 -5.09730261e-03 1.33901581e-01 -3.40275735e-01
6.91417873e-01 1.21490829e-01 5.90931535e-01 1.05120611e+00
1.44857392e-01 9.39014733e-01 8.77338409e-01 1.55604444e-02
-1.03576910e+00 -8.13274145e-01 -2.36936137e-01 1.12706971e+00
-2.31023952e-01 -1.37027144e+00 -2.09673315e-01 -6.28723443e-01
1.41525999e-01 9.56292927e-01 -5.23955643e-01 5.47764421e-01
-6.51200831e-01 -7.38389976e-03 9.11470711e-01 5.02763748e-01
2.84083873e-01 -8.44222307e-01 1.52555376e-01 -7.08903745e-02
-4.70598608e-01 -1.05025303e+00 -7.77700782e-01 1.07673928e-01
-4.09369230e-01 -1.11805761e+00 -5.08050382e-01 -7.15464354e-01
2.71286130e-01 8.97018760e-02 1.21398878e+00 1.25255406e-01
-3.54519814e-01 3.42086256e-01 -3.87980729e-01 -7.25108802e-01
-3.92256469e-01 4.86053824e-01 -3.77705783e-01 -6.12380087e-01
4.03353810e-01 -1.06184557e-02 8.67370237e-03 -9.30445194e-02
-2.35502332e-01 7.56117851e-02 9.94258821e-02 5.13462722e-01
4.81007010e-01 -3.37788612e-01 1.90358371e-01 -1.37149203e+00
1.15222013e+00 -5.83795011e-01 -5.23137748e-01 3.59481663e-01
-6.74677312e-01 -6.21276610e-02 6.94199800e-01 1.34700492e-01
-1.01256096e+00 -4.48603123e-01 5.34943491e-02 -1.04582846e-01
1.89904168e-01 7.69367218e-01 -2.99664289e-02 4.98634487e-01
3.73064071e-01 -3.19221206e-02 -1.86721504e-01 -6.93185568e-01
4.45886105e-01 8.45166147e-01 6.77621484e-01 -1.02972472e+00
7.53356636e-01 -2.00140532e-02 -2.30790779e-01 -7.77560472e-01
-8.74173522e-01 -5.99482119e-01 -5.08210301e-01 3.32242399e-02
6.01588905e-01 -9.23615515e-01 -4.43853438e-01 4.34274405e-01
-1.60231268e+00 -4.81628507e-01 -2.94727296e-01 1.12087265e-01
-1.43005773e-01 3.31925988e-01 -7.37974644e-01 -6.17964208e-01
-4.78071213e-01 -1.06390226e+00 9.53447521e-01 5.19098714e-02
-1.00264859e+00 -1.09656513e+00 1.93574727e-01 6.77545905e-01
4.90476012e-01 -7.50894099e-03 7.27602780e-01 -1.16730428e+00
-2.70865411e-01 -1.54664099e-01 -1.71406537e-01 1.96743622e-01
7.15854112e-03 7.61252999e-01 -7.45967507e-01 -4.20381129e-02
-4.33851361e-01 -3.71253729e-01 1.04685581e+00 1.92123309e-01
1.53950882e+00 -3.08458984e-01 -5.02580941e-01 6.00132704e-01
8.92133534e-01 1.99802984e-02 2.56314635e-01 6.57451630e-01
6.50537848e-01 4.64676291e-01 3.93553495e-01 6.17698312e-01
5.91362536e-01 5.65041780e-01 -2.14442983e-01 4.29908246e-01
-3.60205501e-01 -7.05297664e-02 3.93936098e-01 1.11521375e+00
-1.10947616e-01 -9.12858129e-01 -1.22100532e+00 6.62224054e-01
-1.77705729e+00 -1.09899366e+00 -8.81347656e-01 1.76853132e+00
1.09966207e+00 -2.12898315e-03 -2.21814021e-01 -2.33189508e-01
7.42784142e-01 6.36166036e-01 -2.80583203e-01 -6.55280530e-01
-5.57655573e-01 3.49785358e-01 5.95407188e-01 8.42994928e-01
-1.05810201e+00 9.06558692e-01 6.99382782e+00 6.93287492e-01
-7.08205462e-01 5.33904471e-02 3.19532812e-01 -5.61675906e-01
-7.01524854e-01 2.04848751e-01 -1.24646199e+00 6.83444440e-01
1.17979813e+00 -9.11142588e-01 4.48107958e-01 5.12234747e-01
1.87541768e-01 1.20498940e-01 -1.37190533e+00 6.83517337e-01
3.31335723e-01 -1.57790339e+00 -3.53259295e-02 -6.24050424e-02
7.40423143e-01 1.97614506e-01 8.52121413e-02 4.83993828e-01
6.02340281e-01 -1.33749771e+00 7.59075701e-01 6.22516632e-01
6.12011373e-01 -3.29757243e-01 6.66419327e-01 4.33694422e-02
-8.39224637e-01 2.55881876e-01 3.49418931e-02 -8.96396860e-02
2.83854723e-01 5.29804587e-01 -6.87316120e-01 6.37433827e-01
6.04430079e-01 1.26557219e+00 -7.31094837e-01 7.04190195e-01
-5.93763351e-01 7.85140634e-01 -1.48092166e-01 -1.70242906e-01
-7.47214258e-02 2.95945890e-02 9.65729952e-01 1.71772718e+00
3.62706743e-02 2.02488210e-02 2.15879586e-02 1.02647972e+00
-8.05769861e-01 1.00797564e-01 -2.96694100e-01 -4.59384203e-01
9.18791234e-01 1.24059606e+00 -2.79682398e-01 -5.45214117e-01
-4.36197132e-01 8.12522233e-01 2.77973741e-01 4.16639537e-01
-4.80326325e-01 -6.73798203e-01 6.28406167e-01 -2.23530486e-01
2.18232647e-01 -3.06032568e-01 -7.48780668e-01 -1.19791424e+00
3.13162923e-01 -1.01088166e+00 3.78927529e-01 -7.08869934e-01
-1.45988679e+00 3.30707163e-01 -1.04601616e-02 -5.74973285e-01
-2.61925876e-01 -4.44037586e-01 -7.77274251e-01 1.05905473e+00
-1.02856672e+00 -8.33442926e-01 -3.91844481e-01 3.50587785e-01
9.58354056e-01 -3.71839672e-01 7.16983438e-01 2.30266303e-01
-8.13200951e-01 6.72615230e-01 3.76050264e-01 3.63547236e-01
1.34369218e+00 -1.57711649e+00 9.05644119e-01 9.55164909e-01
1.38983190e-01 1.16358435e+00 9.80058074e-01 -9.92784798e-01
-1.68292880e+00 -8.60755384e-01 1.50177217e+00 -9.14330006e-01
1.10556483e+00 -4.81042951e-01 -9.27641451e-01 1.26271844e+00
6.91944122e-01 -3.67633194e-01 7.18006611e-01 6.15931511e-01
-2.49810383e-01 3.62710178e-01 -5.55006564e-01 4.65469748e-01
9.83439505e-01 -5.88486314e-01 -1.01637697e+00 8.35579693e-01
7.70849168e-01 -8.09162855e-01 -1.00221086e+00 -1.19921096e-01
4.73586798e-01 -1.23312324e-01 6.08462691e-01 -6.11538231e-01
9.39633429e-01 -5.31741381e-02 -1.05338126e-01 -1.20050132e+00
-3.16463560e-01 -8.43506873e-01 -1.53219491e-01 1.67487049e+00
8.79999220e-01 -4.75691438e-01 4.96482491e-01 9.90715265e-01
-5.09776473e-01 -3.14800471e-01 -9.28768754e-01 -8.68564010e-01
4.14288968e-01 -6.68914974e-01 2.20249653e-01 1.06620145e+00
3.01267773e-01 6.19664311e-01 -9.07247961e-02 -2.41332605e-01
5.39512098e-01 3.52432542e-02 9.51475561e-01 -1.09020126e+00
-1.43963113e-01 -8.16804647e-01 9.61789787e-02 -9.65225697e-01
4.50451136e-01 -1.31586289e+00 -4.29818034e-02 -1.99539769e+00
4.46983397e-01 -3.34901661e-02 -1.05838940e-01 6.88286364e-01
-2.05700874e-01 -1.46748230e-01 4.73235279e-01 4.28630590e-01
-9.65135336e-01 2.48164907e-01 8.55647445e-01 -4.88339305e-01
-6.94783479e-02 -3.24274540e-01 -1.00163817e+00 2.09245771e-01
7.58842945e-01 -5.00558615e-01 -2.20137928e-02 -8.06388080e-01
4.03333962e-01 -1.54459283e-01 2.20361143e-01 -5.37485242e-01
7.54199147e-01 -1.62475392e-01 2.32830927e-01 -8.30340445e-01
2.57553905e-01 -1.25926659e-02 -1.49416164e-01 -1.54422089e-01
-9.08562660e-01 1.89135164e-01 3.88287276e-01 4.85568717e-02
-2.26732269e-01 -3.30948323e-01 1.91192195e-01 -1.75203159e-01
-5.35789728e-01 4.07931656e-02 -5.25616288e-01 4.57913160e-01
7.50642121e-01 1.33575022e-01 -1.03642666e+00 -3.44982833e-01
-3.76320690e-01 8.05866957e-01 3.71849775e-01 6.32392585e-01
5.23061514e-01 -7.66786575e-01 -1.07218981e+00 -2.29257122e-01
-5.36712185e-02 7.06613883e-02 6.80419877e-02 7.88497329e-01
-3.75066996e-01 6.74976826e-01 1.85760319e-01 -1.12064116e-01
-1.42259312e+00 7.79468790e-02 -1.48627013e-01 -2.61727393e-01
-5.75865805e-01 1.08668566e+00 -3.84552628e-01 -3.63310784e-01
2.34704763e-01 -1.38209850e-01 -2.24362522e-01 1.48306876e-01
8.40757489e-01 6.36689126e-01 1.10246569e-01 -3.91908795e-01
-5.77568591e-01 2.43692085e-01 -5.37859678e-01 -3.37993234e-01
1.53537655e+00 -1.96457416e-01 -5.30796111e-01 9.90843713e-01
1.38407743e+00 3.99434239e-01 -5.88633895e-01 -3.74707490e-01
2.86375999e-01 -4.39365476e-01 2.22353920e-01 -9.55052316e-01
-7.67506897e-01 4.99141514e-01 -2.81429201e-01 8.39738250e-02
4.58521813e-01 2.12458014e-01 6.95031047e-01 3.96863639e-01
-1.51083633e-01 -1.33920026e+00 -2.89108783e-01 1.06941521e+00
1.19848299e+00 -1.09373403e+00 5.12804925e-01 -4.12640363e-01
-6.79442644e-01 1.41083193e+00 5.05447090e-01 1.23008229e-01
3.00414801e-01 4.87000704e-01 -5.56980446e-02 -4.35827225e-01
-1.22404087e+00 3.92135948e-01 2.58625716e-01 3.00417125e-01
1.03128529e+00 -1.13782249e-01 -6.42223358e-01 8.48157823e-01
-6.47666156e-01 -8.38626772e-02 8.55553925e-01 9.92796659e-01
-1.21847168e-01 -1.26508522e+00 -2.13614419e-01 6.67248487e-01
-7.41962314e-01 -3.25895131e-01 -9.93733823e-01 7.60890067e-01
-2.82458484e-01 1.02332032e+00 2.03785390e-01 -1.35053292e-01
3.93394709e-01 1.54712334e-01 2.69811898e-01 -9.30913210e-01
-8.23695660e-01 -2.91737467e-01 7.54030406e-01 -5.34535348e-01
8.38542636e-03 -1.35633183e+00 -1.30254793e+00 -5.21919787e-01
2.90181041e-01 3.33291411e-01 8.30171645e-01 8.32727075e-01
3.68223757e-01 7.74336398e-01 1.36728510e-01 -3.83400857e-01
-4.95852888e-01 -1.23426425e+00 -4.94641781e-01 1.72129735e-01
2.95667261e-01 -2.05725476e-01 -7.32609987e-01 3.89544755e-01] | [11.995792388916016, 9.444365501403809] |
ce7cb414-d319-496e-83a1-1698bbca5310 | towards-alphachem-chemical-synthesis-planning | 1702.00020 | null | http://arxiv.org/abs/1702.00020v1 | http://arxiv.org/pdf/1702.00020v1.pdf | Towards "AlphaChem": Chemical Synthesis Planning with Tree Search and Deep Neural Network Policies | Retrosynthesis is a technique to plan the chemical synthesis of organic
molecules, for example drugs, agro- and fine chemicals. In retrosynthesis, a
search tree is built by analysing molecules recursively and dissecting them
into simpler molecular building blocks until one obtains a set of known
building blocks. The search space is intractably large, and it is difficult to
determine the value of retrosynthetic positions. Here, we propose to model
retrosynthesis as a Markov Decision Process. In combination with a Deep Neural
Network policy learned from essentially the complete published knowledge of
chemistry, Monte Carlo Tree Search (MCTS) can be used to evaluate positions. In
exploratory studies, we demonstrate that MCTS with neural network policies
outperforms the traditionally used best-first search with hand-coded
heuristics. | ['Mike Preuß', 'Marwin Segler', 'Mark P. Waller'] | 2017-01-31 | null | null | null | null | ['retrosynthesis'] | ['medical'] | [ 4.62896079e-01 1.23681501e-01 -5.86320758e-01 1.19424984e-01
-6.17579401e-01 -1.28447211e+00 5.89534461e-01 4.74445134e-01
-5.02473593e-01 1.37554622e+00 -3.37725133e-02 -8.79187167e-01
-3.46331447e-02 -9.59764540e-01 -8.00942600e-01 -7.87030697e-01
-1.27498478e-01 7.52850533e-01 3.15348874e-03 9.78157446e-02
6.18580282e-01 9.12504494e-01 -1.14092541e+00 1.28594428e-01
8.50292265e-01 9.24553394e-01 5.21476626e-01 7.59466767e-01
-1.26951486e-01 6.80950582e-01 -4.10467833e-01 -4.00518239e-01
3.04114848e-01 -4.85021532e-01 -1.06790948e+00 -3.54734927e-01
-1.73283681e-01 -1.70360371e-01 1.45684004e-01 1.07839644e+00
3.67385477e-01 4.31554526e-01 6.33252800e-01 -6.11185968e-01
-2.13010609e-01 9.64771390e-01 -1.78429519e-03 7.29378983e-02
2.93175370e-01 3.09159726e-01 1.08271348e+00 -6.25049055e-01
6.97316706e-01 1.25235105e+00 4.16536689e-01 4.31474060e-01
-1.27586031e+00 -6.81967676e-01 1.99481323e-01 6.94412813e-02
-1.18181586e+00 -4.31573153e-01 3.28555524e-01 -4.14530396e-01
1.30158091e+00 7.75399953e-02 9.82071638e-01 1.03546345e+00
5.74000716e-01 3.28846127e-01 8.46561491e-01 -3.31685901e-01
8.91602159e-01 -3.39914441e-01 -7.05870748e-01 7.81718194e-01
1.15197562e-02 5.03851831e-01 -3.66429865e-01 -1.09054662e-01
5.51550508e-01 -2.85581946e-02 -1.47465821e-02 -3.47470701e-01
-1.07208109e+00 1.02310169e+00 5.77913165e-01 2.77802080e-01
-5.18162310e-01 5.74146807e-01 4.04646546e-01 -8.10739622e-02
-1.15406094e-02 1.42254150e+00 -7.75420070e-01 -1.24984095e-03
-1.04875720e+00 2.72829771e-01 1.07715225e+00 7.23935246e-01
6.79983437e-01 6.11090027e-02 1.59202218e-01 1.92801863e-01
1.67842194e-01 6.84000598e-03 2.91178167e-01 -1.25420332e+00
1.26111731e-01 1.14592232e-01 3.18251371e-01 -3.56112748e-01
-4.46699917e-01 -2.60065585e-01 -5.57921052e-01 9.71990600e-02
4.74615276e-01 -4.08511788e-01 -1.20552814e+00 1.43793201e+00
4.35051203e-01 -3.17720085e-01 9.44721028e-02 4.00889844e-01
5.55704117e-01 1.09644461e+00 3.90166312e-01 -6.03274643e-01
9.61958766e-01 -9.28105116e-01 -3.87011528e-01 4.16345485e-02
5.04731536e-01 -6.17494106e-01 2.29454264e-01 9.63706851e-01
-1.39715648e+00 -1.65559128e-01 -1.06798410e+00 3.48886549e-02
-9.33445692e-01 -3.15512866e-01 9.71149206e-01 8.16386759e-01
-1.00310874e+00 1.42127752e+00 -7.69957125e-01 1.90563519e-02
5.32809794e-01 6.90907955e-01 3.71823870e-02 1.19389124e-01
-1.13094509e+00 9.87682402e-01 1.18361986e+00 2.46376038e-01
-1.82655489e+00 -6.79266393e-01 -6.28924072e-01 2.95122713e-01
9.63829160e-01 -6.36145949e-01 1.63949621e+00 -8.18598509e-01
-1.91855371e+00 3.77021968e-01 9.77418479e-03 -5.77045739e-01
3.90593380e-01 1.95997715e-01 4.79593799e-02 1.46692827e-01
-5.68016917e-02 1.02503181e+00 7.22987056e-01 -1.13286257e+00
-5.76147914e-01 -4.23583351e-02 3.21860939e-01 5.57173193e-02
4.71928090e-01 2.70496104e-02 7.34254643e-02 -2.39404112e-01
-9.91763547e-02 -8.61784816e-01 -9.46804702e-01 -3.85368139e-01
-5.95681369e-01 -2.24082008e-01 -1.98895286e-04 -4.65316862e-01
1.11004603e+00 -1.35069966e+00 4.25630152e-01 3.28002870e-01
9.85989273e-02 1.08446218e-01 -1.27050206e-01 8.43975484e-01
-3.63566965e-01 4.32798117e-01 -1.64516106e-01 4.70994920e-01
-1.77455097e-01 1.42583713e-01 -2.43809834e-01 1.09875605e-01
4.10876200e-02 9.07640815e-01 -1.50473499e+00 -3.64683211e-01
4.38550860e-02 -6.76534846e-02 -7.10151970e-01 -2.19023868e-01
-1.09111869e+00 4.96605903e-01 -6.69385433e-01 9.63990569e-01
2.31979713e-01 -3.73794466e-01 8.23937476e-01 3.28648053e-02
-4.99265730e-01 5.82927048e-01 -7.26609588e-01 1.85766637e+00
-4.23257887e-01 8.00337046e-02 -2.96146423e-01 -9.74978387e-01
4.76030529e-01 3.83498043e-01 4.41471845e-01 -5.78613877e-01
2.29239941e-01 2.44975194e-01 2.39888430e-01 -1.15766354e-01
4.62111235e-01 -5.47747433e-01 3.14609110e-02 2.98955053e-01
1.42296255e-02 -4.87003386e-01 8.02033246e-01 -4.28360887e-02
1.13091147e+00 5.49911022e-01 8.68948400e-01 -1.86227426e-01
4.74790394e-01 3.67978007e-01 5.42636573e-01 8.59021187e-01
3.34814005e-02 5.24742827e-02 6.80092633e-01 -7.98299730e-01
-1.25550354e+00 -1.00050330e+00 1.91351414e-01 1.20657194e+00
-2.35962316e-01 -5.15709043e-01 -8.16265464e-01 -7.55715132e-01
-3.20784688e-01 7.19920993e-01 -4.75402087e-01 1.02655543e-02
-5.32148778e-01 -5.63676119e-01 2.40005821e-01 3.49763215e-01
2.88879219e-02 -1.41360247e+00 -6.62675917e-01 7.83869684e-01
3.12522471e-01 -6.17365062e-01 -2.24828422e-01 9.86713767e-01
-8.73996079e-01 -1.06149375e+00 -5.34885466e-01 -5.35426497e-01
5.14108360e-01 -7.35417977e-02 1.09441185e+00 -7.42427781e-02
-4.00189906e-01 -3.19169074e-01 -1.04756713e-01 -4.77079898e-01
-8.32025528e-01 3.04657608e-01 -1.62850291e-01 -7.63988912e-01
-3.30199987e-01 -6.08452737e-01 -8.31133246e-01 -1.42157942e-01
-6.94829762e-01 -6.66582063e-02 6.66804790e-01 8.36518049e-01
7.23202467e-01 4.41894144e-01 4.22308743e-01 -6.05404317e-01
5.43311656e-01 -3.20090562e-01 -1.22692132e+00 3.65087926e-01
-6.21786952e-01 5.60187161e-01 8.99114847e-01 -3.82624924e-01
-8.78869176e-01 4.99767363e-01 7.86280353e-03 -1.73965886e-01
-1.14967069e-02 7.06658602e-01 -1.41343653e-01 -1.24654636e-01
5.28222263e-01 1.37791470e-01 -4.07979310e-01 -2.28008971e-01
6.66881979e-01 -1.31607100e-01 2.65858918e-01 -9.61027682e-01
3.05439919e-01 5.93447983e-02 5.26661873e-01 -5.11789620e-01
-8.40144396e-01 -1.04709834e-01 -6.52961373e-01 5.34247868e-02
9.78233457e-01 -5.14042377e-01 -1.27178597e+00 -2.87172467e-01
-1.29885447e+00 -6.70042455e-01 -1.38503730e-01 2.38994792e-01
-8.33810270e-01 2.31891081e-01 -4.07876313e-01 -7.99644947e-01
-1.91236794e-01 -1.33009553e+00 8.31075907e-01 2.45836258e-01
-4.17860061e-01 -8.93870592e-01 2.53643751e-01 1.06683843e-01
-2.34028041e-01 2.79683799e-01 1.26231802e+00 -6.68222725e-01
-1.15516579e+00 1.27434969e-01 5.99004067e-02 -4.80675958e-02
2.26500273e-01 2.21763358e-01 -7.48326004e-01 -8.49550962e-02
-5.37828922e-01 -3.97211581e-01 8.24136198e-01 6.75350249e-01
1.56909335e+00 -6.16923988e-01 -5.80468714e-01 4.60440040e-01
1.32827020e+00 1.09963548e+00 5.38022399e-01 4.45474803e-01
5.37846506e-01 6.89950645e-01 4.85506624e-01 4.38059568e-01
-3.12787056e-01 4.73271543e-03 5.92708707e-01 3.98508877e-01
5.01353502e-01 -5.44840515e-01 3.04330103e-02 1.72525328e-02
-4.11077797e-01 -5.18602967e-01 -8.62003446e-01 1.69615522e-01
-1.37500560e+00 -1.12065578e+00 5.72787404e-01 2.16861796e+00
1.12168622e+00 2.26927251e-01 1.79405838e-01 1.87276721e-01
4.71189022e-01 3.83768007e-02 -7.60255814e-01 -8.41398180e-01
5.18302798e-01 7.20869303e-01 1.08548725e+00 5.91196716e-01
-1.02034152e+00 1.42206252e+00 7.87222195e+00 9.75327313e-01
-1.03594279e+00 -5.05556285e-01 6.23967409e-01 -2.54632592e-01
-2.26089969e-01 4.75413829e-01 -6.08526230e-01 3.64057899e-01
1.12715888e+00 -3.16858888e-01 9.71100450e-01 6.87153995e-01
4.53205258e-01 -3.40201378e-01 -1.51237655e+00 5.20456851e-01
-7.39148796e-01 -2.00639558e+00 2.08770752e-01 1.85285568e-01
8.68192315e-01 -2.08630696e-01 -1.19826339e-01 1.21539228e-01
1.03678107e+00 -1.35213029e+00 7.79146075e-01 1.24567240e-01
4.77120787e-01 -1.16635430e+00 1.11460723e-01 2.14258090e-01
-9.02590692e-01 -4.28692222e-01 -1.39818296e-01 1.00575358e-01
7.09819645e-02 2.00127855e-01 -1.37129068e+00 6.23389125e-01
2.66423702e-01 4.53420818e-01 -5.00019416e-02 1.06914127e+00
-5.24746358e-01 2.51254976e-01 -1.47586122e-01 -5.60891807e-01
9.01521683e-01 -6.01278424e-01 8.49136040e-02 8.94635260e-01
3.37803245e-01 1.67990401e-01 3.42248976e-01 9.79391932e-01
-2.22077996e-01 -1.04905114e-01 -5.68965077e-01 -7.99881279e-01
5.72444499e-01 1.01341271e+00 -1.33866608e+00 -5.95515788e-01
2.05459401e-01 6.11760020e-01 1.35967387e-02 4.69701111e-01
-5.33803940e-01 -1.78483084e-01 2.15881541e-01 -4.24763411e-01
6.97076797e-01 -2.64779866e-01 -3.33236828e-02 -5.23803413e-01
-7.84504771e-01 -1.12617815e+00 3.47963452e-01 -8.70217562e-01
-6.28792107e-01 1.67167768e-01 -4.96651083e-02 -5.55446982e-01
-3.15466791e-01 -7.70870686e-01 -5.04975855e-01 6.88491046e-01
-1.22621608e+00 -4.19187754e-01 4.96645331e-01 -6.33543655e-02
8.05018425e-01 1.43590467e-02 7.26968884e-01 -3.11929852e-01
-5.97720385e-01 -5.16101047e-02 3.90348822e-01 -5.01907647e-01
2.67777532e-01 -1.43478668e+00 3.59848082e-01 4.47842002e-01
-5.23040518e-02 7.70875931e-01 9.71131682e-01 -8.12863648e-01
-1.39905918e+00 -7.91339457e-01 6.00439608e-01 -1.53099611e-01
9.12715197e-01 -2.35493556e-01 -4.77652401e-01 3.29443961e-01
2.66623795e-01 -4.90511000e-01 5.49963415e-01 -4.11640182e-02
-6.85328171e-02 1.00191250e-01 -1.10584772e+00 8.69836569e-01
9.66419280e-01 -2.01107323e-01 -2.50589520e-01 6.47293210e-01
8.20740163e-01 -3.81720543e-01 -9.09014046e-01 9.38579291e-02
8.60146046e-01 -7.20085621e-01 1.15267313e+00 -8.68002474e-01
8.01109612e-01 -1.37121558e-01 2.92005111e-02 -1.27267659e+00
-2.03599885e-01 -1.17585540e+00 2.57115010e-02 4.32920426e-01
7.74612129e-01 -3.49271804e-01 9.68838751e-01 2.57798731e-01
-8.18307791e-03 -7.64196038e-01 -8.03002238e-01 -8.90564144e-01
3.37644398e-01 1.20461151e-01 8.79398406e-01 7.60423481e-01
-9.89763141e-02 5.97096920e-01 -5.81184626e-02 4.12081257e-02
3.99783671e-01 2.55419910e-01 2.07253963e-01 -1.08050275e+00
-2.85962641e-01 -8.00534248e-01 3.38740170e-01 -8.65241885e-01
1.65248483e-01 -8.67532194e-01 3.04325759e-01 -1.47880816e+00
1.03054255e-01 -3.19768041e-01 -3.05171430e-01 3.60154897e-01
2.14730740e-01 -4.64564919e-01 -1.98750924e-02 -3.66545796e-01
-5.20688951e-01 4.28844094e-01 1.54739487e+00 -3.12230766e-01
-4.04177278e-01 -1.26297966e-01 -6.74800456e-01 6.49699748e-01
8.75987768e-01 -7.42580771e-01 -4.98781085e-01 2.99504369e-01
8.68704855e-01 6.43697262e-01 -7.44092464e-02 -5.87292671e-01
1.41618475e-01 -7.67068863e-01 5.61164439e-01 -9.06784654e-01
1.64581820e-01 -5.73780179e-01 2.62013912e-01 9.74113941e-01
-6.06052458e-01 -1.65120289e-01 1.56117573e-01 5.99477768e-01
3.02074432e-01 -7.26879656e-01 5.33447683e-01 -9.48596120e-01
-3.86258274e-01 5.78491390e-01 -9.15313840e-01 -3.31244141e-01
8.97500992e-01 -3.34404558e-01 1.25463888e-01 4.43116017e-02
-8.72639060e-01 3.41273099e-01 3.24908257e-01 -7.28809834e-02
3.65396112e-01 -9.22255099e-01 6.57277927e-02 -5.36737323e-01
-2.46416390e-01 3.50998759e-01 -1.57878220e-01 2.65397817e-01
-8.78525198e-01 8.75633955e-01 -1.42837495e-01 -1.26467928e-01
-1.00516832e+00 1.24573386e+00 5.20776212e-01 -5.62610507e-01
1.28188645e-02 7.12095082e-01 3.04919899e-01 -3.30521613e-01
1.92785755e-01 -7.39946663e-01 5.37327863e-02 1.90386951e-01
2.65659213e-01 4.08778340e-01 -1.12991050e-01 3.03732872e-01
-3.37067455e-01 1.95259243e-01 -7.14964420e-02 -2.20676824e-01
1.27000523e+00 2.70197451e-01 -1.62649453e-01 -2.89522707e-02
8.11192989e-01 -3.26514244e-01 -1.26952946e+00 3.16996992e-01
2.85039544e-01 5.04187271e-02 6.35984242e-02 -1.05179060e+00
-4.55281198e-01 7.19038069e-01 -2.56856550e-02 2.44104207e-01
7.38637567e-01 -2.18256325e-01 5.09469390e-01 1.08935189e+00
3.16534370e-01 -1.25083232e+00 1.77052945e-01 3.96987826e-01
8.05901468e-01 -1.00337613e+00 3.65385681e-01 -1.38197199e-01
-2.10764989e-01 1.37171888e+00 1.23859704e-01 1.45451412e-01
1.82241112e-01 -2.08586045e-02 -6.12815201e-01 -2.16236115e-01
-9.41506207e-01 -2.05248237e-01 -1.96376830e-01 6.06920123e-01
3.03691655e-01 -5.97502664e-02 -2.20027626e-01 -4.78332676e-02
-4.33117867e-01 7.29331821e-02 4.15527433e-01 1.08803058e+00
-7.20016599e-01 -1.42692888e+00 -3.34066331e-01 8.78664607e-04
-6.59512222e-01 -4.94660228e-01 -7.55908489e-01 6.39625371e-01
2.65629053e-01 7.97545075e-01 -2.93096662e-01 1.72540158e-01
-1.73638374e-01 8.17637518e-02 1.11170733e+00 -6.76953077e-01
-5.77472925e-01 2.02244997e-01 5.72635770e-01 -6.75244391e-01
-3.26378495e-01 -4.48537946e-01 -1.15256703e+00 -2.24399909e-01
-1.83324262e-01 4.67758834e-01 8.09377849e-01 1.11057138e+00
-6.66434318e-02 6.68980062e-01 7.45189011e-01 -1.01550615e+00
-4.23469543e-01 -3.11403096e-01 -2.69027203e-01 -6.75871968e-01
3.52506429e-01 -5.02409279e-01 7.70398825e-02 1.76050290e-01] | [4.528843402862549, 6.066320896148682] |
29d94fc8-284d-4b1c-8402-542601a5b3da | loss-aversively-fair-classification | 2105.04273 | null | https://arxiv.org/abs/2105.04273v1 | https://arxiv.org/pdf/2105.04273v1.pdf | Loss-Aversively Fair Classification | The use of algorithmic (learning-based) decision making in scenarios that affect human lives has motivated a number of recent studies to investigate such decision making systems for potential unfairness, such as discrimination against subjects based on their sensitive features like gender or race. However, when judging the fairness of a newly designed decision making system, these studies have overlooked an important influence on people's perceptions of fairness, which is how the new algorithm changes the status quo, i.e., decisions of the existing decision making system. Motivated by extensive literature in behavioral economics and behavioral psychology (prospect theory), we propose a notion of fair updates that we refer to as loss-averse updates. Loss-averse updates constrain the updates to yield improved (more beneficial) outcomes to subjects compared to the status quo. We propose tractable proxy measures that would allow this notion to be incorporated in the training of a variety of linear and non-linear classifiers. We show how our proxy measures can be combined with existing measures for training nondiscriminatory classifiers. Our evaluation using synthetic and real-world datasets demonstrates that the proposed proxy measures are effective for their desired tasks. | ['Krishna P. Gummadi', 'Adish Singla', 'Muhammad Bilal Zafar', 'Junaid Ali'] | 2021-05-10 | null | null | null | null | ['classification'] | ['methodology'] | [ 1.59725711e-01 1.93753377e-01 -5.70240080e-01 -9.72810388e-01
3.77310663e-02 -3.49755108e-01 5.50333083e-01 4.54031467e-01
-8.55882883e-01 1.06492639e+00 1.19937569e-01 -6.61602378e-01
-1.51566625e-01 -8.31526756e-01 -2.64802039e-01 -4.47748899e-01
1.35515749e-01 1.96990415e-01 -2.76961565e-01 -1.10514618e-01
7.39664495e-01 2.71466762e-01 -1.54445255e+00 -8.55674744e-02
1.11766756e+00 1.07472289e+00 -6.03480816e-01 3.32159281e-01
2.38311827e-01 7.65439808e-01 -5.41345000e-01 -1.13200021e+00
7.18409002e-01 -2.96968728e-01 -5.18890440e-01 -3.63474399e-01
4.71910834e-01 -6.66272104e-01 1.02918811e-01 1.37176943e+00
5.32921851e-01 2.16872916e-01 8.58641267e-01 -1.56279111e+00
-8.83529782e-01 7.72023082e-01 -5.64657986e-01 1.73652664e-01
2.45518714e-01 2.40085453e-01 1.15012348e+00 -2.93363273e-01
2.84929246e-01 1.75118399e+00 6.59730613e-01 7.64162779e-01
-1.25205863e+00 -1.02021003e+00 7.85449743e-02 1.83768034e-01
-9.72114921e-01 -9.04403985e-01 6.41887367e-01 -5.61084807e-01
2.83323377e-01 5.67583561e-01 4.11958158e-01 7.54316092e-01
5.15690327e-01 2.72301078e-01 1.72337890e+00 -4.74093229e-01
6.37905836e-01 4.91222501e-01 2.32520312e-01 3.68919939e-01
9.15100098e-01 7.07047880e-01 -2.57823795e-01 -5.70698977e-01
3.79759431e-01 -1.84420705e-01 1.43350154e-01 -4.28365052e-01
-5.34999311e-01 1.14564526e+00 4.18325484e-01 -1.15918271e-01
-5.50108314e-01 7.29108304e-02 3.07619840e-01 5.11004031e-01
5.56917727e-01 7.98520863e-01 -1.10595264e-01 5.16852885e-02
-7.27931976e-01 6.32303178e-01 7.68832862e-01 1.88064754e-01
6.47814035e-01 -1.51245698e-01 -7.57450223e-01 5.88877082e-01
2.38667652e-01 3.06601733e-01 3.81154746e-01 -1.37221658e+00
3.64339620e-01 4.34239686e-01 5.79703748e-01 -1.14206123e+00
-2.18427390e-01 -2.76274502e-01 -5.37454069e-01 6.33186638e-01
7.09357858e-01 -5.04249632e-01 -4.62303162e-01 2.05446744e+00
4.08737540e-01 -3.89778614e-01 -2.72728324e-01 9.98587191e-01
-3.42945680e-02 1.30439252e-01 6.10949636e-01 -4.34169114e-01
1.04738998e+00 -2.43961379e-01 -6.98834777e-01 -3.15912575e-01
4.37797844e-01 -5.77250063e-01 8.85618687e-01 8.91123638e-02
-1.01719975e+00 -1.54956222e-01 -8.30388784e-01 2.92759269e-01
-1.05931386e-01 -4.19419199e-01 1.10276294e+00 1.44632304e+00
-6.76514089e-01 8.72022271e-01 -1.37807027e-01 -3.01240474e-01
8.75177622e-01 2.47739196e-01 1.73228905e-01 2.21475184e-01
-1.33411682e+00 1.26902473e+00 3.93557064e-02 -1.90909728e-01
-5.79176724e-01 -7.31005669e-01 -5.41275263e-01 3.30568165e-01
4.84806478e-01 -9.03082132e-01 1.25392842e+00 -1.54681706e+00
-1.40796268e+00 1.01973855e+00 2.74283409e-01 -7.75263488e-01
9.84038413e-01 -2.71907747e-02 -1.49071515e-01 -2.66434669e-01
3.03092360e-01 6.36549175e-01 8.81544292e-01 -1.00694382e+00
-6.39185607e-01 -6.03260875e-01 3.22996616e-01 5.43064892e-01
-4.03930455e-01 4.63386387e-01 1.00486374e+00 -7.14748442e-01
-5.00724256e-01 -7.84397066e-01 -4.20021206e-01 3.72072071e-01
-3.87631595e-01 -7.12748915e-02 3.03883344e-01 -4.20656055e-01
1.14784515e+00 -1.82859921e+00 -7.64671683e-01 3.80090415e-01
2.17877224e-01 3.63539398e-01 1.45676630e-02 -4.05830704e-02
-4.90940325e-02 4.34766978e-01 -1.89647377e-01 1.80856250e-02
3.70306939e-01 -1.31120071e-01 -1.14397414e-01 6.72975957e-01
-2.17877150e-01 6.19799852e-01 -5.91204286e-01 -4.00806576e-01
6.83039725e-02 -2.55100459e-01 -7.05047190e-01 1.61764100e-01
3.02221149e-01 -2.84502879e-02 -4.30397004e-01 5.53684711e-01
7.95526147e-01 3.81436676e-01 -3.25937718e-02 4.78103310e-01
-1.21498503e-01 4.51557636e-01 -1.04342258e+00 6.19879603e-01
3.94209288e-03 2.68734634e-01 8.58472777e-04 -1.14554334e+00
8.92663181e-01 1.91116817e-02 2.15714276e-01 -7.12156892e-01
3.74826193e-01 1.89764183e-02 3.96576583e-01 -2.27324590e-01
6.28280938e-01 -8.59685183e-01 -9.15697515e-02 7.28853524e-01
-4.54088509e-01 1.75163999e-01 -1.05089426e-01 4.06524818e-03
6.74527764e-01 -3.31080854e-01 8.79863977e-01 -6.46026194e-01
3.55065554e-01 -2.75522500e-01 9.83576238e-01 1.10255098e+00
-1.10593677e+00 1.61396842e-02 6.06082201e-01 -4.04504120e-01
-9.21135962e-01 -1.08503664e+00 -3.53290141e-01 1.36498249e+00
1.09467804e-01 3.62988204e-01 -6.97691441e-01 -6.73090637e-01
6.43800795e-01 1.14287686e+00 -8.16183150e-01 -6.12723112e-01
5.86892590e-02 -1.07417190e+00 5.13617218e-01 3.26137334e-01
5.11448801e-01 -6.79246485e-01 -9.52551186e-01 -1.29103854e-01
1.44044891e-01 -4.83956248e-01 -4.38674420e-01 -2.83894897e-01
-8.23922694e-01 -9.69532788e-01 -5.85409343e-01 -2.61988863e-02
6.00025773e-01 1.13058925e-01 9.32582021e-01 1.04292914e-01
-3.88801061e-02 2.83571184e-01 -1.16592012e-01 -9.52396095e-01
-3.75499040e-01 -4.69696760e-01 5.01757562e-01 2.18817547e-01
6.23695016e-01 -3.43016207e-01 -7.50073195e-01 2.83250988e-01
-5.81913710e-01 -3.84267598e-01 3.51600617e-01 6.62584305e-01
-1.45747453e-01 -5.11974953e-02 9.79102612e-01 -1.12736189e+00
1.23790491e+00 -5.17189205e-01 -6.71612144e-01 2.96381772e-01
-1.25281668e+00 7.78898876e-03 2.97887385e-01 -5.70813298e-01
-1.48378265e+00 -4.33637708e-01 2.69042224e-01 1.55647799e-01
2.50680111e-02 2.26224497e-01 -1.04789764e-01 -3.38121831e-01
8.33683848e-01 -5.29100657e-01 3.02367359e-02 -1.83489114e-01
3.72292161e-01 8.15325677e-01 3.38815033e-01 -8.54863405e-01
6.61987960e-01 1.85921043e-01 -1.83372591e-02 -1.89609349e-01
-7.40861356e-01 -3.65985781e-02 -1.06395327e-01 -3.18908274e-01
5.28573573e-01 -5.46299100e-01 -1.21294832e+00 3.70805502e-01
-7.05206811e-01 4.35338356e-02 -4.92982000e-01 5.03924489e-01
-4.33501005e-01 1.84264749e-01 -3.48934680e-01 -1.36340451e+00
-2.58699059e-01 -7.48173773e-01 1.74935788e-01 4.96938735e-01
-5.16846538e-01 -8.93583894e-01 1.45389270e-02 4.78156090e-01
5.90509832e-01 1.19663209e-01 9.48211908e-01 -8.10677826e-01
-2.97373403e-02 -3.05159092e-01 -2.79628068e-01 4.29949343e-01
-2.35250499e-02 1.42218126e-03 -1.09387076e+00 -1.79914713e-01
1.53328598e-01 -4.06961292e-01 8.26512456e-01 6.73951924e-01
9.81778085e-01 -8.79404128e-01 -3.41770835e-02 2.39418939e-01
1.11481833e+00 3.01358998e-01 5.29790521e-01 3.62187415e-01
1.27763122e-01 1.11046970e+00 6.68126941e-01 8.41699064e-01
6.04922652e-01 5.76142907e-01 3.67403537e-01 9.17505771e-02
4.87593412e-01 -3.63511175e-01 3.81039858e-01 -1.96636081e-01
-2.35482454e-01 2.00623170e-01 -6.53965473e-01 1.46099880e-01
-1.82035124e+00 -1.18561649e+00 3.51489365e-01 2.69698334e+00
9.40470695e-01 2.70397961e-01 3.56080920e-01 1.16532065e-01
1.10625696e+00 6.38252199e-02 -8.79516184e-01 -1.16239309e+00
7.90345669e-02 -7.03368783e-02 7.91053951e-01 5.70405483e-01
-1.12359357e+00 6.22118831e-01 6.88356400e+00 6.81927085e-01
-9.31739867e-01 -6.20746352e-02 1.52402234e+00 -1.06368825e-01
-5.09281993e-01 1.98564768e-01 -4.10212696e-01 5.41486561e-01
8.74455452e-01 -9.29396331e-01 1.84993193e-01 8.23989451e-01
4.32681561e-01 -4.74279612e-01 -1.24576819e+00 6.52523518e-01
-2.53158718e-01 -9.28411663e-01 -5.43970391e-02 1.71722040e-01
7.78462470e-01 -8.23367059e-01 2.76000142e-01 3.96991760e-01
7.80314207e-01 -1.16688883e+00 1.04580808e+00 5.48830509e-01
4.85823244e-01 -7.91005909e-01 7.91974425e-01 3.55217069e-01
-2.80159324e-01 -4.24447119e-01 -5.87961793e-01 -9.64298785e-01
-2.04815269e-01 9.20848429e-01 -8.02993059e-01 8.53256583e-02
3.48058432e-01 1.52399898e-01 -5.37436604e-01 9.82254088e-01
-1.29512787e-01 6.10558331e-01 -9.47488844e-03 -2.46759862e-01
-1.40434327e-02 -1.83880389e-01 4.21982050e-01 8.21679950e-01
1.46373123e-01 2.45746553e-01 -1.59159988e-01 1.19893765e+00
-2.64127970e-01 3.39725584e-01 -5.93686521e-01 1.53148368e-01
7.59124577e-01 9.63464379e-01 -4.19636190e-01 -2.24135309e-01
-6.57282397e-02 3.95681053e-01 4.29065451e-02 2.07238004e-01
-6.04234517e-01 -8.08029175e-02 1.06697190e+00 1.16443537e-01
-4.63431835e-01 3.42073619e-01 -9.68695581e-01 -1.06313372e+00
-1.44825518e-01 -8.04344654e-01 5.71492016e-01 -2.99025118e-01
-1.62796998e+00 -2.45529890e-01 9.10464972e-02 -8.61577928e-01
-7.97772035e-02 -2.95017719e-01 -7.59547651e-01 9.63677049e-01
-1.42232144e+00 -4.75826532e-01 3.66568714e-01 2.83015460e-01
-1.19579598e-01 -1.48035422e-01 4.32258278e-01 4.96350713e-02
-5.89076281e-01 8.66386116e-01 -1.24742836e-03 -5.80047220e-02
1.03473711e+00 -1.12826216e+00 3.11802924e-02 8.95086646e-01
-4.30122316e-01 5.27170658e-01 8.83510530e-01 -5.21944880e-01
-5.87276876e-01 -7.63425291e-01 1.07946825e+00 -3.61904770e-01
3.66109729e-01 -1.10441588e-01 -5.62314153e-01 4.87286001e-01
-1.42809778e-01 -2.03306392e-01 8.94910574e-01 3.36560249e-01
-4.37558442e-01 -4.02348220e-01 -1.89250922e+00 6.73590302e-01
8.54586840e-01 -2.68366843e-01 -7.64362574e-01 -1.58268452e-01
3.31659913e-01 1.90505832e-01 -4.20783222e-01 2.72905916e-01
9.46460605e-01 -1.19178987e+00 9.65734601e-01 -9.80094492e-01
5.54669201e-01 3.43444347e-01 -1.19060792e-01 -1.45367384e+00
-5.36107838e-01 -5.52437365e-01 3.76040280e-01 1.20166039e+00
2.31733769e-01 -1.06946254e+00 5.96820891e-01 1.54896998e+00
4.69696820e-01 -4.98757541e-01 -1.08865201e+00 -6.79644525e-01
6.11480594e-01 -1.33384407e-01 8.13914716e-01 1.32756853e+00
8.24165568e-02 -5.54213650e-04 -6.38756096e-01 -1.78275451e-01
9.10726190e-01 1.04912035e-01 4.98463333e-01 -1.51575482e+00
-2.27218065e-02 -7.70838976e-01 -5.00195622e-01 -3.98099482e-01
2.50053883e-01 -6.35461092e-01 -1.31603509e-01 -7.06576347e-01
4.63942647e-01 -7.24998713e-01 -4.34281528e-01 4.31821197e-01
-6.24627650e-01 -1.67168871e-01 5.49281597e-01 4.27971222e-02
-3.13795686e-01 5.88643968e-01 9.93984282e-01 -7.33155683e-02
-5.19200005e-02 3.83138001e-01 -1.57833672e+00 9.11108255e-01
8.51967990e-01 -6.87072635e-01 -2.05415606e-01 7.02123120e-02
2.35049278e-01 1.00497320e-01 4.08611357e-01 -7.82022655e-01
9.52696055e-02 -9.73026335e-01 3.80137086e-01 5.04832268e-01
-4.50983495e-02 -6.07905567e-01 -2.05827594e-01 7.26133704e-01
-1.03552735e+00 -1.50827870e-01 -2.27725387e-01 3.39417368e-01
2.02675939e-01 -4.22401100e-01 1.10685599e+00 -7.94266686e-02
-2.05344766e-01 1.62273332e-01 -3.94003689e-01 1.67261973e-01
1.22127450e+00 -1.87134922e-01 -5.07971287e-01 -7.43231356e-01
-2.74463147e-01 2.05921575e-01 4.73657489e-01 2.17332572e-01
2.34624982e-01 -1.39477491e+00 -9.93913352e-01 -1.70159370e-01
-3.82526894e-03 -9.19682026e-01 -7.73366094e-02 5.18409967e-01
-1.75745144e-01 4.02172744e-01 -8.43088031e-01 1.97996601e-01
-1.15409875e+00 4.12929416e-01 5.53157985e-01 -2.16939554e-01
1.66995928e-01 6.10653937e-01 3.89926940e-01 -3.73758435e-01
1.33657426e-01 -3.28640454e-02 -1.60751179e-01 1.90514520e-01
5.41472912e-01 8.14905167e-01 -2.32523844e-01 -4.38401103e-01
-3.51182014e-01 -1.09952755e-01 8.93209968e-03 -2.80574739e-01
1.00246310e+00 -3.63203973e-01 2.14203363e-04 3.25046331e-01
5.16621232e-01 -1.34762330e-02 -1.11342561e+00 -2.01456308e-01
1.70824707e-01 -1.17910528e+00 4.94102389e-02 -1.05076575e+00
-7.43024111e-01 6.69950128e-01 6.20728850e-01 3.02760988e-01
1.03459585e+00 -7.00481713e-01 2.23438308e-01 4.66634572e-01
4.80942249e-01 -1.36112761e+00 -4.02435243e-01 -6.55773580e-02
7.11628497e-01 -1.53619289e+00 3.41841787e-01 -2.75319338e-01
-8.16877604e-01 7.11809933e-01 5.56616724e-01 -2.24550739e-02
6.59703076e-01 -1.85925484e-01 1.22471362e-01 3.94940376e-01
-7.37402916e-01 -4.84141111e-02 1.82952330e-01 6.79560959e-01
4.01601106e-01 8.34199250e-01 -1.10536110e+00 8.55068803e-01
-3.79575878e-01 2.52568662e-01 8.31909299e-01 6.65057123e-01
-5.65185368e-01 -1.05950069e+00 -6.36554420e-01 9.91370797e-01
-7.54238546e-01 -3.43050286e-02 -4.66043502e-01 4.03773218e-01
1.71186119e-01 1.02472889e+00 9.16901380e-02 -1.61772281e-01
1.37415215e-01 3.11336704e-02 3.00598562e-01 -4.69343871e-01
-6.25631392e-01 -6.12755716e-01 3.52361619e-01 -5.78614831e-01
-4.72527593e-01 -9.35355544e-01 -5.94875395e-01 -9.04552460e-01
-4.26892787e-02 1.39924124e-01 3.24416995e-01 9.77810681e-01
6.83901906e-02 -1.80016130e-01 8.84793282e-01 -4.47964370e-01
-1.38407302e+00 -6.56290233e-01 -8.71254385e-01 6.85231566e-01
1.65949628e-01 -8.48954082e-01 -4.58459616e-01 -3.84305209e-01] | [8.907069206237793, 5.303138732910156] |
cc219936-ab2f-42cf-b0c2-06eaae0f1af4 | guiding-physical-intuition-with-neural | null | null | https://openreview.net/forum?id=BylctiCctX | https://openreview.net/pdf?id=BylctiCctX | Guiding Physical Intuition with Neural Stethoscopes | Model interpretability and systematic, targeted model adaptation present central challenges in deep learning. In the domain of intuitive physics, we study the task of visually predicting stability of block towers with the goal of understanding and influencing the model's reasoning. Our contributions are two-fold. Firstly, we introduce neural stethoscopes as a framework for quantifying the degree of importance of specific factors of influence in deep networks as well as for actively promoting and suppressing information as appropriate. In doing so, we unify concepts from multitask learning as well as training with auxiliary and adversarial losses. Secondly, we deploy the stethoscope framework to provide an in-depth analysis of a state-of-the-art deep neural network for stability prediction, specifically examining its physical reasoning. We show that the baseline model is susceptible to being misled by incorrect visual cues. This leads to a performance breakdown to the level of random guessing when training on scenarios where visual cues are inversely correlated with stability. Using stethoscopes to promote meaningful feature extraction increases performance from 51% to 90% prediction accuracy. Conversely, training on an easy dataset where visual cues are positively correlated with stability, the baseline model learns a bias leading to poor performance on a harder dataset. Using an adversarial stethoscope, the network is successfully de-biased, leading to a performance increase from 66% to 88%. | ['Alex Bewley', 'Markus Wulfmeier', 'Ingmar Posner', 'Fabian Fuchs', 'Andrea Vedaldi', 'Oliver Groth', 'Adam Kosiorek'] | 2019-05-01 | null | null | null | iclr-2019-5 | ['physical-intuition'] | ['reasoning'] | [ 1.69776738e-01 4.04814810e-01 1.60217449e-01 -3.52371112e-02
-4.49861586e-01 -6.34615600e-01 5.41655481e-01 1.44364089e-01
-2.52873033e-01 3.97828460e-01 6.76761344e-02 -5.03446758e-01
-1.26494244e-01 -6.63714647e-01 -1.23562038e+00 -7.02130914e-01
-4.05713022e-02 1.05619140e-01 2.83107638e-01 -5.31931996e-01
2.00082570e-01 4.56200182e-01 -1.25735795e+00 2.20748216e-01
6.19895697e-01 1.13456917e+00 -1.62380621e-01 9.03077841e-01
6.18150771e-01 1.11403942e+00 -4.97974038e-01 -2.87641019e-01
3.84860009e-01 -4.98695895e-02 -7.44744599e-01 -5.60656130e-01
8.70016813e-01 -4.50477809e-01 -5.31006634e-01 9.55049694e-01
5.50274491e-01 4.80677653e-03 7.70485818e-01 -1.20730734e+00
-7.06988215e-01 5.72626472e-01 -5.14240324e-01 3.47854406e-01
6.92368671e-02 8.91192257e-01 1.04605067e+00 -3.64088267e-01
2.57031590e-01 1.20687509e+00 8.98017466e-01 5.20309806e-01
-1.61454201e+00 -6.67450666e-01 2.09027797e-01 2.88641840e-01
-9.71781969e-01 -4.25509334e-01 7.91074753e-01 -7.00069308e-01
9.40820634e-01 2.08192661e-01 7.56397843e-01 1.48479986e+00
3.40335220e-01 6.42821908e-01 1.32102108e+00 -2.41522565e-01
1.86000034e-01 1.11738183e-01 3.80794019e-01 7.58755565e-01
1.64436772e-01 7.34689534e-01 -7.28110135e-01 2.52696693e-01
9.60316002e-01 -6.76275671e-01 -2.91921020e-01 -5.09407818e-01
-1.04119182e+00 5.68091810e-01 9.28153336e-01 -1.26239628e-01
-1.96502641e-01 5.65976441e-01 1.85718149e-01 3.83918166e-01
-4.02346291e-02 1.10238194e+00 -5.27200997e-01 3.51274991e-03
-7.31400549e-01 3.64168435e-01 5.50764680e-01 4.61239994e-01
4.53344375e-01 4.25816387e-01 -3.89133632e-01 2.86401570e-01
2.83280581e-01 6.15320504e-01 2.37788647e-01 -1.01519704e+00
3.70577186e-01 2.15189114e-01 3.84821117e-01 -1.13484192e+00
-8.69079351e-01 -7.74762630e-01 -7.36122489e-01 9.95752692e-01
8.02157700e-01 -9.64295119e-02 -8.70185673e-01 2.24095535e+00
-6.54294789e-02 -1.54623333e-02 -8.00342932e-02 8.89722645e-01
6.22156858e-01 3.28567445e-01 2.00356126e-01 2.05791950e-01
1.16226006e+00 -9.18860078e-01 -2.12018982e-01 -5.56786418e-01
2.86880165e-01 -4.40411597e-01 1.67131352e+00 6.41803265e-01
-1.23701894e+00 -8.46416950e-01 -1.39683723e+00 -2.77722895e-01
-3.84821832e-01 -5.29394299e-02 6.15037680e-01 6.96031809e-01
-1.27022624e+00 1.03720284e+00 -1.02613282e+00 8.70051757e-02
5.92732549e-01 3.74187589e-01 1.03321858e-01 5.49561620e-01
-1.15457773e+00 1.40521216e+00 2.05857903e-01 -1.98692515e-01
-1.03265452e+00 -1.20570636e+00 -6.37177825e-01 4.51676190e-01
1.36841536e-01 -7.90410280e-01 1.33689463e+00 -9.92027819e-01
-1.41332769e+00 5.84037244e-01 1.94590107e-01 -8.23803425e-01
7.49027729e-01 -3.78676385e-01 -1.83691934e-01 1.52575210e-01
-2.90119171e-01 7.91332066e-01 1.13949788e+00 -1.37741649e+00
-1.23367822e-02 4.81867529e-02 3.37226570e-01 -5.09742759e-02
-3.97306114e-01 -5.63768327e-01 1.55667320e-01 -6.02050304e-01
-1.01743482e-01 -1.00357640e+00 1.55813217e-01 4.19834763e-01
-5.04547358e-01 3.89994621e-01 5.12925863e-01 -7.29291558e-01
1.06042218e+00 -1.87868571e+00 1.54872417e-01 2.68456668e-01
7.65313923e-01 2.62299091e-01 -6.23411760e-02 -1.27305493e-01
-4.48805660e-01 1.34380832e-01 1.40408948e-01 -1.85454786e-01
1.70484453e-01 -4.02825981e-01 -6.33048713e-01 2.36344144e-01
3.76666963e-01 1.11130190e+00 -8.24576676e-01 -8.42628442e-03
3.57588887e-01 2.97771603e-01 -4.82339352e-01 -1.66125625e-01
-1.76342472e-01 3.64775628e-01 -5.55456467e-02 5.27560532e-01
5.27652562e-01 -6.44963443e-01 -1.66320801e-01 -4.94686872e-01
1.19638070e-01 3.14215481e-01 -6.92261279e-01 1.12349331e+00
-6.30935848e-01 1.07623899e+00 -1.79197997e-01 -5.01965821e-01
5.42489111e-01 -4.86979028e-03 1.66857347e-01 -7.53418326e-01
1.57122657e-01 -1.03294156e-01 4.01715189e-01 -3.60050321e-01
3.77964526e-01 -1.15744397e-01 2.01388057e-02 3.38287592e-01
-2.09571257e-01 -3.30797642e-01 -4.85104531e-01 3.66716444e-01
1.26183403e+00 2.62258321e-01 1.08481251e-01 -3.16667795e-01
3.84078287e-02 -1.74897075e-01 -1.21057779e-01 1.04259002e+00
-4.58944112e-01 6.00327551e-01 7.18021214e-01 -6.14434123e-01
-1.23781466e+00 -1.45887077e+00 -2.20974050e-02 8.88705373e-01
8.90715569e-02 -2.65152901e-01 -5.05597532e-01 -3.12936366e-01
1.95878983e-01 9.74403083e-01 -1.05944622e+00 -7.91005909e-01
-2.30723500e-01 -7.74495959e-01 6.97543442e-01 6.88108563e-01
5.27575731e-01 -7.65205920e-01 -1.01688802e+00 -1.19981624e-01
-6.09266348e-02 -9.22390640e-01 5.92971854e-02 2.78842717e-01
-5.92260301e-01 -1.14324355e+00 -4.29357320e-01 -3.63233010e-03
5.04131436e-01 1.01338916e-01 1.32396460e+00 2.14596450e-01
-1.00799523e-01 1.80006415e-01 2.58552253e-01 -3.73400450e-01
-4.32958275e-01 7.22735152e-02 1.57203093e-01 -4.18063074e-01
-1.42618986e-02 -6.76443458e-01 -7.01873899e-01 1.48631111e-01
-6.07163310e-01 4.81769562e-01 5.62029123e-01 6.64429963e-01
7.98518956e-02 -4.37851071e-01 1.58136457e-01 -4.29246098e-01
6.58228695e-01 -2.59509176e-01 -5.21650851e-01 2.13829204e-01
-8.76419127e-01 4.74230796e-01 7.08452761e-01 -7.10541546e-01
-8.46036434e-01 -9.98936146e-02 1.53870672e-01 -6.25730395e-01
1.94626465e-01 8.86980519e-02 2.75744110e-01 -5.97580433e-01
1.39153802e+00 -1.04092702e-01 -1.39468089e-01 -4.23510186e-03
3.28549862e-01 -3.66542898e-02 4.76577103e-01 -4.64746684e-01
9.40175235e-01 3.12181085e-01 4.19294298e-01 -4.41771299e-01
-1.14448273e+00 1.73336893e-01 -4.37759936e-01 -5.01389563e-01
7.35715449e-01 -8.35747600e-01 -1.17793536e+00 8.07909131e-01
-8.32681656e-01 -1.03775871e+00 -1.56887218e-01 1.29392564e-01
-6.81143641e-01 1.31275624e-01 -5.98163009e-01 -7.18682349e-01
-1.77576348e-01 -1.20440912e+00 8.44096065e-01 2.15758070e-01
-5.49004495e-01 -9.38222587e-01 -1.81533098e-01 3.41064721e-01
5.37893951e-01 5.10785043e-01 1.16800356e+00 -3.81735593e-01
-6.38571918e-01 1.00237653e-01 -3.06653857e-01 1.34920791e-01
-2.93020159e-01 1.35292083e-01 -1.44963944e+00 -2.25547194e-01
-8.17727223e-02 -9.94275630e-01 1.25556540e+00 5.48031569e-01
1.26875496e+00 -1.12365812e-01 -1.48974240e-01 6.96809947e-01
1.26844585e+00 -2.98155546e-01 7.68549562e-01 5.13600528e-01
8.06118309e-01 4.64608699e-01 9.90679935e-02 -5.75418444e-03
3.54309499e-01 5.73497593e-01 6.18105233e-01 -6.78301677e-02
-2.21424684e-01 -4.09224153e-01 3.09885114e-01 2.85648331e-02
-3.52009982e-02 -1.03730597e-01 -1.20933139e+00 1.63023666e-01
-1.64920557e+00 -9.73854303e-01 -2.29392245e-01 2.20616794e+00
7.57161260e-01 9.27887857e-01 1.63145572e-01 2.30034828e-01
3.08802575e-01 3.12708765e-01 -9.02096272e-01 -3.56879562e-01
-1.01614699e-01 1.68952689e-01 7.25000560e-01 8.76456439e-01
-1.15779376e+00 9.43410695e-01 6.63249397e+00 6.84913099e-01
-1.38735986e+00 -1.14681505e-01 8.98404717e-01 -5.15620410e-01
-2.27144137e-01 -1.10516965e-01 -4.75895226e-01 3.87081206e-01
1.08616221e+00 1.57144576e-01 7.37617910e-01 6.27746046e-01
3.48302752e-01 -3.09843034e-01 -1.16245890e+00 7.40668237e-01
-7.04238936e-02 -1.28858805e+00 -2.18171045e-01 -1.57666013e-01
6.89031959e-01 -7.45956227e-02 7.53110290e-01 3.81331056e-01
4.60416853e-01 -1.37578869e+00 1.20240939e+00 8.91469121e-01
6.71654820e-01 -4.51699674e-01 1.03448272e-01 2.58722156e-01
-8.07743728e-01 -1.24274358e-01 -1.34885460e-01 -4.47353542e-01
-8.84609967e-02 4.13115293e-01 -7.59791553e-01 -5.17603308e-02
7.89636016e-01 3.84963065e-01 -8.62508357e-01 7.59254634e-01
-4.73697007e-01 6.96881711e-01 -4.78217661e-01 -1.20014779e-01
1.49629429e-01 3.85643661e-01 4.46605533e-01 8.27006817e-01
-1.38473406e-01 -2.55922645e-01 -4.53216493e-01 1.24675512e+00
5.73314028e-03 -5.63222647e-01 -4.06005532e-01 2.41271749e-01
3.11327457e-01 1.01074731e+00 -4.98778969e-01 -4.20645267e-01
3.99375074e-02 7.15975523e-01 7.19221532e-01 5.17148912e-01
-1.06882250e+00 -1.31563827e-01 4.74920452e-01 1.12639233e-01
2.16366753e-01 -1.63411349e-01 -1.00401211e+00 -9.88101006e-01
-6.42539710e-02 -8.47058773e-01 -1.76592201e-01 -1.37109840e+00
-9.68526959e-01 3.04059982e-01 -4.24124673e-02 -9.77782309e-01
-2.52885707e-02 -9.97831166e-01 -8.70071888e-01 1.05805087e+00
-1.41924143e+00 -1.26319563e+00 -5.26852846e-01 2.38939732e-01
3.36353891e-02 -2.44703870e-02 5.73100328e-01 -2.49060676e-01
-2.53126562e-01 7.14948535e-01 -2.28617098e-02 -1.22795992e-01
7.40856886e-01 -1.58620811e+00 8.77588630e-01 8.18978250e-01
-5.62679060e-02 7.19722807e-01 1.30732322e+00 -4.56124246e-01
-9.10233378e-01 -7.38034964e-01 2.87239760e-01 -9.29285526e-01
9.40181196e-01 -4.02320027e-01 -9.64261174e-01 5.84359527e-01
-2.10220605e-01 -1.63197637e-01 3.34716350e-01 2.52452731e-01
-6.89398348e-01 -3.02302148e-02 -1.05956769e+00 9.38073277e-01
9.67232883e-01 -7.11969435e-01 -4.69067752e-01 2.95227230e-01
5.29133558e-01 -7.21349120e-01 -5.45424938e-01 2.45069608e-01
7.89058805e-01 -1.31775117e+00 1.30007982e+00 -8.29420984e-01
9.48937654e-01 1.26947640e-02 2.30889842e-01 -1.56815588e+00
-5.90518534e-01 -5.83569586e-01 -4.47425723e-01 5.86647809e-01
5.90126514e-01 -4.62919056e-01 7.29677260e-01 7.62949526e-01
1.52738858e-02 -8.68635297e-01 -6.95925057e-01 -4.94616210e-01
4.93193150e-01 -5.48179805e-01 1.79228604e-01 5.79030454e-01
-7.87943974e-02 3.47266734e-01 -4.41509157e-01 2.71996707e-01
5.33953011e-01 -2.20777363e-01 5.92928588e-01 -1.08956420e+00
-6.00877762e-01 -8.19831848e-01 -2.31331393e-01 -1.09051836e+00
-4.32075895e-02 -4.63108778e-01 -1.00519828e-01 -1.02894092e+00
-3.89768779e-02 -2.74729818e-01 -2.16291904e-01 3.13819975e-01
-2.76810735e-01 8.28271210e-02 5.03405511e-01 1.43510655e-01
-3.48603070e-01 4.05243099e-01 1.30583143e+00 -1.94834858e-01
-4.22628969e-02 8.43979344e-02 -8.69831920e-01 7.97943711e-01
8.66358578e-01 -7.66906366e-02 -5.07006586e-01 -4.18106705e-01
7.90003419e-01 -1.08480364e-01 1.22021461e+00 -1.35846639e+00
-2.32103281e-02 2.76586674e-02 8.22978437e-01 -2.18511760e-01
7.14313388e-01 -5.90345144e-01 -2.20713228e-01 7.29544222e-01
-7.78691232e-01 4.73776832e-03 7.70204484e-01 5.28126955e-01
4.43203241e-01 -6.65235193e-03 1.01424527e+00 -6.56607673e-02
-5.72024524e-01 -1.18286327e-01 -2.56834775e-01 1.93201631e-01
7.63058484e-01 -4.82807346e-02 -7.63303995e-01 -7.05447078e-01
-8.25691044e-01 2.28460759e-01 5.79851508e-01 2.36327961e-01
3.63962144e-01 -1.06943023e+00 -1.76138401e-01 1.76754504e-01
-1.65975466e-01 -3.84514362e-01 3.34086090e-01 8.55983436e-01
-5.15203357e-01 2.60915875e-01 -5.21771252e-01 -5.51526129e-01
-9.59408998e-01 5.86500168e-01 7.63781607e-01 -1.75766110e-01
-2.44227797e-01 1.03304875e+00 2.21617416e-01 5.54956438e-04
2.51395017e-01 -8.80114377e-01 5.33674695e-02 -1.85523301e-01
2.71109968e-01 2.19134673e-01 -3.25123109e-02 -2.34927133e-01
-5.23637012e-02 5.05395174e-01 2.50504836e-02 -1.96004897e-01
1.08001316e+00 1.05648823e-01 5.08523464e-01 6.56029344e-01
9.09333289e-01 -3.92071232e-02 -1.86753273e+00 1.56935781e-01
-4.99591172e-01 -1.45643994e-01 1.79590255e-01 -1.29185343e+00
-7.29182422e-01 1.18005884e+00 7.77383864e-01 2.97976375e-01
7.91788638e-01 -2.99818009e-01 4.07384396e-01 5.67164183e-01
1.90717764e-02 -1.06856108e+00 6.00320041e-01 5.77745140e-01
1.00953305e+00 -1.51397240e+00 -1.31796643e-01 1.08932942e-01
-7.38483071e-01 1.02159381e+00 9.94344056e-01 -2.16977030e-01
5.63387334e-01 1.77087545e-01 1.94229800e-02 -2.90706187e-01
-6.69217288e-01 6.85008839e-02 3.93201023e-01 4.76814300e-01
-6.43317029e-02 -1.07912868e-01 6.15381300e-01 5.15391827e-01
-7.07399130e-01 -1.09503075e-01 5.19126415e-01 5.62611759e-01
-5.54598331e-01 -2.21214920e-01 -3.90172988e-01 6.24499261e-01
-1.09960511e-01 -3.22238922e-01 -3.90057921e-01 8.18795085e-01
-7.08796233e-02 5.93626678e-01 6.06127903e-02 -5.93521476e-01
2.71064460e-01 1.34492666e-01 6.04448259e-01 -1.96693391e-01
-6.71703041e-01 -4.11124617e-01 9.18925479e-02 -5.05800962e-01
4.97184508e-02 -4.78942007e-01 -1.00526631e+00 -5.06669879e-01
-7.40933344e-02 -5.99754632e-01 3.65794241e-01 9.98114586e-01
1.37938140e-02 1.00596416e+00 2.40331978e-01 -1.14239967e+00
-1.03073609e+00 -9.18943346e-01 -8.13323110e-02 4.10588592e-01
9.32390630e-01 -9.65391636e-01 -6.09737337e-01 -1.45541579e-01] | [10.3177490234375, 2.178366184234619] |
7557257c-b5b9-4d08-8f54-58c3e0bf5cac | n-beats-neural-basis-expansion-analysis-for | 1905.10437 | null | https://arxiv.org/abs/1905.10437v4 | https://arxiv.org/pdf/1905.10437v4.pdf | N-BEATS: Neural basis expansion analysis for interpretable time series forecasting | We focus on solving the univariate times series point forecasting problem using deep learning. We propose a deep neural architecture based on backward and forward residual links and a very deep stack of fully-connected layers. The architecture has a number of desirable properties, being interpretable, applicable without modification to a wide array of target domains, and fast to train. We test the proposed architecture on several well-known datasets, including M3, M4 and TOURISM competition datasets containing time series from diverse domains. We demonstrate state-of-the-art performance for two configurations of N-BEATS for all the datasets, improving forecast accuracy by 11% over a statistical benchmark and by 3% over last year's winner of the M4 competition, a domain-adjusted hand-crafted hybrid between neural network and statistical time series models. The first configuration of our model does not employ any time-series-specific components and its performance on heterogeneous datasets strongly suggests that, contrarily to received wisdom, deep learning primitives such as residual blocks are by themselves sufficient to solve a wide range of forecasting problems. Finally, we demonstrate how the proposed architecture can be augmented to provide outputs that are interpretable without considerable loss in accuracy. | ['Dmitri Carpov', 'Boris N. Oreshkin', 'Nicolas Chapados', 'Yoshua Bengio'] | 2019-05-24 | null | https://openreview.net/forum?id=r1ecqn4YwB | https://openreview.net/pdf?id=r1ecqn4YwB | iclr-2020-1 | ['univariate-time-series-forecasting'] | ['time-series'] | [-4.37148102e-02 -9.93452445e-02 -8.40199217e-02 -8.43066514e-01
-4.51812059e-01 -6.30030990e-01 9.83783543e-01 -2.45999135e-02
-1.66686848e-01 4.99015123e-01 1.03241220e-01 -8.59025061e-01
-3.50351244e-01 -5.46997070e-01 -8.96407008e-01 -6.44284070e-01
-7.45404065e-01 4.91934866e-01 -6.26559332e-02 -8.62269759e-01
1.24528557e-01 7.89509416e-01 -1.65607417e+00 4.51945215e-01
6.42259598e-01 1.86342740e+00 -2.74799138e-01 4.16291445e-01
-5.67594655e-02 1.06036866e+00 -5.43087184e-01 -2.50904679e-01
5.13177693e-01 2.15557486e-01 -5.50122142e-01 -2.03150123e-01
1.81589231e-01 -2.75712579e-01 -2.11006105e-01 2.05416590e-01
1.83293551e-01 2.02356920e-01 6.27551317e-01 -1.40629458e+00
-8.36704373e-01 5.78608930e-01 -2.37696797e-01 3.88951480e-01
-1.45379409e-01 2.41979301e-01 9.99385118e-01 -7.26866424e-01
2.74283201e-01 1.13258958e+00 1.11675966e+00 1.11137949e-01
-1.26768208e+00 -4.85768020e-01 5.01609087e-01 2.58735567e-01
-8.64517808e-01 -2.67308295e-01 7.70330608e-01 -4.84130740e-01
1.35798025e+00 4.90876317e-01 2.33329266e-01 1.32510424e+00
3.25859159e-01 5.24453998e-01 1.02283311e+00 -5.76184951e-02
3.66418391e-01 -1.66745558e-01 2.23011121e-01 1.97124287e-01
-2.18449473e-01 3.06897134e-01 -6.53606728e-02 -9.39107388e-02
4.92590338e-01 2.84208804e-01 -5.82475295e-05 6.92633986e-02
-1.33701873e+00 7.35931396e-01 6.65475130e-01 4.42676336e-01
-6.71356082e-01 2.65922338e-01 7.74743080e-01 7.08889306e-01
9.44978833e-01 5.14312744e-01 -1.12183094e+00 -6.06676750e-02
-9.17007804e-01 2.37961128e-01 8.74895632e-01 5.93974829e-01
3.07136029e-01 5.64810872e-01 7.08029643e-02 5.12090325e-01
-5.52909262e-02 4.19035047e-01 8.16539168e-01 -8.24133039e-01
5.28460681e-01 5.66524267e-01 1.27461404e-01 -1.12975669e+00
-8.22200656e-01 -8.54720473e-01 -1.28013980e+00 1.70604199e-01
4.00920749e-01 -3.32791150e-01 -1.08971024e+00 1.75738859e+00
-1.09582409e-01 4.68115479e-01 5.19223809e-02 7.49251068e-01
8.69874716e-01 1.03336072e+00 -9.54313949e-03 -3.47503684e-02
1.02573526e+00 -8.83421600e-01 -3.08473825e-01 -1.80356592e-01
6.26981556e-01 -4.12988782e-01 9.37244594e-01 4.97150928e-01
-9.27975714e-01 -8.63969445e-01 -9.80366409e-01 -1.94111884e-01
-7.85419166e-01 9.90314707e-02 8.82412016e-01 6.18484765e-02
-1.26964474e+00 8.98416579e-01 -8.23092997e-01 -2.71748453e-01
1.51424244e-01 4.62668508e-01 -2.78718919e-01 4.22799915e-01
-1.30836546e+00 9.69585061e-01 3.77753824e-01 2.80763030e-01
-7.57499754e-01 -8.78431201e-01 -5.63636184e-01 3.17985296e-01
-1.90352902e-01 -6.58731639e-01 1.35968983e+00 -1.15141439e+00
-1.40179586e+00 5.89448929e-01 -1.69477895e-01 -1.04370856e+00
4.99094307e-01 -3.06492060e-01 -1.05318856e+00 -2.96442151e-01
-2.08790526e-01 4.77815419e-01 8.57663512e-01 -8.27126384e-01
-6.53518021e-01 -3.07589501e-01 7.54426047e-02 -3.76306891e-01
-4.32611376e-01 -1.92055032e-01 1.09775431e-01 -9.47432995e-01
-1.93534710e-03 -8.97902489e-01 -5.67568064e-01 -3.28183383e-01
-4.49170172e-03 -3.91690046e-01 8.76314878e-01 -7.64238000e-01
1.10509074e+00 -2.07427311e+00 -5.23567200e-03 2.72091568e-01
6.45196661e-02 4.16631140e-02 -4.00606841e-01 7.27058530e-01
-5.66893280e-01 4.82591651e-02 -2.35905826e-01 -3.98927867e-01
1.14324249e-01 3.47938865e-01 -1.14219022e+00 3.93701851e-01
3.06968749e-01 9.50408578e-01 -6.29145861e-01 3.77585113e-01
2.62325168e-01 5.06562293e-01 -2.68531561e-01 4.74638678e-02
-4.41597879e-01 5.90085745e-01 -1.71931326e-01 3.71994972e-01
5.52707613e-01 -5.85909486e-01 -2.17286181e-02 2.22197287e-02
-1.75420567e-01 4.29025024e-01 -7.46223629e-01 1.47873151e+00
-7.18642831e-01 7.42861152e-01 -4.27712440e-01 -1.37093627e+00
1.27819848e+00 3.35045487e-01 5.96708536e-01 -8.71324897e-01
-9.68908295e-02 5.53768575e-01 -1.04286358e-01 -3.12629163e-01
4.82921451e-01 -2.23595202e-02 -1.79867819e-01 3.20717365e-01
3.19515131e-02 2.18901679e-01 6.44231960e-02 -3.62070382e-01
9.22206998e-01 1.07010312e-01 7.05600306e-02 -4.04434204e-01
6.13569260e-01 -8.04013088e-02 4.40145522e-01 5.38124561e-01
2.85000384e-01 4.20691043e-01 5.82705796e-01 -1.47275531e+00
-1.20516944e+00 -6.74593627e-01 -1.72801226e-01 1.54657185e+00
-4.00026053e-01 -1.70395970e-01 -7.74453431e-02 -4.10778493e-01
1.53895065e-01 1.00982237e+00 -8.97589266e-01 6.32054508e-02
-6.91273630e-01 -6.78092778e-01 5.26137412e-01 7.45178640e-01
3.08383852e-01 -1.03144133e+00 -5.99695027e-01 3.53874207e-01
-2.46044435e-02 -1.01847875e+00 1.19386941e-01 5.93720734e-01
-1.22326136e+00 -6.32439971e-01 -7.07336426e-01 -5.18307686e-01
2.50619143e-01 4.16988023e-02 1.54955888e+00 -1.82582483e-01
3.32541972e-01 -6.59499913e-02 -2.34589383e-01 -5.73486567e-01
-2.70710528e-01 4.62718040e-01 9.37576592e-02 2.01829940e-01
3.47094268e-01 -1.08561218e+00 -6.65750623e-01 4.52906728e-01
-8.44323039e-01 -1.90221295e-01 2.86864817e-01 7.83149123e-01
2.02840790e-01 -1.10648483e-01 9.06693816e-01 -5.39274931e-01
5.24681270e-01 -9.48418021e-01 -7.82944798e-01 5.33295237e-02
-6.33629441e-01 1.84223235e-01 1.22024179e+00 -3.19364339e-01
-7.04297125e-01 -2.06069753e-01 -2.92477220e-01 -4.32261199e-01
-2.35442072e-01 7.66307712e-01 4.29089516e-01 2.17810512e-01
8.44727159e-01 3.67883235e-01 -9.48300809e-02 -7.17531562e-01
3.75391990e-01 4.27786648e-01 7.03885138e-01 -4.30219442e-01
6.18469179e-01 5.98519862e-01 3.17892551e-01 -5.35709023e-01
-7.58995771e-01 -2.05271035e-01 -5.67111135e-01 4.71549146e-02
5.37578106e-01 -1.03978109e+00 -9.62431610e-01 4.31171775e-01
-1.19648874e+00 -4.50420469e-01 -2.80508220e-01 3.24726254e-01
-5.75017154e-01 4.26583327e-02 -4.63713646e-01 -6.80150390e-01
-5.12523353e-01 -8.29758465e-01 1.08624327e+00 -2.95357615e-01
-2.16408536e-01 -1.40337050e+00 -3.50698903e-02 -1.68420538e-01
9.12864447e-01 8.39481831e-01 1.01107812e+00 -1.09902918e+00
-2.71760792e-01 -3.56757671e-01 -1.16558395e-01 4.96107042e-01
-6.86372668e-02 -3.70073952e-02 -1.19322741e+00 -3.13455373e-01
8.34226385e-02 -1.58720791e-01 1.08475244e+00 4.96667683e-01
1.67072117e+00 -4.38185453e-01 -3.37724611e-02 9.21621859e-01
1.22273338e+00 3.99748117e-01 5.83913386e-01 6.06111646e-01
5.69589019e-01 5.79693258e-01 1.71576947e-01 6.72901750e-01
5.82327604e-01 5.86319923e-01 8.39790523e-01 -2.33282834e-01
4.36858147e-01 -1.88076298e-03 3.81429493e-01 8.20733070e-01
-5.73135078e-01 -3.69537681e-01 -1.17468739e+00 5.98831952e-01
-2.02295661e+00 -1.05846643e+00 -3.39950532e-01 2.02850413e+00
3.33772153e-01 2.36481920e-01 3.08575034e-01 2.25990325e-01
2.40250990e-01 4.84396875e-01 -7.07999885e-01 -6.34729922e-01
-2.93495566e-01 3.03601086e-01 4.98518497e-01 2.66747363e-02
-1.32091880e+00 5.50584972e-01 6.89365292e+00 5.75123608e-01
-1.73618138e+00 -6.86853603e-02 9.90302980e-01 -1.70609774e-03
-4.00054961e-01 -2.84314811e-01 -4.57132727e-01 5.47980309e-01
1.66090298e+00 -1.62324548e-01 3.97413403e-01 9.93508399e-01
2.68483311e-01 7.30345368e-01 -1.42593217e+00 8.67803752e-01
-4.09397840e-01 -1.68191504e+00 -7.35921711e-02 3.28505822e-02
6.89018369e-01 5.64320326e-01 4.27278280e-01 7.81343699e-01
3.65347415e-01 -1.42935836e+00 8.25717390e-01 5.35269260e-01
5.72292268e-01 -6.47962570e-01 8.70879471e-01 4.15793926e-01
-1.12272716e+00 -5.06424069e-01 -2.21254706e-01 -5.25502920e-01
1.54963657e-01 6.00103974e-01 -5.62501907e-01 7.91491807e-01
8.73382688e-01 1.03926361e+00 -4.80010271e-01 7.09948659e-01
2.77350366e-01 8.14793050e-01 -6.17689013e-01 2.27528334e-01
7.84197748e-01 9.48129818e-02 2.80083090e-01 1.16757035e+00
7.24908888e-01 -8.56968611e-02 -9.75460187e-02 4.49611872e-01
7.32155964e-02 -1.39587462e-01 -7.03575134e-01 8.55266601e-02
1.29319936e-01 1.01038444e+00 -3.82517219e-01 -3.77470732e-01
-4.63667691e-01 6.04670763e-01 2.15321496e-01 5.23912191e-01
-9.88316536e-01 -1.89604580e-01 6.20773315e-01 1.72153432e-02
6.21920884e-01 -2.70551264e-01 -5.92599154e-01 -1.15290391e+00
9.10382196e-02 -9.39285159e-01 5.14983237e-01 -7.45098770e-01
-1.61372256e+00 1.19804168e+00 -6.61300123e-02 -1.60251939e+00
-7.92432785e-01 -8.54932845e-01 -7.06038058e-01 1.05522096e+00
-1.62865984e+00 -1.18785059e+00 -3.28107327e-01 6.11512363e-01
5.42312980e-01 -4.84177500e-01 9.33162093e-01 3.92379820e-01
-2.30518833e-01 4.08881038e-01 4.68471736e-01 -1.81811839e-01
4.25919890e-01 -1.21774721e+00 1.15189815e+00 6.36677802e-01
-1.37170115e-02 4.70864683e-01 7.24567652e-01 -8.31179246e-02
-1.25943804e+00 -1.37658787e+00 1.02926791e+00 -5.27701318e-01
8.53315771e-01 -2.81529188e-01 -1.12426424e+00 9.30979669e-01
2.17700675e-01 1.12979271e-01 4.81675148e-01 4.22977865e-01
-5.41663289e-01 -5.45894682e-01 -8.13487411e-01 3.54823619e-01
8.94477129e-01 -2.75803506e-01 -7.18346477e-01 5.79399109e-01
8.19330156e-01 -4.35227752e-01 -1.05423570e+00 5.77745795e-01
5.40354073e-01 -1.00464821e+00 1.09072804e+00 -1.01828897e+00
8.21844459e-01 1.13767693e-02 -3.01571131e-01 -1.37254798e+00
-5.97851336e-01 -6.78299606e-01 -3.35302711e-01 8.93553197e-01
5.40395677e-01 -9.50463057e-01 3.21963966e-01 5.52606583e-01
-4.27301645e-01 -8.70540380e-01 -8.75338197e-01 -1.14050937e+00
2.21293896e-01 -7.74200141e-01 1.21856940e+00 1.16377544e+00
-3.31397116e-01 2.70697236e-01 -5.99367499e-01 1.81841642e-01
1.77243695e-01 3.94889712e-01 7.73533583e-01 -1.61951637e+00
-1.73365504e-01 -6.49502456e-01 -3.48750859e-01 -1.16820455e+00
2.74829119e-01 -9.15798545e-01 -4.87785429e-01 -1.29396284e+00
-7.00556219e-01 -5.36803305e-01 -7.14918256e-01 6.58400118e-01
3.29363763e-01 9.13704485e-02 7.51265585e-02 4.37452346e-01
-1.87707186e-01 5.17961502e-01 7.21401691e-01 -9.28776264e-02
-1.93748891e-01 3.51783246e-01 -5.76598227e-01 6.62683129e-01
8.35186303e-01 -4.09873240e-02 -4.17426318e-01 -7.04835355e-01
2.62028515e-01 1.62060380e-01 5.87049901e-01 -8.74129891e-01
-2.83999201e-02 -2.15389878e-01 4.38986003e-01 -7.38611042e-01
3.03903371e-01 -1.09140730e+00 4.41905975e-01 4.52070385e-01
-4.85654652e-01 6.43150210e-01 5.75588524e-01 3.86418700e-01
-4.32673126e-01 4.37175721e-01 3.67829829e-01 -1.25789074e-02
-1.10508823e+00 3.72811049e-01 -6.33905455e-02 -2.51946360e-01
8.81746709e-01 -4.54949252e-02 -4.08926487e-01 -6.07261837e-01
-5.77925622e-01 2.58600175e-01 -8.78761634e-02 7.57620931e-01
1.29479200e-01 -1.49579453e+00 -8.51623118e-01 2.64544666e-01
1.38395548e-01 -1.99401081e-01 1.81370258e-01 7.83013821e-01
-5.72400928e-01 8.20127964e-01 -3.91123682e-01 -8.48926306e-01
-4.00277644e-01 8.08301449e-01 2.85925120e-01 -5.53131700e-01
-7.74622619e-01 6.09225452e-01 2.95810908e-01 -6.17657900e-01
2.36594215e-01 -9.95962620e-01 -1.10352777e-01 -2.33962014e-03
4.10842240e-01 4.20608759e-01 4.11714524e-01 -3.80966336e-01
-4.76519674e-01 3.59719902e-01 1.09629385e-01 1.01303995e-01
2.00595856e+00 8.11962560e-02 4.95483465e-02 7.96034575e-01
1.27039254e+00 -5.18433928e-01 -1.18589091e+00 -2.40763426e-01
1.09188221e-01 2.81314850e-02 1.26771638e-02 -9.77739930e-01
-1.20948315e+00 9.65635002e-01 4.08406645e-01 8.24576497e-01
1.56288028e+00 -3.44548672e-01 8.63568962e-01 5.54216027e-01
2.45806441e-01 -6.60056591e-01 -3.68902147e-01 7.97068715e-01
1.34060109e+00 -1.34244144e+00 -2.46071965e-01 2.93665230e-01
-6.09591663e-01 1.32815576e+00 2.66412288e-01 -4.28995013e-01
7.59314418e-01 7.37395957e-02 6.65408373e-02 -9.86670181e-02
-1.35617077e+00 6.17408119e-02 6.93422318e-01 3.54679465e-01
5.55261850e-01 4.57927138e-02 -3.70352827e-02 6.04701757e-01
-3.88851255e-01 1.82991624e-01 8.24825540e-02 4.76445436e-01
1.19846128e-02 -7.03193069e-01 -1.10437155e-01 4.34859455e-01
-4.34751093e-01 -2.02236734e-02 1.51949264e-02 1.02268529e+00
-7.82940313e-02 7.94654787e-01 2.94146717e-01 -4.62512821e-01
3.81212354e-01 8.35725665e-02 -1.81438670e-01 1.09650478e-01
-1.06282413e+00 -2.60766357e-01 6.94903508e-02 -6.23616517e-01
-4.18711334e-01 -4.36301202e-01 -7.69308925e-01 -7.29965985e-01
4.42508101e-01 -1.39021501e-01 6.58777893e-01 9.27682400e-01
7.91746795e-01 6.29433095e-01 8.33820045e-01 -1.14376235e+00
-8.12658072e-01 -9.53283310e-01 -3.70239943e-01 5.22239089e-01
7.94114292e-01 -4.58477706e-01 -3.39297056e-01 3.50479558e-02] | [6.995205402374268, 3.0527544021606445] |
e1379b9b-4dd0-4d47-b4a0-13fb0d42b697 | whc-weighted-hybrid-criterion-for-filter | 2302.08185 | null | https://arxiv.org/abs/2302.08185v1 | https://arxiv.org/pdf/2302.08185v1.pdf | WHC: Weighted Hybrid Criterion for Filter Pruning on Convolutional Neural Networks | Filter pruning has attracted increasing attention in recent years for its capacity in compressing and accelerating convolutional neural networks. Various data-independent criteria, including norm-based and relationship-based ones, were proposed to prune the most unimportant filters. However, these state-of-the-art criteria fail to fully consider the dissimilarity of filters, and thus might lead to performance degradation. In this paper, we first analyze the limitation of relationship-based criteria with examples, and then introduce a new data-independent criterion, Weighted Hybrid Criterion (WHC), to tackle the problems of both norm-based and relationship-based criteria. By taking the magnitude of each filter and the linear dependence between filters into consideration, WHC can robustly recognize the most redundant filters, which can be safely pruned without introducing severe performance degradation to networks. Extensive pruning experiments in a simple one-shot manner demonstrate the effectiveness of the proposed WHC. In particular, WHC can prune ResNet-50 on ImageNet with more than 42% of floating point operations reduced without any performance loss in top-5 accuracy. | ['Lei Huang', 'Weize Sun', 'Shaowu Chen'] | 2023-02-16 | null | null | null | null | ['neural-network-compression', 'neural-network-compression'] | ['methodology', 'miscellaneous'] | [ 2.81353388e-02 -4.75415915e-01 2.26577967e-01 -4.07222331e-01
6.61864206e-02 -1.28498688e-01 9.08672512e-02 3.98923993e-01
-9.78031874e-01 6.32435203e-01 -1.11216851e-01 -3.61292005e-01
-4.86213595e-01 -9.69635844e-01 -4.88449097e-01 -6.05861843e-01
1.35871917e-01 -2.44502559e-01 7.56486952e-01 -2.07877919e-01
2.94145942e-01 6.32350564e-01 -1.87041116e+00 3.40353161e-01
1.06957006e+00 1.12446070e+00 4.12591696e-01 1.58131108e-01
-9.13351774e-02 6.02264345e-01 -6.89378321e-01 -5.34174442e-01
4.11991239e-01 -2.65079230e-01 -4.08688128e-01 -2.84962296e-01
2.57291138e-01 -2.57449389e-01 -5.07292509e-01 1.50903237e+00
5.11358500e-01 1.06899090e-01 2.03713045e-01 -7.68319786e-01
-2.83589005e-01 6.60641670e-01 -5.71700096e-01 6.32982969e-01
-2.73742139e-01 -1.27235457e-01 7.18238771e-01 -8.49440753e-01
3.05131197e-01 1.20806134e+00 5.63511074e-01 3.82523000e-01
-9.26912189e-01 -8.75685096e-01 3.89887512e-01 5.36081433e-01
-1.51810229e+00 -3.01228553e-01 5.67599297e-01 -2.07250550e-01
8.86411548e-01 4.08686012e-01 7.76705503e-01 3.24956328e-01
5.39393052e-02 6.41373158e-01 8.46646070e-01 -4.16453958e-01
2.61417717e-01 -9.15676579e-02 4.63199437e-01 7.99952745e-01
6.99871600e-01 -7.35798897e-03 -4.15472955e-01 1.05237350e-01
6.55454397e-01 1.23538584e-01 -4.75836456e-01 -2.54032146e-02
-8.75405490e-01 5.86341202e-01 6.72908425e-01 2.67673224e-01
-2.28007555e-01 -1.31725222e-01 5.18270731e-01 1.84326634e-01
3.11106056e-01 3.54340523e-01 -2.50901729e-01 -3.08879302e-04
-1.07922816e+00 1.44759536e-01 4.39893633e-01 8.40049028e-01
6.34633183e-01 2.37814859e-01 -2.15155631e-01 8.84710193e-01
7.74957985e-02 1.67417377e-01 6.05274141e-01 -4.59324837e-01
5.00313342e-01 8.35666001e-01 -3.06331426e-01 -1.37770391e+00
-4.29821402e-01 -8.03817332e-01 -1.07332194e+00 2.57568330e-01
2.85662562e-01 -8.81251246e-02 -9.17444289e-01 1.53039432e+00
1.02226906e-01 2.13276833e-01 -1.23368561e-01 9.39468205e-01
1.04518700e+00 3.83369178e-01 -3.22517864e-02 -3.18662941e-01
1.37589502e+00 -8.45785260e-01 -5.52058399e-01 -1.71503216e-01
3.65750343e-01 -8.41429293e-01 8.05693507e-01 3.18283826e-01
-1.19115579e+00 -6.54345989e-01 -1.50280273e+00 1.64305478e-01
-2.84716219e-01 3.38304609e-01 5.12247622e-01 6.23368680e-01
-8.21442723e-01 7.52711177e-01 -7.88837254e-01 3.12596909e-03
5.62100053e-01 4.06786054e-01 -7.10189939e-02 -1.45537093e-01
-1.18539417e+00 9.10252333e-01 6.85651958e-01 2.43016988e-01
-4.09081668e-01 -7.60864675e-01 -4.44195449e-01 4.47241813e-01
4.81227696e-01 -2.99552381e-01 1.06924272e+00 -6.01658821e-01
-1.16117501e+00 3.24134052e-01 -2.03236967e-01 -7.27054894e-01
6.13120198e-01 -2.48517975e-01 -7.86877334e-01 2.86551088e-01
-2.95827597e-01 5.99123895e-01 6.50274456e-01 -7.98391163e-01
-9.42752302e-01 -1.69507727e-01 1.91150174e-01 1.18044019e-01
-7.90097356e-01 4.85552922e-02 -6.26490414e-01 -7.47320294e-01
4.75291193e-01 -5.38805366e-01 -2.42499724e-01 1.13246523e-01
-2.84492642e-01 -1.65681422e-01 6.67852044e-01 -2.75015593e-01
1.59757960e+00 -2.28268695e+00 -1.38233185e-01 1.06459901e-01
4.38611388e-01 1.02865875e+00 -9.61870030e-02 -1.24080211e-01
-1.01500629e-02 1.63271502e-01 -1.54269502e-01 -3.06061842e-02
-4.39252406e-01 1.03375539e-01 -1.66077852e-01 3.42569709e-01
3.76441330e-01 4.62917238e-01 -8.22990179e-01 -4.26844716e-01
2.91390479e-01 5.64935863e-01 -6.08635485e-01 -2.31423005e-01
1.34717152e-01 -2.50353128e-01 -2.58355618e-01 3.77567559e-01
9.36109602e-01 1.16329372e-01 1.09513946e-01 -5.97349823e-01
-2.67838597e-01 1.00846313e-01 -1.21202219e+00 1.06683242e+00
-2.05509409e-01 5.41037917e-01 -3.28508243e-02 -9.69957948e-01
1.09143591e+00 -1.59158278e-02 1.47614762e-01 -7.93190598e-01
3.62344593e-01 4.31383401e-01 5.35537183e-01 -1.97689056e-01
4.84858423e-01 7.59357288e-02 3.89037460e-01 3.01374774e-02
-5.13215512e-02 3.88590932e-01 6.16176307e-01 2.04656143e-02
1.13912666e+00 -4.37519103e-01 2.31910124e-01 -5.83831251e-01
8.51253033e-01 -2.70193458e-01 9.77874875e-01 5.42468786e-01
-2.05349505e-01 6.37269735e-01 4.83763993e-01 -6.41577959e-01
-6.56709015e-01 -5.90692401e-01 -1.89939693e-01 6.88584864e-01
4.22502130e-01 -6.38971210e-01 -6.38842642e-01 -4.74005789e-01
1.18505456e-01 5.42961955e-01 -3.37213427e-01 -4.85166222e-01
-6.94489121e-01 -9.26099300e-01 6.24321640e-01 4.88144755e-01
9.38546419e-01 -7.44747221e-01 -1.12869895e+00 3.41469109e-01
6.91750348e-02 -1.01238012e+00 -1.92538902e-01 2.77003795e-01
-1.00726259e+00 -1.17979395e+00 -7.46393561e-01 -7.28364825e-01
9.49967444e-01 7.34842896e-01 7.38230526e-01 3.72345001e-01
-3.31973702e-01 -4.24356997e-01 -3.57670546e-01 -3.82451296e-01
-2.09815223e-02 7.68382400e-02 -3.22145829e-03 1.36589527e-01
4.33872312e-01 -5.34547389e-01 -7.39898145e-01 5.01551032e-01
-9.29407656e-01 6.98534250e-02 7.74348080e-01 6.78792655e-01
5.75008154e-01 5.12989581e-01 5.37875593e-01 -8.42159867e-01
7.69907057e-01 -1.18657425e-01 -6.59201324e-01 3.03181469e-01
-6.21364057e-01 1.44201770e-01 9.71023917e-01 -6.13117278e-01
-8.99748921e-01 -5.98580167e-02 -1.38056025e-01 -6.05226636e-01
2.09795490e-01 5.13150811e-01 -2.13927522e-01 -2.37861633e-01
5.96978068e-01 2.73487508e-01 -1.77900136e-01 -6.54497504e-01
7.87950400e-03 2.93090582e-01 5.95140755e-01 -2.83114046e-01
5.74474275e-01 3.73438954e-01 1.30255148e-01 -7.77972043e-01
-6.65807903e-01 -3.49306732e-01 -3.59314144e-01 -1.55115411e-01
4.84313071e-01 -8.74210775e-01 -6.74164176e-01 4.49805975e-01
-1.11112916e+00 2.49696463e-01 -1.00412957e-01 6.07190728e-01
2.15847731e-01 2.92065263e-01 -6.01175904e-01 -6.57581568e-01
-3.92880768e-01 -1.42716265e+00 1.92839339e-01 5.48891008e-01
1.54661238e-01 -3.96963358e-01 -7.19791114e-01 -1.68149561e-01
6.09379113e-01 4.04994674e-02 1.07919312e+00 -5.21441758e-01
-4.75057721e-01 -2.87656307e-01 -6.86216533e-01 6.18660748e-01
-1.48840044e-02 -1.60119240e-03 -5.96854866e-01 -3.16939831e-01
7.85853416e-02 -8.01893137e-03 1.40176952e+00 1.91639781e-01
1.36555409e+00 -1.21594332e-01 -2.01580822e-01 7.63508201e-01
1.57060003e+00 3.95225763e-01 7.01159000e-01 3.85700971e-01
3.92358124e-01 1.84266195e-01 5.43086052e-01 4.55911517e-01
-1.87773108e-01 3.10346276e-01 5.30233502e-01 -1.77916765e-01
-2.25980744e-01 2.57265456e-02 -1.48697451e-01 1.01304591e+00
-2.53736109e-01 -2.26159781e-01 -6.61768615e-01 2.76578188e-01
-1.80839396e+00 -7.43786454e-01 -2.27248058e-01 2.11924434e+00
7.10851133e-01 6.56962812e-01 -2.66924709e-01 4.75850135e-01
1.02533042e+00 1.81034818e-01 -6.73144698e-01 -2.54356533e-01
-2.57829666e-01 3.11708838e-01 6.21571362e-01 1.58167869e-01
-9.98688757e-01 6.19492352e-01 5.40061617e+00 1.08492839e+00
-1.18190980e+00 -5.78412227e-03 6.28865838e-01 -3.78420830e-01
1.85856503e-02 -2.65311338e-02 -1.16711307e+00 5.72973549e-01
5.47038794e-01 -3.19222569e-01 2.58583844e-01 8.98030758e-01
-2.89618084e-03 -1.73006684e-01 -7.56345451e-01 1.05590105e+00
3.14830140e-05 -1.36210001e+00 2.22348273e-01 -1.78783536e-01
6.11567259e-01 -1.78323478e-01 -1.10685289e-01 2.19058380e-01
-1.25971109e-01 -6.92537785e-01 9.49814379e-01 3.99111748e-01
6.22324705e-01 -1.14710271e+00 9.32881594e-01 2.69428074e-01
-1.25530136e+00 -2.73881555e-01 -9.90377307e-01 -1.96437001e-01
-8.09872597e-02 1.07677889e+00 -2.18838125e-01 5.18059909e-01
8.79402101e-01 5.72543025e-01 -6.56306684e-01 1.53972149e+00
-2.28539214e-01 4.96537507e-01 -4.21874255e-01 -3.58617753e-01
2.83243746e-01 -2.68827640e-02 4.31717426e-01 1.15980816e+00
4.08242136e-01 3.66753876e-01 -8.98539275e-02 6.28175378e-01
-9.61349905e-02 1.00492053e-01 9.99217629e-02 2.26676285e-01
7.24899888e-01 1.23343897e+00 -1.05579996e+00 -3.56688440e-01
-3.86232913e-01 5.66730738e-01 4.39322650e-01 2.00026467e-01
-8.77467692e-01 -7.74796426e-01 6.21986270e-01 1.33295864e-01
4.94641900e-01 -1.32450119e-01 -3.34363073e-01 -9.60399866e-01
2.36236259e-01 -7.84189105e-01 1.99002013e-01 -1.80345982e-01
-7.57938266e-01 8.67857099e-01 -6.76933005e-02 -1.30049098e+00
5.02131104e-01 -5.43493927e-01 -6.38005912e-01 7.70404816e-01
-1.59889936e+00 -2.39297971e-01 -5.04791915e-01 3.49513382e-01
5.68198144e-01 -2.86851734e-01 4.37260389e-01 6.31767809e-01
-9.81316328e-01 7.88158476e-01 9.69762579e-02 -6.44688979e-02
4.57685202e-01 -6.84225559e-01 2.27205828e-01 1.23698473e+00
-1.49376139e-01 9.12305474e-01 5.67424834e-01 -4.72102076e-01
-1.03158164e+00 -1.18787491e+00 6.96041882e-01 6.23553693e-01
2.43623570e-01 -1.51566252e-01 -1.12899387e+00 -7.51345158e-02
-1.69915006e-01 2.09517881e-01 3.17067146e-01 -7.58514777e-02
-4.14659381e-01 -5.75026989e-01 -1.09160948e+00 7.14033008e-01
1.19535649e+00 7.13233948e-02 -5.45221508e-01 1.52432770e-01
7.03229427e-01 -3.79193217e-01 -5.03196657e-01 7.87253499e-01
4.74043429e-01 -1.36404419e+00 1.02944887e+00 -2.29810223e-01
3.75976980e-01 -4.54441816e-01 -3.69522162e-02 -9.48729515e-01
-5.68729579e-01 -2.50858814e-01 -2.14550227e-01 1.14276183e+00
3.23692232e-01 -7.65031338e-01 6.27870798e-01 2.24317715e-01
-2.39343807e-01 -1.10745513e+00 -9.52879608e-01 -1.02206457e+00
-2.49109000e-01 -3.05927843e-01 8.52744997e-01 4.94053632e-01
-2.31855080e-01 9.31161717e-02 -1.45465642e-01 7.34843984e-02
2.41098344e-01 -1.23201795e-01 2.10003212e-01 -1.31085479e+00
-1.44349828e-01 -9.08937037e-01 -6.31059289e-01 -1.11128187e+00
-3.33514541e-01 -6.79608941e-01 1.10305190e-01 -1.35122955e+00
-9.26980004e-03 -4.78238255e-01 -6.39545441e-01 4.46220994e-01
-4.00754452e-01 2.85067409e-01 3.45656931e-01 3.03898841e-01
-3.72137159e-01 5.21886230e-01 1.06002867e+00 -4.97558154e-02
-1.00040391e-01 -8.92832875e-03 -7.23988652e-01 9.31425631e-01
8.68391514e-01 -4.62813348e-01 -6.42555356e-01 -7.46609271e-01
6.32290393e-02 -4.04177487e-01 1.40462965e-01 -1.61650538e+00
5.82618356e-01 8.32019001e-02 4.91413325e-01 -7.25589752e-01
1.51148498e-01 -6.88036919e-01 -4.77765314e-02 7.75798917e-01
-1.93050131e-01 1.90129325e-01 4.49572176e-01 5.88328123e-01
-3.99895549e-01 -5.43380439e-01 1.03925037e+00 -8.76221955e-02
-9.04587686e-01 1.39760822e-01 -2.23867923e-01 -7.78505951e-02
9.75479603e-01 -2.67180353e-01 -4.41543519e-01 1.65501699e-01
-3.31995338e-01 1.62082538e-01 3.73861901e-02 2.73272514e-01
8.59012842e-01 -1.08393729e+00 -6.75245464e-01 4.91927803e-01
-1.67797789e-01 8.78198594e-02 3.33706915e-01 6.79533839e-01
-8.32377195e-01 4.00262266e-01 -2.80303240e-01 -3.85391474e-01
-1.49212992e+00 5.87311089e-01 2.00933859e-01 -2.10003793e-01
-6.49232328e-01 1.20401430e+00 1.34365444e-04 2.06961349e-01
5.49370944e-01 -5.07333755e-01 -2.52650082e-01 7.29404837e-02
7.78398335e-01 4.96695489e-01 4.54546630e-01 -2.48551130e-01
-4.46127981e-01 4.31084841e-01 -3.32309097e-01 2.85368115e-01
1.17968190e+00 2.43940786e-01 -6.20223656e-02 -2.17657406e-02
1.00705576e+00 -3.13104957e-01 -1.06187940e+00 -4.14526761e-01
-3.48923467e-02 -5.24069548e-01 3.40210706e-01 -6.32265210e-01
-1.51249528e+00 9.91717756e-01 5.65290511e-01 1.31813645e-01
1.62462759e+00 -6.88192844e-01 9.51262951e-01 4.10161346e-01
2.81518221e-01 -1.06744361e+00 -2.29970247e-01 5.92378497e-01
6.46518290e-01 -7.88300037e-01 4.05708224e-01 -7.14211106e-01
-2.03263909e-01 1.31733680e+00 9.11558807e-01 -1.28534973e-01
7.46492982e-01 2.24076003e-01 -2.65060157e-01 1.06286228e-01
-6.28893614e-01 -2.88382143e-01 2.94340879e-01 1.86394170e-01
2.20416635e-01 -1.22088790e-01 -8.71179402e-01 5.85619211e-01
-1.89190030e-01 -1.16198793e-01 1.78462029e-01 1.07262230e+00
-7.30077267e-01 -8.59599233e-01 -8.39510486e-02 8.19792211e-01
-3.59528959e-01 -2.65227169e-01 3.27858441e-02 7.29075253e-01
5.06595790e-01 9.13236618e-01 1.51848838e-01 -7.47498214e-01
5.85361183e-01 -2.14197010e-01 2.57139593e-01 -3.26048881e-01
-6.12419963e-01 3.29458565e-02 -1.38805583e-01 -3.06252778e-01
-2.14472130e-01 -2.65403688e-01 -1.27346814e+00 -5.27379215e-01
-7.49848008e-01 8.87022465e-02 4.77513254e-01 7.99801826e-01
4.83516842e-01 8.11498642e-01 3.35653812e-01 -4.68925923e-01
-6.55693948e-01 -8.25570881e-01 -3.60170841e-01 2.16444403e-01
-2.41659246e-02 -7.26422846e-01 -3.01872909e-01 -4.14799094e-01] | [8.556328773498535, 3.055878162384033] |
e6982e0e-47ad-4bc4-9089-485859ca5ef6 | 3d-zef-a-3d-zebrafish-tracking-benchmark-1 | 2006.08466 | null | https://arxiv.org/abs/2006.08466v1 | https://arxiv.org/pdf/2006.08466v1.pdf | 3D-ZeF: A 3D Zebrafish Tracking Benchmark Dataset | In this work we present a novel publicly available stereo based 3D RGB dataset for multi-object zebrafish tracking, called 3D-ZeF. Zebrafish is an increasingly popular model organism used for studying neurological disorders, drug addiction, and more. Behavioral analysis is often a critical part of such research. However, visual similarity, occlusion, and erratic movement of the zebrafish makes robust 3D tracking a challenging and unsolved problem. The proposed dataset consists of eight sequences with a duration between 15-120 seconds and 1-10 free moving zebrafish. The videos have been annotated with a total of 86,400 points and bounding boxes. Furthermore, we present a complexity score and a novel open-source modular baseline system for 3D tracking of zebrafish. The performance of the system is measured with respect to two detectors: a naive approach and a Faster R-CNN based fish head detector. The system reaches a MOTA of up to 77.6%. Links to the code and dataset is available at the project page https://vap.aau.dk/3d-zef | ['Stefan Hein Bengtson', 'Thomas B. Moeslund', 'Malte Pedersen', 'Joakim Bruslund Haurum'] | 2020-06-15 | 3d-zef-a-3d-zebrafish-tracking-benchmark | http://openaccess.thecvf.com/content_CVPR_2020/html/Pedersen_3D-ZeF_A_3D_Zebrafish_Tracking_Benchmark_Dataset_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Pedersen_3D-ZeF_A_3D_Zebrafish_Tracking_Benchmark_Dataset_CVPR_2020_paper.pdf | cvpr-2020-6 | ['3d-object-detection-from-stereo-images', '3d-multi-object-tracking'] | ['computer-vision', 'computer-vision'] | [-4.33278948e-01 -3.11843812e-01 6.03097200e-01 -9.64131579e-02
-4.32256997e-01 -6.80154622e-01 1.14493541e-01 1.00822791e-01
-1.10301840e+00 3.16739738e-01 -2.90005654e-01 2.39829421e-01
2.44701505e-01 -2.09500968e-01 -6.99500501e-01 -6.86349452e-01
-1.96708918e-01 1.57353863e-01 7.90316641e-01 -1.30988076e-01
3.13473761e-01 7.69836724e-01 -1.48551297e+00 -1.91798478e-01
1.81237683e-01 8.34221840e-01 2.77977109e-01 9.38837171e-01
3.42146724e-01 3.02294791e-01 -5.21425128e-01 -3.54153901e-01
4.75726515e-01 -2.47443780e-01 -4.96908069e-01 -3.62796217e-01
5.83814621e-01 -5.58669508e-01 -2.31364518e-01 9.82098103e-01
1.00438511e+00 -1.50510948e-02 3.01727891e-01 -1.18489850e+00
-1.06081836e-01 2.36177042e-01 -7.21927822e-01 6.69464052e-01
6.23679578e-01 6.51614249e-01 4.10711527e-01 -5.44896662e-01
6.14564598e-01 1.51282096e+00 8.50044847e-01 1.11323035e+00
-9.92708445e-01 -9.99816656e-01 -1.67335376e-01 -1.64385960e-01
-1.36085236e+00 -4.53527659e-01 6.49477839e-02 -9.09066379e-01
7.03707218e-01 -2.23827869e-01 9.71703351e-01 9.04450297e-01
3.30788821e-01 2.28583843e-01 9.75967467e-01 1.26510739e-01
2.43102655e-01 -3.35323274e-01 -2.57973783e-02 8.51181149e-01
3.31703514e-01 -4.67673428e-02 -5.27860641e-01 -1.40064180e-01
1.11632323e+00 1.53824091e-01 -1.32206514e-01 -2.76105553e-01
-1.08575368e+00 4.91084784e-01 4.05263573e-01 -1.17479071e-01
2.07985312e-01 4.78118569e-01 2.91755736e-01 7.34499395e-02
1.25359163e-01 1.20223142e-01 -5.16118288e-01 -1.77558154e-01
-3.92349124e-01 5.88402569e-01 3.49427998e-01 1.03990328e+00
4.58953112e-01 -2.30041351e-02 3.51523049e-02 3.62614870e-01
7.87071586e-01 8.38866532e-01 5.60692728e-01 -1.09914231e+00
8.97032395e-02 8.60483527e-01 1.00275606e-01 -5.14108121e-01
-9.92597818e-01 1.87596008e-01 -3.03179085e-01 6.78203225e-01
8.65279257e-01 -3.40177238e-01 -7.07166433e-01 1.49457788e+00
7.95452952e-01 3.36353451e-01 -2.84008622e-01 1.00335836e+00
1.30538285e+00 1.94615752e-01 9.71574858e-02 3.58426660e-01
1.41616869e+00 -5.31035721e-01 -1.97543815e-01 1.99262232e-01
6.20660782e-01 -5.63626111e-01 7.23881841e-01 -5.46923093e-03
-9.85756099e-01 -8.54883939e-02 -6.80889964e-01 -1.42615244e-01
-4.53018874e-01 4.10673395e-02 3.41585517e-01 5.80268323e-01
-1.20268977e+00 5.97830236e-01 -1.21094835e+00 -7.74416149e-01
7.41234899e-01 7.07383156e-01 -8.54331374e-01 4.07296270e-01
-3.06474805e-01 7.31778800e-01 4.78289388e-02 -2.47290209e-02
-1.24119878e+00 -8.00972998e-01 -8.35914314e-01 -4.06991363e-01
-1.12189882e-01 -4.39096749e-01 1.48000526e+00 1.17467843e-01
-1.58519506e+00 1.16295779e+00 2.15824410e-01 -3.70402515e-01
8.26508760e-01 -5.04807830e-01 2.13366225e-01 2.85368562e-01
2.75766850e-01 9.28557932e-01 2.89010435e-01 -4.47952509e-01
-7.13718057e-01 -8.43366206e-01 3.11372161e-01 9.97045189e-02
-5.09978738e-03 4.26982909e-01 -4.83261108e-01 -3.63594443e-01
-1.62992850e-01 -1.23005414e+00 -3.88172656e-01 8.29664946e-01
-9.16290283e-02 -1.23106472e-01 7.25814164e-01 -2.82374412e-01
5.89928627e-01 -2.10363293e+00 2.97836103e-02 -5.16155362e-01
2.96489656e-01 3.18519354e-01 -2.51885086e-01 2.76179731e-01
5.64511865e-02 3.85944173e-02 1.74083486e-02 -2.93398142e-01
-1.20283656e-01 -2.50266552e-01 1.82236135e-01 1.05014896e+00
3.84106785e-02 5.36118686e-01 -1.01055193e+00 -3.08094621e-01
2.71979392e-01 5.15932858e-01 -7.55213320e-01 3.11693549e-01
1.39175758e-01 9.78292584e-01 -3.95270020e-01 1.09059584e+00
6.32868707e-01 -9.50880423e-02 -3.81368756e-01 -1.29718408e-02
-6.24058902e-01 -3.44598502e-01 -1.01241064e+00 1.97569883e+00
1.97765872e-01 5.34102499e-01 -5.85794672e-02 -2.59333283e-01
5.04418015e-01 2.69244313e-01 5.93996167e-01 -3.92367303e-01
4.90257233e-01 1.57331809e-01 -8.71522948e-02 -7.81553984e-01
1.49044573e-01 2.51738906e-01 3.91850844e-02 4.26054075e-02
4.12371069e-01 1.24600492e-01 5.50253868e-01 -6.97816312e-02
1.32679927e+00 6.62865281e-01 2.39056006e-01 -3.91549051e-01
2.88341701e-01 -1.03445552e-01 8.16129148e-01 4.50756609e-01
-7.16615021e-01 5.11157811e-01 5.04967511e-01 -7.68246472e-01
-8.90714049e-01 -6.50491238e-01 -3.83008510e-01 8.58292401e-01
4.87502843e-01 -1.99774161e-01 -8.21178913e-01 -4.50547338e-01
1.92267135e-01 -3.95573169e-01 -8.44727099e-01 1.69206917e-01
-4.92272675e-01 -6.98891580e-01 1.07265651e+00 4.13756728e-01
5.26169121e-01 -7.73364246e-01 -1.47160363e+00 -4.95090000e-02
2.70914823e-01 -1.07487082e+00 -1.39687285e-01 7.02993646e-02
-8.32072139e-01 -1.43809283e+00 -1.06716108e+00 -5.36361575e-01
4.16409492e-01 3.70594054e-01 5.10520875e-01 1.50751576e-01
-6.84975684e-01 7.12657794e-02 -2.90586144e-01 -5.44126868e-01
3.73536684e-02 -4.79737997e-01 3.52572024e-01 -6.70541644e-01
3.44693363e-01 -3.60642225e-01 -8.80472064e-01 4.11298722e-01
-8.86348128e-01 -2.76266187e-01 1.92069098e-01 4.10559028e-01
5.92460990e-01 -7.14082897e-01 4.52581514e-03 -2.76667625e-01
-1.87167197e-01 -3.35737467e-01 -1.42962456e+00 -1.62012890e-01
3.61899704e-01 -2.75053561e-01 2.70755231e-01 -4.77381885e-01
-3.20671707e-01 7.34543085e-01 -2.88165599e-01 -5.29950380e-01
-6.34488583e-01 -2.02063367e-01 7.51377940e-02 -6.70999587e-01
6.12333953e-01 -3.25145900e-01 5.75085543e-02 -6.70205593e-01
-6.33787364e-02 3.70027512e-01 2.57821023e-01 4.90417443e-02
8.11599255e-01 8.73331964e-01 3.33473086e-01 -8.90232563e-01
-6.38236940e-01 -6.56217694e-01 -7.00795770e-01 -5.30696988e-01
1.05440271e+00 -1.30658793e+00 -1.33529985e+00 9.35326993e-01
-1.07587755e+00 -6.81728959e-01 1.58852801e-01 5.99183619e-01
-2.56245673e-01 4.70591813e-01 -8.24500442e-01 -7.07365155e-01
-3.07621866e-01 -1.48711419e+00 1.53812087e+00 6.20033860e-01
7.54026845e-02 -6.47522271e-01 5.08727133e-01 1.33887246e-01
2.34769762e-01 7.24607587e-01 5.55561893e-02 -6.02804363e-01
-5.34371734e-01 -1.81633845e-01 -9.67079550e-02 -4.59536701e-01
-1.21816210e-01 2.96427548e-01 -1.11710727e+00 -5.04538417e-01
-4.80199307e-01 -4.46375579e-01 6.85919344e-01 4.16796297e-01
4.89166617e-01 2.01083213e-01 -2.68832922e-01 6.79945409e-01
1.49694681e+00 3.56531233e-01 1.56519100e-01 5.80082715e-01
3.44191730e-01 4.96165246e-01 5.36894679e-01 7.79215634e-01
1.97228774e-01 9.07523215e-01 8.72670889e-01 1.45921126e-01
3.53061147e-02 9.83285531e-02 2.90094584e-01 2.08353475e-01
-3.72733533e-01 -4.62018438e-02 -1.07153082e+00 4.51607823e-01
-1.65381896e+00 -9.01050448e-01 -4.36899334e-01 2.25354075e+00
2.92646497e-01 -4.65927720e-01 7.53514349e-01 -9.28796455e-02
7.97541082e-01 -6.43769979e-01 -6.58159316e-01 3.39686424e-01
-6.49723336e-02 6.45810664e-02 8.30823421e-01 1.39691144e-01
-1.43905091e+00 9.18890715e-01 5.89592266e+00 1.06802642e-01
-1.21838725e+00 -1.70293555e-01 6.21435186e-03 -2.66046613e-01
6.65132642e-01 -3.12949538e-01 -1.29451013e+00 4.93608177e-01
6.60277128e-01 -9.28880938e-04 8.03949386e-02 7.19595432e-01
2.93391734e-01 -2.96545029e-01 -9.52129960e-01 1.12748122e+00
-5.20105921e-02 -1.02459145e+00 -1.74138159e-01 3.09059829e-01
2.84169137e-01 4.30794448e-01 9.92628932e-03 -1.53076455e-01
1.57807291e-01 -6.97644532e-01 1.04476619e+00 3.75672698e-01
8.47638845e-01 -4.10432339e-01 9.69177008e-01 2.40170732e-01
-1.37244916e+00 -8.63969103e-02 -6.90906823e-01 -8.56536627e-02
3.05534303e-02 -3.31846178e-02 -5.37869453e-01 1.10544287e-01
1.49260461e+00 9.46974158e-01 -9.33776617e-01 2.07193136e+00
1.53481930e-01 -2.36861929e-02 -5.40639639e-01 -3.23565900e-01
8.55846927e-02 -1.43470332e-01 7.09043324e-01 1.20121121e+00
6.08260691e-01 1.87397718e-01 -3.19576025e-01 4.23161030e-01
-6.61265552e-02 2.00529769e-02 -7.66699791e-01 9.23879072e-02
2.79756308e-01 1.24405038e+00 -1.09980178e+00 7.33442605e-02
-2.76929229e-01 7.83607721e-01 2.49469057e-01 -1.41995832e-01
-8.89320552e-01 -4.47667629e-01 1.07620168e+00 -1.72030240e-01
4.42198008e-01 -1.27812684e-01 3.07541043e-01 -1.12031293e+00
-3.07547361e-01 -2.62580276e-01 3.31172436e-01 -5.91449797e-01
-9.87418950e-01 6.47003472e-01 -1.35701746e-02 -1.78563368e+00
3.21109861e-01 -8.80880654e-01 -2.70449430e-01 4.23069715e-01
-1.52099264e+00 -9.87899363e-01 -6.10012829e-01 5.44150531e-01
3.85154605e-01 -1.33256882e-01 8.51230979e-01 6.47514880e-01
-6.96252823e-01 3.56209338e-01 3.01973280e-02 2.72331208e-01
6.27756059e-01 -1.18017232e+00 2.84968108e-01 7.73665905e-01
-1.89920276e-01 5.37258625e-01 6.75530851e-01 -5.42886317e-01
-1.56351137e+00 -1.24994040e+00 3.30185771e-01 -6.62335813e-01
7.39274025e-01 -3.42557520e-01 -5.64239144e-01 5.73396802e-01
4.50201035e-02 4.05283779e-01 8.75873029e-01 -8.08780909e-01
-6.97628111e-02 3.69992554e-02 -1.31394422e+00 5.22479117e-01
9.87959325e-01 -6.05716044e-03 -1.42925814e-01 3.02051157e-01
4.62150097e-01 -9.68377650e-01 -8.30327690e-01 2.96066970e-01
9.76624310e-01 -1.03020298e+00 8.36505234e-01 -4.74381864e-01
2.98021555e-01 -9.87229049e-01 1.26023769e-01 -8.56096804e-01
-8.40945840e-02 -5.03479838e-01 4.39587981e-01 8.59385967e-01
7.77982101e-02 -3.40907544e-01 8.03024828e-01 3.52819443e-01
-2.25923210e-01 -3.28771293e-01 -1.26578689e+00 -9.16260004e-01
-2.52697542e-02 -7.34342113e-02 3.06455702e-01 5.04287183e-01
9.64032710e-02 6.72981888e-02 -2.59419233e-01 5.61459482e-01
8.44578683e-01 1.05954558e-02 1.04494929e+00 -1.37824416e+00
1.95902154e-01 -3.85894239e-01 -1.30963993e+00 -9.08615172e-01
-2.73828536e-01 -5.81114531e-01 6.68563601e-03 -1.29028738e+00
1.09488048e-01 -3.75964008e-02 7.43463412e-02 6.14249706e-01
2.49842629e-01 7.70570457e-01 1.92829985e-02 -1.46295995e-01
-9.57811356e-01 2.86711901e-01 9.63327408e-01 1.08516589e-01
-6.03289492e-02 6.09333999e-03 3.93814454e-03 8.96969497e-01
7.87907302e-01 -5.80323160e-01 3.01449478e-01 -7.10829616e-01
-6.46037683e-02 -2.51167528e-02 6.58787251e-01 -1.35487163e+00
4.42377448e-01 1.87736899e-01 2.97209442e-01 -4.29853141e-01
5.58411598e-01 -9.08499718e-01 5.99989116e-01 8.62562895e-01
-1.38988525e-01 2.60810167e-01 5.95389247e-01 6.12669647e-01
1.88661903e-01 -1.60008535e-01 1.20755649e+00 -3.91299516e-01
-6.65921211e-01 4.08605635e-01 -5.22374868e-01 1.00985870e-01
1.32891202e+00 -3.06267470e-01 -4.74768162e-01 3.39762121e-01
-3.92921209e-01 2.32977003e-01 8.66927743e-01 1.48134723e-01
7.48018742e-01 -9.42947447e-01 -4.83099520e-01 -9.91214067e-02
2.76764423e-01 1.65448800e-01 3.18016797e-01 8.87057364e-01
-1.47023213e+00 2.87521094e-01 -6.64627135e-01 -8.37601960e-01
-1.83731902e+00 2.29032338e-02 4.11669165e-01 4.84736502e-01
-6.61343098e-01 1.15006828e+00 1.26837835e-01 -3.15743089e-01
4.43269432e-01 -6.46376431e-01 -5.16738117e-01 -2.12331548e-01
8.38276327e-01 6.82813227e-01 -1.28167912e-01 -9.13435936e-01
-5.99756718e-01 1.08641756e+00 4.25090343e-01 2.72633463e-01
1.69383287e+00 -6.37839884e-02 -1.02403443e-02 3.29692036e-01
8.79679084e-01 -2.87395120e-01 -1.55500484e+00 2.60058969e-01
1.36490073e-02 -6.07801437e-01 -4.14930820e-01 -2.27356806e-01
-1.05603015e+00 1.01268518e+00 1.15406454e+00 9.32345632e-03
6.12147212e-01 1.37202322e-01 5.13250291e-01 3.92538697e-01
5.00238299e-01 -4.82137889e-01 -5.50045166e-03 6.44589365e-01
6.41805410e-01 -1.35289383e+00 -2.15286687e-02 -1.28951058e-01
-2.29358554e-01 1.03956473e+00 8.56029630e-01 -1.67203933e-01
3.52955520e-01 4.34734613e-01 4.29785669e-01 -2.46157467e-01
-5.12672782e-01 -3.92757893e-01 -2.51066595e-01 7.46455789e-01
3.98642480e-01 -5.10567248e-01 1.86185703e-01 1.96415424e-01
-6.97934479e-02 -3.14169936e-02 5.99889874e-01 1.08834493e+00
-4.13371980e-01 -6.52989984e-01 -2.58916557e-01 -7.50790350e-03
-8.78903091e-01 2.16514483e-01 -1.66344494e-01 7.57123828e-01
1.77005112e-01 7.47443795e-01 -2.09813669e-01 -2.14953035e-01
5.70687532e-01 -3.91354084e-01 3.72787267e-01 -6.01205766e-01
-7.41228163e-01 2.34172761e-01 -4.86508608e-01 -8.97120178e-01
-9.89772022e-01 -6.83534265e-01 -1.29050481e+00 -4.21197712e-02
-3.17893654e-01 -4.55355607e-02 8.10038269e-01 5.67570508e-01
1.63131133e-01 2.16809660e-02 3.03748965e-01 -1.42043245e+00
8.78801271e-02 -8.72219503e-01 -6.15580738e-01 1.04497448e-01
4.78230298e-01 -8.91876221e-01 -3.91526043e-01 4.20627952e-01] | [7.678689002990723, -1.002183437347412] |
a95c6e4c-0f20-433b-b176-d6db8b2c4dad | multi-target-regression-via-input-space | 1211.6581 | null | http://arxiv.org/abs/1211.6581v5 | http://arxiv.org/pdf/1211.6581v5.pdf | Multi-Target Regression via Input Space Expansion: Treating Targets as Inputs | In many practical applications of supervised learning the task involves the
prediction of multiple target variables from a common set of input variables.
When the prediction targets are binary the task is called multi-label
classification, while when the targets are continuous the task is called
multi-target regression. In both tasks, target variables often exhibit
statistical dependencies and exploiting them in order to improve predictive
accuracy is a core challenge. A family of multi-label classification methods
address this challenge by building a separate model for each target on an
expanded input space where other targets are treated as additional input
variables. Despite the success of these methods in the multi-label
classification domain, their applicability and effectiveness in multi-target
regression has not been studied until now. In this paper, we introduce two new
methods for multi-target regression, called Stacked Single-Target and Ensemble
of Regressor Chains, by adapting two popular multi-label classification methods
of this family. Furthermore, we highlight an inherent problem of these methods
- a discrepancy of the values of the additional input variables between
training and prediction - and develop extensions that use out-of-sample
estimates of the target variables during training in order to tackle this
problem. The results of an extensive experimental evaluation carried out on a
large and diverse collection of datasets show that, when the discrepancy is
appropriately mitigated, the proposed methods attain consistent improvements
over the independent regressions baseline. Moreover, two versions of Ensemble
of Regression Chains perform significantly better than four state-of-the-art
methods including regularization-based multi-task learning methods and a
multi-objective random forest approach. | ['Eleftherios Spyromitros-Xioufis', 'Grigorios Tsoumakas', 'Ioannis Vlahavas', 'William Groves'] | 2012-11-28 | null | null | null | null | ['multi-target-regression'] | ['miscellaneous'] | [ 7.73022771e-01 -1.96309134e-01 -6.80982232e-01 -6.55915678e-01
-1.29392195e+00 -3.09434533e-01 6.35205388e-01 2.09937811e-01
-2.67551064e-01 1.14674079e+00 -1.95035949e-01 -1.48716167e-01
-2.93971181e-01 -4.15195018e-01 -5.60370266e-01 -1.02015567e+00
2.55430341e-01 8.00189734e-01 1.36763304e-01 9.53614805e-03
1.73519835e-01 1.95573509e-01 -1.64417434e+00 3.19675297e-01
7.66458452e-01 1.09394550e+00 -1.00733280e-01 4.33623970e-01
-3.65672074e-02 7.97952831e-01 -3.42496455e-01 -2.45579764e-01
1.37700930e-01 -3.25615525e-01 -6.64045274e-01 -4.21307683e-02
5.21937609e-01 4.20102447e-01 6.65522575e-01 7.54615247e-01
3.07738990e-01 1.83341414e-01 1.08014894e+00 -1.29100239e+00
-2.52053022e-01 3.12213242e-01 -9.13027227e-01 -2.23010387e-02
8.30650553e-02 -3.96443397e-01 1.33003354e+00 -8.83129358e-01
2.80660301e-01 1.16956747e+00 8.66031349e-01 3.55449617e-01
-1.77757955e+00 -8.90946805e-01 2.98636228e-01 8.78987983e-02
-1.09724414e+00 -3.17909986e-01 7.53422856e-01 -7.15664923e-01
7.50876069e-01 3.93685937e-01 -5.39618805e-02 1.21986854e+00
3.09616208e-01 5.40472746e-01 1.81593430e+00 -8.57766330e-01
1.06748017e-02 5.30517757e-01 4.70526636e-01 4.75922704e-01
3.26671265e-02 3.14327300e-01 -4.60224330e-01 -5.14368117e-01
1.41191930e-01 2.60203052e-02 2.05563828e-02 -4.27996099e-01
-1.00072742e+00 1.23728549e+00 9.19838473e-02 4.25947338e-01
-3.35047692e-01 3.55689861e-02 4.26057458e-01 3.05822939e-01
1.12033737e+00 1.81020096e-01 -8.40187609e-01 3.61377269e-01
-1.07432365e+00 2.25179076e-01 6.86257064e-01 6.27103090e-01
9.67376411e-01 -1.89074069e-01 -2.05774873e-01 1.15238822e+00
4.13311601e-01 3.53471220e-01 4.88495320e-01 -3.26265216e-01
5.88025212e-01 6.39028311e-01 9.36096720e-03 -6.98768497e-01
-7.30832815e-01 -6.84498966e-01 -9.73267555e-01 4.07286167e-01
5.75975180e-01 -6.02011150e-03 -9.57735598e-01 1.59834468e+00
5.66446841e-01 2.01003715e-01 -8.19823593e-02 5.28109729e-01
4.25732225e-01 4.48982567e-01 6.17835283e-01 -6.77621007e-01
1.11049283e+00 -1.30736685e+00 -5.40979683e-01 -5.27663469e-01
6.57546580e-01 -7.74053216e-01 6.04874313e-01 4.90645111e-01
-5.06876051e-01 -6.48757756e-01 -7.36280024e-01 2.90756941e-01
-4.18140769e-01 2.86358148e-01 5.19096792e-01 6.87484384e-01
-4.79796052e-01 6.10422492e-01 -4.07216609e-01 -1.03947315e-02
2.09521711e-01 6.29791856e-01 -3.95330906e-01 -8.62675235e-02
-1.08136511e+00 1.27429044e+00 3.46522421e-01 -6.95938542e-02
-5.55740476e-01 -6.96415305e-01 -6.37502193e-01 -2.23361164e-01
4.16244298e-01 -3.14446360e-01 1.22560537e+00 -1.27513814e+00
-1.24863911e+00 9.31106865e-01 -2.66507596e-01 -8.04836825e-02
4.26481545e-01 -3.29217970e-01 -3.78281653e-01 -4.64714587e-01
2.05860704e-01 1.30462155e-01 1.18251586e+00 -1.49323320e+00
-9.26441371e-01 -5.71457326e-01 -3.81052107e-01 6.33826703e-02
-7.06487969e-02 4.01661426e-01 8.47524628e-02 -7.27860868e-01
-1.51565254e-01 -1.25188315e+00 -5.00925660e-01 -4.85386342e-01
-4.42530841e-01 -2.80803561e-01 8.83827090e-01 -4.90579218e-01
1.17017472e+00 -1.84731817e+00 5.27874470e-01 2.00056240e-01
1.44628838e-01 1.26506865e-01 -7.72219002e-02 2.68940330e-01
-5.78756273e-01 1.10289820e-01 -3.02483559e-01 -6.02226079e-01
-3.48959893e-01 2.00411335e-01 -1.71814710e-01 5.77214599e-01
2.06998810e-01 6.58369541e-01 -6.39726460e-01 -6.87365294e-01
1.56606123e-01 4.62862432e-01 -1.01707620e-03 1.86869249e-01
-3.25280994e-01 8.07274163e-01 -5.11117220e-01 7.60824561e-01
4.39435244e-01 -4.01927471e-01 2.66937256e-01 1.40953660e-02
9.15741641e-03 3.42490673e-02 -1.21005630e+00 1.20363152e+00
-7.28353918e-01 2.27056205e-01 -1.84944719e-01 -1.28484941e+00
1.03157878e+00 4.46436852e-01 7.00492501e-01 -3.35418254e-01
-8.26217160e-02 3.46880496e-01 -1.28437147e-01 -1.34623721e-01
1.41385287e-01 -7.47961342e-01 -1.08437717e-01 4.44999605e-01
2.74030957e-02 1.54518172e-01 1.98138226e-02 -5.02047956e-01
8.76699865e-01 4.99867976e-01 8.18765640e-01 2.11011879e-02
6.82018876e-01 7.89129063e-02 8.02147269e-01 5.59011757e-01
3.96743082e-02 4.28976446e-01 3.19617599e-01 -3.94019395e-01
-7.98431516e-01 -5.48867285e-01 -4.65430200e-01 1.68802989e+00
-3.82400811e-01 -1.80778429e-01 -8.24865773e-02 -1.18401897e+00
2.57403672e-01 7.62838006e-01 -8.74852359e-01 1.68307960e-01
-6.12256110e-01 -1.54551303e+00 2.06433877e-01 4.40126359e-01
-5.23932800e-02 -8.77619922e-01 -3.61544758e-01 4.34319705e-01
-2.49142498e-01 -9.40843940e-01 3.42287533e-02 9.96988595e-01
-9.44291592e-01 -1.21732712e+00 -6.25246823e-01 -5.21542966e-01
3.68173450e-01 6.09529465e-02 1.22153485e+00 -1.14343263e-01
2.00376395e-04 -5.92259578e-02 -4.91077304e-01 -5.27442276e-01
-6.25559807e-01 1.98448688e-01 -9.69306678e-02 2.65108854e-01
3.69594991e-01 -5.97188115e-01 1.14597343e-01 6.02256835e-01
-5.72052598e-01 1.53827041e-01 7.93618858e-01 1.19383585e+00
7.62833714e-01 -8.69270489e-02 7.85153151e-01 -1.50325298e+00
2.40129367e-01 -7.38195419e-01 -4.38981086e-01 5.89833200e-01
-1.11338639e+00 2.80844003e-01 6.79739296e-01 -7.45336533e-01
-9.30478036e-01 2.80438542e-01 2.12183520e-01 -4.27185357e-01
-2.08611652e-01 6.45966470e-01 4.34499979e-02 -1.20882094e-01
6.05388165e-01 -8.86792839e-02 -1.67609215e-01 -6.55759335e-01
1.32379591e-01 6.10237002e-01 -1.57482654e-01 -5.01053810e-01
7.28865683e-01 -2.96812765e-02 5.60270190e-01 -3.63470763e-01
-1.35591662e+00 -8.80958140e-01 -1.01775050e+00 -2.84820765e-01
6.62115276e-01 -7.90704727e-01 -2.96780407e-01 4.52507645e-01
-9.84429300e-01 -1.81224823e-01 2.11635493e-02 4.15065348e-01
-5.05107880e-01 6.53078333e-02 -2.00538114e-01 -9.17450249e-01
-2.58208543e-01 -1.24623346e+00 1.21977115e+00 -3.09724994e-02
-2.27659285e-01 -1.30701470e+00 4.20789182e-01 6.82361603e-01
3.28934848e-01 4.53755289e-01 1.28773880e+00 -1.17830896e+00
-8.63295346e-02 -2.37445131e-01 -1.03227749e-01 2.48653576e-01
2.63171762e-01 -2.19954297e-01 -1.09737217e+00 -2.23728403e-01
-9.39970464e-02 -6.07610226e-01 1.06164706e+00 4.69433010e-01
8.20260584e-01 -3.39886211e-02 -6.27711415e-01 2.70779043e-01
1.70842922e+00 3.74165922e-02 1.73588797e-01 4.98538524e-01
6.79736376e-01 8.27228725e-01 9.63547945e-01 2.90053785e-01
2.18122423e-01 1.22064114e+00 4.35390353e-01 -1.60168380e-01
1.12896182e-01 1.99427947e-01 2.25637689e-01 5.73838532e-01
-3.92737895e-01 -1.12054981e-01 -9.15229559e-01 1.25353456e-01
-2.14205647e+00 -8.34028959e-01 -5.63999593e-01 2.28845692e+00
9.82002378e-01 -2.36880705e-02 3.17675680e-01 3.44661504e-01
7.61538625e-01 2.05441907e-01 -5.34507513e-01 -2.94326693e-01
-7.06144571e-02 3.42340350e-01 6.27238989e-01 4.33560938e-01
-1.56353986e+00 7.10003614e-01 6.38637495e+00 9.40710187e-01
-1.05022562e+00 4.58683193e-01 8.08750033e-01 6.97223172e-02
1.66459084e-01 8.12389627e-02 -1.36183083e+00 2.13544950e-01
1.18045902e+00 2.79197156e-01 1.72373772e-01 9.28223848e-01
1.08756954e-02 -3.00324470e-01 -1.19523096e+00 5.19693792e-01
1.37838110e-01 -6.86452270e-01 -4.67317045e-01 1.01835594e-01
7.00995743e-01 -6.33359104e-02 -4.83054854e-02 5.23653269e-01
3.20264965e-01 -1.24313235e+00 4.61770326e-01 4.85000551e-01
8.21431756e-01 -5.09711087e-01 6.60585523e-01 5.46679437e-01
-1.10243738e+00 -2.57833064e-01 -6.26585409e-02 1.13168871e-02
3.42388116e-02 9.02986348e-01 -6.39031172e-01 8.30459893e-01
3.56434733e-01 8.29512358e-01 -7.72862613e-01 8.08833539e-01
-1.78489730e-01 8.60407531e-01 -7.48010129e-02 1.52742609e-01
5.64544834e-02 -2.81121075e-01 3.54514152e-01 1.28677881e+00
6.44674003e-02 -3.33322585e-01 3.97527874e-01 3.79852593e-01
3.27618718e-01 4.90081191e-01 -4.64929581e-01 3.62441748e-01
-6.95346668e-02 1.44605672e+00 -4.47432488e-01 -3.97583656e-02
-7.38908470e-01 5.41162670e-01 5.97350121e-01 2.40705505e-01
-1.01196265e+00 3.05965245e-01 2.88891613e-01 -1.51153862e-01
3.94523919e-01 1.04461804e-01 -5.24876773e-01 -9.57376182e-01
-6.04811125e-02 -1.05639148e+00 5.23076713e-01 -3.14032972e-01
-1.42650521e+00 4.85595554e-01 1.65299729e-01 -1.24993312e+00
-4.71356392e-01 -6.40194833e-01 -3.14767540e-01 1.05850720e+00
-1.75540686e+00 -1.72423267e+00 -3.95972580e-02 4.59542632e-01
5.79906881e-01 -2.48599440e-01 1.21123242e+00 2.93190390e-01
-5.74744463e-01 3.80214244e-01 5.11000216e-01 -3.93111497e-01
1.07793581e+00 -1.31047583e+00 -2.61680603e-01 3.13401312e-01
3.29598844e-01 5.42057399e-03 7.03974128e-01 -6.19512081e-01
-7.95561194e-01 -1.25174665e+00 1.07701766e+00 -5.11147439e-01
5.04160285e-01 -2.90914685e-01 -8.98237109e-01 9.25413072e-01
-1.66547015e-01 1.65675223e-01 1.22532487e+00 7.05720842e-01
-5.60726404e-01 -1.18139401e-01 -9.08457875e-01 -1.77579045e-01
4.53547716e-01 -1.51693746e-01 -4.12499994e-01 5.81862867e-01
2.09616601e-01 -1.06388189e-01 -1.05115640e+00 7.18641043e-01
6.75841510e-01 -9.02856827e-01 9.43822026e-01 -8.23018849e-01
5.33714652e-01 -1.82661526e-02 -3.60604435e-01 -1.57322359e+00
-4.58297521e-01 -5.67526296e-02 -1.99901789e-01 1.30255604e+00
7.17512608e-01 -5.92463553e-01 6.31269395e-01 5.07774055e-01
1.51267886e-01 -1.02683306e+00 -1.04841137e+00 -6.46926105e-01
2.77586132e-01 -3.84939879e-01 6.83492571e-02 9.55247045e-01
-4.80520934e-01 7.54777074e-01 -9.04663682e-01 -2.27792021e-02
8.58244956e-01 4.66797590e-01 6.69296145e-01 -1.74652755e+00
-4.29924071e-01 -2.63671935e-01 -1.76028281e-01 -4.29615140e-01
6.07761323e-01 -9.28351045e-01 -1.00411884e-01 -1.16698074e+00
4.13938671e-01 -9.23485816e-01 -5.02149045e-01 8.04767847e-01
-4.53889012e-01 1.84458435e-01 1.60018042e-01 4.62435097e-01
-3.98254722e-01 3.93594056e-01 8.54898810e-01 -1.81904063e-01
-2.09721699e-01 7.44717300e-01 -5.98199069e-01 6.26591384e-01
6.14495099e-01 -1.06554377e+00 -4.35178690e-02 2.91668922e-01
2.99833361e-02 2.50489295e-01 3.52306008e-01 -7.61172175e-01
-1.56741560e-01 -6.27982080e-01 4.14742827e-01 -4.71110433e-01
5.05791605e-01 -1.03341329e+00 4.19462591e-01 3.04835588e-01
-5.36784470e-01 -7.23163113e-02 1.71399638e-02 8.08494985e-01
-2.08378732e-01 -4.45955604e-01 1.02395916e+00 1.19882353e-01
-4.51275587e-01 -1.98281743e-02 -1.28754705e-01 -3.06581818e-02
1.39104772e+00 3.78277972e-02 -1.04590811e-01 -6.83980659e-02
-8.90689909e-01 4.08451371e-02 -2.58545671e-02 4.54383254e-01
1.86974779e-01 -1.20492721e+00 -9.05606389e-01 -7.22642317e-02
1.79580197e-01 -2.06653282e-01 -1.95696294e-01 1.16164827e+00
4.52003717e-01 5.84688783e-01 -4.84621450e-02 -5.97052336e-01
-1.74953043e+00 5.64101398e-01 3.37521046e-01 -1.26063025e+00
-1.27503067e-01 6.25615895e-01 2.20871940e-01 -6.00019097e-01
-8.52173311e-04 1.34601574e-02 -5.14252603e-01 4.91598666e-01
1.89965472e-01 4.54441249e-01 1.91166893e-01 -8.49227428e-01
-3.15102547e-01 7.86325514e-01 -1.88608140e-01 1.80891335e-01
1.63481033e+00 1.36269191e-02 -1.29307866e-01 1.04616225e+00
1.06404376e+00 -2.65924074e-02 -8.79956663e-01 -4.24372911e-01
4.22685266e-01 -3.16237301e-01 2.76702959e-02 -1.16803527e+00
-9.12754297e-01 6.70787215e-01 4.70631897e-01 7.78343379e-02
1.22980332e+00 -1.08804896e-01 1.48300678e-01 1.00169107e-02
3.33917320e-01 -8.82733166e-01 -6.78672567e-02 3.77942711e-01
8.03426147e-01 -1.77182961e+00 2.66228706e-01 -6.11927867e-01
-7.36714661e-01 1.20453537e+00 5.68808198e-01 1.67177260e-01
7.11630046e-01 5.74575402e-02 -1.46732023e-02 2.02067606e-02
-9.56377387e-01 -1.69311598e-01 6.92353547e-01 4.12984759e-01
6.53687418e-01 5.58998212e-02 -3.17743570e-01 5.02081454e-01
3.47329944e-01 -4.73544151e-02 -6.46322267e-03 7.94612050e-01
-2.03752682e-01 -1.63315225e+00 -6.30325139e-01 6.90837622e-01
-8.03778589e-01 -6.53461888e-02 -2.12332413e-01 9.15927351e-01
1.73138037e-01 1.11934459e+00 -6.33378446e-01 -3.63892466e-01
2.35748872e-01 5.92339098e-01 3.71756852e-01 -7.91321278e-01
-1.07372415e+00 9.71826836e-02 3.12062293e-01 -1.69657007e-01
-1.05276394e+00 -8.73818099e-01 -7.84642339e-01 2.86029041e-01
-8.88952076e-01 -9.73198842e-03 9.12092745e-01 1.29085934e+00
-1.02649152e-01 4.70601231e-01 8.48727167e-01 -9.00278568e-01
-9.07875776e-01 -1.27821922e+00 -6.13615870e-01 3.58917296e-01
4.27574396e-01 -1.13967001e+00 -4.64875817e-01 4.02946621e-02] | [9.158917427062988, 4.2946858406066895] |
3a2c5ddc-5db1-46c3-943b-c05a58194150 | orb-based-slam-accelerator-on-soc-fpga | 2207.08405 | null | https://arxiv.org/abs/2207.08405v1 | https://arxiv.org/pdf/2207.08405v1.pdf | ORB-based SLAM accelerator on SoC FPGA | Simultaneous Localization and Mapping (SLAM) is one of the main components of autonomous navigation systems. With the increase in popularity of drones, autonomous navigation on low-power systems is seeing widespread application. Most SLAM algorithms are computationally intensive and struggle to run in real-time on embedded devices with reasonable accuracy. ORB-SLAM is an open-sourced feature-based SLAM that achieves high accuracy with reduced computational complexity. We propose an SoC based ORB-SLAM system that accelerates the computationally intensive visual feature extraction and matching on hardware. Our FPGA system based on a Zynq-family SoC runs 8.5x, 1.55x and 1.35x faster compared to an ARM CPU, Intel Desktop CPU, and a state-of-the-art FPGA system respectively, while averaging a 2x improvement in accuracy compared to prior work on FPGA. | ['Deming Chen', 'Vibhakar Vemulapati'] | 2022-07-18 | null | null | null | null | ['simultaneous-localization-and-mapping'] | ['computer-vision'] | [-2.74303079e-01 -6.53342962e-01 -6.73494413e-02 -5.56142509e-01
-4.56578016e-01 -7.05702841e-01 3.75329137e-01 1.86337065e-02
-6.49018228e-01 5.73695660e-01 -4.95311022e-01 -5.42822719e-01
-1.60458043e-01 -6.77920938e-01 -6.31682813e-01 -6.14446588e-02
-3.80754441e-01 4.41687346e-01 4.54843879e-01 -5.54159641e-01
4.04036105e-01 5.71876347e-01 -1.80878150e+00 -5.93995571e-01
7.74302423e-01 1.11283624e+00 7.98265561e-02 7.31966794e-01
4.95179385e-01 2.73077905e-01 -3.73351872e-01 3.49670172e-01
8.55959237e-01 6.52464703e-02 -1.76009029e-01 -4.60339576e-01
7.60448098e-01 -3.29982072e-01 -5.87999940e-01 1.06120920e+00
5.19730687e-01 -1.08992472e-01 -1.11596249e-01 -1.70041955e+00
2.28856206e-01 -1.80650607e-01 -6.15126908e-01 -7.97429762e-04
5.42815566e-01 -3.97573411e-02 3.99685025e-01 -8.44213605e-01
4.70509738e-01 8.58177245e-01 1.03445435e+00 -2.19828397e-01
-8.49762321e-01 -1.03278017e+00 -6.40426815e-01 -7.75009170e-02
-1.96114564e+00 -4.02248263e-01 2.07131542e-02 -1.37569025e-01
1.25320446e+00 2.98980381e-02 1.08071291e+00 2.76055425e-01
1.27444017e+00 -1.47203088e-01 9.35688078e-01 -1.07782431e-01
2.56326407e-01 -2.61666775e-01 -4.51129749e-02 1.06487107e+00
1.19205856e+00 3.94918293e-01 -9.89693820e-01 -1.69699788e-01
6.48708463e-01 2.21642956e-01 -2.00857967e-02 -6.38204515e-01
-1.50027776e+00 7.92858481e-01 7.42436945e-01 -5.00955105e-01
-3.02638531e-01 9.72932994e-01 3.24859262e-01 5.31361699e-01
-2.01939151e-01 4.15800124e-01 -2.59850651e-01 -5.79023540e-01
-1.10280323e+00 3.85108948e-01 4.30797279e-01 1.49513042e+00
1.15834606e+00 2.49238133e-01 8.32458258e-01 5.31696454e-02
5.97089589e-01 1.07462883e+00 5.56166470e-01 -6.85165346e-01
1.51187107e-01 7.22801983e-01 2.57174671e-01 -9.38600838e-01
-6.43534899e-01 -3.88689786e-01 -3.51299852e-01 5.97076952e-01
-1.29861146e-01 -2.67215937e-01 -8.07978451e-01 9.40584064e-01
4.47371662e-01 2.64059827e-02 1.93999797e-01 1.03643608e+00
5.12078047e-01 4.58064467e-01 -5.43750525e-01 5.15960634e-01
1.40890682e+00 -9.25963819e-01 -6.11649692e-01 -6.90636218e-01
6.72463000e-01 -1.08458114e+00 4.32820320e-01 3.89382869e-01
-1.93141058e-01 -4.65090990e-01 -2.02433324e+00 -2.57167041e-01
-2.78065294e-01 4.55381751e-01 1.18312955e+00 6.94882870e-01
-1.12480342e+00 3.61578703e-01 -1.09830570e+00 -7.51612782e-01
-1.13872541e-02 8.58862519e-01 -7.28085399e-01 3.63902412e-02
-7.47105479e-01 1.16849196e+00 3.74656379e-01 -2.02479716e-02
-8.50934505e-01 -5.89041770e-01 -1.04097056e+00 -4.96446759e-01
1.91128939e-01 -5.86691678e-01 1.16045821e+00 -5.53205848e-01
-1.57437396e+00 5.69805384e-01 -1.58481240e-01 -9.95393634e-01
2.58991271e-01 -6.28773153e-01 -5.60204923e-01 -1.06324852e-01
3.93512994e-01 8.54537249e-01 6.65893316e-01 -3.97614330e-01
-1.07111692e+00 -3.82736802e-01 -1.58172861e-01 2.38257885e-01
5.87994695e-01 -4.07593369e-01 2.30517983e-02 4.76910248e-02
8.51186275e-01 -1.47323573e+00 -3.69026542e-01 4.48507041e-01
1.61235675e-01 4.74260479e-01 1.21684766e+00 -4.27826159e-02
5.92452168e-01 -2.09584594e+00 -5.56282282e-01 1.96105912e-01
-2.73888241e-02 -1.18060059e-04 5.58643103e-01 6.03942037e-01
5.83930612e-01 -7.08268583e-01 3.38688910e-01 2.23539487e-01
-7.26271942e-02 2.56499141e-01 -3.50353181e-01 1.28196883e+00
-3.78976196e-01 5.97838044e-01 -8.43816876e-01 -1.47347569e-01
3.88978690e-01 4.63070422e-01 -4.44963187e-01 -2.86160409e-01
4.85302866e-01 5.10842130e-02 -1.18103668e-01 1.01498139e+00
7.34212816e-01 9.85074639e-02 2.30175063e-01 -1.05109893e-01
-9.56593812e-01 4.72297579e-01 -1.43721282e+00 2.23537493e+00
-5.08699596e-01 1.24381328e+00 1.22330956e-01 -9.75716412e-02
1.20493484e+00 -7.12249279e-02 -8.44560042e-02 -5.92830062e-01
5.27618647e-01 9.56035376e-01 1.17921360e-01 3.36394906e-01
1.24033177e+00 1.72504678e-01 -4.63803917e-01 1.46346301e-01
1.61075935e-01 -4.48806047e-01 5.33443969e-03 1.39722124e-01
1.21505773e+00 4.10239309e-01 9.70573068e-01 -1.01608872e+00
3.30023021e-01 7.48870850e-01 3.78110707e-01 4.38353121e-01
-1.99501380e-01 -1.16682146e-02 -2.75286287e-01 -7.58764029e-01
-7.89918721e-01 -8.64600420e-01 -2.40266129e-01 4.48301882e-01
1.10663903e+00 -7.13970065e-01 -6.56669885e-02 1.21298574e-01
5.82777679e-01 -6.80169612e-02 2.29309518e-02 -1.10878721e-02
-2.62281001e-01 -2.53708214e-01 6.19804621e-01 2.31606081e-01
8.86069417e-01 -1.04759648e-01 -1.84023225e+00 3.79675537e-01
5.72965860e-01 -9.31002319e-01 -8.80285501e-02 3.38343769e-01
-8.65380943e-01 -9.71753597e-01 2.00229645e-01 -6.86131537e-01
7.28900254e-01 9.47197318e-01 7.02141166e-01 6.98830485e-02
-6.13241673e-01 9.59469229e-02 -2.27468491e-01 -5.10235965e-01
4.39874530e-01 -1.15543239e-01 7.73232937e-01 -4.10902649e-01
2.91075736e-01 -4.57465440e-01 -6.19102240e-01 2.74893194e-01
-4.88392673e-02 1.49766654e-01 8.15896034e-01 6.29843473e-01
6.22863591e-01 -1.15880556e-01 -1.40365124e-01 -2.31511295e-01
-7.26073096e-03 -1.48533106e-01 -1.50155056e+00 -3.96083653e-01
-8.57569814e-01 2.44840950e-01 3.20470691e-01 5.60182668e-02
-3.71721089e-01 7.78152645e-01 3.74325275e-01 1.72448382e-02
3.66539001e-01 3.65083665e-01 2.54276842e-01 -1.20683849e+00
6.08117104e-01 8.19524601e-02 2.83167988e-01 7.34925270e-02
2.04296634e-01 7.29954600e-01 9.02551293e-01 -9.69287157e-02
1.10106397e+00 7.43436217e-01 5.84924698e-01 -8.15237582e-01
-1.78036705e-01 -7.14085937e-01 -4.45420712e-01 -1.03114270e-01
3.29793364e-01 -1.62077165e+00 -9.12136316e-01 1.76485881e-01
-6.51881814e-01 9.03730318e-02 3.20826679e-01 8.12489390e-01
-3.55134994e-01 7.97466040e-02 7.27236792e-02 -4.68341470e-01
-6.12726271e-01 -1.26391804e+00 1.13596582e+00 4.96523917e-01
-2.93752521e-01 -2.83388495e-01 3.02565813e-01 -2.60484785e-01
6.01027906e-01 3.54885221e-01 -2.88171887e-01 -3.31228338e-02
-1.09090436e+00 -6.90030158e-01 -3.11080992e-01 -7.33066082e-01
3.77129801e-02 -1.83125615e-01 -4.41798687e-01 -7.88319707e-01
-3.21671247e-01 -1.59191653e-01 2.85202771e-01 6.34806678e-02
-2.22521663e-01 1.57859072e-01 -7.75332153e-01 1.10685730e+00
1.96912336e+00 2.49376401e-01 2.56411910e-01 7.00394809e-01
3.79233271e-01 -1.69571459e-01 1.45228839e+00 4.59060311e-01
3.56924385e-01 9.01167274e-01 7.37696528e-01 3.05179030e-01
2.18698844e-01 -3.99695009e-01 3.67129982e-01 3.83750558e-01
2.18316972e-01 3.13821733e-01 -1.01096261e+00 3.50835919e-01
-1.74061155e+00 -2.52992243e-01 -4.90501225e-01 2.39913559e+00
1.25423923e-01 1.80044919e-01 -4.46505934e-01 -5.97188994e-03
4.41693604e-01 2.31924914e-02 -2.20216230e-01 -3.25257331e-01
2.70013511e-01 2.31672570e-01 1.51068115e+00 6.89622581e-01
-1.18371975e+00 1.05658305e+00 5.24062920e+00 1.78159237e-01
-1.58519185e+00 -5.33146635e-02 -7.45596230e-01 -7.49829412e-02
4.29795861e-01 7.12579131e-01 -1.21782362e+00 3.58477771e-01
1.33007193e+00 -5.41716337e-01 2.02352196e-01 1.54038918e+00
-2.74450123e-01 -7.10948110e-01 -8.25644612e-01 1.41624928e+00
-2.86530107e-02 -1.30053985e+00 -3.85648400e-01 4.61412936e-01
4.91603374e-01 4.32428688e-01 -2.91741431e-01 1.26041770e-01
2.47093990e-01 -8.13551486e-01 9.78960633e-01 -1.08498484e-01
1.05285501e+00 -1.13396740e+00 1.18546796e+00 1.62150368e-01
-1.69274116e+00 2.22534701e-01 -5.99146545e-01 -7.57656336e-01
3.26721966e-01 3.31786335e-01 -1.04195976e+00 5.20645678e-01
8.61740112e-01 6.26863301e-01 -6.75422251e-01 1.24248612e+00
-2.01544985e-01 -1.79327533e-01 -6.53171957e-01 -3.30688208e-01
4.41453606e-01 -7.00111836e-02 3.93654227e-01 6.52316451e-01
8.84809017e-01 -9.44854133e-03 3.09792668e-01 3.92096676e-02
3.56125236e-01 -2.70997375e-01 -8.37197781e-01 2.89529145e-01
6.28408730e-01 1.38981271e+00 -7.63847589e-01 -2.80450702e-01
-2.54670918e-01 8.04437339e-01 -1.19109526e-01 -6.12574637e-01
-1.02543104e+00 -1.19702709e+00 1.19305074e+00 -6.38617016e-03
2.57811248e-02 -1.07527208e+00 -1.67663589e-01 -8.48635197e-01
-1.67565390e-01 -5.32733262e-01 -2.63985544e-01 -6.52806818e-01
-2.62515932e-01 7.55955577e-01 -4.11088526e-01 -1.89684165e+00
-2.42838383e-01 -6.28286541e-01 1.18779793e-01 6.81647420e-01
-1.31296432e+00 -1.00202703e+00 -1.00912619e+00 3.91937941e-01
3.86368573e-01 -2.44223535e-01 8.28056753e-01 2.06775114e-01
1.51481643e-01 2.03308150e-01 1.23648100e-01 -3.30520809e-01
9.21606421e-01 -7.85568655e-01 8.06232750e-01 1.34130514e+00
1.48587048e-01 9.10927773e-01 8.74291241e-01 -9.79032278e-01
-2.49849749e+00 -9.59166884e-01 8.77800822e-01 -2.89239645e-01
6.36804521e-01 -5.65115690e-01 2.20895279e-03 9.81025934e-01
1.70782566e-01 3.64658356e-01 1.86749339e-01 -4.28764910e-01
-4.07788903e-02 -4.47865397e-01 -1.14176083e+00 4.85966623e-01
9.80381727e-01 -3.81846309e-01 -3.36392492e-01 2.32473224e-01
4.56390738e-01 -1.24273837e+00 -4.65831637e-01 2.71437496e-01
9.81780767e-01 -9.85407174e-01 6.41184092e-01 3.93782109e-01
-6.14821553e-01 -1.20632589e+00 -5.13051689e-01 -8.46546650e-01
-2.35764623e-01 -9.77337003e-01 3.62730861e-01 4.88328010e-01
4.73001320e-03 -1.16083062e+00 7.25166261e-01 -1.25535447e-02
-3.23292524e-01 -8.84753689e-02 -1.25129282e+00 -9.62497354e-01
-1.17558169e+00 -2.16835797e-01 6.44701898e-01 4.16288406e-01
5.80541193e-02 2.56041288e-01 -4.19594377e-01 8.64437163e-01
7.98597872e-01 3.13068032e-01 1.41986191e+00 -1.36008227e+00
3.25861931e-01 3.33371609e-01 -1.48366654e+00 -1.05365646e+00
-2.08679333e-01 -6.69799805e-01 1.38146490e-01 -1.19360125e+00
-4.65064645e-01 -3.72885495e-01 4.47666198e-01 4.44979221e-01
5.77603698e-01 6.62229061e-01 -1.27753735e-01 1.86407715e-01
-7.71505654e-01 2.70526677e-01 2.36897379e-01 1.43909857e-01
-2.67820582e-02 -2.03561261e-01 -1.24307461e-01 5.63166916e-01
4.99046415e-01 -6.60551846e-01 -3.63822520e-01 -4.10420954e-01
4.15451914e-01 3.91343273e-02 3.53935987e-01 -1.86761463e+00
8.04974675e-01 2.26795942e-01 4.40171927e-01 -1.02974653e+00
6.25199080e-01 -1.14311790e+00 6.62671030e-01 1.33208954e+00
7.09581196e-01 7.82516122e-01 4.42998022e-01 4.38803345e-01
-3.57606232e-01 1.50030240e-01 6.04546785e-01 7.53521472e-02
-1.42295659e+00 9.01067406e-02 -5.16226709e-01 -5.98795354e-01
1.37852526e+00 -4.46723253e-01 -5.72043061e-01 -1.95432767e-01
3.06527019e-01 2.68091023e-01 1.05227840e+00 3.78953904e-01
6.82083189e-01 -1.43457448e+00 -1.49756297e-01 6.55399084e-01
2.42048711e-01 -7.65406638e-02 1.93716753e-02 7.13877499e-01
-1.84218121e+00 1.01438975e+00 -8.49355876e-01 -9.81523693e-01
-1.33462751e+00 2.93623051e-03 5.24953902e-02 5.04466653e-01
-6.68202400e-01 6.85038447e-01 -5.43667078e-01 -3.40487182e-01
-2.54960090e-01 -3.45948339e-01 6.08817101e-01 -3.65788758e-01
5.50372422e-01 6.28239751e-01 3.88601810e-01 -6.54936433e-01
-1.19811499e+00 9.86052752e-01 3.33664179e-01 -3.43206912e-01
8.70414793e-01 -1.40969992e-01 -1.89677209e-01 1.90306067e-01
8.67385089e-01 3.35461825e-01 -1.07468343e+00 4.34210867e-01
1.97188836e-02 -7.65883267e-01 2.95019567e-01 -1.94642231e-01
-4.77063596e-01 3.32019687e-01 1.03046715e+00 -5.91350615e-01
7.10589647e-01 -6.83169782e-01 6.16384983e-01 7.22655833e-01
1.60838580e+00 -7.60753989e-01 -4.33710158e-01 7.31653690e-01
3.78122300e-01 -1.42065358e+00 9.90135133e-01 -3.64006549e-01
-3.04012179e-01 1.18633914e+00 5.79667330e-01 -8.12825143e-01
4.42368418e-01 5.63323975e-01 3.27927679e-01 -2.01461539e-01
-5.12378633e-01 -1.09489979e-02 1.50009431e-02 5.21916151e-01
1.35741487e-01 1.20878562e-01 -4.35416073e-01 5.52039519e-02
-7.88741827e-01 -6.23466261e-02 6.67241991e-01 1.68421197e+00
-8.26596975e-01 -8.18456411e-01 -3.38941991e-01 -1.78067591e-02
-2.20376756e-02 -1.07345015e-01 -2.96998620e-02 1.21080303e+00
7.60922730e-02 7.27735221e-01 2.23487839e-01 -7.25461125e-01
1.24111861e-01 -2.82993704e-01 4.58677560e-01 -3.99095863e-01
-5.74045479e-01 -2.77981132e-01 9.45740417e-02 -1.29281139e+00
-3.81174721e-02 -3.95013005e-01 -1.42661929e+00 -4.68593806e-01
-3.29502195e-01 4.98542152e-02 1.45427012e+00 2.67491430e-01
8.96180809e-01 1.25098020e-01 2.30218351e-01 -1.02932358e+00
-2.25291997e-01 -5.26617825e-01 -7.39083648e-01 -7.01668501e-01
4.51127976e-01 -9.54879522e-01 -1.42226815e-01 -5.55904448e-01] | [7.407505989074707, -2.085824728012085] |
f14f6cce-2191-4442-a448-0d615c6bbf26 | a-zero-shot-framework-for-sketch-based-image-1 | null | null | http://openaccess.thecvf.com/content_ECCV_2018/html/Sasikiran_Yelamarthi_A_Zero-Shot_Framework_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Sasikiran_Yelamarthi_A_Zero-Shot_Framework_ECCV_2018_paper.pdf | A Zero-Shot Framework for Sketch based Image Retrieval | Sketch-based image retrieval (SBIR) is the task of retrieving images from a natural image database that correspond to a given hand-drawn sketch. Ideally, an SBIR model should learn to associate components in the sketch (say, feet, tail, etc.) with the corresponding components in the image. However, current evaluation methods simply focus only on coarse-grained evaluation where the focus is on retrieving images which belong to the same class as the sketch but not necessarily having the same components as in the sketch. As a result, existing methods simply learn to associate sketches with classes seen during training and hence fail to generalize to unseen classes. In this paper, we propose a new bench mark for zero-shot SBIR where the model is evaluated on novel classes that are not seen during training. We show through extensive experiments that existing models for SBIR which are trained in a discriminative setting learn only class specific mappings and fail to generalize to the proposed zero-shot setting. To circumvent this, we propose a generative approach for the SBIR task by proposing deep conditional generative models which take the sketch as an input and fill the missing information stochastically. Experiments on this new benchmark created from the "Sketchy" dataset, which is a large-scale database of sketch-photo pairs demonstrate that the performance of these generative models is significantly better than several state-of-the-art approaches in the proposed zero-shot framework of the coarse-grained SBIR task. | ['Ashish Mishra', 'Anurag Mittal', 'Shiva Krishna Reddy', 'Sasi Kiran Yelamarthi'] | 2018-09-01 | null | null | null | eccv-2018-9 | ['sketch-based-image-retrieval'] | ['computer-vision'] | [ 3.52382183e-01 -4.13508832e-01 -1.98050946e-01 -3.67178679e-01
-1.02703094e+00 -6.36934996e-01 1.00985634e+00 -3.21206003e-01
-2.90782209e-02 4.97538835e-01 8.12252867e-04 1.27573252e-01
-2.61454731e-01 -9.78443503e-01 -9.33400989e-01 -7.49455988e-01
4.03574347e-01 8.92118633e-01 2.40463048e-01 -1.46636456e-01
2.60693341e-01 5.11431515e-01 -1.88110840e+00 5.34578562e-01
2.34374687e-01 8.91012728e-01 2.86871135e-01 6.07640386e-01
-2.44634464e-01 4.64332163e-01 -6.81776404e-01 -5.32224119e-01
3.47417593e-01 -4.13932145e-01 -5.41195512e-01 2.41845384e-01
1.07450461e+00 -6.87480509e-01 -5.79165220e-01 8.97637308e-01
3.82766277e-01 3.26979250e-01 9.65293705e-01 -1.39305902e+00
-1.15174043e+00 4.21274565e-02 -3.76595318e-01 -1.18141428e-01
3.36230487e-01 -3.69699210e-01 1.24063218e+00 -1.55199218e+00
9.84941304e-01 1.25609732e+00 2.54724115e-01 8.25435996e-01
-1.25698495e+00 -6.95402861e-01 1.55273438e-01 1.16018079e-01
-1.76064885e+00 -3.76047850e-01 9.24794137e-01 -1.94750607e-01
5.13693571e-01 2.68687129e-01 3.09019506e-01 1.19754434e+00
-3.16967458e-01 1.10270655e+00 7.74984598e-01 -4.08628225e-01
3.03794920e-01 2.41281703e-01 -2.03512330e-02 5.84585726e-01
4.33277749e-02 7.28198153e-04 -7.17451990e-01 -2.56507665e-01
9.69906211e-01 6.68558240e-01 -2.06092820e-02 -8.99756312e-01
-1.14818358e+00 9.05821323e-01 4.67306912e-01 3.76092702e-01
-2.64535904e-01 2.94872075e-01 9.41470712e-02 3.17470014e-01
3.50683928e-01 4.29570228e-02 -6.57857880e-02 3.47235829e-01
-1.17827916e+00 3.79938215e-01 7.27361202e-01 1.37209141e+00
1.05153739e+00 -3.52275103e-01 -4.83603925e-01 1.15072584e+00
7.76352435e-02 6.99010134e-01 2.22600058e-01 -6.85435414e-01
3.18097264e-01 4.76294279e-01 2.17896685e-01 -8.45382214e-01
6.24414980e-01 -2.21908391e-01 -7.66185105e-01 2.02416390e-01
-5.16184885e-03 6.67542636e-01 -1.25676727e+00 1.72391713e+00
7.21206591e-02 3.14787090e-01 1.21083250e-02 8.97147596e-01
1.08816648e+00 6.73895597e-01 -1.02666356e-01 2.55531222e-01
1.21239758e+00 -1.02726412e+00 -3.88580263e-01 -6.19049259e-02
-1.83287755e-01 -8.63146782e-01 1.33490467e+00 2.78920621e-01
-9.03564453e-01 -7.50634253e-01 -1.16636312e+00 -9.53351632e-02
-6.70203149e-01 3.94958019e-01 2.41764098e-01 5.45409441e-01
-1.03257966e+00 5.09742916e-01 -2.90146947e-01 -5.89038193e-01
3.91993999e-01 -3.24269384e-02 -5.56065619e-01 -6.58447325e-01
-8.22703481e-01 5.15071392e-01 3.54644582e-02 -2.12926283e-01
-1.39418793e+00 -5.24387538e-01 -6.59300447e-01 2.88069576e-01
4.00620639e-01 -7.52447009e-01 1.00676036e+00 -9.62230682e-01
-1.13156712e+00 1.00935197e+00 -3.47752303e-01 6.56852731e-03
4.23592538e-01 -8.96413177e-02 -2.10774332e-01 3.23736370e-01
2.01871842e-01 9.11781371e-01 1.32461596e+00 -1.74314606e+00
-5.09173751e-01 -3.09968293e-01 2.45743588e-01 1.70592288e-03
-3.28285486e-01 -1.79529235e-01 -7.85208046e-01 -8.12409103e-01
4.06126529e-02 -8.77040386e-01 2.74242520e-01 3.58924240e-01
-1.17190108e-01 -3.87014866e-01 1.24419284e+00 -2.81946510e-01
9.96589005e-01 -2.15353155e+00 1.06897473e-01 1.71689585e-01
-1.80506092e-02 3.49514395e-01 -4.52534348e-01 8.00141931e-01
-7.41234869e-02 6.95548505e-02 -1.66154876e-01 -6.08441055e-01
1.78142369e-01 5.50017297e-01 -8.25001121e-01 7.81065971e-02
2.50701040e-01 1.04348648e+00 -1.06767392e+00 -4.35228080e-01
3.33678633e-01 6.13533974e-01 -3.26102555e-01 3.94308209e-01
-1.16336860e-01 7.41270408e-02 -2.62964189e-01 8.19189668e-01
7.74216592e-01 -3.03979963e-01 -3.67101654e-03 -1.69331506e-01
4.60883439e-01 -2.16671422e-01 -1.15609205e+00 1.99277532e+00
-5.08872688e-01 3.90611261e-01 -2.53517419e-01 -9.49901819e-01
1.06519234e+00 3.21951091e-01 1.75079629e-01 -6.12467825e-01
-4.64813441e-01 1.34077951e-01 -5.49774587e-01 -1.35233656e-01
5.30217946e-01 -4.06321883e-01 -9.45321545e-02 5.66267669e-01
4.14733917e-01 -1.66137353e-01 9.15969238e-02 4.83650148e-01
9.08569753e-01 1.36217311e-01 9.40634310e-02 -7.37028047e-02
4.45163608e-01 -3.62203151e-01 1.05418161e-01 1.15151942e+00
1.93590403e-01 1.05464327e+00 3.61873657e-02 -5.61164141e-01
-1.33893299e+00 -1.67376149e+00 -3.54337762e-03 1.13634288e+00
3.43617886e-01 -4.82239544e-01 -3.01356584e-01 -7.01564252e-01
2.41369568e-02 4.84632760e-01 -7.87388861e-01 -8.83753076e-02
-1.09849997e-01 -1.17179655e-01 2.71086633e-01 5.21262884e-01
3.31009001e-01 -1.22641492e+00 -3.68785292e-01 3.11854631e-02
-2.92473733e-02 -8.61254573e-01 -6.43627763e-01 -3.44180256e-01
-6.33507311e-01 -9.47741628e-01 -1.23168468e+00 -9.97052312e-01
9.68529224e-01 7.23821878e-01 1.33350897e+00 3.39300901e-01
-5.91851056e-01 8.34083438e-01 -3.90520126e-01 -2.17438057e-01
-8.82258564e-02 -2.46862859e-01 -2.71548867e-01 3.68970722e-01
2.80993313e-01 -5.62455833e-01 -8.27900469e-01 4.98395890e-01
-1.34639049e+00 -1.23437487e-01 7.69405305e-01 1.25888205e+00
7.85566628e-01 -9.54227224e-02 6.53591156e-01 -8.37612152e-01
4.91723806e-01 -4.52915668e-01 -2.46542037e-01 7.81373680e-01
-4.45322841e-01 1.39002725e-01 4.14851248e-01 -6.47356570e-01
-9.88156915e-01 1.02416128e-01 1.07138090e-01 -1.06991398e+00
-2.04062015e-01 2.02664673e-01 -1.81151673e-01 7.44677261e-02
3.10366899e-01 6.01084054e-01 -1.65044382e-01 -5.64296663e-01
5.58067858e-01 5.46290994e-01 5.30800641e-01 -9.01515663e-01
8.86193275e-01 6.70187235e-01 8.01461283e-03 -8.15560758e-01
-6.90407634e-01 -8.47810447e-01 -5.65339684e-01 -7.47896880e-02
5.20041168e-01 -9.32531059e-01 -2.28822231e-01 1.35563910e-01
-1.01789010e+00 6.70076311e-02 -3.76641273e-01 -1.12505667e-02
-6.93261981e-01 4.54415143e-01 -2.25780979e-01 -9.67508197e-01
-4.97479796e-01 -9.32093382e-01 1.75958276e+00 1.07236564e-01
1.11960776e-01 -6.82988405e-01 1.36466548e-01 9.23251268e-03
3.51197839e-01 -1.27464771e-01 1.09812009e+00 -5.14177501e-01
-9.63114262e-01 -4.60937440e-01 -4.07181978e-01 4.02430058e-01
2.22222969e-01 3.19002345e-02 -1.03238833e+00 -4.14386868e-01
-4.64004934e-01 -4.96398985e-01 1.02878761e+00 -3.89850661e-02
1.18823898e+00 -1.58887744e-01 -3.73454899e-01 2.79183835e-01
1.76225495e+00 1.82303056e-01 9.01247084e-01 -2.89599746e-01
4.10242319e-01 4.35866088e-01 7.36302137e-01 3.72066051e-01
4.46041785e-02 8.38070631e-01 3.24030131e-01 5.30633703e-02
-3.68417531e-01 -8.00937891e-01 -1.72597859e-02 3.83357823e-01
8.97249207e-02 -3.81022453e-01 -4.91502434e-01 9.97683942e-01
-1.94354331e+00 -1.23678172e+00 4.03380990e-01 2.56507492e+00
5.81471264e-01 -2.88222522e-01 -2.94685680e-02 -6.25980198e-02
7.36726701e-01 1.62711501e-01 -3.21101397e-01 -3.12918462e-02
-1.95535496e-02 6.47753239e-01 -1.37944192e-01 3.46555322e-01
-1.03623724e+00 1.04520261e+00 5.78399467e+00 1.04060256e+00
-8.84263813e-01 1.46098003e-01 3.91916215e-01 -6.18673116e-02
-4.64562088e-01 1.39817029e-01 -7.22177625e-01 2.92701602e-01
3.07080418e-01 -8.74204785e-02 6.92994773e-01 9.65213776e-01
-5.80162764e-01 4.23233956e-02 -1.50413370e+00 1.30099487e+00
5.27410805e-01 -1.28162706e+00 6.10197365e-01 -1.03095062e-01
9.77666676e-01 -3.80262375e-01 3.45612764e-01 4.82030481e-01
1.42049920e-02 -1.01947224e+00 6.31727815e-01 8.68554652e-01
1.16593599e+00 -6.14759624e-01 3.71774912e-01 3.34481269e-01
-1.23908913e+00 1.71331257e-01 -7.79958427e-01 3.12090516e-01
-2.08827481e-01 8.98919329e-02 -8.35771263e-01 5.61220706e-01
6.01029515e-01 6.35738850e-01 -5.96385121e-01 9.74952877e-01
-1.52490392e-01 1.10375874e-01 -4.93170135e-02 5.09340800e-02
1.63672164e-01 -1.33644700e-01 4.17321533e-01 9.86699879e-01
5.54557621e-01 1.22674755e-04 4.93624695e-02 1.09055662e+00
-2.61380196e-01 -9.17656869e-02 -1.08489966e+00 -5.32998070e-02
4.88656074e-01 1.22866702e+00 -5.09064436e-01 -7.76400626e-01
-4.88669544e-01 1.35969675e+00 3.41114372e-01 5.94209194e-01
-5.08245766e-01 -4.37212527e-01 6.80612445e-01 1.14610568e-02
7.92301893e-01 6.10202178e-02 8.33135545e-02 -1.20773602e+00
1.71780318e-01 -5.52955806e-01 3.48288059e-01 -1.08769524e+00
-1.77371764e+00 5.46584129e-01 9.47952121e-02 -1.40772033e+00
-4.12499040e-01 -5.13964891e-01 -6.30594134e-01 9.85213935e-01
-1.38571167e+00 -1.48097682e+00 -4.91069317e-01 8.75985861e-01
8.76020372e-01 -2.95377374e-01 1.11499023e+00 4.29606348e-01
1.29722238e-01 6.61339343e-01 2.04692766e-01 1.69463232e-01
9.30848479e-01 -1.06071758e+00 3.58108640e-01 6.09673738e-01
7.95360565e-01 9.03444946e-01 3.58429164e-01 -6.07387960e-01
-1.41588783e+00 -1.18612850e+00 9.01516199e-01 -6.66674018e-01
3.56477469e-01 -6.08036399e-01 -8.05574596e-01 5.99556506e-01
1.10973127e-03 3.78200799e-01 3.68843257e-01 -4.42565680e-02
-7.80283213e-01 -2.93197066e-01 -1.10612667e+00 4.17803466e-01
1.19333661e+00 -1.08950663e+00 -6.89552963e-01 2.46471629e-01
2.85459161e-01 1.86208934e-02 -6.49590433e-01 2.11660609e-01
9.82761860e-01 -7.57952571e-01 1.51851010e+00 -8.10612619e-01
6.57391131e-01 -3.23665828e-01 -4.75558400e-01 -1.08806968e+00
-2.30472520e-01 -5.62089495e-02 -3.09407078e-02 1.18563867e+00
3.66804786e-02 -1.19432487e-01 8.23197842e-01 3.63817394e-01
2.34784275e-01 -5.21024168e-01 -8.64302456e-01 -1.03094006e+00
-1.59347579e-01 -1.57855898e-02 6.91601515e-01 7.27599859e-01
-5.35587311e-01 4.27534342e-01 -6.88947916e-01 -1.18906992e-02
7.94909716e-01 6.35550797e-01 1.01094890e+00 -1.27172232e+00
-3.03290039e-01 -2.83911288e-01 -4.72348511e-01 -9.90404665e-01
1.22414507e-01 -7.98870444e-01 1.94105566e-01 -1.61823928e+00
7.02629447e-01 -4.68665630e-01 -4.72695261e-01 3.61817420e-01
-3.05268794e-01 5.54509580e-01 3.74018490e-01 4.06000644e-01
-7.07319856e-01 5.86532474e-01 1.07395816e+00 -5.67734241e-01
3.11567783e-01 -2.63801776e-02 -3.67242634e-01 2.08271682e-01
2.82490700e-01 -4.12495613e-01 -7.94519663e-01 -3.70165527e-01
2.18815934e-02 1.32580161e-01 8.58834922e-01 -6.88826203e-01
2.68160969e-01 -1.80730168e-02 4.39849645e-01 -8.44283879e-01
7.41671860e-01 -1.06175673e+00 4.56307799e-01 7.99941868e-02
-3.35934937e-01 -1.71996146e-01 -2.25463092e-01 8.92554581e-01
-3.95113736e-01 -3.38050067e-01 5.51564753e-01 -2.82061011e-01
-8.64875734e-01 6.44576371e-01 1.88995406e-01 -2.48946860e-01
8.66574407e-01 -1.68980926e-01 -2.89229572e-01 -5.99462807e-01
-7.66656637e-01 -2.63026595e-01 6.31410956e-01 6.62237644e-01
1.08206141e+00 -1.75582743e+00 -6.09718382e-01 3.86433810e-01
7.88648546e-01 -2.22343594e-01 4.06997949e-01 1.59466401e-01
-4.30640429e-02 4.89582449e-01 -3.23405534e-01 -5.76835811e-01
-1.19166744e+00 1.04322720e+00 5.56088462e-02 -1.53654471e-01
-6.22257948e-01 8.19338858e-01 7.37968624e-01 -4.45630670e-01
2.91819155e-01 1.63990948e-02 2.16559589e-01 -1.17759243e-01
6.03561759e-01 3.01706314e-04 1.38137594e-01 -4.64927822e-01
-3.08983117e-01 6.42217159e-01 -1.30159050e-01 -2.24618226e-01
1.21808267e+00 1.02872141e-01 5.44531643e-02 5.31044781e-01
1.31411552e+00 -1.23703517e-01 -1.14722085e+00 -6.07054949e-01
-1.91727594e-01 -9.54332888e-01 -2.48851597e-01 -8.64352643e-01
-1.03197956e+00 1.08056533e+00 5.65778673e-01 -2.07239643e-01
9.82883096e-01 2.88509727e-01 6.06511533e-01 5.90621412e-01
6.52080953e-01 -8.73574793e-01 5.12732208e-01 1.47506401e-01
1.22081029e+00 -1.24076700e+00 -2.21206561e-01 -1.47974193e-01
-4.81690168e-01 1.14099956e+00 4.01022047e-01 -4.37205076e-01
5.59534550e-01 -1.75917476e-01 -2.94023275e-01 -2.61062652e-01
-8.48549128e-01 -2.45683864e-01 5.34734011e-01 5.38816333e-01
1.12828761e-01 1.46555558e-01 -5.01004197e-02 3.51110518e-01
4.00221407e-01 1.05928101e-01 9.35899764e-02 1.00356197e+00
-2.48976260e-01 -1.43672371e+00 -2.23937422e-01 4.18146491e-01
6.67050034e-02 -1.55549496e-01 -6.92256749e-01 6.84923530e-01
-1.34762481e-01 6.35090590e-01 1.17451146e-01 -2.44071856e-01
2.83914149e-01 2.62979388e-01 8.16390693e-01 -7.81345844e-01
-1.91358775e-02 -4.69136275e-02 -2.70829946e-01 -4.00131553e-01
-4.47096586e-01 -4.58106786e-01 -5.93357682e-01 -6.88972622e-02
-5.63517250e-02 4.42179339e-03 6.75472081e-01 6.93655372e-01
3.69048208e-01 1.06548019e-01 5.85397601e-01 -8.82962823e-01
-6.02974832e-01 -8.47939909e-01 -8.17145407e-01 7.75178671e-01
2.86466360e-01 -8.94071877e-01 -1.37445718e-01 1.00207433e-01] | [11.623329162597656, 0.6544802188873291] |
5f386c7f-7fa0-44f4-bf77-d8039603efd3 | on-confidence-intervals-for-precision | 2208.11977 | null | https://arxiv.org/abs/2208.11977v1 | https://arxiv.org/pdf/2208.11977v1.pdf | On confidence intervals for precision matrices and the eigendecomposition of covariance matrices | The eigendecomposition of a matrix is the central procedure in probabilistic models based on matrix factorization, for instance principal component analysis and topic models. Quantifying the uncertainty of such a decomposition based on a finite sample estimate is essential to reasoning under uncertainty when employing such models. This paper tackles the challenge of computing confidence bounds on the individual entries of eigenvectors of a covariance matrix of fixed dimension. Moreover, we derive a method to bound the entries of the inverse covariance matrix, the so-called precision matrix. The assumptions behind our method are minimal and require that the covariance matrix exists, and its empirical estimator converges to the true covariance. We make use of the theory of U-statistics to bound the $L_2$ perturbation of the empirical covariance matrix. From this result, we obtain bounds on the eigenvectors using Weyl's theorem and the eigenvalue-eigenvector identity and we derive confidence intervals on the entries of the precision matrix using matrix inversion perturbation bounds. As an application of these results, we demonstrate a new statistical test, which allows us to test for non-zero values of the precision matrix. We compare this test to the well-known Fisher-z test for partial correlations, and demonstrate the soundness and scalability of the proposed statistical test, as well as its application to real-world data from medical and physics domains. | ['Matthew B. Blaschko', 'Wacha Bounliphone', 'Aleksei Tiulpin', 'Teodora Popordanoska'] | 2022-08-25 | null | null | null | null | ['topic-models'] | ['natural-language-processing'] | [ 4.95291166e-02 1.97662830e-01 6.36393204e-02 -3.88231911e-02
-8.29742491e-01 -7.68974960e-01 2.97945112e-01 2.01081127e-01
-3.16490084e-01 6.74695313e-01 1.72992021e-01 -6.66813970e-01
-6.83762908e-01 -6.92377031e-01 -7.46059000e-01 -9.61599529e-01
-2.14973629e-01 3.44764471e-01 -4.10392508e-02 9.55107212e-02
2.03554958e-01 4.99023080e-01 -1.27139592e+00 -3.87923151e-01
7.92949915e-01 1.11731005e+00 -5.18038809e-01 6.17436886e-01
9.55611765e-02 1.21235594e-01 -4.57666218e-01 -2.03985140e-01
2.16669559e-01 -3.18976462e-01 -5.97504914e-01 -7.74398521e-02
-4.04441915e-02 8.34690058e-04 1.84671089e-01 1.46278095e+00
2.07584202e-01 2.43691266e-01 1.06835306e+00 -1.30816472e+00
-1.71015009e-01 7.62755156e-01 -7.54544735e-01 -1.82238761e-02
4.75587547e-01 -4.98909354e-01 9.42967653e-01 -7.82491565e-01
4.91469145e-01 1.06151724e+00 7.89189696e-01 4.01720293e-02
-1.46830261e+00 -5.78064561e-01 -4.00622964e-01 -9.84690264e-02
-1.85741603e+00 -2.96018124e-01 4.42830622e-01 -8.04478645e-01
2.79255390e-01 3.95859897e-01 2.46189758e-01 5.48640430e-01
3.16491276e-01 3.93345177e-01 1.00500906e+00 -7.24068880e-01
4.76839811e-01 3.33070159e-01 4.11040902e-01 7.15287983e-01
7.48279154e-01 1.89052880e-01 -2.65251070e-01 -5.47015488e-01
4.61758375e-01 -2.51550615e-01 -3.46166849e-01 -6.90908134e-01
-1.10331047e+00 9.48797464e-01 -1.18766092e-01 3.00354362e-01
-2.92258799e-01 1.57399684e-01 1.72988743e-01 -2.12513581e-01
4.13500279e-01 -4.68120426e-02 -8.11122358e-02 -3.36798988e-02
-7.94807374e-01 1.43165305e-01 1.11293805e+00 7.46787786e-01
5.70956051e-01 -3.51255864e-01 -1.01128414e-01 4.08139169e-01
5.26956260e-01 9.89476562e-01 3.96664962e-02 -8.17744434e-01
1.85026214e-01 1.80648729e-01 3.24198872e-01 -9.51482296e-01
-3.29849601e-01 -3.87343615e-01 -8.46141338e-01 9.70697030e-02
7.17023492e-01 -1.98933169e-01 -4.28528279e-01 1.96534503e+00
4.82214659e-01 5.24188280e-02 -1.35554336e-02 5.92795849e-01
6.67011067e-02 3.66929948e-01 -2.48207688e-01 -5.64671755e-01
1.14190853e+00 4.86247540e-02 -6.29467130e-01 3.05568010e-01
5.49733520e-01 -5.95980048e-01 6.70289099e-01 4.72195238e-01
-7.34954119e-01 -4.99975197e-02 -1.03002799e+00 4.03951019e-01
-2.52238102e-02 1.40065504e-02 3.93903106e-01 1.21243703e+00
-6.95158839e-01 5.31622648e-01 -8.99733961e-01 -1.55098766e-01
-1.57924786e-01 2.18019009e-01 -5.06820619e-01 5.92482202e-02
-1.01323199e+00 6.46395624e-01 3.07767659e-01 3.73192608e-01
-3.12019497e-01 -7.55364478e-01 -6.98783338e-01 2.65630543e-01
2.69319117e-01 -4.43338096e-01 1.00698161e+00 -2.49231517e-01
-1.27208030e+00 3.69264871e-01 -1.94699466e-01 -2.81685203e-01
3.82392734e-01 -6.34785369e-02 -1.74981385e-01 1.71621382e-01
2.39385396e-01 -3.45468849e-01 8.12942326e-01 -9.56355095e-01
-2.85712481e-01 -5.80749452e-01 -3.22975427e-01 -3.41391206e-01
-4.83408608e-02 -2.88980722e-01 -1.41353324e-01 -2.67286658e-01
6.54604375e-01 -1.01499724e+00 -3.92045140e-01 -5.22327662e-01
-7.34176934e-01 4.12510373e-02 6.82785213e-02 -4.53474283e-01
1.33850217e+00 -2.35181975e+00 3.18104327e-02 1.19735909e+00
1.78621709e-01 -4.74229157e-01 3.11320871e-01 6.01683438e-01
-2.24539176e-01 1.10313237e-01 -3.68706346e-01 2.86716461e-01
3.19086552e-01 -2.22444236e-01 -4.78209823e-01 9.87946153e-01
-1.00216173e-01 2.19327390e-01 -6.38889134e-01 -2.02791587e-01
1.45461664e-01 3.81001651e-01 -5.39713442e-01 -1.75820187e-01
9.49840173e-02 1.84236407e-01 -3.75213712e-01 9.69809219e-02
8.60091329e-01 -3.15227032e-01 4.29067761e-01 -2.98323333e-01
-1.73818156e-01 -3.16447057e-02 -1.82313967e+00 1.14983332e+00
-3.36944282e-01 2.77720928e-01 1.50022104e-01 -9.68525290e-01
6.41652584e-01 1.97360754e-01 6.22358322e-01 7.29684755e-02
4.12938207e-01 2.70698577e-01 8.89015123e-02 6.42536357e-02
3.32843781e-01 -4.65943068e-01 -1.69709548e-01 5.95996439e-01
-1.13735706e-01 -1.42256260e-01 5.02265930e-01 4.96905029e-01
1.03281403e+00 -2.62465060e-01 5.35842597e-01 -7.36922145e-01
6.44949079e-01 -4.54556584e-01 2.41985932e-01 6.43103540e-01
1.70171008e-01 3.11598986e-01 1.09604740e+00 1.68859407e-01
-8.36053967e-01 -1.40563691e+00 -7.01316118e-01 3.53367925e-01
-1.63373649e-01 -5.03279924e-01 -7.41300881e-01 -3.13287020e-01
2.80888706e-01 7.39809752e-01 -8.50912631e-01 -6.16375869e-03
1.12545609e-01 -1.02034187e+00 2.13014379e-01 2.72628158e-01
-5.40060103e-02 -2.61409059e-02 -4.75324273e-01 -1.83359206e-01
-2.67307088e-02 -9.63663280e-01 -2.35747829e-01 1.63593873e-01
-7.08594143e-01 -1.33250463e+00 -5.69890976e-01 1.31003812e-01
7.47692227e-01 -6.09672777e-02 6.69716239e-01 -4.50827479e-01
-5.66254137e-03 8.49809945e-01 -9.27158594e-02 -3.86065543e-01
-4.89549398e-01 -3.61233503e-01 4.36509550e-01 1.94258764e-01
1.33071721e-01 -5.94608009e-01 -4.35876936e-01 4.61690396e-01
-8.94728363e-01 -5.81099510e-01 4.93074000e-01 7.37981200e-01
6.82923079e-01 3.40313762e-01 1.96257532e-01 -7.50477195e-01
8.57426345e-01 -5.19135118e-01 -1.05705440e+00 2.81002909e-01
-7.94699371e-01 5.92019439e-01 1.37982890e-01 -2.79940933e-01
-7.07774282e-01 1.23289727e-01 2.80807287e-01 -2.84307122e-01
2.75287867e-01 9.95555639e-01 -1.32351696e-01 3.44361216e-02
6.79972470e-01 -6.23368248e-02 -8.43695924e-02 -4.34306085e-01
6.13635242e-01 5.00952125e-01 6.47434473e-01 -9.02096927e-01
8.00386488e-01 5.92884362e-01 6.52075052e-01 -9.79720533e-01
-8.94923508e-01 -7.24679768e-01 -4.71314788e-01 -1.64231062e-01
6.00884199e-01 -4.95709151e-01 -1.19875956e+00 -1.86662287e-01
-9.50707257e-01 1.86262980e-01 -2.74428278e-01 8.86170328e-01
-5.86744785e-01 5.96625209e-01 -2.08627403e-01 -1.38461244e+00
-8.63307342e-02 -8.81830931e-01 8.13934803e-01 -5.78078777e-02
-2.03955472e-01 -8.30353022e-01 3.80649924e-01 -1.32020012e-01
3.76808234e-02 3.00962985e-01 1.02168834e+00 -6.13603473e-01
-1.93703815e-01 -6.32904768e-01 -3.05888891e-01 2.82184184e-01
-2.64782727e-01 2.18727693e-01 -5.76378524e-01 -9.15538669e-02
2.64759362e-01 3.69466782e-01 6.88900948e-01 7.95626700e-01
8.51574302e-01 -5.28689958e-02 -3.93543094e-01 2.56806284e-01
1.27259374e+00 -8.64163488e-02 2.78443038e-01 2.96043232e-02
2.02840805e-01 5.74844539e-01 4.17686254e-01 8.39027464e-01
-1.70082748e-01 4.36051935e-01 -6.50110170e-02 6.54837191e-01
4.92183030e-01 -6.49438500e-02 3.12075526e-01 8.83049369e-01
4.46834490e-02 2.59930585e-02 -1.05432820e+00 3.08009654e-01
-1.70016122e+00 -9.39008176e-01 -4.19416398e-01 2.82931399e+00
6.31088316e-01 1.13635413e-01 1.42195091e-01 3.63263011e-01
6.87447727e-01 -3.73471320e-01 -3.20634693e-01 -2.09294230e-01
1.89624041e-01 3.54857266e-01 8.43650997e-01 7.77861536e-01
-9.61765468e-01 1.08538188e-01 7.22548532e+00 6.86147213e-01
-6.43483579e-01 -3.92498635e-02 1.85404629e-01 2.05698937e-01
-4.15953875e-01 9.18457955e-02 -8.01121175e-01 3.90582532e-01
1.23750377e+00 -5.73011279e-01 2.73235410e-01 8.48049223e-01
2.68156361e-02 -6.14097953e-01 -9.69780266e-01 9.04357016e-01
-1.99536934e-01 -8.77724409e-01 -4.80946004e-01 4.99655843e-01
7.24994004e-01 -3.58925790e-01 2.06095994e-01 3.19452770e-03
2.71275222e-01 -8.19463134e-01 5.18676877e-01 6.24697208e-01
8.71504009e-01 -1.06943357e+00 8.59320700e-01 1.62047401e-01
-9.90828276e-01 7.27069303e-02 -3.08986247e-01 7.82199204e-02
1.22123607e-01 1.52015448e+00 -9.12746251e-01 5.06368876e-01
2.23663703e-01 8.62619057e-02 -1.18108228e-01 1.06348300e+00
-2.06109345e-01 6.10991418e-01 -8.91065180e-01 8.25970694e-02
-2.19385311e-01 -7.35396624e-01 5.97721517e-01 8.70622754e-01
7.89977610e-01 1.10629037e-01 -1.38607964e-01 8.65153909e-01
1.19906679e-01 2.93766707e-01 -4.06937063e-01 -3.12150985e-01
4.17836726e-01 9.43117976e-01 -9.59941983e-01 -1.25132367e-01
-2.67872840e-01 4.16581333e-01 -1.53700396e-01 3.67537826e-01
-4.58794981e-01 -4.28086519e-01 6.66647315e-01 -5.76806143e-02
4.22909230e-01 -2.86105484e-01 -4.09524918e-01 -9.40015614e-01
1.81651056e-01 -6.77599728e-01 2.56680727e-01 -3.04454118e-01
-1.23537755e+00 3.31527948e-01 4.56775904e-01 -1.10326171e+00
-7.68242240e-01 -8.47112358e-01 -2.77796268e-01 1.07048082e+00
-6.74341619e-01 -4.80819017e-01 3.03220421e-01 5.09028196e-01
-6.22839689e-01 5.48844300e-02 8.46235514e-01 6.89688697e-02
-6.37665033e-01 3.46555889e-01 6.82030737e-01 8.92101228e-02
4.27843183e-01 -1.16825426e+00 -1.19976647e-01 1.08515298e+00
2.05523834e-01 1.00373316e+00 1.26717257e+00 -7.66163826e-01
-1.47815168e+00 -3.83020699e-01 5.22285521e-01 -4.91938859e-01
1.13861406e+00 -1.66503340e-01 -4.98156607e-01 6.34969056e-01
-5.46017289e-01 4.02713232e-02 1.03341317e+00 6.94435120e-01
-4.15094614e-01 6.57223761e-02 -1.05204904e+00 2.10148185e-01
3.96006554e-01 -5.10192275e-01 -3.57257396e-01 3.38964105e-01
3.62745136e-01 -3.00462872e-01 -1.00834441e+00 4.50698763e-01
9.52920079e-01 -1.05111885e+00 8.54953408e-01 -5.84857106e-01
1.86090857e-01 -4.26005214e-01 -5.59945226e-01 -1.15821385e+00
-6.66539446e-02 -5.23350298e-01 -6.36109635e-02 1.01099479e+00
4.32164043e-01 -5.41563451e-01 7.33721972e-01 9.24614727e-01
3.75549823e-01 -5.15458286e-01 -1.18092561e+00 -1.04575956e+00
9.48280934e-03 -9.10704970e-01 2.81502545e-01 7.23207057e-01
3.29186648e-01 1.82934254e-01 -1.56629220e-01 5.13058424e-01
8.93503964e-01 6.26909658e-02 5.51189959e-01 -1.44454849e+00
-5.53764105e-01 -3.66496295e-01 -6.78092599e-01 -5.65192640e-01
9.02022347e-02 -6.47080600e-01 1.80704072e-01 -1.11175823e+00
3.11225802e-01 -3.24671596e-01 -2.74009824e-01 -1.38353959e-01
1.29656032e-01 -3.38688269e-02 -2.10655816e-02 -8.85143727e-02
-2.34554991e-01 4.17786211e-01 6.21063173e-01 1.01203941e-01
-1.47205457e-01 3.45174432e-01 -6.14924431e-01 8.59711587e-01
4.14067537e-01 -6.97847784e-01 -4.06850338e-01 3.75341117e-01
6.87943220e-01 3.24029177e-01 9.19131115e-02 -1.00389874e+00
7.08061606e-02 1.37685537e-01 2.49672025e-01 -6.95305049e-01
-2.53883321e-02 -7.24148750e-01 2.82395035e-01 4.92474437e-01
-2.13853836e-01 -1.87897786e-01 1.61672279e-01 8.19054067e-01
1.08182225e-02 -6.48186028e-01 5.03994346e-01 3.59015554e-01
-6.79590479e-02 -7.30640516e-02 -4.13088620e-01 2.03878284e-02
7.28160083e-01 3.23845744e-01 5.81577383e-02 -5.80090106e-01
-7.62875557e-01 -9.20815095e-02 6.37533218e-02 -2.90364534e-01
4.27656770e-01 -1.30368805e+00 -5.69738209e-01 1.16566293e-01
-1.09542392e-01 -5.00096798e-01 3.70837092e-01 1.23287380e+00
-2.63765246e-01 5.28682768e-01 1.96432561e-01 -7.19629407e-01
-1.12654459e+00 6.88233733e-01 -3.33825722e-02 -1.82352617e-01
-4.84781973e-02 5.63496232e-01 2.16376767e-01 -2.45992735e-01
1.12146109e-01 -3.81042540e-01 1.77026525e-01 3.28408480e-02
7.82059669e-01 5.51508963e-01 7.52420798e-02 -4.21179563e-01
-5.48533380e-01 6.00695550e-01 2.96677440e-01 -8.02555025e-01
1.00332093e+00 -2.38673612e-01 -4.74501163e-01 6.21748090e-01
1.27690828e+00 5.63064873e-01 -5.57111740e-01 -1.76188380e-01
2.21392304e-01 -3.58901620e-01 9.90845338e-02 -5.03149033e-01
-5.55496752e-01 8.11430514e-01 6.62696064e-01 3.17773134e-01
8.50321054e-01 9.62009355e-02 1.02156520e-01 5.94444871e-01
2.46838659e-01 -7.94428527e-01 -5.39566040e-01 3.31727505e-01
7.64500916e-01 -9.81350660e-01 3.44392538e-01 -5.44014513e-01
-1.97082192e-01 1.10822546e+00 -1.92521453e-01 1.17408849e-01
1.16771889e+00 3.52949917e-01 -2.45631695e-01 -2.74729542e-02
-3.51452798e-01 -2.41298378e-01 5.60552120e-01 3.08721155e-01
5.09342730e-01 3.44514638e-01 -6.15768552e-01 7.57616460e-01
-5.06005883e-01 -1.39312342e-01 5.71623206e-01 3.83657098e-01
-3.93653959e-01 -9.59538817e-01 -7.61288822e-01 4.03633505e-01
-4.30161089e-01 -5.07554263e-02 -5.43464944e-02 6.28028691e-01
-4.22965676e-01 9.00505960e-01 -1.13171294e-01 -3.43163759e-01
1.21687315e-01 3.19832534e-01 5.64937055e-01 -3.33134174e-01
3.95748407e-01 2.30848134e-01 -1.28394678e-01 -4.08915877e-01
-3.25321734e-01 -8.46140027e-01 -9.95088398e-01 -4.52041507e-01
-4.86750692e-01 6.95561469e-01 1.00398648e+00 1.04604685e+00
5.89209646e-02 2.35686377e-01 4.74027306e-01 -4.40179288e-01
-1.01046968e+00 -8.86422575e-01 -1.10123742e+00 2.46705431e-02
7.50500932e-02 -6.68519080e-01 -6.60488486e-01 -3.59149754e-01] | [7.298923969268799, 4.2880988121032715] |
74e05fc4-8938-41af-bb17-8590b3fc9016 | the-influence-of-regional-pronunciation | null | null | https://aclanthology.org/2021.conll-1.52 | https://aclanthology.org/2021.conll-1.52.pdf | The Influence of Regional Pronunciation Variation on Children’s Spelling and the Potential Benefits of Accent Adapted Spellcheckers | A child who is unfamiliar with the correct spelling of a word often employs a “sound it out” approach: breaking the word down into its constituent sounds and then choosing letters to represent the identified sounds. This often results in a misspelling that is orthographically very different to the intended target. Recently, efforts have been made to develop phonetic based spellcheckers to tackle the more deviant nature of children’s misspellings. However, little work has been done to investigate the potential of spelling correction tools that incorporate regional pronunciation variation. If a child must first identify the sounds that make up a word, it stands to reason their pronunciation would influence this process. We investigate this hypothesis along with the feasibility and potential benefits of adapting spelling correction tools to more specific language variants - particularly Irish Accented English. We use misspelling data from schoolchildren across Ireland to adapt an existing English phonetic-based spellchecker and demonstrate improvements in performance. These results not only prompt consideration of language varieties in the development of spellcheckers but also contribute to existing literature on the role of regional accent in the acquisition of writing proficiency. | ['Julie Carson-Berndsen', 'Anthony Ventresque', 'Joe Kenny', 'Emma O’Neill'] | null | null | null | null | conll-emnlp-2021-11 | ['spelling-correction'] | ['natural-language-processing'] | [ 4.05868143e-01 -4.68187273e-01 2.11608693e-01 -4.26336616e-01
-8.38918149e-01 -7.79414356e-01 -8.95314366e-02 6.70941114e-01
-6.80169821e-01 3.22186917e-01 6.98951364e-01 -6.31775677e-01
-2.36613795e-01 -4.43680793e-01 -6.12513483e-01 -4.79307264e-01
7.75092006e-01 3.85885745e-01 4.54779297e-01 -2.03691378e-01
4.59600180e-01 3.34061056e-01 -1.73412490e+00 1.74661547e-01
1.05082715e+00 -2.88245082e-01 7.18105435e-01 7.63298571e-01
-3.36612016e-01 4.80045646e-01 -9.98219788e-01 -4.40901816e-01
-1.14090472e-01 -5.20212471e-01 -6.64497793e-01 -1.73074696e-02
7.51675487e-01 -1.61328211e-01 4.57956046e-01 1.12862527e+00
5.58865190e-01 1.52547136e-01 3.70223433e-01 -3.11996639e-01
-7.59290338e-01 8.47617984e-01 -1.12985782e-01 4.05301213e-01
5.42536378e-01 -5.64780496e-02 6.31075501e-01 -8.28903854e-01
2.20229894e-01 1.10713911e+00 7.44186044e-01 5.46801865e-01
-1.30312622e+00 -9.22383130e-01 -1.17850779e-02 -7.38796070e-02
-1.40620387e+00 -5.65716684e-01 4.14426684e-01 -6.49032593e-01
1.27707267e+00 2.43360907e-01 1.04822934e+00 4.72328722e-01
4.82507749e-03 2.74711609e-01 1.36832583e+00 -1.07012427e+00
-1.52119538e-02 1.89902142e-01 2.39713993e-02 -1.02119204e-02
4.27517503e-01 1.51034072e-01 -6.50487661e-01 4.43660408e-01
3.56247514e-01 -6.49596155e-01 -4.32890169e-02 1.56491905e-01
-9.03502882e-01 4.56302732e-01 -2.44605839e-01 6.31308258e-01
-3.70728225e-02 -2.25885808e-01 5.31520128e-01 3.48723829e-01
6.70396164e-02 7.38388360e-01 -4.78860468e-01 -7.78709531e-01
-1.15622580e+00 3.58708441e-01 4.62250084e-01 5.20447075e-01
9.97019261e-02 1.70571804e-01 2.28847131e-01 1.57233560e+00
3.45336586e-01 5.04173100e-01 5.41551709e-01 -6.33796453e-01
3.93153131e-01 2.92568058e-01 -7.66299739e-02 -6.46247208e-01
8.66488218e-02 -1.65751263e-01 1.33956373e-01 4.52716291e-01
6.84270501e-01 -3.26251425e-02 -8.63518000e-01 1.71275210e+00
2.61591166e-01 -2.45868489e-01 5.64482845e-02 4.36281025e-01
4.87122089e-01 5.68046153e-01 3.75828743e-01 -1.53798178e-01
1.06697655e+00 -6.98725820e-01 -5.17626703e-01 -2.50381976e-01
7.54994333e-01 -1.42076671e+00 1.27944863e+00 6.54003799e-01
-1.42034197e+00 -6.70072019e-01 -1.14168596e+00 1.46468326e-01
-3.52408558e-01 -3.37705672e-01 1.72966480e-01 1.39810109e+00
-1.05611825e+00 5.04795253e-01 -7.61398613e-01 -2.97145009e-01
-3.15593272e-01 4.37714010e-01 -4.51033503e-01 9.95928794e-02
-9.86121356e-01 1.13902473e+00 3.94681305e-01 -1.24213345e-01
-1.26082703e-01 -7.77909338e-01 -8.91405642e-01 -1.42262623e-01
-1.24492213e-01 2.02259943e-01 1.44070840e+00 -1.19486642e+00
-1.37326157e+00 9.71899807e-01 1.48975506e-01 1.48849159e-01
-6.41065359e-04 -1.44238248e-01 -9.03718829e-01 -1.30795389e-01
5.25505483e-01 3.38710219e-01 6.60267472e-01 -5.55611670e-01
-1.03565967e+00 -2.53601253e-01 -6.19116366e-01 4.03642207e-01
7.11193457e-02 8.76812279e-01 1.25468314e-01 -1.12589812e+00
1.18600398e-01 -7.75711715e-01 2.29649886e-01 -7.12557375e-01
1.75784305e-01 -3.68497998e-01 -1.65659748e-02 -9.37998831e-01
1.42343187e+00 -2.24942923e+00 -5.71198523e-01 3.23587626e-01
-5.02350509e-01 7.17944741e-01 -4.93237637e-02 5.57373643e-01
-2.68400580e-01 3.21569629e-02 1.49441689e-01 2.23490879e-01
3.40122171e-02 2.78754532e-01 1.29589006e-01 2.50858635e-01
1.96139589e-01 4.25668150e-01 -9.99646902e-01 -1.92509517e-01
3.43005866e-01 5.11032939e-01 -5.68132579e-01 -1.53373495e-01
2.36736640e-01 2.37235650e-01 3.28327090e-01 3.26493442e-01
4.65067416e-01 7.84769952e-01 1.79482549e-01 7.15662241e-01
-9.24994707e-01 1.37876427e+00 -1.29658496e+00 1.03367174e+00
-5.46951175e-01 6.16468966e-01 4.85546254e-02 -6.36122406e-01
7.93153882e-01 4.34615016e-01 -2.99829304e-01 -6.14574194e-01
-1.38381928e-01 9.44306076e-01 8.20710480e-01 -4.34683144e-01
6.36503756e-01 -4.86464560e-01 1.60571679e-01 2.13666752e-01
-2.10837945e-01 -4.64275628e-01 8.85630921e-02 -3.28769565e-01
7.11241186e-01 2.52408057e-01 2.97288001e-01 -7.00021803e-01
3.23323131e-01 7.94556290e-02 3.50540698e-01 7.51837671e-01
-3.21966014e-03 5.67597330e-01 -2.47764617e-01 1.18086971e-01
-7.87154138e-01 -1.18630779e+00 -3.61149669e-01 1.37984240e+00
-5.47901511e-01 -4.03726518e-01 -9.29932535e-01 -1.63600877e-01
-1.79574445e-01 1.27227557e+00 -1.55433372e-01 -2.21026987e-01
-8.12751174e-01 -1.25076443e-01 6.01306975e-01 7.26867914e-01
-1.66722298e-01 -1.36086452e+00 -6.95504010e-01 7.54839957e-01
2.70941228e-01 -5.09680510e-01 -7.20391512e-01 2.34234184e-01
-2.93885708e-01 -6.57237113e-01 -6.83484495e-01 -1.30556917e+00
5.14334738e-01 2.10115492e-01 7.26286054e-01 3.01782161e-01
1.46841660e-01 4.38357562e-01 -6.70556009e-01 -8.87553155e-01
-1.06450343e+00 -5.38624488e-02 1.48406953e-01 -7.19368637e-01
1.02096856e+00 -4.49720770e-01 -1.01169869e-01 -4.63170297e-02
-6.88743353e-01 -6.75486447e-03 3.58872205e-01 4.39760894e-01
4.56525743e-01 2.08144233e-01 6.91446304e-01 -6.76994085e-01
5.44768512e-01 -1.40293524e-01 -4.35444295e-01 4.09548655e-02
-4.72108215e-01 -1.54201731e-01 6.36345685e-01 -5.57270348e-01
-9.23666060e-01 -7.44109824e-02 -8.43182862e-01 5.13719261e-01
-7.73639381e-01 3.73593599e-01 -1.27846256e-01 -2.39945561e-01
5.15391469e-01 1.90714017e-01 -2.25188676e-02 -6.97668493e-01
-2.67613560e-01 8.50955963e-01 9.32834864e-01 -6.52484596e-01
6.14206314e-01 -5.28793991e-01 -4.51806515e-01 -1.16582227e+00
-4.49436307e-01 -7.05484271e-01 -5.80562830e-01 -8.66552964e-02
6.95642352e-01 -6.61266029e-01 -2.55498439e-01 8.49136412e-01
-7.98582137e-01 -4.12704736e-01 -3.14835370e-01 9.05565798e-01
-3.14316750e-01 1.74807087e-01 -1.52477667e-01 -5.42981088e-01
-6.95190132e-02 -1.28795600e+00 5.16958058e-01 3.43072802e-01
-8.57312441e-01 -8.36091757e-01 3.35197955e-01 2.90772498e-01
2.25752831e-01 -3.23872507e-01 1.24280214e+00 -9.49068010e-01
1.12366520e-01 -8.26710910e-02 2.91556090e-01 8.22722018e-01
4.46123809e-01 1.26120433e-01 -6.22906148e-01 -1.10053807e-01
-1.15947694e-01 2.16104444e-02 2.38906726e-01 1.11210361e-01
2.64131784e-01 -4.40650322e-02 2.52255082e-01 2.01287210e-01
1.15802920e+00 5.81247568e-01 3.03362876e-01 4.69159931e-01
6.47414148e-01 7.84716427e-01 5.35667539e-01 -2.38463849e-01
5.37540674e-01 7.61686742e-01 -5.14359355e-01 3.55571568e-01
-5.08098125e-01 -5.94349325e-01 6.79525495e-01 1.60395503e+00
1.39834374e-01 9.99789685e-02 -1.21362746e+00 1.30962133e+00
-8.39243591e-01 -7.39847302e-01 -3.76791835e-01 2.48491573e+00
1.14450252e+00 2.96087563e-01 2.92539328e-01 8.59983563e-01
8.22787583e-01 -1.61061183e-01 1.85152441e-01 -1.43798900e+00
1.86366603e-01 1.06509161e+00 4.45976764e-01 7.91902423e-01
-4.59037155e-01 1.34670007e+00 6.04703712e+00 7.37127900e-01
-1.09758830e+00 -2.59167105e-01 -2.39637420e-02 1.56179085e-01
-4.57786739e-01 -8.27926323e-02 -1.01582897e+00 6.31873310e-01
1.10313368e+00 -1.67521313e-02 3.80840421e-01 3.80684674e-01
4.65117961e-01 -4.32656616e-01 -9.20817077e-01 5.17221868e-01
7.15040192e-02 -7.25367725e-01 -2.87374496e-01 -3.57193768e-01
7.79697657e-01 -1.68185547e-01 -1.27131939e-01 1.73141286e-01
4.00971204e-01 -1.17420471e+00 1.26475251e+00 5.34463786e-02
7.02961624e-01 -8.95568728e-01 2.96733111e-01 5.13952561e-02
-9.74544406e-01 3.11308861e-01 -1.54613301e-01 -7.11229563e-01
-5.92428260e-02 -1.86338335e-01 -1.26900327e+00 -4.23435360e-01
6.05635583e-01 -5.11881039e-02 -6.69651687e-01 1.40078962e+00
-5.77983916e-01 1.34303105e+00 -5.16326010e-01 -2.98036158e-01
2.17166707e-01 1.11801259e-01 5.74810743e-01 1.34085870e+00
7.12330997e-01 1.51945660e-02 -3.37691814e-01 2.74678707e-01
6.69206619e-01 7.68907309e-01 -1.68939203e-01 -1.88338757e-01
9.56004918e-01 3.96571785e-01 -8.32834780e-01 3.00708134e-02
-8.08813453e-01 9.86468911e-01 1.67054191e-01 -4.45961533e-03
-5.93400188e-02 -4.13869917e-01 7.40783095e-01 4.08770859e-01
4.12647545e-01 -2.87174493e-01 -3.58895183e-01 -2.97325522e-01
-4.87302989e-02 -1.20831108e+00 7.09699020e-02 -6.09181821e-01
-8.19920778e-01 -2.96967346e-02 5.95805570e-02 -5.06101012e-01
-4.02782887e-01 -6.56920791e-01 -7.42900550e-01 1.44283271e+00
-8.33243489e-01 -6.43808007e-01 3.97867620e-01 6.19232245e-02
8.55360270e-01 5.37052304e-02 8.84293973e-01 4.39135469e-02
-7.50483051e-02 7.91661561e-01 1.34641230e-01 -1.87829062e-02
7.83362389e-01 -1.53249109e+00 7.93646157e-01 1.09589589e+00
3.58579069e-01 7.58755684e-01 9.14225996e-01 -9.31206346e-01
-4.97578502e-01 -6.70168519e-01 1.59757102e+00 -4.20452595e-01
6.04134977e-01 -1.16993479e-01 -1.02024639e+00 4.15821463e-01
2.25831345e-01 -7.58764803e-01 9.61478531e-01 -6.83738962e-02
5.18375486e-02 1.25504702e-01 -8.99989426e-01 6.59370840e-01
7.73056209e-01 -5.24975836e-01 -1.15176797e+00 -1.97465450e-01
4.70848083e-01 -3.20713550e-01 -6.92591250e-01 1.26466930e-01
7.33999252e-01 -9.71768796e-01 6.13141418e-01 -2.71487564e-01
1.25956669e-01 -2.80966580e-01 2.19648913e-01 -1.87871492e+00
-5.04083931e-01 -6.42989516e-01 9.93628800e-01 1.61286116e+00
4.18006957e-01 -3.76936972e-01 4.57535654e-01 6.62447751e-01
-4.55994040e-01 -1.88394681e-01 -7.12334454e-01 -6.70853019e-01
5.52759528e-01 -8.25168967e-01 7.92771161e-01 7.42351890e-01
1.87302977e-01 -1.64300784e-01 2.07076311e-01 2.39361599e-01
3.70653309e-02 -4.63395000e-01 1.80996940e-01 -1.11184013e+00
-3.53942424e-01 -6.35451198e-01 -5.02153456e-01 -4.23763007e-01
-1.05949081e-01 -8.25414956e-01 4.00126487e-01 -1.18364251e+00
-5.92990637e-01 -4.95579660e-01 -1.91451073e-01 4.26356107e-01
-4.37480122e-01 2.86687791e-01 4.62267995e-01 -2.97785759e-01
2.39128992e-02 -3.23581696e-01 7.35967219e-01 3.96497637e-01
-5.85745871e-01 4.08268809e-01 -8.94088209e-01 9.20871019e-01
9.94773865e-01 -7.11884081e-01 -4.22019631e-01 -5.67916095e-01
4.01727945e-01 -2.15036109e-01 4.90364172e-02 -1.13490987e+00
1.35237187e-01 -3.81930441e-01 3.09625626e-01 -3.86779994e-01
-3.89417619e-01 -6.96383715e-01 2.14948699e-01 2.45742127e-01
-3.87413651e-01 6.58061326e-01 6.02093101e-01 -5.41389287e-01
6.28237724e-02 -9.58448410e-01 9.57725883e-01 -1.74267679e-01
-7.58592367e-01 -4.48622644e-01 -1.12821245e+00 3.16583991e-01
6.99498713e-01 -6.84753180e-01 2.53979027e-01 -5.26598841e-02
-5.12859166e-01 -2.08265513e-01 8.75098348e-01 5.72501600e-01
2.37981066e-01 -1.18250573e+00 -6.59233034e-01 7.95722008e-01
-9.99524742e-02 -2.10697755e-01 6.78225830e-02 5.78895032e-01
-1.11099470e+00 4.83200282e-01 -2.78206468e-01 1.27735317e-01
-1.52018857e+00 2.57046342e-01 1.96825013e-01 5.09459555e-01
-5.68423450e-01 1.28606820e+00 -1.09306552e-01 -3.57944489e-01
2.13898897e-01 -3.47890824e-01 -2.66924948e-01 -5.62037081e-02
7.98528910e-01 6.08176172e-01 2.95218468e-01 -1.15243959e+00
-4.60166931e-01 5.11088550e-01 -2.86324799e-01 -4.43578154e-01
8.85403454e-01 -1.45058200e-01 -2.12520570e-03 7.00282037e-01
8.02702308e-01 1.10709572e+00 -8.81154895e-01 2.14740694e-01
1.48181051e-01 -6.44700885e-01 -1.21413901e-01 -1.04090869e+00
-3.48055780e-01 6.54108822e-01 3.56331646e-01 9.23982337e-02
9.24144387e-01 -1.49704143e-01 6.39931977e-01 -1.83082074e-01
-5.39440475e-02 -1.62438035e+00 -7.00443625e-01 6.85771823e-01
6.53843284e-01 -5.85777283e-01 -3.93247515e-01 -3.09829056e-01
-3.45248312e-01 1.08525717e+00 4.72615153e-01 -1.08548060e-01
3.02241206e-01 3.39571625e-01 4.45968837e-01 1.11601114e-01
-3.77913803e-01 -3.24654251e-01 3.41520935e-01 9.25306618e-01
8.56506348e-01 4.13776815e-01 -9.65987206e-01 5.81417203e-01
-1.15406466e+00 -1.82190120e-01 6.78524554e-01 7.67121017e-01
-6.03916407e-01 -1.53391421e+00 -9.04037714e-01 4.48795199e-01
-8.55933130e-01 -6.66529596e-01 -5.03356278e-01 7.47535348e-01
6.71116114e-01 9.52980101e-01 1.40012354e-01 -1.78124532e-01
3.90302092e-01 4.39571053e-01 7.05614865e-01 -1.18845332e+00
-9.18176472e-01 1.34141281e-01 1.73417121e-01 1.72227830e-01
-1.61479473e-01 -1.30892837e+00 -1.18316567e+00 -7.95007870e-02
-3.00035447e-01 2.89285481e-01 7.54215598e-01 1.22928858e+00
-4.56520706e-01 3.22409064e-01 -2.53631976e-02 -2.93177754e-01
-3.37283552e-01 -6.64205432e-01 -4.86902922e-01 4.47285958e-02
2.61722863e-01 -3.34485054e-01 -1.21299364e-01 -1.41722575e-01] | [11.081284523010254, 10.431405067443848] |
1da7515f-3dc6-48dc-9ead-07502fbcc057 | you-can-t-see-the-forest-for-its-trees | 2112.01955 | null | https://arxiv.org/abs/2112.01955v2 | https://arxiv.org/pdf/2112.01955v2.pdf | Revisiting Neuron Coverage for DNN Testing: A Layer-Wise and Distribution-Aware Criterion | Various deep neural network (DNN) coverage criteria have been proposed to assess DNN test inputs and steer input mutations. The coverage is characterized via neurons having certain outputs, or the discrepancy between neuron outputs. Nevertheless, recent research indicates that neuron coverage criteria show little correlation with test suite quality. In general, DNNs approximate distributions, by incorporating hierarchical layers, to make predictions for inputs. Thus, we champion to deduce DNN behaviors based on its approximated distributions from a layer perspective. A test suite should be assessed using its induced layer output distributions. Accordingly, to fully examine DNN behaviors, input mutation should be directed toward diversifying the approximated distributions. This paper summarizes eight design requirements for DNN coverage criteria, taking into account distribution properties and practical concerns. We then propose a new criterion, NeuraL Coverage (NLC), that satisfies all design requirements. NLC treats a single DNN layer as the basic computational unit (rather than a single neuron) and captures four critical properties of neuron output distributions. Thus, NLC accurately describes how DNNs comprehend inputs via approximated distributions. We demonstrate that NLC is significantly correlated with the diversity of a test suite across a number of tasks (classification and generation) and data formats (image and text). Its capacity to discover DNN prediction errors is promising. Test input mutation guided by NLC results in a greater quality and diversity of exposed erroneous behaviors. | ['Shuai Wang', 'Qi Pang', 'Yuanyuan Yuan'] | 2021-12-03 | null | null | null | null | ['dnn-testing'] | ['adversarial'] | [ 3.97707462e-01 -8.14431682e-02 -3.36479515e-01 -5.18026710e-01
-2.04892695e-01 -6.52382493e-01 2.16242105e-01 -1.54607296e-01
-4.10185680e-02 9.62392628e-01 -1.56197771e-01 -6.46413922e-01
-4.18460310e-01 -9.65204418e-01 -8.89444590e-01 -5.37135243e-01
3.28414947e-01 4.35223073e-01 2.21260920e-01 1.02029629e-01
1.94120735e-01 6.93639934e-01 -1.99382448e+00 7.34484553e-01
1.13919568e+00 1.33279705e+00 2.18242645e-01 4.62873131e-01
-2.24551216e-01 7.44392395e-01 -1.04136062e+00 -5.64435720e-01
1.35525197e-01 -7.80138850e-01 -6.27067029e-01 -1.82531685e-01
3.90204102e-01 -8.96685049e-02 -1.17142297e-01 1.56815898e+00
5.41692376e-01 -2.64213562e-01 8.40540588e-01 -1.54305279e+00
-7.86158204e-01 1.08582532e+00 1.48273692e-01 3.56347859e-01
-1.16343036e-01 5.13784051e-01 8.51719141e-01 -5.25052905e-01
3.73591095e-01 1.14705122e+00 6.93756640e-01 1.00316179e+00
-1.35006428e+00 -6.02761090e-01 -4.21602800e-02 1.54555542e-02
-1.12060630e+00 -1.50546774e-01 4.91829306e-01 -4.71606135e-01
1.01778305e+00 3.89888972e-01 8.60651433e-01 1.40692997e+00
4.17575777e-01 1.02964699e+00 6.95362508e-01 -2.54047334e-01
7.39512026e-01 1.36677250e-01 2.45440945e-01 2.49979809e-01
1.05282760e+00 1.39779121e-01 -3.88843417e-01 1.36797830e-01
6.51360095e-01 -1.29029348e-01 -3.07592183e-01 -3.88926640e-02
-7.99880147e-01 5.68156898e-01 1.36268198e-01 1.18272573e-01
-2.08194345e-01 3.33419859e-01 5.91353357e-01 5.52547634e-01
2.61788648e-02 8.70952606e-01 -5.94599187e-01 -4.40073162e-01
-7.73083568e-01 2.94124454e-01 7.31669605e-01 1.35894966e+00
5.35665751e-01 6.36964023e-01 -5.24393082e-01 8.51994693e-01
3.47338431e-02 5.17669857e-01 5.69232523e-01 -8.20247173e-01
4.30190682e-01 1.05139506e+00 -5.26185572e-01 -6.71760321e-01
-3.58620584e-01 -8.03334653e-01 -9.44478691e-01 4.17096168e-01
3.43898892e-01 -4.11514431e-01 -9.78349626e-01 1.91756654e+00
-3.19503903e-01 -3.12981069e-01 3.86812150e-01 5.40804207e-01
7.52392411e-01 4.05458182e-01 -3.19613993e-01 1.84250474e-01
8.96230578e-01 -4.07825470e-01 -5.24616241e-01 -2.76351482e-01
7.66355276e-01 -1.21673599e-01 1.24721062e+00 7.19389021e-01
-1.23896253e+00 -6.35158300e-01 -1.39816344e+00 4.56316173e-01
-3.95132780e-01 1.30243212e-01 2.02385962e-01 9.11980033e-01
-1.27343273e+00 6.76730216e-01 -5.57329893e-01 -1.57302320e-01
7.04310894e-01 5.48534095e-01 2.00664625e-01 -1.91749156e-01
-1.08984339e+00 8.47956181e-01 7.36806035e-01 6.43785596e-02
-1.17784512e+00 -7.86386132e-01 -6.71565890e-01 3.29334110e-01
-1.92951158e-01 -5.37015676e-01 1.19210982e+00 -1.06670558e+00
-1.11743999e+00 2.81635642e-01 3.55618596e-01 -6.17560923e-01
3.36328059e-01 2.81846225e-01 -7.73713887e-01 -1.77375361e-01
-2.95574486e-01 9.46740210e-01 2.35994041e-01 -1.23498726e+00
-7.60330558e-01 -2.95355618e-02 -7.91453943e-02 -1.48767978e-01
-4.63957131e-01 -4.87855971e-01 -2.09293559e-01 -6.11772299e-01
5.51845804e-02 -5.34874320e-01 2.41982210e-02 -1.24364831e-01
-7.79675484e-01 -2.36315325e-01 5.80570638e-01 7.24536851e-02
1.43366158e+00 -2.03975415e+00 -1.44017518e-01 6.05859697e-01
3.62658918e-01 3.27174932e-01 -3.67788464e-01 -7.97764435e-02
-2.89376769e-02 4.18794066e-01 -2.21046180e-01 2.24946141e-01
3.42208862e-01 3.21369439e-01 -1.70141205e-01 1.86283201e-01
5.11065245e-01 9.39263821e-01 -6.12870753e-01 -3.04464772e-02
-2.36429736e-01 2.05222949e-01 -8.59951496e-01 -4.79112826e-02
-5.62393665e-01 -1.77503854e-01 -2.50929654e-01 1.02831697e+00
6.74850285e-01 -5.56843542e-02 -6.52474090e-02 -1.51433989e-01
1.84379891e-01 2.12989688e-01 -6.66079938e-01 9.57952976e-01
-1.64250135e-02 1.11063063e+00 -6.42789423e-01 -9.91265535e-01
1.47340298e+00 6.61128163e-02 1.42656527e-02 -9.93304789e-01
2.33067080e-01 6.03192687e-01 7.97570586e-01 -4.51734692e-01
1.97440743e-01 2.44588807e-01 -1.25652939e-01 2.41927058e-01
3.70751143e-01 1.04271553e-01 4.35795814e-01 -4.29719627e-01
1.36545348e+00 -2.93838143e-01 -9.73691121e-02 -6.19766712e-01
1.22254547e-02 1.18688747e-01 7.43025899e-01 1.00663590e+00
-7.18562901e-02 6.33992791e-01 1.17519057e+00 -4.01452959e-01
-1.35388637e+00 -1.16463578e+00 -5.61838984e-01 5.21813810e-01
-3.92773747e-02 1.52454793e-01 -1.12368095e+00 -5.64271212e-01
2.11278778e-02 1.24950361e+00 -7.56015003e-01 -8.21523905e-01
-1.72124520e-01 -6.52718663e-01 1.34725189e+00 9.63903189e-01
5.63364029e-01 -1.36723423e+00 -8.33690464e-01 7.50594884e-02
3.22714716e-01 -7.32494771e-01 -1.33527413e-01 7.05720186e-01
-1.09198558e+00 -9.41306233e-01 -5.95577300e-01 -7.18947589e-01
7.89887130e-01 -4.53500271e-01 1.21627450e+00 1.07720859e-01
-1.17164209e-01 9.28166434e-02 -3.28147262e-01 -4.76051718e-01
-6.06144190e-01 1.79299131e-01 1.74661875e-01 -4.12579566e-01
7.64294922e-01 -6.11248016e-01 -2.56954193e-01 4.52845037e-01
-1.23149347e+00 -9.42478850e-02 9.73783672e-01 8.34869444e-01
6.63908184e-01 2.31687069e-01 8.09372187e-01 -8.55328560e-01
1.06035304e+00 -4.86529857e-01 -5.29758573e-01 5.60859978e-01
-7.98006058e-01 3.60914916e-01 8.20178866e-01 -8.20324183e-01
-6.10247612e-01 -4.39032316e-01 -2.00720891e-01 -5.53679109e-01
-3.83967936e-01 6.19642198e-01 -5.51739514e-01 2.30362371e-01
1.02754045e+00 4.92360473e-01 3.08514237e-02 -2.09709406e-01
-3.20854306e-01 4.01913226e-01 4.48706388e-01 -8.53718460e-01
1.82111725e-01 -2.47341603e-01 -1.63847804e-01 -3.70466530e-01
-3.57439905e-01 3.74852806e-01 -3.07373047e-01 -4.30256724e-01
4.95720267e-01 -3.83571059e-01 -8.64894092e-01 3.51342648e-01
-1.14320683e+00 -6.18758678e-01 -5.91515005e-01 5.06377995e-01
-4.71615553e-01 -4.05371964e-01 -4.96176749e-01 -7.24227130e-01
-1.76956609e-01 -1.55688906e+00 3.80730301e-01 1.77102104e-01
-4.33931291e-01 -8.80063355e-01 -2.38600209e-01 -4.66141343e-01
5.23752391e-01 3.12098503e-01 1.59564364e+00 -1.12212980e+00
-2.43387669e-01 -1.41223550e-01 -2.64578938e-01 7.93233275e-01
-1.64324760e-01 6.39516488e-02 -9.50644612e-01 -5.91975823e-02
-1.19775392e-01 -2.71803975e-01 8.59434664e-01 6.65913403e-01
1.49120998e+00 -5.41012228e-01 -1.87279493e-01 7.72958934e-01
1.62919879e+00 7.30388880e-01 1.03154576e+00 3.07448387e-01
5.23101807e-01 5.11750579e-01 1.27749273e-03 4.69673991e-01
-2.84022719e-01 -3.20360139e-02 6.54436588e-01 1.89594656e-01
-1.32311851e-01 -3.16466421e-01 6.05422735e-01 5.76198578e-01
1.35402218e-01 -1.04034686e+00 -1.13629961e+00 3.86410832e-01
-1.49799919e+00 -8.13707829e-01 -1.21344984e-01 1.96455312e+00
7.74076402e-01 7.00337946e-01 -5.50970770e-02 3.00560027e-01
7.81422079e-01 -4.59400713e-01 -1.34923732e+00 -6.59722388e-01
-6.71318829e-01 6.25317097e-02 4.66268271e-01 2.40429435e-02
-3.13664466e-01 4.42007005e-01 6.92367887e+00 9.05020833e-01
-9.81489420e-01 -3.75241250e-01 1.05278981e+00 -3.68150949e-01
-9.98745799e-01 -5.96959591e-01 -1.21053171e+00 6.68462992e-01
1.00672698e+00 -1.22089639e-01 3.32391739e-01 1.00827682e+00
-1.45641759e-01 2.48112187e-01 -1.64091408e+00 8.66177022e-01
-9.83316600e-02 -1.62230146e+00 7.41968036e-01 1.94145858e-01
1.09033728e+00 5.21949641e-02 4.45749760e-01 4.65657026e-01
4.47901458e-01 -1.36248159e+00 1.14312696e+00 7.78760672e-01
1.10507071e+00 -1.13114190e+00 1.12620533e+00 1.88850015e-01
-5.61927915e-01 -3.12583476e-01 -8.23117971e-01 1.66763902e-01
-5.19577742e-01 6.93359256e-01 -7.41049826e-01 -2.44248167e-01
8.48549247e-01 3.66827130e-01 -5.94770670e-01 1.23050821e+00
1.69069115e-02 7.30337560e-01 -8.54576528e-02 -6.28731191e-01
1.81632161e-01 1.11648448e-01 3.62441868e-01 1.32289064e+00
8.01887095e-01 -2.51544058e-01 -4.23881650e-01 1.66349876e+00
-2.91611433e-01 -3.68430227e-01 -7.20529020e-01 -1.12638988e-01
7.55138874e-01 5.52872181e-01 -7.79781103e-01 -1.60831198e-01
-1.47286087e-01 5.87602913e-01 1.27875015e-01 6.30694211e-01
-8.85005951e-01 -5.96066177e-01 1.05737412e+00 6.40829951e-02
2.61367649e-01 4.28316414e-01 -1.00924325e+00 -5.60018003e-01
9.58723873e-02 -1.11079955e+00 7.24092573e-02 -6.93613708e-01
-1.29491007e+00 1.10918343e+00 -7.49602392e-02 -1.44131100e+00
-2.18934968e-01 -1.12634730e+00 -6.70649230e-01 6.66967332e-01
-9.54091489e-01 -2.22990483e-01 -1.83611631e-01 3.83080184e-01
5.87503433e-01 -5.92449963e-01 5.98144233e-01 1.37258604e-01
-9.09612834e-01 1.19551039e+00 1.34164378e-01 9.45678875e-02
-3.73831972e-05 -1.10939336e+00 2.91143000e-01 9.32771802e-01
-3.03599209e-01 5.03862560e-01 6.17033064e-01 -6.62908614e-01
-1.09701836e+00 -1.40184772e+00 4.53859299e-01 -1.47151351e-01
1.07219286e-01 -4.65403348e-01 -9.48849857e-01 4.21712697e-01
8.99582077e-03 -2.45731205e-01 6.77767575e-01 -1.69875249e-01
-8.03665593e-02 -3.79536986e-01 -1.36774886e+00 8.40473771e-01
1.13103867e+00 -3.19623441e-01 -1.35914266e-01 -2.57519573e-01
8.54732931e-01 -2.71516114e-01 -5.93315601e-01 6.89161777e-01
3.84572119e-01 -1.16649902e+00 4.28700179e-01 -6.00218832e-01
6.24312878e-01 -2.51766413e-01 -4.99338895e-01 -1.27531826e+00
-3.22984278e-01 -2.11501002e-01 -1.65474489e-01 1.09921145e+00
1.22554195e+00 -5.62456548e-01 9.62328970e-01 8.19006145e-01
-7.56327152e-01 -1.24141157e+00 -6.61621213e-01 -1.01427042e+00
-5.97957475e-03 -9.80468214e-01 1.05326045e+00 5.70196450e-01
-1.47714512e-02 -1.07224964e-01 2.39328966e-01 8.64838585e-02
2.52267450e-01 -5.32482147e-01 1.65648371e-01 -1.28338706e+00
-1.31588310e-01 -1.15461016e+00 -7.21980035e-01 -1.06073582e+00
1.98186580e-02 -8.76923144e-01 1.38267368e-01 -1.24439061e+00
-4.47372906e-04 -3.41638654e-01 -4.31337565e-01 5.01031935e-01
3.73453200e-01 1.87950402e-01 -2.60232538e-01 1.48440143e-02
-2.60317177e-01 3.36416632e-01 1.11886668e+00 -3.99309367e-01
-1.26744643e-01 2.20271870e-01 -9.56273139e-01 7.21001089e-01
1.01146579e+00 -4.14567739e-01 -8.99849653e-01 -6.47019804e-01
5.89722216e-01 -2.91741431e-01 2.73613840e-01 -1.52574778e+00
4.36633199e-01 -2.37249266e-02 8.02188277e-01 -8.55773926e-01
-2.44935796e-01 -6.51880741e-01 1.71777576e-01 6.82085514e-01
-6.96824133e-01 4.38877404e-01 5.95346928e-01 2.21638083e-01
-2.28477985e-01 -5.94207287e-01 7.09024906e-01 2.55513251e-01
-7.21351504e-01 2.87930459e-01 -6.21880531e-01 2.27404192e-01
7.37661123e-01 -9.62421179e-01 -3.82113278e-01 3.82447499e-03
-4.19109464e-01 4.42195237e-02 2.31894746e-01 1.60012081e-01
1.09210658e+00 -1.55208588e+00 -6.24900281e-01 6.46540105e-01
1.51508749e-01 1.26513734e-01 7.83769265e-02 4.74118948e-01
-8.20294738e-01 4.64542866e-01 -4.64027166e-01 -8.10441852e-01
-6.28397107e-01 3.55724424e-01 7.07125366e-01 1.03816360e-01
-2.04657406e-01 1.13510621e+00 1.97398722e-01 -3.98022711e-01
6.60915852e-01 -1.07485497e+00 -6.16475679e-02 -1.59662992e-01
5.99600136e-01 4.52033907e-01 2.84489363e-01 2.55199760e-01
-1.45873547e-01 9.71400514e-02 2.35399511e-02 4.17031460e-02
1.22778809e+00 3.15635234e-01 1.26306802e-01 4.52147901e-01
1.15053642e+00 -5.96986413e-01 -1.55510175e+00 6.44785017e-02
9.37905312e-02 1.26760319e-01 -1.20832853e-01 -1.06846511e+00
-1.28065681e+00 7.74388075e-01 5.43459475e-01 3.05564553e-01
1.46031725e+00 -1.46759763e-01 4.32935745e-01 5.68821609e-01
1.60386264e-01 -1.01036680e+00 7.77282640e-02 7.15533316e-01
8.33092511e-01 -6.55580759e-01 -5.42407930e-01 1.97994009e-01
-6.64089143e-01 1.36392784e+00 1.22501242e+00 -2.10584849e-02
4.71445620e-01 7.58800387e-01 -2.60413617e-01 -1.73586309e-01
-1.09107172e+00 1.63862482e-01 3.34001452e-01 9.03946698e-01
1.69824541e-01 -2.80875620e-02 2.58452147e-01 9.63393688e-01
-3.96708131e-01 -1.56135961e-01 3.42860878e-01 4.28493649e-01
-6.57078326e-01 -7.10778296e-01 -1.12451144e-01 9.39469934e-01
-1.15479879e-01 -4.82896984e-01 -4.03933495e-01 7.73892164e-01
5.74830711e-01 5.87293327e-01 6.04918003e-01 -1.04295373e+00
4.55180287e-01 -3.92180644e-02 2.69884318e-01 -5.15761316e-01
-5.40166259e-01 -3.53302807e-01 -3.89109142e-02 -2.30503306e-01
2.59065539e-01 -4.83034253e-01 -1.22826421e+00 -6.30823612e-01
-3.09140980e-01 -5.00946753e-02 3.79298836e-01 8.52097988e-01
3.30026180e-01 9.60272968e-01 2.57293314e-01 -2.81606704e-01
-5.97798824e-01 -8.86389256e-01 -7.54271746e-01 1.00179007e-02
1.64240152e-01 -3.67956609e-01 -3.43688428e-01 -2.40149722e-01] | [6.566366195678711, 7.6336750984191895] |
23238c13-a436-4883-8505-caa8948986e0 | context-aware-bayesian-network-actor-critic | 2306.01920 | null | https://arxiv.org/abs/2306.01920v1 | https://arxiv.org/pdf/2306.01920v1.pdf | Context-Aware Bayesian Network Actor-Critic Methods for Cooperative Multi-Agent Reinforcement Learning | Executing actions in a correlated manner is a common strategy for human coordination that often leads to better cooperation, which is also potentially beneficial for cooperative multi-agent reinforcement learning (MARL). However, the recent success of MARL relies heavily on the convenient paradigm of purely decentralized execution, where there is no action correlation among agents for scalability considerations. In this work, we introduce a Bayesian network to inaugurate correlations between agents' action selections in their joint policy. Theoretically, we establish a theoretical justification for why action dependencies are beneficial by deriving the multi-agent policy gradient formula under such a Bayesian network joint policy and proving its global convergence to Nash equilibria under tabular softmax policy parameterization in cooperative Markov games. Further, by equipping existing MARL algorithms with a recent method of differentiable directed acyclic graphs (DAGs), we develop practical algorithms to learn the context-aware Bayesian network policies in scenarios with partial observability and various difficulty. We also dynamically decrease the sparsity of the learned DAG throughout the training process, which leads to weakly or even purely independent policies for decentralized execution. Empirical results on a range of MARL benchmarks show the benefits of our approach. | ['Qi Zhang', 'Dingyang Chen'] | 2023-06-02 | null | null | null | null | ['multi-agent-reinforcement-learning'] | ['methodology'] | [-2.74027348e-01 2.15037659e-01 -6.06170416e-01 -1.08744361e-01
-5.61897576e-01 -4.47129548e-01 5.46345413e-01 -1.23297714e-01
-3.82087022e-01 1.13126493e+00 4.10435855e-01 -4.40667957e-01
-6.49721384e-01 -5.46535909e-01 -6.55525327e-01 -9.20867085e-01
-6.41733527e-01 7.73044288e-01 3.01452667e-01 -2.47963935e-01
5.02740964e-02 2.69263446e-01 -9.76214170e-01 -1.80999622e-01
7.58371830e-01 5.35260737e-01 1.13924295e-01 8.59548569e-01
3.93268824e-01 1.58771408e+00 -5.77636480e-01 -3.03447485e-01
5.33434629e-01 -4.52840239e-01 -7.87508488e-01 3.25989664e-01
-2.82955736e-01 -8.16229463e-01 -4.52776670e-01 1.15406215e+00
2.92477012e-01 1.80637032e-01 5.58004677e-01 -1.54901040e+00
-1.29561394e-01 1.29939413e+00 -8.90685380e-01 1.74443990e-01
1.99221730e-01 3.52665693e-01 1.24807453e+00 2.28952542e-01
4.30107713e-01 1.51625466e+00 2.49088228e-01 4.79214907e-01
-1.13039112e+00 -3.55661601e-01 6.50635958e-01 3.48000228e-01
-9.41636503e-01 -3.01907212e-02 7.28563607e-01 -2.39555851e-01
8.41226816e-01 -4.96771708e-02 7.68560290e-01 1.08977282e+00
3.33054841e-01 1.07990074e+00 1.44012785e+00 -4.67067301e-01
4.15203184e-01 -3.38679552e-01 -1.04979612e-01 7.88501978e-01
4.10868704e-01 3.11707705e-01 -6.22256935e-01 -4.96420741e-01
8.81278396e-01 3.89291458e-02 6.18114062e-02 -5.84598422e-01
-1.04113412e+00 7.68051624e-01 -1.42354527e-02 4.24058002e-04
-6.72401011e-01 6.19688272e-01 3.39402914e-01 7.20327914e-01
1.33582741e-01 4.30975288e-01 -5.20976901e-01 -4.52977806e-01
-2.40637347e-01 3.69487494e-01 1.25641549e+00 7.77197838e-01
7.70436108e-01 1.48774877e-01 -2.25540668e-01 3.31629336e-01
4.51771855e-01 6.30835533e-01 -2.20996831e-02 -1.65391135e+00
4.83661592e-01 3.28800082e-01 5.63188493e-01 -8.91710520e-01
-4.07968760e-01 -2.67506897e-01 -7.32779562e-01 2.37343863e-01
6.44484699e-01 -6.11918271e-01 -2.26109236e-01 1.94625831e+00
3.70774209e-01 1.80016369e-01 2.06954211e-01 6.43009007e-01
-2.22846881e-01 3.82723927e-01 -1.50002018e-02 -5.77646375e-01
9.23509598e-01 -8.12110603e-01 -6.87660992e-01 -1.20521292e-01
7.04590380e-01 -2.54830182e-01 6.69249713e-01 4.62280154e-01
-8.38505387e-01 2.02781633e-01 -5.84176719e-01 7.88449585e-01
2.97408760e-01 -3.31671417e-01 1.12433267e+00 5.81768990e-01
-1.07130814e+00 5.40575385e-01 -1.25122714e+00 -6.33816347e-02
2.73360014e-01 5.42902708e-01 6.86237067e-02 1.31442070e-01
-1.08337927e+00 9.72050488e-01 4.26015079e-01 -6.45778030e-02
-1.48334289e+00 -2.74470985e-01 -4.67746109e-01 5.73766567e-02
1.31987214e+00 -6.07372880e-01 1.67017376e+00 -9.76864040e-01
-2.11176896e+00 -2.63882093e-02 3.28701109e-01 -6.83660328e-01
5.94712853e-01 -2.50327915e-01 1.64882541e-01 2.66100168e-01
1.29078761e-01 2.66164422e-01 8.30587029e-01 -1.21355081e+00
-8.91538680e-01 -8.64667073e-02 8.85744870e-01 5.69302201e-01
-3.10921699e-01 4.97500226e-03 5.81937283e-03 -1.43202171e-01
-5.99496245e-01 -1.06775749e+00 -8.95277321e-01 -6.79566979e-01
-2.67368883e-01 -5.59856296e-01 3.89683455e-01 -2.44579375e-01
1.01100862e+00 -1.79684961e+00 3.73254359e-01 5.29583156e-01
3.75383407e-01 -2.83178478e-01 -2.36779258e-01 7.28687882e-01
5.32452404e-01 -1.37787670e-01 1.85573757e-01 -2.82757636e-02
4.12703484e-01 5.81049562e-01 -3.19556743e-01 6.80477500e-01
-2.81891525e-01 7.72186697e-01 -1.29208636e+00 -4.65491295e-01
2.81934869e-02 -2.44470999e-01 -9.02444303e-01 2.84335196e-01
-5.31611860e-01 5.65377653e-01 -9.58267689e-01 5.08108199e-01
7.25117847e-02 -4.03392583e-01 1.07529533e+00 3.77732575e-01
7.32593536e-02 1.52402937e-01 -1.25793290e+00 1.26679611e+00
-2.25989699e-01 1.74382418e-01 4.29596990e-01 -1.17735660e+00
5.04322827e-01 3.85574132e-01 9.49955344e-01 -4.12438869e-01
2.11537674e-01 -3.64814587e-02 3.79340470e-01 -2.74214506e-01
2.87278384e-01 1.56543538e-01 -2.10415214e-01 8.26681674e-01
-7.23169670e-02 -1.85478665e-02 5.01015902e-01 4.57664549e-01
1.32969677e+00 2.84088135e-01 5.50705850e-01 -2.97656655e-01
1.71286359e-01 -5.65266572e-02 9.37218249e-01 1.46169066e+00
-3.80347252e-01 -5.95179737e-01 1.28150165e+00 -3.61776173e-01
-7.22878218e-01 -6.72243059e-01 5.89682281e-01 1.37558031e+00
1.50541037e-01 -5.54835260e-01 -5.84660769e-01 -8.60589921e-01
-2.60790568e-02 3.88742983e-01 -6.10654891e-01 6.49136901e-02
-5.98332226e-01 -8.41295004e-01 3.46987933e-01 4.31940585e-01
4.94792730e-01 -9.25346315e-01 -8.46959949e-01 4.16584045e-01
1.10164434e-01 -1.01778567e+00 -4.19277966e-01 3.85652810e-01
-6.98271632e-01 -1.40399694e+00 -3.57917339e-01 -2.57172018e-01
6.23070419e-01 2.68054277e-01 1.03143990e+00 6.36369875e-03
4.06279862e-01 1.06980014e+00 -3.04557204e-01 -2.50895470e-01
-5.92179537e-01 1.53335512e-01 4.14994866e-01 -1.68706000e-01
8.98774117e-02 -7.29421496e-01 -5.61317563e-01 3.74944836e-01
-6.38750553e-01 1.05838165e-01 6.78117394e-01 9.94928658e-01
1.23737887e-01 1.58163145e-01 6.06401861e-01 -8.78027380e-01
9.44060624e-01 -4.75814730e-01 -1.21661425e+00 4.41797823e-01
-5.27948558e-01 3.04444015e-01 7.33730197e-01 -5.22703588e-01
-1.13521969e+00 -7.76799768e-02 5.04152000e-01 -2.16648340e-01
9.41013470e-02 6.72580540e-01 7.24717975e-02 6.98012412e-02
4.69319463e-01 6.71385676e-02 2.26060599e-01 6.99749365e-02
5.29003382e-01 3.67427975e-01 3.17090675e-02 -1.29022110e+00
5.97226977e-01 4.07413930e-01 3.67882222e-01 -5.19583166e-01
-8.67437661e-01 -1.85689390e-01 -2.13016570e-01 -5.33050001e-01
3.87069851e-01 -8.13762426e-01 -1.68317056e+00 4.04171705e-01
-8.44331682e-01 -9.68203545e-01 -1.33970767e-01 7.10403204e-01
-9.76692259e-01 5.47677457e-01 -7.47998714e-01 -1.10653985e+00
2.24937856e-01 -1.27002358e+00 5.14986634e-01 3.01095694e-01
1.38253108e-01 -1.17796803e+00 3.93902034e-01 1.26938552e-01
2.90749222e-01 -1.71606570e-01 7.02680409e-01 -6.94848537e-01
-8.92566442e-01 2.52927631e-01 2.14744791e-01 1.38032377e-01
1.46574914e-01 -4.41797040e-02 -2.61383146e-01 -5.72049320e-01
-2.26122469e-01 -6.17148161e-01 3.43772143e-01 5.82648933e-01
8.12694073e-01 -6.54712617e-01 -3.15888643e-01 3.72687057e-02
1.09272635e+00 2.21441582e-01 1.35164380e-01 4.24966931e-01
4.86696482e-01 5.19617319e-01 6.41373575e-01 1.00845695e+00
6.83558941e-01 5.56444764e-01 6.19404972e-01 4.41781551e-01
4.63165909e-01 -2.67817855e-01 9.06445622e-01 8.07481468e-01
-4.59229499e-01 -1.49280531e-02 -7.34213531e-01 2.45691493e-01
-2.37191033e+00 -1.11528313e+00 3.10789049e-01 2.08304095e+00
1.14229488e+00 5.74795417e-02 5.29171646e-01 -3.86061162e-01
4.70247924e-01 1.41340241e-01 -7.04540491e-01 -2.31268302e-01
6.95827305e-02 -1.65163159e-01 9.31757331e-01 6.24244511e-01
-9.77335215e-01 1.02488136e+00 6.76914978e+00 8.66994202e-01
-5.96170723e-01 1.47443548e-01 4.76289779e-01 -1.41252493e-02
-4.79037166e-02 2.98907191e-01 -8.87539983e-01 1.95400894e-01
9.41434205e-01 -2.64861166e-01 1.00673127e+00 1.05514574e+00
5.12609422e-01 -4.39922690e-01 -1.01131630e+00 6.05239332e-01
-3.70735556e-01 -1.20991814e+00 -4.30347115e-01 3.14127743e-01
1.12280381e+00 5.40899523e-02 -2.25506678e-01 4.68894869e-01
1.67401361e+00 -6.28496885e-01 4.75999236e-01 2.77910471e-01
1.59747481e-01 -9.58590031e-01 4.99374986e-01 6.68035388e-01
-9.07734275e-01 -5.17499804e-01 -3.35068911e-01 -4.28829551e-01
-9.92210465e-04 1.09293312e-01 -8.61461163e-01 6.34212971e-01
5.19183457e-01 8.01608145e-01 -1.19023375e-01 7.28251398e-01
-6.65030956e-01 7.62777925e-01 -4.70746338e-01 -3.06807071e-01
5.46206951e-01 -4.17861670e-01 6.58135533e-01 6.80263937e-01
-2.39652097e-01 1.22520857e-01 8.53662729e-01 3.82289350e-01
2.27505956e-02 -1.14721939e-01 -5.13177514e-01 -1.83166847e-01
5.37937760e-01 1.07944310e+00 -7.87037790e-01 -3.03525269e-01
-5.30370772e-01 3.98229927e-01 6.69696927e-01 6.23524487e-01
-7.82657087e-01 2.54916847e-01 7.08293736e-01 -4.28910226e-01
3.01568329e-01 -5.18648863e-01 1.87251642e-01 -1.20320785e+00
-1.69561073e-01 -1.35860753e+00 5.07535398e-01 -9.94242504e-02
-1.34532440e+00 9.55898389e-02 2.79795319e-01 -9.89684403e-01
-8.61637235e-01 -4.37800705e-01 -4.79037791e-01 2.59390682e-01
-1.43053079e+00 -9.24297690e-01 3.07820648e-01 8.13566446e-01
2.18144119e-01 -2.91070282e-01 6.88766956e-01 -9.80798900e-02
-7.13368297e-01 1.85680553e-01 3.02538723e-01 2.18586028e-02
5.70004165e-01 -1.51933670e+00 -3.73346627e-01 8.35171163e-01
9.49387066e-03 3.95560861e-01 8.30099642e-01 -7.80876458e-01
-1.77280080e+00 -7.07599699e-01 -6.68874308e-02 -1.22986771e-01
1.22284400e+00 -6.27462752e-03 -3.93803984e-01 9.03144538e-01
5.75318694e-01 -3.30173463e-01 5.57581365e-01 3.73477340e-01
-9.63288173e-02 -2.01341420e-01 -5.69995761e-01 9.84055698e-01
8.05170357e-01 -1.93133906e-01 -4.17269498e-01 6.18455112e-01
5.08711159e-01 -3.42803985e-01 -8.00861597e-01 2.47178487e-02
3.13164711e-01 -8.15967441e-01 5.40934086e-01 -1.02439272e+00
1.66185305e-01 -3.50727767e-01 -2.05127612e-01 -1.52428937e+00
-1.62408903e-01 -1.34277391e+00 -4.68012571e-01 7.06793189e-01
9.14420113e-02 -7.89925873e-01 7.99463272e-01 5.69238544e-01
6.83295205e-02 -6.18334770e-01 -8.88298750e-01 -1.01607239e+00
-4.29130206e-03 -2.35146120e-01 3.36436629e-01 7.33754218e-01
2.85555333e-01 5.22257499e-02 -9.89554524e-01 3.44064176e-01
8.80859077e-01 1.20736159e-01 1.05702245e+00 -8.54911864e-01
-9.64603782e-01 -4.84750211e-01 2.85028480e-02 -1.47505426e+00
6.03322625e-01 -5.32872975e-01 1.42834499e-01 -1.31333315e+00
4.82157320e-01 -4.60134447e-01 -2.53384024e-01 5.55402815e-01
-7.25108087e-02 -5.76008141e-01 3.07261527e-01 3.09271455e-01
-1.43710768e+00 6.89247906e-01 1.35069036e+00 -1.36064300e-02
-3.43310237e-01 2.36460403e-01 -6.86691761e-01 8.41634274e-01
8.37236166e-01 -6.50641024e-01 -5.29844999e-01 -1.55099750e-01
5.24838984e-01 4.88314956e-01 1.87860206e-01 -4.17016238e-01
4.56864655e-01 -9.32851136e-01 -5.01846135e-01 -2.56198883e-01
7.72564635e-02 -7.78423667e-01 3.38335112e-02 7.76536644e-01
-5.23096263e-01 -5.61946817e-02 -3.70688379e-01 9.67879832e-01
5.74732870e-02 -2.54568994e-01 3.88888001e-01 -1.54067531e-01
-5.92266023e-01 3.30684334e-01 -8.27016652e-01 2.29618341e-01
1.14352548e+00 3.78050447e-01 -3.17610323e-01 -9.08376932e-01
-6.11530125e-01 8.69907498e-01 2.60208517e-01 -9.36068445e-02
2.23149240e-01 -9.61880505e-01 -6.24144435e-01 -4.34918970e-01
-4.18235511e-01 -2.75816172e-01 1.25095516e-01 1.19552898e+00
-1.38280720e-01 3.15246254e-01 -2.00666845e-01 -4.64910984e-01
-1.01308393e+00 3.51334661e-01 4.89228278e-01 -8.71846020e-01
-4.02719378e-01 2.26999402e-01 -1.30246654e-01 -1.59654617e-01
3.55558008e-01 -9.45349336e-02 -8.56242701e-02 -2.05583200e-01
3.95751297e-01 4.15830046e-01 -5.28390229e-01 -2.45181639e-02
-1.18258342e-01 -9.90018770e-02 -2.90082186e-01 -3.27889591e-01
1.43000984e+00 -2.37767696e-01 -2.58227974e-01 2.57737815e-01
1.00497380e-01 -5.14817536e-02 -1.84117675e+00 -5.30422986e-01
2.03941122e-01 -4.68195021e-01 -9.86569375e-02 -4.53242362e-01
-1.11847401e+00 2.10158944e-01 -1.88495472e-01 5.73777020e-01
7.03619659e-01 4.92604449e-02 4.66662906e-02 1.13456011e+00
9.23392177e-01 -1.30284238e+00 2.85233438e-01 8.13313007e-01
5.09293079e-01 -1.10305250e+00 1.53432056e-01 -1.03331819e-01
-9.49827015e-01 1.05785477e+00 5.57705402e-01 -1.98597118e-01
6.38674617e-01 3.97306979e-01 -2.36339360e-01 -1.54609829e-01
-1.17847383e+00 -2.80804425e-01 -4.84982818e-01 7.11128414e-01
-6.29034266e-02 4.63265181e-01 -2.05689073e-01 4.98838454e-01
1.77273154e-01 -6.80350065e-02 9.03944850e-01 1.16081500e+00
-4.01409298e-01 -1.30806017e+00 -2.43947998e-01 5.17273366e-01
-4.77866530e-01 1.28882393e-01 -7.00045601e-02 7.84851849e-01
-6.79142118e-01 1.06407011e+00 -3.11576575e-01 3.48749869e-02
-4.50066067e-02 -3.78164828e-01 6.70060515e-01 -4.48124409e-01
-4.66017574e-01 3.04984659e-01 1.45559326e-01 -7.38402784e-01
-8.29938710e-01 -7.07382083e-01 -9.88624513e-01 -5.55745065e-01
-3.81821282e-02 2.89806783e-01 6.30746186e-02 1.19745040e+00
2.15703785e-01 4.58913416e-01 8.14878702e-01 -6.90204084e-01
-1.38158870e+00 -8.65139484e-01 -7.20839381e-01 -4.40925546e-02
5.23657762e-02 -9.40295100e-01 -3.53054672e-01 -2.95448184e-01] | [3.782569646835327, 2.0672452449798584] |
71daf80c-c6b4-48e7-8a6f-18a9214e2bd6 | spot-spatiotemporal-modeling-for-3d-object | 2207.05856 | null | https://arxiv.org/abs/2207.05856v1 | https://arxiv.org/pdf/2207.05856v1.pdf | SpOT: Spatiotemporal Modeling for 3D Object Tracking | 3D multi-object tracking aims to uniquely and consistently identify all mobile entities through time. Despite the rich spatiotemporal information available in this setting, current 3D tracking methods primarily rely on abstracted information and limited history, e.g. single-frame object bounding boxes. In this work, we develop a holistic representation of traffic scenes that leverages both spatial and temporal information of the actors in the scene. Specifically, we reformulate tracking as a spatiotemporal problem by representing tracked objects as sequences of time-stamped points and bounding boxes over a long temporal history. At each timestamp, we improve the location and motion estimates of our tracked objects through learned refinement over the full sequence of object history. By considering time and space jointly, our representation naturally encodes fundamental physical priors such as object permanence and consistency across time. Our spatiotemporal tracking framework achieves state-of-the-art performance on the Waymo and nuScenes benchmarks. | ['Leonidas J Guibas', 'Yanchao Yang', 'Vitor Guizilini', 'Sergey Zakharov', 'Rares Ambrus', 'Jie Li', 'Davis Rempe', 'Colton Stearns'] | 2022-07-12 | null | null | null | null | ['3d-object-tracking', '3d-multi-object-tracking'] | ['computer-vision', 'computer-vision'] | [-3.36304188e-01 -7.78343678e-01 -4.92466390e-01 -1.97992325e-02
-6.11576974e-01 -9.08812106e-01 8.67936313e-01 1.74979955e-01
-3.53490412e-01 5.19026995e-01 2.87854016e-01 1.07834348e-02
-1.59193560e-01 -5.96648037e-01 -8.82238686e-01 -3.36575955e-01
-4.45588827e-01 4.46578205e-01 8.27688634e-01 2.53695875e-01
-1.60960555e-01 8.85133803e-01 -1.49855733e+00 -1.06762402e-01
3.41322780e-01 1.03812528e+00 -2.66656396e-03 9.82303202e-01
1.10418126e-01 8.00545275e-01 -3.19755703e-01 -1.38833493e-01
3.79816651e-01 1.00249518e-02 -3.65835398e-01 1.18260235e-01
9.10049021e-01 -5.11364877e-01 -1.10359180e+00 7.26979494e-01
1.07254811e-01 4.40527350e-01 3.06607932e-01 -1.53821957e+00
-6.01788938e-01 7.08594397e-02 -8.05981636e-01 6.69006348e-01
3.42701137e-01 3.31062615e-01 9.29420352e-01 -6.28111601e-01
9.50024366e-01 1.19139743e+00 8.64594817e-01 4.71360117e-01
-1.07832265e+00 -6.11364007e-01 7.64424443e-01 1.24068558e-01
-1.51210785e+00 -4.22607839e-01 5.65234363e-01 -7.41260231e-01
8.17979515e-01 2.21313149e-01 8.36683154e-01 9.68023956e-01
1.67962819e-01 7.92132616e-01 3.84082526e-01 1.87176183e-01
-1.41036421e-01 -4.28835630e-01 2.00935036e-01 6.99528754e-01
4.71912324e-01 2.48678416e-01 -7.29538262e-01 -2.96069711e-01
7.98793435e-01 6.14122987e-01 1.52334750e-01 -6.20159924e-01
-1.45909083e+00 3.12880546e-01 3.93682063e-01 7.60810599e-02
-3.56148869e-01 7.64805079e-01 2.23736376e-01 -3.57348025e-01
6.21561408e-01 -4.16243076e-01 -3.42731744e-01 -2.56763607e-01
-1.12481081e+00 7.43522942e-01 1.99895442e-01 1.51433706e+00
7.39691615e-01 -1.44664377e-01 -6.52066827e-01 4.85962741e-02
3.94233257e-01 1.07065034e+00 -3.98248404e-01 -1.14279497e+00
4.85002637e-01 5.10493517e-01 6.78572297e-01 -9.41398025e-01
-2.09424108e-01 -1.09165967e-01 -4.71984208e-01 9.48450193e-02
6.40418172e-01 -7.17968866e-02 -1.04228485e+00 2.03626490e+00
8.53302062e-01 1.16119707e+00 -5.58845997e-01 8.86457384e-01
5.67843199e-01 6.56109333e-01 3.49168897e-01 -2.99060941e-02
1.44418395e+00 -8.42520833e-01 -6.50443196e-01 -3.49161923e-02
3.35518450e-01 -2.94521987e-01 3.35903198e-01 -1.93365395e-01
-1.25454152e+00 -6.41218603e-01 -5.80655575e-01 -6.60477802e-02
-3.70886773e-01 -1.56565920e-01 5.40152669e-01 5.37665606e-01
-8.31412256e-01 3.58771086e-01 -1.42672276e+00 -2.50304043e-01
6.41671419e-01 3.08521539e-01 -1.88136727e-01 -8.01302865e-02
-8.68397176e-01 6.32387757e-01 6.46672919e-02 -2.11733058e-02
-1.14932930e+00 -1.19505286e+00 -8.24654877e-01 -1.73340961e-01
4.96053249e-01 -9.22745824e-01 1.22226346e+00 5.76778948e-02
-8.57822359e-01 8.29104960e-01 -7.23275244e-01 -6.42438591e-01
7.50036061e-01 -5.63645482e-01 -6.68428659e-01 7.23338723e-02
3.01997840e-01 3.86461049e-01 4.94135767e-01 -1.23006809e+00
-1.06183207e+00 -3.91349256e-01 1.53598383e-01 -7.95136169e-02
7.00654387e-02 1.80512577e-01 -1.14903688e+00 -6.58137679e-01
-1.42878488e-01 -1.19197404e+00 -8.64634588e-02 5.93395114e-01
-2.88753808e-01 -1.50059924e-01 1.11221063e+00 -4.65243280e-01
1.34978676e+00 -2.01301837e+00 1.65289029e-01 -1.25809293e-02
5.64653933e-01 -1.10754535e-01 1.75406858e-01 6.47941977e-02
4.52125341e-01 2.96136104e-02 1.82234541e-01 -7.52377927e-01
2.35656783e-01 1.58147216e-01 -3.11427116e-01 8.70204329e-01
9.04698223e-02 1.13144708e+00 -1.25025475e+00 -7.58177876e-01
4.46301639e-01 8.25690508e-01 -5.35581231e-01 -6.42248243e-02
-5.52414238e-01 7.34309912e-01 -7.70239770e-01 8.51203322e-01
7.89762259e-01 -6.27843022e-01 -2.28251472e-01 -1.07875496e-01
-4.59515095e-01 8.32060650e-02 -1.25502586e+00 2.03954124e+00
-1.05447352e-01 5.23911119e-01 -5.82466014e-02 -7.65272006e-02
3.20121199e-01 2.91006982e-01 1.35314953e+00 -5.18928945e-01
-2.04584911e-01 -2.98103511e-01 -6.30632997e-01 -2.47800022e-01
9.18383241e-01 1.95873708e-01 -2.60883719e-01 4.84861493e-01
-1.87810987e-01 5.65576255e-01 2.14369461e-01 3.49702567e-01
1.37075627e+00 6.57226741e-01 2.67850924e-02 2.37212762e-01
2.10389853e-01 2.33427867e-01 8.33265662e-01 1.00413573e+00
-6.02461278e-01 3.99694920e-01 -7.00963736e-02 -4.32565302e-01
-1.13522947e+00 -1.42801905e+00 -8.44124181e-04 1.16294718e+00
6.65803075e-01 -5.83276749e-01 -7.00576901e-02 -6.23548210e-01
4.43919837e-01 1.23901673e-01 -8.00832331e-01 1.80621117e-01
-1.09088266e+00 -3.67062420e-01 5.00840425e-01 7.34783292e-01
2.63714582e-01 -4.35884684e-01 -8.89818549e-01 3.54175717e-01
-4.98284139e-02 -1.53428721e+00 -9.88169551e-01 -4.65979427e-01
-7.49047935e-01 -9.57942545e-01 -7.21258044e-01 -1.55212298e-01
2.36676753e-01 6.84961021e-01 1.22462165e+00 1.06127262e-01
-2.62390494e-01 6.90504730e-01 -1.48771316e-01 -2.36985266e-01
1.59043148e-01 -2.26426452e-01 2.12651089e-01 1.66830599e-01
1.98770657e-01 -3.40068758e-01 -8.45460832e-01 3.76197368e-01
-3.74832511e-01 2.90296879e-02 6.46662787e-02 1.09421380e-01
9.77021575e-01 -6.79552630e-02 -9.90663618e-02 -5.24584472e-01
-2.74238259e-01 -7.82018006e-01 -9.37232852e-01 3.85585606e-01
8.95585567e-02 -2.83799022e-01 4.88996245e-02 -7.47511566e-01
-1.01674712e+00 1.78004712e-01 4.38237309e-01 -9.42624152e-01
-1.52243137e-01 -3.82823870e-02 -2.02444553e-01 3.03521194e-03
1.96902722e-01 -1.50223440e-02 -5.93117356e-01 -5.62827766e-01
5.68541169e-01 -1.48919374e-01 7.94432104e-01 -8.32921743e-01
1.09854579e+00 1.18927407e+00 1.87812567e-01 -4.46789622e-01
-9.77148712e-01 -8.69150102e-01 -8.43759716e-01 -6.39771223e-01
1.06236410e+00 -1.18546069e+00 -1.18166614e+00 3.65959793e-01
-1.19203508e+00 -3.98725659e-01 -2.45125383e-01 5.06146252e-01
-3.68951768e-01 1.56321153e-01 -4.61041957e-01 -9.11999702e-01
1.88904017e-01 -8.05983901e-01 1.65497077e+00 2.21084028e-01
-8.15332383e-02 -1.18534243e+00 3.74015689e-01 -1.24703333e-01
1.15533806e-01 7.78239846e-01 2.93758631e-01 -4.56314534e-03
-1.36546040e+00 -3.06013525e-01 -2.77983636e-01 -6.04140639e-01
1.82169497e-01 3.51789631e-02 -8.24446321e-01 -2.13735446e-01
-4.01393622e-01 4.12852377e-01 8.60892713e-01 7.01249003e-01
1.11251247e+00 -1.51559591e-01 -1.09376585e+00 7.38202274e-01
1.46134913e+00 8.86036176e-03 5.75554855e-02 1.62412271e-01
1.05086744e+00 1.92981794e-01 7.66415417e-01 6.40780091e-01
6.50041103e-01 9.87143517e-01 4.55315858e-01 3.42874318e-01
-4.01695460e-01 -3.70311350e-01 1.38511464e-01 2.17146069e-01
-3.14332724e-01 -4.85365421e-01 -1.01283538e+00 8.48489523e-01
-2.16144443e+00 -1.57881987e+00 -2.42276460e-01 2.05846047e+00
3.36277872e-01 2.55308002e-01 6.66682661e-01 -4.25000638e-01
9.40900326e-01 5.55305004e-01 -9.44098353e-01 8.21406543e-01
-1.18155360e-01 -2.85380572e-01 9.55523670e-01 4.49881285e-01
-1.39032030e+00 7.96193719e-01 6.59375143e+00 3.65853935e-01
-7.14122236e-01 3.53214324e-01 -9.09268484e-02 -6.58357501e-01
-1.53554648e-01 -1.21700214e-02 -1.41411173e+00 6.68107450e-01
9.31726635e-01 -2.19276026e-01 1.62477672e-01 5.09553850e-01
2.72391170e-01 2.44894382e-02 -1.12899804e+00 8.49147677e-01
-2.99706697e-01 -1.72821534e+00 -1.60056174e-01 3.41028482e-01
8.36044014e-01 4.04311240e-01 1.53283954e-01 7.72751570e-02
5.73121727e-01 -6.01967812e-01 1.31371212e+00 8.24862659e-01
6.55472100e-01 -2.66972542e-01 -5.45605123e-02 1.65303648e-01
-2.09916306e+00 3.36997919e-02 1.48381069e-01 9.61170420e-02
7.54240632e-01 3.24392319e-01 -1.96287110e-01 6.18893802e-01
8.42131913e-01 1.31686103e+00 -4.27976936e-01 1.41986716e+00
2.93635398e-01 3.82288873e-01 -7.90072858e-01 3.21264058e-01
2.40755945e-01 1.70680899e-02 7.91949451e-01 1.20837629e+00
3.72828752e-01 2.04613060e-01 4.92587745e-01 7.69328058e-01
-1.00221440e-01 -7.81614721e-01 -6.39091671e-01 2.06388161e-01
8.38280976e-01 9.03641164e-01 -6.97935283e-01 -4.69174296e-01
-8.16164255e-01 5.08400977e-01 1.65972650e-01 4.44325298e-01
-1.45132148e+00 3.26294214e-01 1.30764759e+00 2.78021365e-01
7.66129434e-01 -8.49442601e-01 -2.34138206e-01 -1.25827765e+00
1.29622266e-01 3.97082418e-02 6.34820938e-01 -6.76090360e-01
-1.07971096e+00 1.33294240e-01 3.24290901e-01 -1.43581486e+00
1.56347364e-01 -3.69946003e-01 -4.74247992e-01 7.36484528e-01
-1.62650502e+00 -1.63248444e+00 -4.67701375e-01 8.61994922e-01
3.53684694e-01 3.09266061e-01 2.41950154e-01 6.94343030e-01
-4.92019564e-01 1.78656682e-01 1.68754160e-02 2.42150739e-01
4.56909627e-01 -1.07258356e+00 8.55449021e-01 1.05027246e+00
2.58740723e-01 7.12083042e-01 6.59179688e-01 -1.13968468e+00
-1.75398135e+00 -1.48921132e+00 7.11089909e-01 -1.26456976e+00
9.14385617e-01 -4.97998148e-01 -8.27431560e-01 1.21083236e+00
-3.63659829e-01 7.28070080e-01 3.23814481e-01 1.28207207e-01
-4.10803884e-01 9.81942415e-02 -8.41400981e-01 5.35923004e-01
1.67045248e+00 -6.88006222e-01 -3.16402465e-01 2.11863264e-01
9.81348515e-01 -9.30517197e-01 -1.05384815e+00 1.92240581e-01
7.96372116e-01 -3.70144755e-01 1.55817127e+00 -7.98776567e-01
-2.72680193e-01 -7.82168806e-01 -1.84562564e-01 -5.01986027e-01
-4.95822877e-01 -7.17532635e-01 -1.00980675e+00 1.16860223e+00
-7.04674348e-02 3.62890512e-02 1.05935585e+00 7.92911053e-01
-7.77121931e-02 -3.71377796e-01 -9.68493521e-01 -1.02006161e+00
-3.40295523e-01 -7.35820174e-01 9.66852307e-01 6.92716122e-01
-6.31527424e-01 -3.36762607e-01 -6.82322919e-01 7.34161377e-01
1.05968058e+00 4.74965841e-01 9.27615106e-01 -1.35579836e+00
4.32642736e-02 -2.97370046e-01 -4.69185084e-01 -1.49015963e+00
2.66977042e-01 -6.20609581e-01 1.10001281e-01 -1.27313721e+00
2.95713693e-01 -7.83433735e-01 -3.21811706e-01 2.03551739e-01
-2.61856079e-01 1.98026523e-01 3.41537416e-01 3.12213600e-01
-1.37942624e+00 3.22771043e-01 1.13728869e+00 -1.52907133e-01
-1.32833928e-01 2.37079374e-02 -1.38716727e-01 5.45229137e-01
3.61306638e-01 -7.67773390e-01 -1.00165427e-01 -8.21744561e-01
1.92166492e-02 2.51277775e-01 9.86357749e-01 -1.03224134e+00
6.88332617e-01 -5.56347668e-01 4.00066346e-01 -1.35919273e+00
8.55056167e-01 -1.12737823e+00 6.19406641e-01 1.96715251e-01
-9.55289528e-02 2.72126943e-01 4.25040007e-01 1.27147305e+00
2.47252360e-01 5.09169459e-01 3.60185176e-01 -5.04257046e-02
-1.10142231e+00 1.24096048e+00 1.20484969e-02 1.90763354e-01
1.28532708e+00 -4.38219666e-01 -3.31206918e-01 8.56695250e-02
-8.36429536e-01 5.72939694e-01 7.80011177e-01 5.99095702e-01
3.18243951e-01 -1.69285965e+00 -4.29885626e-01 -1.82755366e-01
1.87295929e-01 9.81800705e-02 5.98976135e-01 8.12966645e-01
-1.45274132e-01 5.51948071e-01 -1.78068683e-01 -1.09251857e+00
-1.29562938e+00 7.18319178e-01 3.54986966e-01 -8.04068446e-02
-1.06815648e+00 6.30370736e-01 5.51933348e-02 2.87645124e-02
3.69897276e-01 -5.05248129e-01 1.89454302e-01 -1.48039237e-01
6.02726281e-01 6.13242805e-01 -3.11818212e-01 -1.09495425e+00
-6.84626341e-01 1.03425920e+00 1.57013372e-01 -2.23848477e-01
1.18817461e+00 -4.15145189e-01 3.94448251e-01 6.30140543e-01
9.73804712e-01 1.09473512e-01 -2.07942271e+00 -5.15592217e-01
1.35531932e-01 -8.93409252e-01 -1.34617716e-01 -3.05167168e-01
-9.58428919e-01 5.12569845e-01 3.85202497e-01 -3.93588282e-03
5.47347188e-01 2.72893876e-01 1.05142927e+00 5.95163368e-02
6.09045029e-01 -6.08347893e-01 -8.96155089e-02 5.24321914e-01
2.03039870e-01 -1.21871102e+00 1.10934235e-01 -3.39377999e-01
-2.71497458e-01 5.44675887e-01 4.57694471e-01 -1.03915885e-01
8.87175441e-01 3.45150054e-01 -5.70452332e-01 -3.69573981e-01
-7.60489583e-01 -5.25553763e-01 5.04851818e-01 6.83782160e-01
7.97614232e-02 -7.59686679e-02 5.09599805e-01 3.70213062e-01
2.62927979e-01 1.44004107e-01 -1.25885159e-02 1.28165781e+00
-3.44794929e-01 -7.00004399e-01 -5.51799119e-01 1.78675443e-01
-3.95575404e-01 4.41261292e-01 8.68030116e-02 9.79718387e-01
3.33836198e-01 7.47306645e-01 3.73697221e-01 -8.35766271e-02
4.45252448e-01 -2.09764615e-01 7.27365077e-01 -5.06645501e-01
-3.24376404e-01 -6.90010637e-02 -1.44987866e-01 -8.57784569e-01
-7.19213009e-01 -1.23456132e+00 -1.30957341e+00 -7.21515000e-01
-1.17016330e-01 -1.83570862e-01 4.47858632e-01 7.99691439e-01
5.41521788e-01 5.62012374e-01 1.28021851e-01 -1.16525412e+00
1.59558639e-01 -3.74723017e-01 -4.40031826e-01 5.75012445e-01
9.60929275e-01 -1.08825743e+00 1.11769348e-01 3.74671727e-01] | [6.318627834320068, -2.089017868041992] |
07487d68-e181-4505-b94b-d4b77fd9d5ee | buol-a-bottom-up-framework-with-occupancy-1 | 2306.00965 | null | https://arxiv.org/abs/2306.00965v1 | https://arxiv.org/pdf/2306.00965v1.pdf | BUOL: A Bottom-Up Framework with Occupancy-aware Lifting for Panoptic 3D Scene Reconstruction From A Single Image | Understanding and modeling the 3D scene from a single image is a practical problem. A recent advance proposes a panoptic 3D scene reconstruction task that performs both 3D reconstruction and 3D panoptic segmentation from a single image. Although having made substantial progress, recent works only focus on top-down approaches that fill 2D instances into 3D voxels according to estimated depth, which hinders their performance by two ambiguities. (1) instance-channel ambiguity: The variable ids of instances in each scene lead to ambiguity during filling voxel channels with 2D information, confusing the following 3D refinement. (2) voxel-reconstruction ambiguity: 2D-to-3D lifting with estimated single view depth only propagates 2D information onto the surface of 3D regions, leading to ambiguity during the reconstruction of regions behind the frontal view surface. In this paper, we propose BUOL, a Bottom-Up framework with Occupancy-aware Lifting to address the two issues for panoptic 3D scene reconstruction from a single image. For instance-channel ambiguity, a bottom-up framework lifts 2D information to 3D voxels based on deterministic semantic assignments rather than arbitrary instance id assignments. The 3D voxels are then refined and grouped into 3D instances according to the predicted 2D instance centers. For voxel-reconstruction ambiguity, the estimated multi-plane occupancy is leveraged together with depth to fill the whole regions of things and stuff. Our method shows a tremendous performance advantage over state-of-the-art methods on synthetic dataset 3D-Front and real-world dataset Matterport3D. Code and models are available in https://github.com/chtsy/buol. | ['Jiaqi Wang', 'Qiong Liu', 'Pan Zhang', 'Tao Chu'] | 2023-06-01 | buol-a-bottom-up-framework-with-occupancy | http://openaccess.thecvf.com//content/CVPR2023/html/Chu_BUOL_A_Bottom-Up_Framework_With_Occupancy-Aware_Lifting_for_Panoptic_3D_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Chu_BUOL_A_Bottom-Up_Framework_With_Occupancy-Aware_Lifting_for_Panoptic_3D_CVPR_2023_paper.pdf | cvpr-2023-1 | ['panoptic-segmentation', '3d-scene-reconstruction', '3d-reconstruction'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 1.96214736e-01 5.40089197e-02 1.77921429e-02 -2.46350646e-01
-6.63389266e-01 -6.05864406e-01 4.44834620e-01 7.99863562e-02
-1.18478779e-02 2.41281480e-01 -9.53602418e-02 -3.31011683e-01
7.97525048e-02 -9.35345352e-01 -5.97635567e-01 -5.41562617e-01
-8.73738620e-03 1.14729309e+00 7.37663031e-01 5.32349907e-02
3.22385043e-01 7.67668784e-01 -1.69193435e+00 2.64712423e-01
8.07548285e-01 1.08097458e+00 4.10976827e-01 4.35023218e-01
-7.72840261e-01 -1.88387513e-01 -1.66570470e-01 2.64524668e-01
8.41121674e-01 -1.53554618e-01 -5.94198406e-01 5.89144826e-01
6.44127846e-01 -4.03152853e-01 3.42536628e-01 9.47708011e-01
1.71396494e-01 -6.34625405e-02 5.45840502e-01 -8.28799546e-01
3.64771307e-01 -9.62721705e-02 -1.08394420e+00 -2.70306975e-01
1.54004872e-01 1.96077555e-01 8.24551523e-01 -1.03354144e+00
7.14188039e-01 1.26367414e+00 4.27053034e-01 1.48563877e-01
-1.26962185e+00 -6.05450749e-01 4.92634088e-01 -3.92729342e-01
-1.24066961e+00 1.50022000e-01 9.13860977e-01 -7.65860200e-01
9.75633025e-01 4.33841825e-01 1.06956184e+00 2.37950906e-01
9.87457708e-02 4.40022707e-01 1.44434333e+00 -1.02998681e-01
3.42531741e-01 -1.51226029e-01 1.85214937e-01 6.33130968e-01
1.36959925e-01 1.21803209e-01 -5.68687737e-01 2.61503663e-02
1.19788337e+00 1.10497274e-01 -1.92094028e-01 -5.77176690e-01
-1.05237079e+00 6.62667513e-01 4.99726921e-01 -1.75338596e-01
-2.16097951e-01 -1.43333599e-01 1.15285881e-01 -1.75679773e-01
9.01644349e-01 6.84821248e-01 -6.65537477e-01 1.91667199e-01
-1.05336452e+00 5.42825162e-01 5.72352648e-01 9.50154901e-01
1.28118122e+00 -2.63173193e-01 2.92637885e-01 6.48445666e-01
3.16537172e-01 6.31552100e-01 -2.27871343e-01 -1.16946018e+00
4.57866281e-01 8.29777539e-01 1.55254066e-01 -6.42065167e-01
-7.25448728e-01 -4.44192380e-01 -5.78730822e-01 5.02950907e-01
3.31376046e-01 2.17743665e-01 -1.42343259e+00 1.06375217e+00
1.02854085e+00 1.32951513e-01 -3.51396948e-01 1.19706821e+00
7.22708762e-01 8.03754985e-01 -4.79485452e-01 -1.47463053e-01
1.57021487e+00 -1.06519008e+00 -2.35648289e-01 -4.27544892e-01
1.88474149e-01 -8.97073388e-01 1.09058630e+00 4.05894756e-01
-1.02426338e+00 -4.76109594e-01 -7.26647079e-01 -3.20556872e-02
-1.89730749e-01 -5.11929691e-01 7.47135222e-01 3.30895364e-01
-6.71801388e-01 2.10036963e-01 -7.09382534e-01 -1.90918669e-01
3.82184684e-01 8.11030492e-02 -7.10872263e-02 -2.76052356e-01
-7.56929100e-01 4.34771866e-01 2.29360476e-01 -1.82992071e-02
-9.43634987e-01 -1.19799650e+00 -8.57859910e-01 -1.29576504e-01
8.23045909e-01 -9.16670978e-01 1.12732995e+00 -2.12866813e-01
-1.20663369e+00 1.02060175e+00 -2.56756991e-01 -1.45570204e-01
5.38198650e-01 -1.43795475e-01 1.75427958e-01 2.28578299e-01
4.06752378e-01 9.36476469e-01 6.88862205e-01 -1.88013673e+00
-7.29579329e-01 -6.78716421e-01 7.46443421e-02 6.62776053e-01
6.60830498e-01 -5.29936731e-01 -7.55399048e-01 -2.44784907e-01
1.01510739e+00 -6.86128497e-01 -6.68614864e-01 -2.22534989e-03
-6.92447901e-01 2.51425773e-01 7.18789339e-01 -2.99065411e-01
8.88785303e-01 -2.02091908e+00 6.93885237e-02 1.99170858e-01
3.89512569e-01 -3.37598503e-01 1.78531945e-01 7.88458809e-02
3.52383554e-01 1.20742731e-01 -8.37184727e-01 -6.46362066e-01
-1.07465617e-01 3.19763809e-01 -4.61393595e-01 4.15083110e-01
-1.17586695e-01 5.16104281e-01 -7.55642653e-01 -4.24459755e-01
7.46051311e-01 3.29796612e-01 -9.39428329e-01 2.11972892e-01
-6.88570023e-01 9.15849507e-01 -6.75050497e-01 8.65136027e-01
1.47757018e+00 -1.96236670e-01 2.06776336e-02 -2.84385383e-01
-8.06043386e-01 4.48308766e-01 -1.43788493e+00 1.76460862e+00
-4.59304065e-01 5.72556444e-02 4.44877476e-01 -5.32307029e-01
7.65427411e-01 -1.11718185e-01 6.41971350e-01 -6.79711521e-01
-1.86533362e-01 3.77825320e-01 -4.59513336e-01 -2.08539233e-01
6.53173268e-01 -3.41536582e-01 -2.03121424e-01 2.54617482e-01
-3.37399989e-01 -1.40716553e+00 -1.38212293e-01 7.39439726e-02
5.23824811e-01 3.56280088e-01 1.15744904e-01 -3.41741115e-01
4.41789329e-02 4.85308915e-01 6.58130169e-01 8.09933841e-01
2.08552137e-01 1.15208852e+00 4.75868464e-01 -6.35870516e-01
-1.13315523e+00 -1.12216818e+00 -5.90429068e-01 2.45930195e-01
7.29804337e-01 -4.37333345e-01 -7.19745338e-01 -4.55239475e-01
4.81898226e-02 5.63408494e-01 -5.11001527e-01 6.89773500e-01
-5.73163390e-01 -9.41249132e-01 -2.44202584e-01 -1.33942124e-02
5.23621559e-01 -6.52065992e-01 -1.12919104e+00 2.32445940e-01
-1.61565125e-01 -1.22828782e+00 -2.82959968e-01 4.61609781e-01
-1.13587856e+00 -1.07592094e+00 -4.76518184e-01 -1.46716788e-01
7.72615552e-01 5.03203034e-01 1.23472095e+00 -1.60916999e-01
-4.14444059e-01 2.26774082e-01 -2.52293259e-01 -3.83899897e-01
6.93389550e-02 -1.17215365e-01 -1.18810870e-01 -1.89714223e-01
-1.89096332e-01 -6.75996065e-01 -8.72724175e-01 6.04759932e-01
-8.15147877e-01 8.95134509e-01 1.58424243e-01 4.54022646e-01
1.33239865e+00 1.23080425e-01 -3.87647778e-01 -9.43373799e-01
-2.65723020e-01 -4.15061802e-01 -1.18060184e+00 -2.11187452e-01
-3.29395354e-01 -1.15876861e-01 1.67960033e-01 -3.06823738e-02
-9.86583054e-01 1.17611274e-01 -2.00435832e-01 -5.90990961e-01
-4.25818622e-01 1.92992195e-01 -1.83174998e-01 3.05000067e-01
2.80190498e-01 5.77044934e-02 -4.10902292e-01 -9.01961267e-01
2.09073886e-01 1.98384300e-01 2.27871701e-01 -6.72432601e-01
6.10851943e-01 9.59225535e-01 2.70699024e-01 -8.09730589e-01
-1.17779469e+00 -6.03821754e-01 -8.63745391e-01 -3.99022192e-01
1.12888861e+00 -9.17931259e-01 -2.67902493e-01 4.21704799e-01
-1.08891153e+00 -6.98307216e-01 -4.87895459e-01 4.20867741e-01
-3.89667511e-01 1.72886968e-01 -2.87361264e-01 -7.47598946e-01
-1.26346156e-01 -1.60934615e+00 1.57421732e+00 8.77020694e-03
-2.83169188e-02 -5.51149905e-01 -4.74844053e-02 5.84205508e-01
8.58524740e-02 3.76935780e-01 9.98480856e-01 1.45387247e-01
-1.18349397e+00 3.60809594e-01 -2.39723712e-01 -2.19813779e-01
-1.11646704e-01 -8.31382498e-02 -9.90506649e-01 1.83602244e-01
1.78847730e-01 -2.81584319e-02 8.82325470e-01 8.10759544e-01
1.16367722e+00 1.05230294e-01 -3.58441025e-01 1.13523436e+00
1.57703984e+00 -4.38442715e-02 4.46259618e-01 2.86876529e-01
7.83333540e-01 7.95324504e-01 8.15398097e-01 5.73289096e-01
4.99436080e-01 9.56540465e-01 8.98053408e-01 -3.07119936e-01
-2.89041311e-01 -4.32450712e-01 -2.73392797e-01 4.78036374e-01
-5.21357730e-03 -2.21657842e-01 -1.11181903e+00 4.17983294e-01
-1.63531780e+00 -3.78997177e-01 -6.39803171e-01 2.27916217e+00
4.99387801e-01 2.92221665e-01 -1.57641232e-01 -2.18915239e-01
3.59730959e-01 4.48015302e-01 -6.83912456e-01 -1.88388988e-01
-2.39478797e-02 1.64102852e-01 6.69656157e-01 1.06585252e+00
-8.72393548e-01 1.12802410e+00 5.18904638e+00 9.29252565e-01
-1.12807620e+00 1.72463298e-01 8.33797753e-01 -1.00466631e-01
-6.57552600e-01 3.88224810e-01 -1.29463696e+00 3.65706950e-01
2.03713939e-01 5.55068612e-01 2.44498029e-01 5.92681289e-01
3.39330584e-01 -7.90198088e-01 -7.72798419e-01 1.12578011e+00
-1.72429800e-01 -1.37195444e+00 2.82850146e-01 2.28965774e-01
9.43063736e-01 3.46707374e-01 -1.20931454e-01 -3.72208953e-02
5.94433434e-02 -8.45623851e-01 9.96884704e-01 1.58034354e-01
1.10465050e+00 -4.21045154e-01 2.42243633e-01 6.03163719e-01
-1.47684693e+00 3.30499142e-01 -2.50487268e-01 -9.18395296e-02
7.22771227e-01 1.14884245e+00 -7.50418901e-01 7.28748024e-01
8.80964994e-01 3.94250125e-01 -1.22071289e-01 9.42951858e-01
-1.83768556e-01 3.52399200e-01 -7.25977838e-01 6.91036403e-01
5.53590834e-01 -5.83882034e-01 8.70805144e-01 9.07335699e-01
4.15701151e-01 3.58339280e-01 4.87027109e-01 1.19915271e+00
2.32919052e-01 -1.68760955e-01 -5.38439393e-01 6.84614599e-01
2.53288180e-01 1.22744811e+00 -1.22693324e+00 -3.78340185e-01
-1.53710827e-01 7.78627694e-01 -7.22315684e-02 2.21157879e-01
-8.13341200e-01 2.02435344e-01 7.20801294e-01 9.14000392e-01
2.94455767e-01 -3.79044145e-01 -8.73281837e-01 -1.21041954e+00
-4.75708507e-02 -3.19617391e-01 2.26596683e-01 -8.43771040e-01
-9.66937602e-01 4.50385481e-01 4.58733082e-01 -1.31059909e+00
3.30230922e-01 -5.82021534e-01 -2.86131412e-01 9.17265356e-01
-1.64879310e+00 -8.88096631e-01 -5.03008068e-01 1.72586277e-01
9.40064073e-01 6.19796693e-01 6.23854458e-01 9.07782465e-02
-1.67366728e-01 -2.89963990e-01 -2.52032727e-01 -4.02936280e-01
2.99790829e-01 -1.13485360e+00 4.61318046e-01 6.39698863e-01
-6.74393922e-02 1.03489354e-01 6.37240827e-01 -9.03539240e-01
-1.14655674e+00 -1.04750478e+00 6.36752665e-01 -4.76044238e-01
1.70143589e-01 -6.87452614e-01 -9.59106266e-01 3.53613764e-01
-2.21719474e-01 1.43126860e-01 2.03819603e-01 -3.98482084e-02
-1.74632475e-01 -7.05753788e-02 -1.25848544e+00 4.20884192e-01
1.24943411e+00 -2.21196011e-01 -2.58226365e-01 2.74876505e-01
8.10404480e-01 -9.56118107e-01 -5.71428120e-01 8.29040349e-01
4.12353247e-01 -1.28306139e+00 1.03090310e+00 1.39665321e-01
5.19770265e-01 -8.53939533e-01 -2.64674544e-01 -1.02739644e+00
1.25626281e-01 -5.02106428e-01 1.24423683e-01 4.90915060e-01
4.20987964e-01 -5.46290278e-01 7.25089669e-01 5.57806134e-01
-6.14241600e-01 -8.73689711e-01 -1.09064114e+00 -3.55464071e-01
-1.47831514e-01 -8.96887898e-01 7.48756766e-01 8.82997751e-01
-6.44085944e-01 2.92519271e-01 -4.79559526e-02 5.41389465e-01
8.31393659e-01 9.98434126e-01 8.93233120e-01 -1.26591551e+00
-2.54182249e-01 -3.17742795e-01 8.91273767e-02 -1.79484701e+00
-3.71686667e-01 -7.45938957e-01 9.41029787e-02 -1.64375019e+00
1.67871788e-01 -9.22416508e-01 3.42599094e-01 2.35564038e-01
1.01929002e-01 3.67527336e-01 1.16348259e-01 2.68206596e-01
-3.33038390e-01 4.39208508e-01 1.68523443e+00 1.94125697e-01
-4.78732258e-01 -9.31017026e-02 -2.76217014e-01 9.93534684e-01
6.22574508e-01 -4.66528893e-01 -3.66881222e-01 -7.83620536e-01
2.85408735e-01 3.29136193e-01 4.92703736e-01 -8.90271723e-01
-1.33234918e-01 -4.25410271e-01 3.28095198e-01 -1.33045161e+00
7.98109531e-01 -9.07164574e-01 3.16341311e-01 2.56748348e-01
2.98093140e-01 -2.41342410e-01 3.12560976e-01 3.58997077e-01
-8.99416059e-02 -5.19319959e-02 9.24479723e-01 -7.50028133e-01
-7.36128509e-01 6.43795431e-01 -9.83504727e-02 3.31219169e-03
8.72232556e-01 -4.69624490e-01 7.25132152e-02 -3.10443211e-02
-8.01252365e-01 2.95868635e-01 8.37689459e-01 -1.42813502e-02
6.70106947e-01 -7.24020600e-01 -5.88676274e-01 6.03978872e-01
-2.55519971e-02 1.18574512e+00 6.53571427e-01 8.40017319e-01
-8.05149436e-01 3.98606688e-01 2.07278877e-01 -1.23072302e+00
-9.57310259e-01 1.51737556e-01 5.69111943e-01 -3.48882824e-01
-1.10130203e+00 9.15732980e-01 1.09256518e+00 -8.40688527e-01
-1.61872566e-01 -7.65132666e-01 1.86102986e-01 5.87973334e-02
2.39437297e-01 6.99253157e-02 6.25404269e-02 -4.22148675e-01
-2.02077940e-01 1.16285837e+00 2.32675299e-01 -2.97487199e-01
1.43521142e+00 -2.87932009e-01 -3.05432882e-02 4.98438835e-01
7.57838666e-01 1.84151202e-01 -1.65808880e+00 -3.12168002e-02
-6.06182754e-01 -8.28560054e-01 1.27688840e-01 -6.85453176e-01
-1.01245463e+00 1.08120406e+00 3.45657349e-01 -3.77092846e-02
9.16883469e-01 3.26435089e-01 7.58990228e-01 -3.02542210e-01
6.98293626e-01 -7.78073013e-01 -2.30746478e-01 7.77842820e-01
7.06113398e-01 -1.08531320e+00 2.62311071e-01 -9.45110977e-01
-5.34862816e-01 9.88119900e-01 8.05995703e-01 8.74549970e-02
7.36658096e-01 4.25502747e-01 -6.49890974e-02 -6.97883785e-01
-3.46639872e-01 -1.91665322e-01 2.10219145e-01 1.78861126e-01
-2.83495914e-02 2.47483134e-01 7.13698333e-03 2.73333043e-01
-1.70858920e-01 -4.94286746e-01 3.31203461e-01 6.60471737e-01
-5.85658491e-01 -8.75299513e-01 -7.18744576e-01 2.81676590e-01
1.19032793e-01 -1.96301043e-01 7.29641095e-02 8.08335304e-01
5.79287171e-01 5.12715101e-01 6.86830282e-01 4.41817120e-02
3.42260331e-01 -2.16119379e-01 5.03894627e-01 -1.08622265e+00
-1.65799081e-01 5.80610394e-01 1.04121700e-01 -6.79000020e-01
-2.72467226e-01 -5.95079064e-01 -1.59180713e+00 2.96772271e-02
-8.87781456e-02 1.19048953e-01 8.27906668e-01 5.85123718e-01
3.85153413e-01 2.01919824e-01 5.94494522e-01 -1.44092023e+00
1.32841870e-01 -6.84793115e-01 -6.20948255e-01 3.18085253e-02
2.74532318e-01 -8.27628314e-01 -5.56080282e-01 -1.96177945e-01] | [8.75183391571045, -2.8965470790863037] |
579b8350-75aa-43bd-833e-4363db68eabb | learning-deformable-object-manipulation-from | 2207.10148 | null | https://arxiv.org/abs/2207.10148v1 | https://arxiv.org/pdf/2207.10148v1.pdf | Learning Deformable Object Manipulation from Expert Demonstrations | We present a novel Learning from Demonstration (LfD) method, Deformable Manipulation from Demonstrations (DMfD), to solve deformable manipulation tasks using states or images as inputs, given expert demonstrations. Our method uses demonstrations in three different ways, and balances the trade-off between exploring the environment online and using guidance from experts to explore high dimensional spaces effectively. We test DMfD on a set of representative manipulation tasks for a 1-dimensional rope and a 2-dimensional cloth from the SoftGym suite of tasks, each with state and image observations. Our method exceeds baseline performance by up to 12.9% for state-based tasks and up to 33.44% on image-based tasks, with comparable or better robustness to randomness. Additionally, we create two challenging environments for folding a 2D cloth using image-based observations, and set a performance benchmark for them. We deploy DMfD on a real robot with a minimal loss in normalized performance during real-world execution compared to simulation (~6%). Source code is on github.com/uscresl/dmfd | ['Gaurav S. Sukhatme', 'Marcus Dominguez-Kuhne', 'I-Chun Arthur Liu', 'Gautam Salhotra'] | 2022-07-20 | null | null | null | null | ['deformable-object-manipulation'] | ['robots'] | [-0.04497123 0.09541973 0.10112944 -0.1789748 -0.6397549 -0.96108437
0.43519664 -0.3868405 -0.4656408 0.7828922 0.01961283 -0.28966525
-0.09119611 -0.27984285 -1.1811731 -0.51399857 -0.6733937 0.48869297
0.3517969 -0.25711516 0.17058751 0.48760036 -1.4756556 0.0660296
0.8972451 0.5711319 0.6746643 1.2288389 0.6116731 0.47448406
-0.5830301 0.32548144 0.67810255 -0.01773547 -0.84188855 0.04756641
0.5777283 -0.86101544 -0.68979865 0.89950645 0.6063771 0.37013268
0.47070834 -1.4099785 -0.5405987 0.3843318 -0.17203411 -0.15781853
0.58421904 1.1151476 0.34701687 -0.6571459 0.9782756 1.6860602
0.40399963 0.8713616 -1.331858 -0.47476962 0.2926386 -0.3031468
-0.63353294 -0.36870217 0.1325868 -0.54988015 1.2455288 -0.0351013
0.8595647 1.560926 0.49437052 0.8015201 1.3622451 0.02007128
0.3741971 -0.4425119 -0.07813097 1.1705133 0.15095536 0.54020804
-0.36419168 -0.19194885 1.0450194 -0.13589264 -0.47031587 -0.82919407
-1.7287015 0.2675212 0.555075 -0.50446695 -0.33264953 0.7525213
0.37401447 0.6952017 -0.14774235 0.7881703 -0.66366976 -0.51085514
-0.03973454 0.9508315 1.2667444 1.249483 0.3808978 0.00872282
-0.12027549 0.05788469 0.2159365 0.8943702 0.1275953 -1.6678238
0.6615255 0.20993087 0.6955696 -0.624124 -0.43021303 0.2422007
-0.18773597 1.0291862 0.62954396 -0.5316558 -1.3785598 1.8312125
0.5353487 -0.16498989 0.20099507 1.2627052 0.64192605 0.60802484
-0.20437083 0.371645 0.7502606 -1.3872565 -0.37159398 -0.12897056
0.47020936 -0.33236983 1.5819086 0.4421458 -1.0107715 -0.5461105
-1.071699 0.09509617 -0.12540811 0.06505033 0.53497297 -0.02176408
-1.0111378 1.1472377 -1.5393986 -0.55680376 0.12617372 0.46135256
-0.3494373 -0.01871085 -0.5978417 1.0931762 0.15105587 0.03056352
-1.9328564 -0.7626799 -0.92986023 -0.44062468 0.5047712 -0.7161407
1.4885641 -0.10655934 -2.0156863 0.6308013 0.24861513 -0.3312002
0.996796 -0.7736643 0.21712087 0.14115822 -0.03306514 0.9391405
0.84442824 -1.5590867 -0.19686538 -0.04587339 0.45804375 0.3920697
0.21228631 -0.48933262 -0.20510404 -0.36403647 -0.07217561 -1.7164762
-0.36287078 0.7192111 -0.497822 -0.09762999 1.2778513 -0.5923071
0.1998015 -1.8219991 0.9092495 -0.1864902 0.28038636 0.04891918
-0.37834054 0.6187494 0.47582027 -0.11067431 -0.24508193 -0.3499357
0.43102798 0.54541975 -0.26983663 0.52175045 0.21468277 1.1245054
-1.2934971 -0.16952062 0.4276781 0.17862584 -0.46929407 0.5313307
-0.6609739 0.865097 -0.57410866 0.9173112 0.36097112 -0.12380838
0.19325458 0.05235055 -0.1442864 0.0691688 -0.90743583 2.293473
-0.525448 0.6006868 0.56007504 -0.58254075 0.7265865 -0.0284424
0.1274168 -0.28080577 -0.03914763 0.2732902 0.17306319 -1.1646401
0.24600385 0.51114494 -0.1972002 0.3673374 0.09894735 -0.954624
0.16027781 0.11273392 1.7494255 0.9581545 -0.27672356 -0.54797816
-0.38555318 0.31971842 0.13336319 0.86673796 -0.41908702 0.26274514
0.27541578 -0.5199398 -1.1091316 -1.4444897 0.220309 0.9511886
0.7579991 -0.17041402 -0.6498573 -0.7821902 0.6533478 0.5274496
-0.6867718 0.02720222 -0.78593385 -0.11507392 0.3543888 0.4906727
0.5611897 -1.3670437 -1.4391764 -0.16551824 0.26839805 -0.94913554
-0.3827748 0.3916112 -0.72349274 -1.1805649 -0.82724196 -0.8073939
0.7663289 0.2666719 0.9720096 -0.08955901 -0.5737255 0.8512173
-0.35121408 -0.09467166 -0.684773 0.09926616 0.40133464 -1.110858
-0.54828143 -0.64740604 -0.6677887 0.49714398 -0.34187633 0.21145876
0.6672006 0.9174658 0.27023292 -0.45609078 0.26434475 -0.14519565
0.79753375 -0.29637098 -0.8112482 0.01123707 -0.22707324 0.24346158
0.45711318 -0.971079 -0.73897964 0.27743167 0.31502193 -0.9648282
-0.18444206 0.00842568 0.19555941 -0.3230389 0.9049376 -0.08934424
0.49763006 -0.30834424 0.5350353 0.17609821 0.59768635 -1.2833681
1.0093834 0.3194519 -0.0780201 -0.4495154 -0.21964562 0.1649695
-0.62457746 -0.26483902 0.7816663 -0.7978697 -1.2184336 0.5082941
-0.89188814 -1.5937577 -0.38550243 0.49972436 -0.92485774 0.18645008
-0.7591562 -0.6143972 -0.15239607 -1.5680017 1.3421185 0.2593846
-0.26904956 -0.6508217 0.11197449 -0.02284058 0.45945507 0.82746667
0.6227831 -0.06083662 -0.81936765 0.04351055 -0.01612055 0.12520407
0.11830871 -0.09595597 -0.5016404 -0.94023585 -0.34677386 -1.1697668
0.76601905 0.09829643 1.1213998 -0.33256823 -0.5762199 0.48894054
1.0233183 0.08825164 0.2003269 0.25679836 0.6299819 0.36836162
0.9627933 0.47035035 0.31090516 0.58518684 0.90523183 0.09145465
-0.15861787 -0.32181993 0.74879456 0.4673115 -0.22303922 -0.29039446
-0.7678381 0.552017 -2.0406256 -0.5916884 0.23285645 2.0165694
0.6865548 0.17075025 0.1778307 -0.48760572 0.36422318 -0.01652658
-1.326915 -0.19958027 0.37775344 0.01095969 0.5123223 0.55051875
-1.0081627 0.8777214 6.319835 0.2803314 -0.9707036 -0.30741453
-0.05666665 -0.34475714 0.14031924 -0.05795499 -0.392215 0.50005215
0.42537558 -0.05513437 0.90464896 1.0448577 0.3144082 -0.2933374
-1.4642168 0.6793725 -0.265023 -0.9288809 -0.38446492 0.08967032
0.78091276 0.38288283 0.21546361 0.510246 1.0291973 -1.0478294
0.91679645 0.31494525 0.7843294 -0.09532715 0.27935615 0.48956433
-0.9689826 -0.10580532 -0.08836488 -0.30413485 0.23924793 -0.13051352
-0.89192116 0.1672361 1.0443491 0.5972951 -0.06791115 0.5668715
-0.34051925 0.2692209 -0.5167411 -0.34094453 0.30626702 0.01236295
0.9109187 1.0472581 0.21973842 0.05142824 0.7737678 0.9977859
0.03425204 -0.81778586 -0.7736296 0.04566011 0.5268062 1.0983375
-0.47142342 -0.3265314 0.32550937 1.2057253 0.41135463 0.4429585
-1.1872319 -0.49934232 1.0095744 -0.13426772 0.2960959 -0.96854776
0.39219975 -1.0842172 0.20076987 -1.0954014 -0.25628653 -0.76651204
-1.1198093 0.2715688 0.29408634 -1.1352367 -0.12242544 -0.75747925
-0.3593569 0.6254582 -1.252159 -0.9385155 -0.91913414 0.19577363
0.8900613 0.2412469 0.8819671 -0.21037863 -0.27192757 0.2247018
0.09546835 -0.10828733 0.72320116 -1.5821705 0.9114641 0.32278803
-0.591929 0.5742599 0.88076365 -0.8128903 -2.2409768 -0.9644556
-0.4002731 -0.8931554 0.68325394 -0.5450091 -0.7423582 0.77938306
0.2912175 0.20049874 -0.29164845 -0.23258092 -0.15180922 0.47331953
-1.3868275 0.83441955 1.7522026 -0.1645712 -0.633061 0.6804206
1.1102244 -1.2998679 -1.0043707 0.5523442 0.96394193 -0.53169996
0.83342093 -0.7543709 0.7448475 -0.51085424 -0.3455219 -1.7700343
-0.06172218 -1.0940177 -0.5907637 0.38678735 0.25595424 -0.62239045
0.65628386 0.50510925 -0.30837134 -1.0543963 -0.60063744 -1.3243775
0.02107397 0.19828773 0.22059314 0.6491917 0.11590277 -0.01235542
-0.32029355 0.29708922 0.73600894 0.34587193 1.4283565 -0.4716651
-0.60521495 -0.01508155 -0.09247334 -1.5558348 0.323886 -0.7670811
0.67477983 -1.6700506 0.10491722 -0.60147727 0.3607397 0.6182454
0.10151449 -0.28437287 0.5738627 0.26833063 -0.67901784 0.7661712
1.7759827 -0.1382899 -0.4623077 -0.21453004 0.12464838 0.71725345
0.6832295 -0.10619159 -0.49174154 -0.9714504 -0.30951452 0.2888954
0.72613704 -1.2017268 0.00748598 -0.41688773 0.19644938 -0.32284737
0.72432786 -0.6119234 0.02522736 1.140311 -0.63016945 0.3449965
0.3865806 0.86855584 0.6243766 0.22925726 0.5841152 -0.41899866
-0.83642894 0.1176304 -0.5003743 0.09041594 1.24968 -0.02804428
-0.83016485 -0.31674898 -0.9138198 0.7713685 1.0251353 0.65553516
0.6319243 -0.94915277 -0.4664354 -0.14907327 -0.18024847 0.32524922
0.04477783 0.5662903 -0.95156825 -0.12133309 -0.4344374 -0.8048902
-1.258544 0.42504984 0.2871027 0.00916873 -1.0095247 0.8194099
-0.00887547 -0.87893426 0.46873084 -0.9637737 0.6280386 -0.93894637
-0.04151016 0.65147436 -0.38132495 0.10526538 -0.27497712 0.5701145
0.10337973 -0.17368609 1.2798773 0.21250616 0.29474926 0.40224755
0.9035147 -0.3825884 -2.1331494 0.22391304 -0.4084903 -0.57676905
-0.49034968 -1.0390074 -0.59992087 0.6198138 0.62758106 0.05037548
0.55403036 0.02333184 0.73876804 1.1730434 0.783358 -0.8495932
0.7115223 0.5574593 1.390361 -1.3793753 0.09617203 -0.27338904
-0.60654896 0.98279935 1.0309547 -0.77922493 0.35839677 0.7325903
-0.0528166 -0.2171846 -0.91932774 0.13249849 -0.3343872 0.7918142
-0.38686886 0.15300491 0.45614126 -0.08934535 -0.03417934 0.01090398
0.4115253 1.5260619 -0.53263 -0.62251425 -0.18773341 0.25400543
0.35896847 0.61294526 -0.34810776 1.1004577 -0.20631944 0.6810736
-0.17894691 -0.5720804 0.4364899 -0.32475126 1.0289463 -0.5950928
-0.34147206 -0.4918383 0.21066721 -1.1049753 -0.27137566 -0.53087294
-1.3702221 -0.31460673 -0.1973349 -0.33238748 0.7652112 0.62150836
0.6911197 0.5479991 0.31804204 -1.859427 -1.263852 -0.8560824
-0.05872006 0.33029914 0.58190715 -1.0587977 -0.4429913 0.01641227] | [4.717764377593994, 0.6166514754295349] |
6e188428-7e55-4806-a994-c883df86c3ba | unified-multimodal-model-with-unlikelihood | 2211.13235 | null | https://arxiv.org/abs/2211.13235v1 | https://arxiv.org/pdf/2211.13235v1.pdf | Unified Multimodal Model with Unlikelihood Training for Visual Dialog | The task of visual dialog requires a multimodal chatbot to answer sequential questions from humans about image content. Prior work performs the standard likelihood training for answer generation on the positive instances (involving correct answers). However, the likelihood objective often leads to frequent and dull outputs and fails to exploit the useful knowledge from negative instances (involving incorrect answers). In this paper, we propose a Unified Multimodal Model with UnLikelihood Training, named UniMM-UL, to tackle this problem. First, to improve visual dialog understanding and generation by multi-task learning, our model extends ViLBERT from only supporting answer discrimination to holding both answer discrimination and answer generation seamlessly by different attention masks. Specifically, in order to make the original discriminative model compatible with answer generation, we design novel generative attention masks to implement the autoregressive Masked Language Modeling (autoregressive MLM) task. And to attenuate the adverse effects of the likelihood objective, we exploit unlikelihood training on negative instances to make the model less likely to generate incorrect answers. Then, to utilize dense annotations, we adopt different fine-tuning methods for both generating and discriminating answers, rather than just for discriminating answers as in the prior work. Finally, on the VisDial dataset, our model achieves the best generative results (69.23 NDCG score). And our model also yields comparable discriminative results with the state-of-the-art in both single-model and ensemble settings (75.92 and 76.17 NDCG scores). | ['Changjun Jiang', 'Junli Wang', 'ZiHao Wang'] | 2022-11-23 | null | null | null | null | ['visual-dialogue', 'visual-dialogue', 'answer-generation'] | ['computer-vision', 'natural-language-processing', 'natural-language-processing'] | [ 1.62640139e-01 1.81019649e-01 -1.69477873e-02 -4.98200119e-01
-1.09296215e+00 -7.59703815e-01 7.83936739e-01 -3.42276037e-01
-4.65467274e-01 7.28648961e-01 3.06220829e-01 -3.29572290e-01
4.00291890e-01 -6.15203023e-01 -3.55759948e-01 -5.55798113e-01
5.76627195e-01 6.12664282e-01 1.94849432e-01 -2.24865377e-01
2.29148477e-01 -9.58861187e-02 -1.25877845e+00 8.29504490e-01
1.18956685e+00 8.49903882e-01 4.84096587e-01 9.28554893e-01
-5.43896377e-01 1.30824482e+00 -9.62551713e-01 -8.96574616e-01
-2.91464981e-02 -9.65728581e-01 -9.00068223e-01 9.64277685e-02
4.77484971e-01 -6.93190694e-01 -2.03832030e-01 7.61524141e-01
6.95010364e-01 9.13835540e-02 8.70310843e-01 -1.25353336e+00
-1.30761576e+00 4.76226538e-01 -6.61953568e-01 -1.10685408e-01
6.12961471e-01 5.35428464e-01 1.10435688e+00 -1.20975780e+00
4.80421901e-01 1.69772100e+00 2.18841761e-01 9.15333271e-01
-1.19776106e+00 -6.62666321e-01 2.69218534e-01 2.16681421e-01
-1.04299176e+00 -4.94566560e-01 5.97444475e-01 -3.62852424e-01
8.08446288e-01 4.84765857e-01 -1.14848698e-02 1.35927427e+00
-3.80818471e-02 1.09942627e+00 1.07663560e+00 -2.44646907e-01
-1.45518184e-01 4.72134143e-01 -4.98431772e-02 7.59022772e-01
-4.43595618e-01 -2.88692176e-01 -4.67181802e-01 -5.28551452e-02
6.45876646e-01 -8.84870589e-02 -2.86389768e-01 1.04579240e-01
-1.12385833e+00 1.12129414e+00 4.42306072e-01 7.88578764e-02
-2.61007458e-01 1.03799753e-01 1.68497086e-01 3.80393445e-01
3.12396973e-01 3.79548728e-01 -1.31363928e-01 4.14076261e-02
-5.26594102e-01 2.96768695e-01 6.51105881e-01 1.04187727e+00
8.03804338e-01 4.81090620e-02 -1.10520208e+00 1.06620002e+00
4.53348547e-01 5.31450987e-01 3.69680971e-01 -1.13855958e+00
7.95114994e-01 7.04338729e-01 1.48213923e-01 -7.53915310e-01
-1.91671580e-01 -3.12440842e-01 -8.83500636e-01 3.44793722e-02
6.34791255e-01 -3.45209002e-01 -9.92649138e-01 1.88223946e+00
1.44987792e-01 -4.40504402e-01 1.34693444e-01 1.05688334e+00
1.46284795e+00 7.94052780e-01 2.33206734e-01 -2.75478736e-02
1.37465239e+00 -1.20167935e+00 -8.53971183e-01 -4.79137927e-01
5.60468376e-01 -8.50421846e-01 1.70778918e+00 2.21709851e-02
-9.79395986e-01 -7.76329041e-01 -6.23576701e-01 -3.43329430e-01
2.48105656e-02 5.05583763e-01 3.67413521e-01 4.80181456e-01
-1.04183447e+00 -1.40956283e-01 -6.18314855e-02 -1.27900511e-01
2.42356643e-01 1.12515420e-01 -1.25703467e-02 -1.75971225e-01
-1.16265869e+00 9.52690840e-01 3.61540727e-02 -4.28037420e-02
-8.41781437e-01 -3.44225317e-01 -8.56083155e-01 9.84545052e-02
4.80993181e-01 -9.72535014e-01 1.38335669e+00 -1.00726926e+00
-1.62129748e+00 8.88997853e-01 -5.27599156e-01 -2.32176900e-01
7.90043890e-01 -1.76769555e-01 5.43907396e-02 2.52592057e-01
1.47398248e-01 1.32413173e+00 8.91838431e-01 -1.50987232e+00
-2.21628159e-01 -1.23087540e-01 5.04093349e-01 3.64964932e-01
-3.42316478e-01 -2.73139272e-02 -6.44319654e-01 -5.91865242e-01
-1.49816185e-01 -7.31938243e-01 -6.27303421e-02 -2.47453302e-01
-4.29653436e-01 -4.95255291e-01 7.22736537e-01 -7.62751043e-01
1.20329869e+00 -2.04213810e+00 1.46864653e-01 -2.43113667e-01
3.42001170e-01 3.44219394e-02 -4.73747283e-01 3.62927645e-01
3.09563935e-01 1.38305515e-01 -1.72628790e-01 -7.01852739e-01
1.27858967e-01 2.89588720e-01 -4.42677230e-01 -1.01117142e-01
5.03632426e-01 1.26958668e+00 -5.95060647e-01 -6.94014192e-01
-2.82335244e-02 1.66665614e-01 -7.24943161e-01 6.52916789e-01
-5.12009323e-01 7.10628390e-01 -3.39773566e-01 6.44329906e-01
6.03628635e-01 -6.41801000e-01 2.05020085e-01 -9.72978175e-02
1.17214777e-01 3.87459062e-02 -7.71938264e-01 1.47367013e+00
-6.75651431e-01 4.55213845e-01 1.35671586e-01 -4.39073175e-01
9.97352898e-01 3.43483150e-01 -1.65574968e-01 -8.25019121e-01
8.71327445e-02 3.34670916e-02 5.54304533e-02 -7.48111308e-01
6.33484483e-01 -1.42461151e-01 -2.25992963e-01 3.74638826e-01
3.24077338e-01 -1.52646959e-01 1.84612006e-01 5.35068750e-01
6.57574654e-01 7.13913217e-02 1.38524458e-01 2.97177225e-01
6.53141201e-01 -2.15421081e-01 3.06243986e-01 1.04519463e+00
-1.25551596e-01 9.09530103e-01 7.32613921e-01 9.14682299e-02
-5.99704981e-01 -1.11271191e+00 4.30183589e-01 1.44946456e+00
2.21777469e-01 -2.13913798e-01 -6.38936818e-01 -1.09547758e+00
-1.22435004e-01 9.17679369e-01 -4.26247835e-01 -1.10919930e-01
-4.93161619e-01 -6.78437710e-01 5.23158550e-01 5.74944317e-01
8.83552194e-01 -1.30771971e+00 -3.33170325e-01 -8.87845978e-02
-6.29130244e-01 -9.85289693e-01 -6.94158614e-01 -1.14564538e-01
-3.53702486e-01 -8.01032603e-01 -1.09919333e+00 -7.32488513e-01
4.78726000e-01 2.24720553e-01 1.31109440e+00 3.28054070e-01
1.19164579e-01 4.33556348e-01 -3.69999111e-01 -1.98228151e-01
-4.46434259e-01 -1.66841820e-01 -5.92041731e-01 1.17842868e-01
1.38342932e-01 -7.77541026e-02 -6.17158592e-01 4.28506196e-01
-7.59332836e-01 2.36887172e-01 7.86807060e-01 1.13261139e+00
2.60423154e-01 -8.29603314e-01 1.06670165e+00 -7.68591702e-01
9.40069973e-01 -5.38587987e-01 -2.05460027e-01 5.79522431e-01
-3.47667664e-01 -1.35767767e-02 5.05934298e-01 -6.09742403e-01
-1.46504998e+00 -1.63569704e-01 -3.26625794e-01 -4.05871332e-01
-9.33642462e-02 2.40588769e-01 -2.62019455e-01 2.50132412e-01
5.12355447e-01 2.12019175e-01 -1.42061384e-02 -3.31557572e-01
6.18421257e-01 7.32338488e-01 5.14428139e-01 -4.46616232e-01
4.36990112e-01 1.71692401e-01 -5.15699208e-01 -4.18489754e-01
-8.49317014e-01 -3.28412205e-01 -2.03752488e-01 -4.54763621e-01
1.26670539e+00 -8.18929195e-01 -1.00920331e+00 3.31980854e-01
-1.62417567e+00 -4.31123853e-01 1.53865471e-01 1.29569113e-01
-4.73777026e-01 4.05959725e-01 -7.59346426e-01 -1.24247324e+00
-3.54451686e-01 -1.23616385e+00 1.04866982e+00 3.82183969e-01
-4.53272164e-01 -7.44734645e-01 -1.49745554e-01 7.46623933e-01
4.70428377e-01 -9.49653834e-02 1.02417684e+00 -6.28120303e-01
-7.20036745e-01 1.59341976e-01 -6.30664766e-01 1.96922272e-01
-6.12109862e-02 -2.90385097e-01 -1.26865137e+00 -3.19418237e-02
6.06097355e-02 -8.66832018e-01 1.28092778e+00 7.99929723e-02
1.09379923e+00 -4.59771037e-01 4.24370766e-02 5.35232760e-02
9.28035676e-01 2.45764539e-01 6.00412488e-01 -1.11348070e-01
7.56070435e-01 9.04815793e-01 5.74005425e-01 3.94478887e-01
7.35497415e-01 6.64898872e-01 4.63101804e-01 -1.93812475e-01
-2.92683810e-01 -3.54711711e-01 4.94443625e-01 3.54508072e-01
2.84276158e-01 -7.36767292e-01 -6.54635489e-01 3.98841411e-01
-2.03365469e+00 -1.00094080e+00 -3.30533206e-01 1.93361807e+00
9.23499525e-01 -1.64472893e-01 2.32658774e-01 -4.42696095e-01
7.08930135e-01 3.59552860e-01 -4.16506708e-01 -3.45609605e-01
-3.38945270e-01 -2.42869370e-02 -6.74844757e-02 6.29338026e-01
-9.36926782e-01 1.01159143e+00 5.72178221e+00 8.84859920e-01
-7.57432580e-01 4.25358981e-01 9.01142538e-01 -9.49156731e-02
-6.94764376e-01 -6.58550709e-02 -7.70921171e-01 4.14710402e-01
4.78882432e-01 2.99354166e-01 1.91258758e-01 6.18383169e-01
7.30491281e-02 -3.13829362e-01 -9.76278841e-01 1.01369882e+00
2.99621224e-01 -1.18135965e+00 1.87914371e-01 -2.34055176e-01
6.32407546e-01 -4.99835253e-01 3.79241377e-01 7.72399783e-01
3.98030162e-01 -1.13426411e+00 8.41834962e-01 6.24576390e-01
7.79304266e-01 -5.97274601e-01 7.75421083e-01 5.40122211e-01
-9.48420525e-01 -2.33240426e-01 -1.10220566e-01 -1.26905600e-02
4.09562171e-01 2.98080564e-01 -9.58037019e-01 3.84379774e-01
4.19710845e-01 2.81698197e-01 -7.20320463e-01 6.63178325e-01
-5.60493946e-01 2.98901945e-01 2.07855418e-01 -3.16378355e-01
3.23675126e-01 -5.24381585e-02 3.70327681e-01 1.18006468e+00
2.78629363e-01 -3.64319384e-02 3.09428781e-01 1.25043559e+00
-3.18134964e-01 1.01335077e-02 -5.50438285e-01 7.71333724e-02
3.98472250e-01 1.30676234e+00 -2.90968627e-01 -4.80810225e-01
-4.00352448e-01 1.36925197e+00 5.24858654e-01 7.77085900e-01
-9.39600110e-01 -1.28344849e-01 3.00959349e-01 -3.17756325e-01
1.18726864e-01 -3.58395763e-02 -5.01755953e-01 -9.96546328e-01
-1.62813775e-02 -9.53646123e-01 5.77939570e-01 -8.96217406e-01
-1.47928655e+00 6.57309711e-01 -8.68303925e-02 -9.82203305e-01
-6.98725820e-01 -4.86000687e-01 -8.90725553e-01 1.15578806e+00
-1.40411735e+00 -1.40629292e+00 -3.93286258e-01 4.79894131e-01
8.99376929e-01 -7.44827762e-02 6.52080238e-01 3.54042858e-01
-4.38929051e-01 8.37790668e-01 -5.12415051e-01 1.04486644e-01
1.19981253e+00 -1.29811764e+00 -3.33195552e-02 6.21174872e-01
3.50113921e-02 5.83479404e-01 5.40111125e-01 -6.33103251e-01
-1.04945683e+00 -9.41359520e-01 9.48776305e-01 -6.12911165e-01
3.19150597e-01 -4.35259342e-01 -9.38951075e-01 4.87748295e-01
6.55175269e-01 -6.16494060e-01 6.27314091e-01 1.21165635e-02
-4.40674424e-01 2.14297935e-01 -1.08171523e+00 8.06501806e-01
7.96240747e-01 -6.43083751e-01 -4.24276352e-01 2.04209343e-01
9.01110053e-01 -3.20837855e-01 -2.41161585e-01 4.59279507e-01
4.34012592e-01 -1.09710383e+00 8.29390526e-01 -5.55986345e-01
7.29034960e-01 -2.10481837e-01 -2.30345875e-01 -1.06492114e+00
-2.37071499e-01 -5.13269663e-01 -2.89060384e-01 1.48681808e+00
7.08173454e-01 -3.89661103e-01 6.13281429e-01 4.95061278e-01
-2.38723904e-01 -8.55177999e-01 -7.12225497e-01 -3.75750661e-01
-5.59913404e-02 -1.57212511e-01 3.58486891e-01 6.57580793e-01
-3.17215323e-01 7.32937455e-01 -7.73821771e-01 -1.53662428e-01
1.90129682e-01 4.54182774e-01 1.06256700e+00 -6.85810864e-01
-6.15097582e-01 -5.45404971e-01 4.13648486e-01 -1.65122700e+00
2.84537166e-01 -7.57795215e-01 1.58767551e-01 -1.76597822e+00
5.00838518e-01 -1.67473257e-02 1.65176839e-01 3.75496566e-01
-7.74234712e-01 3.88832510e-01 6.59340203e-01 1.70606703e-01
-9.09613848e-01 7.70941019e-01 1.59105372e+00 -2.72535950e-01
-1.50944307e-01 -5.48052229e-02 -8.88104081e-01 5.86967468e-01
6.31324351e-01 -1.58834681e-01 -5.21800756e-01 -5.85404634e-01
1.74107641e-01 2.59650260e-01 7.33586788e-01 -3.82925332e-01
6.99357837e-02 -1.53165951e-01 4.83895242e-01 -8.00061762e-01
5.81568778e-01 -2.49170467e-01 -3.00000727e-01 2.02486008e-01
-5.53335428e-01 2.13252276e-01 -2.94457972e-02 5.21529973e-01
-2.77421325e-01 -3.34691107e-01 6.01216495e-01 -2.40293592e-01
-7.43644953e-01 -1.03532253e-02 -3.75105083e-01 2.21534044e-01
6.93068922e-01 -3.69137377e-02 -7.14882433e-01 -1.02149737e+00
-6.30748570e-01 7.52147794e-01 1.53486535e-01 5.32045007e-01
9.81918573e-01 -1.34936714e+00 -8.79970133e-01 -7.96272755e-02
2.35561490e-01 -2.98297346e-01 6.51923537e-01 9.44049120e-01
-9.09586344e-03 2.76301593e-01 2.12815747e-01 -6.39641047e-01
-1.31146491e+00 3.62401068e-01 2.16700152e-01 -3.91481549e-01
-9.03101191e-02 1.08793056e+00 7.51603782e-01 -6.49285316e-01
2.84453511e-01 1.08176932e-01 -3.93274963e-01 2.67296493e-01
4.36263382e-01 2.19082847e-01 -2.26256609e-01 -3.01234603e-01
-1.86812550e-01 2.57195294e-01 -6.41557276e-02 -3.66224468e-01
7.22686946e-01 -3.60930264e-01 1.80195905e-02 4.56963271e-01
1.05712759e+00 2.59863675e-01 -1.31179607e+00 -1.22045456e-02
-1.66821048e-01 -3.86749744e-01 -5.56839526e-01 -1.16195846e+00
-8.08920801e-01 1.08355165e+00 2.92143911e-01 1.60320416e-01
1.00898945e+00 3.44995856e-01 6.60256863e-01 3.45254391e-01
-3.50104310e-02 -8.74307513e-01 7.77387083e-01 8.09812546e-01
1.24165666e+00 -1.65779924e+00 -5.86363316e-01 -4.23289806e-01
-1.36878157e+00 8.34082723e-01 1.14843106e+00 3.08446676e-01
-1.57054678e-01 3.73143405e-02 4.10298437e-01 -1.56070575e-01
-9.34155226e-01 -3.30031037e-01 4.69533205e-01 4.95788842e-01
7.32234538e-01 -2.13026404e-01 -2.38356635e-01 7.32123673e-01
6.24175519e-02 -4.11068648e-01 1.93981543e-01 4.01321143e-01
-4.00079757e-01 -1.02392685e+00 -3.49172950e-01 3.93249750e-01
-2.88932532e-01 -2.68176079e-01 -9.00971591e-01 6.77550375e-01
-6.90573156e-02 1.48133051e+00 3.04227578e-03 -4.66559589e-01
1.87786028e-01 2.48177052e-01 3.81103575e-01 -6.89212203e-01
-7.44499624e-01 1.22828230e-01 3.70649219e-01 -2.44327232e-01
-2.93405175e-01 -1.55119285e-01 -1.25063610e+00 -4.57586944e-02
-3.16529781e-01 -6.53312430e-02 3.41567881e-02 9.31985140e-01
3.50999892e-01 6.68339312e-01 5.88723660e-01 -6.31129444e-01
-8.41852784e-01 -1.30040252e+00 -3.54286656e-02 3.92022789e-01
1.85402453e-01 -5.41717768e-01 -2.03596875e-01 -2.16346204e-01] | [10.916325569152832, 1.4904075860977173] |
e91c0543-cde1-465c-ae32-b59d9fe9ea16 | a-clip-hitchhiker-s-guide-to-long-video | 2205.08508 | null | https://arxiv.org/abs/2205.08508v1 | https://arxiv.org/pdf/2205.08508v1.pdf | A CLIP-Hitchhiker's Guide to Long Video Retrieval | Our goal in this paper is the adaptation of image-text models for long video retrieval. Recent works have demonstrated state-of-the-art performance in video retrieval by adopting CLIP, effectively hitchhiking on the image-text representation for video tasks. However, there has been limited success in learning temporal aggregation that outperform mean-pooling the image-level representations extracted per frame by CLIP. We find that the simple yet effective baseline of weighted-mean of frame embeddings via query-scoring is a significant improvement above all prior temporal modelling attempts and mean-pooling. In doing so, we provide an improved baseline for others to compare to and demonstrate state-of-the-art performance of this simple baseline on a suite of long video retrieval benchmarks. | ['Andrew Zisserman', 'Gül Varol', 'Arsha Nagrani', 'Max Bain'] | 2022-05-17 | null | null | null | null | ['zero-shot-action-recognition'] | ['computer-vision'] | [ 4.16111238e-02 -6.13156140e-01 -3.68580252e-01 -1.96508214e-01
-1.50378263e+00 -4.75809038e-01 1.08202350e+00 -1.37844637e-01
-8.13526869e-01 3.05663615e-01 6.11592770e-01 1.96244135e-01
-8.82749707e-02 -8.15466344e-02 -8.48834574e-01 -6.57359362e-01
-5.19293427e-01 3.51263657e-02 4.22421396e-01 8.89563337e-02
3.13808948e-01 2.14700520e-01 -1.64012456e+00 7.60589123e-01
-1.25143901e-01 1.08784425e+00 -7.02115893e-02 1.09371579e+00
1.73570797e-01 1.06846499e+00 -5.91394186e-01 -4.93573040e-01
3.08702707e-01 -2.35086799e-01 -7.27016211e-01 -7.33893644e-03
1.31456876e+00 -9.49572504e-01 -1.11498320e+00 4.51968849e-01
5.71697831e-01 2.58684367e-01 7.52064407e-01 -1.16174841e+00
-8.47502053e-01 1.22819170e-01 -6.54168606e-01 7.34022081e-01
6.88044608e-01 1.07922375e-01 1.35744941e+00 -1.24562442e+00
1.01644671e+00 1.40760183e+00 5.66593945e-01 3.93073648e-01
-1.15288687e+00 -5.36028922e-01 2.40422249e-01 4.56338674e-01
-1.62505686e+00 -6.87084258e-01 2.56645709e-01 -4.79792535e-01
1.52691472e+00 -4.09310125e-02 5.72756648e-01 1.29246199e+00
5.25654018e-01 1.16600657e+00 3.46111774e-01 -3.35749716e-01
-2.93685019e-01 -2.24652961e-01 3.70858759e-01 6.08589590e-01
5.09113409e-02 1.07700750e-01 -9.92788613e-01 -2.28288904e-01
6.28565252e-01 1.43618032e-01 -2.47381166e-01 -4.02926385e-01
-1.39582789e+00 5.75908363e-01 2.03182667e-01 2.55979538e-01
-3.66513342e-01 8.89843404e-01 8.19588244e-01 5.85017383e-01
7.22881913e-01 1.69702455e-01 -2.93361962e-01 -4.64233488e-01
-1.66094124e+00 5.90702713e-01 5.58345079e-01 1.12306559e+00
5.48091054e-01 -1.88370466e-01 -7.72648335e-01 5.96656501e-01
7.73238242e-02 4.09841150e-01 5.17686546e-01 -1.28403413e+00
3.09390277e-01 -1.23859420e-01 1.36861682e-01 -8.23144257e-01
2.43722890e-02 7.36376047e-02 -9.28245485e-02 4.32242230e-02
1.39238477e-01 1.82858184e-01 -1.18952703e+00 1.57650578e+00
-4.29380506e-01 6.20966971e-01 -1.98886432e-02 7.63606668e-01
6.06299341e-01 1.05375409e+00 3.44519824e-01 -3.44844580e-01
1.17086482e+00 -1.48518395e+00 -8.62716854e-01 1.38845649e-02
6.89675391e-01 -9.23373282e-01 8.29279840e-01 2.05642924e-01
-1.42477298e+00 -5.87225199e-01 -1.06744802e+00 -4.55844790e-01
-4.86029863e-01 -9.72321630e-02 5.09716511e-01 2.72018522e-01
-1.53993082e+00 6.28175795e-01 -8.11359584e-01 -8.54703426e-01
2.73900777e-01 1.70602679e-01 -5.48307419e-01 -2.41612583e-01
-1.03252375e+00 8.07947755e-01 1.68224454e-01 -3.61173242e-01
-1.12458730e+00 -1.04908907e+00 -6.96672857e-01 2.78170332e-02
3.48041892e-01 -7.21879601e-01 1.51012266e+00 -7.59942591e-01
-9.80229914e-01 1.02612078e+00 -4.34147388e-01 -7.28156149e-01
3.98379385e-01 -7.60323703e-01 -3.05790991e-01 8.23176205e-01
-3.90114821e-02 1.34878063e+00 1.23914099e+00 -8.74193490e-01
-5.62543571e-01 3.12714744e-03 1.77910849e-01 2.79487699e-01
-6.36315882e-01 5.42876244e-01 -1.16397285e+00 -8.23335171e-01
-4.94054586e-01 -9.50591445e-01 2.76096702e-01 3.90409499e-01
4.73424911e-01 -4.84140396e-01 1.19211149e+00 -7.07492530e-01
1.78694451e+00 -2.27542281e+00 1.28004164e-01 -3.86141598e-01
3.02490629e-02 2.22558022e-01 -5.27271867e-01 7.31413662e-01
1.58326409e-04 4.01926994e-01 3.70514780e-01 -6.31339848e-01
-2.84812953e-02 -5.93525916e-02 -6.10385835e-01 4.56087202e-01
4.18959409e-01 1.15496588e+00 -8.52825642e-01 -1.02623129e+00
3.07704955e-01 7.59649098e-01 -6.25660121e-01 1.55459285e-01
-2.98102736e-01 -4.06826228e-01 -3.05469304e-01 5.67154288e-01
1.95050642e-01 -3.22591364e-01 -2.33342275e-01 -2.78641701e-01
1.03364266e-01 4.52911742e-02 -5.31237543e-01 2.32358146e+00
-4.56809662e-02 1.32428038e+00 -3.18078846e-01 -5.39893687e-01
3.59600112e-02 6.53137326e-01 8.42603803e-01 -6.57576084e-01
-2.09206551e-01 -3.49306166e-02 -5.28345883e-01 -7.34604537e-01
1.00005341e+00 3.84723902e-01 2.55271811e-02 4.05010462e-01
5.30039728e-01 -1.67710647e-01 5.67693710e-01 7.43803024e-01
1.23163438e+00 5.73654890e-01 -1.75268993e-01 -1.54604062e-01
3.03986251e-01 -5.97178861e-02 -7.69359097e-02 1.00432944e+00
-3.46829921e-01 1.14391100e+00 2.83923715e-01 -4.45548624e-01
-1.36481571e+00 -1.09434891e+00 7.66977668e-03 1.49673557e+00
-7.34354630e-02 -1.01083422e+00 -5.98852336e-01 -5.59475720e-01
-1.10726217e-02 1.95149153e-01 -6.98342502e-01 1.12798894e-02
-7.20487535e-01 -4.20146585e-01 7.42895424e-01 6.85440302e-01
1.39519095e-01 -8.72465909e-01 -4.61474985e-01 2.02314556e-01
-9.54926237e-02 -1.40722573e+00 -9.61726546e-01 -3.28495860e-01
-7.05126524e-01 -8.62278402e-01 -1.35089469e+00 -7.63647139e-01
6.75422847e-02 7.53737926e-01 1.53817129e+00 3.12054336e-01
-5.29314339e-01 1.18464923e+00 -5.10444403e-01 -1.71383724e-01
1.60824120e-01 -5.43236248e-02 -5.07615618e-02 -3.25680524e-01
6.22883618e-01 -1.04164802e-01 -9.04778302e-01 1.50022283e-01
-1.24770713e+00 -4.19227183e-01 4.98118848e-01 6.38378084e-01
4.70542341e-01 -3.60486686e-01 1.59868851e-01 -1.86721861e-01
6.16226315e-01 -2.98733145e-01 -2.61020035e-01 5.80928981e-01
-4.24449623e-01 -1.36259988e-01 -9.88664776e-02 -5.54373145e-01
-6.71226025e-01 -3.61495137e-01 2.85778999e-01 -1.09146214e+00
2.87061602e-01 3.98961633e-01 7.65635967e-01 -1.01164065e-01
4.13484901e-01 9.83197913e-02 -4.32997048e-02 -1.80032074e-01
5.60772657e-01 4.35603857e-01 4.56609011e-01 -5.49400985e-01
4.19865519e-01 5.41179299e-01 -1.74749732e-01 -9.50347483e-01
-7.15086520e-01 -9.46765900e-01 -6.28190696e-01 -2.28633687e-01
1.21466672e+00 -1.45906591e+00 -1.86680019e-01 3.60727310e-01
-1.08839118e+00 -2.47166738e-01 -1.27834633e-01 4.27861452e-01
-7.60097444e-01 5.86263776e-01 -1.02922511e+00 -5.79433620e-01
-3.97026777e-01 -1.14558125e+00 1.69031239e+00 -1.16543472e-01
-7.68584982e-02 -7.79354274e-01 1.15659364e-01 7.74208978e-02
5.84556997e-01 -1.94844261e-01 3.91592234e-01 -3.69381815e-01
-1.01298976e+00 -4.58071142e-01 -4.14884299e-01 3.08275610e-01
-4.05397713e-01 4.54489052e-01 -9.60507631e-01 -6.85784280e-01
-4.25750375e-01 -6.42274261e-01 1.42961001e+00 6.80073142e-01
1.00281048e+00 9.09485593e-02 -3.90880585e-01 3.88552576e-01
1.52851117e+00 -1.32258059e-02 8.62773180e-01 3.39226723e-01
3.03581983e-01 1.45740882e-01 7.23205388e-01 2.60231853e-01
2.14266181e-01 9.09703076e-01 -1.54908046e-01 2.79752672e-01
-4.92759854e-01 -3.35942209e-02 7.21648157e-01 5.43740392e-01
8.12923312e-02 -6.00067377e-01 -6.84964001e-01 7.95812130e-01
-2.11731434e+00 -1.37998235e+00 3.34273130e-01 1.84700072e+00
5.33342183e-01 1.49750516e-01 9.73786339e-02 -3.95977676e-01
3.57110471e-01 8.30097616e-01 -8.00063163e-02 -1.72148749e-01
-1.61182597e-01 7.18541667e-02 4.90083486e-01 3.46517801e-01
-1.36492884e+00 1.05749166e+00 7.88927889e+00 9.80404496e-01
-8.61051261e-01 1.77868709e-01 3.66157115e-01 -7.89693594e-01
1.28916740e-01 -7.07699358e-02 -8.79627049e-01 1.93543479e-01
1.29655695e+00 -3.58311206e-01 3.08777541e-01 5.31618595e-01
1.76327035e-01 -2.17310950e-01 -1.60840261e+00 1.25104117e+00
7.48070955e-01 -1.49984205e+00 4.39020365e-01 -1.44957993e-02
9.68652129e-01 3.09674114e-01 1.55731887e-01 4.74454224e-01
-2.48562962e-01 -7.20198989e-01 6.81753635e-01 7.40428507e-01
8.04012775e-01 -1.87563524e-01 5.17098069e-01 -3.96246672e-01
-1.23527753e+00 -1.50265619e-01 -3.00673783e-01 2.99259752e-01
4.74026889e-01 6.05684146e-02 -2.33790278e-01 4.01125312e-01
9.33407962e-01 1.07310903e+00 -8.70999515e-01 1.28626776e+00
3.67286682e-01 3.67939800e-01 -1.78149402e-01 1.98627144e-01
6.67535007e-01 2.64883935e-01 4.04077381e-01 1.81104803e+00
3.34938079e-01 4.05563377e-02 2.71032572e-01 3.49034406e-02
-2.86341280e-01 1.53077561e-02 -8.72706652e-01 -5.36544979e-01
4.13171388e-02 8.19590449e-01 -2.97600687e-01 -7.96955705e-01
-9.33426917e-01 9.67432141e-01 3.75279449e-02 7.89568126e-01
-9.43872333e-01 -2.56936044e-01 8.17139268e-01 -1.54168857e-02
6.21306002e-01 -4.52645540e-01 4.01213259e-01 -1.23121107e+00
1.36596784e-01 -5.83454370e-01 6.54512227e-01 -1.38193500e+00
-1.23753262e+00 4.03514743e-01 4.57783133e-01 -1.13288748e+00
-5.85332751e-01 -5.35338938e-01 -2.58680105e-01 5.30369282e-01
-1.49967861e+00 -1.12103331e+00 -2.39705682e-01 6.44655824e-01
1.13404703e+00 -1.29522324e-01 6.73158109e-01 4.87028122e-01
-1.26942247e-01 5.50282657e-01 1.97078630e-01 9.15994123e-02
1.35715783e+00 -8.62162530e-01 2.64352500e-01 8.18030775e-01
4.25130934e-01 5.44022501e-01 5.84115505e-01 -4.47466403e-01
-1.61635363e+00 -7.85488367e-01 7.87863791e-01 -8.19339216e-01
1.08046830e+00 -2.40211666e-01 -7.35506833e-01 1.04176331e+00
8.56304884e-01 1.12471007e-01 3.53201002e-01 -1.80290222e-01
-5.54246902e-01 -1.33238912e-01 -5.36828935e-01 6.67975307e-01
9.08071816e-01 -1.12399578e+00 -7.43675411e-01 5.74690163e-01
1.04215014e+00 -5.31941392e-02 -9.50656593e-01 4.23188776e-01
1.06274509e+00 -7.17821419e-01 1.29726744e+00 -7.87534118e-01
6.75591290e-01 -1.99816283e-02 -3.21533114e-01 -6.31420255e-01
-3.17547172e-01 -7.38025844e-01 -4.02686566e-01 1.02595627e+00
1.12432577e-01 2.62156993e-01 8.10608387e-01 5.93569100e-01
6.98032370e-03 -6.06632710e-01 -7.42391825e-01 -8.78322721e-01
1.66497827e-01 -4.87180710e-01 -1.61666572e-01 4.29407179e-01
-6.32447973e-02 5.53543940e-02 -5.63244343e-01 -1.96328342e-01
5.11398792e-01 -3.87574464e-01 7.85068393e-01 -6.42299652e-01
-1.81631558e-02 -6.42310083e-01 -7.08401322e-01 -1.36030018e+00
3.35035264e-01 -4.58681107e-01 -8.96839611e-03 -1.56347275e+00
6.41359091e-01 3.84657800e-01 -7.10647702e-01 2.25196127e-02
-1.08296320e-01 6.96820498e-01 4.51961964e-01 4.37579542e-01
-1.35089231e+00 2.98670739e-01 9.59485412e-01 -3.42400551e-01
1.43950105e-01 -7.97177792e-01 -5.64199574e-02 2.65748382e-01
1.28369704e-01 -2.83496857e-01 -5.03563464e-01 -9.78657544e-01
3.85163026e-03 9.34124924e-03 4.07054335e-01 -1.09735954e+00
4.14407641e-01 2.93446541e-01 3.31230253e-01 -9.20414448e-01
8.43572915e-01 -7.40548313e-01 -1.43124431e-01 3.82334478e-02
-6.21169209e-01 5.28986394e-01 3.53155553e-01 7.57490277e-01
-5.28931737e-01 -1.72392935e-01 2.26212904e-01 -1.96515009e-01
-1.17384958e+00 5.12142599e-01 -4.98808622e-01 4.69257534e-02
8.81564677e-01 -1.51745915e-01 -5.30312955e-01 -6.90821350e-01
-6.01537108e-01 4.04973328e-01 5.63038945e-01 8.84144306e-01
6.62701368e-01 -1.33566630e+00 -7.58494854e-01 -1.54277951e-01
3.45414966e-01 -7.95778394e-01 2.26177663e-01 8.15284252e-01
-4.88385439e-01 9.19373155e-01 8.84530097e-02 -9.58983362e-01
-1.57538819e+00 8.94270122e-01 -4.79964539e-03 -3.20405215e-01
-6.02718353e-01 8.19760680e-01 2.17084229e-01 6.77352488e-01
7.88528442e-01 -1.51508108e-01 1.85983464e-01 2.53881782e-01
8.69603515e-01 1.18244670e-01 -1.84200853e-02 -5.08682430e-01
-2.48431206e-01 8.61188889e-01 -4.63402838e-01 -4.53595906e-01
1.17646027e+00 -3.34960222e-01 1.23515226e-01 5.44297397e-01
1.73511481e+00 -4.39392447e-01 -1.23810112e+00 -1.95322663e-01
1.66113246e-02 -6.43454075e-01 1.18301444e-01 -4.53325927e-01
-7.10912526e-01 1.04398453e+00 7.00248420e-01 2.74274759e-02
1.07389009e+00 1.05130240e-01 6.78330481e-01 8.40884387e-01
3.76715273e-01 -1.09098577e+00 4.45525318e-01 5.39746761e-01
9.09951389e-01 -1.24650073e+00 4.33075607e-01 1.32677123e-01
-4.26530033e-01 1.14198565e+00 2.75493443e-01 -5.63760579e-01
8.29302192e-01 1.39610589e-01 -1.72533020e-02 -2.21552223e-01
-1.34084225e+00 -1.42657027e-01 4.50253189e-01 1.85908526e-01
7.09417105e-01 -5.52620351e-01 -3.42390805e-01 -3.32942158e-01
5.24230838e-01 2.69984603e-01 3.07523489e-01 1.19790614e+00
-2.70344436e-01 -8.77125263e-01 -1.26743689e-01 4.54276174e-01
-1.05096400e+00 -4.58232492e-01 -2.66190737e-01 1.08491671e+00
-6.45486653e-01 7.78850555e-01 3.50446641e-01 -2.56058544e-01
7.43700191e-02 4.74829108e-01 8.01975548e-01 -2.99096555e-01
-4.16725546e-01 5.07489383e-01 1.22250810e-01 -1.03152990e+00
-9.44237530e-01 -6.88925982e-01 -7.17864275e-01 -1.51671782e-01
-2.78901160e-01 2.04269662e-02 5.28958678e-01 6.82771146e-01
4.46510106e-01 3.57587010e-01 1.82893232e-01 -1.40677571e+00
-3.89153779e-01 -9.74053621e-01 -4.03805017e-01 7.16834843e-01
5.60247600e-01 -4.96655107e-01 -3.25737357e-01 5.18589497e-01] | [10.284184455871582, 0.878159761428833] |
6da75d63-0a80-41a4-a96a-63dc3c42fd86 | human-parsing-based-texture-transfer-from | null | null | http://proceedings.neurips.cc/paper/2020/hash/a516a87cfcaef229b342c437fe2b95f7-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/a516a87cfcaef229b342c437fe2b95f7-Paper.pdf | Human Parsing Based Texture Transfer from Single Image to 3D Human via Cross-View Consistency | This paper proposes a human parsing based texture transfer model via cross-view consistency learning to generate the texture of 3D human body from a single image. We use the semantic parsing of human body as input for providing both the shape and pose information to reduce the appearance variation of human image and preserve the spatial distribution of semantic parts. Meanwhile, in order to improve the prediction for textures of invisible parts, we explicitly enforce the consistency across different views of the same subject by exchanging the textures predicted by two views to render images during training. The perception loss and total variation regularization are optimized to maximize the similarity between rendered and input images, which does not necessitate extra 3D texture supervision. Experimental results on pedestrian images and fashion photos demonstrate that our method can produce higher quality textures with convincing details than other texture generation methods. | ['Ling Shao', 'Kaihao Zhang', 'Shengcai Liao', 'Fang Zhao'] | 2020-12-01 | null | null | null | neurips-2020-12 | ['human-parsing', 'image-to-3d'] | ['computer-vision', 'computer-vision'] | [ 1.97632790e-01 4.89377528e-01 4.98105250e-02 -5.00878274e-01
-2.50958413e-01 -3.91529769e-01 2.78063208e-01 -6.60402536e-01
1.81494549e-01 3.48834604e-01 1.44713402e-01 3.80845577e-01
3.45705032e-01 -8.64250124e-01 -9.57923949e-01 -7.26044834e-01
5.56899369e-01 3.61103535e-01 2.53392875e-01 -2.23091364e-01
-2.02244133e-01 1.03188053e-01 -1.42469978e+00 5.93501568e-01
5.86063802e-01 1.16741860e+00 3.62698376e-01 2.17411891e-01
1.12352863e-01 4.75522041e-01 -3.02659005e-01 -6.87461555e-01
6.99601531e-01 -5.12610674e-01 -5.25025547e-01 9.41192627e-01
8.73126626e-01 -5.11963487e-01 -2.40058199e-01 9.26477909e-01
4.83788341e-01 3.18072475e-02 4.26823258e-01 -9.31660473e-01
-8.68219495e-01 -1.66678727e-02 -9.66397703e-01 -4.76319015e-01
4.31400687e-01 1.99521467e-01 6.03250444e-01 -5.66723406e-01
8.94764721e-01 1.76018918e+00 3.91780376e-01 1.01713157e+00
-1.39731324e+00 -4.52661246e-01 3.81962359e-01 -1.18107170e-01
-9.23698843e-01 -2.38144413e-01 1.16572726e+00 -2.91291416e-01
9.64576378e-02 2.56552845e-01 9.37301636e-01 1.26136303e+00
3.05895686e-01 7.95168102e-01 1.39332628e+00 -3.11033130e-01
-1.06596790e-01 1.70752406e-01 -4.67946231e-01 1.20097113e+00
-5.07531827e-03 1.33151263e-01 -6.48394883e-01 4.82661165e-02
1.49114704e+00 -6.99154213e-02 -7.28522763e-02 -9.58136618e-01
-1.11642349e+00 4.31118280e-01 2.55061179e-01 -3.67775530e-01
-1.08069107e-01 -3.05600520e-02 2.87582755e-01 1.35910273e-01
7.91689932e-01 -1.34766653e-01 -3.65948200e-01 4.27504927e-01
-1.46852911e-01 2.40117550e-01 4.75980431e-01 1.15764093e+00
6.72524393e-01 7.35457987e-02 -1.19216286e-01 1.03333735e+00
3.03542644e-01 9.29031014e-01 3.35220955e-02 -1.11549890e+00
7.00954795e-01 7.04277337e-01 9.94224325e-02 -1.29423714e+00
-3.73240337e-02 1.96536779e-02 -9.20199156e-01 3.49644482e-01
4.44809616e-01 5.44865318e-02 -1.17486596e+00 1.67525101e+00
8.92553210e-01 -3.94958764e-01 -4.15228337e-01 1.41924214e+00
7.67476201e-01 3.50758523e-01 5.48721254e-02 2.85878479e-01
1.43980324e+00 -1.03364837e+00 -5.81945062e-01 -3.32439810e-01
-6.55532554e-02 -8.53789091e-01 1.34082592e+00 2.78589576e-01
-1.32821965e+00 -9.38096404e-01 -8.62509906e-01 -2.53781110e-01
3.05486143e-01 3.74724120e-01 6.14922404e-01 6.14566565e-01
-5.93897521e-01 4.32003498e-01 -9.44080591e-01 -2.02831671e-01
3.83529902e-01 1.67236492e-01 -5.94273984e-01 -1.81541339e-01
-7.54524350e-01 6.00072622e-01 3.31277214e-02 1.67504102e-01
-7.41941392e-01 -4.86107439e-01 -1.10769677e+00 -3.41198981e-01
2.26114944e-01 -1.24089813e+00 6.70390129e-01 -1.25371981e+00
-1.84343660e+00 1.41673982e+00 -1.24265835e-01 2.35878602e-01
7.87768960e-01 -2.66550064e-01 -8.36675465e-02 1.70108184e-01
2.44511932e-01 9.39563394e-01 1.12626934e+00 -1.50955307e+00
-2.68779784e-01 -5.80721855e-01 -1.32636234e-01 5.22586942e-01
2.50840068e-01 -3.34022105e-01 -8.70000720e-01 -8.68307590e-01
4.88913894e-01 -9.89599288e-01 -9.97058302e-02 4.12501365e-01
-4.69934523e-01 2.54558235e-01 5.84743917e-01 -1.07871246e+00
3.01442623e-01 -2.19247866e+00 5.11184752e-01 3.41425300e-01
4.51650321e-02 -4.91841614e-01 -2.61889070e-01 -1.96659327e-01
1.87861443e-01 -2.82021940e-01 5.71826287e-02 -5.25391161e-01
-7.43178725e-02 4.52371150e-01 -3.09398860e-01 4.58377391e-01
8.58212784e-02 8.76112401e-01 -6.56445146e-01 -6.16702914e-01
4.27979439e-01 5.58093846e-01 -8.01954031e-01 4.61854905e-01
-2.97318220e-01 1.08016145e+00 -7.35785604e-01 4.92531300e-01
8.77811253e-01 -3.41786683e-01 3.66681606e-01 -5.63319325e-01
5.77863097e-01 2.82037426e-02 -1.01489353e+00 2.01608491e+00
-4.72887248e-01 -1.84389576e-01 1.61452666e-01 -7.35592544e-01
9.03596938e-01 1.70741141e-01 4.68614906e-01 -1.12003982e+00
1.28385037e-01 -1.69834286e-01 -3.93307000e-01 -4.45163429e-01
-7.71807740e-03 -2.90559500e-01 -9.24883559e-02 3.54185075e-01
-1.03708908e-01 -2.83369154e-01 -1.70856342e-01 -1.91605017e-01
1.99070558e-01 7.50830710e-01 -2.95994610e-01 -1.05183795e-01
4.21375871e-01 -3.85569274e-01 8.12229455e-01 2.74861306e-01
2.12140307e-01 1.09159005e+00 3.45255792e-01 -8.31853032e-01
-1.46394718e+00 -1.58895350e+00 1.15935363e-01 9.33915198e-01
5.91723561e-01 -3.45881954e-02 -9.31315243e-01 -8.17228556e-01
-4.65012006e-02 -5.73294237e-02 -8.77240777e-01 1.13472052e-01
-6.86210394e-01 -3.88860106e-01 -1.30435440e-03 4.95597422e-01
7.63907671e-01 -9.14432824e-01 -5.40366709e-01 -1.12676822e-01
-6.71094060e-01 -1.18842232e+00 -9.38451767e-01 -4.10279483e-01
-7.98842907e-01 -1.11758924e+00 -8.71355653e-01 -9.31599796e-01
1.22816312e+00 4.28295359e-02 1.03147674e+00 -2.11841613e-02
-4.96462822e-01 3.75716120e-01 -1.61271647e-01 8.31931531e-02
-2.78233320e-01 -4.71751869e-01 7.32829515e-03 2.95254409e-01
-2.78304070e-01 -5.82250655e-01 -8.31758678e-01 6.84155405e-01
-5.90424776e-01 7.48593688e-01 3.97930264e-01 1.12263215e+00
9.10109282e-01 -1.33990422e-01 -1.00434914e-01 -7.21966505e-01
3.54284905e-02 2.22827554e-01 -5.18440187e-01 4.55451071e-01
-1.35616377e-01 1.75674371e-02 3.12930584e-01 -5.66549659e-01
-1.66140878e+00 2.29721442e-01 1.77866682e-01 -5.83043575e-01
-2.08755776e-01 -3.68637472e-01 -5.42373955e-01 -2.26933584e-02
2.30623350e-01 2.88287193e-01 3.43348533e-01 -5.70167422e-01
4.66752499e-01 4.44792584e-02 3.77904415e-01 -9.63560343e-01
6.84553862e-01 8.64709198e-01 -4.06398736e-02 -3.73496652e-01
-1.11753404e+00 3.69530097e-02 -7.88865149e-01 -2.58157253e-01
1.19889462e+00 -1.16400075e+00 -7.14816391e-01 5.19787550e-01
-1.00976408e+00 -3.61129194e-01 -2.38816753e-01 1.68644145e-01
-9.18739080e-01 4.71200824e-01 -8.62201333e-01 -4.75022763e-01
-3.23543191e-01 -1.03276801e+00 1.73082149e+00 3.78977247e-02
-1.38434872e-01 -8.96481872e-01 -3.53830904e-01 8.85804057e-01
-7.80385584e-02 3.79053026e-01 1.04153287e+00 4.28142697e-01
-7.17055321e-01 3.43036890e-01 -1.30248830e-01 4.29876983e-01
4.13820475e-01 -5.23636520e-01 -8.38113487e-01 -2.00237021e-01
-8.11066665e-03 -4.34938103e-01 4.38501269e-01 5.29940844e-01
1.18780172e+00 -3.93592983e-01 -1.18721172e-01 7.88709044e-01
9.99259710e-01 -2.08486449e-02 6.31870508e-01 -1.17917620e-02
1.07196462e+00 1.14868486e+00 5.57656348e-01 3.60951871e-01
3.09762180e-01 7.72509098e-01 -2.44111326e-02 -4.41695511e-01
-4.78685439e-01 -8.19740176e-01 1.78500071e-01 6.94889843e-01
-2.33921245e-01 1.47937182e-02 -3.21851611e-01 7.25028142e-02
-1.62620354e+00 -7.94152498e-01 2.32575819e-01 2.31732392e+00
7.48216569e-01 -7.82914758e-02 1.32140428e-01 -3.35171789e-01
7.64581382e-01 -8.38920847e-02 -4.44968849e-01 -7.84927085e-02
-1.58628598e-01 8.37686956e-02 2.95182616e-01 4.15121853e-01
-1.03664649e+00 1.03273642e+00 6.19679213e+00 6.88958347e-01
-9.25373316e-01 -4.78449203e-02 1.09515166e+00 -1.50661469e-01
-4.53983158e-01 -2.82465816e-01 -5.06878972e-01 4.82591599e-01
-8.82337838e-02 4.20192122e-01 3.84450108e-01 6.75793350e-01
1.59485519e-01 -6.14240253e-03 -9.47099209e-01 1.18474448e+00
1.56966627e-01 -9.88802314e-01 4.68920410e-01 1.05363972e-01
9.36282754e-01 -6.80604637e-01 3.88090074e-01 -2.55760044e-01
1.88066900e-01 -7.43526578e-01 9.98956203e-01 6.43046558e-01
7.99094081e-01 -5.48017740e-01 3.39571834e-01 8.38224962e-02
-1.07407701e+00 2.62400389e-01 -4.88028020e-01 5.69206774e-02
3.35395634e-01 5.71260810e-01 -2.02318817e-01 5.65174580e-01
9.29402947e-01 6.23143911e-01 -4.49564099e-01 4.56926346e-01
-2.66604662e-01 3.10720205e-02 -1.16843976e-01 5.27047038e-01
-3.76851052e-01 -6.06943846e-01 2.85498857e-01 4.20780301e-01
8.17463621e-02 1.15143172e-01 4.62658852e-01 1.15003908e+00
2.72580117e-01 1.06105514e-01 -4.22294199e-01 4.74418879e-01
3.00286934e-02 1.02842653e+00 -6.46023393e-01 -2.59994388e-01
-2.37739801e-01 1.59487653e+00 1.69900134e-01 4.65021580e-01
-8.37467134e-01 2.93985575e-01 3.99284601e-01 4.43629086e-01
3.26732367e-01 -1.41184963e-02 -4.60378170e-01 -1.32932949e+00
4.31787491e-01 -8.07827890e-01 1.45972818e-01 -1.06237078e+00
-1.67076623e+00 3.48015010e-01 -1.04433037e-01 -1.42178965e+00
5.51880198e-03 -5.34345210e-01 -1.84432954e-01 1.08821154e+00
-8.82911384e-01 -1.77940309e+00 -2.96697140e-01 6.16425991e-01
4.37076300e-01 8.13273117e-02 5.72859287e-01 1.73255101e-01
-2.02624694e-01 5.96200347e-01 -5.69347799e-01 1.15186319e-01
8.74153435e-01 -1.06627500e+00 5.25802612e-01 2.64036179e-01
-1.41377330e-01 4.30538386e-01 4.15850222e-01 -9.19259965e-01
-1.35159206e+00 -9.69976783e-01 4.46022928e-01 -5.96798837e-01
-3.84521820e-02 -6.00009859e-01 -5.89688778e-01 5.87503254e-01
4.46155071e-02 1.12752900e-01 3.27412605e-01 4.14884184e-03
-5.26735306e-01 -1.76421955e-01 -1.13479805e+00 8.01906765e-01
1.32573318e+00 -3.69430870e-01 -3.51948589e-01 9.62337926e-02
3.94606173e-01 -9.83170867e-01 -9.37210083e-01 5.18586636e-01
8.99489045e-01 -1.00793219e+00 1.21273148e+00 -4.29896265e-01
6.86092377e-01 -3.04292321e-01 -6.95901439e-02 -1.11705887e+00
-2.95882761e-01 -3.03772092e-01 3.34310532e-01 1.07010472e+00
1.04345731e-01 -3.83553892e-01 1.04225874e+00 7.82566786e-01
1.14615403e-01 -6.27140820e-01 -7.90969849e-01 -6.75241888e-01
-7.70132011e-03 1.67548835e-01 4.12411451e-01 7.40285277e-01
-2.22919747e-01 2.30276331e-01 -9.32849050e-01 1.27820820e-01
1.04042852e+00 5.53555608e-01 1.14045405e+00 -8.92036676e-01
-5.81249297e-01 9.06414762e-02 -4.28072125e-01 -1.23851991e+00
1.54014096e-01 -5.91193855e-01 2.94552203e-02 -1.15738499e+00
4.89538312e-01 -4.91063088e-01 1.21623993e-01 3.97879660e-01
-2.45978460e-01 4.79871541e-01 1.12052649e-01 -7.70838708e-02
-4.10645902e-01 8.81023407e-01 2.22999048e+00 -9.14762318e-02
5.61149344e-02 -1.19847924e-01 -4.41250950e-01 8.50403965e-01
4.00048196e-01 -1.45219341e-01 -6.50461793e-01 -6.92647934e-01
3.71892331e-03 3.02142292e-01 8.01778972e-01 -6.30612731e-01
-3.74902189e-01 -3.10347468e-01 1.17691350e+00 -3.62200260e-01
4.96530652e-01 -8.23496819e-01 3.78195584e-01 2.44334280e-01
-3.20408791e-01 -6.02991171e-02 8.62149224e-02 8.21550250e-01
-5.14804665e-03 6.47888422e-01 8.75278592e-01 -4.21231419e-01
-4.35034782e-01 5.90956986e-01 1.46246552e-01 -2.53977608e-02
7.18411565e-01 -2.76701421e-01 1.67478453e-02 -2.79337704e-01
-1.13616765e+00 6.85513839e-02 8.81387174e-01 7.12207079e-01
8.66284609e-01 -1.52800643e+00 -5.33608913e-01 6.34880602e-01
5.18997982e-02 9.88484025e-02 7.91454196e-01 4.89866853e-01
-4.54088271e-01 6.04963824e-02 -6.01882994e-01 -7.78015912e-01
-1.36013138e+00 4.76546496e-01 5.43832898e-01 -8.90308153e-03
-1.11175907e+00 6.67224884e-01 1.17498875e+00 -7.88543999e-01
7.93232098e-02 -1.10919155e-01 3.76256555e-01 -7.69000292e-01
3.57448012e-01 5.35128191e-02 -3.30339968e-01 -5.46315849e-01
-4.33151573e-02 1.04914069e+00 -3.81348259e-03 -2.02609524e-01
1.06548870e+00 -5.56785047e-01 1.20424358e-02 2.54687577e-01
9.73667085e-01 1.39929041e-01 -1.86035681e+00 -9.27819237e-02
-6.63111925e-01 -8.30531597e-01 -4.01031494e-01 -9.17388439e-01
-1.42336214e+00 8.41439843e-01 8.48371625e-01 -5.45632005e-01
1.08011377e+00 9.14496928e-02 8.65197301e-01 -2.05110312e-01
8.59179795e-01 -1.17008007e+00 4.85478044e-01 8.39423388e-02
1.05437291e+00 -1.22527170e+00 -1.39202373e-02 -1.05161119e+00
-9.57546294e-01 7.89180756e-01 1.05967879e+00 -2.53795087e-01
4.51481074e-01 4.04860359e-03 2.04195693e-01 -2.06332594e-01
-5.32773137e-01 8.34407136e-02 7.31018722e-01 9.31753695e-01
3.18708748e-01 8.67308080e-02 -1.28535494e-01 4.08971786e-01
-1.02539487e-01 -3.30687016e-01 -1.49847388e-01 4.95865345e-01
-1.22987246e-02 -1.32030785e+00 -3.91761601e-01 5.91749996e-02
-4.22038645e-01 1.49241939e-01 -3.94930422e-01 6.49908721e-01
3.60499561e-01 5.91736317e-01 2.73679495e-02 -3.24792475e-01
5.20326793e-01 -1.40581563e-01 1.22992492e+00 -5.66211164e-01
-1.28882200e-01 4.86899525e-01 1.10130347e-01 -7.96458721e-01
-4.37136352e-01 -3.83826077e-01 -9.32745636e-01 -2.77542502e-01
5.06051555e-02 -3.42055142e-01 1.74513862e-01 7.43830025e-01
3.44858617e-01 3.29495102e-01 6.59422100e-01 -8.65547955e-01
-2.74216026e-01 -5.66432059e-01 -9.64724481e-01 1.09545863e+00
9.72614437e-02 -8.59245598e-01 -7.41965622e-02 5.89560330e-01] | [11.954845428466797, -0.8686598539352417] |
3fe4c804-e1af-47c8-a464-57180b86ba33 | deep-scattering-transform-applied-to-note | 1703.09775 | null | http://arxiv.org/abs/1703.09775v1 | http://arxiv.org/pdf/1703.09775v1.pdf | Deep scattering transform applied to note onset detection and instrument recognition | Automatic Music Transcription (AMT) is one of the oldest and most
well-studied problems in the field of music information retrieval. Within this
challenging research field, onset detection and instrument recognition take
important places in transcription systems, as they respectively help to
determine exact onset times of notes and to recognize the corresponding
instrument sources. The aim of this study is to explore the usefulness of
multiscale scattering operators for these two tasks on plucked string
instrument and piano music. After resuming the theoretical background and
illustrating the key features of this sound representation method, we evaluate
its performances comparatively to other classical sound representations. Using
both MIDI-driven datasets with real instrument samples and real musical pieces,
scattering is proved to outperform other sound representations for these AMT
subtasks, putting forward its richer sound representation and invariance
properties. | ['O. Adam', 'D. Cazau', 'G. Revillon'] | 2017-03-28 | null | null | null | null | ['instrument-recognition', 'music-transcription'] | ['audio', 'music'] | [ 5.82062900e-01 -4.38378304e-01 -1.34758621e-01 2.68487990e-01
-9.09050047e-01 -8.80729795e-01 4.95629579e-01 2.40997508e-01
-1.19864270e-01 3.21041346e-01 3.14206094e-01 1.28227443e-01
-7.78879941e-01 -3.35756868e-01 -9.54563245e-02 -7.60848165e-01
-2.61989892e-01 2.87181288e-01 -1.02419212e-01 -3.04884374e-01
4.76550579e-01 4.59879309e-01 -1.83860588e+00 2.32323959e-01
5.10572374e-01 1.03514504e+00 -4.98079062e-02 7.99102962e-01
-1.69576451e-01 6.50856793e-01 -7.30712056e-01 -3.08987588e-01
1.52519092e-01 -6.74842536e-01 -7.04460859e-01 -2.88167953e-01
3.06096315e-01 3.28857034e-01 -6.27241507e-02 9.46702540e-01
7.57248700e-01 5.49958229e-01 8.75964940e-01 -7.68279314e-01
-2.02641651e-01 1.10375786e+00 -4.04741734e-01 3.45327497e-01
7.32169509e-01 -4.63396221e-01 1.53218555e+00 -8.41698229e-01
3.59693468e-01 8.31768394e-01 8.45554650e-01 -7.55122118e-03
-1.29494452e+00 -7.48615861e-01 -5.02580822e-01 5.11711597e-01
-1.42503309e+00 -5.84917128e-01 1.08144045e+00 -4.33512211e-01
4.96791363e-01 6.55573010e-01 7.30391622e-01 9.43917096e-01
-1.11218974e-01 8.69931638e-01 1.00124574e+00 -6.80766404e-01
8.67095664e-02 -4.90051448e-01 2.84228712e-01 1.09442823e-01
-2.22397834e-01 3.65373679e-02 -1.05760789e+00 -2.17084453e-01
7.44693220e-01 -2.56123722e-01 -5.82162201e-01 3.71529385e-02
-1.50661111e+00 4.18054730e-01 1.61992729e-01 7.74896443e-01
-4.61083710e-01 5.56266308e-02 6.24007642e-01 3.46215755e-01
7.49528110e-02 8.05068612e-01 -2.87795030e-02 -6.62361741e-01
-1.23848593e+00 6.63720906e-01 6.56369984e-01 3.40604305e-01
5.57717159e-02 2.66141862e-01 -5.33103645e-01 9.92134213e-01
-1.63159713e-01 4.42968309e-01 6.83215797e-01 -7.45672643e-01
3.70404959e-01 2.21879885e-01 2.07211614e-01 -1.10750329e+00
-3.58226627e-01 -8.40432167e-01 -9.39809978e-01 -2.31933892e-02
6.25216126e-01 3.66647571e-01 -1.75413176e-01 1.43585229e+00
-1.32053614e-01 4.64163572e-01 -1.23958871e-01 1.01729846e+00
7.38582194e-01 8.06116223e-01 -2.68916696e-01 -4.62035388e-01
1.62313342e+00 -5.74602723e-01 -8.60565484e-01 3.61587375e-01
1.37504712e-02 -1.28113091e+00 9.98488247e-01 9.23589528e-01
-1.06780875e+00 -9.17227507e-01 -9.19686079e-01 -2.16436554e-02
1.55757710e-01 4.94866222e-01 6.70607686e-01 5.33492744e-01
-3.54326963e-01 9.82891142e-01 -3.24221224e-01 1.36329472e-01
-2.44230442e-02 1.07661806e-01 -2.66257091e-03 3.69543940e-01
-1.05881166e+00 1.76672652e-01 7.59486333e-02 1.28614992e-01
-3.75548899e-01 -7.21553326e-01 -2.26847664e-01 2.95383066e-01
2.62035966e-01 -3.90241653e-01 1.52649033e+00 -8.05098534e-01
-1.85824549e+00 6.09845757e-01 -6.34717569e-02 -5.31000257e-01
2.11834937e-01 -5.78371763e-01 -4.92676347e-01 1.56676650e-01
-8.71341825e-02 -1.61188394e-01 1.32509887e+00 -6.34112000e-01
-3.23486328e-01 -3.19556147e-01 -4.58001733e-01 1.82693437e-01
-2.27990791e-01 1.73875585e-01 -1.93505779e-01 -1.36030483e+00
3.95961285e-01 -9.30712104e-01 1.97555870e-01 -6.83385253e-01
-5.75729489e-01 -3.68266612e-01 3.11898619e-01 -7.00848460e-01
1.65229714e+00 -2.37985325e+00 5.91247737e-01 4.09465164e-01
-1.65813163e-01 1.77975267e-01 -1.29537452e-02 7.50955880e-01
-2.09109530e-01 -4.71858948e-01 -1.09841399e-01 -1.01045385e-01
2.88279921e-01 -1.51953742e-01 -9.18162048e-01 1.24753892e-01
-2.39742070e-01 7.92633057e-01 -6.52200401e-01 -1.56685233e-01
4.49244566e-02 4.52472627e-01 -2.12110877e-01 1.08936243e-02
-8.08648691e-02 7.31500149e-01 -3.76677155e-01 4.67025936e-01
1.31915435e-01 4.16781753e-01 1.02221742e-02 -3.49955976e-01
-3.24156582e-01 5.30350506e-01 -1.53778183e+00 1.78231001e+00
-2.93855160e-01 6.49503887e-01 -8.28972906e-02 -8.50318611e-01
1.22042680e+00 5.32726943e-01 5.97400486e-01 -5.20449340e-01
1.27861574e-01 4.73036230e-01 1.79913715e-01 -4.49150175e-01
6.74733877e-01 -3.32933992e-01 -2.17852771e-01 4.76902694e-01
-6.33121356e-02 -2.78898805e-01 1.75422430e-01 -4.07548487e-01
8.50642264e-01 1.66975811e-01 3.71157438e-01 -4.49967803e-03
6.81039631e-01 -4.16146398e-01 2.78831959e-01 4.78562444e-01
2.20891804e-01 7.76565909e-01 2.85647005e-01 -3.01126659e-01
-4.69956666e-01 -1.01399124e+00 -1.50259569e-01 1.40189874e+00
-1.99473560e-01 -7.59048462e-01 -6.21158302e-01 1.91002972e-02
-7.50867426e-02 4.40206796e-01 -4.17776287e-01 4.47406061e-02
-6.44733071e-01 -3.41607451e-01 8.96697044e-01 4.90278572e-01
2.25418419e-01 -1.24459410e+00 -5.92795789e-01 3.93874556e-01
-6.13466978e-01 -5.94730377e-01 -4.63323116e-01 2.10610792e-01
-8.76754224e-01 -1.09884727e+00 -1.05690157e+00 -5.89409411e-01
-3.37040573e-01 1.87293857e-01 9.17482316e-01 -3.09251279e-01
-4.62509632e-01 3.72376561e-01 -5.86889625e-01 -5.34460604e-01
-3.17496657e-01 3.60595971e-01 1.86371505e-01 4.82937634e-01
-8.55685622e-02 -1.01979959e+00 -4.45268184e-01 2.76547253e-01
-9.40905809e-01 -1.35391116e-01 5.46038508e-01 4.73273247e-01
7.29722798e-01 1.73766315e-01 5.30814409e-01 -4.47303355e-01
1.01185715e+00 3.91469598e-02 -2.80911446e-01 -9.48915407e-02
-2.26708964e-01 -1.08047999e-01 5.55551231e-01 -6.36905670e-01
-7.10045397e-01 -1.31492183e-01 -2.82490045e-01 -3.29382211e-01
-8.41380376e-03 5.61243474e-01 1.74964946e-02 1.87804788e-01
7.10922837e-01 2.81770766e-01 -3.55407476e-01 -9.69167888e-01
1.03127673e-01 4.82066631e-01 8.91563714e-01 -7.41191983e-01
8.19722176e-01 1.55450329e-01 2.97848135e-01 -1.11386526e+00
-8.26492488e-01 -7.71122634e-01 -7.42190301e-01 -2.67173439e-01
5.44825256e-01 -3.95064265e-01 -9.69944656e-01 2.84095019e-01
-9.10814226e-01 1.34337440e-01 -6.27967060e-01 6.33409381e-01
-7.54270852e-01 3.95137936e-01 -5.99714577e-01 -9.84127462e-01
-5.41928649e-01 -8.94137144e-01 1.14619434e+00 -9.62458551e-02
-6.70265973e-01 -5.09917200e-01 5.71868420e-01 2.79299319e-01
2.34355807e-01 1.65607810e-01 9.90067720e-01 -3.72366488e-01
-2.85339683e-01 -3.87337953e-01 2.88752407e-01 2.02892557e-01
1.66638568e-01 -7.79989883e-02 -1.27291012e+00 -8.32008943e-02
6.84898347e-02 5.67523167e-02 1.00510633e+00 3.61361206e-01
9.42823172e-01 -2.09871866e-02 1.35665700e-01 5.40895164e-01
1.00172734e+00 9.66447368e-02 5.15124857e-01 1.91654280e-01
3.03663582e-01 4.72288996e-01 6.20097280e-01 6.32515073e-01
-4.85931009e-01 1.18949044e+00 8.99845585e-02 4.59271401e-01
-2.30498075e-01 -2.74803579e-01 4.66849476e-01 1.47902203e+00
-7.79814005e-01 1.54802963e-01 -6.69224918e-01 1.84591308e-01
-1.71913588e+00 -1.27345943e+00 -4.05313700e-01 2.38401508e+00
6.94548130e-01 -1.03138834e-01 5.00265181e-01 1.26745546e+00
5.09994209e-01 2.45468274e-01 -1.82034224e-01 -2.75703907e-01
-1.95892036e-01 8.25028598e-01 -1.19533964e-01 -7.16300821e-03
-1.08693516e+00 5.29176414e-01 5.70340157e+00 1.03887808e+00
-1.01671815e+00 -5.37061207e-02 -1.47174716e-01 -1.10081710e-01
3.92070645e-03 -2.40644768e-01 -2.98174560e-01 1.96558386e-01
7.82486796e-01 8.79276171e-02 7.72457123e-01 3.22987080e-01
2.33339697e-01 2.01976642e-01 -9.91759062e-01 1.41736031e+00
-2.44826265e-02 -1.11486459e+00 5.87689737e-03 -2.20242694e-01
4.64680195e-01 -3.82694930e-01 2.85838038e-01 2.31177598e-01
-6.79561138e-01 -1.12794912e+00 1.06444383e+00 7.58332312e-01
7.48181224e-01 -8.92334342e-01 5.62562704e-01 1.77301556e-01
-1.56018364e+00 -2.39397883e-01 -1.13742545e-01 -3.76975417e-01
1.63686648e-02 5.39354920e-01 -4.78697896e-01 8.70663345e-01
5.31758487e-01 7.72821963e-01 -2.42849812e-01 1.34655178e+00
-2.44464621e-01 1.06563652e+00 -1.93826154e-01 7.62069300e-02
-1.10271618e-01 -4.12053943e-01 9.24945474e-01 1.24913442e+00
6.13011360e-01 -1.21158779e-01 -1.19891018e-01 7.32874930e-01
9.18261632e-02 5.69834173e-01 -2.35205054e-01 -3.45993280e-01
2.29349315e-01 1.12247229e+00 -8.65477979e-01 2.67060697e-02
2.31841788e-01 7.98026800e-01 -2.44058847e-01 2.51133084e-01
-6.73633873e-01 -4.89674211e-01 6.92671835e-01 4.88119945e-02
2.77407676e-01 -4.42516595e-01 -1.49469733e-01 -8.20425689e-01
-1.02663495e-01 -9.97570574e-01 2.78559506e-01 -7.22327352e-01
-1.05242050e+00 4.73551691e-01 -2.80763775e-01 -1.79427254e+00
-4.73720253e-01 -4.84485179e-01 -6.58955932e-01 7.29968667e-01
-1.22100139e+00 -8.35102022e-01 5.22493944e-03 5.72893023e-01
6.05713725e-01 -1.06577069e-01 1.27917683e+00 3.35966557e-01
-2.88907558e-01 3.68156880e-01 4.02497500e-01 1.63670927e-01
7.11744130e-01 -1.15706861e+00 1.78955302e-01 3.33912164e-01
1.28723466e+00 6.09533489e-01 7.75558352e-01 -2.42695108e-01
-1.53526676e+00 -4.87731040e-01 6.97116852e-01 -2.24694982e-01
6.84007347e-01 -3.40239331e-02 -8.52737308e-01 1.67844057e-01
1.91256795e-02 -4.38722283e-01 8.99928570e-01 3.97060186e-01
-4.47041512e-01 -2.50591904e-01 -1.93018422e-01 4.38140124e-01
8.76001477e-01 -7.40219653e-01 -8.36409569e-01 -1.35357380e-02
8.76616091e-02 -2.26409495e-01 -8.64648759e-01 3.41667920e-01
8.80362034e-01 -1.15868485e+00 1.19274807e+00 -2.39809051e-01
1.34503841e-01 -4.40435499e-01 -7.86562338e-02 -1.13233685e+00
-4.32812780e-01 -1.31317723e+00 -1.18962504e-01 1.43525565e+00
-9.55352411e-02 -1.77214220e-01 3.37053299e-01 -5.69077253e-01
-5.76534607e-02 -3.53020161e-01 -1.00874615e+00 -7.24009514e-01
-3.57319891e-01 -1.04388011e+00 4.74652141e-01 7.69508779e-01
2.74222381e-02 6.87727630e-01 -4.64654565e-01 -6.03066385e-02
5.50841093e-01 8.34221780e-01 8.30412388e-01 -1.69704175e+00
-7.43169487e-01 -8.34478855e-01 -5.20023644e-01 -6.80997431e-01
-1.52855972e-02 -1.08676112e+00 -9.27930474e-02 -1.03556609e+00
-1.81110501e-01 -2.15706199e-01 -7.07174480e-01 1.55169651e-01
8.29617456e-02 6.56146944e-01 4.64910865e-01 5.54417431e-01
-3.43120188e-01 4.24446166e-01 1.06242263e+00 -1.89763308e-01
-4.15896267e-01 8.47579777e-01 -2.62173057e-01 7.63362646e-01
5.99433839e-01 -2.86019087e-01 -4.06395018e-01 -3.63518135e-03
6.14908397e-01 3.45924318e-01 3.57460767e-01 -1.41162264e+00
4.72526513e-02 2.61119634e-01 1.58957601e-01 -5.10321617e-01
5.61660945e-01 -5.28663874e-01 3.34466934e-01 1.36183307e-01
-5.54834902e-01 -5.51760159e-02 2.30373293e-01 5.37834942e-01
-5.63476920e-01 -4.84603703e-01 3.90823692e-01 1.77341953e-01
-3.81011158e-01 -1.17948696e-01 -4.39617932e-01 7.21816570e-02
4.08358991e-01 -2.16976956e-01 3.58031511e-01 -3.75090629e-01
-8.57053578e-01 -7.79657900e-01 -1.67168766e-01 4.30688024e-01
5.90119362e-01 -1.26353395e+00 -6.65519714e-01 2.92432487e-01
4.54005525e-02 -4.70581800e-01 4.09636378e-01 1.07298386e+00
-2.65139431e-01 4.04751033e-01 -1.97163865e-01 -5.19600034e-01
-1.60153401e+00 3.65057856e-01 5.61069176e-02 -1.79226041e-01
-8.74464512e-01 8.75323892e-01 -4.02882583e-02 1.12051815e-01
5.49909413e-01 -7.51256108e-01 -5.41578650e-01 4.12453860e-01
6.79948866e-01 6.89571440e-01 2.24226177e-01 -6.02396309e-01
-7.76210949e-02 9.66024220e-01 4.57226902e-01 -3.54993045e-01
1.15115058e+00 2.46292278e-01 -2.06916854e-01 1.13612652e+00
6.85813487e-01 6.57094955e-01 -5.05614817e-01 -1.45098850e-01
3.66871357e-01 -2.88751006e-01 -1.17172478e-02 -6.83010161e-01
-4.86954540e-01 1.10882115e+00 4.48123544e-01 5.70770025e-01
1.24772203e+00 -1.29708976e-01 9.36024189e-01 3.65213931e-01
4.94951040e-01 -8.02591503e-01 8.09645504e-02 4.58422571e-01
1.12110746e+00 -4.39631402e-01 -7.75847062e-02 -3.62718463e-01
-3.42697859e-01 1.41107714e+00 -3.76903355e-01 -7.21750930e-02
3.95449817e-01 1.03939772e-01 3.44581250e-03 4.36902195e-02
-1.41507104e-01 -3.13095927e-01 8.42116833e-01 2.24450842e-01
7.09604204e-01 1.52466998e-01 -1.41727164e-01 1.05802023e+00
-7.72083104e-01 -5.41297235e-02 -2.32317336e-02 3.99974525e-01
-4.06036228e-01 -1.25172257e+00 -7.88458169e-01 1.50273159e-01
-7.27917433e-01 2.95992456e-02 -5.42013168e-01 3.37064534e-01
3.95198427e-02 7.80130327e-01 -2.87996292e-01 -3.96260560e-01
4.97601002e-01 4.18140858e-01 6.84907854e-01 -3.86263430e-01
-9.28379357e-01 4.95659590e-01 -1.70469508e-01 -2.73140132e-01
-5.46917140e-01 -6.75538778e-01 -1.16248286e+00 1.49978712e-01
-2.55969286e-01 5.59441805e-01 5.73507011e-01 7.36666322e-01
1.53656542e-01 9.70313489e-01 5.61476052e-01 -1.25534165e+00
-6.97744370e-01 -1.08012962e+00 -9.48720932e-01 5.25433719e-01
2.15852663e-01 -6.60185754e-01 -4.84724268e-02 9.62281898e-02] | [15.861384391784668, 5.309008598327637] |
f25ee06a-b3a8-42ad-a205-cd18f9bad85d | interpretable-ecg-classification-via-a-query | 2111.07386 | null | https://arxiv.org/abs/2111.07386v2 | https://arxiv.org/pdf/2111.07386v2.pdf | Interpretable ECG classification via a query-based latent space traversal (qLST) | Electrocardiography (ECG) is an effective and non-invasive diagnostic tool that measures the electrical activity of the heart. Interpretation of ECG signals to detect various abnormalities is a challenging task that requires expertise. Recently, the use of deep neural networks for ECG classification to aid medical practitioners has become popular, but their black box nature hampers clinical implementation. Several saliency-based interpretability techniques have been proposed, but they only indicate the location of important features and not the actual features. We present a novel interpretability technique called qLST, a query-based latent space traversal technique that is able to provide explanations for any ECG classification model. With qLST, we train a neural network that learns to traverse in the latent space of a variational autoencoder trained on a large university hospital dataset with over 800,000 ECGs annotated for 28 diseases. We demonstrate through experiments that we can explain different black box classifiers by generating ECGs through these traversals. | ['René van Es', 'Erik Bekkers', 'Rutger J. Hassink', 'Pieter A. Doevendans', 'Rutger R. van de Leur', 'Sharvaree P. Vadgama', 'Melle B. Vessies'] | 2021-11-14 | null | null | null | null | ['ecg-classification', 'electrocardiography-ecg'] | ['medical', 'methodology'] | [ 4.85763878e-01 5.45865417e-01 1.11058086e-01 -6.04243934e-01
-7.71128595e-01 -4.16740328e-01 -1.46920413e-01 4.61846411e-01
7.00872913e-02 5.49378335e-01 1.30516380e-01 -4.96617824e-01
-4.08571094e-01 -4.08368975e-01 -4.23051059e-01 -5.98924100e-01
-2.90198684e-01 7.88251042e-01 -2.90983349e-01 -1.94824338e-02
1.25357807e-01 3.64939719e-01 -1.09487283e+00 6.03651881e-01
8.76281440e-01 8.50448608e-01 -2.44329348e-01 9.99468207e-01
2.37073451e-01 7.61536360e-01 -9.31861460e-01 -2.67242521e-01
-1.61054656e-01 -9.30803776e-01 -9.91241157e-01 -1.45729452e-01
-2.42918711e-02 -3.25470090e-01 -3.96285690e-02 7.52097547e-01
7.72364378e-01 -3.12102139e-01 8.22229743e-01 -1.33928740e+00
-6.35199428e-01 8.09928656e-01 -1.37345776e-01 6.34802222e-01
1.09559171e-01 -5.22449277e-02 1.15242827e+00 -7.61472046e-01
4.33261663e-01 9.73157406e-01 8.88839304e-01 8.58791471e-01
-1.27509248e+00 -2.62996316e-01 -3.36286515e-01 4.55349505e-01
-1.20785522e+00 -5.55454046e-02 1.05673397e+00 -4.15118277e-01
7.62857378e-01 5.79393744e-01 1.03577077e+00 1.28041446e+00
4.73343790e-01 7.90384889e-01 6.57575786e-01 -3.85435969e-01
3.91113013e-01 2.88781896e-02 3.19009006e-01 7.57942617e-01
1.03183948e-01 -1.13583758e-01 -5.19244313e-01 -4.69486266e-01
9.54581022e-01 3.34466815e-01 -3.91575456e-01 -3.77007723e-01
-1.37172914e+00 1.12835240e+00 4.60740983e-01 2.05001101e-01
-5.47701776e-01 4.16286588e-01 6.60364985e-01 2.22065836e-01
4.86763269e-01 9.15648937e-01 -5.50606191e-01 -1.37434229e-01
-9.55449581e-01 2.04721063e-01 5.74037015e-01 2.61821151e-01
1.01571955e-01 2.26041690e-01 -5.07573962e-01 1.84639513e-01
3.61877352e-01 2.14405626e-01 5.67456603e-01 -1.19780374e+00
3.04468200e-02 6.37077153e-01 -1.63314492e-01 -9.30445433e-01
-6.22432828e-01 -7.48798192e-01 -1.10614479e+00 4.51915227e-02
-2.08731648e-02 -3.05506140e-01 -7.58096933e-01 1.45393038e+00
6.60233721e-02 3.67239177e-01 2.14683697e-01 1.12315893e+00
8.34565163e-01 3.64711642e-01 1.80027768e-01 -1.62029788e-01
1.43070674e+00 -6.48028910e-01 -8.80378008e-01 1.00749999e-01
6.31583273e-01 -1.54326692e-01 1.02784479e+00 4.34995443e-01
-9.56271291e-01 -6.27471387e-01 -1.16539800e+00 -1.14880271e-01
6.15333840e-02 -3.99895124e-02 4.83114064e-01 5.89518428e-01
-9.54342067e-01 9.56671596e-01 -1.15409577e+00 2.27804333e-02
6.58273816e-01 2.66404182e-01 -9.47071705e-04 5.93969703e-01
-1.31588125e+00 6.51692748e-01 2.79783517e-01 2.88035512e-01
-1.00258601e+00 -6.62851334e-01 -7.79673338e-01 5.22434890e-01
3.79351191e-02 -1.04900157e+00 1.05452776e+00 -1.17873859e+00
-1.21760058e+00 7.47269094e-01 -2.00868219e-01 -8.02925646e-01
6.00025773e-01 -3.13899279e-01 -1.71077043e-01 5.81944704e-01
1.07917361e-01 7.49396324e-01 1.10459781e+00 -8.38730514e-01
-1.52345791e-01 -2.75447667e-01 -1.49392784e-01 7.27081075e-02
-9.57601517e-02 -2.95554191e-01 1.78219691e-01 -7.50747979e-01
3.43741328e-01 -7.69948781e-01 -3.49947572e-01 1.15502946e-01
-6.45679533e-01 -4.31119591e-01 6.83292449e-01 -7.12263703e-01
1.13975394e+00 -2.18757653e+00 3.97388428e-01 8.98360088e-02
9.65649545e-01 -8.92005954e-03 1.06575213e-01 1.21525452e-01
-4.47217554e-01 3.71891230e-01 -5.99856675e-01 -1.33498088e-01
-6.34136572e-02 2.74713248e-01 -4.92896259e-01 2.41809458e-01
4.07309413e-01 1.29911709e+00 -9.65287209e-01 -4.94077265e-01
2.36071974e-01 6.05434358e-01 -6.24854982e-01 2.36045733e-01
-1.54101506e-01 8.02510738e-01 -5.18579364e-01 3.51373374e-01
1.55759826e-01 -7.73047209e-01 2.79645592e-01 -1.40583232e-01
5.60046256e-01 1.62039138e-02 -6.04787230e-01 1.78844714e+00
7.56557584e-02 9.87453103e-01 -8.23316097e-01 -1.29184437e+00
6.42199814e-01 8.40888560e-01 5.88477373e-01 -1.80625856e-01
1.35144189e-01 1.62914544e-01 1.10867769e-01 -9.56702292e-01
-5.73925935e-02 -2.87822247e-01 -8.82116891e-03 5.01612306e-01
-2.01980144e-01 2.55775183e-01 -4.02979761e-01 1.79078519e-01
1.13980103e+00 5.12389988e-02 3.29806775e-01 -1.92774802e-01
1.90179199e-01 4.73880358e-02 5.04570663e-01 9.72303748e-01
-3.00690234e-01 1.08348954e+00 7.99998641e-01 -1.09044814e+00
-9.94730711e-01 -1.15293622e+00 -1.42972574e-01 4.34535772e-01
-4.04281504e-02 -2.72899956e-01 -1.05325389e+00 -9.58461642e-01
-2.39834473e-01 7.76368618e-01 -8.87061179e-01 -4.20449823e-01
-4.68317419e-01 -7.12650955e-01 8.42216909e-01 8.68443370e-01
1.63446069e-01 -1.49114776e+00 -1.51920331e+00 3.89148265e-01
-6.22519314e-01 -8.06533337e-01 -1.44007087e-01 2.10873291e-01
-1.21943939e+00 -1.13484693e+00 -8.87283146e-01 -5.88083446e-01
7.50416875e-01 -4.74575430e-01 1.46025264e+00 3.96716177e-01
-8.11852455e-01 3.15851390e-01 -2.93431461e-01 -7.77136445e-01
-4.29421544e-01 1.24973953e-01 -7.51765892e-02 1.41083449e-01
3.56589675e-01 -4.64093626e-01 -1.02098536e+00 -2.02825159e-01
-7.25733459e-01 3.25777888e-01 3.53678972e-01 1.13168883e+00
6.73814476e-01 -4.35504794e-01 7.91700363e-01 -1.18643939e+00
8.76492441e-01 -3.06630224e-01 -7.81628191e-02 2.32884422e-01
-6.20339334e-01 3.59550476e-01 3.95741642e-01 -1.48831904e-01
-3.99580210e-01 -8.65385756e-02 -3.96048605e-01 -7.88875222e-01
-3.59672338e-01 6.48216069e-01 3.14228117e-01 5.43883622e-01
9.73524153e-01 2.15214133e-01 -2.23605663e-01 -3.62753868e-01
1.09577939e-01 5.89972436e-01 4.84692156e-01 -1.22614605e-02
3.68803352e-01 4.19244468e-01 2.14761272e-01 -6.06521845e-01
-9.77872491e-01 -1.92864817e-02 -6.70644701e-01 -4.00375053e-02
1.20768619e+00 -4.43311870e-01 -8.34397137e-01 -2.14493439e-01
-1.41165519e+00 -7.89725333e-02 -6.30544960e-01 2.81536579e-01
-5.72308183e-01 3.62727851e-01 -5.96611917e-01 -6.97506368e-01
-7.92149782e-01 -1.11537528e+00 1.39309752e+00 -2.02392220e-01
-8.06093633e-01 -1.08313715e+00 -3.79118510e-02 1.90338254e-01
1.35102585e-01 7.62302935e-01 1.43045366e+00 -6.94030583e-01
-5.08631647e-01 -6.83014095e-02 -6.35662153e-02 2.27313727e-01
1.57216132e-01 -3.37140709e-01 -1.10098886e+00 -1.25345528e-01
1.40605271e-01 -4.34382958e-03 6.31604552e-01 8.86117816e-01
1.83552539e+00 -3.20713401e-01 -2.56727397e-01 7.40236461e-01
1.13413417e+00 8.90651867e-02 4.96124476e-01 -9.05957669e-02
7.74255693e-01 4.68412727e-01 2.02591345e-01 3.82833779e-01
1.96033522e-01 3.69771302e-01 4.67301011e-01 -6.06030762e-01
3.48148257e-01 7.91327357e-02 -9.86683518e-02 7.53889859e-01
-2.35200584e-01 -1.00942247e-01 -9.87334907e-01 4.36630577e-01
-2.00602818e+00 -7.79621482e-01 -2.95850486e-01 1.76222920e+00
6.12472236e-01 1.03867598e-01 -1.22745693e-01 6.08844638e-01
3.93389136e-01 -2.08648399e-01 -8.17764342e-01 -6.42632723e-01
1.83338687e-01 4.11830306e-01 -1.35167882e-01 2.39042744e-01
-1.07131135e+00 3.07264864e-01 6.98148632e+00 2.06858024e-01
-1.24084496e+00 1.99768111e-01 9.89547074e-01 1.97385371e-01
-4.07395869e-01 -3.38724345e-01 1.94943458e-01 2.94362575e-01
1.04023194e+00 6.69623986e-02 -3.79053578e-02 7.45895445e-01
3.63576949e-01 4.54075217e-01 -1.47344971e+00 1.37456346e+00
1.32270485e-01 -1.57378209e+00 1.93983734e-01 -3.41458082e-01
3.44429582e-01 -2.72797912e-01 2.60446161e-01 -6.76988512e-02
-4.71190751e-01 -1.36575282e+00 5.41387379e-01 5.97892344e-01
9.32612121e-01 -6.18422389e-01 9.43340659e-01 2.00377166e-01
-7.16905117e-01 -3.30489613e-02 -8.68570060e-02 2.89050713e-02
4.45904173e-02 3.72416049e-01 -1.06801295e+00 3.58572721e-01
6.93347514e-01 7.22774804e-01 -5.48890412e-01 7.91594028e-01
-4.26621556e-01 1.03813601e+00 2.13171795e-01 9.99658182e-02
2.18685251e-02 2.73680031e-01 7.83862054e-01 8.99691463e-01
3.64175707e-01 3.46310213e-02 3.04066315e-02 1.27759802e+00
1.88720420e-01 -1.04960930e-02 -5.18968344e-01 7.07665132e-03
-9.70860720e-02 1.05832601e+00 -1.00687563e+00 -7.25809216e-01
2.49571025e-01 1.40488183e+00 -1.35495082e-01 3.90386641e-01
-1.13553071e+00 -5.32077432e-01 2.67682493e-01 1.84767157e-01
-1.05367109e-01 3.33614707e-01 -7.41555929e-01 -1.17594230e+00
-2.91909613e-02 -9.65764523e-01 5.85155070e-01 -1.11916745e+00
-1.03513074e+00 1.05867517e+00 -9.23498049e-02 -1.24295509e+00
-5.68274379e-01 -4.34275806e-01 -4.31819320e-01 9.73998308e-01
-1.26273966e+00 -8.14559877e-01 -3.84164840e-01 2.75838584e-01
8.92424583e-01 -1.28302440e-01 1.32897294e+00 1.02114566e-01
-3.61253291e-01 4.02254403e-01 -2.73208320e-01 3.80809188e-01
1.53754860e-01 -1.71552634e+00 5.31028748e-01 4.61263478e-01
5.67281544e-01 6.49122834e-01 8.52355480e-01 -3.69602233e-01
-8.17710817e-01 -1.04159141e+00 9.91841853e-01 -7.70154238e-01
-9.59827229e-02 -9.12255719e-02 -1.04014266e+00 7.48896360e-01
2.25526750e-01 -8.48041091e-04 9.04434025e-01 6.39346894e-03
2.00737402e-01 6.78063333e-02 -9.08208132e-01 5.63177586e-01
7.38318503e-01 -5.60741067e-01 -9.34083939e-01 3.25725913e-01
7.42053211e-01 -5.24313748e-01 -7.17435002e-01 4.70635504e-01
6.59668863e-01 -7.36290395e-01 9.80654120e-01 -8.97180200e-01
7.10197389e-01 -1.67914957e-01 4.36183900e-01 -1.26747179e+00
-2.85278887e-01 -5.85763335e-01 -3.28477442e-01 2.78992772e-01
4.03254390e-01 -6.38115585e-01 9.11922812e-01 4.83855724e-01
-1.22249022e-01 -9.84587193e-01 -7.98458517e-01 4.07697223e-02
-3.43294173e-01 -3.41850609e-01 6.16446793e-01 1.07716918e+00
-9.57992747e-02 5.43659627e-01 -3.22710246e-01 2.07980931e-01
6.88920975e-01 1.71646982e-01 1.14966474e-01 -1.50248432e+00
-4.82030809e-01 -1.12810895e-01 -6.05681717e-01 -5.32732844e-01
-1.92098275e-01 -1.05161500e+00 -1.96522459e-01 -1.61949790e+00
1.83600128e-01 -1.12895213e-01 -7.54862547e-01 6.32933497e-01
-4.41288114e-01 4.83303636e-01 -2.29594067e-01 1.41616076e-01
-4.24447149e-01 3.06005746e-01 1.03089154e+00 -2.67484993e-01
-2.81437457e-01 5.74197359e-02 -7.41490722e-01 7.66373575e-01
8.74088883e-01 -8.49728465e-01 -6.66514695e-01 -5.51472902e-01
3.98245364e-01 5.08790374e-01 7.94757068e-01 -9.86999989e-01
-2.69985527e-01 3.76450241e-01 7.98914373e-01 -5.49908757e-01
1.56163454e-01 -6.85809016e-01 3.03160399e-01 9.06349897e-01
-8.89764547e-01 3.78759056e-01 5.01462668e-02 5.65969050e-01
-2.84293503e-01 -1.47520199e-01 3.86309981e-01 -1.81846946e-01
-3.81480485e-01 2.09284350e-01 -5.54996610e-01 1.18307136e-01
7.27575362e-01 -2.30832189e-01 4.06825095e-01 -4.20516044e-01
-1.20230484e+00 -6.24297708e-02 -3.71658713e-01 2.68536776e-01
1.09835196e+00 -1.11855161e+00 -1.01465869e+00 3.63036513e-01
1.78343117e-01 2.25164331e-02 4.17375028e-01 8.84164810e-01
-1.11847365e+00 3.16070020e-01 -1.64249972e-01 -1.15757585e+00
-1.33591855e+00 3.83939773e-01 6.59206033e-01 -2.08591059e-01
-1.24914134e+00 6.84972644e-01 2.68551201e-01 -6.03184737e-02
1.78298146e-01 -6.29465461e-01 -5.19851506e-01 -1.88913122e-01
3.59634876e-01 1.14185698e-01 -1.48271739e-01 -1.35250792e-01
-3.46990854e-01 3.99143428e-01 3.47013384e-01 -1.64805949e-01
1.27622068e+00 5.04516102e-02 -5.62524796e-02 6.32897019e-01
8.83747995e-01 -6.83694363e-01 -8.29308271e-01 2.45489925e-01
1.14212319e-01 -4.36374657e-02 -2.13711355e-02 -7.50890851e-01
-9.66550469e-01 1.61765456e+00 1.18019712e+00 4.39787120e-01
1.19993484e+00 -1.51153490e-01 7.55372882e-01 3.13009292e-01
-7.68970326e-02 -7.47587800e-01 2.68397719e-01 -2.12950543e-01
8.81839812e-01 -1.07561839e+00 -3.11173707e-01 -2.38873854e-01
-8.74155283e-01 1.27229571e+00 9.38204750e-02 -1.12686999e-01
5.48127592e-01 -1.18649431e-01 4.79995519e-01 -6.30026698e-01
-6.63963139e-01 3.64443898e-01 4.98275489e-01 4.70143288e-01
4.76804942e-01 2.67592698e-01 -6.35597110e-02 7.72654295e-01
-1.99088547e-02 1.07458137e-01 4.06387597e-01 6.15352690e-01
-4.44212072e-02 -8.52556348e-01 -1.33558869e-01 5.89924157e-01
-8.97408605e-01 -2.62460709e-01 -1.07087724e-01 1.99943319e-01
2.07844689e-01 5.75372338e-01 4.31509130e-02 -6.17393479e-02
5.85921220e-02 4.25279945e-01 3.04914683e-01 -6.80773795e-01
-5.81269622e-01 9.52189788e-03 -3.57522100e-01 -2.48862058e-01
-2.40297332e-01 -4.32669133e-01 -1.39318538e+00 3.43451321e-01
-1.77959558e-02 4.83499140e-01 3.97686779e-01 1.06210446e+00
5.74973702e-01 1.31643856e+00 3.56580406e-01 -5.00040054e-01
-5.10022640e-01 -8.25717151e-01 -5.44520676e-01 6.23768926e-01
8.51939976e-01 -3.58772516e-01 -7.95322359e-02 5.51302552e-01] | [14.293721199035645, 3.2916736602783203] |
451a2e80-dd09-40ee-bc00-c1cf8bfdb018 | rovist-learning-robust-metrics-for-visual-1 | 2205.03774 | null | https://arxiv.org/abs/2205.03774v1 | https://arxiv.org/pdf/2205.03774v1.pdf | RoViST:Learning Robust Metrics for Visual Storytelling | Visual storytelling (VST) is the task of generating a story paragraph that describes a given image sequence. Most existing storytelling approaches have evaluated their models using traditional natural language generation metrics like BLEU or CIDEr. However, such metrics based on n-gram matching tend to have poor correlation with human evaluation scores and do not explicitly consider other criteria necessary for storytelling such as sentence structure or topic coherence. Moreover, a single score is not enough to assess a story as it does not inform us about what specific errors were made by the model. In this paper, we propose 3 evaluation metrics sets that analyses which aspects we would look for in a good story: 1) visual grounding, 2) coherence, and 3) non-redundancy. We measure the reliability of our metric sets by analysing its correlation with human judgement scores on a sample of machine stories obtained from 4 state-of-the-arts models trained on the Visual Storytelling Dataset (VIST). Our metric sets outperforms other metrics on human correlation, and could be served as a learning based evaluation metric set that is complementary to existing rule-based metrics. | ['Josiah Poon', 'Caren Han', 'Eileen Wang'] | 2022-05-08 | null | null | null | null | ['visual-storytelling'] | ['natural-language-processing'] | [ 1.46267235e-01 3.65485936e-01 4.25431505e-02 -1.97880760e-01
-7.02314973e-01 -6.04516923e-01 1.39625001e+00 6.69809937e-01
-1.13458961e-01 8.22450459e-01 7.82551050e-01 -1.57841772e-01
-1.56685337e-01 -8.36102068e-01 -6.18433475e-01 -3.71446401e-01
2.09157273e-01 5.07796586e-01 4.80689555e-01 -2.87469000e-01
6.27880454e-01 -1.22887447e-01 -1.84854424e+00 7.90135145e-01
7.28672028e-01 6.91360354e-01 4.22287762e-01 6.39225721e-01
-3.81520927e-01 1.54038346e+00 -9.30822194e-01 -4.81421351e-01
-2.48844996e-01 -9.90042865e-01 -9.60231245e-01 4.12497446e-02
4.64179993e-01 -2.49198433e-02 3.97448950e-02 7.66266525e-01
2.64794976e-01 -1.04735129e-01 9.95448649e-01 -1.20159745e+00
-5.56086361e-01 9.46283519e-01 -1.39485925e-01 -3.97277363e-02
1.04968500e+00 2.74417788e-01 1.21503246e+00 -7.19828427e-01
1.26639962e+00 1.18417454e+00 4.72960293e-01 2.54302174e-01
-9.83459890e-01 -1.01088434e-01 -3.49991888e-01 5.02169490e-01
-9.57277775e-01 -2.01411754e-01 5.78576803e-01 -9.59132254e-01
8.55451584e-01 3.87191534e-01 7.87440836e-01 1.25873280e+00
1.34596705e-01 7.14922309e-01 1.40400982e+00 -5.54781616e-01
3.85010123e-01 2.50797212e-01 -2.19087571e-01 3.83399218e-01
1.52829379e-01 2.46794075e-02 -8.39619339e-01 1.44609183e-01
6.21067941e-01 -7.39577830e-01 -1.59019485e-01 -4.66354162e-01
-1.68969214e+00 9.14283514e-01 1.81943074e-01 7.42489874e-01
-4.05859917e-01 1.19239837e-01 6.30233645e-01 3.26805979e-01
2.08255187e-01 8.83768439e-01 1.96741760e-01 -4.09287214e-01
-1.28815985e+00 7.24790037e-01 6.21603727e-01 7.13194728e-01
3.53751540e-01 -3.89206521e-02 -8.54063094e-01 6.75324619e-01
1.46440148e-01 1.50140241e-01 6.60295844e-01 -8.20069432e-01
2.88469404e-01 7.55267203e-01 1.35601595e-01 -1.08927929e+00
-1.42440155e-01 -2.06635535e-01 -4.86165792e-01 5.32937169e-01
4.62194324e-01 1.71271056e-01 -5.43177605e-01 1.52726591e+00
-1.11554630e-01 -2.90947288e-01 4.67820615e-02 9.74174500e-01
1.43020606e+00 7.30211616e-01 -1.80112153e-01 -2.68211007e-01
1.22179198e+00 -9.23822343e-01 -7.56495893e-01 -1.94339290e-01
5.39408147e-01 -1.08763456e+00 1.33812761e+00 6.37731194e-01
-1.17468894e+00 -7.10051775e-01 -1.35221326e+00 9.93800014e-02
-4.32513565e-01 -2.78904792e-02 3.05296481e-01 4.34764415e-01
-1.13772559e+00 7.36507177e-01 -6.92778304e-02 -6.60994828e-01
-1.18022442e-01 -4.95881110e-01 -2.90421605e-01 1.45916119e-01
-1.00898147e+00 1.28955007e+00 6.62588716e-01 -6.13811910e-01
-1.00602925e+00 -2.25271255e-01 -6.20570779e-01 -1.12280875e-01
3.37858886e-01 -6.97077990e-01 1.32269871e+00 -8.04398894e-01
-1.19415414e+00 1.15232134e+00 4.64809984e-02 -6.00444138e-01
8.63464177e-01 1.86348483e-02 -2.17667267e-01 1.97018817e-01
2.86234945e-01 7.21131086e-01 5.45342386e-01 -1.61907244e+00
-3.99634600e-01 1.66886836e-01 2.00695783e-01 1.37385592e-01
2.07243185e-03 1.27933770e-01 -8.64486769e-02 -8.58321011e-01
-1.72552690e-01 -4.67851043e-01 1.28380015e-01 -2.75526792e-01
-4.11782026e-01 -2.39652634e-01 4.89622444e-01 -7.27754712e-01
1.38555324e+00 -1.75777960e+00 7.03429207e-02 1.00052748e-02
2.38060635e-02 1.07525513e-01 -1.30677491e-01 9.04504538e-01
3.43305543e-02 1.72057852e-01 -1.89732298e-01 -1.53852075e-01
1.23810112e-01 -1.56879485e-01 -3.15592498e-01 -2.11297765e-01
1.17488600e-01 6.76672101e-01 -1.23605061e+00 -9.54806507e-01
3.09935927e-01 4.14321214e-01 -1.89836696e-01 1.37147710e-01
-5.92023373e-01 3.04014087e-01 -5.08525781e-02 1.57365113e-01
5.81327500e-03 -1.14808269e-01 5.86865097e-02 7.69361332e-02
-1.10284112e-01 4.93787885e-01 -9.39127803e-01 1.71915352e+00
-2.72533804e-01 1.08959126e+00 -9.66843486e-01 -4.77287740e-01
1.24521792e+00 5.54503143e-01 9.47277769e-02 -8.76116455e-01
1.49066627e-01 1.18885428e-01 -9.12420452e-02 -6.56307042e-01
8.17578673e-01 -6.21071905e-02 -1.07383989e-01 5.99638224e-01
3.00336897e-01 -5.90283096e-01 7.53667235e-01 3.70592415e-01
1.22808731e+00 3.90959054e-01 6.34784162e-01 -1.87847286e-01
3.12912434e-01 2.45060265e-01 -8.09212551e-02 8.58676791e-01
1.95996508e-01 1.10084999e+00 8.64557207e-01 -3.44286054e-01
-1.37232697e+00 -1.10190952e+00 2.40647957e-01 5.48610210e-01
-4.70757075e-02 -9.29854453e-01 -9.37387884e-01 -5.27694046e-01
-5.00442743e-01 1.33749890e+00 -7.26570666e-01 9.27507579e-02
-4.22085822e-02 -2.71544456e-01 4.30427253e-01 1.10978946e-01
3.08791786e-01 -1.33328736e+00 -1.13434339e+00 1.92199931e-01
-5.51798940e-01 -9.77620065e-01 -9.61303711e-02 -3.11977357e-01
-6.00278080e-01 -1.01099575e+00 -5.93898654e-01 -4.64216888e-01
3.26529145e-01 1.93378553e-01 1.54120481e+00 8.44584256e-02
1.22289911e-01 1.85270578e-01 -1.02487338e+00 -4.70820397e-01
-1.04567289e+00 -2.20091552e-01 -3.24752599e-01 -1.06415041e-01
8.70528966e-02 -4.91707116e-01 -3.30049455e-01 3.04573983e-01
-1.08184314e+00 5.19182324e-01 6.37142956e-01 6.16267264e-01
2.81020850e-01 -5.58210090e-02 2.79947072e-01 -6.89767480e-01
1.03656244e+00 -2.21700475e-01 -6.44393861e-02 4.70202804e-01
-5.76075792e-01 1.79365322e-01 3.23880106e-01 -3.28133613e-01
-9.35766160e-01 -2.90033221e-01 1.35179982e-01 -9.83146504e-02
-2.11965591e-01 6.35856211e-01 1.77387372e-01 3.78284365e-01
1.01931787e+00 2.41467670e-01 -3.17267865e-01 -4.20360044e-02
3.68699312e-01 3.08892429e-01 5.51368952e-01 -3.46242577e-01
7.97991455e-01 1.10270604e-01 -1.85350686e-01 -6.76211774e-01
-6.18798196e-01 -3.06854457e-01 -3.46432686e-01 -9.79256094e-01
9.41263378e-01 -5.69806695e-01 -4.07142580e-01 1.30600363e-01
-1.54087412e+00 -2.06128478e-01 -4.91678506e-01 4.15245652e-01
-9.85871136e-01 2.90227264e-01 -1.54501557e-01 -9.50620294e-01
-1.37802258e-01 -7.84795702e-01 9.54933822e-01 -2.04574745e-02
-1.15661991e+00 -7.10797429e-01 4.26897049e-01 4.19840693e-01
3.96359116e-01 8.75662029e-01 8.79058957e-01 -3.86474073e-01
-3.27769846e-01 -3.09885025e-01 -1.21706322e-01 6.30399734e-02
-2.11367488e-01 1.46614596e-01 -8.82848680e-01 1.96689114e-01
-8.48351046e-02 -4.70422208e-01 7.68827200e-01 2.89404690e-01
5.51963389e-01 -4.23193693e-01 7.81283230e-02 -2.13944882e-01
1.45531225e+00 2.88889349e-01 1.00976920e+00 6.69498444e-01
4.00159448e-01 9.65888560e-01 8.27820718e-01 4.74165559e-01
3.11888754e-01 8.50870788e-01 5.13674438e-01 1.20605305e-01
-4.52992797e-01 -7.40720570e-01 5.46041906e-01 6.76908970e-01
-3.92592818e-01 -5.62475681e-01 -1.11952627e+00 6.67053759e-01
-2.08869314e+00 -1.30714869e+00 -4.86288130e-01 2.17262435e+00
5.86627841e-01 5.69810033e-01 2.83274263e-01 6.31167769e-01
5.80177724e-01 4.19777870e-01 2.65829831e-01 -6.68651462e-01
-5.30970991e-01 -9.74091142e-02 -2.92491913e-02 3.51573080e-01
-6.85829520e-01 7.17212617e-01 6.31038189e+00 1.05544615e+00
-6.88051343e-01 1.71317011e-01 4.38856214e-01 -1.24674551e-01
-5.42954385e-01 2.02375934e-01 -1.34443462e-01 3.73894602e-01
6.93359554e-01 -3.21661979e-01 1.28328189e-01 7.68288553e-01
3.86826962e-01 -4.81395662e-01 -1.18869019e+00 9.59929228e-01
4.75485176e-01 -1.41814709e+00 4.29328978e-01 -1.17971987e-01
8.54348958e-01 -4.03904229e-01 -2.47319490e-01 7.06188455e-02
4.09402907e-01 -1.13483214e+00 1.48508298e+00 8.64273787e-01
5.20250499e-01 -4.89955366e-01 9.20427501e-01 2.71560282e-01
-7.88494229e-01 4.52476233e-01 5.13744205e-02 -2.29483396e-01
4.11436260e-01 4.86811012e-01 -1.15070951e+00 5.99029064e-01
4.75016624e-01 3.58402908e-01 -9.16464210e-01 1.24024439e+00
-6.60145819e-01 5.07662833e-01 2.56424636e-01 -5.39506078e-01
2.74444282e-01 1.23968616e-01 8.74681294e-01 1.41639328e+00
6.64647341e-01 -3.16894412e-01 -3.30429345e-01 9.79514241e-01
4.00186151e-01 4.97547865e-01 -8.60773444e-01 -1.80137411e-01
2.34136239e-01 9.24947679e-01 -9.92217183e-01 -3.93765032e-01
-1.45570293e-01 8.91042471e-01 -4.76080477e-02 -1.20349474e-01
-7.07819819e-01 -2.11689785e-01 5.99889383e-02 4.94785190e-01
1.22981302e-01 -7.30215162e-02 -5.71459949e-01 -7.31179178e-01
1.52297825e-01 -9.33638394e-01 5.80617860e-02 -1.53505886e+00
-9.55123007e-01 9.07039165e-01 1.71979472e-01 -1.60064209e+00
-5.97276032e-01 -2.46912763e-01 -8.11336935e-01 3.60143214e-01
-8.02361012e-01 -1.18285739e+00 -4.62640941e-01 9.42717046e-02
5.88635743e-01 -1.86802864e-01 7.33000219e-01 -1.79239556e-01
1.70563027e-01 1.42680123e-01 -3.47677618e-01 -1.79948494e-01
7.47583508e-01 -1.36937106e+00 4.74129856e-01 9.70456898e-01
4.85130221e-01 1.26946747e-01 1.47339678e+00 -6.06871247e-01
-5.43718517e-01 -6.61882579e-01 1.40092945e+00 -6.59520745e-01
5.75679719e-01 -1.01809293e-01 -7.61778235e-01 1.43122196e-01
5.80353916e-01 -1.00689781e+00 5.23586929e-01 -7.57121518e-02
-4.15977538e-01 3.06811213e-01 -8.23330104e-01 6.25221074e-01
1.01622856e+00 -3.26917082e-01 -8.91318619e-01 3.26616347e-01
5.71929216e-01 -1.27880752e-01 -6.05490565e-01 1.34901404e-01
5.78799605e-01 -1.42535508e+00 6.46767318e-01 -3.64669681e-01
1.35787308e+00 -4.80829686e-01 -1.12575762e-01 -1.32941091e+00
-2.88828492e-01 -4.67458338e-01 2.28243530e-01 1.41729057e+00
4.90064204e-01 -2.48068701e-02 4.44355339e-01 3.15472633e-02
-1.13941438e-01 -3.52236241e-01 -7.87326992e-01 -9.57080841e-01
-2.16747150e-01 -6.28826201e-01 7.32923150e-01 1.04445624e+00
3.57849777e-01 6.27034068e-01 -5.70356011e-01 -5.85333467e-01
3.77864778e-01 -2.88619280e-01 1.02658629e+00 -1.21346307e+00
-3.35927755e-01 -9.17239249e-01 -6.79153085e-01 -2.08869472e-01
-2.93293834e-01 -7.89226413e-01 1.46360368e-01 -2.20211887e+00
3.89510781e-01 2.00451970e-01 7.68797845e-02 2.36707494e-01
2.60422118e-02 2.11069405e-01 4.84155923e-01 3.08929414e-01
-8.12131345e-01 3.48657191e-01 1.14456105e+00 -1.89803779e-01
2.27433518e-02 -5.34634531e-01 -5.58999538e-01 6.53160214e-01
7.22166121e-01 -3.88876826e-01 -4.35075432e-01 -3.29278201e-01
8.00479710e-01 1.84659690e-01 6.03013098e-01 -1.52328873e+00
1.12599500e-01 -1.17145695e-01 8.84140879e-02 -4.83181596e-01
1.98839515e-01 -4.48703796e-01 5.48680186e-01 3.57596993e-01
-4.16672081e-01 6.17514737e-02 -2.02130839e-01 1.36802405e-01
-5.24151683e-01 -4.45804924e-01 4.28403646e-01 -2.73782492e-01
-7.93338299e-01 -4.50073451e-01 -6.54795349e-01 -1.10855848e-01
1.02176154e+00 -6.06224775e-01 -4.78798658e-01 -8.59379053e-01
-4.25000429e-01 -1.32608740e-02 7.53379703e-01 7.24164784e-01
6.90009832e-01 -1.58512723e+00 -1.24486387e+00 -5.64772308e-01
6.19586289e-01 -4.78292525e-01 3.45496880e-03 4.73679960e-01
-7.87585855e-01 4.79251921e-01 -5.53025723e-01 -5.02771676e-01
-1.37498546e+00 4.62358028e-01 -1.47793636e-01 -6.61557674e-01
-2.86396980e-01 4.77468431e-01 -2.33289208e-02 1.47594035e-01
-3.89764085e-02 -1.62547767e-01 -6.49274230e-01 3.78691405e-01
6.81441486e-01 3.50990593e-01 -3.80495489e-02 -7.79646397e-01
-1.13635190e-01 4.62412536e-01 2.97163278e-01 -7.14202523e-01
1.12749219e+00 1.28397584e-01 6.33796602e-02 9.11604643e-01
6.54147029e-01 -2.82116890e-01 -8.09584379e-01 -1.19870529e-01
2.58981913e-01 -4.26479727e-01 -1.53997123e-01 -9.35955524e-01
-4.34538931e-01 7.52818644e-01 3.56728613e-01 7.68836915e-01
8.35907102e-01 1.99110046e-01 4.12295580e-01 1.81425288e-01
4.44724858e-01 -1.28906357e+00 5.09521902e-01 4.42991346e-01
1.56145108e+00 -1.20288455e+00 -4.87693101e-02 -1.83175907e-01
-1.01436400e+00 1.10874844e+00 3.25312197e-01 6.67365342e-02
-3.59700955e-02 -6.05764613e-02 -3.98816578e-02 -3.31728667e-01
-8.45381618e-01 -4.38846141e-01 5.62815547e-01 6.28972292e-01
7.32875288e-01 2.16207832e-01 -8.17140818e-01 3.70415717e-01
-9.59143937e-01 9.40136835e-02 7.01112509e-01 7.12443531e-01
-6.20933294e-01 -8.80659580e-01 -5.20485103e-01 3.44673425e-01
-1.03839971e-01 1.71749622e-01 -9.77231383e-01 7.70964265e-01
2.70922452e-01 1.23687875e+00 -1.35327920e-01 -7.34608531e-01
4.03858602e-01 1.35344435e-02 7.35066354e-01 -6.36651456e-01
-6.80302262e-01 -2.93153644e-01 5.53799212e-01 -3.31749290e-01
-6.94449067e-01 -8.89214754e-01 -7.40583420e-01 -4.70165014e-01
2.58502197e-02 9.50867385e-02 6.75443053e-01 1.03104973e+00
-2.14649737e-01 5.95015824e-01 1.59979120e-01 -6.78953707e-01
2.21944287e-01 -1.28928292e+00 -2.17560470e-01 7.10722744e-01
-1.11150846e-01 -4.80763346e-01 -1.30661398e-01 4.20820147e-01] | [11.718422889709473, 8.807698249816895] |
793090ea-fe89-4fa8-98d9-80bc559c5517 | no-reference-image-quality-assessment-via-1 | 2108.06858 | null | https://arxiv.org/abs/2108.06858v2 | https://arxiv.org/pdf/2108.06858v2.pdf | No-Reference Image Quality Assessment via Transformers, Relative Ranking, and Self-Consistency | The goal of No-Reference Image Quality Assessment (NR-IQA) is to estimate the perceptual image quality in accordance with subjective evaluations, it is a complex and unsolved problem due to the absence of the pristine reference image. In this paper, we propose a novel model to address the NR-IQA task by leveraging a hybrid approach that benefits from Convolutional Neural Networks (CNNs) and self-attention mechanism in Transformers to extract both local and non-local features from the input image. We capture local structure information of the image via CNNs, then to circumvent the locality bias among the extracted CNNs features and obtain a non-local representation of the image, we utilize Transformers on the extracted features where we model them as a sequential input to the Transformer model. Furthermore, to improve the monotonicity correlation between the subjective and objective scores, we utilize the relative distance information among the images within each batch and enforce the relative ranking among them. Last but not least, we observe that the performance of NR-IQA models degrades when we apply equivariant transformations (e.g. horizontal flipping) to the inputs. Therefore, we propose a method that leverages self-consistency as a source of self-supervision to improve the robustness of NRIQA models. Specifically, we enforce self-consistency between the outputs of our quality assessment model for each image and its transformation (horizontally flipped) to utilize the rich self-supervisory information and reduce the uncertainty of the model. To demonstrate the effectiveness of our work, we evaluate it on seven standard IQA datasets (both synthetic and authentic) and show that our model achieves state-of-the-art results on various datasets. | ['Kris M. Kitani', 'Saba Dadsetan', 'S. Alireza Golestaneh'] | 2021-08-16 | null | null | null | null | ['no-reference-image-quality-assessment'] | ['computer-vision'] | [ 2.26403788e-01 -1.08927749e-01 -1.91314612e-02 -4.55001473e-01
-9.27314103e-01 -6.61834598e-01 3.52756232e-01 -1.36078879e-01
-2.50851929e-01 2.89517671e-01 2.23646283e-01 -4.09115665e-02
-1.84410438e-01 -6.89502299e-01 -9.64013577e-01 -6.50267005e-01
2.21246809e-01 -2.90520847e-01 1.29333332e-01 -2.27947712e-01
1.81898460e-01 4.16178226e-01 -1.44684327e+00 3.08547169e-01
1.14669633e+00 1.53854823e+00 -9.61569026e-02 4.55756545e-01
4.56356227e-01 7.26516128e-01 -5.02450466e-01 -5.50207078e-01
6.00022376e-01 -3.80828619e-01 -6.90777421e-01 3.00165147e-01
6.75955772e-01 -4.18989480e-01 -3.89932901e-01 1.31184566e+00
4.26703095e-01 1.12950571e-01 6.01663232e-01 -1.23760247e+00
-1.04044271e+00 2.59567946e-01 -6.45585775e-01 1.50108427e-01
2.25411609e-01 6.21200264e-01 1.27099240e+00 -9.01234388e-01
4.29005474e-01 1.15599632e+00 4.41478640e-01 1.40862525e-01
-1.21991134e+00 -7.14982212e-01 1.42782509e-01 3.87163013e-01
-1.32932758e+00 -6.38822019e-01 9.18547213e-01 -2.81111807e-01
5.12309015e-01 1.09982021e-01 5.27171552e-01 1.05468476e+00
3.00258547e-01 5.37560701e-01 1.26276898e+00 -2.44842306e-01
2.69443423e-01 7.73039982e-02 1.77836772e-02 8.03738773e-01
-1.09685436e-01 3.08142990e-01 -5.50658286e-01 2.72726834e-01
7.91251719e-01 -8.98562595e-02 -4.08024371e-01 -5.35880327e-01
-1.16580760e+00 4.68815565e-01 8.00346494e-01 1.62536129e-01
-3.30801547e-01 -4.06881832e-02 1.77284732e-01 5.02051532e-01
1.35553718e-01 4.63292450e-01 -3.21286380e-01 8.40079114e-02
-7.54757881e-01 -1.41052067e-01 2.95572668e-01 6.98025584e-01
1.00018871e+00 9.96550079e-03 -5.62513828e-01 6.99138820e-01
9.26090479e-02 5.61472535e-01 5.20156264e-01 -1.07302678e+00
6.15147531e-01 7.25205421e-01 2.05483362e-01 -1.36664939e+00
-1.52443260e-01 -7.94365585e-01 -1.09659851e+00 5.08178830e-01
4.70866442e-01 3.46978575e-01 -9.34973121e-01 1.95577121e+00
1.34803906e-01 1.24575384e-01 7.25178234e-03 1.02750146e+00
4.74610984e-01 4.33253974e-01 -1.39678260e-02 -5.56093380e-02
1.23599470e+00 -1.09097767e+00 -5.43970704e-01 -3.49988379e-02
2.20136419e-01 -5.15034258e-01 1.54549789e+00 3.78196955e-01
-1.08580554e+00 -1.07334375e+00 -1.31405675e+00 -2.24897012e-01
-3.54013473e-01 4.30098087e-01 -3.41195315e-02 4.10739213e-01
-1.05924332e+00 8.11556041e-01 -6.42476380e-01 -1.51443565e-02
4.66826051e-01 3.26753587e-01 -5.89071393e-01 -1.18885972e-01
-1.22620881e+00 6.75508738e-01 1.73221707e-01 2.12264851e-01
-9.12219822e-01 -6.06778562e-01 -8.66389036e-01 3.50699902e-01
4.91529137e-01 -7.30769634e-01 7.80010343e-01 -1.39823747e+00
-1.75068128e+00 5.60428143e-01 2.63058618e-02 -2.27977991e-01
4.92331445e-01 -1.96514860e-01 -4.72103268e-01 1.91019997e-01
7.67440870e-02 6.08799398e-01 1.18240559e+00 -1.41355789e+00
-5.56993961e-01 -3.18529218e-01 2.94331491e-01 1.24115258e-01
-4.73218322e-01 -2.76901603e-01 -6.03640378e-01 -6.95856690e-01
1.53598800e-01 -8.67385983e-01 7.42025599e-02 1.45937100e-01
-4.98802036e-01 1.21396273e-01 5.35455823e-01 -7.63513386e-01
1.16110218e+00 -2.41939807e+00 4.52997945e-02 4.13281471e-01
2.72827774e-01 1.47109285e-01 -5.84410071e-01 -7.71800727e-02
-1.96467698e-01 1.09804496e-01 -2.67881662e-01 -3.24585497e-01
8.08031335e-02 3.81928682e-02 -2.38243535e-01 4.30038631e-01
6.14731610e-01 1.05622053e+00 -9.13969755e-01 -4.45324659e-01
2.45729551e-01 4.30822790e-01 -5.16691744e-01 4.85331684e-01
1.30398914e-01 5.52602708e-01 -1.95762292e-01 5.02880216e-01
7.51997530e-01 -4.94862318e-01 -1.10678755e-01 -9.41162467e-01
1.03315271e-01 1.73292741e-01 -1.26391852e+00 1.80781519e+00
-6.59171581e-01 2.46494710e-01 -1.39103025e-01 -8.18887293e-01
7.77762055e-01 1.65632173e-01 3.92086923e-01 -1.12629914e+00
1.61775768e-01 4.86316206e-03 8.78217965e-02 -3.37866545e-01
3.64619136e-01 1.50694147e-01 1.15261562e-01 3.89791042e-01
3.25883716e-01 2.16185898e-01 1.54189738e-02 8.20915028e-02
9.91747975e-01 8.96057114e-02 2.97718138e-01 -4.93700318e-02
7.30206251e-01 -4.47439820e-01 6.44213796e-01 7.95712173e-01
-4.02424157e-01 8.96418869e-01 4.08964217e-01 -3.27177256e-01
-1.09284472e+00 -1.26967192e+00 7.84113854e-02 9.54778671e-01
4.40006256e-01 -2.44473636e-01 -6.97104871e-01 -8.57427239e-01
-2.09999517e-01 3.24755162e-01 -8.82272840e-01 -5.88171422e-01
-3.29173356e-01 -5.18382668e-01 2.66716242e-01 5.83899021e-01
8.72979760e-01 -9.25699115e-01 -4.38254714e-01 3.47319841e-02
-3.07613879e-01 -1.25173819e+00 -6.76351070e-01 1.19359056e-02
-5.13619006e-01 -9.86938655e-01 -4.56439704e-01 -3.51825148e-01
6.82010710e-01 2.76846737e-01 1.12749290e+00 -3.86002399e-02
1.20989218e-01 1.86862484e-01 -2.02965036e-01 2.11385936e-01
-3.22304219e-01 8.95512700e-02 -1.17240325e-01 6.19476795e-01
-1.77701294e-01 -6.59425080e-01 -9.72822189e-01 4.80476350e-01
-1.01720071e+00 -1.25063409e-03 8.34680259e-01 1.00268137e+00
6.37260020e-01 1.53107569e-01 5.42765439e-01 -5.71242392e-01
5.60450137e-01 -1.99851871e-01 -5.59082270e-01 4.15796757e-01
-6.95580065e-01 1.55921191e-01 8.33761334e-01 -4.75051522e-01
-9.37595069e-01 3.73066477e-02 2.21695732e-02 -6.63393557e-01
2.55841818e-02 2.59628445e-01 -5.68214417e-01 -1.91054806e-01
5.27010322e-01 9.52409208e-02 -1.71120986e-01 -2.44523033e-01
5.83543420e-01 5.19474864e-01 8.48525941e-01 -5.02870560e-01
1.21147656e+00 5.02872169e-01 1.41386464e-02 -1.05897471e-01
-1.15062916e+00 -3.13948870e-01 -5.79074204e-01 -2.32346460e-01
7.86598921e-01 -9.77865160e-01 -8.58358681e-01 4.27569807e-01
-9.64128494e-01 -1.70638338e-01 -3.58300328e-01 2.21989200e-01
-5.23349643e-01 3.66876990e-01 -5.53640187e-01 -4.85541910e-01
-4.63581890e-01 -1.32484639e+00 1.14049947e+00 2.26934388e-01
8.46709833e-02 -6.56886578e-01 -1.23698965e-01 3.37746769e-01
4.78399783e-01 1.42635554e-01 9.55671132e-01 -2.88834870e-01
-6.78082168e-01 4.74538393e-02 -6.27793431e-01 7.72521257e-01
3.92419457e-01 -3.61781269e-02 -1.12279570e+00 -4.29106086e-01
7.52482787e-02 -3.95350218e-01 7.72337496e-01 1.08509861e-01
1.31433201e+00 -4.44715381e-01 2.94720441e-01 7.86951602e-01
1.41906488e+00 -6.27244934e-02 7.41853356e-01 4.84751761e-01
7.52805531e-01 4.66312885e-01 5.86889803e-01 2.52613693e-01
3.85330617e-01 7.41726458e-01 5.39860427e-01 -4.11185175e-01
-2.98992306e-01 -2.96929777e-01 4.44031924e-01 8.37941289e-01
4.85299304e-02 -1.51426405e-01 -4.90709335e-01 5.32941639e-01
-1.70172751e+00 -8.09056282e-01 3.67452860e-01 2.25583291e+00
8.58629405e-01 1.97878391e-01 -1.19113460e-01 2.69807309e-01
5.87472081e-01 1.89349070e-01 -7.74113119e-01 -7.84398392e-02
-2.23255545e-01 1.74023375e-01 3.41160178e-01 3.17579329e-01
-1.12636125e+00 7.13947833e-01 5.81806660e+00 7.79330909e-01
-1.35380769e+00 3.05085108e-02 9.20557082e-01 1.04365595e-01
-1.36245057e-01 -1.82822272e-01 -8.02332833e-02 4.49739248e-01
6.89772189e-01 2.65050754e-02 6.41386151e-01 5.94521821e-01
2.88444817e-01 6.60205930e-02 -1.26372576e+00 9.26742792e-01
3.83623727e-02 -9.27846968e-01 2.36356065e-01 -8.35676640e-02
8.96986961e-01 -2.50764906e-01 6.23039782e-01 1.58802435e-01
1.97814047e-01 -1.02001059e+00 9.29483652e-01 5.86977959e-01
9.57537055e-01 -6.41581357e-01 7.91246235e-01 -2.70957835e-02
-1.11302853e+00 -2.68355399e-01 -2.12319613e-01 2.49717146e-01
-1.55049071e-01 6.54069901e-01 -3.30838293e-01 8.80466163e-01
8.69304419e-01 7.52090931e-01 -1.05282462e+00 7.23109603e-01
-2.97172159e-01 3.96059245e-01 3.10407057e-02 7.20996380e-01
5.46427593e-02 -2.49698803e-01 2.68347025e-01 7.57725775e-01
4.03386712e-01 -1.23698972e-01 -5.04939295e-02 1.14314318e+00
-4.29536581e-01 3.10931560e-02 -2.70069122e-01 3.52060974e-01
2.33248979e-01 1.33532000e+00 -2.97674149e-01 -2.79042184e-01
-4.09346104e-01 1.01557529e+00 4.22822207e-01 5.05574167e-01
-9.17235613e-01 -2.88024008e-01 6.01079881e-01 3.29290405e-02
3.13737422e-01 6.99766492e-03 -3.64438772e-01 -1.28298378e+00
4.09724355e-01 -1.23733175e+00 2.18417659e-01 -1.01820600e+00
-1.36897135e+00 7.06420064e-01 -3.61512572e-01 -1.59952092e+00
-1.46753848e-01 -4.44817781e-01 -5.20923138e-01 8.07033837e-01
-1.93476903e+00 -1.23757660e+00 -6.42583787e-01 8.98640454e-01
9.76186395e-02 6.81933761e-02 4.16918486e-01 2.92802036e-01
-5.45844018e-01 9.45508242e-01 -1.14898290e-02 4.01002109e-01
1.08876228e+00 -1.20503271e+00 2.39770845e-01 1.08202708e+00
1.52363658e-01 6.50740862e-01 3.85701746e-01 -2.19215319e-01
-1.34597957e+00 -1.26707590e+00 2.72245735e-01 -3.41919273e-01
6.85962677e-01 -2.36145109e-01 -1.09252477e+00 4.66515452e-01
1.55109853e-01 4.78695720e-01 2.41595894e-01 -1.49854660e-01
-8.51662397e-01 -6.93623543e-01 -1.06090319e+00 5.53556919e-01
9.75194693e-01 -8.81995618e-01 -3.35639089e-01 -2.15859652e-01
8.60071898e-01 -2.49945313e-01 -9.99219775e-01 5.92152178e-01
6.51587009e-01 -1.24267602e+00 9.85460401e-01 -2.11265907e-01
6.00488365e-01 -6.52468860e-01 -1.32478088e-01 -1.37920332e+00
-4.42696571e-01 -3.78250480e-01 8.78107995e-02 1.38151085e+00
3.10830444e-01 -5.74941218e-01 4.25135732e-01 4.95902777e-01
-7.79857486e-02 -5.81782103e-01 -9.21643555e-01 -7.52308428e-01
-1.28786221e-01 -2.43792161e-01 8.48693013e-01 8.04695904e-01
-2.41965190e-01 2.88346738e-01 -5.25365114e-01 4.75009143e-01
6.18494391e-01 2.16535330e-01 7.05108285e-01 -7.75181949e-01
-3.95272970e-01 -3.45166385e-01 -4.19841141e-01 -9.12358344e-01
1.09936923e-01 -5.89573443e-01 1.19339049e-01 -1.03575456e+00
2.74361610e-01 -2.81681418e-01 -8.72268796e-01 5.19824147e-01
-3.00805897e-01 4.62526381e-01 3.76253456e-01 2.22332910e-01
-8.33086550e-01 7.84023404e-01 1.51066577e+00 -4.83144462e-01
-1.69308096e-01 -3.59714627e-01 -8.21290672e-01 3.59175771e-01
5.85458994e-01 -3.38151634e-01 -4.04056579e-01 -5.01168013e-01
2.25792810e-01 4.83580939e-02 6.32589519e-01 -1.22946250e+00
2.11866364e-01 -1.09264620e-01 4.66615617e-01 -1.74076781e-01
1.32061809e-01 -9.45084095e-01 1.46510340e-02 1.02595344e-01
-4.79406983e-01 9.34425816e-02 -7.98929110e-02 5.39185047e-01
-5.10686874e-01 1.73518717e-01 9.81723130e-01 1.23259671e-01
-4.51924086e-01 4.69649404e-01 3.03435147e-01 -8.72470737e-02
6.35204852e-01 -6.27647489e-02 -4.85150009e-01 -4.26881045e-01
-5.47395527e-01 4.89794537e-02 7.41644382e-01 3.97226244e-01
5.90679288e-01 -1.49116838e+00 -5.17179787e-01 3.13123375e-01
4.52827066e-01 -4.23988290e-02 4.46479201e-01 8.64131093e-01
1.43761467e-02 8.03238153e-02 -5.59769869e-01 -6.67391002e-01
-8.71837974e-01 8.70470047e-01 4.28399056e-01 -4.38118935e-01
-4.03463870e-01 2.05177575e-01 4.50678676e-01 -2.00158671e-01
2.25257531e-01 -4.86231565e-01 -2.12272897e-01 -2.72263110e-01
4.17573988e-01 1.42127171e-01 2.51242816e-01 -7.11453795e-01
-1.65663153e-01 7.48055100e-01 3.23164724e-02 -1.72368288e-01
1.19188714e+00 -4.22687978e-01 5.71113601e-02 2.56682813e-01
1.41201854e+00 3.38246934e-02 -1.57831585e+00 -3.72087419e-01
-1.99457452e-01 -6.11402333e-01 9.03750956e-02 -9.64298844e-01
-1.51333034e+00 9.01339114e-01 9.94267344e-01 -5.11075817e-02
1.66419458e+00 -2.68793762e-01 6.55463219e-01 2.16253504e-01
6.40038103e-02 -1.00111043e+00 5.18575609e-01 1.51614085e-01
9.79089618e-01 -1.43592012e+00 -1.65673092e-01 -2.52174973e-01
-5.49833834e-01 9.39213216e-01 6.44743323e-01 -1.09135345e-01
4.15792197e-01 -1.07525982e-01 2.34737545e-01 -7.07581034e-03
-6.21793389e-01 -2.41513759e-01 6.12058997e-01 5.73723197e-01
4.99926582e-02 -1.45418137e-01 1.52758852e-01 6.59521818e-01
-8.15527439e-02 4.20229189e-04 3.95544857e-01 3.79222572e-01
-1.47976831e-03 -8.58993292e-01 -2.47968435e-01 1.36236072e-01
-2.96392947e-01 -8.73373523e-02 -2.74216346e-02 5.38343668e-01
2.25247294e-01 1.12311029e+00 -1.52906328e-01 -7.43373036e-01
4.95156199e-01 -2.66352654e-01 2.29326218e-01 -2.06401348e-01
-6.59447014e-01 -8.20425600e-02 -3.78676146e-01 -9.54421520e-01
-4.81384218e-01 -2.86620736e-01 -8.42011690e-01 -1.35889128e-01
-2.32954323e-01 -2.21501186e-01 4.72281456e-01 8.53971958e-01
4.88985658e-01 7.09994614e-01 1.16396117e+00 -5.76241672e-01
-7.38835335e-01 -8.22730124e-01 -5.49607813e-01 9.48970497e-01
5.22270739e-01 -6.27573490e-01 -4.71233189e-01 1.87991366e-01] | [11.853885650634766, -1.826348900794983] |
ecf13cca-b598-4a3f-993e-73364082af26 | real-time-neural-radiance-talking-portrait | 2211.12368 | null | https://arxiv.org/abs/2211.12368v1 | https://arxiv.org/pdf/2211.12368v1.pdf | Real-time Neural Radiance Talking Portrait Synthesis via Audio-spatial Decomposition | While dynamic Neural Radiance Fields (NeRF) have shown success in high-fidelity 3D modeling of talking portraits, the slow training and inference speed severely obstruct their potential usage. In this paper, we propose an efficient NeRF-based framework that enables real-time synthesizing of talking portraits and faster convergence by leveraging the recent success of grid-based NeRF. Our key insight is to decompose the inherently high-dimensional talking portrait representation into three low-dimensional feature grids. Specifically, a Decomposed Audio-spatial Encoding Module models the dynamic head with a 3D spatial grid and a 2D audio grid. The torso is handled with another 2D grid in a lightweight Pseudo-3D Deformable Module. Both modules focus on efficiency under the premise of good rendering quality. Extensive experiments demonstrate that our method can generate realistic and audio-lips synchronized talking portrait videos, while also being highly efficient compared to previous methods. | ['Jingdong Wang', 'Gang Zeng', 'Jingtuo Liu', 'Tianshu Hu', 'Dongliang He', 'Xiaokang Chen', 'Hang Zhou', 'Kaisiyuan Wang', 'Jiaxiang Tang'] | 2022-11-22 | null | null | null | null | ['talking-face-generation'] | ['computer-vision'] | [ 1.67924836e-01 3.40353660e-02 2.25315422e-01 -1.49786085e-01
-1.01395273e+00 -4.02537107e-01 5.62852561e-01 -6.67822182e-01
2.36708820e-01 5.19043446e-01 3.70495856e-01 -6.34064944e-03
3.50700766e-02 -8.08138669e-01 -5.56026459e-01 -8.28879893e-01
-6.28370121e-02 2.92303741e-01 -4.89418991e-02 -1.45723581e-01
-2.93110639e-01 6.19195640e-01 -2.02251410e+00 2.82917500e-01
8.40957105e-01 1.17669106e+00 1.01741888e-02 9.66289997e-01
1.00466251e-01 7.06735373e-01 -5.48669755e-01 -2.79641062e-01
1.92884132e-01 -3.31709802e-01 -2.52594441e-01 2.81074136e-01
7.30995595e-01 -4.44087833e-01 -4.96325642e-01 3.27027380e-01
8.19252431e-01 2.18688577e-01 6.26133740e-01 -1.42646134e+00
-3.22014779e-01 -3.62653409e-05 -3.57263386e-01 -4.27071303e-01
7.09663510e-01 1.14931807e-01 6.21812761e-01 -7.85616040e-01
6.05293870e-01 1.40250766e+00 1.14382148e+00 7.32604384e-01
-1.12688112e+00 -7.76965857e-01 -2.94727716e-03 -1.59678400e-01
-1.73352110e+00 -5.87690711e-01 1.05113709e+00 -1.44502357e-01
7.31400132e-01 7.40250647e-01 1.04112566e+00 1.11303687e+00
2.74571896e-01 6.21851444e-01 1.07857406e+00 -9.72478464e-02
2.52846807e-01 3.41340378e-02 -6.43636167e-01 7.53008127e-01
-6.44997656e-01 -7.79338256e-02 -9.71642435e-01 -3.78827870e-01
1.00582647e+00 -3.84835035e-01 -3.92813683e-01 -1.95940524e-01
-9.32793856e-01 4.78283167e-01 3.48188013e-01 2.22782984e-01
-3.07547241e-01 4.15891320e-01 1.27119765e-01 -2.44805172e-01
7.82857180e-01 -1.68864682e-01 1.35907754e-01 -2.13271052e-01
-1.33101141e+00 5.10964274e-01 6.71285868e-01 7.62323499e-01
4.05761033e-01 3.17422777e-01 -2.94695050e-02 8.97138596e-01
2.64882654e-01 7.48344600e-01 1.93018749e-01 -1.24421835e+00
5.36262691e-02 2.89137423e-01 3.34841460e-02 -1.20426691e+00
-3.72232348e-01 -2.52745867e-01 -1.12207174e+00 3.04100007e-01
1.03630144e-02 -1.66632805e-03 -5.84903717e-01 1.82819009e+00
9.77966845e-01 6.91382885e-01 -6.39178976e-02 1.11260712e+00
7.87984490e-01 9.56706762e-01 -4.80804071e-02 -1.14229657e-01
1.27881110e+00 -6.23373210e-01 -8.39157581e-01 2.64677614e-01
2.31560782e-01 -6.80543363e-01 1.13388002e+00 5.31740129e-01
-1.35761487e+00 -5.85918069e-01 -7.55667448e-01 -4.63093936e-01
-9.76744946e-03 6.50063008e-02 7.83408284e-01 6.36578619e-01
-1.40319884e+00 3.45793605e-01 -5.80704570e-01 -5.53093292e-03
1.34765014e-01 2.35180244e-01 -2.22831279e-01 3.33991498e-02
-1.04196095e+00 5.13908446e-01 -3.35168451e-01 7.00351894e-02
-8.00340056e-01 -1.13402486e+00 -9.16874528e-01 7.68826678e-02
-5.14434502e-02 -9.98668313e-01 1.15318620e+00 -7.76832283e-01
-2.06253457e+00 6.94543302e-01 -1.95359945e-01 -2.19340175e-01
6.11401558e-01 1.79147571e-01 -4.63788778e-01 3.41096997e-01
-2.50932127e-01 1.00293088e+00 8.91054630e-01 -1.49850392e+00
-2.37390891e-01 -1.44150689e-01 -3.75351906e-02 5.95704734e-01
-4.85398203e-01 -3.30805212e-01 -4.24761236e-01 -8.07007194e-01
-4.61715385e-02 -8.92709851e-01 9.08730179e-02 6.27303243e-01
-3.58952135e-01 2.50040814e-02 1.04930663e+00 -6.10930204e-01
1.15379632e+00 -2.04397249e+00 1.11049436e-01 5.77942431e-02
2.59260535e-01 -1.91690084e-02 5.30329458e-02 2.38712877e-01
1.03351615e-01 5.31321838e-02 -2.02192679e-01 -8.86727095e-01
2.11567178e-01 1.41047910e-01 -3.90204638e-01 5.06163776e-01
8.85243267e-02 6.81996405e-01 -7.15082765e-01 -6.93094492e-01
3.99583548e-01 1.68879044e+00 -8.01225364e-01 1.06319197e-01
-1.47064149e-01 5.30521929e-01 -4.16987181e-01 6.87405884e-01
9.10260379e-01 1.04621844e-02 2.25533452e-02 -4.30008739e-01
-2.36120075e-01 2.65501112e-01 -1.36513078e+00 1.93105805e+00
-8.72107029e-01 6.49733782e-01 6.62537992e-01 -2.87886947e-01
7.39898860e-01 6.18707120e-01 6.94510341e-01 -7.09580541e-01
4.07151505e-02 -1.47501513e-01 -7.93454945e-01 -3.41813594e-01
6.48413718e-01 -3.36482316e-01 -1.95484638e-01 3.54107082e-01
-9.97832268e-02 -7.38597751e-01 -5.14790356e-01 8.37573931e-02
7.66018629e-01 3.95070910e-01 -2.52564400e-01 -2.55542129e-01
3.54609489e-01 -3.54315042e-01 5.03289461e-01 1.22570693e-01
2.44386047e-01 9.46480930e-01 1.56342119e-01 -3.82568330e-01
-9.74944592e-01 -1.15197492e+00 -2.22659588e-01 6.71893537e-01
7.58242235e-03 -6.58899903e-01 -1.18448544e+00 -1.13701798e-01
-1.96205318e-01 6.30404651e-01 -7.05877244e-01 1.87129378e-01
-6.41073108e-01 -4.42803621e-01 6.65493131e-01 2.22420186e-01
5.62172711e-01 -5.98215997e-01 -7.29693055e-01 8.40166137e-02
-3.83222401e-01 -1.14142501e+00 -6.80326819e-01 -3.17733377e-01
-5.76385021e-01 -6.42023325e-01 -9.36869383e-01 -5.21068215e-01
4.45252925e-01 3.23689908e-01 1.07811403e+00 -1.23841465e-01
-4.22824889e-01 6.58625543e-01 1.26247123e-01 -1.72234356e-01
-4.00013506e-01 -2.04426408e-01 2.51021296e-01 2.71525383e-01
-4.88758445e-01 -9.94825780e-01 -7.00206459e-01 4.46427405e-01
-9.54240263e-01 6.91817999e-01 -8.91540349e-02 5.07626832e-01
7.88196921e-01 9.50672030e-02 1.89957693e-01 -1.09956391e-01
4.39477473e-01 -2.77632356e-01 -4.42126155e-01 5.55596761e-02
6.47655129e-02 -3.24330956e-01 6.63642168e-01 -7.29414046e-01
-1.23798990e+00 1.08859941e-01 -4.76118207e-01 -5.96386433e-01
-6.57117739e-02 -1.29606456e-01 -4.12355751e-01 -2.22786665e-01
4.13257927e-01 1.80795744e-01 -1.25440300e-01 -3.35333824e-01
1.84333608e-01 6.42330408e-01 6.92822456e-01 -5.75617313e-01
8.69480193e-01 6.94904625e-01 1.09135129e-01 -1.02970660e+00
-4.80639011e-01 8.97524133e-02 -2.91880220e-01 -8.36580873e-01
8.60377192e-01 -9.88578737e-01 -1.16513479e+00 5.65710545e-01
-9.00892615e-01 -6.97360098e-01 -4.57547337e-01 1.74661741e-01
-8.24699759e-01 1.27210245e-01 -4.59460497e-01 -1.03768671e+00
-3.29401016e-01 -1.09180653e+00 1.76834464e+00 2.45259255e-01
-2.68573523e-01 -9.73924816e-01 2.34329686e-01 5.24077415e-01
5.58152854e-01 7.01178432e-01 5.75210750e-01 4.92439300e-01
-6.33848965e-01 -1.43661378e-02 1.94838077e-01 -3.52394022e-02
-5.22803143e-02 2.07606964e-02 -1.48215258e+00 -3.66294347e-02
5.96190020e-02 -3.61013651e-01 2.11304724e-01 4.64206457e-01
1.17608273e+00 -3.66758198e-01 -2.72141457e-01 1.00066853e+00
1.26463938e+00 -1.51205793e-01 5.46111166e-01 -3.41049820e-01
7.37754464e-01 5.35594344e-01 2.57215172e-01 6.95444822e-01
7.15552032e-01 1.08856165e+00 3.16676319e-01 -2.76986808e-01
-6.79013431e-01 -4.40837324e-01 3.13084126e-01 9.71010327e-01
-1.60458069e-02 -5.21743059e-01 -7.65465617e-01 2.52266288e-01
-1.42188811e+00 -8.55982721e-01 1.98146835e-01 2.00304365e+00
7.38799810e-01 -2.95288771e-01 1.57058120e-01 3.83211285e-01
3.78994077e-01 1.51993677e-01 -6.09709799e-01 -3.89994115e-01
-2.23175749e-01 2.48456597e-02 -1.30630434e-01 7.18291163e-01
-6.36251211e-01 7.60194898e-01 6.37761068e+00 1.02024591e+00
-1.21663666e+00 5.89704327e-02 8.56319427e-01 -4.48822200e-01
-8.98774147e-01 -4.21737015e-01 -8.73301148e-01 3.47789824e-01
1.08196044e+00 -5.49223088e-02 5.03897011e-01 6.43673837e-01
5.23660898e-01 -8.40837210e-02 -6.91510201e-01 1.24602222e+00
2.85456926e-01 -1.45917559e+00 1.03395604e-01 2.28060573e-01
4.91803944e-01 -4.21015292e-01 2.94021308e-01 3.65055166e-02
-8.71855095e-02 -1.13146091e+00 9.71941113e-01 7.60006726e-01
1.41380560e+00 -9.62893009e-01 -1.62854806e-01 3.89649063e-01
-1.40475190e+00 3.18648666e-01 3.96413580e-02 3.01372975e-01
4.31324989e-01 5.89761198e-01 -6.52422190e-01 4.20656323e-01
8.02389681e-01 4.03131783e-01 5.51134385e-02 7.74629235e-01
2.67757267e-01 3.96845698e-01 -7.18708873e-01 1.91385552e-01
1.39924929e-01 -6.66743740e-02 6.57780111e-01 1.15746331e+00
6.15794539e-01 4.02776390e-01 -1.26738757e-01 8.35915089e-01
-6.74201772e-02 -5.63275963e-02 -5.75794220e-01 4.32403952e-01
2.86253363e-01 1.37763298e+00 -4.99593526e-01 -1.23671487e-01
5.74117154e-02 1.04555178e+00 -1.17737204e-01 2.10396260e-01
-1.31121516e+00 -5.16945086e-02 9.63421762e-01 5.28654635e-01
1.43383816e-01 -3.91398460e-01 -1.64719284e-01 -1.06925297e+00
-2.89182272e-02 -7.69209445e-01 -2.00247601e-01 -1.29859734e+00
-8.51199389e-01 9.52333272e-01 -3.13933529e-02 -1.17545593e+00
-3.20584178e-01 -2.10275561e-01 -4.12416667e-01 7.32196152e-01
-1.47505057e+00 -1.55928719e+00 -6.59863174e-01 9.06352997e-01
4.62550372e-01 4.23283070e-01 1.20089436e+00 3.91982555e-01
-3.55725735e-01 7.29000211e-01 -7.10459203e-02 -3.80535573e-01
4.55045044e-01 -8.20793986e-01 3.86566043e-01 2.78208494e-01
-6.99036196e-02 6.12824298e-02 6.48027003e-01 -3.45333129e-01
-1.79386139e+00 -9.62431610e-01 7.11423635e-01 -1.59292623e-01
1.16469622e-01 -9.57713783e-01 -7.90314496e-01 7.87531585e-02
2.04932928e-01 -6.58737198e-02 7.82745361e-01 -3.74629021e-01
-3.23591650e-01 -3.50259423e-01 -1.60967040e+00 7.87070811e-01
1.15943527e+00 -5.43328226e-01 1.14737600e-01 4.24454987e-01
4.84326631e-01 -7.93226898e-01 -1.14020658e+00 2.88564652e-01
7.92647362e-01 -1.03476620e+00 1.19838560e+00 3.04087609e-01
1.01378433e-01 -3.53069067e-01 -2.31192440e-01 -1.11971951e+00
3.25414948e-02 -1.27820551e+00 -1.60712108e-01 1.25623786e+00
-2.66481370e-01 -2.76839584e-01 8.03539217e-01 7.07596362e-01
-1.55357078e-01 -7.68013418e-01 -1.31903911e+00 -3.40495646e-01
-1.36319995e-01 -7.97875404e-01 7.58243024e-01 7.17098653e-01
-2.71534890e-01 1.38005316e-01 -6.81242168e-01 1.78579152e-01
6.99244440e-01 8.43259133e-03 7.47354746e-01 -1.11440670e+00
-1.65484518e-01 -1.10789999e-01 -3.03182721e-01 -1.09771430e+00
2.79049158e-01 -5.40269852e-01 -5.30040544e-03 -1.32971776e+00
-9.26782638e-02 -7.98856080e-01 3.77764434e-01 3.88806015e-01
2.90275604e-01 6.94911838e-01 1.62471712e-01 -1.14044122e-01
-3.40670198e-01 7.85226882e-01 1.34847665e+00 4.46779430e-02
-2.11864233e-01 -1.92896187e-01 -3.43541265e-01 7.12358832e-01
5.31633139e-01 -2.06479028e-01 -6.21300459e-01 -5.47417581e-01
1.47854477e-01 4.15507406e-01 6.25920594e-01 -1.46472585e+00
2.62572289e-01 1.53122768e-01 4.82977897e-01 -7.13660598e-01
1.18080056e+00 -9.07272041e-01 6.55319452e-01 -7.73173720e-02
-2.04432964e-01 5.30294850e-02 5.49386799e-01 2.43600428e-01
-3.01654208e-02 6.19628847e-01 9.41985905e-01 1.46354124e-01
-7.27649406e-02 4.03387517e-01 -5.41259289e-01 -9.13079679e-02
8.63610566e-01 -4.59953219e-01 2.69212723e-01 -9.22853589e-01
-6.44811094e-01 -1.50263965e-01 6.28021181e-01 3.20125550e-01
7.03151405e-01 -1.55033231e+00 -5.87481916e-01 3.99850458e-01
-4.36525226e-01 1.65894210e-01 8.43514979e-01 4.54616100e-01
-3.93512487e-01 1.35725811e-01 -2.28828434e-02 -8.32547486e-01
-1.37801564e+00 1.35956481e-01 5.87656260e-01 -7.93281272e-02
-9.08253729e-01 8.20888937e-01 3.58247310e-01 -4.20686841e-01
5.30054986e-01 -3.12318861e-01 1.62200481e-01 -5.67848459e-02
7.24291205e-01 1.94614664e-01 3.54639105e-02 -9.59472120e-01
-2.29391500e-01 1.06461763e+00 5.66231072e-01 -3.31478596e-01
1.40957129e+00 -2.24637374e-01 3.09608161e-01 3.42684686e-01
1.28397274e+00 1.99793220e-01 -1.63353717e+00 1.30450144e-01
-9.13909316e-01 -4.21464443e-01 4.34224665e-01 -6.79911256e-01
-1.06314707e+00 8.54017079e-01 5.38950026e-01 1.26174644e-01
1.57161248e+00 -3.09296995e-01 1.18421698e+00 -1.53607428e-01
3.74793530e-01 -7.63271749e-01 -1.31273687e-01 2.26583377e-01
9.65315282e-01 -5.10518968e-01 -1.72765687e-01 -6.96818471e-01
-5.51516294e-01 1.03778732e+00 2.22695097e-01 -3.27194519e-02
6.13604188e-01 9.35850143e-01 -1.91756934e-02 5.74335866e-02
-8.43744099e-01 3.50202233e-01 4.18900609e-01 6.81587815e-01
2.95646667e-01 -1.81753650e-01 4.40751195e-01 2.72479773e-01
-5.42284191e-01 1.39926061e-01 1.76528454e-01 6.32398307e-01
8.49355161e-02 -7.33021379e-01 -5.87866008e-01 -2.60447830e-01
-2.84079343e-01 4.74114117e-04 -1.09600879e-01 8.16379607e-01
3.09625357e-01 7.64377892e-01 3.42766315e-01 -4.50863123e-01
2.93706656e-01 1.20637575e-02 6.46984220e-01 8.63132346e-03
-8.53648067e-01 4.11829412e-01 6.32436126e-02 -8.18686843e-01
-3.28609377e-01 -4.14441407e-01 -1.09341693e+00 -6.41295552e-01
-9.16432962e-03 1.32442281e-01 1.03128624e+00 3.73531431e-01
7.54720867e-01 5.66903591e-01 7.66794562e-01 -1.40058768e+00
7.97826946e-02 -6.00058615e-01 -4.46853250e-01 -1.09648779e-02
4.81495202e-01 -5.60537755e-01 -3.47998530e-01 1.55755892e-01] | [13.122381210327148, -0.45722895860671997] |
3051570c-76ce-4a54-901e-7f219648ad97 | distilling-motion-planner-augmented-policies | 2111.06383 | null | https://arxiv.org/abs/2111.06383v1 | https://arxiv.org/pdf/2111.06383v1.pdf | Distilling Motion Planner Augmented Policies into Visual Control Policies for Robot Manipulation | Learning complex manipulation tasks in realistic, obstructed environments is a challenging problem due to hard exploration in the presence of obstacles and high-dimensional visual observations. Prior work tackles the exploration problem by integrating motion planning and reinforcement learning. However, the motion planner augmented policy requires access to state information, which is often not available in the real-world settings. To this end, we propose to distill a state-based motion planner augmented policy to a visual control policy via (1) visual behavioral cloning to remove the motion planner dependency along with its jittery motion, and (2) vision-based reinforcement learning with the guidance of the smoothed trajectories from the behavioral cloning agent. We evaluate our method on three manipulation tasks in obstructed environments and compare it against various reinforcement learning and imitation learning baselines. The results demonstrate that our framework is highly sample-efficient and outperforms the state-of-the-art algorithms. Moreover, coupled with domain randomization, our policy is capable of zero-shot transfer to unseen environment settings with distractors. Code and videos are available at https://clvrai.com/mopa-pd | ['Youngwoon Lee', 'Peter Englert', 'Joseph J. Lim', 'Gaurav S. Sukhatme', 'Shagun Uppal', 'I-Chun Arthur Liu'] | 2021-11-11 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [ 3.70021313e-02 -3.77442800e-02 -2.88076997e-01 2.96371192e-01
-6.19364202e-01 -6.56496823e-01 8.15345705e-01 -1.30278096e-01
-7.66002178e-01 9.34787452e-01 2.64688522e-01 -3.63666594e-01
-2.78129447e-02 -4.51001495e-01 -1.01201248e+00 -6.58733666e-01
-3.04746389e-01 5.38370550e-01 4.65365827e-01 -2.65051216e-01
2.62955725e-01 4.41894442e-01 -1.43964052e+00 -2.88463145e-01
9.22423124e-01 6.18511856e-01 6.28702760e-01 8.86000931e-01
2.27151662e-01 8.45344365e-01 -6.98900670e-02 4.43511486e-01
4.85243857e-01 -2.28850186e-01 -6.08470380e-01 2.35690922e-01
1.56821936e-01 -7.63640344e-01 -6.57083035e-01 8.71387124e-01
4.98028815e-01 7.33405411e-01 5.43439567e-01 -1.42943275e+00
-4.64734405e-01 1.23597205e-01 -5.47291934e-01 1.40383586e-01
4.00618762e-01 9.54475105e-01 5.33756196e-01 -5.25462747e-01
9.77483094e-01 1.45827007e+00 2.15072222e-02 6.30015373e-01
-1.26577616e+00 -2.80171901e-01 5.19598961e-01 3.62429827e-01
-8.60422552e-01 -4.50754285e-01 5.37144959e-01 -5.60663104e-01
1.00861549e+00 -2.64912456e-01 6.84007466e-01 1.72335958e+00
1.34675667e-01 9.44528162e-01 1.05621839e+00 -1.63220838e-01
5.79977870e-01 -3.02496433e-01 -5.04565597e-01 9.24171269e-01
6.10734597e-02 7.24383652e-01 -3.04312617e-01 -6.07750490e-02
9.83388186e-01 -4.55171242e-02 -2.56692976e-01 -1.17958093e+00
-1.39476871e+00 7.67420053e-01 4.88844633e-01 -1.76230311e-01
-5.35182834e-01 5.00124335e-01 2.28194967e-01 2.73833215e-01
-4.83981743e-02 5.85097015e-01 -3.32013041e-01 -4.80024248e-01
-3.03027481e-01 7.90716708e-01 7.08527267e-01 1.22116876e+00
6.47018075e-01 3.20909560e-01 -3.53401005e-01 4.10127997e-01
1.81842819e-01 6.57405674e-01 2.98798472e-01 -1.47989142e+00
5.81569374e-01 6.78440630e-02 7.37038553e-01 -5.83913207e-01
-3.81071597e-01 -4.87083383e-02 -3.08141977e-01 9.60160434e-01
4.74391609e-01 -4.03769761e-01 -1.08261502e+00 1.75321579e+00
7.00102270e-01 3.51783186e-01 1.52931688e-02 1.13449454e+00
2.39801779e-01 4.68566626e-01 5.14162146e-02 3.85963991e-02
7.88952172e-01 -1.49962580e+00 -6.80975616e-01 -5.73814452e-01
3.92801821e-01 -3.78313750e-01 1.28306997e+00 1.52639359e-01
-9.62789953e-01 -5.03140271e-01 -8.78376663e-01 6.40485510e-02
-7.29021132e-02 2.36243829e-02 3.28597337e-01 1.73900500e-02
-8.95640433e-01 7.33518600e-01 -1.36951184e+00 -4.57055211e-01
4.38419193e-01 2.08087608e-01 -4.28375453e-01 -1.48437619e-01
-8.10710251e-01 1.11981070e+00 4.39712197e-01 -1.05612844e-01
-1.67491400e+00 -3.41699749e-01 -1.12881196e+00 -1.53989866e-01
1.09254825e+00 -7.61345625e-01 1.58680212e+00 -3.99441123e-01
-2.03245926e+00 4.59230840e-02 -3.65862474e-02 -2.85749018e-01
9.47552383e-01 -5.21100104e-01 3.05429369e-01 1.76294222e-01
2.29581535e-01 7.42661536e-01 9.41250086e-01 -1.46231222e+00
-7.50431776e-01 -1.15122952e-01 3.13486487e-01 6.72146320e-01
2.35754028e-01 -6.01009667e-01 -5.71149647e-01 -3.69376093e-01
-3.43675435e-01 -1.32014048e+00 -6.57728732e-01 2.74711519e-01
-2.06114978e-01 1.45282298e-01 8.58845532e-01 -4.78146285e-01
7.10740149e-01 -1.99265182e+00 6.39637649e-01 -1.46278426e-01
-3.24065089e-02 2.60522336e-01 -3.82768989e-01 6.86173379e-01
4.49208885e-01 -2.97968358e-01 -2.37825051e-01 -3.64183307e-01
2.28816956e-01 4.05670226e-01 -3.33030075e-01 6.27110004e-01
-1.21173896e-02 1.00761831e+00 -1.45104480e+00 -3.16872269e-01
4.99510914e-01 2.73106366e-01 -6.76988244e-01 4.77743030e-01
-5.97638369e-01 1.15622926e+00 -7.46693432e-01 6.08528078e-01
3.29293638e-01 1.85423139e-02 2.35536039e-01 4.49308962e-01
-3.36193025e-01 3.87409329e-02 -1.16340601e+00 2.11055970e+00
-3.62945378e-01 3.39799494e-01 3.62796098e-01 -7.07078397e-01
5.00110447e-01 2.30198056e-02 3.25532407e-01 -7.52742112e-01
-5.79048656e-02 2.72262320e-02 -1.17240109e-01 -8.51032615e-01
5.26791513e-01 2.08668634e-01 -4.04898115e-02 2.89328188e-01
9.52782556e-02 -4.01122242e-01 1.42667338e-01 6.70290142e-02
1.40969408e+00 9.01789904e-01 2.88972557e-01 1.54686809e-01
1.26150325e-01 2.07757339e-01 6.35847986e-01 1.07553828e+00
-5.58312356e-01 2.26541311e-01 3.57155889e-01 -1.66129440e-01
-1.14424396e+00 -1.28511870e+00 4.45059657e-01 1.01912320e+00
5.01861334e-01 -1.51289657e-01 -4.48978812e-01 -6.24632597e-01
2.36965135e-01 6.99793398e-01 -5.91264427e-01 -1.00333579e-01
-8.39157462e-01 2.66076177e-02 -3.13418289e-03 5.12872398e-01
4.34336931e-01 -1.31021607e+00 -1.15211213e+00 3.36172849e-01
-1.40979171e-01 -1.22601759e+00 -4.94827479e-01 3.35797705e-02
-6.25551343e-01 -1.03587532e+00 -6.85389400e-01 -5.83414733e-01
4.35604155e-01 3.63959372e-01 6.78624034e-01 -3.04950744e-01
-2.21831903e-01 7.66798437e-01 -4.32685882e-01 -1.88579798e-01
-3.35743010e-01 -3.75126079e-02 7.47507662e-02 -4.68308568e-01
-1.97708890e-01 -5.77200234e-01 -7.35508919e-01 1.24521703e-01
-6.90577269e-01 1.06822565e-01 5.78138590e-01 1.03333330e+00
4.98914480e-01 -3.81202996e-01 1.93603680e-01 -3.60814929e-01
5.83020926e-01 -4.84158784e-01 -8.80468726e-01 -7.56624416e-02
-2.90629238e-01 2.85075903e-01 4.95112181e-01 -7.45462596e-01
-1.14565969e+00 3.34810674e-01 2.69185632e-01 -6.45585954e-01
-3.86290342e-01 1.51418895e-01 5.40175810e-02 -1.48925528e-01
6.97363317e-01 1.06090195e-01 9.43487361e-02 -3.37489307e-01
6.87361181e-01 2.65910149e-01 6.09178424e-01 -8.33198369e-01
9.18552756e-01 7.21001863e-01 4.76547721e-04 -8.04214835e-01
-3.31898838e-01 -4.01655883e-01 -5.78185618e-01 -3.09659570e-01
9.15467203e-01 -8.09128165e-01 -7.82262921e-01 3.32576305e-01
-1.03694808e+00 -1.10361433e+00 -3.01650614e-01 6.85839653e-01
-1.24661958e+00 4.78479058e-01 -5.43125689e-01 -8.95932615e-01
1.35413006e-01 -1.45269430e+00 1.05249548e+00 2.61997551e-01
-1.50029105e-03 -6.87148631e-01 2.90279746e-01 -3.46524641e-02
3.39069068e-01 6.14419341e-01 6.20670915e-01 -1.14052229e-01
-9.22998667e-01 1.81186959e-01 1.31115213e-01 -1.58926100e-01
6.52346984e-02 -3.32757145e-01 -5.50006986e-01 -6.32446051e-01
-2.73190230e-01 -7.19770670e-01 7.73243845e-01 3.67284507e-01
8.47813606e-01 -3.31107050e-01 -4.77451086e-01 6.09107554e-01
1.21590745e+00 3.82110476e-01 3.32905322e-01 7.31442690e-01
5.91577053e-01 5.39323151e-01 9.67270255e-01 7.39665329e-01
4.78984296e-01 8.24756563e-01 8.62708986e-01 1.60166025e-01
6.36893362e-02 -5.68256438e-01 4.47252154e-01 1.55496448e-01
-5.63426316e-02 -1.82017341e-01 -8.50765824e-01 7.27537274e-01
-2.37709546e+00 -9.52764511e-01 3.59564155e-01 2.06451702e+00
4.67634976e-01 5.27516082e-02 2.73696184e-01 -4.24869567e-01
3.92251819e-01 3.32334697e-01 -1.05125523e+00 -1.73559397e-01
3.54900301e-01 -8.49844441e-02 5.42245269e-01 6.74405336e-01
-1.24120486e+00 1.30852282e+00 5.63969994e+00 4.17614788e-01
-8.56184661e-01 1.36701027e-02 -1.67300656e-01 -3.10832173e-01
2.12074760e-02 1.35100484e-01 -4.78895158e-01 3.48783672e-01
4.60498929e-01 -5.85531890e-02 8.42754662e-01 9.60011423e-01
4.49667305e-01 -4.42420453e-01 -1.08226693e+00 6.22103035e-01
-4.24333930e-01 -1.00647688e+00 -4.14864659e-01 1.70142055e-01
6.75111592e-01 2.60888517e-01 1.20633818e-01 6.91410542e-01
9.81654823e-01 -9.67833519e-01 8.05326045e-01 4.12097514e-01
5.45594811e-01 -5.09799182e-01 8.55708867e-02 7.25238264e-01
-1.04820573e+00 -3.94467294e-01 -2.76206076e-01 -2.01758936e-01
4.75522608e-01 -2.73328990e-01 -8.30925465e-01 4.79454994e-01
5.32729864e-01 6.80013657e-01 -6.57608062e-02 1.22395456e+00
-5.98253846e-01 2.03924254e-01 -2.23996848e-01 -1.39115095e-01
6.67590916e-01 -2.32738882e-01 9.35690939e-01 8.59728575e-01
2.69109607e-01 -2.41617803e-02 7.61451364e-01 8.36734772e-01
3.51879805e-01 -4.43965822e-01 -9.88842368e-01 2.62752473e-02
4.96503592e-01 8.85084271e-01 -4.70554024e-01 -1.78155214e-01
-2.18030632e-01 1.10380530e+00 7.14400351e-01 8.09817672e-01
-8.68847489e-01 -1.53449267e-01 9.37420964e-01 -1.97657138e-01
7.19347835e-01 -7.89400697e-01 2.61664420e-01 -1.15276194e+00
1.66639999e-01 -8.48837614e-01 3.45894657e-02 -6.47109032e-01
-8.89332175e-01 3.09134215e-01 2.28564516e-01 -1.24178660e+00
-5.28022468e-01 -4.30108070e-01 -4.12093341e-01 7.32811809e-01
-1.81388712e+00 -9.52307343e-01 -3.78311336e-01 5.96732616e-01
7.86649883e-01 -1.06733561e-01 5.97588956e-01 -2.15948984e-01
-2.80233860e-01 4.87015918e-02 2.14219376e-01 -1.07596867e-01
7.29927421e-01 -1.28997386e+00 5.51986217e-01 7.28776932e-01
-3.32100600e-01 2.43024081e-01 8.14722300e-01 -8.23298156e-01
-1.68491626e+00 -1.04989934e+00 3.78510426e-03 -3.79277468e-01
7.67085075e-01 -4.31147188e-01 -9.00958598e-01 7.85596669e-01
2.58036494e-01 1.81038842e-01 -7.45817721e-02 -3.40949982e-01
-3.33547331e-02 3.95946920e-01 -9.05535460e-01 9.82051373e-01
1.19929039e+00 -2.37291798e-01 -6.50087953e-01 2.09724188e-01
7.85143495e-01 -8.69720280e-01 -3.06459308e-01 3.21889997e-01
6.51705027e-01 -6.49330497e-01 1.02243567e+00 -9.02559102e-01
3.53008956e-01 -4.19977516e-01 -2.16522049e-02 -1.63846397e+00
-4.47893292e-01 -8.16779137e-01 -3.87923568e-01 4.72226411e-01
9.82208252e-02 -4.17533338e-01 6.78100169e-01 3.28722447e-01
-9.73718315e-02 -5.66488981e-01 -9.62632895e-01 -1.14032865e+00
1.09441988e-01 -1.44732311e-01 1.59011170e-01 5.01489758e-01
-4.72452454e-02 1.25531077e-01 -6.93673193e-01 3.13262761e-01
7.07842350e-01 1.56235695e-02 1.18923295e+00 -5.71513593e-01
-5.94491899e-01 -2.82783628e-01 -7.83749148e-02 -1.32180798e+00
4.59902376e-01 -4.60781306e-01 6.31057680e-01 -1.69714570e+00
-2.13591367e-01 -2.89382279e-01 1.29032016e-01 4.17948455e-01
-1.91853613e-01 -5.49769044e-01 5.34240782e-01 1.38308123e-01
-6.82993233e-01 1.17758679e+00 1.68036711e+00 -1.35894820e-01
-4.41209972e-01 -1.11192733e-01 -1.62162781e-02 6.68187618e-01
8.54296803e-01 -3.83285075e-01 -6.92331493e-01 -6.44761920e-01
-3.72358501e-01 4.76282120e-01 4.47459877e-01 -1.09010732e+00
2.34030396e-01 -6.96117163e-01 7.35752434e-02 -2.76924044e-01
5.57685912e-01 -6.96273446e-01 -1.52856201e-01 7.75560260e-01
-3.21423471e-01 1.28463507e-01 2.22551450e-01 1.14085793e+00
2.88526118e-01 -8.73977989e-02 6.18546903e-01 -4.09586400e-01
-1.15253460e+00 4.68824893e-01 -6.64348006e-01 2.40449801e-01
1.29250395e+00 -1.92169845e-02 -4.61953700e-01 -5.06638825e-01
-7.02675343e-01 6.99106157e-01 7.10551023e-01 6.02423131e-01
7.03060210e-01 -1.14635730e+00 -4.86409396e-01 -2.20452133e-03
7.31190443e-02 1.37264803e-01 2.35254303e-01 6.60785675e-01
-4.19597119e-01 2.73357302e-01 -4.70346004e-01 -5.04519224e-01
-9.15059865e-01 8.08075428e-01 1.95203140e-01 -2.76307255e-01
-8.81246865e-01 3.68119031e-01 2.70110041e-01 -6.24206722e-01
4.76956636e-01 -3.18613142e-01 -1.49745420e-01 -5.48065662e-01
3.66109669e-01 5.33312380e-01 -5.68703711e-01 -2.28853688e-01
-1.85061410e-01 3.80291164e-01 7.43216798e-02 -5.86117029e-01
1.20497823e+00 -2.05935821e-01 4.92471427e-01 3.76111358e-01
8.00564468e-01 -3.07114840e-01 -2.21462035e+00 -2.70323902e-01
1.33899748e-02 -7.43552685e-01 -1.53917864e-01 -6.80525720e-01
-4.62761164e-01 7.20890641e-01 4.53130752e-01 -3.69148344e-01
6.13485515e-01 -2.46604726e-01 6.24736249e-01 7.33329594e-01
6.84015572e-01 -1.22917020e+00 5.35538852e-01 8.75782669e-01
1.06136525e+00 -1.45041263e+00 -2.72734582e-01 -2.17422098e-02
-9.72605050e-01 8.42193186e-01 9.47524369e-01 -3.51987571e-01
2.73578376e-01 2.82996744e-01 -9.40462400e-04 1.50372535e-01
-9.29281175e-01 -6.11287057e-01 -6.95864782e-02 1.03176177e+00
-3.74783099e-01 -8.17857161e-02 -9.19775572e-03 -7.56758973e-02
2.11219400e-01 5.64951822e-02 5.92213452e-01 1.51564324e+00
-5.62994421e-01 -9.60952103e-01 -2.27890044e-01 -6.17178157e-02
7.09120408e-02 2.76047766e-01 -9.96398926e-02 1.06937122e+00
-1.80899769e-01 8.69382501e-01 -1.54746443e-01 -1.72342464e-01
5.15288293e-01 -2.29336664e-01 5.80154479e-01 -6.39190435e-01
-2.13839933e-01 1.80405349e-01 6.77805245e-02 -1.08641875e+00
-1.12624824e-01 -8.21506858e-01 -1.22641265e+00 -7.59250596e-02
1.32853508e-01 -2.07907602e-01 5.65512121e-01 8.80716503e-01
4.66600776e-01 6.41616583e-01 3.99804801e-01 -1.45890260e+00
-1.02529967e+00 -7.05314875e-01 -2.18019381e-01 3.21945399e-01
8.74136388e-01 -1.12661099e+00 -1.49394721e-01 -2.39237830e-01] | [4.546651840209961, 0.9940419793128967] |
87b7f3aa-ef71-4d5a-831d-4d2acd8d9c1d | gender-prediction-using-limited-twitter-data | 2010.02005 | null | https://arxiv.org/abs/2010.02005v1 | https://arxiv.org/pdf/2010.02005v1.pdf | Gender prediction using limited Twitter Data | Transformer models have shown impressive performance on a variety of NLP tasks. Off-the-shelf, pre-trained models can be fine-tuned for specific NLP classification tasks, reducing the need for large amounts of additional training data. However, little research has addressed how much data is required to accurately fine-tune such pre-trained transformer models, and how much data is needed for accurate prediction. This paper explores the usability of BERT (a Transformer model for word embedding) for gender prediction on social media. Forensic applications include detecting gender obfuscation, e.g. males posing as females in chat rooms. A Dutch BERT model is fine-tuned on different samples of a Dutch Twitter dataset labeled for gender, varying in the number of tweets used per person. The results show that finetuning BERT contributes to good gender classification performance (80% F1) when finetuned on only 200 tweets per person. But when using just 20 tweets per person, the performance of our classifier deteriorates non-steeply (to 70% F1). These results show that even with relatively small amounts of data, BERT can be fine-tuned to accurately help predict the gender of Twitter users, and, consequently, that it is possible to determine gender on the basis of just a low volume of tweets. This opens up an operational perspective on the swift detection of gender. | ['Stephan Raaijmakers', 'Maaike H. T. de Boer', 'Maaike Burghoorn'] | 2020-09-29 | null | null | null | null | ['gender-prediction'] | ['computer-vision'] | [-2.23764941e-01 3.16041172e-01 -1.12404794e-01 -6.55944645e-01
-7.18953788e-01 -8.43173265e-01 7.07579315e-01 6.98002815e-01
-8.27555120e-01 6.65796757e-01 8.04357231e-02 -5.00892758e-01
1.43236652e-01 -1.00948465e+00 -2.28449330e-01 -4.84015286e-01
1.46962926e-01 9.86718416e-01 1.45663125e-02 -2.54147798e-01
1.59645319e-01 4.25733894e-01 -1.49243402e+00 1.76343486e-01
4.74172115e-01 6.09588623e-01 -2.55240679e-01 7.47041225e-01
-2.25104198e-01 1.07188568e-01 -9.92887318e-01 -1.10755599e+00
3.46153602e-02 -6.05542995e-02 -6.50540829e-01 -3.97913754e-01
5.57866096e-01 -3.29982430e-01 6.37787506e-02 7.14708626e-01
6.80090785e-01 -1.40547425e-01 6.99813545e-01 -1.25837100e+00
-8.22452009e-02 8.67515862e-01 -4.71435755e-01 3.32558095e-01
3.22295338e-01 8.59714001e-02 1.08277023e+00 -6.69685245e-01
4.69022810e-01 1.46312273e+00 8.79982650e-01 5.75976908e-01
-1.47196603e+00 -1.09314203e+00 -3.22065562e-01 -1.77202988e-02
-1.47951961e+00 -3.90148163e-01 2.20678404e-01 -5.37026346e-01
7.97226071e-01 4.88753945e-01 7.03969002e-01 1.21076393e+00
2.22654659e-02 2.73983926e-01 1.14301276e+00 -2.18807399e-01
-1.45610079e-01 6.40304983e-01 2.20306337e-01 4.04810727e-01
4.79233772e-01 -3.16433460e-02 -5.75149655e-01 -3.33372593e-01
1.25232473e-01 -4.34758633e-01 3.47991139e-01 2.53137648e-01
-9.62551236e-01 1.24454129e+00 -1.72165297e-02 4.99457389e-01
2.08866835e-01 -1.90566763e-01 8.24610472e-01 2.70581990e-01
7.98478663e-01 7.95843840e-01 -5.15858531e-01 -6.25406146e-01
-1.26711881e+00 5.13275445e-01 1.05329287e+00 4.26475406e-01
8.86348426e-01 -1.34533614e-01 -2.12135628e-01 9.08376276e-01
-1.51759550e-01 7.05926597e-01 4.15855289e-01 -5.68720579e-01
6.71006203e-01 7.73368537e-01 -5.66490972e-03 -1.27862024e+00
-6.26694798e-01 -3.08831066e-01 -4.23767537e-01 -2.34390229e-01
9.57607806e-01 -5.06681621e-01 -4.94548857e-01 1.55375361e+00
4.64868695e-01 -6.24682717e-02 -3.62943828e-01 5.04790366e-01
8.24401557e-01 5.87534249e-01 1.72415614e-01 6.70831534e-04
1.54870689e+00 -1.32373348e-01 -2.99884677e-01 -4.48840469e-01
8.75086546e-01 -7.88122833e-01 1.21493638e+00 2.73014158e-01
-8.03474367e-01 -2.96372652e-01 -7.56241918e-01 1.74204949e-02
-7.08579123e-01 -1.77082978e-02 7.03318775e-01 1.41905761e+00
-6.32795572e-01 7.71017134e-01 -5.99082649e-01 -4.61256504e-01
3.24500024e-01 8.74107480e-01 -4.81979519e-01 -5.41922189e-02
-1.17847812e+00 8.73950720e-01 2.25145668e-01 1.97501779e-02
-2.04506740e-01 -9.38078880e-01 -8.35534334e-01 3.28539833e-02
-2.37377789e-02 -2.73607373e-01 1.11659336e+00 -5.69529951e-01
-9.66020882e-01 1.15470719e+00 -2.05755964e-01 -4.16529834e-01
7.01992989e-01 1.60383388e-01 -4.50428963e-01 8.26798677e-02
3.47265720e-01 6.50158107e-01 7.79091001e-01 -8.34545970e-01
-7.76616693e-01 -5.26693523e-01 -1.96872652e-02 -2.15633690e-01
-8.46283197e-01 4.04499203e-01 4.01884690e-02 -3.17178607e-01
-4.63581741e-01 -9.77154195e-01 1.05035476e-01 -7.13025987e-01
-3.24096531e-01 -4.12642658e-01 7.35368192e-01 -7.70554245e-01
1.38188493e+00 -2.04844666e+00 -4.30127025e-01 4.00676489e-01
3.98164690e-01 3.24181706e-01 2.01137699e-02 5.30622780e-01
9.00583938e-02 4.83089656e-01 2.23672062e-01 -4.32212949e-01
3.27549964e-01 2.22844526e-01 -1.86173111e-01 4.39565510e-01
5.29151373e-02 7.18962431e-01 -8.39691341e-01 -5.54070771e-01
2.45854139e-01 4.03470844e-01 -6.62904680e-01 3.78107056e-02
2.99188733e-01 2.58519590e-01 -5.29159866e-02 4.68274802e-01
6.59587562e-01 1.57314643e-01 2.38972262e-01 -3.94526124e-02
-7.23140612e-02 2.86606938e-01 -1.08454084e+00 5.00999987e-01
-6.09421730e-01 9.34641182e-01 6.85279891e-02 -8.62816751e-01
1.12807143e+00 -6.01563044e-02 3.71604055e-01 -5.66167295e-01
3.63093883e-01 4.04982060e-01 2.85234272e-01 -4.09240156e-01
6.71156704e-01 -3.32813174e-01 -6.24932468e-01 6.03410006e-01
6.32003099e-02 -1.97726682e-01 8.48238826e-01 1.29807843e-02
9.89395916e-01 -6.93938434e-01 -8.63603130e-02 -8.02398995e-02
4.19351637e-01 -3.60746635e-03 3.90318453e-01 6.50941908e-01
-1.08070284e-01 5.08377612e-01 8.38069379e-01 -3.95953417e-01
-1.13765299e+00 -7.27033675e-01 -4.51914370e-01 1.56257331e+00
-4.03121948e-01 -7.27834404e-01 -9.35972571e-01 -6.39316142e-01
2.99787760e-01 7.17715442e-01 -7.87244439e-01 -1.56799063e-01
-5.34502149e-01 -1.20280516e+00 1.07578170e+00 2.22506359e-01
3.10770512e-01 -8.30185831e-01 -1.97389722e-01 6.54350519e-02
-2.56934553e-01 -1.24616313e+00 -2.34916627e-01 2.44912761e-03
-4.60058421e-01 -1.21108437e+00 -4.55596298e-01 -5.07445693e-01
5.68270743e-01 -2.75456727e-01 1.14969671e+00 1.13678932e-01
-1.78452536e-01 2.44387388e-01 -2.30850711e-01 -6.21036828e-01
-5.62939942e-01 6.09033525e-01 8.26335028e-02 6.53353035e-02
9.04991925e-01 -4.76107180e-01 -3.13686699e-01 4.18714672e-01
-5.06667852e-01 -4.48833555e-01 2.61483699e-01 7.54386246e-01
-1.55607656e-01 7.45402202e-02 5.66171408e-01 -1.31641603e+00
9.44944322e-01 -4.22866881e-01 -3.26665401e-01 -1.61331773e-01
-6.68403327e-01 -1.38842121e-01 5.85984886e-01 -7.29496300e-01
-6.69795156e-01 -3.90201569e-01 -4.10020173e-01 6.84193447e-02
-7.57116079e-02 3.69827181e-01 2.63895363e-01 -6.38375506e-02
7.69673169e-01 -1.50867894e-01 1.82501003e-02 -4.69573498e-01
-3.70010808e-02 9.65536952e-01 1.07316680e-01 -4.87067848e-01
8.82197320e-01 1.30146414e-01 -1.01307377e-01 -8.96296501e-01
-7.02908516e-01 -4.67705488e-01 -6.77092612e-01 -3.69120799e-02
5.60908377e-01 -6.94917202e-01 -1.05082464e+00 2.60818779e-01
-5.71628451e-01 -4.25910652e-01 -6.43744767e-02 3.51301342e-01
-1.28739715e-01 1.13415159e-01 -6.63408101e-01 -7.88703501e-01
-3.04128766e-01 -7.90223658e-01 1.09497404e+00 -7.10864589e-02
-1.06370234e+00 -1.20139337e+00 -4.62463200e-02 6.91370249e-01
1.24868229e-01 1.60232574e-01 9.72099185e-01 -1.12450671e+00
2.04672828e-01 -6.55490577e-01 -2.63784558e-01 1.03741929e-01
-1.00978740e-01 2.61992961e-01 -1.08632123e+00 -2.53004521e-01
-4.00380760e-01 -3.52726132e-01 6.37603879e-01 -1.82026960e-02
1.05488217e+00 -5.25586903e-01 -2.96578616e-01 2.68910885e-01
9.55655336e-01 -3.66783917e-01 3.79172981e-01 3.48446995e-01
6.83866918e-01 7.61262774e-01 6.30025923e-01 5.61370194e-01
6.42417133e-01 5.84154069e-01 1.05954468e-01 1.92703664e-01
1.21749058e-01 -3.57522070e-01 2.34195188e-01 3.69840950e-01
-4.41856496e-02 1.05091080e-01 -1.16529799e+00 5.42557299e-01
-1.35658967e+00 -9.44981158e-01 -2.78929025e-01 2.26129007e+00
1.02673531e+00 3.50512147e-01 6.46632254e-01 5.56324363e-01
7.63951242e-01 1.57211125e-01 -9.57015604e-02 -9.88712192e-01
2.19282061e-01 4.91958737e-01 7.33123839e-01 4.85536307e-01
-9.75164354e-01 7.82775164e-01 6.52470112e+00 1.10845840e+00
-1.45785272e+00 4.54796813e-02 7.12960899e-01 -2.59035587e-01
-1.14221193e-01 -2.31390566e-01 -1.23954558e+00 8.34976792e-01
1.35118783e+00 -2.57814437e-01 3.67727339e-01 6.41313612e-01
3.65804791e-01 -2.26970479e-01 -1.01985621e+00 1.24206388e+00
-2.21441127e-02 -1.18693829e+00 -2.10922509e-01 4.38226491e-01
3.16373974e-01 -2.00024545e-01 6.06273524e-02 7.74878144e-01
2.53077541e-02 -1.19163990e+00 5.22673428e-01 -1.47253722e-01
9.19762552e-01 -1.14342415e+00 1.06727016e+00 5.97069681e-01
-7.56926894e-01 -4.05809969e-01 -2.98682064e-01 -2.22719267e-01
-1.34747803e-01 8.13386858e-01 -1.46752799e+00 -1.58978358e-01
6.85187519e-01 2.76345760e-01 -1.00173032e+00 5.80466390e-01
-4.86062951e-02 9.40680802e-01 -5.88853359e-01 -4.17644292e-01
8.54433849e-02 5.70336133e-02 2.17236653e-01 1.53005922e+00
3.56012553e-01 -1.40629917e-01 -1.77136242e-01 5.00113487e-01
-9.63118076e-02 3.14081401e-01 -3.87274981e-01 -3.48702997e-01
6.01318777e-01 1.54566419e+00 -7.32650697e-01 -2.37013847e-01
-3.78795229e-02 3.21900576e-01 3.65092158e-01 2.63858736e-02
-5.97115517e-01 -4.26546097e-01 5.81184566e-01 7.77690411e-01
1.60603300e-01 -1.36048645e-01 -5.21574020e-01 -8.19356501e-01
-2.62807369e-01 -8.03222835e-01 4.60284591e-01 -3.51552665e-01
-1.18710244e+00 1.39840096e-01 5.89230545e-02 -7.79866278e-01
-6.34474277e-01 -7.35994399e-01 -5.12368798e-01 7.38947630e-01
-1.00986540e+00 -1.26560926e+00 -6.95869178e-02 2.88971335e-01
4.79987934e-02 -1.60683498e-01 1.01787508e+00 5.44659615e-01
-5.92375159e-01 1.14425063e+00 -1.33151412e-01 4.96455818e-01
9.41802084e-01 -1.34298611e+00 2.39962086e-01 2.58114308e-01
9.01458785e-02 7.84764111e-01 9.78280723e-01 -5.87461174e-01
-1.07522857e+00 -1.01747346e+00 1.60261822e+00 -6.42934024e-01
8.15506279e-01 -8.26385617e-01 -5.67853093e-01 5.03092527e-01
-2.54395574e-01 -1.93426505e-01 1.04514945e+00 7.18971789e-01
-4.26491201e-01 -2.55197853e-01 -1.50201142e+00 5.02191246e-01
6.37990713e-01 -5.49503148e-01 -4.54380244e-01 3.08069646e-01
1.02509156e-01 -3.86725068e-01 -1.09471643e+00 1.29199088e-01
7.33031154e-01 -1.07193863e+00 8.93721938e-01 -2.79974788e-01
4.18196321e-01 3.28856945e-01 1.76497936e-01 -1.30083990e+00
-2.05963366e-02 -3.98313731e-01 6.33265600e-02 1.64524472e+00
4.71806228e-01 -8.40294421e-01 1.02761817e+00 9.59198654e-01
4.21818584e-01 -7.64042914e-01 -8.13239515e-01 -6.79855168e-01
4.12760586e-01 -4.78855461e-01 7.61291265e-01 9.19438303e-01
1.50666043e-01 5.44972241e-01 -6.70246184e-02 -2.17430592e-01
3.99167448e-01 -1.15792982e-01 1.14705765e+00 -1.49788940e+00
-5.55551238e-02 -5.19465566e-01 -6.96249187e-01 -2.11161852e-01
4.03341502e-01 -7.98052251e-01 -4.61711943e-01 -9.75246966e-01
1.97460413e-01 -7.11900294e-01 3.72545213e-01 5.23247957e-01
-1.55695513e-01 8.16047490e-01 2.55690634e-01 -6.72756210e-02
-9.64892060e-02 -1.14529822e-02 7.46261001e-01 -1.80525839e-01
-1.18447505e-01 2.79211193e-01 -9.37612891e-01 6.54437602e-01
8.55296791e-01 -6.90081477e-01 -1.84910834e-01 2.82376427e-02
6.69431329e-01 -4.47946072e-01 2.76523978e-01 -7.15626836e-01
-1.14680320e-01 -1.50799155e-02 6.14957809e-01 -4.79467779e-01
5.94821870e-01 -4.81465638e-01 -2.46879965e-01 2.93893546e-01
-6.13565929e-02 -1.03664398e-01 3.88434559e-01 1.31562203e-01
-1.48794010e-01 -4.09838170e-01 5.66322327e-01 -9.64124948e-02
-3.59454751e-01 1.95514008e-01 -3.56173366e-01 4.20031250e-01
7.79623926e-01 -5.59464395e-01 -1.30817980e-01 -2.27921262e-01
-7.70817578e-01 1.07380198e-02 2.87270814e-01 3.44161183e-01
4.98108827e-02 -1.00841498e+00 -6.80749834e-01 3.96002531e-01
9.32618156e-02 -4.70784754e-01 8.81197210e-03 6.42111957e-01
-4.05020118e-01 3.34489167e-01 -6.26400039e-02 -5.92768669e-01
-1.66037929e+00 1.66084617e-01 5.19395098e-02 -3.86408955e-01
-1.70578405e-01 8.81212950e-01 -1.29242972e-01 -8.98616135e-01
-1.37788281e-01 -1.00980885e-01 -4.83792871e-01 9.17808115e-01
5.99803388e-01 5.35828054e-01 3.23392540e-01 -6.89624429e-01
-3.62893492e-01 3.68427157e-01 3.59815988e-03 -3.64095271e-01
1.44365144e+00 2.10450813e-02 -1.91321045e-01 5.30423343e-01
1.32423842e+00 5.42797625e-01 -7.13365972e-01 4.23981905e-01
-7.49822631e-02 -7.28823423e-01 -2.35381916e-01 -5.18740058e-01
-8.25921595e-01 1.03178096e+00 2.56084591e-01 4.37820435e-01
5.18241882e-01 -2.31940344e-01 8.30824554e-01 2.71687359e-01
3.07259351e-01 -1.32425869e+00 -8.18916112e-02 5.86693347e-01
4.65478629e-01 -1.21168149e+00 6.27719015e-02 -3.98555189e-01
-4.81417626e-01 1.16269350e+00 4.17877674e-01 4.44890372e-02
5.85534215e-01 2.23241121e-01 -1.64278236e-03 -1.83701947e-01
-3.65919799e-01 4.22223732e-02 6.91870451e-02 7.50252187e-01
4.50255394e-01 3.54504764e-01 -4.65345651e-01 7.39285171e-01
-1.19784868e+00 -3.62298042e-01 6.07886374e-01 3.36675942e-01
-2.64722109e-01 -1.31484199e+00 -5.81648409e-01 1.02484679e+00
-7.80564904e-01 -9.06069502e-02 -5.82887411e-01 9.08377826e-01
4.57707107e-01 1.07541096e+00 2.99852788e-01 -4.75500971e-01
1.65622339e-01 2.15504423e-01 5.39841712e-01 -8.62268507e-01
-1.00592172e+00 -4.83418822e-01 7.18439698e-01 -1.39717489e-01
4.69899215e-02 -8.74948561e-01 -8.47186208e-01 -1.06686354e+00
-1.50511697e-01 2.65078783e-01 5.62694728e-01 9.66149688e-01
1.13335572e-01 -3.18053029e-02 5.29904068e-01 -8.40479851e-01
-5.45490980e-01 -1.12926722e+00 -7.61387587e-01 3.44893277e-01
1.36181772e-01 -6.85142994e-01 -6.43511832e-01 -2.78966129e-01] | [9.36807632446289, 10.3450927734375] |
198a2003-b714-41ca-be38-44cccdc12120 | domain-adaptive-person-re-identification-via-1 | 2011.03363 | null | https://arxiv.org/abs/2011.03363v1 | https://arxiv.org/pdf/2011.03363v1.pdf | Domain Adaptive Person Re-Identification via Coupling Optimization | Domain adaptive person Re-Identification (ReID) is challenging owing to the domain gap and shortage of annotations on target scenarios. To handle those two challenges, this paper proposes a coupling optimization method including the Domain-Invariant Mapping (DIM) method and the Global-Local distance Optimization (GLO), respectively. Different from previous methods that transfer knowledge in two stages, the DIM achieves a more efficient one-stage knowledge transfer by mapping images in labeled and unlabeled datasets to a shared feature space. GLO is designed to train the ReID model with unsupervised setting on the target domain. Instead of relying on existing optimization strategies designed for supervised training, GLO involves more images in distance optimization, and achieves better robustness to noisy label prediction. GLO also integrates distance optimizations in both the global dataset and local training batch, thus exhibits better training efficiency. Extensive experiments on three large-scale datasets, i.e., Market-1501, DukeMTMC-reID, and MSMT17, show that our coupling optimization outperforms state-of-the-art methods by a large margin. Our method also works well in unsupervised training, and even outperforms several recent domain adaptive methods. | ['Shiliang Zhang', 'Xiaobin Liu'] | 2020-11-06 | null | null | null | null | ['unsupervised-person-re-identification'] | ['computer-vision'] | [-9.55431387e-02 -3.02615047e-01 -2.31180876e-01 -6.51348770e-01
-9.03755307e-01 -4.43172127e-01 4.70565915e-01 -1.93741545e-01
-8.26849341e-01 8.17203462e-01 2.21520230e-01 1.63769722e-01
-1.27349645e-01 -3.85029852e-01 -3.83072525e-01 -5.85637808e-01
4.33610529e-01 7.90297925e-01 3.06571624e-03 -4.10764217e-02
-6.71316385e-02 5.52128106e-02 -1.18816745e+00 -1.58906490e-01
1.09780979e+00 7.19519019e-01 9.07062218e-02 2.02214301e-01
1.91874608e-01 3.97443026e-01 -3.97510111e-01 -6.65893376e-01
5.72874606e-01 -3.19715768e-01 -8.86066496e-01 2.22343907e-01
5.98531544e-01 -4.55285639e-01 -4.84362215e-01 1.13875294e+00
9.37137067e-01 4.07594800e-01 6.93861485e-01 -1.29929209e+00
-1.10856688e+00 3.43405515e-01 -6.54966533e-01 -3.52346487e-02
3.31546992e-01 1.21745719e-02 6.22988105e-01 -1.00998795e+00
5.53670228e-01 1.21183634e+00 8.36328864e-01 9.68711495e-01
-1.26693523e+00 -1.03240120e+00 3.91783208e-01 3.40353906e-01
-1.77515578e+00 -3.71810734e-01 6.47075832e-01 -5.22858202e-01
6.47475123e-01 -1.16261996e-01 -5.47710457e-04 1.12221110e+00
-7.63936758e-01 8.40195537e-01 1.29827118e+00 -4.53778774e-01
9.58279669e-02 4.85840410e-01 3.13031137e-01 4.67580169e-01
7.60075599e-02 2.20907018e-01 -5.50344169e-01 -1.41824767e-01
6.37929797e-01 5.48386481e-04 -1.87115029e-01 -3.72074634e-01
-1.32251394e+00 5.73544800e-01 2.52566308e-01 -6.53005093e-02
4.80677262e-02 -4.22966182e-01 4.34216887e-01 3.57301891e-01
5.40537238e-01 1.15475178e-01 -6.39694273e-01 1.15917683e-01
-6.92284763e-01 3.06321144e-01 7.09093571e-01 1.37459159e+00
9.12320554e-01 -3.90220553e-01 -2.81816959e-01 1.26861250e+00
2.91082591e-01 5.21425426e-01 7.10146248e-01 -8.48243237e-01
8.00790250e-01 7.33688116e-01 2.82122463e-01 -6.59395397e-01
-4.80906487e-01 -4.76492643e-01 -9.10222411e-01 -1.90544873e-01
6.17500603e-01 -3.39279801e-01 -7.71517992e-01 1.91844618e+00
5.17692804e-01 5.16469121e-01 1.06461115e-01 1.13770509e+00
8.84157121e-01 1.92688510e-01 3.71358633e-01 5.64405099e-02
1.33840132e+00 -1.15105963e+00 -5.15927255e-01 -4.47961777e-01
8.50415111e-01 -4.27252620e-01 9.74983931e-01 2.68927775e-02
-7.89840877e-01 -7.70855367e-01 -9.37364697e-01 -1.44000247e-01
-5.35052419e-01 5.32735586e-01 3.01802665e-01 8.04213107e-01
-9.68601108e-01 2.99735814e-01 -1.67249247e-01 -7.13577092e-01
3.90954167e-01 4.97310430e-01 -6.82813704e-01 -3.19762856e-01
-1.17141974e+00 7.98930705e-01 5.37798226e-01 -2.25608096e-01
-5.63796878e-01 -7.31778145e-01 -9.56267476e-01 -1.52507737e-01
3.35790604e-01 -7.71573126e-01 1.21046090e+00 -9.05383646e-01
-1.69028521e+00 1.39406228e+00 -2.69111693e-01 -2.84318447e-01
6.72676265e-01 -2.42286190e-01 -6.32674634e-01 -1.53402135e-01
4.11710411e-01 9.13169980e-01 6.39861882e-01 -1.23677814e+00
-7.24542022e-01 -5.46999037e-01 -1.99578792e-01 4.74929929e-01
-7.25534201e-01 1.76034153e-01 -8.11491907e-01 -8.23434412e-01
-1.57921910e-01 -1.02276230e+00 -1.89657733e-01 -1.94116384e-01
-4.26966310e-01 -3.96617442e-01 6.50942206e-01 -7.67626882e-01
1.15143561e+00 -2.14122391e+00 9.72421169e-02 2.41035223e-01
2.44561940e-01 3.33687514e-01 -2.06739843e-01 1.15328692e-01
-2.39513740e-01 -1.25579074e-01 -2.96386719e-01 -8.16357911e-01
2.04777777e-01 2.09539339e-01 9.55790207e-02 6.29482508e-01
-4.98471521e-02 9.66534913e-01 -9.18286502e-01 -7.62931168e-01
5.37752435e-02 6.02400489e-02 -4.14989620e-01 3.14167649e-01
2.38617346e-01 8.89793217e-01 -3.53544176e-01 7.47876346e-01
1.01017010e+00 -4.22904491e-01 2.13675037e-01 -6.74783885e-02
7.35511929e-02 -1.07635021e-01 -1.41249335e+00 1.86126983e+00
-7.85592571e-02 1.83492288e-01 -5.78522719e-02 -1.33136129e+00
9.13772345e-01 2.19692379e-01 4.62517381e-01 -9.58304405e-01
9.72068608e-02 1.09470531e-01 -4.85812783e-01 -2.79501438e-01
3.88148010e-01 1.47784948e-01 -2.57024139e-01 3.98249298e-01
2.29889333e-01 8.12164068e-01 1.12630591e-01 1.25252500e-01
7.04812348e-01 1.72401264e-01 1.57430381e-01 -2.62897670e-01
8.39581370e-01 -8.13176036e-02 1.09233963e+00 7.43899524e-01
-6.26231730e-01 7.24519551e-01 -1.42775208e-01 -2.38896951e-01
-9.06082094e-01 -1.01826286e+00 -1.51979372e-01 1.46604168e+00
7.40599990e-01 -2.27510214e-01 -9.25503314e-01 -1.20606709e+00
2.18862727e-01 3.96013767e-01 -5.43255210e-01 -1.18443154e-01
-6.99936092e-01 -8.45727026e-01 5.69044292e-01 7.22861588e-01
1.13665688e+00 -5.48082352e-01 4.22735035e-01 6.15402572e-02
-3.78478497e-01 -1.45289958e+00 -1.09500635e+00 -2.35247597e-01
-5.66067696e-01 -1.01728880e+00 -1.14575040e+00 -1.17041361e+00
1.02293754e+00 4.32153136e-01 9.38559115e-01 -2.27123439e-01
-9.20408964e-02 6.08282328e-01 -4.43394184e-01 -2.07214683e-01
-1.80661138e-02 1.98683158e-01 4.85276312e-01 2.62459010e-01
9.38724637e-01 -2.91466266e-01 -6.81981623e-01 1.06422079e+00
-3.99927437e-01 -1.26188293e-01 5.59962809e-01 1.07360482e+00
6.64735854e-01 -1.25917226e-01 7.79252946e-01 -9.42022681e-01
5.16014159e-01 -5.71069896e-01 -4.87414330e-01 5.09614885e-01
-8.48799169e-01 -7.39541873e-02 4.97907966e-01 -7.28508890e-01
-1.37002075e+00 3.31320167e-01 1.51602656e-01 -3.96315992e-01
-3.62544298e-01 4.94817756e-02 -5.50521612e-01 -2.49409989e-01
8.16002429e-01 2.43418008e-01 -1.07787229e-01 -8.46762478e-01
3.07330251e-01 9.27634776e-01 8.04928362e-01 -6.68583572e-01
1.24359453e+00 4.72147346e-01 -4.94988203e-01 -3.38810772e-01
-8.36166501e-01 -9.11919534e-01 -1.01766765e+00 3.27361040e-02
8.77923965e-01 -1.30884075e+00 -6.03927433e-01 7.17465103e-01
-8.37135673e-01 -3.51941288e-01 -1.73178211e-01 5.80622137e-01
-2.89248645e-01 6.45946741e-01 -3.51997346e-01 -4.44507033e-01
-2.36593828e-01 -9.48155880e-01 1.01703703e+00 4.74345386e-01
-8.14222172e-02 -1.14290404e+00 3.85802761e-02 7.01286674e-01
1.76400289e-01 -1.78217903e-01 4.47080135e-01 -1.00335705e+00
-3.16821277e-01 -2.20094636e-01 -6.01971209e-01 4.44678545e-01
1.44764960e-01 -8.71832252e-01 -1.02052271e+00 -6.28241718e-01
-4.06958014e-01 -3.88996959e-01 7.48792529e-01 1.44125834e-01
1.22359693e+00 -1.35757118e-01 -6.37513518e-01 8.81677151e-01
1.22007167e+00 -4.73920209e-03 4.36897546e-01 6.60815358e-01
8.56975198e-01 5.87961197e-01 9.15956557e-01 4.32964385e-01
9.95407283e-01 1.02336693e+00 -1.81354582e-01 -2.69293904e-01
-2.00390339e-01 -3.41241688e-01 3.51399094e-01 5.80898821e-01
-3.59086655e-02 2.80817915e-02 -8.37216198e-01 5.40658116e-01
-2.16012621e+00 -8.34487855e-01 2.38341764e-01 2.29914284e+00
8.85762453e-01 -3.22406650e-01 5.19725144e-01 -2.94078380e-01
1.06048119e+00 -2.95872420e-01 -8.74893665e-01 1.97171926e-01
-3.11418802e-01 -2.43247226e-01 9.13076580e-01 1.83973730e-01
-1.49703169e+00 1.02715123e+00 6.39873743e+00 9.00596917e-01
-5.27937889e-01 4.42133516e-01 4.73217547e-01 8.95747840e-02
2.16469079e-01 -1.53159395e-01 -1.31543100e+00 7.41013050e-01
6.07255876e-01 -1.64947927e-01 4.89089251e-01 1.02049482e+00
-1.66649923e-01 2.38515586e-01 -1.22468865e+00 1.49818635e+00
1.56924352e-01 -9.20002460e-01 -2.22076684e-01 1.17929399e-01
9.45967376e-01 -1.40014768e-01 1.37728751e-01 6.26556754e-01
5.73312223e-01 -9.07214701e-01 4.12091136e-01 4.41042393e-01
1.01220286e+00 -6.47232771e-01 8.50929022e-01 4.98206526e-01
-1.25351334e+00 -3.74089479e-01 -4.91806567e-01 1.90973520e-01
-1.55038228e-02 3.46154392e-01 -6.20944262e-01 6.30756021e-01
1.02698171e+00 9.23799872e-01 -6.25206470e-01 9.31637824e-01
-3.44171971e-02 4.42138672e-01 -1.19413733e-01 4.56878841e-01
-1.92621142e-01 -2.48084828e-01 3.83387506e-01 1.38581276e+00
1.22607857e-01 1.28499433e-01 4.98703897e-01 6.84213221e-01
-3.30327451e-01 1.45823047e-01 -3.35837364e-01 3.91174704e-01
7.15853870e-01 1.12044942e+00 -4.70142663e-02 -4.16617006e-01
-7.34275401e-01 1.39341784e+00 5.15089273e-01 6.22361660e-01
-8.30949664e-01 -3.34038883e-01 8.90018940e-01 -6.74091429e-02
8.77246335e-02 -1.44580543e-01 -2.22246438e-01 -1.39243507e+00
1.06983572e-01 -8.32981765e-01 8.03626120e-01 -2.59538054e-01
-1.87650955e+00 1.96956024e-01 1.06270924e-01 -1.56429696e+00
-9.78659391e-02 -5.68727851e-01 -3.56971145e-01 1.10260820e+00
-1.88475716e+00 -1.37671220e+00 -5.60113788e-01 1.03572166e+00
3.95463020e-01 -6.13208652e-01 7.39893258e-01 7.51419127e-01
-1.01904380e+00 1.37825596e+00 4.17207927e-01 5.40027261e-01
1.50207627e+00 -1.31640756e+00 3.97499979e-01 7.34830678e-01
-3.46381932e-01 6.09717011e-01 1.27062336e-01 -6.79017544e-01
-1.00671017e+00 -1.39221644e+00 8.40721011e-01 -5.94024539e-01
2.47810960e-01 -3.48328829e-01 -8.72392774e-01 8.56835723e-01
-1.03864573e-01 4.93692681e-02 1.02815402e+00 1.54733628e-01
-6.15735471e-01 -2.70914048e-01 -1.41069818e+00 4.39869285e-01
1.62844276e+00 -5.20838857e-01 -5.04638851e-01 5.25405407e-01
5.31388283e-01 -5.06586313e-01 -1.04274094e+00 2.70682901e-01
3.60785931e-01 -5.91944695e-01 1.29118478e+00 -5.25591016e-01
-3.04540277e-01 -3.62316281e-01 -9.13568437e-02 -1.38562238e+00
-6.86030746e-01 -5.21266103e-01 6.43096864e-02 1.73239219e+00
2.24456832e-01 -1.05602026e+00 8.63162875e-01 9.96173918e-01
1.32720515e-01 -1.74758777e-01 -8.97186279e-01 -1.18429554e+00
9.92912725e-02 -4.13075946e-02 8.50264788e-01 1.30257595e+00
-2.19618037e-01 1.97654068e-01 -5.74470401e-01 4.19875085e-01
1.01319885e+00 -1.36535570e-01 1.05274892e+00 -1.36404824e+00
-2.13894561e-01 -2.85145015e-01 -3.20508480e-01 -1.46806836e+00
4.43182349e-01 -1.00641060e+00 -1.43302351e-01 -1.12482643e+00
4.89305556e-01 -7.27228045e-01 -4.91579711e-01 6.31104231e-01
-4.13096517e-01 2.57897526e-01 3.15381736e-02 5.60429037e-01
-9.29006755e-01 5.35969377e-01 1.08599401e+00 -3.11979800e-01
-2.73915201e-01 -1.07250690e-01 -9.66227591e-01 5.97806811e-01
7.21509278e-01 -4.33452189e-01 -2.81097680e-01 -5.52473068e-01
-2.73567915e-01 -6.16113067e-01 4.20090258e-01 -9.47880447e-01
5.76293111e-01 -1.11925259e-01 6.91491842e-01 -4.05726850e-01
1.83536887e-01 -7.87994802e-01 -1.33292049e-01 -7.95589462e-02
-3.14679086e-01 -1.62760675e-01 -4.39452231e-02 6.58257544e-01
-1.33406430e-01 -1.16561532e-01 9.42571580e-01 -3.78508866e-02
-1.15763474e+00 6.67773604e-01 2.65468389e-01 2.10342407e-01
1.03539419e+00 -5.09182096e-01 -4.66246575e-01 -1.82440966e-01
-7.33404100e-01 8.61934900e-01 5.62212586e-01 5.16327262e-01
3.65987331e-01 -1.63292480e+00 -9.15485382e-01 1.50572702e-01
4.87905592e-01 6.00968674e-02 4.28064972e-01 8.49117517e-01
3.07084620e-02 3.00832778e-01 -1.11235511e-02 -6.16790771e-01
-1.28122711e+00 5.28721213e-01 4.83401775e-01 -3.22279125e-01
-4.13305968e-01 8.87972236e-01 4.66497481e-01 -1.13126791e+00
4.22346681e-01 4.06825751e-01 -3.17081213e-01 -1.64789155e-01
7.44914472e-01 5.93371093e-01 -1.80375010e-01 -9.07874346e-01
-5.07172704e-01 9.37395215e-01 -4.01319236e-01 1.24162830e-01
9.48653579e-01 -4.96280819e-01 1.71827093e-01 -1.29343584e-01
1.30549192e+00 -2.39282742e-01 -1.39075947e+00 -9.98119473e-01
1.04893319e-01 -6.01849854e-01 -2.61152476e-01 -8.92116904e-01
-8.26844573e-01 4.22205031e-01 9.65916216e-01 -2.54212916e-01
1.18698204e+00 5.99983595e-02 1.00381267e+00 4.11899358e-01
3.71921688e-01 -1.73475206e+00 1.63559243e-01 4.40579951e-01
5.05104959e-01 -1.78297114e+00 -4.05673385e-02 -4.47836399e-01
-9.29567873e-01 6.58874452e-01 1.00039434e+00 2.15082496e-01
5.93557656e-01 -2.05566555e-01 1.81254223e-01 2.86192536e-01
-7.81266689e-02 -4.90931004e-01 4.36682671e-01 1.10425019e+00
3.59205268e-02 1.18410736e-01 -3.96384895e-02 9.09853160e-01
7.85145443e-03 -6.75186887e-02 -2.39045054e-01 7.15920508e-01
-1.95480078e-01 -1.48304927e+00 -4.65089321e-01 1.75581664e-01
-1.28890380e-01 1.64938033e-01 -4.87934440e-01 8.11872602e-01
4.16378886e-01 1.00145817e+00 -1.62385523e-01 -5.74745595e-01
5.01886547e-01 3.85504551e-02 2.84734279e-01 -5.61129332e-01
-4.27163780e-01 -3.14426780e-01 3.69002260e-02 -3.98386896e-01
-4.82455373e-01 -8.07882190e-01 -1.13380039e+00 -3.33301365e-01
-3.81203085e-01 1.11999204e-02 3.94234866e-01 1.09575784e+00
5.81287503e-01 -7.79628754e-04 8.00643861e-01 -6.38569832e-01
-6.95247829e-01 -8.86159897e-01 -6.82527244e-01 8.40083539e-01
8.64226073e-02 -8.62329900e-01 -1.03916951e-01 1.74328595e-01] | [14.782818794250488, 1.0685265064239502] |
590012d0-c406-46b9-97c4-d3caeb29491a | boosting-cross-lingual-transferability-in | 2305.15233 | null | https://arxiv.org/abs/2305.15233v1 | https://arxiv.org/pdf/2305.15233v1.pdf | Boosting Cross-lingual Transferability in Multilingual Models via In-Context Learning | Existing cross-lingual transfer (CLT) prompting methods are only concerned with monolingual demonstration examples in the source language. In this paper, we propose In-CLT, a novel cross-lingual transfer prompting method that leverages both source and target languages to construct the demonstration examples. We conduct comprehensive evaluations on multilingual benchmarks, focusing on question answering tasks. Experiment results show that In-CLT prompt not only improves multilingual models' cross-lingual transferability, but also demonstrates remarkable unseen language generalization ability. In-CLT prompting, in particular, improves model performance by 10 to 20\% points on average when compared to prior cross-lingual transfer approaches. We also observe the surprising performance gain on the other multilingual benchmarks, especially in reasoning tasks. Furthermore, we investigate the relationship between lexical similarity and pre-training corpora in terms of the cross-lingual transfer gap. | ['Jinsik Lee', 'Yireun Kim', 'Dayeon Ki', 'Sunkyoung Kim'] | 2023-05-24 | null | null | null | null | ['cross-lingual-transfer'] | ['natural-language-processing'] | [-2.44423807e-01 -2.08337575e-01 -3.40644926e-01 -5.13223350e-01
-1.57003069e+00 -9.87141728e-01 7.79224098e-01 7.92861134e-02
-6.93450332e-01 8.51825953e-01 1.23837776e-01 -7.84367502e-01
2.03865007e-01 -4.31477159e-01 -1.02217555e+00 -1.96877748e-01
2.10271716e-01 5.82381904e-01 4.67706099e-02 -5.94645143e-01
-3.61962020e-01 -6.14171177e-02 -8.64773512e-01 5.94193459e-01
1.44067216e+00 6.27058864e-01 2.15637729e-01 3.44442278e-01
-3.01501930e-01 9.70773757e-01 -6.72069490e-01 -7.29154229e-01
1.21081611e-02 -4.73179430e-01 -8.14077199e-01 -5.99930406e-01
1.08948112e+00 -4.15102124e-01 -1.46159247e-01 8.05290878e-01
4.86777663e-01 6.00959398e-02 7.38732517e-01 -1.33027434e+00
-1.10780632e+00 1.05962181e+00 -4.18527603e-01 2.88101077e-01
6.90672636e-01 4.48318630e-01 1.22192025e+00 -1.19787943e+00
6.06758714e-01 1.57591426e+00 7.29975104e-01 6.10706210e-01
-1.31327295e+00 -9.53070462e-01 3.58547658e-01 3.61301005e-01
-1.14535069e+00 -2.12952048e-01 6.89738333e-01 -3.68579268e-01
1.03498828e+00 -5.01878560e-02 1.73424240e-02 1.45334101e+00
-3.13926511e-03 1.26438725e+00 1.45610082e+00 -5.37397504e-01
-3.58798981e-01 4.07018304e-01 2.23550558e-01 6.74565732e-01
-2.16682673e-01 2.53766924e-01 -7.70109296e-01 1.80622548e-01
4.08148736e-01 -6.39385939e-01 -4.32921261e-01 -2.31828853e-01
-1.59715772e+00 7.67244518e-01 4.35876161e-01 3.62985820e-01
-7.95745105e-02 4.70703058e-02 7.79398263e-01 1.00393867e+00
4.05974746e-01 4.23903853e-01 -8.50646675e-01 -1.40612908e-02
-2.36078829e-01 1.27332419e-01 5.62820613e-01 1.38623047e+00
6.27569973e-01 1.03197582e-01 -5.00740469e-01 1.07464278e+00
-3.55868787e-02 8.62306476e-01 6.94650352e-01 -6.28779054e-01
1.09443426e+00 4.51300412e-01 -1.61919832e-01 -1.22756369e-01
-6.38829125e-03 -4.02808934e-01 -4.43417251e-01 -2.46169418e-01
6.36028767e-01 -3.89055461e-01 -2.99934238e-01 2.23069787e+00
1.98948473e-01 -1.84876040e-01 5.85335493e-01 5.61398149e-01
9.85196710e-01 5.50788999e-01 4.75582689e-01 7.72754550e-02
1.34439778e+00 -1.44459951e+00 -6.49791300e-01 -5.34363240e-02
1.17208683e+00 -1.07102263e+00 2.06860185e+00 3.00618075e-02
-8.11800778e-01 -9.25174892e-01 -7.99677253e-01 -4.03398871e-01
-5.94689488e-01 3.99868757e-01 4.43878084e-01 2.61682987e-01
-8.60666811e-01 7.59532228e-02 -3.72409672e-01 -5.65683722e-01
-8.63080546e-02 -2.57346869e-01 -2.58408129e-01 -4.66717690e-01
-1.36870229e+00 1.14095092e+00 3.83099526e-01 -4.28184181e-01
-1.03468764e+00 -1.33925319e+00 -1.10485768e+00 -1.27213165e-01
1.79939345e-01 -5.12956500e-01 1.72410285e+00 -6.23064756e-01
-1.45431483e+00 8.46519172e-01 -1.84583534e-02 -3.57617110e-01
7.15809286e-01 -6.24814570e-01 -4.36754704e-01 -2.02115685e-01
5.47138631e-01 1.07426465e+00 2.83524066e-01 -1.01059997e+00
-7.09072530e-01 1.52610257e-01 1.97546795e-01 5.56939602e-01
-5.34615159e-01 -7.63710886e-02 -2.24270031e-01 -8.47864330e-01
-5.44793367e-01 -8.63205135e-01 2.57805228e-01 -2.58751154e-01
-1.45660892e-01 -1.19814718e+00 7.60546446e-01 -6.45071745e-01
8.69190931e-01 -1.98154402e+00 -1.01755194e-01 -4.89623815e-01
-3.37082446e-01 1.30223826e-01 -6.06859267e-01 6.54899001e-01
-8.04155767e-02 -1.81963041e-01 2.09376998e-02 -1.21346615e-01
4.78294700e-01 2.12608799e-01 -7.73239791e-01 2.12517157e-02
1.70106381e-01 1.29875553e+00 -1.09444976e+00 -6.17198229e-01
7.31850490e-02 1.36283547e-01 -6.04131579e-01 5.57032466e-01
-5.11646509e-01 7.42682457e-01 -2.90315419e-01 4.16732758e-01
1.23420186e-01 -8.34118351e-02 9.48222056e-02 -9.39499438e-02
7.33871982e-02 7.69952655e-01 -2.18606368e-01 2.35176730e+00
-9.34978962e-01 6.63794637e-01 -5.49934447e-01 -6.61329806e-01
8.92171144e-01 6.00508690e-01 2.14166790e-02 -1.17205012e+00
-1.31744340e-01 6.40895545e-01 3.06932360e-01 -6.28753006e-01
3.84618640e-01 -2.27577224e-01 -4.05076623e-01 6.06771588e-01
2.14878589e-01 -2.55017996e-01 3.15137565e-01 3.60828042e-01
6.02331340e-01 5.96568227e-01 9.91667658e-02 -6.59603894e-01
7.27495372e-01 1.42934337e-01 1.49491951e-01 5.25536418e-01
-2.52846897e-01 -1.05954476e-01 1.19382150e-01 -8.25880840e-02
-7.74538219e-01 -1.30986750e+00 -1.30250871e-01 1.54500520e+00
4.91832048e-02 -4.45789665e-01 -8.80152285e-01 -1.18654370e+00
3.43066871e-01 1.27461588e+00 -4.76655394e-01 -1.32760674e-01
-8.48733962e-01 -2.57985860e-01 9.24887657e-01 6.64976060e-01
6.55661702e-01 -1.19160604e+00 -1.84895083e-01 1.72315806e-01
-4.72912490e-01 -1.71974730e+00 -8.90202284e-01 6.36392459e-02
-7.47494698e-01 -8.19830060e-01 -8.51919889e-01 -1.26871181e+00
5.44909537e-01 1.82264447e-01 1.52083373e+00 -1.78344488e-01
2.30302602e-01 6.25122190e-01 -3.82846802e-01 -2.64220029e-01
-5.75525582e-01 4.66262817e-01 1.81798130e-01 -4.49125528e-01
6.75174117e-01 -4.64678466e-01 -1.61345944e-01 4.26733673e-01
-4.22194660e-01 -3.47970128e-02 5.91814220e-01 8.12595844e-01
4.70012963e-01 -5.63902199e-01 1.19518924e+00 -8.02974522e-01
1.02385044e+00 -3.75401914e-01 -3.96607250e-01 7.22465992e-01
-4.86715287e-01 2.72396058e-01 8.63883495e-01 -7.08067358e-01
-1.23300004e+00 -6.31065071e-01 -4.95608300e-02 -4.02784377e-01
-2.59088635e-01 6.19470596e-01 -1.18400902e-01 1.88278541e-01
8.01975012e-01 3.48707616e-01 -4.07826692e-01 -7.52732158e-01
7.07414210e-01 4.52283442e-01 6.54513657e-01 -1.49073112e+00
7.93943048e-01 -1.72224328e-01 -7.61954665e-01 -6.21959507e-01
-1.02289212e+00 -2.46461153e-01 -6.21150076e-01 -3.12596969e-02
6.77384019e-01 -1.24644363e+00 -6.87243342e-01 3.13324988e-01
-1.36266100e+00 -1.02499628e+00 -2.53469259e-01 6.89250529e-01
-5.06863236e-01 1.24625869e-01 -9.61026132e-01 -2.52040416e-01
-3.68604958e-01 -1.21305990e+00 9.83636737e-01 -1.71661034e-01
-2.66601294e-01 -1.40913534e+00 2.55915672e-01 4.73591745e-01
3.29204470e-01 -1.74996555e-01 1.48314583e+00 -9.49698448e-01
-4.70796883e-01 2.41812691e-01 -2.76849002e-01 5.60373485e-01
1.79593205e-01 -6.75823033e-01 -7.38155842e-01 -5.29865205e-01
-1.80573538e-01 -1.01891851e+00 3.44677866e-01 -2.33566120e-01
6.89810216e-01 -2.30904073e-01 -5.25575057e-02 3.58365148e-01
1.08881509e+00 1.35321826e-01 1.35127246e-01 3.20998788e-01
7.94538081e-01 7.79246628e-01 9.38070118e-01 -4.56258088e-01
9.48096812e-01 8.38757634e-01 -3.56042266e-01 -8.47133398e-02
-6.64622128e-01 -8.40320528e-01 9.63999331e-01 1.80394340e+00
4.56930578e-01 1.42816022e-01 -9.25678849e-01 8.03293765e-01
-1.61795843e+00 -4.19694573e-01 -1.25431046e-01 1.95671761e+00
1.44517553e+00 -1.41820163e-01 1.78401675e-02 -5.80024064e-01
3.92687201e-01 -2.18084395e-01 -6.86525941e-01 -2.39906833e-01
-1.63048685e-01 2.02485874e-01 1.02194585e-01 6.87500179e-01
-6.88327968e-01 1.50433445e+00 6.08624601e+00 1.15271497e+00
-1.13223219e+00 5.28841734e-01 2.59521157e-01 2.47627884e-01
-4.90060836e-01 -1.73664898e-01 -9.12755013e-01 2.41174817e-01
7.49921441e-01 -4.44811553e-01 1.38118058e-01 8.81760299e-01
-2.66786575e-01 2.41351411e-01 -1.65132439e+00 9.02491868e-01
2.54199505e-01 -6.80003047e-01 3.98335516e-01 -2.43907079e-01
9.70151782e-01 3.46325487e-01 1.47021189e-01 1.10021448e+00
5.01884043e-01 -7.82877982e-01 7.93837845e-01 -1.86804146e-01
1.04440963e+00 -7.20539987e-01 3.70671540e-01 5.63649654e-01
-1.19764972e+00 3.65948468e-01 -1.44140273e-01 8.85338485e-02
2.19822228e-01 -1.84694812e-01 -1.04300475e+00 6.95100009e-01
7.06917226e-01 7.24348366e-01 -6.38220429e-01 4.92726833e-01
-8.29608738e-01 8.46984148e-01 -7.75595084e-02 -5.46007156e-02
4.16121870e-01 -9.94194597e-02 2.28674099e-01 1.36562395e+00
3.19840908e-01 -3.00592989e-01 5.40140271e-01 7.39929199e-01
-3.52503628e-01 6.48572624e-01 -7.99317718e-01 1.76020041e-01
4.66444612e-01 8.28411520e-01 4.20649238e-02 -6.41375065e-01
-6.04684293e-01 9.56854165e-01 8.19025874e-01 6.00331604e-01
-9.00299549e-01 -3.65703493e-01 4.98867452e-01 -1.77109957e-01
6.29669651e-02 -4.37655658e-01 7.13958056e-04 -1.39977932e+00
2.24151775e-01 -1.42400897e+00 4.86224532e-01 -8.30987632e-01
-1.78252280e+00 9.01394129e-01 2.07894310e-01 -1.14693308e+00
-4.17628378e-01 -7.49315381e-01 -4.24536198e-01 1.02196014e+00
-2.02946639e+00 -1.51527858e+00 2.24302360e-03 9.92049694e-01
1.02978480e+00 -5.11072576e-01 8.34464431e-01 6.48918867e-01
-3.77774268e-01 1.29120624e+00 -6.78981915e-02 2.80540526e-01
1.36284411e+00 -1.34827995e+00 5.93073070e-01 5.98671377e-01
3.91584724e-01 8.34584653e-01 4.73944306e-01 -4.20975864e-01
-1.19773626e+00 -1.13500130e+00 1.28840339e+00 -8.64241838e-01
1.14396966e+00 -4.16109532e-01 -1.06133771e+00 1.23758912e+00
8.89533997e-01 -3.26592147e-01 7.42201626e-01 4.89412189e-01
-1.04828882e+00 -5.16273715e-02 -6.55139208e-01 9.74089622e-01
8.67154360e-01 -1.18107665e+00 -9.27294970e-01 5.81665754e-01
1.20982397e+00 -3.52016032e-01 -1.13692212e+00 6.09999537e-01
3.49687874e-01 -5.64163625e-01 9.12479818e-01 -8.99797082e-01
3.32321525e-01 -5.93942180e-02 -2.94637799e-01 -1.63913226e+00
1.95924193e-01 -4.94576275e-01 4.25840974e-01 1.35539162e+00
7.71529019e-01 -7.93491066e-01 1.98007509e-01 -1.53861880e-01
-3.88206333e-01 -7.16796041e-01 -7.80805230e-01 -1.35829377e+00
1.04017806e+00 -5.42428911e-01 4.26444590e-01 1.39061928e+00
2.86087066e-01 1.07894838e+00 -1.24855772e-01 -8.22314993e-02
6.76371336e-01 2.36409292e-01 1.04001498e+00 -6.55246139e-01
-3.91403675e-01 -3.98883075e-01 4.11545038e-01 -1.66753149e+00
6.99291527e-01 -1.53220022e+00 1.31631941e-01 -1.08874917e+00
-5.87253133e-03 -7.56453872e-01 -2.96706945e-01 5.74626446e-01
-5.09911716e-01 2.28717402e-02 1.84918553e-01 1.72423631e-01
-6.87471449e-01 8.35184813e-01 1.50821865e+00 -1.17624611e-01
1.98198691e-01 -5.05265594e-01 -4.40844566e-01 5.61962545e-01
7.45894253e-01 -4.41652775e-01 -6.51602566e-01 -1.12583125e+00
-2.21061319e-01 -9.79687124e-02 1.28005311e-01 -7.67379105e-01
1.34769663e-01 8.08844194e-02 -2.60876596e-01 -4.79375094e-01
2.34156281e-01 -5.07698298e-01 -6.91573620e-01 3.63817096e-01
-8.17534089e-01 3.58806372e-01 5.82986116e-01 1.90886676e-01
-4.19769526e-01 3.11546680e-02 4.47441578e-01 1.73536032e-01
-6.78757846e-01 5.75436577e-02 -5.14234640e-02 5.77342689e-01
6.62938893e-01 4.48960662e-01 -8.02675784e-01 -3.84320289e-01
-2.27980912e-01 6.11629307e-01 1.55660778e-01 9.47670579e-01
1.11964077e-01 -1.88474512e+00 -1.14609754e+00 -9.08742696e-02
7.15933561e-01 -2.95349926e-01 1.73368528e-01 1.04290521e+00
-1.40233994e-01 1.07725871e+00 -1.18080303e-01 -7.86288321e-01
-1.07158792e+00 6.58298016e-01 2.48348072e-01 -4.07109112e-01
-4.50810164e-01 8.83470714e-01 7.06126750e-01 -1.25219893e+00
3.55973989e-01 -5.34120440e-01 1.76184505e-01 -2.17403248e-01
2.23237619e-01 -5.88498116e-02 -1.26196176e-01 -5.17512441e-01
-1.12405993e-01 6.76253259e-01 -4.49499726e-01 -2.39961222e-01
7.26018190e-01 6.57114461e-02 1.34866163e-01 7.54192829e-01
1.42960179e+00 2.35211596e-01 -1.01996660e+00 -7.60480165e-01
2.79459089e-01 4.93977033e-02 -5.72947383e-01 -1.17709255e+00
-6.56493247e-01 1.10949111e+00 1.71411589e-01 -1.84495971e-01
6.28617525e-01 2.63655245e-01 1.09261680e+00 8.27847421e-01
6.28353477e-01 -7.49711215e-01 3.97702575e-01 7.54767895e-01
1.26825261e+00 -1.40917516e+00 -5.51109433e-01 -4.07451212e-01
-7.18819499e-01 6.14647508e-01 1.21980476e+00 -3.57960202e-02
2.19817534e-01 -1.63644366e-02 6.23382688e-01 5.35135530e-02
-1.05417776e+00 -6.47814199e-02 3.30924362e-01 4.82780874e-01
9.86262858e-01 1.94005460e-01 -2.91306645e-01 5.96078694e-01
-6.17285371e-01 -3.14181983e-01 -1.88600540e-01 7.30744720e-01
-1.50858417e-01 -1.42391217e+00 -9.28310752e-02 -2.48652175e-01
-1.37738332e-01 -5.68103254e-01 -3.43393207e-01 1.21330523e+00
-3.95619050e-02 7.81374693e-01 -1.82885796e-01 -1.52603611e-01
6.67061627e-01 4.05901045e-01 8.36269140e-01 -6.01204097e-01
-8.01693916e-01 -8.58171731e-02 2.01009989e-01 -2.49148220e-01
-1.12594604e-01 -6.11724317e-01 -1.06226134e+00 -2.26062164e-03
-1.14838071e-01 5.20978212e-01 4.93965298e-01 1.13541734e+00
1.56359941e-01 6.89323604e-01 1.65189147e-01 -1.36360496e-01
-9.19801772e-01 -1.39714134e+00 5.32993861e-02 4.93105471e-01
1.69093788e-01 -5.52952528e-01 -3.13106865e-01 1.31813779e-01] | [11.082542419433594, 9.638372421264648] |
09547fb9-a3f5-4766-9b1e-2ef2134e83a3 | dgst-discriminator-guided-scene-text-detector | 2002.12509 | null | https://arxiv.org/abs/2002.12509v1 | https://arxiv.org/pdf/2002.12509v1.pdf | DGST : Discriminator Guided Scene Text detector | Scene text detection task has attracted considerable attention in computer vision because of its wide application. In recent years, many researchers have introduced methods of semantic segmentation into the task of scene text detection, and achieved promising results. This paper proposes a detector framework based on the conditional generative adversarial networks to improve the segmentation effect of scene text detection, called DGST (Discriminator Guided Scene Text detector). Instead of binary text score maps generated by some existing semantic segmentation based methods, we generate a multi-scale soft text score map with more information to represent the text position more reasonably, and solve the problem of text pixel adhesion in the process of text extraction. Experiments on standard datasets demonstrate that the proposed DGST brings noticeable gain and outperforms state-of-the-art methods. Specifically, it achieves an F-measure of 87% on ICDAR 2015 dataset. | ['Cunzhao Shi', 'Baihua Xiao', 'Yanna Wang', 'Fuxi Jia', 'Jinyuan Zhao', 'Chunheng Wang'] | 2020-02-28 | null | null | null | null | ['scene-text-detection'] | ['computer-vision'] | [ 7.04133689e-01 -2.29944184e-01 2.31073827e-01 -3.15085351e-01
-6.74252629e-01 -2.44887546e-01 5.82522810e-01 -3.44029069e-02
-4.12142813e-01 2.55322009e-01 9.45360735e-02 -7.24721998e-02
5.49830079e-01 -9.73836482e-01 -4.72129524e-01 -8.62668216e-01
8.65403295e-01 4.49253440e-01 8.62297654e-01 9.37489048e-02
5.12765586e-01 2.36097015e-02 -1.01389050e+00 2.68233508e-01
1.25507689e+00 8.77062798e-01 4.24302846e-01 4.74974513e-01
-7.16548979e-01 8.43177259e-01 -9.57319140e-01 -5.26179254e-01
2.30804399e-01 -6.77554131e-01 -4.40432608e-01 4.45745438e-01
3.75936031e-01 -4.27029282e-01 -5.35579681e-01 1.46525967e+00
5.63369989e-01 -7.66599551e-02 8.14945281e-01 -1.09852612e+00
-6.87764049e-01 9.10074294e-01 -8.36485386e-01 9.96584445e-02
1.34458795e-01 -3.32921706e-02 7.24310696e-01 -7.78446794e-01
3.83191049e-01 1.32648611e+00 5.36603451e-01 3.16615224e-01
-9.20622170e-01 -7.13660240e-01 2.48968303e-01 -5.41924164e-02
-1.36281872e+00 2.34571868e-03 9.33982134e-01 -5.57953238e-01
3.58416259e-01 2.55489081e-01 2.83674747e-01 1.09781766e+00
1.35521650e-01 1.44046950e+00 9.41809475e-01 -4.07800794e-01
1.24742620e-01 5.27731255e-02 -4.13830066e-03 6.99707091e-01
4.66607660e-01 -6.21781528e-01 -2.67822146e-01 1.59463748e-01
7.94039369e-01 2.88758963e-01 -1.54650107e-01 -5.64272664e-02
-1.04263771e+00 9.89027441e-01 3.47441614e-01 3.81913483e-01
-4.67621572e-02 -1.00143723e-01 4.73081142e-01 -2.81695068e-01
7.61331618e-01 -5.46013471e-03 -7.82346651e-02 2.10871980e-01
-1.22481167e+00 1.01266868e-01 5.20683706e-01 8.16741705e-01
3.62071961e-01 4.71488327e-01 -5.61617017e-01 9.98649180e-01
4.20052588e-01 9.93876278e-01 7.55854785e-01 -2.15297788e-01
6.86949611e-01 9.20282066e-01 -3.63971829e-01 -1.11041689e+00
-1.10973679e-01 -2.18769073e-01 -9.71460402e-01 -1.57149192e-02
1.80851579e-01 -2.53352582e-01 -1.59942818e+00 1.00654531e+00
3.03509057e-01 9.81334448e-02 2.25868188e-02 8.65457416e-01
1.10695732e+00 9.44661677e-01 1.54319152e-01 1.43796101e-01
1.11903918e+00 -1.08494353e+00 -8.55875790e-01 -6.55503273e-01
2.79838175e-01 -1.02511644e+00 8.53137970e-01 4.60961640e-01
-6.32731080e-01 -6.50042951e-01 -1.07696354e+00 -3.26262191e-02
-4.30276126e-01 3.48851353e-01 3.35300356e-01 8.18206251e-01
-7.40506530e-01 5.07279336e-02 -8.35687160e-01 -4.81492102e-01
5.96659660e-01 9.50232893e-02 2.74505615e-01 5.20375110e-02
-9.48332489e-01 3.99180174e-01 8.08823049e-01 -4.35881503e-02
-7.48287976e-01 5.13757989e-02 -7.74486125e-01 -5.58792837e-02
5.93985915e-01 -4.78504300e-01 1.09781528e+00 -1.05815196e+00
-1.36828482e+00 9.54286993e-01 4.14096490e-02 -4.77028698e-01
8.65904450e-01 -3.00239086e-01 -3.24366838e-01 3.38327199e-01
5.36262572e-01 6.16000354e-01 1.17398441e+00 -1.03050888e+00
-6.81836843e-01 -3.78505737e-01 -5.36617994e-01 3.29821736e-01
-4.73846525e-01 1.58558354e-01 -9.82029915e-01 -1.16101336e+00
3.86150569e-01 -6.65250480e-01 -2.28515491e-01 -1.08269855e-01
-9.51297045e-01 -1.56127408e-01 1.28555381e+00 -7.55068421e-01
1.00320590e+00 -2.03249454e+00 -1.15028888e-01 -1.02072805e-01
5.95179833e-02 4.78632450e-01 2.24497780e-01 1.57932058e-01
3.34365517e-01 3.21615428e-01 -5.75327218e-01 -3.75556767e-01
-2.92009134e-02 -1.88233647e-02 -4.63615239e-01 2.67992675e-01
-4.07203333e-03 8.94513786e-01 -4.81441885e-01 -1.08197868e+00
6.32253885e-01 3.66790563e-01 -2.45584428e-01 1.81593299e-01
-4.91238356e-01 1.27319708e-01 -1.10763121e+00 6.64603114e-01
9.53992903e-01 -2.53681540e-01 -2.50612378e-01 9.31976140e-02
1.73704103e-01 -2.08689705e-01 -1.19381893e+00 1.36866724e+00
2.38525659e-01 8.16475809e-01 -1.36913702e-01 -9.60224569e-01
1.04876721e+00 -1.58507273e-01 3.25520188e-01 -6.80125952e-01
5.32410741e-01 -1.76831916e-01 -4.20434117e-01 -4.33348000e-01
7.80615151e-01 4.21666056e-02 -2.20587745e-01 -2.27743611e-02
-3.31972778e-01 -5.04534006e-01 1.03036083e-01 4.15971279e-01
8.64037871e-01 1.49182409e-01 -8.40916764e-03 6.50326442e-03
5.82236230e-01 2.13566095e-01 4.77541596e-01 7.54697561e-01
-2.21408933e-01 1.01938736e+00 2.80388147e-01 -1.54085299e-02
-9.27884817e-01 -1.03798473e+00 -2.50671268e-01 1.02879190e+00
6.07045770e-01 -1.99823961e-01 -1.31619728e+00 -8.25312257e-01
-6.38269931e-02 5.69553792e-01 -5.82546055e-01 1.88056603e-02
-5.46901166e-01 -9.88209188e-01 8.63852859e-01 6.63498223e-01
1.35124242e+00 -1.23154879e+00 -1.82767376e-01 1.55304447e-01
-3.70262355e-01 -1.55734038e+00 -6.29690230e-01 1.10263349e-02
-7.84959853e-01 -7.57869959e-01 -1.09741890e+00 -1.17724442e+00
7.23992169e-01 4.07724768e-01 5.76292872e-01 -1.58778548e-01
-5.48047602e-01 6.01926334e-02 -5.85366607e-01 -4.90300477e-01
-5.41705668e-01 -8.77138600e-02 -5.38649797e-01 2.71258563e-01
3.51728886e-01 2.80758739e-01 -6.02426112e-01 3.49681467e-01
-1.23503649e+00 3.14992249e-01 7.05956697e-01 6.53557241e-01
6.27334356e-01 4.93388683e-01 3.13846111e-01 -1.24059987e+00
4.84895080e-01 -3.84560451e-02 -6.48935199e-01 5.79117052e-02
-6.76002502e-01 -1.22391813e-01 7.61869490e-01 -2.33729690e-01
-1.37815011e+00 1.96661457e-01 -2.51866549e-01 -2.89268136e-01
-3.46145809e-01 2.44803444e-01 -2.88185507e-01 4.23039682e-02
4.74581063e-01 8.11028242e-01 -3.37806493e-01 -3.02386940e-01
1.97952092e-01 1.04843819e+00 6.50092125e-01 -1.85403869e-01
8.90008450e-01 6.31960869e-01 -3.58599275e-01 -8.63597810e-01
-8.61650348e-01 -6.43038988e-01 -3.86853456e-01 1.11644298e-01
1.56746936e+00 -9.97526586e-01 2.28719376e-02 1.12264109e+00
-1.00040841e+00 -3.25592816e-01 1.41548216e-01 7.04815313e-02
-3.24467331e-01 7.46007800e-01 -7.15182364e-01 -6.79266095e-01
-6.69390678e-01 -1.10451543e+00 1.57729018e+00 4.85372514e-01
4.17715460e-01 -9.76450622e-01 -3.08697075e-01 6.41480923e-01
4.15121019e-02 4.07862961e-01 5.65593421e-01 -8.15372050e-01
-7.36686468e-01 -4.53173339e-01 -4.72220391e-01 4.25648659e-01
-1.19554296e-01 5.80670265e-03 -1.00452614e+00 -3.52414548e-02
1.24557670e-02 1.58656299e-01 1.06077981e+00 6.03909492e-01
1.32355225e+00 -1.04785256e-01 -5.73961496e-01 6.21503770e-01
1.57077253e+00 4.13350254e-01 8.19809675e-01 4.43994224e-01
1.14597344e+00 1.23516083e-01 8.01392674e-01 3.75658244e-01
9.52017605e-02 3.83133650e-01 3.20066512e-01 -3.91717196e-01
-2.94312596e-01 -3.83776426e-01 1.63499638e-01 4.76930052e-01
5.54504752e-01 -8.17895591e-01 -9.81321275e-01 3.42448890e-01
-1.92278111e+00 -7.43423939e-01 -4.60163862e-01 1.63398814e+00
5.41536450e-01 6.48682117e-01 -3.12404558e-02 1.85657203e-01
1.31812406e+00 3.90775561e-01 -6.59788191e-01 -9.56715643e-02
-3.90099645e-01 3.07511389e-02 6.77644610e-01 6.00910522e-02
-1.48141181e+00 1.54612839e+00 5.80540657e+00 1.30191278e+00
-1.06657803e+00 1.38489129e-02 7.67708004e-01 6.49105668e-01
1.90731272e-01 -2.31242731e-01 -9.56055939e-01 8.53598237e-01
1.57711849e-01 -8.03652555e-02 -2.72048730e-02 1.05827415e+00
2.32096806e-01 -2.51830310e-01 -4.88222390e-01 1.13598049e+00
4.85392332e-01 -9.06354129e-01 3.77832234e-01 -1.66811749e-01
1.08334351e+00 5.59887895e-03 1.59274697e-01 1.89412102e-01
2.56846815e-01 -8.92256618e-01 6.78640664e-01 2.25743711e-01
7.13748872e-01 -5.76354563e-01 7.70431340e-01 4.51069951e-01
-1.29713249e+00 1.36662170e-01 -5.41102946e-01 3.43775392e-01
-7.58439628e-03 6.17661655e-01 -1.06769550e+00 4.09578621e-01
6.41797721e-01 7.32166111e-01 -8.72455180e-01 1.14220095e+00
-4.57591206e-01 9.66946781e-01 -1.71474963e-01 -3.18305135e-01
3.76748592e-01 -1.95997775e-01 4.67713118e-01 1.36489797e+00
7.75029063e-02 -1.18573055e-01 5.46603024e-01 9.60287213e-01
-2.47326717e-01 4.36270684e-01 -4.06718373e-01 -1.09582081e-01
2.35090926e-01 1.18213880e+00 -1.45631719e+00 -5.99538386e-01
-2.70011246e-01 1.44271314e+00 -4.46810961e-01 3.24612439e-01
-9.66886997e-01 -7.93993771e-01 -2.19823703e-01 -3.41006368e-02
4.55360383e-01 3.62037234e-02 -8.04630280e-01 -1.14029229e+00
1.09901831e-01 -8.19050014e-01 3.60438228e-01 -1.04713130e+00
-9.05083776e-01 4.07108933e-01 -3.67226005e-01 -1.15032887e+00
7.80622512e-02 -5.84435821e-01 -8.43241394e-01 7.85359383e-01
-1.05530560e+00 -1.28501308e+00 -6.12774670e-01 6.14295006e-01
1.26022422e+00 -1.84571296e-01 3.61030370e-01 5.47509221e-03
-7.43571103e-01 6.28781438e-01 4.35346484e-01 8.38803649e-01
6.90433145e-01 -1.26035619e+00 6.48768961e-01 1.38368976e+00
6.22149967e-02 6.50619343e-02 7.31114686e-01 -9.91756082e-01
-8.97053003e-01 -1.44413185e+00 2.43634209e-01 -3.71242553e-01
3.55031580e-01 -3.98851305e-01 -8.08042109e-01 5.17954886e-01
2.00292408e-01 -3.69783044e-01 1.19522503e-02 -6.49198949e-01
-3.52112576e-02 1.29823551e-01 -1.31846595e+00 7.07513213e-01
7.40094125e-01 -2.49702230e-01 -7.39910841e-01 4.36590403e-01
8.61238539e-01 -5.89224458e-01 -1.74753323e-01 4.70001847e-01
2.81486455e-02 -8.07819426e-01 9.38296676e-01 2.04099014e-01
4.39283758e-01 -3.61991405e-01 8.16463307e-02 -7.93048799e-01
-5.25786206e-02 -2.83868551e-01 4.06555891e-01 1.37940967e+00
3.13326158e-03 -6.30859137e-01 1.00712848e+00 1.41342312e-01
-3.12292993e-01 -3.15741688e-01 -5.94913244e-01 -4.65990901e-01
1.22104853e-01 -2.47848839e-01 2.51064688e-01 8.21444750e-01
-4.80394393e-01 4.84925061e-01 -2.30806127e-01 8.06392580e-02
6.07616901e-01 1.28391892e-01 9.08739865e-01 -1.16015720e+00
-9.86424387e-02 -6.72485471e-01 -6.59968078e-01 -1.39936137e+00
4.98016644e-03 -8.41675758e-01 4.04387981e-01 -1.93054712e+00
5.44085801e-01 1.24499379e-02 4.20230068e-02 4.05047297e-01
-5.90293288e-01 4.60642070e-01 1.46707952e-01 8.94642919e-02
-8.70793164e-01 4.74239349e-01 1.44137728e+00 -5.59650779e-01
-4.44946662e-02 -1.11099727e-01 -6.18615925e-01 1.06028223e+00
7.69361258e-01 -4.22396898e-01 -1.90682516e-01 -3.17837656e-01
-2.74032205e-01 -9.68620852e-02 3.36984456e-01 -1.15623665e+00
2.79728413e-01 -5.24893366e-02 7.53314376e-01 -1.26941860e+00
4.39769700e-02 -6.00548744e-01 -4.55571502e-01 3.78335983e-01
-2.37476453e-01 -5.28162956e-01 1.28015384e-01 7.19800711e-01
-1.78721622e-01 -3.21006149e-01 9.65882957e-01 -7.68316677e-03
-9.61826563e-01 2.96882510e-01 -3.72528791e-01 3.61996263e-01
9.96228576e-01 -4.23367381e-01 -4.99967933e-01 -1.79833934e-01
-3.66209030e-01 2.04185262e-01 4.49280769e-01 4.95662570e-01
7.83021092e-01 -9.03966844e-01 -6.59070849e-01 1.55628011e-01
9.95553583e-02 4.11014587e-01 5.96127063e-02 4.15399849e-01
-8.59755754e-01 2.77177364e-01 2.55688101e-01 -9.24769819e-01
-1.27725697e+00 4.34851706e-01 1.97759986e-01 -2.08591729e-01
-9.65101540e-01 7.01962709e-01 7.11413562e-01 -7.34384209e-02
1.49018332e-01 -2.46014267e-01 -1.32869184e-01 -1.98270932e-01
2.11254030e-01 2.87976921e-01 -3.20110507e-02 -6.31328583e-01
-9.68946218e-02 9.07142460e-01 -2.19235092e-01 -5.22568375e-02
8.67752433e-01 -1.81119338e-01 1.93154246e-01 1.88365713e-01
9.52383339e-01 7.29916170e-02 -1.09876537e+00 -3.18385452e-01
-2.43093267e-01 -4.04806793e-01 2.37191737e-01 -9.74113107e-01
-1.11707199e+00 8.75205815e-01 7.48647988e-01 5.48536897e-01
1.15080118e+00 -5.28904460e-02 1.09124434e+00 2.54991740e-01
4.51809205e-02 -1.35264087e+00 4.56237346e-01 4.84997958e-01
4.42919433e-01 -1.39592791e+00 -7.55557790e-02 -6.23850882e-01
-7.77997553e-01 1.02862716e+00 6.95098996e-01 -3.64148796e-01
1.51456475e-01 3.78596276e-01 2.31307015e-01 -5.25696017e-02
2.33478278e-01 -4.56472009e-01 1.99877515e-01 5.05479574e-01
2.09906116e-01 1.17285572e-01 -2.57124871e-01 4.89076823e-01
-6.55274317e-02 -4.27457362e-01 3.74092460e-01 7.68975019e-01
-9.25923467e-01 -8.64006460e-01 -7.19545364e-01 6.79643631e-01
-8.64627421e-01 -3.76112908e-01 -7.19804108e-01 6.83753729e-01
-5.73308319e-02 1.00689840e+00 -5.73529191e-02 -3.38088453e-01
2.01845132e-02 4.18920070e-02 -2.64765695e-02 -6.69760287e-01
-2.46577397e-01 6.02016509e-01 -3.58636260e-01 -4.07183170e-02
-2.77235597e-01 -5.83674908e-01 -1.66795993e+00 -1.52885923e-02
-6.73163414e-01 -9.13972929e-02 6.92095935e-01 8.99413168e-01
-7.93623254e-02 8.13079119e-01 6.95319593e-01 -6.49432003e-01
-2.42087185e-01 -1.12406075e+00 -5.88659167e-01 6.10876560e-01
-8.31735879e-03 -2.62351990e-01 -4.69910115e-01 4.03914958e-01] | [12.036837577819824, 2.26973557472229] |
1da35451-c575-4a42-879c-737543a74ba3 | phrase-based-neural-unsupervised-machine-1 | null | null | https://aclanthology.org/D18-1549 | https://aclanthology.org/D18-1549.pdf | Phrase-Based \& Neural Unsupervised Machine Translation | Machine translation systems achieve near human-level performance on some languages, yet their effectiveness strongly relies on the availability of large amounts of parallel sentences, which hinders their applicability to the majority of language pairs. This work investigates how to learn to translate when having access to only large monolingual corpora in each language. We propose two model variants, a neural and a phrase-based model. Both versions leverage a careful initialization of the parameters, the denoising effect of language models and automatic generation of parallel data by iterative back-translation. These models are significantly better than methods from the literature, while being simpler and having fewer hyper-parameters. On the widely used WMT{'}14 English-French and WMT{'}16 German-English benchmarks, our models respectively obtain 28.1 and 25.2 BLEU points without using a single parallel sentence, outperforming the state of the art by more than 11 BLEU points. On low-resource languages like English-Urdu and English-Romanian, our methods achieve even better results than semi-supervised and supervised approaches leveraging the paucity of available bitexts. Our code for NMT and PBSMT is publicly available. | ["Marc{'}Aurelio Ranzato", 'Alexis Conneau', 'Myle Ott', 'Ludovic Denoyer', 'Guillaume Lample'] | 2018-10-01 | null | null | null | emnlp-2018-10 | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [ 7.51928613e-02 -2.32636303e-01 -5.49955726e-01 -3.65893304e-01
-1.53244126e+00 -8.94572854e-01 9.37680960e-01 -8.06212425e-03
-7.90303946e-01 1.16176534e+00 2.50024140e-01 -8.64505649e-01
2.78515756e-01 -4.74277556e-01 -8.55234683e-01 -3.19854915e-01
2.90153474e-01 9.29750085e-01 -2.06174940e-01 -6.92409277e-01
7.51092583e-02 -3.33522223e-02 -7.67262280e-01 3.46395880e-01
1.09355569e+00 4.46957529e-01 2.38303661e-01 4.59371030e-01
-2.12494820e-01 3.38876843e-01 -3.13390881e-01 -8.97506833e-01
5.06661773e-01 -5.42537332e-01 -8.28206480e-01 -2.89627820e-01
5.76857567e-01 -2.42442161e-01 -7.46184736e-02 8.76180768e-01
7.60775864e-01 -2.80018806e-01 5.23490489e-01 -6.29324853e-01
-9.28302944e-01 9.72806513e-01 -4.96787697e-01 2.80919224e-01
4.63941365e-01 1.12190962e-01 1.28558159e+00 -1.22305298e+00
7.84065008e-01 1.07614446e+00 7.75937915e-01 4.63188022e-01
-1.30831492e+00 -5.75160325e-01 -2.35941947e-01 1.43800592e-02
-1.28828228e+00 -7.68483281e-01 2.65846729e-01 -2.09569901e-01
1.53713942e+00 1.77755862e-01 3.53964597e-01 1.49470544e+00
3.29492986e-01 7.38308072e-01 1.42992437e+00 -7.88459778e-01
-1.42727613e-01 1.82777166e-01 -1.86181262e-01 4.97394294e-01
9.41809043e-02 1.06892675e-01 -6.45941019e-01 -1.25712886e-01
4.48550552e-01 -3.83070767e-01 -1.16340481e-01 2.00177982e-01
-1.61693931e+00 7.61487842e-01 3.59938145e-02 4.36423987e-01
-3.72806519e-01 -2.43042454e-01 5.35082817e-01 7.86187172e-01
7.41037965e-01 4.53005403e-01 -6.92013502e-01 -3.94114077e-01
-1.21135271e+00 1.61493048e-01 1.03790045e+00 1.25934529e+00
7.20225215e-01 -2.06414282e-01 -1.40424445e-01 1.16236460e+00
6.20911904e-02 9.79952991e-01 6.48421109e-01 -4.93846059e-01
1.22628021e+00 2.69822508e-01 2.14740448e-02 -5.63087523e-01
-1.22464113e-01 -4.93396938e-01 -7.49349713e-01 -4.87411827e-01
5.13443768e-01 -2.98894733e-01 -7.50963092e-01 1.81032133e+00
1.70890195e-03 -3.13357532e-01 2.85742790e-01 6.34415209e-01
4.73741651e-01 8.78229916e-01 -2.13811904e-01 -3.11901420e-01
1.24213135e+00 -1.34555233e+00 -5.61384559e-01 -6.23935640e-01
7.47188210e-01 -1.47166049e+00 1.36739337e+00 2.03016758e-01
-1.37040019e+00 -4.59492803e-01 -8.74901175e-01 -2.96134442e-01
-2.38525212e-01 2.59786010e-01 4.44073081e-01 5.60327768e-01
-1.29664111e+00 6.42481089e-01 -7.89536417e-01 -6.44959867e-01
4.63317521e-02 3.90164763e-01 -4.79747832e-01 -3.49170357e-01
-1.30598032e+00 1.31931937e+00 2.86024302e-01 3.99080664e-02
-5.91625154e-01 -5.24642885e-01 -7.04891443e-01 -2.79509187e-01
4.60544080e-02 -7.79390335e-01 1.48059678e+00 -9.39302623e-01
-1.80010676e+00 9.78105962e-01 -2.87565738e-01 -4.17662531e-01
8.97002399e-01 -3.56368393e-01 -2.88765699e-01 -1.84363618e-01
3.04922849e-01 7.46700943e-01 4.48013246e-01 -7.79205859e-01
-6.11141920e-01 -1.16347581e-01 -9.57128555e-02 4.24529284e-01
-3.62602413e-01 3.90351921e-01 -6.42166495e-01 -6.91984475e-01
-1.45844042e-01 -1.08616304e+00 -2.37209529e-01 -6.18393660e-01
-4.56944495e-01 -1.19189598e-01 2.36681804e-01 -1.17823946e+00
1.16220224e+00 -1.66303909e+00 3.67190182e-01 -1.28213689e-01
-3.25789899e-01 2.70505250e-01 -6.08952105e-01 9.40756083e-01
2.75779456e-01 1.42953560e-01 -2.39796862e-01 -5.84236264e-01
-2.75235493e-02 3.66717339e-01 -2.65027940e-01 2.93826431e-01
1.91765100e-01 1.08051026e+00 -9.71935570e-01 -3.65437448e-01
-1.44129187e-01 3.21678698e-01 -4.81865704e-01 9.61451828e-02
-1.46921605e-01 4.84603733e-01 -8.86554047e-02 6.24454141e-01
3.50634664e-01 -7.62717724e-02 3.45355153e-01 2.85306543e-01
-2.45430887e-01 1.00216603e+00 -5.22269607e-01 1.97607946e+00
-7.73988068e-01 5.40916741e-01 -1.54396981e-01 -5.28388500e-01
7.95779467e-01 4.95814443e-01 3.11273299e-02 -8.71885419e-01
6.46106005e-02 9.52548265e-01 1.95240527e-01 -2.32926935e-01
5.31822979e-01 -2.37296015e-01 -1.92428693e-01 7.28745043e-01
1.63340062e-01 -2.41795495e-01 7.16175735e-01 -9.78541933e-03
1.03100932e+00 2.68339068e-01 2.77016819e-01 -4.61011618e-01
4.66497868e-01 1.33398727e-01 5.60186684e-01 6.84561014e-01
1.94537804e-01 6.12199485e-01 1.66412428e-01 -3.16917807e-01
-1.46194303e+00 -1.00795496e+00 -3.85261923e-02 1.19075704e+00
-3.20210904e-01 -6.13891721e-01 -8.51445138e-01 -5.97681761e-01
-2.64983922e-01 8.44244361e-01 -1.98194265e-01 1.16154715e-01
-9.55522656e-01 -1.03805089e+00 7.64042556e-01 3.02257508e-01
2.88056910e-01 -8.83676648e-01 9.30983126e-02 3.97029489e-01
-7.59899855e-01 -1.23327148e+00 -7.17073441e-01 6.21841513e-02
-8.91710997e-01 -5.02461374e-01 -9.27059531e-01 -8.85543048e-01
4.39644247e-01 -4.50096130e-02 1.56376648e+00 -2.27247655e-01
3.10885370e-01 -2.37685010e-01 -2.55458415e-01 -2.46005192e-01
-8.65574241e-01 7.86231577e-01 3.03416759e-01 -3.67326707e-01
5.96257329e-01 -6.03818655e-01 -2.63991296e-01 2.31722102e-01
-4.78183478e-01 4.10940111e-01 9.53733683e-01 1.11110580e+00
4.33359027e-01 -7.65248656e-01 4.04575616e-01 -9.32684839e-01
7.87352860e-01 -5.10013402e-01 -3.83637428e-01 3.44895363e-01
-7.97273099e-01 9.01573896e-02 9.47788954e-01 -4.40837920e-01
-8.11296165e-01 -3.10795277e-01 -2.84054458e-01 4.89476547e-02
2.75457297e-02 5.76618552e-01 1.03433743e-01 2.22231090e-01
7.41090119e-01 4.26843762e-01 -1.83926120e-01 -6.95290327e-01
4.14215297e-01 8.82906973e-01 4.21348244e-01 -7.41186261e-01
7.93680251e-01 -4.04315181e-02 -4.90235955e-01 -5.65676630e-01
-6.76033854e-01 -3.11067462e-01 -7.81654119e-01 2.85472751e-01
4.17915225e-01 -1.07162166e+00 6.03162088e-02 3.43127340e-01
-1.37718248e+00 -4.04860437e-01 7.15431198e-02 7.01408803e-01
-6.60749137e-01 4.08250988e-01 -1.18111634e+00 -2.20478192e-01
-8.31734955e-01 -1.27873421e+00 1.09750235e+00 -3.01459491e-01
-4.73093033e-01 -1.07786608e+00 3.99751544e-01 4.64954555e-01
5.54853737e-01 -3.25448960e-01 1.02808309e+00 -7.94008136e-01
-2.90215939e-01 -1.20315947e-01 -5.63897938e-02 5.00731766e-01
2.09085360e-01 -2.42656901e-01 -5.82947552e-01 -6.12697423e-01
-2.11422235e-01 -4.93810087e-01 5.24343550e-01 -2.50889622e-02
2.22277433e-01 -4.25375104e-01 -7.50442669e-02 4.64651614e-01
1.34915853e+00 -5.40352426e-02 2.67971188e-01 4.27428246e-01
5.64948499e-01 5.70865571e-01 2.04864219e-01 -4.30629179e-02
6.57826722e-01 7.01462865e-01 -1.87355354e-01 -8.92726108e-02
-1.64637521e-01 -4.22294170e-01 8.16577733e-01 1.87721050e+00
-1.89611807e-01 -2.61921972e-01 -1.20414281e+00 6.90522909e-01
-1.73761725e+00 -6.42819405e-01 -2.26206526e-01 2.17292476e+00
1.41203284e+00 2.02336594e-01 -4.53631692e-02 -2.57021993e-01
4.53919232e-01 -4.64220494e-02 -1.68963343e-01 -6.95506155e-01
-3.22738379e-01 3.67141604e-01 6.19888842e-01 7.69946754e-01
-8.49234581e-01 1.32723546e+00 6.84034777e+00 8.28018844e-01
-1.16222513e+00 3.88467580e-01 6.13225043e-01 -2.39570960e-01
-3.65300000e-01 6.27416372e-02 -9.77964044e-01 4.63876605e-01
1.43097961e+00 -1.74110249e-01 8.33549380e-01 3.30732793e-01
3.09272617e-01 1.85602129e-01 -1.37923765e+00 8.21136892e-01
1.26925692e-01 -1.13705802e+00 1.73124090e-01 1.94176342e-02
1.03442597e+00 8.00174832e-01 2.08216868e-02 4.55695778e-01
5.75398982e-01 -1.04333949e+00 7.98726141e-01 2.35478655e-01
9.56834018e-01 -5.50884545e-01 7.13435888e-01 6.37227893e-01
-6.96020901e-01 3.25085908e-01 -4.23325092e-01 -1.80691257e-01
2.91776031e-01 5.61152399e-01 -8.00402105e-01 7.84519255e-01
4.88067985e-01 6.61943674e-01 -5.62603712e-01 5.25272846e-01
-6.12713397e-01 9.90964949e-01 -5.67833364e-01 -1.78920720e-02
5.89473426e-01 -3.68999630e-01 3.90588582e-01 1.50950122e+00
5.67051291e-01 -4.40892667e-01 1.83109596e-01 4.45201427e-01
-2.76359499e-01 7.03362226e-01 -6.00596368e-01 -1.49075255e-01
3.70197833e-01 1.05870092e+00 -2.77361691e-01 -3.49837989e-01
-7.77150571e-01 1.26800001e+00 6.73238218e-01 3.36652935e-01
-5.42974651e-01 -1.60160124e-01 4.10465062e-01 -9.88605991e-02
9.99963656e-03 -4.37149167e-01 -3.48963857e-01 -1.40374398e+00
3.67700875e-01 -1.43861818e+00 9.35068429e-02 -3.44659775e-01
-1.58796239e+00 1.06790173e+00 -1.53782845e-01 -1.11120725e+00
-8.52197826e-01 -6.68675244e-01 -3.15597266e-01 1.35238349e+00
-1.51875997e+00 -1.39569676e+00 5.02278864e-01 3.25509906e-01
6.76941872e-01 -4.73080635e-01 1.07537127e+00 5.68445027e-01
-5.79674006e-01 7.76001751e-01 5.46070099e-01 1.79294750e-01
1.15069556e+00 -1.17401505e+00 1.10535967e+00 9.04584527e-01
5.16997755e-01 8.07999909e-01 5.88476419e-01 -5.60397744e-01
-1.49842966e+00 -8.65913272e-01 1.96351540e+00 -7.56318510e-01
8.39418232e-01 -6.74400687e-01 -6.08939290e-01 8.56082380e-01
7.86293030e-01 -4.03406411e-01 6.60039186e-01 3.37626368e-01
-3.87851834e-01 3.28462683e-02 -6.91782653e-01 9.64533091e-01
1.10712004e+00 -6.08439863e-01 -7.41338968e-01 6.37302637e-01
4.65641320e-01 -4.24047649e-01 -8.70737851e-01 2.44861439e-01
4.99659717e-01 -6.91348493e-01 7.34603226e-01 -7.09607184e-01
8.48717332e-01 4.28529121e-02 -3.64899516e-01 -1.65709758e+00
-2.37567797e-01 -9.24517214e-01 2.03412592e-01 1.14702177e+00
9.60654914e-01 -7.04517663e-01 3.49235147e-01 5.84584475e-02
-3.99353681e-03 -8.41440916e-01 -1.02169812e+00 -9.77785647e-01
7.48432577e-01 -3.20319116e-01 4.80653316e-01 9.80326593e-01
1.65233493e-03 9.12150443e-01 -5.29458344e-01 -1.76571116e-01
4.26231265e-01 4.33793999e-02 7.77950883e-01 -7.14179158e-01
-5.60043573e-01 -5.18070161e-01 1.96202546e-01 -1.19489443e+00
2.62134850e-01 -1.33443344e+00 -5.42029738e-02 -1.46612132e+00
4.26560789e-01 -4.56573457e-01 -4.90840040e-02 6.40742719e-01
-3.00819337e-01 5.95136881e-01 1.34033009e-01 4.92995083e-01
-1.96298569e-01 4.73892212e-01 1.06210947e+00 -1.14349395e-01
-4.84425798e-02 -1.91039756e-01 -5.75104117e-01 4.35981750e-01
9.02286172e-01 -5.53515851e-01 -1.98149338e-01 -1.19251764e+00
3.44880372e-01 -4.25266102e-02 -1.95705190e-01 -6.35295153e-01
1.70711711e-01 2.19069924e-02 1.10816367e-01 -3.23772073e-01
1.74035519e-01 -3.45890343e-01 7.35708401e-02 3.62585753e-01
-2.50710368e-01 8.86262357e-01 2.45639011e-01 1.22155733e-01
-3.01043838e-01 -1.33925438e-01 5.75163662e-01 -2.43660167e-01
-2.49031022e-01 1.88140586e-01 -4.36258048e-01 3.17435920e-01
5.41903794e-01 1.79924443e-01 -2.30784118e-01 -3.16838831e-01
-1.77044496e-01 1.00381643e-01 5.02198339e-01 7.62943864e-01
4.74327877e-02 -1.43751085e+00 -1.35165977e+00 1.95616469e-01
7.94808418e-02 -4.04712617e-01 -2.89444506e-01 1.03108370e+00
-6.16104007e-01 7.14671195e-01 -1.43668994e-01 -5.49542248e-01
-9.22649622e-01 2.08787680e-01 1.05874762e-01 -7.08301425e-01
-3.76467317e-01 6.65870667e-01 -2.93555647e-01 -1.05576706e+00
-1.90570548e-01 -1.39099792e-01 3.44385713e-01 -1.97066143e-01
3.03192079e-01 2.22914517e-01 3.93603981e-01 -8.42779398e-01
-2.05227211e-01 3.61162573e-01 -4.01609033e-01 -5.69974542e-01
1.14343393e+00 -1.53542385e-01 -4.20261443e-01 4.86618847e-01
1.15263963e+00 3.68368775e-01 -8.13354313e-01 -5.12580037e-01
2.80380815e-01 -1.87725857e-01 -2.51519084e-01 -1.14726877e+00
-4.97121185e-01 9.47009504e-01 9.16715637e-02 -3.77885759e-01
8.80747199e-01 -1.95916563e-01 1.11707354e+00 5.69202423e-01
6.58171952e-01 -1.07153678e+00 -4.04184967e-01 1.03225398e+00
8.37399900e-01 -1.24518859e+00 -1.95402846e-01 -2.25190863e-01
-4.01345611e-01 9.04102504e-01 1.55736908e-01 1.39670685e-01
2.18662173e-01 3.69662791e-01 5.48478901e-01 4.62800205e-01
-9.13247943e-01 1.33819252e-01 4.49858785e-01 1.83605105e-01
8.90607238e-01 1.58770874e-01 -7.53825843e-01 4.61429268e-01
-6.93294287e-01 -1.99478224e-01 1.43801659e-01 8.31894457e-01
3.53055857e-02 -1.78621340e+00 -2.17726707e-01 2.88439810e-01
-8.20313394e-01 -7.74906993e-01 -4.58560556e-01 7.52570152e-01
-1.89106077e-01 1.00542235e+00 -1.22613557e-01 -2.40043551e-01
2.16686502e-01 2.76026189e-01 6.31400406e-01 -7.31776297e-01
-8.32760096e-01 9.03543681e-02 3.69219899e-01 -3.05757463e-01
-2.10051224e-01 -8.08561683e-01 -6.84116066e-01 -6.88929617e-01
-1.28513232e-01 1.33164361e-01 5.86061239e-01 1.07719195e+00
2.61137366e-01 -1.23272009e-01 5.21085322e-01 -7.52175212e-01
-1.05236554e+00 -1.33631444e+00 4.07983782e-03 2.57383764e-01
-5.38481474e-02 -4.68990244e-02 1.39770685e-02 1.00304214e-02] | [11.607590675354004, 10.252205848693848] |
64e8cb74-1c4f-4606-b84d-64ede41662a0 | code-generation-as-a-dual-task-of-code | 1910.05923 | null | https://arxiv.org/abs/1910.05923v1 | https://arxiv.org/pdf/1910.05923v1.pdf | Code Generation as a Dual Task of Code Summarization | Code summarization (CS) and code generation (CG) are two crucial tasks in the field of automatic software development. Various neural network-based approaches are proposed to solve these two tasks separately. However, there exists a specific intuitive correlation between CS and CG, which have not been exploited in previous work. In this paper, we apply the relations between two tasks to improve the performance of both tasks. In other words, exploiting the duality between the two tasks, we propose a dual training framework to train the two tasks simultaneously. In this framework, we consider the dualities on probability and attention weights, and design corresponding regularization terms to constrain the duality. We evaluate our approach on two datasets collected from GitHub, and experimental results show that our dual framework can improve the performance of CS and CG tasks over baselines. | ['Zhiyi Fu', 'Zhi Jin', 'Ge Li', 'Xin Xia', 'Bolin Wei'] | 2019-10-14 | code-generation-as-a-dual-task-of-code-1 | http://papers.nips.cc/paper/8883-code-generation-as-a-dual-task-of-code-summarization | http://papers.nips.cc/paper/8883-code-generation-as-a-dual-task-of-code-summarization.pdf | neurips-2019-12 | ['code-summarization'] | ['computer-code'] | [ 1.29102111e-01 6.16571829e-02 -1.72090575e-01 -2.79548138e-01
-5.26360035e-01 -4.01352078e-01 5.27823031e-01 -1.72910132e-02
-1.76074654e-01 3.80297154e-01 2.19824642e-01 -3.75512779e-01
1.37064874e-01 -4.94660795e-01 -7.51138628e-01 -4.49633509e-01
3.10191154e-01 -2.82968342e-01 2.02069357e-01 -1.26646623e-01
4.63380933e-01 -1.78478360e-01 -1.35089135e+00 2.26956859e-01
1.40916121e+00 7.74708271e-01 5.54707706e-01 3.73182625e-01
-2.25113556e-01 7.59279430e-01 -2.90881604e-01 -5.25025308e-01
6.67173564e-02 -4.87070322e-01 -7.82852292e-01 -1.08022667e-01
-1.62042920e-02 -2.50418913e-02 -1.39633894e-01 1.28965402e+00
3.85535955e-01 -7.03504384e-02 4.67185050e-01 -1.17488897e+00
-9.60878789e-01 8.76063704e-01 -9.08642709e-01 1.04067065e-01
2.57178068e-01 8.74646381e-02 1.19821310e+00 -8.79480362e-01
1.07503004e-01 1.07750785e+00 6.99484885e-01 5.28022349e-01
-1.02688265e+00 -7.32789457e-01 5.23821294e-01 2.38807108e-02
-1.21876502e+00 -5.32552063e-01 8.98954391e-01 -6.26668155e-01
9.36534286e-01 -3.69358025e-02 4.72467184e-01 9.89376664e-01
-6.74085468e-02 1.15876913e+00 7.05342650e-01 -5.57276189e-01
4.07192297e-03 1.68796286e-01 2.88893431e-01 7.59695470e-01
3.90960783e-01 -3.16351712e-01 -3.21819365e-01 -5.95178222e-03
6.15053773e-01 -1.52793765e-01 -4.93838102e-01 -3.00148129e-01
-1.23746467e+00 8.70432019e-01 3.41500103e-01 3.22072327e-01
-5.92581853e-02 2.75779396e-01 5.78824222e-01 -3.94239789e-03
6.29284799e-01 5.54099560e-01 -3.89264107e-01 -1.86162725e-01
-6.73253894e-01 2.27556691e-01 7.21308112e-01 1.26649475e+00
5.85159302e-01 -2.35915720e-01 -4.23224896e-01 9.00304019e-01
4.94714200e-01 1.01499073e-01 7.58366346e-01 -5.30780315e-01
8.78190994e-01 7.56009936e-01 -2.48602033e-02 -1.03227901e+00
-1.60772577e-01 -5.36786258e-01 -7.31962800e-01 -3.35506082e-01
9.10933167e-02 -3.23062152e-01 -7.05534518e-01 1.96090698e+00
-1.96509180e-03 3.66798818e-01 -2.53594923e-03 7.60234058e-01
9.06757951e-01 5.16139567e-01 -3.04372031e-02 -5.80698252e-02
1.23362994e+00 -1.47759235e+00 -8.17181706e-01 -4.57708895e-01
7.66206563e-01 -7.17514992e-01 1.35791624e+00 1.51495308e-01
-1.15579212e+00 -5.41131973e-01 -1.25673258e+00 -1.96396559e-01
-3.90613861e-02 4.96102720e-01 8.20184290e-01 4.11426187e-01
-8.03584635e-01 4.56460804e-01 -1.03828812e+00 -1.91240489e-01
4.22550708e-01 2.14081556e-02 1.44911155e-01 1.27467006e-01
-1.20072234e+00 5.17630279e-01 3.99327099e-01 3.27288032e-01
-5.13674974e-01 -5.48259735e-01 -8.17160010e-01 2.61834532e-01
4.77020621e-01 -7.05292106e-01 1.54697096e+00 -9.70543921e-01
-1.60021639e+00 8.21047246e-01 -1.02255158e-01 -2.15796828e-01
5.05560756e-01 -4.61455256e-01 -1.46356244e-02 -3.64609033e-01
3.29818994e-01 3.89551103e-01 4.05085087e-01 -1.12772834e+00
-6.57240570e-01 -6.86345622e-02 2.88289160e-01 1.25696957e-01
-7.02241480e-01 1.36914998e-01 -9.95224178e-01 -7.17204094e-01
-1.37784988e-01 -7.50612617e-01 -2.63446212e-01 -2.45668992e-01
-4.86416608e-01 -5.21761000e-01 3.41579407e-01 -7.29984581e-01
1.59182966e+00 -2.40295005e+00 3.98657590e-01 -1.46846637e-01
2.26972327e-01 3.73169005e-01 -2.27135301e-01 1.50311515e-01
5.82882166e-02 2.03960881e-01 -4.40495282e-01 -5.38587809e-01
2.23446861e-01 6.25617057e-02 -3.47339392e-01 1.09381661e-01
3.30215514e-01 1.01083875e+00 -9.64208484e-01 -4.81554687e-01
-4.74524081e-01 2.27977246e-01 -8.80863905e-01 5.56890190e-01
-4.28041250e-01 2.84534693e-01 -7.88766086e-01 3.85902971e-01
7.10328400e-01 -3.93144578e-01 2.30913788e-01 4.33129445e-02
-9.83289927e-02 5.06267130e-01 -9.31293011e-01 1.97945261e+00
-6.34024978e-01 3.69713604e-01 -2.99303383e-02 -1.27032411e+00
7.71904290e-01 2.30730176e-01 8.03540051e-02 -6.39218211e-01
6.79328218e-02 1.62408441e-01 1.28068123e-02 -7.87887275e-01
4.01719987e-01 3.27333882e-02 -7.29230344e-02 6.50091290e-01
-8.43237713e-02 1.51402459e-01 4.02757794e-01 9.23298895e-02
9.69567358e-01 3.13815057e-01 3.35247695e-01 -2.34405473e-01
7.43631780e-01 -2.73873121e-01 9.08628821e-01 6.74625456e-01
-3.17030214e-02 6.51759446e-01 1.03076482e+00 -1.61912367e-01
-6.87688231e-01 -7.02975333e-01 8.33528861e-02 1.09341633e+00
2.45373696e-01 -6.47170365e-01 -1.02440667e+00 -9.68234539e-01
-2.08121076e-01 5.36737025e-01 -6.52454972e-01 -3.24206710e-01
-7.42096663e-01 -8.88250887e-01 6.34647787e-01 6.62950575e-01
6.12496138e-01 -9.10313249e-01 -5.02254903e-01 2.09496439e-01
-3.55294555e-01 -1.31394482e+00 -7.00994313e-01 1.37313660e-02
-6.68491423e-01 -1.10447812e+00 -7.46096075e-01 -9.70617652e-01
6.27770185e-01 5.30453205e-01 1.27135825e+00 5.56437552e-01
1.57708615e-01 7.84932077e-02 -5.35426497e-01 -3.55798274e-01
-4.48740482e-01 4.80060846e-01 -3.25223714e-01 -7.25138187e-02
1.86123565e-01 -6.48776710e-01 -4.07974422e-01 1.84971411e-02
-1.01739228e+00 5.85844398e-01 8.77915800e-01 8.15772831e-01
1.65326342e-01 -1.72117263e-01 6.60489082e-01 -1.09256792e+00
9.62648869e-01 -6.33428395e-01 -7.20638573e-01 5.17017245e-01
-6.98874712e-01 3.69675577e-01 6.77007854e-01 -3.46927762e-01
-1.17087185e+00 3.10762823e-02 -1.67654544e-01 -2.08421677e-01
3.39836143e-02 1.09989548e+00 -4.98721719e-01 9.51580331e-02
1.85486853e-01 2.17323989e-01 -2.13797718e-01 -5.34718812e-01
3.23582888e-01 9.19097006e-01 4.39327747e-01 -8.10170114e-01
6.03913248e-01 7.69318715e-02 -4.34002340e-01 -2.96242505e-01
-1.12463200e+00 -1.58970118e-01 -4.47958052e-01 3.09768915e-01
8.54658723e-01 -8.86193335e-01 -4.24938977e-01 3.80305231e-01
-1.45103347e+00 -2.02593431e-01 2.28443429e-01 2.79211402e-01
-4.97627765e-01 6.91199839e-01 -3.85738790e-01 -6.14727497e-01
-3.27432185e-01 -1.53026474e+00 8.80914330e-01 2.72112578e-01
5.26813697e-03 -9.77674723e-01 2.32672453e-01 2.02214211e-01
3.54375631e-01 1.02061614e-01 9.71434236e-01 -4.95999843e-01
-4.62069184e-01 -7.17372494e-03 -5.75072527e-01 4.30703372e-01
2.46734574e-01 1.25338957e-01 -8.20785522e-01 -1.31221265e-01
1.36772186e-01 -2.04379320e-01 1.04829323e+00 1.23403020e-01
1.41504145e+00 -3.56949747e-01 -3.49739403e-01 7.71373153e-01
1.16689038e+00 9.78030264e-02 5.49308240e-01 3.49165320e-01
8.36506307e-01 5.70797563e-01 4.21077579e-01 4.23484027e-01
7.58653581e-01 8.32807720e-01 2.87117571e-01 -8.93402100e-03
6.13781475e-02 -2.07473204e-01 5.53248763e-01 1.16006124e+00
-2.10834742e-02 -1.07223071e-01 -1.05063975e+00 5.16726553e-01
-2.19531703e+00 -5.86369276e-01 -1.37649909e-01 1.98580956e+00
1.18673003e+00 1.78936005e-01 -7.17262458e-03 -2.53292501e-01
7.82580853e-01 2.08912119e-01 -3.96766216e-01 -2.35658139e-01
2.25840762e-01 1.32207153e-02 9.14090723e-02 2.09404081e-01
-1.15231705e+00 8.33623290e-01 5.91891003e+00 8.20997775e-01
-1.17746854e+00 1.61327377e-01 4.59950507e-01 2.17244085e-02
-4.97622401e-01 1.75281242e-01 -6.31041825e-01 6.89317703e-01
7.18415797e-01 -4.12797540e-01 4.38791633e-01 8.44563127e-01
6.21207943e-03 2.72433072e-01 -1.33471978e+00 8.90861988e-01
1.65836662e-01 -1.05162799e+00 -1.68697283e-01 -1.53676465e-01
7.75546253e-01 -1.65876299e-01 2.03369581e-03 6.41929984e-01
4.56449866e-01 -8.63533378e-01 7.92133212e-01 2.65299112e-01
4.28452194e-01 -5.74115098e-01 7.58295655e-01 3.71113986e-01
-1.22902906e+00 2.65460368e-02 -2.11434484e-01 -1.38792932e-01
3.60090435e-02 6.98587298e-01 -3.51272821e-01 8.39575887e-01
4.79873300e-01 9.84006107e-01 -7.77110338e-01 1.06529641e+00
-5.37582219e-01 6.36833608e-01 1.26170993e-01 -1.51654845e-02
2.16891915e-01 -2.17819944e-01 3.74226600e-01 1.33563149e+00
3.48796248e-01 -1.57481372e-01 1.80601567e-01 1.47268558e+00
-2.60134786e-01 -9.72103421e-03 -3.63961816e-01 -3.81573737e-01
2.60919124e-01 1.29850709e+00 -6.17414296e-01 -2.47411698e-01
-8.31748128e-01 8.65810692e-01 6.29202008e-01 3.66977364e-01
-1.10772192e+00 -7.07682431e-01 6.04098082e-01 -1.41695395e-01
3.60692799e-01 -2.64743239e-01 -4.68805701e-01 -1.61861193e+00
4.71974015e-01 -8.39878142e-01 1.95580035e-01 -3.39465261e-01
-1.12843990e+00 5.16567111e-01 -1.27822042e-01 -1.28680933e+00
-1.21802136e-01 -4.53860074e-01 -9.27688062e-01 8.76370549e-01
-1.85577130e+00 -1.10546243e+00 -4.46545690e-01 1.40479863e-01
4.83196199e-01 -5.81585281e-02 4.41125900e-01 4.78914559e-01
-1.08611155e+00 7.81913877e-01 1.06376879e-01 1.89529717e-01
5.90060472e-01 -1.34869802e+00 7.25742400e-01 1.17069221e+00
8.14604480e-03 9.35501993e-01 4.71481770e-01 -4.70602572e-01
-1.18030334e+00 -1.09033442e+00 7.65420020e-01 -1.20923191e-01
6.76642060e-01 -5.66118777e-01 -1.07520056e+00 5.98020494e-01
3.73191148e-01 -1.37508392e-01 6.71725690e-01 2.26877987e-01
-4.51807559e-01 2.49522239e-01 -4.20421541e-01 4.91530538e-01
9.87901270e-01 -4.13365841e-01 -6.25924468e-01 3.71696025e-01
1.09952295e+00 -4.58599597e-01 -5.16775370e-01 3.71339172e-01
3.97521615e-01 -1.01686609e+00 7.78784394e-01 -6.36796355e-01
1.11124682e+00 -2.45232642e-01 5.78890358e-05 -1.33426762e+00
-2.58254915e-01 -5.31657994e-01 -1.87091485e-01 1.62743628e+00
3.63580078e-01 -5.49003482e-01 4.32078898e-01 4.06437516e-01
-4.95863676e-01 -9.79455233e-01 -4.48588163e-01 -7.64922440e-01
3.27847242e-01 -2.48836562e-01 7.52778530e-01 1.00889611e+00
1.21424437e-01 5.67999601e-01 -2.43509412e-01 9.92156193e-02
2.26208717e-01 4.67062861e-01 8.06212664e-01 -1.09051287e+00
-7.39140689e-01 -8.96193564e-01 1.13371834e-01 -1.38430929e+00
5.20561993e-01 -1.09097707e+00 2.73797810e-01 -1.52782655e+00
5.26957154e-01 -4.20634359e-01 -3.54339272e-01 5.32299638e-01
-7.75565565e-01 -3.56150419e-01 7.44488835e-02 3.49987805e-01
-7.73384571e-01 8.94049287e-01 9.64149177e-01 -3.80988419e-02
-2.35361144e-01 8.70379619e-03 -1.29612970e+00 7.78218806e-01
7.08692253e-01 -4.04941976e-01 -3.96431983e-01 -9.77467060e-01
4.12702024e-01 -5.38828298e-02 1.39354184e-01 -7.64373839e-01
2.14754835e-01 -2.44013712e-01 -1.98827624e-01 -2.05734834e-01
-3.60409647e-01 -4.07293558e-01 -3.75410885e-01 3.42311084e-01
-4.51621056e-01 2.79740319e-02 7.72598758e-02 5.94549179e-01
-4.24817741e-01 -5.12188554e-01 6.31191194e-01 -1.36005327e-01
-4.54740554e-01 1.80457994e-01 -5.14572784e-02 1.76064640e-01
7.43278027e-01 9.99245718e-02 -1.82179317e-01 -1.09954350e-01
-2.80312955e-01 5.52984297e-01 3.21116656e-01 6.44288242e-01
5.03920138e-01 -1.36439490e+00 -6.34633839e-01 1.66972250e-01
2.08093345e-01 3.53818029e-01 2.29106750e-03 1.09117854e+00
-3.48627836e-01 4.42827165e-01 -9.86331999e-02 -2.93881744e-01
-9.22076464e-01 5.82627773e-01 2.87312269e-01 -5.89536250e-01
-3.76676500e-01 8.93772542e-01 7.06300557e-01 -6.16734445e-01
3.55914474e-01 -2.98072815e-01 -5.05436420e-01 -1.08060673e-01
4.52479273e-01 2.18057670e-02 3.88455112e-03 -1.92522049e-01
-2.23846391e-01 4.40248460e-01 -4.83360469e-01 2.04562381e-01
1.51177287e+00 -1.93270460e-01 -3.89992028e-01 3.01489025e-01
1.12046409e+00 -5.56689240e-02 -1.17315829e+00 -3.62579703e-01
3.25876802e-01 -4.81148422e-01 1.28287468e-02 -4.36863989e-01
-1.24778247e+00 1.11633909e+00 1.42080709e-01 4.21009839e-01
1.21918035e+00 -1.40154645e-01 8.81836653e-01 1.39822677e-01
-8.09212849e-02 -8.89360011e-01 2.28860602e-01 6.37503564e-01
7.93377221e-01 -1.19580281e+00 -1.35294095e-01 -6.11337781e-01
-5.71475267e-01 1.13398242e+00 9.85389054e-01 -2.97888238e-02
2.21152842e-01 1.01193413e-01 -2.99281389e-01 -7.69495070e-02
-8.45372379e-01 -1.75209060e-01 3.36010307e-01 2.16681153e-01
1.01374757e+00 -3.01622570e-01 -6.55453861e-01 1.30615616e+00
1.22525983e-01 1.18735798e-01 5.84628463e-01 1.12398112e+00
-2.20570996e-01 -1.21877384e+00 3.94109376e-02 3.28222036e-01
-6.48179054e-01 -3.11168313e-01 -3.42429161e-01 5.05692363e-01
1.51295692e-03 7.51915395e-01 -2.24833295e-01 -4.97798175e-01
3.07654440e-01 -1.89290732e-01 2.03706220e-01 -7.19429493e-01
-5.85382640e-01 1.71893001e-01 -8.30301866e-02 -3.55177641e-01
-4.14853752e-01 -4.86194104e-01 -1.10633481e+00 4.26882580e-02
-5.50021470e-01 1.64902315e-01 3.90569329e-01 1.06367397e+00
4.66459632e-01 8.26392770e-01 6.00471139e-01 -6.78654015e-01
-7.26973295e-01 -8.76577973e-01 -2.52113432e-01 3.54242474e-01
2.02294588e-01 -6.91226363e-01 -2.68664271e-01 9.58117172e-02] | [7.627533435821533, 7.937437534332275] |
d5a3ade4-93b9-4017-815a-27361a391fb7 | spanish-datasets-for-sensitive-entity | null | null | https://aclanthology.org/2022.lrec-1.400 | https://aclanthology.org/2022.lrec-1.400.pdf | Spanish Datasets for Sensitive Entity Detection in the Legal Domain | The de-identification of sensible data, also known as automatic textual anonymisation, is essential for data sharing and reuse, both for research and commercial purposes. The first step for data anonymisation is the detection of sensible entities. In this work, we present four new datasets for named entity detection in Spanish in the legal domain. These datasets have been generated in the framework of the MAPA project, three smaller datasets have been manually annotated and one large dataset has been automatically annotated, with an estimated error rate of around 14%. In order to assess the quality of the generated datasets, we have used them to fine-tune a battery of entity-detection models, using as foundation different pre-trained language models: one multilingual, two general-domain monolingual and one in-domain monolingual. We compare the results obtained, which validate the datasets as a valuable resource to fine-tune models for the task of named entity detection. We further explore the proposed methodology by applying it to a real use case scenario. | ['Maite Melero', 'Montse Cuadros', 'Aitor García Pablos', 'Ona de Gibert Bonet'] | null | null | null | null | lrec-2022-6 | ['de-identification'] | ['natural-language-processing'] | [ 5.62803112e-02 5.29195547e-01 9.25637111e-02 -4.36864823e-01
-8.09453368e-01 -8.34984779e-01 9.79385257e-01 6.88918710e-01
-9.36954677e-01 1.10805154e+00 4.87977982e-01 -1.65641785e-01
-6.42274991e-02 -6.40097320e-01 -5.23090959e-01 -2.19010696e-01
1.83875293e-01 8.96582246e-01 1.90505594e-01 -1.13144375e-01
1.88960075e-01 8.47958744e-01 -1.38666010e+00 5.82985878e-01
1.02044988e+00 4.32261020e-01 -2.57687986e-01 1.87454984e-01
-2.22541705e-01 6.05744362e-01 -8.43260825e-01 -9.32801604e-01
4.52399641e-01 -5.07044345e-02 -1.14664781e+00 -2.75937259e-01
3.52562875e-01 2.61677742e-01 1.28901526e-01 9.62736130e-01
5.44314504e-01 -2.31306404e-01 6.12000048e-01 -1.20072162e+00
-1.78061813e-01 8.28840435e-01 -2.13704295e-02 -9.34381410e-02
4.66197401e-01 -1.02309689e-01 7.03758776e-01 -4.32098806e-01
1.25999606e+00 1.07309771e+00 7.33813643e-01 2.88370788e-01
-1.26670194e+00 -5.10716081e-01 -4.63751137e-01 -9.08081084e-02
-1.62572086e+00 -8.91358554e-01 2.34531790e-01 -6.98121309e-01
6.83753014e-01 3.02141041e-01 1.37024239e-01 1.15809298e+00
-2.10033238e-01 3.95636320e-01 1.52178299e+00 -6.46769285e-01
2.93714136e-01 8.50206792e-01 -6.41719848e-02 4.90347780e-02
7.25436568e-01 -1.28957123e-01 -2.34965146e-01 -4.44982678e-01
2.33020499e-01 -6.41939044e-01 -1.78534195e-01 -7.74471998e-01
-1.12484694e+00 7.87789226e-01 -4.25090231e-02 8.77505243e-01
-3.15514177e-01 -4.36759323e-01 8.15161884e-01 4.20802563e-01
5.74881196e-01 6.66913271e-01 -5.83516777e-01 -6.68762401e-02
-1.06071150e+00 3.49049300e-01 1.21444595e+00 1.14498663e+00
7.27674544e-01 -3.73889387e-01 -2.39596546e-01 7.45199680e-01
3.03605556e-01 3.77531886e-01 6.77443922e-01 -3.50571275e-01
8.06974411e-01 9.26506698e-01 6.90065324e-01 -1.02844584e+00
-4.70231205e-01 -1.03101462e-01 -5.16606688e-01 1.98550206e-02
7.21712828e-01 -2.78490812e-01 -5.43412983e-01 1.64022779e+00
4.93592501e-01 -3.44825417e-01 3.64027888e-01 3.67255539e-01
5.44328272e-01 1.54460326e-01 3.32102954e-01 -1.83528379e-01
1.34028792e+00 -3.05361509e-01 -7.87594378e-01 2.72813141e-01
1.08173287e+00 -1.03618932e+00 5.98925948e-01 1.06011972e-01
-6.50258303e-01 -3.36352378e-01 -7.14165449e-01 3.55090871e-02
-8.90222490e-01 5.03191292e-01 2.22643375e-01 1.19629669e+00
-8.36182833e-01 4.34068471e-01 -6.27803028e-01 -8.94403219e-01
2.97642410e-01 2.05533639e-01 -9.44655061e-01 3.23105663e-01
-1.44892633e+00 9.42558348e-01 9.17280495e-01 -8.43929499e-02
-3.78885627e-01 -4.72784728e-01 -8.21497560e-01 -2.28648812e-01
1.37918308e-01 -2.54070342e-01 9.20340955e-01 -6.65084898e-01
-9.57963765e-01 1.63628566e+00 5.44466032e-03 -8.12735081e-01
1.13726640e+00 -2.36393481e-01 -9.71033633e-01 -2.85829782e-01
6.20467305e-01 2.77988464e-01 4.03100222e-01 -1.21711898e+00
-6.82251871e-01 -4.18033332e-01 -1.30209282e-01 -2.75211513e-01
-3.29936028e-01 4.17089701e-01 -4.72571403e-02 -9.21409965e-01
-3.93353224e-01 -9.71036613e-01 -5.97365312e-02 -5.55011332e-01
-4.85348165e-01 1.43731296e-01 4.95151877e-01 -1.10996008e+00
1.44971776e+00 -1.97227871e+00 -4.51845787e-02 4.04435247e-01
-1.50209129e-01 6.32284343e-01 2.18495250e-01 7.53378749e-01
-3.02127779e-01 4.09101486e-01 -5.00994742e-01 -3.10002744e-01
1.47105724e-01 5.29808328e-02 -2.13730678e-01 4.76685077e-01
2.24730790e-01 5.58352947e-01 -7.28744447e-01 -5.58533251e-01
8.24016556e-02 3.23737949e-01 -2.38627136e-01 6.88532665e-02
-9.25408155e-02 5.61262846e-01 -3.66698563e-01 1.26682073e-01
8.00200105e-01 4.74881649e-01 4.10959452e-01 -1.36192009e-01
-4.11271036e-01 2.00547531e-01 -1.69901776e+00 1.65475774e+00
-5.67136645e-01 4.12381798e-01 -2.93283965e-02 -6.10809863e-01
1.04668963e+00 5.98703444e-01 4.71789688e-01 -7.12993801e-01
1.46125808e-01 6.12917840e-01 -3.72962981e-01 -6.40661895e-01
9.02470410e-01 7.88089484e-02 -3.84869754e-01 5.15573502e-01
2.28857368e-01 3.48916322e-01 6.77455544e-01 1.32323638e-01
7.49559939e-01 1.11418150e-01 5.69607437e-01 -3.96016240e-01
1.07485843e+00 1.13537073e-01 5.27683139e-01 4.46765512e-01
-1.31463453e-01 3.90235037e-01 5.63420296e-01 -3.85675430e-01
-1.39277124e+00 -6.06307685e-01 -3.32498252e-01 6.17669463e-01
-3.89794469e-01 -3.14134777e-01 -1.11505258e+00 -9.13318455e-01
-4.94626500e-02 1.03756487e+00 -4.69629794e-01 1.49320096e-01
-5.75995386e-01 -7.36245036e-01 1.06775999e+00 -8.35280046e-02
3.45360696e-01 -1.31604624e+00 -5.09640753e-01 2.78177559e-01
-2.21754268e-01 -1.37172341e+00 -2.49313302e-02 -9.13472921e-02
-4.01427954e-01 -1.23984051e+00 -5.88138103e-01 -4.20036733e-01
4.38700616e-01 -7.09271848e-01 1.24792075e+00 -2.99275964e-01
-2.24120915e-01 1.36050746e-01 -4.76334631e-01 -4.47239339e-01
-1.20217347e+00 6.98163509e-01 -9.10858531e-03 2.43087903e-01
5.74019670e-01 -3.03039670e-01 -1.91515926e-02 4.31332260e-01
-1.29350173e+00 -6.12789273e-01 5.53253531e-01 2.84063846e-01
2.68827051e-01 -1.49680629e-01 4.88072455e-01 -1.73864675e+00
7.19825625e-01 -4.39773560e-01 -6.96479619e-01 5.16242862e-01
-4.61320519e-01 3.34145367e-01 5.47540486e-01 2.66688671e-02
-1.16532338e+00 3.47953200e-01 -4.02320683e-01 1.85826421e-01
-8.55845332e-01 1.77009329e-01 -7.05124557e-01 -1.39141589e-01
8.37829471e-01 -2.88723856e-01 -4.24179822e-01 -9.31797385e-01
7.07180560e-01 8.96902025e-01 4.13899243e-01 -4.26145613e-01
8.03371131e-01 2.63867050e-01 -3.46848041e-01 -7.83597767e-01
-3.34770977e-01 -7.28784025e-01 -1.19572651e+00 2.13293344e-01
8.04895639e-01 -1.02089906e+00 -1.22841850e-01 6.77088320e-01
-1.14381313e+00 4.82599400e-02 -5.14033973e-01 2.84641385e-01
-3.52902234e-01 4.10129875e-01 -1.71253905e-01 -6.08966827e-01
-2.40751401e-01 -7.94296026e-01 9.54367876e-01 -3.93318161e-02
-6.01407409e-01 -1.00027966e+00 4.66116697e-01 3.61708820e-01
2.36204535e-01 5.90784073e-01 8.25974166e-01 -1.36330569e+00
-4.89778370e-02 -3.53226036e-01 -9.03315693e-02 1.08825348e-01
1.58042967e-01 -1.52514681e-01 -1.10928345e+00 -3.58129263e-01
-1.09987512e-01 -4.59683537e-02 4.73391026e-01 -4.26860303e-01
3.44597340e-01 -4.08952028e-01 -2.88795948e-01 3.07924271e-01
1.38544261e+00 -7.40197077e-02 8.18605125e-01 6.82434499e-01
5.21615803e-01 7.28393137e-01 6.39297545e-01 3.84734809e-01
3.24935645e-01 9.10144210e-01 -1.86711457e-02 -4.42873687e-03
-1.48374622e-03 -3.07949334e-01 5.64281940e-02 5.17989635e-01
1.86261132e-01 -4.05245662e-01 -1.11975276e+00 8.68343890e-01
-1.68374777e+00 -9.46539283e-01 -4.97636527e-01 2.46666670e+00
7.08957851e-01 -2.67774556e-02 3.51774901e-01 7.80535638e-02
8.84961843e-01 -9.47268978e-02 1.99779153e-01 -6.28799975e-01
-2.51365483e-01 2.85729796e-01 8.04505587e-01 4.17971194e-01
-1.39342546e+00 9.68696177e-01 5.57902527e+00 6.87744558e-01
-1.11235988e+00 1.86146900e-01 1.75565109e-01 3.99860799e-01
-2.17117012e-01 1.70608446e-01 -9.27133322e-01 6.03060305e-01
1.28103220e+00 -1.95027724e-01 -4.50160950e-02 7.88257718e-01
2.58327693e-01 3.11556887e-02 -1.01141071e+00 5.05244493e-01
1.37070781e-02 -1.06415951e+00 1.00908473e-01 1.81097567e-01
7.04064786e-01 -1.29081145e-01 -4.31181848e-01 9.50771049e-02
3.42701942e-01 -7.85081685e-01 8.46769810e-01 7.04334915e-01
7.28465319e-01 -7.64944851e-01 1.11409152e+00 2.09565997e-01
-8.30835581e-01 1.24442667e-01 -1.01706289e-01 4.53832567e-01
3.97519112e-01 5.52668154e-01 -1.07013011e+00 1.06507194e+00
5.33582091e-01 3.10659498e-01 -1.01694930e+00 1.15534389e+00
-2.01086700e-01 2.57441163e-01 -4.48386282e-01 2.37679154e-01
-1.11382954e-01 -7.45908394e-02 5.41675091e-01 1.42778683e+00
3.21570784e-01 -5.66984773e-01 -2.53571391e-01 6.78820491e-01
-3.20972323e-01 7.06308186e-01 -5.83954453e-01 -2.08165497e-01
5.46458006e-01 1.28805709e+00 -5.79330504e-01 -1.57003388e-01
-2.62741566e-01 1.07632017e+00 3.60252053e-01 -2.94530438e-03
-5.08713126e-01 -6.85078561e-01 3.84961426e-01 4.26089585e-01
2.80295402e-01 5.50123975e-02 -2.64756614e-03 -1.25644648e+00
2.45805353e-01 -1.00530696e+00 4.97434765e-01 -4.71739173e-01
-1.18025005e+00 8.06720912e-01 1.81117624e-01 -1.08766282e+00
-6.84371293e-01 -5.22801995e-01 -1.09221891e-01 9.89498258e-01
-1.05835760e+00 -1.32063997e+00 -1.34984910e-01 3.12781125e-01
-2.25203991e-01 -3.35202962e-01 1.03894353e+00 9.34496224e-01
-5.27747691e-01 6.93126857e-01 3.64204735e-01 4.62006211e-01
9.54586327e-01 -1.14235437e+00 7.48111010e-01 9.84829545e-01
3.76630723e-01 7.46229529e-01 7.95336902e-01 -7.23922729e-01
-5.22105277e-01 -1.24829793e+00 1.66614902e+00 -7.38910556e-01
7.19041765e-01 -6.47665620e-01 -1.02518237e+00 7.17644989e-01
1.36079356e-01 -2.67814904e-01 7.49164104e-01 -1.78579152e-01
-2.44925648e-01 9.01174396e-02 -1.68969190e+00 2.94936690e-02
7.55831182e-01 -6.21463895e-01 -7.01961100e-01 3.22009236e-01
3.58253926e-01 -2.40858406e-01 -1.32328534e+00 1.20675102e-01
3.34516138e-01 -1.11971605e+00 6.54784560e-01 -6.88464046e-01
-2.36291900e-01 -4.41492736e-01 5.85605763e-02 -1.13559699e+00
2.58170605e-01 -5.21805227e-01 4.37511176e-01 2.23036122e+00
6.53805912e-01 -7.46243536e-01 6.85512364e-01 7.85722554e-01
3.16090047e-01 2.06803188e-01 -1.02248859e+00 -6.08576238e-01
1.00817427e-01 -1.62023515e-01 8.83332789e-01 1.25365973e+00
-3.45974535e-01 1.41443864e-01 -2.83397168e-01 3.08103144e-01
2.87294745e-01 -2.37503722e-01 9.84150112e-01 -1.41434145e+00
2.40886524e-01 -6.15796968e-02 -7.58270323e-01 -8.36721957e-02
2.92716622e-01 -9.28554416e-01 -4.46437001e-01 -1.06369400e+00
-8.47705454e-02 -7.01075852e-01 4.62801829e-02 4.10917699e-01
1.55500546e-01 1.32089362e-01 1.57996297e-01 2.24518791e-01
-3.66777748e-01 1.99455157e-01 7.94813260e-02 2.10856587e-01
-2.16451317e-01 8.73916522e-02 -5.20742297e-01 5.73237777e-01
6.27031028e-01 -9.89921093e-01 9.05357078e-02 -2.66492337e-01
2.85944194e-01 -5.46813548e-01 1.46319732e-01 -1.08303487e+00
8.81401356e-03 2.30175570e-01 1.74453765e-01 -3.76807302e-01
-2.45539412e-01 -1.08161438e+00 6.35360658e-01 3.12145978e-01
-3.42047274e-01 9.20321867e-02 1.90073296e-01 3.26054901e-01
-3.32658559e-01 -6.06606960e-01 7.16850221e-01 -2.41770700e-01
-9.04354036e-01 -3.34643386e-02 -1.97652355e-01 5.95077693e-01
1.26472461e+00 4.09481023e-03 -2.00431179e-02 -7.34235793e-02
-8.78576219e-01 4.73694205e-02 1.01556861e+00 4.73593444e-01
-2.79478371e-01 -1.12330651e+00 -7.32642114e-01 2.02637449e-01
6.08198762e-01 -3.90591741e-01 1.26563370e-01 4.55071270e-01
-8.99282157e-01 6.82013929e-01 -5.64413846e-01 -2.01381624e-01
-1.27505362e+00 6.10586941e-01 1.65891632e-01 -5.10295391e-01
-8.97481814e-02 1.97291806e-01 -5.06765068e-01 -1.05420768e+00
-5.08771911e-02 -6.64416477e-02 -6.18262649e-01 4.48435068e-01
3.09351623e-01 5.26922584e-01 5.15618503e-01 -1.34848154e+00
-4.01548594e-01 2.89729506e-01 1.05184510e-01 -2.03073993e-01
1.33591402e+00 -2.26446986e-01 -3.93330276e-01 2.50244528e-01
1.19427752e+00 6.55369461e-01 -2.81581521e-01 -5.47433496e-02
8.73556495e-01 -3.53783965e-01 -4.61039990e-01 -6.75754011e-01
-7.21784174e-01 5.24162710e-01 7.17681170e-01 3.75564426e-01
7.71649897e-01 -2.98333824e-01 2.96746999e-01 3.37256163e-01
5.26774347e-01 -1.09467113e+00 -8.84241760e-01 2.48448819e-01
7.01342046e-01 -1.12952232e+00 7.27281496e-02 -4.74348515e-01
-5.96034527e-01 9.41017210e-01 1.14013560e-01 3.06345016e-01
3.87578368e-01 -1.02062775e-02 3.03201437e-01 -4.99441698e-02
1.57885142e-02 -1.32021949e-01 2.10630074e-01 6.99329734e-01
5.94034553e-01 -5.21152727e-02 -8.13650429e-01 3.55253309e-01
-3.27569515e-01 2.27092922e-01 5.16753137e-01 7.65332758e-01
1.49349317e-01 -1.76062655e+00 -4.16058660e-01 1.97145820e-01
-8.17467809e-01 -5.36672920e-02 -8.84810507e-01 1.26614869e+00
4.52992499e-01 8.28156769e-01 -1.75575227e-01 4.96663675e-02
7.21810102e-01 3.25890392e-01 6.14886731e-02 -5.34868240e-01
-1.04162014e+00 -7.22024024e-01 7.05041289e-01 -3.72944891e-01
-6.30232334e-01 -9.46984708e-01 -7.23072648e-01 -2.52121687e-01
-2.18776852e-01 4.75705475e-01 7.82156825e-01 7.95948327e-01
5.34020185e-01 5.42208776e-02 3.09842825e-01 -2.46740982e-01
-3.23708594e-01 -9.71645355e-01 -3.50449502e-01 1.09335780e+00
-2.30120286e-01 -3.29490036e-01 -2.13330612e-01 2.76912183e-01] | [9.610657691955566, 9.50782585144043] |
947bcd28-0196-4035-a042-f3ae084fe9b9 | parametric-scattering-networks | 2107.09539 | null | https://arxiv.org/abs/2107.09539v4 | https://arxiv.org/pdf/2107.09539v4.pdf | Parametric Scattering Networks | The wavelet scattering transform creates geometric invariants and deformation stability. In multiple signal domains, it has been shown to yield more discriminative representations compared to other non-learned representations and to outperform learned representations in certain tasks, particularly on limited labeled data and highly structured signals. The wavelet filters used in the scattering transform are typically selected to create a tight frame via a parameterized mother wavelet. In this work, we investigate whether this standard wavelet filterbank construction is optimal. Focusing on Morlet wavelets, we propose to learn the scales, orientations, and aspect ratios of the filters to produce problem-specific parameterizations of the scattering transform. We show that our learned versions of the scattering transform yield significant performance gains in small-sample classification settings over the standard scattering transform. Moreover, our empirical results suggest that traditional filterbank constructions may not always be necessary for scattering transforms to extract effective representations. | ['Michael Eickenberg', 'Irina Rish', 'Muawiz Chaudhary', 'Guy Wolf', 'Eugene Belilovsky', 'Laurent Alsène-Racicot', 'Benjamin Thérien', 'Shanel Gauthier'] | 2021-07-20 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Gauthier_Parametric_Scattering_Networks_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Gauthier_Parametric_Scattering_Networks_CVPR_2022_paper.pdf | cvpr-2022-1 | ['small-data'] | ['computer-vision'] | [ 4.65516806e-01 5.72521947e-02 -1.25512347e-01 -3.26904327e-01
-1.15074420e+00 -5.78811288e-01 4.60007548e-01 -2.73161501e-01
7.31019676e-02 4.84423876e-01 5.23714602e-01 1.02986738e-01
-3.31062496e-01 -6.94224000e-01 -6.38748586e-01 -1.03273082e+00
-4.15579081e-01 1.18445776e-01 3.24996375e-02 -2.45743528e-01
4.29991968e-02 6.34813964e-01 -1.41734135e+00 4.48822200e-01
3.10631126e-01 1.03823328e+00 6.80946857e-02 6.91186428e-01
3.22487950e-01 1.64944366e-01 -5.25460243e-01 -1.13305345e-01
6.05528712e-01 -3.11180979e-01 -5.44386268e-01 2.65302360e-01
7.47923732e-01 -1.69479400e-01 -2.85505950e-01 7.98067451e-01
4.21464860e-01 4.01903361e-01 7.05710292e-01 -6.85334265e-01
-6.65511787e-01 5.99616349e-01 -3.81824702e-01 4.76317406e-01
2.98186958e-01 -2.21855342e-01 1.46606541e+00 -9.39303160e-01
5.03681302e-01 1.27815437e+00 8.92730892e-01 2.55928785e-01
-1.55104947e+00 -5.31637192e-01 -1.42044857e-01 -1.71026476e-02
-1.04386723e+00 -5.84379554e-01 1.13146734e+00 -2.74230033e-01
6.29396319e-01 3.22603196e-01 6.41005158e-01 1.36423492e+00
2.67450720e-01 3.89583886e-01 1.19810224e+00 -6.48195624e-01
1.16367437e-01 -2.32636228e-01 1.85401484e-01 7.32870102e-01
4.97445077e-01 9.52476636e-02 -7.58053660e-01 -4.45596814e-01
1.11923718e+00 -9.04405564e-02 -5.23191690e-01 -3.39666694e-01
-1.25199711e+00 1.10303926e+00 1.42430574e-01 5.71696103e-01
-4.12128657e-01 5.55184364e-01 1.72704846e-01 6.54899657e-01
8.97463679e-01 6.50714695e-01 -4.46693778e-01 1.09986335e-01
-7.66949713e-01 2.71722406e-01 7.24559486e-01 4.97059345e-01
7.68349826e-01 4.01840389e-01 7.74598196e-02 6.98330879e-01
-1.80958718e-01 5.90421200e-01 1.07833773e-01 -1.05434203e+00
2.05183044e-01 -2.16501251e-01 1.58974156e-02 -1.14674258e+00
-4.17514145e-01 -8.75380576e-01 -6.63127124e-01 -1.64833769e-01
5.53972244e-01 -7.43084401e-02 -5.53424776e-01 1.68367517e+00
5.63371368e-02 6.45547509e-01 -1.76113825e-02 8.92047048e-01
3.34938169e-01 4.61559951e-01 -2.42622763e-01 -2.06094980e-01
1.20129478e+00 -1.55966654e-01 -5.67952096e-01 -1.66215003e-01
4.00528729e-01 -1.20890689e+00 1.09190381e+00 3.38233232e-01
-9.57500219e-01 -6.06802166e-01 -1.08230877e+00 2.79210508e-01
1.90822318e-01 3.03030789e-01 8.32322896e-01 6.32861495e-01
-6.45317554e-01 6.92190945e-01 -9.90980804e-01 -1.02107428e-01
2.80229539e-01 9.72858295e-02 -2.35084906e-01 -4.55358066e-02
-8.56624484e-01 8.50510538e-01 -8.05912092e-02 -8.30424801e-02
-5.17262995e-01 -8.85827959e-01 -7.12609231e-01 8.71040076e-02
2.24292222e-02 -5.55091858e-01 8.19930136e-01 -7.19494224e-01
-1.28863609e+00 6.73262358e-01 -6.65793866e-02 -5.74300230e-01
-1.93487816e-02 -3.08652997e-01 -2.16045335e-01 6.36621773e-01
-1.34849563e-01 1.11311395e-02 1.49184871e+00 -1.15718257e+00
-1.09105319e-01 -9.15828869e-02 1.62618861e-01 -1.72443002e-01
-5.97882450e-01 -1.94189534e-01 2.72016406e-01 -1.19612169e+00
5.37855327e-01 -9.41497445e-01 -2.58907199e-01 -1.61780000e-01
-4.40673493e-02 6.80065975e-02 9.72610474e-01 -7.38107443e-01
7.60202050e-01 -2.29931450e+00 2.17362285e-01 4.78709906e-01
1.04230642e-01 -2.04911828e-01 -2.39221469e-01 4.36401725e-01
-2.13627040e-01 -1.50040925e-01 5.22101074e-02 -8.90381783e-02
-1.96047485e-01 2.70304024e-01 -6.41376615e-01 1.00616825e+00
3.69192004e-01 4.74294960e-01 -5.61982572e-01 -1.00269891e-01
1.02793537e-01 7.44906306e-01 -7.95305371e-01 -1.04792625e-01
-1.12329819e-03 6.18399143e-01 -7.07484066e-01 4.63088334e-01
4.85504776e-01 -2.04205126e-01 7.39033474e-03 -7.01625168e-01
3.93354565e-01 2.89244860e-01 -1.09965646e+00 1.52904570e+00
-7.41049290e-01 7.70951867e-01 1.89346924e-01 -1.68826151e+00
1.21775568e+00 4.08717453e-01 8.79937708e-01 -2.26559982e-01
4.69604097e-02 1.84449077e-01 -1.37038916e-01 -5.04401922e-01
-2.05543879e-02 -4.99651581e-01 -1.31765306e-01 3.66530925e-01
2.62044728e-01 -2.85361618e-01 -1.48182839e-01 -3.46399099e-01
1.26254487e+00 -9.78699625e-02 2.36867532e-01 -6.43872917e-01
3.85320038e-01 -3.66498917e-01 5.30912340e-01 5.16278386e-01
2.52359778e-01 8.65217984e-01 1.43720940e-01 -6.88813806e-01
-1.03295970e+00 -1.23061979e+00 -3.49099398e-01 1.02832711e+00
-1.59117967e-01 -4.41790223e-01 -4.72356468e-01 -3.42867851e-01
-4.38152738e-02 4.34796870e-01 -3.66955757e-01 -3.77351075e-01
-1.10693431e+00 -6.82733357e-01 4.92727697e-01 5.58217764e-01
2.95576125e-01 -3.82076293e-01 -8.90068114e-01 2.61630625e-01
-5.53039685e-02 -1.54645717e+00 -4.99720693e-01 2.40007252e-01
-9.30976868e-01 -1.22045958e+00 -4.40637261e-01 -7.07725704e-01
6.28483057e-01 4.02912170e-01 8.56611729e-01 -1.72096808e-02
-3.95298213e-01 7.25414574e-01 -6.08445942e-01 -1.41202316e-01
-2.32626066e-01 6.61572888e-02 7.55725503e-02 2.62664765e-01
-1.74658328e-01 -1.08085644e+00 -6.24011397e-01 2.36220196e-01
-8.98882091e-01 -2.13760197e-01 5.58908701e-01 1.15439248e+00
4.37225252e-01 3.20381314e-01 5.05748093e-01 -7.71374762e-01
5.92614830e-01 -1.53473884e-01 -2.32090965e-01 -1.25305541e-02
-8.57876837e-02 3.11670363e-01 8.61200809e-01 -7.78003752e-01
-8.25520337e-01 -4.67600115e-02 -2.66700033e-02 -5.14669716e-01
1.59515202e-01 4.98246461e-01 3.47792119e-01 -5.29484689e-01
8.42962503e-01 1.10083073e-01 6.03908338e-02 -5.76238871e-01
9.45147797e-02 1.09192366e-02 2.91689783e-01 -1.20768690e+00
1.28661227e+00 7.09784269e-01 3.47697467e-01 -1.17694318e+00
-9.65880454e-01 -3.10612082e-01 -5.89933038e-01 -1.42060556e-02
5.68768024e-01 -7.91048229e-01 -4.34025556e-01 -2.41702989e-01
-9.81902540e-01 3.78419310e-02 -5.07515848e-01 8.73944521e-01
-8.10257494e-01 8.58145878e-02 -4.85626578e-01 -7.91359484e-01
-1.64491430e-01 -9.74627554e-01 1.33384395e+00 -2.62422681e-01
-1.96012348e-01 -1.03791487e+00 -1.40784323e-01 1.13670819e-01
6.28547847e-01 6.59303665e-01 1.03936982e+00 -4.64062303e-01
-4.59994525e-01 -1.94505751e-01 2.93890148e-01 5.28115273e-01
3.47750992e-01 -1.32765621e-01 -9.94979680e-01 -6.67826951e-01
2.89064646e-01 -1.85150698e-01 1.07830441e+00 6.01505518e-01
1.15653920e+00 -2.44254887e-01 -8.73630028e-03 8.92041981e-01
1.20193899e+00 -4.56462242e-02 1.59312189e-01 -3.24085951e-02
3.31144780e-01 5.62436044e-01 2.97576129e-01 4.46970284e-01
-3.63877982e-01 7.40353703e-01 1.52564328e-03 5.39626628e-02
-2.33134165e-01 1.17256297e-02 5.14127254e-01 9.22632515e-01
-5.08504510e-01 3.37502748e-01 -5.73089361e-01 1.06450446e-01
-1.33322549e+00 -1.08103061e+00 2.99211681e-01 2.11713314e+00
5.52359104e-01 1.64269984e-01 -3.78801748e-02 4.21787351e-01
4.37568218e-01 5.70884526e-01 -1.62526160e-01 2.44908687e-02
-3.11581403e-01 1.15746057e+00 5.96804619e-01 4.46171463e-01
-1.15040147e+00 6.55894339e-01 7.00803614e+00 5.78386307e-01
-1.21396124e+00 1.95528343e-01 2.85965681e-01 -3.15896496e-02
-5.42259097e-01 9.11111757e-02 -6.75370753e-01 5.39662912e-02
7.99210846e-01 -2.17841446e-01 4.35380489e-01 5.78017652e-01
1.29605100e-01 5.93970001e-01 -1.02930295e+00 9.28059518e-01
-1.14720993e-01 -1.42396927e+00 6.16858201e-03 -1.29107526e-02
3.44200581e-01 -4.11428005e-01 2.84941047e-01 -1.58841722e-03
-5.31161502e-02 -1.06001139e+00 6.34248257e-01 3.60097080e-01
6.33453131e-01 -7.26872265e-01 4.44746524e-01 -4.16046903e-02
-1.48993444e+00 -4.42790389e-02 -6.19674087e-01 8.77210498e-02
-9.23972651e-02 8.05839956e-01 -7.13757575e-01 4.24250424e-01
6.03152514e-01 7.75273740e-01 -2.58643091e-01 5.98360181e-01
1.83184847e-01 1.09443152e+00 -4.98028427e-01 2.47345164e-01
1.05855733e-01 -2.66076833e-01 7.46915102e-01 9.86052215e-01
6.34789944e-01 1.25726014e-01 5.41203201e-01 5.67450881e-01
1.16969250e-01 -1.00783467e-01 -6.91331029e-01 -1.81498706e-01
3.84569556e-01 9.68208909e-01 -8.02301884e-01 2.45444521e-01
-3.94989938e-01 5.09286523e-01 -1.02971263e-01 5.53990424e-01
-6.28123701e-01 7.37699345e-02 9.18587446e-01 4.14478660e-01
6.47084057e-01 -7.29929507e-01 -3.95133533e-02 -1.26306987e+00
7.84437880e-02 -9.91373539e-01 3.80388945e-01 -2.55211264e-01
-1.50424647e+00 4.77118999e-01 4.14000690e-01 -1.25044513e+00
-2.49976382e-01 -8.07719588e-01 -5.91834247e-01 5.53488612e-01
-1.19537461e+00 -1.24611497e+00 -3.82795371e-02 6.66826546e-01
6.11607552e-01 -1.92900077e-01 1.05385053e+00 5.46911769e-02
-5.31828962e-02 4.47503090e-01 8.03357512e-02 2.14987397e-01
5.37359416e-01 -9.82113004e-01 3.01293552e-01 6.37657523e-01
5.08838117e-01 7.83724725e-01 1.16540897e+00 -2.57416368e-01
-1.92473400e+00 -6.75167680e-01 2.62921248e-02 -6.72087371e-02
7.42736459e-01 -4.96332228e-01 -8.45411718e-01 8.36813390e-01
-2.04847440e-01 6.02284372e-01 8.48442137e-01 1.85173422e-01
-6.23554945e-01 -3.48471284e-01 -1.03545833e+00 2.21270725e-01
1.20464444e+00 -6.03652775e-01 -6.68759286e-01 4.45064753e-01
5.05629182e-01 -2.90613025e-01 -1.09741366e+00 5.13913572e-01
5.85337579e-01 -8.20818126e-01 1.62351274e+00 -6.39997840e-01
1.86358422e-01 1.17497437e-01 -7.07289636e-01 -1.42133868e+00
-5.88722467e-01 -8.24723184e-01 -3.99361812e-02 7.36905217e-01
8.60949904e-02 -8.76660407e-01 8.58022511e-01 -1.29958823e-01
-1.80530325e-01 -8.73389602e-01 -1.30535710e+00 -8.65724862e-01
1.68951690e-01 -3.63108039e-01 3.86105776e-01 9.19068813e-01
-4.86432910e-01 2.17354208e-01 -2.79626638e-01 5.03641427e-01
8.62009883e-01 4.59258258e-01 5.44061601e-01 -1.38982415e+00
-7.19738126e-01 -2.50437170e-01 -6.68325067e-01 -9.90250707e-01
5.60434043e-01 -8.21825504e-01 -3.35162461e-01 -8.61594558e-01
-2.84026533e-01 -3.91854644e-01 -2.43546635e-01 7.60014206e-02
1.22989610e-01 2.68842667e-01 -7.07864761e-04 2.21402332e-01
2.31588498e-01 4.20421779e-01 1.43196297e+00 -1.04602113e-01
1.41355872e-01 2.48347670e-01 -6.64909601e-01 7.17990637e-01
7.95058429e-01 -2.02060610e-01 -6.31678343e-01 -2.06517935e-01
5.88437319e-02 8.96243751e-02 3.76720339e-01 -1.06865883e+00
-2.77395368e-01 -3.01110923e-01 5.36393583e-01 -8.28278884e-02
6.57854915e-01 -7.42761910e-01 2.56113917e-01 3.79529148e-01
-5.16825736e-01 -1.27639011e-01 -1.25744790e-01 6.63803637e-01
-1.46418586e-01 -8.97681937e-02 8.98952901e-01 -5.28166443e-02
-2.45393530e-01 1.83230191e-01 -1.76184714e-01 3.01328301e-02
7.70317018e-01 -2.00641349e-01 -1.02674554e-03 -5.36889553e-01
-6.78927958e-01 -5.18440783e-01 1.02037057e-01 1.92800164e-01
7.17691123e-01 -1.39073670e+00 -8.25699747e-01 4.89498496e-01
-1.80973276e-01 -2.49252975e-01 -2.66675502e-01 7.07853794e-01
-4.29311812e-01 1.51876509e-01 -1.30265385e-01 -4.48642641e-01
-1.28523540e+00 2.69742403e-02 2.71100044e-01 -1.24986529e-01
-1.23211062e+00 8.16355586e-01 8.30025598e-02 4.69127260e-02
-1.39658108e-01 -5.94024062e-01 1.70281082e-01 1.75428297e-02
3.58327419e-01 1.92107052e-01 1.92499161e-01 -6.11206293e-01
-1.58128649e-01 8.88748407e-01 1.35462672e-01 -2.24216115e-02
1.75132871e+00 2.90312856e-01 2.20375061e-01 3.54929835e-01
1.50649011e+00 3.26272845e-01 -1.10997427e+00 -3.60026956e-01
-1.01684323e-02 -7.04770386e-01 2.18717769e-01 1.59866482e-01
-1.30765748e+00 5.36216497e-01 2.89923042e-01 5.10413706e-01
1.19089866e+00 1.78027138e-01 1.03978765e+00 6.09688044e-01
5.83909333e-01 -7.79881775e-01 3.61874819e-01 2.59628981e-01
1.06521058e+00 -8.34718704e-01 2.33610392e-01 -7.60505736e-01
3.95934209e-02 1.67291474e+00 -1.19869011e-02 -7.36937582e-01
8.09593439e-01 5.05240977e-01 -1.70351669e-01 -2.39993528e-01
-5.89677095e-01 -1.92919318e-02 5.73392868e-01 6.15342140e-01
5.22168100e-01 1.06941620e-02 -1.24294236e-01 4.18615311e-01
-7.50423014e-01 -5.77551365e-01 2.79996574e-01 9.07177210e-01
-6.21943176e-01 -1.13772035e+00 -7.38126040e-01 5.91156423e-01
-5.73824823e-01 2.29394019e-01 -8.46751556e-02 7.34615207e-01
-1.51649162e-01 6.81005538e-01 -6.03659302e-02 -4.11489546e-01
3.08320016e-01 2.28358567e-01 1.00964046e+00 -6.48794830e-01
-3.08424532e-01 3.18417788e-01 1.72671005e-01 -4.50651407e-01
-6.98556423e-01 -6.78780973e-01 -1.12289584e+00 -2.03692168e-01
-3.15412968e-01 1.32436335e-01 3.13793540e-01 9.74990487e-01
6.49308935e-02 4.60078478e-01 8.03202033e-01 -1.08014560e+00
-1.05029857e+00 -7.96649158e-01 -5.74720263e-01 6.65198982e-01
5.70205927e-01 -1.14057934e+00 -7.52686501e-01 2.24657655e-01] | [15.356672286987305, 5.62111234664917] |
773b2b29-b3d9-479b-bc85-8b70232496cb | mmd-aggregated-two-sample-test | 2110.15073 | null | https://arxiv.org/abs/2110.15073v3 | https://arxiv.org/pdf/2110.15073v3.pdf | MMD Aggregated Two-Sample Test | We propose two novel nonparametric two-sample kernel tests based on the Maximum Mean Discrepancy (MMD). First, for a fixed kernel, we construct an MMD test using either permutations or a wild bootstrap, two popular numerical procedures to determine the test threshold. We prove that this test controls the probability of type I error non-asymptotically. Hence, it can be used reliably even in settings with small sample sizes as it remains well-calibrated, which differs from previous MMD tests which only guarantee correct test level asymptotically. When the difference in densities lies in a Sobolev ball, we prove minimax optimality of our MMD test with a specific kernel depending on the smoothness parameter of the Sobolev ball. In practice, this parameter is unknown and, hence, the optimal MMD test with this particular kernel cannot be used. To overcome this issue, we construct an aggregated test, called MMDAgg, which is adaptive to the smoothness parameter. The test power is maximised over the collection of kernels used, without requiring held-out data for kernel selection (which results in a loss of test power), or arbitrary kernel choices such as the median heuristic. We prove that MMDAgg still controls the level non-asymptotically, and achieves the minimax rate over Sobolev balls, up to an iterated logarithmic term. Our guarantees are not restricted to a specific type of kernel, but hold for any product of one-dimensional translation invariant characteristic kernels. We provide a user-friendly parameter-free implementation of MMDAgg using an adaptive collection of bandwidths. We demonstrate that MMDAgg significantly outperforms alternative state-of-the-art MMD-based two-sample tests on synthetic data satisfying the Sobolev smoothness assumption, and that, on real-world image data, MMDAgg closely matches the power of tests leveraging the use of models such as neural networks. | ['Arthur Gretton', 'Benjamin Guedj', 'Béatrice Laurent', 'Mélisande Albert', 'Ilmun Kim', 'Antonin Schrab'] | 2021-10-28 | null | null | null | null | ['hypothesis-testing', 'hypothesis-testing'] | ['methodology', 'miscellaneous'] | [-7.92823136e-02 -2.05561131e-01 -2.71696806e-01 -7.55030587e-02
-9.53675926e-01 -6.98362291e-01 1.29561409e-01 2.71951228e-01
-5.98513484e-01 8.30173075e-01 -4.92205918e-01 -6.36763871e-01
-4.02932942e-01 -9.17160451e-01 -8.56455624e-01 -1.01982749e+00
-3.88114572e-01 3.32627207e-01 6.60774291e-01 2.35267311e-01
4.23550814e-01 5.11279643e-01 -1.40077782e+00 -3.99014592e-01
1.27429986e+00 1.22608614e+00 -1.90066501e-01 7.90962279e-01
3.80066037e-01 -8.16620514e-02 -4.35695559e-01 -2.29864538e-01
3.80533785e-01 -4.20868576e-01 -4.79623556e-01 -1.35748118e-01
4.74040687e-01 -2.52134711e-01 3.06969285e-01 1.26862824e+00
5.61209917e-01 2.67656714e-01 1.01936376e+00 -1.22701585e+00
-6.65606141e-01 1.53675988e-01 -4.92530376e-01 1.30229503e-01
-7.80410618e-02 -1.57432295e-02 7.39423215e-01 -9.99247551e-01
5.45053519e-02 7.93979883e-01 1.01473045e+00 2.46589556e-01
-1.65882134e+00 -4.40325409e-01 -4.69538540e-01 -7.10638016e-02
-1.63255644e+00 -7.28145763e-02 3.09936255e-01 -6.86543941e-01
3.32294643e-01 3.38439256e-01 4.37314868e-01 7.54115522e-01
3.34539920e-01 2.34770656e-01 1.36260426e+00 -5.29829621e-01
7.68238068e-01 2.53723472e-01 1.76833987e-01 8.33704710e-01
6.11268222e-01 -7.31866583e-02 -2.33141985e-02 -7.99862802e-01
1.26337576e+00 -4.67738569e-01 -5.56809425e-01 -7.22477674e-01
-1.07774115e+00 1.05877542e+00 -2.30618894e-01 2.10192114e-01
-6.79698959e-02 -1.20397814e-01 2.48000056e-01 4.07028556e-01
7.55063832e-01 2.85839230e-01 -6.81071341e-01 -1.48751764e-02
-8.58941436e-01 1.68360829e-01 1.07168770e+00 6.07149780e-01
6.33598208e-01 -2.30764344e-01 -4.04486090e-01 9.49291587e-01
-8.99712518e-02 8.00435722e-01 3.23436916e-01 -7.03367174e-01
7.19907507e-02 1.41492769e-01 4.17528629e-01 -6.77252233e-01
-3.24539781e-01 -3.32994759e-01 -8.52516592e-01 5.47223628e-01
9.24540997e-01 -4.96909693e-02 -6.61939323e-01 1.88011563e+00
4.68721330e-01 5.13218701e-01 -2.06721991e-01 8.31229568e-01
-7.07629472e-02 2.12018132e-01 -2.13588655e-01 -2.99452364e-01
1.24091053e+00 -3.83943170e-01 -2.44402662e-01 2.56166875e-01
8.60255003e-01 -4.89310950e-01 1.46611512e+00 4.60063726e-01
-1.02601087e+00 -2.43280515e-01 -1.20120108e+00 4.56751585e-01
-2.45597154e-01 2.17909683e-02 1.77147329e-01 9.55290377e-01
-1.00067449e+00 6.56707466e-01 -7.39511132e-01 -1.69919983e-01
1.92453578e-01 2.48560876e-01 -4.51507479e-01 1.51326016e-01
-9.51134443e-01 7.22969115e-01 2.30781838e-01 -3.33313346e-02
-3.61463875e-01 -1.07140827e+00 -7.95408845e-01 4.90449183e-02
2.46285677e-01 -3.78388226e-01 8.77553463e-01 -5.48099637e-01
-1.54835248e+00 6.78033829e-01 1.01935126e-01 -5.27369440e-01
7.66879797e-01 3.16735320e-02 -1.76989332e-01 1.43215746e-01
5.14428876e-02 -5.64880185e-02 9.16415095e-01 -8.20037842e-01
-4.74979222e-01 -2.19587877e-01 -2.98782170e-01 -2.70766586e-01
-3.14437717e-01 -2.22835124e-01 -1.89587787e-01 -5.95493853e-01
-1.29108876e-01 -8.31484735e-01 -7.81175122e-02 1.02117963e-01
-3.35210592e-01 -1.40596822e-01 4.77472872e-01 -4.83137816e-01
1.11504853e+00 -2.32539439e+00 -3.93491030e-01 6.43640161e-01
-9.24850255e-02 1.08553834e-01 -1.57955922e-02 1.21695707e-02
2.34211180e-02 2.65534490e-01 -7.35778570e-01 1.14991240e-01
1.39635965e-01 -1.21271219e-02 -1.81208178e-01 1.05724227e+00
2.08996370e-01 4.35770422e-01 -7.27022529e-01 -2.87608355e-01
4.94424514e-02 4.15123671e-01 -6.86064005e-01 1.29278600e-01
1.93623826e-01 1.94413707e-01 -3.19690198e-01 3.37996155e-01
1.05689192e+00 -3.37092727e-01 -2.93698013e-01 1.46868769e-02
-1.12154782e-01 -2.54086375e-01 -1.45282626e+00 9.34100986e-01
-4.48156923e-01 4.58894223e-01 4.02607694e-02 -1.29128087e+00
9.47252214e-01 9.25773159e-02 1.79887876e-01 -1.53428942e-01
-6.38596192e-02 7.78905451e-01 -3.03123266e-01 -4.37173903e-01
-3.03423285e-01 -3.01067561e-01 3.78926992e-02 4.05502580e-02
-3.88423912e-02 6.39595762e-02 1.14592724e-02 -4.28488076e-01
1.30805147e+00 -2.38140509e-01 5.67429185e-01 -9.11340237e-01
5.56516588e-01 -4.02677983e-01 4.68142807e-01 1.01993752e+00
-1.58492357e-01 6.20087922e-01 9.13256824e-01 -5.18288091e-02
-9.72772002e-01 -1.44194067e+00 -9.85255837e-01 8.16484749e-01
8.96876603e-02 4.97702032e-01 -8.23639154e-01 -6.42064691e-01
4.54289466e-01 6.51993692e-01 -9.67903554e-01 -1.06839985e-01
-2.11766183e-01 -9.12896991e-01 5.98731041e-01 2.79967338e-01
6.83776438e-01 -5.71459353e-01 -5.28176248e-01 -5.16396947e-02
3.50159734e-01 -9.99760568e-01 -7.14352131e-01 2.90710926e-01
-8.08111370e-01 -1.26093495e+00 -1.02724206e+00 -6.94234848e-01
7.35426188e-01 4.20415448e-03 8.19943309e-01 -2.46058464e-01
-1.41309733e-02 4.78948295e-01 -1.28942445e-01 4.94967736e-02
-3.11031163e-01 -1.39972165e-01 5.99828400e-02 1.47377580e-01
2.00168431e-01 -5.52088797e-01 -7.88116753e-01 7.32634544e-01
-9.89289284e-01 -4.07115191e-01 5.40259898e-01 1.05512059e+00
5.60477614e-01 1.58748448e-01 8.37433040e-01 -6.60119295e-01
6.98290944e-01 -5.74612737e-01 -1.17224634e+00 2.85669148e-01
-6.45202219e-01 2.09207729e-01 7.52226949e-01 -9.57914710e-01
-3.65411580e-01 -3.83223325e-01 2.56803900e-01 -4.68439281e-01
-5.86713366e-02 5.82301795e-01 -1.53651694e-03 -5.44842124e-01
7.85374641e-01 1.21465109e-01 8.40540454e-02 -3.12494427e-01
-6.42904118e-02 6.40181661e-01 6.06351554e-01 -8.10075700e-01
6.94840372e-01 3.42678726e-01 3.31372023e-01 -9.91226375e-01
-6.56730473e-01 -5.30584037e-01 -3.49317372e-01 6.59371242e-02
6.72829926e-01 -4.00630355e-01 -9.84258592e-01 4.76513326e-01
-7.50178397e-01 -7.72457063e-01 -3.74837697e-01 8.30183387e-01
-8.88695061e-01 3.81566644e-01 -4.44835395e-01 -1.06368351e+00
1.53787574e-02 -9.58649874e-01 9.06284869e-01 -3.78196314e-02
9.76903364e-02 -1.29624867e+00 1.16356328e-01 -2.57375509e-01
3.41475189e-01 5.30986667e-01 1.10414064e+00 -9.96319652e-01
-1.47980914e-01 -3.14030021e-01 -2.86220968e-01 6.91959441e-01
1.25529826e-01 -1.72180131e-01 -6.06398523e-01 -4.42162603e-01
2.26176754e-01 -9.09453779e-02 6.29852474e-01 7.99034715e-01
1.34206545e+00 -5.50871432e-01 -1.35090891e-02 6.80142820e-01
1.51343966e+00 -1.99992031e-01 3.56505781e-01 3.12022358e-01
1.81810006e-01 4.89438325e-01 3.75267714e-01 5.02203465e-01
-6.36378825e-02 6.40611410e-01 1.30601779e-01 -1.85453016e-02
4.41567481e-01 1.25235334e-01 3.17413807e-01 3.57786506e-01
1.38698399e-01 7.52274916e-02 -6.44867897e-01 5.86351275e-01
-1.79395676e+00 -6.60582364e-01 -1.39703795e-01 3.28245902e+00
1.14044654e+00 1.09717578e-01 4.08411056e-01 1.82170033e-01
9.05061364e-01 -5.14193892e-01 -5.93685508e-01 -4.74155247e-01
-1.49973840e-01 5.78818679e-01 1.05472827e+00 6.22305989e-01
-1.01534045e+00 3.27341944e-01 6.06942463e+00 1.13951468e+00
-9.13251996e-01 4.77900952e-02 5.41948915e-01 2.58376688e-01
-9.43242610e-02 1.28947878e-02 -5.84053814e-01 7.01804101e-01
9.98720348e-01 -2.88970649e-01 3.34439009e-01 1.04655766e+00
3.43161851e-01 -3.35580736e-01 -8.79305899e-01 6.70885980e-01
-3.03149581e-01 -9.55936491e-01 -4.42631245e-01 3.75491232e-01
6.39082074e-01 -3.99005860e-01 1.73298389e-01 1.31921843e-01
1.14625379e-01 -9.57181752e-01 5.27187705e-01 2.89702773e-01
9.02741015e-01 -9.86955523e-01 9.84177589e-01 2.85632610e-01
-9.45011377e-01 1.18744925e-01 -5.20775855e-01 3.82567316e-01
-1.89216435e-01 1.31476080e+00 -6.93902254e-01 1.56287938e-01
4.93620843e-01 -1.04069360e-01 -2.97327191e-01 1.57527828e+00
1.44743770e-01 9.12992597e-01 -5.59141636e-01 -5.88254258e-02
-2.16595046e-02 -4.11599308e-01 5.48510969e-01 1.32539773e+00
8.04982424e-01 -7.25178719e-02 2.19728991e-01 8.16066682e-01
1.20990977e-01 3.89507949e-01 -4.18767184e-01 4.07038540e-01
6.12836421e-01 9.53722298e-01 -7.36979425e-01 -1.58699363e-01
-4.04572099e-01 9.70124722e-01 2.15834051e-01 5.97918451e-01
-9.26113069e-01 -6.82003617e-01 6.52921081e-01 3.02036375e-01
5.29688656e-01 -2.60363638e-01 -3.47519696e-01 -9.16869938e-01
2.94957727e-01 -5.41274846e-01 4.84098881e-01 -2.88175017e-01
-1.61965799e+00 -2.56691594e-02 3.16291600e-01 -1.12972534e+00
-1.88959599e-01 -9.16243076e-01 -6.09650373e-01 1.02672505e+00
-1.31490874e+00 -5.12351036e-01 -5.12770638e-02 4.87129569e-01
-1.43340334e-01 2.53699452e-01 7.36402810e-01 3.31983864e-02
-3.25128257e-01 8.18960369e-01 3.70704442e-01 1.27108216e-01
9.01814401e-01 -1.46723092e+00 -5.42044044e-02 6.97375774e-01
-3.80067766e-01 3.88142377e-01 8.62491131e-01 -5.45624495e-01
-1.06977320e+00 -8.76856744e-01 4.08254236e-01 -2.18636077e-02
1.05726373e+00 -3.74536693e-01 -1.30480981e+00 2.32684538e-01
-5.39250791e-01 5.89588583e-01 7.45154798e-01 6.42824620e-02
-2.92935342e-01 -5.40064648e-02 -1.46739256e+00 6.29318237e-01
6.84744477e-01 -2.73985714e-01 -2.29618579e-01 2.46015310e-01
3.59216183e-01 -1.26501039e-01 -1.15479505e+00 8.21596205e-01
6.06702447e-01 -1.06601501e+00 8.01327586e-01 -4.47825849e-01
-1.19108975e-01 -3.89713109e-01 -8.15211535e-02 -1.27246404e+00
-2.42070034e-01 -6.92741454e-01 -7.94877633e-02 9.26523268e-01
3.82620752e-01 -1.22086596e+00 5.27824938e-01 3.81426007e-01
1.35527939e-01 -9.66162801e-01 -1.38813972e+00 -1.35912442e+00
3.27029347e-01 -4.13289577e-01 1.69602290e-01 9.03987408e-01
9.77384765e-03 -3.76514643e-01 -7.88200349e-02 5.50303996e-01
9.57464814e-01 -1.86817527e-01 6.91922605e-01 -1.34772992e+00
-5.82366526e-01 -6.67678654e-01 -7.86722481e-01 -8.75898898e-01
1.81716189e-01 -5.19857645e-01 2.22346276e-01 -8.65672052e-01
-3.04681156e-03 -6.03763223e-01 -2.24306479e-01 2.96464413e-01
-2.69479632e-01 1.80147484e-01 -6.09727621e-01 -1.05045319e-01
8.35839137e-02 3.79007399e-01 1.07884979e+00 2.54168302e-01
-4.95797366e-01 2.21307173e-01 -4.11911666e-01 6.92028165e-01
8.34243059e-01 -3.01119089e-01 -3.75778228e-01 4.34675962e-01
2.06291843e-02 8.71576145e-02 6.22515559e-01 -1.03506160e+00
9.15516019e-02 -3.38613570e-01 3.09626549e-01 -9.62548852e-02
-1.52208924e-01 -4.69594359e-01 -1.83808237e-01 3.05106282e-01
-2.68343478e-01 -2.32262760e-01 2.17912152e-01 5.77045858e-01
1.93568438e-01 -4.57256019e-01 1.24431837e+00 6.22045040e-01
-7.32048005e-02 2.18135729e-01 -2.81745374e-01 2.89623410e-01
1.02398443e+00 -2.77911425e-01 -3.17214072e-01 -1.81404606e-01
-4.22024757e-01 -5.52430265e-02 5.88351727e-01 -2.44875312e-01
4.21831846e-01 -1.49202049e+00 -7.73034155e-01 3.31136703e-01
1.71340823e-01 -1.34873703e-01 3.56121138e-02 1.29010701e+00
-2.86093414e-01 6.34467006e-02 1.98263645e-01 -6.30828321e-01
-8.47487926e-01 5.13346851e-01 4.42315131e-01 -8.73056948e-02
-5.87767363e-01 4.96713310e-01 3.70387495e-01 -6.40306845e-02
-6.35114592e-03 -6.18868887e-01 2.40531012e-01 -2.72517979e-01
6.45352304e-01 4.74765748e-01 1.89318433e-01 -6.75813556e-02
-2.10850671e-01 6.36905074e-01 4.26948577e-01 -3.66274327e-01
9.55130279e-01 1.60356075e-01 -6.36941344e-02 6.19173169e-01
1.38487995e+00 4.71956760e-01 -1.46400642e+00 1.02069125e-01
1.26568857e-03 -5.45248032e-01 -1.29826799e-01 -4.30858552e-01
-6.55802727e-01 3.51374716e-01 5.52115798e-01 6.79385006e-01
1.12162650e+00 -9.97751355e-02 3.82700592e-01 1.55804321e-01
3.05603951e-01 -9.95946646e-01 -1.38904080e-01 3.49839866e-01
6.81158781e-01 -1.01076913e+00 -2.47879997e-01 -4.61304754e-01
-1.54966503e-01 1.14803886e+00 2.75526106e-01 -4.65565115e-01
8.36160958e-01 4.08152580e-01 -2.62943774e-01 3.57598126e-01
-2.90680200e-01 -1.19396351e-01 5.84038496e-01 5.85015297e-01
1.62179992e-01 1.32421166e-01 -6.57889307e-01 4.96850789e-01
-7.27276504e-02 -9.76763573e-03 4.48294252e-01 5.90179265e-01
-7.12431192e-01 -7.15901911e-01 -5.13862014e-01 5.64033151e-01
-4.18307513e-01 -3.73485722e-02 1.39076725e-01 9.38027680e-01
-1.79059908e-01 7.87939310e-01 1.79852501e-01 -3.93870063e-02
2.00327724e-01 1.42564952e-01 5.34772933e-01 -1.95683256e-01
2.37608090e-01 -7.58079886e-02 -2.19789118e-01 -3.80468488e-01
-1.27355874e-01 -4.21574503e-01 -1.01478612e+00 -5.12181580e-01
-5.43950021e-01 3.85937005e-01 4.25689548e-01 1.00667298e+00
3.23968977e-02 -1.46972239e-01 6.88443959e-01 -5.57512403e-01
-9.08743978e-01 -8.21932197e-01 -9.15607274e-01 9.61718708e-02
4.32427943e-01 -8.89680088e-01 -1.10397506e+00 -2.78722942e-01] | [7.488466262817383, 4.173100471496582] |
4e8ef89b-27dd-4dec-aca7-31032717d375 | lexical-simplification-using-multi-level-and | 2302.01823 | null | https://arxiv.org/abs/2302.01823v1 | https://arxiv.org/pdf/2302.01823v1.pdf | Lexical Simplification using multi level and modular approach | Text Simplification is an ongoing problem in Natural Language Processing, solution to which has varied implications. In conjunction with the TSAR-2022 Workshop @EMNLP2022 Lexical Simplification is the process of reducing the lexical complexity of a text by replacing difficult words with easier to read (or understand) expressions while preserving the original information and meaning. This paper explains the work done by our team "teamPN" for English sub task. We created a modular pipeline which combines modern day transformers based models with traditional NLP methods like paraphrasing and verb sense disambiguation. We created a multi level and modular pipeline where the target text is treated according to its semantics(Part of Speech Tag). Pipeline is multi level as we utilize multiple source models to find potential candidates for replacement, It is modular as we can switch the source models and their weight-age in the final re-ranking. | ['Pawan Kumar Rajpoot', 'Nikita Katyal'] | 2023-02-03 | null | null | null | null | ['lexical-simplification'] | ['natural-language-processing'] | [ 4.47345763e-01 5.91615915e-01 1.25660315e-01 -4.42722857e-01
-7.03549504e-01 -6.15325868e-01 5.53571343e-01 6.81025147e-01
-8.13847899e-01 8.25940013e-01 1.00015450e+00 -2.00176194e-01
-1.78979799e-01 -4.83412325e-01 -2.33823940e-01 8.81438479e-02
6.04728818e-01 7.79318511e-01 1.68553934e-01 -7.97281802e-01
2.82511026e-01 2.57048488e-01 -1.44081426e+00 7.47452557e-01
8.97999883e-01 7.48010864e-03 5.20216763e-01 6.13405824e-01
-7.70677507e-01 7.37960458e-01 -7.67425299e-01 -4.54750955e-01
3.27442318e-01 -2.49990627e-01 -1.34592628e+00 -4.65938389e-01
4.35845226e-01 5.08478343e-01 -1.44960545e-02 1.22271252e+00
5.27633250e-01 4.71374512e-01 2.79372126e-01 -7.67967939e-01
-2.97479868e-01 1.32750189e+00 -2.86262333e-01 3.11222345e-01
8.39709222e-01 -3.94634902e-01 1.20135772e+00 -9.59114313e-01
1.08763194e+00 1.72883475e+00 9.18003619e-01 5.89508891e-01
-1.16860223e+00 -4.17209774e-01 2.88760483e-01 2.65267789e-01
-1.31013739e+00 -4.12057281e-01 1.69882387e-01 -1.75846621e-01
1.87342501e+00 5.73405147e-01 5.70931673e-01 6.72836304e-01
1.99866265e-01 7.83632517e-01 7.62703657e-01 -9.81399894e-01
-2.90614069e-01 1.27932817e-01 6.15369380e-01 4.90173340e-01
3.06788892e-01 -6.24334335e-01 -6.77726626e-01 -1.02968097e-01
-1.30794153e-01 -2.19785303e-01 -8.66772234e-02 3.26237589e-01
-1.19614720e+00 5.68871498e-01 -1.78011522e-01 7.06938505e-01
-4.55561221e-01 -9.80876088e-02 7.16082156e-01 5.62058508e-01
3.30136865e-01 1.07811761e+00 -9.71176088e-01 -1.15922078e-01
-1.00456011e+00 6.74169004e-01 9.64690924e-01 1.16683590e+00
7.08330035e-01 -2.74345875e-01 -6.22586548e-01 1.13992774e+00
2.09929734e-01 2.88845927e-01 9.04400051e-01 -7.86216080e-01
4.65295106e-01 8.79629135e-01 1.17383912e-01 -5.91621697e-01
-6.40873194e-01 -1.89329952e-01 -4.33389038e-01 -1.89642161e-01
-7.77477678e-03 -1.41145438e-01 -1.04210520e+00 1.50693429e+00
2.08314687e-01 -6.09155834e-01 -1.74460430e-02 2.01655805e-01
9.22695816e-01 6.99810028e-01 6.62691295e-01 -3.49994093e-01
1.70820761e+00 -1.06756330e+00 -1.13833213e+00 -5.82278490e-01
9.75792885e-01 -1.40280795e+00 1.24880457e+00 5.53902745e-01
-1.16126275e+00 -3.65082383e-01 -9.61043715e-01 -7.85370469e-01
-9.96362925e-01 -3.95865664e-02 4.59644496e-01 4.40675586e-01
-9.58076239e-01 8.81508410e-01 -2.61722445e-01 -9.02853012e-01
-9.69180390e-02 2.97222674e-01 -4.34541941e-01 1.98998451e-02
-1.58741021e+00 1.65085006e+00 8.52107882e-01 -5.11737585e-01
-1.47445217e-01 -1.02098453e+00 -1.04028559e+00 -5.77627011e-02
3.04747790e-01 -9.96233046e-01 1.49977255e+00 -7.59490967e-01
-1.27875245e+00 1.05661881e+00 -7.03897715e-01 -5.34732640e-01
3.84931177e-01 -6.22466207e-01 -4.61510509e-01 -5.43089390e-01
4.91233855e-01 6.35176778e-01 5.97928107e-01 -6.94769382e-01
-1.17698717e+00 -2.00496525e-01 -1.55411839e-01 5.72267175e-01
-5.77840768e-02 5.91286361e-01 -1.99345574e-01 -1.03864408e+00
1.27157480e-01 -8.01411748e-01 -1.93034768e-01 -7.38175213e-01
-2.62280613e-01 -5.03979981e-01 5.45151114e-01 -1.06345463e+00
1.88763654e+00 -1.73191214e+00 3.94006222e-01 -2.68662758e-02
1.65731192e-01 5.17414093e-01 3.42677310e-02 8.45254242e-01
-4.57022548e-01 5.83466768e-01 -1.67450249e-01 -6.35706544e-01
1.61290094e-01 2.80868202e-01 -2.85361886e-01 -9.42165405e-02
8.99447035e-03 8.36804450e-01 -1.12858963e+00 -5.70574224e-01
3.76045704e-01 1.03012003e-01 -3.21299940e-01 -1.70905530e-01
-1.37730300e-01 -2.11203769e-01 -1.51313126e-01 3.89204651e-01
5.90580821e-01 5.76429486e-01 1.74566269e-01 -2.64533073e-01
-4.02987599e-01 9.19740498e-01 -1.24579024e+00 1.74256051e+00
-6.60357952e-01 6.67090058e-01 4.10936214e-02 -3.95694792e-01
5.55344641e-01 1.98970005e-01 1.11937582e-01 -4.81281042e-01
-3.51746827e-02 1.84068590e-01 1.00820251e-02 -5.51188886e-01
1.12479234e+00 -3.76740217e-01 -2.81485289e-01 1.68333635e-01
2.39800006e-01 -6.26893222e-01 7.19182551e-01 3.67467523e-01
1.34063160e+00 3.94064456e-01 9.91692245e-01 -6.14731014e-01
6.35240912e-01 4.46463436e-01 6.16804421e-01 8.25004697e-01
-2.14360934e-02 3.51654977e-01 2.16510490e-01 -3.69349062e-01
-1.18430281e+00 -7.24108815e-01 -8.39735009e-03 1.49639690e+00
-3.46253812e-01 -1.15972590e+00 -6.97178662e-01 -5.37439466e-01
-5.82829751e-02 1.38117838e+00 -5.10818899e-01 -2.43757553e-02
-8.86793911e-01 -5.31583786e-01 8.79173040e-01 2.96735968e-02
-2.82381307e-02 -1.22193003e+00 -3.12789053e-01 4.76994574e-01
-3.63370150e-01 -9.32410955e-01 -2.75479674e-01 4.68913585e-01
-5.39643526e-01 -8.42780173e-01 -4.60007675e-02 -8.66580307e-01
3.35961908e-01 9.24179256e-02 1.40012860e+00 -1.35601135e-02
-3.53574723e-01 -1.29008487e-01 -6.93870187e-01 -1.09376776e+00
-7.21853614e-01 2.60149240e-01 -4.58442532e-02 -6.89691722e-01
7.47654080e-01 -2.44283602e-01 2.07680743e-02 -4.96776223e-01
-1.04308951e+00 1.74344108e-01 4.03270274e-01 6.32037520e-01
6.28615081e-01 6.28164113e-02 2.01756865e-01 -1.34972346e+00
1.15763044e+00 -2.25825951e-01 4.46690246e-02 4.26353753e-01
-5.78233004e-01 2.69506693e-01 7.18064487e-01 4.13394012e-02
-1.28215897e+00 1.54516295e-01 -4.62986946e-01 2.44513959e-01
-1.03111468e-01 5.32173514e-01 -2.55873531e-01 2.94696629e-01
7.31319666e-01 -1.89641893e-01 -4.36837465e-01 -7.76071489e-01
6.29207373e-01 5.89252949e-01 2.22031191e-01 -4.77490246e-01
7.06534326e-01 1.79655060e-01 -1.64178103e-01 -8.84679437e-01
-1.08820891e+00 -6.27072275e-01 -7.71225035e-01 2.37355426e-01
7.55566299e-01 -7.24826574e-01 -1.11845005e-02 2.47422963e-01
-1.61916041e+00 3.23154628e-02 -6.63101971e-01 1.89830497e-01
1.64038949e-02 5.81525087e-01 -3.85236502e-01 -4.37258244e-01
-6.92208648e-01 -7.10801125e-01 1.07089269e+00 2.18390927e-01
-1.33386481e+00 -9.29667711e-01 8.21789280e-02 1.99069425e-01
3.67184430e-01 -7.24497885e-02 1.34690988e+00 -1.18429983e+00
3.67295593e-01 -1.43102750e-01 5.67380153e-02 3.42704087e-01
2.08134532e-01 2.93975435e-02 -6.50459409e-01 1.74660504e-01
4.15530056e-02 3.10113519e-01 7.99103141e-01 1.55987427e-01
6.32596552e-01 -2.53286362e-01 -4.34729815e-01 2.69152343e-01
1.12227285e+00 -2.96465494e-02 6.90646529e-01 7.39201307e-01
8.06977153e-01 9.66052711e-01 8.07321668e-01 2.46896312e-01
3.72471303e-01 3.22248071e-01 -2.22208783e-01 7.46969804e-02
-4.62583572e-01 -3.65406483e-01 4.30170119e-01 1.07995176e+00
5.18377960e-01 -3.19206804e-01 -1.13774323e+00 6.19597733e-01
-1.97384369e+00 -8.12076271e-01 -6.37092471e-01 2.01853728e+00
1.18091488e+00 2.68980324e-01 -3.10458422e-01 1.42053356e-02
6.73024356e-01 4.37905006e-02 7.00543970e-02 -1.01745617e+00
-4.69989449e-01 5.73214471e-01 6.12589419e-01 1.17153871e+00
-9.83666599e-01 1.69436574e+00 6.51896334e+00 1.19325316e+00
-6.06928468e-01 2.56579787e-01 -5.84550872e-02 -1.84072211e-01
-6.28050685e-01 3.52775991e-01 -1.13983238e+00 1.25727355e-01
7.58305490e-01 -7.51104116e-01 4.33663011e-01 4.39616412e-01
4.63002712e-01 -2.37733930e-01 -1.11000109e+00 7.74475098e-01
2.68907368e-01 -1.08401918e+00 6.39591992e-01 -5.92540145e-01
4.17858183e-01 1.75579965e-01 -7.93516874e-01 7.33245671e-01
5.17602742e-01 -1.07028127e+00 8.20425034e-01 6.00326478e-01
1.17074721e-01 -6.84369683e-01 7.73323596e-01 4.09262240e-01
-1.13930726e+00 4.24984060e-02 -4.33457613e-01 -3.21585029e-01
3.54380548e-01 5.38132131e-01 -7.58828044e-01 6.55616224e-01
7.08940446e-01 5.40246725e-01 -8.87630880e-01 7.83867538e-01
-4.89236921e-01 1.33623794e-01 -3.52967381e-01 -1.74423158e-01
1.24553237e-02 -1.12630308e-01 1.19409311e+00 2.00856423e+00
6.69660838e-03 5.48404828e-02 1.28690764e-01 3.59056622e-01
-5.42685650e-02 6.63190544e-01 -4.43338245e-01 1.15492232e-01
7.06232309e-01 1.24555123e+00 -5.03179729e-01 -7.59470582e-01
-3.98194976e-02 9.88647103e-01 3.44498962e-01 -1.51445508e-01
-4.34341013e-01 -8.88530433e-01 6.69600904e-01 4.62894104e-02
-4.64253239e-02 9.77964997e-02 -6.66517198e-01 -1.20560658e+00
4.27271724e-02 -1.16722476e+00 5.62852085e-01 -8.75845909e-01
-1.20213592e+00 6.21220708e-01 2.45992869e-01 -6.80989981e-01
1.59247629e-02 -5.93792856e-01 -3.84117126e-01 1.22340453e+00
-1.32043111e+00 -1.02503920e+00 3.12269758e-02 2.66345620e-01
1.03263986e+00 -1.73748541e-03 1.08979332e+00 3.63722682e-01
-4.40709472e-01 3.86192918e-01 -1.51461944e-01 -2.35681370e-01
9.32434857e-01 -1.50518763e+00 7.22019315e-01 1.12911081e+00
-2.06144433e-02 1.12240732e+00 1.22490776e+00 -1.03889537e+00
-1.03450119e+00 -1.05447316e+00 2.00604177e+00 -6.68058574e-01
8.78258467e-01 -2.76576221e-01 -9.75991130e-01 6.88950777e-01
6.40750647e-01 -9.32224572e-01 6.93673015e-01 1.72818065e-01
-1.10108256e-01 3.26204374e-02 -1.19520998e+00 9.25383091e-01
1.12502456e+00 -4.51489031e-01 -1.58563268e+00 5.73720515e-01
1.00839746e+00 -4.91027564e-01 -5.58176935e-01 2.27900416e-01
2.47377574e-01 -3.39501053e-01 5.00534058e-01 -8.74905467e-01
4.50208001e-02 -5.23200512e-01 -5.53227216e-02 -1.57232833e+00
-5.37506163e-01 -1.07524705e+00 3.37849915e-01 1.53248572e+00
7.69669116e-01 -2.80447543e-01 2.09531575e-01 6.92303061e-01
-4.69129711e-01 -2.78546721e-01 -8.20225239e-01 -5.03744006e-01
1.35868207e-01 -6.42079473e-01 4.52265829e-01 9.91001725e-01
3.92566711e-01 9.40770686e-01 -8.00899267e-02 -2.99039245e-01
1.68894604e-01 -5.05393326e-01 4.39322919e-01 -1.37219191e+00
2.90293302e-02 -6.37480557e-01 -2.14587256e-01 -7.01380014e-01
3.09223950e-01 -1.15225387e+00 1.61284441e-03 -1.97790170e+00
3.55014391e-02 1.12152740e-01 -3.41316573e-02 7.74313748e-01
-2.71473587e-01 -1.73098281e-01 4.36977118e-01 -4.00879607e-02
-5.78106284e-01 2.02608496e-01 6.40556693e-01 -2.82008201e-02
-3.38321716e-01 -4.11683828e-01 -9.91108716e-01 9.96264577e-01
6.45492733e-01 -1.04988027e+00 -9.28404704e-02 -7.89954245e-01
8.28638494e-01 -6.75941586e-01 -2.08501905e-01 -7.00093031e-01
2.72064328e-01 -1.05457999e-01 1.02851465e-01 -5.35143435e-01
-9.08671971e-03 -6.91949666e-01 9.29878727e-02 3.94209117e-01
-4.66139287e-01 6.00209594e-01 5.03851116e-01 -1.10615008e-02
-1.00635327e-01 -7.99051285e-01 7.82841027e-01 -5.45235693e-01
-8.31918895e-01 -3.45131099e-01 -7.87580252e-01 3.67297709e-01
8.65086377e-01 -2.55031615e-01 -1.57234922e-01 1.51852630e-02
-8.89337718e-01 3.23298305e-01 3.52310658e-01 8.20110321e-01
1.84110075e-01 -7.98806369e-01 -8.92216027e-01 -1.12636171e-01
1.23794280e-01 -6.08395180e-03 -1.47647455e-01 4.43977922e-01
-7.88214684e-01 5.48567176e-01 -8.47213492e-02 2.18596384e-01
-1.53935575e+00 2.53437996e-01 3.85449976e-01 -5.87755084e-01
-4.16949987e-01 8.80556643e-01 -3.84776801e-01 -6.32112324e-01
1.31409347e-01 -5.58136880e-01 -4.84232038e-01 3.83338958e-01
7.07687914e-01 6.05423212e-01 6.66922867e-01 -7.01840043e-01
-6.09876812e-01 2.26158053e-01 -4.66864318e-01 -2.15236232e-01
1.34590197e+00 -5.15973389e-01 -8.11666965e-01 4.39418167e-01
8.51405501e-01 5.13845265e-01 7.55825778e-03 -1.94742039e-01
8.56086433e-01 -9.94021744e-02 -7.62891471e-02 -1.04982495e+00
-8.68969038e-02 4.20486718e-01 1.31754637e-01 1.84015855e-01
7.86490500e-01 -1.34679690e-01 8.14718902e-01 8.53508115e-01
5.84216556e-03 -1.74696326e+00 -6.86851859e-01 1.38189268e+00
1.01162851e+00 -9.09117877e-01 3.64892036e-01 -6.51083052e-01
-6.70201063e-01 1.07992601e+00 4.79725659e-01 -8.05197433e-02
6.09470844e-01 2.54353613e-01 1.10244676e-01 -2.44364932e-01
-8.10875773e-01 -2.93433666e-01 2.46208787e-01 5.97111225e-01
9.97249663e-01 1.53569415e-01 -1.20282447e+00 8.79850566e-01
-7.44235098e-01 -3.74292046e-01 5.24525166e-01 9.86738503e-01
-6.23847604e-01 -1.63716066e+00 -1.69146746e-01 5.34897089e-01
-7.12548137e-01 -8.93465161e-01 -1.00413907e+00 7.55599737e-01
4.74438787e-01 9.70409691e-01 -3.76137704e-01 -1.70846060e-01
9.40850019e-01 4.90781426e-01 2.70186901e-01 -1.42470706e+00
-1.29871702e+00 -2.37766504e-01 6.36699021e-01 -3.41145873e-01
-1.04564719e-01 -8.41089487e-01 -1.51331687e+00 -2.73717791e-01
-1.01378396e-01 2.69573092e-01 6.81037366e-01 1.29460716e+00
4.22045916e-01 6.35802984e-01 -1.17185660e-01 -6.40775442e-01
-4.11951035e-01 -1.23397160e+00 -2.61891723e-01 6.32544279e-01
-4.61213626e-02 -2.20661715e-01 -1.05834424e-01 2.03020021e-01] | [10.919851303100586, 10.398530960083008] |
1c5f2dc6-2ff7-4735-82c0-d08821006406 | distributed-control-design-and-safety | 2303.12610 | null | https://arxiv.org/abs/2303.12610v1 | https://arxiv.org/pdf/2303.12610v1.pdf | Distributed Control Design and Safety Verification for Multi-Agent Systems | We propose distributed iterative algorithms for safe control design and safety verification for networked multi-agent systems. These algorithms rely on distributing a control barrier function (CBF) related quadratic programming (QP) problem. The proposed distributed algorithm addresses infeasibility issues of existing schemes by dynamically allocating auxiliary variables across iterations. The resulting control input is guaranteed to be optimal, and renders the system safe. Furthermore, a truncated algorithm is proposed to facilitate computational implementation, with probabilistically guaranteed constraint satisfication, while generating a Lipschitz continuous control input. We further develop a distributed safety verification algorithm to quantify safety for a multi-agent system by means of CBFs in probability. Both upper and lower bounds on the probability of safety are obtained using the so called scenario approach. Both the scenario sampling and safety verification procedures are fully distributed. The efficacy of our algorithms is demonstrated by an example on multi-robot collision avoidance. | ['Kostas Margellos', 'Antonis Papachristodoulou', 'Han Wang'] | 2023-03-22 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [ 3.28904837e-02 5.02878845e-01 -2.10824728e-01 8.98880735e-02
-1.16306460e+00 -7.46035099e-01 4.34089750e-01 5.77977180e-01
-4.84438092e-01 1.13829708e+00 -3.55106533e-01 -5.09048164e-01
-5.74068308e-01 -9.35819983e-01 -7.75887072e-01 -9.81824815e-01
-6.21252239e-01 4.20244366e-01 1.85048178e-01 -2.75053352e-01
6.51619881e-02 3.28384936e-01 -1.16922712e+00 -4.59897697e-01
1.05061734e+00 9.21468556e-01 -1.22081913e-01 7.40801871e-01
8.51825237e-01 5.62876701e-01 -5.46812415e-01 3.86650637e-02
5.18700898e-01 -3.11818361e-01 -8.57824743e-01 1.08495884e-01
-1.24245390e-01 -6.77091181e-01 2.46897221e-01 1.47675896e+00
1.97471455e-01 3.73488784e-01 5.04867554e-01 -2.29959011e+00
-8.17797855e-02 7.40104437e-01 -6.16324246e-01 -4.18321401e-01
2.30015650e-01 2.33860672e-01 8.54406774e-01 -2.17082068e-01
3.69605035e-01 1.28912044e+00 5.01908541e-01 5.52267432e-01
-1.22424591e+00 -4.70320135e-01 2.15008393e-01 -1.09981097e-01
-1.65278018e+00 1.09227542e-02 2.95831829e-01 -3.25976998e-01
6.21681273e-01 4.80256528e-01 4.37495440e-01 5.01013637e-01
6.59347177e-01 5.50097942e-01 9.11043108e-01 -3.52316320e-01
8.17632079e-01 2.33217314e-01 -2.17027348e-02 6.40943587e-01
6.37396932e-01 4.99473840e-01 1.46542475e-01 -8.30061436e-01
4.38848108e-01 -1.18186459e-01 -2.03204647e-01 -5.13072491e-01
-1.10257363e+00 1.05824685e+00 5.89909637e-03 -3.80453259e-01
-3.89478534e-01 4.07091230e-01 6.24732435e-01 2.90073127e-01
2.81588912e-01 5.65023795e-02 -4.62388247e-01 1.84462711e-01
-3.57483715e-01 7.83148408e-01 1.04529059e+00 1.18214571e+00
5.40295184e-01 6.77082837e-02 -1.19512655e-01 2.20239371e-01
6.22701287e-01 7.28801012e-01 -5.52226305e-01 -1.38835490e+00
4.91738588e-01 1.16818160e-01 6.83650255e-01 -9.56525743e-01
-1.77008525e-01 1.23829067e-01 -7.33446360e-01 8.56427431e-01
3.51594031e-01 -6.61581993e-01 -3.03867459e-01 2.05690384e+00
8.09319258e-01 -1.62412208e-02 1.89741448e-01 8.80676270e-01
-6.53688908e-01 1.00526297e+00 1.32237315e-01 -7.71067679e-01
9.42572653e-01 -5.47985554e-01 -7.40643799e-01 3.30226064e-01
7.27463901e-01 -8.69200155e-02 5.35036206e-01 5.65848410e-01
-1.12565660e+00 2.44499549e-01 -9.99817669e-01 5.15886724e-01
4.85566780e-02 -3.35139692e-01 2.77817279e-01 8.33308220e-01
-1.12690568e+00 2.26354510e-01 -9.54250336e-01 -1.17366143e-01
1.66786075e-01 6.50959551e-01 -1.09539032e-01 3.35280299e-01
-1.05552185e+00 8.96965504e-01 3.96123111e-01 2.14400142e-01
-1.46137738e+00 -9.16877985e-01 -1.06826317e+00 -6.91973865e-02
7.13371277e-01 -6.71682954e-01 1.37053204e+00 -4.83877897e-01
-1.72889543e+00 1.05593884e-02 4.11411136e-01 -5.90663016e-01
7.04525650e-01 1.19202472e-01 1.50595516e-01 3.43099684e-01
1.70699671e-01 3.10828388e-01 7.19506621e-01 -1.53886318e+00
-9.75309253e-01 -4.48339507e-02 5.29598236e-01 5.58438860e-02
-2.17612073e-01 1.65270254e-01 4.84755844e-01 -3.15823466e-01
-6.89705729e-01 -9.65585232e-01 -1.07427967e+00 3.10994655e-01
-2.63868660e-01 -9.96282324e-02 6.91196740e-01 -3.90150964e-01
1.03104818e+00 -1.91890085e+00 2.01819152e-01 6.88911617e-01
-8.33531376e-03 -1.60654292e-01 3.92347910e-02 6.25846922e-01
4.18715209e-01 7.27838501e-02 -4.75686491e-01 -1.05407424e-01
4.17503566e-01 1.55688345e-01 -4.66704518e-01 1.11814272e+00
1.45828739e-01 1.65089682e-01 -8.56068015e-01 -4.50177163e-01
2.88881004e-01 5.36098592e-02 -8.63995492e-01 1.30962655e-01
-4.41450149e-01 1.77327871e-01 -6.60882473e-01 2.15375513e-01
9.17494118e-01 4.03779477e-01 3.94968390e-01 2.12315589e-01
-4.41465080e-01 -3.17727149e-01 -1.54373538e+00 1.03929782e+00
-6.22401059e-01 -1.97947115e-01 9.40496862e-01 -8.89526188e-01
4.44494933e-01 5.11351347e-01 5.93976021e-01 -7.82321114e-03
5.85274994e-01 2.04714406e-02 -3.49753797e-01 -1.72403723e-01
5.40407121e-01 -2.21972674e-01 -6.36092305e-01 5.14518797e-01
-5.29930174e-01 -2.95874625e-01 -3.64196301e-02 3.58752340e-01
9.77203012e-01 -1.91922024e-01 5.01993477e-01 -8.68618250e-01
7.93816745e-01 3.44722271e-01 7.16465473e-01 5.97093821e-01
-5.65839350e-01 -2.98831671e-01 9.09774542e-01 8.43567327e-02
-1.00955892e+00 -7.68764019e-01 8.25491846e-02 7.13236272e-01
4.69625115e-01 -1.80174157e-01 -7.45157480e-01 -5.37463367e-01
2.02771276e-01 7.79964864e-01 -6.44670129e-01 -2.48061374e-01
-3.66744220e-01 -4.69099164e-01 5.21138370e-01 2.91824073e-01
4.70919684e-02 -4.63288248e-01 -9.16750610e-01 3.43236029e-01
2.58490354e-01 -7.14311421e-01 -6.81103945e-01 -7.30807544e-04
-4.28926468e-01 -1.17807806e+00 -2.27185771e-01 -7.10000038e-01
1.00183332e+00 7.14998618e-02 2.45640889e-01 -6.85673580e-02
-5.32739088e-02 6.16069496e-01 1.01798587e-01 -2.72510171e-01
-7.38814473e-01 -3.89948577e-01 3.62423003e-01 -9.42901298e-02
-5.49159825e-01 -1.23636045e-01 -3.74890119e-01 3.60714972e-01
-1.07285094e+00 -3.15605134e-01 -1.36333657e-02 7.71075368e-01
6.26149178e-01 3.96500349e-01 8.87563288e-01 -4.11410451e-01
9.18622911e-01 -7.10970461e-01 -1.55276775e+00 2.72933155e-01
-5.35918176e-01 7.23603601e-03 9.02440310e-01 -4.29778039e-01
-8.06119442e-01 2.54910529e-01 3.45987499e-01 -3.20635438e-01
4.81160805e-02 4.46040928e-01 -2.32428402e-01 -2.17340261e-01
2.75404662e-01 -1.84518307e-01 5.70766747e-01 2.92926341e-01
4.00797874e-01 3.78849596e-01 1.77904889e-01 -1.19039857e+00
8.57478201e-01 5.46020091e-01 3.76124084e-01 -4.68725920e-01
-1.12992205e-01 -2.98934299e-02 1.03255384e-01 -2.82558799e-01
3.49344403e-01 -6.79111779e-01 -1.84943378e+00 2.41737664e-01
-1.07783222e+00 -5.61671734e-01 -2.10685939e-01 2.53353387e-01
-9.00039732e-01 5.42420149e-01 -5.55949628e-01 -1.63002622e+00
-1.36466533e-01 -1.24914467e+00 9.51496601e-01 -4.94291224e-02
2.82540694e-02 -1.04256320e+00 4.44342792e-01 -2.57044077e-01
3.15210432e-01 7.33726799e-01 5.99296093e-01 -2.73037285e-01
-5.24584293e-01 -1.89425841e-01 6.91998675e-02 1.86533347e-01
-9.69250649e-02 1.79237455e-01 -3.05162609e-01 -7.24104285e-01
5.01077808e-02 -3.42471212e-01 4.25302982e-03 3.26996237e-01
8.69316399e-01 -1.12908268e+00 -3.18148583e-01 2.24282257e-02
1.84168029e+00 1.81434140e-01 1.05391510e-01 3.64152223e-01
2.11240560e-01 6.78877354e-01 1.18450940e+00 1.07293773e+00
5.43092549e-01 5.08543015e-01 8.88261676e-01 3.80158037e-01
7.80643940e-01 1.54841766e-01 8.02720547e-01 1.76309943e-01
2.09176213e-01 -1.61543220e-01 -8.80914629e-01 8.13817739e-01
-2.15391636e+00 -7.79490471e-01 -1.11338757e-01 2.46817708e+00
9.57003474e-01 -1.89395666e-01 4.76503760e-01 1.16674729e-01
9.12197173e-01 -4.44739372e-01 -4.12110835e-01 -7.36310303e-01
4.03937727e-01 -1.30324930e-01 8.45106304e-01 1.12565303e+00
-1.04725516e+00 3.88541579e-01 6.47282314e+00 8.62140000e-01
-6.27055526e-01 1.94775879e-01 5.12384653e-01 1.16900958e-01
-2.16340631e-01 2.06258312e-01 -5.78333557e-01 1.89237460e-01
1.19765115e+00 -9.13040221e-01 5.09756267e-01 1.14502645e+00
7.86509335e-01 -3.19119126e-01 -9.13619161e-01 3.44010115e-01
-4.97870833e-01 -9.83900189e-01 -3.42793792e-01 8.24322104e-02
7.85237849e-01 -5.52188873e-01 1.13064334e-01 4.89868522e-02
8.01716745e-01 -6.45377815e-01 1.06086898e+00 -1.00023642e-01
5.82837820e-01 -1.44312739e+00 6.66309953e-01 3.27180713e-01
-1.26876366e+00 -4.20558661e-01 -1.81309879e-02 -1.85935766e-01
4.84985650e-01 3.08916390e-01 -6.21597528e-01 7.53619969e-01
2.89129108e-01 1.29964888e-01 2.52753913e-01 1.03041089e+00
-3.25209796e-02 1.45103708e-01 -7.29893506e-01 -1.91486567e-01
3.73263747e-01 -3.49300414e-01 8.07574630e-01 8.64091992e-01
-1.21078256e-03 8.31659511e-02 6.85821891e-01 8.65769207e-01
4.24529761e-01 -1.35504544e-01 -6.07610881e-01 4.05814737e-01
6.24134362e-01 1.22444510e+00 -5.79259157e-01 -1.14866085e-01
-4.84232232e-02 4.39573497e-01 -1.56248640e-02 3.39347214e-01
-1.38234234e+00 -6.37859702e-01 9.06841516e-01 -4.32154655e-01
9.51577425e-02 -4.38917160e-01 -3.07621837e-01 -4.30769950e-01
1.58624668e-02 -8.04258883e-01 5.23818135e-01 -7.95283318e-02
-1.30393386e+00 2.11803973e-01 4.26506251e-01 -1.21489716e+00
-4.61569548e-01 -3.41301739e-01 -6.60513878e-01 5.76081276e-01
-1.27732658e+00 -9.83624876e-01 2.48624876e-01 8.00255656e-01
8.17601755e-02 3.88203442e-01 6.23521268e-01 3.10138077e-01
-8.87077868e-01 5.53394496e-01 5.72879100e-03 -4.21213299e-01
3.68048251e-01 -1.20785284e+00 -4.71903890e-01 1.03631675e+00
-1.24466610e+00 5.98340571e-01 1.08103359e+00 -6.40196323e-01
-1.88618410e+00 -1.61292541e+00 1.09843217e-01 -1.27600744e-01
1.21348870e+00 -2.41337568e-01 -4.44416672e-01 6.13436460e-01
2.04541773e-01 -1.43307522e-01 1.75264016e-01 -4.16830659e-01
3.62625010e-02 -1.20273672e-01 -1.72289050e+00 6.65153563e-01
5.16763687e-01 -6.13024086e-02 -2.31695786e-01 4.31277812e-01
9.13417459e-01 -1.90923780e-01 -1.00285482e+00 2.18962297e-01
5.39233796e-02 -1.42093956e-01 6.11061573e-01 -4.43889946e-01
-8.02403763e-02 -7.41404891e-01 -2.57100523e-01 -1.13174951e+00
-3.43245640e-03 -1.34595013e+00 -1.56720385e-01 1.07400012e+00
3.97480279e-01 -8.90501082e-01 3.53258967e-01 9.20346797e-01
-1.43516749e-01 -5.20088077e-01 -1.28795362e+00 -1.26316476e+00
3.01142812e-01 -3.80819827e-01 2.86398560e-01 7.78464198e-01
7.30946124e-01 -2.81783938e-01 -2.99392521e-01 1.03619671e+00
1.10262692e+00 -1.80936158e-01 4.12802190e-01 -4.81659949e-01
-1.82670996e-01 -3.14941347e-01 -1.23153307e-01 -2.32356384e-01
4.09730524e-01 -4.49306041e-01 5.49326062e-01 -1.05490196e+00
5.43308519e-02 -6.39371216e-01 1.25405723e-02 6.64447367e-01
8.88046399e-02 -2.92781889e-01 1.46732688e-01 -2.82629162e-01
-9.89807069e-01 8.16695333e-01 8.32309663e-01 -1.19287096e-01
-3.06678534e-01 -3.34817432e-02 -3.75515312e-01 4.77156788e-01
9.10410285e-01 -4.99939978e-01 -6.12697601e-01 -1.15828954e-01
1.94822997e-01 6.41750157e-01 5.89098871e-01 -6.72022939e-01
4.07209516e-01 -9.37660336e-01 -8.05765331e-01 -5.03558099e-01
1.05652846e-02 -1.25660682e+00 4.71718460e-01 1.15183914e+00
-3.65310669e-01 1.59506593e-02 2.38697112e-01 9.04418409e-01
5.17841093e-02 -1.20834030e-01 9.34940815e-01 5.09867549e-01
-2.56783724e-01 4.05500531e-01 -8.90411735e-01 -2.18727797e-01
1.79115248e+00 3.39120179e-01 -2.42226720e-01 -3.80864233e-01
-5.02223909e-01 9.36341345e-01 3.32120299e-01 -4.06503938e-02
5.45791030e-01 -1.18351877e+00 -6.57239377e-01 -1.79674476e-01
-2.19820932e-01 -2.67130993e-02 2.31275678e-01 9.81615484e-01
-5.58712900e-01 2.40100607e-01 -1.11443408e-01 -3.93645912e-01
-9.90534067e-01 8.90115917e-01 4.17188585e-01 -4.09144834e-02
-2.87474483e-01 3.07746589e-01 9.27312076e-02 -2.36728087e-01
2.42193386e-01 -5.33384442e-01 2.88562506e-01 -3.72477174e-01
6.19307458e-01 8.05522144e-01 -3.41299415e-01 -1.48596421e-01
-6.43683255e-01 2.50584215e-01 1.69782132e-01 -6.18576705e-01
1.24977529e+00 -4.51070398e-01 -5.13261914e-01 -1.46135688e-01
8.69197369e-01 -4.07904126e-02 -1.38860917e+00 3.65020841e-01
-2.00826168e-01 -2.18861073e-01 -5.03881127e-02 -3.42481911e-01
-8.35898042e-01 9.36350077e-02 1.79499969e-01 4.97937262e-01
1.08735168e+00 -3.46049041e-01 4.90884662e-01 3.19781542e-01
9.55159307e-01 -1.01142585e+00 -4.51307595e-01 5.80721200e-01
8.20649385e-01 -8.18498313e-01 -3.71083990e-03 -6.50660515e-01
-6.06288791e-01 8.92612398e-01 6.24182224e-01 -5.65184891e-01
6.05917335e-01 7.90529132e-01 -4.49307531e-01 3.14588487e-01
-9.00826395e-01 2.05012247e-01 -5.29778838e-01 6.84084356e-01
-5.10613739e-01 1.73602298e-01 -7.23134398e-01 8.46167862e-01
2.79609352e-01 -1.22074328e-01 1.05066288e+00 1.24519277e+00
-5.01936674e-01 -8.68484437e-01 -6.97302938e-01 -2.90785223e-01
-3.46727073e-01 3.27780157e-01 1.93546236e-01 8.29579473e-01
-2.80111253e-01 1.24622750e+00 -2.36540064e-01 -4.42871451e-02
7.84341842e-02 -4.90738720e-01 2.57897586e-01 -4.14645910e-01
-5.61524272e-01 -1.01297386e-01 2.92429000e-01 -6.36831045e-01
-2.44136199e-01 -4.58747238e-01 -1.48418152e+00 -5.94977379e-01
-5.39517283e-01 4.66968209e-01 4.78809208e-01 7.44959474e-01
2.93854535e-01 3.06728601e-01 1.10133958e+00 -8.17934930e-01
-1.32029426e+00 -2.57824570e-01 -5.32425463e-01 -3.26797187e-01
4.71537918e-01 -5.76912999e-01 -6.46557927e-01 -5.21774776e-02] | [4.76751708984375, 2.244967222213745] |
78347e23-7289-4b96-8b13-16e9ba717400 | transformer-based-approach-towards-music | 2101.02051 | null | https://arxiv.org/abs/2101.02051v1 | https://arxiv.org/pdf/2101.02051v1.pdf | Transformer-based approach towards music emotion recognition from lyrics | The task of identifying emotions from a given music track has been an active pursuit in the Music Information Retrieval (MIR) community for years. Music emotion recognition has typically relied on acoustic features, social tags, and other metadata to identify and classify music emotions. The role of lyrics in music emotion recognition remains under-appreciated in spite of several studies reporting superior performance of music emotion classifiers based on features extracted from lyrics. In this study, we use the transformer-based approach model using XLNet as the base architecture which, till date, has not been used to identify emotional connotations of music based on lyrics. Our proposed approach outperforms existing methods for multiple datasets. We used a robust methodology to enhance web-crawlers' accuracy for extracting lyrics. This study has important implications in improving applications involved in playlist generation of music based on emotions in addition to improving music recommendation systems. | ['Vinoo Alluri', 'Ramaguru Guru Ravi Shanker', 'Yudhik Agrawal'] | 2021-01-06 | null | null | null | null | ['music-emotion-recognition'] | ['music'] | [ 2.15861097e-01 -4.69575763e-01 -1.53781297e-02 -8.46653581e-02
-7.90094316e-01 -8.40649486e-01 3.18279237e-01 1.73263595e-01
-2.67305881e-01 4.19755608e-01 5.52580476e-01 3.68476361e-01
-6.32372618e-01 -4.90630358e-01 4.67771944e-03 -6.82173967e-01
1.17179103e-01 1.85608506e-01 -1.30223528e-01 -1.90371200e-01
7.93608308e-01 3.54509056e-01 -2.30007100e+00 5.73103964e-01
1.27988249e-01 1.21436536e+00 4.03645709e-02 8.59644830e-01
-4.16718185e-01 7.80059755e-01 -7.06543863e-01 -2.72576451e-01
8.06547776e-02 -7.46210754e-01 -6.11721456e-01 -2.74811000e-01
7.01644868e-02 5.04684746e-01 3.23789418e-01 8.29041839e-01
9.57996130e-01 1.94456309e-01 6.80303276e-01 -9.62461889e-01
-2.09573749e-02 9.60914433e-01 -2.13881955e-01 -2.74813436e-02
6.59030914e-01 -5.10142505e-01 1.41750348e+00 -5.67992806e-01
5.49477994e-01 6.16761148e-01 9.38887298e-01 2.43566632e-01
-7.39959478e-01 -1.05921388e+00 -5.76715112e-01 5.28091609e-01
-1.29332268e+00 -3.40837359e-01 1.27727938e+00 -5.33713400e-01
8.79999638e-01 6.04430676e-01 1.08129966e+00 8.42529953e-01
-1.30007416e-01 6.69870019e-01 1.39917171e+00 -8.65332127e-01
1.77875102e-01 3.45703065e-01 5.32596819e-02 1.20201156e-01
-2.73643255e-01 -2.06673846e-01 -1.21208441e+00 -2.84071356e-01
2.69799501e-01 -4.10483241e-01 5.92071600e-02 4.11881618e-02
-9.75889444e-01 7.27159262e-01 -2.42681041e-01 9.51490164e-01
-7.88176239e-01 1.04996540e-01 8.15656841e-01 5.86856008e-01
5.66652775e-01 8.66793275e-01 -5.05235136e-01 -1.02091992e+00
-1.30413747e+00 2.06107333e-01 8.29169750e-01 2.90668517e-01
2.95438975e-01 5.94121702e-02 3.14148873e-01 1.32135892e+00
3.42324972e-01 1.79445937e-01 8.43714118e-01 -8.05438221e-01
-3.75257164e-01 7.63975739e-01 -3.24229300e-01 -1.11499977e+00
-2.69598722e-01 -5.14155686e-01 -2.89574653e-01 1.88231692e-01
2.05611438e-01 1.35903195e-01 -1.97851914e-03 1.40409458e+00
1.79294392e-01 3.96749169e-01 3.87121476e-02 6.70499325e-01
8.94974113e-01 3.86571050e-01 -8.48123729e-02 -2.76391208e-01
1.36264050e+00 -3.36143583e-01 -7.70321071e-01 5.95161319e-01
4.11916167e-01 -1.53661680e+00 1.12757111e+00 1.25757062e+00
-9.34728980e-01 -5.39171815e-01 -1.15939629e+00 3.74507397e-01
-5.89793026e-01 2.73638546e-01 6.95207417e-01 1.10945201e+00
-6.55839324e-01 7.67046750e-01 -2.77048558e-01 -6.27450049e-01
7.46348500e-02 5.20082474e-01 -1.30582079e-01 6.58349335e-01
-9.67394054e-01 5.56172311e-01 1.40275899e-02 -2.21264318e-01
-3.39822233e-01 -4.35970485e-01 -2.31176361e-01 -3.10772628e-01
2.22042799e-02 -1.48583204e-01 9.57102299e-01 -1.47600651e+00
-1.83809876e+00 9.59415615e-01 1.10054359e-01 -2.51198620e-01
-2.11500630e-01 -2.64781028e-01 -9.20582652e-01 1.60776734e-01
3.09874467e-03 2.16770872e-01 7.03211069e-01 -9.51297104e-01
-7.00068772e-01 -4.67027038e-01 -2.93212235e-01 2.33729482e-01
-6.81458652e-01 6.58374012e-01 -2.51097977e-01 -9.46362376e-01
3.06482404e-01 -1.10737097e+00 3.82775754e-01 -7.60296345e-01
-1.25539079e-01 -4.01074708e-01 6.69287086e-01 -5.34217775e-01
1.66678500e+00 -2.12561965e+00 -1.53753430e-01 4.74333882e-01
-3.01429629e-01 8.91016647e-02 1.73222497e-01 7.27084696e-01
-2.57160157e-01 -4.13922481e-02 3.31348598e-01 2.32537970e-01
2.72098511e-01 -1.86046600e-01 -4.06970590e-01 2.10757956e-01
-3.95171463e-01 4.72024471e-01 -5.17102182e-01 -6.46134853e-01
2.22022787e-01 7.17169821e-01 -3.65218490e-01 1.01086095e-01
1.25163227e-01 2.46789768e-01 -4.74293411e-01 6.99216604e-01
1.17711402e-01 4.04963672e-01 1.26224514e-02 -4.68196005e-01
-4.18817520e-01 5.44659972e-01 -1.44241297e+00 2.04303765e+00
-5.27330697e-01 7.41096973e-01 -2.76220232e-01 -6.93153501e-01
1.28510153e+00 8.79516125e-01 1.09859586e+00 -2.46873021e-01
3.26594710e-01 4.06692594e-01 8.52817520e-02 -7.29612947e-01
6.93595052e-01 -4.11497355e-01 -1.46214738e-01 7.46705532e-01
4.29885909e-02 1.85285527e-02 1.24336772e-01 -2.79990822e-01
9.74522054e-01 5.35605133e-01 3.38540375e-01 -1.43907174e-01
7.91829824e-01 4.65155803e-02 2.80254185e-01 4.54091519e-01
-3.56249399e-02 4.41354990e-01 9.24771279e-02 -1.88780874e-01
-7.82023728e-01 -5.92860699e-01 -2.71376520e-01 1.40694845e+00
-3.46631974e-01 -9.75317597e-01 -5.40202975e-01 -3.47929001e-01
-3.68323296e-01 4.40990865e-01 -3.02497149e-01 9.50465426e-02
-1.50689274e-01 -5.92571497e-01 1.11369157e+00 -1.95582621e-02
1.67813748e-02 -1.64803243e+00 -6.69772506e-01 5.05046368e-01
-2.48634815e-01 -4.32385236e-01 8.80127996e-02 4.33526337e-01
-6.94329381e-01 -9.93867934e-01 -4.57279384e-01 -6.51774406e-01
-2.91910231e-01 -1.76973090e-01 1.17451108e+00 -8.45362693e-02
-5.20746708e-01 7.05043972e-01 -8.80431950e-01 -9.08385873e-01
-2.83440322e-01 3.48129034e-01 8.74844287e-03 1.24174252e-01
9.69416738e-01 -9.51962352e-01 -3.93874794e-01 1.62914336e-01
-8.19878638e-01 -4.72704500e-01 3.21519524e-01 1.89879701e-01
4.51774776e-01 3.71348828e-01 1.06203294e+00 -6.55203640e-01
8.57597351e-01 -2.81111181e-01 1.43189341e-01 -1.64141998e-01
-6.79580867e-01 -3.51993799e-01 1.59010783e-01 -4.72442865e-01
-7.49412358e-01 7.59229958e-02 -3.21931511e-01 -3.03068478e-02
-4.86030549e-01 4.76044029e-01 7.86561053e-03 -1.30268216e-01
5.44230402e-01 -1.16336316e-01 -1.89015225e-01 -7.23944604e-01
1.47954851e-01 1.26046932e+00 4.76046950e-01 -6.74801350e-01
3.19620371e-01 3.34689081e-01 -3.71170044e-02 -9.70169902e-01
-7.61413336e-01 -1.19035602e+00 -5.41367888e-01 -9.39312637e-01
9.46372628e-01 -7.73889899e-01 -9.58503127e-01 2.16193989e-01
-5.97516656e-01 4.51140195e-01 -3.80041778e-01 9.07751918e-01
-7.10222840e-01 2.23846123e-01 -5.03638864e-01 -1.30819333e+00
-5.98350644e-01 -6.97575510e-01 9.19812024e-01 6.47047684e-02
-9.18728292e-01 -5.82307100e-01 6.02253377e-01 5.01612127e-01
2.86204726e-01 2.15898395e-01 7.18056798e-01 -6.19572520e-01
1.76021438e-02 -2.75884598e-01 3.73121917e-01 2.43087351e-01
1.95192829e-01 -2.97640041e-02 -1.38604188e+00 3.49412173e-01
3.32078971e-02 -6.05622470e-01 5.21159530e-01 2.35921025e-01
8.72313023e-01 2.43245497e-01 3.12276840e-01 2.10929871e-01
1.38343823e+00 2.92682081e-01 5.83237171e-01 8.41214895e-01
3.97624642e-01 6.64931357e-01 8.94717693e-01 7.57584095e-01
-1.13795541e-01 7.06302285e-01 1.63965687e-01 4.07465965e-01
-1.78898320e-01 -6.56047603e-04 6.89117670e-01 1.31448519e+00
-2.49166772e-01 1.12582944e-01 -5.94600618e-01 4.52443451e-01
-1.71246088e+00 -1.23346102e+00 -2.83857793e-01 2.00426459e+00
7.28248119e-01 -6.94581494e-02 3.73650163e-01 1.01096344e+00
5.76030910e-01 1.05007291e-02 -1.69498786e-01 -7.01508462e-01
-1.83987290e-01 6.89766943e-01 8.59820247e-02 -1.42871842e-01
-1.00263882e+00 8.73012066e-01 5.78105640e+00 9.55011308e-01
-1.15367270e+00 6.68622181e-02 -2.67463177e-01 -5.02288997e-01
9.97344870e-03 6.06101891e-03 -5.10548770e-01 2.20438853e-01
9.31090176e-01 2.68810660e-01 6.23181820e-01 7.95573175e-01
4.15407479e-01 -1.80985162e-03 -5.93755782e-01 1.23309505e+00
5.97611368e-01 -9.14799213e-01 -2.82303691e-01 7.81976152e-03
5.99865615e-01 1.59855634e-02 -1.76585555e-01 2.09005579e-01
-3.01901370e-01 -6.97259665e-01 6.47381485e-01 6.85511410e-01
4.06438321e-01 -1.12747121e+00 6.79340959e-01 -6.00124970e-02
-1.23791492e+00 1.18672602e-01 -5.06250039e-02 -3.08889508e-01
2.42322087e-02 3.59257728e-01 -9.50941026e-01 4.18411344e-01
9.22029912e-01 9.59064543e-01 -3.94218832e-01 1.05024242e+00
2.04431042e-01 1.13052058e+00 -2.01620519e-01 -1.39822423e-01
1.99520569e-02 -3.50800335e-01 6.20993614e-01 1.22686672e+00
6.85613275e-01 -4.10891533e-01 -2.08104119e-01 4.62449610e-01
3.72071058e-01 1.11569536e+00 -4.57375079e-01 -4.87691969e-01
1.87314793e-01 1.55746257e+00 -9.77085471e-01 -6.55124262e-02
-1.90611511e-01 7.22648025e-01 -4.10297930e-01 -1.62283748e-01
-5.59451282e-01 -4.15873140e-01 6.56451643e-01 2.66185794e-02
1.38947204e-01 1.16881929e-01 -2.85327584e-01 -7.25876272e-01
-3.39816660e-01 -1.20924604e+00 5.17413318e-01 -1.02934790e+00
-1.34954226e+00 7.48549759e-01 -5.79311967e-01 -1.55986881e+00
-3.69143099e-01 -3.83502424e-01 -4.32059795e-01 5.88876903e-01
-8.57217371e-01 -1.20320284e+00 -8.02161768e-02 6.27761602e-01
4.97156799e-01 -6.45682156e-01 1.30361545e+00 5.23480475e-01
-7.41636008e-02 1.64287522e-01 1.12204056e-03 -2.28674039e-01
1.23277235e+00 -1.22023892e+00 -7.38306463e-01 1.51023254e-01
9.70343292e-01 3.55234355e-01 8.36527526e-01 -5.28744519e-01
-1.25198352e+00 -4.97300923e-01 1.16758358e+00 -4.19852704e-01
7.95909107e-01 3.11912745e-02 -3.55235487e-01 3.24299447e-02
5.15372276e-01 -9.51591134e-01 1.68579638e+00 7.85953045e-01
-2.83861399e-01 -1.90518290e-01 -7.09894776e-01 3.20915371e-01
5.76417387e-01 -8.93986762e-01 -6.22088730e-01 -8.45250562e-02
-1.82961211e-01 3.68802100e-01 -1.23906350e+00 2.77249992e-01
1.10117853e+00 -1.11891329e+00 8.41094911e-01 -3.52147311e-01
2.73723722e-01 -5.41468918e-01 -3.84033859e-01 -9.79993403e-01
-1.66450639e-03 -4.85902160e-01 4.77891177e-01 1.62231052e+00
2.88221598e-01 1.24985009e-01 8.67499828e-01 -9.45162997e-02
5.56138977e-02 -2.20728084e-01 -5.80320418e-01 -3.66211027e-01
-5.80943584e-01 -1.30960596e+00 5.66522002e-01 1.15482104e+00
3.17980975e-01 6.15990579e-01 -6.39969945e-01 -4.29632217e-01
2.39332989e-01 5.03609478e-01 8.93114924e-01 -1.49992967e+00
-5.40964782e-01 -7.73861766e-01 -7.96284139e-01 4.48068529e-02
2.50939071e-01 -1.24486375e+00 -2.31963277e-01 -1.07585680e+00
1.98237792e-01 -3.56737256e-01 -8.98463190e-01 1.48198456e-01
4.07957494e-01 1.09828269e+00 1.94003418e-01 2.61356533e-01
-6.70423925e-01 6.40410036e-02 6.84173644e-01 1.78868935e-01
-3.61069888e-01 1.33736953e-01 -7.28324413e-01 1.01812553e+00
8.97314310e-01 -7.64706314e-01 -4.03800011e-01 5.01968980e-01
1.11033511e+00 -3.20434064e-01 -7.28116632e-02 -1.32373369e+00
6.52626678e-02 2.55013853e-01 2.19890311e-01 -7.74677277e-01
4.66993958e-01 -8.29217970e-01 5.85915565e-01 -1.72946472e-02
-6.32120788e-01 1.99693695e-01 -3.17388475e-02 2.23374858e-01
-5.14139950e-01 -5.70272624e-01 3.24000031e-01 -4.99853417e-02
-5.54802418e-01 -2.87793815e-01 -6.68994009e-01 -2.65542954e-01
5.53626835e-01 -4.46707010e-01 4.61654693e-01 -5.35871327e-01
-1.02065885e+00 -8.19503963e-01 2.75736362e-01 6.26349449e-01
1.76206753e-01 -1.51061475e+00 -5.00552475e-01 -8.07072148e-02
3.81720573e-01 -1.11669767e+00 2.17928901e-01 7.99864888e-01
-4.14786965e-01 1.53679013e-01 -3.56180400e-01 -3.28223079e-01
-1.95606780e+00 1.98395520e-01 -2.50778764e-01 -7.14119226e-02
-3.11246753e-01 5.65500498e-01 -5.00714719e-01 -2.81602293e-01
4.87613469e-01 2.28703618e-01 -8.91181469e-01 6.48095191e-01
4.43522155e-01 4.63046461e-01 2.18116134e-01 -9.14882660e-01
-1.71885535e-01 7.83547938e-01 4.55969423e-01 -5.78484893e-01
1.51876140e+00 -1.82155892e-01 -3.43071282e-01 1.32219660e+00
8.77566040e-01 6.31303489e-01 -1.21396147e-01 1.95765004e-01
4.53169703e-01 -2.81980693e-01 3.97555530e-01 -1.11130655e+00
-7.76669383e-01 5.87386250e-01 9.42676187e-01 4.33606684e-01
1.32915616e+00 -1.85732350e-01 7.03363895e-01 1.90955639e-01
1.79110110e-01 -1.63145375e+00 9.85463150e-03 2.97840863e-01
6.98993862e-01 -7.82857835e-01 -1.46373346e-01 8.64423811e-03
-6.77768528e-01 1.21088457e+00 -5.27921803e-02 -2.25969572e-02
9.88723278e-01 2.14720979e-01 5.08560777e-01 -3.14209163e-01
-6.69210374e-01 -5.91271162e-01 6.00559711e-01 2.26044610e-01
1.32139766e+00 1.48591504e-01 -8.82495880e-01 9.14592564e-01
-6.94583595e-01 2.26745397e-01 1.81811050e-01 6.74213350e-01
-4.79701757e-01 -1.84354174e+00 -6.27703965e-01 4.78423566e-01
-1.22132075e+00 -2.66494662e-01 -8.70281279e-01 4.31846112e-01
6.26117170e-01 1.25603640e+00 -3.42032611e-01 -7.20165312e-01
5.60382660e-03 5.43450713e-01 5.53796351e-01 -3.31069767e-01
-1.41765368e+00 6.97997868e-01 4.18086886e-01 -3.53178948e-01
-1.21261930e+00 -9.64234889e-01 -9.76380825e-01 6.85365945e-02
-3.62616569e-01 5.41030049e-01 1.29623044e+00 7.22739160e-01
2.45942175e-01 6.02021158e-01 5.42242050e-01 -5.52447498e-01
5.79042137e-02 -1.14940345e+00 -1.09966660e+00 5.00411332e-01
-5.66903055e-01 -4.78155643e-01 -1.86799586e-01 1.71690986e-01] | [15.931672096252441, 5.219769477844238] |
6b7f6b30-06fa-489d-904f-06d9f79c78cc | studying-the-role-of-named-entities-for | 2206.09676 | null | https://arxiv.org/abs/2206.09676v1 | https://arxiv.org/pdf/2206.09676v1.pdf | Studying the role of named entities for content preservation in text style transfer | Text style transfer techniques are gaining popularity in Natural Language Processing, finding various applications such as text detoxification, sentiment, or formality transfer. However, the majority of the existing approaches were tested on such domains as online communications on public platforms, music, or entertainment yet none of them were applied to the domains which are typical for task-oriented production systems, such as personal plans arrangements (e.g. booking of flights or reserving a table in a restaurant). We fill this gap by studying formality transfer in this domain. We noted that the texts in this domain are full of named entities, which are very important for keeping the original sense of the text. Indeed, if for example, someone communicates the destination city of a flight it must not be altered. Thus, we concentrate on the role of named entities in content preservation for formality text style transfer. We collect a new dataset for the evaluation of content similarity measures in text style transfer. It is taken from a corpus of task-oriented dialogues and contains many important entities related to realistic requests that make this dataset particularly useful for testing style transfer models before using them in production. Besides, we perform an error analysis of a pre-trained formality transfer model and introduce a simple technique to use information about named entities to enhance the performance of baseline content similarity measures used in text style transfer. | ['Alexander Panchenko', 'Irina Krotova', 'Varvara Logacheva', 'David Dale', 'Nikolay Babakov'] | 2022-06-20 | null | null | null | null | ['text-style-transfoer'] | ['natural-language-processing'] | [ 4.64224339e-01 2.38105446e-01 -3.39231715e-02 -5.85666001e-01
-3.98748040e-01 -7.69569397e-01 8.85592341e-01 5.83220661e-01
-7.99541056e-01 1.14105916e+00 5.64343810e-01 -2.92196423e-01
1.38291597e-01 -8.21964622e-01 -7.11777925e-01 -2.02455252e-01
1.01484813e-01 7.26446986e-01 3.44517708e-01 -1.00351393e+00
4.47450042e-01 1.67587385e-01 -1.11810255e+00 6.25484109e-01
8.00328851e-01 5.95618725e-01 2.19441086e-01 2.51550108e-01
-5.90216041e-01 4.96599942e-01 -1.12375724e+00 -6.97373986e-01
-7.79242516e-02 -6.01319432e-01 -1.24451184e+00 6.46030605e-02
8.33259150e-02 -8.93401727e-02 -7.14733377e-02 9.24228311e-01
6.80790782e-01 5.25869131e-01 1.05948400e+00 -1.13634932e+00
-6.00298464e-01 9.54567969e-01 -1.09623864e-01 1.30029842e-01
6.15315139e-01 -2.30149314e-01 1.18653309e+00 -6.83335423e-01
8.74131203e-01 1.29244065e+00 3.04795146e-01 5.76540530e-01
-1.06820118e+00 -3.90198141e-01 -9.76745621e-04 -7.65799880e-02
-9.42820549e-01 -2.99469262e-01 5.44527888e-01 -2.06509188e-01
7.63697445e-01 3.03900242e-01 2.34189987e-01 1.32956553e+00
1.50120601e-01 8.97349477e-01 9.62168157e-01 -5.81036448e-01
1.79318517e-01 7.67621458e-01 -9.43494067e-02 1.81749508e-01
-1.58072546e-01 -4.68901426e-01 -6.07480764e-01 5.32388426e-02
3.84769768e-01 -3.54043990e-01 -3.33710700e-01 -1.89377367e-01
-1.54403615e+00 1.00351489e+00 3.71975839e-01 7.38883674e-01
-8.86360556e-03 -2.97434419e-01 9.23281491e-01 7.67140150e-01
6.41651094e-01 7.53380358e-01 -6.10115886e-01 -3.76288921e-01
-5.76832533e-01 3.48387301e-01 1.36933219e+00 1.36078584e+00
6.38156891e-01 -4.79950666e-01 -4.01425302e-01 9.54804361e-01
-2.01746657e-01 2.50212580e-01 6.54503524e-01 -4.26959544e-01
7.75791526e-01 5.46710014e-01 1.99248627e-01 -9.54742670e-01
-2.61952877e-01 -1.97334811e-01 -9.04689074e-01 -3.79400998e-01
4.68633890e-01 -4.92491901e-01 -1.78411096e-01 1.60221338e+00
2.61176765e-01 -2.92859346e-01 3.23174506e-01 6.26883984e-01
8.53006661e-01 7.68795371e-01 4.20299433e-02 -1.18722498e-01
1.49590397e+00 -9.68201339e-01 -6.80181205e-01 -6.99381903e-02
1.06894183e+00 -1.07011437e+00 1.39715433e+00 -3.19314711e-02
-8.03682923e-01 -5.17211497e-01 -6.48091018e-01 -6.64428622e-02
-8.58325899e-01 -5.17735817e-02 3.21427107e-01 4.61893827e-01
-6.33174181e-01 8.56958151e-01 -1.78272069e-01 -8.26571822e-01
-1.50192842e-01 5.94843589e-02 -5.18990219e-01 1.11201771e-01
-1.67529297e+00 9.42063928e-01 3.92265737e-01 -4.00327533e-01
-6.36375844e-02 -6.60281539e-01 -1.00431502e+00 1.48306653e-01
3.80786240e-01 -6.17987394e-01 1.50810087e+00 -1.31232095e+00
-1.51087177e+00 1.05300593e+00 9.97664686e-03 -3.55689019e-01
4.84477729e-01 -2.83163697e-01 -5.34560502e-01 -1.92336321e-01
3.56998175e-01 6.92858338e-01 7.47719944e-01 -9.34590459e-01
-7.38424897e-01 2.96046846e-02 5.66681683e-01 3.94232810e-01
-6.65325463e-01 2.27108836e-01 -9.86718163e-02 -9.62353349e-01
-5.70118725e-01 -9.67076182e-01 1.18779749e-01 -1.97574720e-01
-3.92553896e-01 -2.98319519e-01 7.59133577e-01 -5.16683280e-01
1.17747605e+00 -2.18876624e+00 3.91165465e-01 1.12725280e-01
-4.08782065e-01 2.51965851e-01 -8.12769458e-02 1.12625349e+00
-1.75960734e-02 2.54963100e-01 -2.37663046e-01 -2.54320771e-01
2.82303602e-01 1.53373137e-01 -4.21199709e-01 -4.47022505e-02
1.91299498e-01 7.39364266e-01 -8.10223043e-01 -3.41980219e-01
-5.76471239e-02 1.14923716e-01 -5.70837736e-01 1.84801370e-01
-3.83603364e-01 5.75367451e-01 -4.95516360e-01 1.10816453e-02
1.25814870e-01 -9.16260257e-02 6.23455197e-02 -2.79797852e-01
-2.20989101e-02 6.46712244e-01 -9.80655968e-01 1.81350636e+00
-8.67885232e-01 5.43409944e-01 -1.47345498e-01 -8.09148729e-01
7.70718455e-01 2.92263776e-01 4.09503043e-01 -6.74612403e-01
2.49241367e-01 7.70678297e-02 1.87943522e-02 -2.80855089e-01
1.03474724e+00 -3.57254595e-01 -4.56863016e-01 7.34265983e-01
-5.95008582e-03 -5.69174349e-01 5.23424149e-01 3.73980850e-01
9.33966041e-01 1.14320517e-02 3.98300856e-01 -3.86387199e-01
6.76231742e-01 -1.74697246e-02 -1.71684418e-02 3.14246178e-01
1.41170606e-01 5.60085654e-01 4.17410105e-01 -9.18011591e-02
-9.38656151e-01 -6.64253950e-01 1.55301625e-03 1.42147744e+00
1.21621341e-01 -7.22963333e-01 -8.45782876e-01 -1.04272890e+00
-1.41745389e-01 1.20126498e+00 -6.32793903e-01 -2.54436553e-01
-4.50889707e-01 -3.20374250e-01 2.60254204e-01 1.60108373e-01
4.26276028e-01 -1.38737464e+00 -2.82743335e-01 4.08006132e-01
-5.53418696e-01 -1.07506227e+00 -9.73550200e-01 -5.43960091e-03
-5.76741755e-01 -7.88040876e-01 -9.04861689e-01 -7.23697543e-01
5.58544636e-01 1.00766093e-01 1.37573099e+00 -6.33516833e-02
1.63823023e-01 5.95532298e-01 -8.25382113e-01 -7.52447128e-01
-8.04289222e-01 6.80782557e-01 6.20694309e-02 1.13995016e-01
3.52686763e-01 -4.89937603e-01 -1.86771110e-01 5.39444327e-01
-1.23245573e+00 -5.43022081e-02 3.58907789e-01 8.81732464e-01
3.94217856e-03 -1.96195543e-02 7.30561435e-01 -1.26031423e+00
1.22304869e+00 -5.66378355e-01 -1.32062197e-01 2.26993933e-01
-1.61448419e-01 2.22675905e-01 7.09366977e-01 -5.07828534e-01
-1.03999031e+00 -4.41650748e-01 -2.43050784e-01 2.32712373e-01
-3.17994326e-01 4.85685825e-01 -5.21363365e-03 2.56434619e-01
7.47347653e-01 2.36278012e-01 -1.21673197e-01 -3.84700924e-01
2.36408919e-01 8.39485526e-01 -8.37765541e-03 -7.40440607e-01
8.41547072e-01 2.93184668e-02 -2.95544177e-01 -1.08491051e+00
-8.41096282e-01 -6.28274918e-01 -5.33967912e-01 4.68080305e-02
5.26698470e-01 -4.33265477e-01 -5.79710007e-01 9.73976627e-02
-1.23409903e+00 -4.78560835e-01 -6.15818977e-01 3.80217791e-01
-6.35918498e-01 3.38827342e-01 -4.33201700e-01 -1.43251806e-01
-3.90298665e-01 -8.25179636e-01 1.12459362e+00 -3.77383269e-02
-6.29375935e-01 -1.41271377e+00 1.46435425e-01 1.12167843e-01
5.88946342e-01 2.49259751e-02 1.25951278e+00 -8.82927597e-01
9.31550860e-02 6.40190989e-02 1.07724667e-02 3.06030095e-01
5.14472902e-01 -2.32409492e-01 -5.88258862e-01 -3.37831140e-01
-2.84619838e-01 -5.11942923e-01 4.23116595e-01 -1.92816198e-01
8.21377397e-01 -4.22242582e-01 -1.25666708e-01 -5.58507070e-02
8.30349684e-01 -9.46219042e-02 6.96937442e-01 4.33295637e-01
4.30369884e-01 1.14991617e+00 9.65171158e-01 5.70159197e-01
3.96402121e-01 9.62554932e-01 5.59940822e-02 -1.00753024e-01
1.49329007e-03 -2.89487362e-01 4.92011964e-01 1.01691365e+00
1.72593728e-01 -5.72725296e-01 -5.47407746e-01 5.69220304e-01
-1.69224572e+00 -7.35040188e-01 2.26100683e-01 2.13148594e+00
1.30811656e+00 1.78802058e-01 2.69997045e-02 -3.01496014e-02
6.97648585e-01 4.03883830e-02 -2.73902059e-01 -5.42685151e-01
5.30523360e-02 1.34652510e-01 1.70616701e-01 3.07298303e-01
-9.08536613e-01 1.08611524e+00 4.87723160e+00 1.01304245e+00
-9.28964376e-01 -1.74804532e-03 4.35503155e-01 2.51050979e-01
-4.24709886e-01 -2.01287851e-01 -9.33472574e-01 4.85544354e-01
8.56689990e-01 -3.96876335e-01 2.02967495e-01 6.01436794e-01
1.47824496e-01 6.56814724e-02 -1.50845122e+00 8.11700821e-01
8.82446766e-02 -9.73146617e-01 3.34671348e-01 -2.41258487e-01
6.00631118e-01 -3.17423582e-01 -9.03214514e-02 5.78435242e-01
1.42551541e-01 -8.00103426e-01 6.38981342e-01 1.85542598e-01
8.04467261e-01 -7.46761620e-01 8.25826585e-01 4.70187247e-01
-7.40103364e-01 3.19716424e-01 -3.55341941e-01 3.42197865e-02
3.63264173e-01 4.64237392e-01 -1.10413921e+00 7.08947659e-01
5.72215378e-01 6.53778076e-01 -4.23591673e-01 6.40694797e-01
-1.89218819e-01 3.88597459e-01 -2.15232223e-01 -4.30627197e-01
4.54666704e-01 -1.57142028e-01 5.81578255e-01 1.28176641e+00
4.25273299e-01 1.31308157e-02 -6.40223101e-02 5.45024216e-01
-5.66815555e-01 8.12739909e-01 -9.02528286e-01 -1.92535117e-01
1.07536018e-02 1.17483222e+00 -6.35900855e-01 -1.75613984e-01
-5.51935315e-01 1.28802216e+00 1.75554842e-01 1.49221838e-01
-7.21034169e-01 -7.96689451e-01 5.68580985e-01 3.52293193e-01
1.49479449e-01 -1.75467595e-01 2.65969187e-01 -1.19764423e+00
5.21620288e-02 -1.11858714e+00 2.39400923e-01 -7.43452072e-01
-1.51071191e+00 7.81891227e-01 2.02627197e-01 -1.34937549e+00
-5.51829934e-01 -4.24277961e-01 -5.82392633e-01 6.62287295e-01
-1.48305500e+00 -9.06257868e-01 -4.41841930e-02 7.68910587e-01
9.27195668e-01 -3.49891871e-01 9.95896280e-01 4.23273116e-01
-1.38422614e-02 6.24921739e-01 6.54850304e-02 1.46476895e-01
1.44530380e+00 -1.26060951e+00 5.18865168e-01 3.68951499e-01
2.28548154e-01 6.30127370e-01 8.27846348e-01 -5.18216729e-01
-9.54016507e-01 -1.14919913e+00 1.14083529e+00 -3.36944908e-01
8.01361740e-01 -5.12722313e-01 -9.27673995e-01 6.80268884e-01
7.83742011e-01 -6.04288876e-01 8.10207009e-01 9.20412913e-02
1.28397182e-01 8.61871168e-02 -1.07885087e+00 8.47221732e-01
9.54397678e-01 -2.87353963e-01 -7.94743061e-01 7.19114304e-01
8.19737494e-01 -3.27953428e-01 -7.58719623e-01 1.79748371e-01
1.05547950e-01 -7.04483211e-01 5.18986285e-01 -9.24167275e-01
5.67991614e-01 1.30849276e-02 8.22889507e-02 -1.91674984e+00
1.26062512e-01 -7.82298625e-01 4.63508397e-01 1.68690693e+00
6.49110734e-01 -6.41469777e-01 4.58526343e-01 5.51354587e-01
-4.62469384e-02 -2.17575029e-01 -6.23824954e-01 -4.41638857e-01
8.96654129e-02 -2.23427210e-02 4.78792936e-01 1.16408598e+00
1.66077450e-01 9.10253882e-01 -2.51305908e-01 -4.81612951e-01
-3.89867462e-02 2.82857791e-02 9.85777676e-01 -9.80024278e-01
-2.91842937e-01 -2.32392520e-01 -1.02435291e-01 -1.20368099e+00
4.03663129e-01 -9.37268138e-01 -7.48324469e-02 -1.40079618e+00
-2.59979367e-01 -6.28300130e-01 2.69509524e-01 3.60856533e-01
6.06706515e-02 6.68476745e-02 2.83477008e-01 -6.78664073e-02
-4.76409256e-01 8.59951854e-01 1.38514662e+00 -3.86748374e-01
-2.74154425e-01 5.27257502e-01 -6.86393261e-01 6.15317941e-01
9.31723356e-01 -4.95933026e-01 -5.45139074e-01 -2.45675743e-01
4.64048088e-01 2.80962754e-02 -2.95732826e-01 -7.01526165e-01
1.01499721e-01 -1.03068642e-01 -2.42609665e-01 -7.79675096e-02
2.14969531e-01 -1.18022084e+00 -3.39252889e-01 2.21685261e-01
-5.99753261e-01 1.14920393e-01 2.43443727e-01 2.85454780e-01
-6.16481781e-01 -6.15046203e-01 5.60353875e-01 -1.75605580e-01
-6.97190821e-01 5.43280393e-02 -5.71350396e-01 4.53006506e-01
9.12585080e-01 -1.07655197e-01 -1.92947447e-01 -9.08253133e-01
-5.56198120e-01 1.03056259e-01 3.43405336e-01 8.75881672e-01
2.99144983e-01 -1.20073891e+00 -8.44876707e-01 -1.87215079e-02
3.91567171e-01 -1.01768881e-01 2.28292011e-02 5.22879660e-01
-5.15226364e-01 3.60443950e-01 -4.36857909e-01 -2.33605072e-01
-1.24691641e+00 4.09986198e-01 -2.73255575e-02 -5.06680906e-01
-3.96291047e-01 3.72162879e-01 3.66965085e-01 -8.52886379e-01
1.97449893e-01 -5.03549874e-01 -4.74091351e-01 2.72218108e-01
5.17854393e-01 1.96034029e-01 2.54475415e-01 -5.19214928e-01
-4.77598459e-02 3.62905085e-01 -1.95247784e-01 -2.09642515e-01
1.21086049e+00 -3.62766922e-01 -2.42945001e-01 6.81232929e-01
1.19902837e+00 3.02008778e-01 -5.23992479e-01 -4.47355628e-01
6.37019053e-02 -1.11328363e-01 -5.83758354e-01 -6.70985341e-01
-6.30041718e-01 9.51320052e-01 5.64221106e-02 5.01449525e-01
9.05378163e-01 -3.76222804e-02 8.20780575e-01 8.83004367e-01
5.31629384e-01 -1.04339874e+00 1.47280246e-01 1.12594259e+00
1.23979259e+00 -1.12882447e+00 -2.77030677e-01 -4.96301621e-01
-9.45679843e-01 1.12943983e+00 4.59350616e-01 3.21098298e-01
5.56472421e-01 -3.26518230e-02 -5.67002818e-02 8.69633034e-02
-5.33207655e-01 3.61879286e-03 1.91253945e-01 5.70563853e-01
6.78163350e-01 -1.71445772e-01 -5.76072633e-01 3.79572481e-01
-4.87593621e-01 -2.22748250e-01 5.64354658e-01 9.02793705e-01
-2.55331784e-01 -1.42542338e+00 -4.74821925e-02 2.75435060e-01
-6.22691810e-01 -3.13032776e-01 -6.39048040e-01 7.66598165e-01
-3.28126401e-01 7.62566030e-01 -7.54524469e-02 -1.12211354e-01
7.17936754e-01 1.62402615e-01 3.36090773e-01 -1.10103667e+00
-1.06757951e+00 -3.14537704e-01 4.40307766e-01 -2.21602231e-01
-6.08067691e-01 -4.65164006e-01 -1.10028672e+00 -3.93491983e-01
-1.85149521e-01 5.70921719e-01 6.73437357e-01 9.85924661e-01
1.15715317e-01 5.81493735e-01 5.06661773e-01 -8.23536873e-01
-5.06364882e-01 -1.37995839e+00 -8.11892092e-01 8.93867910e-01
-1.60311669e-01 -5.05580962e-01 -1.08974680e-01 3.42303753e-01] | [11.60770034790039, 9.217864036560059] |
2565f9de-eaf9-40f2-abe1-8f01b07c0814 | interpretable-graph-neural-networks-for | 2207.00813 | null | https://arxiv.org/abs/2207.00813v2 | https://arxiv.org/pdf/2207.00813v2.pdf | Interpretable Graph Neural Networks for Connectome-Based Brain Disorder Analysis | Human brains lie at the core of complex neurobiological systems, where the neurons, circuits, and subsystems interact in enigmatic ways. Understanding the structural and functional mechanisms of the brain has long been an intriguing pursuit for neuroscience research and clinical disorder therapy. Mapping the connections of the human brain as a network is one of the most pervasive paradigms in neuroscience. Graph Neural Networks (GNNs) have recently emerged as a potential method for modeling complex network data. Deep models, on the other hand, have low interpretability, which prevents their usage in decision-critical contexts like healthcare. To bridge this gap, we propose an interpretable framework to analyze disorder-specific Regions of Interest (ROIs) and prominent connections. The proposed framework consists of two modules: a brain-network-oriented backbone model for disease prediction and a globally shared explanation generator that highlights disorder-specific biomarkers including salient ROIs and important connections. We conduct experiments on three real-world datasets of brain disorders. The results verify that our framework can obtain outstanding performance and also identify meaningful biomarkers. All code for this work is available at https://github.com/HennyJie/IBGNN.git. | ['Carl Yang', 'Lifang He', 'Xiaoxiao Li', 'Yanqiao Zhu', 'Wei Dai', 'Hejie Cui'] | 2022-06-30 | null | null | null | null | ['disease-prediction'] | ['medical'] | [-6.21271245e-02 3.29980344e-01 -2.47057214e-01 -4.10572469e-01
3.09142739e-01 -1.33845493e-01 2.66510993e-01 4.70250919e-02
1.19590327e-01 5.71347952e-01 2.26317853e-01 -2.12000147e-01
-4.83324736e-01 -6.21764004e-01 -1.22619547e-01 -4.59170014e-01
-3.47976297e-01 3.56851190e-01 1.67177513e-01 -8.54154602e-02
1.25505343e-01 5.58380961e-01 -8.25357616e-01 1.84695378e-01
1.10213315e+00 8.05585504e-01 2.12630659e-01 5.80830649e-02
-1.30513087e-01 6.34825170e-01 -3.02778035e-01 -2.09980249e-01
-2.09065154e-02 -6.40499473e-01 -7.20442772e-01 -2.82880247e-01
-2.25910008e-01 3.10592577e-02 -5.82916319e-01 1.44276726e+00
3.40046197e-01 -3.68449539e-01 4.07133430e-01 -1.50445509e+00
-5.25550008e-01 6.17390275e-01 -6.14725947e-01 5.27868629e-01
-1.00491002e-01 2.80148119e-01 1.07924163e+00 -5.41318357e-01
6.27526999e-01 1.35996926e+00 3.04962367e-01 6.33683383e-01
-1.13516843e+00 -7.61236310e-01 2.55987525e-01 6.78831756e-01
-1.20882678e+00 -2.59394050e-01 7.36406326e-01 -6.35091722e-01
6.34598196e-01 2.21585140e-01 1.00751793e+00 1.03923738e+00
5.35035074e-01 4.31239903e-01 8.71572852e-01 2.60791004e-01
-7.36189075e-03 -2.41989389e-01 2.68911660e-01 8.26671183e-01
5.02476573e-01 -2.94780731e-01 -3.72269660e-01 -1.12741828e-01
8.74433160e-01 5.78930974e-01 -5.42876422e-01 -1.12577803e-01
-1.41161704e+00 6.38417959e-01 1.04152858e+00 5.18592238e-01
-3.50844234e-01 1.75430998e-01 4.24516171e-01 2.25212779e-02
5.84883451e-01 1.95543125e-01 -1.64457574e-01 3.62830043e-01
-7.74946809e-01 -4.63223793e-02 4.52731282e-01 4.90610540e-01
3.93520176e-01 -1.18601970e-01 -4.56311963e-02 9.27586496e-01
4.64089900e-01 5.82816079e-02 4.65132713e-01 -4.70729858e-01
3.24001998e-01 1.18832850e+00 -6.06539607e-01 -1.34583604e+00
-8.93784940e-01 -6.23017609e-01 -1.35838425e+00 -2.02179238e-01
1.56690300e-01 -1.43728763e-01 -3.87006283e-01 1.78792644e+00
3.40669811e-01 3.69739711e-01 -6.96284235e-01 1.04559743e+00
8.57808232e-01 2.56395131e-01 1.73024014e-01 3.73245813e-02
1.48488796e+00 -8.73625934e-01 -7.36567557e-01 -2.05078587e-01
4.52282399e-01 -2.53327221e-01 7.38697827e-01 2.66329665e-02
-7.78577149e-01 -1.99657246e-01 -6.98257208e-01 -8.47031176e-02
-2.58770019e-01 1.09211527e-01 6.80981576e-01 2.03065593e-02
-1.13401747e+00 5.24788976e-01 -8.80260348e-01 -5.61880469e-01
9.95544970e-01 3.63478273e-01 -3.42226207e-01 -7.54721016e-02
-1.12691164e+00 8.69476199e-01 2.96718538e-01 4.92701471e-01
-7.75026321e-01 -7.33078122e-01 -4.48725611e-01 2.60604680e-01
2.33622015e-01 -8.18942487e-01 5.40576994e-01 -5.65550208e-01
-8.08493555e-01 7.32248425e-01 -8.20585862e-02 -1.89122200e-01
2.74309725e-01 8.85181278e-02 -7.28278399e-01 3.23670805e-01
4.78084162e-02 6.24278486e-01 4.72030073e-01 -6.48478150e-01
-1.61819115e-01 -7.00660169e-01 -2.21003279e-01 -4.69925664e-02
-5.13463795e-01 1.70149386e-01 -8.01231712e-02 -5.52313864e-01
3.86643797e-01 -7.60692239e-01 -4.10720557e-01 4.24932808e-01
-1.06058776e+00 -1.63300008e-01 7.12864578e-01 -7.68948138e-01
1.27894998e+00 -1.85215712e+00 3.18938762e-01 4.47910488e-01
1.14464462e+00 6.68592528e-02 1.12215448e-02 3.40607047e-01
-4.78592515e-01 3.43461126e-01 -2.38810197e-01 1.71870321e-01
-2.55503595e-01 -2.44929567e-01 7.66221508e-02 6.32702529e-01
2.41708964e-01 1.08472049e+00 -9.10109818e-01 -4.18735325e-01
6.86762184e-02 4.41896945e-01 -3.89250219e-01 -2.18189758e-04
-1.44584686e-01 6.85914099e-01 -8.14511240e-01 6.63446069e-01
3.65730107e-01 -7.06180573e-01 4.69085097e-01 -3.66262138e-01
1.54061183e-01 2.09133178e-01 -6.09859228e-01 1.42784548e+00
1.96479306e-01 6.23507917e-01 1.75794065e-01 -1.30981040e+00
9.61165726e-01 3.47433597e-01 6.68105543e-01 -4.92774040e-01
2.67421216e-01 1.11905403e-01 8.84447515e-01 -6.72621667e-01
-4.00245577e-01 1.86150819e-01 3.61047149e-01 4.93080020e-01
-2.88273524e-02 5.13116658e-01 1.08487912e-01 3.50697547e-01
1.57900262e+00 -3.84294003e-01 4.15292114e-01 -5.41285276e-01
5.03859460e-01 -1.80906519e-01 7.37931967e-01 -1.07995495e-01
-4.82917964e-01 2.75631517e-01 1.06845701e+00 -5.04143655e-01
-8.73557925e-01 -8.28174710e-01 -2.27241352e-01 5.28541148e-01
1.31378785e-01 -2.78892875e-01 -8.46150458e-01 -4.12069768e-01
-7.05412999e-02 1.52548507e-01 -7.16137290e-01 -3.57833922e-01
-2.70287514e-01 -8.46427023e-01 5.13929605e-01 2.10792586e-01
4.84687150e-01 -1.30911994e+00 -2.27185830e-01 4.59978431e-02
-1.88808948e-01 -9.70788240e-01 -3.46970618e-01 -3.23121756e-01
-1.03724241e+00 -1.58817065e+00 -5.76872408e-01 -7.27370858e-01
1.14151561e+00 2.89415866e-01 9.08233523e-01 5.24659932e-01
-8.01179528e-01 -4.32089344e-02 -6.36443198e-02 -1.49087250e-01
-1.19159468e-01 1.44503489e-01 6.99555799e-02 1.37533814e-01
3.36705714e-01 -9.53729391e-01 -1.09615839e+00 3.98616195e-01
-8.40846002e-01 4.38668430e-01 6.59520566e-01 6.71976984e-01
5.24311900e-01 -2.89657265e-01 9.43535626e-01 -9.71671402e-01
9.54157174e-01 -1.11273170e+00 -4.76041257e-01 1.99522451e-01
-4.57918912e-01 -3.02530617e-01 6.23676062e-01 -5.66684157e-02
-6.14835382e-01 -1.98341057e-01 -3.39833722e-02 -2.67451316e-01
-2.47394979e-01 6.38801336e-01 -3.76096129e-01 1.21923968e-01
3.64426792e-01 -2.98870932e-02 2.20429763e-01 -5.53194046e-01
1.50907964e-01 3.93118829e-01 1.85771883e-01 -2.57669091e-01
3.68191659e-01 3.76656711e-01 2.17608765e-01 -6.99231386e-01
-5.90042830e-01 -3.44900876e-01 -6.35176957e-01 -4.97579873e-01
8.66543174e-01 -5.60703397e-01 -8.53650272e-01 7.75146484e-02
-1.30410659e+00 -1.49037153e-01 1.64860651e-01 3.81123424e-01
2.67984858e-03 1.36467114e-01 -6.20146334e-01 -1.39693201e-01
-5.17435789e-01 -1.08949363e+00 4.62775975e-01 3.10725659e-01
-4.00464118e-01 -1.18697810e+00 1.50204867e-01 3.21279526e-01
3.23127240e-01 4.54181522e-01 1.33260500e+00 -7.14600205e-01
-5.98876297e-01 1.37842178e-01 -6.55919492e-01 -5.37406988e-02
3.01716238e-01 -1.91237852e-02 -6.75221384e-01 1.43088559e-02
-2.00784430e-01 1.76734123e-02 8.37509930e-01 3.95295084e-01
1.53023040e+00 -3.39250684e-01 -5.84849298e-01 5.18176138e-01
1.16715550e+00 8.68652388e-02 5.73078692e-01 -6.80903271e-02
8.80511045e-01 8.95557761e-01 -1.54392809e-01 3.43421400e-01
3.31069916e-01 4.40867871e-01 7.26209223e-01 -2.37012684e-01
-1.92874610e-01 3.81835327e-02 2.29898691e-01 1.03568685e+00
-1.58372268e-01 -1.21052094e-01 -1.31603265e+00 5.86693585e-01
-1.97303450e+00 -1.01034820e+00 -5.08756816e-01 1.64364254e+00
5.61760724e-01 -9.28552449e-02 6.81960136e-02 -1.63566992e-01
1.10908508e+00 -6.66385219e-02 -1.04607725e+00 -1.72591627e-01
3.80178615e-02 1.80842523e-02 -1.53139666e-01 4.01654728e-02
-5.08442521e-01 7.50520349e-01 5.13214636e+00 4.16238785e-01
-9.91537631e-01 3.04672182e-01 9.71275210e-01 -2.56758720e-01
-2.78334945e-01 -1.87238649e-01 -2.90365726e-01 6.69307828e-01
7.00351298e-01 -2.81938881e-01 5.79333007e-01 6.07474923e-01
7.06529438e-01 2.72825092e-01 -1.00136340e+00 9.41874325e-01
-3.49762291e-01 -1.44863045e+00 3.90522368e-02 2.92232782e-01
5.25903106e-01 2.64804333e-01 -6.37539923e-02 -3.36635500e-01
-2.28763185e-02 -1.20009232e+00 2.13522539e-01 8.82873356e-01
7.84925222e-01 -4.02532697e-01 5.12859583e-01 2.26657182e-01
-1.08756006e+00 6.70407154e-03 -4.43538934e-01 1.01606719e-01
-5.51872961e-02 1.02366984e+00 -7.25902021e-01 2.68676579e-01
6.76822543e-01 1.16813791e+00 -5.94370127e-01 1.36564434e+00
-4.46620315e-01 6.46805704e-01 1.64386675e-01 -6.65633827e-02
-3.22441831e-02 -4.70690906e-01 4.76635039e-01 9.29208100e-01
3.48914087e-01 1.35906175e-01 -7.00974166e-02 1.65896749e+00
-3.45219344e-01 2.27424234e-01 -7.20726609e-01 -2.90520549e-01
5.06594837e-01 1.84746730e+00 -1.11327517e+00 -2.98587792e-02
-3.52539062e-01 5.86590827e-01 4.64027762e-01 2.39671648e-01
-7.62677491e-01 -2.27470189e-01 8.40799332e-01 2.30900139e-01
-3.77600759e-01 -1.73812620e-02 -4.36747730e-01 -1.20450258e+00
-1.42057966e-02 -6.70979083e-01 2.23426953e-01 -5.95985293e-01
-1.49784625e+00 7.81800866e-01 -3.62112701e-01 -1.17653906e+00
2.47367591e-01 -4.81303960e-01 -1.03604233e+00 8.53955269e-01
-1.23938215e+00 -8.32597733e-01 -5.61089575e-01 7.59062231e-01
1.89136371e-01 -2.34803215e-01 7.31658936e-01 4.46246654e-01
-1.28001857e+00 7.47413486e-02 -1.55100683e-02 2.35699460e-01
3.45694751e-01 -8.08768928e-01 2.59893626e-01 6.37223065e-01
-1.05344757e-01 8.63745093e-01 2.61604786e-01 -7.89955020e-01
-9.86032307e-01 -1.26681495e+00 6.59694791e-01 6.15003565e-03
1.07206810e+00 -4.99150336e-01 -9.89917517e-01 5.83128214e-01
2.00455412e-01 1.99324891e-01 8.07695329e-01 9.63655263e-02
-1.90804396e-02 -2.21515819e-01 -1.12745464e+00 7.51714826e-01
1.27722394e+00 -3.42367351e-01 -2.11709946e-01 6.26526654e-01
5.36790967e-01 2.11376250e-01 -8.83129299e-01 9.72613618e-02
4.60204750e-01 -9.93588507e-01 7.34760523e-01 -6.92857027e-01
6.76169753e-01 -2.23412171e-01 3.73071820e-01 -1.53087008e+00
-4.45052594e-01 -3.45055580e-01 -1.59336582e-01 1.05090439e+00
4.34225649e-01 -9.39338565e-01 5.73476791e-01 5.95203280e-01
-3.09673876e-01 -1.19727981e+00 -7.25942075e-01 -4.12109554e-01
-2.28066251e-01 -1.21326603e-01 6.37912214e-01 1.03098285e+00
4.68791515e-01 4.50376421e-01 -4.42130417e-02 -2.14339010e-02
5.39413989e-01 -2.14006472e-02 1.90343678e-01 -1.49097550e+00
1.14270717e-01 -8.82936120e-01 -5.34525037e-01 -3.80793065e-01
1.79892793e-01 -1.36719847e+00 -4.57231611e-01 -1.84019244e+00
5.41381717e-01 -4.42270339e-01 -6.45594001e-01 7.36858368e-01
9.15395543e-02 1.14527745e-02 3.09511069e-02 3.00734997e-01
-5.72997928e-01 4.99419093e-01 1.53888071e+00 -1.75305396e-01
4.47437428e-02 -2.56970286e-01 -9.52763200e-01 9.58979905e-01
1.35278320e+00 -5.00031352e-01 -5.93637586e-01 -3.64317447e-01
1.26036569e-01 -1.08510084e-01 6.27011001e-01 -8.82018924e-01
4.21861470e-01 -1.30144358e-01 4.81462330e-01 -1.47571906e-01
1.27675638e-01 -7.64995873e-01 2.51060903e-01 7.94129312e-01
-3.59984010e-01 2.16320649e-01 -1.02196271e-02 4.01011139e-01
-1.14603035e-01 1.88848510e-01 6.91415548e-01 -1.03607975e-01
-3.84484291e-01 7.68470824e-01 -1.98570803e-01 1.43039227e-01
1.07396364e+00 7.37141147e-02 -7.82764435e-01 -2.14533970e-01
-7.17505813e-01 3.39559048e-01 1.91964116e-02 4.13074285e-01
9.53943908e-01 -1.21220326e+00 -6.60535932e-01 4.48519140e-02
-9.65790302e-02 -2.92615861e-01 5.84572077e-01 1.25110209e+00
-5.40773988e-01 5.96030414e-01 -6.23519301e-01 -4.43585455e-01
-1.04922032e+00 2.65615374e-01 4.20475900e-01 -2.80302942e-01
-6.81961179e-01 6.27324700e-01 4.83594000e-01 -2.63215482e-01
3.42300348e-02 -3.14476132e-01 -4.53271568e-01 -1.16568901e-01
4.93331641e-01 5.64177096e-01 -2.05327466e-01 -6.72121704e-01
-5.56558728e-01 7.46690109e-02 -1.07737586e-01 5.01908243e-01
1.69174397e+00 -4.98245237e-03 -7.95209348e-01 1.46537170e-01
8.90018702e-01 -4.44571316e-01 -7.30675876e-01 -4.79054116e-02
1.40247419e-01 -9.97944996e-02 2.85640936e-02 -8.26182663e-01
-1.62836242e+00 1.15422988e+00 4.29957569e-01 4.00020987e-01
9.86456752e-01 1.94067419e-01 8.20889831e-01 1.14455983e-01
2.63047606e-01 -5.75791538e-01 -4.40056110e-03 1.51645884e-01
1.16655099e+00 -9.65400159e-01 4.00096811e-02 -5.40997446e-01
-2.55612701e-01 1.21054959e+00 8.75786722e-01 -2.35596150e-01
1.00794649e+00 -1.68320403e-01 -3.27644795e-01 -7.29334652e-01
-7.44307935e-01 6.44638157e-03 4.84298885e-01 3.62162530e-01
5.24686277e-01 2.34344676e-01 -4.05055106e-01 1.00073898e+00
6.51900843e-02 -1.84146538e-01 1.35494649e-01 1.81006461e-01
-3.87713313e-01 -1.01396406e+00 -8.54593441e-02 1.04802728e+00
-3.64834994e-01 -2.46355057e-01 -6.77970231e-01 6.08060598e-01
2.55008250e-01 7.24454701e-01 -1.84231803e-01 -5.08540332e-01
9.14276913e-02 -8.61654729e-02 7.69638717e-02 -7.25040376e-01
-4.36867237e-01 -1.60488319e-02 -2.31650695e-01 -8.60193431e-01
-2.34694257e-01 -4.96826470e-01 -1.53720856e+00 -5.09005427e-01
-3.19484882e-02 -1.78037956e-01 4.11311239e-01 8.57207954e-01
8.62114847e-01 8.86409879e-01 4.48287487e-01 -4.97068197e-01
8.00883621e-02 -8.86793256e-01 -7.93137968e-01 1.81583866e-01
8.23806599e-02 -6.81320310e-01 5.60968146e-02 -9.13249403e-02] | [12.42697525024414, 3.3840749263763428] |
48ec56c8-d346-40ee-9bb4-0c09026457be | adaptive-exploration-for-unsupervised-person | 1907.04194 | null | https://arxiv.org/abs/1907.04194v2 | https://arxiv.org/pdf/1907.04194v2.pdf | Adaptive Exploration for Unsupervised Person Re-Identification | Due to domain bias, directly deploying a deep person re-identification (re-ID) model trained on one dataset often achieves considerably poor accuracy on another dataset. In this paper, we propose an Adaptive Exploration (AE) method to address the domain-shift problem for re-ID in an unsupervised manner. Specifically, in the target domain, the re-ID model is inducted to 1) maximize distances between all person images and 2) minimize distances between similar person images. In the first case, by treating each person image as an individual class, a non-parametric classifier with a feature memory is exploited to encourage person images to move far away from each other. In the second case, according to a similarity threshold, our method adaptively selects neighborhoods for each person image in the feature space. By treating these similar person images as the same class, the non-parametric classifier forces them to stay closer. However, a problem of the adaptive selection is that, when an image has too many neighborhoods, it is more likely to attract other images as its neighborhoods. As a result, a minority of images may select a large number of neighborhoods while a majority of images have only a few neighborhoods. To address this issue, we additionally integrate a balance strategy into the adaptive selection. We evaluate our methods with two protocols. The first one is called "target-only re-ID", in which only the unlabeled target data is used for training. The second one is called "domain adaptive re-ID", in which both the source data and the target data are used during training. Experimental results on large-scale re-ID datasets demonstrate the effectiveness of our method. Our code has been released at https://github.com/dyh127/Adaptive-Exploration-for-Unsupervised-Person-Re-Identification. | ['Yuhang Ding', 'Mingliang Xu', 'Yi Yang', 'Hehe Fan'] | 2019-07-09 | null | null | null | null | ['unsupervised-person-re-identification'] | ['computer-vision'] | [-2.60060467e-02 -1.21333197e-01 -2.11730689e-01 -4.30526704e-01
-3.50570887e-01 -3.46476167e-01 4.42596525e-01 7.54838437e-02
-6.97146773e-01 5.98322332e-01 1.82436984e-02 1.54450327e-01
-3.20833400e-02 -9.96507108e-01 -5.71139514e-01 -8.98954451e-01
2.49163285e-01 6.98989928e-01 9.24336091e-02 9.69271809e-02
8.14366564e-02 1.73817545e-01 -1.49685431e+00 -3.50383520e-02
9.29082096e-01 6.76434636e-01 2.54462630e-01 1.56266153e-01
-3.78156267e-02 1.94244593e-01 -6.17175937e-01 -2.27724075e-01
5.93207896e-01 -6.59744322e-01 -7.19779968e-01 1.68825373e-01
3.83621842e-01 -3.93758625e-01 -3.45505804e-01 1.44455314e+00
6.49655342e-01 5.05178630e-01 5.65496922e-01 -1.22236419e+00
-7.80368805e-01 6.38191521e-01 -7.72890329e-01 2.26420298e-01
2.35485762e-01 2.35033691e-01 5.52370667e-01 -9.29090202e-01
4.49886203e-01 1.21983171e+00 4.01733816e-01 9.63465154e-01
-1.39296174e+00 -1.02944863e+00 4.24075246e-01 -4.51226309e-02
-1.78244603e+00 -3.82164150e-01 8.83064568e-01 -2.64696747e-01
2.81156600e-01 1.35343492e-01 6.52962685e-01 9.59183931e-01
-4.47687984e-01 8.64612341e-01 1.04473436e+00 -3.57073665e-01
3.55215311e-01 6.39191031e-01 2.81808376e-01 2.08539784e-01
2.60285944e-01 3.45489711e-01 -3.91251594e-01 -1.30412295e-01
6.36881828e-01 2.88089097e-01 -1.71161801e-01 -5.34957647e-01
-9.99730110e-01 7.76463926e-01 5.39953649e-01 3.62909675e-01
-3.27258795e-01 -3.00544381e-01 2.14564502e-01 2.85490543e-01
2.87966460e-01 3.68594110e-01 6.66234549e-03 5.31528443e-02
-8.04243684e-01 4.83688056e-01 4.12392855e-01 6.24779582e-01
1.02725613e+00 -4.19294804e-01 -1.72656178e-01 1.12411571e+00
2.12720111e-01 3.86599749e-01 9.00922179e-01 -5.94604433e-01
4.42643315e-01 8.38057578e-01 2.07638547e-01 -1.05758810e+00
-1.28665552e-01 -4.71754402e-01 -9.67748284e-01 1.80314302e-01
5.72714388e-01 -1.98969647e-01 -8.81776214e-01 2.04660702e+00
4.93737042e-01 2.42475107e-01 5.33029810e-02 1.17240036e+00
6.24139309e-01 4.95496958e-01 1.30924344e-01 -1.14777513e-01
1.20832992e+00 -8.46084237e-01 -3.18274736e-01 -2.42330432e-01
4.40005749e-01 -3.07101369e-01 1.09560096e+00 1.31427988e-01
-8.68941486e-01 -8.21559012e-01 -9.56372261e-01 5.13277531e-01
-4.09747869e-01 9.45274010e-02 -7.29957037e-03 5.58508396e-01
-7.86518216e-01 4.87276942e-01 -3.38436514e-01 -5.61039865e-01
4.68159795e-01 4.09175068e-01 -3.47526461e-01 -1.25695765e-01
-1.24451900e+00 2.91372597e-01 5.26535869e-01 -2.43889928e-01
-8.38247061e-01 -5.47233999e-01 -5.74506342e-01 1.19800001e-01
2.32361674e-01 -4.52327371e-01 9.10017490e-01 -1.41094625e+00
-1.20179868e+00 1.01918709e+00 -2.97028840e-01 -3.44024360e-01
7.88711369e-01 -4.95633595e-02 -6.03787124e-01 -6.79659322e-02
4.34944749e-01 9.01096404e-01 8.81710052e-01 -1.47961521e+00
-7.39727318e-01 -5.45832217e-01 -9.11336765e-02 5.60841203e-01
-8.06661606e-01 -1.00588128e-01 -6.53992414e-01 -7.42071867e-01
-8.69477692e-04 -1.13580322e+00 -2.12727994e-01 -2.20475435e-01
-5.99864781e-01 -3.78959000e-01 8.20665956e-01 -2.12402344e-01
1.41303039e+00 -2.45312524e+00 2.45971270e-02 5.78105748e-01
3.12486351e-01 2.75853246e-01 -2.22899362e-01 -2.09364388e-02
-3.00004631e-01 3.06155145e-01 -2.00554281e-01 -4.27631408e-01
-3.47713768e-01 -6.44363686e-02 -4.92633618e-02 2.91350067e-01
-1.15300477e-01 5.05691230e-01 -9.80641901e-01 -4.34429765e-01
1.90324798e-01 1.25164390e-01 -4.29071844e-01 2.67582148e-01
2.57475197e-01 6.25474572e-01 -5.77333391e-01 4.14395958e-01
9.62231815e-01 -2.97810644e-01 1.19644761e-01 -1.03795156e-01
-1.17300369e-01 -1.46697074e-01 -1.47148883e+00 1.23143804e+00
4.91355546e-02 1.43133327e-01 -2.11290434e-01 -9.53514934e-01
9.19236779e-01 2.54775491e-02 5.55881143e-01 -6.63134158e-01
-3.69132310e-03 1.47700965e-01 -6.52966788e-03 -1.07129499e-01
3.20024848e-01 7.08581880e-02 -1.40368016e-02 7.79784024e-01
-3.09875697e-01 6.08856678e-01 1.93904117e-01 1.81870952e-01
7.66369164e-01 -1.57089323e-01 1.57093197e-01 -2.33648568e-01
7.22915947e-01 -1.33006079e-02 8.80190074e-01 1.04614127e+00
-6.52655959e-01 6.89153314e-01 7.76667371e-02 -5.96335411e-01
-9.77867246e-01 -1.12621772e+00 -9.25834179e-02 1.08385980e+00
6.65715814e-01 -1.89969480e-01 -8.33713591e-01 -1.03555846e+00
1.65938377e-01 5.87829232e-01 -7.27842987e-01 -3.48724186e-01
-6.13895178e-01 -7.73726225e-01 4.44352299e-01 3.95322889e-01
1.01300454e+00 -1.06794572e+00 -3.37900311e-01 8.25110003e-02
-2.60126859e-01 -4.98096675e-01 -8.59897673e-01 -1.23843752e-01
-6.64950490e-01 -9.17147756e-01 -1.14880383e+00 -8.82771671e-01
1.13956976e+00 5.02337933e-01 6.82693660e-01 1.97510675e-01
9.85921621e-02 1.62347600e-01 -3.63975674e-01 -2.36932561e-01
-3.24353009e-01 4.09008786e-02 4.64771509e-01 4.10285115e-01
9.06114042e-01 -3.42376798e-01 -8.75275731e-01 7.34865427e-01
-7.17613339e-01 -1.75556168e-01 3.81779701e-01 9.06140327e-01
7.15613484e-01 4.30301368e-01 6.50626719e-01 -7.75290549e-01
7.05815256e-01 -6.12518489e-01 -3.53679538e-01 1.48171008e-01
-7.52323151e-01 -1.42723665e-01 6.11275792e-01 -1.02787483e+00
-9.04827416e-01 3.02872926e-01 2.37258494e-01 -5.30259609e-01
-3.77258867e-01 2.40975484e-01 -4.15876478e-01 2.02279776e-01
7.27091908e-01 4.15705860e-01 1.41877785e-01 -4.72460270e-01
2.38926094e-02 8.27737927e-01 5.57367861e-01 -5.37956417e-01
1.02973914e+00 4.75581676e-01 -6.39180720e-01 -4.96720076e-01
-5.45607090e-01 -6.14337981e-01 -5.15050888e-01 -2.75401264e-01
8.26186121e-01 -9.78947997e-01 -4.07181889e-01 6.36592269e-01
-6.49055362e-01 -3.45922559e-01 -4.64743346e-01 5.52588165e-01
5.89539595e-02 3.17363381e-01 -1.80626094e-01 -7.22497880e-01
-1.63113579e-01 -1.18754089e+00 7.18449116e-01 8.39049876e-01
-3.68363678e-01 -7.37912536e-01 4.41656224e-02 2.60555565e-01
1.38708487e-01 -1.53688461e-01 6.28068805e-01 -1.10511100e+00
-2.81628340e-01 -2.00948238e-01 -2.43628547e-01 2.58513272e-01
4.49635416e-01 -5.04846036e-01 -7.85184205e-01 -5.80287576e-01
-1.88670292e-01 -2.58623332e-01 7.26094961e-01 2.25655660e-01
1.33137012e+00 -2.62748063e-01 -7.09867477e-01 5.52718341e-01
1.14809835e+00 4.17098492e-01 4.27519888e-01 6.59688413e-01
6.26325309e-01 6.48577988e-01 6.53452039e-01 4.31298643e-01
4.22640502e-01 6.73820496e-01 2.68809367e-02 -2.86426395e-01
4.89912108e-02 -4.99698490e-01 1.81090936e-01 1.48407042e-01
-2.08177208e-03 -2.53960997e-01 -8.38324666e-01 7.06281304e-01
-1.82464516e+00 -1.03008556e+00 4.09152389e-01 2.72854877e+00
8.48353446e-01 -5.23656569e-02 5.43779433e-01 -7.50985369e-02
1.21428812e+00 -8.45450908e-02 -9.69672799e-01 1.62748739e-01
-1.37936905e-01 -3.45741183e-01 4.22733635e-01 2.86264956e-01
-1.21281087e+00 8.77382517e-01 5.04806137e+00 7.91018128e-01
-1.05316484e+00 -2.64324923e-03 8.71178746e-01 -1.71995729e-01
-1.71966776e-01 4.29034233e-02 -1.06724322e+00 9.21043992e-01
4.10676569e-01 -2.70742893e-01 5.19633710e-01 9.10916150e-01
1.42277196e-01 -1.15071245e-01 -1.05773962e+00 1.15990150e+00
4.36524227e-02 -8.79132211e-01 1.80639252e-01 3.22250456e-01
7.34223664e-01 -2.96708494e-01 1.89633772e-01 3.19770992e-01
3.63813460e-01 -9.47671533e-01 6.09705985e-01 2.55446076e-01
7.32264102e-01 -8.93805921e-01 5.79215109e-01 4.97066796e-01
-1.20629978e+00 -4.05387402e-01 -4.27704841e-01 2.75499701e-01
-4.73043434e-02 5.66210508e-01 -5.14767289e-01 3.89160395e-01
1.13820362e+00 4.97919172e-01 -5.87220728e-01 8.60577166e-01
4.40182984e-02 3.31859648e-01 -3.29327315e-01 1.63065940e-01
-4.92134616e-02 -2.93117672e-01 7.49874771e-01 8.81676555e-01
2.61902183e-01 1.13432080e-01 4.68092501e-01 9.40269470e-01
-8.09974074e-02 1.31877288e-01 -5.98950803e-01 3.35349351e-01
8.03098619e-01 1.02528429e+00 -5.85171103e-01 -5.74076414e-01
-2.50060588e-01 1.10463810e+00 3.84202421e-01 6.08865440e-01
-7.03081667e-01 -2.84825325e-01 6.63931072e-01 4.06146407e-01
-8.38983990e-03 1.95139468e-01 -4.48741242e-02 -1.03908122e+00
-8.23886395e-02 -8.83137047e-01 7.81224012e-01 -3.95301044e-01
-1.59666061e+00 5.58008313e-01 2.11679742e-01 -1.39237165e+00
-9.42864269e-02 -1.76367879e-01 -5.61968327e-01 9.62463856e-01
-1.19822216e+00 -8.19884598e-01 -5.26987016e-01 8.79199684e-01
4.26252067e-01 -4.43072736e-01 4.06329155e-01 4.40017879e-01
-8.03873181e-01 1.07938540e+00 7.23947510e-02 5.17834783e-01
9.53331947e-01 -1.06894112e+00 1.52291246e-02 8.54067087e-01
-7.19409287e-02 9.33278799e-01 3.89358312e-01 -8.64272535e-01
-8.07839990e-01 -1.24205208e+00 7.45937765e-01 -1.82043552e-01
6.87693357e-02 -3.08568388e-01 -1.09235930e+00 5.86305141e-01
-2.95602083e-01 -7.04494268e-02 7.35522747e-01 2.95341443e-02
-1.64144516e-01 -3.80418181e-01 -1.42793894e+00 7.83047616e-01
1.08507180e+00 -3.19850802e-01 -4.88809466e-01 1.94740936e-01
3.35876465e-01 -1.26569077e-01 -6.30846977e-01 3.76665801e-01
3.59588295e-01 -9.57289517e-01 1.00864220e+00 -2.24777505e-01
4.21210006e-02 -6.44263268e-01 9.53705534e-02 -1.41228831e+00
-5.61869204e-01 -2.92623222e-01 7.42369443e-02 1.50241113e+00
2.06697121e-01 -1.04361057e+00 9.91486073e-01 8.26733351e-01
4.27108318e-01 -4.59251910e-01 -7.99754560e-01 -9.35809851e-01
8.82537365e-02 5.02702035e-02 9.27527010e-01 9.79036272e-01
-2.40090251e-01 8.14102069e-02 -3.04120809e-01 2.51038164e-01
7.04947770e-01 1.74359947e-01 9.13060963e-01 -1.16008496e+00
-1.73986435e-01 -4.94953185e-01 -1.22955598e-01 -1.20053124e+00
6.62845075e-02 -7.89406896e-01 3.18160616e-02 -9.66237605e-01
6.87887907e-01 -9.14317906e-01 -5.33989429e-01 5.42143822e-01
-4.22693551e-01 1.29090801e-01 2.14367136e-01 6.46945119e-01
-4.40141976e-01 4.61922884e-01 9.68317151e-01 -2.58858025e-01
-6.93929017e-01 1.47129953e-01 -9.70529974e-01 5.70554733e-01
9.69122767e-01 -5.28726816e-01 -5.02277732e-01 -2.33504295e-01
-3.36716354e-01 -4.31814313e-01 5.15471995e-01 -1.10942841e+00
4.69662040e-01 -5.41793406e-02 8.10486078e-01 -5.09435117e-01
8.42764676e-02 -6.79209828e-01 1.54189676e-01 4.88051593e-01
-4.04194325e-01 -5.11446549e-03 -5.97110949e-02 4.77383405e-01
-1.79938346e-01 -2.32246712e-01 1.07315564e+00 -2.50810295e-01
-8.59423876e-01 5.36042988e-01 -2.34765276e-01 3.69548909e-02
1.17885852e+00 -6.08801901e-01 -8.15970078e-02 -3.64833474e-01
-6.73977852e-01 4.67807561e-01 8.40782762e-01 4.75555778e-01
6.98648572e-01 -1.46627402e+00 -7.41012096e-01 4.50282246e-01
3.66946578e-01 1.18502297e-01 4.04924095e-01 5.13893783e-01
1.35762319e-01 -4.04457971e-02 -6.16012774e-02 -7.49586999e-01
-1.35570645e+00 6.97646916e-01 5.54772019e-01 -2.80590594e-01
-5.74004233e-01 8.24099183e-01 7.55750299e-01 -6.43674016e-01
2.25098565e-01 3.83648187e-01 -3.16911370e-01 -1.19057015e-01
7.75622427e-01 2.74112105e-01 -5.06974459e-01 -7.98328996e-01
-4.28320795e-01 5.44504881e-01 -4.07237321e-01 -2.10511848e-01
9.00587559e-01 -3.31758112e-01 -1.26345437e-02 1.85302034e-01
1.02730989e+00 2.86171045e-02 -1.27052677e+00 -4.02478546e-01
-2.03759938e-01 -7.46023715e-01 -2.21248180e-01 -6.97181582e-01
-1.05638313e+00 4.23778534e-01 1.09766424e+00 5.03523350e-02
1.26861596e+00 3.69730359e-03 6.51812315e-01 6.49589077e-02
2.26043269e-01 -1.34799719e+00 3.06961298e-01 2.24649936e-01
6.09304547e-01 -1.42267621e+00 -1.04050666e-01 -1.18599206e-01
-9.06301022e-01 7.34314680e-01 1.14226043e+00 3.34621710e-03
6.01066649e-01 -1.25289127e-01 1.36421323e-01 5.28665148e-02
-1.31882921e-01 -1.02223344e-01 5.27855568e-02 6.99538827e-01
-5.52439429e-02 2.02928573e-01 -4.02001105e-03 6.53859794e-01
-7.15740174e-02 -5.10058925e-02 7.14253709e-02 7.18575418e-01
-1.86521143e-01 -1.18985486e+00 -6.22019589e-01 3.19462150e-01
-8.94789770e-02 3.73467952e-02 -4.24836367e-01 7.08666325e-01
4.65578824e-01 9.53185380e-01 3.24419230e-01 -6.05672300e-01
2.42966965e-01 -1.38582394e-01 3.88077646e-02 -5.60854912e-01
-6.07244790e-01 -6.66492879e-02 -3.64361674e-01 -2.73126662e-01
-2.41714567e-01 -7.49543786e-01 -1.27830362e+00 -2.98050702e-01
-1.28121331e-01 2.66551167e-01 2.28126779e-01 6.00500524e-01
4.21088248e-01 5.76985665e-02 8.83805096e-01 -5.65123916e-01
-6.04713917e-01 -7.71927059e-01 -5.80922365e-01 7.63974547e-01
1.57344431e-01 -6.67751670e-01 -3.74665469e-01 -4.83298227e-02] | [14.825148582458496, 1.102555513381958] |
c857c244-8f51-4c5d-a814-69f868133aa1 | multiobjective-bilevel-evolutionary-approach | 2106.07318 | null | https://arxiv.org/abs/2106.07318v1 | https://arxiv.org/pdf/2106.07318v1.pdf | Multiobjective Bilevel Evolutionary Approach for Off-Grid Direction-of-Arrival Estimation | The source number identification is an essential step in direction-of-arrival (DOA) estimation. Existing methods may provide a wrong source number due to inferior statistical properties (in low SNR or limited snapshots) or modeling errors (caused by relaxing sparse penalties), especially in impulsive noise. To address this issue, we propose a novel idea of simultaneous source number identification and DOA estimation. We formulate a multiobjective off-grid DOA estimation model to realize this idea, by which the source number can be automatically identified together with DOA estimation. In particular, the source number is properly exploited by the $l_0$ norm of impinging signals without relaxations, guaranteeing accuracy. Furthermore, we design a multiobjective bilevel evolutionary algorithm to solve the proposed model. The source number identification and sparse recovery are simultaneously optimized at the on-grid (lower) level. A forward search strategy is developed to further refine the grid at the off-grid (upper) level. This strategy does not need linear approximations and can eliminate the off-grid gap with low computational complexity. Simulation results demonstrate the outperformance of our method in terms of source number and root mean square error. | ['Xin Yao', 'J. Andrew Zhang', 'Jin Zhang', 'Qi Zhao', 'Bai Yan'] | 2021-06-14 | null | null | null | null | ['direction-of-arrival-estimation'] | ['audio'] | [ 7.02725947e-02 -3.56423885e-01 2.08565563e-01 2.87573785e-01
-8.63888085e-01 -4.73271847e-01 -7.82486051e-02 -1.77638084e-02
7.87102655e-02 9.92281437e-01 1.77756086e-01 7.25130178e-03
-7.30176866e-01 -8.09354186e-01 -1.85009763e-01 -1.20244813e+00
-1.46705478e-01 -2.78528407e-02 -3.14075381e-01 -7.01859817e-02
9.54335779e-02 4.20158476e-01 -1.12813890e+00 -5.55647194e-01
1.38034558e+00 1.06819558e+00 2.30417803e-01 3.24682206e-01
8.15317482e-02 5.77857494e-01 -7.59739697e-01 2.19891340e-01
3.86427104e-01 -8.47179949e-01 2.38007516e-01 1.96114004e-01
-1.82030931e-01 -1.73601195e-01 -7.22375587e-02 1.40424967e+00
9.12791193e-01 1.61617428e-01 4.28101361e-01 -1.16421676e+00
-1.20185673e-01 3.43850553e-01 -9.45447266e-01 3.23289186e-01
1.44650638e-01 -5.72437271e-02 4.95437771e-01 -1.12711191e+00
1.65746287e-01 8.00918937e-01 8.74086797e-01 -3.30471665e-01
-9.82261121e-01 -6.62074983e-01 -1.63944170e-01 -5.83644435e-02
-2.05467463e+00 -6.14954710e-01 1.19034147e+00 -3.34380925e-01
6.70954958e-02 3.37525845e-01 6.85348809e-01 2.13100612e-01
-1.43184930e-01 3.87217075e-01 1.06211436e+00 -5.40938139e-01
2.36885771e-01 -2.22625449e-01 -1.68033689e-01 6.99949503e-01
7.35922992e-01 1.50097767e-02 -6.00098312e-01 -3.93986046e-01
8.97754133e-01 -3.77472073e-01 -6.99463606e-01 -1.83090717e-01
-1.28895593e+00 6.49693966e-01 1.03879571e-01 6.07124686e-01
-7.77765632e-01 -7.13287741e-02 -7.97724575e-02 1.09314032e-01
3.40480268e-01 5.15547931e-01 5.04794456e-02 3.65698114e-02
-1.24071348e+00 3.43498588e-02 6.66842818e-01 6.69865727e-01
6.20455444e-01 7.81980753e-01 -2.91551232e-01 9.11753774e-01
3.64827067e-01 9.54002976e-01 1.97102815e-01 -7.92241693e-01
4.18188989e-01 1.49311751e-01 5.56400180e-01 -1.62591863e+00
-3.92174065e-01 -1.30249155e+00 -1.26277113e+00 -1.15459315e-01
5.00138581e-01 -7.50306487e-01 -3.26568127e-01 1.52243757e+00
5.70013463e-01 7.35701621e-01 2.28756387e-02 1.22096264e+00
3.85068029e-01 8.92274082e-01 -4.87210453e-01 -1.09634233e+00
1.20063472e+00 -5.18274367e-01 -1.14287281e+00 -7.93911740e-02
3.62689912e-01 -9.25744951e-01 3.68275851e-01 4.82516050e-01
-1.08450282e+00 -4.39133406e-01 -9.77365792e-01 7.94890583e-01
3.39886069e-01 6.20141447e-01 2.55575091e-01 7.90433168e-01
-7.09642351e-01 1.30040748e-02 -4.54233110e-01 2.21522361e-01
-1.90515220e-02 -1.38591034e-02 2.51092374e-01 9.40586207e-04
-1.14672542e+00 4.79782283e-01 1.33465350e-01 6.59870863e-01
-3.84165585e-01 -9.37653363e-01 -6.32621706e-01 1.74090177e-01
6.72812343e-01 -6.12480044e-01 5.82456052e-01 -8.22416425e-01
-1.45085061e+00 -7.40733817e-02 -4.20243025e-01 -1.37945741e-01
3.56327832e-01 1.32720500e-01 -6.78533196e-01 2.64862031e-01
2.35859126e-01 -2.44767264e-01 8.96655440e-01 -1.38760233e+00
-4.98056650e-01 -9.29239672e-03 -5.13759971e-01 3.76002461e-01
-2.22141892e-01 -2.34175339e-01 -1.50122449e-01 -1.18148863e+00
7.07978368e-01 -6.07148528e-01 -5.07930160e-01 -2.77056843e-01
-3.12232107e-01 2.03089595e-01 3.96333098e-01 -8.20122719e-01
1.52885377e+00 -2.20147777e+00 7.06410408e-02 6.51423812e-01
8.09176713e-02 2.01232076e-01 5.66087489e-04 4.98125672e-01
5.49448729e-02 -2.04367712e-01 -4.01861459e-01 9.39729512e-02
-3.56364161e-01 -2.64810562e-01 -1.06873892e-01 7.60731757e-01
-4.80948091e-02 1.04400568e-01 -9.36762750e-01 -4.10273045e-01
1.65366650e-01 4.41697121e-01 -6.41818762e-01 1.22201540e-01
2.63735145e-01 7.50408530e-01 -7.88453162e-01 7.31222987e-01
1.05056560e+00 -1.57685474e-01 1.68695539e-01 -5.61070025e-01
-5.48272669e-01 -2.23589763e-01 -2.05447102e+00 1.45060086e+00
-6.67268336e-01 2.50295132e-01 6.95429862e-01 -1.21170437e+00
1.11699176e+00 4.19723332e-01 9.32367086e-01 -4.86923039e-01
1.49205834e-01 5.51533043e-01 1.70709103e-01 -3.90339434e-01
7.34794214e-02 -1.98917493e-01 2.44633351e-02 2.00770423e-01
-3.25034231e-01 -6.98429123e-02 1.72770426e-01 -2.36703709e-01
6.26866400e-01 -3.75654250e-01 6.21106625e-01 -5.84457934e-01
9.32917237e-01 -5.12228645e-02 9.35557902e-01 7.21495688e-01
6.68589696e-02 4.02578562e-01 1.93454787e-01 2.14281529e-02
-6.12770557e-01 -4.33369488e-01 -1.00017779e-01 2.38541394e-01
4.07619536e-01 -1.08824261e-01 -4.42181885e-01 -7.95645174e-03
-1.34089202e-01 7.41567969e-01 -1.88240275e-01 -4.24258374e-02
-7.55307615e-01 -1.04766405e+00 3.17311883e-01 -6.62774146e-02
7.34504580e-01 -2.44190872e-01 -4.24875408e-01 5.91251493e-01
-3.09463382e-01 -8.88407946e-01 -3.78099710e-01 5.27167600e-03
-6.21730149e-01 -8.03891659e-01 -1.02701032e+00 -6.30891323e-01
7.72598267e-01 5.73920727e-01 5.94634891e-01 7.71503076e-02
1.75743073e-01 2.74901390e-01 -3.82254124e-01 -2.37813175e-01
-2.47664414e-02 -2.98720390e-01 2.98147827e-01 6.71866298e-01
-4.07621086e-01 -7.71892726e-01 -5.60929298e-01 5.21204054e-01
-5.25480747e-01 -1.50943071e-01 6.06544375e-01 9.13672268e-01
6.41533136e-01 7.53636301e-01 9.99204516e-01 -8.76293983e-03
7.60509908e-01 -7.37160802e-01 -1.05313587e+00 9.57341939e-02
-3.76159161e-01 -3.15049052e-01 8.57908905e-01 -2.98554748e-01
-1.09562898e+00 3.09453905e-03 -7.66521767e-02 -4.90499467e-01
2.14048415e-01 9.55923378e-01 -3.54092896e-01 -4.93256927e-01
2.94253379e-01 6.90508723e-01 -1.40344158e-01 -5.27856648e-01
-1.81559213e-02 4.50385243e-01 2.26174489e-01 -3.56248468e-01
1.13725114e+00 3.53649974e-01 4.70638901e-01 -1.03128111e+00
-6.81997597e-01 -4.60392207e-01 -3.05941552e-01 -3.05994451e-01
2.15016976e-01 -9.60781395e-01 -7.63827801e-01 4.56543267e-01
-9.34898138e-01 2.23227397e-01 -7.81160742e-02 1.05327117e+00
-1.44021854e-01 4.84407037e-01 1.56476013e-02 -1.20408595e+00
-1.75313085e-01 -9.15858328e-01 5.18719018e-01 2.86500245e-01
6.13520965e-02 -7.51290321e-01 -5.32023497e-02 -8.95352885e-02
5.18155515e-01 4.52220052e-01 5.10136127e-01 -1.11561231e-01
-5.45756042e-01 -5.20576090e-02 -1.75492331e-01 4.06286083e-02
2.75058478e-01 -4.59531873e-01 -2.09363833e-01 -3.28910917e-01
4.81891125e-01 3.64777505e-01 3.22176814e-01 8.20264578e-01
8.21650207e-01 -5.66187620e-01 -3.56376499e-01 8.84374380e-01
1.78970706e+00 4.61137205e-01 3.15219373e-01 1.66113108e-01
5.76264203e-01 2.44227752e-01 7.64240026e-01 1.02092481e+00
5.83973080e-02 8.49152803e-01 2.63442904e-01 -8.27013925e-02
-2.70714164e-01 -2.25882865e-02 -2.15770062e-02 1.07311332e+00
6.56925887e-02 -3.94462883e-01 -7.73885965e-01 6.29772246e-01
-1.57183433e+00 -8.93225551e-01 -4.61502612e-01 2.34434867e+00
7.19101548e-01 -3.72017294e-01 7.57851154e-02 5.10248303e-01
9.19575691e-01 2.21896529e-01 -3.41054887e-01 1.77644446e-01
-3.03504109e-01 -1.91742346e-01 6.19067073e-01 6.43595815e-01
-7.34853327e-01 1.32591903e-01 5.36627865e+00 1.19504750e+00
-1.20708764e+00 3.06674540e-02 3.35008562e-01 9.92192551e-02
-5.33975661e-01 -5.35299294e-02 -6.44526184e-01 8.61411095e-01
3.15930218e-01 -4.50771242e-01 4.86676455e-01 4.16795433e-01
7.12056577e-01 -3.63133878e-01 -2.90741950e-01 1.17137420e+00
6.57205135e-02 -1.39969695e+00 -2.75612533e-01 -1.40121328e-02
9.17388618e-01 -6.94002688e-01 -2.54699975e-01 -1.91214070e-01
-1.89040840e-01 -7.02077031e-01 7.47975886e-01 6.54112339e-01
7.31863499e-01 -9.53139186e-01 6.93642974e-01 6.17489934e-01
-1.36641276e+00 -2.56960988e-01 -1.63716435e-01 3.97324674e-02
6.69423819e-01 1.36247754e+00 -5.93900025e-01 1.15223181e+00
2.80523002e-01 5.55618823e-01 1.11570261e-01 1.56859076e+00
-1.92291155e-01 7.90777504e-01 -5.63581169e-01 -8.94515216e-02
1.97339997e-01 -7.66207576e-01 1.30880034e+00 8.16169024e-01
1.15321004e+00 7.17945039e-01 4.24088269e-01 6.94695473e-01
2.82207280e-01 3.26477259e-01 -2.97146440e-01 3.43035072e-01
9.81595993e-01 8.70982051e-01 -7.09194541e-01 -5.82756102e-02
-2.57098287e-01 3.74701917e-01 -5.61411500e-01 6.78641737e-01
-7.91236520e-01 -5.23520291e-01 3.35654557e-01 1.55122250e-01
3.68795305e-01 -4.37285751e-01 -4.47389632e-01 -1.04594815e+00
7.75531447e-03 -9.58862603e-01 2.19985574e-01 -4.74100471e-01
-8.95776868e-01 4.41236615e-01 -1.74944207e-01 -1.74102378e+00
-1.52564973e-01 1.41754314e-01 -6.26545429e-01 9.01080906e-01
-1.56974173e+00 -6.06815338e-01 -2.46578366e-01 5.48368573e-01
3.46491724e-01 -1.20792009e-01 4.89875853e-01 7.50142395e-01
-6.53409481e-01 5.09156108e-01 4.54909116e-01 -8.87013301e-02
2.39385441e-01 -5.38744807e-01 -6.93531215e-01 1.26746380e+00
-1.96448147e-01 3.60886455e-01 1.03485334e+00 -6.72635257e-01
-1.62477148e+00 -7.91032851e-01 7.01232076e-01 5.32881796e-01
5.12504697e-01 5.72900251e-02 -5.79617023e-01 -1.27867970e-03
-2.69352868e-02 1.06630810e-01 5.98871291e-01 -3.17142695e-01
3.54428172e-01 -4.01902437e-01 -1.15356827e+00 4.93218452e-01
6.61876380e-01 4.82164882e-02 -1.62644193e-01 3.53768855e-01
3.21911603e-01 -4.42564756e-01 -8.53070378e-01 5.84390819e-01
2.09023416e-01 -8.36130738e-01 1.14043140e+00 2.08939567e-01
-7.30837137e-02 -9.91479456e-01 -2.26625592e-01 -1.47732949e+00
-4.46341902e-01 -9.88691151e-01 -1.61239818e-01 1.36670697e+00
5.78422882e-02 -7.96382606e-01 4.61769402e-01 -2.44881317e-01
-1.46011859e-01 -6.33739769e-01 -1.18160057e+00 -1.08144927e+00
-4.99824315e-01 -1.63372517e-01 5.68628669e-01 1.13699996e+00
-1.73791066e-01 4.50544804e-02 -7.57912159e-01 9.62813556e-01
1.01830673e+00 3.39328796e-01 5.99657655e-01 -9.99472439e-01
-3.97880584e-01 -4.03122962e-01 1.68583825e-01 -1.13401031e+00
-2.94820428e-01 -5.51234245e-01 6.29685745e-02 -1.54237914e+00
-3.37727845e-01 -8.94635201e-01 -2.08804548e-01 -2.83778645e-02
-2.47249365e-01 1.77774519e-01 1.50964200e-01 3.31944555e-01
-1.76000178e-01 6.99180067e-01 1.29489243e+00 4.83836234e-02
-4.13547844e-01 3.81614268e-01 -5.67761898e-01 8.57114255e-01
6.28006697e-01 -6.44697011e-01 -3.13364178e-01 -4.49473411e-01
2.81079859e-01 8.44604373e-01 2.08407313e-01 -1.37583017e+00
3.04961205e-01 -1.65554434e-01 1.52864262e-01 -6.81521237e-01
2.75266826e-01 -9.55723345e-01 5.32139719e-01 5.62898874e-01
1.30130440e-01 -4.24187630e-01 -3.72919962e-02 6.31500781e-01
-5.08345127e-01 -4.23974067e-01 6.82724714e-01 1.87568665e-01
-3.43970060e-01 8.52908641e-02 -5.10733008e-01 -1.78288251e-01
9.67111409e-01 -2.39314064e-01 1.02126703e-01 -7.31161654e-01
-4.71585453e-01 2.61865318e-01 5.04855346e-03 -2.99550623e-01
4.37139213e-01 -1.48483264e+00 -9.88047600e-01 1.62523746e-01
-3.45728606e-01 -1.83067858e-01 5.85866094e-01 1.29239225e+00
-3.16782027e-01 2.95398206e-01 1.33336142e-01 -5.02414882e-01
-7.69836426e-01 1.86254084e-01 4.56300944e-01 -2.28542551e-01
-1.51252076e-01 7.93258548e-01 -1.75295308e-01 -6.20603040e-02
-1.04411401e-01 1.21359855e-01 -4.03142571e-01 4.00044650e-01
4.89686847e-01 8.00098777e-01 -4.14523296e-02 -6.68645144e-01
-4.76333797e-01 9.96632278e-01 9.26291764e-01 -7.38245174e-02
1.26247084e+00 -5.31521678e-01 -3.15057725e-01 -3.62135284e-02
1.16179502e+00 1.06177354e+00 -1.02172160e+00 -1.91450492e-01
-5.04222393e-01 -8.39659572e-01 3.70864153e-01 -6.65991545e-01
-1.23681593e+00 3.74931306e-01 4.60264504e-01 3.20921540e-01
1.52083874e+00 -5.16552150e-01 8.10464621e-01 1.65164739e-01
4.33756590e-01 -8.32739055e-01 -1.08348496e-01 2.48357087e-01
9.46442783e-01 -7.39759147e-01 2.09975988e-01 -7.21343696e-01
-1.60342172e-01 1.04355240e+00 1.67341217e-01 -1.86759278e-01
6.40057147e-01 3.68156493e-01 -7.50710219e-02 1.46230385e-02
1.95179507e-02 -1.75066635e-01 2.18266189e-01 3.54092658e-01
1.63132802e-01 -4.61250618e-02 -8.21102500e-01 7.63555944e-01
4.78762947e-02 -2.26954930e-02 6.18640780e-01 5.43953061e-01
-6.11381233e-01 -8.32863331e-01 -1.14647341e+00 2.86355108e-01
-4.27898914e-01 -2.33652651e-01 5.08391201e-01 3.30971599e-01
1.90976709e-01 1.23786342e+00 -2.96246052e-01 2.66420413e-02
3.89875561e-01 -4.28145856e-01 8.30589831e-02 -2.05766052e-01
-1.17993169e-01 6.41878605e-01 2.88372517e-01 -1.60343155e-01
-3.65007013e-01 -7.68467784e-01 -9.82674599e-01 -3.16830613e-02
-7.66223490e-01 7.33418107e-01 4.17828113e-01 8.23667109e-01
4.04481262e-01 6.10811293e-01 1.17085779e+00 -5.78123331e-01
-4.47231412e-01 -5.85923374e-01 -8.03361475e-01 -3.77154589e-01
3.30999732e-01 -6.53627276e-01 -6.48330748e-01 -1.16766684e-01] | [6.465387344360352, 1.3497486114501953] |
d3a1bf7c-4a52-443a-835c-fd4c56ad1b0b | an-inter-observer-consistent-deep-adversarial | 2211.07336 | null | https://arxiv.org/abs/2211.07336v2 | https://arxiv.org/pdf/2211.07336v2.pdf | An Inter-observer consistent deep adversarial training for visual scanpath prediction | The visual scanpath is a sequence of points through which the human gaze moves while exploring a scene. It represents the fundamental concepts upon which visual attention research is based. As a result, the ability to predict them has emerged as an important task in recent years. In this paper, we propose an inter-observer consistent adversarial training approach for scanpath prediction through a lightweight deep neural network. The adversarial method employs a discriminative neural network as a dynamic loss that is better suited to model the natural stochastic phenomenon while maintaining consistency between the distributions related to the subjective nature of scanpaths traversed by different observers. Through extensive testing, we show the competitiveness of our approach in regard to state-of-the-art methods. | ['Alessandro Bruno', 'Aladine Chetouani', 'Marouane Tliba', 'Mohamed Amine Kerkouri'] | 2022-11-14 | null | null | null | null | ['scanpath-prediction'] | ['computer-vision'] | [ 1.22384220e-01 -7.78013319e-02 -1.94741949e-01 -5.25205553e-01
-3.92070591e-01 -4.41041559e-01 5.89573145e-01 -1.12257637e-01
-4.68298048e-01 2.97784179e-01 1.64525323e-02 -2.33111575e-01
-1.18488729e-01 -3.75260442e-01 -8.77144933e-01 -4.71293539e-01
-6.39947876e-02 4.10831779e-01 4.60715473e-01 -2.42968455e-01
4.24575180e-01 6.36156440e-01 -1.53071606e+00 1.84358191e-02
6.81066871e-01 1.07453060e+00 -2.05354214e-01 5.11693239e-01
6.40754029e-02 1.00572634e+00 -4.32775348e-01 -5.03658712e-01
3.64890575e-01 -5.10230005e-01 -8.46566796e-01 -1.67869344e-01
8.42207193e-01 -4.55608442e-02 -6.79075658e-01 1.28794348e+00
1.81980610e-01 4.41958368e-01 6.29248798e-01 -1.25342858e+00
-6.82597339e-01 1.06629483e-01 -8.09632361e-01 6.27904654e-01
5.23888111e-01 5.23760855e-01 9.97815192e-01 -5.52391410e-01
6.02408051e-01 1.18612278e+00 5.34991026e-01 6.45821095e-01
-1.44281673e+00 -5.14184475e-01 2.65007704e-01 8.68972182e-01
-1.31592667e+00 -1.90075383e-01 8.93175602e-01 -4.71352518e-01
4.18647915e-01 2.78750360e-01 6.16302371e-01 1.64424992e+00
6.03326917e-01 8.90513122e-01 1.07096922e+00 -2.49129102e-01
9.99070182e-02 1.18970364e-01 -1.04160450e-01 4.98088747e-01
-2.05207765e-02 4.69996125e-01 -4.73743260e-01 1.26687407e-01
4.78713989e-01 -1.51004139e-02 -3.36412042e-01 -6.16064966e-01
-6.32358968e-01 7.02935696e-01 7.38886654e-01 2.71996874e-02
-2.27414966e-01 2.96772242e-01 2.84196466e-01 1.07848883e-01
5.10678053e-01 4.49387938e-01 1.19090699e-01 -2.31702998e-01
-1.05324304e+00 3.93058687e-01 6.88344181e-01 7.56588280e-01
4.88712013e-01 1.05914004e-01 -4.84515429e-01 4.26417232e-01
3.20430458e-01 4.88674462e-01 4.28393781e-01 -8.49874973e-01
1.55995399e-01 6.17265582e-01 1.44385612e-02 -1.18131232e+00
-2.85890013e-01 -3.59918654e-01 -5.82673490e-01 6.76784098e-01
4.28193033e-01 3.42262179e-01 -9.22419131e-01 1.87776065e+00
3.30456048e-01 2.67367929e-01 -5.81610620e-01 9.87254441e-01
3.31559688e-01 3.13612342e-01 3.17051500e-01 3.49229932e-01
1.06612778e+00 -1.00495863e+00 -6.07032180e-01 -1.99110448e-01
-2.05152154e-01 -5.80537438e-01 1.34914041e+00 3.22039217e-01
-1.21040702e+00 -6.43896222e-01 -8.43616903e-01 -1.49405509e-01
-2.61478096e-01 -5.10894179e-01 3.18894446e-01 5.09128034e-01
-1.07156813e+00 6.17562652e-01 -9.65742469e-01 -2.47816846e-01
6.90244198e-01 6.59288466e-02 -2.98984855e-01 -7.03891460e-03
-1.04052150e+00 1.12774539e+00 1.18597783e-01 -1.23435147e-01
-1.14496589e+00 -7.28882432e-01 -7.52652824e-01 1.31292284e-01
3.01741660e-01 -5.63519418e-01 1.27918994e+00 -1.21281564e+00
-1.40761375e+00 1.18333054e+00 -4.03869927e-01 -6.72128916e-01
9.09884512e-01 -3.19974363e-01 -4.75647956e-01 1.40084565e-01
1.02369236e-02 5.28766632e-01 1.34154320e+00 -1.32335114e+00
-6.39872551e-01 -2.88877517e-01 2.42427856e-01 9.97589007e-02
9.40419137e-02 -5.72306067e-02 -4.80576068e-01 -5.62032104e-01
-2.06337824e-01 -1.08260918e+00 -8.77246186e-02 4.28838313e-01
-5.50363660e-01 -3.79559100e-01 7.66678393e-01 -4.28620696e-01
1.12099445e+00 -2.17487788e+00 2.81300545e-01 2.55686849e-01
4.81681675e-01 2.90724397e-01 -1.17905447e-02 1.69097409e-01
-1.99793220e-01 8.34718570e-02 -2.37300768e-01 -5.54646790e-01
8.78146943e-03 1.06500201e-01 -4.00196046e-01 8.03994715e-01
2.13688865e-01 9.39707458e-01 -9.20920014e-01 -2.65641958e-01
1.80016294e-01 3.81351352e-01 -3.51723790e-01 6.40361369e-01
-3.38452458e-01 8.53198886e-01 -1.87768355e-01 6.07273459e-01
6.93710208e-01 -3.76482695e-01 -4.69348788e-01 1.56439897e-02
4.01467122e-02 -1.63448781e-01 -5.73010623e-01 1.63815928e+00
-3.40287089e-01 1.08415163e+00 -4.79189128e-01 -7.15660989e-01
5.37732780e-01 -1.21055424e-01 1.78712502e-01 -1.08035028e+00
3.60920280e-01 -1.70946941e-01 7.91622698e-02 -7.80380130e-01
6.22662187e-01 -1.01959705e-01 1.91414937e-01 4.02120143e-01
-1.28948659e-01 1.28367692e-01 2.17495710e-02 -2.71537099e-02
1.05297399e+00 7.57186115e-02 4.51257139e-01 -6.41581938e-02
6.56178713e-01 -1.34463340e-01 2.44027272e-01 7.70006061e-01
-6.92842603e-01 6.91874623e-01 5.52625895e-01 -6.93499625e-01
-1.13988328e+00 -1.33038807e+00 -1.71862051e-01 9.68421340e-01
6.15011811e-01 3.04070145e-01 -7.49975920e-01 -7.70852923e-01
-7.24249557e-02 9.81430590e-01 -1.26947892e+00 -4.63604569e-01
-5.42766511e-01 1.88735537e-02 6.10142112e-01 5.20632267e-01
3.45308900e-01 -1.26316226e+00 -8.38971019e-01 -2.34266371e-01
-6.63005486e-02 -1.10270250e+00 -5.87274253e-01 1.60837267e-02
-3.43616366e-01 -1.33439064e+00 -6.39712393e-01 -2.65447408e-01
5.12455404e-01 2.71016434e-02 1.32255244e+00 -1.79416806e-01
-4.13030088e-01 6.45954669e-01 -2.94007719e-01 -5.39552331e-01
-3.70086610e-01 8.53825137e-02 -1.67387262e-01 2.94345587e-01
5.84934890e-01 -5.75794339e-01 -9.29208338e-01 2.48324499e-01
-9.82440233e-01 -2.53176361e-01 3.47365469e-01 6.46247387e-01
7.56647110e-01 -2.02259868e-01 -2.69169390e-01 -9.12543476e-01
5.15383422e-01 -9.77195084e-01 -7.46288836e-01 1.47292420e-01
-6.09842777e-01 2.03504637e-01 6.48275673e-01 -5.14864206e-01
-1.01780093e+00 -2.89999098e-01 -2.70649400e-02 -8.51987302e-01
-4.62808609e-01 1.07066527e-01 1.03354067e-01 -4.41479236e-01
7.14940310e-01 3.07708442e-01 -8.63394141e-02 -1.83919042e-01
3.27940196e-01 4.41605523e-02 5.52096188e-01 -8.54541436e-02
1.12845075e+00 8.40431690e-01 3.93791169e-01 -5.97039878e-01
-1.13917553e+00 -5.93079686e-01 -5.35305917e-01 -4.04733270e-01
1.07373977e+00 -5.29684186e-01 -9.36998427e-01 3.83574456e-01
-1.19815910e+00 -4.51148152e-01 -4.03931469e-01 2.30768085e-01
-7.87393808e-01 2.76316196e-01 4.46789227e-02 -6.95522904e-01
-5.02593257e-02 -1.11123013e+00 8.59291911e-01 5.48311770e-01
-4.26931053e-01 -1.32878435e+00 5.33704638e-01 1.55780226e-01
5.10599673e-01 6.06571615e-01 6.87267721e-01 -6.55342877e-01
-8.83689106e-01 -2.68679738e-01 -2.51072258e-01 3.26576710e-01
-1.40937641e-01 -5.94697222e-02 -1.01875353e+00 -3.11904609e-01
1.48907259e-01 -3.09146196e-01 7.00851142e-01 4.36600029e-01
1.68400586e+00 -3.65927629e-02 -2.72128969e-01 1.06292093e+00
1.52725899e+00 3.20429020e-02 7.34702408e-01 5.60313940e-01
8.56989443e-01 5.14876902e-01 4.89536405e-01 2.63621688e-01
3.19791764e-01 7.62618899e-01 1.12807131e+00 -1.39467672e-01
-9.31012407e-02 -4.00063396e-01 5.96358366e-02 -1.15884012e-02
-4.72051837e-02 -6.24018848e-01 -9.87789035e-01 5.62086046e-01
-1.64743376e+00 -1.12728906e+00 -3.81886475e-02 2.20842981e+00
3.36877733e-01 7.24795833e-02 3.55021283e-02 -6.01411089e-02
5.57144880e-01 4.90154356e-01 -1.00522792e+00 -5.93540967e-01
1.05708659e-01 3.82919103e-01 5.29155374e-01 1.91498131e-01
-1.04223192e+00 7.79129982e-01 6.28083467e+00 6.80145860e-01
-1.16671991e+00 1.06135853e-01 4.51519728e-01 -8.44189525e-02
-1.81720585e-01 -2.21478865e-01 -4.90256071e-01 7.47658134e-01
8.62475276e-01 -2.27860302e-01 5.54421604e-01 9.99143839e-01
2.39326864e-01 -1.47999212e-01 -1.37198615e+00 9.24908102e-01
4.00576472e-01 -8.28734815e-01 7.76624084e-02 -5.93791204e-03
8.06252480e-01 2.84806937e-01 7.97096610e-01 -8.79714731e-03
2.06724152e-01 -1.45620847e+00 9.73484159e-01 7.99097300e-01
8.41408432e-01 -6.49221599e-01 5.71320176e-01 2.44389236e-01
-9.90688622e-01 -6.41210601e-02 -2.37578750e-01 2.16200024e-01
2.90643275e-01 1.64123297e-01 -5.31957686e-01 2.34269440e-01
8.32430601e-01 5.78488290e-01 -6.80191755e-01 1.30877328e+00
-5.21920264e-01 6.15264654e-01 8.96181911e-02 9.40281451e-02
4.53348488e-01 1.59590825e-01 9.10572171e-01 8.93356442e-01
8.35147500e-02 -3.27554643e-01 -1.14814237e-01 1.17804933e+00
-2.25836545e-01 -1.36206448e-01 -7.02041030e-01 3.70024860e-01
1.99389681e-01 8.66051316e-01 -3.99390310e-01 1.16137393e-01
-3.50907505e-01 1.09641159e+00 6.37026548e-01 4.65143621e-01
-1.07672560e+00 -1.90156773e-01 7.76784837e-01 3.94126028e-01
3.21480304e-01 4.64307964e-02 -2.35689327e-01 -1.09367049e+00
2.13525519e-01 -7.92071402e-01 2.49887943e-01 -7.15334654e-01
-1.51339579e+00 8.29039872e-01 -1.04987942e-01 -1.56376326e+00
-1.34790689e-01 -4.90822822e-01 -8.55203867e-01 1.01368082e+00
-1.88659632e+00 -1.01435590e+00 -7.68004477e-01 7.51260459e-01
6.25444055e-01 -7.92955682e-02 5.76203823e-01 -4.75297347e-02
-2.51473993e-01 9.00559902e-01 -7.55383596e-02 1.11456187e-02
5.45354545e-01 -1.34123552e+00 6.37378931e-01 1.02191782e+00
2.99944520e-01 5.38419306e-01 9.37871635e-01 -2.38883734e-01
-7.44495749e-01 -1.00112998e+00 5.65457702e-01 -7.45326638e-01
6.74587250e-01 3.36804353e-02 -1.28799844e+00 7.75636017e-01
3.78136247e-01 2.83148170e-01 6.97173834e-01 -1.55967355e-01
-4.99135345e-01 -1.09699927e-01 -1.10620058e+00 7.15155244e-01
9.11167502e-01 -6.45957589e-01 -5.46743035e-01 -9.83471498e-02
5.05458891e-01 -7.57949412e-01 -3.73485684e-01 7.61196986e-02
6.27674878e-01 -1.40153039e+00 8.68481874e-01 -7.89984167e-01
6.41153634e-01 -6.64950013e-02 1.71149194e-01 -1.42106450e+00
-4.03428584e-01 -5.37902057e-01 -2.66146153e-01 6.51999652e-01
1.09880827e-01 -4.88716692e-01 9.44481075e-01 6.36293292e-01
1.51834533e-01 -6.46257102e-01 -1.06865358e+00 -7.36402333e-01
-7.16924444e-02 -3.84171367e-01 4.43396538e-01 6.00665629e-01
-4.82935905e-01 -1.16090231e-01 -5.21569371e-01 2.00975865e-01
9.27006006e-01 -2.48238593e-01 8.62762153e-01 -1.23760366e+00
-1.94915667e-01 -6.03784561e-01 -7.62264967e-01 -9.92447078e-01
6.21936500e-01 -7.45173097e-01 1.20171033e-01 -8.65038633e-01
2.15596497e-01 -8.46746787e-02 -5.95371306e-01 4.88910861e-02
-3.30182165e-01 1.91209942e-01 2.02549860e-01 2.61119902e-01
-7.18638420e-01 6.90441132e-01 1.34042776e+00 -1.77467480e-01
4.97581102e-02 3.73680681e-01 -4.03160691e-01 8.53585482e-01
8.67289722e-01 -7.76849508e-01 -4.92351353e-01 -5.70416212e-01
3.11042160e-01 -1.90998971e-01 6.01827741e-01 -1.15643764e+00
3.70949566e-01 -2.36945868e-01 1.89734265e-01 -3.85775208e-01
3.11407506e-01 -1.13002598e+00 -6.08532131e-02 4.15374815e-01
-5.79085469e-01 1.77873492e-01 2.80084491e-01 1.03393447e+00
-2.95717448e-01 -1.17621444e-01 1.13311374e+00 -7.52285272e-02
-9.34168100e-01 7.04214633e-01 -1.32878855e-01 5.07623851e-01
1.27389002e+00 -3.49960268e-01 -1.60637796e-02 -4.70402628e-01
-5.91772795e-01 1.38385728e-01 6.46710098e-01 5.98434627e-01
6.90912545e-01 -1.07945967e+00 -4.37812865e-01 1.92271337e-01
4.34109569e-01 -1.41734585e-01 4.85933155e-01 8.56768906e-01
-6.57306194e-01 2.09716111e-01 -5.24512589e-01 -8.30470443e-01
-1.20985174e+00 6.91648960e-01 3.80563855e-01 -3.93616021e-01
-4.00318682e-01 9.83295023e-01 2.44519219e-01 1.85299471e-01
3.00369769e-01 -5.70745394e-03 -4.07396793e-01 -2.98522115e-01
6.11640394e-01 5.71681082e-01 -2.75775492e-01 -7.12208033e-01
-2.57257342e-01 5.26675582e-01 -8.56622830e-02 5.85056357e-02
1.13514245e+00 -4.51211721e-01 7.99368247e-02 7.30220020e-01
1.18603075e+00 -9.75890234e-02 -1.60555339e+00 -2.62310743e-01
-4.40013595e-02 -7.88217723e-01 -7.36247376e-02 -6.43815041e-01
-1.00215662e+00 1.06211114e+00 8.47597539e-01 2.45561346e-01
1.02544963e+00 2.57283859e-02 6.80409193e-01 -5.39204553e-02
2.77093858e-01 -6.52689934e-01 2.60974228e-01 3.75485420e-01
7.66912460e-01 -1.64528072e+00 -3.38790059e-01 -4.11892310e-02
-7.44932532e-01 9.91042137e-01 7.67607093e-01 -3.04793000e-01
5.65284669e-01 -2.07771689e-01 1.55005112e-01 -3.39625657e-01
-4.65605140e-01 -2.57179707e-01 5.59876621e-01 8.40413392e-01
1.77907962e-02 -1.17323659e-01 9.51011330e-02 4.15852182e-02
-2.24540718e-02 -2.15192825e-01 4.54161674e-01 4.74409044e-01
-7.32975230e-02 -8.70691955e-01 -1.91823274e-01 1.65816605e-01
-5.54697335e-01 1.16746931e-03 -9.47091132e-02 9.27978396e-01
8.74862820e-02 5.35431325e-01 1.85069978e-01 -2.54937679e-01
4.25663441e-01 -3.17348510e-01 4.13760275e-01 -5.76802254e-01
-3.81277293e-01 -7.51190484e-01 -6.57047272e-01 -1.14858913e+00
-4.33417946e-01 -5.02122641e-01 -7.74254084e-01 -4.68702406e-01
1.28024280e-01 -2.72342980e-01 4.16663826e-01 8.02780032e-01
1.30388886e-01 6.45219445e-01 8.61943603e-01 -8.67043197e-01
-6.10705972e-01 -7.56755888e-01 -6.03990793e-01 8.43596399e-01
7.11784959e-01 -9.10845518e-01 -5.44696867e-01 4.87184562e-02] | [10.085067749023438, 1.2011289596557617] |
43963cda-24ff-4259-bb5e-6967e0b2083b | nsurl-2019-task-8-semantic-question | null | null | https://aclanthology.org/2019.nsurl-1.1 | https://aclanthology.org/2019.nsurl-1.1.pdf | NSURL-2019 Task 8: Semantic Question Similarity in Arabic | null | ['Hussein T. Al-Natsheh', 'Wael Farhan', 'Hesham Al-Bataineh', 'Ahmad Mustafa', 'Haitham Seelawi'] | null | null | null | null | nsurl-2019-9 | ['question-similarity'] | ['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] | [-7.231813907623291, 3.763392210006714] |
71d53075-f015-4460-8893-afafd33367aa | pushing-the-limits-of-3d-shape-generation-at | 2306.11510 | null | https://arxiv.org/abs/2306.11510v1 | https://arxiv.org/pdf/2306.11510v1.pdf | Pushing the Limits of 3D Shape Generation at Scale | We present a significant breakthrough in 3D shape generation by scaling it to unprecedented dimensions. Through the adaptation of the Auto-Regressive model and the utilization of large language models, we have developed a remarkable model with an astounding 3.6 billion trainable parameters, establishing it as the largest 3D shape generation model to date, named Argus-3D. Our approach addresses the limitations of existing methods by enhancing the quality and diversity of generated 3D shapes. To tackle the challenges of high-resolution 3D shape generation, our model incorporates tri-plane features as latent representations, effectively reducing computational complexity. Additionally, we introduce a discrete codebook for efficient quantization of these representations. Leveraging the power of transformers, we enable multi-modal conditional generation, facilitating the production of diverse and visually impressive 3D shapes. To train our expansive model, we leverage an ensemble of publicly-available 3D datasets, consisting of a comprehensive collection of approximately 900,000 objects from renowned repositories such as ModelNet40, ShapeNet, Pix3D, 3D-Future, and Objaverse. This diverse dataset empowers our model to learn from a wide range of object variations, bolstering its ability to generate high-quality and diverse 3D shapes. Extensive experimentation demonstrate the remarkable efficacy of our approach in significantly improving the visual quality of generated 3D shapes. By pushing the boundaries of 3D generation, introducing novel methods for latent representation learning, and harnessing the power of transformers for multi-modal conditional generation, our contributions pave the way for substantial advancements in the field. Our work unlocks new possibilities for applications in gaming, virtual reality, product design, and other domains that demand high-quality and diverse 3D objects. | ['Yanwei Fu', 'Bo Zhao', 'Tiejun Huang', 'Jingyang Huo', 'Xuelin Qian', 'Wang Yu'] | 2023-06-20 | null | null | null | null | ['3d-shape-generation', 'quantization'] | ['computer-vision', 'methodology'] | [-8.25454369e-02 -2.67057344e-02 -2.70787696e-03 -5.14702313e-02
-8.57611775e-01 -9.88383651e-01 8.56538653e-01 -3.07689101e-01
4.25065368e-01 3.24932992e-01 4.81839299e-01 -3.31777662e-01
1.77838039e-02 -1.12724364e+00 -8.07987690e-01 -4.11315233e-01
5.51811494e-02 6.00387156e-01 -1.31444275e-01 -4.74658549e-01
1.08775921e-01 9.30314243e-01 -1.81981647e+00 3.30102533e-01
7.07343102e-01 9.75349724e-01 1.31884351e-01 3.82691383e-01
-3.01952660e-01 7.87870809e-02 -4.10969079e-01 -4.24138874e-01
4.32014734e-01 2.94710714e-02 -3.44046980e-01 2.43225917e-01
5.07480800e-01 -4.26727027e-01 -7.12043643e-02 3.77410322e-01
7.53895402e-01 -9.52027515e-02 9.15795386e-01 -1.08871603e+00
-1.29528964e+00 2.50532955e-01 -3.15655589e-01 -4.66159523e-01
4.89837557e-01 5.32557368e-01 1.12543905e+00 -1.21686947e+00
8.16786408e-01 1.38169181e+00 7.16763318e-01 7.66635180e-01
-1.42913771e+00 -7.21827745e-01 1.61161851e-02 -4.99264389e-01
-1.42397821e+00 -4.77553576e-01 8.59985292e-01 -5.78347385e-01
1.08715773e+00 7.27618560e-02 9.51190591e-01 1.37001348e+00
-3.30769643e-03 9.81476665e-01 8.50571156e-01 -2.16935262e-01
1.69010296e-01 -1.37617931e-01 -7.14523196e-01 6.62214875e-01
8.58616605e-02 3.45616519e-01 -4.60715324e-01 -2.56778449e-01
1.32081139e+00 -1.41877025e-01 2.40458444e-01 -7.49042094e-01
-1.24733150e+00 7.14412212e-01 3.17437053e-01 8.47894177e-02
-2.92416245e-01 2.55503923e-01 -4.44665458e-03 -2.26745322e-01
5.72806001e-01 5.94218433e-01 -4.64150131e-01 -2.72667348e-01
-7.18532264e-01 7.58167863e-01 6.20397329e-01 1.35075188e+00
4.89022642e-01 4.36324418e-01 -2.68286914e-01 8.28382015e-01
3.55835289e-01 9.54263270e-01 1.52590992e-02 -1.03340137e+00
4.30718303e-01 7.68518090e-01 1.38615638e-01 -8.89934599e-01
-1.81868598e-01 -3.92226875e-01 -7.49682724e-01 4.18101639e-01
-4.58913147e-02 9.10975933e-02 -1.15653574e+00 1.58042753e+00
3.83347809e-01 3.43714394e-02 -1.89175323e-01 6.53541744e-01
9.01096821e-01 7.26641417e-01 1.37719363e-01 4.00786847e-01
8.52513254e-01 -4.65760559e-01 -1.33131176e-01 3.51963103e-01
4.35374141e-01 -9.20204282e-01 1.29632521e+00 3.71282369e-01
-1.32094586e+00 -7.13394523e-01 -1.03251493e+00 -7.60589242e-02
-3.68479371e-01 -1.28842071e-01 1.21426678e+00 7.58164942e-01
-1.03218114e+00 4.05943006e-01 -6.25424445e-01 6.47365823e-02
9.93910253e-01 3.13384145e-01 -4.24839675e-01 -1.80027068e-01
-7.19336033e-01 5.84629476e-01 6.92266524e-02 -1.83567017e-01
-9.01752651e-01 -1.15339756e+00 -9.54037368e-01 -2.45830908e-01
1.35781601e-01 -1.22490883e+00 1.00563729e+00 -1.59958526e-01
-1.65305388e+00 7.75401771e-01 1.67585656e-01 -3.60121839e-02
1.93160534e-01 -4.84325364e-02 -2.25270525e-01 -2.67346799e-01
1.39681771e-01 1.06750929e+00 8.55970442e-01 -1.44145715e+00
-2.62346804e-01 -2.15698630e-01 8.15514550e-02 1.79778278e-01
-2.07626387e-01 -5.55459321e-01 -4.51394945e-01 -7.36715853e-01
8.95083174e-02 -8.24239731e-01 -3.86504263e-01 2.19525304e-02
-3.57695371e-01 -2.99071580e-01 7.56680489e-01 -5.20762093e-02
7.46636212e-01 -2.30075264e+00 2.45605335e-01 1.66466787e-01
4.08541799e-01 1.63788944e-01 -4.04216617e-01 4.42424923e-01
8.27514529e-02 5.54466724e-01 -7.54280835e-02 -4.80188847e-01
4.24818218e-01 2.04203025e-01 -5.36901116e-01 -1.75284892e-01
5.64033210e-01 1.59831023e+00 -7.61386037e-01 -2.67446429e-01
4.10876721e-01 7.44666398e-01 -9.32497203e-01 1.90096959e-01
-5.40136576e-01 4.65471298e-01 -5.89722812e-01 8.38588893e-01
5.97524464e-01 -3.63551229e-01 -2.97332406e-01 -2.13004872e-01
9.78150591e-03 1.67644173e-01 -1.05368912e+00 2.04233027e+00
-6.83435917e-01 5.07996157e-02 -3.27007771e-01 -3.19883972e-01
1.25049555e+00 6.70986623e-02 6.96624815e-01 -8.41892302e-01
-1.36232272e-01 1.87495232e-01 -5.68540573e-01 -1.42010450e-01
7.33905613e-01 -3.79464418e-01 -4.20697361e-01 5.34667909e-01
2.68340290e-01 -1.03734803e+00 -1.95107330e-02 3.77520800e-01
6.99093342e-01 6.78725183e-01 -1.95435211e-01 4.68379259e-02
-1.39412150e-01 -2.27721915e-01 6.73120320e-02 5.46331763e-01
3.85897517e-01 8.41543436e-01 2.38702700e-01 -5.20609260e-01
-1.48086262e+00 -1.71514237e+00 -1.74034104e-01 6.59809411e-01
-1.78002462e-01 -5.99518239e-01 -2.45424807e-01 -5.52537024e-01
5.17922640e-01 6.06630504e-01 -6.12437725e-01 -1.36556029e-01
-4.35735434e-01 -5.40691733e-01 5.02679527e-01 5.34891427e-01
1.08120598e-01 -9.13620830e-01 -1.17613502e-01 1.94702938e-01
1.16157075e-02 -1.13017917e+00 -3.47850204e-01 -2.68729597e-01
-1.09516096e+00 -6.54986084e-01 -7.93251932e-01 -5.82761526e-01
4.73863512e-01 1.32672951e-01 1.53443635e+00 -2.45327309e-01
-5.04849911e-01 6.43399775e-01 -2.41381228e-01 -4.18478698e-01
-5.66379189e-01 2.49044299e-01 1.20362110e-01 -3.40949118e-01
-1.15776666e-01 -9.80795085e-01 -3.60749543e-01 2.53303349e-02
-9.59925711e-01 3.14442813e-01 5.85721195e-01 5.76747298e-01
8.39301765e-01 -1.91892490e-01 6.68076634e-01 -7.20534742e-01
5.74216664e-01 -5.56190193e-01 -6.58687174e-01 -9.12152603e-02
-3.50004315e-01 2.08765849e-01 4.37415421e-01 -4.59451377e-01
-9.98426855e-01 -2.50901771e-03 -4.02465433e-01 -4.99767423e-01
-1.62490696e-01 2.53669471e-01 -2.42993042e-01 8.53832662e-02
7.11428046e-01 2.60820627e-01 -3.16243432e-02 -6.50322556e-01
1.16992545e+00 3.83107632e-01 4.16728616e-01 -9.17793870e-01
1.17109597e+00 3.98138434e-01 1.09817818e-01 -6.38660431e-01
-6.22384310e-01 4.92610969e-02 -5.80651581e-01 2.78322622e-02
7.12005198e-01 -1.07726729e+00 -6.83959246e-01 4.77401137e-01
-9.59478736e-01 -3.14033568e-01 -7.50653327e-01 -1.92352068e-02
-8.64477038e-01 1.06305644e-01 -3.03700507e-01 -6.40189826e-01
-3.39390606e-01 -9.89943624e-01 1.55989170e+00 -4.41438220e-02
-2.63554573e-01 -7.93969214e-01 -1.42065343e-02 3.39305878e-01
5.86908638e-01 5.16527951e-01 1.34594321e+00 2.94295996e-02
-1.03164852e+00 -6.22052439e-02 -1.11638717e-01 2.70175159e-01
3.20337057e-01 -9.12587391e-04 -7.15634942e-01 8.39373842e-03
-6.15057945e-01 -6.05480611e-01 5.63430369e-01 2.29488656e-01
1.12515068e+00 2.22408418e-02 -1.93812370e-01 8.87846947e-01
1.18138206e+00 -9.95493680e-02 6.41622901e-01 -1.35761589e-01
9.59466398e-01 1.66756079e-01 2.30472788e-01 6.14005506e-01
6.70392811e-01 7.88319945e-01 4.84799713e-01 -1.73841819e-01
-4.61290985e-01 -8.02869499e-01 -5.32502308e-02 1.08154559e+00
-1.95420176e-01 -2.15784442e-02 -8.14125061e-01 5.48332989e-01
-1.33165014e+00 -8.29133391e-01 1.57349408e-01 2.02395749e+00
9.62703228e-01 1.56596288e-01 4.36832011e-02 -1.92931190e-01
1.20533906e-01 2.24363089e-01 -5.63562036e-01 -3.50375563e-01
-3.04877967e-01 5.79259872e-01 2.39491425e-02 2.30561286e-01
-9.37898040e-01 1.13758695e+00 7.11973047e+00 1.00808394e+00
-8.70441020e-01 -3.34300220e-01 6.30353034e-01 -3.05236965e-01
-1.11935019e+00 -2.65426993e-01 -1.07177424e+00 1.86469883e-01
5.51855922e-01 -1.27433240e-01 4.47372913e-01 9.06947613e-01
7.45329857e-02 2.91989952e-01 -9.65093374e-01 1.24984264e+00
1.62656695e-01 -1.89906156e+00 7.11326063e-01 2.57558852e-01
1.14828193e+00 -1.09330468e-01 6.25293076e-01 5.11936784e-01
5.52565038e-01 -1.42604482e+00 8.72122645e-01 6.77237868e-01
1.38535213e+00 -8.48772168e-01 2.15664646e-03 1.36137560e-01
-1.26024497e+00 1.42322078e-01 -2.27675810e-01 1.34981111e-01
3.54856342e-01 6.28810346e-01 -8.06811035e-01 6.02782309e-01
4.51663077e-01 8.73237133e-01 -4.55028266e-01 7.24929333e-01
8.20372161e-03 3.05175960e-01 -4.70213205e-01 9.89083052e-02
1.47437051e-01 -1.11355551e-01 5.43518126e-01 8.65728140e-01
4.84518081e-01 1.00250155e-01 8.72518718e-02 1.48043156e+00
-2.41374642e-01 -6.62203282e-02 -1.02670240e+00 -3.61597121e-01
5.98256946e-01 1.04269636e+00 -3.31874758e-01 -1.69006631e-01
-1.33420989e-01 7.69008577e-01 1.22674733e-01 2.76715726e-01
-7.43889928e-01 -7.80136511e-02 6.93822801e-01 2.43971556e-01
5.77802360e-01 -8.35456789e-01 -6.90363586e-01 -9.67296302e-01
-3.58350971e-03 -8.94224703e-01 -2.00780645e-01 -9.99426842e-01
-1.53665721e+00 5.29017389e-01 4.62691709e-02 -1.32185280e+00
-3.36205244e-01 -5.60769975e-01 -6.60930648e-02 9.65756953e-01
-1.34583819e+00 -1.80845845e+00 -8.26276392e-02 5.15795708e-01
4.19919461e-01 -4.29651320e-01 1.00981498e+00 2.31953323e-01
1.48203358e-01 4.81656879e-01 -2.34271586e-01 -1.83511630e-01
4.00836527e-01 -1.04071343e+00 1.10271645e+00 1.98108897e-01
5.70167959e-01 5.77896059e-01 1.14801496e-01 -7.04288304e-01
-1.64815605e+00 -9.88134444e-01 5.13373137e-01 -1.15771592e+00
3.65876704e-01 -8.02227199e-01 -4.20833796e-01 4.87763226e-01
-3.56408536e-01 -8.99994001e-02 7.68088579e-01 2.15683222e-01
-6.65205896e-01 1.21887743e-01 -8.27184141e-01 7.84825981e-01
1.53744864e+00 -5.01483560e-01 -4.73165751e-01 2.47201566e-02
1.13780546e+00 -6.28174484e-01 -1.09877217e+00 6.09959543e-01
7.45468259e-01 -7.65978336e-01 1.44597566e+00 -6.32222712e-01
8.28930676e-01 -1.37785506e-02 -3.19715858e-01 -1.18157661e+00
-5.62578738e-01 -6.62595868e-01 -3.40677321e-01 1.30287695e+00
5.68056881e-01 -1.73980162e-01 9.56972599e-01 5.62162161e-01
-3.27279776e-01 -1.14587021e+00 -7.49477148e-01 -6.36773825e-01
4.31768358e-01 -8.89552057e-01 1.29702497e+00 5.95592558e-01
-5.59995651e-01 2.18745530e-01 -2.49908999e-01 -2.78003603e-01
6.28875077e-01 3.95085841e-01 1.15611851e+00 -1.14931262e+00
-3.84191155e-01 -4.73186314e-01 -3.90088886e-01 -1.61586523e+00
-3.56214233e-02 -1.17448926e+00 -4.16698694e-01 -1.60359538e+00
2.63271034e-02 -1.08235610e+00 7.20646158e-02 5.56775272e-01
1.90531701e-01 5.79701841e-01 4.88360971e-01 2.02208403e-02
-4.14845854e-01 1.01569366e+00 1.85143626e+00 -1.26309872e-01
-3.60125661e-01 -8.78568459e-03 -1.07108581e+00 4.89125311e-01
4.87078011e-01 -1.53360680e-01 -5.43516457e-01 -6.74772441e-01
5.00995934e-01 -2.53273100e-01 4.68948066e-01 -7.21142948e-01
-2.34901026e-01 -1.36190757e-01 7.98016250e-01 -7.97830582e-01
6.90883875e-01 -6.01629555e-01 1.75094277e-01 -1.56984448e-01
9.41385608e-03 -1.67318523e-01 4.87581939e-01 4.38513368e-01
1.81840062e-01 4.57204342e-01 4.81978476e-01 -2.75150448e-01
-5.64683020e-01 7.99175501e-01 -2.03078501e-02 1.64114803e-01
8.53096962e-01 -4.94558066e-01 -7.28981383e-03 -3.47289771e-01
-7.44182587e-01 -7.75148422e-02 5.13456762e-01 1.03368294e+00
8.99918914e-01 -1.92124259e+00 -7.99820423e-01 7.23776400e-01
3.39186907e-01 2.16561541e-01 3.28743815e-01 1.04941539e-01
-1.56366751e-01 2.67623484e-01 -2.64297485e-01 -8.30814421e-01
-6.51286781e-01 3.30904424e-01 4.72010747e-02 -8.48166794e-02
-7.12406218e-01 8.25691998e-01 1.66568920e-01 -8.61108720e-01
-2.25372359e-01 -3.05701375e-01 8.74964893e-02 -2.44572878e-01
2.60236293e-01 4.72757481e-02 -4.69604395e-02 -5.07482767e-01
2.37705410e-02 7.34404743e-01 2.10181862e-01 -9.91832390e-02
1.62940919e+00 1.42335624e-01 1.75138518e-01 3.15925866e-01
1.03798032e+00 3.22557002e-01 -1.58193374e+00 9.41724852e-02
-4.45123881e-01 -5.18145263e-01 -1.85972407e-01 -9.51205254e-01
-9.15032029e-01 7.85049319e-01 2.42993444e-01 -6.88821748e-02
8.50280941e-01 2.23252892e-01 1.05508363e+00 1.18923150e-02
9.09929991e-01 -6.55569255e-01 5.33401072e-01 7.26244271e-01
1.11731398e+00 -8.72974336e-01 -1.52942538e-01 -3.47880483e-01
-5.87677956e-01 8.56397152e-01 3.22275132e-01 -1.81013688e-01
6.96237862e-01 5.50881386e-01 -1.06883608e-01 -2.36534983e-01
-8.04776669e-01 -8.75635594e-02 5.41381180e-01 1.03459167e+00
2.61721522e-01 1.99596480e-01 3.63663405e-01 6.95137322e-01
-5.47044933e-01 3.04421838e-02 1.45241410e-01 6.05753005e-01
-1.16901539e-01 -1.50304854e+00 -9.94914100e-02 3.71596217e-01
1.14259720e-02 -2.59298772e-01 -1.49119765e-01 7.03885913e-01
2.84656554e-01 5.32182813e-01 -1.41687989e-01 -4.58563566e-01
6.43063426e-01 -4.91871238e-02 8.10933352e-01 -7.85877824e-01
-1.56138420e-01 -8.55455622e-02 -7.40142465e-02 -7.29536355e-01
-2.40247995e-02 -4.94941890e-01 -9.82648134e-01 -4.47263509e-01
-4.37173583e-02 -3.19034815e-01 7.97367215e-01 5.62095046e-01
9.87086833e-01 4.01294291e-01 7.06370413e-01 -1.32621086e+00
-5.16580224e-01 -4.46550429e-01 -5.18733740e-01 5.19507647e-01
-4.62443270e-02 -8.74249876e-01 -2.00744197e-02 1.34076059e-01] | [8.976791381835938, -3.6089279651641846] |
1bff07e5-ead4-440f-83dc-360500273ccd | building-and-evaluation-of-a-real-room | 1811.06795 | null | http://arxiv.org/abs/1811.06795v2 | http://arxiv.org/pdf/1811.06795v2.pdf | Building and Evaluation of a Real Room Impulse Response Dataset | This paper presents BUT ReverbDB - a dataset of real room impulse responses
(RIR), background noises and re-transmitted speech data. The retransmitted data
includes LibriSpeech test-clean, 2000 HUB5 English evaluation and part of 2010
NIST Speaker Recognition Evaluation datasets. We provide a detailed description
of RIR collection (hardware, software, post-processing) that can serve as a
"cook-book" for similar efforts. We also validate BUT ReverbDB in two sets of
automatic speech recognition (ASR) experiments and draw conclusions for
augmenting ASR training data with real and artificially generated RIRs. We show
that a limited number of real RIRs, carefully selected to match the target
environment, provide results comparable to a large number of artificially
generated RIRs, and that both sets can be combined to achieve the best ASR
results. The dataset is distributed for free under a non-restrictive license
and it currently contains data from 8 rooms, which is growing. The distribution
package also contains a Kaldi-based recipe for augmenting publicly available
AMI close-talk meeting data and test the results on an AMI single distant
microphone set, allowing it to reproduce our experiments. | [] | 2019-05-30 | null | null | null | null | ['room-impulse-response'] | ['audio'] | [ 2.74553984e-01 -2.06196904e-01 8.36627185e-01 -6.76854134e-01
-1.60921502e+00 -5.74741304e-01 4.81140673e-01 -1.96686924e-01
-5.22093415e-01 5.08672953e-01 7.58299530e-01 -5.85016787e-01
-2.66084131e-02 -5.73156849e-02 -5.26025534e-01 -7.74611056e-01
-2.09218800e-01 3.16504031e-01 1.58044040e-01 -6.75484478e-01
-2.06427649e-01 5.19475400e-01 -1.52424955e+00 4.48201716e-01
2.93259710e-01 5.84123969e-01 5.23119748e-01 1.29651082e+00
4.67731327e-01 7.15109110e-01 -1.17006731e+00 4.86516468e-02
3.50655615e-01 -2.62212843e-01 -5.40393472e-01 -2.68413603e-01
4.56092507e-01 -3.77559245e-01 -6.94176733e-01 5.19653738e-01
1.34596598e+00 5.78940749e-01 2.22560093e-01 -8.04912269e-01
-2.73551196e-01 9.11424398e-01 2.65330166e-01 4.72395152e-01
6.43979549e-01 1.26923686e-02 5.31639695e-01 -9.66370404e-01
2.17411220e-02 1.13558841e+00 7.19923794e-01 6.38418913e-01
-1.05136347e+00 -7.33687580e-01 -2.11378798e-01 -1.28022015e-01
-1.60511422e+00 -1.38855362e+00 5.74261963e-01 3.10678966e-02
1.35805702e+00 1.10888243e+00 1.28342032e-01 1.66176808e+00
-6.57893181e-01 2.98211902e-01 1.03200269e+00 -6.63175344e-01
3.74844611e-01 3.68247002e-01 3.21764797e-01 -2.00597703e-01
-4.54621375e-01 3.71773154e-01 -3.74825031e-01 -3.03723544e-01
2.90606230e-01 -6.04574144e-01 -7.29677379e-01 4.45395738e-01
-1.49191034e+00 1.97148666e-01 1.64566204e-01 6.18886411e-01
-3.01252216e-01 -9.93960351e-02 3.17470849e-01 6.67877436e-01
3.40208441e-01 2.96600282e-01 -5.15478373e-01 -4.42215294e-01
-7.62881815e-01 2.60080546e-01 1.10511589e+00 1.10725033e+00
2.80385345e-01 4.88217503e-01 -2.05585167e-01 1.81411338e+00
3.93283308e-01 1.00073814e+00 5.97254515e-01 -1.03416634e+00
6.70351267e-01 -5.81597567e-01 2.78741390e-01 -4.47010398e-01
-3.29619378e-01 -2.30334044e-01 -6.62781179e-01 -5.25572710e-03
3.81349742e-01 -1.98143244e-01 -9.20956552e-01 1.47441447e+00
-2.53607128e-02 1.00246154e-01 3.42028469e-01 8.02546859e-01
1.13590467e+00 1.02516747e+00 -5.95745385e-01 -2.04673842e-01
1.16999221e+00 -8.75015259e-01 -1.04308856e+00 -1.59592435e-01
4.91368413e-01 -1.19877434e+00 1.13499510e+00 6.54599428e-01
-1.15495086e+00 -7.11030185e-01 -1.14515865e+00 1.91636473e-01
-3.29475462e-01 2.39683297e-02 -4.27210964e-02 1.18947756e+00
-1.49843073e+00 2.28203297e-01 -6.19020700e-01 -3.21690023e-01
-4.55671400e-01 3.56756657e-01 -4.68134224e-01 -2.10079864e-01
-1.17906535e+00 7.38430262e-01 -2.14908883e-01 4.98025119e-01
-9.24200773e-01 -4.35314685e-01 -9.72456396e-01 -3.36304873e-01
2.17825249e-02 3.20748538e-02 1.75060534e+00 -3.96463454e-01
-1.70232880e+00 6.46423697e-01 -2.96988487e-01 -4.01005268e-01
3.61227453e-01 -1.25529692e-01 -1.08406568e+00 -2.23531336e-01
-3.87448967e-01 1.91554457e-01 5.43627679e-01 -1.30330360e+00
-1.04597673e-01 -5.08291088e-02 -4.12610888e-01 1.89856544e-01
-5.60140759e-02 5.77161491e-01 -2.11826980e-01 -7.22677469e-01
8.96927435e-03 -8.46883893e-01 -2.50870407e-01 -1.10056603e+00
-5.87358594e-01 2.89747000e-01 3.49562168e-01 -1.05290854e+00
1.13066745e+00 -2.49581575e+00 -4.03350115e-01 4.03867871e-01
-3.87297690e-01 3.73178095e-01 -3.89445484e-01 4.71556783e-01
-4.98186290e-01 1.21018507e-01 -1.80703729e-01 -8.24559748e-01
1.17666073e-01 4.88139838e-02 -4.48433220e-01 5.81324518e-01
-3.17045301e-01 2.24603578e-01 -4.54633147e-01 1.67297319e-01
6.67663097e-01 7.01892912e-01 -3.59666944e-01 5.56622505e-01
6.13278806e-01 5.33130348e-01 1.44071192e-01 3.34624738e-01
7.88229406e-01 7.10412920e-01 -4.04022902e-01 -1.65702820e-01
-3.25038552e-01 9.12161469e-01 -1.51669931e+00 1.44919562e+00
-7.32625306e-01 9.26322639e-01 7.08257377e-01 -6.18871272e-01
1.21619844e+00 8.18159282e-01 1.04525974e-02 -5.90459943e-01
5.47282621e-02 6.29392445e-01 1.49485737e-01 -4.47139829e-01
6.02057457e-01 1.48109370e-03 7.16201682e-03 1.75431311e-01
-1.94915935e-01 -7.06801593e-01 -1.19107805e-01 4.19440269e-02
1.42998767e+00 -6.66823089e-01 3.47922184e-02 -2.22393036e-01
7.74377942e-01 -5.68573356e-01 5.13504781e-02 1.05903411e+00
-1.77740335e-01 1.13986278e+00 -4.89904970e-01 2.22914636e-01
-1.27579010e+00 -1.28427327e+00 -4.23607469e-01 9.76642966e-01
-5.66531122e-01 -2.48630971e-01 -6.96028769e-01 1.09678991e-01
-5.91217935e-01 1.30840230e+00 -2.05855131e-01 2.82297492e-01
-7.84193575e-01 -4.50311631e-01 1.07566822e+00 4.40676808e-01
1.58467785e-01 -1.28462911e+00 3.10294002e-01 1.30826414e-01
-4.56997395e-01 -1.21550179e+00 -7.36271024e-01 6.30220175e-01
-1.37221083e-01 -3.75707418e-01 -9.46703970e-01 -9.39216554e-01
3.47475857e-01 5.60531497e-01 9.63128567e-01 -2.74687503e-02
-1.28749892e-01 8.70167136e-01 -6.45640492e-01 -3.48333061e-01
-1.25496948e+00 -3.58659148e-01 4.95900422e-01 -2.75146723e-01
1.91096246e-01 -6.52204633e-01 -2.08990350e-01 7.26372063e-01
-4.99497980e-01 -4.10117030e-01 9.88028347e-02 6.23194814e-01
2.14913189e-01 -1.37766495e-01 7.27921903e-01 -2.25243241e-01
6.42076552e-01 -1.32443741e-01 -4.11771089e-01 -3.90586779e-02
5.80612794e-02 -3.22372854e-01 5.87584436e-01 -2.84072906e-01
-1.21899295e+00 -3.43328983e-01 -1.07368028e+00 -2.11706877e-01
-6.33626282e-01 2.02201419e-02 -5.10733962e-01 3.25141400e-01
9.33885932e-01 2.93866545e-01 -2.44254529e-01 -7.31434882e-01
3.01990181e-01 1.72362387e+00 8.48099947e-01 -4.53425944e-01
8.03023577e-01 3.15419510e-02 -7.96650887e-01 -1.45289791e+00
-1.12748869e-01 -9.00304377e-01 -1.58564955e-01 4.36962545e-02
5.33278763e-01 -1.06843281e+00 -4.63148206e-01 6.30695164e-01
-1.07290351e+00 -8.11141729e-01 -4.63785410e-01 9.28558350e-01
-5.68567932e-01 1.63472161e-01 -6.14563942e-01 -1.29707503e+00
-2.17947915e-01 -1.21368575e+00 1.11779130e+00 -2.23281205e-01
-2.85892099e-01 -6.83641315e-01 3.58083218e-01 6.63483322e-01
7.75216997e-01 -4.73072410e-01 2.25195169e-01 -1.19113064e+00
-6.00021854e-02 -3.91934544e-01 3.92118245e-01 1.01019406e+00
3.06818306e-01 -1.26417279e-01 -1.77294993e+00 -4.13991570e-01
3.19316715e-01 -2.45344967e-01 5.44680536e-01 5.46109915e-01
8.39802921e-01 -3.84313405e-01 1.01378776e-01 3.46844286e-01
6.05315387e-01 4.03074920e-01 1.03490531e+00 1.35174558e-01
3.96296710e-01 5.86402595e-01 3.56363207e-01 3.04815739e-01
6.38077222e-03 9.71746266e-01 -4.02288139e-02 -1.02220476e-01
-3.61587971e-01 2.12719277e-01 7.17170238e-01 1.62650418e+00
9.78314504e-03 -4.58275080e-01 -8.43605936e-01 5.05088210e-01
-1.08011568e+00 -1.13114262e+00 -3.72586221e-01 2.81112194e+00
8.52802098e-01 -1.24014772e-01 1.83440924e-01 6.66306913e-01
6.54235244e-01 7.84609839e-02 1.94297686e-01 -7.26371169e-01
-2.95479596e-01 3.58011216e-01 4.11192685e-01 1.17238212e+00
-8.16752791e-01 3.15739661e-01 6.90691376e+00 6.13872707e-01
-8.27381492e-01 2.40114450e-01 4.42732692e-01 -1.01801708e-01
-3.48043650e-01 -6.68145180e-01 -5.68125248e-01 1.89032301e-01
1.79449821e+00 -4.26468179e-02 9.91065979e-01 6.18207574e-01
6.77381039e-01 1.94155708e-01 -1.19619727e+00 1.08511078e+00
1.94976449e-01 -7.78481543e-01 -7.44924188e-01 3.93496305e-02
2.48267189e-01 6.92288101e-01 1.21725619e-01 5.70000112e-01
3.02486867e-01 -1.22923565e+00 7.19334006e-01 5.96328042e-02
7.41814792e-01 -4.84969258e-01 7.57971883e-01 1.77415103e-01
-1.09721637e+00 -6.76310658e-02 -3.17227870e-01 1.83040172e-01
1.79680437e-01 3.19354028e-01 -1.10902071e+00 5.23011446e-01
8.75400722e-01 1.20139979e-01 -5.69206238e-01 1.11943781e+00
2.38024473e-01 1.02182531e+00 -5.84818780e-01 1.27106458e-01
-2.21124142e-01 1.39369905e-01 9.23126400e-01 1.66567314e+00
5.15228689e-01 5.95713444e-02 -2.03790486e-01 4.56868708e-01
1.87840581e-01 1.32160366e-01 -6.82818472e-01 3.08471769e-01
8.46006930e-01 1.15598464e+00 -2.40025043e-01 -2.49702245e-01
-2.39459664e-01 9.02652800e-01 -3.40958625e-01 7.13399529e-01
-8.92997444e-01 -5.44423401e-01 5.25572717e-01 -3.39104496e-02
1.38155222e-01 -2.23665059e-01 7.93494880e-02 -8.67295682e-01
-4.59037274e-02 -1.32777536e+00 -1.73707545e-01 -1.23972845e+00
-1.19703329e+00 1.12358487e+00 2.12339446e-01 -1.20022917e+00
-3.75700504e-01 -6.34844184e-01 -5.22590399e-01 1.25357187e+00
-9.49514508e-01 -5.39189219e-01 -4.96341914e-01 5.69962561e-01
7.43026972e-01 -2.00171947e-01 1.24987400e+00 7.33866036e-01
-4.30344939e-01 8.90346587e-01 3.62153769e-01 1.12245068e-01
8.23975384e-01 -1.30652535e+00 7.43706048e-01 7.04890370e-01
2.54836321e-01 6.96346939e-01 1.20456338e+00 -1.30804433e-02
-1.03877199e+00 -9.00972724e-01 5.86835265e-01 -7.65961766e-01
5.05917192e-01 -8.96497130e-01 -1.05818236e+00 7.50582457e-01
3.10954064e-01 -1.78605825e-01 7.56287754e-01 7.52998283e-03
-2.28922471e-01 -3.24384928e-01 -1.03442287e+00 4.77606833e-01
8.87429535e-01 -5.43878376e-01 -8.73640120e-01 4.13909972e-01
1.04077649e+00 -5.58266342e-01 -7.08692491e-01 2.76089430e-01
2.31430843e-01 -9.32038724e-01 1.16663015e+00 -1.51482993e-03
-3.24223220e-01 -4.62871909e-01 -9.13075626e-01 -1.55767548e+00
6.04789890e-02 -1.05308115e+00 3.74021977e-01 1.65302563e+00
6.81109190e-01 -8.69360387e-01 6.63282201e-02 5.66116512e-01
-7.87254155e-01 8.52989554e-02 -1.24247956e+00 -1.01756656e+00
-2.44555026e-01 -1.08970952e+00 4.53712910e-01 5.23803055e-01
-9.22600776e-02 2.73486704e-01 -4.60321158e-01 3.04201901e-01
4.00451779e-01 -8.58747125e-01 1.01760542e+00 -5.44877350e-01
-5.22718072e-01 -1.84157506e-01 -2.39214823e-01 -1.05229807e+00
-1.21107556e-01 -7.18328416e-01 6.36674404e-01 -1.29011178e+00
-5.10918677e-01 -7.23981738e-01 -2.72795856e-01 3.33043426e-01
7.58403316e-02 1.85814351e-01 8.70929733e-02 -1.07528858e-01
-1.06660403e-01 5.02432883e-01 6.61376297e-01 -1.49447322e-01
-4.86765385e-01 5.19045115e-01 -3.49778116e-01 3.67678910e-01
6.10403121e-01 -2.85414666e-01 -3.77828032e-01 -1.64264143e-01
-4.50206101e-01 4.44299690e-02 3.53276938e-01 -1.28019953e+00
8.32355674e-03 3.49425077e-01 1.74581707e-02 -5.89470863e-01
9.34658408e-01 -8.79017830e-01 2.34381244e-01 -3.29205580e-02
-6.14622116e-01 1.82931311e-02 5.93054056e-01 2.80206829e-01
-8.16065520e-02 -1.78176120e-01 6.45017207e-01 1.70995891e-01
-1.82513967e-01 -3.80484998e-01 -6.39808357e-01 8.46754313e-02
2.72591352e-01 -4.72343415e-02 -3.11808378e-01 -9.07235682e-01
-7.22718298e-01 -2.75358975e-01 1.35529324e-01 5.82617700e-01
4.79494721e-01 -1.23243797e+00 -1.14688063e+00 4.36462224e-01
2.28205267e-02 -1.07540891e-01 4.19466436e-01 7.97668159e-01
-3.44485998e-01 3.38130713e-01 5.16977787e-01 -5.41305125e-01
-1.60718787e+00 4.37110603e-01 6.36045158e-01 2.49966487e-01
-7.53827512e-01 8.90005291e-01 1.45051861e-02 -7.99516141e-01
7.10599184e-01 -5.14620423e-01 4.10981849e-02 -4.71329123e-01
1.04807353e+00 3.52719784e-01 5.72583497e-01 -7.80082524e-01
-5.58117688e-01 -7.57996440e-02 1.22333236e-01 -7.58106589e-01
1.13141167e+00 -3.90101552e-01 1.77061900e-01 8.91795576e-01
1.29153812e+00 7.84304798e-01 -5.68267584e-01 -1.52585600e-02
-4.45582598e-01 -2.32986137e-01 -1.60161983e-02 -9.91971493e-01
-4.98778403e-01 7.13644087e-01 8.86592209e-01 3.49797338e-01
1.13500798e+00 -1.82100665e-02 5.06395698e-01 6.33136153e-01
2.71285266e-01 -7.86598921e-01 -1.49327084e-01 5.54777086e-01
1.36213291e+00 -9.34708238e-01 -4.63842213e-01 -2.12529495e-01
-6.33527935e-01 8.15393746e-01 1.91288173e-01 2.62741476e-01
6.45832181e-01 9.13124382e-01 6.26633704e-01 4.66160268e-01
-5.24158537e-01 6.84963539e-02 -9.33235977e-03 1.05929339e+00
6.50819182e-01 1.72250077e-01 4.54109550e-01 6.97773635e-01
-8.44195545e-01 -5.35094500e-01 6.45051122e-01 5.36257744e-01
-2.34743670e-01 -1.03101778e+00 -1.03028738e+00 7.51794577e-02
-4.02336657e-01 -4.89654422e-01 -1.39451474e-01 6.12563610e-01
-5.17169356e-01 1.67324865e+00 -6.74663112e-02 -5.69684744e-01
7.55476832e-01 1.10952221e-01 2.51605034e-01 -4.22633916e-01
-8.38442564e-01 4.05607671e-01 7.38667190e-01 -2.15928108e-01
-5.52447550e-02 -5.52656114e-01 -9.49297428e-01 -3.58109266e-01
-4.21584576e-01 3.82711500e-01 1.05630612e+00 6.96509540e-01
2.30565537e-02 1.00842071e+00 8.07415545e-01 -1.12507725e+00
-7.60020256e-01 -1.55174494e+00 -6.43481195e-01 3.37813616e-01
5.54673612e-01 -3.95300835e-02 -9.29124236e-01 -4.24074046e-02] | [14.982332229614258, 5.993803024291992] |
51b2868a-d6fc-4db3-8d48-de1574975b8a | lbl2vec-an-embedding-based-approach-for | 2210.06023 | null | https://arxiv.org/abs/2210.06023v1 | https://arxiv.org/pdf/2210.06023v1.pdf | Lbl2Vec: An Embedding-Based Approach for Unsupervised Document Retrieval on Predefined Topics | In this paper, we consider the task of retrieving documents with predefined topics from an unlabeled document dataset using an unsupervised approach. The proposed unsupervised approach requires only a small number of keywords describing the respective topics and no labeled document. Existing approaches either heavily relied on a large amount of additionally encoded world knowledge or on term-document frequencies. Contrariwise, we introduce a method that learns jointly embedded document and word vectors solely from the unlabeled document dataset in order to find documents that are semantically similar to the topics described by the keywords. The proposed method requires almost no text preprocessing but is simultaneously effective at retrieving relevant documents with high probability. When successively retrieving documents on different predefined topics from publicly available and commonly used datasets, we achieved an average area under the receiver operating characteristic curve value of 0.95 on one dataset and 0.92 on another. Further, our method can be used for multiclass document classification, without the need to assign labels to the dataset in advance. Compared with an unsupervised classification baseline, we increased F1 scores from 76.6 to 82.7 and from 61.0 to 75.1 on the respective datasets. For easy replication of our approach, we make the developed Lbl2Vec code publicly available as a ready-to-use tool under the 3-Clause BSD license. | ['Florian Matthes', 'Daniel Braun', 'Tim Schopf'] | 2022-10-12 | null | null | null | null | ['document-classification', 'unsupervised-text-classification'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.69465834e-01 2.18717530e-02 -4.42460924e-01 -5.00387490e-01
-1.25629997e+00 -8.77764404e-01 8.24160039e-01 6.17102921e-01
-5.87558568e-01 6.56907678e-01 8.62771571e-02 -2.43905321e-01
-2.85947621e-01 -6.68609619e-01 -4.26501274e-01 -6.92281842e-01
2.03386188e-01 6.81325197e-01 1.70723543e-01 1.49589390e-01
3.56646985e-01 2.42496908e-01 -1.66515863e+00 2.10753396e-01
7.99487650e-01 1.16069114e+00 1.99828297e-01 5.10135770e-01
-4.83475775e-01 3.71576369e-01 -5.76526582e-01 -1.95614651e-01
9.66198444e-02 -2.11484395e-02 -8.29876304e-01 2.56941855e-01
4.37641650e-01 -8.79681185e-02 -1.83668211e-01 8.91707301e-01
2.92606950e-01 3.30244333e-01 1.01213181e+00 -1.01251960e+00
-6.34316742e-01 5.77209294e-01 -3.89269501e-01 -3.13906670e-02
2.48376340e-01 -5.76012611e-01 1.43930840e+00 -9.96651351e-01
6.40933335e-01 7.99268603e-01 1.35576546e-01 1.48139507e-01
-1.06382430e+00 -6.57430053e-01 1.99115455e-01 1.14120287e-03
-1.54577804e+00 -2.61783928e-01 6.61557615e-01 -3.90357494e-01
8.95762563e-01 3.22425157e-01 8.15370232e-02 9.61209714e-01
-1.51091322e-01 6.91282809e-01 9.17257547e-01 -7.17377305e-01
1.72037080e-01 6.42912686e-01 5.87009966e-01 3.41360182e-01
5.98464012e-01 -2.98011601e-01 -3.60572398e-01 -3.38059992e-01
-3.74674089e-02 1.47786885e-01 -2.92847246e-01 -4.01179731e-01
-1.00944257e+00 9.50405121e-01 7.00502247e-02 6.58558547e-01
-2.44864404e-01 -1.23202711e-01 3.64664704e-01 1.61006376e-01
6.28233671e-01 3.76340002e-01 -4.96076614e-01 8.14761072e-02
-1.23083150e+00 8.71655997e-03 7.94002354e-01 1.08617353e+00
8.55351448e-01 -3.86559308e-01 -1.54753685e-01 8.90800416e-01
4.05331522e-01 5.68606377e-01 7.74279058e-01 -3.55099738e-01
5.44540882e-01 5.92836320e-01 2.18273968e-01 -1.09021819e+00
-2.65725672e-01 -6.21448934e-01 -4.86244082e-01 -6.94829941e-01
-7.27551356e-02 1.07141972e-01 -1.04992235e+00 1.40100360e+00
1.68144777e-01 -2.19554752e-01 3.33509445e-01 6.07260346e-01
7.83428907e-01 9.05632675e-01 1.25929369e-02 -1.86571762e-01
1.39927709e+00 -8.47631991e-01 -7.11295962e-01 -5.32980338e-02
7.28465796e-01 -8.78783822e-01 1.02939463e+00 5.43778002e-01
-6.00283027e-01 -2.60634273e-01 -1.01473546e+00 -4.27210480e-02
-7.86707401e-01 4.57102388e-01 5.94839752e-01 9.18155670e-01
-9.74272072e-01 4.81745377e-02 -5.14744341e-01 -4.94338304e-01
2.10955262e-01 3.41548502e-01 -3.51317018e-01 -3.95862788e-01
-1.14326668e+00 5.51168680e-01 8.07652891e-01 -3.62175077e-01
-8.54451716e-01 -3.55078518e-01 -7.80406356e-01 2.82380432e-01
4.17649835e-01 -8.97316113e-02 9.77250874e-01 -6.88566864e-01
-1.00173998e+00 7.85135865e-01 -1.73441127e-01 -3.77732068e-01
1.58560395e-01 -4.33357120e-01 -5.61736465e-01 4.77101564e-01
2.84046948e-01 6.53037846e-01 6.96658075e-01 -1.29016364e+00
-7.95881689e-01 -3.15764904e-01 -8.32522213e-02 1.18699551e-01
-1.06537306e+00 1.95337925e-02 -1.02205634e+00 -8.64320755e-01
1.19996957e-01 -1.00946844e+00 1.49065048e-01 -3.51968944e-01
-4.72328365e-01 -3.40910524e-01 8.38484764e-01 -5.34278512e-01
1.18862200e+00 -2.09754872e+00 -6.02867603e-02 5.03244281e-01
-7.59813115e-02 1.37864366e-01 -9.93581489e-02 5.50732315e-01
1.83682531e-01 1.86840028e-01 -2.50259548e-01 -1.07004836e-01
5.97878769e-02 -7.79552981e-02 -5.39704502e-01 4.64095592e-01
-6.38900772e-02 5.53749263e-01 -8.57696772e-01 -5.21928489e-01
-6.26836568e-02 5.93956828e-01 -3.23785633e-01 3.05530708e-02
-1.12361789e-01 4.52941507e-02 -5.64074516e-01 6.31434798e-01
4.95687813e-01 -3.23313802e-01 3.14575016e-01 4.05062782e-03
2.31788114e-01 1.81464389e-01 -1.18215346e+00 1.54532349e+00
-6.42254293e-01 7.35657990e-01 -4.04935956e-01 -1.20923734e+00
1.08871806e+00 6.17458344e-01 6.38041794e-01 -4.75199401e-01
2.14247014e-02 2.72650659e-01 -4.77910876e-01 -3.03070694e-01
6.52273595e-01 -5.35288528e-02 -2.03200534e-01 7.27197170e-01
3.94566864e-01 -1.21837810e-01 4.44190562e-01 4.09465760e-01
7.19691515e-01 -2.51525372e-01 1.79424837e-01 -3.15776199e-01
6.47067845e-01 5.92070818e-02 4.87902649e-02 8.38469625e-01
3.26659769e-01 5.02684593e-01 1.49278596e-01 4.63048741e-03
-7.44497657e-01 -7.17193365e-01 -4.28478628e-01 1.15690792e+00
3.59067135e-02 -6.38010144e-01 -3.01525086e-01 -9.02589381e-01
5.81685901e-02 9.16386962e-01 -5.48722625e-01 -1.10655591e-01
-1.44311711e-02 -4.95229363e-01 5.09043813e-01 4.03933853e-01
7.05909505e-02 -6.57965481e-01 -2.56893545e-01 -2.80228425e-02
-2.04540312e-01 -1.06841207e+00 -4.28884596e-01 4.26282734e-01
-7.31948853e-01 -8.81252706e-01 -1.01999688e+00 -9.02026474e-01
9.00753200e-01 4.41309124e-01 9.15440500e-01 -4.95998934e-02
-1.69633821e-01 6.76245570e-01 -6.37188137e-01 -3.60647023e-01
6.45926297e-02 4.18749094e-01 -1.61039997e-02 4.57106680e-02
6.08937144e-01 -3.86423729e-02 -4.35590595e-01 2.65829891e-01
-1.29259408e+00 -4.31457043e-01 3.97964656e-01 7.40783989e-01
5.94455242e-01 4.86604422e-01 5.61645627e-01 -1.03372991e+00
6.80709362e-01 -6.77868128e-01 -6.65005445e-01 3.75172973e-01
-1.01653326e+00 1.08940095e-01 5.83360612e-01 -4.00921732e-01
-9.49849248e-01 -3.78880017e-02 2.79016316e-01 -1.49687424e-01
-2.71556139e-01 7.30202079e-01 -2.04345621e-02 3.35189790e-01
4.47030604e-01 3.97002697e-01 -5.67798555e-01 -5.04768848e-01
4.79209095e-01 1.13726127e+00 2.48529449e-01 -6.07436717e-01
7.08437026e-01 4.87986833e-01 -3.94254774e-01 -6.98163569e-01
-1.00257373e+00 -1.17510605e+00 -4.80120569e-01 4.75278310e-02
4.83448178e-01 -9.16602194e-01 -1.82417244e-01 4.83685546e-02
-7.23297954e-01 2.11166769e-01 -1.79694053e-02 7.05948770e-01
-1.26353174e-01 5.38532674e-01 -1.23349026e-01 -7.78457522e-01
-5.14227271e-01 -9.42931652e-01 1.20618176e+00 -4.98617627e-02
-2.97916025e-01 -9.94860351e-01 2.24608388e-02 3.22984606e-01
3.39997113e-01 -3.03946465e-01 9.75368798e-01 -1.22663891e+00
-1.12173721e-01 -8.16703975e-01 -1.80972993e-01 4.11209524e-01
5.79113901e-01 -1.17105864e-01 -1.03761971e+00 -5.71213663e-01
-1.73639238e-01 -4.29767340e-01 1.18535984e+00 1.45913675e-01
1.20710623e+00 -3.00533682e-01 -4.75334793e-01 1.15924954e-01
1.44245613e+00 3.66563201e-01 2.60295540e-01 5.19025087e-01
4.78488773e-01 8.71065080e-01 7.85462201e-01 5.14767766e-01
1.87947586e-01 6.28678918e-01 9.50512011e-03 2.57815182e-01
2.58106202e-01 -1.82317421e-01 2.21487492e-01 7.07150161e-01
4.09559458e-01 -6.94069386e-01 -1.09952950e+00 7.97597528e-01
-1.35396075e+00 -6.67847753e-01 1.77399054e-01 2.28717470e+00
9.58154202e-01 1.81928441e-01 -1.66729838e-01 3.66370916e-01
7.00676560e-01 1.12718627e-01 -2.04536125e-01 -1.53821260e-01
1.17322139e-03 2.21913919e-01 4.42985117e-01 5.36212683e-01
-1.35985720e+00 8.55689704e-01 5.74892187e+00 8.77019882e-01
-1.10533905e+00 -9.58199948e-02 4.10472691e-01 -1.60500765e-01
-5.61056852e-01 7.36162886e-02 -1.02259004e+00 4.69574213e-01
1.13759887e+00 -4.48340476e-01 -8.57053697e-02 9.43722427e-01
-1.52629793e-01 -1.07469738e-01 -9.04938996e-01 7.13978112e-01
4.05098140e-01 -9.75687027e-01 3.77787858e-01 1.24003954e-01
7.81055331e-01 -2.52440661e-01 2.63637334e-01 3.44233394e-01
6.77193105e-02 -9.26115870e-01 4.38762516e-01 2.75686353e-01
9.15512443e-01 -8.57113361e-01 8.73281837e-01 2.40890384e-01
-1.00433600e+00 6.61669299e-02 -4.58644450e-01 5.25424659e-01
-1.79159746e-01 8.05429995e-01 -9.19550776e-01 6.18827999e-01
7.05273092e-01 5.77549219e-01 -5.72571874e-01 9.04622793e-01
-1.15994103e-01 8.06255341e-01 -3.10550153e-01 -1.75052017e-01
4.86584991e-01 2.88811382e-02 2.05810696e-01 1.45381224e+00
3.75071049e-01 -2.07386583e-01 1.64985627e-01 1.61211088e-01
-2.75701284e-01 6.57643318e-01 -5.27471781e-01 -3.01864266e-01
4.45264816e-01 1.17763722e+00 -1.02065921e+00 -5.96429825e-01
-4.92380857e-01 7.64393747e-01 2.40606666e-02 4.65428442e-01
-6.22463882e-01 -8.60341549e-01 -3.32660973e-02 -2.47995153e-01
7.24849045e-01 -3.74725871e-02 4.71227206e-02 -1.18237305e+00
1.86029300e-01 -7.33745456e-01 5.38348734e-01 -5.26705444e-01
-1.14785337e+00 6.79753006e-01 1.08141124e-01 -1.30394351e+00
-2.34955013e-01 -5.06131589e-01 -6.54142573e-02 5.78601539e-01
-1.56345725e+00 -8.61068249e-01 -2.46548370e-01 4.70536321e-01
4.53979313e-01 -4.36641335e-01 1.06001902e+00 3.96806687e-01
-3.28824282e-01 6.25694513e-01 7.31233418e-01 8.83492306e-02
1.04553711e+00 -1.21133053e+00 -1.15097016e-01 6.81704700e-01
7.64602125e-01 8.71016502e-01 4.61912364e-01 -5.19917965e-01
-1.15954375e+00 -1.17240775e+00 1.19334126e+00 -3.02597165e-01
5.75863540e-01 -3.14778179e-01 -8.52647960e-01 5.54648578e-01
2.71369070e-01 -2.25187123e-01 1.21321499e+00 1.38714865e-01
-6.33913338e-01 -1.83755621e-01 -1.04253364e+00 2.78901905e-01
2.90388048e-01 -6.20844543e-01 -7.31145382e-01 6.01452708e-01
6.03249788e-01 4.77222279e-02 -7.64573514e-01 1.90268949e-01
5.18249035e-01 -2.11734965e-01 8.14581811e-01 -7.26777375e-01
2.39863083e-01 -2.96561390e-01 -5.85594773e-01 -1.01138401e+00
-2.21671443e-02 -2.35726833e-01 -1.50790706e-01 1.40573752e+00
6.95100069e-01 -5.86976171e-01 5.38334131e-01 3.91230881e-01
2.09022030e-01 -5.97509742e-01 -6.86895967e-01 -7.60256112e-01
6.49212226e-02 -6.58688486e-01 3.37487668e-01 8.86153102e-01
-8.49409699e-02 2.20110252e-01 -2.21665978e-01 2.94470042e-01
3.93401086e-01 4.54712540e-01 5.36572754e-01 -1.37001967e+00
-3.10381819e-02 -8.49826708e-02 -3.24849188e-01 -9.23746943e-01
3.66715103e-01 -1.24624383e+00 4.36238162e-02 -1.52161217e+00
4.98221427e-01 -4.73820686e-01 -8.70580554e-01 4.86540705e-01
-6.10692985e-02 3.95912647e-01 -1.01080917e-01 4.62345183e-01
-6.87627077e-01 4.53904361e-01 5.16062379e-01 -5.73041379e-01
-1.16007045e-01 8.95005092e-03 -9.40712512e-01 5.03008783e-01
9.13643420e-01 -7.57298887e-01 -6.12882435e-01 -3.69793475e-01
6.75485656e-02 -3.47340345e-01 -8.06176011e-03 -7.82617331e-01
1.83502823e-01 1.01198368e-01 3.00229669e-01 -7.04306543e-01
2.10320607e-01 -1.03203666e+00 -3.30553800e-01 1.66156873e-01
-7.64977098e-01 -3.30545902e-01 2.25652620e-01 8.04916620e-01
-4.68716383e-01 -5.63862383e-01 4.21968669e-01 1.16489455e-01
-6.35957837e-01 7.90110305e-02 -5.36994636e-01 1.21324904e-01
1.05789053e+00 8.00817832e-02 -1.10991240e-01 -5.23107290e-01
-5.18439412e-01 8.97799432e-02 3.80034357e-01 5.63929200e-01
4.96001899e-01 -1.16595387e+00 -5.35477281e-01 -4.02039476e-02
5.72735965e-01 -3.77891183e-01 -2.05201879e-02 4.44298387e-01
-2.01129153e-01 1.23770869e+00 2.28966668e-01 -5.75795293e-01
-1.13062871e+00 5.80376804e-01 -2.25035280e-01 -2.79778570e-01
-4.90748912e-01 6.41625583e-01 6.59599602e-02 -3.87071609e-01
5.09515285e-01 -2.69696563e-01 -3.62795293e-01 5.63929200e-01
4.51378971e-01 1.16951980e-01 4.19543475e-01 -7.41501153e-01
-4.94622618e-01 6.02259815e-01 -4.05550361e-01 -3.62565130e-01
1.31527615e+00 -9.77779366e-03 9.62713584e-02 4.55114245e-01
1.52668810e+00 2.65661746e-01 -5.42604983e-01 -3.53689641e-01
3.14457178e-01 -3.26478422e-01 3.68405163e-01 -7.31563568e-01
-1.06337452e+00 6.91411138e-01 6.23276174e-01 2.73453236e-01
1.16776955e+00 1.88151881e-01 6.04893923e-01 7.42608011e-01
3.17598671e-01 -1.11997128e+00 -2.07197629e-02 3.27521801e-01
5.64498901e-01 -1.32419658e+00 2.79914796e-01 -2.39730418e-01
-6.96646869e-01 9.86803591e-01 2.03945309e-01 1.65479094e-01
7.21733034e-01 -1.30486220e-01 1.42110720e-01 -2.57338226e-01
-7.01376975e-01 -2.18552008e-01 7.15599179e-01 3.09618801e-01
6.78801656e-01 -1.21003084e-01 -4.20457244e-01 6.07786059e-01
-1.15359880e-01 -1.83933884e-01 2.27360830e-01 1.10038555e+00
-4.07493204e-01 -1.15242231e+00 -3.02004814e-01 6.02828801e-01
-7.40483999e-01 -2.06422061e-01 -3.82534474e-01 6.43067718e-01
-3.23339552e-01 1.18863642e+00 1.68650784e-02 -9.65662077e-02
5.35056666e-02 3.83070439e-01 -9.36696902e-02 -9.50488865e-01
-1.82693660e-01 2.07507849e-01 -4.52457219e-02 -1.57475814e-01
-5.02756596e-01 -4.93326902e-01 -1.02739298e+00 3.21242034e-01
-6.91230834e-01 5.67372680e-01 6.89770103e-01 8.44776869e-01
3.83030951e-01 1.89701840e-01 6.80299461e-01 -3.87566447e-01
-3.57070565e-01 -9.69683707e-01 -6.33210957e-01 2.90905565e-01
1.01494871e-01 -6.82179451e-01 -4.13956106e-01 2.54178971e-01] | [10.455607414245605, 7.74520206451416] |
3027d3fa-e852-4e0d-b1d1-1b6a8aff8b04 | rudas-synthetic-datasets-for-rule-learning | 1909.07095 | null | https://arxiv.org/abs/1909.07095v2 | https://arxiv.org/pdf/1909.07095v2.pdf | RuDaS: Synthetic Datasets for Rule Learning and Evaluation Tools | Logical rules are a popular knowledge representation language in many domains, representing background knowledge and encoding information that can be derived from given facts in a compact form. However, rule formulation is a complex process that requires deep domain expertise,and is further challenged by today's often large, heterogeneous, and incomplete knowledge graphs. Several approaches for learning rules automatically, given a set of input example facts,have been proposed over time, including, more recently, neural systems. Yet, the area is missing adequate datasets and evaluation approaches: existing datasets often resemble toy examples that neither cover the various kinds of dependencies between rules nor allow for testing scalability. We present a tool for generating different kinds of datasets and for evaluating rule learning systems, including new performance measures. | ['Veronika Thost', 'Cristina Cornelio'] | 2019-09-16 | null | null | null | null | ['inductive-knowledge-graph-completion'] | ['knowledge-base'] | [ 2.12866828e-01 3.89872104e-01 -6.81824267e-01 -5.51304042e-01
-1.65850833e-01 -6.67889714e-01 7.60824800e-01 5.95420897e-01
1.01310067e-01 1.38048172e+00 7.62106897e-03 -5.66206932e-01
-7.52826452e-01 -1.21034765e+00 -7.16054618e-01 -7.01376945e-02
-2.40265921e-01 7.68773317e-01 6.72318518e-01 -3.59203011e-01
3.33415419e-02 4.41892684e-01 -1.93406928e+00 5.79278886e-01
1.08206177e+00 1.15513134e+00 -2.59061843e-01 2.96757340e-01
-5.54806948e-01 1.36375022e+00 -6.29956365e-01 -7.44673789e-01
-1.73946936e-02 -3.27655345e-01 -1.01279545e+00 -3.88419181e-01
3.48886490e-01 1.59153953e-01 -2.98043996e-01 8.88650179e-01
-1.01375148e-01 1.35895565e-01 5.79923689e-01 -1.29321945e+00
-7.84425318e-01 9.04518485e-01 1.82785004e-01 9.89267379e-02
7.51694560e-01 -1.14879943e-01 9.53571320e-01 -4.88380998e-01
8.55559409e-01 1.13509917e+00 5.61095357e-01 4.42965090e-01
-1.10435569e+00 -5.10521889e-01 1.78065315e-01 7.50327289e-01
-1.35529757e+00 -2.08295539e-01 6.89859986e-01 -4.18519855e-01
1.13574851e+00 3.75611693e-01 7.29581714e-01 1.20298469e+00
-2.14903682e-01 6.30341172e-01 9.95392442e-01 -6.95971787e-01
3.28226537e-01 4.71134663e-01 4.53493357e-01 6.93923414e-01
1.01890182e+00 2.03783721e-01 -5.41746795e-01 -2.34697387e-01
7.06340671e-01 -1.89038053e-01 -2.66163707e-01 -4.77756262e-01
-8.15485716e-01 8.89303744e-01 1.43741295e-01 6.02272928e-01
-2.18791351e-01 -1.75866812e-01 3.76868427e-01 5.29089689e-01
3.88984233e-02 7.78753519e-01 -7.14516878e-01 -1.04273640e-01
-7.79602230e-01 6.80194139e-01 1.33711064e+00 1.08276129e+00
7.74078786e-01 2.10478995e-02 -1.18033007e-01 6.50772691e-01
8.74714740e-03 2.57460535e-01 4.42298830e-01 -6.59285903e-01
5.00743806e-01 1.21757174e+00 1.92652106e-01 -1.05697560e+00
-4.59468096e-01 -1.92439556e-01 -7.00367332e-01 5.54344654e-02
5.50053120e-01 -1.96755052e-01 -8.07430565e-01 1.57759368e+00
2.18545288e-01 2.06051335e-01 3.28326821e-01 4.46983039e-01
1.01047730e+00 5.44036865e-01 -1.57882515e-02 -2.57000446e-01
1.02134693e+00 -4.84772652e-01 -6.23541713e-01 -4.12670434e-01
5.95600188e-01 -1.74498290e-01 7.74612904e-01 5.70509970e-01
-1.05146492e+00 -3.45440269e-01 -1.01690745e+00 1.00390635e-01
-9.94625211e-01 -1.75480783e-01 1.15553689e+00 6.25847936e-01
-6.63881242e-01 5.21774828e-01 -3.71753603e-01 -2.56725729e-01
7.16328561e-01 4.55247164e-01 -4.72004950e-01 -5.44193566e-01
-1.69171393e+00 1.38467586e+00 1.17302954e+00 -1.61422089e-01
-4.66703832e-01 -7.74208426e-01 -1.05231714e+00 1.15089647e-01
1.13301921e+00 -7.16802895e-01 1.05753469e+00 -7.49400496e-01
-1.23759985e+00 5.41089356e-01 8.39093104e-02 -6.12170815e-01
3.15013498e-01 -5.69310412e-02 -1.03330159e+00 -2.32217878e-01
-2.43338674e-01 2.04142421e-01 4.50864434e-01 -1.31940019e+00
-7.31837690e-01 -2.62007385e-01 4.97213900e-01 -2.40014911e-01
-7.32138306e-02 -3.42207514e-02 -1.23438723e-01 -4.15085196e-01
-2.49281660e-01 -4.79895473e-01 -1.54793173e-01 -3.77057314e-01
-5.02185285e-01 -3.73647839e-01 7.33162582e-01 -3.48271966e-01
1.52832234e+00 -1.61695349e+00 -5.03279865e-02 6.17643774e-01
-5.16884550e-02 6.35419309e-01 3.04612190e-01 2.96703249e-01
-1.85235575e-01 4.18141007e-01 -2.44337320e-01 6.63244367e-01
2.08184242e-01 6.20118022e-01 -4.91436899e-01 -4.92357641e-01
3.71073574e-01 9.47222412e-01 -9.57335174e-01 -5.71570694e-01
1.55442178e-01 1.08850822e-01 -3.26005727e-01 5.40669262e-02
-7.39140332e-01 -6.52036443e-02 -5.67050517e-01 6.56282067e-01
1.94891259e-01 -3.76685381e-01 6.36985958e-01 1.43921934e-03
3.05075586e-01 4.17999446e-01 -1.50630367e+00 1.40182972e+00
-3.03119242e-01 3.84491235e-01 -5.86249828e-01 -1.29731512e+00
9.75044429e-01 3.92149687e-01 2.12164328e-01 -4.77398872e-01
1.47777155e-01 2.09007412e-01 1.30814105e-01 -8.14919770e-01
2.41237864e-01 -3.24607819e-01 3.91833717e-03 2.52251923e-01
1.86488330e-01 -2.04649523e-01 9.79359865e-01 2.65296618e-03
1.29083884e+00 1.57927781e-01 8.14167202e-01 8.49863291e-02
5.40887535e-01 5.18347442e-01 6.51424468e-01 7.10093379e-01
3.93533796e-01 2.41532433e-03 6.83287859e-01 -1.03557956e+00
-6.57512009e-01 -7.54329026e-01 -1.64324120e-01 6.91071391e-01
-1.81575954e-01 -6.51227057e-01 -1.58289477e-01 -8.89892578e-01
3.52625191e-01 9.27936554e-01 -5.21988928e-01 -2.00452238e-01
-4.91833061e-01 -4.96866882e-01 7.02416062e-01 6.48660243e-01
3.50421816e-01 -1.36328888e+00 -6.32823169e-01 3.98495674e-01
-7.51931369e-02 -1.34394705e+00 4.91231203e-01 2.77020037e-01
-9.47193325e-01 -1.74580169e+00 3.11678976e-01 -5.96663713e-01
5.26553392e-01 -2.77028918e-01 1.62320054e+00 3.77527624e-01
-3.39395553e-01 1.27754077e-01 -4.85006332e-01 -8.23475480e-01
-4.23628420e-01 -3.24371904e-02 -3.34757194e-02 -3.68498027e-01
5.39098978e-01 -5.62195420e-01 1.73899531e-01 6.35633916e-02
-1.12902510e+00 -1.04188034e-03 7.08862305e-01 7.68236101e-01
4.79098260e-01 6.13692582e-01 8.08465719e-01 -1.45474494e+00
8.20392549e-01 -4.46673095e-01 -5.19775510e-01 8.76711965e-01
-7.03917921e-01 3.48699391e-01 8.58698964e-01 -2.29421020e-01
-1.15954447e+00 -2.10170597e-01 1.90229431e-01 -1.03117831e-01
-6.90300643e-01 1.08763885e+00 -4.19371545e-01 7.85421208e-02
1.02527416e+00 -7.27715641e-02 -2.80573189e-01 -1.42388612e-01
4.94098902e-01 2.44853020e-01 4.92563993e-01 -1.08502996e+00
7.32980430e-01 1.35350693e-02 5.32760769e-02 -4.90725845e-01
-1.13875210e+00 1.69659734e-01 -6.96616471e-01 2.20784441e-01
2.64535964e-01 -4.04819876e-01 -5.28923094e-01 -2.70026356e-01
-1.07563043e+00 -3.55742335e-01 -6.91845655e-01 4.75994945e-01
-4.01937872e-01 -6.43006489e-02 -1.50024340e-01 -6.15895033e-01
-6.31328160e-03 -5.75930119e-01 1.34644404e-01 2.15021074e-01
-5.14795184e-01 -1.19139946e+00 1.27736092e-01 2.01454669e-01
3.97028118e-01 4.76530254e-01 1.56627417e+00 -9.36778009e-01
-3.61572176e-01 -4.97420996e-01 -2.34968029e-02 3.01127344e-01
3.40379924e-01 2.16109306e-01 -7.43098497e-01 2.14850515e-01
-4.09470499e-01 -6.11818790e-01 5.01020074e-01 1.57280326e-01
1.45917726e+00 -6.42791569e-01 -5.16156137e-01 9.07015949e-02
1.24033821e+00 4.31454659e-01 6.47890329e-01 2.03173727e-01
2.68032342e-01 5.58079004e-01 4.12325501e-01 1.75391763e-01
4.46720392e-01 5.87708950e-01 1.73482776e-01 3.61453265e-01
1.08858876e-01 -2.76726902e-01 4.77871485e-02 3.12903017e-01
-6.40610754e-01 -1.17149048e-01 -1.13224447e+00 4.43060905e-01
-2.06082797e+00 -1.40323997e+00 1.35610715e-01 2.03408384e+00
1.15380979e+00 5.37245691e-01 3.18175629e-02 7.14214742e-01
5.40829062e-01 -2.51869261e-01 -6.02828264e-01 -4.12678123e-01
-3.66890967e-01 5.46756327e-01 -1.01858877e-01 3.73017102e-01
-9.20689344e-01 9.15648103e-01 7.09028053e+00 6.71861708e-01
-8.70384753e-01 -4.23771203e-01 1.71030238e-01 5.74286282e-02
-4.35471296e-01 9.53448117e-02 -6.42583370e-01 1.70848802e-01
8.24786901e-01 -5.07798791e-01 3.56445581e-01 8.93823624e-01
-2.73489356e-01 -2.64030136e-02 -1.40996039e+00 7.95118630e-01
6.86731115e-02 -1.86104226e+00 4.66771990e-01 -2.61926770e-01
8.48347604e-01 -5.14405429e-01 -4.90797818e-01 7.04029083e-01
7.97603786e-01 -1.29403293e+00 3.65128160e-01 6.57915831e-01
5.86524189e-01 -7.28984118e-01 8.22260678e-01 3.65415215e-01
-9.40395474e-01 -3.94570529e-01 -3.04127723e-01 -4.28029120e-01
-2.40226597e-01 9.04778063e-01 -8.52302134e-01 1.06683528e+00
5.39949894e-01 7.10415423e-01 -8.94246340e-01 1.02898717e+00
-6.23400509e-01 6.53769851e-01 -2.79357284e-01 -2.06667647e-01
-1.52043164e-01 9.83517095e-02 -2.12821644e-02 1.20271635e+00
2.61734366e-01 2.93712527e-01 1.81723595e-01 7.99812198e-01
-5.05157784e-02 4.80067497e-03 -1.10710979e+00 -1.25073284e-01
6.40294433e-01 9.93650496e-01 -5.11451244e-01 -5.25814772e-01
-5.72631538e-01 -6.27057999e-02 6.68223262e-01 4.42446113e-01
-6.11112297e-01 -3.05067986e-01 4.12641525e-01 2.01802015e-01
1.09891444e-01 2.32357625e-03 -2.25031912e-01 -1.17033029e+00
2.96970814e-01 -1.12130332e+00 8.64937901e-01 -7.06394911e-01
-1.36831689e+00 4.51961428e-01 4.03288484e-01 -8.85743618e-01
-6.45386577e-01 -8.78648281e-01 -5.64327478e-01 3.70617300e-01
-1.57650161e+00 -9.79070663e-01 -3.54648203e-01 8.58532548e-01
8.86798799e-02 -3.78469080e-01 1.16216719e+00 2.02738360e-01
-4.28573191e-01 4.09719229e-01 -4.49547172e-01 2.00247496e-01
3.54197711e-01 -1.26120412e+00 -1.95026785e-01 5.57036161e-01
4.47267711e-01 7.63224125e-01 5.97370028e-01 -5.17970264e-01
-1.37076199e+00 -1.06100309e+00 8.50338280e-01 -6.42140031e-01
7.94609606e-01 -3.28657776e-02 -1.23516011e+00 9.33526695e-01
5.79009466e-02 3.45658779e-01 9.36512589e-01 5.13284445e-01
-6.87623799e-01 -3.25273007e-01 -1.13348567e+00 4.40683246e-01
1.23852026e+00 -3.14468831e-01 -1.13307559e+00 1.81192055e-01
2.52571166e-01 -2.97703058e-01 -1.00910449e+00 6.94545031e-01
3.00610602e-01 -8.94950509e-01 8.64870191e-01 -1.35419905e+00
5.82858324e-01 -4.56976920e-01 -7.33928531e-02 -1.39207566e+00
-3.47590774e-01 -4.24255580e-01 -7.57667720e-01 1.07542348e+00
9.22303259e-01 -3.47178966e-01 7.96088874e-01 9.04959857e-01
3.58229578e-02 -9.56006229e-01 -7.47126877e-01 -9.47028041e-01
-1.03150964e-01 -7.39738643e-01 9.36371267e-01 1.29934955e+00
6.38146043e-01 6.01716340e-01 -8.49111676e-02 -4.15474661e-02
2.38088876e-01 5.15799701e-01 6.53699577e-01 -1.82438993e+00
-7.47111589e-02 -6.37488246e-01 -7.15294778e-01 -3.12464118e-01
2.75672346e-01 -1.19467807e+00 -4.11575794e-01 -1.93618464e+00
-3.61222848e-02 -5.73224783e-01 -3.02151620e-01 1.06094301e+00
-8.63103643e-02 -4.31251466e-01 -1.65377170e-01 -2.94603646e-01
-5.97907126e-01 9.35411081e-02 1.15880048e+00 -3.07813138e-01
-2.88762189e-02 -2.04087734e-01 -9.44817662e-01 9.23923492e-01
6.96216226e-01 -2.45298609e-01 -7.95818031e-01 -1.35552809e-01
8.55187356e-01 -5.21972887e-02 4.05974686e-01 -1.09486830e+00
3.91691267e-01 -7.04061568e-01 4.62779611e-01 -1.44295052e-01
2.37529948e-02 -1.06318915e+00 4.40514237e-01 3.25732440e-01
-3.03647131e-01 -6.59402162e-02 3.60884994e-01 3.87908518e-01
-5.73868215e-01 -2.67613322e-01 3.52046162e-01 -1.57044888e-01
-1.06846631e+00 1.88439667e-01 1.51040986e-01 4.43891883e-01
1.17764592e+00 -1.98635966e-01 -5.99908292e-01 -8.64538252e-02
-7.89126754e-01 2.33737260e-01 5.07954583e-02 5.61249495e-01
7.61637986e-01 -1.37230480e+00 -6.58206105e-01 2.07975879e-01
3.09983522e-01 2.10562304e-01 -1.62942141e-01 2.74364799e-01
-2.80902177e-01 5.40858328e-01 -3.65498424e-01 -2.64372062e-02
-8.04644525e-01 8.10263395e-01 2.57339925e-01 -5.78959823e-01
-6.14594877e-01 3.69441092e-01 -4.30862635e-01 -4.08017069e-01
1.90704495e-01 -6.65625334e-01 -6.26396418e-01 -5.56176268e-02
5.93895316e-01 2.33613551e-01 2.43029445e-01 4.98743877e-02
-3.27219337e-01 2.61066526e-01 6.97585046e-02 4.58694667e-01
1.46504462e+00 6.57697976e-01 5.40923104e-02 4.80456680e-01
2.08521232e-01 -2.75677979e-01 -6.30051911e-01 -4.39498395e-01
5.87570071e-01 -4.37348753e-01 -4.53079373e-01 -1.19551063e+00
-9.75258350e-01 5.47418177e-01 -5.52216098e-02 7.02309549e-01
8.52391303e-01 3.86372544e-02 1.85231149e-01 1.03865921e+00
6.54239774e-01 -1.22774673e+00 -2.68776298e-01 6.99586689e-01
9.64210987e-01 -1.31210506e+00 2.40682662e-01 -7.45559037e-01
-5.35647094e-01 1.19751573e+00 7.52641082e-01 1.31671041e-01
7.55051374e-01 5.71460605e-01 -1.08146690e-01 -4.17048484e-01
-1.08288836e+00 -2.54096717e-01 2.62872517e-01 7.76237965e-01
4.81266409e-01 4.51415591e-02 -1.96710929e-01 9.21376765e-01
-3.30393434e-01 5.78334868e-01 2.41870031e-01 1.06206298e+00
-4.32474226e-01 -1.33210468e+00 -2.29104400e-01 1.05121028e+00
-2.43511736e-01 1.20991863e-01 -7.02319562e-01 1.10832012e+00
4.97596681e-01 9.25702810e-01 -4.55648124e-01 -3.47866535e-01
7.24558473e-01 4.53109086e-01 8.05470288e-01 -8.03805232e-01
-3.52056712e-01 -9.48778689e-01 5.61399758e-01 -2.84218043e-01
-4.98174071e-01 -4.73803103e-01 -1.37250209e+00 -2.43566319e-01
-1.94894001e-01 5.55143356e-01 1.52132154e-01 1.31087768e+00
2.45329887e-01 5.56953609e-01 -4.25779708e-02 -1.94281995e-01
-1.79529965e-01 -7.55941987e-01 -4.99610752e-01 6.77200675e-01
-2.05499247e-01 -1.04977190e+00 3.94176170e-02 1.38511688e-01] | [9.049936294555664, 7.293849945068359] |
7ab4ed52-2962-4c30-b171-aa1065d74450 | cross-architecture-distillation-using | null | null | https://openreview.net/forum?id=o9DnX55PEAo | https://openreview.net/pdf?id=o9DnX55PEAo | Cross-Architecture Distillation Using Bidirectional CMOW Embeddings | Large pretrained language models (PreLMs) are revolutionizing natural language processing across all benchmarks. However, their sheer size is prohibitive for small laboratories or deployment on mobile devices. Approaches like pruning and distillation reduce the model size but typically retain the same model architecture. In contrast, we explore distilling PreLMs into a different, more efficient architecture CMOW, which embeds each word as a matrix and uses matrix multiplication to encode sequences. We extend the CMOW architecture and its CMOW/CBOW-Hybrid variant with a bidirectional component, per-token representations for distillation during pretraining, and a two-sequence encoding scheme that facilitates downstream tasks on sentence pairs such as natural language inferencing. Our results show that the embedding-based models yield scores comparable to DistilBERT on QQP and RTE, while using only half of its parameters and providing three times faster inference speed. We match or exceed the scores of ELMo, and only fall behind more expensive models on linguistic acceptability. Still, our distilled bidirectional CMOW/CBOW-Hybrid model more than doubles the scores on linguistic acceptability compared to previous cross-architecture distillation approaches. Furthermore, our experiments confirm the positive effects of bidirection and the two-sequence encoding scheme. | ['Ansgar Scherp', 'Angelina Sonderecker', 'Henrik Ferdinand Nölscher', 'Christoph Meyer', 'Isabelle Cuber', 'Lukas Paul Achatius Galke'] | 2021-09-29 | null | null | null | null | ['linguistic-acceptability'] | ['natural-language-processing'] | [-2.62152050e-02 5.74430943e-01 -2.53624171e-01 -6.02134407e-01
-9.98106480e-01 -4.88420814e-01 4.58127707e-01 3.10584515e-01
-1.06194627e+00 6.38873279e-01 4.36765879e-01 -9.40865815e-01
4.03611928e-01 -9.42948699e-01 -8.90321791e-01 -1.93042025e-01
2.10238732e-02 6.45237267e-01 -1.00795865e-01 -5.65101624e-01
5.52402735e-02 -1.10347278e-01 -1.21888542e+00 5.79744041e-01
9.27600920e-01 8.22964489e-01 6.49470985e-02 9.38512981e-01
-4.23584640e-01 7.16477811e-01 -3.35567385e-01 -9.62408304e-01
1.85924590e-01 1.63639951e-02 -9.01565194e-01 -9.03707325e-01
6.40197396e-01 -5.28291702e-01 -1.67222530e-01 6.31323934e-01
6.00065053e-01 2.00638264e-01 3.91193092e-01 -9.64788258e-01
-9.91080642e-01 1.46217120e+00 -3.67091537e-01 1.33761674e-01
4.88289028e-01 3.39629114e-01 1.65784883e+00 -1.06751883e+00
4.07713205e-01 1.69039583e+00 1.02505922e+00 7.43420064e-01
-1.59810162e+00 -5.95429599e-01 4.86048669e-01 1.20308243e-01
-1.24253488e+00 -5.25623024e-01 2.71857798e-01 -3.57790217e-02
2.02437735e+00 4.25156355e-01 4.94017571e-01 1.19308138e+00
2.79048383e-01 1.00150597e+00 8.81178260e-01 -5.70992649e-01
1.47732228e-01 2.15167448e-01 4.83892679e-01 7.87872732e-01
3.52548808e-01 -3.32662128e-02 -4.16884094e-01 -9.89328101e-02
8.69645849e-02 -2.81808287e-01 1.83789268e-01 1.07989006e-01
-9.93876755e-01 9.51543689e-01 3.91048282e-01 2.49699488e-01
5.59087396e-02 4.36471254e-01 6.31601930e-01 5.73009133e-01
5.52420676e-01 8.73626947e-01 -6.11842036e-01 -3.87868643e-01
-1.00736690e+00 2.53452092e-01 9.35403228e-01 8.39718640e-01
7.97215164e-01 1.38057336e-01 -2.28412777e-01 9.23025489e-01
1.46793008e-01 3.43733221e-01 6.79062486e-01 -9.29826379e-01
8.66348445e-01 5.21907449e-01 -3.82030904e-01 -6.97999299e-01
-3.80331397e-01 -1.44386172e-01 -6.16751790e-01 -5.26937284e-02
1.87957495e-01 -4.07060772e-01 -7.17078805e-01 2.04204202e+00
6.64538145e-02 -6.78633247e-03 9.59226713e-02 3.54519278e-01
6.50299907e-01 9.22673643e-01 3.11486125e-01 2.75752574e-01
1.38761830e+00 -1.07489157e+00 -3.81878972e-01 -8.08238387e-01
1.38670027e+00 -4.88390326e-01 1.66407692e+00 5.86006761e-01
-1.50176132e+00 -6.63646519e-01 -1.17220736e+00 -9.08134341e-01
-4.80684996e-01 -9.10684690e-02 8.51951718e-01 8.57792675e-01
-1.32487416e+00 6.36190176e-01 -8.95666897e-01 -2.75831670e-02
1.30953506e-01 5.46773911e-01 -2.94174850e-01 -2.04932049e-01
-1.53348064e+00 1.29773760e+00 5.99021375e-01 5.44924214e-02
-3.87907118e-01 -1.11139190e+00 -1.34672928e+00 2.08150804e-01
-2.16005109e-02 -9.08368945e-01 1.35017312e+00 -4.32777584e-01
-1.65397835e+00 7.79164612e-01 -3.24561298e-01 -7.92473376e-01
1.77444816e-01 -6.44589484e-01 -2.90793598e-01 -4.22792882e-01
-1.54112935e-01 1.02914894e+00 2.88530886e-01 -6.69042945e-01
-3.85678232e-01 -4.50798683e-02 5.09899557e-01 1.82721213e-01
-4.99995887e-01 -1.43862635e-01 -1.81891605e-01 -4.68826532e-01
-3.51480663e-01 -8.01637530e-01 -2.84784913e-01 -2.15034515e-01
-5.15271008e-01 -3.93960416e-01 2.90037245e-01 -7.37242520e-01
1.72199273e+00 -2.01592493e+00 3.42963547e-01 -2.41786074e-02
1.97685361e-01 4.00458992e-01 -5.50526321e-01 5.00448823e-01
-9.84130800e-02 4.62989867e-01 -4.63872224e-01 -9.20165777e-01
3.43332499e-01 5.93928754e-01 -2.75171429e-01 -1.40954480e-01
3.92543405e-01 1.27237380e+00 -7.87043095e-01 -4.43768740e-01
1.53361067e-01 3.04953367e-01 -1.33435833e+00 1.22729294e-01
-3.40055585e-01 -4.55083132e-01 2.38656282e-01 1.68459639e-01
6.30711615e-01 1.19144984e-01 2.57134825e-01 -1.67390294e-02
7.99158514e-02 1.16484416e+00 -1.07490718e+00 2.01759505e+00
-1.29654264e+00 5.63196480e-01 7.37936720e-02 -7.71174967e-01
6.56774044e-01 1.23386353e-01 -1.95163921e-01 -6.97293043e-01
-1.31340444e-01 3.35572273e-01 3.24960321e-01 -3.66043806e-01
8.56582284e-01 -2.86262006e-01 -2.74650365e-01 5.71907520e-01
4.51254323e-02 -4.24504846e-01 3.78218085e-01 4.64385927e-01
1.03019977e+00 2.62353532e-02 1.98392078e-01 -2.13307366e-01
3.53409916e-01 -4.99873698e-01 4.63400543e-01 7.17611730e-01
2.22046375e-01 2.95734107e-01 7.07547486e-01 -3.63293797e-01
-9.87859428e-01 -1.15235591e+00 -1.39758497e-01 1.58612967e+00
-3.17152917e-01 -8.99818540e-01 -8.24941635e-01 -6.05335057e-01
1.74375281e-01 1.28102887e+00 -4.16797578e-01 -4.57256019e-01
-9.50185895e-01 -9.63039756e-01 1.16693807e+00 8.12045455e-01
1.15908846e-01 -8.03307295e-01 -3.59497845e-01 2.27122471e-01
-3.32177103e-01 -9.33274746e-01 -2.13860571e-01 4.19972450e-01
-8.40638995e-01 -3.62741083e-01 -1.09112710e-01 -6.54422641e-01
3.87440979e-01 -3.65929127e-01 1.58991468e+00 9.96562690e-02
-1.45239249e-01 -3.11099619e-01 -5.89118265e-02 -3.23681712e-01
-4.19733852e-01 4.66733187e-01 1.85587183e-01 -6.89432323e-01
6.54882729e-01 -5.74928999e-01 -4.02874947e-01 -2.95609772e-01
-6.89594150e-01 5.63222356e-02 6.93965495e-01 1.04608345e+00
8.93966481e-02 -5.54914415e-01 3.71853262e-01 -1.29681015e+00
8.43428552e-01 -5.80579996e-01 -2.02620506e-01 5.97308241e-02
-8.48274350e-01 5.45299828e-01 6.69898868e-01 -2.43279696e-01
-1.13801014e+00 -3.69774759e-01 -8.17010760e-01 2.33011752e-01
2.14830209e-02 5.19629598e-01 -6.65185899e-02 4.85536367e-01
7.47260332e-01 -2.91759372e-01 -1.33376956e-01 -4.54866260e-01
1.05788565e+00 6.25988185e-01 3.31816703e-01 -9.56123471e-01
5.34855545e-01 -1.58188981e-03 -5.59706330e-01 -6.80944264e-01
-8.38429511e-01 -1.32095173e-01 -4.10966247e-01 7.24684715e-01
8.81615996e-01 -9.57580447e-01 -6.43320262e-01 1.18380405e-01
-1.47960556e+00 -5.68478346e-01 -5.29549658e-01 4.07582283e-01
-1.76058725e-01 4.39573616e-01 -1.18964577e+00 -7.08571672e-01
-6.07779622e-01 -1.13542342e+00 1.19712543e+00 -3.98576766e-01
-8.93165112e-01 -1.19902968e+00 2.56264627e-01 2.59819567e-01
6.29925966e-01 -3.75359416e-01 1.47083557e+00 -5.51265895e-01
-1.39762089e-01 1.18337665e-02 -1.97358876e-01 5.57016909e-01
-3.60949010e-01 -1.01844780e-01 -1.14627492e+00 -1.36492386e-01
-2.23650545e-01 -6.69365883e-01 1.18193316e+00 2.89237738e-01
1.14934158e+00 -5.71340919e-01 -1.28114522e-01 8.25643659e-01
1.20645905e+00 -1.61229521e-01 6.22320294e-01 2.73054302e-01
6.39251709e-01 4.08139080e-01 7.92250782e-02 1.27735630e-01
5.69140375e-01 5.27808607e-01 1.23936817e-01 -1.49067894e-01
-7.12922439e-02 -4.02914137e-01 8.52154016e-01 1.24941289e+00
1.19757548e-01 -3.14759403e-01 -9.30936158e-01 5.02285004e-01
-1.57561731e+00 -7.70063281e-01 7.81196728e-02 1.85873950e+00
1.35465348e+00 4.66257691e-01 -1.77799374e-01 1.62357107e-01
2.63584256e-02 2.22982526e-01 -3.28653604e-01 -1.59139240e+00
-8.23313091e-03 8.31736028e-01 3.62296253e-01 1.21204937e+00
-6.58441067e-01 1.17474639e+00 6.77895546e+00 8.64349127e-01
-8.19963634e-01 2.01721922e-01 8.55591774e-01 -5.04204631e-01
-8.43693674e-01 -1.22009806e-01 -1.07320416e+00 2.85827637e-01
1.42628014e+00 1.14414170e-01 4.67312455e-01 8.95729303e-01
-1.58304229e-01 -1.23103805e-01 -1.64720762e+00 7.89383292e-01
6.13884479e-02 -1.29128873e+00 2.13638440e-01 -2.91480094e-01
4.54646319e-01 3.92428935e-01 2.02076450e-01 9.51259255e-01
7.17762887e-01 -1.25765312e+00 5.78782022e-01 7.80202299e-02
8.73138785e-01 -7.73165107e-01 7.96534538e-01 3.16233248e-01
-9.87556100e-01 -1.58251107e-01 -4.91857141e-01 -6.05192721e-01
3.13663304e-01 5.58302104e-01 -7.29361534e-01 1.89655200e-01
5.14453173e-01 3.54731232e-01 -6.02405190e-01 2.27925226e-01
-4.05436933e-01 7.14488328e-01 -3.83333415e-01 -1.52290791e-01
4.49237525e-01 -1.54518366e-01 2.24904567e-01 1.82276225e+00
1.05744302e-01 -1.69748217e-01 -2.17143223e-01 8.31124425e-01
-2.97862798e-01 -8.39997381e-02 -6.91141427e-01 5.68505228e-02
6.10179365e-01 1.10305083e+00 7.30769858e-02 -6.15901470e-01
-4.16728079e-01 9.17105258e-01 9.14058268e-01 2.06342176e-01
-8.82923961e-01 -6.90728664e-01 1.10109866e+00 -4.69536334e-02
5.61520681e-02 -3.89104545e-01 -6.90898061e-01 -1.23536026e+00
1.33707136e-01 -9.30769563e-01 2.68447548e-01 -4.45972204e-01
-1.15444100e+00 7.99358070e-01 9.87880584e-03 -3.83078009e-01
-7.77681947e-01 -9.33538735e-01 -6.16178930e-01 1.11762047e+00
-1.53965127e+00 -9.89233375e-01 2.07313061e-01 1.81494236e-01
5.58496654e-01 1.46860331e-01 1.18632185e+00 4.33635950e-01
-8.16603184e-01 1.10981882e+00 -8.92322510e-02 1.68543473e-01
6.27782047e-01 -1.56861055e+00 1.13300776e+00 8.22923362e-01
3.28949362e-01 1.23254812e+00 4.89275843e-01 -2.61842191e-01
-1.34277713e+00 -1.05144227e+00 1.37333500e+00 -8.51993978e-01
7.17015803e-01 -8.57706964e-01 -9.05293047e-01 1.09634936e+00
5.03799438e-01 -3.14910352e-01 7.06185460e-01 9.14784491e-01
-7.77588308e-01 -1.11149184e-01 -7.99236834e-01 8.64220262e-01
1.10929012e+00 -7.81642616e-01 -9.79511201e-01 1.07820496e-01
1.17831469e+00 -3.86370480e-01 -7.35924482e-01 1.09709986e-01
7.01064646e-01 -8.38860273e-01 1.08287680e+00 -9.39029157e-01
9.05060232e-01 3.44292849e-01 -2.32317805e-01 -1.56315219e+00
-3.17830592e-01 -5.12508810e-01 -2.06584573e-01 1.21853709e+00
1.03233802e+00 -7.13130236e-01 7.89188802e-01 9.06867683e-01
-3.09786320e-01 -1.07183349e+00 -8.50232124e-01 -8.08044195e-01
6.77183986e-01 -9.10013020e-01 6.70325279e-01 7.04143703e-01
4.02564883e-01 7.52804756e-01 -1.28551751e-01 -3.15675288e-01
1.53741166e-01 -2.13198140e-01 4.94945526e-01 -6.84797287e-01
-8.18612337e-01 -6.36954665e-01 -4.23919447e-02 -1.31413901e+00
4.72871602e-01 -1.28716648e+00 3.39022018e-02 -1.52688646e+00
-2.54288055e-02 -6.92398667e-01 -2.61340082e-01 6.20311201e-01
-4.02179658e-01 2.42067575e-01 1.21185720e-01 -5.31093001e-01
-2.16648296e-01 4.52199578e-01 4.33946937e-01 -2.65618116e-01
-1.77500322e-01 -4.99973297e-01 -9.60264206e-01 8.14525247e-01
7.79416203e-01 -2.92755902e-01 -5.67948937e-01 -1.36280143e+00
6.91159904e-01 -4.67767179e-01 -3.58471051e-02 -8.03002417e-01
7.78196305e-02 1.41920507e-01 1.78401899e-02 -9.24697742e-02
7.34916925e-01 -3.40672404e-01 -4.47590798e-01 5.56311786e-01
-7.78579950e-01 5.48668981e-01 5.88860273e-01 2.13517725e-01
-1.20028205e-01 -2.35570699e-01 5.84195793e-01 -1.49066404e-01
-5.43302417e-01 -4.06302735e-02 -3.85932505e-01 4.50330347e-01
5.32828867e-01 -3.81672122e-02 -3.10854703e-01 -4.77129482e-02
-4.98945504e-01 2.24119395e-01 -2.87643448e-02 4.11366940e-01
6.17412448e-01 -1.09041524e+00 -7.01369524e-01 4.45335418e-01
4.47908677e-02 1.70500830e-01 -6.50449703e-03 6.42166793e-01
-6.09825075e-01 7.05456913e-01 1.04375770e-02 -2.47339517e-01
-1.11563623e+00 3.63529086e-01 1.50723964e-01 -5.33987522e-01
-4.25212115e-01 1.46147728e+00 1.72724128e-01 -1.04070342e+00
1.86178416e-01 -1.04542685e+00 3.02520633e-01 -1.94689352e-02
4.65379745e-01 4.28554833e-01 3.31831753e-01 -1.26414169e-02
-3.14395130e-01 2.43049204e-01 -3.30243856e-01 -2.69901931e-01
1.28567457e+00 4.99357581e-02 -2.73202807e-01 5.02195418e-01
1.29260135e+00 2.03124881e-01 -6.92366421e-01 -1.99456081e-01
1.21168993e-01 -3.72851938e-02 -4.49878126e-02 -7.41460860e-01
-5.63657522e-01 1.39591908e+00 1.67999677e-02 -1.19423784e-01
6.90942645e-01 -2.40812808e-01 1.33181322e+00 8.53658020e-01
1.02844857e-01 -1.16567910e+00 -2.13806286e-01 8.28003764e-01
6.78534865e-01 -1.14915979e+00 -2.32230231e-01 -1.61683425e-01
-6.24653399e-01 7.34100044e-01 8.56996596e-01 -7.86124542e-02
4.81103659e-01 7.50768363e-01 -8.76763463e-02 2.75010139e-01
-1.28375316e+00 1.58819586e-01 6.87329322e-02 3.30083162e-01
1.07831442e+00 2.79597253e-01 -2.53089905e-01 6.52536392e-01
-7.81595886e-01 -3.00347090e-01 2.07239062e-01 5.65705538e-01
-3.94616067e-01 -1.43689454e+00 1.54046178e-01 4.65221077e-01
-3.80270064e-01 -9.93208766e-01 -2.23674297e-01 8.49879742e-01
2.07409844e-01 8.73212278e-01 3.24055761e-01 -4.80239093e-01
3.08347076e-01 3.46536368e-01 3.28081191e-01 -9.88263190e-01
-8.76594543e-01 -6.47935867e-01 6.04396999e-01 -8.94798756e-01
3.83440197e-01 -2.39566296e-01 -1.31245887e+00 -7.49397755e-01
-2.59194791e-01 1.94277495e-01 6.01961017e-01 8.53437781e-01
5.61083376e-01 5.24109364e-01 1.15119591e-01 -6.59660220e-01
-9.12891567e-01 -1.06309330e+00 -1.12836197e-01 1.92076623e-01
3.82327020e-01 -2.10702077e-01 -1.51651084e-01 -3.15006375e-01] | [10.755297660827637, 8.623562812805176] |
60931dd7-6e77-4ac9-a6fc-fd6a9a003044 | svnr-spatially-variant-noise-removal-with | 2306.16052 | null | https://arxiv.org/abs/2306.16052v1 | https://arxiv.org/pdf/2306.16052v1.pdf | SVNR: Spatially-variant Noise Removal with Denoising Diffusion | Denoising diffusion models have recently shown impressive results in generative tasks. By learning powerful priors from huge collections of training images, such models are able to gradually modify complete noise to a clean natural image via a sequence of small denoising steps, seemingly making them well-suited for single image denoising. However, effectively applying denoising diffusion models to removal of realistic noise is more challenging than it may seem, since their formulation is based on additive white Gaussian noise, unlike noise in real-world images. In this work, we present SVNR, a novel formulation of denoising diffusion that assumes a more realistic, spatially-variant noise model. SVNR enables using the noisy input image as the starting point for the denoising diffusion process, in addition to conditioning the process on it. To this end, we adapt the diffusion process to allow each pixel to have its own time embedding, and propose training and inference schemes that support spatially-varying time maps. Our formulation also accounts for the correlation that exists between the condition image and the samples along the modified diffusion process. In our experiments we demonstrate the advantages of our approach over a strong diffusion model baseline, as well as over a state-of-the-art single image denoising method. | ['Dani Lischinski', 'Daniel Cohen-Or', 'Alex Rav Acha', 'Assaf Zomet', 'Dana Berman', 'Yaron Brodsky', 'Naama Pearl'] | 2023-06-28 | null | null | null | null | ['image-denoising'] | ['computer-vision'] | [ 5.17995834e-01 -1.08074732e-01 6.06318951e-01 -2.03630000e-01
-7.15407014e-01 -5.10853767e-01 1.05531180e+00 -3.14228415e-01
-5.50338745e-01 4.04901922e-01 3.83154005e-01 5.34169376e-03
-1.04580827e-01 -8.16156805e-01 -6.23000979e-01 -1.34307289e+00
1.48838878e-01 2.31479630e-01 2.44027480e-01 -2.83510327e-01
1.74994767e-02 4.77500737e-01 -1.08416617e+00 1.81470439e-01
6.86287284e-01 6.12666070e-01 3.99320036e-01 9.36319828e-01
2.08440602e-01 7.55569398e-01 -6.20619476e-01 -3.46224517e-01
3.38159382e-01 -6.77320957e-01 -2.88456291e-01 3.07675242e-01
4.89489108e-01 -4.44356412e-01 -7.58706629e-01 1.13817751e+00
5.73599160e-01 3.61971647e-01 7.08520532e-01 -6.33751690e-01
-9.99395490e-01 5.31592011e-01 -6.18870318e-01 2.83277214e-01
-1.20538414e-01 4.88238335e-01 5.73186278e-01 -5.23494601e-01
9.30439174e-01 1.45181608e+00 7.88544059e-01 7.76361883e-01
-1.75419712e+00 -1.67249754e-01 1.65710241e-01 -4.34069559e-02
-9.94158149e-01 -5.75123131e-01 9.42794740e-01 -2.69232988e-01
7.72776306e-01 2.57212073e-02 5.14856040e-01 1.49893475e+00
3.69757056e-01 6.95822716e-01 1.33016920e+00 -3.28253686e-01
4.49755788e-01 -3.75241607e-01 -1.28923450e-02 2.47975186e-01
1.77921616e-02 2.10293531e-01 -2.98643291e-01 2.73117051e-02
1.05497766e+00 5.11587001e-02 -4.03316647e-01 -2.28746504e-01
-1.14949763e+00 6.82631254e-01 5.30433714e-01 5.12159526e-01
-7.78804302e-01 5.38688302e-01 1.45985842e-01 4.03187186e-01
8.45854938e-01 9.62417722e-02 -2.98075788e-02 2.70731207e-02
-1.31591594e+00 2.75629848e-01 6.82887912e-01 4.78281677e-01
6.30929410e-01 2.78984427e-01 -3.04513216e-01 6.55103743e-01
1.20777130e-01 5.91988325e-01 3.07037026e-01 -1.40640521e+00
-8.07352737e-03 -2.30504915e-01 1.45515695e-01 -8.01999152e-01
-5.01645654e-02 -4.83943641e-01 -1.46086442e+00 6.39054835e-01
6.38648152e-01 -7.28759542e-02 -1.38041437e+00 1.89973319e+00
1.36184439e-01 5.20903289e-01 6.17387183e-02 7.57051885e-01
8.35412666e-02 8.21563363e-01 -4.95107397e-02 -3.93704802e-01
8.73708367e-01 -7.97150433e-01 -9.62635756e-01 -2.53341049e-01
5.87878972e-02 -7.19794393e-01 7.85504341e-01 7.27809548e-01
-1.44650757e+00 -6.47319376e-01 -8.27030480e-01 -2.82053977e-01
-1.55282691e-01 -3.36699367e-01 3.77340794e-01 5.13108194e-01
-1.45920491e+00 1.11144447e+00 -1.35555458e+00 -2.94550151e-01
3.94813687e-01 5.22949584e-02 -2.08351955e-01 -5.11326492e-01
-9.11830783e-01 9.15249884e-01 -2.59373635e-01 4.83825058e-01
-1.36389017e+00 -6.27710760e-01 -7.66079366e-01 -4.97229733e-02
1.57654062e-01 -9.90896642e-01 1.04970133e+00 -9.67807353e-01
-1.52355838e+00 6.34556234e-01 -3.65313590e-01 -6.16407454e-01
9.56151247e-01 -2.41950199e-01 -2.24650279e-01 2.42413372e-01
-8.19947049e-02 4.82129991e-01 1.51090705e+00 -1.53308463e+00
-1.09403266e-03 -3.17080230e-01 -1.02428541e-01 -8.39341730e-02
4.21291566e-04 -2.18293339e-01 -5.33194542e-01 -1.04372680e+00
1.73641860e-01 -7.07484603e-01 -6.44842863e-01 2.50982732e-01
-2.07933947e-01 3.39715004e-01 8.60332549e-01 -9.55043137e-01
8.90652716e-01 -2.35163093e+00 5.88457942e-01 2.94082642e-01
3.96692872e-01 1.41097501e-01 -4.40200388e-01 4.13779885e-01
1.83395650e-02 3.28855067e-02 -7.56282568e-01 -9.63248372e-01
-3.97598967e-02 7.03673065e-01 -3.39588404e-01 6.59232616e-01
2.43223920e-01 8.02036524e-01 -9.47315395e-01 -1.95886716e-02
3.02940339e-01 1.16546023e+00 -4.02689457e-01 8.65233168e-02
-1.16775222e-01 7.68872619e-01 -4.40550409e-02 1.13569550e-01
8.25826228e-01 2.35565212e-02 -1.03880294e-01 -3.46859276e-01
2.03659832e-02 -1.47809833e-01 -1.27469611e+00 1.82718444e+00
-5.49644232e-01 6.02278709e-01 6.54801130e-01 -9.93587494e-01
5.80321491e-01 3.75776172e-01 3.61722499e-01 -6.05997264e-01
5.32610230e-02 1.58403605e-01 -1.25120413e-02 -3.63845587e-01
1.53797001e-01 -5.27651012e-01 4.30734545e-01 4.95525181e-01
2.24571422e-01 -6.23449981e-01 3.23428631e-01 3.64068955e-01
1.37000704e+00 1.83915213e-01 -2.72845834e-01 -1.77328780e-01
3.62143219e-01 -5.04622281e-01 3.31589639e-01 1.11548841e+00
-6.78671673e-02 9.86746788e-01 2.83278376e-01 -1.47940159e-01
-1.21960092e+00 -1.24846709e+00 9.12017077e-02 4.33511823e-01
-1.83857810e-02 -1.21497840e-01 -1.05755794e+00 -3.66942197e-01
-2.37559319e-01 7.79264987e-01 -9.45673883e-01 -2.10753188e-01
-7.31254816e-01 -8.87541413e-01 4.15035456e-01 3.28921378e-01
5.55643022e-01 -8.21602583e-01 -2.28611544e-01 3.72716993e-01
-3.70629281e-02 -8.57554913e-01 -5.27703106e-01 4.31053579e-01
-9.83225584e-01 -5.41323066e-01 -1.18004334e+00 -5.95059574e-01
6.36327088e-01 3.16935390e-01 1.04905808e+00 5.63067943e-02
-3.00652266e-01 5.09214938e-01 3.17698643e-02 9.21894517e-03
-6.41670763e-01 -3.61369967e-01 -2.20006913e-01 2.87958056e-01
-1.84463143e-01 -9.97574508e-01 -8.99235427e-01 -6.59334706e-03
-1.54114807e+00 -1.50878936e-01 6.32657230e-01 9.88172770e-01
7.06954241e-01 4.08809334e-01 1.84497848e-01 -8.59974325e-01
8.13656390e-01 -3.49677682e-01 -3.67597938e-01 2.57296823e-02
-5.40031493e-01 2.53271013e-01 4.88500446e-01 -8.11102390e-01
-1.37976348e+00 1.66544840e-02 -4.17381406e-01 -5.44078529e-01
-1.21750146e-01 3.25673848e-01 -1.68262348e-02 -1.05850575e-02
7.76174724e-01 3.99400055e-01 2.43596062e-01 -7.34039068e-01
8.03559601e-01 3.92360613e-02 7.85531521e-01 -4.94900733e-01
1.11462629e+00 1.01990175e+00 1.00977361e-01 -1.04298341e+00
-6.84310019e-01 -3.61515135e-01 -7.45091081e-01 -4.38062511e-02
9.30443704e-01 -7.35879481e-01 -2.41131574e-01 8.87325406e-01
-1.15342557e+00 -8.85533392e-01 -6.73469901e-01 2.21940011e-01
-6.40078783e-01 5.41753471e-01 -1.11473799e+00 -7.89058566e-01
-1.01399105e-02 -1.10292196e+00 1.06923866e+00 -1.27477631e-01
-4.51983362e-02 -1.53780437e+00 1.43520027e-01 -1.28100604e-01
8.30799103e-01 1.85290769e-01 6.67153895e-01 -3.00682522e-02
-6.14881098e-01 -1.69932500e-01 -8.93329903e-02 9.64880943e-01
7.57510364e-02 -5.15684672e-02 -9.79259193e-01 -2.45437115e-01
8.27268004e-01 2.78823748e-02 1.52623570e+00 8.42229307e-01
8.10845912e-01 -1.58624165e-02 -8.89239535e-02 7.37236202e-01
1.50975263e+00 -1.29395455e-01 1.05401516e+00 1.56255260e-01
6.20415151e-01 4.08910275e-01 -1.82888582e-01 3.15003060e-02
1.94211584e-02 3.00967097e-01 2.97622174e-01 -4.30295378e-01
-6.88866198e-01 5.67892268e-02 4.37273741e-01 7.85483778e-01
-2.55313694e-01 -4.37428594e-01 -5.48603475e-01 5.87650836e-01
-1.59430671e+00 -1.35285902e+00 -2.77356744e-01 1.96525264e+00
9.17726755e-01 2.45735437e-01 -2.90823072e-01 2.28088006e-01
3.67660195e-01 4.81746197e-01 -5.41287601e-01 -3.31717916e-02
-5.58434725e-01 5.14766276e-01 4.17267293e-01 9.60428298e-01
-1.02630734e+00 6.28042638e-01 6.80910349e+00 7.58897364e-01
-1.08878255e+00 3.78052384e-01 5.90001822e-01 -4.17052861e-03
-4.66346949e-01 -2.88405437e-02 -2.30395451e-01 2.93220490e-01
8.25879276e-01 1.05673872e-01 8.42318892e-01 1.49884835e-01
5.88839531e-01 -1.83090821e-01 -1.03330243e+00 8.92230392e-01
6.88972231e-03 -1.08467400e+00 4.72165011e-02 3.02535705e-02
9.34944749e-01 6.90398142e-02 3.06967378e-01 -1.19009674e-01
5.30450761e-01 -9.52073574e-01 7.79922843e-01 9.93934155e-01
2.65598178e-01 -2.02678367e-01 5.53895533e-01 4.69190538e-01
-6.33658886e-01 1.84033006e-01 -2.84642458e-01 -2.05981154e-02
6.07959628e-01 1.14415348e+00 -1.43576674e-02 2.91678846e-01
5.99120677e-01 9.43923116e-01 -4.58022982e-01 9.13917482e-01
-4.90962744e-01 8.85532439e-01 -4.47002292e-01 7.12927997e-01
2.33783200e-01 -6.03902459e-01 5.31934798e-01 1.25927973e+00
5.27139843e-01 1.91336557e-01 -1.97430164e-01 9.01128113e-01
-2.90123932e-02 -5.08121073e-01 -4.94193405e-01 1.93010002e-01
-2.04479232e-01 1.11523736e+00 -8.29084098e-01 -4.83525693e-01
-3.30254734e-01 1.60588586e+00 -3.86810862e-02 1.01432204e+00
-5.66082656e-01 2.87463749e-03 6.44439340e-01 -2.45556813e-02
5.64630747e-01 -5.47045112e-01 -3.53838831e-01 -1.29386759e+00
8.72051052e-04 -7.73471832e-01 -7.45143965e-02 -9.02595222e-01
-1.63530385e+00 5.49159169e-01 -2.35881433e-01 -7.97900677e-01
-6.10061884e-02 -3.89962286e-01 -7.64726102e-01 1.21948826e+00
-1.58471954e+00 -1.22903144e+00 -2.09524259e-01 6.65433407e-01
5.05354464e-01 5.31604171e-01 5.50505757e-01 2.76328951e-01
-2.79977262e-01 -2.25339010e-02 5.40327072e-01 -2.27365553e-01
8.51137161e-01 -1.41041994e+00 6.83490753e-01 1.18919480e+00
1.46465540e-01 7.84218073e-01 1.08526218e+00 -7.69344985e-01
-1.33724678e+00 -8.45349193e-01 4.53855723e-01 -4.65259850e-01
8.22515070e-01 -3.82546663e-01 -1.27779830e+00 6.48865402e-01
5.59770703e-01 1.09019682e-01 1.14427261e-01 -3.52367938e-01
-3.05756241e-01 -1.19561918e-01 -1.06718540e+00 7.45945573e-01
1.05401754e+00 -5.52639186e-01 -3.44328076e-01 3.44253927e-01
3.70867729e-01 -2.84076929e-01 -7.52559066e-01 -1.96660142e-02
2.71858245e-01 -9.13034201e-01 1.23569953e+00 -9.74484980e-02
4.12246585e-01 -3.68930757e-01 -2.27859281e-02 -1.66938066e+00
-4.95373368e-01 -8.95828843e-01 -2.43662566e-01 1.16035688e+00
2.02662975e-01 -4.31332767e-01 4.76608515e-01 5.31652391e-01
-3.58257964e-02 -2.17919275e-01 -7.76517630e-01 -7.60291398e-01
3.56135845e-01 -5.92679620e-01 4.30961549e-02 6.58215106e-01
-7.34144151e-01 1.44099936e-01 -5.50486565e-01 2.77144939e-01
1.08842003e+00 -4.73373920e-01 4.99413192e-01 -9.98691916e-01
-5.49865723e-01 -5.00310421e-01 -1.26720205e-01 -1.36353981e+00
-8.33964199e-02 -5.11311293e-01 3.89424056e-01 -1.77905607e+00
-1.02621615e-01 -1.73476607e-01 -1.92753688e-01 1.46675155e-01
-2.73850173e-01 4.95120674e-01 1.79291219e-02 3.05443972e-01
-1.35604143e-01 5.68655312e-01 1.59178269e+00 -4.41998243e-01
5.48752490e-04 -8.65524169e-03 -5.26769459e-01 6.59300148e-01
4.60138589e-01 -4.34945881e-01 -5.10654390e-01 -8.47267747e-01
3.99515359e-03 -1.09210469e-01 5.89975893e-01 -8.60975564e-01
3.77105176e-01 7.63325840e-02 4.82948542e-01 -5.94458058e-02
5.77109814e-01 -7.76510477e-01 3.50141376e-01 4.20928270e-01
-3.54842007e-01 -1.65723905e-01 2.88476478e-02 8.95534813e-01
-2.66574591e-01 -2.60204464e-01 9.37176049e-01 -3.64676833e-01
-4.80594486e-01 1.97408199e-01 -5.52497327e-01 -1.80733860e-01
6.56438410e-01 -1.31844252e-01 -1.58151656e-01 -7.10186660e-01
-1.27053308e+00 -1.96377605e-01 6.94943130e-01 -3.79731245e-02
5.15079498e-01 -9.32238817e-01 -7.23711848e-01 2.40374237e-01
-5.26810169e-01 -3.63835059e-02 5.72727144e-01 9.49936032e-01
-3.89883339e-01 -4.29779291e-01 8.44663084e-02 -5.09858549e-01
-8.85961294e-01 8.45099747e-01 4.13418919e-01 -2.95956194e-01
-9.29994047e-01 8.37051988e-01 3.57469052e-01 -5.87798934e-03
1.44747660e-01 -5.09938776e-01 2.43630156e-01 -1.37929827e-01
6.77055776e-01 2.69637495e-01 1.00740634e-01 -4.89406824e-01
1.74713477e-01 6.69249833e-01 8.67451876e-02 -6.17534995e-01
1.58560503e+00 -4.87742007e-01 -4.02466387e-01 4.47391272e-01
1.07281196e+00 6.74582049e-02 -1.79703486e+00 -3.15752387e-01
-2.87660271e-01 -3.22374135e-01 4.64397132e-01 -8.54606032e-01
-1.19177306e+00 1.11952400e+00 7.43344486e-01 3.65459830e-01
1.34896028e+00 -1.98919713e-01 8.30908895e-01 1.11545749e-01
-5.41808717e-02 -9.38019276e-01 8.80573168e-02 3.67149323e-01
9.47027981e-01 -1.08401167e+00 -4.95344177e-02 -3.37835342e-01
-2.91442364e-01 1.02633488e+00 -1.16475306e-01 -4.99850482e-01
8.70658636e-01 5.10373235e-01 3.41374099e-01 -1.87755495e-01
-6.56771660e-01 -1.72158733e-01 8.36050659e-02 8.93346608e-01
1.18769690e-01 -3.28576237e-01 -9.16228816e-02 5.75419627e-02
2.76899070e-01 4.29222696e-02 5.28352737e-01 9.34636533e-01
-1.92181826e-01 -1.23738027e+00 -4.97908056e-01 1.62266776e-01
-4.23789948e-01 -2.80776739e-01 -6.59566596e-02 4.63687778e-01
1.17092445e-01 9.31459427e-01 -4.50734571e-02 1.17353283e-01
3.54233414e-01 -5.53711057e-02 6.80220068e-01 -4.04844552e-01
-5.73391080e-01 5.49040318e-01 -3.10501128e-01 -5.04041135e-01
-6.36802018e-01 -7.48951912e-01 -8.09519053e-01 -3.18321735e-01
3.27624679e-02 -2.47418255e-01 6.45096123e-01 9.40051734e-01
8.41405690e-02 6.77842557e-01 2.61595726e-01 -1.44862211e+00
-6.67847574e-01 -9.42814291e-01 -6.64806426e-01 8.88489962e-01
7.17032552e-01 -3.26918215e-01 -6.36605918e-01 6.04487002e-01] | [11.63367748260498, -2.323256254196167] |
3783eb96-efa9-4b32-ba2e-57dbe529e729 | geometry-aware-approaches-for-balancing | 2306.14872 | null | https://arxiv.org/abs/2306.14872v1 | https://arxiv.org/pdf/2306.14872v1.pdf | Geometry-Aware Approaches for Balancing Performance and Theoretical Guarantees in Linear Bandits | This paper is motivated by recent developments in the linear bandit literature, which have revealed a discrepancy between the promising empirical performance of algorithms such as Thompson sampling and Greedy, when compared to their pessimistic theoretical regret bounds. The challenge arises from the fact that while these algorithms may perform poorly in certain problem instances, they generally excel in typical instances. To address this, we propose a new data-driven technique that tracks the geometry of the uncertainty ellipsoid, enabling us to establish an instance-dependent frequentist regret bound for a broad class of algorithms, including Greedy, OFUL, and Thompson sampling. This result empowers us to identify and ``course-correct" instances in which the base algorithms perform poorly. The course-corrected algorithms achieve the minimax optimal regret of order $\tilde{\mathcal{O}}(d\sqrt{T})$, while retaining most of the desirable properties of the base algorithms. We present simulation results to validate our findings and compare the performance of our algorithms with the baselines. | ['Mohsen Bayati', 'Yuwei Luo'] | 2023-06-26 | null | null | null | null | ['thompson-sampling'] | ['methodology'] | [ 1.95260838e-01 1.74381316e-01 -6.51524663e-01 -3.85524601e-01
-1.10903955e+00 -9.09587026e-01 1.87123075e-01 1.78062335e-01
-2.63481885e-01 1.18121314e+00 1.32638782e-01 -6.82171166e-01
-9.47095871e-01 -5.86256623e-01 -9.24515069e-01 -7.97516584e-01
-1.07172221e-01 6.26017869e-01 -2.35664651e-01 2.11611494e-01
5.87449968e-01 3.32502425e-01 -1.27972209e+00 6.31216215e-03
1.15055084e+00 1.51114333e+00 -2.99365550e-01 2.14425161e-01
-8.27081427e-02 7.49846995e-01 -1.73449650e-01 -8.05401206e-01
4.93338287e-01 -4.61311936e-01 -8.78149927e-01 6.32945448e-02
1.23743847e-01 -2.73324192e-01 3.14693823e-02 1.17767751e+00
2.06934035e-01 3.05082381e-01 6.73734307e-01 -1.12038875e+00
-2.58685827e-01 7.90047288e-01 -9.76688027e-01 3.17799181e-01
2.13657752e-01 -3.39683652e-01 1.35453057e+00 -3.14443767e-01
4.67568487e-01 1.20001733e+00 6.89192355e-01 1.16941214e-01
-1.37443352e+00 -6.26354516e-01 4.90001827e-01 3.62163559e-02
-1.30827963e+00 -4.88987565e-01 5.96016228e-01 -2.40188181e-01
4.18981165e-01 6.86118841e-01 6.11925662e-01 7.38111019e-01
1.58948433e-02 1.17002571e+00 1.24710703e+00 -5.50176680e-01
5.82640946e-01 2.15257809e-01 1.81525901e-01 4.01612222e-01
7.52927959e-01 4.16398227e-01 -4.13557589e-01 -4.68609333e-01
6.47022545e-01 5.71383797e-02 -4.37556356e-01 -8.04185331e-01
-7.47390091e-01 1.03600228e+00 2.34800100e-01 -1.05041042e-01
-4.47879195e-01 3.96587849e-01 1.54404595e-01 2.09954634e-01
8.02613795e-01 2.96026707e-01 -4.43928719e-01 -3.65124911e-01
-1.17559576e+00 6.04464114e-01 8.47606242e-01 1.07612813e+00
4.91977662e-01 -3.03470522e-01 -3.26327264e-01 6.01084590e-01
2.32779473e-01 3.49512517e-01 -1.69335619e-01 -1.14875412e+00
7.90143251e-01 2.49864832e-01 9.39846277e-01 -9.08967376e-01
-2.57002771e-01 -8.01970005e-01 -5.34022987e-01 -2.71476656e-02
7.49252260e-01 -2.46041298e-01 -5.09932876e-01 1.55573511e+00
3.68086755e-01 -2.46968150e-01 -2.36012414e-01 7.88559675e-01
-2.46220872e-01 4.05349880e-01 -3.53138268e-01 -8.47347915e-01
7.59081602e-01 -7.39332318e-01 -5.73698103e-01 -2.12795883e-01
6.14706755e-01 -6.29000425e-01 8.49780083e-01 5.08159220e-01
-1.38029253e+00 2.29280248e-01 -8.37922812e-01 5.28289557e-01
2.73411572e-01 -3.40861291e-01 9.82035398e-01 1.09270668e+00
-6.73355460e-01 7.36635089e-01 -7.48701930e-01 -1.53148770e-01
5.42784095e-01 1.00911722e-01 3.12871069e-01 -1.24240920e-01
-5.88984013e-01 7.39607751e-01 3.49562943e-01 2.01734439e-01
-6.19298100e-01 -7.61879206e-01 -3.95447671e-01 1.52326375e-01
8.27274144e-01 -4.79484081e-01 1.63388586e+00 -1.00494719e+00
-1.20989156e+00 4.63682353e-01 -2.31139868e-01 -7.03169584e-01
1.11743224e+00 -3.01574290e-01 2.28510097e-01 -5.99255413e-02
1.46435425e-01 -3.27097625e-02 3.54113877e-01 -1.19657958e+00
-1.01261854e+00 -4.53933358e-01 3.40114743e-01 2.90070951e-01
-1.30891791e-02 -8.95222425e-02 -1.66352242e-01 -5.20944178e-01
2.46034339e-01 -9.22975063e-01 -5.58279097e-01 -3.09102058e-01
-5.44303954e-01 -7.27988109e-02 -1.48056060e-01 4.54833033e-03
1.41123497e+00 -1.93824017e+00 -1.62114888e-01 5.98806679e-01
-2.29187340e-01 -3.46971065e-01 3.76267165e-01 5.24937034e-01
5.45308776e-02 2.53753632e-01 -6.41758591e-02 2.56433506e-02
2.54953921e-01 2.33077630e-01 -4.56304699e-01 8.32434714e-01
-3.90930146e-01 6.35833323e-01 -9.16194558e-01 -3.11272502e-01
-9.80988797e-03 -4.75713044e-01 -8.09635937e-01 -7.85626993e-02
-3.82350802e-01 1.34925306e-01 -7.43183851e-01 5.70868611e-01
6.24739528e-01 -4.58458990e-01 4.35688585e-01 2.53315061e-01
-2.94132441e-01 2.88090289e-01 -1.48393917e+00 1.27060568e+00
-3.04090958e-02 8.13853741e-02 2.23724514e-01 -1.24094260e+00
4.55333591e-01 1.11697160e-01 5.64602196e-01 -5.05107045e-01
2.81456649e-01 1.92083597e-01 -3.19463849e-01 -3.09566051e-01
2.66946763e-01 -3.87245923e-01 -5.48897088e-02 8.33916128e-01
-5.69322348e-01 4.83461805e-02 1.89991429e-01 -1.10261820e-01
8.67251635e-01 3.77703123e-02 5.04002810e-01 -6.61510170e-01
-6.60412535e-02 2.11045131e-01 6.96806431e-01 1.39881039e+00
-2.09965676e-01 1.71680003e-01 7.82885969e-01 -4.74053264e-01
-9.25985277e-01 -1.07167041e+00 -3.71888310e-01 1.17475307e+00
3.88315767e-01 7.13667423e-02 -6.38774216e-01 -5.08654714e-01
2.96957582e-01 1.02008116e+00 -8.64920616e-01 5.85178696e-02
-1.08772822e-01 -1.16229403e+00 1.22369900e-01 4.62309152e-01
1.31649092e-01 -2.01814577e-01 -6.16488159e-01 2.27878854e-01
-2.81657964e-01 -7.20019579e-01 -2.81995445e-01 3.31887662e-01
-1.00189507e+00 -1.20927632e+00 -4.89836365e-01 -1.09621935e-01
7.42897987e-01 3.36281776e-01 9.91087377e-01 -4.06877100e-01
8.73765722e-02 2.66921043e-01 -4.98879373e-01 -8.02201688e-01
-1.38360914e-02 6.34822343e-03 8.51130337e-02 4.99998359e-03
2.41117552e-01 -4.38219696e-01 -8.00741792e-01 3.11430991e-01
-6.43770814e-01 -1.05785027e-01 5.41006923e-01 7.98440933e-01
7.33401597e-01 1.78616643e-01 5.20309210e-01 -1.07492471e+00
6.40826583e-01 -5.72164178e-01 -1.04629242e+00 4.62662488e-01
-1.03142858e+00 1.78080693e-01 3.02018523e-01 -1.23525433e-01
-1.04878914e+00 -1.46479651e-01 2.71771610e-01 -2.17962742e-01
3.06034029e-01 7.11833537e-01 2.80975103e-01 5.55338897e-03
6.51745439e-01 1.36120990e-01 -2.00261757e-01 -5.13409793e-01
3.17701578e-01 6.05536461e-01 2.50347704e-01 -9.59890544e-01
5.33383608e-01 6.13476276e-01 1.06769493e-02 -2.23705307e-01
-1.62653506e+00 -2.75078058e-01 -9.89709049e-02 -1.02752686e-01
1.97171956e-01 -6.37701333e-01 -9.93108034e-01 -2.13763982e-01
-6.88533425e-01 -6.81091249e-02 -4.44363505e-01 4.85124350e-01
-1.11404335e+00 2.68356651e-01 -1.62816256e-01 -1.63202941e+00
-1.45547226e-01 -9.13231075e-01 6.36253476e-01 1.73186436e-01
-1.91171587e-01 -8.53132665e-01 6.02651425e-02 5.03710389e-01
7.80778751e-02 4.57457602e-01 8.94721687e-01 -6.46969199e-01
-7.91278660e-01 -2.68742532e-01 -3.47792864e-01 5.79300337e-02
-3.78032736e-02 -2.68076360e-01 -5.13052762e-01 -5.50905704e-01
5.82335740e-02 -3.49651933e-01 7.46042609e-01 8.43667626e-01
1.25022435e+00 -5.94302654e-01 -5.28935134e-01 5.88455439e-01
1.51769018e+00 4.13374633e-01 2.84469396e-01 5.91143012e-01
-1.48140594e-01 3.67728531e-01 9.39012110e-01 9.22448397e-01
2.99972326e-01 5.40578365e-01 5.93187153e-01 3.66042614e-01
8.01457942e-01 -1.31019160e-01 -9.03604925e-02 1.09536089e-01
-1.58828661e-01 -2.68083662e-01 -7.40879118e-01 8.00923824e-01
-2.19362092e+00 -1.06641591e+00 6.23753853e-02 2.54057550e+00
9.39179122e-01 3.22365135e-01 3.66106808e-01 2.63194531e-01
6.85168147e-01 9.75744054e-02 -5.54299891e-01 -6.84878647e-01
1.83086693e-01 3.36978175e-02 9.25382793e-01 4.25695390e-01
-9.73614037e-01 5.47115445e-01 7.17950201e+00 9.80948925e-01
-6.19376481e-01 -1.03809536e-01 7.51603484e-01 -8.17362726e-01
-4.39653605e-01 1.82581738e-01 -7.84111798e-01 5.21034777e-01
9.31976616e-01 -6.94605410e-01 7.96743751e-01 1.02725518e+00
4.06738132e-01 -4.66814071e-01 -1.29252172e+00 8.58107507e-01
-2.20410779e-01 -1.45112765e+00 -3.16956997e-01 1.10533550e-01
1.09982598e+00 -1.57182053e-01 1.64574713e-01 1.09072253e-01
8.49273205e-01 -8.93433630e-01 9.91950929e-01 3.40926766e-01
5.71043193e-01 -1.10377884e+00 5.54458201e-01 5.61671853e-01
-5.92589676e-01 -4.22875911e-01 -4.40499306e-01 -3.45881820e-01
-6.78715855e-02 9.15648341e-01 -8.01674008e-01 7.42728055e-01
8.09878349e-01 1.84478119e-01 2.54909754e-01 1.38103795e+00
3.46044265e-02 5.92720866e-01 -5.87256908e-01 -3.39008182e-01
5.08929312e-01 -3.79130483e-01 4.78642255e-01 8.46903622e-01
2.79867023e-01 1.88097700e-01 7.47365654e-02 7.86716640e-01
1.38154700e-02 2.26032868e-01 -3.11856687e-01 -5.83457239e-02
6.77601993e-01 6.46086276e-01 -7.28467643e-01 -6.52376562e-02
-2.16806471e-01 4.74487662e-01 4.11283404e-01 3.65938067e-01
-9.36931431e-01 -2.36411586e-01 6.52978718e-01 6.95692450e-02
6.51231527e-01 2.67680939e-02 -7.01513112e-01 -9.55271542e-01
3.66595834e-01 -5.36821723e-01 6.07050061e-01 -2.91435361e-01
-1.24802053e+00 -1.04905285e-01 2.39102855e-01 -1.06628966e+00
-3.09079140e-01 -3.24547410e-01 -1.47180498e-01 5.51856101e-01
-1.42191029e+00 -5.58945119e-01 2.12628707e-01 4.40094352e-01
1.82608232e-01 3.76831084e-01 4.65057552e-01 5.27062900e-02
-4.94363427e-01 7.45412529e-01 1.13123429e+00 -2.50729620e-01
2.72109509e-01 -1.16285753e+00 -1.44645214e-01 6.82878256e-01
-7.69657195e-02 5.16280293e-01 1.06486678e+00 -4.63000029e-01
-1.39234400e+00 -9.10028577e-01 5.67870855e-01 -4.22389209e-01
8.93527627e-01 -5.44145703e-02 -1.32695675e-01 1.03740251e+00
-3.13780606e-01 -1.49229392e-01 7.28612363e-01 6.11181259e-01
-2.74850428e-01 -3.55016798e-01 -1.25662470e+00 6.30649984e-01
1.14634323e+00 -1.47151025e-02 -4.45622504e-01 5.96873224e-01
2.66630471e-01 -5.98328412e-01 -7.73869038e-01 4.41906929e-01
7.70539761e-01 -1.28761160e+00 7.28585243e-01 -8.47226799e-01
3.35883707e-01 2.38688722e-01 -3.79423738e-01 -1.25901639e+00
-2.12251261e-01 -8.72631788e-01 -1.33412004e-01 9.86993372e-01
4.66559678e-01 -6.94594264e-01 1.02914345e+00 8.96707296e-01
1.92366332e-01 -9.85361099e-01 -1.35729897e+00 -1.02821481e+00
3.54434103e-01 -5.79190850e-01 5.78086913e-01 8.25205266e-01
2.26502866e-01 -2.65420824e-01 -5.19775689e-01 -1.16472961e-02
9.23042536e-01 7.68920302e-01 5.47291815e-01 -1.20873129e+00
-3.72932315e-01 -4.78614211e-01 2.46954318e-02 -1.23886037e+00
-2.74677217e-01 -5.81054091e-01 -1.67115983e-02 -1.32584214e+00
5.19488990e-01 -7.69564927e-01 -4.89339471e-01 1.87274456e-01
-1.25818491e-01 -1.18785888e-01 3.70280296e-02 1.02631941e-01
-8.51191759e-01 3.44329983e-01 1.13702774e+00 6.91242069e-02
-1.59696296e-01 4.07071024e-01 -1.36609185e+00 7.02507317e-01
5.13849318e-01 -5.92731237e-01 -4.00913537e-01 -4.36287999e-01
7.07549036e-01 5.43992996e-01 6.29014671e-02 -5.62672496e-01
8.07071030e-02 -7.94958413e-01 2.23740235e-01 -6.93344355e-01
4.64151725e-02 -8.91073108e-01 1.67172089e-01 3.74618709e-01
-5.90339839e-01 -2.46543273e-01 -4.66100918e-03 9.11258876e-01
2.15186566e-01 -4.08025175e-01 7.74421930e-01 -2.22221345e-01
1.44534126e-01 1.63550779e-01 -2.19444200e-01 3.65800381e-01
1.18903410e+00 -3.51447672e-01 -3.50611985e-01 -6.49103284e-01
-4.89172846e-01 3.69855970e-01 3.34199220e-01 -1.24889649e-01
-3.36129032e-02 -1.11204445e+00 -5.58883250e-01 -2.42317840e-01
-5.51154390e-02 9.25068408e-02 1.94866598e-01 1.17379105e+00
-3.09561968e-01 8.36143613e-01 1.89342558e-01 -3.70980740e-01
-7.53386796e-01 6.96238816e-01 2.78457344e-01 -4.15416420e-01
-2.34316781e-01 9.83598232e-01 1.04692638e-01 -2.66018454e-02
4.11550820e-01 -3.53944808e-01 4.23113376e-01 8.04062933e-02
3.94412339e-01 7.28451967e-01 3.72544378e-02 6.59129396e-02
-2.70297587e-01 1.06814034e-01 -3.25691909e-01 -6.62076399e-02
1.42026544e+00 -3.51398677e-01 6.73682317e-02 3.43129188e-01
7.49617219e-01 8.34127888e-02 -1.27131784e+00 -2.45216712e-01
3.00286412e-01 -8.95477355e-01 1.33558512e-01 -8.99782538e-01
-7.32571661e-01 3.99771333e-01 2.76540667e-01 6.77284360e-01
9.80198205e-01 -1.24896809e-01 4.53507036e-01 4.11221564e-01
7.81745613e-01 -1.21501338e+00 -4.90158141e-01 2.22744331e-01
5.86135924e-01 -1.02489674e+00 3.30983698e-01 -3.21697265e-01
-4.14637297e-01 6.61952317e-01 9.99739543e-02 -1.63709745e-01
4.02741849e-01 1.17553167e-01 -2.71675467e-01 8.04516599e-02
-8.37539256e-01 -1.30784601e-01 -1.11503236e-01 1.92916155e-01
1.20873109e-01 1.28212228e-01 -7.37146795e-01 7.21549153e-01
-3.28455597e-01 2.20573947e-01 5.02947927e-01 1.03897572e+00
-5.51003695e-01 -8.36937129e-01 -5.90455472e-01 7.21615493e-01
-9.01434898e-01 5.00866883e-02 -3.11754465e-01 8.33037734e-01
-4.20974553e-01 1.07908630e+00 7.13951215e-02 5.48081882e-02
2.64063656e-01 -2.99493410e-02 7.16319501e-01 -1.38951555e-01
-2.44485527e-01 2.62970924e-01 3.64839435e-01 -5.56312442e-01
-4.41657543e-01 -1.00557685e+00 -6.58550560e-01 -6.50169849e-01
-4.77517366e-01 6.17121041e-01 4.72609818e-01 1.09359753e+00
3.65026556e-02 8.80674794e-02 8.37765276e-01 -5.33400297e-01
-1.21829212e+00 -5.51622450e-01 -9.26513731e-01 1.77698120e-01
3.44427198e-01 -8.21945548e-01 -4.85173464e-01 -5.18713772e-01] | [4.538644313812256, 3.3029773235321045] |
c358d6ec-d032-44f0-a8be-2c7b952ba2cc | 190412619 | 1904.12619 | null | https://arxiv.org/abs/1904.12619v2 | https://arxiv.org/pdf/1904.12619v2.pdf | Multiple receptive fields and small-object-focusing weakly-supervised segmentation network for fast object detection | Object detection plays an important role in various visual applications. However, the precision and speed of detector are usually contradictory. One main reason for fast detectors' precision reduction is that small objects are hard to be detected. To address this problem, we propose a multiple receptive field and small-object-focusing weakly-supervised segmentation network (MRFSWSnet) to achieve fast object detection. In MRFSWSnet, multiple receptive fields block (MRF) is used to pay attention to the object and its adjacent background's different spatial location with different weights to enhance the feature's discriminability. In addition, in order to improve the accuracy of small object detection, a small-object-focusing weakly-supervised segmentation module which only focuses on small object instead of all objects is integrated into the detection network for auxiliary training to improve the precision of small object detection. Extensive experiments show the effectiveness of our method on both PASCAL VOC and MS COCO detection datasets. In particular, with a lower resolution version of 300x300, MRFSWSnet achieves 80.9% mAP on VOC2007 test with an inference speed of 15 milliseconds per frame, which is the state-of-the-art detector among real-time detectors. | ['Haifeng Shen', 'Yuan Zhao', 'Yingjie Yin', 'Xingang Wang', 'Siyang Sun', 'De Xu'] | 2019-04-19 | null | null | null | null | ['small-object-detection'] | ['computer-vision'] | [ 6.39306474e-03 -3.37934107e-01 -1.92629680e-01 -4.17870939e-01
-2.93372840e-01 -2.69286543e-01 1.97798371e-01 7.87826777e-02
-9.94638443e-01 1.92177534e-01 -3.44783098e-01 -3.48917283e-02
4.43549454e-01 -7.81753659e-01 -7.61570752e-01 -8.44494522e-01
4.25318152e-01 9.08473581e-02 1.48494256e+00 -1.13803089e-01
2.23641902e-01 3.98982048e-01 -1.44052851e+00 4.20607179e-01
7.61239171e-01 1.11253262e+00 1.01451004e+00 4.94638532e-01
-4.68232622e-03 5.77554822e-01 -7.43180633e-01 9.20015760e-03
2.72807479e-01 -2.41500214e-02 -4.11003321e-01 -6.61257580e-02
4.17228043e-01 -5.94734669e-01 -2.48378977e-01 1.37162101e+00
5.09148717e-01 -8.23799521e-02 5.47326148e-01 -1.02214432e+00
-3.07639509e-01 5.21462440e-01 -1.05027223e+00 9.85242844e-01
-1.80814788e-01 4.03106451e-01 4.92779166e-01 -8.20386827e-01
2.12657273e-01 1.27278793e+00 2.18099549e-01 6.11651063e-01
-1.05513275e+00 -9.93846953e-01 2.62880772e-01 3.89829874e-01
-1.63707590e+00 -2.28287533e-01 2.58725047e-01 -2.62837231e-01
8.44789267e-01 2.18890831e-01 5.20731151e-01 4.98328954e-01
9.74567980e-02 1.13143551e+00 8.03767920e-01 -9.29322019e-02
-2.27340478e-02 3.51828188e-01 1.58804074e-01 5.23903847e-01
4.29958850e-01 1.38415275e-02 -1.06539458e-01 3.90864819e-01
9.39543843e-01 2.96785772e-01 -1.79471508e-01 7.30904117e-02
-1.22785521e+00 6.32125676e-01 8.82053614e-01 3.45854312e-01
-1.51651308e-01 2.03670248e-01 4.15772527e-01 -2.20971718e-01
1.30526200e-01 -9.47599933e-02 -3.32241625e-01 1.59101412e-01
-7.13952243e-01 2.76329871e-02 2.38359839e-01 9.65611696e-01
5.11454225e-01 -6.69762939e-02 -5.28216004e-01 8.14886451e-01
3.60711664e-01 7.80752778e-01 4.76146847e-01 -5.77192843e-01
6.13192499e-01 8.63300622e-01 4.09777872e-02 -8.92831326e-01
-2.85310119e-01 -5.99309206e-01 -9.10718739e-01 2.39306256e-01
5.65707147e-01 1.89541698e-01 -1.11579180e+00 1.30171609e+00
4.27868456e-01 2.02144668e-01 -1.59035712e-01 1.40297687e+00
1.07471216e+00 9.48187649e-01 3.61852556e-01 -1.89835727e-01
1.74150479e+00 -1.01416469e+00 -2.89148420e-01 -4.63025451e-01
4.71426785e-01 -7.83747077e-01 9.74572837e-01 1.23689346e-01
-7.93449640e-01 -9.73304927e-01 -1.02499580e+00 -1.54788308e-02
-1.81004137e-01 6.18396878e-01 4.32403505e-01 4.42657709e-01
-4.30873811e-01 1.00235760e-01 -8.59760761e-01 -3.05061042e-01
8.10807049e-01 3.89472932e-01 6.19801180e-03 -1.04387186e-01
-9.17895436e-01 7.09007084e-01 8.92071843e-01 1.62765428e-01
-8.63808990e-01 -3.32520545e-01 -4.81224447e-01 3.07197124e-01
6.62546396e-01 -2.94249564e-01 1.08400750e+00 -9.32078004e-01
-9.67612982e-01 7.59803593e-01 -5.66776544e-02 -4.57786709e-01
5.01013875e-01 9.89125296e-02 -4.41873431e-01 1.96373418e-01
3.19019169e-01 9.26466346e-01 9.05945063e-01 -8.26433659e-01
-1.25351465e+00 -3.31412435e-01 -2.15586443e-02 2.82208413e-01
-2.23064199e-01 2.19584897e-01 -8.49225104e-01 -4.20966446e-01
2.63061970e-01 -6.17212057e-01 -1.99744135e-01 1.02331825e-01
-3.31226826e-01 -7.44568467e-01 1.06365430e+00 -2.42048949e-01
1.30397427e+00 -2.44509101e+00 -3.25732023e-01 -1.33226201e-01
1.71084985e-01 7.84234226e-01 -1.86340883e-01 -4.16860282e-01
2.88552821e-01 -1.67050257e-01 -2.61554378e-03 3.51924822e-02
-4.42607760e-01 1.18177071e-01 -1.56702250e-01 4.87561345e-01
3.14667404e-01 7.77622163e-01 -8.32799435e-01 -9.58219469e-01
4.13409799e-01 2.29695275e-01 -2.85818458e-01 1.16417758e-01
2.36746157e-03 9.96027812e-02 -6.74508810e-01 7.72682965e-01
9.82150316e-01 -1.68108806e-01 -3.72839928e-01 -4.72109795e-01
-3.78679276e-01 -1.37065232e-01 -1.53836644e+00 1.28625298e+00
-2.59256493e-02 4.93681282e-01 1.08686261e-01 -8.84602129e-01
8.14212382e-01 -1.24012187e-01 -9.98560712e-02 -1.01236689e+00
2.58619457e-01 2.97809392e-01 3.35052133e-01 -5.83274364e-01
1.87187240e-01 1.65891781e-01 3.13270420e-01 2.64895540e-02
-1.03454448e-01 1.22785352e-01 5.11294365e-01 2.37378195e-01
8.03542554e-01 -1.32115260e-01 2.95836002e-01 -3.52083862e-01
6.60489619e-01 -2.54438389e-02 8.71450424e-01 8.27501059e-01
-5.08105516e-01 5.94261527e-01 1.58607602e-01 -3.50373179e-01
-7.56390989e-01 -1.04263377e+00 -4.25755143e-01 1.28493261e+00
7.58316100e-01 -2.50558704e-01 -7.93076873e-01 -7.50863671e-01
4.73842770e-03 3.47184092e-01 -4.35010046e-01 -1.15717433e-01
-7.09835768e-01 -8.90154719e-01 4.56056952e-01 9.82089162e-01
8.75866175e-01 -1.31639469e+00 -1.14546502e+00 1.76005363e-01
-1.92573518e-02 -1.11077118e+00 -7.34822989e-01 7.26478547e-02
-7.46402085e-01 -1.21896207e+00 -7.62772501e-01 -1.08417785e+00
8.10490787e-01 7.32825518e-01 6.29704952e-01 1.91514343e-01
-5.60895205e-01 -3.27853590e-01 -1.80012748e-01 -4.82619882e-01
-8.70395005e-02 -8.10495540e-02 -1.97428092e-02 -7.14285895e-02
5.47167778e-01 1.22127593e-01 -1.00240910e+00 7.26388931e-01
-1.06861329e+00 1.17750898e-01 9.94332075e-01 6.82618737e-01
6.82106674e-01 5.00536822e-02 3.31129700e-01 -6.03886724e-01
1.58361688e-01 -1.13129020e-01 -9.56790566e-01 2.12258607e-01
-2.45521322e-01 -8.64231065e-02 6.97820961e-01 -8.31713855e-01
-1.00036561e+00 3.47635865e-01 -7.36374110e-02 -2.60306150e-01
-1.19060963e-01 -1.66837797e-01 -9.83791053e-02 -8.29407647e-02
6.80233479e-01 5.92393637e-01 -2.17911974e-01 -3.23092043e-01
-1.71424877e-02 9.55824614e-01 5.72409034e-01 -1.05310325e-02
6.11579359e-01 4.22237515e-01 -2.13379025e-01 -8.02977800e-01
-7.47633755e-01 -8.78603220e-01 -4.08196956e-01 -2.11271778e-01
1.03162241e+00 -1.09256017e+00 -9.71833527e-01 5.46454906e-01
-1.18852782e+00 -1.22429512e-01 -1.23963431e-02 6.55663013e-01
7.05652386e-02 1.76597446e-01 -6.76162004e-01 -7.41019011e-01
-4.79707360e-01 -1.19375515e+00 1.11791086e+00 8.05039465e-01
4.62373853e-01 -2.81697214e-01 -6.20564520e-01 1.49121910e-01
4.02707070e-01 -2.49550149e-01 2.69415677e-01 -4.46103275e-01
-7.83147156e-01 -3.13856363e-01 -9.94711637e-01 4.97652560e-01
-1.41462669e-01 -4.98422831e-02 -8.61579835e-01 -3.14434618e-01
-7.91647062e-02 -2.58018285e-01 1.31549263e+00 5.55552065e-01
1.26708686e+00 5.48673719e-02 -7.31691062e-01 5.41768074e-01
1.48331654e+00 3.53187114e-01 5.29128551e-01 3.00943702e-01
7.09811926e-01 1.80637971e-01 9.58081841e-01 2.88112998e-01
1.26497358e-01 5.72647214e-01 3.68539959e-01 -1.55557409e-01
-2.37992480e-01 -5.53310197e-03 1.44358814e-01 3.13363999e-01
-3.23249847e-01 1.78681605e-03 -6.49139166e-01 3.29372138e-01
-1.77825201e+00 -9.66517806e-01 -3.02276462e-01 2.23826146e+00
6.30879462e-01 7.70239294e-01 2.28876397e-01 -9.61745977e-02
1.00932443e+00 9.39766243e-02 -7.15145290e-01 1.74029425e-01
-8.48445445e-02 -1.68790475e-01 8.68109405e-01 -4.08063531e-02
-1.20631611e+00 9.96388018e-01 4.88205576e+00 1.39165127e+00
-1.19194925e+00 2.21476212e-01 6.21123850e-01 -3.50582935e-02
4.55696195e-01 -1.90230370e-01 -1.44369054e+00 7.65369177e-01
2.17254981e-01 1.80793077e-01 1.10350214e-01 1.17978549e+00
1.80479765e-01 -4.24759060e-01 -7.76147187e-01 1.08671403e+00
-1.33690819e-01 -1.09375417e+00 -1.30237177e-01 -2.23573744e-01
3.82955551e-01 1.33323237e-01 -2.25498900e-01 4.45306361e-01
-3.16253185e-01 -7.16101468e-01 7.92321861e-01 3.13150175e-02
7.74307907e-01 -7.47720540e-01 8.72623205e-01 7.45972812e-01
-1.60012484e+00 -2.85058498e-01 -9.57216382e-01 1.26973405e-01
-2.23982614e-02 5.04614174e-01 -6.46723747e-01 -7.71696717e-02
1.09351301e+00 4.74393845e-01 -7.23338068e-01 1.25628078e+00
-2.56780535e-01 6.46078587e-01 -6.56401992e-01 -4.43995148e-01
3.32074910e-01 1.27053529e-01 4.45040226e-01 1.47609365e+00
-6.27424046e-02 4.05433059e-01 5.09457469e-01 8.87177646e-01
-9.26912203e-03 -1.39564294e-02 8.13804846e-03 2.27043182e-01
5.17385662e-01 1.51185632e+00 -1.20578086e+00 -6.40164018e-01
-3.94127637e-01 9.32509840e-01 1.81045130e-01 1.81858167e-01
-1.04069042e+00 -7.26155818e-01 1.48986787e-01 3.42149377e-01
5.92719138e-01 -3.91888097e-02 5.28186224e-02 -1.01612103e+00
2.67808102e-02 -4.67468828e-01 4.27293897e-01 -5.80886066e-01
-8.21967423e-01 5.16623020e-01 -1.25247583e-01 -1.22378445e+00
3.59217703e-01 -6.30998194e-01 -7.51216173e-01 7.28716731e-01
-1.47616601e+00 -1.10327232e+00 -6.15718484e-01 5.95183492e-01
8.40352476e-01 -2.89762896e-02 1.86187074e-01 3.65557790e-01
-8.60818803e-01 5.73618472e-01 -9.23257507e-03 4.07560110e-01
6.02780819e-01 -1.00786555e+00 3.03423226e-01 9.64937270e-01
-7.41054714e-02 3.44386011e-01 3.67981762e-01 -5.44972718e-01
-1.13620543e+00 -1.28524315e+00 3.14503014e-01 -9.52025130e-02
2.26375505e-01 -4.04656470e-01 -1.08216977e+00 2.63565063e-01
-1.75229728e-01 6.30263507e-01 -5.28664440e-02 -4.07075793e-01
-2.46271610e-01 -4.76805627e-01 -1.03790605e+00 4.37031269e-01
9.42703485e-01 -8.02897438e-02 -6.08849347e-01 4.41488564e-01
6.81476176e-01 -4.71961349e-01 -2.04991400e-01 5.47445774e-01
3.62800121e-01 -9.91511226e-01 9.73302424e-01 -8.28449130e-02
4.63142246e-02 -7.67768085e-01 7.12996647e-02 -7.40573764e-01
-5.15115380e-01 4.05566059e-02 -1.76695548e-02 1.06605422e+00
3.43650789e-03 -6.64583862e-01 8.35665643e-01 -2.34645256e-03
-7.90458247e-02 -6.49530053e-01 -8.48264813e-01 -7.93844998e-01
-4.59816188e-01 -2.95704454e-01 2.69760549e-01 3.26353252e-01
-4.61712837e-01 5.23111343e-01 1.18763968e-01 2.83957630e-01
6.59307122e-01 2.88074166e-01 5.90807140e-01 -1.01344430e+00
-2.40944356e-01 -5.85491180e-01 -6.10684752e-01 -1.53305912e+00
-4.59385157e-01 -4.96164113e-01 3.44692528e-01 -1.40151715e+00
6.68201864e-01 -4.04902011e-01 -5.25210500e-01 3.46439660e-01
-4.34482723e-01 5.66876471e-01 2.70807713e-01 2.90910095e-01
-1.08093739e+00 2.23283783e-01 1.47205257e+00 -2.58181065e-01
-2.71668881e-01 2.56524801e-01 -2.56080657e-01 8.34786057e-01
5.91890514e-01 -5.86865664e-01 -1.43482670e-01 -2.88702965e-01
-2.81505167e-01 -2.16530740e-01 4.26069289e-01 -1.25134945e+00
6.34589016e-01 -1.19529903e-01 8.34868014e-01 -9.64079022e-01
7.23533556e-02 -4.51476991e-01 -4.55436021e-01 1.07027709e+00
-4.37871851e-02 -3.40836197e-01 3.42804134e-01 6.04191303e-01
-1.15156263e-01 -3.45620751e-01 1.20234036e+00 -1.41986102e-01
-1.28933263e+00 3.27053994e-01 -2.05149919e-01 8.34041461e-03
1.30261946e+00 -3.75431955e-01 -3.42444986e-01 3.84020478e-01
-2.55071193e-01 5.36954641e-01 1.13046363e-01 4.60994244e-01
8.98779094e-01 -9.81173456e-01 -7.80305684e-01 3.69513243e-01
3.56232852e-01 4.45876032e-01 3.86873692e-01 9.23391700e-01
-5.63326418e-01 3.70498091e-01 -1.44666284e-01 -9.24812555e-01
-1.45720649e+00 5.99635303e-01 3.55411500e-01 2.64340013e-01
-6.36265397e-01 1.13398194e+00 6.87321186e-01 1.32258579e-01
3.24084103e-01 -5.72542489e-01 -3.76376569e-01 -2.05492690e-01
9.36309159e-01 3.97700399e-01 -3.29121649e-01 -4.74002242e-01
-6.55818760e-01 5.59820950e-01 -3.85629803e-01 4.73023623e-01
8.21610391e-01 2.58041155e-02 8.57355967e-02 1.13909073e-01
8.93980980e-01 -1.76609531e-01 -1.46734786e+00 -2.45228469e-01
-3.47855657e-01 -5.67831099e-01 1.95397750e-01 -5.97147942e-01
-1.16010964e+00 9.19065297e-01 1.04328883e+00 6.34987727e-02
1.10754740e+00 2.31770694e-01 7.52527416e-01 4.15214062e-01
1.72566742e-01 -1.21347868e+00 1.68395102e-01 2.48817250e-01
5.33750176e-01 -1.73762298e+00 1.17287096e-02 -5.50128639e-01
-6.32700145e-01 1.08177090e+00 1.19797266e+00 -2.70167530e-01
3.06771070e-01 2.99749583e-01 -1.03176348e-01 8.00753012e-02
-1.97049752e-01 -4.65723604e-01 4.11434144e-01 3.83326530e-01
1.02450497e-01 6.45111427e-02 -4.07892436e-01 5.30851007e-01
5.27773559e-01 -1.82480171e-01 2.00803608e-01 7.99315035e-01
-1.08055508e+00 -4.89677697e-01 -3.94026786e-01 6.61790788e-01
-3.81514668e-01 -1.02410339e-01 1.38711333e-01 6.97125137e-01
5.40049732e-01 8.87870193e-01 3.58503670e-01 -2.73937434e-01
3.58406276e-01 -7.26578355e-01 2.52656251e-01 -6.58698261e-01
-3.64818513e-01 3.59752476e-01 -4.98220414e-01 -6.36344850e-01
-2.98210025e-01 -3.87089372e-01 -1.75138712e+00 -1.55591354e-01
-7.67094612e-01 2.88166255e-02 5.80186129e-01 9.13231432e-01
7.06861988e-02 5.72433889e-01 3.96412820e-01 -6.45204902e-01
-6.97591782e-01 -1.11075425e+00 -7.01195300e-01 3.90482396e-01
1.32223696e-01 -5.51079214e-01 -2.26444975e-02 -6.11426309e-02] | [8.731654167175293, -0.5827187299728394] |
32812ca2-544a-4ef5-a75e-948fac5a6ca6 | analyse-de-la-r-egulation-de-la-longueur-dans | null | null | https://aclanthology.org/2020.jeptalnrecital-recital.5 | https://aclanthology.org/2020.jeptalnrecital-recital.5.pdf | Analyse de la r\'egulation de la longueur dans un syst\`eme neuronal de compression de phrase : une \'etude du mod\`ele LenInit (Investigating Length Regulation in a Sentence Compression Neural System : a Study on the LenInit Model) | La simplification de phrase vise {\`a} r{\'e}duire la complexit{\'e} d{'}une phrase tout en retenant son sens initial et sa grammaticalit{\'e}. En pratique, il est souvent attendu que la phrase produite soit plus courte que la phrase d{'}origine, et les mod{\`e}les qui int{\`e}grent un contr{\^o}le explicite de la longueur de sortie rev{\^e}tent un int{\'e}r{\^e}t particulier. Dans la continuit{\'e} de la litt{\'e}rature d{\'e}di{\'e}e {\`a} la compr{\'e}hension du comportement des syst{\`e}mes neuronaux, nous examinons dans cet article les m{\'e}canismes de r{\'e}gulation de longueur d{'}un encodeur-d{\'e}codeur RNN appliqu{\'e} {\`a} la compression de phrase, en {\'e}tudiant sp{\'e}cifiquement le cas du mod{\`e}le LenInit. Notre analyse met en {\'e}vidence la coexistence de deux influences distinctes au cours du d{\'e}codage : celle du contr{\^o}le explicite de la longueur, et celle du mod{\`e}le de langue du d{\'e}codeur. | ['Fran{\\c{c}}ois Buet'] | 2020-06-01 | null | null | null | jeptalnrecital-2020-6 | ['sentence-compression'] | ['natural-language-processing'] | [ 3.14862758e-01 3.23681444e-01 3.47662359e-01 -9.01326463e-02
-1.83984861e-01 -1.21254098e+00 4.20403898e-01 2.74288625e-01
-7.04841673e-01 8.49388123e-01 -7.83198401e-02 -4.65290278e-01
-3.74097556e-01 -1.06371093e+00 -7.66641259e-01 -7.45373368e-01
-2.68403143e-01 3.01485658e-01 -1.17965117e-02 -7.67837226e-01
1.86805606e-01 4.46751803e-01 -1.52734590e+00 2.83993363e-01
5.04226029e-01 1.21273470e+00 3.50953728e-01 6.58949971e-01
-1.14961788e-02 6.27046704e-01 -6.17505789e-01 -5.82816124e-01
6.89944804e-01 -7.96253622e-01 -5.90589285e-01 -3.06311905e-01
4.37057093e-02 5.31367123e-01 -3.24553490e-01 1.79436672e+00
7.30015412e-02 1.87946767e-01 1.54489040e+00 -7.73686647e-01
-7.85758257e-01 6.83664262e-01 -3.73282105e-01 9.72081125e-01
4.05725181e-01 7.85766467e-02 1.29980361e+00 -7.10603237e-01
6.15654171e-01 4.93703246e-01 8.65552425e-01 4.66422468e-01
-9.20205712e-01 -6.70763850e-01 2.09310070e-01 -6.15067720e-01
-1.36380780e+00 -1.50057957e-01 6.60908073e-02 -3.68279308e-01
8.75977457e-01 4.45357233e-01 3.48572165e-01 6.46020174e-01
4.43061352e-01 7.26051569e-01 1.40533674e+00 -2.95790702e-01
2.04858080e-01 1.95600212e-01 -1.57710686e-02 8.67658734e-01
-1.76743418e-01 4.16585267e-01 1.36982203e-01 -1.04501463e-01
1.22111475e+00 6.49834350e-02 -2.09807694e-01 1.04341674e+00
-9.02480125e-01 1.00896561e+00 2.45023265e-01 8.01982701e-01
-2.18459085e-01 3.29795808e-01 3.35777879e-01 3.91213983e-01
-1.58521801e-01 5.03974020e-01 -2.96411604e-01 -5.69867849e-01
-6.84307218e-01 4.58737910e-01 8.29057157e-01 1.24235654e+00
4.74002123e-01 8.62949863e-02 2.12511227e-01 5.96638560e-01
-1.24843232e-01 2.78058350e-01 8.96303654e-02 -1.06671846e+00
2.64594764e-01 3.71147066e-01 -1.38771653e-01 -3.33743274e-01
-5.11931837e-01 -7.84403563e-01 -1.03393471e+00 2.42031783e-01
5.96459568e-01 -8.75164717e-02 -2.91121006e-01 2.11529279e+00
-6.57558501e-01 -4.00980651e-01 -2.46413555e-02 6.78874612e-01
3.44362147e-02 6.27801061e-01 3.58465612e-01 -8.99693191e-01
1.83854270e+00 -4.01114039e-02 -7.58103142e-03 -6.04149932e-03
5.58166802e-01 -1.15739846e+00 8.69447649e-01 6.45736456e-01
-1.46712804e+00 -3.90621632e-01 -9.10631120e-01 4.00053672e-02
5.51590556e-03 -2.97241639e-02 -1.15737893e-01 6.02394521e-01
-1.04132974e+00 5.20687521e-01 -2.13522419e-01 -4.44469191e-02
-4.64858525e-02 5.93998909e-01 -4.09653038e-01 4.15370405e-01
-1.05255234e+00 8.16351295e-01 8.38641107e-01 4.18763518e-01
-5.38855553e-01 -1.04390737e-02 -7.33428478e-01 1.36279583e-01
7.60823786e-01 -3.64946455e-01 1.18375731e+00 -1.00497079e+00
-6.24334931e-01 8.87899995e-01 -3.38510603e-01 -4.09810126e-01
6.90976918e-01 4.82064486e-01 -6.61374211e-01 1.73781887e-01
6.29047677e-02 5.51634312e-01 2.50704110e-01 -1.26525831e+00
-1.18349075e+00 -2.96885997e-01 7.00425744e-01 2.53298074e-01
-1.52412117e-01 4.77452248e-01 6.72049150e-02 -1.42749906e+00
4.54660028e-01 -1.31004286e+00 3.04457515e-01 -1.51398882e-01
-5.42112350e-01 -8.36854801e-02 7.50400603e-01 -2.43975505e-01
1.66331697e+00 -2.02335048e+00 5.57684183e-01 3.27270895e-01
7.64452636e-01 6.45370483e-01 -1.74627379e-01 2.53028840e-01
-2.88945943e-01 2.53103107e-01 -8.35533917e-01 2.90548772e-01
3.49289142e-02 4.07439798e-01 1.23551590e-02 1.67740181e-01
-6.25035882e-01 2.88850069e-01 -6.87329412e-01 -1.70221105e-01
-4.16367412e-01 1.38900027e-01 -7.78899729e-01 -2.20684528e-01
-7.00751841e-01 3.74403656e-01 -8.34346652e-01 -2.18392015e-01
4.20642883e-01 -2.55759001e-01 6.61085695e-02 -5.87492287e-02
-7.70529509e-01 1.81059942e-01 -1.73184896e+00 1.13594341e+00
4.36752513e-02 -2.28623748e-02 5.98256290e-01 -9.16302800e-01
1.22844529e+00 4.34985995e-01 8.18372965e-01 -5.16325593e-01
6.54311359e-01 5.50459385e-01 1.38678744e-01 3.57245594e-01
5.07766485e-01 -5.93238354e-01 -4.00739200e-02 1.41657269e+00
-4.61246490e-01 -2.29363829e-01 6.94180965e-01 -2.66885132e-01
1.43849540e+00 1.35581911e-01 3.82351764e-02 -5.04431844e-01
9.93736207e-01 -5.72316408e-01 7.05066979e-01 5.65754831e-01
-2.51231819e-01 1.35281339e-01 8.24674964e-01 -3.10642153e-01
-1.66674519e+00 -8.65521252e-01 -3.73205513e-01 1.12807930e+00
-2.03887865e-01 -4.98824716e-01 -1.02543640e+00 -4.77488339e-01
-4.49195057e-01 1.08858311e+00 -6.17366970e-01 -5.40971160e-02
-8.40236485e-01 -1.01709306e+00 1.04401112e+00 8.30133200e-01
7.57318616e-01 -1.09847105e+00 -8.04943681e-01 3.32437932e-01
-6.06441736e-01 -7.77652144e-01 -6.80605412e-01 6.21168375e-01
-4.38389063e-01 -4.58016634e-01 -6.93467319e-01 -2.93387920e-01
6.32583439e-01 -3.25133473e-01 1.32531822e+00 -1.83823779e-01
-3.50489110e-01 -2.32031569e-01 -4.21251565e-01 -7.20081985e-01
-1.01348174e+00 -3.82387266e-02 2.44969472e-01 -3.17709476e-01
4.88470107e-01 -5.95501244e-01 -8.81239772e-01 6.80656016e-01
-1.09507358e+00 -3.32786232e-01 2.61075318e-01 4.50852156e-01
8.62513185e-01 2.00102255e-01 8.70831311e-02 -6.06760502e-01
8.70731235e-01 -2.14570761e-01 -3.89242321e-01 -4.03259546e-01
-3.97368133e-01 1.53349668e-01 1.19580925e+00 1.50857925e-01
-6.31140232e-01 -6.87209606e-01 -4.98851538e-01 -3.18359822e-01
-5.16174018e-01 5.41709185e-01 -1.67952403e-01 1.03707588e+00
9.86827672e-01 5.10960221e-01 -5.87926686e-01 -5.46345532e-01
2.33085006e-01 3.64449829e-01 8.80244136e-01 -7.54931152e-01
4.53152061e-01 1.34428352e-01 -4.51137200e-02 -3.18130165e-01
-3.79973441e-01 4.17580605e-01 -8.81491959e-01 1.73071027e-01
1.41212273e+00 -6.07447684e-01 -1.00758100e+00 -1.64418519e-01
-1.08763969e+00 2.90583204e-02 -5.38889110e-01 7.53168046e-01
-1.01399231e+00 -7.41577521e-02 -8.85184944e-01 -4.46410894e-01
-1.04397565e-01 -1.12576377e+00 8.54382634e-01 1.59374133e-01
-5.75698674e-01 -6.23565972e-01 4.71168431e-03 1.54688060e-01
-2.85405934e-01 -1.98860407e-01 1.42115927e+00 -3.07376832e-01
-8.39018404e-01 -1.48976669e-01 -3.03922772e-01 5.95810592e-01
1.51332498e-01 -1.91631705e-01 -4.57287490e-01 -3.38158548e-01
-1.73999608e-01 7.19336033e-01 5.00511110e-01 8.81655440e-02
7.73809254e-01 -7.37301946e-01 1.55388147e-01 5.05059898e-01
1.63530397e+00 8.22852671e-01 6.36817634e-01 7.28453472e-02
-2.37576619e-01 1.89859524e-01 5.82696021e-01 4.79154773e-02
3.22570413e-01 8.26462269e-01 2.04625696e-01 6.32667243e-01
4.52059567e-01 1.54667377e-01 5.51582873e-01 1.22709715e+00
-1.07895601e+00 -4.74987149e-01 -1.05194104e+00 4.96993929e-01
-1.40271699e+00 -7.63198555e-01 -3.35506409e-01 2.17781162e+00
8.55281293e-01 6.52590930e-01 1.93866625e-01 3.43590677e-01
6.63810194e-01 2.39625111e-01 8.34486037e-02 -1.00582182e+00
-5.19982755e-01 4.66179729e-01 2.32627183e-01 5.33570051e-01
-5.46486199e-01 9.77950096e-01 4.19461918e+00 1.12572181e+00
-1.02056134e+00 -7.32434448e-04 4.62465197e-01 8.45936537e-02
-9.24789608e-01 -1.14383720e-01 -9.77834761e-01 8.16789508e-01
9.56504881e-01 3.40128615e-02 3.88150662e-01 3.42683971e-01
1.11343354e-01 -2.98067182e-01 -1.20903003e+00 6.72066152e-01
3.39607894e-01 -1.23913682e+00 -6.88565731e-01 1.64226860e-01
4.31876987e-01 -4.99475226e-02 2.52731323e-01 3.51931602e-01
9.01292503e-01 -1.34134662e+00 1.06399512e+00 2.41132572e-01
1.34563029e+00 -7.01987803e-01 -2.88421601e-01 5.63930452e-01
-1.33610940e+00 -1.83946073e-01 -3.31578135e-01 -1.49974197e-01
4.23099250e-01 2.31494844e-01 8.99967388e-04 7.86760092e-01
5.03441930e-01 7.56005850e-03 -1.25527024e-01 4.72746998e-01
-3.03737819e-01 6.51264727e-01 -5.63213229e-01 -6.73723519e-02
7.50184238e-01 -8.00489962e-01 1.12416959e+00 1.23105013e+00
5.89423716e-01 1.22972512e+00 -6.34333193e-02 9.86250043e-01
-2.12069497e-01 1.14482678e-01 -3.06505501e-01 2.72988342e-02
-5.67968078e-02 1.49830803e-01 -1.36501753e+00 7.95004293e-02
-1.48106068e-01 6.84892416e-01 1.25518680e-01 1.14348277e-01
-9.37027395e-01 -6.30975902e-01 6.45464778e-01 1.30276784e-01
1.93478152e-01 -3.74550909e-01 -3.34438473e-01 -7.62133956e-01
-1.14313029e-01 -8.61683071e-01 3.82874727e-01 -8.10955465e-01
-6.96432233e-01 8.64112914e-01 2.06793651e-01 -9.84519422e-01
-7.55978525e-01 -4.95873302e-01 -2.44050533e-01 9.59189892e-01
-5.21077037e-01 -4.10212785e-01 8.62547696e-01 8.66966784e-01
1.04798950e-01 -5.40116906e-01 1.29813707e+00 4.67669994e-01
-3.66799772e-01 2.40388215e-01 3.87828559e-01 3.82253379e-01
5.48674986e-02 -1.21152651e+00 4.37372401e-02 5.80147028e-01
3.76262099e-01 7.50939965e-01 8.13090265e-01 -7.41450608e-01
-6.69949114e-01 -8.68431032e-01 1.14233375e+00 -2.57233799e-01
5.15542388e-01 -3.55704010e-01 -6.45439982e-01 1.71914339e+00
-2.70490069e-02 -3.12459499e-01 6.18902206e-01 -1.64115295e-01
-2.42100239e-01 -7.78996646e-02 -7.71838546e-01 9.03594971e-01
6.18072927e-01 -5.97676516e-01 -5.44844449e-01 -4.80982922e-02
1.11190878e-01 -3.09130341e-01 -1.29591894e+00 5.73431611e-01
2.29717821e-01 -1.42975605e+00 2.30815947e-01 -1.30031094e-01
5.60191751e-01 -6.13515496e-01 2.80989930e-02 -7.26180851e-01
2.00572107e-02 -7.75413096e-01 3.12877536e-01 9.80350554e-01
2.57967263e-01 -2.86099583e-01 4.82740462e-01 7.24140257e-02
-3.51210862e-01 -8.77912045e-02 -1.61353374e+00 -3.87642026e-01
8.14659119e-01 -1.08487022e+00 4.69826788e-01 7.74034381e-01
-5.25729001e-01 3.04569188e-03 -1.12333260e-01 -3.73919666e-01
9.69653204e-02 6.66039512e-02 2.09034905e-02 -8.75365913e-01
-7.37499177e-01 -7.39639103e-01 -2.11702764e-01 -1.38020229e+00
-6.40490055e-02 -1.28004456e+00 -5.35894454e-01 -1.09525502e+00
1.10913239e-01 -6.52082503e-01 -5.06674170e-01 5.33288717e-01
3.68654281e-01 5.77368259e-01 8.11408460e-01 2.58316725e-01
-4.93620545e-01 7.45809227e-02 1.65451336e+00 2.89551228e-01
1.48662239e-01 -2.01277789e-02 -8.02003562e-01 1.36400175e+00
2.89749235e-01 -6.99798346e-01 -4.95169312e-01 -5.72323263e-01
1.01433694e+00 6.17763221e-01 6.90887645e-02 -7.97292113e-01
3.56702536e-01 1.96575467e-02 5.08220971e-01 -4.80341703e-01
7.32371211e-01 -5.60961008e-01 2.91775674e-01 2.87127551e-02
-2.37457216e-01 5.08197308e-01 1.99917629e-01 3.24113257e-02
-9.59199220e-02 -3.83752435e-01 1.02774191e+00 -6.52413607e-01
-4.71520185e-01 4.69395369e-01 -8.67594481e-01 3.04327875e-01
1.04868054e+00 -5.34809947e-01 8.43286216e-02 -1.42673150e-01
-1.31385291e+00 -7.12942705e-02 2.64510572e-01 2.47274917e-02
3.33258770e-02 -9.57858384e-01 -1.20181239e+00 2.42788881e-01
-2.36896321e-01 1.87020808e-01 6.20992899e-01 6.45905018e-01
-7.94227660e-01 2.57272124e-01 -4.42135572e-01 -6.93590641e-01
-1.20194364e+00 4.12733585e-01 3.66888911e-01 -5.83913140e-02
-7.39599586e-01 1.37536144e+00 -5.52162975e-02 -1.38340265e-01
-4.31694835e-01 -1.93944708e-01 -4.46324646e-02 1.73634347e-02
-2.83120591e-02 6.00697339e-01 -3.17269385e-01 -1.25530827e+00
-6.93449471e-03 6.39000475e-01 -2.10958958e-01 -4.28252459e-01
7.71749377e-01 -2.98176736e-01 -1.01701176e+00 1.77283958e-01
1.16607261e+00 3.20540726e-01 -1.13793051e+00 6.39774740e-01
-3.11746746e-01 1.82709023e-01 -8.09589267e-01 -9.87804353e-01
-6.57203734e-01 6.22292817e-01 5.29698372e-01 6.01229481e-02
8.80343556e-01 -1.23343319e-01 5.49387336e-01 3.51242393e-01
8.66112173e-01 -1.42400491e+00 -4.41510499e-01 1.03665352e+00
8.22360396e-01 -5.09115279e-01 -2.73973078e-01 -2.89849728e-01
-6.40282035e-01 1.18008947e+00 -3.41787279e-01 -1.94476843e-01
7.92429864e-01 4.07260090e-01 1.07638940e-01 -1.03141747e-01
-3.85111243e-01 -2.04032049e-01 3.54331620e-02 -8.24516360e-03
7.33577192e-01 4.13074642e-02 -1.13089883e+00 1.10670114e+00
-1.17005694e+00 2.52016962e-01 5.35503626e-01 1.07436407e+00
-3.76826227e-01 -1.34260571e+00 -3.87500286e-01 -1.97367016e-02
-6.85581744e-01 -2.23780155e-01 -4.32864055e-02 7.67013609e-01
1.02011740e+00 8.80367696e-01 1.63991630e-01 -2.82093197e-01
6.04910851e-01 2.08429903e-01 4.67763066e-01 -5.11073232e-01
-1.32569611e+00 8.54473054e-01 -1.53606772e-01 -1.33864842e-02
-4.22350556e-01 -6.86517894e-01 -1.40379429e+00 -8.70740712e-01
2.36452803e-01 5.56067705e-01 1.02085859e-01 1.11497235e+00
-3.16268265e-01 5.15053093e-01 5.89373447e-02 -1.63550690e-01
-5.68424881e-01 -4.06786054e-01 -1.36845672e+00 4.72675562e-01
-6.76090568e-02 -2.39695176e-01 -1.04024954e-01 1.04945730e-02] | [14.103391647338867, 13.317919731140137] |
544425e9-1e7c-4335-af37-81755eb30987 | vehicle-occurrence-based-parking-space | 2306.09940 | null | https://arxiv.org/abs/2306.09940v1 | https://arxiv.org/pdf/2306.09940v1.pdf | Vehicle Occurrence-based Parking Space Detection | Smart-parking solutions use sensors, cameras, and data analysis to improve parking efficiency and reduce traffic congestion. Computer vision-based methods have been used extensively in recent years to tackle the problem of parking lot management, but most of the works assume that the parking spots are manually labeled, impacting the cost and feasibility of deployment. To fill this gap, this work presents an automatic parking space detection method, which receives a sequence of images of a parking lot and returns a list of coordinates identifying the detected parking spaces. The proposed method employs instance segmentation to identify cars and, using vehicle occurrence, generate a heat map of parking spaces. The results using twelve different subsets from the PKLot and CNRPark-EXT parking lot datasets show that the method achieved an AP25 score up to 95.60\% and AP50 score up to 79.90\%. | ['Rodrigo A. Krauel', 'João V. Fröhlich', 'Andre Gustavo Hochuli', 'Luiz S. Oliveira', 'Jeovane Honório Alves', 'Paulo R. Lisboa de Almeida'] | 2023-06-16 | null | null | null | null | ['management'] | ['miscellaneous'] | [-3.09413940e-01 3.67833786e-02 -1.83456555e-01 -3.08814436e-01
-4.83368576e-01 -1.90990075e-01 6.82627141e-01 -9.39689018e-03
-5.52345097e-01 7.09059179e-01 -2.68939257e-01 -3.91673923e-01
1.54116526e-01 -1.03382194e+00 -3.27346355e-01 -5.24283409e-01
1.80993095e-01 6.17136598e-01 5.51246583e-01 -1.78337917e-01
5.79758108e-01 4.83048946e-01 -2.11293316e+00 -1.89096093e-01
9.06864047e-01 8.83659542e-01 7.47293651e-01 2.90347129e-01
-2.77712941e-01 1.12458482e-01 -2.58191198e-01 2.52581444e-02
4.73765761e-01 2.24449616e-02 -2.72381663e-01 2.84046680e-01
-4.10586707e-02 5.16135544e-02 8.97095874e-02 8.69167626e-01
3.12557936e-01 -1.74002603e-01 4.75549161e-01 -1.23399699e+00
1.82421327e-01 -3.23958782e-04 -3.04340661e-01 2.39997312e-01
-1.28341123e-01 3.72291088e-01 7.43603408e-01 -1.30564964e+00
4.04991448e-01 9.05320883e-01 4.10064906e-01 -1.63444262e-02
-1.16350389e+00 -6.12803400e-01 -5.81489921e-01 6.68308675e-01
-1.65028536e+00 -3.98717165e-01 6.93347871e-01 -4.98726815e-01
8.27641726e-01 9.74727944e-02 6.35815442e-01 -7.26274475e-02
-1.60876587e-01 8.04987311e-01 1.32220376e+00 -5.63120425e-01
4.33522016e-01 5.51699698e-01 1.09086268e-01 2.01700315e-01
3.13188761e-01 3.63306031e-02 6.62118867e-02 3.75049978e-01
2.60719359e-01 1.70157924e-01 4.39273477e-01 -6.50774002e-01
-9.58171904e-01 1.17793679e+00 3.74301255e-01 4.04848963e-01
-6.22956753e-01 -3.65886867e-01 2.40540341e-01 -3.87232274e-01
3.85293551e-02 6.38631582e-02 7.40510076e-02 -1.55717701e-01
-1.21228373e+00 2.59039074e-01 3.53628486e-01 7.18824327e-01
1.35974848e+00 -4.15202342e-02 2.08390579e-01 7.12433636e-01
3.40364516e-01 9.36231077e-01 1.54278442e-01 -8.34062397e-01
4.73633349e-01 7.52018869e-01 3.10096025e-01 -1.03690886e+00
-3.35553318e-01 2.29145184e-01 -3.66873831e-01 1.87091514e-01
1.83707878e-01 2.37259433e-01 -9.01468992e-01 8.52837741e-01
2.31074274e-01 3.02279852e-02 -1.44980863e-01 8.00733209e-01
5.42756379e-01 6.09381318e-01 2.91061640e-01 1.49223983e-01
1.24013484e+00 -6.02169096e-01 -4.78476077e-01 -6.17252588e-01
3.77052307e-01 -7.01576948e-01 9.53797460e-01 7.11590201e-02
-8.82990241e-01 -7.64772952e-01 -9.18052614e-01 4.93655771e-01
-8.85194957e-01 5.45471907e-01 2.19190970e-01 7.62238622e-01
-1.06472492e+00 8.72347280e-02 -5.45095921e-01 -5.28923094e-01
4.26331222e-01 4.09929991e-01 -1.13946341e-01 -1.72487706e-01
-1.18947315e+00 1.17769897e+00 5.52145660e-01 -1.08443998e-01
-5.37683010e-01 -1.98117882e-01 -8.17009032e-01 5.34395017e-02
1.41448334e-01 -9.99149233e-02 7.54026771e-01 -1.57009289e-01
-8.87623310e-01 9.48048234e-01 -2.16912523e-01 -8.45282197e-01
2.58105785e-01 1.51318267e-01 -3.86984408e-01 6.02396540e-02
6.51911259e-01 9.45424855e-01 4.70157534e-01 -1.16486573e+00
-1.16947651e+00 -5.67017317e-01 -3.99909526e-01 7.27666318e-02
2.20281616e-01 -9.30354372e-02 -6.61800802e-01 2.86361307e-01
-1.30987078e-01 -1.14358675e+00 -4.71594900e-01 -6.95707321e-01
-2.94817954e-01 -2.54772693e-01 1.18042850e+00 -6.30765557e-01
1.03901386e+00 -2.14400673e+00 -6.74147248e-01 3.40464264e-01
-1.02983452e-01 6.60607457e-01 2.96176881e-01 2.41626412e-01
1.44816786e-01 -3.43434125e-01 -1.45409003e-01 -4.66979623e-01
7.80236721e-02 5.73166668e-01 -1.47840217e-01 3.56889516e-01
8.49033892e-02 8.83390784e-01 -5.22887290e-01 -7.67207205e-01
1.39151478e+00 4.95882422e-01 -1.54466406e-01 -2.34140337e-01
5.95796406e-02 7.58198500e-02 -2.72267610e-01 6.78184748e-01
9.82730091e-01 9.63722244e-02 -3.03897206e-02 1.56093225e-01
-6.55297458e-01 1.07719101e-01 -1.32169056e+00 8.13531399e-01
-5.06080627e-01 7.75536716e-01 1.11948200e-01 -1.26678503e+00
1.34438038e+00 -1.99494511e-01 6.57081187e-01 -1.39508867e+00
3.19443084e-02 4.02387768e-01 -5.13731658e-01 -2.49556005e-01
1.19402528e+00 4.12331894e-02 -2.55977035e-01 8.58434215e-02
-6.16523802e-01 2.14010924e-02 4.04949933e-01 -3.43227037e-03
7.86004961e-01 -2.46547267e-01 1.04194269e-01 -1.71087738e-02
8.27575862e-01 5.68582714e-01 3.16307366e-01 3.47336739e-01
-5.36253273e-01 3.51621807e-01 -1.74300715e-01 -2.83756733e-01
-1.47842622e+00 -1.18028843e+00 -1.35321453e-01 4.28654552e-01
4.00448561e-01 -5.99157400e-02 -8.45799923e-01 -1.57471299e-01
3.47662836e-01 9.39157963e-01 -1.89946204e-01 7.28745610e-02
-5.61485827e-01 -3.75058144e-01 -4.19163480e-02 2.66398519e-01
7.14982808e-01 -1.24878311e+00 -1.11122084e+00 3.24228585e-01
-2.80421108e-01 -1.28306293e+00 2.01731697e-01 6.16842806e-02
-7.85647154e-01 -1.02164400e+00 -5.58142543e-01 -6.24894440e-01
6.74718380e-01 7.27228701e-01 9.41982508e-01 -2.98936963e-01
-6.16829216e-01 1.84244439e-01 -2.53856927e-02 -2.73230225e-01
-3.91806722e-01 -8.43928903e-02 -9.53977406e-02 -9.70673040e-02
1.06201458e+00 -7.97589198e-02 -6.76440954e-01 9.10224140e-01
-4.49103117e-01 -1.20216548e-01 8.00867200e-01 5.32952607e-01
9.90957677e-01 3.85896087e-01 5.48945785e-01 -1.71211973e-01
3.68655145e-01 -6.15687490e-01 -1.05695319e+00 -3.32708061e-01
-1.12328219e+00 -1.39844194e-01 4.37805891e-01 3.97513300e-01
-5.72660625e-01 6.92135930e-01 1.64079890e-01 -3.44813943e-01
-6.39566660e-01 2.97187753e-02 -2.73072749e-01 2.08291203e-01
2.56407768e-01 4.50183660e-01 3.29733044e-01 -3.59667957e-01
4.07510161e-01 9.44356620e-01 7.34737456e-01 2.06974164e-01
7.65338957e-01 7.91160941e-01 6.38830736e-02 -1.04742420e+00
-2.25690007e-02 -1.24390244e+00 -6.96735859e-01 -4.54947710e-01
9.28069711e-01 -9.27672088e-01 -6.33582652e-01 -7.58503452e-02
-6.47743404e-01 1.02107443e-01 -2.66650110e-01 4.39143479e-01
-6.83706522e-01 3.55980545e-01 3.83811414e-01 -1.11478746e+00
-2.61326700e-01 -1.20250595e+00 1.04923189e+00 3.79559755e-01
-6.04285114e-02 -4.92787600e-01 3.46167944e-02 7.64786541e-01
5.66505075e-01 -5.55660110e-03 1.60230145e-01 -2.93027729e-01
-8.59270096e-01 -5.99036634e-01 -4.77052480e-01 2.30159745e-01
2.10048668e-02 -3.73098046e-01 -9.88893330e-01 2.03595012e-01
-6.15820169e-01 3.83836806e-01 8.45640242e-01 7.74621010e-01
5.05638182e-01 1.57989100e-01 -7.35715330e-01 -2.44779348e-01
1.47615016e+00 2.95007735e-01 1.09990907e+00 9.12170947e-01
1.22481048e-01 6.36906683e-01 1.30151403e+00 5.53580165e-01
6.52954638e-01 9.42156732e-01 9.30557787e-01 -1.83833078e-01
1.07078969e-01 -2.34969065e-01 3.10172915e-01 2.61375103e-02
2.21945569e-01 4.79936182e-01 -1.05715573e+00 1.32022405e+00
-1.86935985e+00 -1.19205225e+00 -3.80337775e-01 2.29267859e+00
6.26620725e-02 3.95116001e-01 7.46511281e-01 5.75805008e-01
1.11688077e+00 -1.27090022e-01 -1.71465755e-01 -5.62634528e-01
-2.54431590e-02 1.06909499e-01 1.24131000e+00 3.22922140e-01
-1.12729764e+00 1.05448544e+00 5.70877457e+00 7.33187556e-01
-9.82186854e-01 -1.63882837e-01 3.55695218e-01 2.83142656e-01
8.62125382e-02 1.60305753e-01 -1.06578052e+00 9.54149544e-01
1.23955417e+00 -9.96595547e-02 1.74805984e-01 1.29573762e+00
9.67619956e-01 -7.62364626e-01 -3.16636920e-01 1.17373776e+00
-1.10849380e-01 -1.37451208e+00 -5.69370031e-01 6.10208333e-01
4.36824590e-01 4.04306084e-01 9.60419998e-02 1.10110104e-01
2.21866101e-01 -9.83479381e-01 4.64453250e-01 5.04067063e-01
5.23040235e-01 -9.96203542e-01 8.77495944e-01 3.52555275e-01
-1.43106854e+00 -1.67069092e-01 -2.96356916e-01 -1.66504636e-01
5.93090475e-01 6.84610367e-01 -1.86752832e+00 2.00620010e-01
8.07842553e-01 4.67093319e-01 -7.76629508e-01 1.39769232e+00
4.18178737e-02 3.96968812e-01 -6.47193372e-01 -2.37249002e-01
4.26164299e-01 -3.96303505e-01 3.19208741e-01 1.14680743e+00
4.69981492e-01 -5.42055309e-01 2.50002235e-01 1.07317662e+00
6.67828679e-01 2.48190034e-02 -9.45617139e-01 2.11855829e-01
6.06865048e-01 1.43349934e+00 -8.35575938e-01 -2.18173102e-01
-3.31908792e-01 3.45015913e-01 -4.82873857e-01 -1.51478842e-01
-7.97032297e-01 -3.85992467e-01 6.16514981e-01 6.11989796e-01
9.02824819e-01 -4.54544783e-01 -5.68724155e-01 -3.07946265e-01
-1.28007203e-01 1.08866118e-01 1.09911926e-01 -8.07614207e-01
-5.02599061e-01 3.26896384e-02 -4.54711914e-03 -1.37432146e+00
-4.18114632e-01 -5.79669118e-01 -5.84233105e-01 7.88900554e-01
-1.81945944e+00 -8.91650140e-01 -4.29967374e-01 5.15974104e-01
5.31862438e-01 -2.82865673e-01 3.93165320e-01 6.55175805e-01
-3.42994153e-01 -1.34673463e-02 1.64011940e-01 -2.40681749e-02
1.67962298e-01 -1.05881977e+00 3.05819422e-01 7.87712157e-01
-2.38507211e-01 -2.62768241e-03 9.58597898e-01 -5.81249535e-01
-1.11693823e+00 -1.16981244e+00 1.28898335e+00 -3.45833272e-01
3.23576301e-01 8.66031200e-02 -3.92502874e-01 8.39340240e-02
-7.89931342e-02 -1.99854746e-01 1.10297978e-01 -5.43447971e-01
4.61741954e-01 -3.06538343e-01 -1.38352120e+00 3.70810181e-01
2.13972047e-01 -4.18241650e-01 -4.60834414e-01 -1.05738431e-01
-1.69145897e-01 1.79286957e-01 -4.21506673e-01 1.70457318e-01
1.03059217e-01 -1.21476197e+00 1.11208522e+00 4.16478723e-01
-2.31257342e-02 -7.46704519e-01 -2.86267608e-01 -8.31084490e-01
4.65997320e-04 1.98162034e-01 4.69833404e-01 1.08291507e+00
3.68043870e-01 -3.99028957e-01 1.36162293e+00 6.62299693e-01
-2.60189146e-01 -2.21962705e-01 -1.22161901e+00 -6.91373646e-01
-3.39127690e-01 -5.81001699e-01 4.29265678e-01 4.67656672e-01
-2.00277090e-01 2.29720548e-01 -1.20370671e-01 2.23712288e-02
7.93445826e-01 2.23551318e-01 1.17485964e+00 -1.47994435e+00
5.85316479e-01 -4.62995648e-01 -8.71655524e-01 -5.16474009e-01
8.56899023e-02 -6.84592068e-01 9.29360539e-02 -1.88398433e+00
-3.01163271e-02 -6.63826108e-01 -1.00327302e-02 2.98665255e-01
3.34853202e-01 3.75847399e-01 1.63156837e-01 4.01879370e-01
-7.49449432e-01 2.48940617e-01 4.97699350e-01 -2.21274123e-01
-2.42470115e-01 1.51799738e-01 -1.90318227e-01 4.39200491e-01
1.15864253e+00 -6.00452647e-02 -7.97261894e-02 3.21332544e-01
-2.58601964e-01 -3.32010895e-01 6.19682968e-01 -1.27409256e+00
1.02862351e-01 -1.87194258e-01 9.94558409e-02 -1.50160825e+00
6.32737696e-01 -1.15077519e+00 3.51544291e-01 7.56159127e-01
4.94383991e-01 2.25223554e-03 2.27653459e-01 4.43610609e-01
-3.82482946e-01 -8.39073118e-03 8.62458408e-01 -1.48221865e-01
-1.57183063e+00 -1.67182699e-01 -5.13442755e-01 -4.57275838e-01
1.54028499e+00 -1.00668240e+00 -1.00449428e-01 -1.97450459e-01
-5.91939270e-01 4.29788142e-01 6.53186679e-01 4.43831861e-01
6.05655491e-01 -1.35435534e+00 -2.70225316e-01 3.70013982e-01
4.25437897e-01 -8.92403349e-02 1.72779605e-01 8.78438294e-01
-4.79711950e-01 1.11682439e+00 -4.11840886e-01 -1.03714418e+00
-1.42537630e+00 6.04309082e-01 4.87617142e-02 3.09456070e-03
-5.92674315e-01 -1.31475031e-01 -3.80279899e-01 -3.53522092e-01
-2.59482473e-01 7.56827295e-02 -5.91843188e-01 1.29360378e-01
1.97114527e-01 6.26737297e-01 1.83433890e-01 -1.29070890e+00
-6.87301993e-01 7.38430500e-01 4.83399659e-01 -5.06871752e-02
1.21102524e+00 -7.08188176e-01 3.25850844e-01 1.20976023e-01
7.61009634e-01 -1.37389064e-01 -9.17000055e-01 -1.15394004e-01
4.22745109e-01 -5.34853995e-01 5.72165176e-02 -4.80031669e-01
-8.22707713e-01 7.97363460e-01 1.07790792e+00 4.21236426e-01
8.61826301e-01 2.71566510e-01 8.72442663e-01 1.01256920e-02
5.52530885e-01 -1.76609719e+00 -5.44266522e-01 4.02111888e-01
5.52718788e-02 -1.53399777e+00 -2.33024970e-01 -1.22315176e-01
-8.02142203e-01 8.61197472e-01 2.84702003e-01 4.06893641e-02
5.68422794e-01 -1.80909391e-02 -2.34822407e-02 -2.64560312e-01
-1.65407404e-01 -8.82026851e-01 -2.88843066e-01 8.11640918e-01
-1.93753883e-01 6.37364686e-01 -5.32784939e-01 9.26544890e-02
-2.73219973e-01 2.38538146e-01 4.03675854e-01 9.16275740e-01
-1.18799376e+00 -9.75099742e-01 -8.45132768e-01 3.00813347e-01
2.34514624e-02 1.81472093e-01 -1.68778732e-01 7.79380739e-01
3.68952692e-01 1.16287577e+00 3.55827898e-01 -3.36960375e-01
3.94813627e-01 3.91459540e-02 -3.09383869e-01 -2.09505230e-01
2.34750852e-01 -8.75171497e-02 1.94440693e-01 -3.97632807e-01
-4.70787406e-01 -7.36652255e-01 -1.56803739e+00 -4.22653317e-01
-2.80761421e-01 5.45496285e-01 1.32846558e+00 7.41379142e-01
6.50026381e-01 1.93277635e-02 1.07151127e+00 -1.03299463e+00
-1.40387312e-01 -5.74586987e-01 -7.27712810e-01 3.26140046e-01
-1.73126776e-02 -1.05698264e+00 -3.36545974e-01 -1.42661065e-01] | [8.043051719665527, -1.134684681892395] |
c0cec9dc-c77e-49cb-8513-c3326b24db6f | adding-context-to-source-code-representations | 2208.00203 | null | https://arxiv.org/abs/2208.00203v1 | https://arxiv.org/pdf/2208.00203v1.pdf | Adding Context to Source Code Representations for Deep Learning | Deep learning models have been successfully applied to a variety of software engineering tasks, such as code classification, summarisation, and bug and vulnerability detection. In order to apply deep learning to these tasks, source code needs to be represented in a format that is suitable for input into the deep learning model. Most approaches to representing source code, such as tokens, abstract syntax trees (ASTs), data flow graphs (DFGs), and control flow graphs (CFGs) only focus on the code itself and do not take into account additional context that could be useful for deep learning models. In this paper, we argue that it is beneficial for deep learning models to have access to additional contextual information about the code being analysed. We present preliminary evidence that encoding context from the call hierarchy along with information from the code itself can improve the performance of a state-of-the-art deep learning model for two software engineering tasks. We outline our research agenda for adding further contextual information to source code representations for deep learning. | ['Christoph Treude', 'Fuwei Tian'] | 2022-07-30 | null | null | null | null | ['code-classification', 'vulnerability-detection'] | ['computer-code', 'miscellaneous'] | [ 1.02535477e-02 3.33109587e-01 -3.38229150e-01 -5.41563451e-01
-4.06934112e-01 -5.11725008e-01 4.05006409e-01 9.54635918e-01
1.31272361e-01 1.11948945e-01 4.92673606e-01 -1.08717859e+00
1.10210240e-01 -8.30839217e-01 -6.44204140e-01 2.26537645e-01
-2.38020316e-01 -2.51976758e-01 2.41586730e-01 -2.32357413e-01
5.32859683e-01 2.38619000e-01 -1.58973002e+00 9.15564895e-01
5.31003296e-01 4.84874696e-01 2.28594262e-02 7.48791933e-01
-6.58114552e-01 1.37402523e+00 -1.06713986e+00 -3.10002565e-01
-2.78504491e-01 -1.95425078e-01 -1.31819379e+00 -4.17300582e-01
4.38865244e-01 -3.02837700e-01 8.26863572e-02 9.79723930e-01
-8.21121689e-03 -1.99991062e-01 4.52385336e-01 -1.14528465e+00
-6.25483930e-01 1.06317496e+00 -4.20534104e-01 3.65761787e-01
5.81716418e-01 1.97471268e-02 1.26204085e+00 -4.84981388e-01
6.88048542e-01 1.16457212e+00 8.86463225e-01 5.07867038e-01
-1.11759496e+00 -2.15088353e-01 2.73834258e-01 2.05059201e-01
-5.82259953e-01 -2.67385155e-01 8.25146079e-01 -9.25834060e-01
1.74961698e+00 1.63009152e-01 5.29098809e-01 8.36072266e-01
4.72012848e-01 8.25386345e-01 5.02130806e-01 -7.88502276e-01
2.41707221e-01 -1.36677891e-01 8.26044500e-01 9.11509931e-01
4.01479900e-01 -1.20433472e-01 -1.21718273e-01 -5.19096136e-01
3.86362493e-01 -9.12006199e-02 1.35015726e-01 -3.97609085e-01
-8.20225239e-01 1.08967257e+00 6.61177397e-01 6.22039914e-01
-1.13720514e-01 7.96618164e-01 1.08156395e+00 5.15746713e-01
3.00020248e-01 8.44155908e-01 -7.45802462e-01 -4.71569359e-01
-8.14447761e-01 2.47384727e-01 9.55297887e-01 8.72887552e-01
9.43276525e-01 3.38897616e-01 -8.15150440e-02 6.20892704e-01
5.61803758e-01 -3.72384995e-01 4.17393416e-01 -8.49475503e-01
6.50318146e-01 1.22581792e+00 -2.00497165e-01 -9.16489542e-01
-3.28132391e-01 -4.10406202e-01 -7.43772928e-03 4.97714251e-01
1.66671872e-01 -1.13074780e-01 -9.55306172e-01 1.46803188e+00
-1.40396312e-01 -8.62493962e-02 2.42496878e-02 2.12271228e-01
1.16792345e+00 5.03364325e-01 8.50280747e-02 3.81767124e-01
1.18898463e+00 -5.84739804e-01 -2.78900564e-01 -6.24031842e-01
1.58284235e+00 -5.12045383e-01 1.02628994e+00 2.19721675e-01
-8.91726375e-01 -5.20568490e-01 -1.27450955e+00 -2.88478971e-01
-5.61418295e-01 2.97794435e-02 9.38786983e-01 8.72458279e-01
-1.27723932e+00 5.59056878e-01 -9.35999751e-01 -3.55649918e-01
4.51252729e-01 2.75868207e-01 -2.66300946e-01 -8.68432689e-03
-8.81189167e-01 7.20348120e-01 6.29129887e-01 -1.55066341e-01
-1.04757082e+00 -5.11974275e-01 -1.44191694e+00 4.78111833e-01
2.09028006e-01 -3.60000163e-01 1.58923233e+00 -1.01891637e+00
-7.79955566e-01 7.47917652e-01 -1.81163736e-02 -4.60360050e-01
-2.36925408e-01 -1.58262551e-01 -1.48682818e-01 -3.59193265e-01
-4.42482904e-02 1.75548777e-01 6.01387501e-01 -1.10051441e+00
-5.52795410e-01 -3.03152114e-01 8.18048954e-01 -5.83076298e-01
-3.13872427e-01 5.94469607e-01 7.24037141e-02 -4.83488053e-01
-4.17490900e-01 -5.85568428e-01 -1.76477283e-01 -3.75858992e-01
-3.05904567e-01 -3.75131845e-01 8.81081879e-01 -9.31719244e-01
1.58334196e+00 -2.21047258e+00 1.54771268e-01 4.50416915e-02
5.28650701e-01 4.37137753e-01 -3.02700430e-01 5.32767296e-01
-3.75529706e-01 6.24721885e-01 -2.58794993e-01 2.09342614e-01
2.53790289e-01 1.68438867e-01 -1.88780993e-01 2.10122854e-01
4.67627048e-01 9.75239456e-01 -8.39699984e-01 -2.03069612e-01
-3.21941599e-02 2.53859550e-01 -8.91753614e-01 1.47033468e-01
-6.39570951e-01 -2.41126269e-01 -5.26836038e-01 6.28858149e-01
2.64790565e-01 -1.10303879e-01 3.07597723e-02 3.49681258e-01
-2.53372043e-02 9.24071193e-01 -8.51943791e-01 1.52809024e+00
-9.21887279e-01 1.09561634e+00 -9.52945948e-02 -1.03562379e+00
9.75217164e-01 3.50284815e-01 -4.76083942e-02 -4.75832850e-01
-1.50230870e-01 2.01388642e-01 3.63794506e-01 -5.77975571e-01
4.78769630e-01 1.58357978e-01 -5.13767064e-01 7.18356252e-01
2.17225149e-01 1.05751060e-01 3.22014630e-01 1.48474023e-01
1.45875251e+00 3.12845409e-01 4.63009715e-01 -2.41915971e-01
5.78075111e-01 -1.49929170e-02 4.29679573e-01 6.16786361e-01
1.65831283e-01 1.73414648e-01 1.27953708e+00 -8.36034834e-01
-9.38827515e-01 -4.42588359e-01 9.69517827e-02 1.29520905e+00
-5.77586293e-01 -1.23839748e+00 -9.98769641e-01 -1.03226340e+00
-1.14683472e-01 8.18712354e-01 -9.12083626e-01 -5.08570850e-01
-8.50255311e-01 -3.99124265e-01 7.78385520e-01 1.07662368e+00
-4.70558181e-02 -1.23053157e+00 -1.17256999e+00 4.74697143e-01
2.93093294e-01 -4.17723119e-01 5.80889210e-02 6.13944888e-01
-1.08196247e+00 -1.21467042e+00 -1.44539550e-01 -7.57731736e-01
4.47383642e-01 -3.70960653e-01 1.59607291e+00 7.60239720e-01
-3.51026863e-01 2.00400665e-01 -6.13691688e-01 -4.82054532e-01
-1.09059358e+00 3.67911130e-01 -7.66793847e-01 -6.40434265e-01
5.17108262e-01 -2.24949256e-01 -3.10953185e-02 -2.56143332e-01
-1.06272256e+00 -3.03029895e-01 3.26053381e-01 8.34469080e-01
-1.42176941e-01 2.78270524e-02 4.04291272e-01 -1.20689917e+00
8.29353511e-01 -6.54890478e-01 -4.22330141e-01 2.73765564e-01
-2.43273824e-01 3.76385361e-01 7.30204582e-01 -2.68936390e-03
-9.81028736e-01 -2.19965622e-01 -4.24311280e-01 7.08541125e-02
-3.73115242e-01 1.27051795e+00 -1.28451675e-01 -4.43147197e-02
9.39493418e-01 -1.64353013e-01 -2.30446890e-01 -5.72474658e-01
3.40049446e-01 7.28819728e-01 1.41241746e-02 -8.55761051e-01
2.87291884e-01 -1.58167928e-01 -1.07888401e-01 -6.58546388e-01
-2.90739059e-01 -2.34028012e-01 -8.13152254e-01 2.54474252e-01
5.85103035e-01 -4.05691117e-01 -2.09078491e-01 2.24926561e-01
-1.45015419e+00 -4.66646075e-01 -2.02218339e-01 -1.53271854e-01
-4.66871887e-01 3.83184820e-01 -6.88503921e-01 -6.05057061e-01
-4.57954593e-03 -1.45277858e+00 9.71567571e-01 -1.23893812e-01
-6.89703226e-01 -1.32259333e+00 1.96698662e-02 -8.01530629e-02
4.93939519e-01 5.79957187e-01 1.77271616e+00 -9.49593365e-01
-4.86652672e-01 -3.05585235e-01 -1.88506972e-02 3.63859177e-01
1.59540609e-01 3.51117700e-01 -1.04157913e+00 -2.05284193e-01
-4.44134995e-02 -3.46202940e-01 8.68241131e-01 6.59117848e-02
1.20687497e+00 -3.69596362e-01 -3.79768550e-01 6.33472502e-01
1.45261562e+00 3.36471319e-01 5.83588123e-01 6.18480504e-01
9.10930693e-01 6.79683328e-01 1.22092620e-01 3.52157265e-01
3.19885105e-01 3.57503921e-01 8.50665748e-01 9.42031071e-02
-8.44331831e-02 7.23828152e-02 4.36402351e-01 4.38313425e-01
3.15843493e-01 -1.34919323e-02 -1.46894491e+00 8.22172701e-01
-1.72391748e+00 -7.42637455e-01 -4.20991987e-01 1.94031894e+00
7.23429143e-01 2.09716067e-01 4.65096496e-02 4.15344328e-01
3.88140738e-01 2.87519336e-01 -2.62448788e-01 -1.22146881e+00
6.47335112e-01 2.85826445e-01 6.61856905e-02 3.99520576e-01
-8.63109648e-01 7.29479194e-01 6.66057968e+00 2.61373371e-01
-1.07691991e+00 8.20343196e-03 3.24079067e-01 4.91286308e-01
-6.38239443e-01 3.60484034e-01 -4.94213432e-01 1.53217748e-01
1.45097017e+00 -1.32374197e-01 1.69197500e-01 1.15193880e+00
-3.93771648e-01 -6.72785565e-03 -1.68750334e+00 3.14459175e-01
-2.09664911e-01 -1.50982392e+00 -5.12210652e-02 8.62097964e-02
4.48876411e-01 -9.51597188e-03 -2.23587453e-01 6.87442720e-01
4.69916254e-01 -1.16789865e+00 6.96324348e-01 1.30885258e-01
5.50285637e-01 -7.93779433e-01 9.63455677e-01 1.15584187e-01
-1.19203317e+00 -4.05987769e-01 -3.58290344e-01 -3.09858710e-01
-4.15473789e-01 3.75800788e-01 -1.07753813e+00 5.14730930e-01
5.45590699e-01 1.00022042e+00 -1.23100913e+00 1.00342536e+00
-2.22018749e-01 7.76883423e-01 1.76132917e-01 -1.65454775e-01
4.73674923e-01 6.28403604e-01 2.37944305e-01 1.67920840e+00
1.79983065e-01 -6.17570758e-01 3.05358525e-02 1.20610523e+00
2.45691705e-02 -1.46115780e-01 -8.11569810e-01 -5.02570629e-01
1.71758428e-01 7.11995900e-01 -6.48294210e-01 -4.23235238e-01
-8.70209813e-01 2.57837027e-01 3.20448041e-01 2.96426833e-01
-4.49630648e-01 -1.02056313e+00 7.53299654e-01 1.48037180e-01
3.84616107e-01 -1.97010428e-01 -5.12621820e-01 -1.08875453e+00
1.90366104e-01 -1.01163995e+00 5.31530917e-01 -7.30693579e-01
-6.73230648e-01 7.26867676e-01 1.34311616e-01 -7.89392114e-01
-7.38285899e-01 -7.57705033e-01 -9.55590367e-01 1.07009983e+00
-1.30040574e+00 -9.64622676e-01 -1.00736864e-01 1.57928556e-01
4.10675228e-01 -3.21562499e-01 8.73357415e-01 2.58094054e-02
-3.40387464e-01 5.49365997e-01 -1.99485213e-01 4.81010675e-01
6.89091310e-02 -1.67231870e+00 1.26310217e+00 1.12841105e+00
2.58359939e-01 1.15990841e+00 6.75335169e-01 -5.75977743e-01
-1.39367259e+00 -1.11075807e+00 8.20103347e-01 -7.32780278e-01
7.44846046e-01 -3.97193998e-01 -1.35332763e+00 1.13974047e+00
3.20469201e-01 1.58995479e-01 7.42745936e-01 5.26129842e-01
-7.87640750e-01 2.15128765e-01 -8.59071493e-01 2.16889620e-01
6.01877749e-01 -9.31062281e-01 -9.70328987e-01 -3.81315500e-02
7.51125693e-01 -4.47844267e-01 -8.45533907e-01 2.69497707e-02
1.85352698e-01 -1.03296828e+00 7.37298310e-01 -1.10002184e+00
8.53242755e-01 -1.09521575e-01 -8.09989274e-02 -1.35974550e+00
-3.73275787e-01 -4.76885140e-01 -5.83042026e-01 1.25183702e+00
4.11643833e-01 -3.58104169e-01 5.95539153e-01 5.06049454e-01
-5.71166992e-01 -5.29639244e-01 -4.83486205e-01 -5.71958959e-01
4.76486087e-01 -6.43431127e-01 1.00361633e+00 9.86138284e-01
4.49266106e-01 1.58820540e-01 2.38493100e-01 -4.95932065e-02
1.44404635e-01 1.19098403e-01 5.75493336e-01 -1.68869710e+00
-2.99040556e-01 -7.57251203e-01 -7.21680284e-01 -4.40907210e-01
5.78866184e-01 -1.25112784e+00 -6.68895692e-02 -2.13135147e+00
-8.38522613e-02 -1.80052102e-01 -2.79871196e-01 9.28607166e-01
-1.20778106e-01 -4.83773023e-01 1.43210247e-01 -1.25702634e-01
-3.85551155e-01 -5.65217696e-02 4.39628065e-01 -4.05481368e-01
3.07393074e-02 -1.06803551e-01 -9.13113117e-01 6.10636652e-01
7.78612316e-01 -6.66376770e-01 -4.05569166e-01 -7.81011105e-01
6.30631149e-01 1.92589074e-01 2.98824549e-01 -8.24135900e-01
-6.42164126e-02 -7.10460469e-02 6.79029599e-02 3.90972011e-03
-2.91249782e-01 -5.43988109e-01 -1.74324036e-01 6.46078169e-01
-6.54114664e-01 2.00577229e-01 6.64086342e-01 2.44900912e-01
-3.52196455e-01 -9.73604381e-01 4.21233296e-01 -4.95138347e-01
-1.14542520e+00 -1.91822574e-01 -7.25379288e-01 1.22293137e-01
6.77248180e-01 -2.56839544e-01 -4.51766372e-01 5.19622676e-02
-5.95025539e-01 -9.96689871e-02 7.92800605e-01 7.84419060e-01
6.05186224e-01 -1.13297141e+00 -4.90380079e-01 3.83239150e-01
5.89094579e-01 -2.44023159e-01 -2.13704765e-01 1.38711184e-01
-7.92272091e-01 6.24072552e-01 -4.66234297e-01 -3.53744745e-01
-1.19684446e+00 8.01692307e-01 3.99487317e-01 -2.09183410e-01
-6.96188688e-01 7.11237967e-01 1.03366241e-01 -5.10264456e-01
2.95315623e-01 -1.03371906e+00 -4.24978286e-01 -1.17454529e-02
7.23613918e-01 2.25866169e-01 5.71233511e-01 -3.46914262e-01
-6.01647079e-01 2.94762850e-01 -3.53993654e-01 3.09131414e-01
1.48506641e+00 5.18081307e-01 -5.59221923e-01 4.79997307e-01
1.35839021e+00 -1.72625542e-01 -1.02587438e+00 -1.04456231e-01
8.71644616e-01 -4.19810712e-01 1.70054570e-01 -9.06231582e-01
-1.02991414e+00 1.41402245e+00 7.45839477e-02 6.70171380e-01
7.58123457e-01 1.38538569e-01 4.38211203e-01 5.72552383e-01
4.91756856e-01 -6.08516455e-01 1.16656899e-01 8.44383538e-01
7.71485031e-01 -9.24548030e-01 -2.05755025e-01 1.52067223e-03
-9.29211527e-02 1.64349842e+00 7.15996981e-01 -1.81043327e-01
5.96130013e-01 6.45941317e-01 -1.17272168e-01 -4.35265809e-01
-1.05308533e+00 8.81107673e-02 2.72085965e-01 8.42930913e-01
1.17250538e+00 -2.24981502e-01 1.95444264e-02 5.90754807e-01
4.51236255e-02 6.96894825e-02 1.20354509e+00 1.55705738e+00
-5.64050257e-01 -1.36635184e+00 -2.55736947e-01 6.49706244e-01
-7.88857460e-01 -2.69330412e-01 -6.54334366e-01 7.41686940e-01
7.58286417e-02 7.71406353e-01 -3.32093472e-03 -3.82791370e-01
3.26238304e-01 2.53823072e-01 3.41435015e-01 -1.59102690e+00
-1.24954331e+00 -3.83387327e-01 2.85608292e-01 -5.15590250e-01
-4.86585870e-02 -5.09134531e-01 -1.29803765e+00 4.99406888e-04
-9.29537937e-02 1.87085763e-01 6.69676661e-01 9.13369179e-01
4.26063448e-01 1.06939721e+00 -4.65569682e-02 -5.90509176e-01
-3.47216129e-01 -7.69905627e-01 -1.45037606e-01 9.60337669e-02
9.07598436e-01 -4.33117300e-01 -1.23081952e-01 2.15051576e-01] | [7.614378929138184, 7.822826862335205] |
125139d6-d618-46a5-b262-f642a78da37d | mining-local-process-models | 1606.06066 | null | http://arxiv.org/abs/1606.06066v2 | http://arxiv.org/pdf/1606.06066v2.pdf | Mining Local Process Models | In this paper we describe a method to discover frequent behavioral patterns
in event logs. We express these patterns as \emph{local process models}. Local
process model mining can be positioned in-between process discovery and episode
/ sequential pattern mining. The technique presented in this paper is able to
learn behavioral patterns involving sequential composition, concurrency, choice
and loop, like in process mining. However, we do not look at start-to-end
models, which distinguishes our approach from process discovery and creates a
link to episode / sequential pattern mining. We propose an incremental
procedure for building local process models capturing frequent patterns based
on so-called process trees. We propose five quality dimensions and
corresponding metrics for local process models, given an event log. We show
monotonicity properties for some quality dimensions, enabling a speedup of
local process model discovery through pruning. We demonstrate through a real
life case study that mining local patterns allows us to get insights in
processes where regular start-to-end process discovery techniques are only able
to learn unstructured, flower-like, models. | ['Wil M. P. van der Aalst', 'Reinder Haakma', 'Natalia Sidorova', 'Niek Tax'] | 2016-06-20 | null | null | null | null | ['model-discovery', 'sequential-pattern-mining'] | ['miscellaneous', 'natural-language-processing'] | [ 4.85185683e-01 3.81922632e-01 -8.63597170e-02 -6.74155802e-02
-1.34226531e-01 -5.02712429e-01 8.75121951e-01 6.70706332e-01
-3.14682946e-02 3.54886979e-01 1.60084087e-02 -4.90423679e-01
-8.30479622e-01 -1.11310267e+00 -1.68975428e-01 -2.37475947e-01
-1.02236784e+00 9.53966141e-01 6.16533935e-01 4.84380960e-01
2.16998979e-01 6.33508563e-01 -1.65670562e+00 6.40293300e-01
2.57241517e-01 5.56551635e-01 -1.38750017e-01 8.54797602e-01
-6.46074831e-01 1.21617472e+00 -3.54333431e-01 -4.82011251e-02
2.73112953e-01 -6.07194424e-01 -9.61777329e-01 3.87539357e-01
-3.99446011e-01 1.75302908e-01 2.63895124e-01 4.43837255e-01
-1.56380966e-01 -2.29638979e-01 4.10820872e-01 -1.58539200e+00
3.35157126e-01 8.96588981e-01 -7.01853633e-01 4.52529877e-01
8.82566810e-01 2.25958362e-01 9.95740473e-01 -6.30203784e-01
8.75326514e-01 1.30199254e+00 8.57668400e-01 3.60987246e-01
-1.60383856e+00 -4.07636881e-01 2.24764213e-01 -3.79874967e-02
-1.03452027e+00 -2.49848828e-01 1.81300640e-01 -4.52957034e-01
1.12619543e+00 5.77363014e-01 9.77442861e-01 9.63577569e-01
3.13149661e-01 1.07045627e+00 1.20799530e+00 -4.82000798e-01
6.61516905e-01 -2.60278344e-01 4.50201571e-01 6.08669221e-01
4.93639797e-01 7.64708593e-02 -6.87926471e-01 -8.92760694e-01
1.08416820e+00 6.50805831e-01 6.32375553e-02 -3.82508278e-01
-1.27594805e+00 5.48676252e-01 -5.73962390e-01 5.04175305e-01
-4.98804867e-01 2.17545822e-01 3.66527677e-01 8.98074627e-01
-1.40123412e-01 3.51521015e-01 -4.00440991e-01 -6.16654098e-01
-8.13802302e-01 3.94648880e-01 1.83477461e+00 1.03969872e+00
8.82106841e-01 -3.81394088e-01 -1.72852367e-01 1.40135229e-01
2.83219218e-01 4.13335897e-02 2.63076484e-01 -8.06898594e-01
-1.28639698e-01 1.04600489e+00 2.56009459e-01 -4.69746619e-01
-4.98184055e-01 -3.18174921e-02 -7.10229516e-01 -7.93527141e-02
5.20683587e-01 5.17220274e-02 -5.14474869e-01 1.07262182e+00
2.22809523e-01 2.44666830e-01 -1.79757655e-01 8.86335373e-02
-1.58680126e-01 4.27244574e-01 1.42130002e-01 -1.06025946e+00
1.45483613e+00 -7.19170332e-01 -7.07607329e-01 2.83995569e-02
4.19826955e-01 -3.99009913e-01 8.60004544e-01 7.44412959e-01
-1.15720201e+00 -2.98885107e-01 -5.72441339e-01 9.03328478e-01
2.57282615e-01 -6.37019157e-01 1.11921549e+00 7.13388324e-01
-9.20524895e-01 8.69211853e-01 -1.40527809e+00 -7.62367189e-01
1.19708925e-01 3.62358302e-01 -1.96972236e-01 1.54042557e-01
-3.70383382e-01 4.01332378e-01 3.46414953e-01 -4.90324914e-01
-1.08456945e+00 -6.46806240e-01 -5.03500462e-01 3.49195868e-01
1.00778985e+00 -6.87539697e-01 1.60788953e+00 -7.09987700e-01
-1.02308738e+00 5.99258304e-01 -7.51960814e-01 -6.53889954e-01
6.37939036e-01 6.23516627e-02 -6.08055174e-01 6.27120808e-02
-4.86478806e-02 -2.98673064e-01 7.36877859e-01 -1.44888306e+00
-1.23360825e+00 -2.26340994e-01 -2.91680932e-01 -5.26949406e-01
1.55220330e-01 3.75262737e-01 -8.46799091e-02 1.28412265e-02
3.40569079e-01 -6.55555487e-01 -7.97139287e-01 -5.29533923e-01
-2.57216841e-01 -5.52165687e-01 6.98501170e-01 8.57012048e-02
1.58785462e+00 -1.79083109e+00 -3.24111074e-01 1.07093692e+00
6.08941376e-01 -5.19784808e-01 2.23371640e-01 1.14966834e+00
8.46686140e-02 4.86640662e-01 -2.92225689e-01 -3.96792084e-01
1.91543758e-01 6.73850477e-01 -3.87510329e-01 1.46471903e-01
2.21280798e-01 5.95059931e-01 -1.01067066e+00 -4.00597423e-01
-2.45080926e-02 -2.69414872e-01 -2.00250849e-01 2.55526900e-01
-2.97087908e-01 2.86065608e-01 -3.65591407e-01 7.19629467e-01
4.01206911e-01 -4.76909906e-01 7.64586806e-01 8.97760749e-01
-5.32643378e-01 8.31044316e-02 -1.65184486e+00 1.10222864e+00
-2.96623409e-01 -9.07333568e-02 1.73789650e-01 -4.28695410e-01
1.01196182e+00 5.29135108e-01 8.28112066e-01 -4.76955801e-01
-4.19689059e-01 1.44131988e-01 5.97423268e-03 -3.49880427e-01
7.01644197e-02 -3.51950169e-01 -1.22382373e-01 1.26877904e+00
-1.24904387e-01 5.23021817e-01 7.75064230e-01 -5.53434789e-02
2.07189155e+00 -1.97908014e-01 9.42084610e-01 -4.34815764e-01
4.49893773e-01 7.73967728e-02 7.39421606e-01 1.17743993e+00
9.21136886e-02 1.65287182e-01 1.28479862e+00 -1.06297219e+00
-9.40500617e-01 -1.33958852e+00 2.71632850e-01 9.41684842e-01
-1.77220985e-01 -1.26999915e+00 -8.68226662e-02 -6.39984727e-01
-2.21632242e-01 3.08777034e-01 -5.64102113e-01 3.81425977e-01
-7.58470535e-01 -6.11082256e-01 4.62836683e-01 3.69272828e-01
2.29875930e-02 -1.23004568e+00 -1.04103172e+00 7.93011427e-01
4.87602234e-01 -7.87647188e-01 1.71461046e-01 6.44557238e-01
-1.32877040e+00 -1.46488535e+00 2.35817045e-01 -4.69112873e-01
3.70280415e-01 -2.17846006e-01 1.43057644e+00 3.63343954e-02
-1.74391448e-01 5.47380924e-01 -3.46364617e-01 -7.38060951e-01
-6.37226939e-01 -6.75713569e-02 -7.10013807e-02 1.11826308e-01
7.24974036e-01 -1.20796168e+00 -3.31290036e-01 4.32418346e-01
-1.03114855e+00 -2.64223784e-01 7.72353530e-01 3.10861737e-01
7.34263539e-01 3.13165277e-01 3.00658315e-01 -1.09572697e+00
8.65806937e-01 -5.69023848e-01 -2.86196202e-01 2.79336721e-01
-8.92182827e-01 1.38723448e-01 5.80141187e-01 -4.58008111e-01
-9.25276995e-01 1.91953078e-01 3.37695956e-01 -2.68736154e-01
-8.03473175e-01 4.88645941e-01 -1.55428380e-01 5.57997048e-01
5.51578641e-01 3.82998884e-01 4.78759632e-02 -3.52822810e-01
-1.84475467e-01 -1.21745534e-01 -1.94708556e-01 -7.43610203e-01
8.65691483e-01 9.44887042e-01 2.87317574e-01 -7.48019338e-01
-3.25920768e-02 -9.11274254e-01 -4.66056883e-01 -1.17270708e-01
2.68899977e-01 -3.93160760e-01 -1.12013388e+00 -2.90891603e-02
-9.04897630e-01 -4.13846165e-01 -9.72990513e-01 4.25789505e-02
-8.03697586e-01 2.47174442e-01 -8.20101798e-01 -1.26993001e+00
-1.66885644e-01 -3.50232631e-01 7.28936851e-01 -1.01497434e-01
-9.33195412e-01 -1.05151308e+00 6.14952326e-01 -4.96181339e-01
2.23344252e-01 2.79738188e-01 8.65528345e-01 -1.31785131e+00
-7.87791073e-01 -1.41413644e-01 3.21410686e-01 -3.28024864e-01
2.26645529e-01 -6.79643154e-02 -6.06710732e-01 -1.42217308e-01
3.08014870e-01 3.73594522e-01 3.06078672e-01 -1.05570130e-01
9.47191358e-01 -4.89190966e-01 -6.66858792e-01 2.19234720e-01
1.41203856e+00 4.02997047e-01 7.68230736e-01 4.80664402e-01
4.77383941e-01 6.87966168e-01 3.43424469e-01 7.48489499e-01
1.44023925e-01 9.03464109e-02 -7.56425187e-02 3.11207652e-01
6.59874618e-01 -4.10371691e-01 4.00577813e-01 3.80917192e-01
-6.27961457e-01 3.24199349e-01 -1.36135375e+00 5.76570809e-01
-2.25274611e+00 -1.58662605e+00 -7.51369953e-01 2.09164524e+00
5.62839448e-01 4.70297962e-01 5.55290222e-01 5.50843358e-01
3.90463948e-01 -1.42715901e-01 2.19467163e-01 -8.08767855e-01
3.52935284e-01 5.99620521e-01 3.18906069e-01 4.18735236e-01
-8.89830947e-01 3.65922898e-01 6.68578005e+00 5.15965402e-01
-4.98763174e-01 1.59673184e-01 1.90396100e-01 -1.38530761e-01
-4.57535118e-01 4.25001740e-01 -8.37980628e-01 1.25490734e-02
1.33280933e+00 -3.56095076e-01 5.80447055e-02 1.00968874e+00
5.93700886e-01 -1.49807706e-01 -1.83823371e+00 7.71496892e-01
-5.41197181e-01 -1.26058495e+00 1.09579504e-01 5.60521543e-01
7.16281831e-01 -3.61085474e-01 -7.99610853e-01 2.99775779e-01
5.88655293e-01 -1.25644982e+00 5.01944661e-01 6.01796925e-01
3.43687204e-03 -8.82382631e-01 4.33137327e-01 7.75361121e-01
-1.60562623e+00 -4.14103270e-01 2.13255603e-02 -6.74904943e-01
2.53911555e-01 9.58840072e-01 -1.16491580e+00 7.01744616e-01
5.95273316e-01 5.96072733e-01 -2.05494836e-01 1.13234472e+00
-4.45658388e-03 8.83321583e-01 -5.60539544e-01 6.12276532e-02
1.90140456e-01 -2.51144290e-01 7.01542020e-01 1.48314452e+00
3.51840019e-01 -4.31405336e-01 6.21360660e-01 8.23064387e-01
5.10766208e-01 4.95982869e-03 -8.07163358e-01 2.49361053e-01
3.66496921e-01 1.09376037e+00 -1.14840281e+00 -2.64816344e-01
-3.94600421e-01 5.74732959e-01 -2.36650139e-01 1.40825704e-01
-2.56123126e-01 1.74981803e-01 5.02243221e-01 7.98139930e-01
4.05952811e-01 -3.78169388e-01 -4.35968816e-01 -8.86736751e-01
1.95576161e-01 -1.09764564e+00 7.39545166e-01 1.62634164e-01
-1.40328956e+00 2.56877691e-01 3.16451907e-01 -1.03149664e+00
-5.58927894e-01 -1.67848513e-01 -1.17537904e+00 6.50990665e-01
-8.97594333e-01 -1.02691090e+00 1.25455707e-02 7.66905844e-01
5.55086851e-01 1.68174833e-01 1.06614661e+00 -3.39981735e-01
-9.11914185e-02 -1.15208037e-01 -2.64925301e-01 -2.76627779e-01
3.43699381e-02 -1.59591353e+00 5.70435822e-01 7.76005864e-01
3.38461071e-01 1.00054204e+00 7.84626126e-01 -8.87315571e-01
-1.57206368e+00 -8.79486501e-01 1.11772907e+00 -5.49732447e-01
1.01242495e+00 -3.55791658e-01 -1.04772234e+00 1.12369204e+00
1.06779985e-01 -2.89227426e-01 1.03364670e+00 6.74228489e-01
-1.50544837e-01 -4.96889465e-02 -9.57755804e-01 5.19278705e-01
1.56070018e+00 -2.79291868e-01 -7.91685283e-01 -2.54828662e-01
5.35091385e-02 4.17415142e-01 -1.02272964e+00 2.12390602e-01
4.65050012e-01 -1.20280600e+00 5.52785337e-01 -5.37617564e-01
1.26514032e-01 -4.04960096e-01 4.38095570e-01 -7.67096817e-01
-3.59901547e-01 -1.40231705e+00 -8.85045052e-01 1.04059458e+00
4.51016635e-01 -4.80940014e-01 8.74110222e-01 2.94675708e-01
3.38472724e-01 -7.67462969e-01 -9.19788599e-01 -8.87406826e-01
-4.67242420e-01 -8.19458544e-01 6.25730038e-01 8.73952866e-01
4.01557028e-01 3.74521315e-01 -9.66415852e-02 1.13034705e-02
8.87005448e-01 4.32395548e-01 9.60134327e-01 -1.75757146e+00
-7.84571409e-01 -5.77342391e-01 -2.74185956e-01 -7.94150889e-01
-3.84829938e-01 -5.88723838e-01 -3.31755102e-01 -1.65217841e+00
3.13173503e-01 -4.91166919e-01 -2.83026858e-03 3.95336598e-01
1.66276515e-01 -5.50784171e-01 -3.01389936e-02 6.01499021e-01
-9.77830589e-01 -7.75395557e-02 1.01314759e+00 2.84854949e-01
-6.32772565e-01 3.48250359e-01 -4.96087283e-01 8.48201931e-01
8.08036447e-01 -7.57250845e-01 -5.10119081e-01 5.12549818e-01
5.90178490e-01 3.34023625e-01 2.88762718e-01 -1.07661235e+00
4.88202125e-01 -4.92631882e-01 2.84640670e-01 -3.19557875e-01
-2.08446324e-01 -1.00947034e+00 8.43341351e-01 9.05883551e-01
-2.55092919e-01 3.03042978e-01 -1.61226675e-01 1.00751889e+00
-2.83594996e-01 -2.45015934e-01 1.21563487e-01 -7.25197077e-01
-7.99618721e-01 2.76957512e-01 -1.02637887e+00 -3.64160627e-01
1.27327168e+00 -6.47394180e-01 2.17868075e-01 -2.23894715e-01
-1.33271432e+00 3.14513296e-01 5.01684546e-01 1.60964187e-02
4.64194447e-01 -9.52519178e-01 -5.33794940e-01 3.13441902e-01
2.90675431e-01 2.02960774e-01 -1.17400907e-01 1.10050094e+00
-4.45813775e-01 3.92638482e-02 -2.42304236e-01 -8.36520672e-01
-1.43912530e+00 6.79044366e-01 5.19942679e-02 -1.18608975e+00
-9.19315457e-01 6.48431331e-02 -2.67341495e-01 -2.53685594e-01
4.30195667e-02 -6.71570539e-01 8.55871215e-02 6.08174205e-02
9.13086712e-01 4.89726514e-01 -7.36950412e-02 5.43112338e-01
-3.16938043e-01 -1.67049989e-02 7.76327923e-02 -2.50458241e-01
1.36935043e+00 -3.78427319e-02 -5.74905396e-01 1.14622200e+00
2.86246121e-01 3.02420408e-01 -1.09541452e+00 -7.02932701e-02
1.21089554e+00 -2.73464829e-01 -9.64956105e-01 -3.64428997e-01
-3.32484096e-01 4.32915330e-01 1.07702583e-01 1.09010124e+00
8.45308483e-01 4.07571375e-01 4.12310958e-01 6.27242982e-01
7.15288579e-01 -8.21349978e-01 -1.44643867e-02 7.91697502e-01
4.84464765e-01 -6.78757489e-01 -2.78401554e-01 -5.58247328e-01
-4.52394247e-01 1.21317184e+00 3.08924168e-01 -2.60420442e-01
7.31471479e-01 9.96811867e-01 -3.90687853e-01 -4.62902695e-01
-1.36782396e+00 -1.63336873e-01 -5.46640515e-01 7.09876359e-01
1.40421674e-01 2.76457459e-01 -3.53435248e-01 7.73184299e-01
-6.12314232e-02 1.90671146e-01 6.67443693e-01 1.78445673e+00
-7.50399053e-01 -1.56031847e+00 -5.43093383e-01 8.33737850e-01
-6.58852160e-01 1.63947150e-01 -4.43451792e-01 1.01061821e+00
8.79898816e-02 9.08105731e-01 4.93871063e-01 -3.41525376e-01
6.31356239e-01 6.95277929e-01 5.92768252e-01 -8.22430789e-01
-9.60088968e-01 2.02718392e-01 5.14315486e-01 -6.82709336e-01
-3.88567179e-01 -8.85412693e-01 -1.14984250e+00 -6.16027474e-01
2.84453362e-01 4.20706064e-01 1.16977487e-02 9.15098667e-01
5.16823307e-02 4.11102027e-01 4.17239785e-01 -2.73669332e-01
-3.26125741e-01 -8.92870307e-01 -7.20201910e-01 4.17184293e-01
-3.28148529e-02 -5.51421493e-02 -2.33401403e-01 2.75436521e-01] | [8.563179969787598, 6.049996852874756] |
a142cd74-6ef2-4c0f-a2b4-ba7284dffa98 | translate-to-disambiguate-zero-shot | 2304.13803 | null | https://arxiv.org/abs/2304.13803v1 | https://arxiv.org/pdf/2304.13803v1.pdf | Translate to Disambiguate: Zero-shot Multilingual Word Sense Disambiguation with Pretrained Language Models | Pretrained Language Models (PLMs) learn rich cross-lingual knowledge and can be finetuned to perform well on diverse tasks such as translation and multilingual word sense disambiguation (WSD). However, they often struggle at disambiguating word sense in a zero-shot setting. To better understand this contrast, we present a new study investigating how well PLMs capture cross-lingual word sense with Contextual Word-Level Translation (C-WLT), an extension of word-level translation that prompts the model to translate a given word in context. We find that as the model size increases, PLMs encode more cross-lingual word sense knowledge and better use context to improve WLT performance. Building on C-WLT, we introduce a zero-shot approach for WSD, tested on 18 languages from the XL-WSD dataset. Our method outperforms fully supervised baselines on recall for many evaluation languages without additional training or finetuning. This study presents a first step towards understanding how to best leverage the cross-lingual knowledge inside PLMs for robust zero-shot reasoning in any language. | ['Luke Zettlemoyer', 'Terra Blevins', 'Haoqiang Kang'] | 2023-04-26 | null | null | null | null | ['word-sense-disambiguation'] | ['natural-language-processing'] | [ 1.85329542e-01 -6.22459799e-02 -9.93094146e-01 -4.04020995e-01
-1.41421413e+00 -9.63690519e-01 8.65523577e-01 4.59406823e-01
-8.11577201e-01 7.26718903e-01 6.02385402e-01 -6.36402249e-01
1.66028053e-01 -6.78179085e-01 -7.44981706e-01 -9.50936452e-02
5.17539859e-01 6.21405005e-01 1.06491864e-01 -6.73484981e-01
-7.44367093e-02 -1.52447358e-01 -8.84435654e-01 5.01898170e-01
1.10676849e+00 2.47060955e-01 5.11153936e-01 1.17843844e-01
-5.45420349e-01 3.00900966e-01 -3.54212403e-01 -6.06030583e-01
8.24791119e-02 -7.46310800e-02 -9.62781489e-01 -4.28436071e-01
7.46208787e-01 3.40903878e-01 2.13237152e-01 9.83868062e-01
6.31395340e-01 2.14853585e-01 4.45295781e-01 -6.18545890e-01
-9.68181849e-01 1.20140553e+00 -3.65698189e-01 4.11378145e-01
4.12510782e-01 2.93661982e-01 1.44280446e+00 -1.10633087e+00
1.08940077e+00 1.62621403e+00 6.97421134e-01 5.79009354e-01
-1.39573574e+00 -5.25566757e-01 3.95805329e-01 1.02233164e-01
-1.22256291e+00 -5.07874310e-01 2.17643306e-01 -1.95830733e-01
1.86719227e+00 -6.99386746e-02 2.36176640e-01 1.47217834e+00
3.08455348e-01 1.02466393e+00 1.39974427e+00 -8.81928384e-01
5.40826283e-02 -7.21189901e-02 3.36251795e-01 5.02902329e-01
5.56544781e-01 -9.84281152e-02 -9.16611671e-01 -4.72825021e-02
2.38852561e-01 -4.04041946e-01 -1.14252433e-01 6.91851825e-02
-1.59675157e+00 6.47672117e-01 2.12072328e-01 7.77397752e-01
-1.68853611e-01 -6.50765523e-02 4.70970333e-01 4.92338985e-01
6.96492434e-01 1.18235350e+00 -1.00429773e+00 -1.05485938e-01
-1.01310527e+00 2.03364529e-02 6.97299361e-01 8.92253876e-01
9.93284404e-01 -1.28679991e-01 -6.23756349e-01 1.18599093e+00
1.06292650e-01 1.04837024e+00 8.33005965e-01 -6.59962058e-01
6.71499908e-01 6.03093028e-01 -8.00726563e-02 -3.69121581e-01
-2.51750082e-01 -6.31739199e-01 -1.76132858e-01 -3.69044393e-01
1.30735978e-01 -2.75011569e-01 -1.05277908e+00 2.11925459e+00
6.38458086e-03 8.66264701e-02 3.92841697e-01 7.48583019e-01
4.97842789e-01 4.50246543e-01 5.26258528e-01 -5.31082600e-02
1.70782423e+00 -8.46949041e-01 -6.71762049e-01 -1.29978120e+00
1.13235593e+00 -1.13647485e+00 1.76395464e+00 -5.93099147e-02
-8.53888869e-01 -3.26021284e-01 -1.20468009e+00 -3.68429095e-01
-7.13494837e-01 -3.58995162e-02 6.15911543e-01 3.33769619e-01
-9.69088137e-01 2.48493508e-01 -8.28443348e-01 -8.28393102e-01
2.86293030e-02 -1.98888123e-01 -2.21345067e-01 -4.30463523e-01
-2.00629520e+00 1.54869330e+00 5.09046793e-01 -5.11975586e-01
-6.31648123e-01 -9.33421731e-01 -1.07850266e+00 -2.62821496e-01
6.50297463e-01 -7.89602458e-01 1.48454630e+00 -6.58446848e-01
-8.90074492e-01 1.39488661e+00 -5.41061938e-01 -5.23090363e-01
-4.46250029e-02 -4.61378574e-01 -5.25725722e-01 -4.61339593e-01
7.22632647e-01 5.68719923e-01 4.03161466e-01 -1.02909493e+00
-7.47878015e-01 -2.77364582e-01 2.08847634e-02 3.71192306e-01
-3.43849473e-02 6.80707097e-02 -3.75866175e-01 -8.73723805e-01
-1.27350256e-01 -8.65510225e-01 -9.18257907e-02 -7.19951630e-01
-1.67823687e-01 -6.01701081e-01 3.72803211e-01 -3.04935694e-01
1.37164176e+00 -1.75171328e+00 2.86781024e-02 -2.23970607e-01
-9.73884910e-02 5.68534911e-01 -4.82353717e-01 5.92763066e-01
2.13390827e-01 3.44360918e-01 -2.64827132e-01 -5.03640592e-01
3.55718285e-01 6.62251234e-01 -5.68963885e-01 -1.44032463e-01
3.55330557e-01 1.41424739e+00 -1.41642845e+00 -3.48584950e-01
1.35041876e-02 2.31870428e-01 -2.92778850e-01 -2.37096101e-01
-5.84081113e-01 -1.27925903e-01 -3.40383828e-01 5.95899701e-01
1.69214174e-01 -2.08060458e-01 5.05887032e-01 -1.30340427e-01
3.49682309e-02 1.03048396e+00 -7.44560003e-01 2.18453574e+00
-1.02607799e+00 4.72052038e-01 -2.93199986e-01 -3.77235860e-01
6.40634537e-01 2.67216623e-01 -1.49310939e-02 -9.78769839e-01
-2.48307243e-01 6.43421769e-01 -7.10329935e-02 -4.00442988e-01
7.40929008e-01 -4.80946660e-01 -5.09114265e-01 7.87985861e-01
4.28097934e-01 -2.75954992e-01 3.65725040e-01 3.00781727e-01
9.20295417e-01 1.68873921e-01 6.43711627e-01 -6.07156277e-01
2.43646890e-01 4.02869731e-01 5.54890752e-01 8.88419688e-01
1.80434510e-02 -1.19659482e-02 -1.01231178e-02 -2.30383009e-01
-7.66472340e-01 -1.13154614e+00 -4.89156619e-02 1.81791401e+00
1.16638541e-01 -8.01376402e-01 -6.51659608e-01 -6.05422378e-01
9.24441442e-02 1.37365139e+00 -3.37884575e-01 -3.23833793e-01
-6.76026762e-01 -7.92747378e-01 9.36263442e-01 4.71158415e-01
2.21651509e-01 -8.99679184e-01 -2.62019038e-01 4.79513943e-01
-5.23880124e-01 -1.37204766e+00 -7.76888013e-01 4.58265156e-01
-5.41267157e-01 -7.34396458e-01 -2.52558887e-01 -8.98602903e-01
7.66429976e-02 3.77970338e-01 1.91359615e+00 -1.40726805e-01
6.56270655e-03 3.12497079e-01 -3.00609529e-01 -4.96050477e-01
-5.16512692e-01 6.74293816e-01 4.18192893e-01 -6.58665061e-01
1.14812303e+00 -2.51375288e-01 -1.36890873e-01 1.86003178e-01
-8.69191647e-01 -1.07467853e-01 6.32647812e-01 7.16362417e-01
7.84967840e-01 -6.63376212e-01 4.49118078e-01 -1.11322284e+00
1.16156483e+00 -4.65522408e-01 -2.35208064e-01 7.54808068e-01
-8.76967728e-01 7.28045881e-01 1.42559290e-01 -4.24024433e-01
-1.12197971e+00 -5.36196291e-01 1.16444007e-02 4.42090109e-02
1.28334522e-01 7.79151320e-01 -2.35801302e-02 2.92511314e-01
9.62374747e-01 6.52514622e-02 -6.03629053e-01 -5.25280893e-01
9.21387374e-01 5.72090507e-01 5.04143119e-01 -1.14075696e+00
5.23729503e-01 3.72619927e-02 -5.64105093e-01 -4.33238238e-01
-1.59546196e+00 -4.87339795e-01 -7.39942908e-01 3.39583665e-01
1.01484656e+00 -1.15132403e+00 4.91370335e-02 8.87265801e-02
-1.16134989e+00 -6.17239833e-01 -1.09217480e-01 4.89613116e-01
-2.89121509e-01 -3.93840671e-02 -6.44707739e-01 -3.11452270e-01
-5.24784625e-01 -1.01482117e+00 1.51632810e+00 -1.23210825e-01
-7.26994038e-01 -1.54247117e+00 3.22473854e-01 1.65225133e-01
5.37481129e-01 -2.53620267e-01 1.31939733e+00 -8.19757402e-01
-1.89415485e-01 3.65074843e-01 -1.16199385e-02 4.64523435e-02
2.53886938e-01 -5.49961329e-01 -8.64142299e-01 -2.95020550e-01
-2.32162014e-01 -6.39830232e-01 1.08372843e+00 1.39152139e-01
8.44556466e-02 -2.95758188e-01 -4.63793188e-01 4.52715516e-01
1.48457587e+00 -3.72905135e-01 2.10699178e-02 5.75446010e-01
5.99791169e-01 1.89147621e-01 8.63424003e-01 -2.01824501e-01
8.21908057e-01 5.93311250e-01 -2.26909354e-01 -8.33126754e-02
-4.98940319e-01 -5.22280276e-01 5.53473234e-01 1.09136391e+00
3.06067109e-01 -1.97755218e-01 -1.32544470e+00 7.99583852e-01
-1.69944918e+00 -5.17121553e-01 2.46364370e-01 2.09618831e+00
1.56558728e+00 3.26702476e-01 -4.55087334e-01 -5.15662909e-01
4.25059497e-01 5.40597916e-01 -4.53159779e-01 -5.58145881e-01
-5.76132357e-01 4.63360310e-01 7.11139917e-01 1.09904718e+00
-7.48237252e-01 1.94905746e+00 6.04040384e+00 1.09143329e+00
-1.30400634e+00 5.69330394e-01 1.26904666e-01 -4.30200770e-02
-7.95495331e-01 1.07052855e-01 -1.13751459e+00 8.17465484e-02
8.42824996e-01 -3.15930873e-01 4.65335190e-01 5.24656713e-01
-8.83090422e-02 -1.05560375e-02 -1.12749791e+00 7.43714631e-01
1.95738658e-01 -1.12146127e+00 2.43098393e-01 -3.53424400e-01
9.76809680e-01 7.65788257e-01 -1.29485846e-01 7.59874105e-01
8.89009953e-01 -9.36116338e-01 5.22050738e-01 2.06632465e-01
1.05410051e+00 -4.82723057e-01 5.72404504e-01 4.13671821e-01
-1.07760847e+00 4.87605274e-01 -5.17480493e-01 -2.55794711e-02
1.38787001e-01 4.80389327e-01 -1.12710607e+00 4.46836501e-01
1.28623441e-01 7.49472857e-01 -6.96105540e-01 2.18022346e-01
-7.50492811e-01 7.78887868e-01 -1.10011369e-01 -5.10508195e-02
5.88249087e-01 3.15522850e-01 7.82290816e-01 1.66348922e+00
2.86419511e-01 -1.08483888e-01 5.05447567e-01 6.83841884e-01
-2.87806302e-01 2.40333661e-01 -4.50752884e-01 -2.94341594e-01
8.49871755e-01 9.31820810e-01 -3.16728145e-01 -6.86043084e-01
-4.07376677e-01 9.62870598e-01 6.23023272e-01 4.77024138e-01
-3.14426392e-01 -4.00168914e-03 1.22187853e+00 -1.07984748e-02
-3.93596366e-02 -5.83161712e-01 -5.10961115e-01 -1.51461029e+00
-1.34836912e-01 -9.32570994e-01 5.63174605e-01 -8.06806684e-01
-1.53201652e+00 6.62697077e-01 -2.32649632e-02 -8.02450478e-01
-6.00656450e-01 -7.70191550e-01 -2.13527292e-01 1.04905033e+00
-1.87964404e+00 -1.41360116e+00 3.11250925e-01 4.06539857e-01
8.83714557e-01 -1.35850385e-01 1.11295927e+00 2.49867160e-02
-2.42485344e-01 7.35211849e-01 -1.42595261e-01 8.41647908e-02
1.41139531e+00 -1.40380943e+00 9.90895152e-01 1.09382033e+00
6.41265750e-01 1.21177495e+00 7.26654947e-01 -1.05669129e+00
-1.42798650e+00 -1.18993926e+00 1.82312953e+00 -8.74708533e-01
1.04460156e+00 -4.25764591e-01 -9.32769954e-01 9.43523049e-01
5.84153473e-01 -1.06712848e-01 8.26735675e-01 7.96887517e-01
-9.17289257e-01 9.46453363e-02 -8.08337212e-01 8.06609869e-01
1.27490389e+00 -1.06578386e+00 -1.43280053e+00 2.43607461e-01
1.24817955e+00 -3.50807846e-01 -7.72198141e-01 2.94648886e-01
2.13446587e-01 -2.64596850e-01 8.61354530e-01 -9.56693947e-01
2.07398161e-01 -2.48453200e-01 -3.90632093e-01 -2.02023911e+00
-2.50125319e-01 -6.39689028e-01 1.52797788e-01 1.06547618e+00
9.21643376e-01 -6.29143059e-01 -1.79118529e-01 2.44179577e-01
-2.75249541e-01 -6.20488167e-01 -8.55226457e-01 -9.41675305e-01
5.44807911e-01 -9.04700756e-01 7.75810897e-01 1.34317613e+00
1.85963690e-01 8.48187685e-01 2.10025278e-03 -6.36079684e-02
4.50622737e-01 1.24820113e-01 1.91721782e-01 -8.67474675e-01
-2.92109936e-01 -4.51267153e-01 -7.88725987e-02 -9.22040522e-01
5.59975445e-01 -1.49190736e+00 1.01120681e-01 -1.63248622e+00
4.77768853e-02 -3.87569457e-01 -6.17245317e-01 1.01290858e+00
-8.25057447e-01 1.78542063e-01 2.07557440e-01 3.28607634e-02
-7.45409429e-01 1.33000344e-01 1.09737551e+00 -1.66261464e-01
-5.36136068e-02 -5.77180684e-01 -8.92467439e-01 6.00554645e-01
5.04685521e-01 -5.74177146e-01 -3.06045473e-01 -1.23948300e+00
6.00884795e-01 -2.73135513e-01 -1.87455788e-01 -4.07694846e-01
2.65752614e-01 -4.04447466e-01 6.33455291e-02 -1.54825658e-01
1.29682096e-02 -2.54252374e-01 -5.25065243e-01 1.68797359e-01
-5.58966219e-01 2.74736822e-01 3.55766743e-01 1.20928034e-01
-9.81513932e-02 5.63497134e-02 5.43054998e-01 -4.91173297e-01
-1.23400080e+00 -1.46214776e-02 -2.02960789e-01 8.65980208e-01
4.36383098e-01 2.34530494e-01 -5.60162425e-01 6.00996502e-02
-4.49995458e-01 3.87814462e-01 6.54041231e-01 1.12690699e+00
6.84306696e-02 -1.53897130e+00 -8.08011532e-01 1.92880452e-01
6.54910982e-01 -3.93139333e-01 -2.98999161e-01 7.09690392e-01
2.13822171e-01 8.68127763e-01 1.17241777e-01 -4.46596175e-01
-7.65953481e-01 4.78005856e-01 3.95482630e-01 -4.51991796e-01
-2.11750284e-01 1.01677799e+00 -5.69345094e-02 -8.28082979e-01
-1.67686060e-01 -6.09570205e-01 2.81105548e-01 2.39061520e-01
5.40875494e-01 -7.57228732e-02 3.96993101e-01 -5.93795121e-01
-6.42969668e-01 5.14951408e-01 -1.35874823e-01 -3.66975576e-01
1.04879689e+00 -4.48157310e-01 -2.35535353e-01 8.51537585e-01
8.81591618e-01 3.56922001e-01 -8.27352047e-01 -9.76857543e-01
7.71936417e-01 -1.74414292e-01 7.82712996e-02 -1.37893534e+00
-3.57537270e-01 7.38943517e-01 1.70150593e-01 -4.24548954e-01
6.72845900e-01 2.53420174e-01 1.20918155e+00 6.10592127e-01
7.32637346e-01 -1.32500684e+00 -2.35903531e-01 1.25573087e+00
5.36926210e-01 -1.40906346e+00 -4.91354227e-01 2.03660652e-02
-7.90936887e-01 7.26256251e-01 5.49393296e-01 6.97853416e-02
2.45614767e-01 4.15803492e-01 6.42518282e-01 -1.08547240e-01
-1.12607706e+00 -6.44334793e-01 7.42764652e-01 3.56828749e-01
9.23647285e-01 5.09595573e-01 -5.31071246e-01 7.51848400e-01
-5.15406489e-01 -2.74349660e-01 -1.19803464e-02 8.04903626e-01
-7.15450525e-01 -1.52103686e+00 -7.54774883e-02 1.49330825e-01
-4.50156689e-01 -9.18161929e-01 -6.38494015e-01 5.53062558e-01
2.14835525e-01 9.99621153e-01 -1.41476467e-01 -2.18919903e-01
2.38333255e-01 6.47193611e-01 5.38012445e-01 -1.17545688e+00
-4.55601484e-01 -2.84011122e-02 3.72586101e-01 -6.92021668e-01
-2.12300479e-01 -4.99046355e-01 -1.27973056e+00 -5.49163818e-02
6.62332028e-02 5.07180989e-02 4.54255670e-01 1.56189919e+00
2.73818582e-01 2.81766862e-01 -6.85483515e-02 -1.54356077e-01
-5.31141996e-01 -1.13618028e+00 -1.37010202e-01 4.11888421e-01
2.57867068e-01 -5.13032138e-01 -4.19325605e-02 -1.52349710e-01] | [10.860685348510742, 9.735904693603516] |
bd967c3b-eaed-4590-bff5-6ed9ab8b963f | trans-inpainter-a-transformer-model-for-high | 2305.05385 | null | https://arxiv.org/abs/2305.05385v1 | https://arxiv.org/pdf/2305.05385v1.pdf | Trans-Inpainter: A Transformer Model for High Accuracy Image Inpainting from Channel State Information | Radio Frequency (RF) signal-based multimodal image inpainting has recently emerged as a promising paradigm to enhance the capability of distortion-free image restoration by integrating wireless and visual information from the identical physical environment and has potential applications in fields like security and surveillance systems. In this paper, we aim to implement an RF-based image inpainting system that enables image restoration in a complex environment while maintaining high robustness and accuracy. This requires accurately converting RF signals into meaningful visual information and overcoming the challenges of RF signals in complex environments, such as multipath interference, signal attenuation, and noise. To tackle this problem, we propose Trans-Inpainter, a novel image inpainting method that utilizes the Channel State Information (CSI) of WiFi signals in combination with transformer networks to generate high-quality reconstructed images. This approach is the first to use CSI for image inpainting, which allows for extracting visual information from WiFi signals to fill in missing regions in images. To further improve Trans-Inpainter's performance, we investigate the impact of variations in CSI data on RF-based imaging ability, i.e., analyzing how the location of the CSI sensors, the combination of CSI from different sensors, and changes in temporal or frequency dimensions of CSI matrix affect the imaging quality. We compare the performance of Trans-Inpainter with RF-Inpainter, the state-of-the-art technology for RF-based multimodal image inpainting, under more realistic experimental scenarios, and with single-modality image inpainting models when only RF or image data is available, respectively. The results show that Trans-Inpainter outperforms other baseline methods in all cases. | ['Mohamed Wahib', 'Jihong Park', 'Mehdi Bennis', 'Takayuki Nishio', 'Shoki Ohta', 'Cheng Chen'] | 2023-05-09 | null | null | null | null | ['image-inpainting'] | ['computer-vision'] | [ 5.61865389e-01 -3.73922765e-01 2.44909167e-01 -6.69977069e-02
-8.02617848e-01 -6.95417821e-01 2.08702937e-01 -5.58555007e-01
-1.89104512e-01 6.79742873e-01 4.39608455e-01 -2.61689454e-01
-3.33855510e-01 -6.10915124e-01 -9.00058508e-01 -1.04413414e+00
-3.79122607e-02 -3.76005918e-01 -1.60622403e-01 -1.04120426e-01
-8.78404900e-02 5.93314230e-01 -1.25055277e+00 3.03262681e-01
9.53369856e-01 1.11193812e+00 4.96074349e-01 6.65480554e-01
9.77233425e-02 6.40885949e-01 -9.53058243e-01 6.31316379e-02
5.12329757e-01 -3.48597854e-01 5.34519292e-02 -1.69278160e-01
2.46774241e-01 -6.32970452e-01 -9.42846835e-01 5.69478810e-01
7.73075879e-01 -3.24313700e-01 2.45094299e-01 -1.06709576e+00
-2.08124608e-01 3.71196717e-01 -6.48587763e-01 -1.38149234e-02
7.89417148e-01 8.54628235e-02 -2.55100846e-01 -5.81670225e-01
6.13858700e-01 1.03440499e+00 7.89976418e-01 2.34487444e-01
-1.22381806e+00 -1.04865313e+00 -4.57006037e-01 1.97055995e-01
-1.39027131e+00 -6.01326644e-01 8.80969942e-01 -2.63351612e-02
4.21878040e-01 4.63700861e-01 5.70323110e-01 9.57647145e-01
5.43651819e-01 2.60520339e-01 1.09229076e+00 -6.22234583e-01
1.33975390e-02 3.72958109e-02 -5.83199084e-01 1.66991353e-01
4.08408139e-03 3.93517047e-01 -8.53425145e-01 -3.03125560e-01
8.77098978e-01 1.99809775e-01 -8.99153769e-01 -2.08955016e-02
-1.32875562e+00 1.64527491e-01 5.18155277e-01 4.84224379e-01
-5.58544278e-01 4.00259495e-01 -3.25886518e-01 5.28498948e-01
2.71440804e-01 4.75776136e-01 -1.24531105e-01 1.37787551e-01
-1.31959474e+00 -1.08308256e-01 3.53336096e-01 7.72116959e-01
3.70262563e-01 2.95695037e-01 -5.17821312e-01 6.74173415e-01
2.58651644e-01 1.22642052e+00 8.30079615e-02 -1.06914008e+00
5.31748593e-01 -3.12681586e-01 3.51229250e-01 -1.04681468e+00
-4.23979193e-01 -7.27130771e-01 -8.86278570e-01 -5.70058711e-02
-1.60538301e-01 -5.95704377e-01 -9.57780898e-01 1.82370675e+00
1.32985219e-01 6.39177382e-01 2.09528953e-01 9.24507022e-01
8.20463836e-01 8.51684213e-01 -4.43064600e-01 -4.81472880e-01
1.08647287e+00 -4.18749809e-01 -1.00513148e+00 5.96631207e-02
-4.74327430e-02 -1.16624355e+00 5.69956303e-01 5.23971379e-01
-9.62088466e-01 -6.66934729e-01 -1.04366672e+00 5.39699256e-01
3.20891529e-01 1.25257924e-01 4.35051411e-01 9.45056677e-01
-1.14015615e+00 1.59135148e-01 -4.20229107e-01 -1.58630773e-01
1.50037929e-01 3.21982473e-01 -3.90240222e-01 -6.88914001e-01
-1.18304598e+00 6.85883164e-01 -5.15424430e-01 8.75804946e-03
-8.22454214e-01 -1.17745924e+00 -5.43657124e-01 -7.55462870e-02
2.27185756e-01 -7.82785058e-01 8.53069067e-01 -9.60846841e-01
-1.39068162e+00 -4.14615870e-02 -2.88146198e-01 -3.23927522e-01
2.30744377e-01 1.98499002e-02 -8.43869746e-01 5.18932581e-01
9.36441794e-02 3.72460663e-01 1.15318108e+00 -1.53983748e+00
-7.05470219e-02 -2.49247581e-01 -1.30670846e-01 7.73364827e-02
-1.19471606e-02 -2.06553206e-01 -4.86480981e-01 -7.97338665e-01
3.34318757e-01 -9.95505273e-01 -7.63976797e-02 4.80059274e-02
-2.98221052e-01 8.61367345e-01 1.01760924e+00 -7.35234678e-01
8.97339761e-01 -2.33866978e+00 -2.47198984e-01 4.29929525e-01
-6.27688244e-02 1.86087042e-01 -2.45186061e-01 5.74696779e-01
6.55768663e-02 -1.44597769e-01 -9.49902013e-02 -1.03041366e-01
-6.38200939e-01 1.97324872e-01 -3.07388455e-01 4.60452437e-01
-2.02029124e-01 5.89820147e-01 -5.86162865e-01 -1.04951590e-01
4.06651437e-01 9.42207336e-01 -3.12547177e-01 5.99806271e-02
4.50747281e-01 1.19906437e+00 -4.05655771e-01 7.01799273e-01
1.25819492e+00 2.78186381e-01 2.66624570e-01 -7.82679915e-01
-1.79654330e-01 -1.65220827e-01 -1.05664849e+00 1.87200320e+00
-9.66156363e-01 7.05319643e-01 5.91997683e-01 -6.49958730e-01
7.17674732e-01 7.25232124e-01 8.71261477e-01 -1.46249592e+00
-5.74143790e-02 -2.54001170e-02 -2.38176137e-01 -5.83047450e-01
2.54546732e-01 -1.53020650e-01 -1.04688108e-01 3.68320197e-01
2.94289850e-02 -9.17523578e-02 -2.38658905e-01 2.85042256e-01
1.35900342e+00 -3.55078205e-02 -4.73993808e-01 7.62966573e-02
2.25469664e-01 -2.68460393e-01 3.51530910e-01 9.54071224e-01
3.68600518e-01 8.06893647e-01 -1.96841404e-01 9.00040269e-02
-7.95697212e-01 -1.26250482e+00 -2.05136940e-01 2.06242800e-01
5.93553305e-01 -1.82504356e-01 -5.24391651e-01 -1.78956315e-01
-7.58892531e-03 7.00519562e-01 -1.84060559e-01 -1.82164893e-01
-3.25955570e-01 -7.27680147e-01 5.41715503e-01 -1.41590968e-01
7.05488443e-01 -4.57894355e-01 -3.16255063e-01 4.20410246e-01
-7.67816603e-01 -1.37046814e+00 -2.69687086e-01 -1.08473092e-01
-6.97781801e-01 -8.75525773e-01 -1.02484560e+00 -2.46848881e-01
7.58595705e-01 1.04551518e+00 8.03714633e-01 -5.20343557e-02
-4.89889532e-01 8.67977202e-01 -4.71795917e-01 -2.42635697e-01
-3.16442192e-01 -5.21968186e-01 -4.45222966e-02 3.96232456e-01
-6.08787656e-01 -8.62084866e-01 -9.38330412e-01 4.94903564e-01
-1.15333772e+00 8.43876824e-02 8.46791267e-01 5.62729597e-01
3.86858940e-01 2.55657315e-01 4.21596467e-01 -3.33340049e-01
5.06178975e-01 -4.04728383e-01 -3.56711090e-01 2.00888574e-01
-2.29999229e-01 3.01236212e-02 5.22792518e-01 -4.81022745e-01
-1.29881918e+00 7.43329972e-02 -1.69795111e-01 -3.43310863e-01
2.10936561e-01 5.31576753e-01 -2.26485893e-01 -7.31397748e-01
4.58426178e-01 4.18892980e-01 3.34536098e-02 -3.13807577e-01
3.43576103e-01 7.45712638e-01 6.78596079e-01 -2.13302493e-01
9.90412474e-01 8.63621950e-01 2.25561112e-01 -1.04952478e+00
-2.29171395e-01 -3.67450804e-01 1.44864321e-01 -6.39756083e-01
3.98655355e-01 -1.19202256e+00 -4.67489541e-01 4.68723476e-01
-1.03653944e+00 -1.54784620e-01 5.51524796e-02 8.78755987e-01
-2.56353170e-01 4.34025228e-01 -3.54633689e-01 -6.04489565e-01
-2.89424032e-01 -1.13813019e+00 9.26679015e-01 4.34021980e-01
4.43753242e-01 -3.63540053e-01 1.32663250e-01 4.92924869e-01
9.12777722e-01 3.15741092e-01 5.17180800e-01 6.76972151e-01
-7.89987206e-01 -5.19929715e-02 -8.79209936e-02 3.06073744e-02
2.67003149e-01 -6.44961715e-01 -9.49179411e-01 -4.33763266e-01
7.91806355e-02 1.72279209e-01 7.64295340e-01 7.60943234e-01
7.23224938e-01 -2.35898450e-01 -5.75025141e-01 9.65194821e-01
1.52254283e+00 2.34385729e-01 1.35527635e+00 5.63944243e-02
3.31203520e-01 8.92297551e-02 7.97234118e-01 5.60859025e-01
-1.38245691e-02 1.11251807e+00 3.95478457e-01 -5.73948383e-01
-4.25500035e-01 -1.38728231e-01 4.59539115e-01 2.27023408e-01
9.13105011e-02 -4.72253501e-01 -1.47031441e-01 2.48856664e-01
-1.46808434e+00 -9.82275724e-01 -4.13011052e-02 2.39208198e+00
4.76431102e-01 -4.90489364e-01 -3.42248350e-01 -2.28234343e-02
6.28163397e-01 -2.00475305e-01 -4.31541711e-01 -1.54662803e-02
-2.48441905e-01 4.04386550e-01 1.04707003e+00 3.92830938e-01
-5.83165348e-01 2.20572978e-01 5.86190462e+00 7.46119320e-01
-1.60384846e+00 3.35445315e-01 1.98520631e-01 -1.93187758e-01
-6.04057670e-01 -6.43693432e-02 -1.08085655e-01 6.05680108e-01
7.91796684e-01 3.76501769e-01 6.87720597e-01 -3.52222510e-02
7.65729845e-01 -5.45877457e-01 -3.67214739e-01 1.17044783e+00
2.29403913e-01 -1.28935838e+00 -1.94399223e-01 8.70366916e-02
4.94092077e-01 -1.44289866e-01 2.23083988e-01 -2.65921652e-01
-2.99752980e-01 -7.73797870e-01 6.27872229e-01 9.32757556e-01
1.01798594e+00 -8.10224295e-01 6.33647621e-01 1.17025658e-01
-1.12831545e+00 -1.30656108e-01 -7.51629919e-02 3.85618269e-01
4.82241064e-01 1.19504178e+00 -7.46900439e-01 1.01082456e+00
8.91107559e-01 1.71119347e-01 -1.52403578e-01 1.18165457e+00
-1.23303719e-01 6.54103458e-01 -5.71689844e-01 6.82591498e-01
-2.47868508e-01 1.09971948e-02 6.76057696e-01 8.48936021e-01
9.34545815e-01 2.72372216e-01 -1.40452951e-01 7.61808157e-01
1.54576138e-01 -4.55786854e-01 -6.42361522e-01 3.71747583e-01
6.50027156e-01 1.19956756e+00 -3.18948060e-01 2.17342466e-01
-3.36846739e-01 9.88166511e-01 -7.87095666e-01 7.25295126e-01
-1.05689764e+00 -4.26726848e-01 5.27447581e-01 5.11395276e-01
2.87692070e-01 -5.72971046e-01 2.25511059e-01 -9.18523073e-01
1.88476607e-01 -6.47832811e-01 -2.59655267e-01 -1.29561019e+00
-5.93249857e-01 5.86679637e-01 -1.08679786e-01 -1.56431973e+00
-1.45129502e-01 1.67044982e-01 -3.75042379e-01 7.53550887e-01
-1.80850828e+00 -1.41417634e+00 -6.09155834e-01 9.36013281e-01
-4.07077707e-02 1.02428518e-01 6.57948256e-01 7.60428369e-01
-1.93889569e-02 6.94524586e-01 6.90190971e-01 -2.62653261e-01
1.05565369e+00 -3.10905457e-01 -1.27795234e-01 9.10719752e-01
5.43923080e-02 4.23813671e-01 8.62159550e-01 -6.38510108e-01
-2.09418726e+00 -9.70905244e-01 1.70478255e-01 3.72660130e-01
-1.97234973e-01 -3.47852647e-01 -2.30510935e-01 5.35721369e-02
1.56252265e-01 8.61334056e-02 5.79003930e-01 -6.30277634e-01
-1.38518214e-01 -6.77446306e-01 -1.67555034e+00 3.13900441e-01
7.88770497e-01 -3.90900731e-01 1.86070114e-01 1.84418663e-01
6.48391366e-01 -3.32656056e-01 -7.09498167e-01 4.72755104e-01
8.63352180e-01 -9.13950324e-01 1.39469099e+00 5.77637494e-01
-2.11887315e-01 -7.55587399e-01 -3.29344749e-01 -1.32937932e+00
-2.15020087e-02 -1.06174171e+00 3.02054852e-01 1.28984427e+00
3.75648171e-01 -6.99879169e-01 5.34812570e-01 1.63560539e-01
1.36724459e-02 -1.30015224e-01 -1.01210582e+00 -6.79151297e-01
-7.57636368e-01 -5.27748406e-01 5.31417608e-01 5.21129310e-01
-2.91071624e-01 -2.76265681e-01 -8.80892396e-01 4.06175822e-01
8.06702197e-01 -6.23407289e-02 8.73913229e-01 -7.82983422e-01
-5.68253279e-01 6.45281494e-01 -2.89043367e-01 -1.11091757e+00
-2.70304292e-01 -4.72384065e-01 1.57672018e-02 -1.62393510e+00
-2.67158505e-02 -8.28274965e-01 -1.13620140e-01 2.03138977e-01
3.19819719e-01 6.91166937e-01 4.44475681e-01 1.91200659e-01
-2.54138798e-01 5.13574421e-01 1.21900666e+00 -2.83349574e-01
-2.13628054e-01 1.54794604e-01 -6.97490871e-01 1.23525709e-01
5.49727798e-01 -6.02364242e-01 -4.89192188e-01 -5.00278533e-01
1.16998248e-01 7.86884606e-01 5.23156047e-01 -1.46193898e+00
3.43508124e-01 9.96264964e-02 7.31886506e-01 -3.35462749e-01
5.91214001e-01 -1.27196229e+00 8.80895972e-01 4.30692375e-01
1.07893348e-01 -1.64898664e-01 4.60122496e-01 6.60146952e-01
-1.85376391e-01 1.72539279e-01 6.21785939e-01 4.32066992e-02
-2.24222019e-01 -5.37511036e-02 -6.40341520e-01 -3.76247138e-01
6.73583865e-01 -3.03188026e-01 -3.48800242e-01 -1.12923193e+00
-4.03841883e-01 -2.39767238e-01 4.22806859e-01 2.19594344e-01
8.87552440e-01 -1.23721075e+00 -5.93909919e-01 3.81744534e-01
-2.42038235e-01 -6.20285928e-01 8.15964818e-01 1.09461725e+00
-2.63592482e-01 2.75640815e-01 -1.99971542e-01 -7.25993633e-01
-1.37561536e+00 2.23576486e-01 2.40477577e-01 -7.86996931e-02
-4.80959266e-01 3.44632268e-01 7.75793865e-02 -1.28539771e-01
3.88247147e-02 -2.13511065e-02 8.44963491e-02 -5.04588127e-01
6.76705599e-01 6.88457638e-02 3.28163296e-01 -4.13466990e-01
-3.20927352e-01 9.52795327e-01 3.78684849e-01 -4.70705777e-01
1.18903255e+00 -5.22399247e-01 1.50967449e-01 -3.01943511e-01
9.46200371e-01 5.40844083e-01 -1.18074167e+00 -7.57008865e-02
-7.90514410e-01 -7.21887112e-01 4.96134669e-01 -1.23087478e+00
-1.44168091e+00 4.60242003e-01 1.17175746e+00 -9.03712884e-02
1.71438396e+00 -1.37899652e-01 9.93819654e-01 -8.99441987e-02
8.42679918e-01 -5.63058615e-01 -3.82831949e-03 -1.71178848e-01
8.60408127e-01 -5.51599026e-01 1.72182620e-01 -5.22206724e-01
-2.77210474e-01 9.55004513e-01 -1.72577351e-01 3.02455723e-01
6.92751169e-01 7.17327535e-01 2.34200284e-01 2.38798484e-01
-1.74023956e-01 9.46510285e-02 -5.20722680e-02 1.06171441e+00
1.11483291e-01 -1.30893782e-01 -6.94765896e-02 4.51437563e-01
1.03391804e-01 2.62800813e-01 6.33963704e-01 1.05412722e+00
-1.21521086e-01 -1.34975815e+00 -9.83417451e-01 3.13309193e-01
-5.37229538e-01 -4.95221801e-02 -3.59073840e-02 3.98210049e-01
3.31124246e-01 1.54637384e+00 -3.98782700e-01 -5.58025002e-01
3.77914965e-01 -7.02150941e-01 8.12436998e-01 1.41194806e-01
-4.59829897e-01 3.44485313e-01 8.62447619e-02 -7.20937371e-01
-6.25607669e-01 -4.24455971e-01 -1.11149037e+00 -3.36250365e-01
-4.46897626e-01 9.26771313e-02 1.26421952e+00 8.16446543e-01
9.84206319e-01 7.39481032e-01 1.01956820e+00 -1.04998219e+00
1.39327511e-01 -4.96687233e-01 -6.86653793e-01 -5.96175976e-02
4.83044356e-01 -6.76737368e-01 -4.17287856e-01 -3.04010898e-01] | [10.551823616027832, -2.6252968311309814] |
5bd9e196-d8be-4a09-b1c4-ad20145d7126 | tofg-a-unified-and-fine-grained-environment | 2305.20068 | null | https://arxiv.org/abs/2305.20068v1 | https://arxiv.org/pdf/2305.20068v1.pdf | TOFG: A Unified and Fine-Grained Environment Representation in Autonomous Driving | In autonomous driving, an accurate understanding of environment, e.g., the vehicle-to-vehicle and vehicle-to-lane interactions, plays a critical role in many driving tasks such as trajectory prediction and motion planning. Environment information comes from high-definition (HD) map and historical trajectories of vehicles. Due to the heterogeneity of the map data and trajectory data, many data-driven models for trajectory prediction and motion planning extract vehicle-to-vehicle and vehicle-to-lane interactions in a separate and sequential manner. However, such a manner may capture biased interpretation of interactions, causing lower prediction and planning accuracy. Moreover, separate extraction leads to a complicated model structure and hence the overall efficiency and scalability are sacrificed. To address the above issues, we propose an environment representation, Temporal Occupancy Flow Graph (TOFG). Specifically, the occupancy flow-based representation unifies the map information and vehicle trajectories into a homogeneous data format and enables a consistent prediction. The temporal dependencies among vehicles can help capture the change of occupancy flow timely to further promote model performance. To demonstrate that TOFG is capable of simplifying the model architecture, we incorporate TOFG with a simple graph attention (GAT) based neural network and propose TOFG-GAT, which can be used for both trajectory prediction and motion planning. Experiment results show that TOFG-GAT achieves better or competitive performance than all the SOTA baselines with less training time. | ['JianPing Wang', 'Xinhong Chen', 'Yifan Zhang', 'Zihao Wen'] | 2023-05-31 | null | null | null | null | ['trajectory-prediction', 'graph-attention', 'motion-planning'] | ['computer-vision', 'graphs', 'robots'] | [-1.27634317e-01 4.87532876e-02 -7.02111661e-01 -6.53574467e-01
-4.65309709e-01 -2.47452214e-01 7.23699033e-01 9.72000957e-02
-1.82121158e-01 4.85768437e-01 4.34381008e-01 -7.92383373e-01
-1.65361747e-01 -1.04826665e+00 -8.11057031e-01 -4.85728145e-01
-1.61278665e-01 4.15599942e-01 6.64309859e-01 -3.70092690e-01
5.44782132e-02 5.05956650e-01 -1.57710481e+00 5.80485016e-02
9.55346704e-01 7.46845603e-01 5.23657918e-01 4.55664873e-01
-3.02641153e-01 8.15951943e-01 1.47102177e-01 -1.67265490e-01
9.91318226e-02 1.35901555e-01 -4.70992565e-01 -2.07435265e-01
1.40175819e-01 -6.65564477e-01 -9.77055967e-01 7.11182714e-01
-2.55334936e-02 3.49347681e-01 4.99826312e-01 -1.81546772e+00
-2.24801809e-01 5.57658434e-01 -2.58189648e-01 3.00544709e-01
-5.70834540e-02 3.80619138e-01 7.43445635e-01 -9.48142529e-01
5.15972197e-01 1.23385119e+00 4.88445848e-01 3.15173864e-01
-9.32631135e-01 -7.50691593e-01 7.05659449e-01 6.90061212e-01
-1.46524549e+00 -5.66677928e-01 8.06521952e-01 -4.66142178e-01
1.24366152e+00 2.90583074e-01 6.57460928e-01 8.80467355e-01
5.37664115e-01 9.29914832e-01 3.84072840e-01 3.02586347e-01
2.13355701e-02 -1.28455564e-01 3.92643094e-01 4.64291602e-01
1.17773078e-01 4.18274611e-01 -3.98157537e-01 3.59081417e-01
2.47748137e-01 2.54477531e-01 -2.94538494e-02 -2.81137377e-01
-1.01271403e+00 6.66953981e-01 5.71098566e-01 -1.44890293e-01
-4.36059177e-01 4.48018402e-01 3.20038557e-01 -9.71428379e-02
3.72515351e-01 -1.59520432e-01 -1.57559112e-01 -3.00206125e-01
-7.79630542e-01 5.97103417e-01 4.11258489e-01 1.43452370e+00
1.07982135e+00 2.14372888e-01 -3.42191160e-01 2.66227096e-01
3.81882191e-01 6.93627536e-01 4.30706851e-02 -7.73499072e-01
8.89286518e-01 6.00158989e-01 1.05558261e-01 -1.28767037e+00
-5.89161396e-01 -1.75781012e-01 -7.14818954e-01 -2.30421588e-01
8.39267001e-02 8.65672305e-02 -1.14833200e+00 1.83814156e+00
2.99241275e-01 6.98352158e-01 -6.30665477e-03 8.87034893e-01
8.56898487e-01 9.13017571e-01 2.76030838e-01 7.37210736e-02
1.05614483e+00 -1.17775249e+00 -9.41994250e-01 -5.85321188e-01
9.50919092e-01 -1.33470803e-01 7.28650928e-01 -3.14337701e-01
-7.64999866e-01 -6.79746211e-01 -9.72554386e-01 -2.22646728e-01
-6.83542788e-01 -2.79516816e-01 4.55613345e-01 1.64748400e-01
-9.39974368e-01 3.71386737e-01 -1.06238830e+00 -2.28792682e-01
4.13983494e-01 3.30433279e-01 -2.78366715e-01 -3.83608341e-01
-1.37181973e+00 9.23148870e-01 5.45101166e-01 2.57829368e-01
-8.69010925e-01 -7.61972904e-01 -1.27634108e+00 1.80328548e-01
4.01844025e-01 -3.95638555e-01 1.11177313e+00 -9.28293765e-02
-1.00727248e+00 1.56807136e-02 -7.78574467e-01 -5.94339430e-01
4.14145976e-01 -9.64953527e-02 -7.61565924e-01 -3.19211394e-01
1.74426705e-01 8.25133741e-01 4.09620106e-01 -1.18294799e+00
-1.13259435e+00 -2.54360288e-01 -1.22876607e-01 1.49151057e-01
2.58068722e-02 -3.68464381e-01 -1.00838161e+00 -5.98521680e-02
1.19066484e-01 -1.03926837e+00 -3.52861106e-01 -5.47059894e-01
-3.72356147e-01 -4.08849061e-01 1.25647998e+00 -6.13438189e-01
1.76366568e+00 -2.22363281e+00 -2.23945752e-01 1.68161914e-01
4.20707613e-01 2.94857651e-01 -3.11479867e-01 5.81012189e-01
3.05401951e-01 -9.77378245e-03 -6.21080250e-02 -4.52000737e-01
1.69386253e-01 6.51034713e-01 -5.55305660e-01 3.81680489e-01
2.91207165e-01 1.44555366e+00 -1.15341055e+00 -4.69202131e-01
5.64451873e-01 4.98883277e-01 -5.22115290e-01 9.02659446e-02
-2.42825106e-01 4.59534705e-01 -7.21557736e-01 3.57133031e-01
7.09056020e-01 -2.01444188e-03 8.59308839e-02 -8.73866603e-02
-4.08167124e-01 6.72642708e-01 -7.37769127e-01 1.26616943e+00
-3.24618995e-01 1.00296271e+00 -3.70925844e-01 -7.52423942e-01
8.51696551e-01 3.12241074e-02 4.53437686e-01 -1.24595010e+00
-6.71210140e-02 -3.60536985e-02 1.56660769e-02 -4.54308122e-01
1.06598127e+00 4.70921069e-01 -2.13721380e-01 1.05802625e-01
-4.59514827e-01 3.74363273e-01 2.03861460e-01 3.47278237e-01
9.62555230e-01 1.42228872e-01 -1.11746211e-02 -1.19218551e-01
2.52158403e-01 3.84440005e-01 9.40785050e-01 4.66226280e-01
-5.13056159e-01 2.28658855e-01 2.83546090e-01 -6.77008867e-01
-1.11179340e+00 -8.09978306e-01 1.54522834e-02 9.26782489e-01
7.17760921e-01 -5.82139313e-01 -5.23515224e-01 -5.23199022e-01
2.89441139e-01 1.15518773e+00 -4.44450349e-01 -4.31076407e-01
-1.03432143e+00 -3.49483728e-01 2.49405950e-01 8.04592550e-01
3.98389816e-01 -7.19522238e-01 -3.54312778e-01 4.85144734e-01
-3.04355979e-01 -1.33011079e+00 -6.34484172e-01 -1.33873716e-01
-5.28161407e-01 -8.47131789e-01 1.51256800e-01 -4.87538189e-01
4.57112104e-01 8.83052588e-01 1.06665742e+00 -1.44657986e-02
4.41864133e-01 -3.68218347e-02 -2.42529333e-01 -5.23328722e-01
-3.82774323e-01 3.09044659e-01 5.11183105e-02 -3.41401659e-02
8.38930905e-01 -5.79173028e-01 -6.26007378e-01 5.23621142e-01
-5.57137609e-01 5.86992562e-01 3.75300199e-01 4.61205035e-01
6.64264679e-01 8.07660595e-02 6.46023214e-01 -6.62351072e-01
2.25115150e-01 -9.30628657e-01 -6.07910633e-01 -1.59292430e-01
-8.83089721e-01 -1.31210670e-01 6.69541538e-01 -5.21294661e-02
-8.90446186e-01 1.16908282e-01 -1.68300480e-01 -6.29130006e-01
-1.17722452e-01 7.41389394e-01 -3.73690397e-01 3.08555722e-01
1.22572564e-01 5.13680160e-01 -1.11776933e-01 -2.36498684e-01
4.70184445e-01 6.01803124e-01 6.82421744e-01 -9.81286317e-02
8.15292597e-01 3.81159753e-01 1.17053920e-02 -5.95889866e-01
-4.38275099e-01 -6.57366097e-01 -4.76079255e-01 -4.19849157e-01
8.34945440e-01 -1.07192469e+00 -6.58529162e-01 2.13168785e-01
-1.19604516e+00 -5.08246660e-01 3.49235758e-02 5.45269132e-01
-4.91795778e-01 8.11351463e-02 -3.98757011e-01 -8.96681488e-01
5.57806715e-02 -1.41137683e+00 1.17000818e+00 1.17459938e-01
1.59020033e-02 -8.90868723e-01 -1.38709143e-01 9.42792147e-02
3.62699717e-01 1.35078371e-01 9.80360508e-01 -5.19944012e-01
-1.13675010e+00 -2.48012871e-01 -5.05171657e-01 -2.96096683e-01
6.50953203e-02 -1.37521058e-01 -8.36444557e-01 -2.04850808e-01
-4.49203283e-01 4.05518711e-01 9.02465165e-01 3.79636079e-01
1.09961081e+00 -3.09053153e-01 -9.15724814e-01 7.04518199e-01
1.18498671e+00 4.38984960e-01 8.09233904e-01 2.85611898e-01
1.32416022e+00 8.04257393e-01 8.69340718e-01 1.00753509e-01
1.29069448e+00 9.29185808e-01 7.10709870e-01 1.34769231e-02
-2.90328711e-01 -7.31186569e-01 3.75480384e-01 9.75774050e-01
1.88439772e-01 -4.50964630e-01 -1.18515944e+00 1.01555777e+00
-2.41027951e+00 -1.11590683e+00 -5.34339428e-01 1.99002516e+00
1.09798461e-01 2.00939775e-01 1.64212421e-01 -4.49979268e-02
6.03195846e-01 4.32950765e-01 -7.37671316e-01 -3.26830834e-01
1.78037480e-01 -5.96108556e-01 1.05538440e+00 6.79829419e-01
-9.34474170e-01 1.30200982e+00 5.73069572e+00 8.19399476e-01
-1.16720676e+00 1.06046952e-01 4.07270432e-01 -1.04939885e-01
-5.19184053e-01 -3.64278890e-02 -1.25675988e+00 7.83282697e-01
1.36442316e+00 -3.95485044e-01 4.04550672e-01 7.93600261e-01
7.13349104e-01 1.63188323e-01 -1.23409772e+00 7.79692233e-01
-4.79612350e-01 -1.64352787e+00 1.23527795e-01 3.15801591e-01
4.61038619e-01 5.65094233e-01 -2.04894934e-02 6.61889911e-01
2.82202721e-01 -1.11494792e+00 8.35573316e-01 5.19736052e-01
6.01923704e-01 -8.52404237e-01 6.20470107e-01 6.02293015e-01
-1.81040096e+00 -1.76159248e-01 -2.05798790e-01 -1.03632778e-01
7.55770683e-01 2.46421993e-01 -1.02915859e+00 8.08780432e-01
4.51280773e-01 1.04752338e+00 -2.58940965e-01 7.82178283e-01
1.35777369e-01 6.63314044e-01 -2.42178217e-01 1.59477845e-01
7.07632244e-01 -2.99916089e-01 6.70022488e-01 1.22201085e+00
3.89496595e-01 2.93153767e-02 4.90485162e-01 8.90231073e-01
2.37080291e-01 -2.48629689e-01 -1.02809215e+00 5.20290919e-02
8.74066949e-01 9.84914303e-01 -3.23058724e-01 -3.05032402e-01
-6.57586336e-01 2.50927657e-01 3.84538442e-01 6.89700842e-01
-1.08068275e+00 -1.64462551e-01 1.21472538e+00 4.86799598e-01
4.67516810e-01 -5.61690986e-01 -1.55745849e-01 -6.74356580e-01
5.29398099e-02 -2.74615616e-01 3.42962034e-02 -5.25087476e-01
-6.65642738e-01 6.72294915e-01 2.40859717e-01 -1.32141733e+00
-4.58541542e-01 -2.42808089e-01 -7.26528645e-01 9.65197086e-01
-2.07783055e+00 -1.33162439e+00 -4.66647893e-01 3.20353240e-01
6.64782047e-01 1.31776603e-02 2.26920158e-01 5.48851311e-01
-8.69655013e-01 4.89042699e-01 -3.71541716e-02 4.65418771e-02
2.76742339e-01 -9.01544750e-01 1.19524086e+00 1.10134673e+00
-1.22344412e-01 3.12665820e-01 4.99446213e-01 -8.74641538e-01
-1.77510965e+00 -2.09296989e+00 1.09647059e+00 -6.80729866e-01
5.13443053e-01 -2.68279999e-01 -1.13001215e+00 9.19522405e-01
-2.63555050e-01 2.46382095e-02 2.08046317e-01 -9.37005803e-02
-8.98193493e-02 -2.29581028e-01 -6.14984810e-01 8.39667678e-01
1.33864260e+00 -5.67064106e-01 -2.36871064e-01 2.38361843e-02
1.06789708e+00 -5.57436347e-01 -5.67633390e-01 5.99525928e-01
4.48371053e-01 -8.13330472e-01 8.03672075e-01 -6.54246509e-01
2.81979948e-01 -5.93229651e-01 -2.68504024e-01 -1.13366354e+00
-7.08400488e-01 -4.02713567e-01 -6.01492405e-01 1.06805146e+00
4.24247414e-01 -7.59208083e-01 7.38170028e-01 8.89625609e-01
-7.79551089e-01 -9.85643327e-01 -9.16669905e-01 -8.39632988e-01
-1.52716590e-02 -1.05096602e+00 1.37882888e+00 6.59935296e-01
-9.79432836e-02 2.94513613e-01 -4.99049217e-01 5.73587060e-01
3.89385611e-01 4.49053943e-02 1.01464295e+00 -1.03490806e+00
4.05456185e-01 -5.10011792e-01 -5.64900339e-01 -1.46768701e+00
3.43724966e-01 -9.37047124e-01 3.53755981e-01 -1.78371227e+00
-4.84624383e-04 -6.10035658e-01 -3.15312892e-01 3.25222909e-01
-2.79027313e-01 -1.66484132e-01 1.10791363e-01 2.88805097e-01
-6.02613032e-01 6.64624035e-01 1.12003756e+00 -2.29035139e-01
-4.13249224e-01 -2.35755712e-01 -3.98836583e-01 3.84128332e-01
7.31953382e-01 -4.16742742e-01 -8.20057869e-01 -6.36442482e-01
-4.23839465e-02 1.57718167e-01 2.53119081e-01 -7.87833631e-01
5.13631225e-01 -4.45033193e-01 -1.76961780e-01 -1.24173117e+00
4.12800074e-01 -7.86230683e-01 3.78693521e-01 2.42542118e-01
4.34389897e-02 2.67874360e-01 3.58478397e-01 9.53476965e-01
-3.36318403e-01 4.07844216e-01 2.06539959e-01 4.34606820e-01
-1.41333115e+00 1.03076899e+00 -4.48431194e-01 -3.72178406e-01
1.19876075e+00 -5.97857058e-01 -5.42115510e-01 -4.82434571e-01
-1.71461582e-01 1.00235128e+00 4.32123154e-01 8.83932650e-01
5.53891301e-01 -1.58826327e+00 -6.00120425e-01 3.91813815e-01
4.41841453e-01 3.39868158e-01 5.40104926e-01 1.05550897e+00
-2.11779550e-01 9.19098020e-01 5.35161272e-02 -7.87702799e-01
-8.93128932e-01 6.44384086e-01 1.35401145e-01 -1.43955618e-01
-1.02167666e+00 2.32063055e-01 6.10643327e-01 -2.94028372e-01
1.48243522e-02 -5.54559052e-01 -2.88609028e-01 -2.93722212e-01
6.43782675e-01 4.81962949e-01 3.06249578e-02 -1.22371554e+00
-3.50877345e-01 2.32695848e-01 -7.29048029e-02 2.21339375e-01
1.01962221e+00 -5.51614761e-01 3.98677647e-01 3.62260252e-01
1.11093974e+00 -3.34737599e-01 -1.62306082e+00 -2.20455527e-01
-2.74602463e-03 -5.87732136e-01 4.11279768e-01 -2.59388030e-01
-1.07196856e+00 8.88891757e-01 2.47185901e-01 2.78002650e-01
6.72025859e-01 -1.99202836e-01 1.37559295e+00 1.59502640e-01
5.70918620e-01 -9.30023909e-01 -6.64982021e-01 7.29900479e-01
6.42433465e-01 -1.33253658e+00 -3.37851644e-01 -6.62720442e-01
-6.95258558e-01 6.57933414e-01 8.04641068e-01 2.45083526e-01
5.85746527e-01 8.77338126e-02 -1.05646163e-01 -1.37045249e-01
-9.83403862e-01 -3.87430549e-01 5.39946437e-01 7.02445388e-01
-1.24214359e-01 3.62986654e-01 2.00717613e-01 6.39432907e-01
-2.20733762e-01 -3.69967185e-02 2.90344775e-01 7.01119602e-01
-4.91055489e-01 -8.94768775e-01 9.71839055e-02 5.45632184e-01
2.45720714e-01 6.59050196e-02 1.59377232e-01 8.71475816e-01
-3.13391001e-03 1.07989633e+00 4.91470605e-01 -8.44640255e-01
3.60322148e-01 -5.83856134e-03 -1.40262187e-01 -4.01135176e-01
-1.05778903e-01 -2.77029335e-01 3.97709191e-01 -1.02034044e+00
1.35776168e-02 -4.80355620e-01 -1.59110701e+00 -8.07139993e-01
-6.48573488e-02 2.49354672e-02 6.24435246e-01 1.09902751e+00
9.68260527e-01 8.48758519e-01 5.85754156e-01 -1.05904329e+00
-1.68368798e-02 -6.73974633e-01 -3.85572970e-01 3.03805321e-01
6.36613727e-01 -8.70117247e-01 1.73064739e-01 -2.52439141e-01] | [5.963662147521973, 1.0061230659484863] |
808155df-4394-4763-9db0-f8dbaec0c58f | stance-classification-for-rumour-analysis-in | 1901.01911 | null | http://arxiv.org/abs/1901.01911v1 | http://arxiv.org/pdf/1901.01911v1.pdf | Stance Classification for Rumour Analysis in Twitter: Exploiting Affective Information and Conversation Structure | Analysing how people react to rumours associated with news in social media is
an important task to prevent the spreading of misinformation, which is nowadays
widely recognized as a dangerous tendency. In social media conversations, users
show different stances and attitudes towards rumourous stories. Some users take
a definite stance, supporting or denying the rumour at issue, while others just
comment it, or ask for additional evidence related to the veracity of the
rumour. On this line, a new shared task has been proposed at SemEval-2017 (Task
8, SubTask A), which is focused on rumour stance classification in English
tweets. The goal is predicting user stance towards emerging rumours in Twitter,
in terms of supporting, denying, querying, or commenting the original rumour,
looking at the conversation threads originated by the rumour. This paper
describes a new approach to this task, where the use of conversation-based and
affective-based features, covering different facets of affect, has been
explored. Our classification model outperforms the best-performing systems for
stance classification at SemEval-2017 Task 8, showing the effectiveness of the
feature set proposed. | ['Viviana Patti', 'Endang Wahyu Pamungkas', 'Valerio Basile'] | 2019-01-07 | null | null | null | null | ['rumour-detection'] | ['natural-language-processing'] | [-2.39091739e-01 4.43530440e-01 -4.26646650e-01 -2.90621668e-01
-2.08353624e-01 -2.37979531e-01 1.21012735e+00 7.05124319e-01
-1.98136851e-01 7.46040761e-01 9.57506061e-01 -5.57330921e-02
3.57440710e-01 -7.75679171e-01 -1.65044948e-01 -5.04882157e-01
1.95833251e-01 4.54228550e-01 1.67659208e-01 -8.54155183e-01
6.71570837e-01 -1.81647688e-02 -1.48592067e+00 9.41697717e-01
2.72305042e-01 8.91786218e-01 -3.99929941e-01 3.17584634e-01
-3.11733752e-01 1.69935238e+00 -9.39109743e-01 -5.04879296e-01
-5.79716325e-01 -7.02185988e-01 -1.19275677e+00 2.04083845e-01
2.04758152e-01 3.01143639e-02 1.23947017e-01 8.21483016e-01
2.96008587e-01 -1.62287578e-01 5.79876184e-01 -1.08290005e+00
-2.35994369e-01 8.61429930e-01 -2.66028553e-01 5.91227412e-01
8.72851729e-01 -3.52933794e-01 9.47408497e-01 -7.76584566e-01
8.46566200e-01 1.29353464e+00 7.72785008e-01 4.73555803e-01
-1.08906829e+00 -4.96080756e-01 -5.96752428e-02 3.75732481e-01
-6.07277155e-01 -4.30912852e-01 8.85396838e-01 -9.67875421e-01
5.13207972e-01 6.20585084e-01 7.84994721e-01 1.62223661e+00
3.36236984e-01 6.40511632e-01 1.50833476e+00 -3.84688266e-02
1.89186350e-01 8.15966010e-01 6.54516041e-01 1.52517930e-01
-2.70406995e-02 -4.92938221e-01 -8.18579316e-01 -5.37325084e-01
-3.76474321e-01 -2.62826085e-01 -3.59240353e-01 2.01192662e-01
-9.42246675e-01 1.33705842e+00 5.80838501e-01 6.58129632e-01
-8.47051919e-01 -6.96055293e-01 9.48336601e-01 8.80737960e-01
1.53946567e+00 5.25511086e-01 -2.34597828e-02 -1.55549318e-01
-9.67401087e-01 5.10205567e-01 1.40287077e+00 3.94507423e-02
5.06761909e-01 -2.64041156e-01 -1.49346888e-01 1.04922438e+00
-2.06315201e-02 1.61156476e-01 6.31568789e-01 -2.74340570e-01
2.71412283e-01 7.49864757e-01 2.76278079e-01 -1.52359486e+00
-8.17463517e-01 -5.62725782e-01 -6.82380438e-01 2.86709145e-02
1.50115445e-01 -3.43623966e-01 1.44766703e-01 1.39995897e+00
4.44574893e-01 -2.06400469e-01 1.69827733e-02 8.76118720e-01
1.12394917e+00 8.05172741e-01 -3.52661133e-01 -8.11132789e-01
1.66329169e+00 -6.18286371e-01 -1.12496269e+00 -2.26124153e-01
5.65204680e-01 -9.69952047e-01 4.75269765e-01 6.31931484e-01
-1.08857274e+00 -9.44370963e-03 -1.05172586e+00 1.88982889e-01
-4.18227881e-01 -4.94308531e-01 1.02007836e-01 4.76499617e-01
-6.69135809e-01 5.63344717e-01 1.47459328e-01 -4.25733536e-01
1.75455689e-01 -3.94490093e-01 -9.42876190e-02 2.16601059e-01
-1.78389573e+00 1.18595672e+00 2.71293759e-01 -7.02535063e-02
-3.70625913e-01 -4.14941639e-01 -4.59694207e-01 -3.74330819e-01
2.92405099e-01 -3.18807155e-01 1.26917839e+00 -1.26910734e+00
-1.47492850e+00 1.33253300e+00 -9.05812457e-02 -9.23551381e-01
8.20968509e-01 -2.09157124e-01 -6.90610409e-01 -3.81968804e-02
3.03338379e-01 -2.32959092e-01 1.20514822e+00 -8.52562487e-01
-4.11831349e-01 -4.20671344e-01 -7.04685077e-02 1.10267764e-02
-2.94814736e-01 6.49680972e-01 6.67766273e-01 -3.59394342e-01
9.21583995e-02 -9.73560512e-01 4.33315396e-01 -9.50439990e-01
-5.01728475e-01 -5.00484169e-01 9.30262446e-01 -6.49872720e-01
1.37420464e+00 -1.86986578e+00 3.40078808e-02 1.33522470e-02
4.23939139e-01 2.86932677e-01 7.38864005e-01 8.46286714e-01
5.54190166e-02 9.12663266e-02 1.55269936e-01 -1.08591758e-01
-9.80079100e-02 -5.76745346e-02 -7.70035803e-01 7.53586531e-01
-2.51556188e-01 4.11450326e-01 -9.65846837e-01 -1.59715518e-01
-2.79262990e-01 1.93098307e-01 -1.71567917e-01 2.00364575e-01
-2.84424871e-01 4.22792673e-01 -2.62575626e-01 2.90135816e-02
4.96071190e-01 -2.12019578e-01 1.27797440e-01 -5.91192432e-02
-5.02844572e-01 1.07153249e+00 -4.85271007e-01 3.38399619e-01
-4.19505626e-01 1.05629122e+00 2.07718730e-01 -7.47207761e-01
1.13146949e+00 7.04207659e-01 1.89436391e-01 -4.24279541e-01
4.25300330e-01 3.24454725e-01 -2.12232605e-01 -5.92014909e-01
5.51071405e-01 -6.45865977e-01 -2.87469000e-01 9.43070829e-01
-4.15193290e-01 -1.53295502e-01 -5.34167401e-02 3.09951335e-01
6.87238991e-01 -6.12187028e-01 7.87603319e-01 -3.26520234e-01
7.48565018e-01 1.58812739e-02 -4.28580120e-02 4.30811852e-01
-1.33759901e-01 2.73996085e-01 1.11042786e+00 -6.60477102e-01
-6.58809364e-01 -1.75452948e-01 -1.82421729e-01 1.34470427e+00
-2.38475055e-01 -2.62481034e-01 -6.00624800e-01 -5.19273818e-01
-5.15665412e-02 1.08627653e+00 -1.11400390e+00 1.61100090e-01
-3.44507575e-01 -9.69511986e-01 3.86182487e-01 -4.00214106e-01
5.22290230e-01 -1.18877268e+00 -5.47927201e-01 4.00220752e-01
-9.26430643e-01 -9.13925290e-01 3.54056247e-02 -1.17733050e-02
-5.51823199e-01 -9.95442152e-01 -4.65373904e-01 -1.53240457e-01
4.74162516e-04 2.91712165e-01 1.01154041e+00 2.69422799e-01
4.31982011e-01 -1.37520537e-01 -5.08286774e-01 -6.47042930e-01
-1.03472233e+00 -2.18847711e-02 -3.52166221e-02 4.68859524e-01
1.42826006e-01 -3.87786269e-01 -1.33376062e-01 4.78680551e-01
-6.43896341e-01 -1.05293162e-01 -8.54470953e-03 5.92375875e-01
-4.29361194e-01 -4.25141156e-01 1.01489842e+00 -1.41755891e+00
1.16300642e+00 -1.33897150e+00 3.35031509e-01 -6.74874842e-01
-5.28891265e-01 -5.13354957e-01 3.50108385e-01 -2.97202289e-01
-1.04827201e+00 -1.06937814e+00 -4.62455481e-01 3.31659704e-01
-1.45566776e-01 7.17837811e-01 4.01570022e-01 3.51104826e-01
1.12211609e+00 1.42675385e-01 3.24881822e-01 -3.94963413e-01
4.86301295e-02 1.08898401e+00 -4.99995053e-02 6.12998419e-02
2.05248445e-01 7.29226708e-01 -4.16478366e-01 -1.15088797e+00
-1.78003395e+00 -9.72197831e-01 -7.75168017e-02 -6.48419023e-01
5.29641390e-01 -7.50461638e-01 -6.47282720e-01 6.41859531e-01
-1.29541779e+00 1.82601094e-01 2.73894161e-01 1.03616700e-01
-4.83459592e-01 1.04160935e-01 -1.03425956e+00 -9.13760960e-01
-5.31598210e-01 -3.87737989e-01 1.51901931e-01 -5.47638834e-01
-1.11417198e+00 -1.00307000e+00 4.78056967e-01 7.87160456e-01
7.35084653e-01 4.70899999e-01 6.44439638e-01 -1.35000467e+00
4.52781886e-01 -3.14149112e-01 1.57452130e-03 1.85467213e-01
1.93426266e-01 -3.46610427e-01 -1.03656256e+00 -1.14058331e-01
7.87592351e-01 -6.63208127e-01 8.76155615e-01 -1.63785353e-01
3.24722052e-01 -1.10463154e+00 -1.65049713e-02 -4.41083491e-01
6.62880063e-01 -5.67533791e-01 4.99324024e-01 7.59696245e-01
6.92296922e-02 1.12884343e+00 6.34713233e-01 7.20847070e-01
3.28368455e-01 7.34922469e-01 7.37400770e-01 2.28287280e-01
2.10669577e-01 5.89424791e-03 8.85586500e-01 8.35227430e-01
-4.09794003e-02 -1.36560336e-01 -5.85549414e-01 4.41951871e-01
-1.77550185e+00 -1.42664707e+00 -1.00309801e+00 1.81287563e+00
9.04598117e-01 5.90615511e-01 6.31584942e-01 4.39521849e-01
7.07079887e-01 8.90420198e-01 9.54452604e-02 -1.06444883e+00
-1.44780219e-01 -6.42455637e-01 -9.00787115e-02 8.14156473e-01
-9.19308424e-01 4.84254688e-01 5.02843428e+00 3.65177095e-01
-1.40173101e+00 4.52565849e-01 5.20025969e-01 -6.09053634e-02
-1.18341543e-01 -3.73613894e-01 -6.75479174e-01 6.11612380e-01
1.23164415e+00 -1.71823829e-01 9.78856813e-03 8.84202898e-01
6.18328154e-01 -3.31901550e-01 -8.24424207e-01 3.90752316e-01
7.20539033e-01 -1.27348018e+00 -1.03178389e-01 -1.84909672e-01
5.36265790e-01 1.60368592e-01 -4.58586253e-02 3.54559809e-01
-2.49927878e-01 -6.41861260e-01 1.06711972e+00 4.13036346e-01
-1.92129180e-01 -5.87564945e-01 1.26629043e+00 9.16405261e-01
-5.84427938e-02 1.36723563e-01 -1.50221199e-01 -7.73708522e-01
3.83324891e-01 1.09449589e+00 -1.38556969e+00 1.44997358e-01
4.24358487e-01 9.36671913e-01 -3.15785438e-01 4.20504838e-01
-4.04472500e-01 1.05721986e+00 1.50989413e-01 -5.09037614e-01
3.62097919e-01 5.62023632e-02 1.18235421e+00 1.59057641e+00
-1.48566902e-01 -1.26338169e-01 -1.98269397e-01 5.98497808e-01
-1.59019843e-01 3.94214034e-01 -5.68863750e-01 6.58305362e-02
-2.64167078e-02 1.23093331e+00 -4.32405382e-01 -4.89826143e-01
1.70521869e-03 7.89937615e-01 2.47933567e-01 -2.54296333e-01
-6.70333624e-01 4.02130067e-01 3.53389382e-01 7.06531107e-01
-1.50534019e-01 5.15640199e-01 -1.49831727e-01 -8.77361417e-01
-2.49486595e-01 -1.05183959e+00 2.95499682e-01 -6.21629894e-01
-1.32392776e+00 7.53270924e-01 -2.24539533e-01 -1.00928140e+00
-3.81126702e-01 -1.57292858e-02 -8.59842777e-01 6.55500650e-01
-1.43159592e+00 -7.71303535e-01 -1.67904586e-01 3.11382651e-01
6.52903259e-01 1.34833708e-01 5.78058779e-01 4.93346229e-02
-2.47297257e-01 -1.21208750e-01 -4.18438613e-01 -1.07608378e-01
9.81091917e-01 -9.72611904e-01 2.53080148e-02 6.63874496e-04
-3.22926342e-01 3.98697674e-01 1.63261437e+00 -5.71388304e-01
-6.70663357e-01 -7.93312609e-01 1.75733912e+00 -5.67822039e-01
1.31706595e+00 -1.12598583e-01 -1.30342603e+00 4.42919195e-01
5.93681753e-01 -7.04521954e-01 7.52281904e-01 5.09768486e-01
-4.88743931e-01 3.23897868e-01 -1.25817895e+00 4.03219342e-01
4.25365508e-01 -3.45429838e-01 -1.10837269e+00 7.32361376e-01
3.23542655e-01 -7.20504969e-02 -5.48782229e-01 -1.60559058e-01
3.06819022e-01 -1.49491739e+00 4.92595315e-01 -8.05514812e-01
9.57201898e-01 2.47505933e-01 1.79402694e-01 -1.68109989e+00
-1.60610005e-01 -5.98819792e-01 -9.10964310e-02 1.02340031e+00
4.56169963e-01 -9.64397609e-01 4.39999461e-01 -1.51302040e-01
4.66341414e-02 -6.86826944e-01 -7.98111618e-01 -1.68658629e-01
9.04456154e-02 -2.65848458e-01 1.06063135e-01 1.37883747e+00
8.29136848e-01 9.45133388e-01 -9.14088488e-01 -2.91232556e-01
2.70744056e-01 2.19370410e-01 7.56638825e-01 -1.54713082e+00
1.97060063e-01 -7.86450267e-01 -1.85537785e-01 -6.20199800e-01
7.87821487e-02 -7.61123896e-01 -1.23255327e-01 -1.31534779e+00
2.27094576e-01 2.22613737e-01 6.40537366e-02 -6.96069077e-02
2.82619167e-02 2.14265823e-01 7.41032213e-02 6.39929652e-01
-5.58621228e-01 2.35053763e-01 9.35586035e-01 -3.73658761e-02
-6.62346408e-02 5.90275228e-01 -7.26893365e-01 1.18728220e+00
9.12204087e-01 -6.02381587e-01 1.53715849e-01 6.34543180e-01
1.19811094e+00 3.06761682e-01 5.14133632e-01 -5.36427200e-01
6.62758648e-02 4.19998839e-02 -3.88820529e-01 -7.36604214e-01
5.22546530e-01 -3.99962157e-01 -1.85353249e-01 6.71993256e-01
-5.86102605e-01 -1.77999303e-01 -1.51344314e-01 5.65480530e-01
-4.09582287e-01 -1.93478927e-01 1.00828791e+00 -9.11685303e-02
-1.31277993e-01 -3.23679984e-01 -1.11039114e+00 2.41580576e-01
7.23789811e-01 2.73287505e-01 -5.65097690e-01 -9.95836198e-01
-9.67963696e-01 -9.04706568e-02 2.01938987e-01 5.21846056e-01
4.79242802e-01 -8.35543573e-01 -1.51215219e+00 -3.58623952e-01
2.02056885e-01 -9.55511987e-01 2.62273878e-01 1.63869655e+00
-2.46329457e-02 2.82304734e-01 -7.76200518e-02 -3.00706714e-01
-1.56871855e+00 5.96753478e-01 2.24232852e-01 -2.85547048e-01
-5.76560140e-01 5.66688001e-01 -2.93301880e-01 -2.60532081e-01
-2.04842389e-01 5.42014576e-02 -8.86133492e-01 1.14461482e+00
9.48824465e-01 8.25994074e-01 4.34325099e-01 -9.65248227e-01
-1.44089282e-01 -4.02019858e-01 -3.33709061e-01 1.23619027e-02
1.19563532e+00 -4.13100183e-01 -7.42945194e-01 1.25163627e+00
1.10498023e+00 4.63510543e-01 -2.21332535e-01 -4.24385786e-01
3.83163512e-01 -2.49557003e-01 6.04944080e-02 -8.36105824e-01
-3.21353316e-01 4.99273658e-01 -3.40775400e-01 1.30427182e+00
2.26696670e-01 2.48405904e-01 8.26650321e-01 2.07944408e-01
1.51355982e-01 -1.07338881e+00 1.96360603e-01 8.28713000e-01
1.57497656e+00 -1.26236713e+00 1.85276389e-01 -4.40909356e-01
-1.00497615e+00 1.12794006e+00 -7.95259513e-03 -1.55491218e-01
6.28414869e-01 -2.80208379e-01 1.89625025e-01 -6.62234306e-01
-1.01377749e+00 2.24118561e-01 1.58425570e-01 -1.44203752e-01
6.94176495e-01 3.43302369e-01 -9.44513977e-01 4.05316621e-01
-4.66832250e-01 -2.67098755e-01 1.08760178e+00 3.65058631e-01
-9.06286836e-01 -3.58694822e-01 -5.91757834e-01 4.40704972e-01
-9.15001810e-01 3.80858965e-02 -1.09247458e+00 5.83992660e-01
-1.91951990e-01 1.26283109e+00 -2.24772662e-01 -2.24608988e-01
1.17984042e-01 2.72565246e-01 -1.84435412e-01 -6.63856208e-01
-1.10764766e+00 -2.88771152e-01 1.11751568e+00 -3.60909224e-01
-6.93902671e-01 -9.01129246e-01 -6.60466135e-01 -8.03026378e-01
-2.47212008e-01 5.84341943e-01 6.47453308e-01 1.37522972e+00
-1.53868288e-01 3.00067455e-01 9.55080211e-01 -5.01584888e-01
-6.94555998e-01 -1.46568632e+00 -4.39559519e-01 6.83267057e-01
7.80369520e-01 -5.25892794e-01 -7.82919645e-01 -4.40625668e-01] | [8.228943824768066, 10.116671562194824] |
ad7fffd2-22d7-44f5-abdb-9ee405e0055b | apsense-data-driven-algorithm-in-ppg-based | 2306.10863 | null | https://arxiv.org/abs/2306.10863v1 | https://arxiv.org/pdf/2306.10863v1.pdf | ApSense: Data-driven Algorithm in PPG-based Sleep Apnea Sensing | In this paper, we utilized obstructive sleep apnea and cardiovascular disease-related photoplethysmography (PPG) features in constructing the input to deep learning (DL). The features are pulse wave amplitude (PWA), beat-to-beat or RR interval, a derivative of PWA, a derivative of RR interval, systolic phase duration, diastolic phase duration, and pulse area. Then, we develop DL architectures to evaluate the proposed features' usefulness. Eventually, we demonstrate that in human-machine settings where the medical staff only needs to label 20% of the PPG recording length, our proposed features with the developed DL architectures achieve 79.95% and 73.81% recognition accuracy in MESA and HeartBEAT datasets. This simplifies the labelling task of the medical staff during the sleep test yet provides accurate apnea event recognition. | ['Theerawit Wilaiprasitporn', 'Thapanun Sudhawiyangkul', 'Thee Mateepithaktham', 'Phoomraphee Luenam', 'Narin Kunaseth', 'Thitikorn Keawlee', 'Punnawish Thuwajit', 'Guntitat Sawadwuthikul', 'Tanut Choksatchawathi'] | 2023-06-19 | null | null | null | null | ['photoplethysmography-ppg'] | ['medical'] | [ 3.72996144e-02 1.92947686e-02 -7.70660192e-02 -6.94025040e-01
-3.93721461e-01 -4.89133537e-01 -2.57854313e-01 -2.70822421e-02
-3.83173764e-01 9.52581763e-01 -1.98234972e-02 -4.81299102e-01
5.90497404e-02 -6.15952611e-01 8.17941129e-02 -7.57040858e-01
-4.62607831e-01 1.96102262e-01 -3.81675899e-01 4.95460331e-01
1.45897895e-01 4.95255113e-01 -9.35948193e-01 -8.75349641e-02
7.80363023e-01 1.61071217e+00 -4.89077359e-01 1.04900825e+00
1.75526842e-01 5.27051032e-01 -8.21070969e-01 1.27522439e-01
3.44464868e-01 -6.75097764e-01 -4.65244144e-01 -1.35045394e-01
3.19551170e-01 -4.62285936e-01 -3.20853323e-01 4.54178900e-01
1.18112218e+00 -8.58947113e-02 3.52414936e-01 -1.18001068e+00
-2.41603494e-01 2.62934506e-01 -2.32519269e-01 8.84286761e-01
1.82351500e-01 2.14430168e-01 6.01217866e-01 -5.25343955e-01
-4.27877307e-02 5.53090155e-01 1.13615060e+00 5.08900464e-01
-1.00709987e+00 -4.89684612e-01 -9.53945816e-01 3.58130425e-01
-1.31053758e+00 -4.72591311e-01 7.60601819e-01 -3.05851996e-01
9.50261474e-01 3.54683727e-01 1.04857051e+00 5.30099511e-01
5.06255567e-01 1.99025840e-01 1.20783031e+00 -3.94731224e-01
1.43465430e-01 -7.00949952e-02 3.36724609e-01 9.63277638e-01
2.12353230e-01 -2.79081184e-02 -2.28773147e-01 -1.73688799e-01
6.65715814e-01 2.90608019e-01 -4.16329682e-01 2.53181577e-01
-8.81735027e-01 6.57163084e-01 -1.08184861e-02 3.52431297e-01
-4.06652123e-01 2.09659128e-03 4.34405237e-01 2.32357740e-01
1.92322843e-02 4.26774919e-01 -4.18438733e-01 -6.61335409e-01
-1.08254552e+00 -3.31108898e-01 1.25503564e+00 6.10012412e-01
5.86613894e-01 4.23171408e-02 -6.48602307e-01 7.50024736e-01
3.75275791e-01 5.58502257e-01 7.32772827e-01 -1.36101270e+00
-5.26041240e-02 6.79341495e-01 9.46999192e-02 -7.29485393e-01
-1.02793884e+00 -3.02560329e-01 -9.30165112e-01 -2.22839132e-01
3.79749656e-01 -3.71129364e-01 -7.68462777e-01 1.27658570e+00
-4.80015762e-02 4.21166003e-01 -1.16784267e-01 7.65730917e-01
9.57001269e-01 2.13363290e-01 5.31539917e-02 -7.18500316e-01
1.52294743e+00 -7.90652275e-01 -9.38166022e-01 -1.20728411e-01
3.74019742e-01 -4.27203357e-01 8.14495504e-01 4.49579358e-02
-1.16636252e+00 -7.90984273e-01 -9.97940779e-01 -2.51524989e-02
-2.45577291e-01 3.75319064e-01 1.17430583e-01 8.10843766e-01
-1.03636551e+00 1.01909518e+00 -9.94296789e-01 -9.85084996e-02
5.40279686e-01 4.30605173e-01 9.01411846e-02 1.47352457e-01
-1.27240837e+00 9.54114497e-01 -6.39382973e-02 2.74968266e-01
-4.83589321e-01 -8.94498765e-01 -9.06164110e-01 3.07234645e-01
-3.53589326e-01 -8.26854229e-01 1.01792598e+00 3.54718952e-03
-1.43441701e+00 1.11465204e+00 -2.34788135e-01 -6.60596967e-01
2.11059973e-01 -1.24543436e-01 -8.60825121e-01 3.80204141e-01
-2.45860294e-01 1.45556778e-01 7.02902198e-01 -4.54334438e-01
-2.71677077e-01 -5.40496826e-01 -2.67483175e-01 -9.04133394e-02
-1.06509082e-01 -3.33031118e-01 -1.33340672e-01 -6.88163638e-02
-1.18395679e-01 -8.82524014e-01 -5.00095263e-02 1.40172347e-01
-3.20720166e-01 -2.18339249e-01 6.51129305e-01 -7.66897738e-01
1.61959100e+00 -2.42442083e+00 -6.22101068e-01 1.56042933e-01
5.69748223e-01 7.35313892e-01 3.69697243e-01 6.51439354e-02
5.75980311e-03 8.26200619e-02 -4.34671819e-01 -4.16938931e-01
6.92262053e-02 5.42009234e-01 1.34476140e-01 5.74414372e-01
-1.18840180e-01 1.09696865e+00 -7.69528449e-01 -7.02003002e-01
7.50105977e-01 6.83661520e-01 -7.56625235e-02 5.21089852e-01
5.89206338e-01 4.90926415e-01 -1.25332817e-01 7.32888460e-01
4.63244855e-01 -3.27531129e-01 -5.56950681e-02 -4.33603317e-01
-1.73124373e-01 5.03456473e-01 -5.58463752e-01 1.61212194e+00
-4.28349763e-01 8.18610132e-01 -3.69641185e-01 -9.75449324e-01
1.29470003e+00 6.89687371e-01 7.47313976e-01 -8.17825317e-01
5.48863232e-01 2.04942480e-01 4.37414914e-01 -1.06739986e+00
-5.28701067e-01 -4.36760455e-01 2.49069601e-01 6.56720221e-01
-1.45786196e-01 6.04402199e-02 2.49826998e-01 -4.49153841e-01
1.12133658e+00 -3.66608053e-01 8.37275624e-01 -2.58604199e-01
7.81355917e-01 -6.69816852e-01 6.37223840e-01 7.01456070e-01
-8.46770525e-01 7.92032242e-01 4.64366645e-01 -8.68587554e-01
-8.35925996e-01 -1.04831326e+00 -5.04229248e-01 3.77772599e-01
-2.87559718e-01 -1.87562615e-01 -4.18759972e-01 -5.97258210e-01
1.38001695e-01 6.30654812e-01 -5.40357769e-01 -3.84430528e-01
-7.87434280e-01 -8.27264071e-01 8.96973491e-01 9.00885701e-01
6.20317161e-01 -1.35903585e+00 -1.24098694e+00 3.30091864e-01
-1.56485811e-01 -1.03806114e+00 -4.96475458e-01 4.11061943e-02
-9.85233247e-01 -1.35145080e+00 -7.37637758e-01 -6.20866179e-01
2.31334060e-01 -4.45752949e-01 1.17302668e+00 -3.07194069e-02
-1.12986732e+00 3.93804491e-01 3.58280018e-02 -2.23825842e-01
-2.99967498e-01 -2.69076705e-01 1.18103541e-01 -1.47734582e-01
7.11874485e-01 -1.10753071e+00 -1.26955009e+00 1.03999980e-01
-1.94122195e-01 -4.25196439e-01 4.97087598e-01 2.93853223e-01
7.30082989e-01 -5.69040060e-01 7.45564342e-01 -4.18852061e-01
7.61246622e-01 -1.62688598e-01 -6.73773140e-02 1.75546557e-01
-9.79535460e-01 -4.72664416e-01 6.45719290e-01 -8.95305425e-02
-2.92265147e-01 -2.16434479e-01 -2.30864763e-01 -6.28964186e-01
-3.45588535e-01 1.89738750e-01 3.37280095e-01 1.64808214e-01
6.91113532e-01 6.28578067e-01 2.47931048e-01 -4.16029274e-01
-7.88872093e-02 8.81021500e-01 8.16806316e-01 -2.03242719e-01
2.59152591e-01 2.93024361e-01 2.09352419e-01 -7.70500898e-01
-8.30177188e-01 -6.45260155e-01 -9.81746912e-01 -4.06059213e-02
1.01699221e+00 -7.32782543e-01 -1.19600344e+00 3.01572502e-01
-7.49312341e-01 -1.16505384e-01 -6.88542426e-01 5.54430068e-01
-4.49268103e-01 5.45672178e-01 -4.80871856e-01 -8.89528513e-01
-1.16856956e+00 -5.45144677e-01 6.63409352e-01 5.93612492e-01
-4.71107394e-01 -1.02065802e+00 2.89445370e-01 9.35385749e-02
8.38043213e-01 7.61861563e-01 8.41068923e-01 -7.75494754e-01
2.41229355e-01 -4.34364587e-01 -2.10142314e-01 8.47200572e-01
7.83142090e-01 -1.12169221e-01 -9.27306235e-01 -1.09163240e-01
3.72503340e-01 6.49774894e-02 4.41249430e-01 7.77041733e-01
1.31251669e+00 -3.01193178e-01 -1.06429808e-01 8.62520635e-01
1.04406059e+00 8.60587418e-01 7.61318088e-01 -1.08934805e-01
4.81263101e-01 -6.56663924e-02 6.55264035e-02 9.35486555e-01
5.54891407e-01 1.47139758e-01 -1.24929301e-01 -3.25505883e-01
-1.99146718e-01 3.66632789e-01 1.92204908e-01 8.43147635e-01
-2.49534205e-01 1.83525160e-01 -7.96856344e-01 3.57268780e-01
-1.18590415e+00 -6.92603409e-01 -2.12545007e-01 2.11983776e+00
1.10260069e+00 -1.01561628e-01 2.89441049e-01 3.42030168e-01
7.06175923e-01 6.12061545e-02 -8.11940253e-01 -5.76364815e-01
2.83562303e-01 7.91342139e-01 2.84402966e-01 2.51447976e-01
-9.38102067e-01 1.67983249e-02 6.92830801e+00 -3.17361027e-01
-1.22226870e+00 -2.20490262e-01 5.94997108e-01 -1.13876678e-01
4.04967457e-01 -4.99563247e-01 -6.01453781e-01 9.58407938e-01
1.42891109e+00 -3.21278214e-01 5.51771224e-01 7.73846984e-01
4.60848629e-01 1.37883157e-01 -1.43617392e+00 1.36412871e+00
2.62779351e-02 -1.15480995e+00 -7.88870215e-01 -2.20225975e-01
-5.87389339e-03 5.43441884e-02 -2.63052821e-01 3.21576297e-01
-6.41075194e-01 -1.04428160e+00 -3.81057560e-01 7.21399903e-01
1.32584941e+00 -3.33037049e-01 7.98382938e-01 -1.67146802e-01
-1.10225022e+00 -1.30105540e-01 -2.89009750e-01 1.22434422e-02
1.09573632e-01 8.09313953e-01 -1.21924579e+00 2.00407550e-01
6.10181570e-01 8.56007576e-01 -5.49437225e-01 1.34788513e+00
-3.02957028e-01 8.43329370e-01 -4.74264175e-01 3.43673825e-01
-2.22909495e-01 -1.33667931e-01 1.88994199e-01 1.18077946e+00
4.59970146e-01 5.39618194e-01 4.27881293e-02 1.15012455e+00
1.99281909e-02 -1.80074975e-01 -4.34599578e-01 -4.69608419e-02
5.71502328e-01 1.34423923e+00 -3.36640656e-01 -5.09864688e-01
-3.99651915e-01 4.01179671e-01 -3.59420508e-01 1.93964168e-01
-8.99951994e-01 -6.97408557e-01 4.93973494e-01 2.77926661e-02
-7.80764744e-02 2.34603763e-01 -6.03410065e-01 -8.05950999e-01
-1.08005993e-01 -4.31975543e-01 5.99384427e-01 -5.49577773e-01
-1.06323528e+00 5.08056998e-01 -4.04887617e-01 -1.11555421e+00
-4.18775678e-01 -2.48589188e-01 -1.19298673e+00 1.13259912e+00
-1.77314091e+00 -4.84577090e-01 -7.14439750e-01 7.29871392e-01
6.35171458e-02 5.57027385e-02 1.13191116e+00 7.53249943e-01
-7.54088700e-01 5.27195454e-01 -5.55278361e-01 4.02273655e-01
5.47521949e-01 -1.57353306e+00 2.44100273e-01 3.18044871e-01
-3.93772364e-01 6.86507881e-01 3.36716920e-01 -1.68725222e-01
-1.05641675e+00 -9.23417032e-01 1.12148166e+00 -2.57468492e-01
1.88788742e-01 3.07118535e-01 -7.22973645e-01 5.75870275e-01
-9.20020975e-03 5.45750678e-01 1.29135370e+00 -3.83432627e-01
1.77159414e-01 -4.51715171e-01 -1.41513300e+00 -8.40613171e-02
4.81643587e-01 -5.67988813e-01 -9.21923280e-01 4.80333269e-02
1.79574370e-01 -6.06903613e-01 -1.51980162e+00 5.16130507e-01
8.35572004e-01 -8.51915598e-01 8.55076313e-01 -3.37502599e-01
8.00382793e-02 -2.25305855e-01 3.05484086e-01 -8.61734211e-01
-2.35958055e-01 -8.97334397e-01 -5.33440530e-01 1.00335217e+00
7.05803037e-02 -8.20021868e-01 5.59109926e-01 1.00322092e+00
-6.33199990e-01 -7.63652921e-01 -9.68129396e-01 -3.31433117e-01
-3.86232793e-01 5.97784258e-02 3.21338207e-01 9.44304049e-01
1.09515145e-01 1.96960449e-01 -2.89604753e-01 -1.65386908e-02
4.02547866e-01 4.37276602e-01 2.01401249e-01 -1.49620152e+00
-9.86977220e-02 -9.66868922e-02 -1.85558081e-01 -4.09541547e-01
-3.74755502e-01 -7.57810354e-01 2.49728444e-03 -1.64093482e+00
-9.74599496e-02 -3.04054618e-01 -1.01163638e+00 8.86265993e-01
-1.56315029e-01 4.40282851e-01 -2.25629061e-01 -8.56615603e-02
-1.73961967e-01 1.46627173e-01 1.26880050e+00 1.74916446e-01
-6.84343755e-01 3.96470726e-01 -3.96969825e-01 5.56532204e-01
1.04017770e+00 -4.85920727e-01 -3.40170383e-01 2.31590003e-01
-2.54983217e-01 3.35466981e-01 2.86542386e-01 -1.05821657e+00
-3.82604785e-02 3.41785550e-01 7.62476504e-01 -5.36556840e-01
2.89807558e-01 -6.81440771e-01 -1.02166131e-01 7.45845497e-01
-9.69015211e-02 3.78804207e-02 1.82924300e-01 1.02293408e-02
-4.83052917e-02 -8.10705721e-02 9.86793697e-01 -1.70065492e-01
-2.46421099e-01 5.02030671e-01 -3.09863597e-01 3.59101981e-01
8.21463704e-01 -3.39304358e-01 -4.36877161e-01 -6.27110749e-02
-1.19609034e+00 1.71649903e-01 -3.84778380e-02 1.42643541e-01
7.83076584e-01 -1.09391999e+00 -5.83084583e-01 5.21686256e-01
-1.12535311e-02 -1.61991507e-01 3.55040520e-01 1.56547749e+00
-6.82844162e-01 4.94131118e-01 -4.11575675e-01 -6.13353252e-01
-1.18789732e+00 1.32922828e-01 8.44695389e-01 -4.41496111e-02
-1.00800443e+00 6.20644808e-01 -3.75621617e-01 2.39302382e-01
4.09442484e-01 -6.17176414e-01 -3.34088504e-01 6.00174554e-02
6.75774336e-01 7.96015680e-01 5.27241305e-02 -8.66156444e-02
-6.16062284e-01 6.21900856e-01 3.00764859e-01 5.10715723e-01
1.09232199e+00 -2.89850473e-01 -2.60389209e-01 3.87930512e-01
1.35232377e+00 -3.43306780e-01 -1.01785660e+00 8.67811963e-02
-2.03181505e-01 -2.10669190e-01 1.06188953e-02 -1.10047054e+00
-1.21253157e+00 1.23029387e+00 1.27884030e+00 4.91803825e-01
1.19316816e+00 -3.25679868e-01 1.24739861e+00 2.43179753e-01
-1.45998523e-01 -7.45229125e-01 -1.92466259e-01 -1.44620880e-01
4.19592202e-01 -9.28384066e-01 3.23996954e-02 1.02394037e-01
-6.35072112e-01 1.46143210e+00 3.57552320e-01 -1.24221474e-01
9.25473750e-01 2.03698397e-01 4.80180949e-01 -2.49184534e-01
-5.15027821e-01 1.86368152e-02 3.07103068e-01 5.61319888e-01
5.17672062e-01 -2.80511528e-02 -5.80753624e-01 4.35657918e-01
-2.77322203e-01 4.69551682e-01 5.59099555e-01 6.26799881e-01
-5.33358216e-01 -7.20482349e-01 2.30673358e-01 7.80825913e-01
-8.36285293e-01 -1.88471943e-01 4.42659333e-02 5.72262645e-01
3.17795157e-01 9.50371623e-01 2.95509726e-01 -6.88147247e-02
3.43688726e-01 8.49385500e-01 4.10884172e-01 -6.92777991e-01
-4.68300998e-01 5.39689809e-02 4.71563973e-02 -5.34607410e-01
-5.07249653e-01 -2.91802973e-01 -1.54649413e+00 7.07964450e-02
1.46736935e-01 1.98671043e-01 4.31595445e-01 9.30145979e-01
5.27565658e-01 6.38568342e-01 8.33007991e-01 -3.53893876e-01
-3.47040117e-01 -1.31670582e+00 -8.46125722e-01 1.91135272e-01
1.02608454e+00 -2.19580457e-01 -5.88135779e-01 3.20794433e-01] | [13.863937377929688, 3.0156333446502686] |
4fa9fe5c-5102-469e-b853-15361a9137ab | predicting-multiple-sclerosis-disease | 2304.04062 | null | https://arxiv.org/abs/2304.04062v1 | https://arxiv.org/pdf/2304.04062v1.pdf | Predicting multiple sclerosis disease severity with multimodal deep neural networks | Multiple Sclerosis (MS) is a chronic disease developed in human brain and spinal cord, which can cause permanent damage or deterioration of the nerves. The severity of MS disease is monitored by the Expanded Disability Status Scale (EDSS), composed of several functional sub-scores. Early and accurate classification of MS disease severity is critical for slowing down or preventing disease progression via applying early therapeutic intervention strategies. Recent advances in deep learning and the wide use of Electronic Health Records (EHR) creates opportunities to apply data-driven and predictive modeling tools for this goal. Previous studies focusing on using single-modal machine learning and deep learning algorithms were limited in terms of prediction accuracy due to the data insufficiency or model simplicity. In this paper, we proposed an idea of using patients' multimodal longitudinal and longitudinal EHR data to predict multiple sclerosis disease severity at the hospital visit. This work has two important contributions. First, we describe a pilot effort to leverage structured EHR data, neuroimaging data and clinical notes to build a multi-modal deep learning framework to predict patient's MS disease severity. The proposed pipeline demonstrates up to 25% increase in terms of the area under the Area Under the Receiver Operating Characteristic curve (AUROC) compared to models using single-modal data. Second, the study also provides insights regarding the amount useful signal embedded in each data modality with respect to MS disease prediction, which may improve data collection processes. | ['Shayan Shams', 'Elmer V. Bernstam', 'Xiaoqian Jiang', 'John A. Lincoln', 'Kai Zhang'] | 2023-04-08 | null | null | null | null | ['disease-prediction'] | ['medical'] | [ 2.00316757e-01 -3.05984646e-01 -4.00754422e-01 -5.05397618e-01
-8.88451815e-01 -3.46531391e-01 3.65641564e-01 6.73801959e-01
-6.63805962e-01 7.66296029e-01 5.81538022e-01 -2.64916033e-01
-6.05738282e-01 -5.67877591e-01 -1.02288112e-01 -3.18243474e-01
-5.08627534e-01 8.70666385e-01 -1.96439028e-01 9.72121656e-02
1.19644381e-01 7.40112424e-01 -1.04764342e+00 6.69077754e-01
8.61744225e-01 9.56303596e-01 3.99774075e-01 5.93614638e-01
-7.57952482e-02 7.22909153e-01 -3.20287853e-01 3.13477874e-01
-2.98696049e-02 -2.03444019e-01 -7.48064339e-01 -3.05501312e-01
5.33748828e-02 -7.65105128e-01 -3.68461996e-01 5.11903822e-01
1.10077310e+00 -3.47041637e-01 4.83014584e-01 -8.92423928e-01
-3.30063432e-01 2.09353834e-01 -2.95654356e-01 5.19720972e-01
6.66404217e-02 4.70755070e-01 4.95843351e-01 -3.88779432e-01
6.97188377e-01 8.57748628e-01 9.22822356e-01 5.88197589e-01
-1.42697465e+00 -3.93070549e-01 -4.54941064e-01 5.50751686e-01
-8.32085013e-01 -3.02619755e-01 -4.72360849e-02 -8.83988023e-01
1.11983621e+00 3.24364722e-01 7.33938515e-01 9.05760109e-01
5.36617756e-01 5.47206342e-01 1.46826589e+00 -2.33394355e-01
1.28775463e-01 -3.34035873e-01 4.72877234e-01 3.64745289e-01
-1.87363222e-01 1.87542215e-01 -2.28297904e-01 -3.81984383e-01
7.25511849e-01 4.80852187e-01 5.76479547e-03 -4.51165065e-02
-1.64121497e+00 7.28605866e-01 4.56443220e-01 4.64792728e-01
-8.76878440e-01 -9.42007303e-02 7.45244086e-01 4.59219307e-01
3.49536449e-01 3.47783826e-02 -6.28385425e-01 -3.68126959e-01
-1.14559531e+00 1.58215597e-01 3.73015255e-01 3.18006538e-02
2.00319830e-02 -3.06827009e-01 -4.99861240e-01 8.91166151e-01
9.78613347e-02 6.52121067e-01 7.02444673e-01 -8.26425374e-01
2.21466869e-01 8.87185454e-01 -1.96233034e-01 -5.06972969e-01
-1.14863420e+00 -7.09916294e-01 -9.04207349e-01 7.62099996e-02
2.66308695e-01 -3.63358855e-01 -9.44616497e-01 1.54457343e+00
7.45337903e-02 -1.35478765e-01 -1.60870686e-01 9.67125833e-01
4.64523584e-01 -5.32104596e-02 2.22140417e-01 -1.58394575e-01
1.37739301e+00 -3.74086112e-01 -6.53575122e-01 -2.49199364e-02
1.13169265e+00 -2.54662454e-01 9.80368018e-01 4.05981869e-01
-8.71047437e-01 2.04688236e-02 -5.25031507e-01 1.45160675e-01
-3.20546120e-01 2.15605572e-01 5.01616240e-01 3.60808283e-01
-1.06345928e+00 6.29824638e-01 -1.20513916e+00 -7.37414122e-01
7.03607500e-01 5.60211599e-01 -5.49559534e-01 -2.80387163e-01
-1.09317720e+00 1.25439811e+00 2.10544944e-01 -2.28447825e-01
-7.84590721e-01 -1.09806955e+00 -1.67594090e-01 -1.98068216e-01
-2.97567844e-01 -1.04535329e+00 8.87572348e-01 -3.68648320e-01
-8.29914749e-01 7.66160488e-01 -1.85867965e-01 -6.56846404e-01
4.10033226e-01 -2.36586511e-01 -5.76697767e-01 2.86742061e-01
1.31466463e-01 4.40903962e-01 1.46584943e-01 -4.58280504e-01
-7.02412486e-01 -1.22433841e+00 -3.48183125e-01 4.94463705e-02
-3.39987457e-01 4.08883125e-01 3.46310943e-01 -5.63783586e-01
-1.55427292e-01 -7.85713494e-01 -4.88623798e-01 -1.28383279e-01
-1.42720953e-01 -6.12265095e-02 6.13477170e-01 -1.24703133e+00
1.24844885e+00 -1.77903485e+00 1.01910867e-01 7.73512498e-02
6.13659799e-01 3.64411116e-01 1.88929252e-02 4.44712609e-01
-3.00048590e-01 1.22297995e-01 1.33174649e-02 3.23694176e-03
-3.22772503e-01 -6.20491430e-02 2.81264931e-01 4.76219118e-01
5.85738048e-02 1.10508597e+00 -5.77544928e-01 -3.84235987e-03
4.08884525e-01 7.20280051e-01 -2.46841207e-01 -5.36866039e-02
1.52733311e-01 8.44318449e-01 -1.93310529e-01 5.02805173e-01
1.36399046e-01 -5.00419617e-01 6.86264411e-02 -1.34249367e-02
-1.30632326e-01 3.20762753e-01 -5.89001894e-01 1.73450029e+00
-2.70388693e-01 4.92949575e-01 1.09426856e-01 -1.16362226e+00
5.92453480e-01 3.67674410e-01 9.73569155e-01 -9.30212855e-01
7.57144243e-02 3.02433163e-01 2.57182300e-01 -7.88580120e-01
-6.39623404e-01 -2.86592543e-01 3.97266418e-01 7.08589852e-01
-3.23668897e-01 6.48187339e-01 -1.51638575e-02 6.55359542e-03
1.61081171e+00 -4.37984824e-01 2.23332763e-01 6.40509054e-02
3.61720413e-01 3.03393543e-01 3.70300084e-01 6.29917562e-01
-3.77805501e-01 3.42327982e-01 4.40592796e-01 -4.62077975e-01
-1.01175749e+00 -9.53578651e-01 -5.46051443e-01 9.32852924e-01
-7.26429284e-01 -1.46084309e-01 -4.56568897e-01 -3.33370090e-01
1.72462359e-01 4.65711921e-01 -4.48505878e-01 -1.45931453e-01
-3.37913990e-01 -9.81637120e-01 5.91118157e-01 7.05256343e-01
2.50993252e-01 -9.43975747e-01 -6.95451558e-01 2.73638487e-01
-2.19865307e-01 -5.07946432e-01 4.91449535e-02 3.86322476e-02
-1.26179719e+00 -1.15614343e+00 -1.15023124e+00 -6.01198792e-01
1.57294571e-01 -1.29243195e-01 4.39637542e-01 -4.07349110e-01
-5.45410514e-01 5.16729951e-01 -4.37229276e-02 -3.67002010e-01
-3.03589970e-01 6.19214587e-02 3.32148463e-01 -2.26152197e-01
7.30726659e-01 -6.55950725e-01 -1.12454200e+00 -2.69930005e-01
-7.72787154e-01 -1.08165853e-01 1.03952467e+00 5.21251321e-01
7.72184968e-01 -3.65646809e-01 9.75667000e-01 -4.07174557e-01
8.30030560e-01 -8.23351502e-01 1.80517673e-01 2.51568228e-01
-1.11148727e+00 -2.32669339e-01 6.74085245e-02 -2.04010293e-01
-5.27843475e-01 -1.99946299e-01 -4.04806286e-02 -1.58834308e-01
-6.16674066e-01 9.75163817e-01 4.38506007e-01 1.60902575e-01
6.09426141e-01 2.95820713e-01 6.36115909e-01 -7.91238189e-01
7.21336110e-03 1.19178998e+00 3.86283487e-01 9.18499157e-02
-1.91783845e-01 4.98434335e-01 1.77473679e-01 -8.17956448e-01
-4.28622425e-01 -7.29278326e-01 -6.86334431e-01 -1.40928894e-01
8.24677050e-01 -8.01057518e-01 -6.64132714e-01 4.87908930e-01
-8.96040022e-01 -3.15857708e-01 5.55259315e-03 8.32172394e-01
-5.44907987e-01 2.52137393e-01 -6.44780159e-01 -4.12321359e-01
-1.04607320e+00 -9.60975230e-01 8.73803318e-01 -3.00677598e-01
-5.80743074e-01 -9.82008338e-01 4.13150489e-01 4.98502016e-01
8.47917855e-01 4.88926589e-01 1.52219236e+00 -9.73731995e-01
4.93603647e-02 -6.06215537e-01 -3.99091244e-01 2.53863126e-01
2.39332438e-01 -6.45948052e-01 -3.32742035e-01 -4.60293949e-01
-2.93763960e-03 -3.07865858e-01 5.50222099e-01 8.43783140e-01
7.91662812e-01 1.62669614e-01 -2.80748695e-01 2.98248887e-01
1.31303847e+00 2.80118376e-01 4.81562585e-01 6.88008785e-01
4.10607487e-01 6.48233771e-01 1.55944228e-01 5.03584921e-01
4.33225632e-01 5.95711172e-01 1.53600380e-01 5.27961999e-02
-2.61316031e-01 4.78889585e-01 1.49383783e-01 4.34845448e-01
-2.36271750e-02 3.99075568e-01 -1.50286901e+00 5.68563759e-01
-1.70043945e+00 -9.11419570e-01 -4.77892727e-01 2.20648909e+00
7.48701572e-01 -1.24138154e-01 4.57052857e-01 7.09929168e-02
6.86226547e-01 -4.98272717e-01 -8.62990439e-01 -2.08260611e-01
-1.10724054e-01 1.73648119e-01 4.38284516e-01 -9.52947512e-03
-6.73115075e-01 1.72708139e-01 6.76306677e+00 2.90591955e-01
-1.51966166e+00 5.68793297e-01 2.42810190e-01 -6.95314467e-01
1.44968957e-01 -2.74508327e-01 -3.34315717e-01 5.43307722e-01
1.36298060e+00 -9.43767354e-02 5.05922914e-01 2.10790277e-01
9.59300756e-01 1.51997143e-02 -9.07512009e-01 7.96465576e-01
-4.43546772e-01 -1.39983940e+00 -1.72839254e-01 2.64692754e-01
2.17898965e-01 8.95625293e-01 5.44091314e-02 6.65428564e-02
-1.28892228e-01 -1.09478557e+00 1.10142194e-02 1.12692666e+00
1.20045817e+00 -6.00985706e-01 8.27238441e-01 4.87223119e-01
-4.99278694e-01 -2.01945812e-01 2.96541333e-01 -8.68166238e-02
2.76098758e-01 7.50852227e-01 -1.06078088e+00 1.93547964e-01
6.91136837e-01 6.18587315e-01 -4.21533614e-01 1.14378417e+00
3.38658482e-01 7.01037943e-01 -1.19755283e-01 4.97440130e-01
1.36054561e-01 2.84024272e-02 6.37701631e-01 1.06503248e+00
2.35426769e-01 9.70292687e-02 -4.87549715e-02 6.87514842e-01
5.64149439e-01 1.64022878e-01 -4.05314624e-01 -5.34862816e-01
9.27458331e-02 9.65820432e-01 -3.39544088e-01 -8.83604363e-02
-6.94383204e-01 5.13504505e-01 7.36123174e-02 1.39756605e-01
-2.06034899e-01 -2.36960813e-01 2.99193531e-01 5.31826794e-01
-3.12874734e-01 -2.23135352e-01 -7.59398639e-01 -8.12078238e-01
-2.21314251e-01 -8.96202385e-01 5.85649014e-01 -4.90882754e-01
-1.28629172e+00 2.66631156e-01 -4.21139568e-01 -8.15646470e-01
-3.68397892e-01 -3.46641898e-01 -3.56498063e-01 1.35681462e+00
-1.13444829e+00 -1.25537682e+00 -7.99775049e-02 5.86044610e-01
1.95128739e-01 -5.05923688e-01 1.17302227e+00 4.08299059e-01
-6.22501075e-01 1.92180544e-01 5.26852310e-01 -6.17318302e-02
8.27125847e-01 -1.10068071e+00 -2.61788726e-01 3.29098880e-01
-7.95278907e-01 7.93482244e-01 4.06032175e-01 -1.11078322e+00
-1.07473361e+00 -9.71073449e-01 1.04590547e+00 -3.64534497e-01
8.76622140e-01 3.05990696e-01 -7.72802293e-01 5.86373508e-01
-1.52355000e-01 -3.71175557e-01 1.17869878e+00 2.31099099e-01
1.36264563e-01 6.53638095e-02 -1.46750188e+00 2.21467555e-01
6.80518448e-01 -6.05197251e-01 -6.39045715e-01 3.30456585e-01
3.55953157e-01 2.02817887e-01 -1.60403764e+00 8.33590209e-01
8.35517287e-01 -7.53264844e-01 7.70034373e-01 -1.22715104e+00
5.44717610e-01 2.75118202e-01 -1.38766676e-01 -1.24616969e+00
-4.30914342e-01 -6.17314130e-02 -5.03267109e-01 7.57501304e-01
1.39928848e-01 -5.45530796e-01 6.65173948e-01 7.89225757e-01
-3.78532745e-02 -1.11712730e+00 -7.67953515e-01 -3.87077004e-01
2.55647331e-01 -3.82637799e-01 4.50110257e-01 9.20976460e-01
1.13956094e-01 2.53039837e-01 4.70502377e-02 6.14157394e-02
5.08197188e-01 -8.98857117e-02 1.44949839e-01 -1.45966363e+00
-1.06502445e-02 -4.43463445e-01 -6.07694209e-01 -3.90141271e-02
-3.50073963e-01 -1.25597775e+00 -7.78240621e-01 -2.24551654e+00
5.26740134e-01 -5.21347344e-01 -8.17528725e-01 5.91409981e-01
2.34372333e-01 -6.70948848e-02 -1.25543535e-01 5.31266332e-01
-3.07892650e-01 1.04797892e-01 1.03332198e+00 -9.49362889e-02
-3.91653836e-01 1.51955768e-01 -5.52816570e-01 3.51988643e-01
1.05257082e+00 -3.89093250e-01 -2.68958569e-01 -6.85177982e-01
1.12929441e-01 2.55462646e-01 5.72885156e-01 -9.53890145e-01
1.17696285e-01 -5.12073673e-02 4.95602906e-01 -7.70315468e-01
1.16884857e-01 -6.05190933e-01 7.73943067e-02 1.06941330e+00
-4.96496588e-01 -1.03495665e-01 6.97946176e-02 5.38554251e-01
-7.14791659e-03 2.79295206e-01 7.13831663e-01 -2.65171705e-03
-5.99718511e-01 2.85770506e-01 -6.48651540e-01 -2.82085985e-01
1.01829922e+00 -1.39767841e-01 -4.18974042e-01 -1.43574610e-01
-1.41389191e+00 3.11047792e-01 1.04890592e-01 4.64322180e-01
5.57274401e-01 -1.26149142e+00 -8.97820115e-01 -7.56082172e-03
4.72927354e-02 -4.73857433e-01 5.41518509e-01 1.86331916e+00
-5.41125596e-01 1.02990282e+00 -6.81722105e-01 -7.23128080e-01
-1.21533549e+00 2.27506280e-01 3.65030944e-01 -2.61733770e-01
-7.31751800e-01 3.39119494e-01 -3.53540152e-01 -4.25622940e-01
5.16036212e-01 -1.31830247e-02 -4.00704116e-01 3.17790955e-01
8.94238830e-01 8.31233025e-01 4.51046467e-01 -4.25706327e-01
-5.19406438e-01 -4.97257104e-03 -3.18305641e-01 -1.91904470e-01
1.79703975e+00 -3.30926567e-01 -4.39341635e-01 5.22966444e-01
1.12839055e+00 -5.99906683e-01 -5.73096991e-01 -2.35725641e-01
1.93187907e-01 1.21029884e-01 5.75678527e-01 -1.51406872e+00
-7.92063594e-01 9.83714938e-01 1.55419898e+00 1.84492413e-02
9.92334843e-01 5.82637861e-02 1.12136662e+00 2.45227709e-01
2.85556793e-01 -8.25240135e-01 -3.63292545e-01 3.37574780e-01
7.18231618e-01 -1.18150067e+00 -3.81682754e-01 4.09245193e-01
-6.10998988e-01 1.03675127e+00 7.66407251e-02 1.76055491e-01
6.24998271e-01 3.40633728e-02 1.41594172e-01 -5.62094629e-01
-6.85760319e-01 -1.80656031e-01 1.27502516e-01 8.47813606e-01
6.12186432e-01 4.23939109e-01 -8.71486366e-01 1.06020379e+00
2.46065632e-01 7.55891442e-01 6.82408884e-02 9.82748210e-01
-6.13532484e-01 -1.09817255e+00 -1.78151205e-01 1.33840919e+00
-6.96305096e-01 -7.07671940e-02 -5.40325880e-01 3.37718725e-01
1.98041841e-01 7.62894332e-01 -3.69031131e-02 -2.91273862e-01
3.57776433e-01 3.71741980e-01 3.13674271e-01 -6.15978122e-01
-3.38509917e-01 -7.17361318e-03 1.05856985e-01 -5.86374342e-01
-2.65436262e-01 -8.47496271e-01 -1.48205507e+00 -3.26204240e-01
2.20317602e-01 -3.69166166e-01 9.64397550e-01 1.26487112e+00
8.72911394e-01 6.85613215e-01 2.83463746e-01 -3.39093238e-01
-7.33475029e-01 -1.08654773e+00 -6.33253813e-01 5.45027666e-02
3.45877677e-01 -5.22564709e-01 3.92424911e-02 -2.06838593e-01] | [14.25850772857666, -1.7436529397964478] |
d999ed70-5399-4f33-b0c5-cf0ad4bc0902 | unidexgrasp-universal-robotic-dexterous | 2303.00938 | null | https://arxiv.org/abs/2303.00938v2 | https://arxiv.org/pdf/2303.00938v2.pdf | UniDexGrasp: Universal Robotic Dexterous Grasping via Learning Diverse Proposal Generation and Goal-Conditioned Policy | In this work, we tackle the problem of learning universal robotic dexterous grasping from a point cloud observation under a table-top setting. The goal is to grasp and lift up objects in high-quality and diverse ways and generalize across hundreds of categories and even the unseen. Inspired by successful pipelines used in parallel gripper grasping, we split the task into two stages: 1) grasp proposal (pose) generation and 2) goal-conditioned grasp execution. For the first stage, we propose a novel probabilistic model of grasp pose conditioned on the point cloud observation that factorizes rotation from translation and articulation. Trained on our synthesized large-scale dexterous grasp dataset, this model enables us to sample diverse and high-quality dexterous grasp poses for the object point cloud.For the second stage, we propose to replace the motion planning used in parallel gripper grasping with a goal-conditioned grasp policy, due to the complexity involved in dexterous grasping execution. Note that it is very challenging to learn this highly generalizable grasp policy that only takes realistic inputs without oracle states. We thus propose several important innovations, including state canonicalization, object curriculum, and teacher-student distillation. Integrating the two stages, our final pipeline becomes the first to achieve universal generalization for dexterous grasping, demonstrating an average success rate of more than 60\% on thousands of object instances, which significantly outperforms all baselines, meanwhile showing only a minimal generalization gap. | ['He Wang', 'Li Yi', 'Tengyu Liu', 'Jiayi Chen', 'Yijia Weng', 'Haoran Geng', 'Ruicheng Wang', 'Hao Shen', 'Zikang Shan', 'Haoran Liu', 'Jialiang Zhang', 'Weikang Wan', 'Yinzhen Xu'] | 2023-03-02 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Xu_UniDexGrasp_Universal_Robotic_Dexterous_Grasping_via_Learning_Diverse_Proposal_Generation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Xu_UniDexGrasp_Universal_Robotic_Dexterous_Grasping_via_Learning_Diverse_Proposal_Generation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['motion-planning'] | ['robots'] | [-7.41247833e-02 5.22409715e-02 -1.91885307e-01 -1.96939901e-01
-9.48859334e-01 -1.01221228e+00 3.24679196e-01 -1.09466128e-01
-1.94306716e-01 3.34113747e-01 -3.43115509e-01 -2.67637402e-01
-3.21301758e-01 -5.86240947e-01 -1.34131157e+00 -9.97768283e-01
-2.76898712e-01 1.11267495e+00 2.31244802e-01 -1.83938608e-01
2.37102926e-01 9.39829051e-01 -1.47729087e+00 1.69082969e-01
8.44949126e-01 1.08603334e+00 9.15912628e-01 7.20970333e-01
1.07489862e-01 1.57076597e-01 -4.08481389e-01 -1.80386409e-01
5.83627105e-01 4.38462943e-01 -6.94880486e-01 1.02663733e-01
4.22177613e-01 -7.80433774e-01 -4.43612784e-01 9.85090911e-01
3.23402345e-01 1.01563066e-01 7.07721770e-01 -1.21331954e+00
-6.62006259e-01 7.36215353e-01 -1.23626463e-01 -4.85687315e-01
2.37560838e-01 7.77806640e-01 9.49795902e-01 -6.97526455e-01
6.61736786e-01 1.58543432e+00 1.66324407e-01 7.38328278e-01
-1.15141726e+00 -5.68948269e-01 4.70295727e-01 -1.45151809e-01
-7.39341974e-01 2.13381276e-01 4.56564814e-01 -4.66313750e-01
8.03051829e-01 2.09389329e-02 5.90240777e-01 1.79391217e+00
2.10119247e-01 1.29484951e+00 8.81357849e-01 5.24672233e-02
4.51984912e-01 -6.37921393e-01 -5.26812449e-02 5.10072649e-01
2.69709468e-01 2.23048124e-02 -1.78050417e-02 -1.01020999e-01
1.13881576e+00 4.26073939e-01 -9.85266343e-02 -9.35784638e-01
-1.64120710e+00 4.27068561e-01 7.26218760e-01 -5.39105758e-02
-6.62558794e-01 4.94186044e-01 1.40498579e-01 2.29723305e-01
-2.84322351e-01 8.89491022e-01 -9.60417449e-01 -2.01324642e-01
-2.71616340e-01 9.90609050e-01 1.10003412e+00 1.75477719e+00
5.58274209e-01 -2.46254638e-01 -2.47675404e-01 3.66028309e-01
1.12770490e-01 1.13194966e+00 2.77068975e-05 -1.08224523e+00
7.78545082e-01 4.27848607e-01 5.59965730e-01 -5.12664139e-01
-4.04953033e-01 -2.14122415e-01 -4.66248065e-01 3.70643198e-01
5.23571610e-01 -5.11614904e-02 -1.50070274e+00 1.60254776e+00
3.53798896e-01 -1.83699921e-01 1.16315968e-01 1.15388691e+00
3.30733001e-01 5.21322072e-01 2.49266669e-01 3.62697631e-01
1.26330972e+00 -9.90007341e-01 -2.03347221e-01 -1.51010126e-01
2.90452123e-01 -5.54834962e-01 1.22290921e+00 8.45015585e-01
-8.86218429e-01 -5.44838071e-01 -8.59505951e-01 -3.64969037e-02
-1.52229786e-01 3.89803231e-01 1.06020808e+00 -8.42314810e-02
-4.55282420e-01 1.18953574e+00 -1.57014155e+00 -1.59198865e-01
4.31803733e-01 5.33526063e-01 -4.38295633e-01 -4.68131602e-01
-4.36040312e-01 9.53109443e-01 6.06374145e-01 2.86523048e-02
-1.59182227e+00 -6.44241631e-01 -6.13106728e-01 1.50818720e-01
8.45661521e-01 -6.20165825e-01 1.35722578e+00 -1.91379830e-01
-1.71910596e+00 4.32523787e-01 3.22969705e-01 -1.20495476e-01
6.10475361e-01 -7.30557501e-01 5.27530134e-01 1.91603333e-01
-7.97164068e-02 8.65406990e-01 1.05678558e+00 -1.59660995e+00
-4.39395487e-01 -5.34567297e-01 4.33814853e-01 1.59729749e-01
1.54528543e-01 -6.25688732e-01 -6.75758541e-01 -5.88191867e-01
5.98823965e-01 -1.47706103e+00 -3.05798560e-01 1.19805917e-01
-4.41913545e-01 -6.52648747e-01 8.16093266e-01 -5.90257883e-01
5.88864572e-02 -1.89349449e+00 1.03681672e+00 -9.41955820e-02
-8.40967298e-02 6.17258884e-02 -4.20455456e-01 5.31733215e-01
2.70817429e-01 -3.77603620e-01 -1.02462932e-01 -3.10640633e-01
5.63121498e-01 6.82886064e-01 -9.44816411e-01 1.08295955e-01
5.62019646e-01 1.19906414e+00 -1.32459855e+00 1.03318617e-01
2.63112932e-01 2.19796747e-02 -7.48603940e-01 6.80017471e-01
-1.15572524e+00 7.96591282e-01 -8.54033411e-01 1.15615129e+00
7.10201085e-01 -1.27260968e-01 3.60156715e-01 -4.02373523e-01
1.99096069e-01 1.20594382e-01 -1.09694934e+00 2.32119679e+00
-5.00975788e-01 -2.25583509e-01 2.77805805e-01 -8.74336362e-01
9.87335801e-01 1.56535119e-01 4.88119751e-01 6.13656305e-02
1.13710918e-01 4.73334074e-01 3.36656906e-02 -7.66225755e-01
3.59997302e-01 1.93420172e-01 -1.63413957e-01 1.50021642e-01
6.52291477e-01 -9.64670956e-01 -2.79959384e-02 -6.95293248e-02
1.09445930e+00 9.45658624e-01 -3.19431931e-01 -3.11382055e-01
-2.50548840e-01 1.00824513e-01 2.42718309e-01 7.69391835e-01
1.90065056e-01 6.58732057e-01 5.90306282e-01 -3.28686416e-01
-1.06434417e+00 -1.65124512e+00 1.13233000e-01 1.02022970e+00
3.84946674e-01 5.91560453e-02 -5.26805580e-01 -7.67934978e-01
6.64136529e-01 5.62115252e-01 -4.54704553e-01 -9.63193402e-02
-8.83954763e-01 -2.80168027e-01 3.41702551e-02 8.20391357e-01
3.05230096e-02 -1.34587467e+00 -7.34845579e-01 1.70425624e-01
1.67160794e-01 -1.32279968e+00 -1.22933313e-02 5.52228153e-01
-1.09977639e+00 -1.20296621e+00 -1.01368654e+00 -7.25613952e-01
6.84197724e-01 1.22622333e-01 7.50134349e-01 -3.52934301e-01
-4.25908983e-01 5.85282922e-01 -6.33822739e-01 -5.08207440e-01
-3.80407393e-01 3.69208366e-01 3.40432823e-01 -6.08924747e-01
3.96498945e-04 -6.30208731e-01 -5.57796359e-01 4.92648631e-02
-8.51168275e-01 -8.04654732e-02 1.26682293e+00 7.66975522e-01
5.15257716e-01 -4.80736077e-01 3.73116463e-01 -3.08082812e-02
2.07975745e-01 -4.07310426e-01 -7.98380554e-01 3.57649893e-01
-4.68046591e-02 2.83244878e-01 6.62713945e-01 -9.69680965e-01
-5.69058180e-01 2.08056808e-01 -4.31986377e-02 -1.05254102e+00
-3.47431898e-01 5.61144277e-02 -2.97100335e-01 2.12091729e-02
4.91674423e-01 9.07536224e-02 1.72218025e-01 -8.45015287e-01
6.23996913e-01 3.79160553e-01 7.31904030e-01 -1.45758247e+00
8.89563262e-01 2.54959822e-01 1.33434251e-01 -5.12198091e-01
-7.01901615e-01 -2.39182830e-01 -8.54236722e-01 1.59981728e-01
8.85946512e-01 -7.56149471e-01 -1.45154059e+00 4.70605284e-01
-1.36696339e+00 -8.31912339e-01 -3.05276394e-01 6.89593077e-01
-1.06383014e+00 3.23736399e-01 -5.39436519e-01 -6.29074991e-01
-3.20130050e-01 -1.62038922e+00 1.55453992e+00 -8.55952036e-03
1.97918430e-01 -1.33424580e-01 -4.53159332e-01 -9.09693167e-03
1.98760331e-01 4.28299367e-01 9.55713749e-01 -5.86911976e-01
-1.21182346e+00 -1.81449652e-01 -1.04139805e-01 4.91901904e-01
2.12535858e-01 -1.93507865e-01 -5.18207967e-01 -8.28326344e-01
-1.94074377e-01 -8.34122837e-01 7.47408569e-01 8.83456320e-02
1.49240613e+00 -1.23746417e-01 -4.40272659e-01 5.98660827e-01
1.19228995e+00 1.30396606e-02 3.29142123e-01 -8.07338487e-03
8.06465745e-01 4.88685817e-01 9.51145589e-01 2.97724307e-01
1.35084674e-01 6.01763248e-01 1.09868121e+00 6.59066200e-01
6.88231289e-02 -3.40741456e-01 2.71217167e-01 6.00437284e-01
-4.13319826e-01 -5.78164682e-02 -8.80693078e-01 4.70519036e-01
-1.80258596e+00 -4.24920678e-01 4.37377304e-01 2.08040071e+00
6.77602887e-01 1.01007484e-01 -8.29447955e-02 -3.83010149e-01
1.32225171e-01 -1.03650294e-01 -1.09237111e+00 1.63668171e-01
3.27247977e-01 3.37334871e-01 5.64505458e-01 2.36442357e-01
-9.82140303e-01 1.26284456e+00 5.04326963e+00 5.94889998e-01
-1.22959530e+00 -3.01356316e-01 -2.32720762e-01 -1.47656715e-02
-2.74564773e-02 -4.65343632e-02 -8.24854791e-01 2.66374081e-01
2.35284671e-01 4.92769241e-01 8.50117385e-01 1.31994629e+00
-4.31279898e-01 1.96253210e-01 -1.62961257e+00 8.54067743e-01
-3.05353284e-01 -9.03339207e-01 3.30078751e-01 -2.73484528e-01
6.22131884e-01 2.56492347e-01 7.98166096e-02 6.85985029e-01
6.29736125e-01 -8.46216798e-01 8.61137211e-01 4.12217617e-01
6.60018504e-01 -2.39886239e-01 1.69235364e-01 5.51795542e-01
-7.67273247e-01 -5.48806369e-01 -6.16003692e-01 2.65118927e-01
7.55723268e-02 6.08682185e-02 -7.56064057e-01 8.56579125e-01
7.73852289e-01 4.46289480e-01 1.86031312e-01 8.07401061e-01
-5.21616578e-01 5.41527532e-02 -3.31382096e-01 -2.01389834e-01
4.28462327e-01 -1.41496167e-01 8.02660465e-01 7.74346650e-01
4.36199903e-01 3.01810443e-01 5.48051357e-01 1.07410049e+00
1.04308492e-02 -5.23247123e-01 -4.11539823e-01 -2.32023239e-01
4.68539894e-01 1.25913918e+00 -4.41471517e-01 -2.44733706e-01
2.41398945e-01 1.02484906e+00 3.87468338e-01 3.31561267e-01
-6.51557803e-01 -2.51821220e-01 7.63868451e-01 -4.24956411e-01
6.79966509e-01 -8.63437116e-01 -1.11566879e-01 -1.24688113e+00
3.36588204e-01 -1.08908868e+00 -2.53857136e-01 -8.36516201e-01
-1.19859028e+00 4.24981177e-01 3.21932137e-01 -9.55767989e-01
-2.54369646e-01 -1.32361448e+00 -1.69454634e-01 8.44350576e-01
-1.24002492e+00 -1.38114321e+00 -5.43758810e-01 2.26747379e-01
7.55989850e-01 5.43043017e-02 8.42169106e-01 -2.82250226e-01
-3.11236437e-02 1.74253136e-01 -1.33489788e-01 -1.17081381e-01
5.59458554e-01 -1.56716621e+00 6.12171769e-01 2.23142102e-01
-2.66911089e-01 9.14110541e-01 7.34128594e-01 -7.79453158e-01
-2.28272557e+00 -1.05446661e+00 -1.17186338e-01 -8.40845644e-01
7.19241440e-01 -6.41714990e-01 -9.36814487e-01 9.29522872e-01
-4.03409660e-01 -8.53411928e-02 -3.53449643e-01 4.46134843e-02
-3.96644592e-01 -4.80316356e-02 -1.05379438e+00 6.32013083e-01
1.26750374e+00 -5.08864336e-02 -9.33725536e-01 7.50614583e-01
1.05173492e+00 -1.01973975e+00 -1.07624078e+00 8.92749667e-01
7.22580373e-01 -1.56697720e-01 1.14006925e+00 -1.04926813e+00
6.12217605e-01 -5.82051985e-02 -5.01723170e-01 -1.21192873e+00
-1.94524631e-01 -7.45092750e-01 -4.74180162e-01 6.41330004e-01
-5.60077578e-02 -4.39766854e-01 7.89831519e-01 4.15688902e-01
-4.69817638e-01 -1.14402199e+00 -6.71657324e-01 -1.18924296e+00
4.21358228e-01 -1.86313987e-01 7.31731057e-01 4.61014569e-01
-2.05981195e-01 -9.24569443e-02 -3.58850090e-03 7.24741638e-01
7.30905533e-01 7.06705332e-01 1.21808004e+00 -1.12778807e+00
-6.57766640e-01 -2.68553078e-01 -9.28115770e-02 -1.80129385e+00
2.96012461e-01 -9.45890248e-01 6.25822365e-01 -1.47012901e+00
1.83051080e-01 -1.01194477e+00 9.58627686e-02 6.24823749e-01
-2.14340910e-01 -5.59875071e-01 6.21568859e-01 4.17644113e-01
-3.44322681e-01 6.40788436e-01 1.83147204e+00 -2.25331545e-01
-1.61470026e-01 1.56796485e-01 -1.17437080e-01 4.98477280e-01
5.64853251e-01 -1.72537267e-01 -2.38684297e-01 -9.11488593e-01
-2.88433105e-01 2.66980737e-01 8.44886005e-01 -9.78874207e-01
-6.17045462e-02 -5.22650540e-01 2.12788001e-01 -6.72777116e-01
5.76811850e-01 -9.06794250e-01 -2.61903733e-01 7.92935371e-01
-1.57439545e-01 -9.27253142e-02 1.93545476e-01 7.54113376e-01
4.06972677e-01 -1.75610185e-01 4.76733029e-01 -2.96611696e-01
-5.50108910e-01 7.15301216e-01 1.96129367e-01 -3.49705726e-01
1.08104396e+00 4.05027121e-01 -3.50610018e-01 2.14514494e-01
-9.48928654e-01 3.81884485e-01 4.89835262e-01 8.98878157e-01
4.27429616e-01 -9.38934147e-01 -5.23948371e-01 7.32291788e-02
4.01499867e-02 8.24789941e-01 2.48929054e-01 4.86582875e-01
-4.96586919e-01 4.57006663e-01 -3.63846093e-01 -1.02165103e+00
-5.27205467e-01 9.47856367e-01 -1.19593166e-01 -8.61900374e-02
-9.11120415e-01 7.90520549e-01 1.29994646e-01 -8.86764705e-01
5.79032779e-01 -1.00102735e+00 4.87200648e-01 -6.30167603e-01
-4.27440228e-03 2.17619255e-01 -9.07110870e-02 2.22705662e-01
-1.03521951e-01 5.32733917e-01 -8.98839310e-02 1.82456300e-01
1.50907505e+00 6.36075377e-01 2.23576534e-03 3.71381849e-01
8.95160377e-01 -3.75980914e-01 -1.98453009e+00 1.66316897e-01
-1.05120622e-01 -3.32408249e-01 -6.76668286e-01 -1.03057742e+00
-5.70402205e-01 9.74278808e-01 2.95612067e-01 -1.53377786e-01
4.79776114e-01 3.12628716e-01 8.05250287e-01 1.18679011e+00
1.08764064e+00 -7.58957148e-01 3.75988066e-01 8.01714122e-01
1.52598488e+00 -1.27432513e+00 -2.72161424e-01 -3.43552560e-01
-4.70254064e-01 1.23029697e+00 7.52466023e-01 -6.87264144e-01
3.01901609e-01 2.93496460e-01 -6.36595041e-02 -1.54044539e-01
-6.60013378e-01 9.41887945e-02 3.11401784e-01 6.85698450e-01
-3.72113705e-01 2.65938997e-01 1.25269890e-01 5.34733593e-01
-1.03528969e-01 -4.25280109e-02 -1.40770704e-01 1.21052051e+00
-4.89971429e-01 -1.03284585e+00 -1.58817247e-01 3.49473625e-01
-4.74745594e-02 4.14588630e-01 6.19059913e-02 9.95427191e-01
-1.66868344e-01 3.61668378e-01 -5.48907407e-02 -4.01784629e-01
5.98828375e-01 -2.41008215e-02 1.19459474e+00 -8.08027446e-01
-3.10705602e-01 -2.63685137e-01 -6.01201952e-01 -1.04495966e+00
8.31463784e-02 -5.05192399e-01 -1.34423745e+00 2.05385923e-01
-3.12374383e-01 -1.21079460e-01 1.07958519e+00 9.19403255e-01
4.00203466e-01 5.69552422e-01 3.00089628e-01 -1.91691899e+00
-1.80779827e+00 -1.14497507e+00 -4.06415671e-01 5.67092061e-01
3.22332680e-01 -1.10206151e+00 -1.53056100e-01 -2.10976422e-01] | [5.722727298736572, -0.7932260036468506] |
240c8387-0383-427c-9e74-69fd0fa5c273 | feature-selection-with-distance-correlation | 2212.00046 | null | https://arxiv.org/abs/2212.00046v1 | https://arxiv.org/pdf/2212.00046v1.pdf | Feature Selection with Distance Correlation | Choosing which properties of the data to use as input to multivariate decision algorithms -- a.k.a. feature selection -- is an important step in solving any problem with machine learning. While there is a clear trend towards training sophisticated deep networks on large numbers of relatively unprocessed inputs (so-called automated feature engineering), for many tasks in physics, sets of theoretically well-motivated and well-understood features already exist. Working with such features can bring many benefits, including greater interpretability, reduced training and run time, and enhanced stability and robustness. We develop a new feature selection method based on Distance Correlation (DisCo), and demonstrate its effectiveness on the tasks of boosted top- and $W$-tagging. Using our method to select features from a set of over 7,000 energy flow polynomials, we show that we can match the performance of much deeper architectures, by using only ten features and two orders-of-magnitude fewer model parameters. | ['David Shih', 'Gregor Kasieczka', 'Ranit Das'] | 2022-11-30 | null | null | null | null | ['automated-feature-engineering', 'feature-engineering'] | ['methodology', 'methodology'] | [ 1.16334789e-01 -2.72548258e-01 -1.48309126e-01 -9.37716663e-01
-9.16987360e-01 -5.66933692e-01 5.80862761e-01 5.17874241e-01
-5.17967105e-01 6.22706771e-01 -5.55275120e-02 -3.24629635e-01
-6.03207111e-01 -8.35772634e-01 -4.58456755e-01 -6.91539347e-01
-5.29630482e-01 5.67685008e-01 1.44268051e-01 -2.34003574e-01
4.75617051e-01 9.27682698e-01 -1.58020389e+00 4.63052541e-02
5.51904380e-01 1.28212225e+00 -1.12487324e-01 7.09809601e-01
7.70852938e-02 5.11111677e-01 -4.99228165e-02 -2.37086043e-01
2.52371371e-01 -3.21130514e-01 -1.02534652e+00 -4.53690469e-01
3.12921226e-01 9.25183892e-02 -4.45967853e-01 9.42863941e-01
5.39578140e-01 5.82585871e-01 6.87447429e-01 -1.12180126e+00
-4.10157681e-01 4.32694256e-01 -3.62896085e-01 4.84390020e-01
3.06266081e-02 3.03591281e-01 1.37256455e+00 -7.20121980e-01
5.06066799e-01 1.03057635e+00 7.03604460e-01 3.39390665e-01
-1.43545699e+00 -6.39106274e-01 -2.96359807e-01 8.62628222e-02
-1.05413759e+00 -7.46653557e-01 8.87313366e-01 -5.38679540e-01
1.20787716e+00 3.61671358e-01 4.64530230e-01 7.20192134e-01
4.05891329e-01 3.87155533e-01 8.77472401e-01 -4.88717020e-01
1.80443719e-01 1.69343755e-01 5.04288793e-01 9.79599774e-01
2.66191334e-01 3.67822140e-01 -6.65645361e-01 -3.44635725e-01
4.48628962e-01 -2.04075605e-01 -3.30426581e-02 -5.50053358e-01
-1.08029091e+00 1.15486181e+00 4.12982792e-01 2.70437896e-01
-6.03052117e-02 2.35394120e-01 5.57201684e-01 4.77475971e-01
4.09273237e-01 1.10113609e+00 -8.76464307e-01 -2.06424907e-01
-9.00772929e-01 3.98447305e-01 6.24000549e-01 3.91777903e-01
9.81685877e-01 -2.14400478e-02 1.57977909e-01 5.74466169e-01
-7.08928891e-03 1.72497794e-01 5.86623132e-01 -1.01297641e+00
1.36956856e-01 6.23664200e-01 -1.28571004e-01 -9.56846654e-01
-1.02021134e+00 -5.43162763e-01 -7.33844995e-01 6.07661605e-01
5.77386558e-01 -1.80936098e-01 -1.00790858e+00 1.88112330e+00
2.54197925e-01 -5.76999962e-01 -2.95028150e-01 8.89465928e-01
4.94689077e-01 4.54997092e-01 -3.70430574e-02 6.51834756e-02
1.13822567e+00 -5.05185902e-01 -6.98136240e-02 -2.35964566e-01
7.94559300e-01 -6.62219048e-01 8.79840314e-01 4.93676543e-01
-1.00011027e+00 -5.99922955e-01 -1.06311965e+00 -3.76086026e-01
-6.20107949e-01 -7.21640512e-02 1.38272381e+00 5.64664900e-01
-9.98690963e-01 1.31999850e+00 -8.87362719e-01 -9.72894579e-02
5.61026156e-01 9.09331083e-01 -5.74666381e-01 2.76172847e-01
-1.03508759e+00 8.98445725e-01 3.50796342e-01 -2.64180321e-02
-4.96952355e-01 -7.98827410e-01 -8.94850433e-01 3.01869661e-01
1.83646441e-01 -6.41006052e-01 1.11212349e+00 -9.97637928e-01
-1.33097196e+00 6.74066186e-01 -2.06848606e-01 -2.38899872e-01
2.23122574e-02 -2.48897433e-01 -3.99431735e-01 1.75131354e-02
-1.62492856e-01 5.23237050e-01 7.79258728e-01 -5.05281925e-01
-5.47988653e-01 -3.51790905e-01 -6.77707419e-02 -1.54856116e-01
-3.08600456e-01 3.23586941e-01 1.84513107e-01 -3.95064145e-01
2.34849229e-01 -9.07877147e-01 -4.30307865e-01 -1.99613883e-03
-3.06533277e-01 -3.99044394e-01 4.56465513e-01 -4.64056492e-01
9.73068714e-01 -2.04386544e+00 3.15348357e-01 5.77601194e-01
5.22644341e-01 2.98281927e-02 2.81881727e-02 2.00352535e-01
-4.45559442e-01 2.00284451e-01 -2.01710716e-01 5.79230599e-02
9.38975438e-02 -2.29640216e-01 4.14802432e-02 5.78611195e-01
6.22281432e-01 8.07016790e-01 -8.70998859e-01 -2.02131450e-01
2.43567780e-01 2.87790447e-01 -6.63936079e-01 -1.34088755e-01
-2.01111108e-01 2.54476786e-01 -6.17949963e-01 4.26514298e-01
1.93471640e-01 -5.37439764e-01 -4.77979816e-02 -1.51866376e-01
-3.69125940e-02 6.10044658e-01 -1.07328200e+00 1.60194993e+00
-3.88563007e-01 9.31108117e-01 -1.69638228e-02 -1.12003720e+00
8.09988201e-01 4.42552827e-02 8.05084229e-01 -6.77219927e-01
3.51861298e-01 4.02859271e-01 5.73585689e-01 -4.57273841e-01
5.31137049e-01 -3.44724685e-01 -1.50434345e-01 5.07563055e-01
4.77218896e-01 1.04652382e-01 2.46352494e-01 2.37882864e-02
1.29820383e+00 9.61742997e-02 2.06231564e-01 -5.92990398e-01
3.06763113e-01 2.04961225e-01 6.27224803e-01 6.40519619e-01
-2.80172765e-01 2.64161766e-01 4.94096249e-01 -8.38255048e-01
-1.08512020e+00 -6.49964213e-01 -3.37278396e-01 1.37105119e+00
-3.52382660e-01 -4.64726537e-01 -3.66229028e-01 -5.69585562e-01
2.62557149e-01 4.01841134e-01 -5.34070909e-01 -2.94995278e-01
-6.96365654e-01 -1.07543170e+00 2.37526327e-01 4.94809389e-01
1.15179094e-02 -1.12827981e+00 -5.30990541e-01 1.36380732e-01
5.33645093e-01 -5.95024407e-01 -9.09141451e-02 8.50080550e-01
-8.21835279e-01 -1.02205539e+00 -6.55793250e-02 -5.47447801e-01
4.42566782e-01 2.74627749e-02 1.45981085e+00 1.14133283e-01
-5.78142047e-01 -7.77875781e-02 -1.43426761e-01 -3.67609560e-01
-3.16306412e-01 3.85402471e-01 2.88201094e-01 -3.71380329e-01
4.86448646e-01 -5.71423650e-01 -4.94361103e-01 -8.23443979e-02
-6.51525557e-01 -2.08939105e-01 8.15354943e-01 9.85496104e-01
3.34040821e-01 2.84904867e-01 5.56127608e-01 -9.54905689e-01
3.44949335e-01 -2.28844002e-01 -6.36764586e-01 5.50615974e-02
-7.93211102e-01 7.06906080e-01 8.84661913e-01 -1.62594080e-01
-6.21420026e-01 2.60746419e-01 -3.07046026e-01 -2.24129081e-01
-1.40115440e-01 3.71858567e-01 -1.74956903e-01 -5.70580781e-01
8.49441707e-01 -3.59559029e-01 -2.20606960e-02 -5.20155787e-01
4.88093615e-01 1.48150310e-01 2.89640069e-01 -7.60253131e-01
8.99777532e-01 2.05204383e-01 4.90848333e-01 -7.13430464e-01
-8.56842518e-01 -4.35159802e-01 -7.77639568e-01 2.68907070e-01
7.33211756e-01 -5.15338957e-01 -6.64967000e-01 2.28543460e-01
-6.87377632e-01 -9.55179334e-02 -2.92838007e-01 6.02530956e-01
-3.62965167e-01 2.94532832e-02 -4.19659972e-01 -5.69557965e-01
-3.61140221e-01 -1.28071785e+00 6.79429114e-01 3.48654985e-01
-3.66739541e-01 -9.25286889e-01 1.10023052e-01 7.28293881e-02
4.48828340e-01 1.93514496e-01 1.37298775e+00 -1.11907804e+00
-3.87674749e-01 -2.24674463e-01 -8.38500038e-02 2.91264027e-01
3.25565226e-02 1.70800522e-01 -1.12204146e+00 -3.47597748e-01
-5.09971194e-02 -5.39228141e-01 1.19970202e+00 2.96978980e-01
1.52414405e+00 6.63904175e-02 -3.60389620e-01 8.36250901e-01
1.27533722e+00 4.30532619e-02 1.53013498e-01 2.33010888e-01
6.10183358e-01 6.05198503e-01 2.64667720e-01 3.82511020e-02
-3.14861536e-02 4.82962936e-01 9.90440473e-02 -1.01677522e-01
1.65898308e-01 1.01514377e-01 6.51609823e-02 7.50152171e-01
-1.22075103e-01 4.41014282e-02 -9.93918419e-01 3.29800159e-01
-1.54089475e+00 -1.09165156e+00 4.70463419e-03 2.26995158e+00
7.16856599e-01 6.90611959e-01 2.06078380e-01 1.97989419e-01
3.50477338e-01 1.05418019e-01 -8.28059375e-01 -4.77722526e-01
8.74410570e-02 9.47163880e-01 6.50720477e-01 5.28665304e-01
-1.32594645e+00 4.67498779e-01 6.75230026e+00 8.83228898e-01
-1.29183292e+00 -2.08108500e-01 9.93316829e-01 -1.66074425e-01
-3.13751817e-01 1.73006523e-02 -6.91491723e-01 2.63814807e-01
1.24199986e+00 2.33897474e-02 4.89197403e-01 9.13312018e-01
1.32745713e-01 -4.49240655e-02 -1.48824763e+00 9.82740998e-01
-2.08421320e-01 -1.55297494e+00 -3.79926175e-01 1.98668372e-02
5.10780632e-01 3.46637428e-01 5.96449804e-03 4.00681287e-01
2.94882566e-01 -1.33109164e+00 4.47823197e-01 4.57084000e-01
6.93319917e-01 -9.29356337e-01 4.34694201e-01 1.67354897e-01
-9.46411967e-01 -1.16915509e-01 -4.36904371e-01 -6.35638759e-02
-2.08149269e-01 1.00772679e+00 -5.32548785e-01 4.93070960e-01
5.95509648e-01 3.99645209e-01 -7.77102411e-01 9.58218336e-01
1.92354903e-01 5.44642627e-01 -4.87971544e-01 -2.10687757e-01
1.24681242e-01 -1.11162558e-01 2.75984019e-01 1.13993716e+00
1.86212659e-01 1.09009415e-01 -1.86258927e-02 6.53937519e-01
-1.76092103e-01 -4.34832461e-02 -5.70024312e-01 -3.92006904e-01
2.47260109e-01 1.69519556e+00 -7.87863255e-01 -2.35633850e-01
-3.68538290e-01 6.02641582e-01 5.47736943e-01 -6.66315481e-02
-4.49731410e-01 -8.11688781e-01 8.97084296e-01 -4.17816937e-02
1.86812609e-01 -4.26887125e-01 -3.56886774e-01 -9.78268862e-01
-1.08142555e-01 -8.07243347e-01 3.30364227e-01 -3.05418938e-01
-1.52984250e+00 5.34363151e-01 -2.97744513e-01 -7.74654806e-01
-4.74235117e-01 -1.20030248e+00 -7.53267109e-01 9.69826341e-01
-1.28076744e+00 -6.67425454e-01 1.58997610e-01 1.25912473e-01
2.66379207e-01 -2.62594730e-01 9.17358875e-01 3.19277465e-01
-4.08127338e-01 4.17916656e-01 4.11440641e-01 2.22754851e-01
5.65418959e-01 -1.37881708e+00 5.61182022e-01 5.62467158e-01
4.30307031e-01 6.32673979e-01 7.93869019e-01 -2.07093880e-01
-1.66202188e+00 -7.36190975e-01 9.35323298e-01 -5.76663375e-01
8.11350524e-01 -4.91582662e-01 -8.60644519e-01 4.22788471e-01
-1.26085594e-01 1.93355605e-01 7.62632191e-01 6.19685054e-01
-2.40345597e-01 -1.67878598e-01 -1.11937559e+00 1.61214381e-01
1.02024269e+00 -7.03170717e-01 -3.72644275e-01 4.20894474e-01
4.78723258e-01 -2.59762049e-01 -9.92682636e-01 4.15897757e-01
6.89963877e-01 -9.62338984e-01 9.89791274e-01 -1.18691516e+00
3.57888967e-01 -6.19793236e-02 -5.94607517e-02 -1.45582414e+00
-7.05880284e-01 -9.16388512e-01 5.07075191e-02 8.67314100e-01
7.66603172e-01 -4.99877900e-01 9.42035437e-01 9.66373622e-01
-1.31291270e-01 -1.10754514e+00 -8.34133625e-01 -6.71259999e-01
2.82905877e-01 -4.32974666e-01 5.36230028e-01 1.09854734e+00
1.47747388e-02 6.39950633e-01 -2.12344509e-02 -9.85451639e-02
3.78652781e-01 3.64655823e-01 3.54030609e-01 -1.61532331e+00
-4.88580972e-01 -6.89875245e-01 -6.00974739e-01 -5.74160278e-01
7.34526291e-02 -1.13963521e+00 -4.65324111e-02 -1.05752778e+00
1.15209103e-01 -6.47680581e-01 -6.74413264e-01 5.41596234e-01
-1.12234488e-01 -1.70960475e-03 -1.31547347e-01 -1.66214898e-01
-4.85807389e-01 4.13058430e-01 9.32940245e-01 1.06232285e-01
-8.67418572e-02 -1.43409744e-01 -7.45549023e-01 8.34658563e-01
8.84331465e-01 -5.26068211e-01 -1.91489816e-01 -2.92929828e-01
4.36376184e-01 -1.25634760e-01 2.76499838e-01 -1.11624300e+00
-5.86259216e-02 -2.71918714e-01 9.66214538e-01 -2.59249896e-01
3.85170907e-01 -5.42546868e-01 -4.64911275e-02 2.05513522e-01
-5.18317580e-01 3.56161624e-01 1.86081588e-01 2.56261289e-01
7.15604648e-02 -5.77284038e-01 7.17534065e-01 -1.98164538e-01
-7.96640754e-01 3.15900445e-01 -3.93809844e-03 1.22045569e-01
7.18532979e-01 2.55981266e-01 -3.90236288e-01 -1.13584459e-01
-5.81469417e-01 1.01224959e-01 3.65945280e-01 2.61696130e-01
3.10395330e-01 -1.25203657e+00 -4.05547708e-01 4.69273478e-01
-6.97481036e-02 -9.89444330e-02 -1.19524710e-01 6.11596644e-01
-2.72016376e-01 5.11205256e-01 -2.10232601e-01 -3.75292450e-01
-9.63700116e-01 3.44176114e-01 3.41751128e-01 -2.93595016e-01
-5.83837390e-01 1.04238045e+00 -1.23664677e-01 -3.88343066e-01
-1.10111162e-01 -2.95312524e-01 1.44148201e-01 2.02442985e-02
4.14873928e-01 2.78349549e-01 4.20009226e-01 -5.07678866e-01
-4.89885539e-01 3.37541223e-01 -1.60811424e-01 4.16559204e-02
1.63305473e+00 4.16058362e-01 -1.36861950e-01 3.98727119e-01
1.44568491e+00 9.72213522e-02 -1.17193258e+00 -5.61463684e-02
3.13912928e-01 -3.30083042e-01 3.53932798e-01 -7.15041041e-01
-1.07940733e+00 8.91396284e-01 7.18572736e-01 3.96562606e-01
8.13481629e-01 1.53110586e-02 6.30351722e-01 8.97097826e-01
2.71140397e-01 -1.20609128e+00 -1.38494208e-01 4.99736547e-01
3.88538450e-01 -1.51002669e+00 2.84207582e-01 2.63255388e-02
-2.29144230e-01 1.43300831e+00 4.25549448e-01 -2.52305120e-01
8.14291596e-01 8.05866420e-02 -1.89538598e-01 -5.54766059e-01
-7.91786909e-01 -1.05575718e-01 6.22974515e-01 1.57845080e-01
6.54541969e-01 -3.57329100e-02 -1.27256997e-02 5.70630193e-01
-4.27403003e-01 -2.88578182e-01 1.76170707e-01 9.09194350e-01
-7.85067976e-01 -1.34897709e+00 -5.99804036e-02 1.22559011e+00
-6.03973091e-01 -1.03824943e-01 -1.66759446e-01 8.49048555e-01
8.45992342e-02 6.61521912e-01 -9.73043311e-03 -5.59333861e-01
-4.12941463e-02 5.34843445e-01 5.81492066e-01 -5.33350885e-01
-7.57657826e-01 -3.40874016e-01 2.52898365e-01 -6.86023474e-01
-2.68221140e-01 -7.19025731e-01 -1.22537851e+00 -4.41757560e-01
-4.17374372e-01 2.56248087e-01 8.14585805e-01 1.12460589e+00
3.33394349e-01 5.26060998e-01 6.70049369e-01 -1.13309181e+00
-7.24349260e-01 -7.51648247e-01 -5.03517866e-01 4.46386725e-01
5.56775868e-01 -7.84069121e-01 -4.97333527e-01 -3.22121024e-01] | [8.113638877868652, 4.39579963684082] |
726047fb-3dfa-47ff-92ae-708e8dca33c2 | representing-prior-knowledge-using-randomly | 2111.10686 | null | https://arxiv.org/abs/2111.10686v2 | https://arxiv.org/pdf/2111.10686v2.pdf | Representing Prior Knowledge Using Randomly, Weighted Feature Networks for Visual Relationship Detection | The single-hidden-layer Randomly Weighted Feature Network (RWFN) introduced by Hong and Pavlic (2021) was developed as an alternative to neural tensor network approaches for relational learning tasks. Its relatively small footprint combined with the use of two randomized input projections -- an insect-brain-inspired input representation and random Fourier features -- allow it to achieve rich expressiveness for relational learning with relatively low training cost. In particular, when Hong and Pavlic compared RWFN to Logic Tensor Networks (LTNs) for Semantic Image Interpretation (SII) tasks to extract structured semantic descriptions from images, they showed that the RWFN integration of the two hidden, randomized representations better captures relationships among inputs with a faster training process even though it uses far fewer learnable parameters. In this paper, we use RWFNs to perform Visual Relationship Detection (VRD) tasks, which are more challenging SII tasks. A zero-shot learning approach is used with RWFN that can exploit similarities with other seen relationships and background knowledge -- expressed with logical constraints between subjects, relations, and objects -- to achieve the ability to predict triples that do not appear in the training set. The experiments on the Visual Relationship Dataset to compare the performance between RWFNs and LTNs, one of the leading Statistical Relational Learning frameworks, show that RWFNs outperform LTNs for the predicate-detection task while using fewer number of adaptable parameters (1:56 ratio). Furthermore, background knowledge represented by RWFNs can be used to alleviate the incompleteness of training sets even though the space complexity of RWFNs is much smaller than LTNs (1:27 ratio). | ['Theodore P. Pavlic', 'Jinyung Hong'] | 2021-11-20 | null | https://openreview.net/forum?id=iwoNpozn10l | https://openreview.net/pdf?id=iwoNpozn10l | aaai-workshop-clear-2022-2 | ['visual-relationship-detection', 'tensor-networks', 'predicate-detection', 'relational-reasoning'] | ['computer-vision', 'methodology', 'natural-language-processing', 'natural-language-processing'] | [ 1.48553178e-01 5.40694714e-01 -4.09967571e-01 -4.31423873e-01
1.72770828e-01 -1.15037896e-01 7.55192876e-01 -3.66695151e-02
-2.00411722e-01 4.67759073e-01 1.64167523e-01 -3.42987865e-01
-5.78999221e-01 -1.08412743e+00 -5.97146690e-01 -4.04534012e-01
-2.40005270e-01 5.40842772e-01 3.27096581e-01 -3.85931790e-01
-2.31959652e-02 4.92403924e-01 -1.81974649e+00 7.12072313e-01
4.13887739e-01 1.22854590e+00 1.25585824e-01 3.30155015e-01
-2.52078295e-01 1.35447538e+00 -1.47932649e-01 -5.78542531e-01
1.10161394e-01 -9.17828381e-02 -1.15150952e+00 -1.46768734e-01
3.18713129e-01 -2.19963893e-01 -5.44800580e-01 6.64840877e-01
-1.07636839e-01 2.89596498e-01 6.29902720e-01 -1.43603492e+00
-1.05416429e+00 9.03882325e-01 -2.66434342e-01 1.50929138e-01
4.15275395e-01 -1.53896317e-01 1.56116831e+00 -1.03048170e+00
7.92238235e-01 1.51653576e+00 5.33064485e-01 4.02094424e-01
-1.33443189e+00 -4.91047621e-01 2.34545209e-02 6.68396831e-01
-1.37592602e+00 -1.75257131e-01 7.96154022e-01 -3.69617760e-01
1.41545630e+00 3.28866452e-01 6.16459012e-01 8.88300657e-01
-2.22411022e-01 8.33479404e-01 7.27030635e-01 -5.95331609e-01
1.69500381e-01 2.38691896e-01 5.18587112e-01 1.09656215e+00
4.20516431e-01 -5.25659770e-02 -8.42775464e-01 -1.43301934e-01
7.76898444e-01 4.07761149e-02 -1.12286888e-01 -6.38966501e-01
-1.35815144e+00 9.40998435e-01 9.73868847e-01 6.52116060e-01
-1.69257134e-01 1.70989007e-01 4.40325886e-01 3.89411569e-01
1.85782745e-01 5.48834562e-01 -3.75929445e-01 3.16530079e-01
-4.96713728e-01 1.12978093e-01 7.03923464e-01 9.33531702e-01
9.14946079e-01 1.27954617e-01 -1.95483565e-01 7.31505513e-01
2.62032092e-01 2.96426445e-01 4.63557869e-01 -9.18880820e-01
4.96358514e-01 1.04070032e+00 -4.17025656e-01 -1.45493782e+00
-4.65250194e-01 -1.07556023e-01 -8.06583047e-01 -1.71876341e-01
4.01836902e-01 4.34033483e-01 -7.57012486e-01 1.48441362e+00
1.45106480e-01 -3.30353171e-01 2.88843483e-01 8.83072019e-01
1.13856375e+00 5.73482692e-01 -7.15738386e-02 8.39544460e-02
1.53718531e+00 -7.15115547e-01 -6.03367925e-01 -1.02959976e-01
8.84489894e-01 -2.39133239e-01 1.03343070e+00 1.65644377e-01
-7.45298505e-01 -5.37646592e-01 -1.11316681e+00 -2.66792983e-01
-7.46671140e-01 6.46332875e-02 1.53695130e+00 3.73570472e-01
-9.80618238e-01 7.08763778e-01 -5.75483203e-01 -5.69179654e-01
6.99962556e-01 6.06969953e-01 -6.58609092e-01 -1.20494820e-01
-1.23702335e+00 1.03422606e+00 7.35149562e-01 9.32009816e-02
-5.08121312e-01 -6.56442583e-01 -1.07073081e+00 2.40838081e-01
7.18062282e-01 -4.68682438e-01 6.95165157e-01 -6.77367389e-01
-1.21609545e+00 7.13865638e-01 -4.57461998e-02 -5.36545813e-01
6.56013489e-02 8.73108804e-02 -2.17000321e-01 3.18714470e-01
-3.54731008e-02 7.12369204e-01 6.54301047e-01 -1.17196727e+00
-1.42446563e-01 -3.55715692e-01 4.41608608e-01 -1.58636421e-02
-5.57563484e-01 7.55784363e-02 -1.43756941e-01 -3.50598872e-01
4.22026992e-01 -7.92457938e-01 1.19513951e-01 1.15113646e-01
-4.93666023e-01 -6.32021427e-01 1.04539549e+00 -2.15508640e-01
7.97230661e-01 -2.10959625e+00 4.08213556e-01 3.65864754e-01
5.67801356e-01 3.55893850e-01 -1.22855097e-01 2.99936473e-01
-4.62605059e-01 9.34113115e-02 -2.65608281e-02 7.21087456e-02
-7.68801942e-02 6.91524684e-01 -3.51027340e-01 1.35289073e-01
3.87629747e-01 1.10121429e+00 -9.91152465e-01 -5.14183939e-01
2.40985692e-01 4.18443054e-01 -4.40790623e-01 1.76500827e-01
-2.97310680e-01 -7.02167824e-02 -2.50787258e-01 6.96639180e-01
3.35327566e-01 -4.00953472e-01 3.99576962e-01 -7.60396123e-01
2.16500595e-01 8.85412246e-02 -1.08446562e+00 1.32423270e+00
-4.02446926e-01 6.92746282e-01 -5.59139431e-01 -1.36101508e+00
1.07602668e+00 3.86329293e-01 4.40827429e-01 -7.32057691e-01
1.72181576e-01 3.37938406e-02 7.50827193e-02 -5.99954128e-01
1.62191793e-01 -1.00379907e-01 2.94828974e-02 4.12685573e-01
5.97012639e-01 2.46150531e-02 4.18155909e-01 3.69003534e-01
1.07338977e+00 3.14088464e-01 3.57037902e-01 -1.47841007e-01
7.12664902e-01 -2.13680137e-02 5.76739609e-01 4.87069905e-01
2.04264343e-01 1.08875334e-01 8.33928049e-01 -1.10488319e+00
-8.39977086e-01 -1.07012808e+00 -9.18344483e-02 1.04735839e+00
9.90630090e-02 -7.71109641e-01 -1.27346948e-01 -6.73337162e-01
6.26778752e-02 7.16085732e-01 -6.69973910e-01 -1.97376639e-01
-4.88090694e-01 -6.08050764e-01 4.66063142e-01 6.55770600e-01
4.66976255e-01 -1.12634015e+00 -7.65987873e-01 -2.52945751e-01
-8.02617744e-02 -1.39821851e+00 3.92961591e-01 4.26983386e-01
-8.03650439e-01 -1.27567422e+00 -1.48399889e-01 -6.37809873e-01
8.06629300e-01 1.48563713e-01 9.89496350e-01 1.32318541e-01
-3.77814889e-01 2.69343495e-01 -4.87805903e-01 -1.91256121e-01
-1.08526379e-01 -3.06024373e-01 1.83310658e-01 7.37066716e-02
4.44866091e-01 -5.36941886e-01 5.88036105e-02 3.24650139e-01
-9.37274814e-01 2.49819577e-01 3.78224611e-01 1.11021483e+00
1.95477620e-01 2.36208871e-01 1.91004440e-01 -1.05409420e+00
2.62768209e-01 -1.92159832e-01 -4.87976938e-01 4.13297802e-01
-4.90275800e-01 5.74934006e-01 6.32251799e-01 -4.94651169e-01
-1.04495621e+00 -1.47960916e-01 5.47952533e-01 -5.81120014e-01
2.49904633e-01 4.83350903e-01 1.89671908e-02 -2.88767010e-01
7.11071491e-01 -1.86076477e-01 1.95961492e-03 -3.00978363e-01
6.30586445e-01 3.07798803e-01 2.32386336e-01 -5.78625202e-01
8.57494831e-01 4.11230415e-01 3.71056199e-01 -6.57397926e-01
-1.14117599e+00 -2.36000791e-01 -8.50508749e-01 4.07105125e-02
8.42800140e-01 -6.93195045e-01 -1.03664231e+00 -1.31729290e-01
-1.19699907e+00 4.75562271e-03 -4.40840989e-01 5.96979022e-01
-5.11573434e-01 1.10276259e-01 -6.16050363e-01 -7.78649747e-01
-1.11448765e-01 -9.44897950e-01 5.76941729e-01 -5.56548052e-02
-2.85666347e-01 -9.71530020e-01 -4.71452206e-01 6.00542903e-01
4.35493380e-01 1.52013019e-01 1.69068909e+00 -9.28957343e-01
-6.25953972e-01 1.20872902e-02 -6.00825906e-01 3.66259277e-01
5.87048531e-02 -3.93170193e-02 -9.94137883e-01 1.34254590e-01
-1.02603957e-01 -5.60258746e-01 8.72268319e-01 -1.21458419e-01
1.15330243e+00 -4.46720928e-01 -2.08074227e-01 4.49206531e-01
1.42263222e+00 7.73803964e-02 4.50386494e-01 2.65603155e-01
9.83505130e-01 7.94793963e-01 4.67316687e-01 1.77719757e-01
4.88661051e-01 7.31853366e-01 5.39708793e-01 -6.47675339e-03
-2.71079481e-01 -1.49133101e-01 1.89007252e-01 6.75890326e-01
-6.73159122e-01 2.40935057e-01 -9.79101181e-01 2.84807652e-01
-2.03428721e+00 -1.12592483e+00 -7.99723938e-02 2.09172678e+00
7.10052490e-01 1.68738291e-01 -2.88804285e-02 4.24113005e-01
3.67580950e-01 1.26961887e-01 -1.30838826e-01 -5.29786825e-01
-8.45814273e-02 2.79864430e-01 3.78774703e-01 2.36281261e-01
-9.09628749e-01 1.07074273e+00 5.98565960e+00 5.55157363e-01
-8.54815722e-01 1.68992113e-02 3.49755228e-01 4.79230992e-02
-3.26614916e-01 2.10672781e-01 -6.36032760e-01 -1.15296520e-01
9.82382476e-01 1.25859961e-01 6.31497085e-01 8.59681249e-01
-3.17937523e-01 -7.35666752e-02 -1.45780659e+00 1.10764670e+00
2.19121009e-01 -1.62961757e+00 2.49863192e-01 -1.20443910e-01
3.01148564e-01 -2.18820751e-01 -1.16991431e-01 5.37889004e-01
1.78307205e-01 -1.16285872e+00 5.85305393e-01 6.04975581e-01
6.22671485e-01 -6.05644703e-01 9.80281472e-01 -2.33952999e-02
-1.16652739e+00 -4.27194923e-01 -7.71043599e-01 -2.85199940e-01
-2.47083962e-01 4.94614452e-01 -1.03302908e+00 7.99574733e-01
8.03637564e-01 7.10480273e-01 -6.90998793e-01 3.81990492e-01
-2.15036795e-01 2.57847190e-01 -3.07230949e-01 -1.21422574e-01
1.30176723e-01 -3.26590948e-02 3.40628773e-01 8.69822919e-01
-7.90981576e-02 1.15337502e-02 1.95400015e-01 1.22430301e+00
3.02740335e-02 -1.11483902e-01 -9.17705238e-01 -2.33336359e-01
2.40695730e-01 1.34420395e+00 -6.44150674e-01 -2.29365125e-01
-4.64050919e-01 5.17235160e-01 5.46924412e-01 3.60982209e-01
-7.34876990e-01 -3.34556252e-01 3.82520527e-01 1.05490446e-01
4.53762412e-01 -2.15813980e-01 -1.54624298e-01 -1.20776629e+00
-7.74339736e-02 -5.55965304e-01 4.32811916e-01 -8.68546963e-01
-1.13787341e+00 8.14329982e-01 4.78998452e-01 -1.04779541e+00
-2.54851967e-01 -1.12096155e+00 -1.86999857e-01 5.77788115e-01
-1.44445634e+00 -1.51647842e+00 -1.96408018e-01 7.97671258e-01
1.70270666e-01 -4.29421097e-01 1.13666487e+00 5.00519723e-02
-5.75306594e-01 3.92662674e-01 -6.51617348e-01 4.44486797e-01
1.41702443e-01 -1.08370841e+00 -2.16619715e-01 4.66464490e-01
4.21790898e-01 9.06564236e-01 4.46849376e-01 -4.24468219e-01
-1.65753353e+00 -8.08338881e-01 1.12900782e+00 -2.12884501e-01
1.00261819e+00 -3.88829827e-01 -8.90259981e-01 8.62120152e-01
-4.91974130e-03 5.79243898e-01 7.96134412e-01 3.96521300e-01
-8.97417545e-01 -3.48487496e-01 -1.05104506e+00 5.84747434e-01
1.17054236e+00 -7.75481641e-01 -8.59220326e-01 5.34989953e-01
9.69002962e-01 1.55697269e-02 -1.18398213e+00 4.32597578e-01
3.66305679e-01 -9.56855059e-01 1.17219269e+00 -6.89633608e-01
4.58528072e-01 -3.06208193e-01 -4.93203491e-01 -8.21467280e-01
-5.10538936e-01 -2.91294307e-01 -3.90052617e-01 1.13806176e+00
4.28707212e-01 -5.73562741e-01 5.83816230e-01 6.19213223e-01
3.03093106e-01 -9.32781875e-01 -6.99666440e-01 -8.76999617e-01
-5.38500547e-01 -4.39820379e-01 4.66279894e-01 9.57659781e-01
2.71298081e-01 6.88131452e-01 -2.89049417e-01 -2.59178221e-01
5.28844118e-01 1.67905927e-01 3.98411840e-01 -1.44177079e+00
-2.19814390e-01 -9.23163742e-02 -1.00329649e+00 -2.39267945e-01
3.95166397e-01 -1.21228933e+00 -3.91622901e-01 -1.42380619e+00
1.75824851e-01 -5.01950741e-01 -4.33169216e-01 1.16876686e+00
2.43786782e-01 1.91317514e-01 4.65441972e-01 2.13492617e-01
-4.20434594e-01 5.32128453e-01 1.12205112e+00 -3.53819311e-01
-3.63824554e-02 -1.79289415e-01 -5.99507809e-01 7.97519028e-01
3.79457861e-01 -3.55263114e-01 -7.90387392e-01 -2.53119290e-01
5.62769294e-01 -2.77274046e-02 5.49971819e-01 -9.05639112e-01
2.16218114e-01 -3.98224294e-02 4.49759632e-01 -4.69908565e-01
5.25246918e-01 -9.73131835e-01 6.41900375e-02 3.55116934e-01
-3.68516117e-01 3.73161957e-02 -7.13334531e-02 4.87133026e-01
-2.27293447e-01 -3.44403267e-01 3.18186939e-01 -3.93928081e-01
-1.01584792e+00 4.27508987e-02 -7.02113733e-02 -3.23254228e-01
1.02642584e+00 -3.21322680e-01 -4.93667156e-01 -1.08087640e-02
-7.92688251e-01 -1.23986974e-01 -4.56046425e-02 3.47377181e-01
8.72384369e-01 -1.40777075e+00 -2.83159912e-01 3.09649020e-01
3.64153922e-01 9.49750319e-02 -1.17043607e-01 9.01077390e-01
-4.82145846e-01 7.00640798e-01 -3.75690520e-01 -6.13538265e-01
-1.13451362e+00 8.43880236e-01 7.31104463e-02 -3.64501417e-01
-9.47945714e-01 9.85746384e-01 -5.11320755e-02 -5.78222096e-01
2.10586324e-01 -4.24331754e-01 -4.99355078e-01 1.67526841e-01
5.21026969e-01 1.72357827e-01 4.20456678e-02 -5.56224525e-01
-5.76802909e-01 3.85021478e-01 -4.41667512e-02 2.47401357e-01
1.65685475e+00 3.38289618e-01 -6.19724154e-01 6.96871996e-01
1.14500678e+00 -4.57944095e-01 -7.04216301e-01 -7.24092007e-01
4.68313307e-01 -4.85311687e-01 1.38214707e-01 -5.64235806e-01
-9.84198272e-01 7.66795576e-01 2.74288028e-01 2.92309165e-01
8.96455109e-01 2.29937553e-01 3.14915985e-01 6.94351315e-01
5.24334252e-01 -8.38500440e-01 3.35855812e-01 4.94000673e-01
8.32228899e-01 -1.15241146e+00 2.89124638e-01 -7.47857988e-01
-8.90080214e-01 1.49716783e+00 6.25238121e-01 3.66716795e-02
4.55103815e-01 3.63081694e-02 -4.42638665e-01 -5.68646312e-01
-9.03462470e-01 -3.56779069e-01 5.24654031e-01 6.34637415e-01
3.74303997e-01 2.37324350e-02 9.94500220e-02 3.10466349e-01
-2.57655740e-01 -1.33532686e-02 3.11684608e-01 7.51775205e-01
-1.68590158e-01 -1.06369269e+00 -1.51670933e-01 5.24128914e-01
1.12922140e-03 -1.42272726e-01 -1.37029752e-01 8.70797455e-01
3.77865076e-01 8.48858416e-01 1.37094676e-01 -6.09579265e-01
1.77315891e-01 9.59626362e-02 6.07480764e-01 -7.37542927e-01
-4.85602736e-01 -4.85424668e-01 3.41576040e-01 -9.96775508e-01
-7.76380420e-01 -2.88504273e-01 -1.40566599e+00 -2.57813990e-01
-2.91543305e-01 -5.09486198e-02 4.18305546e-01 1.25763607e+00
5.72888292e-02 5.54167092e-01 2.86290020e-01 -5.75549960e-01
-4.69518125e-01 -6.96259558e-01 -4.89783227e-01 5.09685338e-01
6.59952462e-02 -1.04788792e+00 -1.44996911e-01 -6.46754652e-02] | [10.349385261535645, 2.288921594619751] |
cb7e33a3-fb2a-4309-bcd4-7ff7e9a24746 | read-highlight-and-summarize-a-hierarchical | 1910.03177 | null | https://arxiv.org/abs/1910.03177v2 | https://arxiv.org/pdf/1910.03177v2.pdf | Read, Highlight and Summarize: A Hierarchical Neural Semantic Encoder-based Approach | Traditional sequence-to-sequence (seq2seq) models and other variations of the attention-mechanism such as hierarchical attention have been applied to the text summarization problem. Though there is a hierarchy in the way humans use language by forming paragraphs from sentences and sentences from words, hierarchical models have usually not worked that much better than their traditional seq2seq counterparts. This effect is mainly because either the hierarchical attention mechanisms are too sparse using hard attention or noisy using soft attention. In this paper, we propose a method based on extracting the highlights of a document; a key concept that is conveyed in a few sentences. In a typical text summarization dataset consisting of documents that are 800 tokens in length (average), capturing long-term dependencies is very important, e.g., the last sentence can be grouped with the first sentence of a document to form a summary. LSTMs (Long Short-Term Memory) proved useful for machine translation. However, they often fail to capture long-term dependencies while modeling long sequences. To address these issues, we have adapted Neural Semantic Encoders (NSE) to text summarization, a class of memory-augmented neural networks by improving its functionalities and proposed a novel hierarchical NSE that outperforms similar previous models significantly. The quality of summarization was improved by augmenting linguistic factors, namely lemma, and Part-of-Speech (PoS) tags, to each word in the dataset for improved vocabulary coverage and generalization. The hierarchical NSE model on factored dataset outperformed the state-of-the-art by nearly 4 ROUGE points. We further designed and used the first GPU-based self-critical Reinforcement Learning model. | ['Rajeev Bhatt Ambati', 'Prasenjit Mitra', 'Saptarashmi Bandyopadhyay'] | 2019-10-08 | null | null | null | null | ['hard-attention'] | ['methodology'] | [ 4.43804532e-01 2.88821608e-01 -1.69491395e-01 -2.80568093e-01
-7.55224645e-01 -2.95823097e-01 4.53783929e-01 3.49822581e-01
-5.53741693e-01 1.14208198e+00 9.82080519e-01 -1.28822774e-01
1.51749790e-01 -7.62891352e-01 -9.39861834e-01 -5.05394042e-01
6.37640432e-02 4.64916140e-01 1.41161799e-01 -5.79694569e-01
7.42960691e-01 6.42421991e-02 -1.53636098e+00 5.98179162e-01
1.20895624e+00 4.70745355e-01 7.49428928e-01 8.57456684e-01
-7.29719877e-01 8.02653909e-01 -1.00935173e+00 -2.35691831e-01
-2.11578950e-01 -7.88659573e-01 -9.26663280e-01 -1.35521337e-01
4.66227740e-01 -3.72004211e-01 -2.22127214e-01 1.00779140e+00
7.35349238e-01 3.83917511e-01 5.58572769e-01 -5.56153238e-01
-8.91591489e-01 1.10765159e+00 -5.19056082e-01 2.56800324e-01
4.86799270e-01 -2.13175595e-01 1.28632057e+00 -7.31389105e-01
5.92184424e-01 1.20899785e+00 5.57137847e-01 7.03559339e-01
-7.98035562e-01 -2.28880987e-01 1.21739961e-01 2.39866406e-01
-6.69107735e-01 -4.10289019e-01 4.29882139e-01 -1.14109695e-01
1.79783499e+00 5.16288221e-01 5.43755293e-01 1.27253902e+00
7.20580459e-01 1.07797813e+00 4.83393937e-01 -4.42216039e-01
1.87328190e-01 -2.04803720e-02 5.30348897e-01 4.75879043e-01
2.39575326e-01 -4.67939794e-01 -7.09000885e-01 1.28988428e-02
4.25332397e-01 2.76692715e-02 -1.52629122e-01 3.11544836e-01
-1.24178696e+00 1.02180636e+00 4.43495750e-01 6.37222886e-01
-6.09177589e-01 7.87260309e-02 9.25234973e-01 1.66459948e-01
5.21393120e-01 7.21280456e-01 -4.64167058e-01 -2.15021476e-01
-1.16762960e+00 1.11854509e-01 8.54050517e-01 9.87826169e-01
5.91635108e-01 3.35219920e-01 -6.66345000e-01 9.54389274e-01
-2.77883798e-01 2.95349568e-01 1.15374589e+00 -6.07706189e-01
5.67251921e-01 5.35264432e-01 -1.64570525e-01 -8.09746742e-01
-4.51467276e-01 -6.01612806e-01 -1.37387180e+00 -5.73336720e-01
-3.19239229e-01 -2.33969048e-01 -9.21933711e-01 1.57845271e+00
-2.54945457e-01 -2.43047386e-01 3.80845159e-01 7.16625690e-01
1.18158460e+00 1.21131754e+00 -1.27982870e-01 -5.69944263e-01
1.28369534e+00 -1.37795019e+00 -1.01329219e+00 -3.36854994e-01
6.87276900e-01 -5.62143505e-01 1.06874430e+00 2.41972551e-01
-1.26622283e+00 -7.24818707e-01 -9.99017954e-01 -2.40956962e-01
-3.67975026e-01 -1.39547139e-01 4.13553327e-01 2.60644615e-01
-1.23003018e+00 9.78708625e-01 -5.68704963e-01 -7.00988889e-01
1.88416392e-01 4.11366731e-01 -2.79874682e-01 3.02741319e-01
-1.42680728e+00 9.11441505e-01 8.00393283e-01 -5.46202026e-02
-4.92265821e-01 -3.22253466e-01 -9.45381641e-01 5.91697216e-01
2.83926219e-01 -1.18243015e+00 1.27502978e+00 -1.19235742e+00
-1.71114767e+00 3.19486290e-01 -3.89733046e-01 -8.63140166e-01
1.96761042e-02 -5.80069482e-01 -7.71938488e-02 2.22802699e-01
2.37027913e-01 7.89630294e-01 8.32998276e-01 -8.66016686e-01
-4.93472666e-01 -3.08642000e-01 -1.18303008e-01 3.86935353e-01
-5.29053569e-01 9.35240909e-02 -1.25257045e-01 -8.50609004e-01
-3.43239576e-01 -5.51513731e-01 -3.63014162e-01 -1.06438851e+00
-6.32326007e-01 -4.31610376e-01 5.81208646e-01 -1.07654238e+00
1.56789184e+00 -1.74667454e+00 4.03024495e-01 -5.46260357e-01
-1.24192186e-01 6.42135680e-01 -2.85191864e-01 8.92819762e-01
-3.09681543e-03 2.76438296e-01 -4.66915876e-01 -4.78007913e-01
-5.99764697e-02 2.21844748e-01 -5.85960746e-01 -1.70689255e-01
2.33366981e-01 1.19835806e+00 -9.82476175e-01 -5.96670389e-01
-8.67346674e-02 1.56235501e-01 -5.16431987e-01 2.86126018e-01
-3.53337616e-01 1.26901075e-01 -3.53672385e-01 1.81833759e-01
2.82214344e-01 3.83867994e-02 -1.25429288e-01 2.89824158e-02
-1.52406290e-01 4.11682069e-01 -5.18724859e-01 2.06260061e+00
-5.21912992e-01 5.90060234e-01 -2.24299073e-01 -1.27494586e+00
9.59805846e-01 3.30598623e-01 2.29661614e-01 -8.12141359e-01
1.39600322e-01 1.83385953e-01 -7.57407025e-02 -7.26019323e-01
1.19048738e+00 -1.42640740e-01 -2.72551864e-01 3.48063409e-01
2.80961424e-01 -1.44364983e-01 5.52759230e-01 4.33433563e-01
1.13567579e+00 1.17799990e-01 6.76106513e-01 -2.37346321e-01
5.73311687e-01 5.97375855e-02 4.56368417e-01 9.67363298e-01
2.03492060e-01 7.49225378e-01 5.59406161e-01 -1.95287660e-01
-1.19833398e+00 -5.18265903e-01 3.68218333e-01 1.30113256e+00
-2.34144449e-01 -6.40530884e-01 -1.10825264e+00 -6.45529985e-01
-3.99042308e-01 1.13514507e+00 -3.93585205e-01 -2.36857787e-01
-6.65439725e-01 -6.77905560e-01 5.74121416e-01 6.67562127e-01
5.67387462e-01 -1.59811640e+00 -5.57451844e-01 5.97038984e-01
-4.36156422e-01 -9.20819938e-01 -6.81694388e-01 3.04629624e-01
-1.03572524e+00 -4.39997405e-01 -9.10533726e-01 -8.08899939e-01
3.71254086e-01 1.23517908e-01 1.16190827e+00 -1.36979371e-01
5.44678271e-02 -4.95542437e-02 -7.65141010e-01 -4.79288995e-01
-6.14642501e-01 6.14458025e-01 -3.07794660e-03 -1.41614676e-01
1.57557636e-01 -4.90936011e-01 -3.49806696e-01 -3.37070733e-01
-1.15584874e+00 2.41166398e-01 1.04045427e+00 1.17103946e+00
1.32219821e-01 -2.90444493e-01 1.14667261e+00 -9.12756443e-01
1.06295800e+00 -3.17539185e-01 5.35652526e-02 1.38917580e-01
-1.42901346e-01 3.32927465e-01 1.16139913e+00 -1.93253234e-01
-1.06929016e+00 -2.34167099e-01 -4.42856103e-01 -9.15572047e-03
-1.34515136e-01 6.84326112e-01 -2.91576004e-03 6.62285864e-01
4.77159351e-01 6.69322789e-01 5.87954782e-02 -2.96893239e-01
2.00109676e-01 1.07179070e+00 3.82782787e-01 -2.24802926e-01
9.38395932e-02 1.77582055e-02 -2.46252060e-01 -1.27025890e+00
-9.60071862e-01 -5.71525037e-01 -5.89018404e-01 2.97459006e-01
9.26547885e-01 -6.62533939e-01 -4.17375416e-01 3.07360977e-01
-1.55057991e+00 -1.39551818e-01 -3.90811712e-01 3.77453119e-01
-6.54490352e-01 8.06765020e-01 -7.43991733e-01 -6.59229934e-01
-1.15486538e+00 -9.27034914e-01 1.26028919e+00 2.61162549e-01
-5.55425167e-01 -8.62818956e-01 1.00398161e-01 2.63417125e-01
7.11715102e-01 -9.03635174e-02 1.01618111e+00 -9.97299671e-01
-6.00644713e-03 1.68839265e-02 -6.86538965e-02 7.28774905e-01
3.08565851e-02 -2.38359749e-01 -6.45134389e-01 -2.91103274e-01
1.79112956e-01 -2.66297609e-01 1.38609898e+00 6.08485460e-01
1.08766544e+00 -6.44953549e-01 -4.09195684e-02 1.38627589e-01
1.17735291e+00 6.46888241e-02 8.17433476e-01 3.13489437e-01
5.84895015e-01 5.98803937e-01 4.73698616e-01 2.70063281e-01
1.41925424e-01 3.95106167e-01 2.57924914e-01 -5.47693036e-02
-1.80528089e-02 -1.97628871e-01 8.25329602e-01 1.44054806e+00
3.02652139e-02 -6.25512481e-01 -4.83014762e-01 5.84466279e-01
-2.06693673e+00 -1.08499157e+00 -2.70216763e-01 1.97840440e+00
8.78455043e-01 2.08367944e-01 -8.27293247e-02 -1.47206172e-01
7.23192990e-01 3.49593312e-01 -3.97089064e-01 -1.09835434e+00
-2.61213422e-01 1.45588025e-01 4.31442261e-01 4.73727435e-01
-8.00682604e-01 1.17357576e+00 5.66683531e+00 1.18300140e+00
-1.10353887e+00 -1.66254546e-02 3.84998769e-01 -7.76438043e-02
-2.90312678e-01 -2.07394734e-01 -9.20077264e-01 5.41601598e-01
1.27658248e+00 -2.40789518e-01 1.41242325e-01 5.44372916e-01
3.69403630e-01 -2.47169703e-01 -9.66092706e-01 6.74410284e-01
5.43090701e-01 -1.32359874e+00 6.19207978e-01 -1.80122763e-01
8.65815282e-01 4.89136726e-02 -4.07060236e-01 7.00000286e-01
-4.58238041e-03 -1.06346822e+00 3.63954991e-01 5.45181334e-01
5.73986173e-01 -7.37656653e-01 1.20016789e+00 7.49432623e-01
-7.35288799e-01 -6.96713626e-02 -6.64620578e-01 -2.20261425e-01
2.88992316e-01 5.82471430e-01 -7.92394578e-01 1.01700735e+00
3.93360198e-01 8.11552882e-01 -4.63836223e-01 6.66711926e-01
-6.24243431e-02 6.24382436e-01 -3.90797183e-02 -5.32225132e-01
6.86817050e-01 -2.49222875e-01 7.94887781e-01 1.59125173e+00
4.69826341e-01 4.99134790e-03 -4.54715528e-02 5.58891654e-01
-1.38818204e-01 4.05628890e-01 -6.74782515e-01 -1.27412543e-01
-9.25332867e-03 1.14553082e+00 -5.99641800e-01 -7.65239537e-01
-3.20728928e-01 1.37298417e+00 2.96126515e-01 2.95450211e-01
-7.43240058e-01 -7.54350603e-01 1.68948710e-01 -6.88996464e-02
5.31153083e-01 -4.13300693e-02 -1.79081649e-01 -1.20224798e+00
2.49615423e-02 -9.23386276e-01 1.13556445e-01 -7.65045524e-01
-1.04472423e+00 8.79101217e-01 -1.54756799e-01 -8.95386100e-01
-4.16637510e-01 -1.76493421e-01 -7.16069698e-01 6.70446873e-01
-1.31708169e+00 -9.68627930e-01 -1.48742022e-02 1.35730803e-01
1.35958493e+00 -3.13956827e-01 9.68481302e-01 -7.19862878e-02
-6.53241754e-01 2.73355037e-01 3.55516225e-01 -1.02562718e-01
7.71096945e-01 -1.31693649e+00 6.72910035e-01 7.88568377e-01
7.01564029e-02 6.25150263e-01 8.11594009e-01 -8.00571740e-01
-1.21493912e+00 -1.18974876e+00 1.22847295e+00 -7.35565573e-02
5.17467141e-01 -3.30865622e-01 -1.02976549e+00 6.67605519e-01
9.34015572e-01 -7.22185194e-01 5.25122106e-01 -1.45769030e-01
1.86417475e-01 9.15059894e-02 -6.84767663e-01 5.93046010e-01
9.70978796e-01 -1.72928870e-01 -1.22470880e+00 2.44455948e-01
1.18976235e+00 -1.96674630e-01 -4.45000857e-01 3.08555007e-01
2.46278033e-01 -1.10019791e+00 5.02701998e-01 -7.28568077e-01
8.87387991e-01 1.17303088e-01 -2.97282226e-02 -1.79132295e+00
-3.30180049e-01 -6.61701024e-01 -2.61307180e-01 1.33672869e+00
2.99193025e-01 -5.51054358e-01 5.54841757e-01 -1.45652816e-01
-7.25432336e-01 -8.88945162e-01 -6.78399384e-01 -7.32138515e-01
1.55198008e-01 8.29417557e-02 4.14650530e-01 5.14484704e-01
1.79758862e-01 1.10423148e+00 -4.85337794e-01 -3.87839615e-01
2.19508499e-01 -1.20107951e-02 6.13640368e-01 -9.65000391e-01
-1.72214746e-01 -7.07652688e-01 -1.22400805e-01 -1.32038128e+00
5.12459517e-01 -9.02136981e-01 1.71777278e-01 -2.06738329e+00
5.30892074e-01 4.99355614e-01 -2.95685232e-02 2.42969349e-01
-4.69709098e-01 -2.41225570e-01 1.23445563e-01 -3.12375315e-02
-7.44819283e-01 9.59400117e-01 1.13363004e+00 -2.71282434e-01
-2.79939145e-01 -1.23397194e-01 -8.27971339e-01 4.81175840e-01
9.70194519e-01 -3.29287022e-01 -1.80718288e-01 -5.03291786e-01
1.62534386e-01 2.87633359e-01 -1.47339895e-01 -8.34126711e-01
4.97675180e-01 2.90383637e-01 9.01462063e-02 -1.07705963e+00
5.70382550e-02 -3.98801655e-01 -3.15385133e-01 5.71334720e-01
-6.81400776e-01 1.42222300e-01 2.25548625e-01 5.22140205e-01
-5.19935012e-01 -6.74624622e-01 3.74166161e-01 -4.55758065e-01
-6.01761103e-01 -8.70916173e-02 -7.23682165e-01 -2.30475944e-02
7.50366747e-01 -1.86943591e-01 -2.59648323e-01 -4.89001513e-01
-4.50824291e-01 3.21299374e-01 1.53404802e-01 5.33872068e-01
6.03402197e-01 -9.18127716e-01 -1.13360941e+00 5.01951613e-02
-1.91805795e-01 2.27911308e-01 4.02045608e-01 7.96638489e-01
-4.64147836e-01 9.69442725e-01 -2.96859056e-01 -4.14505154e-01
-1.29788578e+00 6.11564338e-01 -2.87958942e-02 -6.41018331e-01
-6.00363314e-01 5.44644475e-01 2.79373884e-01 -3.66546273e-01
2.82395393e-01 -5.28526545e-01 -5.54607749e-01 3.96187425e-01
6.39660776e-01 4.70559001e-01 3.07960004e-01 -5.20814896e-01
-1.04469456e-01 5.05498469e-01 -3.29909116e-01 7.60297254e-02
1.48948216e+00 -1.26327127e-01 -4.98672992e-01 5.59245050e-01
1.21778190e+00 -7.54168108e-02 -5.73016524e-01 -1.09785005e-01
6.62885308e-02 1.40503690e-01 -9.80757177e-02 -5.11200607e-01
-5.92766583e-01 9.72325444e-01 -1.75090596e-01 5.51927328e-01
9.77068365e-01 -2.24521920e-01 1.36217439e+00 5.52115917e-01
8.44419226e-02 -1.23668170e+00 2.23778576e-01 1.03208792e+00
1.06014192e+00 -1.10860443e+00 -5.94555847e-02 4.20691781e-02
-8.89775813e-01 1.26411247e+00 4.75324512e-01 -1.71869189e-01
-1.54276922e-01 7.81428218e-02 -3.94985765e-01 -1.36979325e-02
-8.57323647e-01 -1.73929855e-01 1.41102418e-01 2.77116716e-01
5.94143212e-01 -8.30527097e-02 -6.03155971e-01 8.03419113e-01
-4.77409780e-01 -1.13188155e-01 8.37982059e-01 7.50998259e-01
-8.80023777e-01 -8.85487974e-01 -2.68433094e-01 6.44300461e-01
-6.54695630e-01 -4.53153491e-01 -5.29858053e-01 3.80826056e-01
-3.12401503e-01 9.64320719e-01 1.52644157e-01 -2.13871241e-01
3.74571979e-01 1.53971970e-01 2.66203314e-01 -9.83863294e-01
-9.79363441e-01 -2.96223983e-02 2.38506302e-01 -3.11724454e-01
-3.71975243e-01 -4.54659343e-01 -1.32162678e+00 -2.25617588e-01
-3.58402103e-01 4.90430474e-01 6.14229202e-01 9.63868558e-01
5.36151111e-01 9.72608387e-01 5.65212905e-01 -1.10016429e+00
-7.91253746e-01 -1.48946655e+00 -4.38818723e-01 2.64358521e-01
4.37248826e-01 -2.83214152e-02 -2.66719967e-01 -7.03488588e-02] | [12.39097785949707, 9.416102409362793] |
1fd4e23c-8927-4ddf-8a89-9b88208d4bac | model-based-reinforcement-learning-for-6 | 2304.10000 | null | https://arxiv.org/abs/2304.10000v1 | https://arxiv.org/pdf/2304.10000v1.pdf | Model Based Reinforcement Learning for Personalized Heparin Dosing | A key challenge in sequential decision making is optimizing systems safely under partial information. While much of the literature has focused on the cases of either partially known states or partially known dynamics, it is further exacerbated in cases where both states and dynamics are partially known. Computing heparin doses for patients fits this paradigm since the concentration of heparin in the patient cannot be measured directly and the rates at which patients metabolize heparin vary greatly between individuals. While many proposed solutions are model free, they require complex models and have difficulty ensuring safety. However, if some of the structure of the dynamics is known, a model based approach can be leveraged to provide safe policies. In this paper we propose such a framework to address the challenge of optimizing personalized heparin doses. We use a predictive model parameterized individually by patient to predict future therapeutic effects. We then leverage this model using a scenario generation based approach that is capable of ensuring patient safety. We validate our models with numerical experiments by comparing the predictive capabilities of our model against existing machine learning techniques and demonstrating how our dosing algorithm can treat patients in a simulated ICU environment. | ['Yonatan Mintz', 'Qinyang He'] | 2023-04-19 | null | null | null | null | ['model-based-reinforcement-learning'] | ['reasoning'] | [ 1.09678164e-01 -6.16824403e-02 -4.43532765e-01 5.87960370e-02
-3.80912960e-01 -7.33281136e-01 2.09644020e-01 5.50018668e-01
-2.47880816e-01 1.05838072e+00 1.16401806e-01 -6.95876300e-01
-5.48898816e-01 -6.76924109e-01 -4.41497773e-01 -6.15096331e-01
-2.88630575e-01 8.56438577e-01 -8.82783234e-02 -3.04918379e-01
1.65799618e-01 5.63112617e-01 -1.06016362e+00 1.21135943e-01
7.64101624e-01 7.55842984e-01 -2.14014590e-01 9.39575195e-01
2.54731596e-01 8.49581957e-01 -3.41570675e-01 3.09022099e-01
6.68652296e-01 -5.09011149e-01 -3.55551362e-01 2.30705842e-01
-2.17759684e-01 -4.58250761e-01 -4.23359752e-01 3.28899562e-01
5.81910193e-01 3.81702781e-02 4.16934788e-01 -1.12339914e+00
3.68171155e-01 5.21763563e-01 -1.79186221e-02 7.68882632e-02
1.62570789e-01 4.97387379e-01 4.99451727e-01 2.37642378e-01
1.28408968e-01 9.32812572e-01 5.97605884e-01 6.49306834e-01
-1.31612432e+00 -3.32362294e-01 3.91256243e-01 -5.79938153e-03
-1.00440145e+00 -1.28874600e-01 2.98436612e-01 -6.59896612e-01
7.41015851e-01 2.76501805e-01 9.12433088e-01 6.81201160e-01
7.85642445e-01 4.87896413e-01 1.07023609e+00 -1.81460202e-01
5.00122428e-01 3.15444350e-01 8.84946808e-02 4.52389270e-01
4.35961753e-01 5.41064560e-01 -3.30866814e-01 -5.40806413e-01
8.36653173e-01 5.33692956e-01 -6.34115815e-01 -7.65839875e-01
-9.69005108e-01 7.55634546e-01 -2.61857808e-01 -2.48964876e-01
-6.10474169e-01 6.14466108e-02 2.71095902e-01 3.28984022e-01
-8.94272178e-02 7.73246527e-01 -7.66210616e-01 -1.40131220e-01
-8.03885877e-01 5.11648059e-01 1.41276181e+00 8.25953960e-01
2.58851439e-01 -5.19756675e-02 -2.93191522e-01 2.25859538e-01
-1.02390144e-02 2.31507495e-01 3.71256053e-01 -1.17555046e+00
3.25253159e-01 4.45889860e-01 8.69813859e-01 -3.41458708e-01
-6.26880586e-01 -4.10636544e-01 -4.06115264e-01 2.42764622e-01
5.81242383e-01 -7.27003813e-01 -8.46159339e-01 1.72135854e+00
1.61467269e-01 3.60014498e-01 6.86736330e-02 7.51081467e-01
-3.42995852e-01 4.35789585e-01 2.45887395e-02 -8.82097661e-01
8.65035951e-01 -5.78476965e-01 -6.52201891e-01 -5.60440309e-02
3.34009171e-01 -3.26511919e-01 6.22853220e-01 6.52880192e-01
-1.23180115e+00 1.64939061e-01 -9.75444317e-01 7.54541755e-01
1.11507751e-01 -5.71074009e-01 3.56129646e-01 7.86479950e-01
-8.66756380e-01 1.09094024e+00 -1.30632675e+00 -4.10420209e-01
1.26692533e-01 6.48130059e-01 3.98611635e-01 -1.00617902e-02
-1.10182238e+00 1.19483185e+00 1.97348297e-01 -7.08333850e-02
-1.07836032e+00 -1.28717852e+00 -5.72844267e-01 1.94362208e-01
6.21219337e-01 -1.31587553e+00 1.67113483e+00 -2.31231764e-01
-1.52845955e+00 -9.12332311e-02 1.72945801e-02 -6.48703456e-01
9.85688508e-01 -1.84341222e-01 8.60014111e-02 1.68470919e-01
-4.29681629e-01 -1.91258192e-01 4.85989302e-01 -1.03204179e+00
-6.67750895e-01 -1.86034173e-01 3.39540601e-01 3.53170335e-01
-6.01348095e-02 -3.33937109e-01 -2.45533697e-02 -2.32595816e-01
-1.89206868e-01 -1.04551971e+00 -8.88816893e-01 -7.38457963e-02
-2.28811413e-01 5.02109528e-01 6.73570275e-01 -3.14233452e-01
1.41371262e+00 -1.39304602e+00 -1.62130892e-02 3.98953915e-01
2.11211089e-02 4.45098393e-02 2.99983889e-01 1.07838261e+00
-1.15691766e-01 8.96432623e-02 -3.24781209e-01 2.37400960e-02
-1.68396369e-01 5.54930151e-01 -3.65575075e-01 3.60331833e-01
1.02411240e-01 5.82882643e-01 -8.38831544e-01 -2.23149449e-01
5.49610913e-01 3.07129622e-01 -8.73439431e-01 5.31032324e-01
-3.45132500e-01 7.91117072e-01 -8.63536775e-01 3.96338820e-01
2.28812605e-01 -3.43759596e-01 8.80811274e-01 1.17031395e-01
-1.05487935e-01 -4.47376706e-02 -1.34299648e+00 9.70215917e-01
-6.65732503e-01 -2.63024300e-01 1.05712831e-01 -1.18827558e+00
3.30482841e-01 5.44243276e-01 1.25679314e+00 -9.43105295e-02
4.30785537e-01 -2.16865569e-01 2.04342559e-01 -5.82143068e-01
-2.03879960e-02 -7.73707390e-01 1.22754417e-01 7.60202169e-01
-5.70165694e-01 -3.10083866e-01 1.28007337e-01 2.36757815e-01
1.46233737e+00 -8.20061713e-02 7.33764231e-01 -4.65954840e-01
1.35175958e-01 1.79515287e-01 5.79994738e-01 8.83324206e-01
-3.40411276e-01 4.14896786e-01 6.77034318e-01 -4.81636405e-01
-8.96084428e-01 -8.96034122e-01 -2.77139187e-01 3.11758012e-01
1.29192755e-01 -2.37024903e-01 -5.87743700e-01 -3.03966224e-01
5.34948111e-01 7.27002263e-01 -6.77188873e-01 -2.61835098e-01
-6.64082766e-01 -1.11998737e+00 -1.91724509e-01 4.54259276e-01
-2.87439853e-01 -5.79353154e-01 -1.19594431e+00 9.41385448e-01
1.47720471e-01 -8.51877809e-01 -4.62018132e-01 2.49412358e-01
-1.03270304e+00 -1.59230769e+00 -2.34520957e-01 -3.36915068e-02
3.90923679e-01 -5.40379211e-02 1.00700045e+00 1.46194398e-02
-5.58660924e-01 9.77112949e-01 1.10340171e-01 -7.29749143e-01
-4.75416124e-01 -4.14309800e-01 3.81169885e-01 -3.37693617e-02
-8.90223607e-02 -5.37832439e-01 -1.11368811e+00 3.32786053e-01
-7.97115743e-01 -2.93775983e-02 2.89712816e-01 8.49692345e-01
6.28063858e-01 -1.36584770e-02 6.22256517e-01 -1.26074815e+00
8.34581852e-01 -6.53403223e-01 -7.35594690e-01 3.74876231e-01
-1.09082282e+00 2.37374812e-01 8.29608560e-01 -6.28423870e-01
-7.96105385e-01 4.56306338e-01 3.42686921e-01 -4.97050792e-01
1.29135564e-01 3.98146182e-01 9.79927853e-02 5.83770536e-02
3.03202957e-01 4.72920984e-02 4.15483475e-01 -3.69452566e-01
6.25048354e-02 3.32693130e-01 1.13965660e-01 -9.52501595e-01
4.51158255e-01 4.44036186e-01 3.72403890e-01 -2.87877351e-01
-6.54746890e-01 -3.15834165e-01 -4.59316701e-01 -6.44548014e-02
3.70691866e-01 -7.34715760e-01 -1.28682160e+00 -6.08171057e-03
-4.17324603e-01 -5.51123440e-01 -5.34051120e-01 4.46097940e-01
-1.08291209e+00 3.43888462e-01 -6.49962962e-01 -1.21746957e+00
-1.21650644e-01 -1.29815328e+00 5.62961042e-01 6.66497424e-02
-3.02378356e-01 -1.31998026e+00 3.21571678e-01 -9.64093357e-02
6.16854310e-01 6.36009693e-01 1.04714501e+00 -7.25773573e-01
-7.76446223e-01 -4.31152552e-01 4.17461336e-01 -4.18691672e-02
3.77466023e-01 -4.15290505e-01 -3.75683457e-01 -6.28646731e-01
3.93276721e-01 3.09256800e-02 2.79188007e-01 6.77380741e-01
1.24937344e+00 -5.93286216e-01 -5.76858580e-01 2.96870738e-01
1.69226146e+00 5.21938980e-01 6.59347773e-02 1.47022322e-01
2.38358349e-01 4.11403954e-01 6.68665469e-01 1.15532422e+00
4.58144277e-01 5.43596029e-01 3.91494632e-01 1.71774969e-01
5.43130636e-01 -6.43896237e-02 2.25070287e-02 2.63936967e-01
-5.92322322e-03 -1.55179903e-01 -1.04570889e+00 3.55044872e-01
-2.05941629e+00 -8.07967424e-01 1.59348547e-01 2.48242927e+00
9.22254980e-01 3.53237689e-02 2.36638814e-01 -1.23925671e-01
3.33758652e-01 -4.78720486e-01 -1.00706303e+00 -4.41304773e-01
4.15128052e-01 3.00514638e-01 8.09479237e-01 6.13382816e-01
-7.12060750e-01 1.52533144e-01 7.68513632e+00 -1.00311913e-01
-9.43745017e-01 -2.15914592e-01 6.50522709e-01 -5.92245400e-01
1.88591748e-01 1.58558831e-01 -6.82894230e-01 4.24248993e-01
1.34983683e+00 -7.77091682e-01 4.97538269e-01 6.37857258e-01
9.82716143e-01 -1.45108208e-01 -1.56220591e+00 7.02040136e-01
-1.72300130e-01 -1.14439154e+00 -1.40858695e-01 1.46042585e-01
6.44582152e-01 -3.26633215e-01 -1.31043538e-01 3.11111033e-01
6.25995994e-01 -8.48315537e-01 2.34639183e-01 7.34605730e-01
3.46399486e-01 -6.83955610e-01 5.59781492e-01 7.41445839e-01
-9.27674592e-01 -6.17581844e-01 5.06510399e-02 -2.58294284e-01
5.33207715e-01 2.62545288e-01 -1.01956236e+00 4.02006090e-01
2.35450089e-01 4.49339718e-01 4.82293889e-02 1.39421964e+00
2.49550119e-01 5.53563952e-01 -5.47619164e-01 1.37435392e-01
-6.50948659e-02 -2.32298464e-01 4.02015865e-01 8.38841856e-01
2.62983680e-01 6.80802047e-01 7.10660994e-01 3.46463948e-01
5.29518127e-01 4.98038530e-02 -6.34783685e-01 8.03649127e-02
3.34338844e-01 7.65512228e-01 -2.62960464e-01 -3.98855329e-01
-3.84581089e-01 3.72347504e-01 -2.85927281e-02 4.01408851e-01
-6.34611905e-01 1.14886083e-01 9.63690579e-01 4.72885162e-01
-2.10363474e-02 -2.79674053e-01 -5.49110115e-01 -1.21904445e+00
-3.89168233e-01 -9.55865622e-01 8.70156586e-01 -1.91033646e-01
-1.15027046e+00 5.52854575e-02 1.97381228e-01 -1.40762258e+00
-6.61753774e-01 -5.95081091e-01 -4.57031876e-01 9.63675618e-01
-1.47638810e+00 -4.61554259e-01 1.66923981e-02 6.00966632e-01
5.74151397e-01 2.54714012e-01 6.71194375e-01 -5.62562719e-02
-7.15049148e-01 2.06881508e-01 2.77625203e-01 -6.11289084e-01
5.73485196e-01 -1.33065510e+00 -2.49264643e-01 5.29822469e-01
-7.69081771e-01 9.02181327e-01 1.06644154e+00 -7.96730995e-01
-1.82081318e+00 -1.00092232e+00 1.68144256e-01 -8.29997957e-01
8.07618082e-01 9.56001282e-02 -8.45934868e-01 6.87532365e-01
1.84478443e-02 -3.17889094e-01 7.45708048e-01 -1.51961580e-01
2.77887374e-01 -1.17484592e-02 -1.16118109e+00 5.75849175e-01
6.80556476e-01 -1.51240351e-02 -2.49268621e-01 6.08251333e-01
3.31318140e-01 -5.77524006e-01 -1.17784190e+00 4.63810503e-01
5.55345893e-01 -6.57471359e-01 8.24834704e-01 -1.28700626e+00
-6.67428896e-02 -1.58406168e-01 -1.16287572e-02 -1.54351521e+00
-2.51758285e-02 -9.44069147e-01 -4.34944689e-01 3.53264451e-01
4.67415452e-01 -7.68237293e-01 8.77305567e-01 1.46559978e+00
4.79514562e-02 -1.08117306e+00 -7.25104153e-01 -7.86553323e-01
2.68571615e-01 -1.68826759e-01 6.37565911e-01 7.81415105e-01
5.56486428e-01 -2.22541969e-02 -7.42908418e-01 2.59602249e-01
6.90998256e-01 2.44515553e-01 4.55736518e-01 -9.18485999e-01
-8.58649909e-01 -1.80110350e-01 2.56176200e-03 -6.49727106e-01
-2.76462376e-01 -4.11897451e-01 3.93407345e-02 -1.54447734e+00
4.95711505e-01 -5.78763902e-01 -5.33026278e-01 4.95706648e-01
-2.90753245e-01 -5.15973508e-01 9.82112736e-02 3.77018191e-02
-2.39528984e-01 2.19424129e-01 1.10771322e+00 1.82149023e-01
-5.46847939e-01 4.20897514e-01 -7.79808283e-01 3.48372012e-01
9.82683897e-01 -3.23823631e-01 -7.59485900e-01 6.86174482e-02
-1.82882681e-01 8.16151559e-01 5.40241599e-02 -8.18591833e-01
2.00253919e-01 -8.78686726e-01 1.24204010e-01 -2.83585817e-01
2.58066535e-01 -9.73150313e-01 5.54888248e-01 1.17642415e+00
-2.56860703e-01 2.21617147e-01 3.73208731e-01 8.59937668e-01
2.62096703e-01 1.37407124e-01 9.60643947e-01 -4.15785342e-01
-5.96513525e-02 7.54672408e-01 -5.75862825e-01 9.50652510e-02
1.32486761e+00 5.10643125e-02 -2.40132570e-01 -7.68235922e-01
-1.06251168e+00 6.96874499e-01 5.00446558e-01 2.52945125e-01
3.95283788e-01 -6.41646981e-01 -4.89557922e-01 7.19394535e-02
-1.55823693e-01 -3.44172597e-01 2.06766874e-01 8.57725739e-01
-4.46894288e-01 5.98453939e-01 6.86852783e-02 -4.39944923e-01
-8.75010848e-01 1.08172929e+00 9.24103320e-01 -2.62966007e-01
-5.30043662e-01 -5.53349704e-02 2.43150845e-01 -3.83420512e-02
5.90154305e-02 -1.72729567e-01 1.12371035e-01 -2.61604428e-01
5.97374320e-01 2.10328445e-01 5.53246066e-02 3.65983993e-02
-1.94150835e-01 3.41658711e-01 -1.41387001e-01 4.20106724e-02
1.35372770e+00 -1.56570822e-01 2.58585095e-01 3.48656356e-01
5.71620524e-01 -2.82212436e-01 -1.54725814e+00 -9.29495469e-02
-1.01876959e-01 -4.87581223e-01 -2.13848963e-01 -1.07771266e+00
-8.28124166e-01 7.17196822e-01 6.31147981e-01 1.46792874e-01
1.12545574e+00 -3.57924581e-01 5.69563329e-01 3.79773080e-01
5.37295520e-01 -8.65551233e-01 -1.43589258e-01 2.47354493e-01
5.22473812e-01 -9.15009379e-01 8.46878365e-02 -2.91755646e-01
-7.85567880e-01 9.63407159e-01 5.18772781e-01 -9.22036767e-02
1.02223897e+00 6.54057264e-01 2.84106553e-01 1.90426201e-01
-1.30043602e+00 4.94487397e-02 -3.50689501e-01 3.60017836e-01
1.45083129e-01 1.14242166e-01 -5.07256329e-01 6.51935220e-01
1.30455811e-02 3.16758126e-01 9.93564546e-01 1.45261788e+00
-4.09725219e-01 -1.55550516e+00 -3.62694323e-01 7.97723174e-01
-4.71788853e-01 6.32973984e-02 5.61009012e-02 7.27826953e-01
-4.06852782e-01 9.89094138e-01 -3.27322394e-01 3.24018039e-02
7.37665057e-01 2.23794386e-01 5.63497782e-01 -7.83889711e-01
-5.45153797e-01 6.60430640e-02 5.25517166e-02 -7.06337214e-01
1.14878535e-01 -9.34442461e-01 -1.11721182e+00 -2.63989061e-01
5.83429635e-02 2.29387522e-01 4.83827621e-01 8.88645589e-01
4.13172036e-01 5.90123594e-01 8.67298484e-01 -6.90590084e-01
-1.31474447e+00 -3.85613501e-01 -7.00527251e-01 2.72691816e-01
5.98436475e-01 -8.30108762e-01 -2.22081989e-01 3.56920473e-02] | [4.011250019073486, 2.739718198776245] |
046fe903-2a64-4aad-8b50-6f8cdcb19c4f | multi-objective-conflict-based-search-using | 2108.00745 | null | https://arxiv.org/abs/2108.00745v3 | https://arxiv.org/pdf/2108.00745v3.pdf | Multi-objective Conflict-based Search Using Safe-interval Path Planning | This paper addresses a generalization of the well known multi-agent path finding (MAPF) problem that optimizes multiple conflicting objectives simultaneously such as travel time and path risk. This generalization, referred to as multi-objective MAPF (MOMAPF), arises in several applications ranging from hazardous material transportation to construction site planning. In this paper, we present a new multi-objective conflict-based search (MO-CBS) approach that relies on a novel multi-objective safe interval path planning (MO-SIPP) algorithm for its low-level search. We first develop the MO-SIPP algorithm, show its properties and then embed it in MO-CBS. We present extensive numerical results to show that (1) there is an order of magnitude improvement in the average low level search time, and (2) a significant improvement in the success rates of finding the Pareto-optimal front can be obtained using the proposed approach in comparison with the previous MO-CBS. | ['Maxim Likhachev', 'Howie Choset', 'Sivakumar Rathinam', 'Zhongqiang Ren'] | 2021-08-02 | null | null | null | null | ['multi-agent-path-finding'] | ['playing-games'] | [ 7.21416473e-02 2.20316276e-02 -2.54557788e-01 6.85536787e-02
-7.49365270e-01 -5.79264581e-01 1.97951421e-01 5.39452791e-01
-3.05939585e-01 1.37461376e+00 -2.33768076e-01 -2.78975040e-01
-1.21731126e+00 -9.13536787e-01 -5.69546998e-01 -7.78083026e-01
-6.57320440e-01 8.01982880e-01 4.85617667e-01 -5.16044676e-01
5.42058349e-01 7.32258558e-01 -1.19624460e+00 -3.53237152e-01
1.22723031e+00 9.12587166e-01 2.32991815e-01 4.43430096e-01
-1.43390624e-02 1.41978815e-01 -5.12927830e-01 -1.16685316e-01
1.03523798e-01 -1.36038095e-01 -1.27195776e+00 -9.22911242e-02
-7.64014900e-01 6.18397444e-02 4.24432635e-01 7.77221084e-01
2.01273113e-01 6.33302987e-01 8.20935786e-01 -2.18395591e+00
1.09080888e-01 4.66729999e-01 -7.46369243e-01 -2.92980727e-02
4.38650727e-01 -2.71721631e-01 8.14723790e-01 -4.71212745e-01
6.79725230e-01 1.21826589e+00 2.51176029e-01 2.05910578e-01
-9.94111836e-01 -4.62532304e-02 2.62507468e-01 3.84342551e-01
-1.47454727e+00 2.92642713e-01 4.26830024e-01 1.21419085e-02
1.02896643e+00 6.32700443e-01 5.06559312e-01 4.97728363e-02
7.56064832e-01 4.35588747e-01 1.03338993e+00 -6.10771120e-01
4.52842861e-01 -2.30626747e-01 3.20467614e-02 5.46786368e-01
6.12144172e-01 2.52151221e-01 -1.00371495e-01 -2.43680105e-01
3.81550312e-01 -4.44382250e-01 -1.40561402e-01 -3.01290452e-01
-1.32204854e+00 9.31619823e-01 3.10165286e-01 2.50327528e-01
-5.61228156e-01 3.01341712e-01 -6.28086627e-02 1.23726472e-01
1.74946994e-01 4.42857981e-01 -1.18160278e-01 5.17930416e-03
-3.80397022e-01 6.12849534e-01 8.43896687e-01 8.92444372e-01
6.14517212e-01 7.08035240e-03 -7.06142653e-03 5.45084655e-01
3.70827854e-01 4.25853729e-01 -4.81271863e-01 -1.23212051e+00
5.82945406e-01 2.77841061e-01 7.24031031e-01 -9.99801219e-01
-6.96994245e-01 -1.23126782e-01 -3.52881193e-01 6.77909672e-01
1.41625896e-01 -1.38292164e-01 -5.85338056e-01 1.56089532e+00
5.81755102e-01 -1.79080948e-01 1.65709049e-01 7.56969213e-01
9.97955725e-03 1.16495788e+00 -2.79103696e-01 -5.98906636e-01
9.62815166e-01 -1.23926330e+00 -7.15810597e-01 -1.08767757e-02
3.18214804e-01 -5.87765038e-01 4.53412265e-01 4.02155906e-01
-1.33264661e+00 2.76720256e-01 -1.24545097e+00 6.74598992e-01
-5.82587838e-01 -7.51825929e-01 4.98699397e-01 6.87608123e-01
-9.56844568e-01 3.67468089e-01 -4.77458328e-01 -2.41979405e-01
-8.35287645e-02 7.25561798e-01 -2.74716347e-01 -3.57272834e-01
-1.03879595e+00 1.02018487e+00 5.49324632e-01 -1.31276086e-01
-7.56536067e-01 -5.39218605e-01 -6.17837965e-01 2.12505773e-01
1.02188003e+00 -4.63281214e-01 9.97835100e-01 -2.56811798e-01
-1.60675931e+00 1.87298562e-02 -1.81052983e-01 7.17174485e-02
3.68990034e-01 3.27145725e-01 -4.26678628e-01 2.12686419e-01
3.63695651e-01 3.76779139e-01 1.96183264e-01 -1.62691379e+00
-1.02055979e+00 -4.30618152e-02 4.49482501e-01 2.32649207e-01
2.28687599e-01 1.83227286e-01 -3.42868157e-02 -4.38403845e-01
-1.15712613e-01 -1.01083088e+00 -8.75240862e-01 -3.45995665e-01
-3.70429277e-01 -2.72567332e-01 4.15657908e-01 -1.69831097e-01
1.48775065e+00 -1.49392390e+00 4.06304717e-01 8.72459948e-01
-3.18620324e-01 -2.43197113e-01 -3.24344695e-01 1.13574111e+00
2.35068724e-01 2.22582296e-01 -4.26754236e-01 1.71656832e-02
1.26981869e-01 3.92474264e-01 2.74446219e-01 4.97199595e-01
-1.29346317e-02 5.88369787e-01 -1.12092888e+00 -4.35277402e-01
1.77593231e-02 -1.01616934e-01 -2.23546132e-01 -2.83636034e-01
-2.39727482e-01 1.71365470e-01 -5.89764714e-01 9.95544732e-01
6.73936963e-01 9.07312110e-02 6.63837045e-02 3.28951865e-01
-7.12471485e-01 -3.10535520e-01 -1.45623720e+00 1.25722492e+00
-4.56691712e-01 1.59460381e-01 2.50953257e-01 -9.05348182e-01
5.99485695e-01 2.93195188e-01 9.56593275e-01 -7.02613592e-01
2.40310337e-02 5.27198017e-01 -2.08505560e-02 -2.54534036e-01
6.92126691e-01 -2.79250778e-02 -2.96553165e-01 6.33556604e-01
-4.19427752e-01 1.55849278e-01 7.19261467e-01 -1.79696328e-03
9.88554478e-01 -3.68015170e-01 5.06773710e-01 -8.69279027e-01
8.43513012e-01 4.65651572e-01 6.94835424e-01 5.37962079e-01
-1.62607640e-01 -1.46295846e-01 4.62749571e-01 -9.81493518e-02
-6.22128308e-01 -1.17580652e+00 1.54768959e-01 6.65058553e-01
8.06051075e-01 -3.68208922e-02 -3.86113524e-01 -4.88729119e-01
1.76426336e-01 9.49422181e-01 -3.50727558e-01 2.03185096e-01
-7.55387604e-01 -9.38121855e-01 1.15346663e-01 2.15586677e-01
3.14924747e-01 -5.74512959e-01 -7.21829772e-01 7.46469140e-01
-3.64345282e-01 -7.89924145e-01 -4.14429307e-01 -4.61724326e-02
-5.30133903e-01 -1.02948952e+00 -7.56998777e-01 -8.00870955e-01
7.64384449e-01 5.42742968e-01 6.80816889e-01 6.25675470e-02
-1.21913835e-01 4.16268647e-01 -5.39379895e-01 -5.57697773e-01
-3.39673012e-01 -1.29414676e-02 8.26242566e-03 2.70224977e-02
-3.76711100e-01 -4.54646528e-01 -4.07077402e-01 8.49365234e-01
-7.81389356e-01 -3.52587402e-01 4.25974041e-01 5.42186320e-01
7.19797850e-01 8.55557263e-01 1.23029470e+00 -2.65402377e-01
8.36687505e-01 -6.83302760e-01 -1.04083037e+00 6.83895051e-01
-1.07264519e+00 6.22180756e-03 3.46924841e-01 5.44586629e-02
-9.28295195e-01 -2.52538472e-01 1.39243111e-01 2.01356113e-01
7.92721659e-02 8.08162987e-01 -2.00817302e-01 -7.13243544e-01
-1.40625015e-01 -1.39488623e-01 -3.01534962e-02 -1.95607498e-01
3.10061295e-02 3.05902094e-01 4.74784553e-01 -7.63467968e-01
8.07587266e-01 4.47505027e-01 9.17226493e-01 -2.69639343e-01
-5.82126752e-02 -4.48284686e-01 -1.44120693e-01 -7.50354290e-01
7.32425451e-01 6.39158953e-03 -1.38108528e+00 1.42783687e-01
-1.11547136e+00 -1.27735943e-01 2.30650857e-01 4.83562976e-01
-9.02706444e-01 2.83159524e-01 -1.37363866e-01 -1.30694580e+00
-1.06370740e-01 -1.22027147e+00 4.91250187e-01 2.94130296e-01
3.06654423e-02 -1.15181363e+00 2.41093457e-01 1.25413850e-01
4.22329962e-01 9.65894580e-01 1.24156427e+00 -2.48367906e-01
-8.13761711e-01 -5.18785119e-02 5.13045758e-04 -5.51916301e-01
9.08505768e-02 1.17545656e-03 1.87120005e-01 -3.29697192e-01
-3.40256602e-01 2.81385213e-01 6.60290495e-02 5.30884147e-01
5.12771308e-01 -3.96568209e-01 -7.57765114e-01 4.49309461e-02
2.13114500e+00 9.79890764e-01 5.77157319e-01 1.08489633e+00
-3.25132832e-02 9.67782497e-01 1.48412025e+00 4.63458389e-01
5.62262535e-01 9.66431737e-01 8.41347277e-01 7.77551383e-02
5.28399944e-01 3.69731367e-01 1.41418979e-01 2.93568790e-01
-5.06637990e-01 -8.64602625e-01 -1.03026485e+00 8.54135036e-01
-2.29571557e+00 -8.46929669e-01 -3.21086407e-01 2.15828776e+00
2.14279622e-01 4.07763869e-02 4.98464078e-01 4.79972988e-01
9.25879359e-01 -2.54397929e-01 -2.31694967e-01 -1.07511330e+00
5.70918769e-02 4.80423719e-02 9.04026866e-01 9.63084638e-01
-7.61226594e-01 2.58783817e-01 6.48532772e+00 8.51589382e-01
-6.87744141e-01 4.52362560e-02 2.37802759e-01 3.57146859e-02
-3.69061172e-01 1.56562105e-02 -3.94925863e-01 4.00198847e-01
8.81950200e-01 -9.32452381e-01 6.89751148e-01 5.10542214e-01
5.40377617e-01 -5.68355799e-01 -5.64220369e-01 4.85249311e-01
-3.02746654e-01 -1.25665653e+00 -2.54427016e-01 3.47455770e-01
8.71517777e-01 -6.67435944e-01 -2.41228372e-01 -2.54572898e-01
2.58805066e-01 -8.89098048e-01 7.67727911e-01 7.14452490e-02
2.73267478e-01 -1.69769692e+00 7.93880522e-01 3.18599045e-01
-1.49246895e+00 -2.52909809e-01 3.20833959e-02 1.39329523e-01
1.16049099e+00 3.77084345e-01 -6.42352223e-01 1.44139290e+00
6.06584966e-01 -6.24256656e-02 3.62614453e-01 1.72226107e+00
3.71205285e-02 -9.71281976e-02 -5.19954979e-01 -3.43595296e-01
7.79975057e-01 -4.61722076e-01 1.09274590e+00 7.32205272e-01
5.29149592e-01 1.42587602e-01 2.80458748e-01 5.98356366e-01
4.98355597e-01 1.01673380e-01 -2.88339049e-01 1.66718617e-01
4.98809844e-01 8.88447821e-01 -1.14271152e+00 1.70765936e-01
-1.31400660e-01 5.41920900e-01 -3.39700550e-01 3.06483418e-01
-1.22690690e+00 -8.65462303e-01 5.52455842e-01 -2.34158337e-01
2.29658559e-01 -4.33898568e-01 -4.72675681e-01 -1.95167035e-01
-1.67617500e-02 -2.05199927e-01 4.90363568e-01 -4.92692202e-01
-7.86987484e-01 5.80296755e-01 7.70513713e-01 -1.26455319e+00
-2.77368039e-01 -2.69491285e-01 -7.72258759e-01 7.09238529e-01
-1.90414166e+00 -8.21969390e-01 -3.50835621e-02 4.88222271e-01
5.41164219e-01 1.02895923e-01 5.37526786e-01 4.49662685e-01
-5.16810358e-01 3.36726993e-01 3.93417269e-01 -7.73969591e-01
6.52304292e-02 -1.03087151e+00 -2.13598013e-01 9.84366775e-01
-7.48381615e-01 9.32436585e-02 8.54131877e-01 -6.82592988e-01
-1.44474697e+00 -9.07330096e-01 9.65842664e-01 2.05284938e-01
4.69324201e-01 3.42494816e-01 -2.68834054e-01 3.96880627e-01
2.79214382e-01 -7.11184680e-01 4.38339829e-01 -3.91841263e-01
5.29730976e-01 -1.03586584e-01 -1.60190952e+00 6.06868148e-01
8.33476305e-01 6.41580403e-01 -9.77780372e-02 2.94345081e-01
6.52639091e-01 -1.11881800e-01 -8.46327782e-01 6.73781514e-01
4.61292535e-01 -5.68177581e-01 9.83347833e-01 -3.81611079e-01
5.93945719e-02 -5.73327422e-01 -3.32915068e-01 -1.54866910e+00
-6.66214764e-01 -7.86227882e-01 3.13935071e-01 1.14322007e+00
5.44236064e-01 -1.14725530e+00 4.50514764e-01 6.05958641e-01
-4.54029828e-01 -1.10509658e+00 -1.43259311e+00 -1.61483824e+00
-2.27388218e-02 -5.96868386e-03 9.47998345e-01 5.68684459e-01
2.05044612e-01 -1.09954476e-01 -2.33926952e-01 5.14798701e-01
9.45763648e-01 2.07846388e-01 1.82895541e-01 -1.00939977e+00
-1.15902163e-01 -4.33098376e-01 -5.98605014e-02 -3.64963919e-01
7.28152832e-03 -6.03235960e-01 2.75023431e-01 -2.12100792e+00
-2.97174841e-01 -8.49027693e-01 -3.98287356e-01 2.72202253e-01
2.34242588e-01 -2.79546499e-01 3.03800464e-01 -5.58982193e-02
-5.01238286e-01 4.51077044e-01 1.10915172e+00 -9.86400917e-02
-3.05410922e-01 2.64936984e-01 -3.75437677e-01 2.93980420e-01
8.93594086e-01 -8.44473481e-01 -4.61929590e-01 -2.76165158e-01
2.58039445e-01 7.77948678e-01 -1.05376029e-03 -7.76812255e-01
2.37323701e-01 -1.15070355e+00 -5.47040045e-01 -8.88147533e-01
3.66114050e-01 -1.07968748e+00 6.05062246e-01 8.47772658e-01
1.14120714e-01 5.22956789e-01 1.99415788e-01 6.98248804e-01
-2.14181200e-01 -6.37816072e-01 6.52344704e-01 -1.40464744e-02
-7.54408300e-01 5.52185578e-03 -8.74259710e-01 -5.41207790e-01
1.71349657e+00 -3.51477832e-01 -6.48732960e-01 -2.79230744e-01
-5.06943703e-01 9.41093385e-01 2.46858895e-01 9.82731301e-03
5.82118511e-01 -1.43103337e+00 -5.82618177e-01 -6.53475642e-01
-1.89006060e-01 -3.06142271e-01 2.36617982e-01 1.14345658e+00
-8.61867607e-01 8.61878991e-01 -4.82759058e-01 -1.26057163e-01
-1.07482111e+00 6.98040724e-01 9.24141183e-02 -6.52653039e-01
-2.14397281e-01 5.09552598e-01 -2.47300074e-01 2.84624286e-02
-6.87913522e-02 1.57242715e-01 -3.96766141e-02 -1.43672734e-01
3.06204557e-01 1.34474075e+00 -7.15473443e-02 -4.28395033e-01
-8.60481203e-01 7.17272401e-01 4.72220570e-01 -6.24264538e-01
1.41380358e+00 -4.08315241e-01 -4.12648410e-01 -4.58266102e-02
7.88680494e-01 7.11248294e-02 -5.40557027e-01 4.06789631e-01
1.60797551e-01 -5.77999592e-01 -4.80771884e-02 -1.06062198e+00
-8.70459259e-01 1.05455652e-01 1.47616267e-01 3.62688452e-01
1.39794099e+00 -3.12299341e-01 9.13055420e-01 3.37271541e-01
1.05123854e+00 -1.16444743e+00 -1.12544082e-01 3.68012637e-01
9.49735343e-01 -6.43261015e-01 -3.49461213e-02 -1.06733060e+00
-4.33847070e-01 1.12860405e+00 4.93167132e-01 8.79478455e-03
5.94181001e-01 2.76530236e-01 -4.46759641e-01 -4.45295945e-02
-6.62924469e-01 -3.99703890e-01 -1.71322376e-01 5.55042684e-01
-5.93794286e-01 1.22325487e-01 -1.00985897e+00 5.00162721e-01
2.44344577e-01 -5.04086614e-02 8.36481214e-01 1.48139298e+00
-6.79425895e-01 -1.34946918e+00 -8.11853707e-01 6.31598905e-02
-2.35008046e-01 4.74849164e-01 -1.14436984e-01 9.64570045e-01
-2.04722852e-01 1.46776986e+00 -2.13818148e-01 -9.82182547e-02
3.97050232e-01 -3.40639234e-01 5.83087444e-01 -9.84300300e-02
-3.49665612e-01 -1.09296094e-03 6.90202355e-01 -5.22394478e-01
-5.77673316e-01 -5.31759381e-01 -1.60166204e+00 -5.27829528e-01
-3.63062173e-01 5.86377501e-01 8.79130721e-01 9.65947509e-01
1.41651526e-01 5.57327092e-01 9.15139973e-01 -7.88463056e-01
-3.71958241e-02 -1.18306763e-01 -6.03388309e-01 -4.72097665e-01
1.58872679e-01 -1.20867908e+00 -2.21548930e-01 -6.91701293e-01] | [4.992466926574707, 1.8822332620620728] |
0f0d542f-6884-488c-9c30-7872e8846c8e | tripose-a-weakly-supervised-3d-human-pose | 2105.06599 | null | https://arxiv.org/abs/2105.06599v1 | https://arxiv.org/pdf/2105.06599v1.pdf | TriPose: A Weakly-Supervised 3D Human Pose Estimation via Triangulation from Video | Estimating 3D human poses from video is a challenging problem. The lack of 3D human pose annotations is a major obstacle for supervised training and for generalization to unseen datasets. In this work, we address this problem by proposing a weakly-supervised training scheme that does not require 3D annotations or calibrated cameras. The proposed method relies on temporal information and triangulation. Using 2D poses from multiple views as the input, we first estimate the relative camera orientations and then generate 3D poses via triangulation. The triangulation is only applied to the views with high 2D human joint confidence. The generated 3D poses are then used to train a recurrent lifting network (RLN) that estimates 3D poses from 2D poses. We further apply a multi-view re-projection loss to the estimated 3D poses and enforce the 3D poses estimated from multi-views to be consistent. Therefore, our method relaxes the constraints in practice, only multi-view videos are required for training, and is thus convenient for in-the-wild settings. At inference, RLN merely requires single-view videos. The proposed method outperforms previous works on two challenging datasets, Human3.6M and MPI-INF-3DHP. Codes and pretrained models will be publicly available. | ['Z. Jane Wang', 'Rabab Ward', 'Helge Rhodin', 'Ahmad Rezaei', 'Mohsen Gholami'] | 2021-05-14 | null | null | null | null | ['weakly-supervised-3d-human-pose-estimation'] | ['computer-vision'] | [-7.18478635e-02 -1.07238311e-02 -1.96440935e-01 -4.02538657e-01
-7.92142928e-01 -5.12362659e-01 3.43313962e-01 -4.13219750e-01
-6.94356084e-01 6.04613900e-01 1.83666930e-01 1.30621910e-01
2.86563873e-01 -3.74043643e-01 -9.83578384e-01 -5.84914327e-01
2.50100270e-02 6.78444326e-01 2.38905907e-01 -2.45891083e-02
-1.87361658e-01 4.59979534e-01 -1.42723846e+00 1.12901673e-01
2.98753679e-01 8.82831037e-01 1.51525036e-01 7.77726710e-01
4.81076300e-01 4.12580192e-01 -3.96392375e-01 -3.07408422e-01
5.65139234e-01 -3.55632544e-01 -4.88528222e-01 5.96268952e-01
7.54418731e-01 -7.42983937e-01 -4.11351204e-01 8.50790799e-01
5.92035890e-01 2.79351383e-01 3.96181107e-01 -1.11752605e+00
-2.37060618e-02 -2.39540294e-01 -7.60745406e-01 -1.46762818e-01
7.98928022e-01 3.82114053e-02 7.25801766e-01 -1.08528149e+00
7.26281404e-01 1.25662839e+00 7.88351893e-01 7.78927326e-01
-1.04101133e+00 -4.82955039e-01 2.89932787e-01 4.18221019e-02
-1.46245992e+00 -3.71475786e-01 7.82004654e-01 -5.06437421e-01
7.23750710e-01 2.86415089e-02 8.75943184e-01 1.63098097e+00
6.86259642e-02 6.88872576e-01 9.82283294e-01 -3.13127816e-01
8.36095810e-02 -1.04022920e-01 -3.27733248e-01 1.00942862e+00
2.63539195e-01 -5.94206490e-02 -7.26903379e-01 -6.44615293e-02
1.19792044e+00 4.20473814e-01 -3.89348298e-01 -7.25921273e-01
-1.48616636e+00 7.23116815e-01 3.92215341e-01 -2.20701337e-01
-3.49307656e-01 1.71505883e-01 3.12177688e-01 -7.58830383e-02
5.95277786e-01 -2.57830862e-02 -4.99642283e-01 -4.65603471e-02
-7.46910572e-01 3.86885822e-01 4.77925926e-01 1.02549326e+00
6.26893818e-01 -2.37387970e-01 3.03081185e-01 5.67620993e-01
3.36180300e-01 7.63585627e-01 2.38218367e-01 -1.13511074e+00
7.55699813e-01 4.31425869e-01 4.32631522e-01 -1.17335975e+00
-4.73446429e-01 -2.16380998e-01 -7.85669148e-01 3.79343219e-02
5.63985109e-01 -2.18586251e-01 -9.89108682e-01 1.68324387e+00
6.84670389e-01 1.62539497e-01 -8.02590400e-02 1.36153364e+00
5.95693707e-01 5.09242237e-01 -3.46770108e-01 -6.52798042e-02
1.21557760e+00 -1.03272140e+00 -5.53901970e-01 -3.96772325e-01
4.75044936e-01 -6.26486719e-01 7.18486428e-01 5.32671154e-01
-9.98284101e-01 -7.75387406e-01 -8.78724337e-01 -1.51691481e-01
4.15940247e-02 4.11807656e-01 3.07301432e-01 2.90340811e-01
-6.85015500e-01 4.26282406e-01 -1.25147057e+00 -5.46372950e-01
-2.37053260e-02 3.59231949e-01 -8.02017987e-01 -1.98831335e-01
-1.00258827e+00 8.70185256e-01 3.48133653e-01 5.42549789e-01
-9.44307506e-01 -4.93915714e-02 -1.19844699e+00 -4.30078983e-01
8.08076918e-01 -8.36274803e-01 1.08569431e+00 -4.53107417e-01
-1.39335227e+00 9.58127022e-01 -2.04622209e-01 -1.95770636e-01
9.02940214e-01 -8.90943170e-01 8.00196975e-02 3.76763970e-01
1.44781575e-01 6.72281504e-01 9.35806513e-01 -1.34985042e+00
-4.66659993e-01 -5.70712686e-01 2.94614494e-01 5.39520621e-01
3.24069075e-02 -4.50731575e-01 -1.05196679e+00 -5.70788562e-01
5.17032027e-01 -1.36841559e+00 -3.05786312e-01 1.85154587e-01
-4.94035661e-01 9.28767677e-03 6.42015278e-01 -8.19735289e-01
5.99650443e-01 -1.85077596e+00 7.87597716e-01 1.77694231e-01
9.19393227e-02 3.53776328e-02 1.27488583e-01 1.89023629e-01
1.89577430e-01 -4.26010966e-01 -9.23003815e-03 -8.37259412e-01
-1.25816181e-01 4.10480261e-01 2.87621636e-02 7.51291335e-01
-7.02377409e-04 6.61664307e-01 -9.37648833e-01 -5.00479996e-01
4.95451808e-01 7.84408391e-01 -7.60446846e-01 6.72399044e-01
-1.18168756e-01 8.96662712e-01 -3.04189593e-01 4.31159735e-01
5.20595193e-01 -3.98889661e-01 2.76869953e-01 -5.26718915e-01
6.86129555e-02 3.15627754e-02 -1.38721502e+00 2.18874335e+00
-5.08100688e-01 7.51314610e-02 -6.17237873e-02 -9.08139467e-01
6.61716223e-01 4.97074485e-01 4.61886585e-01 -1.71153411e-01
1.06233723e-01 1.74568921e-01 -4.88362163e-01 -6.48828745e-01
1.30641252e-01 -2.76053380e-02 -2.48666108e-01 1.18691400e-01
2.19772220e-01 -3.87339480e-02 4.65209968e-02 5.31024002e-02
7.15916574e-01 7.61753917e-01 3.58270496e-01 1.85214892e-01
4.70101744e-01 -3.64119649e-01 8.66121173e-01 3.38268876e-01
7.34593272e-02 8.76459479e-01 3.89365464e-01 -6.76192760e-01
-1.06405377e+00 -1.20923543e+00 2.63464481e-01 6.29098713e-01
1.45094991e-01 -5.90374470e-01 -6.65252149e-01 -1.07670295e+00
-5.73143102e-02 1.77164003e-01 -6.29199147e-01 1.05068415e-01
-8.51796746e-01 -3.55209351e-01 2.09350631e-01 6.49942458e-01
4.61142987e-01 -5.56034267e-01 -9.24536824e-01 -5.58608249e-02
-6.68764234e-01 -1.49699664e+00 -7.45748580e-01 -6.18114099e-02
-8.91530335e-01 -1.27408671e+00 -9.99181390e-01 -6.35611832e-01
1.06846809e+00 3.42441231e-01 8.08163106e-01 -6.74451608e-03
-1.05102003e-01 4.62819666e-01 -3.32048118e-01 1.26138732e-01
3.45041603e-02 -8.92735049e-02 5.16832232e-01 2.11956546e-01
1.78387627e-01 -5.32325625e-01 -5.54164350e-01 6.41804814e-01
-4.48744833e-01 3.38297278e-01 4.49054986e-01 9.85682964e-01
8.48735392e-01 -1.95870668e-01 1.32840127e-01 -8.20080459e-01
-2.95207232e-01 -1.25360891e-01 -6.65090263e-01 4.17341329e-02
5.24020717e-02 -1.57002613e-01 5.52140653e-01 -4.27726597e-01
-1.07263231e+00 6.03860080e-01 -1.58377111e-01 -9.67106819e-01
-1.92464218e-01 2.34150961e-01 -3.35571051e-01 1.29484162e-01
4.15423483e-01 -5.96404355e-03 -1.01193808e-01 -6.92115247e-01
1.24628104e-01 2.82331675e-01 6.20379686e-01 -5.47715843e-01
9.35796857e-01 6.50599003e-01 7.35679716e-02 -8.84590328e-01
-1.31500971e+00 -5.75563610e-01 -1.12706590e+00 -3.08251560e-01
1.16739917e+00 -1.32797301e+00 -7.70040214e-01 4.26988691e-01
-1.32954264e+00 -8.95708203e-02 1.48980200e-01 8.55558038e-01
-6.86649323e-01 6.29467547e-01 -5.76283216e-01 -9.23399448e-01
-3.27697881e-02 -1.27167463e+00 1.49850833e+00 -1.54234558e-01
-2.59478152e-01 -8.53484988e-01 -4.78502475e-02 6.78641975e-01
-3.22805405e-01 6.16406143e-01 2.03598246e-01 -1.75833344e-01
-5.10969460e-01 -3.66601080e-01 2.16469198e-01 4.14736718e-01
1.55259684e-01 -4.52974498e-01 -8.02353740e-01 -5.48164666e-01
1.01779699e-01 -7.05264628e-01 7.00562358e-01 3.21883738e-01
1.00765836e+00 -2.46087819e-01 -2.96719372e-01 7.86771059e-01
9.59129333e-01 -2.73226470e-01 2.67435431e-01 8.20307732e-02
1.13104892e+00 6.58132136e-01 7.21323252e-01 5.52009642e-01
5.21788836e-01 1.03045845e+00 3.64571303e-01 -1.47605874e-02
1.66679040e-01 -6.89474761e-01 4.72888827e-01 9.60393548e-01
-5.32367885e-01 -8.72256532e-02 -6.73928320e-01 2.70763516e-01
-1.90446568e+00 -8.69250834e-01 1.35430872e-01 2.47006035e+00
6.25160098e-01 2.82984048e-01 2.73472905e-01 1.55385518e-02
6.89595580e-01 2.27119192e-01 -5.61126232e-01 4.16530758e-01
2.78762400e-01 -9.77010354e-02 3.08443129e-01 5.88500440e-01
-1.15519512e+00 6.71070755e-01 5.13200331e+00 1.52101293e-01
-8.11661005e-01 9.92863253e-02 3.00455362e-01 -4.32780325e-01
1.52213514e-01 -2.48546824e-02 -9.74944949e-01 2.97847241e-01
3.79356891e-01 5.96030235e-01 2.73433685e-01 7.66171694e-01
2.22305536e-01 -8.91132876e-02 -1.34114957e+00 1.23197305e+00
3.34459126e-01 -8.40259850e-01 -5.15361428e-02 4.13439088e-02
6.65866494e-01 -2.08185166e-01 -2.22567871e-01 1.52805403e-01
-8.06269050e-02 -6.60753429e-01 8.36585999e-01 4.65142310e-01
8.16280782e-01 -7.46037841e-01 7.25065947e-01 6.56831145e-01
-1.28489995e+00 2.18150511e-01 -2.97931135e-01 -1.48400096e-02
5.26046336e-01 5.55035233e-01 -5.47266603e-01 6.47710085e-01
9.09729779e-01 9.55302596e-01 -3.23303878e-01 5.54772854e-01
-5.15739977e-01 1.11571595e-01 -5.06937802e-01 3.35557669e-01
-3.93507583e-03 -7.01396838e-02 4.12844926e-01 8.97292435e-01
4.40751821e-01 4.07030806e-02 7.65272319e-01 3.59337360e-01
4.14164513e-02 -3.06497663e-01 -6.18102789e-01 3.41233462e-01
2.21959352e-01 1.19748271e+00 -6.37725770e-01 -3.64939868e-01
-4.88943130e-01 1.46582437e+00 3.83892298e-01 4.03297335e-01
-9.49415267e-01 5.16333915e-02 3.97977740e-01 1.38926357e-01
3.57396424e-01 -6.58509254e-01 4.94599462e-01 -1.80828762e+00
4.06395286e-01 -9.29273009e-01 4.82296348e-01 -6.97589040e-01
-1.19797695e+00 5.44871807e-01 3.88606995e-01 -1.43680036e+00
-5.86597383e-01 -6.24987185e-01 -1.45565912e-01 6.05077386e-01
-1.17894936e+00 -1.21097887e+00 -5.33311963e-01 6.39890134e-01
6.17603481e-01 1.44893706e-01 7.40139365e-01 2.71824390e-01
-4.23368663e-01 4.08623427e-01 -4.79063928e-01 2.88750112e-01
8.92553389e-01 -1.15545642e+00 6.01969838e-01 6.31489694e-01
2.47527868e-01 6.18151665e-01 5.93196273e-01 -6.95704281e-01
-1.50574756e+00 -9.94740665e-01 8.27358842e-01 -7.58360505e-01
1.24793708e-01 -8.24400365e-01 -5.49970806e-01 9.84914720e-01
-1.82151109e-01 3.81683290e-01 4.93145764e-01 8.23556408e-02
-3.90658677e-01 -2.55997777e-02 -9.37828183e-01 4.23529625e-01
1.44268346e+00 -4.90451276e-01 -5.54811776e-01 5.02421916e-01
5.83950460e-01 -8.95470142e-01 -7.21701562e-01 3.54934245e-01
6.07474267e-01 -1.03721523e+00 1.07318735e+00 -3.85238707e-01
1.56609625e-01 -4.64905322e-01 -2.83383965e-01 -1.18588424e+00
3.54654118e-02 -7.06635952e-01 -3.05476427e-01 6.75647557e-01
2.16510855e-02 -3.76262218e-01 1.01179028e+00 2.16298789e-01
1.67688057e-01 -6.89790368e-01 -1.00307882e+00 -7.59209514e-01
-4.66841847e-01 -2.81413823e-01 2.45011058e-02 6.29593074e-01
-2.31388837e-01 4.24590856e-01 -9.55154955e-01 4.39014256e-01
1.04488838e+00 4.68406193e-02 1.21655238e+00 -1.02349031e+00
-6.71608150e-01 5.64193487e-01 -4.79210526e-01 -1.49844825e+00
2.10107714e-01 -5.57898939e-01 2.09317923e-01 -1.23214149e+00
2.06523284e-01 1.19317342e-02 9.23796445e-02 3.41485798e-01
-1.20280705e-01 4.48893100e-01 2.82216996e-01 2.26755127e-01
-7.56129563e-01 6.03794217e-01 1.22358739e+00 2.33991578e-01
4.98747267e-02 1.29128575e-01 -8.24031681e-02 1.18229580e+00
4.77885514e-01 -4.01395977e-01 -4.54954565e-01 -6.44053161e-01
9.01093259e-02 3.35988671e-01 7.60885417e-01 -1.15505421e+00
1.03397965e-01 -8.30693990e-02 8.85188222e-01 -9.05932724e-01
8.71459365e-01 -9.11485851e-01 1.94121182e-01 5.27297199e-01
-1.54157490e-01 3.22650999e-01 -1.87111109e-01 7.82350779e-01
1.48030773e-01 1.10682810e-03 6.28177643e-01 -5.44312894e-01
-4.57783967e-01 4.91896093e-01 4.95373197e-02 1.45194054e-01
8.91978562e-01 -1.42611444e-01 4.25208926e-01 -5.70837855e-01
-9.48425651e-01 2.56745040e-01 6.74514472e-01 4.74695355e-01
8.92030835e-01 -1.51382601e+00 -5.59078336e-01 3.47818583e-01
1.46138957e-02 5.80287099e-01 3.02121788e-01 7.04692662e-01
-4.93508965e-01 3.53342474e-01 -4.94728014e-02 -9.71908092e-01
-1.22858930e+00 5.07892907e-01 2.59630561e-01 -1.64579853e-01
-8.12268138e-01 6.73776269e-01 2.10567847e-01 -7.04855025e-01
4.66426283e-01 -1.52981907e-01 1.36504591e-01 -1.41896501e-01
4.93051797e-01 3.16033870e-01 -7.74895847e-02 -9.34878290e-01
-4.59668517e-01 9.96158361e-01 -5.97231016e-02 -2.36179158e-01
1.39740968e+00 -1.43093899e-01 2.37673029e-01 6.37637317e-01
1.48870850e+00 2.83869891e-03 -1.68840349e+00 -3.01286936e-01
-3.77900571e-01 -7.50011623e-01 -3.48265886e-01 -3.08450013e-01
-1.11828661e+00 9.36638951e-01 3.83990586e-01 -5.49108624e-01
8.64647508e-01 -6.57700896e-02 9.39933598e-01 6.83646083e-01
7.03194499e-01 -1.07285011e+00 5.57032943e-01 4.15310264e-01
8.96573424e-01 -1.28759897e+00 2.70754874e-01 -4.39364761e-01
-5.40362895e-01 1.02759075e+00 9.33949053e-01 -2.01836810e-01
4.44961995e-01 -6.95115477e-02 2.30304934e-02 -8.96754265e-02
-5.32505929e-01 -1.07318005e-02 4.23006505e-01 3.93914133e-01
3.97121042e-01 -1.93081781e-01 7.16481209e-02 2.30666116e-01
3.06340028e-02 4.11977731e-02 3.57247025e-01 9.96066391e-01
-6.55341074e-02 -1.01113713e+00 -5.76166391e-01 3.82326469e-02
-3.94140989e-01 3.55047941e-01 -2.09503740e-01 8.28647017e-01
3.21671724e-01 7.33968139e-01 -7.31271580e-02 -5.24260998e-01
5.69498181e-01 3.01070348e-03 8.50147188e-01 -7.32354164e-01
-5.00911921e-02 3.85887176e-01 8.61498713e-02 -7.48800576e-01
-7.18748629e-01 -6.62340462e-01 -1.17093992e+00 -1.31633535e-01
-4.14824575e-01 6.09830134e-02 3.32622021e-01 9.28269804e-01
2.08969414e-01 1.34515584e-01 5.40804863e-01 -1.51459110e+00
-6.92668378e-01 -7.01258421e-01 -3.39356363e-01 7.01282442e-01
3.98806095e-01 -1.04820466e+00 -4.38762754e-01 3.04772168e-01] | [7.005630016326904, -0.916429340839386] |
ddc18c1b-64a3-46b9-b17d-dd3794d80c10 | nasgec-a-multi-domain-chinese-grammatical | 2305.16023 | null | https://arxiv.org/abs/2305.16023v1 | https://arxiv.org/pdf/2305.16023v1.pdf | NaSGEC: a Multi-Domain Chinese Grammatical Error Correction Dataset from Native Speaker Texts | We introduce NaSGEC, a new dataset to facilitate research on Chinese grammatical error correction (CGEC) for native speaker texts from multiple domains. Previous CGEC research primarily focuses on correcting texts from a single domain, especially learner essays. To broaden the target domain, we annotate multiple references for 12,500 sentences from three native domains, i.e., social media, scientific writing, and examination. We provide solid benchmark results for NaSGEC by employing cutting-edge CGEC models and different training data. We further perform detailed analyses of the connections and gaps between our domains from both empirical and statistical views. We hope this work can inspire future studies on an important but under-explored direction--cross-domain GEC. | ['Min Zhang', 'Fei Huang', 'Chen Li', 'Zhenghua Li', 'Haochen Jiang', 'Bo Zhang', 'Yue Zhang'] | 2023-05-25 | null | null | null | null | ['grammatical-error-correction'] | ['natural-language-processing'] | [-2.75802054e-02 -2.14615669e-02 -6.46039173e-02 -3.38636607e-01
-1.13657486e+00 -6.84507012e-01 2.97376424e-01 3.95083278e-01
-5.35711467e-01 1.00719142e+00 4.77158099e-01 -7.83152044e-01
1.76952705e-01 -4.43703204e-01 -6.15078092e-01 4.17248113e-03
6.61946595e-01 2.50795364e-01 1.82267413e-01 -4.23509836e-01
7.37583458e-01 -2.68517315e-01 -8.52408290e-01 2.74018466e-01
1.93645525e+00 1.35020718e-01 2.58044302e-01 7.01939285e-01
-2.45264411e-01 8.07105124e-01 -1.01695800e+00 -1.20308590e+00
-4.64412928e-01 -6.38122320e-01 -1.02036119e+00 -5.35316765e-01
7.85431862e-01 1.02401294e-01 -5.12891039e-02 1.46557283e+00
4.22833115e-01 -1.40455082e-01 7.67446995e-01 -6.57043755e-01
-1.58833027e+00 1.02147400e+00 -3.36870730e-01 3.52174997e-01
3.39736789e-01 -9.14247185e-02 1.02929151e+00 -1.00104952e+00
9.70229030e-01 1.19550145e+00 1.12483132e+00 1.18982732e+00
-7.18980491e-01 -7.69397140e-01 2.29966700e-01 2.38319620e-01
-1.00347686e+00 -3.68087858e-01 5.08146226e-01 -3.22788090e-01
8.71178329e-01 1.86221585e-01 7.09832087e-02 1.49662936e+00
7.94910789e-02 1.09506297e+00 1.19353759e+00 -9.57512379e-01
-7.01750964e-02 1.74914435e-01 7.69142926e-01 5.80100119e-01
5.92544973e-01 -3.10120910e-01 -6.41503334e-01 -2.78434381e-02
5.54104224e-02 -4.67105985e-01 -5.62390804e-01 6.81329906e-01
-7.62702763e-01 9.37964737e-01 -2.17809886e-01 4.10360694e-01
2.94606626e-01 9.92459208e-02 4.79923010e-01 6.34570956e-01
7.55338967e-01 6.62100911e-01 -8.46897602e-01 -6.63749456e-01
-8.67509246e-01 4.08766180e-01 1.10212886e+00 1.43529117e+00
2.29896441e-01 -7.06298649e-02 -7.35975206e-02 1.04975367e+00
3.38818103e-01 7.19144821e-01 6.87214434e-01 -7.90964305e-01
1.14000213e+00 4.76600647e-01 -7.25467876e-02 -7.90173471e-01
-9.49882939e-02 -4.30867672e-01 -4.52742726e-01 -5.16314268e-01
7.41293252e-01 -4.49669093e-01 -2.68405139e-01 1.65254247e+00
-2.59544462e-01 -1.87280864e-01 9.05356258e-02 7.52581298e-01
1.23854160e+00 2.30698198e-01 4.08637911e-01 8.30564722e-02
1.20441532e+00 -1.00827563e+00 -8.92848909e-01 -3.60716522e-01
1.31861484e+00 -8.67716253e-01 1.44890475e+00 2.82063961e-01
-1.11322379e+00 -1.95304751e-01 -8.80440414e-01 -6.47213221e-01
-2.31126219e-01 5.88271976e-01 1.71315551e-01 7.82483578e-01
-6.38912499e-01 4.80893791e-01 -8.47228825e-01 -3.65605563e-01
1.83006153e-01 -5.27296185e-01 -3.97013426e-02 -2.22057864e-01
-1.42619753e+00 9.98307586e-01 4.99448851e-02 -4.42254245e-01
-3.88765216e-01 -9.43928659e-01 -6.04672551e-01 -1.82156526e-02
8.77825767e-02 -1.70590416e-01 1.54574132e+00 -7.27654874e-01
-1.21981049e+00 1.02805042e+00 -4.83807743e-01 -3.71101528e-01
4.65994745e-01 -6.23053849e-01 -6.49034977e-01 -2.88249999e-01
2.62272745e-01 1.91846900e-02 2.57343531e-01 -8.57768655e-01
-6.07573390e-01 -4.87840831e-01 -4.30097997e-01 1.95401954e-03
-5.08512735e-01 6.63597643e-01 -1.96896389e-01 -1.12874913e+00
-1.33679718e-01 -8.27493310e-01 3.60582143e-01 -5.39340794e-01
-2.83177406e-01 -7.37473309e-01 5.08988678e-01 -1.27993000e+00
1.89254141e+00 -1.88051665e+00 1.68580875e-01 -3.49043041e-01
2.41849706e-01 4.30050462e-01 -9.31917354e-02 2.56571740e-01
8.68657902e-02 7.16887653e-01 -3.26976448e-01 -7.02736914e-01
-8.90097395e-02 -1.76084280e-01 -3.65940720e-01 2.20388830e-01
1.54012054e-01 1.00344646e+00 -1.22692418e+00 -3.79408807e-01
-6.99209154e-01 -5.49991094e-02 -6.37471318e-01 -6.97891861e-02
-1.31154090e-01 2.86733866e-01 -4.32420731e-01 6.76299870e-01
7.17189372e-01 -1.49210110e-01 2.36942992e-01 6.73698068e-01
-2.58464932e-01 9.59764183e-01 -5.89718878e-01 1.73685133e+00
-3.40962768e-01 1.00450373e+00 -6.73731118e-02 -6.09033883e-01
1.09874249e+00 1.60164684e-01 -4.23883438e-01 -6.57067418e-01
6.32629991e-02 8.03589940e-01 1.25433624e-01 -4.43579644e-01
1.11573124e+00 3.37520130e-02 -3.80656928e-01 8.24452281e-01
3.45614970e-01 -1.21322930e-01 2.74595529e-01 5.55520058e-01
1.03961396e+00 -1.36409938e-01 2.38950968e-01 -7.15379775e-01
4.74141836e-01 2.04295456e-01 6.45623088e-01 9.08986211e-01
-6.02924049e-01 6.84795618e-01 5.90258121e-01 9.98386294e-02
-8.40133846e-01 -7.53542185e-01 -4.11292344e-01 1.29280853e+00
-1.51824459e-01 -7.14976072e-01 -1.04408324e+00 -1.33660614e+00
1.41665891e-01 1.07346165e+00 -4.33978915e-01 -1.41718537e-01
-7.98312426e-01 -6.64083362e-01 1.10926008e+00 5.57948411e-01
3.82900923e-01 -8.21499527e-01 1.35919163e-02 1.16039388e-01
-5.28685451e-01 -9.01963949e-01 -7.95236886e-01 -1.52785135e-02
-8.56027424e-01 -1.23270822e+00 -6.32792115e-01 -9.15520191e-01
4.39988583e-01 1.73513576e-01 1.51492906e+00 7.00465024e-01
2.55780876e-01 4.74969298e-01 -7.90575206e-01 -8.43227983e-01
-9.11484182e-01 6.25184357e-01 -3.19757499e-02 -9.83129680e-01
1.09023094e+00 -1.05866641e-02 6.50948510e-02 8.29892382e-02
-2.77402937e-01 -8.74474794e-02 -9.13076177e-02 9.27045763e-01
-1.22980826e-01 -5.79174936e-01 9.96434450e-01 -1.47150171e+00
1.20162284e+00 -7.78838694e-01 -4.05358166e-01 7.38124251e-01
-7.55398929e-01 -1.41268373e-02 6.00440681e-01 -1.90374717e-01
-1.47880328e+00 -7.47891605e-01 -1.55094966e-01 3.86649728e-01
8.10419098e-02 8.00171196e-01 1.60653278e-01 3.21880095e-02
1.00436187e+00 2.30048690e-03 -2.53105074e-01 -6.86397552e-01
1.48415953e-01 1.16825593e+00 7.48601377e-01 -1.10052514e+00
3.71780664e-01 -4.65634555e-01 -8.73973668e-01 -5.76213717e-01
-1.12894416e+00 -2.49307975e-01 -6.42336309e-01 -6.97456524e-02
5.49121618e-01 -1.02299607e+00 -3.07727218e-01 6.36276901e-01
-1.66112828e+00 -4.54468429e-01 2.31696591e-01 3.52030277e-01
1.23949699e-01 4.62951452e-01 -1.18983853e+00 -6.36810303e-01
-3.86832505e-01 -8.27709258e-01 8.13356757e-01 5.01963317e-01
-4.83159781e-01 -1.43429685e+00 2.64795959e-01 4.74972457e-01
3.55397940e-01 -4.39504147e-01 9.71925735e-01 -8.77257586e-01
-2.09667087e-01 1.46998629e-01 -2.85919428e-01 4.71505105e-01
-2.73777217e-01 2.36534119e-01 -6.56151593e-01 -1.23962380e-01
-3.67275141e-02 -5.51051736e-01 1.05921626e+00 -8.82670879e-02
1.36663973e+00 -6.97182193e-02 -8.21366161e-02 4.01157588e-01
1.12403882e+00 8.08531046e-02 5.68132162e-01 4.76231486e-01
6.74798906e-01 4.21190202e-01 6.28169954e-01 2.49974519e-01
8.46169055e-01 2.78506905e-01 -2.29764447e-01 4.85173345e-01
-4.78470445e-01 -5.62543988e-01 4.59272206e-01 1.71926057e+00
-6.60856515e-02 -6.94069147e-01 -1.23743498e+00 8.05035651e-01
-1.65706062e+00 -7.84670889e-01 -7.31281102e-01 1.84884977e+00
1.43316865e+00 -1.77928224e-01 -4.38883662e-01 -1.41661495e-01
8.10740232e-01 -1.72820315e-01 -2.58094192e-01 -6.35451436e-01
-4.60914552e-01 4.42959249e-01 4.56673056e-01 5.54607987e-01
-9.04936373e-01 1.15885997e+00 6.72359467e+00 7.31583714e-01
-8.57252836e-01 4.11407441e-01 4.36730087e-01 2.02169344e-01
-6.23080015e-01 2.24428885e-02 -1.45071590e+00 8.80078673e-01
1.15985751e+00 -3.74334246e-01 2.10641921e-01 8.77913237e-01
-7.01127276e-02 -3.97737361e-02 -7.79850006e-01 5.73517323e-01
4.24763113e-01 -1.21472502e+00 -4.02776033e-01 -2.57545412e-01
1.17934525e+00 2.70019114e-01 -7.05071241e-02 8.04788589e-01
9.23912525e-01 -7.82163858e-01 8.65064144e-01 3.99449557e-01
9.20926750e-01 -3.46955448e-01 7.81145453e-01 5.98381877e-01
-2.86669195e-01 7.24138916e-02 -4.42548603e-01 -2.88026124e-01
-7.33682737e-02 5.15241146e-01 -4.57350999e-01 3.66666496e-01
7.46171117e-01 1.08976912e+00 -1.13799191e+00 6.98226988e-01
-7.87033558e-01 1.39741790e+00 2.98932552e-01 -4.27063584e-01
-1.22814506e-01 -1.02332905e-01 4.42464978e-01 1.60014343e+00
5.07599115e-01 2.22662687e-01 -3.91071081e-01 1.05178642e+00
-5.09354711e-01 6.56978786e-02 -2.15371370e-01 -3.93403806e-02
1.08042145e+00 8.51515234e-01 -1.23957008e-01 -4.50968981e-01
-6.73664749e-01 1.08352757e+00 1.03554237e+00 3.00588280e-01
-8.60187590e-01 -7.70164549e-01 4.48136777e-01 -1.34811565e-01
-1.13144130e-01 -4.58194822e-01 -9.90739405e-01 -1.74967849e+00
-2.14597136e-02 -1.14769089e+00 3.96526009e-01 -4.32395279e-01
-1.76336694e+00 2.73456395e-01 -4.02629077e-01 -8.25469613e-01
-1.28871510e-02 -8.43694746e-01 -8.01889241e-01 1.13671696e+00
-1.67977142e+00 -7.14490354e-01 -2.79766172e-01 3.52699101e-01
5.99924862e-01 -4.53166336e-01 7.64457881e-01 5.78253210e-01
-9.08209383e-01 1.27780974e+00 6.51648819e-01 5.69362283e-01
1.34909856e+00 -1.51398993e+00 7.49860764e-01 1.11805415e+00
3.83218490e-02 9.87275660e-01 3.51916671e-01 -1.03778541e+00
-9.01906133e-01 -1.17667186e+00 1.85178113e+00 -1.09361553e+00
7.84798026e-01 -3.58072490e-01 -1.29998803e+00 9.35045600e-01
3.35219949e-01 -4.07533169e-01 7.05402970e-01 6.72366381e-01
-2.90759325e-01 5.62014341e-01 -9.51424122e-01 5.00742316e-01
1.18146217e+00 -7.74822056e-01 -9.32308435e-01 1.44794494e-01
8.21414530e-01 -7.46922910e-01 -8.81619930e-01 -1.64239019e-01
7.09905028e-02 -6.09911263e-01 2.75470257e-01 -9.48125303e-01
1.17615533e+00 1.50179788e-01 1.51371852e-01 -1.85034025e+00
-3.10236871e-01 -3.35902035e-01 -7.06152096e-02 1.73188448e+00
5.70079446e-01 -6.10595584e-01 4.49141860e-01 8.91065836e-01
-8.30886066e-01 -2.68371820e-01 -7.27886379e-01 -8.36111903e-01
9.87230301e-01 -6.97536170e-01 6.23612523e-01 1.36799848e+00
4.46766883e-01 3.08052301e-01 -3.53007279e-02 -2.01234385e-01
3.48204345e-01 -2.86585867e-01 5.15873671e-01 -1.37754178e+00
1.47235870e-01 -5.20032525e-01 2.67996043e-01 -1.20545149e+00
6.51189089e-01 -1.29513109e+00 1.54917464e-01 -1.03249288e+00
3.17024291e-01 -7.29401112e-01 1.51500389e-01 3.55998307e-01
-8.43094587e-01 -5.69528826e-02 1.45555452e-01 8.37198794e-02
-6.73954844e-01 6.02612793e-01 1.17456639e+00 6.91098869e-02
1.11518458e-01 -3.43398213e-01 -1.22910488e+00 6.28797948e-01
7.76628911e-01 -7.30891526e-01 1.27999082e-01 -1.09977841e+00
2.19490618e-01 -1.85716748e-01 -4.44226619e-03 -6.05789244e-01
3.89511675e-01 -2.83570945e-01 -1.32472720e-03 -1.41549498e-01
-7.51820385e-01 -4.81491126e-02 -6.58344507e-01 1.04011290e-01
-6.69119835e-01 2.79583544e-01 1.18993573e-01 3.01261306e-01
-3.37587953e-01 -8.55655730e-01 7.42073894e-01 -2.81857669e-01
-5.65988362e-01 -2.27442965e-01 -4.19813067e-01 9.85709012e-01
3.14728737e-01 2.43386567e-01 -1.06417227e+00 -2.83562094e-01
-2.77866542e-01 3.03654879e-01 4.39179152e-01 6.45431638e-01
3.57832819e-01 -1.27718115e+00 -1.14785182e+00 3.76676349e-03
2.67138571e-01 -2.70839751e-01 1.68084931e-02 6.82687879e-01
-4.65281308e-01 6.30630791e-01 1.81316212e-01 -1.70095056e-01
-1.24815428e+00 -1.12733804e-01 9.84657779e-02 -2.92986810e-01
-3.34516943e-01 9.71438169e-01 -3.29396158e-01 -9.23257470e-01
3.80561128e-02 -2.61321962e-01 -1.84230506e-01 -1.49230659e-01
6.67364359e-01 7.49770522e-01 3.56215298e-01 -2.81142950e-01
-1.68852568e-01 5.91422617e-02 -2.67774403e-01 1.37199506e-01
1.13589573e+00 -3.71986389e-01 -2.51032472e-01 3.67803842e-01
9.81760442e-01 5.59472263e-01 -6.90889120e-01 -4.47970718e-01
5.26308417e-01 -4.97111589e-01 -6.75233975e-02 -1.25240993e+00
-7.06858635e-01 9.07722414e-01 -5.77044347e-03 -3.04036945e-01
6.88059747e-01 -1.33910328e-01 7.64498949e-01 2.15311185e-01
2.74973780e-01 -1.59103990e+00 2.33363174e-03 1.44776452e+00
8.44270587e-01 -1.43291461e+00 -2.35442623e-01 -1.97955281e-01
-8.41444433e-01 1.35335827e+00 1.04048193e+00 6.37284815e-02
4.69637543e-01 1.09082282e-01 1.46785611e-02 -1.89878732e-01
-7.51991749e-01 3.80727977e-01 3.54031265e-01 4.01009440e-01
1.45905495e+00 2.67004937e-01 -1.07580388e+00 1.38498020e+00
-4.73198473e-01 -5.67665733e-02 1.04864895e+00 5.97080648e-01
-2.72095531e-01 -1.27188933e+00 -2.18697816e-01 5.14711976e-01
-6.93863213e-01 -6.04300857e-01 -5.61162949e-01 7.34397709e-01
-4.83845994e-02 1.02303755e+00 5.75374588e-02 -1.01254821e-01
1.38566166e-01 3.47297311e-01 3.40699226e-01 -8.28972816e-01
-8.15442145e-01 -5.84649563e-01 3.67123038e-01 1.09880082e-01
-6.21946231e-02 -1.16431594e+00 -1.25065267e+00 -7.34730363e-01
-5.69206059e-01 3.36772203e-01 5.07207036e-01 8.57051194e-01
4.66931522e-01 2.64069587e-01 6.57709166e-02 2.29786247e-01
-6.76104665e-01 -1.47354579e+00 -5.59261024e-01 3.83156389e-01
3.54104936e-02 -3.44189316e-01 -5.12550116e-01 -5.79092242e-02] | [11.052448272705078, 10.735769271850586] |
31c373ce-b69b-4a53-8916-aea4b0a6e170 | a-dual-source-approach-for-3d-pose-estimation | 1509.06720 | null | http://arxiv.org/abs/1509.06720v2 | http://arxiv.org/pdf/1509.06720v2.pdf | A Dual-Source Approach for 3D Pose Estimation from a Single Image | One major challenge for 3D pose estimation from a single RGB image is the
acquisition of sufficient training data. In particular, collecting large
amounts of training data that contain unconstrained images and are annotated
with accurate 3D poses is infeasible. We therefore propose to use two
independent training sources. The first source consists of images with
annotated 2D poses and the second source consists of accurate 3D motion capture
data. To integrate both sources, we propose a dual-source approach that
combines 2D pose estimation with efficient and robust 3D pose retrieval. In our
experiments, we show that our approach achieves state-of-the-art results and is
even competitive when the skeleton structure of the two sources differ
substantially. | ['Björn Krüger', 'Hashim Yasin', 'Andreas Weber', 'Umar Iqbal', 'Juergen Gall'] | 2015-09-22 | a-dual-source-approach-for-3d-pose-estimation-1 | http://openaccess.thecvf.com/content_cvpr_2016/html/Yasin_A_Dual-Source_Approach_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Yasin_A_Dual-Source_Approach_CVPR_2016_paper.pdf | cvpr-2016-6 | ['pose-retrieval'] | ['computer-vision'] | [ 1.74440265e-01 -3.12622525e-02 -1.46702215e-01 -1.02056295e-01
-1.32725108e+00 -7.90770590e-01 2.39807785e-01 -1.86563417e-01
-5.97647250e-01 4.20523465e-01 -3.74678075e-02 -9.15080979e-02
2.59303451e-01 -4.62422907e-01 -8.59750211e-01 -3.75752389e-01
2.01243863e-01 8.01876664e-01 5.67838252e-01 3.99304777e-02
8.72494876e-02 7.18830824e-01 -1.46410096e+00 -1.91548288e-01
4.72102165e-01 1.32930517e+00 2.28660718e-01 7.69356132e-01
-1.47192897e-02 2.03637302e-01 -2.95165032e-01 -2.50812531e-01
6.89376950e-01 -3.02841991e-01 -6.37898564e-01 4.38762605e-01
7.10575521e-01 -6.63003802e-01 -9.76787806e-02 8.12508643e-01
7.03986824e-01 -5.85814938e-02 4.50434625e-01 -1.15903187e+00
2.66509682e-01 -2.96688527e-01 -7.71594822e-01 -2.25644395e-01
6.43205225e-01 2.78602075e-02 6.00257456e-01 -1.01202631e+00
9.08977509e-01 9.82757270e-01 6.37984455e-01 5.44471681e-01
-1.16762495e+00 -3.11084718e-01 5.00670746e-02 -2.35976607e-01
-1.48304141e+00 -3.62501740e-01 9.92616355e-01 -3.99836689e-01
6.92874491e-01 9.97426659e-02 9.06004667e-01 1.06420445e+00
-2.87262857e-01 8.88376415e-01 1.08686256e+00 -4.48259532e-01
2.56440520e-01 -1.22294419e-01 -4.24969345e-01 7.45085716e-01
3.14969063e-01 -4.47773235e-03 -6.47766888e-01 -3.14336121e-02
1.19536507e+00 -5.68782538e-02 -2.44099930e-01 -1.01998687e+00
-1.34343553e+00 5.77440739e-01 4.16872501e-01 -2.21012354e-01
-2.99797863e-01 3.54187727e-01 7.16635361e-02 -1.93764046e-01
3.70298117e-01 3.03201020e-01 -5.64455450e-01 -3.47872674e-01
-9.60785210e-01 2.52504557e-01 5.83417296e-01 1.12804091e+00
8.21103454e-01 -3.72489482e-01 5.17854631e-01 4.50460762e-01
5.27347207e-01 7.57981598e-01 2.75842667e-01 -1.27768028e+00
6.06185794e-01 6.49067521e-01 3.06557178e-01 -8.66022527e-01
-4.13990766e-01 -1.86820954e-01 -3.19522530e-01 1.32626757e-01
7.07588911e-01 1.07166141e-01 -1.11033165e+00 1.55104184e+00
7.12443590e-01 2.40756236e-02 -7.76498392e-02 1.15591431e+00
7.23037004e-01 2.66549945e-01 -3.27018678e-01 8.49749194e-04
1.00999117e+00 -9.14190829e-01 -3.36487144e-01 -6.09710574e-01
2.41181448e-01 -8.02673876e-01 8.22324634e-01 2.72009879e-01
-1.12931883e+00 -3.45183462e-01 -1.08117700e+00 -2.74108350e-01
-1.07661225e-01 2.31808335e-01 5.21323860e-01 4.53747243e-01
-7.80754387e-01 5.11131108e-01 -1.21946847e+00 -3.55494887e-01
2.37270713e-01 4.48763281e-01 -7.00494230e-01 -3.15524757e-01
-7.21067667e-01 8.66628587e-01 2.61926174e-01 1.50697082e-01
-7.75476277e-01 -3.90222698e-01 -9.99679685e-01 -5.56705534e-01
6.21019185e-01 -6.70374632e-01 1.33042288e+00 -4.71508592e-01
-1.30111754e+00 1.14334047e+00 -7.01727867e-02 1.39060825e-01
7.71349013e-01 -5.29087007e-01 4.74665970e-01 5.56591570e-01
1.02998883e-01 7.88457572e-01 8.00734282e-01 -1.61498725e+00
-2.62155175e-01 -8.39570940e-01 -6.65395483e-02 4.57232594e-01
-1.78372055e-01 -3.05096477e-01 -1.27464354e+00 -2.95965433e-01
6.82133019e-01 -1.25169063e+00 -3.31528097e-01 4.29456651e-01
-3.49224865e-01 2.73183346e-01 6.64073884e-01 -5.77175140e-01
7.41171539e-01 -1.83315277e+00 4.76122648e-01 2.38963351e-01
1.88083559e-01 8.16177577e-02 -3.49115185e-03 1.50661200e-01
2.46751562e-01 -1.05342038e-01 -4.04420085e-02 -7.15897501e-01
-7.40185976e-02 4.36530232e-01 -1.24335615e-02 4.96847272e-01
1.79069951e-01 7.72122562e-01 -9.55432892e-01 -7.36376464e-01
2.80812085e-01 5.18074632e-01 -5.59910893e-01 4.42778796e-01
-8.76318850e-03 7.76060283e-01 -7.35800683e-01 9.57486868e-01
5.56197822e-01 -4.67610508e-01 1.58931807e-01 -4.90277112e-01
5.32690473e-02 1.37293071e-01 -1.45250213e+00 2.48448563e+00
-1.22992903e-01 2.11911112e-01 6.09233752e-02 -5.85680127e-01
5.30086458e-01 3.18941116e-01 9.24281120e-01 -1.69282332e-01
3.96070004e-01 5.31245470e-01 -5.83935499e-01 -2.31927067e-01
3.66378069e-01 -1.00176781e-01 -2.01726124e-01 4.84850794e-01
1.00982271e-01 -7.24555433e-01 -2.44915765e-02 4.83469628e-02
1.17632008e+00 7.82882988e-01 3.85380834e-01 2.83914238e-01
2.63164431e-01 1.49319366e-01 5.69463015e-01 3.10799956e-01
-2.36789033e-01 1.10611761e+00 3.05429667e-01 -2.53708065e-01
-1.11160612e+00 -1.28819442e+00 1.61357857e-02 4.15534586e-01
4.26881641e-01 -5.81494093e-01 -6.52937531e-01 -7.94898987e-01
2.89327707e-02 -1.21463723e-01 -4.60769206e-01 6.25523776e-02
-5.95437229e-01 -4.25957173e-01 3.50233138e-01 8.99015725e-01
4.61925685e-01 -2.72687078e-01 -1.08106589e+00 -1.93383515e-01
-3.78641158e-01 -1.62282920e+00 -3.95307302e-01 2.79879659e-01
-1.17789972e+00 -1.33583808e+00 -8.72903824e-01 -4.43767846e-01
9.69464123e-01 3.53202105e-01 1.15327144e+00 1.35296509e-01
-2.71370351e-01 5.36984801e-01 -5.30010223e-01 -3.44099432e-01
5.25450893e-02 4.56080474e-02 1.89174280e-01 -3.61588806e-01
-1.66440532e-02 -5.33979356e-01 -6.29771292e-01 4.95819539e-01
-7.93713748e-01 2.01470964e-03 8.45602691e-01 5.58758974e-01
1.17884111e+00 -2.97688216e-01 -1.26374349e-01 -4.74814504e-01
-2.20808417e-01 1.64707512e-01 -6.07136965e-01 4.53377403e-02
-3.17694128e-01 1.22833088e-01 8.93256962e-02 -4.04557079e-01
-6.68277383e-01 1.02938843e+00 -2.46391565e-01 -8.58072639e-01
-3.00316215e-01 3.31478894e-01 -3.39169860e-01 -2.85859138e-01
4.76380467e-01 -4.15748768e-02 7.00525567e-02 -7.85920262e-01
3.89656335e-01 5.27312100e-01 7.53480494e-01 -6.39947653e-01
1.11820030e+00 6.11682236e-01 2.23665372e-01 -5.97428262e-01
-9.53921378e-01 -6.01685762e-01 -1.23253942e+00 -3.64231348e-01
8.70792985e-01 -1.09712744e+00 -3.45197648e-01 5.23209810e-01
-1.04625118e+00 -3.02234679e-01 -2.19910696e-01 7.57593632e-01
-9.36092079e-01 4.34140861e-01 -4.35419053e-01 -6.97568536e-01
-1.51663110e-01 -1.34995651e+00 1.75759482e+00 -2.39672922e-02
-1.11702137e-01 -5.83515882e-01 2.16265887e-01 4.69069093e-01
-2.24280015e-01 5.97803771e-01 3.12535912e-01 -2.61956513e-01
-8.01612794e-01 -6.51967466e-01 -1.57775674e-02 1.20077714e-01
3.06597035e-02 -2.80583173e-01 -9.50521886e-01 -2.96039134e-01
-9.09494981e-02 -7.95461416e-01 5.07007420e-01 2.22287536e-01
8.59917819e-01 2.77267903e-01 -2.79319614e-01 6.27385914e-01
1.40082550e+00 -2.95962274e-01 4.88623440e-01 2.64965564e-01
8.67635429e-01 4.98053163e-01 8.55526745e-01 2.70296127e-01
4.17857736e-01 9.48664665e-01 5.24519801e-01 -4.88625327e-03
-1.91091761e-01 -5.40323436e-01 1.41545296e-01 8.91925752e-01
-4.69661802e-01 8.22416097e-02 -1.01124489e+00 3.01955193e-01
-1.85613620e+00 -5.35533607e-01 1.36158764e-01 2.42611933e+00
8.06015611e-01 9.17041153e-02 2.36749962e-01 2.77926177e-01
3.70495796e-01 -1.34113461e-01 -7.90635645e-01 4.97036874e-01
2.30025142e-01 7.65213966e-02 6.76430702e-01 3.09621900e-01
-1.06448650e+00 7.73149967e-01 7.05971718e+00 4.83431011e-01
-8.88273656e-01 -3.96135896e-02 1.64108813e-01 -4.08781260e-01
-1.08244531e-01 -2.16127671e-02 -6.53699517e-01 1.34215161e-01
4.53393936e-01 4.04162675e-01 4.21177177e-03 9.32990551e-01
-1.19472831e-01 -4.83252823e-01 -1.08343709e+00 1.25555539e+00
2.16899499e-01 -9.14626539e-01 -2.28416175e-01 3.09689999e-01
7.03447461e-01 5.91546595e-02 -2.48536378e-01 -2.23779306e-01
6.23415001e-02 -6.40233338e-01 9.68441248e-01 3.66549492e-01
1.06303442e+00 -5.74486434e-01 6.01913750e-01 5.76015830e-01
-1.30672336e+00 3.26707602e-01 -1.21994533e-01 7.76893198e-02
3.05365264e-01 4.69374299e-01 -4.30548817e-01 7.70331323e-01
7.22985685e-01 7.18490064e-01 -7.08622515e-01 1.06084430e+00
-4.84440029e-01 1.18491100e-02 -7.34345257e-01 2.61359811e-01
-5.06980829e-02 -3.55794728e-02 4.18700725e-01 4.50223386e-01
4.86397058e-01 1.95513979e-01 5.22735715e-01 5.24909317e-01
2.77522784e-02 -1.86656356e-01 -6.44365668e-01 1.45653042e-03
4.59753364e-01 1.32715321e+00 -1.06390560e+00 -7.93081596e-02
-4.45803553e-01 1.18242502e+00 3.10363322e-01 1.32660612e-01
-6.50524378e-01 -5.32496646e-02 4.96891588e-01 1.15964472e-01
3.68873090e-01 -8.37998807e-01 -1.54165745e-01 -1.39492738e+00
1.90572664e-01 -4.65132594e-01 2.89302200e-01 -9.13934588e-01
-1.06276155e+00 4.93562788e-01 2.79077530e-01 -1.47167194e+00
-4.67333734e-01 -9.35783446e-01 -3.22305672e-02 6.34556174e-01
-1.39115179e+00 -1.32705295e+00 -6.69025004e-01 5.89098871e-01
3.56708795e-01 3.53958964e-01 7.94294775e-01 3.01210195e-01
-3.72492194e-01 3.94389153e-01 -3.24058205e-01 2.05888048e-01
6.83369935e-01 -1.26633191e+00 3.20325226e-01 7.84672856e-01
3.77903372e-01 4.90444541e-01 4.81292844e-01 -6.29841506e-01
-1.97245800e+00 -6.21111035e-01 4.54534978e-01 -1.01218915e+00
2.16203153e-01 -3.46943498e-01 -3.45825911e-01 6.42281473e-01
-5.50049186e-01 2.69304335e-01 6.55095398e-01 -1.35047689e-01
-2.35490978e-01 1.19155914e-01 -9.91328657e-01 2.40975529e-01
1.11462426e+00 -4.98225868e-01 -5.60203612e-01 1.37243375e-01
4.86448288e-01 -1.01793909e+00 -9.34553206e-01 5.18137097e-01
6.53113723e-01 -8.26908290e-01 1.30986702e+00 -3.17396343e-01
3.01320940e-01 -4.22942102e-01 -5.37719607e-01 -1.02958274e+00
3.89867127e-01 -2.44124010e-01 -3.20521086e-01 7.17259407e-01
1.02236532e-01 3.92695032e-02 1.25885856e+00 8.14724028e-01
8.18622783e-02 -7.61129081e-01 -9.09080803e-01 -1.00516522e+00
-2.26418540e-01 -6.54736042e-01 3.14688385e-01 4.23764259e-01
-2.07043648e-01 4.10038143e-01 -3.22427779e-01 1.00758366e-01
7.66143501e-01 4.61034149e-01 1.18727231e+00 -1.22120094e+00
-1.12929687e-01 -1.67308841e-02 -6.40275538e-01 -1.48965132e+00
-1.42658921e-02 -5.27608573e-01 4.30131614e-01 -1.54774785e+00
2.43643820e-01 -2.54072458e-01 1.44761235e-01 5.36229610e-01
-1.18449569e-01 7.41154134e-01 1.85403451e-01 4.27440703e-01
-6.01545870e-01 4.86336082e-01 1.26825321e+00 2.90311277e-01
-1.42017782e-01 -9.38013121e-02 -3.58356357e-01 9.57340777e-01
3.47025275e-01 -2.88213462e-01 -4.77911502e-01 -6.03958189e-01
2.14434162e-01 8.92126039e-02 5.14150321e-01 -9.35467124e-01
7.89937191e-03 -1.14027567e-01 7.00303853e-01 -9.95470166e-01
8.42138290e-01 -9.71883357e-01 3.48979910e-03 3.53758574e-01
7.60038495e-02 1.07304551e-01 2.51817371e-04 7.95410931e-01
-1.21466247e-02 -9.95342582e-02 5.78168631e-01 -5.80583036e-01
-7.66124785e-01 5.04316926e-01 1.06462047e-01 7.70820975e-02
8.87047589e-01 -3.96508187e-01 2.07144573e-01 -4.40084934e-01
-6.46306872e-01 8.92317072e-02 9.73845363e-01 3.21807384e-01
8.20525229e-01 -1.61338913e+00 -2.57437319e-01 1.64539859e-01
2.15804771e-01 6.16578221e-01 -2.73051053e-01 8.32537293e-01
-6.25739455e-01 2.67418891e-01 -2.48162851e-01 -8.24692130e-01
-1.22173595e+00 3.47354025e-01 1.57223120e-02 -1.49632692e-01
-4.71233606e-01 7.24458814e-01 -1.19467713e-01 -6.76671445e-01
2.83033818e-01 -1.17516026e-01 2.86102176e-01 2.69315671e-02
4.09180343e-01 2.26920754e-01 2.07696334e-01 -9.73049402e-01
-5.93631923e-01 1.01621807e+00 3.94381583e-01 -4.91955489e-01
1.35939467e+00 -2.05452219e-01 1.48954019e-01 4.83057886e-01
1.34989309e+00 -6.81504682e-02 -1.55245292e+00 -1.23720266e-01
-1.75101340e-01 -1.01089263e+00 8.07827190e-02 -5.44386089e-01
-1.09379113e+00 9.73061621e-01 4.42080706e-01 -3.54022533e-01
1.22492754e+00 2.52672404e-01 9.44901049e-01 4.96791810e-01
8.00740004e-01 -1.13661230e+00 5.06365240e-01 3.28474283e-01
7.02819526e-01 -1.43924010e+00 4.11834270e-01 -6.30226195e-01
-3.55541438e-01 1.10307801e+00 6.78902268e-01 -1.40546430e-02
3.31219971e-01 1.31406337e-01 2.70740211e-01 -3.62837970e-01
-1.90060735e-01 -4.98354435e-01 5.92041790e-01 5.83061635e-01
1.67424649e-01 -3.62196267e-01 -1.34153441e-01 4.38537806e-01
-2.50643194e-02 -1.38936162e-01 2.45480746e-01 1.40702510e+00
-2.58514851e-01 -1.33747792e+00 -6.25631690e-01 -2.53325123e-02
-2.39051878e-01 3.31749767e-01 -6.89299345e-01 8.73640537e-01
-1.48195714e-01 6.08803511e-01 -1.34182587e-01 -5.46563983e-01
4.88859564e-01 1.21638440e-01 1.08183515e+00 -6.25078082e-01
9.41925347e-02 6.40945554e-01 -6.68489560e-02 -1.06828249e+00
-6.80761874e-01 -6.95883691e-01 -1.32719040e+00 8.89170244e-02
-4.26143497e-01 -1.58883914e-01 9.18472290e-01 8.89019787e-01
2.19188020e-01 4.30813320e-02 4.68725979e-01 -1.53460801e+00
-5.32412589e-01 -6.22530341e-01 -5.66122055e-01 3.95828784e-01
3.16055745e-01 -1.05299604e+00 -1.54666618e-01 1.31577939e-01] | [6.95778751373291, -1.1618390083312988] |
4900cd9b-8096-4d47-a03e-01d4c7407769 | knowledge-transfer-for-on-device-speech | 2210.14977 | null | https://arxiv.org/abs/2210.14977v3 | https://arxiv.org/pdf/2210.14977v3.pdf | Knowledge Transfer For On-Device Speech Emotion Recognition with Neural Structured Learning | Speech emotion recognition (SER) has been a popular research topic in human-computer interaction (HCI). As edge devices are rapidly springing up, applying SER to edge devices is promising for a huge number of HCI applications. Although deep learning has been investigated to improve the performance of SER by training complex models, the memory space and computational capability of edge devices represents a constraint for embedding deep learning models. We propose a neural structured learning (NSL) framework through building synthesized graphs. An SER model is trained on a source dataset and used to build graphs on a target dataset. A relatively lightweight model is then trained with the speech samples and graphs together as the input. Our experiments demonstrate that training a lightweight SER model on the target dataset with speech samples and graphs can not only produce small SER models, but also enhance the model performance compared to models with speech samples only and those using classic transfer learning strategies. | ['Björn W. Schuller', 'Kun Qian', 'Thanh Tam Nguyen', 'Zhao Ren', 'Yi Chang'] | 2022-10-26 | null | null | null | null | ['speech-emotion-recognition'] | ['speech'] | [ 2.97150999e-01 3.82660598e-01 -1.79946795e-01 -3.84632409e-01
-5.70054889e-01 6.79492503e-02 2.85685450e-01 -3.80708762e-02
-1.77946314e-01 4.42997426e-01 1.07420221e-01 -3.20171893e-01
4.30886656e-01 -6.99937403e-01 -8.50852907e-01 -3.84692758e-01
-5.09668812e-02 9.82911214e-02 6.75303861e-02 -3.82702500e-02
-3.30538660e-01 3.81581396e-01 -1.51114166e+00 5.73452413e-01
8.25318575e-01 1.57074308e+00 1.91471934e-01 7.48704195e-01
-3.14861000e-01 9.11950588e-01 -8.35768461e-01 -4.64935333e-01
-1.98843226e-01 -5.24150074e-01 -4.25218403e-01 -6.37296289e-02
-5.23643792e-02 8.49337131e-02 -5.98375142e-01 8.70402813e-01
8.77464354e-01 1.74928442e-01 4.42464054e-01 -1.50094533e+00
-6.58365428e-01 6.88360333e-01 -5.92161119e-02 -1.83107272e-01
3.74975771e-01 -2.03616321e-01 6.85783029e-01 -9.75552917e-01
3.75358254e-01 1.04335606e+00 8.17235589e-01 8.72896135e-01
-8.25167000e-01 -8.32050204e-01 1.95078984e-01 3.04270178e-01
-1.29806113e+00 -6.57801867e-01 1.27230608e+00 1.22254275e-01
1.27114856e+00 2.15260133e-01 8.35968673e-01 1.59685588e+00
1.36583701e-01 1.15215993e+00 6.36874259e-01 -4.30934012e-01
6.54687107e-01 5.74144959e-01 2.88178116e-01 6.32851720e-01
-7.09404349e-02 -7.42999837e-02 -8.05113494e-01 -1.91343561e-01
5.24999201e-01 -7.74659663e-02 -2.05338612e-01 -7.96120167e-02
-4.40422237e-01 5.85355878e-01 4.92351651e-01 4.13878143e-01
-4.78612840e-01 1.96559951e-02 7.98542976e-01 3.88584942e-01
6.50318503e-01 1.20042570e-01 -5.15404224e-01 -5.42740107e-01
-7.09566712e-01 -4.11579013e-01 1.02687562e+00 9.30381358e-01
5.30283570e-01 4.82252449e-01 2.89531380e-01 9.14537191e-01
-2.04298440e-02 3.26457441e-01 4.69163835e-01 -9.99875739e-02
5.09751499e-01 9.64561403e-01 -3.92387271e-01 -9.57231641e-01
-2.65449226e-01 -3.29442799e-01 -9.23158169e-01 -1.10344172e-01
-1.01167649e-01 -4.15567368e-01 -8.95607829e-01 1.54313755e+00
6.60940334e-02 5.93552768e-01 1.91967279e-01 5.50640345e-01
1.07263851e+00 9.35845435e-01 2.82748640e-01 -1.64731562e-01
7.97383726e-01 -9.60130453e-01 -9.52683568e-01 -4.91293758e-01
8.94548297e-01 -1.09773666e-01 1.20492578e+00 5.81980824e-01
-9.41721976e-01 -6.77708149e-01 -1.26223278e+00 -5.02835773e-03
-6.83003366e-01 4.05747056e-01 6.29547298e-01 9.59220946e-01
-1.25293863e+00 4.42373812e-01 -8.41095328e-01 -3.90226692e-01
6.65233433e-01 6.98062241e-01 -2.03444600e-01 -6.08287798e-03
-1.20526087e+00 6.17573738e-01 1.47255361e-01 2.83468127e-01
-5.57563961e-01 -4.06000167e-01 -1.07235181e+00 4.17568833e-01
2.72264719e-01 -2.72971928e-01 7.76212752e-01 -1.34132266e+00
-1.95261192e+00 5.04093111e-01 -3.29489931e-02 -4.37299460e-01
1.78091928e-01 -3.35138649e-01 -1.05948567e+00 9.46758036e-03
-6.65052831e-01 3.90734673e-01 1.05224288e+00 -1.21875513e+00
-1.89477414e-01 -2.26624161e-01 -2.25017458e-01 4.39465651e-03
-1.12096262e+00 1.36381045e-01 -4.51764613e-01 -3.55712503e-01
-3.41819644e-01 -8.05301011e-01 -2.44347025e-02 -1.77036479e-01
-2.24208817e-01 -2.41719127e-01 1.25316894e+00 -7.87775159e-01
1.58724964e+00 -2.36124301e+00 4.15367670e-02 3.07227671e-01
1.36801854e-01 6.62542760e-01 -2.22782120e-01 2.78463751e-01
-1.07205085e-01 1.31790444e-01 1.86986312e-01 -5.73418140e-01
-3.77885215e-02 -5.08396327e-03 9.06571373e-02 -3.60954441e-02
3.80144238e-01 1.16814935e+00 -7.68634677e-01 -2.94617146e-01
3.45946103e-01 9.40716982e-01 -3.86996388e-01 4.36569124e-01
-2.03280039e-02 8.51574391e-02 -4.06186521e-01 5.16468346e-01
4.05094028e-01 -5.34258425e-01 2.64771253e-01 -3.45832705e-01
5.27497172e-01 2.26367012e-01 -8.97233129e-01 1.47071314e+00
-9.18897510e-01 7.13202953e-01 8.29452351e-02 -1.13295305e+00
1.28358674e+00 5.97109139e-01 4.57380533e-01 -8.32840800e-01
5.90460241e-01 1.00568300e-02 -1.89736381e-01 -6.18903637e-01
2.74831265e-01 -2.34813541e-02 -1.54871300e-01 1.65383756e-01
2.21515149e-01 1.84313834e-01 -6.10034406e-01 1.87895909e-01
1.21880519e+00 -2.23877370e-01 1.36068417e-02 1.85930863e-01
4.24697876e-01 -4.62961376e-01 3.97902310e-01 1.50672093e-01
-8.46241415e-02 1.88049689e-01 3.21400583e-01 -5.14018059e-01
-8.10736537e-01 -9.38013613e-01 2.92967498e-01 9.18583393e-01
6.93664849e-02 -7.12928295e-01 -9.47428048e-01 -9.86062467e-01
-4.56973284e-01 6.51309907e-01 -4.91344899e-01 -7.64786601e-01
-1.69767365e-01 -4.39958751e-01 4.09271121e-01 9.63982701e-01
5.45689106e-01 -1.39244437e+00 -4.68136162e-01 1.32285774e-01
1.29554465e-01 -1.45596075e+00 -3.43035430e-01 3.35978925e-01
-7.87833095e-01 -5.75539649e-01 -5.82562983e-01 -1.10362363e+00
7.96279967e-01 -6.44006953e-02 8.24525356e-01 1.60125196e-01
-7.35686049e-02 4.31954533e-01 -5.81046164e-01 -7.16818571e-01
-4.62848037e-01 7.94010460e-02 1.39661774e-01 4.61853713e-01
5.72983801e-01 -5.25975823e-01 -2.28988960e-01 1.02554470e-01
-7.93177009e-01 1.54506803e-01 5.77471197e-01 8.95752370e-01
7.64484853e-02 1.75143823e-01 1.03210926e+00 -9.88935947e-01
8.25066686e-01 -4.85517740e-01 -8.30611587e-02 3.32540572e-01
-3.99900079e-01 -7.63511285e-02 7.30001330e-01 -9.30015981e-01
-1.14406669e+00 1.65696323e-01 -2.22907677e-01 -8.22221696e-01
6.16843738e-02 6.83278441e-01 -6.50505424e-01 -3.26573923e-02
6.67670071e-01 -5.68809658e-02 -2.10944954e-02 -8.85912478e-02
9.69074592e-02 1.28788018e+00 1.17597304e-01 -3.26342821e-01
2.79862911e-01 3.00983811e-05 -4.32933807e-01 -1.17545509e+00
-3.01141620e-01 -2.27927148e-01 -3.59026223e-01 -4.83370930e-01
7.63716280e-01 -8.62160504e-01 -5.73895574e-01 4.80360389e-01
-1.06558299e+00 -6.54220343e-01 -9.87735167e-02 4.47250038e-01
-1.89401776e-01 -3.55899185e-02 -6.25903845e-01 -1.22081685e+00
-5.08318126e-01 -9.85156298e-01 1.17888594e+00 3.07308435e-01
-3.19563031e-01 -1.23866260e+00 -3.62729996e-01 1.33066937e-01
4.27518904e-01 9.57050323e-02 8.08055997e-01 -8.96300614e-01
-6.95215091e-02 -6.85651243e-01 -8.33154023e-02 7.12456822e-01
2.85832703e-01 -3.61617953e-01 -1.35844553e+00 -1.42882809e-01
1.69788808e-01 -5.16618907e-01 3.50258470e-01 2.10005075e-01
1.39739692e+00 -8.33291262e-02 -5.99635422e-01 3.98034513e-01
1.05828726e+00 6.60187662e-01 8.73297393e-01 -1.74024746e-01
9.88517761e-01 4.29496646e-01 2.01487944e-01 2.91552603e-01
1.75982922e-01 3.47158283e-01 -2.55255885e-02 -3.05167168e-01
-1.41156036e-02 -4.89502788e-01 7.07480550e-01 1.18700314e+00
1.84838057e-01 -4.18719828e-01 -8.85523558e-01 3.15322995e-01
-1.81951535e+00 -5.96442103e-01 3.03430378e-01 2.00425696e+00
4.41185743e-01 3.15661490e-01 1.87635925e-02 4.16653782e-01
6.84772849e-01 -9.73711312e-02 -7.35075235e-01 -5.22310913e-01
9.24785808e-02 6.52707696e-01 -9.74296499e-03 1.67281523e-01
-7.40963221e-01 9.93511438e-01 6.08928442e+00 7.24127948e-01
-1.68840718e+00 -2.85751224e-02 9.39855218e-01 2.98868846e-02
-2.09706694e-01 -5.56020856e-01 -5.70553541e-01 4.70943749e-01
1.62245023e+00 -8.33630785e-02 3.60144645e-01 1.18216765e+00
1.32850129e-02 3.24830264e-01 -1.24239421e+00 1.63909733e+00
3.19190890e-01 -1.18728089e+00 2.80332882e-02 -1.01405874e-01
4.52678859e-01 -3.54546279e-01 -1.23352930e-02 7.23319530e-01
-1.23491667e-01 -1.16282046e+00 3.73579502e-01 1.37519076e-01
9.47655976e-01 -8.60042274e-01 8.66149962e-01 3.34949374e-01
-1.29705930e+00 -1.45559803e-01 -1.69244081e-01 -1.22103952e-01
1.03399299e-01 4.65968132e-01 -9.59167540e-01 4.73582774e-01
7.79004037e-01 7.32819259e-01 -3.79702568e-01 6.01920724e-01
2.84115244e-02 8.74279976e-01 -1.03975125e-01 -5.99354982e-01
-2.15055987e-01 -8.92129242e-02 6.32590353e-02 1.10293448e+00
3.74050558e-01 5.31971902e-02 2.88095996e-02 6.58733368e-01
-4.05896962e-01 2.91654140e-01 -8.68431330e-01 -6.23294294e-01
4.46347713e-01 1.22183299e+00 -6.25888288e-01 -4.27132100e-01
-5.55780411e-01 1.27700603e+00 4.72361296e-01 4.57074344e-01
-8.71761203e-01 -4.48483884e-01 3.73094738e-01 -6.51623383e-02
1.11271217e-01 -1.52053520e-01 -4.12052035e-01 -9.75254416e-01
1.81650281e-01 -8.06871355e-01 1.59578502e-01 -8.72140467e-01
-1.24752676e+00 1.14619529e+00 -4.04127181e-01 -1.15049541e+00
-8.69572163e-02 -7.33944893e-01 -8.08920026e-01 6.42471075e-01
-1.05899906e+00 -1.28797281e+00 -4.72780824e-01 5.76161921e-01
5.26816547e-01 -3.67415577e-01 9.80621338e-01 4.28586692e-01
-6.50138855e-01 1.00278366e+00 -3.47649425e-01 2.29439035e-01
2.72450030e-01 -8.89906168e-01 4.25435692e-01 4.88037527e-01
3.63123924e-01 4.08013076e-01 3.42272192e-01 -7.21841693e-01
-1.75192225e+00 -1.01819563e+00 6.54013097e-01 -2.56397337e-01
5.22854805e-01 -9.77304816e-01 -1.15103316e+00 6.95538640e-01
1.93916366e-01 1.85136527e-01 7.38659680e-01 2.85429537e-01
-7.59238899e-02 -1.24146827e-01 -9.82675254e-01 7.20507801e-01
1.18385243e+00 -9.35532868e-01 -1.08226679e-01 -3.58315185e-02
8.35747600e-01 -3.84730965e-01 -8.68355036e-01 3.13203692e-01
4.09927756e-01 -4.60832119e-01 6.37571931e-01 -5.37080884e-01
8.85411575e-02 2.86726028e-01 7.34317154e-02 -1.55121505e+00
1.81534186e-01 -8.28701138e-01 -5.02648830e-01 1.22911739e+00
5.97387195e-01 -5.27758300e-01 1.19391000e+00 9.99093175e-01
-3.27928722e-01 -9.48771775e-01 -7.16220081e-01 -7.45396316e-01
-2.91336954e-01 -8.50435197e-01 4.85756308e-01 7.66579270e-01
5.45592904e-01 7.75306940e-01 -4.34249759e-01 1.62412450e-01
1.71975121e-01 -5.81613541e-01 6.43361747e-01 -1.31640840e+00
-2.25425139e-01 -1.26093104e-01 -5.39592087e-01 -8.12969983e-01
4.33317751e-01 -8.72321069e-01 -9.62780789e-02 -1.48252451e+00
-2.16276005e-01 -4.64682788e-01 -4.09884185e-01 3.98697913e-01
-2.22983181e-01 -1.91967130e-01 -1.94548797e-02 -4.77870226e-01
-5.02906024e-01 8.34945321e-01 8.29876602e-01 -1.96359947e-01
-6.05222166e-01 2.36609094e-02 -4.69063908e-01 3.66788119e-01
7.36772478e-01 -1.11750752e-01 -8.48629653e-01 -3.17650400e-02
2.54286140e-01 1.95373312e-01 4.46552113e-02 -1.24219489e+00
5.44377148e-01 5.17366588e-01 4.55426574e-01 -4.20985855e-02
6.89233601e-01 -1.08465958e+00 1.97136268e-01 1.24509394e-01
-2.31965810e-01 -1.34429052e-01 6.46072447e-01 6.00981891e-01
-2.35955477e-01 9.33720395e-02 3.64808500e-01 3.27741027e-01
-8.12298596e-01 4.74725336e-01 -1.80100918e-01 -3.46933275e-01
9.77773070e-01 -4.75178659e-01 6.71448261e-02 -6.85505211e-01
-9.38406706e-01 1.86835844e-02 1.53583894e-02 7.96382010e-01
1.03645146e+00 -1.39741755e+00 -9.85377803e-02 4.97403741e-01
1.50550827e-01 -9.27749053e-02 2.80059934e-01 4.78124261e-01
-1.37794542e-03 2.91165914e-02 -6.72011450e-02 -4.25885916e-01
-1.49253559e+00 6.90622091e-01 2.24436700e-01 -8.52492973e-02
-6.55926049e-01 7.90242195e-01 9.13622081e-02 -3.54759037e-01
6.89122975e-01 -3.01039845e-01 -6.79324344e-02 -3.28081667e-01
6.50286973e-01 2.15452135e-01 3.28152120e-01 -2.45915011e-01
-2.36068323e-01 7.50279352e-02 -1.33106969e-02 -6.66499436e-02
1.56737673e+00 2.05347821e-01 4.18128967e-01 7.00407743e-01
1.55977607e+00 -1.63494721e-01 -1.10335720e+00 -2.10407570e-01
1.65897369e-01 4.00761068e-02 3.18782538e-01 -7.04350829e-01
-1.14340544e+00 1.07777584e+00 7.42191195e-01 1.17970876e-01
1.34611261e+00 -4.44024168e-02 1.00643110e+00 2.80384243e-01
5.83147764e-01 -1.31002712e+00 6.18739307e-01 1.94359660e-01
7.65210390e-01 -1.24540746e+00 -6.08620524e-01 -5.57664871e-01
-1.01415181e+00 1.05811787e+00 6.43820345e-01 -9.82260332e-03
9.44845498e-01 7.97244668e-01 5.87295415e-03 -2.39746213e-01
-6.73537791e-01 9.35844555e-02 3.16677958e-01 8.57958436e-01
3.33222508e-01 -7.04994649e-02 2.36041650e-01 1.14880705e+00
2.84994300e-02 3.67301941e-01 2.87131430e-03 9.24687982e-01
-1.08095400e-01 -1.23967612e+00 4.47710752e-02 6.19228423e-01
-8.32799450e-02 -1.03233188e-01 -7.30883837e-01 5.77939451e-01
-1.68568179e-01 1.04342365e+00 3.17331776e-02 -1.18013036e+00
3.84107411e-01 4.43066150e-01 3.07660729e-01 -6.37128532e-01
-6.89155579e-01 -2.35398799e-01 3.43213260e-01 -4.47523773e-01
-1.07210271e-01 -2.66792923e-01 -1.47729337e+00 5.11176437e-02
-4.98305112e-01 1.76290169e-01 8.90954077e-01 9.20169473e-01
6.72375798e-01 7.79589653e-01 7.25921273e-01 -8.23700130e-01
-9.72015038e-02 -9.90219295e-01 -4.94686633e-01 3.63920778e-01
8.17169845e-02 -4.76977050e-01 -3.94483119e-01 1.16986610e-01] | [13.85413932800293, 5.894458770751953] |
4420d3f4-2cef-4098-b087-4914f22e3cad | fast-point-cloud-generation-with-straight | 2212.01747 | null | https://arxiv.org/abs/2212.01747v1 | https://arxiv.org/pdf/2212.01747v1.pdf | Fast Point Cloud Generation with Straight Flows | Diffusion models have emerged as a powerful tool for point cloud generation. A key component that drives the impressive performance for generating high-quality samples from noise is iteratively denoise for thousands of steps. While beneficial, the complexity of learning steps has limited its applications to many 3D real-world. To address this limitation, we propose Point Straight Flow (PSF), a model that exhibits impressive performance using one step. Our idea is based on the reformulation of the standard diffusion model, which optimizes the curvy learning trajectory into a straight path. Further, we develop a distillation strategy to shorten the straight path into one step without a performance loss, enabling applications to 3D real-world with latency constraints. We perform evaluations on multiple 3D tasks and find that our PSF performs comparably to the standard diffusion model, outperforming other efficient 3D point cloud generation methods. On real-world applications such as point cloud completion and training-free text-guided generation in a low-latency setup, PSF performs favorably. | ['Qiang Liu', 'Vikas Chandra', 'Raghuraman Krishnamoorthi', 'Rakesh Ranjan', 'Yunyang Xiong', 'Xingchao Liu', 'Chengyue Gong', 'Dilin Wang', 'Lemeng Wu'] | 2022-12-04 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wu_Fast_Point_Cloud_Generation_With_Straight_Flows_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wu_Fast_Point_Cloud_Generation_With_Straight_Flows_CVPR_2023_paper.pdf | cvpr-2023-1 | ['point-cloud-completion', 'point-cloud-generation'] | ['computer-vision', 'computer-vision'] | [-7.86946267e-02 -8.16979259e-02 1.48498327e-01 1.08950034e-01
-1.15378511e+00 -7.51016200e-01 9.86312032e-01 2.74995938e-02
-1.29642099e-01 4.88989413e-01 1.09823138e-01 -6.37905955e-01
1.45398915e-01 -9.15726304e-01 -8.34039092e-01 -4.20687020e-01
-9.03121680e-02 7.21715212e-01 3.25369239e-01 -2.90037096e-01
4.53224212e-01 9.28156257e-01 -1.38889837e+00 3.18220146e-02
9.26020980e-01 5.92167616e-01 2.38232404e-01 7.96761394e-01
-4.96067315e-01 3.19480002e-01 -8.34639609e-01 -1.78432733e-01
5.87028086e-01 -1.92015141e-01 -7.06278443e-01 1.78755730e-01
6.51454985e-01 -4.32140052e-01 -2.02841684e-01 6.79423869e-01
7.21189618e-01 8.16807225e-02 7.15831101e-01 -1.27737510e+00
-6.05802655e-01 2.17479020e-01 -6.69955492e-01 -1.21380379e-02
3.78845066e-01 4.01993155e-01 6.65718973e-01 -1.17510808e+00
7.98174143e-01 1.38165903e+00 7.96438575e-01 4.96501565e-01
-1.31675863e+00 -5.38908422e-01 1.03655048e-01 -2.41785586e-01
-1.43968987e+00 -3.87971580e-01 6.65755808e-01 -4.03771937e-01
1.06146002e+00 1.76782757e-01 8.70145798e-01 1.04041982e+00
1.89669997e-01 9.76609766e-01 8.82047236e-01 -2.44405136e-01
5.33266306e-01 -1.62794515e-01 -4.65264171e-01 5.12531698e-01
6.30804598e-02 1.86012298e-01 -6.44485056e-01 -5.00927210e-01
1.06404126e+00 -6.73786104e-02 -1.75445303e-01 -4.22941148e-01
-1.36279762e+00 7.80561745e-01 6.99356139e-01 -2.45211154e-01
-3.67501259e-01 3.91489387e-01 1.34788528e-01 2.77103633e-01
9.71352041e-01 4.18082237e-01 -1.34844080e-01 -5.30808747e-01
-1.29937160e+00 7.94442832e-01 7.64892697e-01 1.44699228e+00
7.30101883e-01 1.71376705e-01 -1.85430378e-01 6.61907256e-01
3.17590207e-01 9.13244545e-01 1.07122779e-01 -1.26577485e+00
5.29960096e-01 4.18827325e-01 7.91656449e-02 -7.80071855e-01
-9.25791636e-02 -3.18493068e-01 -9.81081307e-01 6.04423583e-01
2.22560480e-01 -2.18144968e-01 -1.15127957e+00 1.22382343e+00
6.79835200e-01 4.82033283e-01 -1.26145616e-01 8.60685587e-01
4.50299472e-01 9.31863904e-01 -2.80396402e-01 1.38981536e-01
8.26464355e-01 -9.59865570e-01 -8.00768584e-02 -1.33411393e-01
6.22357249e-01 -1.08406687e+00 1.28758121e+00 5.47891617e-01
-1.17907226e+00 -4.36413020e-01 -7.90619612e-01 -2.49899253e-01
-5.50459558e-03 -3.39650005e-01 7.41757631e-01 4.58732396e-01
-1.36007047e+00 9.03894305e-01 -1.04676402e+00 -3.45795423e-01
6.31768584e-01 7.03963116e-02 -2.03412980e-01 -1.81521863e-01
-4.79316264e-01 4.80860770e-01 -1.08133599e-01 -1.69473067e-01
-8.48460495e-01 -1.25114143e+00 -5.44620752e-01 -1.98034227e-01
1.91987440e-01 -1.26954710e+00 1.49476171e+00 -1.32943898e-01
-1.74443519e+00 5.28825700e-01 -3.65342796e-01 -4.22745228e-01
8.58787477e-01 -4.91235614e-01 2.18612492e-01 6.97398931e-02
3.10804099e-01 9.97122169e-01 1.10039663e+00 -1.39262152e+00
-5.65176249e-01 -1.55827552e-01 -9.71115753e-02 3.34646434e-01
-7.96707496e-02 -4.46043521e-01 -7.36671329e-01 -7.34470487e-01
2.66387522e-01 -1.21832144e+00 -6.85479462e-01 2.97665477e-01
-4.32414174e-01 -3.49503100e-01 1.15104020e+00 -2.84774043e-02
6.98342443e-01 -2.01729751e+00 1.01246625e-01 2.59261876e-01
4.51509833e-01 1.27102911e-01 -1.66695982e-01 6.64449036e-01
1.61032125e-01 3.21653217e-01 -3.18063080e-01 -7.90575624e-01
7.40555599e-02 2.19848186e-01 -7.60969579e-01 2.93766677e-01
3.17456603e-01 9.77928877e-01 -1.05264604e+00 -2.77584523e-01
4.01114106e-01 5.57515681e-01 -8.17643523e-01 8.91254768e-02
-4.55699861e-01 3.44891489e-01 -3.77459407e-01 6.17434859e-01
9.11583245e-01 -4.01978821e-01 -6.02874994e-01 4.86335829e-02
-8.81270617e-02 1.90686479e-01 -1.10074747e+00 2.23381948e+00
-5.80571592e-01 5.84882140e-01 -1.86305568e-02 -1.83703423e-01
9.75171924e-01 4.36535440e-02 4.55662251e-01 -3.33769023e-01
-3.18673640e-01 2.04440966e-01 -5.40305316e-01 5.38560981e-03
9.98321831e-01 2.98973219e-03 9.74060968e-02 6.37494445e-01
-3.45710248e-01 -1.08771169e+00 -1.26624061e-02 7.15261161e-01
1.31733882e+00 3.44379365e-01 -3.06421995e-01 -4.10678834e-02
-1.14334874e-01 4.18512225e-01 1.04004942e-01 8.91586006e-01
1.65945634e-01 1.18359995e+00 3.76209766e-01 -2.39335194e-01
-1.29824209e+00 -1.44936752e+00 2.15580687e-01 3.78352463e-01
1.99358374e-01 -7.98622251e-01 -5.69601953e-01 -6.71760440e-01
2.11463124e-01 6.89607918e-01 -1.94235355e-01 1.15749985e-01
-5.52079797e-01 -6.40473485e-01 4.88205910e-01 3.80954951e-01
2.92917609e-01 -7.49234438e-01 -2.81439483e-01 2.53107131e-01
1.15230836e-01 -1.06615114e+00 -5.06051123e-01 -3.13644528e-01
-1.28579414e+00 -6.67488813e-01 -1.00920844e+00 -3.37521762e-01
6.56253695e-01 8.04692507e-01 1.37344801e+00 1.27599344e-01
-6.27650023e-02 5.36496282e-01 -3.58497113e-01 -4.80891705e-01
-5.29970884e-01 3.90479028e-01 -4.22863802e-03 -4.83754456e-01
2.54509524e-02 -7.69417703e-01 -9.10333574e-01 1.60424531e-01
-1.00773978e+00 2.02963904e-01 6.61331296e-01 4.28425938e-01
9.11678612e-01 -1.48222610e-01 3.73403013e-01 -7.30121195e-01
9.04036880e-01 -6.17852151e-01 -5.76648593e-01 -4.42076832e-01
-7.23623693e-01 3.45533602e-02 4.92744446e-01 -3.20813179e-01
-8.41751695e-01 7.37962313e-03 -4.83205825e-01 -8.22439611e-01
-1.33919984e-01 1.68232664e-01 2.84931630e-01 -1.35177508e-01
8.92421305e-01 2.97325939e-01 5.42133786e-02 -5.31408429e-01
8.27576578e-01 3.70876670e-01 4.96209949e-01 -8.00004900e-01
1.27378058e+00 7.70743489e-01 2.03165844e-01 -1.06948137e+00
-3.87795508e-01 -4.02340740e-01 -5.03312349e-01 -9.71815586e-02
4.77783889e-01 -9.53217626e-01 -4.73041594e-01 3.60482067e-01
-1.44268739e+00 -5.62582076e-01 -5.57839870e-01 4.97151427e-02
-7.58755147e-01 4.68730628e-01 -6.22541130e-01 -5.98128736e-01
-6.46362185e-01 -1.07261682e+00 1.72348130e+00 -5.28338887e-02
-3.05131614e-01 -7.96615303e-01 2.21393436e-01 -7.81701505e-02
5.89376092e-01 2.93505132e-01 6.42519355e-01 -2.92994417e-02
-1.23166800e+00 -2.51408041e-01 -3.07453543e-01 1.03085399e-01
-3.06274854e-02 2.25867167e-01 -7.63096392e-01 -3.41588825e-01
-2.31553495e-01 -8.39243457e-02 6.82889819e-01 2.59786487e-01
1.12323558e+00 5.06079756e-04 -5.42724907e-01 9.59578276e-01
1.26662838e+00 -2.64965892e-01 5.67270219e-01 1.91234216e-01
8.49361002e-01 4.89590019e-02 4.50780839e-01 4.09272850e-01
3.41044933e-01 6.02773726e-01 4.36693341e-01 -3.37901920e-01
-4.62258786e-01 -6.51780009e-01 2.54783630e-01 9.60785508e-01
-4.13824953e-02 -2.84351379e-01 -1.04145586e+00 5.65082312e-01
-1.75737298e+00 -8.29059601e-01 -4.10717368e-01 2.16182351e+00
5.79638660e-01 2.73598015e-01 4.32686247e-02 -4.01030779e-02
7.00443462e-02 2.39562139e-01 -5.09795725e-01 -1.40849441e-01
1.05227619e-01 4.36857074e-01 5.11775017e-01 5.75348973e-01
-7.18016505e-01 1.16640723e+00 7.05679893e+00 1.06896520e+00
-1.23302269e+00 5.31758629e-02 3.38921219e-01 -4.71292496e-01
-5.59255064e-01 9.31257103e-03 -8.76291573e-01 3.60180855e-01
7.87488222e-01 -4.24067438e-01 3.24920297e-01 1.02490115e+00
6.27741873e-01 -2.66064517e-02 -1.03211415e+00 1.16462958e+00
-1.77425787e-01 -1.91476488e+00 4.00774300e-01 2.62674034e-01
9.61428404e-01 4.95863169e-01 1.18150249e-01 2.03184053e-01
6.60400093e-01 -8.18539262e-01 7.68209100e-01 4.05189931e-01
8.45598757e-01 -7.54033387e-01 1.64579883e-01 6.12891614e-01
-9.76057053e-01 5.56362152e-01 -5.39886713e-01 1.09180704e-01
6.99240446e-01 1.01114142e+00 -1.20825326e+00 6.02053821e-01
6.96087241e-01 7.23557472e-01 -4.63006347e-01 1.30179238e+00
-1.83059037e-01 6.01325572e-01 -6.96920514e-01 1.00900486e-01
3.78077507e-01 -3.23690563e-01 8.71582091e-01 1.16948819e+00
8.66088271e-01 -2.10721418e-01 1.24419913e-01 1.11085010e+00
-4.65260819e-02 -1.33261070e-01 -1.07722604e+00 2.36600012e-01
5.81181586e-01 1.16910756e+00 -7.72317350e-01 -3.45777333e-01
-9.89621207e-02 1.37076402e+00 2.80105174e-01 3.83244216e-01
-6.40014052e-01 -4.41852778e-01 8.72681022e-01 4.34679568e-01
2.71884233e-01 -8.08444083e-01 -5.46184123e-01 -8.97627175e-01
5.83828948e-02 -6.51071191e-01 -3.29967648e-01 -9.41756904e-01
-1.41656315e+00 8.18027794e-01 6.16497137e-02 -1.48906195e+00
-3.14779490e-01 -2.37421364e-01 -6.64934874e-01 1.12952197e+00
-1.54941428e+00 -1.16385818e+00 -4.80270386e-01 5.71032584e-01
8.43750715e-01 3.78953032e-02 6.44577563e-01 1.30257398e-01
-8.16568807e-02 1.93744510e-01 7.10582659e-02 -4.65096444e-01
5.99151492e-01 -1.49347746e+00 1.53210700e+00 8.63813818e-01
2.48024464e-01 6.04626238e-01 5.98737061e-01 -9.07191217e-01
-1.50534415e+00 -1.11585498e+00 5.80674529e-01 -8.26761484e-01
5.17896116e-01 -4.66781199e-01 -8.50174546e-01 3.63230020e-01
1.11954048e-01 -1.59515798e-01 1.82880864e-01 -4.39085811e-02
-1.14751890e-01 5.27362302e-02 -9.37546313e-01 9.19551611e-01
1.39265883e+00 -3.18606824e-01 -2.88663208e-01 7.22838163e-01
1.12412810e+00 -9.20702100e-01 -7.04662085e-01 1.12729460e-01
9.16485041e-02 -9.66562629e-01 1.17307687e+00 -1.61771521e-01
7.19112098e-01 -5.18933117e-01 1.56441137e-01 -1.70721173e+00
-3.09233934e-01 -1.28582132e+00 -2.87849903e-01 8.84730458e-01
3.54906440e-01 -3.59525532e-01 1.32671356e+00 3.28720391e-01
-4.29776669e-01 -7.05444038e-01 -8.00428510e-01 -8.77447724e-01
2.63417304e-01 -7.51106024e-01 8.61552238e-01 5.39525688e-01
-4.73193496e-01 3.54098022e-01 -1.00179181e-01 2.09124640e-01
9.13702369e-01 2.64323130e-02 1.52770877e+00 -9.87352729e-01
-1.91006497e-01 -4.33892190e-01 -1.35219693e-01 -1.83703184e+00
-2.63258308e-01 -9.37991560e-01 -1.81341842e-02 -1.83446801e+00
-4.50958759e-01 -9.35033560e-01 4.99792397e-01 6.29545748e-02
-2.50084937e-01 2.15501487e-01 3.62145960e-01 4.44144756e-01
-1.77378565e-01 7.97207117e-01 1.36662376e+00 -3.25864926e-02
-5.17727017e-01 2.58874983e-01 -5.67309797e-01 6.40967309e-01
6.43845856e-01 -5.04607499e-01 -7.91662157e-01 -8.71074438e-01
3.39524329e-01 -5.72092272e-02 2.59482503e-01 -1.21944761e+00
4.31201190e-01 -2.02395573e-01 2.26859972e-01 -1.07635033e+00
5.79608023e-01 -6.46041453e-01 1.86104611e-01 1.33321390e-01
1.06911354e-01 4.15507317e-01 2.98058361e-01 7.17177689e-01
9.37845260e-02 2.29347628e-02 4.64785725e-01 -5.06053008e-02
-5.40830016e-01 7.43790686e-01 -1.47593692e-01 1.31080560e-02
1.01288164e+00 -3.61183375e-01 -2.84568250e-01 -4.12086427e-01
-4.47071046e-01 1.59355462e-01 7.37596869e-01 5.19496977e-01
9.66362119e-01 -1.39658999e+00 -9.08274651e-01 3.82183522e-01
-6.87323604e-03 7.35965490e-01 1.14656322e-01 3.92409652e-01
-1.09201860e+00 1.51007518e-01 3.92557621e-01 -1.09483159e+00
-9.98017848e-01 4.17412400e-01 8.86349604e-02 -4.33122553e-02
-1.13675618e+00 8.50012541e-01 -5.63361868e-02 -4.78218138e-01
-1.04283746e-02 -4.38900888e-01 6.38307035e-01 -4.14684981e-01
4.81387883e-01 3.75345379e-01 3.22806567e-01 -1.46830976e-01
1.08730756e-01 5.55551529e-01 -1.72022641e-01 -4.35864806e-01
1.35072327e+00 -1.97688621e-02 8.12639445e-02 1.96262017e-01
9.77394223e-01 2.66614974e-01 -1.44280326e+00 -1.68136597e-01
-2.60567158e-01 -8.27339232e-01 1.63552508e-01 -3.88854295e-01
-9.06097710e-01 7.83944190e-01 2.99605876e-01 3.56161326e-01
6.93035960e-01 2.78066974e-02 1.21470857e+00 2.00787768e-01
6.28331840e-01 -6.48508310e-01 9.25231352e-02 5.25135934e-01
1.00513327e+00 -9.76666868e-01 1.14947587e-01 -5.16317070e-01
-4.12970692e-01 8.87285233e-01 3.93435806e-01 -3.35054874e-01
6.56032443e-01 4.45975304e-01 1.84697732e-01 -2.72835851e-01
-9.61763978e-01 1.08435735e-01 5.72828427e-02 9.24226820e-01
8.66139978e-02 -1.23193465e-01 1.31506115e-01 -2.24881515e-01
-6.42450452e-01 1.29581660e-01 5.16070306e-01 8.75906289e-01
-4.50145453e-01 -1.28058124e+00 -5.70809901e-01 3.28462154e-01
-2.30312720e-03 -1.24785915e-01 -2.62950718e-01 6.99560225e-01
-2.99774945e-01 8.03678870e-01 2.38160715e-01 -2.75411755e-01
5.81373155e-01 -1.95129067e-01 3.33358467e-01 -6.73584104e-01
-4.41063255e-01 9.36072618e-02 -3.00098240e-01 -8.11301529e-01
-6.34140745e-02 -4.97370839e-01 -1.13297093e+00 -9.40027952e-01
-5.99919185e-02 1.36269361e-01 8.35584760e-01 5.51885903e-01
1.02323222e+00 2.35150039e-01 6.32143140e-01 -1.43960512e+00
-5.73394895e-01 -6.56231225e-01 -2.38351092e-01 3.53447258e-01
3.92586708e-01 -3.55304778e-01 -3.53703558e-01 3.98396049e-03] | [8.943395614624023, -3.623363971710205] |
eb01bc4f-7c3f-49b6-86a7-24af9c046c51 | self-paced-deep-regression-forests-with-1 | 2112.06455 | null | https://arxiv.org/abs/2112.06455v8 | https://arxiv.org/pdf/2112.06455v8.pdf | Self-Paced Deep Regression Forests with Consideration of Ranking Fairness | Deep discriminative models (DDMs), e.g. deep regression forests and deep decision forests, have been extensively studied recently to solve problems such as facial age estimation, head pose estimation, etc.. Due to a shortage of well-labeled data that does not have noise and imbalanced distribution problems, learning DDMs is always challenging. Existing methods usually tackle these challenges through learning more discriminative features or re-weighting samples. We argue that learning DDMs gradually, from easy to hard, is more reasonable, for two reasons. First, this is more consistent with the cognitive process of human beings. Second, noisy as well as underrepresented examples can be distinguished by virtue of previously learned knowledge. Thus, we resort to a gradual learning strategy -- self-paced learning (SPL). Then, a natural question arises: can SPL lead DDMs to achieve more robust and less biased solutions? To answer this question, this paper proposes a new SPL method: easy and underrepresented examples first, for learning DDMs. This tackles the fundamental ranking and selection problem in SPL from a new perspective: fairness. Our idea is fundamental and can be easily combined with a variety of DDMs. Extensive experimental results on three computer vision tasks, i.e., facial age estimation, head pose estimation, and gaze estimation, show our new method gains considerable performance improvement in both accuracy and fairness. Source code is available at https://github.com/learninginvision/SPU. | ['Zenglin Xu', 'Yali Zheng', 'Yazhou Ren', 'Mingming Meng', 'Lili Pan'] | 2021-12-13 | null | null | null | null | ['head-pose-estimation', 'gaze-estimation', 'age-estimation', 'age-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision', 'miscellaneous'] | [-1.09576195e-01 3.78324315e-02 -4.45062369e-01 -7.16211379e-01
-4.21828985e-01 1.32177277e-02 4.10185069e-01 -1.34320006e-01
-4.60999042e-01 9.95646954e-01 1.17022410e-01 5.57116531e-02
-1.08495377e-01 -5.90448439e-01 -3.30093354e-01 -9.13810492e-01
3.18872482e-01 4.19452369e-01 -6.13216236e-02 -9.99014452e-02
2.79632807e-01 3.10358286e-01 -1.85426140e+00 -3.44043732e-01
1.39079809e+00 1.05723774e+00 5.72867431e-02 -9.86346975e-02
-1.55696824e-01 5.12859941e-01 -2.07159892e-01 -7.79521585e-01
9.09597725e-02 -3.02219212e-01 -7.75613606e-01 -6.80636391e-02
4.93612468e-01 -3.71737272e-01 7.91679919e-02 1.15806127e+00
7.98394680e-01 1.00460395e-01 7.71223247e-01 -1.69854450e+00
-3.84997934e-01 3.87095153e-01 -1.08256793e+00 -7.28382096e-02
2.51783878e-01 5.02125546e-02 9.13361728e-01 -9.54955876e-01
1.76001832e-01 1.39689028e+00 5.78999341e-01 1.05166817e+00
-1.08122575e+00 -1.12929058e+00 3.49474102e-01 5.34900308e-01
-1.44778311e+00 -6.65837526e-01 7.74870336e-01 -3.74916464e-01
-6.85794428e-02 3.06631625e-01 6.16623104e-01 1.27348757e+00
-1.18664809e-01 1.14603877e+00 1.70257628e+00 -3.27109128e-01
1.08142667e-01 3.40935588e-01 2.15502590e-01 7.27171779e-01
4.07730013e-01 9.92335305e-02 -7.39662826e-01 -3.96548994e-02
2.80556202e-01 7.33633190e-02 -3.04894000e-01 -4.42012131e-01
-6.07975543e-01 7.96294332e-01 3.56094867e-01 9.30197313e-02
-1.91238120e-01 -2.40385801e-01 2.92867243e-01 1.46629125e-01
6.24192595e-01 8.30809623e-02 -4.36153591e-01 3.95210311e-02
-9.53792870e-01 3.67033422e-01 6.63896918e-01 6.29395843e-01
8.74707937e-01 -4.27056067e-02 -1.94324076e-01 1.21649098e+00
5.30857086e-01 5.73767304e-01 5.88385165e-01 -7.42435277e-01
1.20391279e-01 3.26244473e-01 -1.42029896e-01 -9.28605914e-01
-4.83369023e-01 -4.92357314e-01 -1.14486170e+00 3.37269396e-01
6.05647445e-01 -1.81318685e-01 -7.99674988e-01 2.13908815e+00
6.43034399e-01 9.01882574e-02 -3.50994915e-01 8.91322553e-01
9.15233970e-01 3.05485904e-01 2.71271855e-01 -6.82818294e-01
1.33794093e+00 -7.17555583e-01 -7.85384357e-01 -2.94144422e-01
2.22008079e-01 -6.07809067e-01 1.16319513e+00 6.29807591e-01
-7.91463673e-01 -4.07274693e-01 -7.02630997e-01 2.48508919e-02
-1.49263412e-01 1.00636683e-01 7.14870870e-01 7.61823773e-01
-9.02757108e-01 5.51628768e-01 -3.99550110e-01 -3.78154606e-01
7.62713730e-01 4.43859845e-01 -2.60316849e-01 -1.87398627e-01
-1.26252890e+00 8.88722420e-01 2.06283221e-05 1.16334066e-01
-5.75888753e-01 -3.55266243e-01 -7.54258454e-01 -2.49535084e-01
4.80492175e-01 -8.49802077e-01 1.45039773e+00 -1.23055756e+00
-1.56289566e+00 1.08831894e+00 -3.81902903e-01 -1.43642515e-01
7.25873768e-01 -4.23098922e-01 -9.42443907e-02 -3.09470743e-01
1.12002909e-01 6.30177557e-01 1.17454529e+00 -1.23846269e+00
-6.01542234e-01 -9.18768764e-01 -1.33223400e-01 3.48053038e-01
-5.83757222e-01 5.17394990e-02 -1.49271265e-01 -5.82004488e-01
9.79393721e-02 -9.13316190e-01 -9.91521552e-02 9.78781283e-02
-4.69039053e-01 -7.24028051e-01 4.15115386e-01 -5.64825654e-01
1.43670142e+00 -1.96469712e+00 8.48986059e-02 1.16083995e-01
5.37491560e-01 3.37146699e-01 7.73474947e-02 -2.13292509e-01
-1.92699973e-02 2.94817076e-03 -1.04155354e-02 -5.34170747e-01
7.99303949e-02 7.10755140e-02 2.91418135e-02 3.81047398e-01
-9.41378698e-02 6.42159820e-01 -8.06020379e-01 -8.06413531e-01
-2.46053204e-01 2.26746455e-01 -3.61049324e-01 2.81785727e-01
1.27040714e-01 5.70640385e-01 -3.91203791e-01 8.88219476e-01
1.02289844e+00 4.99633364e-02 -9.10664424e-02 -1.92514777e-01
2.20700093e-02 6.66379333e-02 -1.12753308e+00 1.27510703e+00
-4.50382948e-01 2.52717793e-01 -4.32229936e-02 -1.01421762e+00
9.57874477e-01 4.36968096e-02 2.38846272e-01 -6.91332936e-01
2.51907468e-01 3.64821225e-01 -3.73495445e-02 -6.11230433e-01
2.51584202e-01 -4.03412282e-01 5.92265688e-02 2.71887869e-01
4.49355729e-02 1.07848816e-01 1.76672563e-02 -2.43101969e-01
4.65896368e-01 1.07182749e-01 6.04309678e-01 -1.96731344e-01
6.05886042e-01 -5.86948633e-01 1.10116482e+00 4.54021484e-01
-6.53629124e-01 3.42192054e-01 5.21516025e-01 -4.16089773e-01
-4.53519970e-01 -9.24813151e-01 -2.37520680e-01 1.49452257e+00
1.75614238e-01 -2.39987090e-01 -9.68117476e-01 -8.72096777e-01
4.11611870e-02 5.34235001e-01 -7.19012856e-01 -2.27889001e-01
-3.53077441e-01 -9.22711551e-01 2.34768853e-01 3.96581262e-01
5.97727001e-01 -1.06079197e+00 -3.22163433e-01 -2.31493101e-01
-1.82539180e-01 -6.50962174e-01 -2.62722671e-01 -3.17402966e-02
-9.44174707e-01 -9.17452157e-01 -1.04348481e+00 -7.25184739e-01
7.30159223e-01 3.14760298e-01 1.20364892e+00 1.81343928e-01
-3.45997773e-02 3.43607850e-02 -2.08204940e-01 -7.57201016e-01
7.44902343e-02 1.98909089e-01 5.20815909e-01 2.12503925e-01
7.49719501e-01 -7.91900277e-01 -7.78488219e-01 3.95119667e-01
-5.14756262e-01 7.50105903e-02 8.70338380e-01 1.02198017e+00
4.87941742e-01 -1.12326719e-01 8.05032134e-01 -1.03889751e+00
6.25677943e-01 -4.45623815e-01 -3.29537481e-01 2.58885413e-01
-9.59851682e-01 9.76181254e-02 2.82796413e-01 -4.15256172e-01
-1.16147900e+00 4.60164659e-02 -3.00945759e-01 -2.67710119e-01
-3.51486415e-01 3.20559710e-01 -5.10665417e-01 -1.65558308e-02
5.77481031e-01 1.46211863e-01 3.16675216e-01 -4.13601369e-01
1.02562293e-01 9.22097206e-01 9.70260203e-02 -6.65064037e-01
9.05710936e-01 1.28940225e-01 -8.14942941e-02 -5.48592865e-01
-1.06658387e+00 -1.82883114e-01 -4.22971606e-01 -4.27150398e-01
5.06942630e-01 -8.32775533e-01 -8.40824366e-01 8.56970429e-01
-8.12505007e-01 -1.77541897e-01 -3.46105732e-02 3.78451645e-01
-3.82463753e-01 2.81951249e-01 -2.17359036e-01 -1.03969216e+00
-5.47507286e-01 -9.51287508e-01 9.07837212e-01 7.21675277e-01
-3.08876157e-01 -6.13876879e-01 -7.48192668e-02 4.29344356e-01
3.76080096e-01 5.85476197e-02 7.38749266e-01 -5.49491704e-01
-3.61571789e-01 -1.72645804e-02 -2.24410713e-01 3.40321064e-01
5.48539869e-02 -2.73729749e-02 -1.22864318e+00 -2.66590118e-01
-7.84973055e-02 -6.69949174e-01 1.05480289e+00 4.71329302e-01
1.55176210e+00 -3.86279583e-01 -2.94116825e-01 5.97931027e-01
1.06591821e+00 -1.42220140e-01 5.83006382e-01 2.93963701e-01
6.23874009e-01 8.89021754e-01 8.14762354e-01 6.50702596e-01
7.59383917e-01 6.20911360e-01 3.98873866e-01 3.22018713e-02
9.67463478e-02 -2.60133475e-01 2.97872812e-01 6.99786544e-01
-4.48605984e-01 1.03428073e-01 -8.28128636e-01 1.83190376e-01
-1.83017123e+00 -8.30643475e-01 1.75460875e-01 2.39934039e+00
1.05594683e+00 6.23513199e-02 5.40577531e-01 3.41373593e-01
7.77436137e-01 1.16243668e-01 -7.13241935e-01 -2.15954199e-01
1.02931306e-01 9.83115509e-02 1.04962736e-01 1.50116116e-01
-1.08653915e+00 7.71634698e-01 4.82290745e+00 1.22534835e+00
-1.28224099e+00 2.22483769e-01 1.05969560e+00 3.04262526e-02
-3.10081065e-01 -3.64485919e-01 -1.01610625e+00 6.98081493e-01
4.57522005e-01 -1.74743459e-01 2.84984231e-01 9.52004313e-01
4.44619395e-02 -2.11548880e-01 -9.30978119e-01 1.40802240e+00
1.22373804e-01 -6.47175014e-01 -1.81620896e-01 6.12764852e-04
4.37209934e-01 -3.72729331e-01 3.81957352e-01 5.78430414e-01
1.29477799e-01 -9.89100516e-01 6.59211993e-01 6.77806497e-01
8.44303727e-01 -7.66224384e-01 7.66565323e-01 5.25414824e-01
-8.73053432e-01 -1.90586701e-01 -2.98518211e-01 -4.03565429e-02
-2.56453544e-01 8.00798059e-01 -3.89031529e-01 3.79344672e-01
9.01462674e-01 7.02776670e-01 -6.24167383e-01 1.29972148e+00
-5.56786716e-01 4.94680762e-01 -1.25243440e-01 -2.91662633e-01
-3.64320368e-01 -7.96121359e-03 3.27802211e-01 7.00995624e-01
2.97870100e-01 1.19727269e-01 -3.34926248e-02 5.42166948e-01
-1.00752868e-01 4.27405030e-01 -5.29451072e-01 4.39104646e-01
5.43548167e-01 1.44960487e+00 -4.02507782e-01 1.76701993e-02
-1.58285692e-01 7.74059832e-01 5.46765566e-01 2.32500181e-01
-7.15672731e-01 -2.01747581e-01 5.96550524e-01 2.64355570e-01
-2.35394701e-01 1.47172198e-01 -3.40983540e-01 -1.27627957e+00
-5.89970453e-03 -1.08701038e+00 4.58103597e-01 -4.70091879e-01
-1.61477554e+00 6.70004547e-01 -4.37520109e-02 -1.32382417e+00
-1.53664440e-01 -4.18610424e-01 -6.01859391e-01 7.14821041e-01
-1.81757009e+00 -1.09995711e+00 -6.77540541e-01 5.83403230e-01
4.64853466e-01 -1.86078623e-01 6.42887115e-01 4.23916668e-01
-9.13909197e-01 1.11457860e+00 -1.07683487e-01 -8.80720690e-02
1.02975488e+00 -1.26554382e+00 -2.46737316e-01 4.24229980e-01
-8.59913602e-02 5.16154468e-01 5.48798025e-01 -2.57827103e-01
-8.34065437e-01 -8.39888930e-01 9.90162849e-01 -1.57351375e-01
1.85848325e-01 -3.34520102e-01 -9.16654766e-01 1.46203101e-01
5.69886863e-02 -4.32960242e-02 9.14302230e-01 4.67872441e-01
-3.40876400e-01 -6.32882416e-01 -1.17197084e+00 4.97851074e-01
1.18417370e+00 -2.16631517e-01 -3.49196136e-01 2.46055827e-01
1.81903526e-01 -3.50557446e-01 -4.71495032e-01 7.44283319e-01
8.17379296e-01 -1.23299503e+00 7.24970579e-01 -3.70776385e-01
4.11183059e-01 -8.97763371e-02 1.30664706e-01 -1.42539263e+00
-2.33786911e-01 -5.79872787e-01 -3.00051689e-01 1.44585049e+00
1.26389176e-01 -8.18239927e-01 8.77890289e-01 6.73622787e-01
3.23692560e-01 -1.19327748e+00 -8.71150076e-01 -6.36046767e-01
8.08248073e-02 -1.47957534e-01 7.36489356e-01 9.68374610e-01
-3.30113351e-01 5.40551543e-01 -6.23278737e-01 -1.17278300e-01
8.06941867e-01 2.38147467e-01 8.12291682e-01 -1.72316873e+00
1.12257048e-01 -6.58220232e-01 -2.88896412e-01 -9.78599906e-01
4.01008606e-01 -6.03542984e-01 1.36694193e-01 -1.16084218e+00
4.66236264e-01 -6.81358874e-01 -3.62315327e-01 6.06489837e-01
-5.45875847e-01 1.39527753e-01 7.80956522e-02 2.75698483e-01
-6.78043187e-01 8.03012550e-01 1.20839334e+00 4.06461582e-02
-1.93604659e-02 5.61768413e-01 -1.04099584e+00 1.00040448e+00
9.98764813e-01 -3.89184326e-01 -3.14652056e-01 -1.32066712e-01
1.40948802e-01 -2.11728483e-01 1.94964364e-01 -8.89494717e-01
1.10506140e-01 -3.16879869e-01 5.14434636e-01 -3.42264324e-01
1.69130921e-01 -5.37121058e-01 -3.30848008e-01 4.04287517e-01
-1.95228726e-01 -2.09841356e-01 -1.65212393e-01 4.01846409e-01
-1.80978715e-01 -2.52211511e-01 9.92412925e-01 -4.58406694e-02
-8.28846276e-01 6.28309608e-01 2.88946293e-02 1.76582217e-01
8.63655210e-01 -2.61405826e-01 -1.65783942e-01 -6.45728886e-01
-8.12857926e-01 4.10255045e-01 2.92710394e-01 3.72575700e-01
5.92690349e-01 -1.40167809e+00 -7.66105592e-01 1.39150113e-01
1.41642749e-01 1.12874825e-02 2.44933680e-01 1.14272738e+00
1.14890561e-01 -6.11696392e-02 -2.82925189e-01 -5.03681123e-01
-1.47843742e+00 2.27177590e-01 2.61132777e-01 -1.44915640e-01
1.27439737e-01 1.13775802e+00 2.73518384e-01 -6.13338590e-01
5.29473186e-01 4.61966880e-02 -6.12809718e-01 5.38388789e-01
5.61377287e-01 4.66540158e-01 -1.78087667e-01 -6.38368607e-01
-4.17649418e-01 7.16847003e-01 -1.79814160e-01 4.07400161e-01
1.25886941e+00 -3.35291743e-01 -2.51366377e-01 3.12217981e-01
9.88116264e-01 -8.67955312e-02 -1.05077374e+00 -4.65783864e-01
9.93165374e-02 -6.56762600e-01 -5.77489883e-02 -6.64681315e-01
-1.28619826e+00 1.03065217e+00 9.00817990e-01 4.19939943e-02
1.42227757e+00 -3.48677523e-02 7.70224690e-01 2.13868350e-01
4.97142136e-01 -1.13602340e+00 2.32418254e-01 3.68820637e-01
8.27142298e-01 -1.74130476e+00 1.44473627e-01 -2.90321618e-01
-6.92271233e-01 9.82237637e-01 1.14326501e+00 2.84889668e-01
7.18081355e-01 -1.42088741e-01 5.83941005e-02 7.41034970e-02
-5.28963685e-01 -4.25436080e-01 4.11582679e-01 4.75474507e-01
4.63278979e-01 1.84139267e-01 -5.69970787e-01 8.33359420e-01
-2.46972218e-01 2.23017976e-01 9.70872417e-02 6.35044217e-01
-5.51222622e-01 -1.17736316e+00 -3.60455811e-01 5.77700675e-01
-3.65326852e-01 4.08926308e-02 -2.88676351e-01 4.76195723e-01
5.06425500e-01 6.92260802e-01 -2.33673215e-01 -5.24643600e-01
1.55274406e-01 3.00842412e-02 4.89084274e-01 -4.80311722e-01
-5.36635742e-02 -9.38070193e-02 -8.27462077e-02 -4.63088125e-01
-6.79634511e-01 -7.33589172e-01 -9.12133873e-01 -3.65251392e-01
-5.43078661e-01 1.42980263e-01 4.14713919e-01 8.76107991e-01
8.79599154e-02 1.08301505e-01 9.84532535e-01 -8.16352308e-01
-8.76228750e-01 -9.38432932e-01 -6.72365367e-01 2.78990656e-01
2.78010190e-01 -1.04515135e+00 -4.13716167e-01 -3.72296065e-01] | [13.575993537902832, 0.9156275987625122] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.