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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
d3e8e689-f462-4002-8339-9a3c97f736d6 | deep-representation-learning-of-tissue | 2305.1559 | null | https://arxiv.org/abs/2305.15590v2 | https://arxiv.org/pdf/2305.15590v2.pdf | Deep Representation Learning of Tissue Metabolome and Computed Tomography Images Annotates Non-invasive Classification and Prognosis Prediction of NSCLC | The rich chemical information from tissue metabolomics provides a powerful means to elaborate tissue physiology or tumor characteristics at cellular and tumor microenvironment levels. However, the process of obtaining such information requires invasive biopsies, is costly, and can delay clinical patient management. Conversely, computed tomography (CT) is a clinical standard of care but does not intuitively harbor histological or prognostic information. Furthermore, the ability to embed metabolome information into CT to subsequently use the learned representation for classification or prognosis has yet to be described. This study develops a deep learning-based framework -- tissue-metabolomic-radiomic-CT (TMR-CT) by combining 48 paired CT images and tumor/normal tissue metabolite intensities to generate ten image embeddings to infer metabolite-derived representation from CT alone. In clinical NSCLC settings, we ascertain whether TMR-CT achieves state-of-the-art results in solving histology classification/prognosis tasks in an unseen international CT dataset of 742 patients. TMR-CT non-invasively determines histological classes - adenocarcinoma/ squamous cell carcinoma with an F1-score=0.78 and further asserts patients' prognosis with a c-index=0.72, surpassing the performance of radiomics models and clinical features. Additionally, our work shows the potential to generate informative biology-inspired CT-led features to explore connections between hard-to-obtain tissue metabolic profiles and routine lesion-derived image data. | ['Eric O Aboagye', 'Marco A Calzado', 'Joram M. Posma', 'Richard Lee', 'Ángel Salvatierra', 'Marina Álvarez-Benito', 'Paula Moreno', 'OCTAPUS-AI', 'Mitchell Chen', 'Sumeet Hindocha', 'Kristofer Linton-Reid', 'Marc Boubnovski Martell'] | 2023-05-24 | null | null | null | null | ['computed-tomography-ct'] | ['methodology'] | [ 4.26694453e-01 -9.96633098e-02 -4.99083310e-01 -1.38418168e-01
-1.12051582e+00 -5.21403372e-01 4.84232306e-01 5.87728381e-01
-5.04890501e-01 7.51503289e-01 3.35390866e-01 -5.98718524e-01
-1.67499110e-01 -6.08189106e-01 -1.98507592e-01 -1.30229461e+00
-3.52260023e-01 7.83712029e-01 -5.32761455e-01 3.67777765e-01
-1.99709058e-01 7.84286916e-01 -7.15698957e-01 4.95396793e-01
6.26945257e-01 1.05444551e+00 3.55184942e-01 9.62754667e-01
-5.85515797e-02 5.36833763e-01 -1.16152568e-02 -5.21756001e-02
1.21843353e-01 -5.60942233e-01 -6.88590586e-01 -1.16835497e-01
8.67350250e-02 -1.33374795e-01 -3.36614847e-01 8.66691768e-01
6.28765821e-01 -4.02403980e-01 1.07568848e+00 -8.61806512e-01
-7.00248420e-01 3.43839288e-01 -4.22225177e-01 2.26836368e-01
-2.25665402e-02 5.01384854e-01 7.49109268e-01 -6.26528323e-01
6.49592876e-01 7.09259093e-01 6.52805209e-01 6.59687579e-01
-1.35392535e+00 -3.25661957e-01 -3.21520537e-01 8.94172713e-02
-1.12871110e+00 -1.10957570e-01 2.74824470e-01 -5.23849249e-01
8.47609937e-01 5.89308679e-01 1.10169065e+00 1.33414459e+00
7.05019534e-01 3.84877145e-01 1.34757578e+00 -7.75193870e-02
2.23714322e-01 6.41216412e-02 -2.55023926e-01 1.05603468e+00
3.44908535e-01 2.64568001e-01 -3.80901456e-01 -2.16762777e-02
5.31698227e-01 4.25121874e-01 -6.97299719e-01 -1.07188895e-01
-1.48718977e+00 6.52989566e-01 8.01476419e-01 4.89816338e-01
-4.03024197e-01 3.97249103e-01 5.29052854e-01 1.15037061e-01
3.69596541e-01 4.25075501e-01 -5.14563620e-01 1.79820880e-01
-7.85582125e-01 -3.38742286e-01 5.00129104e-01 3.10977161e-01
1.85087085e-01 -2.54313827e-01 -2.85638999e-02 5.94087064e-01
3.20460945e-01 3.26632321e-01 1.13943493e+00 -5.49784958e-01
-1.52921319e-01 4.68750417e-01 -2.53147364e-01 -3.44225168e-01
-8.59132528e-01 -7.29467690e-01 -1.09043384e+00 -4.18791659e-02
4.33534950e-01 8.95080194e-02 -9.84940231e-01 1.47896397e+00
2.09133312e-01 1.78442165e-01 -1.94914062e-02 1.02149630e+00
5.78823686e-01 1.15701422e-01 4.63222057e-01 -3.20243061e-01
1.53925109e+00 -6.80531800e-01 -1.94445491e-01 1.78984866e-01
9.46800053e-01 -2.97632366e-01 9.14352179e-01 3.34211737e-01
-6.16354585e-01 -2.48751678e-02 -9.84037101e-01 -1.87976331e-01
-5.86842716e-01 -2.52361558e-02 9.17874873e-01 8.52097213e-01
-1.09639943e+00 6.00174665e-01 -1.32089972e+00 -6.13090575e-01
7.89694846e-01 4.35663790e-01 -6.12155199e-01 -5.38066804e-01
-8.10492635e-01 1.03100181e+00 1.49923310e-01 -1.15272313e-01
-1.42598581e+00 -1.54102111e+00 -6.84037745e-01 -6.97160605e-04
-1.33426562e-01 -1.39303923e+00 1.01519036e+00 -5.06970167e-01
-1.63360596e+00 8.80025327e-01 -1.02565989e-01 -3.17978621e-01
6.58538878e-01 4.54005480e-01 -2.75334835e-01 6.27417445e-01
-9.37427655e-02 8.93917084e-01 3.39167416e-01 -7.41398990e-01
-1.57458201e-01 -7.29404926e-01 -5.64193010e-01 2.25356847e-01
-3.88897061e-01 -3.80967408e-01 -4.49766964e-02 -2.27345690e-01
2.01552138e-01 -1.01516092e+00 -6.16147816e-01 5.48588216e-01
-4.16495144e-01 3.35692346e-01 2.88927406e-01 -3.89523715e-01
3.32637995e-01 -1.81497657e+00 1.29497811e-01 2.75875814e-02
5.24685085e-01 -3.30999702e-01 -1.74295813e-01 1.81957453e-01
-3.80149066e-01 7.80344963e-01 -1.63771152e-01 -9.07304790e-03
-3.44355702e-02 -1.67938828e-01 5.38972497e-01 1.06558669e+00
3.10775906e-01 1.28930807e+00 -1.20491111e+00 -5.18192887e-01
4.25681829e-01 6.97081268e-01 -2.67106324e-01 -2.16986924e-01
-1.15343742e-01 5.61995387e-01 -4.72862244e-01 1.05609000e+00
3.26239526e-01 -5.38887680e-01 6.67350829e-01 -3.22948903e-01
1.65806979e-01 9.00681987e-02 7.13761523e-02 1.94341791e+00
-5.58429062e-01 5.13810456e-01 -1.75031729e-03 -6.27766430e-01
1.50083631e-01 3.95842761e-01 1.13100719e+00 -6.25214994e-01
3.54840815e-01 1.63786948e-01 3.57186824e-01 -6.47164524e-01
-3.43980372e-01 -1.03985596e+00 1.71515048e-01 2.77016103e-01
-1.49688005e-01 -3.22944969e-01 -2.62688011e-01 -1.64529368e-01
1.64977622e+00 -2.27825508e-01 2.99964130e-01 -3.93304348e-01
3.46770167e-01 3.24001431e-01 2.35295683e-01 1.34101778e-01
-4.74615067e-01 6.14532530e-01 4.82407719e-01 -5.31423055e-02
-1.00604832e+00 -1.28452194e+00 -5.68935633e-01 5.07180274e-01
-3.97565395e-01 5.41349165e-02 -2.98987448e-01 -6.61868095e-01
2.13241413e-01 3.60568255e-01 -1.16331291e+00 -2.21818641e-01
-4.05141935e-02 -1.57553208e+00 6.87005103e-01 6.32839203e-01
-6.13865145e-02 -4.40590829e-01 -3.52424979e-01 3.49386334e-01
-3.67647363e-03 -8.06613564e-01 -3.16843867e-01 8.64437461e-01
-1.17807090e+00 -1.48641419e+00 -7.94638574e-01 -4.83095109e-01
8.20439219e-01 1.74061343e-01 9.42141414e-01 5.19315377e-02
-9.50326622e-01 2.16754749e-01 -2.03978628e-01 -3.87912869e-01
-4.94453162e-01 -1.18571818e-01 4.77446504e-02 -1.81724891e-01
1.99804932e-01 -6.09394193e-01 -1.05140328e+00 9.68818814e-02
-8.81154954e-01 6.93800822e-02 1.12109280e+00 8.96879435e-01
1.09304166e+00 6.55506328e-02 5.83163917e-01 -7.91199386e-01
1.66467667e-01 -7.43017912e-01 -4.39556036e-03 4.95582372e-02
-6.79952800e-01 -4.08333912e-02 6.05670810e-01 -3.03158373e-01
-6.75744057e-01 1.34902820e-01 -8.22049826e-02 -4.06325251e-01
-1.39197201e-01 8.83959949e-01 -9.06299949e-02 6.20740168e-02
6.67455554e-01 2.42023841e-01 2.25915372e-01 -1.51589707e-01
1.98007941e-01 4.45988923e-01 5.79049945e-01 -4.40345258e-01
5.59750617e-01 7.96769679e-01 6.41121209e-01 -7.97357202e-01
-8.02932978e-01 -7.16823101e-01 -7.00666308e-01 1.04716957e-01
1.13596034e+00 -1.02706146e+00 -7.99280763e-01 9.91428271e-02
-3.93430710e-01 -6.87785566e-01 -1.08062938e-01 8.20285320e-01
-6.26551151e-01 2.88868517e-01 -1.08833563e+00 -3.30875695e-01
-4.97768015e-01 -1.04905581e+00 1.02506709e+00 -2.66872883e-01
-2.11389333e-01 -1.22624385e+00 1.18005559e-01 3.78215075e-01
3.67100358e-01 8.43269825e-01 1.41596532e+00 -5.57110190e-01
-4.96133566e-01 -2.63950527e-01 -2.67614484e-01 1.77749824e-02
4.58400398e-01 -8.50026160e-02 -1.18546903e+00 -1.91422790e-01
3.64961438e-02 -3.84414017e-01 9.32891965e-01 5.12680292e-01
1.44667065e+00 2.41579205e-01 -6.21744275e-01 8.83378625e-01
1.76549840e+00 8.63039345e-02 4.72060829e-01 8.71603042e-02
6.07237697e-01 2.81488210e-01 1.62017092e-01 2.40751907e-01
7.54846334e-02 2.02250302e-01 7.11291194e-01 -8.04573968e-02
-3.13080192e-01 -1.51191846e-01 2.62153774e-01 6.39583290e-01
1.38433293e-01 -4.44809496e-01 -9.16694462e-01 4.95274454e-01
-9.74402487e-01 -8.43736231e-01 -4.96184886e-01 1.77000690e+00
9.25264776e-01 -1.14563003e-01 -3.42198193e-01 1.18558127e-02
2.52096057e-01 -2.47175053e-01 -8.17675650e-01 -2.22398173e-02
1.17796876e-01 3.39391947e-01 6.48507237e-01 1.03311025e-01
-6.81461632e-01 3.23523790e-01 6.45019197e+00 3.72397572e-01
-1.37504709e+00 3.96527708e-01 9.24451530e-01 -3.62232804e-01
-3.20379198e-01 -4.57849801e-01 2.70553958e-03 2.08081782e-01
1.53431654e+00 -1.60801411e-01 2.91273385e-01 5.47657847e-01
5.57533920e-01 -1.01803191e-01 -1.70898986e+00 8.51407588e-01
-9.42258537e-02 -1.37167585e+00 -1.49067447e-01 6.67851329e-01
4.46858257e-01 5.39546907e-01 4.31437612e-01 1.07421577e-01
1.64818436e-01 -1.49819815e+00 2.18173638e-01 4.79971170e-01
1.40216935e+00 -4.81428117e-01 9.02213812e-01 1.40206963e-01
-9.15677845e-01 -1.11501953e-02 -3.05676103e-01 4.68156815e-01
-2.64988393e-01 7.94937253e-01 -1.64883566e+00 7.57681072e-01
2.98623562e-01 7.94995248e-01 -6.51302278e-01 1.00473547e+00
1.08208738e-01 5.19452393e-01 -2.28868887e-01 -6.28086850e-02
3.16966087e-01 1.05519027e-01 8.09538644e-03 1.35707188e+00
4.96817321e-01 1.15430295e-01 8.60535577e-02 8.45322609e-01
-2.86770225e-01 7.67375082e-02 -2.62318671e-01 -3.95569474e-01
2.01697007e-01 1.65372717e+00 -9.55794454e-01 -4.07673955e-01
-2.06511974e-01 1.00799072e+00 -5.40526398e-03 -5.86789884e-02
-8.46499085e-01 3.70323718e-01 8.11717510e-01 1.74753979e-01
-1.73941433e-01 8.01777914e-02 -3.37401569e-01 -1.08565986e+00
-6.98467374e-01 -6.44885302e-01 3.63358110e-01 -7.23636150e-01
-1.51136518e+00 3.20205450e-01 -4.48079914e-01 -1.06254327e+00
2.98326105e-01 -1.06454539e+00 -5.57659566e-01 1.01754355e+00
-1.84431469e+00 -1.03855240e+00 -6.32931292e-01 1.90201744e-01
4.34052080e-01 3.79912972e-01 1.11783528e+00 7.55129606e-02
-7.29173005e-01 3.66531700e-01 3.53903711e-01 -8.60882632e-04
6.62053049e-01 -1.75157297e+00 -2.68054008e-01 7.80870169e-02
-2.70659864e-01 4.17323977e-01 3.34356338e-01 -4.89358425e-01
-1.80145478e+00 -1.58073926e+00 3.44292819e-01 -7.44983375e-01
8.14705253e-01 -3.43348756e-02 -4.71736819e-01 4.13286418e-01
-1.74946878e-02 4.26007777e-01 1.58862019e+00 -3.08381975e-01
-3.22002858e-01 3.84948514e-02 -1.45540750e+00 5.27831256e-01
7.87038684e-01 -7.20572412e-01 -1.65116534e-01 6.72779679e-01
4.36312854e-01 -1.78186327e-01 -1.65410531e+00 1.96712777e-01
6.33200705e-01 -5.93342125e-01 1.06334376e+00 -6.57755017e-01
4.51637059e-01 -1.68368414e-01 -3.79390121e-01 -1.48729801e+00
-5.61451495e-01 -2.80928127e-02 3.64654303e-01 4.95162547e-01
5.22881806e-01 -5.22069812e-01 1.10925782e+00 5.59937060e-01
-5.67663789e-01 -1.01336265e+00 -7.96950161e-01 -6.06113374e-01
5.32110929e-01 -4.46334273e-01 4.89363998e-01 1.07178557e+00
3.18775356e-01 -1.58658817e-01 5.68544567e-01 1.97217062e-01
4.53100115e-01 -1.16946369e-01 5.65886647e-02 -1.15291500e+00
-3.02501172e-01 -6.93996787e-01 -5.05097985e-01 -8.85968283e-02
1.50192171e-01 -1.62074184e+00 -1.33859426e-01 -1.56341219e+00
8.20005834e-01 -5.10279238e-01 -9.50838685e-01 3.94767344e-01
1.38927130e-02 4.90072012e-01 -1.76603153e-01 -1.75860841e-02
-1.34627782e-02 3.19913179e-01 1.56050253e+00 -5.86516440e-01
1.15343526e-01 -5.84001839e-01 -9.55114543e-01 4.16361541e-01
7.53700912e-01 -6.37092233e-01 -4.42106366e-01 1.86211765e-01
-1.40863672e-01 4.57737088e-01 5.33233583e-01 -9.58465576e-01
-1.39177039e-01 -4.14776862e-01 1.24135256e+00 -4.08293039e-01
3.73670578e-01 -8.97548914e-01 5.03220022e-01 9.98937905e-01
-3.55439603e-01 -7.90508389e-02 1.22593112e-01 8.22804511e-01
2.32151896e-01 -1.32828981e-01 7.94686377e-01 -5.88653862e-01
-6.83422908e-02 7.09192097e-01 -6.81639791e-01 -3.23518753e-01
9.77881789e-01 -1.41243637e-01 -5.02775311e-01 2.54089713e-01
-7.67158747e-01 4.96856943e-02 6.57203794e-01 -1.71794727e-01
5.38975596e-01 -1.19490993e+00 -7.24462092e-01 -2.22152948e-01
2.36322790e-01 -3.25445719e-02 3.36706370e-01 1.49537075e+00
-7.86303222e-01 5.96882284e-01 -1.35989666e-01 -8.19356143e-01
-8.40997517e-01 7.22825348e-01 7.53118455e-01 -5.15541315e-01
-6.56179547e-01 7.70223498e-01 3.47790748e-01 -2.78410971e-01
-4.03357685e-01 -7.13384807e-01 3.42938811e-01 1.43575549e-01
3.54337871e-01 1.14677481e-01 3.61685038e-01 -1.52106732e-01
-4.05235916e-01 2.09377676e-01 -2.07172763e-02 -5.72434179e-02
1.40318465e+00 1.32424146e-01 -1.90486342e-01 3.51496398e-01
1.64757121e+00 -3.17291468e-01 -8.91704500e-01 3.74922335e-01
-5.80763519e-02 -2.51061052e-01 4.78184193e-01 -1.50398350e+00
-1.04830623e+00 8.90774667e-01 8.42939734e-01 -2.99877405e-01
9.98483717e-01 -9.08302665e-02 7.11487114e-01 2.54335612e-01
1.11037627e-01 -4.62694466e-01 2.22300142e-01 -1.96715862e-01
7.19472826e-01 -1.19979560e+00 1.66671410e-01 -4.85992968e-01
-2.97749609e-01 1.17280257e+00 2.68667489e-01 3.87982011e-01
6.07299268e-01 3.74120474e-01 2.06479833e-01 -4.07591760e-01
-1.12406683e+00 -6.95669129e-02 3.12744714e-02 7.17782438e-01
6.79758549e-01 5.24835527e-01 2.17564940e-01 4.92211699e-01
-1.93121672e-01 9.41267088e-02 6.50373816e-01 6.85758352e-01
-2.96025872e-01 -1.07609153e+00 -1.00406781e-01 8.23129892e-01
-6.90974474e-01 -1.58663645e-01 -5.66158950e-01 9.27088857e-01
-1.56264797e-01 5.42781293e-01 -9.79788825e-02 -2.25371152e-01
-3.13455909e-01 4.92454886e-01 6.91884518e-01 -8.51346016e-01
-5.22492170e-01 2.43745789e-01 5.98326209e-04 -3.22886050e-01
-3.15588921e-01 -8.37425828e-01 -1.44622910e+00 -1.53723881e-01
-2.05979943e-01 -1.02504022e-01 9.66009855e-01 6.10175610e-01
1.48508191e-01 9.73082662e-01 5.66630602e-01 -8.19228649e-01
-4.53095496e-01 -8.91722798e-01 -7.30329156e-01 5.66724129e-02
5.21450698e-01 -3.37676793e-01 -3.27382147e-01 3.46037969e-02] | [15.012991905212402, -2.8357532024383545] |
4d02896b-dd00-479f-ad68-cf921b1bb8e1 | non-contact-heart-rate-measurement-from | 2304.14789 | null | https://arxiv.org/abs/2304.14789v1 | https://arxiv.org/pdf/2304.14789v1.pdf | Non-Contact Heart Rate Measurement from Deteriorated Videos | Remote photoplethysmography (rPPG) offers a state-of-the-art, non-contact methodology for estimating human pulse by analyzing facial videos. Despite its potential, rPPG methods can be susceptible to various artifacts, such as noise, occlusions, and other obstructions caused by sunglasses, masks, or even involuntary facial contact, such as individuals inadvertently touching their faces. In this study, we apply image processing transformations to intentionally degrade video quality, mimicking these challenging conditions, and subsequently evaluate the performance of both non-learning and learning-based rPPG methods on the deteriorated data. Our results reveal a significant decrease in accuracy in the presence of these artifacts, prompting us to propose the application of restoration techniques, such as denoising and inpainting, to improve heart-rate estimation outcomes. By addressing these challenging conditions and occlusion artifacts, our approach aims to make rPPG methods more robust and adaptable to real-world situations. To assess the effectiveness of our proposed methods, we undertake comprehensive experiments on three publicly available datasets, encompassing a wide range of scenarios and artifact types. Our findings underscore the potential to construct a robust rPPG system by employing an optimal combination of restoration algorithms and rPPG techniques. Moreover, our study contributes to the advancement of privacy-conscious rPPG methodologies, thereby bolstering the overall utility and impact of this innovative technology in the field of remote heart-rate estimation under realistic and diverse conditions. | ['Miguel Bordallo López', 'Olli Silvén', 'Constantino Álvarez Casado', 'Le Nguyen', 'Nhi Nguyen'] | 2023-04-28 | null | null | null | null | ['heart-rate-estimation'] | ['medical'] | [ 5.43549538e-01 -1.38780490e-01 1.37435868e-01 -1.62919641e-01
-6.33609831e-01 -4.67909634e-01 2.66250193e-01 -3.21472436e-01
-1.90457895e-01 7.60414243e-01 3.51834476e-01 1.47151470e-03
1.29104930e-03 -4.89035338e-01 -4.12739038e-01 -8.04121017e-01
-5.74694201e-02 -2.77671903e-01 -1.64800659e-01 1.99341312e-01
3.01133633e-01 6.78269982e-01 -1.56522918e+00 -7.40222260e-02
8.82403553e-01 9.65375185e-01 -3.12061757e-01 4.72537696e-01
2.78845757e-01 3.62946510e-01 -6.65707171e-01 -4.77267504e-01
5.37644565e-01 -5.99583745e-01 -1.57202855e-01 2.70755291e-01
5.82363546e-01 -7.20147729e-01 -3.97398651e-01 6.28005743e-01
1.01677251e+00 -9.49118927e-04 2.14083046e-01 -9.90969837e-01
-2.42036745e-01 -1.70555755e-01 -6.16200626e-01 1.99605539e-01
6.33074880e-01 5.28882146e-01 3.13171387e-01 -6.63486242e-01
4.60679024e-01 8.61179113e-01 8.83592129e-01 6.18870676e-01
-1.42464554e+00 -6.98097408e-01 -4.58319485e-01 1.98753893e-01
-1.27956951e+00 -1.02051353e+00 1.06575072e+00 -2.01958507e-01
2.13493347e-01 6.00713491e-01 6.52806342e-01 1.20012975e+00
9.88153890e-02 3.09836119e-01 1.67021394e+00 -3.86691958e-01
4.36071791e-02 1.91377312e-01 -4.70950574e-01 4.50182587e-01
2.76179254e-01 2.51316994e-01 -7.81143546e-01 -4.58263755e-01
8.47244024e-01 -2.69390404e-01 -6.88552856e-01 -3.66297811e-01
-8.20918381e-01 3.37199897e-01 -3.28869760e-01 2.20652312e-01
-4.86896247e-01 -1.19857453e-02 3.88485223e-01 6.02131151e-02
4.81351584e-01 2.20758542e-01 -1.74164891e-01 -4.42209691e-01
-1.00051010e+00 3.34534980e-02 9.48469877e-01 4.42560643e-01
3.01586956e-01 1.26527488e-01 -3.94439697e-01 9.39093828e-01
2.30941266e-01 4.38869029e-01 3.43290776e-01 -1.21242654e+00
2.19727576e-01 1.73043728e-01 2.16177136e-01 -1.30441368e+00
-2.51572251e-01 -4.52731729e-01 -5.79949915e-01 1.70518592e-01
4.65246171e-01 -8.66400301e-02 -6.67787313e-01 1.58927608e+00
5.49274027e-01 7.61925161e-01 -1.22329816e-01 1.05465066e+00
8.44441473e-01 1.22565702e-01 1.33952528e-01 -7.71594822e-01
1.33615911e+00 -3.78837258e-01 -8.96364927e-01 8.79024193e-02
1.77458987e-01 -8.90655935e-01 9.46266472e-01 5.72549045e-01
-1.00300312e+00 -5.08593738e-01 -9.02996242e-01 1.65360793e-01
2.35426277e-01 3.51689793e-02 4.19003397e-01 1.19864345e+00
-8.95345151e-01 8.89627457e-01 -7.13364184e-01 -3.66679668e-01
4.08547252e-01 1.97098702e-01 -4.20951217e-01 -2.53810734e-01
-9.93725002e-01 8.97262275e-01 -3.12633574e-01 6.35614574e-01
-6.13677800e-01 -9.88326132e-01 -7.36999452e-01 -9.72785577e-02
3.53443414e-01 -4.43059534e-01 6.05570912e-01 -6.96312487e-01
-1.84222937e+00 7.77243674e-01 -1.90210864e-01 -2.20789671e-01
8.34766090e-01 -1.47189304e-01 -5.74299157e-01 3.97158206e-01
-3.90196443e-01 1.49073735e-01 1.20239079e+00 -1.09941030e+00
1.97403595e-01 -4.54657704e-01 -5.67286015e-01 1.77169368e-01
-3.50960225e-01 4.46861759e-02 -2.23895565e-01 -4.59612846e-01
8.03015381e-02 -8.63389492e-01 1.48900568e-01 3.93055469e-01
4.62059341e-02 3.92175674e-01 8.16836238e-01 -1.02145350e+00
1.11300910e+00 -2.17922783e+00 -2.84923911e-01 2.40649313e-01
9.74721462e-02 6.32130563e-01 -1.05286635e-01 4.30670679e-01
1.20540597e-01 9.79622975e-02 -2.52094567e-01 -3.65584761e-01
-2.34169394e-01 2.28605658e-01 -8.39877725e-02 8.21582079e-01
-4.81053405e-02 7.52662480e-01 -5.62020004e-01 -5.46310246e-01
5.61031342e-01 8.95906985e-01 -1.58084914e-01 2.26811424e-01
3.79079729e-01 9.17056859e-01 -1.60320640e-01 7.81393409e-01
7.95396626e-01 3.89409035e-01 2.01203823e-01 -4.32695657e-01
-1.15672812e-01 4.27799933e-02 -1.11301494e+00 1.41217542e+00
-3.28455299e-01 6.29039168e-01 1.39620036e-01 -6.64025784e-01
1.15940332e+00 7.03794122e-01 6.96876884e-01 -7.96150208e-01
1.80560902e-01 2.27249146e-01 -1.82374552e-01 -9.73683953e-01
1.12796545e-01 -5.15520573e-01 6.20628178e-01 2.41941631e-01
-1.71346381e-01 1.66095980e-02 -9.03231651e-02 -1.88239038e-01
8.81342351e-01 2.71698713e-01 2.46490702e-01 -1.37570396e-01
5.91048658e-01 -7.61487961e-01 5.62255919e-01 6.29603803e-01
-6.37357473e-01 9.45882976e-01 4.76210624e-01 -3.65542173e-01
-6.87903285e-01 -8.33469272e-01 -4.44402575e-01 3.56222421e-01
-9.38708633e-02 -2.52138585e-01 -5.16600430e-01 -3.93348366e-01
1.77095234e-01 4.10402387e-01 -4.85252172e-01 -2.99432695e-01
-4.49709028e-01 -1.00373924e+00 6.90796316e-01 1.84599847e-01
5.93664169e-01 -9.34994102e-01 -7.72838116e-01 4.22659367e-01
-6.15375817e-01 -1.20997202e+00 -2.93136686e-01 -4.96730745e-01
-9.33041513e-01 -1.14593220e+00 -7.89606333e-01 1.09954011e-02
5.27799904e-01 4.18156423e-02 8.41349602e-01 1.49564907e-01
-6.15578890e-01 7.95429289e-01 -2.30874434e-01 -2.18011022e-01
-4.47520435e-01 -5.08676231e-01 5.11418805e-02 5.74401438e-01
1.10506870e-01 -9.38806355e-01 -1.20248854e+00 4.48290199e-01
-6.75880373e-01 -2.83864379e-01 2.24389851e-01 4.21032906e-01
2.94820964e-01 -2.18271181e-01 6.07473016e-01 -6.53038919e-01
7.76951551e-01 -1.72398865e-01 -4.82393384e-01 4.21839096e-02
-7.40125000e-01 -4.59771842e-01 4.57516640e-01 -5.47600329e-01
-1.13309443e+00 -2.02198431e-01 -4.76998091e-02 -3.90278012e-01
-2.48949200e-01 2.93814629e-01 4.41990467e-03 -6.09272897e-01
8.00050020e-01 3.86597544e-01 3.83854330e-01 -3.01714718e-01
-1.29737854e-01 6.92152441e-01 6.70815766e-01 -5.17692626e-01
7.67921507e-01 6.10632598e-01 3.88619751e-01 -1.08485794e+00
-3.61517906e-01 -3.12160641e-01 -3.33851546e-01 -7.43978441e-01
4.75009620e-01 -7.36147523e-01 -9.18840528e-01 6.74477279e-01
-6.30006790e-01 -1.03367053e-01 -4.25771028e-02 4.50550526e-01
-4.21354741e-01 1.02217829e+00 -4.95615244e-01 -1.14767718e+00
-5.63918293e-01 -8.35574210e-01 7.13140488e-01 4.30851191e-01
-1.77355006e-01 -8.14700544e-01 1.91554055e-01 7.85140097e-01
8.48381102e-01 7.13315725e-01 3.26769859e-01 2.13206168e-02
-4.53159422e-01 -2.84409165e-01 1.16456173e-01 6.11856282e-01
3.21128190e-01 2.87043322e-02 -1.31106424e+00 -4.34917957e-01
3.19836766e-01 -3.55960369e-01 4.08691764e-01 2.22636893e-01
1.14237869e+00 -1.79810360e-01 1.10877238e-01 6.84241295e-01
1.35410440e+00 2.21524090e-02 1.23052812e+00 -1.51159521e-02
2.93276280e-01 6.69742048e-01 4.93741810e-01 7.01621532e-01
4.58816811e-02 6.48171484e-01 3.81322175e-01 -3.65184009e-01
-2.12842464e-01 -1.68278188e-01 1.38773263e-01 3.05003524e-01
-4.41953927e-01 -1.00547886e-02 -4.50984359e-01 2.26393074e-01
-1.26769710e+00 -8.26379716e-01 -8.93140659e-02 2.65114737e+00
7.61434615e-01 -3.21498781e-01 5.22469245e-02 1.94593892e-01
5.83750188e-01 2.26572111e-01 -4.04580593e-01 -3.36451918e-01
1.00293182e-01 4.29477125e-01 3.48072410e-01 6.96398988e-02
-7.94386625e-01 3.38271886e-01 6.40006351e+00 2.55295187e-01
-1.41675127e+00 -1.37037367e-01 6.36879265e-01 -9.07933787e-02
-1.58547223e-01 -3.31538357e-02 -3.13728690e-01 6.76519454e-01
9.60027039e-01 9.77103412e-02 5.64740479e-01 3.34186584e-01
6.42886758e-01 -2.92557269e-01 -6.32971525e-01 9.86893237e-01
4.72934663e-01 -9.70316291e-01 -5.25316894e-01 4.09607887e-02
2.21921384e-01 -4.77992445e-01 1.79252587e-02 -1.24608152e-01
-7.00479925e-01 -7.79549181e-01 6.70226663e-02 6.02202654e-01
7.98355937e-01 -3.06358606e-01 5.41257203e-01 -9.85946134e-02
-7.78359115e-01 2.08724141e-01 -1.37581483e-01 -3.39978710e-02
2.09049374e-01 7.14188457e-01 -6.16265357e-01 4.99214768e-01
3.61097366e-01 4.26575452e-01 -3.60636592e-01 1.22625697e+00
-2.20120206e-01 6.94559038e-01 -4.46308911e-01 3.56936276e-01
-5.46141565e-01 -3.12636733e-01 6.31823242e-01 7.99560666e-01
3.35832417e-01 3.92794102e-01 -3.30808282e-01 7.53752232e-01
8.02256763e-02 2.38873288e-01 -5.34875989e-01 6.24246970e-02
4.00065511e-01 1.45055926e+00 -4.88893956e-01 -2.16155369e-02
-4.89063203e-01 8.12175632e-01 -1.75205275e-01 5.83832681e-01
-8.17990422e-01 -6.08179867e-02 5.21322548e-01 5.19095838e-01
1.53911086e-02 -8.70626122e-02 -2.66167283e-01 -1.16752493e+00
4.26494002e-01 -9.67018008e-01 4.92240459e-01 -8.34667742e-01
-8.65181088e-01 3.35841179e-01 -1.89351551e-02 -1.33122480e+00
6.57323971e-02 -1.12013064e-01 -7.06160963e-01 6.95915937e-01
-1.84085941e+00 -9.32432055e-01 -5.71980596e-01 5.53841531e-01
1.33808494e-01 3.38234633e-01 8.13173771e-01 4.68488425e-01
-4.82671112e-01 6.01026475e-01 -2.50808448e-01 -3.06126952e-01
9.56018090e-01 -6.30297542e-01 -2.30436578e-01 8.46026719e-01
-1.58856556e-01 6.09071672e-01 8.55942905e-01 -4.97146457e-01
-1.61730480e+00 -6.78455651e-01 6.18325412e-01 -1.56073481e-01
2.70759940e-01 -3.67573917e-01 -9.15059388e-01 2.66124845e-01
-1.16954349e-01 3.31237644e-01 7.37338543e-01 -2.04887241e-01
-6.81831762e-02 -5.79730809e-01 -1.54236746e+00 5.01850605e-01
8.09853315e-01 -4.07485694e-01 -3.58936727e-01 5.59009314e-02
-1.10873960e-01 -5.66183448e-01 -1.22654450e+00 6.77004337e-01
1.04683459e+00 -1.15728951e+00 9.01583135e-01 4.63614203e-02
2.76406318e-01 -9.69393253e-02 3.26304913e-01 -1.01386702e+00
7.92153180e-02 -1.09648645e+00 -1.20237418e-01 1.42355967e+00
-6.27109930e-02 -1.06740355e+00 7.39531457e-01 1.13451636e+00
7.34649673e-02 -4.40372914e-01 -1.05253208e+00 -4.95756745e-01
-3.54449570e-01 -1.95901856e-01 1.26293063e-01 8.94922614e-01
6.26925528e-02 -5.12196362e-01 -8.87293279e-01 1.54123440e-01
7.56377518e-01 1.54666379e-02 8.63141596e-01 -9.49156523e-01
-2.78445393e-01 2.17090860e-01 -3.87537062e-01 -3.24091852e-01
-1.86573882e-02 -2.39768684e-01 -1.02564748e-02 -9.98661399e-01
-2.90065091e-02 -3.29804242e-01 -2.23440811e-01 3.46804231e-01
-2.29492337e-01 7.92242587e-01 1.92103609e-01 1.02545559e-01
8.18992481e-02 4.11132306e-01 1.40122831e+00 3.86160225e-01
-3.92141074e-01 2.98862718e-02 -7.18014956e-01 3.95439386e-01
8.24304044e-01 -4.81634110e-01 -4.67035413e-01 1.00264482e-01
-1.03374995e-01 4.49897200e-01 6.15422845e-01 -1.06055272e+00
1.28833652e-02 5.82084097e-02 4.83932346e-01 -9.73789021e-02
5.01042962e-01 -8.07858467e-01 6.17828131e-01 4.20089811e-01
3.79535668e-02 -4.41634208e-01 3.46499234e-01 2.97383666e-01
-3.59841832e-03 1.73548404e-02 8.71927619e-01 -2.18197867e-01
-2.16656208e-01 1.06827594e-01 -2.98479855e-01 -5.55288233e-02
8.87616158e-01 -6.06274903e-01 -3.56040806e-01 -5.87369204e-01
-6.57549322e-01 -8.90252590e-02 3.28543127e-01 2.20134005e-01
6.36603475e-01 -8.92402649e-01 -8.09594512e-01 6.51966453e-01
-9.78305712e-02 -5.75688362e-01 6.52236700e-01 1.51558101e+00
-4.31823939e-01 -5.78535721e-02 -3.10124010e-01 -4.62861747e-01
-1.47333217e+00 2.88164347e-01 4.75533485e-01 2.01908067e-01
-8.80670071e-01 3.23530793e-01 -2.12256700e-01 -2.11652275e-02
1.88386187e-01 1.72517568e-01 -2.16826722e-02 -1.48062840e-01
4.37213600e-01 5.90188682e-01 2.24297836e-01 -2.22117037e-01
-2.12975621e-01 6.12576008e-01 2.28106081e-01 1.93911433e-01
1.01407862e+00 -3.91999781e-01 -2.44166832e-02 -4.09041792e-02
8.48933399e-01 9.35084149e-02 -1.36234379e+00 1.81174740e-01
-5.04509330e-01 -9.55886304e-01 -7.54063437e-03 -8.95099699e-01
-1.27821684e+00 5.78869462e-01 8.60676765e-01 -1.09511644e-01
1.53891540e+00 -3.84185195e-01 7.12729990e-01 -2.64935285e-01
3.58872056e-01 -8.55609119e-01 -2.14941818e-02 -5.23712039e-01
7.38003850e-01 -8.88908923e-01 3.62275988e-01 -6.54893696e-01
-5.03174245e-01 1.01676309e+00 2.19721854e-01 1.63045198e-01
3.75181586e-01 2.97167003e-01 5.09293854e-01 7.65936002e-02
-4.95458364e-01 8.99384469e-02 1.61344483e-01 6.79816067e-01
4.34537441e-01 -2.58342713e-01 -9.21612740e-01 -8.94450843e-02
1.30253077e-01 4.92158264e-01 7.38508224e-01 8.82374048e-01
3.23873125e-02 -1.13791692e+00 -5.54807544e-01 4.57391351e-01
-7.00724244e-01 1.56348124e-01 -7.95365945e-02 8.53729784e-01
-9.00496319e-02 1.04919314e+00 -4.03200924e-01 7.53131732e-02
5.35970569e-01 1.73187226e-01 5.25083244e-01 3.46608721e-02
-5.44640005e-01 3.10608357e-01 1.29060343e-01 -7.79716611e-01
-6.97959602e-01 -9.16438103e-01 -6.21197641e-01 -3.20371360e-01
-3.35807443e-01 -2.44323447e-01 7.71260619e-01 7.60370731e-01
5.52711070e-01 1.18319750e-01 7.24344909e-01 -5.88835359e-01
-5.13177097e-01 -8.75946462e-01 -6.49917424e-01 6.92492783e-01
2.04644695e-01 -6.21394634e-01 -6.82958961e-01 5.54835889e-03] | [13.865635871887207, 2.6870622634887695] |
26f011f4-9673-4559-ad84-d2a0d4eed190 | damuel-a-large-multilingual-dataset-for | 2306.09288 | null | https://arxiv.org/abs/2306.09288v1 | https://arxiv.org/pdf/2306.09288v1.pdf | DaMuEL: A Large Multilingual Dataset for Entity Linking | We present DaMuEL, a large Multilingual Dataset for Entity Linking containing data in 53 languages. DaMuEL consists of two components: a knowledge base that contains language-agnostic information about entities, including their claims from Wikidata and named entity types (PER, ORG, LOC, EVENT, BRAND, WORK_OF_ART, MANUFACTURED); and Wikipedia texts with entity mentions linked to the knowledge base, along with language-specific text from Wikidata such as labels, aliases, and descriptions, stored separately for each language. The Wikidata QID is used as a persistent, language-agnostic identifier, enabling the combination of the knowledge base with language-specific texts and information for each entity. Wikipedia documents deliberately annotate only a single mention for every entity present; we further automatically detect all mentions of named entities linked from each document. The dataset contains 27.9M named entities in the knowledge base and 12.3G tokens from Wikipedia texts. The dataset is published under the CC BY-SA license at https://hdl.handle.net/11234/1-5047. | ['Milan Straka', 'David Kubeša'] | 2023-06-15 | null | null | null | null | ['entity-linking'] | ['natural-language-processing'] | [-7.97530770e-01 4.96549904e-01 -5.16492605e-01 1.82729468e-01
-8.98482621e-01 -1.03882682e+00 7.13277996e-01 8.04722667e-01
-6.09383881e-01 1.09058762e+00 6.46165788e-01 6.57204241e-02
-1.15528055e-01 -1.10589445e+00 -6.69930458e-01 -1.24969468e-01
3.08654845e-01 7.04376221e-01 4.69444513e-01 -3.21289599e-01
4.22838442e-02 2.61615872e-01 -1.00297022e+00 1.03855833e-01
1.01285219e+00 7.40820527e-01 2.66633689e-01 2.03717694e-01
-8.62431943e-01 9.49049771e-01 -7.53006458e-01 -1.00287509e+00
-1.14750713e-01 -3.33829746e-02 -1.09495366e+00 -4.38785225e-01
5.08766890e-01 2.09980696e-01 -4.60062087e-01 1.00754583e+00
2.78639883e-01 -4.50629383e-01 5.95480382e-01 -1.25466478e+00
-8.93397689e-01 1.15850103e+00 -3.12883854e-01 5.16154543e-02
4.37419772e-01 -3.28460991e-01 1.20967412e+00 -8.53483796e-01
1.48942769e+00 9.04697061e-01 6.42979383e-01 2.09584132e-01
-5.99604428e-01 -6.71589434e-01 -4.03514475e-01 1.05487416e-03
-1.54679525e+00 -4.98119324e-01 1.77317262e-01 -7.38299370e-01
8.92563760e-01 7.68983066e-02 1.92038044e-01 7.68599570e-01
-2.58824885e-01 1.09848745e-01 7.74856925e-01 -6.70393825e-01
-1.69771925e-01 6.34891510e-01 3.81649971e-01 7.62191296e-01
1.07683206e+00 -5.46667814e-01 -4.99552578e-01 -2.62514353e-01
3.88766229e-01 -6.14886165e-01 -1.78981036e-01 -3.64890814e-01
-1.30137670e+00 2.95248359e-01 2.19398364e-01 4.94707704e-01
-4.39486921e-01 -5.11994399e-02 8.68980825e-01 4.94559146e-02
2.64526427e-01 5.87699711e-01 -8.73158693e-01 4.05078307e-02
-3.72144014e-01 2.06558377e-01 1.26312482e+00 1.62211692e+00
1.03908241e+00 -4.54177380e-01 1.08385570e-01 1.04522276e+00
4.03661162e-01 6.06516480e-01 4.00133759e-01 -9.16430354e-01
8.97459090e-01 1.04356301e+00 5.25336862e-01 -1.02146268e+00
-5.91282725e-01 -3.69383097e-01 -1.11832663e-01 -5.65940678e-01
4.34401929e-01 -2.43036255e-01 -4.90190476e-01 1.61080730e+00
5.93084455e-01 -3.09013009e-01 3.51809949e-01 3.69709939e-01
1.62805498e+00 3.13878596e-01 3.64513934e-01 1.39201032e-02
1.74718451e+00 -5.50862491e-01 -9.93165135e-01 7.26641491e-02
1.07369268e+00 -7.91924417e-01 2.91809291e-01 -3.33912939e-01
-9.21067834e-01 -1.52958751e-01 -6.67588949e-01 -2.82312721e-01
-1.22180998e+00 2.33484298e-01 2.77706385e-01 5.23888588e-01
-6.24527454e-01 4.98752818e-02 -4.14791137e-01 -6.25751138e-01
2.05270693e-01 -1.28539085e-01 -7.73532510e-01 2.23311305e-01
-1.70949674e+00 1.25925171e+00 1.04862165e+00 -3.43675524e-01
-3.74814898e-01 -7.22208023e-01 -1.30021918e+00 -3.18725497e-01
7.79504895e-01 -3.80271405e-01 9.78045523e-01 -2.93101013e-01
-5.60190618e-01 1.35289919e+00 1.43768638e-01 -3.72624665e-01
1.79545745e-01 5.96205844e-03 -1.14449608e+00 1.32205024e-01
1.03038239e+00 2.88180828e-01 1.73868820e-01 -1.17747664e+00
-1.06980288e+00 -2.88744569e-01 2.71801442e-01 -1.57988474e-01
-4.04826343e-01 4.52833980e-01 -5.43670416e-01 -5.47443330e-01
-2.62811482e-01 -7.53399551e-01 3.64721894e-01 -4.62921441e-01
-7.65227079e-01 -3.86002600e-01 4.61364001e-01 -1.22550249e+00
1.66603374e+00 -1.77676463e+00 -2.48084828e-01 -4.01160717e-02
3.04046750e-01 1.19679287e-01 2.50476480e-01 9.81830299e-01
4.35785018e-02 5.10721922e-01 -3.65601033e-02 1.91509232e-01
4.97642428e-01 1.43513471e-01 1.12844156e-02 3.97495061e-01
-3.87400165e-02 8.90876472e-01 -1.20851529e+00 -8.37702036e-01
-3.40792000e-01 1.86293259e-01 -9.18073058e-02 -3.39865297e-01
-1.80738553e-01 -2.94669531e-03 -6.64594889e-01 6.32697344e-01
5.12494087e-01 -1.24464519e-01 5.65471947e-01 -5.78903615e-01
-6.43452466e-01 7.86908150e-01 -1.40304947e+00 1.47344303e+00
-4.04900849e-01 5.82300842e-01 2.54398853e-01 -2.66619384e-01
7.35792160e-01 6.58963919e-01 3.81666869e-01 -2.16165170e-01
5.97913414e-02 7.48848915e-01 -4.61412817e-01 -5.14353216e-01
1.04647040e+00 3.91861916e-01 -7.16391206e-01 1.98162198e-01
4.27997708e-01 1.58128113e-01 9.49241817e-01 6.19877934e-01
9.83055592e-01 2.99510837e-01 7.04867244e-01 -2.55992234e-01
6.65578246e-01 5.07111907e-01 7.41913855e-01 3.42050225e-01
1.30945444e-01 -1.91878617e-01 4.73880053e-01 1.24451794e-01
-1.29710019e+00 -9.23659027e-01 -6.35421395e-01 7.59704113e-01
9.06561539e-02 -8.01018357e-01 -6.15626574e-01 -7.50388741e-01
4.84000266e-01 8.79581630e-01 -4.66097742e-01 2.89129853e-01
-4.89569366e-01 -1.44628420e-01 9.66895819e-01 2.30330124e-01
3.81471813e-01 -9.26179588e-01 -1.32734016e-01 3.08057040e-01
-3.77488524e-01 -1.33695555e+00 -3.88364404e-01 -5.85078448e-02
8.08331519e-02 -1.30054533e+00 -5.10337353e-01 -7.97095001e-01
4.99459177e-01 -4.25545633e-01 1.37111592e+00 -1.85397014e-01
-8.74356925e-03 6.02796793e-01 -4.46029603e-01 -1.75898626e-01
-7.68376291e-01 2.23904774e-01 3.76191773e-02 -4.43987787e-01
4.28091884e-01 -3.46439257e-02 4.11114581e-02 1.53854132e-01
-7.50527442e-01 -3.08743149e-01 1.45830989e-01 5.02674222e-01
2.31316134e-01 -2.20925119e-02 7.86986768e-01 -1.26483512e+00
2.28054434e-01 -1.05798471e+00 -6.65319026e-01 5.22452414e-01
-3.47675115e-01 -5.15789427e-02 4.02357042e-01 3.18932198e-02
-1.32168281e+00 -2.84544975e-01 1.12001583e-01 2.94424087e-01
-2.61827558e-01 1.10570467e+00 -5.22009254e-01 2.66406357e-01
5.83583772e-01 -1.14010237e-01 -5.00048637e-01 -7.05748498e-01
8.21640134e-01 8.43788624e-01 9.27421391e-01 -8.24620128e-01
8.22195709e-01 1.05904259e-01 -3.99823666e-01 -5.25632560e-01
-9.61425662e-01 -6.35288715e-01 -1.13189113e+00 -2.35147148e-01
7.44790852e-01 -1.29293811e+00 -2.93406993e-01 3.27403635e-01
-1.02637839e+00 7.45807514e-02 -4.74091172e-01 4.33297157e-01
-9.50981528e-02 2.89642394e-01 -7.89620519e-01 -3.01318049e-01
-2.56773859e-01 -4.00998920e-01 4.09388572e-01 1.77261293e-01
-3.73832077e-01 -1.17327166e+00 1.62803307e-01 3.41417879e-01
8.71273130e-03 5.27429461e-01 8.41787696e-01 -1.12576091e+00
-1.87981948e-01 -4.31281358e-01 -4.02088076e-01 -6.43749982e-02
2.73990035e-01 1.61401913e-01 -3.55144948e-01 5.68600483e-02
-7.59879887e-01 -5.21220006e-02 5.42516172e-01 -3.05639237e-01
-3.32223400e-02 -5.81120253e-01 -6.21311128e-01 1.03698507e-01
1.58727086e+00 -8.31706263e-03 3.30560148e-01 1.06067491e+00
9.15509224e-01 6.58891678e-01 3.87081653e-01 3.37331206e-01
8.69246423e-01 7.45729148e-01 8.45841244e-02 3.69277298e-01
-3.81105572e-01 -3.76720637e-01 2.26231992e-01 8.74759138e-01
1.89855933e-01 -2.22329855e-01 -1.21727240e+00 1.06158781e+00
-1.45755041e+00 -1.07872045e+00 -7.52662778e-01 2.09314942e+00
1.41801620e+00 -9.65588465e-02 7.96553269e-02 -2.48288244e-01
9.60707247e-01 -1.30892396e-01 -8.58525038e-02 8.21020976e-02
-4.43341583e-01 1.64845940e-02 1.11813712e+00 3.08234483e-01
-1.00757861e+00 9.06912923e-01 4.82742691e+00 7.15065360e-01
-6.27200902e-01 5.14547169e-01 -5.39329231e-01 3.51236016e-01
-3.31280768e-01 2.93705910e-01 -1.43415749e+00 6.80919528e-01
1.06301725e+00 -7.87079751e-01 -5.72409220e-02 7.32866228e-01
-1.18056320e-01 -2.73955725e-02 -3.55012804e-01 2.92354345e-01
1.68053862e-02 -1.55004930e+00 -1.79510608e-01 1.60742059e-01
7.63326466e-01 2.39245892e-01 -4.11587298e-01 3.35088134e-01
6.23198926e-01 -1.97549775e-01 1.31103194e+00 4.48749423e-01
1.07175183e+00 -6.58580124e-01 7.96202302e-01 6.78661540e-02
-1.26991260e+00 1.34283796e-01 -2.47432128e-01 4.83804226e-01
3.66991013e-01 7.76162088e-01 -4.90462512e-01 1.11320877e+00
6.77559376e-01 8.08436275e-01 -5.75104415e-01 1.12118304e+00
-5.05968690e-01 4.45386529e-01 -1.90718158e-03 1.43711373e-01
5.02689630e-02 -1.22532241e-01 6.69002056e-01 1.64115596e+00
4.09161836e-01 -1.77666873e-01 7.45165274e-02 5.00505984e-01
-7.77036846e-01 4.59807724e-01 -4.76447523e-01 -4.10607964e-01
1.21098840e+00 1.42530811e+00 -6.62126005e-01 -6.98265254e-01
-6.30848408e-01 5.95430672e-01 5.34263790e-01 4.79943901e-02
-5.43235958e-01 -9.88926053e-01 4.02032912e-01 3.16834241e-01
4.71319079e-01 -1.28597707e-01 3.31722915e-01 -1.06444716e+00
4.84330580e-02 -4.14397448e-01 7.69095600e-01 -9.22682643e-01
-1.24526834e+00 3.23806703e-01 1.26171857e-01 -1.12035823e+00
-1.55783594e-01 -5.61485946e-01 2.24390581e-01 9.69853580e-01
-1.44488227e+00 -1.29132831e+00 -3.47105041e-02 2.16533422e-01
-2.19304889e-01 -2.00698838e-01 8.27553272e-01 9.77878749e-01
-6.93694413e-01 3.19489539e-01 4.22953129e-01 9.49357927e-01
1.17543769e+00 -1.36716807e+00 2.96396047e-01 5.72564900e-01
-1.41282275e-01 8.64811778e-01 6.28890932e-01 -1.19004834e+00
-9.40575361e-01 -1.31458247e+00 1.63764548e+00 -9.39155817e-01
1.53030646e+00 -1.24102078e-01 -9.77879405e-01 1.02111828e+00
4.07332033e-01 -1.09601311e-01 7.32969999e-01 2.33043447e-01
-6.49915516e-01 2.26881608e-01 -1.19222426e+00 1.28421992e-01
9.52519596e-01 -7.03523040e-01 -8.34640324e-01 4.83759463e-01
6.88467503e-01 -8.58596265e-01 -1.61136532e+00 -1.75520733e-01
3.46731663e-01 -8.07588324e-02 7.51338601e-01 -4.37175155e-01
2.74344623e-01 -6.73357308e-01 -8.23195651e-02 -1.15229869e+00
-2.71378338e-01 -8.78069550e-02 -3.06224227e-01 2.22039294e+00
8.81642282e-01 -9.63192046e-01 1.07390478e-01 5.24677873e-01
-5.26744783e-01 1.36133119e-01 -1.00832665e+00 -1.06236374e+00
1.55509576e-01 -1.66705325e-01 7.47528195e-01 1.60237443e+00
4.70343649e-01 3.58167887e-01 -2.10118946e-02 2.43612111e-01
5.26691675e-01 -1.23409554e-01 4.43199188e-01 -1.32899380e+00
3.38808358e-01 -3.41973454e-01 -4.05933827e-01 -1.74278304e-01
3.00206095e-01 -1.23227417e+00 -2.68147707e-01 -1.91459346e+00
8.04605931e-02 -8.27474296e-01 4.04974036e-02 9.06523883e-01
6.99425768e-03 -8.32924917e-02 -5.64215705e-02 5.34063399e-01
-7.34108627e-01 7.16959639e-03 7.17308998e-01 -6.25572875e-02
9.14421901e-02 -6.93199635e-01 -6.63753331e-01 5.76152146e-01
5.42265415e-01 -8.80490959e-01 4.87093538e-01 -3.27323884e-01
7.19532013e-01 -2.49937117e-01 1.05049558e-01 -7.58616209e-01
4.14086968e-01 -1.63399622e-01 -6.26539066e-02 -4.81263816e-01
-5.15104160e-02 -6.83674932e-01 5.54932058e-01 3.20937455e-01
-3.87885600e-01 -1.47926226e-01 1.63683102e-01 3.38248909e-01
-2.13432208e-01 -7.20431387e-01 4.35359031e-01 -3.29593182e-01
-1.07394445e+00 -6.43470064e-02 -1.83502257e-01 6.54680967e-01
8.01959515e-01 1.79631934e-01 -8.16994071e-01 1.84464574e-01
-7.25532115e-01 3.03110421e-01 7.85336852e-01 4.48970556e-01
-3.04985732e-01 -1.46547091e+00 -9.91220772e-01 -5.06251693e-01
6.26819968e-01 -4.55254167e-01 1.80461742e-02 5.40576935e-01
-3.53564441e-01 4.64404970e-01 -1.35172531e-01 2.73688942e-01
-7.71813393e-01 6.13450289e-01 5.12865223e-02 -1.83692634e-01
-4.71555144e-01 2.55977124e-01 -3.74112517e-01 -8.54447126e-01
-2.89667547e-01 1.10773399e-01 -5.83387494e-01 6.56363666e-01
3.15050244e-01 5.47499299e-01 7.16454983e-02 -1.31997192e+00
-5.54414988e-01 2.59908170e-01 8.92421454e-02 -5.74759170e-02
1.22438598e+00 -4.69247878e-01 -6.67863131e-01 5.38266003e-01
1.01735783e+00 8.62441123e-01 -2.43809968e-01 -4.82944429e-01
6.48413956e-01 4.73505445e-02 -2.42359072e-01 -1.14040887e+00
-8.61497104e-01 -1.19594239e-01 -1.37913465e-01 3.15249175e-01
4.20549750e-01 5.48326492e-01 6.91696823e-01 6.39822558e-02
6.30930960e-01 -1.25132036e+00 -7.34451890e-01 7.84117579e-01
7.56612957e-01 -8.35357547e-01 1.28597403e-02 -6.29813373e-01
-4.71892238e-01 7.54455328e-01 2.84458458e-01 4.74956870e-01
6.23513579e-01 1.74922124e-01 1.85169101e-01 -5.12469172e-01
-4.09830838e-01 -5.51058829e-01 2.63544768e-01 6.08179927e-01
5.98695099e-01 1.39616862e-01 -6.71800554e-01 8.00558805e-01
-1.80946663e-01 -1.33892298e-01 7.41102576e-01 1.05788875e+00
-2.84267545e-01 -1.09021139e+00 -2.55212814e-01 4.49345171e-01
-7.65154600e-01 -3.20528865e-01 -2.72632539e-01 1.17163754e+00
4.63599265e-01 8.10819030e-01 1.15611158e-01 1.82670087e-01
5.11665940e-01 1.25890955e-01 1.02489591e-01 -9.46027339e-01
-7.45698035e-01 -4.63827342e-01 1.01917994e+00 1.46929979e-01
-4.84934717e-01 -7.44241893e-01 -1.65498042e+00 -2.86535859e-01
-4.48796570e-01 6.10374629e-01 7.14408755e-01 7.40057111e-01
4.91914779e-01 2.24451125e-01 1.13944910e-01 -4.95791957e-02
3.01915795e-01 -7.15706229e-01 -7.16861486e-01 3.55673701e-01
-1.27153829e-01 -7.17020333e-01 -3.95013869e-01 2.80130059e-01] | [9.497920989990234, 8.938844680786133] |
1d216d73-bb6a-470f-8afc-75372eb5aeab | improving-japanese-semantic-role-labeling | null | null | https://aclanthology.org/Y18-1058 | https://aclanthology.org/Y18-1058.pdf | Improving Japanese semantic-role-labeling performance with transfer learning as case for limited resources of tagged corpora on aggregated language | null | ['Hitoshi Ueda', 'Tatsuhiko Abo', 'Masaya Iizuka', 'Yoshihiko Inada', 'Masahiro Taguchi', 'Yasuhiro Ishihara', 'Koichi Takeuchi', 'Takuya Okamura'] | null | null | null | null | paclic-2018-12 | ['semantic-role-labeling'] | ['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] | [-7.129366397857666, 3.7449145317077637] |
eb45ce21-37ad-42b4-8228-04809a6bd4c9 | controllable-image-enhancement | 2206.08488 | null | https://arxiv.org/abs/2206.08488v1 | https://arxiv.org/pdf/2206.08488v1.pdf | Controllable Image Enhancement | Editing flat-looking images into stunning photographs requires skill and time. Automated image enhancement algorithms have attracted increased interest by generating high-quality images without user interaction. However, the quality assessment of a photograph is subjective. Even in tone and color adjustments, a single photograph of auto-enhancement is challenging to fit user preferences which are subtle and even changeable. To address this problem, we present a semiautomatic image enhancement algorithm that can generate high-quality images with multiple styles by controlling a few parameters. We first disentangle photo retouching skills from high-quality images and build an efficient enhancement system for each skill. Specifically, an encoder-decoder framework encodes the retouching skills into latent codes and decodes them into the parameters of image signal processing (ISP) functions. The ISP functions are computationally efficient and consist of only 19 parameters. Despite our approach requiring multiple inferences to obtain the desired result, experimental results present that the proposed method achieves state-of-the-art performances on the benchmark dataset for image quality and model efficiency. | ['Kyoung Mu Lee', 'Heewon Kim'] | 2022-06-16 | null | null | null | null | ['photo-retouching'] | ['computer-vision'] | [ 9.06635702e-01 -1.37853593e-01 2.75596350e-01 -5.32600701e-01
-7.66823292e-01 -4.91244406e-01 3.66188645e-01 -2.95740485e-01
-4.66951221e-01 4.54042166e-01 -1.78092763e-01 -1.83719277e-01
-1.17431656e-01 -6.53129160e-01 -7.25023150e-01 -6.90441012e-01
3.18693072e-01 -3.58723819e-01 1.35352015e-01 -6.43677786e-02
3.95081520e-01 2.72053450e-01 -1.72466147e+00 2.26769060e-01
1.34042418e+00 9.54867005e-01 5.16465962e-01 1.15587485e+00
2.76800603e-01 6.86806321e-01 -5.24502635e-01 -9.24683452e-01
5.41844130e-01 -5.40907025e-01 -5.02500534e-01 6.98061228e-01
6.22983992e-01 -6.39507890e-01 -1.72908947e-01 1.50440681e+00
5.91903985e-01 -1.53239399e-01 5.06661296e-01 -1.05270255e+00
-1.10215378e+00 2.32723132e-01 -6.86277747e-01 3.84338722e-02
2.30096698e-01 2.80853778e-01 5.57541668e-01 -7.94856966e-01
3.88642728e-01 9.57196593e-01 5.36931932e-01 2.57843763e-01
-1.33017898e+00 -6.66448057e-01 -2.10885316e-01 2.63220668e-01
-1.41105485e+00 -6.31190419e-01 7.17345893e-01 -2.94427335e-01
4.68474478e-01 4.42663163e-01 6.20156586e-01 6.49752557e-01
2.61212707e-01 5.68616509e-01 1.40329969e+00 -4.25765932e-01
4.01009209e-02 4.31162089e-01 -3.76967639e-01 7.37247765e-01
2.27916062e-01 1.93290368e-01 -4.14596170e-01 3.56227607e-01
1.00466871e+00 -2.92578787e-02 -4.19669926e-01 -1.10937431e-01
-1.04953909e+00 3.49764436e-01 2.25097492e-01 1.26139060e-01
-4.85092759e-01 -1.07845604e-01 -3.89180034e-02 3.82831693e-01
1.42941698e-01 5.19184887e-01 -2.33502567e-01 -3.18159223e-01
-1.14895713e+00 -1.60417616e-01 1.72923058e-01 1.06218147e+00
7.00979590e-01 1.32458761e-01 -1.68007687e-01 1.05899239e+00
-2.15951309e-01 5.75976491e-01 3.22457314e-01 -1.17797756e+00
3.19910854e-01 3.04968715e-01 3.18110824e-01 -1.19992423e+00
-6.91559166e-03 -4.63424444e-01 -1.14845860e+00 5.78742504e-01
2.40659684e-01 -1.27314195e-01 -8.29629481e-01 1.38949013e+00
4.81456965e-02 -1.17680229e-01 -3.72877866e-02 1.05743647e+00
3.57266128e-01 8.73244524e-01 5.09385876e-02 -3.82106930e-01
1.50186980e+00 -1.04449272e+00 -1.09727490e+00 -3.01650345e-01
-1.56791117e-02 -1.18199480e+00 1.24848735e+00 7.50923455e-01
-1.71144700e+00 -1.09075272e+00 -1.33995211e+00 -2.23006248e-01
-2.03919765e-02 7.31252968e-01 4.01255608e-01 9.00107861e-01
-1.29608595e+00 5.76007664e-01 -4.45262879e-01 -2.48773508e-02
1.77901253e-01 2.64476389e-01 -3.78391027e-01 -5.74140586e-02
-1.06951833e+00 8.80383432e-01 2.03274548e-01 2.55716592e-01
-6.07673824e-01 -5.51730096e-01 -8.67055476e-01 1.25123262e-01
3.25742364e-01 -5.94763756e-01 1.15829587e+00 -1.40599871e+00
-1.92958057e+00 8.59513342e-01 -2.73899157e-02 -1.53179437e-01
5.23091495e-01 -1.11491263e-01 -7.63838649e-01 4.00152206e-01
-2.96215206e-01 6.30370438e-01 1.37188411e+00 -1.40786362e+00
-8.42140853e-01 2.08581872e-02 -7.76472390e-02 3.22957426e-01
-4.60062116e-01 2.96786785e-01 -8.75880063e-01 -8.06961000e-01
2.13978235e-02 -6.36489630e-01 -9.20818746e-02 4.23417538e-01
-2.14979678e-01 6.74742341e-01 6.14737570e-01 -1.12762797e+00
1.37351024e+00 -2.28310323e+00 -6.58472925e-02 -2.92785577e-02
1.86266080e-01 4.45436865e-01 -1.56494677e-01 -1.74627170e-01
3.30732055e-02 -3.65589112e-02 -2.93974429e-01 -3.19668621e-01
-1.05080314e-01 -1.77941054e-01 1.37457043e-01 3.21015179e-01
2.88631588e-01 7.49092162e-01 -7.39152551e-01 -6.72101080e-01
4.07243758e-01 6.60276175e-01 -7.00974464e-01 4.61712629e-01
4.59650993e-01 7.09759444e-02 3.88886454e-03 6.07264102e-01
1.01389170e+00 -2.22000569e-01 9.66840461e-02 -6.78556323e-01
-1.08398184e-01 -2.05163360e-01 -1.34194589e+00 1.48900592e+00
-5.99965215e-01 8.09243619e-01 2.68679947e-01 -4.73248512e-01
9.09454465e-01 1.56388015e-01 1.47826880e-01 -8.19567025e-01
2.32388958e-01 9.69784781e-02 -1.33933678e-01 -6.41029775e-01
9.83612001e-01 -1.79209873e-01 7.64002576e-02 4.15425390e-01
-3.20887640e-02 -3.17018330e-01 2.45399401e-01 1.29331173e-02
6.53138936e-01 -6.25001267e-02 2.85338163e-01 1.40457526e-01
5.26290655e-01 -4.74491984e-01 4.84002262e-01 4.76873755e-01
-2.06401482e-01 9.97298717e-01 1.83858320e-01 -9.73701105e-02
-1.43807280e+00 -1.22552550e+00 2.67920252e-02 8.33517432e-01
3.32763702e-01 -1.81467697e-01 -9.97208953e-01 -1.70307457e-01
-6.37554169e-01 5.86477816e-01 -3.85526836e-01 -2.59135783e-01
-3.11567992e-01 -6.74940526e-01 4.24578160e-01 2.75750577e-01
8.44999909e-01 -8.83991778e-01 -6.31786048e-01 7.64884800e-02
-3.86478484e-01 -1.30713201e+00 -1.10143459e+00 -2.37830073e-01
-7.38278449e-01 -8.25915694e-01 -1.07600784e+00 -9.36099887e-01
1.17671084e+00 4.71001804e-01 8.63990843e-01 -3.35174538e-02
-3.93477291e-01 8.31915885e-02 -2.25778744e-01 -5.74669540e-02
-6.37483537e-01 -3.12774658e-01 -1.23045236e-01 4.39134955e-01
-1.12027377e-01 -4.92710143e-01 -9.42337096e-01 4.65772390e-01
-1.24142146e+00 4.66079891e-01 1.13286901e+00 7.23907590e-01
7.30659008e-01 6.39915645e-01 2.07537085e-01 -5.99642098e-01
7.89292812e-01 1.68784797e-01 -7.92444766e-01 4.22435194e-01
-7.75415242e-01 -4.09335829e-03 6.45026863e-01 -6.62906945e-01
-1.67103624e+00 1.83793172e-01 1.16599284e-01 -2.12879688e-01
-3.88816148e-02 8.24200176e-03 -3.24574023e-01 -2.92372912e-01
6.01243734e-01 6.13003612e-01 -4.09551822e-02 -3.24316502e-01
5.41972220e-01 9.92883146e-01 1.04587400e+00 -1.81565255e-01
9.42445040e-01 3.10713351e-01 -3.77972960e-01 -7.31654227e-01
-5.06694496e-01 -1.44317478e-01 -5.40385365e-01 -5.92670500e-01
8.39744151e-01 -9.37207878e-01 -5.45991123e-01 8.53433311e-01
-9.26550925e-01 -1.46112397e-01 -1.73103586e-01 3.60270411e-01
-4.87006038e-01 6.28425717e-01 -7.98894882e-01 -6.80866778e-01
-3.54468703e-01 -1.39199984e+00 9.73764956e-01 5.38448036e-01
1.00786664e-01 -5.40000081e-01 -3.53741825e-01 3.66024673e-01
5.25088966e-01 -1.26190022e-01 6.52839780e-01 5.41998863e-01
-6.52774274e-01 -2.94515014e-01 -5.62871635e-01 7.44062662e-01
3.09392720e-01 4.23985794e-02 -1.03289282e+00 -2.08599478e-01
4.96153757e-02 -5.91370314e-02 3.72272044e-01 4.15315866e-01
1.42411816e+00 -4.02566910e-01 2.48944849e-01 8.92679691e-01
1.53907251e+00 4.19886768e-01 1.16070175e+00 2.94283539e-01
3.28120917e-01 4.57887471e-01 7.02205539e-01 4.41919923e-01
2.30323255e-01 7.21531987e-01 8.80621448e-02 -6.69051468e-01
-2.98114717e-01 -3.34854782e-01 5.35965621e-01 7.51542509e-01
-1.73805967e-01 -9.17378962e-02 -3.02112997e-01 4.56290990e-01
-1.12673497e+00 -1.03385317e+00 -1.53405026e-01 2.12759328e+00
1.31549525e+00 1.49288371e-01 -1.70509145e-01 4.01030719e-01
8.52702439e-01 -5.76591166e-03 -3.72165620e-01 -5.62535882e-01
-1.71529278e-01 1.06985994e-01 6.76165640e-01 5.94589829e-01
-8.67712617e-01 7.13841975e-01 6.48511267e+00 8.87141109e-01
-1.10001576e+00 -4.35182005e-02 8.38240504e-01 -4.12995517e-02
-1.93748549e-01 -2.15291232e-01 -4.43063647e-01 5.62297285e-01
6.38978064e-01 -2.21415773e-01 6.55489504e-01 6.03191912e-01
4.59269911e-01 -2.01385617e-01 -7.88794160e-01 1.32033813e+00
2.29755476e-01 -9.92074192e-01 6.34607580e-03 -1.26800179e-01
8.54488671e-01 -8.08297694e-01 4.38995153e-01 -1.06184639e-01
1.51096266e-02 -8.50618184e-01 8.53288531e-01 5.89290440e-01
1.24599433e+00 -6.82438552e-01 4.67584163e-01 -2.59307632e-03
-1.04283512e+00 -5.91844060e-02 -3.98221612e-01 1.76963866e-01
4.40564364e-01 5.45396686e-01 -4.63238269e-01 3.90681237e-01
7.72624254e-01 3.34525734e-01 -8.40518653e-01 1.13914096e+00
-2.80722827e-01 2.89868921e-01 9.11668912e-02 3.37236583e-01
-2.35452414e-01 -2.99935699e-01 2.42957383e-01 1.17749798e+00
6.56826973e-01 4.32880372e-01 -4.61303771e-01 8.71702969e-01
6.97540790e-02 -3.20083089e-02 -1.67266563e-01 -1.23276748e-01
1.86590001e-01 1.39581728e+00 -5.66857159e-01 -3.30343068e-01
-1.84309691e-01 1.54099560e+00 -2.01469749e-01 3.60165119e-01
-9.61807609e-01 -7.23602057e-01 5.71785092e-01 1.35464549e-01
2.28437468e-01 -5.82642108e-03 -3.80182475e-01 -1.10071385e+00
2.04932377e-01 -1.21831548e+00 5.84327430e-02 -1.43241084e+00
-8.56131256e-01 8.75499189e-01 -2.25760162e-01 -1.54500639e+00
-3.31304595e-02 -5.00772774e-01 -2.65823156e-01 7.96395898e-01
-1.62061167e+00 -9.24122572e-01 -6.76920295e-01 5.91870964e-01
6.42347753e-01 6.95349500e-02 4.61912274e-01 6.89597547e-01
-4.64690030e-01 9.38380539e-01 -2.92108338e-02 -1.03466727e-01
8.84910643e-01 -1.08238661e+00 3.12129647e-01 1.42081332e+00
-2.01441810e-01 3.12550396e-01 8.71500552e-01 -4.25917655e-01
-1.15538132e+00 -8.57675612e-01 8.80002737e-01 1.03989147e-01
1.42125383e-01 -2.20493600e-01 -7.66910434e-01 2.46186003e-01
4.65420902e-01 -2.11043775e-01 3.47245902e-01 -5.64649224e-01
-1.30676746e-01 -3.06319624e-01 -1.12539935e+00 7.56123126e-01
8.77975523e-01 -4.69861031e-01 -2.42660016e-01 -2.26303965e-01
4.64811444e-01 -4.93963838e-01 -7.00293481e-01 9.82634351e-02
6.38121367e-01 -1.13841283e+00 1.03949499e+00 6.68940321e-02
6.95248008e-01 -6.02935314e-01 1.50122806e-01 -1.29074931e+00
-4.71675098e-01 -8.99731994e-01 2.75058478e-01 1.20920277e+00
4.67258364e-01 -2.34385967e-01 4.23814803e-01 8.08680594e-01
7.44417310e-02 -3.26134324e-01 -9.82571989e-02 -5.92611909e-01
-6.33064926e-01 -3.27368259e-01 5.92337847e-01 7.10040092e-01
-6.12187646e-02 -2.30972748e-03 -9.46549177e-01 4.25766051e-01
8.55409205e-01 6.15679957e-02 6.92073822e-01 -6.28910720e-01
-4.81055260e-01 -3.61446172e-01 -3.16686213e-01 -1.16037571e+00
-4.17220503e-01 -3.76013994e-01 3.12892079e-01 -1.24007249e+00
3.70645702e-01 -2.15535283e-01 -1.18849330e-01 2.12868795e-01
-4.28342074e-01 5.50167978e-01 2.32952625e-01 -5.24476804e-02
-4.73679990e-01 4.63823348e-01 1.57897174e+00 -2.22508460e-01
-1.68327853e-01 -9.57613513e-02 -9.79811907e-01 6.45188808e-01
7.44731545e-01 -3.01895976e-01 -5.97093582e-01 -6.72747016e-01
2.97232479e-01 2.10171342e-01 3.28322232e-01 -9.88239586e-01
2.76138246e-01 -1.19554169e-01 5.40810108e-01 -3.12473267e-01
3.39700162e-01 -9.17292655e-01 4.13840324e-01 2.77526736e-01
-4.39636707e-01 1.31898612e-01 7.15645626e-02 4.05759633e-01
-5.14659405e-01 -2.44915187e-01 1.24043036e+00 1.46781042e-01
-7.50742137e-01 1.30717933e-01 -4.16272670e-01 -2.65790135e-01
9.28957343e-01 -5.45823574e-01 -9.57265347e-02 -7.20337033e-01
-5.07832766e-01 -1.97037429e-01 6.06980026e-01 3.52474004e-01
9.78049695e-01 -1.25009871e+00 -7.45170772e-01 4.68883455e-01
-8.51801038e-02 -4.26179767e-01 7.39188492e-01 6.02285028e-01
-7.24652886e-01 -5.87683544e-02 -5.41200578e-01 -3.26697350e-01
-1.58486557e+00 5.70578754e-01 4.61953253e-01 -9.98647362e-02
-4.76850003e-01 6.85904443e-01 2.61391640e-01 1.03910804e-01
1.39986381e-01 -3.25312167e-01 -1.85988411e-01 -2.09286988e-01
8.24003398e-01 4.03236032e-01 -1.33089926e-02 -4.88931924e-01
1.44303694e-01 7.41929471e-01 1.10377502e-02 -2.44555116e-01
1.22994578e+00 -6.64030254e-01 1.69005126e-01 -2.05546275e-01
1.16420388e+00 2.77204812e-02 -1.61314595e+00 -5.22547252e-02
-5.65307140e-01 -1.00279105e+00 3.46586138e-01 -9.56099272e-01
-1.29275453e+00 8.34689558e-01 9.80542660e-01 7.44950846e-02
1.93841243e+00 -5.74937403e-01 8.45983684e-01 -2.49929167e-02
2.03822762e-01 -1.29826653e+00 2.28541374e-01 -1.40186593e-01
9.03509796e-01 -1.10674059e+00 2.17944663e-02 -5.77704191e-01
-9.00761008e-01 9.60067511e-01 4.99869227e-01 2.36774683e-01
3.43794733e-01 4.19028848e-01 2.12963164e-01 1.09664746e-01
-5.02295494e-01 -6.21752664e-02 4.00901258e-01 6.60586417e-01
1.31396785e-01 -1.37179524e-01 -2.84887791e-01 5.70642173e-01
-3.17803442e-01 2.57056981e-01 7.95545816e-01 6.12597883e-01
-3.29954743e-01 -9.78284240e-01 -4.95659471e-01 1.43713459e-01
-5.90576291e-01 -1.61022604e-01 5.51843606e-02 4.04832989e-01
1.55750617e-01 1.11680555e+00 -2.00148389e-01 -5.21431088e-01
4.72051084e-01 -3.94223630e-01 5.05957067e-01 -3.06518711e-02
-3.25055331e-01 1.18861675e-01 -7.47562200e-02 -4.40801620e-01
-4.40930367e-01 -4.21036541e-01 -6.99415982e-01 -3.34173679e-01
-3.35489184e-01 -1.27908558e-01 6.99095249e-01 4.97417748e-01
3.54678303e-01 6.69167817e-01 7.59022236e-01 -7.28031516e-01
-4.06936109e-01 -6.75672412e-01 -7.74293423e-01 5.20083129e-01
2.28445917e-01 -2.19951570e-01 -3.27101052e-01 8.66395175e-01] | [11.006376266479492, -2.119178056716919] |
62e0cc2d-9749-4480-a6e1-f91afcf1eefb | self-supervised-contrastive-pre-training-for | 2206.08496 | null | https://arxiv.org/abs/2206.08496v3 | https://arxiv.org/pdf/2206.08496v3.pdf | Self-Supervised Contrastive Pre-Training For Time Series via Time-Frequency Consistency | Pre-training on time series poses a unique challenge due to the potential mismatch between pre-training and target domains, such as shifts in temporal dynamics, fast-evolving trends, and long-range and short-cyclic effects, which can lead to poor downstream performance. While domain adaptation methods can mitigate these shifts, most methods need examples directly from the target domain, making them suboptimal for pre-training. To address this challenge, methods need to accommodate target domains with different temporal dynamics and be capable of doing so without seeing any target examples during pre-training. Relative to other modalities, in time series, we expect that time-based and frequency-based representations of the same example are located close together in the time-frequency space. To this end, we posit that time-frequency consistency (TF-C) -- embedding a time-based neighborhood of an example close to its frequency-based neighborhood -- is desirable for pre-training. Motivated by TF-C, we define a decomposable pre-training model, where the self-supervised signal is provided by the distance between time and frequency components, each individually trained by contrastive estimation. We evaluate the new method on eight datasets, including electrodiagnostic testing, human activity recognition, mechanical fault detection, and physical status monitoring. Experiments against eight state-of-the-art methods show that TF-C outperforms baselines by 15.4% (F1 score) on average in one-to-one settings (e.g., fine-tuning an EEG-pretrained model on EMG data) and by 8.4% (precision) in challenging one-to-many settings (e.g., fine-tuning an EEG-pretrained model for either hand-gesture recognition or mechanical fault prediction), reflecting the breadth of scenarios that arise in real-world applications. Code and datasets: https://github.com/mims-harvard/TFC-pretraining. | ['Marinka Zitnik', 'Theodoros Tsiligkaridis', 'Ziyuan Zhao', 'Xiang Zhang'] | 2022-06-17 | null | null | null | null | ['hand-gesture-recognition', 'hand-gesture-recognition-1', 'gesture-recognition', 'fault-detection'] | ['computer-vision', 'computer-vision', 'computer-vision', 'miscellaneous'] | [ 5.26833057e-01 -2.46190071e-01 -3.51359904e-01 -3.73823613e-01
-9.24508035e-01 -5.41251361e-01 3.11839253e-01 -2.32260585e-01
-2.81652421e-01 5.43694854e-01 4.55040157e-01 -1.93873838e-01
-4.75576848e-01 -3.90661091e-01 -6.82294905e-01 -6.63768947e-01
-6.22608721e-01 1.42524019e-01 9.72942710e-02 -1.67095661e-01
8.21814090e-02 1.57514080e-01 -1.29103243e+00 5.06703317e-01
7.25533009e-01 9.16986465e-01 1.13846339e-01 3.61954838e-01
5.47514677e-01 1.67308271e-01 -5.09546816e-01 3.50923002e-01
8.82766172e-02 -6.19266450e-01 -6.91293240e-01 -6.41193315e-02
2.04178318e-01 -1.38146937e-01 -5.89279294e-01 8.29464316e-01
6.43782973e-01 1.79163337e-01 7.66288817e-01 -1.08622038e+00
-4.77798879e-01 4.14746016e-01 -4.26924855e-01 5.94178438e-01
3.65373224e-01 4.07387465e-01 7.20350981e-01 -7.75609434e-01
6.87718809e-01 8.34793568e-01 9.97045338e-01 6.50782645e-01
-1.53605223e+00 -6.71683192e-01 2.71838009e-01 3.54070276e-01
-1.07990384e+00 -1.84451580e-01 8.88372958e-01 -5.39475858e-01
1.05736208e+00 2.23474085e-01 7.19202578e-01 1.63953674e+00
5.09032190e-01 5.59746921e-01 1.27124524e+00 -2.42860075e-02
2.79544324e-01 -4.50051695e-01 1.49930835e-01 5.02819661e-03
-1.99323162e-01 3.52490008e-01 -8.89811635e-01 -2.04810753e-01
7.79062748e-01 1.69963285e-01 -5.42729318e-01 -9.02023092e-02
-1.38030636e+00 3.64844054e-01 2.81947643e-01 5.47203124e-01
-5.91249526e-01 5.14663793e-02 5.43841779e-01 6.09559894e-01
4.69440281e-01 5.32754004e-01 -7.87792683e-01 -6.58079743e-01
-8.46324861e-01 1.55065089e-01 4.55721140e-01 4.26803023e-01
3.21620584e-01 8.67945552e-02 -4.76034209e-02 8.84611726e-01
-1.95894361e-01 3.00865769e-01 1.11270082e+00 -7.53621757e-01
4.41998661e-01 2.53788382e-01 -2.57972479e-01 -6.99940443e-01
-5.77291131e-01 -7.18226254e-01 -7.44095981e-01 1.93004329e-02
5.58723450e-01 -1.87839374e-01 -1.05230820e+00 2.12316275e+00
2.90136695e-01 6.37371421e-01 -1.37977064e-01 9.76000667e-01
1.49717763e-01 4.10519332e-01 -9.84552577e-02 -4.73128676e-01
1.25497496e+00 -4.02666926e-01 -5.64707994e-01 -4.27728653e-01
5.67605317e-01 -3.84068131e-01 1.39616048e+00 5.17397642e-01
-7.88179338e-01 -5.56229949e-01 -9.03851628e-01 4.55024809e-01
-1.30273834e-01 -2.10102126e-01 2.33140528e-01 1.86501235e-01
-5.26732862e-01 1.17521906e+00 -1.26875615e+00 -3.46247315e-01
1.48370445e-01 3.57708901e-01 -5.17047942e-01 7.44158775e-02
-1.34842443e+00 8.25355351e-01 1.29439935e-01 4.03735749e-02
-7.93558121e-01 -1.15226758e+00 -5.20002425e-01 -2.30559960e-01
8.40975791e-02 -3.82814109e-01 1.06785154e+00 -8.31085026e-01
-1.14815199e+00 6.94405317e-01 -8.55489150e-02 -4.62176323e-01
4.95373756e-01 -4.50015038e-01 -8.96835506e-01 1.57042921e-01
1.11473396e-01 2.07566738e-01 1.06358099e+00 -7.00220823e-01
-3.52602988e-01 -4.71928090e-01 -3.08692724e-01 1.98349327e-01
-5.71137071e-01 -3.20696473e-01 -1.62184715e-01 -1.05210686e+00
2.56119072e-01 -1.06890202e+00 6.73808306e-02 -1.93707243e-01
-1.77534178e-01 -5.93169257e-02 8.87972593e-01 -7.66144276e-01
1.46118069e+00 -2.48619938e+00 1.08626328e-01 3.00322622e-01
6.92487359e-02 1.89850051e-02 -2.59102017e-01 5.10641932e-01
-6.30793393e-01 -2.19245970e-01 -2.50327855e-01 7.99229369e-02
-1.55998722e-01 1.92556188e-01 -2.86630273e-01 7.45943129e-01
3.82472277e-01 5.97636163e-01 -1.06310999e+00 1.23848794e-02
1.32334471e-01 4.93173599e-01 -4.71032172e-01 4.75410558e-02
-2.53363177e-02 7.59249210e-01 -2.33626515e-01 5.30392170e-01
1.52835995e-01 -3.24534982e-01 2.96075702e-01 -3.83545578e-01
3.20066988e-01 6.01036668e-01 -1.11717427e+00 1.81190002e+00
-3.42104495e-01 6.06398106e-01 -2.82557189e-01 -1.30440307e+00
6.16704166e-01 4.06049281e-01 9.79589522e-01 -8.24831545e-01
-7.04571232e-02 5.28358698e-01 4.67395127e-01 -6.68531060e-01
-1.59396842e-01 -2.95008540e-01 1.18986238e-02 5.66052794e-01
1.47468925e-01 1.59198835e-01 -9.40692425e-02 -2.65827209e-01
1.54671574e+00 1.61733210e-01 4.79178242e-02 -2.36561075e-01
-1.09813578e-01 -7.46829584e-02 7.66704798e-01 3.57731253e-01
-2.70323753e-01 7.27312803e-01 4.01783377e-01 -1.99002281e-01
-7.15306997e-01 -1.24044859e+00 -3.75026882e-01 9.77919757e-01
-1.43598365e-02 -5.02738297e-01 -5.70167243e-01 -5.15811741e-01
2.83883631e-01 5.60339630e-01 -8.53205264e-01 -7.54203975e-01
-7.92246103e-01 -5.47253013e-01 4.54672247e-01 8.03911150e-01
2.77656168e-02 -9.64662850e-01 -7.80621767e-01 5.25980771e-01
-2.31473982e-01 -9.30285633e-01 -7.17929244e-01 6.23088181e-01
-1.29955614e+00 -9.30320024e-01 -7.78793395e-01 -5.26969552e-01
4.31708097e-01 -8.69228914e-02 9.27961826e-01 -3.04261863e-01
-2.89202958e-01 5.00046015e-01 -2.60723263e-01 -1.91565320e-01
-8.91170278e-02 -9.37939584e-02 4.53214526e-01 -1.79960672e-02
3.66432965e-01 -1.13867199e+00 -9.04381275e-01 4.61441606e-01
-7.31719375e-01 -2.46851265e-01 5.94637692e-01 1.26830292e+00
7.02899635e-01 -7.42519125e-02 8.37362051e-01 -6.30259275e-01
7.58713126e-01 -7.27307916e-01 1.08576491e-01 2.68411823e-02
-8.89175177e-01 -1.20316461e-01 6.12070024e-01 -1.26088929e+00
-4.53498065e-01 -9.28432718e-02 1.35303512e-01 -8.09893012e-01
-4.76389304e-02 8.79060984e-01 2.27982402e-01 3.54546428e-01
1.09824574e+00 3.63545537e-01 2.57940710e-01 -4.38013911e-01
4.28353250e-02 5.62973022e-01 8.98738503e-01 -7.15175867e-01
7.17639446e-01 2.69162148e-01 -2.08892688e-01 -5.38215995e-01
-5.76259017e-01 -4.75030214e-01 -5.65172791e-01 -3.28774184e-01
5.25066495e-01 -8.25144351e-01 -2.60126948e-01 4.87776667e-01
-7.05841064e-01 -6.85816526e-01 -3.70488018e-01 8.85195851e-01
-6.63419724e-01 1.03517152e-01 -5.74686944e-01 -5.44153333e-01
-2.33387321e-01 -7.65778780e-01 9.38159764e-01 -1.35297894e-01
-7.65424669e-01 -9.44692075e-01 2.02036127e-01 3.21289245e-03
4.08806950e-01 5.07311463e-01 9.37798381e-01 -7.33583987e-01
1.99544072e-01 -1.43945664e-01 3.22412759e-01 5.38392305e-01
5.45120835e-01 -5.11933506e-01 -9.95373845e-01 -4.57160294e-01
3.02279651e-01 -2.22216398e-01 6.09836459e-01 4.79998946e-01
1.19877672e+00 -7.45598376e-02 -4.24993128e-01 4.85754937e-01
1.06982112e+00 3.18436772e-01 5.65332174e-01 2.75653809e-01
3.88209522e-01 5.68120003e-01 7.81256497e-01 3.83920491e-01
1.87788047e-02 6.77499950e-01 4.04051766e-02 9.09407213e-02
-1.30002588e-01 -2.69749731e-01 4.91051167e-01 8.91102135e-01
-1.14711374e-01 2.67104447e-01 -1.07623994e+00 7.66293705e-01
-1.88419652e+00 -9.62214828e-01 1.64019436e-01 2.25519323e+00
1.13894308e+00 4.32292134e-01 4.07689601e-01 4.44935977e-01
5.54653168e-01 -6.52316511e-02 -1.14943898e+00 1.63060248e-01
1.55812457e-01 3.93902272e-01 5.25778644e-02 4.47794423e-02
-9.78709400e-01 3.28261495e-01 5.89087152e+00 7.39437580e-01
-1.71633255e+00 1.90157548e-01 4.92003411e-01 -4.25520331e-01
-3.85188945e-02 -3.02081913e-01 -2.04265922e-01 7.68365026e-01
1.27433074e+00 -3.70747745e-01 5.90084374e-01 5.45888007e-01
4.09281015e-01 1.93587303e-01 -1.49741077e+00 1.10837030e+00
-3.34373116e-01 -1.06163621e+00 -5.81636965e-01 7.37288669e-02
4.78195578e-01 2.14009553e-01 1.65649831e-01 4.70008165e-01
-4.08280730e-01 -1.05952811e+00 7.19295681e-01 3.21114302e-01
1.03915918e+00 -2.10488394e-01 4.69725579e-01 2.86201894e-01
-1.28766632e+00 -1.62409738e-01 1.26762375e-01 2.66426522e-02
7.71405101e-02 6.31672919e-01 -5.18084049e-01 4.65574503e-01
9.00302112e-01 1.09603870e+00 -1.49521500e-01 8.33106101e-01
1.38560355e-01 9.87795234e-01 -4.40170288e-01 2.21800655e-01
-5.53437322e-02 -1.21467495e-02 5.95378160e-01 1.10806513e+00
4.31156158e-01 2.14183271e-01 -1.78407934e-02 6.56794906e-01
3.15588117e-01 -3.52691352e-01 -5.86086512e-01 -1.77732468e-01
5.96336663e-01 7.81894982e-01 -4.02552336e-01 -1.00388616e-01
-3.97724390e-01 9.38432097e-01 3.52368727e-02 5.17268836e-01
-1.02051616e+00 -4.63771164e-01 8.77593219e-01 2.85829663e-01
1.14319846e-01 -1.90734103e-01 -3.32007527e-01 -1.05966687e+00
4.78764713e-01 -1.23406613e+00 6.20311260e-01 -6.49612665e-01
-1.60446751e+00 5.44040740e-01 1.56778559e-01 -1.89857340e+00
-6.42849028e-01 -5.94377637e-01 -6.05269074e-01 8.33090067e-01
-9.97660995e-01 -6.95993245e-01 -1.51815057e-01 8.19697857e-01
5.39403975e-01 8.19285363e-02 8.87512445e-01 4.03990597e-01
-2.61425853e-01 5.68370163e-01 3.36495601e-02 2.75645871e-02
1.00117612e+00 -1.16409755e+00 1.94298625e-01 6.81362689e-01
1.27122551e-01 8.22741389e-01 7.06436217e-01 -6.73394144e-01
-1.61436749e+00 -1.01876163e+00 5.07373750e-01 -4.18762118e-01
1.11506832e+00 -2.98758119e-01 -1.35053623e+00 4.73233581e-01
-2.67632514e-01 2.23830715e-01 6.16444349e-01 3.59340698e-01
-4.85598326e-01 -2.88386315e-01 -8.92660677e-01 5.52229285e-01
1.45457399e+00 -9.97495592e-01 -8.88973713e-01 4.48904216e-01
3.44593585e-01 -6.38226390e-01 -1.36946726e+00 6.46341503e-01
8.41838360e-01 -5.56047261e-01 9.46740389e-01 -7.42477179e-01
4.53745157e-01 -1.87131807e-01 -2.07958743e-01 -1.62475133e+00
-2.60527164e-01 -7.31886625e-01 -4.39599872e-01 9.20252860e-01
3.67189944e-01 -7.99512565e-01 5.65506458e-01 4.49020296e-01
-3.56317818e-01 -1.11641300e+00 -1.03151906e+00 -1.34559357e+00
1.74678653e-01 -7.13311136e-01 3.17402631e-01 1.10931063e+00
5.02474308e-01 1.86265901e-01 -2.45280415e-01 2.76243538e-02
2.30308533e-01 4.12200391e-02 2.72660106e-01 -1.07199824e+00
-6.21487737e-01 -4.44404155e-01 -4.67311054e-01 -8.52338850e-01
-2.71671768e-02 -6.64079607e-01 1.36157215e-01 -1.21214128e+00
-1.19543420e-02 -3.54631424e-01 -8.49199533e-01 7.06152856e-01
-2.05552801e-01 5.47338761e-02 -1.07027911e-01 3.90907377e-01
-2.01106574e-02 3.78688246e-01 1.05935061e+00 -1.33425757e-01
-3.81954491e-01 -2.23473795e-02 -4.17142153e-01 4.69624490e-01
7.77199268e-01 -5.98892987e-01 -7.43643999e-01 -1.81347072e-01
-1.83645397e-01 2.46551171e-01 4.25235301e-01 -1.18088973e+00
4.52566445e-02 -1.74740031e-01 4.19268698e-01 -3.21797878e-01
4.61398423e-01 -6.86225295e-01 3.62513661e-01 6.92962825e-01
-3.39827657e-01 1.92707747e-01 3.92525107e-01 7.49358058e-01
-3.11207265e-01 2.58431256e-01 5.68017900e-01 2.69603223e-01
-6.39323711e-01 1.99318752e-01 -3.74659091e-01 3.52873236e-01
7.42423177e-01 -4.28421468e-01 -5.04182220e-01 -3.02113682e-01
-9.48208630e-01 5.65867648e-02 1.32290557e-01 6.58786952e-01
5.53189337e-01 -1.38842440e+00 -4.39823061e-01 2.41047025e-01
3.08223188e-01 -3.67770493e-01 4.26683962e-01 1.37143672e+00
2.81062394e-01 8.99960101e-02 -3.29974949e-01 -9.13520515e-01
-1.08283329e+00 1.49626821e-01 4.24429476e-01 -1.75102502e-01
-9.57387686e-01 6.17862642e-01 1.64517671e-01 -3.33665997e-01
3.39776605e-01 -6.27167225e-01 2.17859894e-01 -1.63043171e-01
3.44040453e-01 1.43938631e-01 3.61800849e-01 -2.12349206e-01
-7.52619326e-01 6.42192900e-01 1.32730663e-01 -2.23881677e-01
1.50103450e+00 2.74920762e-01 3.07116419e-01 9.54941988e-01
1.35421872e+00 -3.68912697e-01 -1.41274500e+00 -3.22823554e-01
7.80538321e-02 -2.44069114e-01 -1.13190457e-01 -1.01970959e+00
-9.27825451e-01 7.66693115e-01 1.04346800e+00 1.05296098e-01
1.54492664e+00 -1.64505187e-02 7.53467560e-01 2.06750318e-01
4.50685352e-01 -1.23697567e+00 4.33725744e-01 3.08804572e-01
1.02315378e+00 -8.89569521e-01 -1.43321797e-01 -1.15870327e-01
-6.85424626e-01 1.08633435e+00 5.63904524e-01 -2.38753617e-01
9.95684206e-01 2.56112337e-01 5.37271462e-02 -2.03503728e-01
-1.04558921e+00 1.44227326e-01 7.02326655e-01 7.06891119e-01
4.81754005e-01 1.74458712e-01 -1.93172336e-01 8.28465581e-01
-1.47490874e-01 1.34537503e-01 7.34147616e-03 1.06024289e+00
1.46150246e-01 -9.42130268e-01 -2.90382206e-01 8.89955521e-01
-3.68048549e-01 1.79968834e-01 -6.89147413e-02 9.24323440e-01
-5.32516688e-02 7.83485472e-01 1.88799113e-01 -9.39658344e-01
4.96387511e-01 2.33608857e-01 5.08062422e-01 -5.89102805e-01
-6.21521354e-01 3.91844571e-01 1.82397991e-01 -8.58643413e-01
-2.52384752e-01 -9.55050647e-01 -1.46975935e+00 1.91350747e-02
-1.36766076e-01 -1.14812583e-01 3.67567599e-01 8.15764248e-01
5.55194318e-01 7.02990472e-01 5.78967452e-01 -7.55979896e-01
-8.54933381e-01 -1.21152210e+00 -6.84923887e-01 8.02414536e-01
4.47750062e-01 -8.47743094e-01 -5.48974097e-01 1.68140739e-01] | [7.492997169494629, 3.023341417312622] |
ad0ae3fb-b126-4208-a573-97aa28615655 | visual-instruction-tuning | 2304.08485 | null | https://arxiv.org/abs/2304.08485v1 | https://arxiv.org/pdf/2304.08485v1.pdf | Visual Instruction Tuning | Instruction tuning large language models (LLMs) using machine-generated instruction-following data has improved zero-shot capabilities on new tasks, but the idea is less explored in the multimodal field. In this paper, we present the first attempt to use language-only GPT-4 to generate multimodal language-image instruction-following data. By instruction tuning on such generated data, we introduce LLaVA: Large Language and Vision Assistant, an end-to-end trained large multimodal model that connects a vision encoder and LLM for general-purpose visual and language understanding.Our early experiments show that LLaVA demonstrates impressive multimodel chat abilities, sometimes exhibiting the behaviors of multimodal GPT-4 on unseen images/instructions, and yields a 85.1% relative score compared with GPT-4 on a synthetic multimodal instruction-following dataset. When fine-tuned on Science QA, the synergy of LLaVA and GPT-4 achieves a new state-of-the-art accuracy of 92.53%. We make GPT-4 generated visual instruction tuning data, our model and code base publicly available. | ['Yong Jae Lee', 'Qingyang Wu', 'Chunyuan Li', 'Haotian Liu'] | 2023-04-17 | null | null | null | null | ['science-question-answering', 'instruction-following'] | ['miscellaneous', 'natural-language-processing'] | [ 2.07523674e-01 2.56349206e-01 -2.47546166e-01 -3.59519422e-01
-1.38388753e+00 -5.24301708e-01 7.80106664e-01 -1.47152603e-01
-3.75164390e-01 3.37134331e-01 9.07199755e-02 -7.91092515e-01
5.66600561e-01 -3.31463426e-01 -1.35654652e+00 -4.45697844e-01
2.46583149e-01 1.03342736e+00 -3.52516435e-02 -3.63081068e-01
3.71760726e-01 -1.55154288e-01 -1.63368857e+00 9.37147439e-01
9.41653550e-01 5.92397451e-01 4.90719855e-01 1.35102344e+00
-4.45383161e-01 1.11959112e+00 -5.26016355e-01 -5.06347001e-01
-2.22534344e-01 -3.30827713e-01 -7.09445715e-01 -1.22219011e-01
1.16350627e+00 -4.51241702e-01 -2.85631835e-01 6.36944473e-01
6.39838457e-01 -1.07251234e-01 6.46639705e-01 -1.48291314e+00
-1.16072428e+00 5.91155231e-01 -7.25864053e-01 -2.11688474e-01
6.29288554e-01 8.80083084e-01 6.44548416e-01 -1.03256392e+00
7.15548098e-01 1.92533648e+00 3.77142817e-01 9.01960969e-01
-1.40083778e+00 -6.33952856e-01 -3.32827866e-01 -5.81251606e-02
-1.04138350e+00 -5.12665033e-01 1.09585486e-01 -4.19857979e-01
1.59273148e+00 -7.77012110e-02 -3.83728668e-02 1.66599953e+00
6.17759049e-01 1.12057459e+00 1.09438837e+00 -7.54647911e-01
-6.36691824e-02 6.87822402e-02 2.18703359e-01 1.39213812e+00
-3.17791015e-01 1.67784795e-01 -7.28138506e-01 1.61256120e-01
4.76395726e-01 -5.05498111e-01 -7.82181695e-02 -4.69057202e-01
-1.58877838e+00 9.76473212e-01 3.82978320e-01 -1.94674179e-01
1.74504995e-01 5.69023073e-01 6.42444909e-01 5.03775477e-01
-2.33732685e-02 4.46942180e-01 -2.54058510e-01 -4.24170464e-01
-7.40970254e-01 -4.45521474e-02 6.89366698e-01 1.09607303e+00
7.86963999e-01 3.07892114e-01 -4.46964979e-01 6.29488647e-01
3.80210489e-01 8.67047369e-01 6.65324569e-01 -1.09688413e+00
9.25604105e-01 4.81883079e-01 -2.22303808e-01 -3.88758212e-01
-3.81420672e-01 1.17314532e-01 -3.08823347e-01 4.81754392e-01
4.97591496e-01 -7.53883198e-02 -1.26998889e+00 1.60075712e+00
-1.24578774e-01 -3.32022607e-01 4.64838266e-01 6.75222278e-01
1.47540665e+00 1.15466177e+00 5.90314150e-01 4.55984622e-01
1.42547178e+00 -1.47637963e+00 -3.85042310e-01 -6.07548535e-01
9.85911608e-01 -9.84848619e-01 1.92055035e+00 3.13958615e-01
-1.11280787e+00 -1.02846563e+00 -9.76663411e-01 -5.07370412e-01
-4.64523017e-01 4.68804866e-01 5.04423201e-01 5.21534503e-01
-1.44178689e+00 -1.74937308e-01 -6.64392412e-01 -7.17438936e-01
2.09160537e-01 2.27137625e-01 -6.06555581e-01 -3.54082584e-01
-6.24494076e-01 9.87729907e-01 4.50450480e-01 -4.06480372e-01
-1.59428144e+00 -6.21994615e-01 -1.26268721e+00 -2.85559869e-03
3.34515780e-01 -9.50608850e-01 1.72816598e+00 -9.15623128e-01
-1.48448133e+00 1.18366957e+00 -3.65481913e-01 -3.86110336e-01
2.87132174e-01 -3.81381959e-02 -2.93016285e-01 1.50259212e-01
1.37503609e-01 1.61592877e+00 1.00533485e+00 -1.64905465e+00
-4.00448889e-01 -1.00532331e-01 1.29812323e-02 2.05740899e-01
4.67180158e-04 -1.95172071e-01 -6.58195257e-01 -5.55517226e-02
-8.03880930e-01 -1.03739166e+00 1.33320749e-01 -3.11984330e-01
-4.50257361e-01 -2.55904734e-01 8.64201665e-01 -5.41210473e-01
7.55607963e-01 -2.03691792e+00 2.51809478e-01 -3.14755797e-01
1.60376683e-01 3.08837324e-01 -9.57710803e-01 6.27917886e-01
3.88758592e-02 -8.38258937e-02 -1.18917905e-01 -5.95980585e-01
4.46416080e-01 3.04751992e-01 -4.93235469e-01 -1.95055202e-01
3.56123924e-01 1.57176971e+00 -8.70458305e-01 -6.32984877e-01
4.32888746e-01 5.45716465e-01 -5.04730165e-01 4.33524698e-01
-6.40019536e-01 2.87579030e-01 -1.45651102e-01 8.01029980e-01
2.12668642e-01 -3.73505145e-01 -2.15767860e-01 -2.80710131e-01
-2.04536337e-02 -2.35890865e-01 -2.03578010e-01 2.17940283e+00
-7.71928549e-01 1.16504014e+00 6.23833574e-02 -5.06053329e-01
6.90501392e-01 1.32719800e-01 -3.21523041e-01 -1.09618640e+00
2.90390477e-02 1.92034885e-01 -1.45690143e-01 -8.84741485e-01
6.82104766e-01 1.88027486e-01 -4.90014762e-01 3.73666614e-01
6.00140393e-01 -4.91322428e-01 2.98774391e-01 6.41351640e-01
8.70708883e-01 4.67806846e-01 -7.25760981e-02 -4.88170870e-02
3.32162827e-01 5.28661072e-01 -5.51498055e-01 9.17085767e-01
-1.47245884e-01 4.86131579e-01 3.27681065e-01 -2.78329849e-01
-1.01856124e+00 -1.12455773e+00 1.40689820e-01 1.86078107e+00
-3.52076679e-01 -4.82025832e-01 -1.06419325e+00 -7.35191226e-01
5.14161400e-02 1.29421365e+00 -6.74250007e-01 -2.77237147e-01
-3.74228209e-01 -4.81831998e-01 8.25346410e-01 3.13670754e-01
4.65919316e-01 -1.17552447e+00 -6.70758545e-01 -9.58599821e-02
-2.73617268e-01 -1.23602223e+00 -6.61703169e-01 2.24988341e-01
-5.27381480e-01 -8.94840479e-01 -6.72178388e-01 -1.13408005e+00
5.21922648e-01 9.21798721e-02 1.85633171e+00 -1.96695253e-01
-3.20612192e-01 1.06115031e+00 -4.10482883e-02 -2.89817750e-01
-1.09196627e+00 -2.46207928e-03 -3.94138783e-01 -5.39623678e-01
5.75027943e-01 2.15298429e-01 -2.44859874e-01 1.18839130e-01
-8.29280078e-01 4.82177675e-01 9.62138593e-01 1.02183366e+00
3.47350866e-01 -8.75536680e-01 3.35220605e-01 -6.50260806e-01
8.85749221e-01 -2.11743027e-01 -5.20842075e-01 5.90088546e-01
-3.47204119e-01 4.66188341e-01 3.92286241e-01 -5.47809362e-01
-1.19289601e+00 1.39063284e-01 -8.12439099e-02 -5.48990786e-01
-3.88301998e-01 2.70777971e-01 3.96226086e-02 -3.37004602e-01
7.99789727e-01 3.12824339e-01 -2.56193671e-02 -4.02886141e-03
1.10133994e+00 6.30627096e-01 1.08715630e+00 -8.00537109e-01
2.40956530e-01 -7.30251670e-02 -1.12722993e-01 -6.96586490e-01
-4.82439250e-01 1.69369616e-02 -4.38364983e-01 -7.69634470e-02
1.30323482e+00 -1.11812508e+00 -1.03352880e+00 1.78292006e-01
-1.06649804e+00 -9.98673081e-01 4.39516865e-02 1.12530231e-01
-9.84468222e-01 2.15889588e-01 -8.48590016e-01 -3.46773177e-01
-5.86729109e-01 -1.98748100e+00 1.86017346e+00 3.08374912e-01
-3.11106831e-01 -1.05432940e+00 2.89090842e-01 8.93505454e-01
3.98483068e-01 1.95588097e-02 1.18539548e+00 -2.88505852e-01
-6.11007094e-01 1.17380701e-01 -5.50297976e-01 -7.78870881e-02
-4.43695337e-01 2.09502801e-01 -1.24426055e+00 -4.19322848e-01
-6.31485164e-01 -1.27712405e+00 1.00498831e+00 2.47217849e-01
9.12284732e-01 9.50766131e-02 -2.74121374e-01 6.42970920e-01
1.33687747e+00 1.03891879e-01 4.80525047e-01 2.05425009e-01
1.13896346e+00 4.77568328e-01 5.16550958e-01 -6.99924752e-02
7.51474082e-01 4.89033192e-01 5.65271974e-01 -7.06546307e-02
-5.69930196e-01 -6.22118175e-01 9.26918924e-01 7.33265102e-01
4.48534280e-01 -3.56449425e-01 -1.26501751e+00 4.04766053e-01
-1.61770666e+00 -6.01529419e-01 -1.16264544e-01 1.99381351e+00
8.91244829e-01 -1.19753657e-02 -1.79355606e-01 -8.20129693e-01
1.19620614e-01 -9.47342068e-02 -5.36321759e-01 -1.03007352e+00
-2.14328662e-01 1.68893963e-01 3.90512258e-01 8.31642985e-01
-9.94342864e-01 1.32251012e+00 6.58800554e+00 1.09152162e+00
-8.85290682e-01 1.14988275e-01 7.11093485e-01 -1.25631303e-01
-3.00207734e-01 -2.05873474e-01 -9.39792275e-01 1.91857323e-01
1.62198162e+00 1.34403184e-01 5.30678213e-01 6.85665488e-01
-9.08166170e-02 -3.39804620e-01 -1.39339221e+00 1.23980844e+00
4.89974231e-01 -1.23576617e+00 5.78864872e-01 -1.00455642e-01
8.38833809e-01 3.18380862e-01 5.01450419e-01 1.16006315e+00
6.70456886e-01 -1.47743034e+00 6.04095280e-01 4.12905216e-01
1.16919184e+00 -7.02164173e-01 4.32599008e-01 2.86638051e-01
-8.93930674e-01 -3.12145730e-03 -1.34444922e-01 3.62229198e-01
5.35778739e-02 -3.40026170e-01 -1.17724180e+00 1.00206107e-01
6.24253154e-01 3.19389254e-01 -1.14534080e+00 4.80976284e-01
-1.47220001e-01 5.27098954e-01 1.94434866e-01 4.51967455e-02
6.61771238e-01 1.27503216e-01 1.65815309e-01 1.62692976e+00
2.99380600e-01 -5.28513074e-01 2.35745087e-01 8.81264389e-01
-1.88327432e-01 -3.87054794e-02 -1.01314628e+00 -3.92260045e-01
8.33718702e-02 1.36784434e+00 -1.66894287e-01 -7.34059036e-01
-6.55938745e-01 1.23919749e+00 5.88949800e-01 5.31066418e-01
-9.43930387e-01 -2.74372846e-01 4.25092250e-01 -3.18286866e-01
1.07848480e-01 -1.88020051e-01 -5.04536517e-02 -1.10532820e+00
-6.67651117e-01 -1.46271133e+00 2.15298489e-01 -1.63098204e+00
-1.13672435e+00 6.37982070e-01 4.96360622e-02 -8.39345515e-01
-6.95137620e-01 -1.13841987e+00 -4.94646817e-01 6.92786455e-01
-1.19315624e+00 -1.69883549e+00 -4.73456502e-01 6.96812570e-01
1.02430511e+00 -4.07762378e-01 1.03565335e+00 -4.18821797e-02
-5.18846929e-01 9.64737356e-01 6.51325658e-02 -8.83765519e-02
1.25101697e+00 -1.46982670e+00 5.74071527e-01 4.63092595e-01
1.45171285e-01 4.83800888e-01 8.22652996e-01 -5.89510202e-01
-1.98625255e+00 -9.27835822e-01 5.98996699e-01 -1.23698783e+00
7.96507120e-01 -4.55217421e-01 -7.88468301e-01 1.05801809e+00
1.23881948e+00 -4.34642404e-01 5.81709683e-01 -3.66167516e-01
-5.22708118e-01 3.15443397e-01 -8.99514675e-01 8.73143315e-01
6.07491910e-01 -8.54310155e-01 -6.84230387e-01 3.62753034e-01
9.37587976e-01 -5.91425121e-01 -8.89926374e-01 1.47280246e-01
4.26590502e-01 -8.54690552e-01 1.18096828e+00 -6.09004200e-01
7.55048692e-01 -1.74090847e-01 -3.69908780e-01 -1.52773654e+00
9.51267183e-02 -5.08464634e-01 -1.58197060e-01 1.01088715e+00
5.51387787e-01 -2.64072806e-01 4.23793465e-01 3.97164941e-01
-5.97627103e-01 -2.39276886e-01 -6.71657503e-01 -3.89100522e-01
4.11312789e-01 -3.64101052e-01 1.68010980e-01 5.89823723e-01
5.06626740e-02 9.41555083e-01 -3.05010319e-01 -1.52167320e-01
7.73048460e-01 8.37860331e-02 1.30154204e+00 -5.23899555e-01
-4.70372319e-01 -5.21676898e-01 9.24909264e-02 -1.15457487e+00
3.55534792e-01 -9.65644896e-01 2.91620702e-01 -1.45806289e+00
5.56167901e-01 3.37020725e-01 1.35577664e-01 7.38481700e-01
-2.56441146e-01 2.86822706e-01 2.84550101e-01 1.11434013e-02
-1.01083791e+00 5.72328031e-01 1.43166637e+00 -5.70197761e-01
-2.07097288e-02 -8.52687240e-01 -4.80716556e-01 4.11231577e-01
5.62452912e-01 1.36363745e-01 -5.69851995e-01 -9.05667603e-01
2.13557854e-02 1.07782669e-01 3.78579378e-01 -9.22922432e-01
-4.84280195e-03 6.51883632e-02 4.29633707e-01 -6.02422416e-01
4.89309341e-01 -4.13364470e-01 -2.86002278e-01 3.88430178e-01
-5.41182101e-01 4.66983914e-01 8.65361333e-01 2.74454027e-01
-2.02954039e-01 -9.36576873e-02 4.63675618e-01 -2.24118993e-01
-1.50875735e+00 -2.58740604e-01 -4.91807729e-01 -1.03951769e-03
8.79540861e-01 2.17037991e-01 -1.24180245e+00 -6.09464884e-01
-4.23260629e-01 5.83274722e-01 7.35235989e-01 8.39735746e-01
8.75706494e-01 -1.24809539e+00 -7.44564414e-01 3.42515945e-01
8.01201105e-01 -5.42265177e-01 3.32675636e-01 6.85976744e-01
-7.13887990e-01 8.28761041e-01 -3.52545470e-01 -1.16655195e+00
-1.39814901e+00 8.44996870e-01 2.21677184e-01 -1.28189161e-01
-1.77448735e-01 9.29610908e-01 7.33764350e-01 -9.88086224e-01
1.73553094e-01 -4.93642509e-01 9.65892524e-02 -2.58046120e-01
6.20957494e-01 -4.76465710e-02 -1.89380899e-01 -8.14396143e-01
-1.70576014e-02 7.47969985e-01 5.32023096e-03 -2.85297960e-01
6.12553537e-01 -1.07512206e-01 1.58816695e-01 6.75269783e-01
1.33025110e+00 -2.37560526e-01 -1.40122402e+00 1.34548157e-01
-3.06992531e-01 -6.52438253e-02 -9.57864523e-03 -1.20884001e+00
-6.10237241e-01 1.54698861e+00 9.91410613e-01 -2.83343673e-01
8.33545864e-01 1.73361391e-01 7.01323271e-01 6.74073577e-01
2.64812082e-01 -6.49658501e-01 5.19686460e-01 7.56636083e-01
9.48731661e-01 -2.02399325e+00 -6.80372775e-01 2.51027822e-01
-1.07734740e+00 1.07151306e+00 1.06106412e+00 2.96861440e-01
-2.78411448e-01 2.42964000e-01 4.09507573e-01 -2.18770057e-01
-1.19978058e+00 -1.73929900e-01 5.15348077e-01 7.77229965e-01
7.43732870e-01 1.82605237e-01 5.27263045e-01 1.06568299e-01
-2.76389033e-01 7.98095539e-02 3.27267855e-01 8.16606939e-01
-5.21334887e-01 -8.57556045e-01 -6.38974309e-01 1.19168796e-01
-8.08493886e-03 -5.13197243e-01 -3.76527667e-01 1.04095924e+00
-2.16919705e-01 1.08017206e+00 1.59677297e-01 -4.59326774e-01
2.96934009e-01 4.52222914e-01 8.61039162e-01 -5.45885026e-01
-6.63350940e-01 1.73846039e-03 3.23652059e-01 -8.15654695e-01
1.27318457e-01 -1.61137402e-01 -1.41395426e+00 -4.09551471e-01
4.33529824e-01 -1.48437575e-01 7.26563692e-01 5.31944215e-01
5.72398841e-01 7.18637705e-01 -2.17488542e-01 -1.15093756e+00
-3.66337985e-01 -1.10788834e+00 2.87099928e-01 4.47163790e-01
5.19044042e-01 -3.73917043e-01 2.06981488e-02 2.44218379e-01] | [10.961844444274902, 1.5863548517227173] |
7da4ccac-80f6-44ab-86d9-064f288bdc75 | learning-difference-equations-with-structured | 2307.01238 | null | https://arxiv.org/abs/2307.01238v1 | https://arxiv.org/pdf/2307.01238v1.pdf | Learning Difference Equations with Structured Grammatical Evolution for Postprandial Glycaemia Prediction | People with diabetes must carefully monitor their blood glucose levels, especially after eating. Blood glucose regulation requires a proper combination of food intake and insulin boluses. Glucose prediction is vital to avoid dangerous post-meal complications in treating individuals with diabetes. Although traditional methods, such as artificial neural networks, have shown high accuracy rates, sometimes they are not suitable for developing personalised treatments by physicians due to their lack of interpretability. In this study, we propose a novel glucose prediction method emphasising interpretability: Interpretable Sparse Identification by Grammatical Evolution. Combined with a previous clustering stage, our approach provides finite difference equations to predict postprandial glucose levels up to two hours after meals. We divide the dataset into four-hour segments and perform clustering based on blood glucose values for the twohour window before the meal. Prediction models are trained for each cluster for the two-hour windows after meals, allowing predictions in 15-minute steps, yielding up to eight predictions at different time horizons. Prediction safety was evaluated based on Parkes Error Grid regions. Our technique produces safe predictions through explainable expressions, avoiding zones D (0.2% average) and E (0%) and reducing predictions on zone C (6.2%). In addition, our proposal has slightly better accuracy than other techniques, including sparse identification of non-linear dynamics and artificial neural networks. The results demonstrate that our proposal provides interpretable solutions without sacrificing prediction accuracy, offering a promising approach to glucose prediction in diabetes management that balances accuracy, interpretability, and computational efficiency. | ['J. Ignacio Hidalgo', 'Gabriel Kronberger', 'J. Manuel Velasco', 'David Joedicke', 'Daniel Parra'] | 2023-07-03 | null | null | null | null | ['clustering'] | ['methodology'] | [ 2.44072363e-01 4.75794315e-01 -4.20235962e-01 -5.82810760e-01
-2.39600003e-01 -2.17410251e-01 -1.01567216e-01 8.54912519e-01
-8.53947923e-02 8.99587035e-01 3.43832821e-01 -3.01683009e-01
-4.91646618e-01 -7.54469395e-01 -3.68303657e-01 -6.84294045e-01
-5.79418123e-01 8.08731556e-01 -4.89192277e-01 2.77527254e-02
-2.67529115e-02 2.34323576e-01 -1.36268103e+00 4.59530622e-01
1.44231820e+00 1.33987856e+00 -4.30157751e-01 5.54071784e-01
-4.43566069e-02 4.82640415e-01 -4.68749315e-01 -1.13525875e-01
3.64802122e-01 -9.27778542e-01 -9.15031061e-02 2.03486726e-01
-1.78258702e-01 -9.65913683e-02 4.30771470e-01 8.98132503e-01
4.21101153e-01 5.08728288e-02 5.11720419e-01 -1.04236460e+00
-4.40552145e-01 8.74627829e-01 1.38574839e-03 -1.86278507e-01
1.29218429e-01 2.57800341e-01 3.57882947e-01 -2.45543495e-01
-2.52243932e-02 9.26237822e-01 1.20734608e+00 3.98934156e-01
-1.79653692e+00 -9.82243642e-02 1.86887965e-01 2.67794520e-01
-1.42215109e+00 -4.50073391e-01 4.51629251e-01 -4.74345803e-01
9.30335402e-01 8.12632263e-01 1.34360528e+00 8.21190417e-01
3.79529297e-01 1.94364890e-01 9.71771240e-01 -4.44636762e-01
7.67068267e-01 2.72095799e-01 2.19611704e-01 2.85913467e-01
4.90226209e-01 2.61753738e-01 -1.66727584e-02 -1.22269034e-01
6.89909577e-01 2.21363038e-01 -3.61981392e-01 -2.05608550e-03
-1.14947224e+00 9.77843821e-01 8.23618993e-02 -3.12252324e-02
-7.30742872e-01 -3.10719520e-01 4.05580789e-01 1.95336103e-01
6.83409929e-01 4.11393553e-01 -7.60629594e-01 -9.31305885e-02
-8.96650016e-01 -1.18439689e-01 1.15226114e+00 6.73544705e-01
2.58640647e-01 1.73003018e-01 -2.39769537e-02 5.82552910e-01
2.85733163e-01 3.26360047e-01 5.43708026e-01 -8.85790765e-01
-1.05484761e-01 8.78305197e-01 2.58360028e-01 -1.08578134e+00
-9.29976046e-01 -5.15662253e-01 -1.54354239e+00 1.58250839e-01
6.79862857e-01 -4.03809905e-01 -8.40656400e-01 1.39637685e+00
2.37718210e-01 8.02611038e-02 3.32901239e-01 1.05009902e+00
3.02159309e-01 5.29815853e-01 3.82595807e-01 -8.28453541e-01
1.28356969e+00 -7.50681102e-01 -7.13689625e-01 2.24951312e-01
6.06063545e-01 -3.31979334e-01 5.71617544e-01 7.31208563e-01
-1.04949462e+00 -5.13131022e-01 -7.12187052e-01 3.53927255e-01
-2.63448596e-01 7.89921284e-02 4.96933430e-01 7.96166062e-01
-8.14671934e-01 1.08138359e+00 -1.13835382e+00 -4.21514630e-01
-8.15144107e-02 5.40785134e-01 1.26713052e-01 3.35986376e-01
-1.18399060e+00 9.16282177e-01 4.22751933e-01 3.69394958e-01
4.64109145e-02 -1.02655768e+00 -1.00938761e+00 2.20490485e-01
-1.17817275e-01 -1.10534585e+00 8.86178970e-01 -1.34339726e+00
-1.60374427e+00 2.46843234e-01 -3.75256628e-01 -1.04144108e+00
8.75747561e-01 -6.22576475e-02 -5.86087286e-01 2.44441256e-02
-2.42062181e-01 4.72359240e-01 4.24935281e-01 -7.95828879e-01
-5.16747892e-01 -3.60718131e-01 -4.79412228e-01 2.49617025e-02
2.38036320e-01 -4.05029804e-01 1.20944880e-01 -5.08890331e-01
3.23830247e-01 -7.31455982e-01 -7.99284577e-01 -2.40028873e-02
-4.59193140e-01 2.02693529e-02 -5.67583963e-02 -9.16179240e-01
1.25946546e+00 -1.76259780e+00 6.83928281e-02 5.08354306e-01
1.74601182e-01 7.63089210e-02 3.02922428e-01 1.46406189e-01
-4.60091829e-01 2.76158810e-01 -3.06327313e-01 -1.02640472e-01
2.78860658e-01 3.92759293e-01 5.84998056e-02 3.10123175e-01
2.77000099e-01 6.60441697e-01 -7.34219134e-01 -1.07885534e-02
6.50854826e-01 7.19457030e-01 -6.09264612e-01 -1.20154776e-01
-3.72632533e-01 5.36509991e-01 7.95041397e-03 5.67154467e-01
4.71058488e-01 -3.08217436e-01 8.43679667e-01 -3.83564860e-01
-2.76167840e-01 3.71005125e-02 -1.33029509e+00 1.16124952e+00
-5.50866313e-02 6.52491748e-02 -3.65462601e-01 -1.39368546e+00
1.19510603e+00 4.14672047e-01 8.65166724e-01 -8.90794456e-01
1.69710647e-02 3.70459929e-02 1.33532658e-01 -6.69113457e-01
-2.96124548e-01 -2.87315875e-01 2.87203908e-01 -7.11990222e-02
-4.25070822e-01 6.23698235e-01 4.32541788e-01 -5.18408716e-01
6.11849606e-01 2.08900645e-01 7.75398135e-01 -6.28737628e-01
3.90889019e-01 9.64671150e-02 8.95558238e-01 4.81313586e-01
-2.28452116e-01 5.47894180e-01 6.74673796e-01 -1.18553853e+00
-1.12966144e+00 -8.34316134e-01 -4.49309051e-01 3.54908943e-01
-9.77749377e-02 -3.35705340e-01 -6.30199492e-01 -3.90029401e-01
2.16803789e-01 9.69923019e-01 -7.91190684e-01 -1.22324832e-01
-4.15149182e-01 -1.24368775e+00 1.91812236e-02 5.35227001e-01
3.17376047e-01 -4.86653477e-01 -6.24818206e-01 6.97061718e-01
-2.60299385e-01 -6.10522866e-01 -1.07465163e-02 5.78504741e-01
-1.13810503e+00 -9.82730329e-01 -5.53294778e-01 -3.72604012e-01
7.00495183e-01 -4.69516307e-01 1.35075200e+00 -1.71419635e-01
-2.57720143e-01 1.75897889e-02 -2.30368719e-01 -4.31901008e-01
-5.21658659e-01 -2.71673262e-01 3.57621938e-01 4.30576541e-02
7.42665529e-01 -6.62770867e-01 -1.04987741e+00 4.15730566e-01
-4.07593459e-01 3.43094975e-01 3.49831700e-01 6.97665095e-01
1.01894665e+00 7.99105838e-02 6.60373271e-01 -5.38257599e-01
2.11017653e-01 -7.61595964e-01 -5.57970226e-01 1.26807719e-01
-1.06109762e+00 -2.31977608e-02 8.62248778e-01 -5.01209259e-01
-4.63475108e-01 3.18887264e-01 -2.13628244e-02 -9.91995186e-02
-6.06919587e-01 5.81020117e-01 6.67082220e-02 2.86949545e-01
8.81659091e-01 1.57443106e-01 4.18900132e-01 -5.62299073e-01
-1.92814935e-02 2.58495986e-01 1.53926581e-01 -2.31256515e-01
9.22758579e-02 1.57068186e-02 3.56530324e-02 -6.96261346e-01
-4.79660720e-01 -8.53885487e-02 -6.25788152e-01 -8.83068368e-02
8.97105455e-01 -8.13616991e-01 -1.13357520e+00 -4.80470806e-02
-6.98659599e-01 -3.64044666e-01 -5.51396072e-01 8.54154885e-01
-5.55658817e-01 2.44597986e-01 -4.65131968e-01 -8.33397627e-01
-5.20831645e-01 -9.03386176e-01 5.28003573e-01 9.88062099e-02
-9.00109053e-01 -1.37643576e+00 -1.12392403e-01 2.24350858e-02
4.11576927e-01 8.43647957e-01 1.05492115e+00 -7.77788579e-01
-2.12481663e-01 1.48810491e-01 -4.84316312e-02 2.76398659e-01
4.09667224e-01 -9.37029570e-02 -3.96982104e-01 1.57394350e-01
1.41835734e-01 4.28118974e-01 5.44477284e-01 1.12252724e+00
9.57237363e-01 -9.89721715e-01 -1.43506438e-01 6.72020435e-01
1.55423927e+00 6.13012016e-01 7.06052184e-01 3.05899441e-01
1.54521674e-01 5.34097254e-01 1.72528282e-01 8.22430253e-01
5.51502705e-01 5.23149788e-01 4.15741771e-01 -5.12212157e-01
1.25481039e-01 3.32372576e-01 2.22050920e-01 5.92235804e-01
-5.31914473e-01 4.92459759e-02 -8.72451365e-01 3.34035575e-01
-1.98715544e+00 -9.55038190e-01 -5.50640702e-01 2.40925550e+00
9.00065422e-01 -8.69178772e-02 3.93059433e-01 2.34883294e-01
4.81977642e-01 -5.77084541e-01 -6.24053121e-01 -6.89805031e-01
-2.97235753e-02 1.60161301e-01 3.55608493e-01 6.69029951e-01
-1.06646395e+00 9.30660777e-03 6.64483547e+00 -5.87426089e-02
-1.04299033e+00 -2.98609287e-01 1.09077096e+00 -1.74017578e-01
7.36908317e-02 -4.31285709e-01 -6.14728034e-01 9.44562614e-01
1.46501505e+00 -1.43467292e-01 4.76468772e-01 8.33701372e-01
9.26780403e-01 -2.39884123e-01 -1.21853960e+00 8.08342516e-01
-3.60647514e-02 -1.29625940e+00 -1.90766796e-01 -2.48830914e-01
5.20115018e-01 -3.35390687e-01 -3.39620948e-01 7.30982125e-02
7.45892450e-02 -1.12268293e+00 4.22706395e-01 9.06944752e-01
3.33391428e-01 -7.84076214e-01 9.11129653e-01 1.92082509e-01
-9.14122522e-01 -9.97263566e-02 -3.41081619e-01 -5.05458713e-01
2.86376745e-01 1.04866683e+00 -7.77493775e-01 6.15549505e-01
5.02820373e-01 7.11362839e-01 -2.44833842e-01 1.29141152e+00
1.98887512e-01 5.51306725e-01 -6.30195916e-01 4.90721352e-02
2.07837243e-02 -7.91549802e-01 2.90231198e-01 1.09925985e+00
7.47137189e-01 3.87050271e-01 3.02854031e-01 8.08298290e-01
7.42869616e-01 2.90002733e-01 -2.73754328e-01 3.30005616e-01
-3.06641553e-02 7.28529513e-01 -6.63202047e-01 -5.50085604e-01
-2.96773553e-01 8.43366146e-01 -3.74261498e-01 3.87492448e-01
-9.82521772e-01 -4.31582667e-02 7.88003564e-01 3.54638129e-01
-4.45760526e-02 1.05998978e-01 -8.28929484e-01 -1.11847317e+00
-2.17116401e-02 -9.95879412e-01 5.77440441e-01 -3.88425022e-01
-1.23628080e+00 3.73932362e-01 -2.22309381e-01 -1.35799038e+00
-5.05398154e-01 -4.63872135e-01 -3.45042676e-01 8.86021972e-01
-1.31964493e+00 -6.67963207e-01 -3.36931467e-01 3.52592468e-01
4.83876556e-01 1.73902035e-01 1.50568414e+00 3.80583972e-01
-7.33254790e-01 2.63906240e-01 3.02347183e-01 -2.28425056e-01
4.52448964e-01 -1.46372139e+00 -1.36215836e-01 2.93129116e-01
-5.04240870e-01 4.87275571e-01 9.22752976e-01 -6.01419449e-01
-9.86673713e-01 -1.31498277e+00 1.24692214e+00 1.28837004e-02
3.21090966e-01 1.55560806e-01 -8.74873877e-01 6.78113163e-01
1.06247157e-01 -2.53632039e-01 1.25319266e+00 2.86859781e-01
2.67753094e-01 -4.07589018e-01 -1.30951202e+00 5.55817068e-01
5.37889719e-01 3.79939735e-01 -3.90268445e-01 5.82666457e-01
3.16333830e-01 -4.23030168e-01 -1.51575208e+00 5.54963768e-01
6.75712168e-01 -1.16364908e+00 1.01693881e+00 -6.54198110e-01
3.19109142e-01 -3.07967931e-01 -4.74360138e-02 -1.48593462e+00
-4.98967260e-01 -6.47872329e-01 -2.15711176e-01 6.67357624e-01
5.92122436e-01 -9.86890316e-01 7.79326320e-01 1.26319814e+00
-1.48447379e-01 -7.24347949e-01 -5.93366563e-01 -5.86105704e-01
-2.11424440e-01 -2.94222593e-01 6.16981328e-01 1.15254569e+00
5.11227310e-01 4.70031388e-02 -4.64198530e-01 2.81882942e-01
8.02047849e-01 2.38560468e-01 2.68278718e-01 -1.44160414e+00
-3.62970918e-01 -5.81654906e-01 -5.69443166e-01 -4.06000108e-01
-4.10115033e-01 -7.40111411e-01 -3.36895347e-01 -1.45622218e+00
-1.21147104e-01 -4.17836577e-01 -3.31129938e-01 7.60077059e-01
-2.21647955e-02 1.64470196e-01 -7.47918710e-02 -1.36552781e-01
-1.33982405e-01 1.59677222e-01 7.21230388e-01 -1.33967981e-01
-7.12772429e-01 6.80490062e-02 -5.93983114e-01 8.94301653e-01
1.20152152e+00 -2.25106135e-01 -3.72509658e-01 -1.19045533e-01
-6.93005621e-02 2.19416976e-01 3.57911497e-01 -1.08529162e+00
-1.21178463e-01 -3.48231524e-01 1.06728959e+00 -2.61085272e-01
6.45316243e-02 -9.93856668e-01 1.07921040e+00 9.10114408e-01
-1.68021634e-01 -1.66834310e-01 3.49440247e-01 2.93950826e-01
-1.06444418e-01 3.14664841e-02 6.80868089e-01 -2.25099564e-01
-3.55167776e-01 -8.05561915e-02 -6.06003523e-01 -4.40764576e-01
1.10457623e+00 -3.44245791e-01 -2.21622083e-02 -2.03438729e-01
-1.98150790e+00 2.33922243e-01 2.78941989e-01 -2.97356825e-02
2.83643246e-01 -1.30603576e+00 -7.70703912e-01 4.89048988e-01
-2.30896920e-01 -1.43074527e-01 4.68538851e-01 1.27731895e+00
-5.87322652e-01 4.68791008e-01 -1.24173298e-01 -8.39306533e-01
-1.01379836e+00 6.43111169e-01 5.48783660e-01 -2.37389103e-01
-9.88977611e-01 7.70763084e-02 -1.21721961e-01 -9.78265479e-02
4.38632816e-01 -9.89896119e-01 -1.41956553e-01 1.76630408e-01
7.77470946e-01 4.54033583e-01 9.68445763e-02 -3.14714462e-02
-1.22869469e-01 5.84255099e-01 4.73785102e-01 6.45421922e-01
1.39302301e+00 -3.40884954e-01 -7.55455717e-02 3.83865714e-01
7.67506361e-01 -5.13215423e-01 -1.18376684e+00 4.02438432e-01
4.18759957e-02 -2.53995340e-02 -1.94892660e-01 -1.37607944e+00
-9.05687630e-01 4.69146878e-01 9.22639251e-01 6.38191104e-01
1.44298410e+00 -6.43331528e-01 4.63939965e-01 9.03296024e-02
1.95922926e-02 -7.55700350e-01 -6.69473767e-01 7.32314140e-02
8.35831285e-01 -1.18002141e+00 -1.10250473e-01 -4.94453132e-01
-6.22198403e-01 1.37843525e+00 5.95853180e-02 4.90559451e-02
5.01142800e-01 1.15086704e-01 2.93450058e-01 3.64772737e-01
-6.90660894e-01 2.80458145e-02 1.81351945e-01 7.33444810e-01
2.76991159e-01 4.34551388e-01 -5.64407527e-01 9.16837394e-01
-5.80175631e-02 3.52756083e-01 1.32267594e-01 2.69281447e-01
-3.37039590e-01 -8.58316481e-01 -3.79466057e-01 4.92822468e-01
-3.27834010e-01 -2.17433780e-01 1.47867978e-01 8.10178280e-01
2.93511093e-01 8.52379918e-01 4.30290610e-01 -4.88133775e-03
4.33185965e-01 6.62172496e-01 2.47136056e-01 -7.32582882e-02
-5.58271945e-01 5.04045606e-01 4.03130561e-01 -7.61642933e-01
-4.95720297e-01 -7.09636629e-01 -1.28812551e+00 -5.78267992e-01
2.15830401e-01 3.48684698e-01 5.90994179e-01 6.84855103e-01
7.01098919e-01 6.43473268e-01 4.88231957e-01 -7.58030295e-01
-4.86260861e-01 -9.28777099e-01 -7.32948184e-01 3.58442366e-01
2.95008212e-01 -2.46666327e-01 -4.18220580e-01 5.95125675e-01] | [8.124436378479004, 5.622197151184082] |
0a243589-61d9-476f-8422-a200b3ec4b22 | interpretability-of-machine-learning-recent | 2305.00537 | null | https://arxiv.org/abs/2305.00537v1 | https://arxiv.org/pdf/2305.00537v1.pdf | Interpretability of Machine Learning: Recent Advances and Future Prospects | The proliferation of machine learning (ML) has drawn unprecedented interest in the study of various multimedia contents such as text, image, audio and video, among others. Consequently, understanding and learning ML-based representations have taken center stage in knowledge discovery in intelligent multimedia research and applications. Nevertheless, the black-box nature of contemporary ML, especially in deep neural networks (DNNs), has posed a primary challenge for ML-based representation learning. To address this black-box problem, the studies on interpretability of ML have attracted tremendous interests in recent years. This paper presents a survey on recent advances and future prospects on interpretability of ML, with several application examples pertinent to multimedia computing, including text-image cross-modal representation learning, face recognition, and the recognition of objects. It is evidently shown that the study of interpretability of ML promises an important research direction, one which is worth further investment in. | ['Ling Guan', 'Lei Gao'] | 2023-04-30 | null | null | null | null | ['face-recognition'] | ['computer-vision'] | [ 6.69729888e-01 6.24289624e-02 -5.43316483e-01 -5.33940017e-01
-3.74955624e-01 -1.63625732e-01 5.05536497e-01 3.61684471e-01
-1.37864307e-01 6.04885578e-01 7.33246207e-02 -3.56097132e-01
-3.12048882e-01 -6.10922277e-01 -3.52426022e-01 -6.87034369e-01
-1.25942603e-01 2.44834781e-01 -4.48132902e-01 7.86799192e-02
2.68356770e-01 6.93068802e-01 -1.89779627e+00 7.75131702e-01
5.13547301e-01 1.40396726e+00 1.03491463e-01 5.57475030e-01
-4.59262550e-01 1.18477154e+00 -5.60339749e-01 -6.54912114e-01
-3.40644807e-01 -4.29447412e-01 -6.95792019e-01 4.84843642e-01
3.10909152e-01 -5.52240983e-02 -5.00094473e-01 8.91538203e-01
4.00978416e-01 1.95481196e-01 8.46345484e-01 -1.45610261e+00
-9.80056524e-01 5.21963537e-01 -6.49275184e-01 3.02569181e-01
2.60574609e-01 -2.85057038e-01 1.17534292e+00 -1.15044057e+00
2.70390511e-01 1.54544795e+00 4.39003825e-01 4.86639053e-01
-8.54035020e-01 -3.85617465e-01 1.86974421e-01 7.76647806e-01
-1.33244860e+00 -3.58975738e-01 7.91104257e-01 -2.76329249e-01
5.78808427e-01 4.64516610e-01 5.57933450e-01 9.89905477e-01
2.55745143e-01 1.41207254e+00 8.47666562e-01 -7.83838868e-01
4.42373753e-02 1.98542312e-01 1.58398867e-01 8.59693289e-01
2.32867509e-01 -2.56425619e-01 -8.81850600e-01 1.84193775e-01
7.68957913e-01 4.31100093e-02 -1.20876126e-01 -1.06184959e-01
-9.67596710e-01 1.03942907e+00 2.76208341e-01 4.38780159e-01
-7.92530626e-02 2.04757676e-01 4.42305356e-01 2.27467537e-01
5.29052377e-01 1.60263225e-01 -2.34148964e-01 1.10977650e-01
-7.61767030e-01 -7.84846395e-03 5.72830677e-01 7.61944950e-01
6.69995964e-01 5.40287852e-01 3.61719966e-01 9.40578282e-01
4.15213794e-01 3.94533783e-01 6.75867736e-01 -7.41275311e-01
3.42645347e-01 7.51089990e-01 -6.65243149e-01 -1.38125205e+00
-3.42698723e-01 -4.45249349e-01 -1.30750287e+00 2.12823749e-02
2.04714075e-01 1.90550581e-01 -4.79296297e-01 1.41596425e+00
-1.23054080e-03 1.04172662e-01 -5.85803762e-03 6.10784292e-01
1.20993209e+00 8.48631680e-01 2.31245622e-01 -3.60994965e-01
1.23608398e+00 -6.27981007e-01 -9.55590785e-01 -1.99233964e-01
4.57822710e-01 -6.41476512e-01 7.17599094e-01 4.94085401e-01
-8.64756763e-01 -7.60944843e-01 -1.03399146e+00 -1.30078852e-01
-5.23306906e-01 3.11224312e-01 9.78742838e-01 5.32740355e-01
-7.14440703e-01 3.55268866e-01 -4.78796750e-01 -4.37090337e-01
7.14265347e-01 5.41475236e-01 -3.54652584e-01 -3.44795734e-01
-1.22381675e+00 7.40720987e-01 5.79229414e-01 4.02209699e-01
-3.30543816e-01 -2.34984621e-01 -1.07522380e+00 -6.36150548e-03
2.78583199e-01 -5.42595565e-01 9.93209064e-01 -1.57472360e+00
-1.16640174e+00 1.03941572e+00 -3.63351583e-01 -3.77322644e-01
1.26889482e-01 -1.33059651e-01 -8.94527435e-01 2.64885664e-01
-1.33702308e-01 6.83812976e-01 1.15902710e+00 -1.22416949e+00
-5.60585260e-01 -4.57871556e-01 1.01018980e-01 1.19634584e-01
-6.20270669e-01 -4.41595027e-03 -1.60493895e-01 -8.06145072e-01
2.58312345e-01 -5.23425639e-01 1.43066673e-02 2.90465921e-01
-1.51051328e-01 -5.33696234e-01 9.07040000e-01 -3.85135859e-01
1.30973148e+00 -2.12066627e+00 1.72709197e-01 2.63526291e-02
3.27431709e-01 2.02916816e-01 -3.29346836e-01 3.76741976e-01
-3.10801476e-01 2.00256258e-01 -1.06888667e-01 -2.14896455e-01
-1.02724284e-01 4.15712893e-01 -4.21644419e-01 4.06065762e-01
5.30976355e-01 9.48446035e-01 -6.65131390e-01 -6.24156237e-01
4.66644943e-01 5.25891364e-01 -2.30962083e-01 5.70158958e-02
-3.70859981e-01 3.80324662e-01 -6.69577539e-01 7.77670324e-01
3.92652333e-01 -4.43333238e-01 1.45478338e-01 -3.60354632e-01
2.12781988e-02 -2.14873120e-01 -9.93247509e-01 1.45078051e+00
-3.50900143e-01 1.44734883e+00 -2.44924694e-01 -1.56930840e+00
8.18250656e-01 5.05231857e-01 6.67775154e-01 -7.53016293e-01
1.59492165e-01 -1.46558946e-02 7.02882260e-02 -8.26653838e-01
6.09492183e-01 -3.29778373e-01 3.57553512e-02 4.69605863e-01
-8.51882249e-03 2.09071919e-01 1.50534913e-01 4.43175323e-02
3.37928593e-01 -1.98946491e-01 6.36863232e-01 -1.48402872e-02
9.33269978e-01 -2.97028512e-01 1.67289481e-01 5.22190988e-01
-1.27309486e-01 4.32334989e-01 2.75637150e-01 -6.94372237e-01
-6.06917918e-01 -9.66969669e-01 -2.56334126e-01 1.29069066e+00
2.83312481e-02 -2.72869617e-01 -5.37758887e-01 -4.03702497e-01
-1.02632783e-01 4.85027552e-01 -5.20842493e-01 -1.72717318e-01
-6.22136474e-01 -7.04987049e-01 5.84656239e-01 5.00809014e-01
5.24545670e-01 -1.26641238e+00 -1.46689221e-01 1.16341069e-01
-2.54098892e-01 -1.23306048e+00 1.64602473e-01 -1.10432962e-02
-1.18160045e+00 -1.14242625e+00 -4.03304219e-01 -9.46476996e-01
7.15401590e-01 4.04275328e-01 1.15663445e+00 1.60686582e-01
-4.71581548e-01 7.34094143e-01 -2.51354873e-01 -6.33005321e-01
-4.04252172e-01 1.29822316e-02 4.17912379e-02 2.82992333e-01
4.38451201e-01 -2.91621745e-01 -2.45306104e-01 3.93940248e-02
-1.42914891e+00 3.22352238e-02 6.80099964e-01 7.07369089e-01
5.34594715e-01 4.13376778e-01 1.02510810e+00 -7.61382461e-01
6.03109121e-01 -7.23297477e-01 -1.53480992e-01 4.14045781e-01
-1.52232513e-01 -3.94596159e-02 5.53166151e-01 -4.44334358e-01
-1.13337541e+00 -2.62749165e-01 -1.48596749e-01 -1.14514112e-01
-2.40872324e-01 1.04301941e+00 -5.69788754e-01 -9.07009393e-02
4.59994882e-01 2.98753083e-01 3.27452645e-02 -4.00531977e-01
4.29396093e-01 8.28500986e-01 5.04499018e-01 -5.80923975e-01
4.35829669e-01 3.90301675e-01 2.50058293e-01 -1.51202989e+00
-9.98350918e-01 -2.86860108e-01 -6.75363362e-01 -4.84635979e-01
6.63816929e-01 -7.28252351e-01 -6.29822552e-01 3.23993206e-01
-1.28686869e+00 3.65125746e-01 -1.08811036e-01 3.42825443e-01
-5.12751698e-01 5.08010924e-01 -5.10132968e-01 -8.21324229e-01
-8.94739330e-02 -1.12024260e+00 6.96485102e-01 2.50766069e-01
-3.80245537e-01 -1.47051072e+00 -4.59406346e-01 8.33391786e-01
1.08516261e-01 1.32358670e-01 1.34340680e+00 -6.36885703e-01
-8.31809819e-01 -3.96386445e-01 -3.20383847e-01 5.41603744e-01
1.90370396e-01 -2.39934749e-03 -1.33256376e+00 -1.57526683e-03
1.38135478e-01 -4.34247315e-01 7.72521377e-01 4.14453179e-01
1.77677584e+00 -3.06900322e-01 -6.47329018e-02 3.43564540e-01
1.24749494e+00 1.91720739e-01 5.06287098e-01 1.31750390e-01
5.92280567e-01 6.34109318e-01 4.11769360e-01 4.70904142e-01
1.30402684e-01 2.50787199e-01 5.45237720e-01 -1.11022584e-01
-4.89117280e-02 -9.87935346e-03 2.94945538e-01 1.10341775e+00
-1.86054811e-01 -6.71999037e-01 -7.00638652e-01 1.36553571e-01
-1.81441629e+00 -1.16361606e+00 -6.96550831e-02 1.81796563e+00
4.83969301e-01 -1.11470945e-01 -3.94421369e-01 5.83625793e-01
7.96395481e-01 3.40407878e-01 -5.44668436e-01 -2.98781246e-01
-4.34826106e-01 1.65152058e-01 -3.75418097e-01 2.25723594e-01
-1.18844056e+00 6.66106820e-01 6.50080013e+00 1.01727211e+00
-1.28155398e+00 -2.66422570e-01 9.04360175e-01 3.23305756e-01
-3.13465834e-01 -5.58033645e-01 -5.63234866e-01 2.04525203e-01
8.56946170e-01 -1.21447928e-01 1.38943851e-01 8.50494444e-01
2.04267874e-01 -2.03157783e-01 -1.46664596e+00 1.39860439e+00
4.76114988e-01 -1.55019498e+00 7.97093749e-01 -1.57912642e-01
6.43432140e-01 -4.31161374e-01 5.23585320e-01 2.28597730e-01
-5.77877700e-01 -1.28789103e+00 5.68767309e-01 3.30097646e-01
7.62208283e-01 -9.81697500e-01 7.09949076e-01 3.34061593e-01
-1.17228520e+00 -2.08201678e-03 -4.88557994e-01 -2.77867496e-01
-9.99905989e-02 6.42245173e-01 -6.23668432e-01 4.86056954e-01
1.94848076e-01 1.07171297e+00 -3.72928798e-01 8.59284937e-01
-1.17460914e-01 3.97318214e-01 2.08781987e-01 -2.93911770e-02
1.54365554e-01 -8.03317800e-02 4.35463369e-01 1.25393713e+00
-1.01826862e-02 1.50827199e-01 3.34517397e-02 6.71843708e-01
-4.58752185e-01 7.90382251e-02 -5.99876344e-01 -4.33338910e-01
1.17106855e-01 1.08980453e+00 -7.07460463e-01 -3.72376323e-01
-5.28370321e-01 7.01750278e-01 1.81826845e-01 3.37386668e-01
-6.88825369e-01 -3.06256805e-02 6.81496680e-01 -1.69235989e-02
-7.53571019e-02 -3.13204557e-01 -4.01826531e-01 -1.08675683e+00
-1.43555149e-01 -1.01459181e+00 4.70135152e-01 -6.10740781e-01
-1.35544896e+00 4.97026682e-01 1.19185820e-01 -1.07631946e+00
-2.50189692e-01 -9.82905209e-01 -2.75938034e-01 4.04973745e-01
-1.60068417e+00 -1.03582311e+00 -1.29945070e-01 5.15969872e-01
9.08515275e-01 -3.81943524e-01 8.47392261e-01 3.58786821e-01
-6.20272517e-01 5.53081155e-01 2.55631775e-01 1.84855312e-01
3.29443723e-01 -6.64292812e-01 -2.18513057e-01 2.73380458e-01
6.52802706e-01 4.85518754e-01 3.26582193e-01 -9.10511240e-02
-1.57580137e+00 -1.06055188e+00 1.01783204e+00 -2.55807906e-01
6.32252336e-01 2.23635789e-02 -8.78791392e-01 5.89024961e-01
1.39232635e-01 -1.04148224e-01 1.19339907e+00 -3.59262191e-02
-1.74953446e-01 -3.11437994e-01 -8.84285629e-01 6.19635463e-01
6.15640879e-01 -7.46021748e-01 -5.26088953e-01 3.66609544e-01
3.32350820e-01 2.76080444e-02 -6.75168633e-01 3.92545491e-01
6.88362956e-01 -8.65374506e-01 1.15496111e+00 -8.45743954e-01
5.92591107e-01 -5.64785767e-03 -2.47182831e-01 -7.98616171e-01
-1.01275742e-01 -1.68238655e-01 -3.72718394e-01 1.05578458e+00
2.83809695e-02 -2.89724648e-01 8.93990338e-01 4.67307627e-01
-1.59862489e-01 -9.15050089e-01 -8.36729109e-01 -2.89886326e-01
-1.91754505e-01 -8.29921961e-01 1.79358944e-01 8.71211648e-01
-5.95320724e-02 5.62769115e-01 -3.48419368e-01 -2.43342035e-02
7.21848011e-01 1.24169268e-01 2.65789956e-01 -1.46551430e+00
3.82317342e-02 -6.14813983e-01 -6.92238688e-01 -1.20183754e+00
5.59646487e-01 -9.84927773e-01 -3.20684046e-01 -1.34928477e+00
1.91874281e-01 -1.54977322e-01 -4.14355665e-01 2.16285244e-01
-4.02841810e-03 4.25015271e-01 1.94691703e-01 1.26077130e-01
-8.06475759e-01 6.12412632e-01 1.32304251e+00 -5.49333751e-01
2.84705043e-01 2.04402179e-01 -6.39259636e-01 1.21981490e+00
8.98683488e-01 -1.43498808e-01 -6.43518269e-01 -6.28928065e-01
3.19984734e-01 7.90554881e-02 2.45192274e-01 -9.52942073e-01
1.25096634e-01 -3.09548378e-01 7.30418086e-01 -6.48281038e-01
6.07739985e-01 -9.67108846e-01 -1.75547060e-02 2.34398931e-01
-5.76812685e-01 5.11611067e-02 2.06410393e-01 7.10039318e-01
-6.39294267e-01 -4.00122941e-01 7.17162251e-01 -2.37013191e-01
-1.26726711e+00 4.40614939e-01 -8.70445609e-01 1.87941119e-02
7.99152374e-01 -4.44629163e-01 5.53390048e-02 -5.15362680e-01
-7.17305541e-01 -2.03032777e-01 -2.18936592e-01 5.15965760e-01
1.14931726e+00 -1.29585147e+00 -4.91742253e-01 2.67760366e-01
1.37746021e-01 -1.10789582e-01 2.53747702e-01 3.70005518e-01
-4.64033097e-01 6.36208832e-01 -8.07413384e-02 -6.79583251e-01
-1.32589650e+00 2.51867682e-01 6.96757063e-02 -9.43785161e-02
-1.93420827e-01 7.91311860e-01 3.68372917e-01 7.70264193e-02
5.58135390e-01 -1.64053768e-01 -4.27347183e-01 2.12143108e-01
7.54428506e-01 4.96925265e-01 -1.53114796e-01 -7.61863708e-01
-1.85602307e-01 3.96020561e-01 -4.18365672e-02 3.08965921e-01
1.34568226e+00 -5.09704351e-02 -3.94848853e-01 7.65165567e-01
1.37593746e+00 -5.37370980e-01 -7.13020504e-01 -3.42263937e-01
3.60521823e-01 -2.82964677e-01 4.27339710e-02 -4.54460710e-01
-1.16471851e+00 1.25671923e+00 3.20725471e-01 2.18670756e-01
1.23265111e+00 2.85578165e-02 4.51372564e-01 7.56957233e-01
2.89852619e-01 -1.02340925e+00 6.89010680e-01 4.84223932e-01
1.02065694e+00 -1.44659960e+00 2.49888882e-01 -4.42958176e-01
-3.79732311e-01 1.57068002e+00 3.71807069e-01 2.68428981e-01
7.66014457e-01 -5.36433719e-02 -1.91042736e-01 -2.14543995e-02
-5.99163949e-01 7.60809407e-02 6.18170798e-01 6.61967397e-01
7.78624356e-01 -3.02326214e-02 5.95616773e-02 4.10939515e-01
1.95633352e-01 1.81298610e-02 2.57348686e-01 9.08960760e-01
-6.07594132e-01 -1.04172313e+00 -4.14858431e-01 4.53541547e-01
-5.64669311e-01 -2.75923431e-01 -3.06030512e-01 8.02806497e-01
1.72509879e-01 9.47524190e-01 1.05548538e-01 -2.21734330e-01
-1.87593371e-01 2.97363941e-02 5.57124376e-01 -3.73986125e-01
-1.66616663e-01 -3.63463163e-01 2.36895820e-03 -1.98342368e-01
-8.63433242e-01 -4.26444918e-01 -1.25901723e+00 -5.04799366e-01
-2.61251420e-01 6.36235625e-02 6.82528853e-01 1.22500551e+00
9.42131504e-02 4.56869125e-01 4.85446632e-01 -8.41537297e-01
-2.20521137e-01 -7.05422759e-01 -6.50054932e-01 2.05663666e-01
3.93627077e-01 -5.99003315e-01 -2.33392701e-01 3.83983374e-01] | [10.402058601379395, 1.7087620496749878] |
93f7834e-3960-4926-ad06-069875c835db | italic-an-italian-intent-classification | 2306.08502 | null | https://arxiv.org/abs/2306.08502v1 | https://arxiv.org/pdf/2306.08502v1.pdf | ITALIC: An Italian Intent Classification Dataset | Recent large-scale Spoken Language Understanding datasets focus predominantly on English and do not account for language-specific phenomena such as particular phonemes or words in different lects. We introduce ITALIC, the first large-scale speech dataset designed for intent classification in Italian. The dataset comprises 16,521 crowdsourced audio samples recorded by 70 speakers from various Italian regions and annotated with intent labels and additional metadata. We explore the versatility of ITALIC by evaluating current state-of-the-art speech and text models. Results on intent classification suggest that increasing scale and running language adaptation yield better speech models, monolingual text models outscore multilingual ones, and that speech recognition on ITALIC is more challenging than on existing Italian benchmarks. We release both the dataset and the annotation scheme to streamline the development of new Italian SLU models and language-specific datasets. | ['Elena Baralis', 'Luca Cagliero', 'Eliana Pastor', 'Giuseppe Attanasio', 'Luca Colomba', 'Lorenzo Vaiani', 'Moreno La Quatra', 'Alkis Koudounas'] | 2023-06-14 | null | null | null | null | ['classification-1', 'spoken-language-understanding', 'intent-classification', 'spoken-language-understanding'] | ['methodology', 'natural-language-processing', 'natural-language-processing', 'speech'] | [-1.41469106e-01 1.23233356e-01 -3.11893106e-01 -5.53326249e-01
-1.18602610e+00 -8.09035838e-01 7.42258370e-01 1.17209516e-01
-5.96432090e-01 6.07890725e-01 8.57658863e-01 -3.57453495e-01
4.21369880e-01 -2.31018901e-01 -5.11215746e-01 -1.62498981e-01
1.11532308e-01 8.54060411e-01 1.30540863e-01 -2.08619967e-01
-1.32531732e-01 -7.95988813e-02 -1.30958045e+00 8.49630117e-01
6.24787688e-01 7.04722106e-01 1.32491112e-01 8.32721651e-01
-2.85673797e-01 1.01783466e+00 -9.96814489e-01 -4.17044759e-01
-1.86913803e-01 -1.98808774e-01 -1.02381861e+00 7.72952959e-02
5.32880366e-01 -2.54385024e-01 -3.83922219e-01 4.90573198e-01
7.17192233e-01 -3.97725441e-02 7.18052804e-01 -9.64418948e-01
-8.68796885e-01 1.26650655e+00 2.68261552e-01 1.66969642e-01
7.43130028e-01 4.21852879e-02 1.26321375e+00 -1.12743831e+00
7.31016934e-01 1.32192743e+00 6.84118986e-01 5.02139628e-01
-1.05716813e+00 -4.73539442e-01 3.48094910e-01 8.56095701e-02
-1.50765061e+00 -1.07144666e+00 5.39138556e-01 -4.34518874e-01
1.54025877e+00 3.19039255e-01 3.53324294e-01 1.78930199e+00
-3.10114741e-01 1.36216938e+00 8.82752001e-01 -5.36333621e-01
2.17777342e-01 4.10962343e-01 1.11534446e-01 2.10217401e-01
-2.63398707e-01 -1.85846254e-01 -1.04617858e+00 -2.85750255e-02
4.28310931e-02 -5.92491984e-01 -1.37853146e-01 1.16144106e-01
-1.44243360e+00 8.49917710e-01 -1.87438920e-01 4.71506596e-01
-4.28806283e-02 -1.84107050e-01 1.05410302e+00 4.92330015e-01
7.69019604e-01 2.63055027e-01 -9.48796153e-01 -7.09712088e-01
-8.73776317e-01 -1.62520975e-01 1.23378789e+00 1.13093913e+00
4.86699581e-01 2.06156224e-01 1.97091643e-02 1.40663385e+00
2.18245059e-01 7.11072385e-01 8.51815522e-01 -8.58019292e-01
8.75728309e-01 4.11470503e-01 -2.20945746e-01 -4.16980505e-01
-4.78410423e-01 -2.39790708e-01 -2.23345637e-01 -5.21976292e-01
4.63399231e-01 -3.14438850e-01 -6.25230253e-01 1.45337665e+00
-1.18048586e-01 -2.30835661e-01 4.05445129e-01 4.80314076e-01
1.30647075e+00 7.43541181e-01 1.46546438e-01 -2.48766989e-02
1.44447935e+00 -1.09850085e+00 -6.02580070e-01 -5.66262305e-01
1.09474099e+00 -9.20213282e-01 1.50261652e+00 5.27518392e-01
-8.83584678e-01 -4.66282576e-01 -6.04384780e-01 -2.44344324e-01
-7.07849503e-01 6.87244594e-01 4.28109169e-01 1.09406281e+00
-1.01057827e+00 -3.57903302e-01 -5.14085948e-01 -5.11134684e-01
2.33079404e-01 -2.08996143e-02 -5.33029318e-01 1.98204190e-01
-1.31103325e+00 5.69225132e-01 3.66186082e-01 -3.58874947e-01
-9.96759534e-01 -6.97265446e-01 -1.13128662e+00 -1.75686166e-01
1.68689981e-01 2.32665926e-01 1.57715440e+00 -8.65146339e-01
-1.72016740e+00 1.28063786e+00 -2.05648810e-01 -6.57895923e-01
3.53885204e-01 -1.89738721e-01 -7.46658385e-01 -1.16972111e-01
2.32001826e-01 9.07017708e-01 4.92130637e-01 -8.14940989e-01
-8.66094172e-01 -2.26996660e-01 -2.10550740e-01 2.13428408e-01
-6.89259112e-01 3.58667225e-01 -2.95097589e-01 -7.78373063e-01
-2.68138230e-01 -8.94997299e-01 4.70671356e-01 -5.35694838e-01
-3.89824182e-01 -5.65993845e-01 7.15523064e-01 -8.89095724e-01
1.29207969e+00 -2.15640640e+00 -1.97795764e-01 -4.74671334e-01
-2.71196783e-01 3.18986237e-01 -2.71829844e-01 8.59827161e-01
1.21713221e-01 1.80794477e-01 9.74712893e-02 -7.76411474e-01
2.26967603e-01 4.44752991e-01 -5.21463990e-01 2.68835545e-01
4.66094054e-02 9.05805826e-01 -7.04846799e-01 -3.25399816e-01
3.04725736e-01 2.79084146e-01 -6.60391152e-01 -4.46284153e-02
-2.91981608e-01 5.56010067e-01 -6.66033849e-02 6.98699832e-01
1.66076764e-01 2.86263436e-01 6.14272840e-02 1.34426445e-01
-3.15605760e-01 1.26749897e+00 -8.34991276e-01 1.87339044e+00
-1.07737970e+00 9.60464716e-01 1.22140676e-01 -8.06696415e-01
8.55296612e-01 7.83089578e-01 1.13188669e-01 -5.35889268e-01
2.42938120e-02 5.91552973e-01 -1.89424101e-02 -6.26288772e-01
5.31934977e-01 2.16611460e-01 -5.57385445e-01 3.48595709e-01
4.55909997e-01 -4.21783805e-01 5.34687974e-02 1.31048076e-02
1.02204633e+00 -3.23731512e-01 5.50070107e-01 -3.99363846e-01
7.28419006e-01 -1.76280702e-03 1.26590669e-01 7.36112058e-01
-5.47784686e-01 5.70464134e-01 2.41450086e-01 -3.71106535e-01
-7.59094238e-01 -8.18545699e-01 -3.90392900e-01 1.94627512e+00
-6.49409175e-01 -6.88793838e-01 -6.07668519e-01 -8.69069576e-01
-1.23830110e-01 9.49278831e-01 -4.01098996e-01 3.55535924e-01
-5.75283349e-01 -2.51806766e-01 1.24421775e+00 5.36673307e-01
3.59141201e-01 -1.35595226e+00 1.57553613e-01 3.48449260e-01
-5.85103691e-01 -1.57694793e+00 -6.80461884e-01 2.71166056e-01
-2.16487691e-01 -5.76694429e-01 -5.70762157e-01 -1.05793488e+00
1.30362855e-02 -2.04058975e-01 1.45015776e+00 -3.97627920e-01
-2.05677114e-02 5.91032386e-01 -6.48034871e-01 -5.92989922e-01
-9.03640807e-01 8.26190472e-01 2.25315675e-01 -1.66638181e-01
7.27724373e-01 -1.56283826e-01 1.03530593e-01 2.64683932e-01
-3.83618087e-01 -3.31853032e-01 2.85680264e-01 8.47926259e-01
1.63665768e-02 -5.12810647e-01 1.02422190e+00 -8.22726190e-01
5.09501219e-01 -3.46313685e-01 -1.95197552e-01 1.38254657e-01
1.61566600e-01 -2.61317194e-01 7.49799967e-01 -3.74112457e-01
-1.10578859e+00 1.09699138e-01 -9.09342229e-01 1.96837679e-01
-6.69039369e-01 4.28216904e-01 -2.68169999e-01 2.91808695e-01
7.51444519e-01 4.32146072e-01 -3.27770114e-01 -5.34497559e-01
5.41840792e-01 1.52759969e+00 5.35756230e-01 -6.17997944e-01
-2.36761831e-02 2.03368202e-01 -9.27394688e-01 -1.59558320e+00
-8.10477197e-01 -7.19843686e-01 -7.21976936e-01 4.45903726e-02
7.01518357e-01 -1.37016582e+00 -6.74670041e-01 4.88602042e-01
-1.26353228e+00 -5.83190620e-01 -4.04589415e-01 3.88511181e-01
-5.65774739e-01 2.04792291e-01 -8.78276527e-01 -1.03678548e+00
-2.39901617e-01 -1.26614034e+00 1.67631102e+00 -3.68271410e-01
-7.28379190e-01 -1.32453442e+00 2.34936953e-01 7.14414477e-01
1.56052560e-01 -8.00171614e-01 4.86430764e-01 -1.32312942e+00
-2.35035978e-02 -1.23814248e-01 9.47394744e-02 4.24183786e-01
6.04071319e-02 -7.68812746e-02 -1.40857685e+00 -1.42662391e-01
-3.16798329e-01 -8.86352718e-01 6.48335278e-01 1.07606255e-01
7.19746053e-01 -4.24322456e-01 -6.77876249e-02 1.74111694e-01
6.50465727e-01 1.46585211e-01 3.06847572e-01 2.28087544e-01
5.15959740e-01 8.92404079e-01 3.58120412e-01 3.88280302e-01
8.37136030e-01 7.65432060e-01 -4.88960035e-02 3.89035553e-01
-3.52409303e-01 -4.83727068e-01 9.34078038e-01 1.53833675e+00
7.29078352e-01 -4.61732268e-01 -1.19687653e+00 1.17243302e+00
-1.47082281e+00 -4.28972721e-01 9.46668908e-04 1.95623100e+00
1.19017982e+00 -2.22845301e-02 4.32083696e-01 8.77945796e-02
3.05115879e-01 2.40839228e-01 1.93243902e-02 -5.62448740e-01
-3.71867150e-01 -2.52986532e-02 1.91373467e-01 8.58309507e-01
-1.25412977e+00 1.52310300e+00 7.43026209e+00 1.15752661e+00
-1.04332328e+00 4.76797312e-01 6.14210844e-01 3.27109755e-03
-3.86709064e-01 -2.87955105e-01 -1.27113712e+00 4.13339227e-01
1.41486812e+00 3.79024297e-01 3.09206277e-01 9.29272473e-01
1.33050933e-01 4.11649905e-02 -1.11223912e+00 9.17290568e-01
3.52645904e-01 -1.17399693e+00 -3.38430516e-02 -2.82744348e-01
6.49025798e-01 6.32712901e-01 -1.89177483e-01 7.45651305e-01
5.01559079e-01 -9.69590962e-01 1.06311190e+00 -1.04095407e-01
8.38385820e-01 -4.49629694e-01 6.14030242e-01 4.04049903e-01
-1.02575505e+00 -2.65016928e-02 -1.92951202e-01 -2.37979740e-01
2.86011130e-01 1.91304088e-01 -1.16464579e+00 2.34243423e-01
6.26112461e-01 8.25910151e-01 -7.85386443e-01 3.60683322e-01
-2.77846366e-01 1.29056156e+00 -5.89616776e-01 -4.73450273e-01
3.72777373e-01 1.52182698e-01 6.94348156e-01 1.77431679e+00
-2.10342649e-02 -6.35002375e-01 4.33046699e-01 2.93209523e-01
-7.78863505e-02 7.67470121e-01 -8.33973587e-01 -3.09320778e-01
6.42549276e-01 9.59576428e-01 -4.72690940e-01 -4.20591742e-01
-7.86348164e-01 9.38486040e-01 4.94607031e-01 2.21559510e-01
-3.77216697e-01 -1.14786729e-01 7.25892067e-01 -9.36018899e-02
1.09813437e-01 -4.42639858e-01 -5.03619671e-01 -1.28264809e+00
8.38894397e-03 -1.17883170e+00 3.12799573e-01 -6.55092835e-01
-1.33274698e+00 8.39121699e-01 -3.09607774e-01 -7.68586099e-01
-6.47993445e-01 -9.72821593e-01 -3.00341368e-01 5.28390467e-01
-1.26764941e+00 -1.55916893e+00 2.32799426e-01 4.59548563e-01
1.34074187e+00 -7.75936425e-01 1.06379783e+00 3.98462266e-01
-4.60253030e-01 7.30926454e-01 -3.08067221e-02 3.87317419e-01
1.02103722e+00 -1.24455655e+00 7.41905868e-01 5.60605586e-01
8.27981651e-01 3.12718749e-01 5.59410393e-01 -4.76085037e-01
-8.76107156e-01 -9.03402448e-01 1.52199042e+00 -1.07034886e+00
1.18462563e+00 -9.76115346e-01 -7.28134573e-01 1.03106058e+00
5.20096958e-01 -2.87221670e-01 7.81309307e-01 4.96880233e-01
-3.35490316e-01 9.54569578e-02 -7.44547963e-01 6.34402394e-01
1.08869302e+00 -1.20421505e+00 -5.96959114e-01 5.81176519e-01
9.56664324e-01 -2.56856471e-01 -8.79506290e-01 6.36961386e-02
5.78969538e-01 -7.72128642e-01 8.21830690e-01 -6.57235920e-01
-5.47166616e-02 9.28373709e-02 -5.06234050e-01 -1.39558196e+00
2.92662799e-01 -5.88825226e-01 3.13057393e-01 1.59309435e+00
8.36664140e-01 -8.38126063e-01 5.64565539e-01 -1.04346843e-02
-5.24714112e-01 -2.91495025e-01 -1.30474067e+00 -8.00120413e-01
4.63136196e-01 -1.08462846e+00 3.74107391e-01 8.14975917e-01
3.83738130e-01 5.88410020e-01 -2.19121665e-01 8.83879978e-03
2.07953826e-02 -4.23226833e-01 7.27517426e-01 -1.05440533e+00
-2.35212311e-01 -3.08710665e-01 -4.31422949e-01 -1.23649871e+00
8.15853298e-01 -9.59391117e-01 2.19241858e-01 -1.13272297e+00
-3.17701012e-01 -5.07350385e-01 4.88277584e-01 8.22130978e-01
2.99692135e-02 3.07797462e-01 -4.94953319e-02 1.43549725e-01
-7.00421453e-01 7.48482466e-01 5.68594038e-01 -3.78361106e-01
-2.49696195e-01 1.18440963e-01 -4.29943532e-01 7.66165137e-01
7.13777363e-01 -2.83542782e-01 -3.87608081e-01 -5.66506386e-01
6.26157075e-02 -1.44150063e-01 -1.39489351e-02 -1.02555001e+00
2.28770494e-01 2.77078718e-01 -1.15215674e-01 -6.21406436e-01
3.84238243e-01 -5.98975480e-01 -4.71935332e-01 1.32596985e-01
-5.67512631e-01 -1.97531462e-01 3.62223864e-01 5.03180325e-02
-4.64488447e-01 -4.24096853e-01 5.09301305e-01 3.58188496e-04
-8.04325759e-01 -9.03155878e-02 -1.08584380e+00 6.81574047e-01
6.37466252e-01 2.07479358e-01 -3.45357418e-01 -7.94471145e-01
-8.20412397e-01 -1.75186768e-02 1.38465643e-01 9.78377879e-01
5.19612394e-02 -1.08482027e+00 -9.65074837e-01 5.12175560e-01
6.75259054e-01 -3.74779493e-01 9.55751911e-02 5.99609017e-01
-4.32697326e-01 1.04660642e+00 3.61637026e-01 -6.82504773e-01
-1.13435018e+00 1.41553327e-01 3.57235014e-01 -1.56233326e-01
-3.60568315e-01 1.01090610e+00 2.06043616e-01 -1.33919787e+00
4.27338004e-01 -5.21993220e-01 -1.02929644e-01 2.39248782e-01
5.44846058e-01 2.84622103e-01 3.44641387e-01 -1.03226566e+00
-6.06129944e-01 2.00063996e-02 -1.26310423e-01 -6.24788642e-01
8.33262146e-01 -4.89866942e-01 1.35943040e-01 9.79595602e-01
1.15668857e+00 5.61135888e-01 -8.78184438e-01 -1.57811940e-01
1.19283959e-01 1.48457602e-01 -3.84191237e-02 -9.70701635e-01
-4.14029986e-01 8.90013576e-01 2.43122473e-01 2.80188352e-01
5.56561887e-01 3.35408896e-01 7.32450366e-01 7.25411355e-01
4.07433301e-01 -1.46944177e+00 -2.86247097e-02 1.19912064e+00
1.09932554e+00 -1.44464266e+00 -6.47651374e-01 -4.04658407e-01
-9.78144288e-01 7.83410132e-01 3.92921895e-01 3.23620856e-01
7.31202126e-01 5.18457115e-01 5.78577638e-01 1.06332853e-01
-8.37803841e-01 -2.52240479e-01 2.93279737e-01 8.95117283e-01
1.00528789e+00 3.45487505e-01 -3.43680903e-02 9.33199346e-01
-8.04664195e-01 -3.15465838e-01 3.57559770e-01 5.48586965e-01
-2.70459682e-01 -1.22500646e+00 -2.67763734e-01 2.39138812e-01
-4.42014605e-01 -5.04749417e-01 -5.44693112e-01 7.12796688e-01
-7.82638788e-02 1.34317434e+00 1.81361720e-01 -1.56549513e-01
2.90068328e-01 5.58137655e-01 3.17256711e-02 -1.01085937e+00
-6.87898815e-01 5.72958030e-02 7.88433135e-01 -2.74673343e-01
-1.39591202e-01 -1.01386654e+00 -9.77614582e-01 -3.76934968e-02
-1.34971470e-01 3.66889425e-02 7.34119356e-01 1.13037181e+00
3.65976900e-01 3.17824751e-01 3.15003961e-01 -8.78437698e-01
-3.04489821e-01 -1.44094431e+00 -4.93408948e-01 4.91036326e-02
3.46562624e-01 -3.22728097e-01 -5.65432906e-01 1.35128528e-01] | [14.135594367980957, 6.935575008392334] |
3c006b22-fc97-42e9-9504-283f749519a2 | speech-dereverberation-with-a-reverberation | 2210.11089 | null | https://arxiv.org/abs/2210.11089v6 | https://arxiv.org/pdf/2210.11089v6.pdf | Speech Dereverberation with a Reverberation Time Shortening Target | This work proposes a new learning target based on reverberation time shortening (RTS) for speech dereverberation. The learning target for dereverberation is usually set as the direct-path speech or optionally with some early reflections. This type of target suddenly truncates the reverberation, and thus it may not be suitable for network training. The proposed RTS target suppresses reverberation and meanwhile maintains the exponential decaying property of reverberation, which will ease the network training, and thus reduce signal distortion caused by the prediction error. Moreover, this work experimentally study to adapt our previously proposed FullSubNet speech denoising network to speech dereverberation. Experiments show that RTS is a more suitable learning target than direct-path speech and early reflections, in terms of better suppressing reverberation and signal distortion. FullSubNet is able to achieve outstanding dereverberation performance. | ['Xiaofei Li', 'Wenye Zhu', 'Rui Zhou'] | 2022-10-20 | null | null | null | null | ['speech-denoising', 'speech-dereverberation'] | ['speech', 'speech'] | [-1.35048598e-01 -2.07055047e-01 3.74876171e-01 -1.34209841e-01
-4.14405853e-01 -2.25234732e-01 1.43716618e-01 -1.39445022e-01
-3.15510809e-01 7.14471579e-01 3.61956447e-01 -5.05250514e-01
-1.56816602e-01 -5.99779427e-01 -4.07983720e-01 -9.46702659e-01
-2.79000819e-01 -4.06967491e-01 1.58055335e-01 -5.03336251e-01
-4.17435169e-01 5.65219700e-01 -1.60187650e+00 1.65731892e-01
1.18250918e+00 8.62925768e-01 5.61570406e-01 7.25295782e-01
2.41324320e-01 6.44336164e-01 -1.36793661e+00 1.88359931e-01
1.55274063e-01 -6.87956989e-01 -9.23585817e-02 -3.81383955e-01
2.04339206e-01 -3.40961933e-01 -5.50179482e-01 9.64057267e-01
1.07864094e+00 6.19803309e-01 1.47560582e-01 -8.26610088e-01
-3.30371819e-02 8.97155285e-01 -9.70877632e-02 6.29245996e-01
7.13735521e-02 -3.70964105e-03 5.68092167e-01 -7.37423241e-01
8.23852140e-03 1.09967256e+00 1.03842926e+00 4.61188167e-01
-8.82677853e-01 -1.02745664e+00 1.51388809e-01 6.70318961e-01
-1.38798892e+00 -7.89568067e-01 1.04751015e+00 1.51038766e-01
8.43785405e-01 6.58131480e-01 6.21636212e-01 1.19463742e+00
6.95165545e-02 5.01250267e-01 1.18652368e+00 -4.01504993e-01
1.85572088e-01 -7.75793493e-02 3.24806184e-01 4.16439436e-02
-9.06446278e-02 7.41878390e-01 -3.47775549e-01 1.17768474e-01
5.09691417e-01 -4.70319033e-01 -9.59022820e-01 6.52543306e-01
-7.69107699e-01 1.08807348e-01 7.03666687e-01 7.62537658e-01
-2.26374343e-01 -1.20126847e-02 2.68210679e-01 9.42049921e-01
7.77570367e-01 2.26514369e-01 -5.35302103e-01 -2.75368057e-02
-8.90959322e-01 2.04836335e-02 7.89117634e-01 5.07722914e-01
3.17191035e-01 9.72511411e-01 -1.54626340e-01 1.24149609e+00
1.96850076e-01 7.38759100e-01 2.71401078e-01 -5.72452128e-01
3.38928729e-01 -5.83741486e-01 3.17522027e-02 -8.23093116e-01
-6.43768668e-01 -1.40085208e+00 -1.27676940e+00 3.69855583e-01
3.68980199e-01 -3.90272319e-01 -8.33192945e-01 1.72465563e+00
3.46707463e-01 5.38935125e-01 2.93577880e-01 1.08670080e+00
1.03178251e+00 1.10194790e+00 -2.88449168e-01 -8.97553861e-01
8.06985438e-01 -8.44425261e-01 -1.26759636e+00 -2.56791729e-02
1.57556981e-01 -9.97271657e-01 1.05442715e+00 1.00818753e+00
-9.34023857e-01 -1.08647597e+00 -1.16338670e+00 4.38377500e-01
2.92663202e-02 -1.64182723e-01 -3.48229185e-02 1.17645037e+00
-1.10357964e+00 6.95106447e-01 -6.95721447e-01 2.85328656e-01
-3.26866359e-01 -3.51796225e-02 5.16950823e-02 1.20405026e-01
-1.55092812e+00 8.38599443e-01 -7.96987489e-02 5.17028630e-01
-9.73486483e-01 -9.09637272e-01 -4.74601030e-01 1.52228609e-01
1.44652903e-01 -3.54978561e-01 1.29518056e+00 -1.24393833e+00
-1.66830003e+00 -1.22095291e-02 -1.66215405e-01 -5.53534746e-01
5.89136541e-01 -2.73996502e-01 -1.20985115e+00 -1.27681077e-01
-6.53363883e-01 -3.21153790e-01 1.29793990e+00 -1.53643394e+00
-3.44375223e-01 -1.58005074e-01 -2.27477670e-01 2.28053227e-01
-6.08040571e-01 -9.82566252e-02 1.71507820e-01 -1.13229823e+00
3.84084880e-01 -4.52305734e-01 -1.30615130e-01 -5.15454054e-01
-2.49911860e-01 1.26257196e-01 9.86949503e-01 -1.20617235e+00
1.52001381e+00 -2.31656384e+00 -4.53702122e-01 2.78103977e-01
-1.25606120e-01 7.97520936e-01 -3.25289756e-01 2.11528167e-01
-4.82313663e-01 -1.70448154e-01 7.38348216e-02 -3.16246867e-01
-3.22561860e-01 3.93164381e-02 -4.15481746e-01 5.60967803e-01
-3.69696528e-01 -1.25785530e-01 -7.67600536e-01 5.30633889e-03
2.04994142e-01 9.35704112e-01 -3.63003045e-01 3.10776889e-01
2.09845945e-01 5.47658026e-01 7.30367452e-02 2.15371490e-01
1.05509675e+00 8.02349687e-01 -1.82311893e-01 -4.08176064e-01
-3.93302530e-01 4.49277222e-01 -1.27415693e+00 1.01385915e+00
-9.51848507e-01 7.19784141e-01 7.63371348e-01 -8.06348205e-01
1.32902133e+00 5.95026612e-01 1.89686015e-01 -7.72074342e-01
1.61774635e-01 4.41562533e-01 3.99987251e-01 -4.92773265e-01
5.66122979e-02 -1.60894558e-01 7.74521112e-01 2.18853578e-02
-2.11611837e-01 -1.95707947e-01 -2.80449092e-01 -3.48248214e-01
1.06651676e+00 -1.37559384e-01 -3.99985015e-02 -3.61432374e-01
7.96429992e-01 -8.23045611e-01 7.65120685e-01 6.77086771e-01
-1.16783656e-01 6.99009955e-01 -2.18236730e-01 -1.26208887e-01
-6.40077293e-01 -1.26254368e+00 -2.32891321e-01 9.93556499e-01
-3.84341963e-02 -1.94927156e-01 -9.08260643e-01 -3.34319681e-01
-4.60189074e-01 1.17650032e+00 5.12544625e-02 -4.17709261e-01
-8.39046419e-01 -5.14959872e-01 6.54874742e-01 1.65079489e-01
6.68160915e-01 -9.15650904e-01 1.25934780e-01 3.46798122e-01
-5.08340240e-01 -5.56361914e-01 -7.49984920e-01 5.48231363e-01
-7.33256638e-01 -6.87520742e-01 -8.23542237e-01 -7.79416919e-01
4.02261525e-01 7.42857635e-01 7.30024040e-01 1.58180892e-01
1.47217900e-01 6.61110878e-02 -6.29577279e-01 -5.37425280e-01
-8.96099269e-01 -2.09779501e-01 2.18621895e-01 5.29838018e-02
-3.10942858e-01 -1.18072176e+00 -6.63051188e-01 4.65984881e-01
-6.89630270e-01 -3.45504016e-01 2.32545584e-01 8.56157184e-01
2.04278961e-01 7.94584930e-01 8.30073714e-01 -9.87677202e-02
9.85217273e-01 -2.07063124e-01 -4.72954124e-01 -6.10769144e-04
-4.86900806e-01 -5.39493561e-01 1.35564101e+00 -6.64180100e-01
-1.60730076e+00 -5.75328708e-01 -7.95167446e-01 -4.39553857e-01
-1.69114798e-01 3.47107053e-01 -2.95558363e-01 -8.09224620e-02
9.83422577e-01 2.97463745e-01 -1.07562222e-01 -7.38193393e-01
-2.45988935e-01 9.24232244e-01 3.90832961e-01 -1.01330012e-01
9.51257706e-01 1.29259855e-01 -1.76662073e-01 -1.41279733e+00
-3.69123340e-01 -4.64871407e-01 6.98362663e-02 -3.68566692e-01
2.64924616e-01 -8.52076054e-01 -5.86776674e-01 7.39552021e-01
-1.43261182e+00 -4.21543568e-01 -1.38319716e-01 7.89642930e-01
-1.58388481e-01 4.19182807e-01 -8.19013536e-01 -1.15099907e+00
-6.41817868e-01 -9.00990009e-01 1.29092321e-01 1.48044840e-01
-3.78154218e-02 -7.51986504e-01 -7.71805122e-02 -1.22534242e-02
7.67325580e-01 -7.03233406e-02 5.71660519e-01 -3.66370916e-01
-1.64121091e-01 1.76267233e-02 4.06602293e-01 1.09518158e+00
2.78304577e-01 -2.09725887e-01 -1.30097461e+00 -5.29686213e-01
7.17014551e-01 2.08255365e-01 7.98393488e-01 6.45615399e-01
1.12899995e+00 -2.36936375e-01 1.32654324e-01 6.68041587e-01
1.05073202e+00 6.76910877e-01 1.04934537e+00 2.01980636e-01
2.82553375e-01 5.49979210e-01 5.73835850e-01 3.80658358e-01
-1.94306195e-01 4.07983840e-01 3.68617445e-01 -6.51510835e-01
-7.39639938e-01 -1.47479698e-02 6.89402640e-01 1.36982906e+00
-1.94718838e-01 -4.98366207e-01 -5.85015357e-01 3.12478811e-01
-1.24922574e+00 -1.06226528e+00 -6.15071356e-01 2.33191729e+00
9.62457418e-01 4.27656211e-02 -1.08698778e-01 6.90525115e-01
6.54219568e-01 4.05582190e-01 -3.75651330e-01 -5.53657591e-01
-2.89933383e-01 5.54473758e-01 3.61473709e-01 8.02072823e-01
-6.79269850e-01 3.99097502e-01 5.79098463e+00 1.09940505e+00
-1.33557749e+00 3.34641814e-01 6.04244769e-01 3.69791500e-02
-3.27311486e-01 -3.95480424e-01 -4.93839920e-01 1.74588427e-01
9.76125121e-01 -1.56186372e-01 7.08263814e-01 5.84536076e-01
1.02619636e+00 2.06304610e-01 -6.27425671e-01 9.80701327e-01
5.39110452e-02 -5.34268200e-01 -3.54275435e-01 -4.85513985e-01
5.06187320e-01 -9.07315835e-02 4.95555997e-02 4.36984122e-01
-1.38743594e-01 -8.60263944e-01 7.86381364e-01 4.14349675e-01
4.88130867e-01 -1.03262579e+00 6.56695485e-01 5.69156229e-01
-1.13867378e+00 -3.51423711e-01 -3.54626775e-01 1.90908164e-02
-1.17554404e-02 1.18819058e+00 -9.12390530e-01 6.80559039e-01
7.16502666e-01 2.17449039e-01 8.00367817e-02 1.40542650e+00
-4.77215052e-01 1.09897387e+00 -3.78932327e-01 7.43488222e-02
-1.59527943e-01 -9.22724083e-02 1.06345284e+00 1.31769443e+00
6.95389152e-01 1.28995597e-01 -5.55078723e-02 2.95304090e-01
1.78227454e-01 3.87240127e-02 -2.34534502e-01 6.66758657e-01
5.89636981e-01 8.06347072e-01 -2.56579131e-01 -8.98931399e-02
-1.48129591e-03 7.14533687e-01 -5.26856959e-01 9.70738709e-01
-9.59489763e-01 -8.72060895e-01 7.61113644e-01 1.26223549e-01
1.78885892e-01 -1.56419694e-01 -1.67492002e-01 -7.47011304e-01
9.12274942e-02 -1.17896104e+00 -1.48716557e-03 -7.77656198e-01
-9.36529994e-01 1.03066027e+00 -2.40188658e-01 -1.40759408e+00
1.44864246e-01 -3.18272024e-01 -8.57719481e-01 9.82842743e-01
-1.45852244e+00 -6.42186701e-01 -3.25636119e-01 6.98573411e-01
6.75437272e-01 5.09542339e-02 7.82048941e-01 6.51966393e-01
-3.97851378e-01 8.12393546e-01 3.22319567e-01 -5.32600701e-01
9.18895125e-01 -1.01610398e+00 6.90310970e-02 9.70611632e-01
-8.83834586e-02 5.22746444e-01 1.40256488e+00 -4.35187250e-01
-9.69017565e-01 -1.15718150e+00 8.39765251e-01 4.13368016e-01
2.58345246e-01 -3.53348166e-01 -1.05716956e+00 1.13519549e-01
2.36918539e-01 -1.92273095e-01 3.80374938e-01 2.00987279e-01
-3.12619328e-01 -1.13337123e+00 -1.06528413e+00 7.83138037e-01
9.98857200e-01 -3.52226645e-01 -5.33611059e-01 2.60759145e-01
9.36829388e-01 -3.10031652e-01 -5.02384365e-01 4.57540572e-01
2.28097424e-01 -1.24075305e+00 1.06672692e+00 1.79031163e-01
-4.57472727e-02 -4.23744917e-01 2.79924385e-02 -2.09028339e+00
-2.63333946e-01 -1.13406050e+00 -1.53656930e-01 1.44592786e+00
3.91429275e-01 -9.40362275e-01 2.86842316e-01 -3.32360864e-01
-7.85655141e-01 -2.44193777e-01 -1.10176778e+00 -1.37195921e+00
-8.03063512e-02 -7.06574023e-01 5.26060939e-01 6.89172506e-01
-2.08001748e-01 3.77604812e-02 -5.05720794e-01 4.97400790e-01
7.00293839e-01 -3.59494299e-01 5.02617419e-01 -7.60713756e-01
-4.59694028e-01 -3.46777946e-01 2.82500386e-01 -1.39498270e+00
2.04925369e-02 -5.09530127e-01 6.51979029e-01 -1.37482023e+00
-8.61821234e-01 -4.21612412e-01 -6.39630377e-01 -1.22090457e-02
-1.71008319e-01 2.10439879e-02 -1.06750578e-02 -3.62007529e-01
1.52485967e-01 7.62401938e-01 1.34299624e+00 -1.31549433e-01
-6.77140772e-01 7.40091205e-01 -3.57593089e-01 7.96900213e-01
8.52184474e-01 -3.94241005e-01 -8.83219182e-01 -2.83471972e-01
-3.40297669e-01 3.63804936e-01 1.46779776e-01 -1.21008945e+00
1.64008692e-01 2.07623184e-01 6.89260289e-02 -5.80883324e-01
2.95710295e-01 -1.05047798e+00 3.74595940e-01 6.70422733e-01
-1.46288961e-01 -5.36866844e-01 4.26555395e-01 4.44800079e-01
-4.45220768e-01 -4.85019013e-03 1.01075792e+00 1.27531782e-01
-1.67501390e-01 8.95495061e-03 -7.72310853e-01 -4.16701347e-01
4.96561795e-01 -2.71314293e-01 -1.29166588e-01 -7.09580302e-01
-8.30476522e-01 -1.34929657e-01 -3.68198037e-01 2.89738685e-01
7.45485246e-01 -9.49775338e-01 -9.28679407e-01 1.66966781e-01
-6.39469504e-01 -4.41336095e-01 6.50714278e-01 8.70688438e-01
-2.24384576e-01 -3.90758552e-02 1.89302102e-01 -3.85073870e-01
-1.69882154e+00 3.01006347e-01 6.86896563e-01 -3.36165726e-02
-6.70373797e-01 1.02411532e+00 9.50640887e-02 -1.87031701e-01
7.30368495e-01 -4.68563616e-01 -3.11581761e-01 -1.71669528e-01
8.76280069e-01 7.28323519e-01 5.41111410e-01 -2.48984531e-01
-8.14478099e-02 1.63129643e-01 1.82442278e-01 -5.22009954e-02
1.20716643e+00 -3.77893269e-01 -1.37458906e-01 4.17004615e-01
1.14241850e+00 3.72933030e-01 -1.05073452e+00 -9.35748369e-02
-3.86550158e-01 -3.26934844e-01 5.28310955e-01 -1.00540066e+00
-1.00427747e+00 6.91422284e-01 8.76024306e-01 4.87607211e-01
1.84708333e+00 -6.58459127e-01 1.11346364e+00 4.25096512e-01
4.21047360e-01 -1.03782189e+00 1.30116776e-01 7.68295705e-01
1.30808425e+00 -5.04037917e-01 -3.87224704e-01 -3.79109681e-01
7.98557326e-02 1.34532320e+00 5.04685104e-01 5.57682142e-02
9.52333212e-01 5.04617929e-01 4.41458553e-01 5.47025204e-01
-3.90007406e-01 -2.40579359e-02 4.79524955e-02 8.58075738e-01
5.00128090e-01 -1.33812815e-01 -4.71561193e-01 7.32431650e-01
-6.42742991e-01 -4.88741845e-01 3.18938822e-01 3.75917792e-01
-6.60365760e-01 -8.94912481e-01 -1.00483429e+00 1.31396577e-01
-4.14560229e-01 -2.88242102e-01 -1.25266299e-01 3.89625847e-01
3.27731162e-01 1.69485056e+00 -2.47375052e-02 -4.46954519e-01
9.96640682e-01 -2.74559915e-01 3.03130686e-01 -4.61636260e-02
-1.07643175e+00 6.48716629e-01 4.39158410e-01 -1.63422570e-01
-1.37024447e-01 -3.13597411e-01 -1.12532580e+00 -4.95646447e-01
-6.92143977e-01 2.60852247e-01 6.11889184e-01 8.31762135e-01
-2.00875312e-01 1.15090263e+00 1.18716061e+00 -5.88936925e-01
-7.67557919e-01 -1.23399770e+00 -8.04359913e-01 1.43315494e-01
9.36276793e-01 -1.68091446e-01 -8.70693505e-01 -3.93557027e-02] | [15.091310501098633, 5.902318477630615] |
8eb28b3a-6df1-4829-987d-c899f29a5481 | machine-learning-model-interpretability-for | 1610.09045 | null | http://arxiv.org/abs/1610.09045v1 | http://arxiv.org/pdf/1610.09045v1.pdf | Machine Learning Model Interpretability for Precision Medicine | Interpretability of machine learning models is critical for data-driven
precision medicine efforts. However, highly predictive models are generally
complex and are difficult to interpret. Here using Model-Agnostic Explanations
algorithm, we show that complex models such as random forest can be made
interpretable. Using MIMIC-II dataset, we successfully predicted ICU mortality
with 80% balanced accuracy and were also were able to interpret the relative
effect of the features on prediction at individual level. | [] | 2016-10-28 | null | null | null | null | ['icu-mortality'] | ['medical'] | [ 1.58799425e-01 6.54647946e-01 -3.32412750e-01 -6.22745335e-01
-2.01577663e-01 -3.55732650e-01 1.13200076e-01 4.94102895e-01
1.39865577e-01 1.24478698e+00 3.49652052e-01 -1.09296668e+00
-6.66808486e-01 -4.85735387e-01 -5.38781941e-01 -2.84351766e-01
-2.28892446e-01 1.05718613e+00 -3.44981283e-01 7.24965148e-03
3.11116934e-01 4.33256298e-01 -9.22348380e-01 5.23282051e-01
8.34331632e-01 6.02018237e-01 -3.49330187e-01 1.05455720e+00
1.18532382e-01 1.24032891e+00 -2.63214439e-01 -1.33741170e-01
9.58466902e-02 -5.82557917e-01 -9.08080578e-01 -2.48884454e-01
-2.15668738e-01 -1.46058366e-01 -7.77983665e-02 1.27427831e-01
1.33758545e-01 -4.09460813e-01 1.17372978e+00 -1.45114040e+00
-4.79056418e-01 8.29475820e-01 -7.38090351e-02 2.11652756e-01
2.56996214e-01 4.24591303e-01 8.60328078e-01 -5.13615727e-01
2.21322402e-01 1.23442709e+00 8.99320483e-01 6.22033477e-01
-1.36504388e+00 -5.94040394e-01 8.07550475e-02 2.72067815e-01
-8.95085037e-01 -5.53042218e-02 2.15578437e-01 -7.30954826e-01
9.36183274e-01 7.10030258e-01 6.68486059e-01 8.02726150e-01
6.92245960e-01 8.57820082e-03 1.44536996e+00 -4.67880458e-01
1.02258459e-01 3.62280868e-02 7.09350228e-01 9.98739719e-01
7.29531944e-01 4.29172128e-01 -3.89476359e-01 -3.76065612e-01
6.54048443e-01 6.23856127e-01 -3.20850939e-01 3.05968225e-02
-1.25281477e+00 1.01080596e+00 4.79720056e-01 -2.16825083e-01
-5.32243013e-01 3.67052108e-01 2.39012539e-01 1.31752910e-02
-7.74321482e-02 7.47105002e-01 -1.15210545e+00 -8.06259662e-02
-4.23275620e-01 3.01252771e-02 8.71862710e-01 6.83370709e-01
2.47061715e-01 -1.03776909e-01 -2.03604311e-01 1.98798507e-01
4.68361706e-01 3.22778970e-01 4.64134216e-01 -8.22779715e-01
4.56962883e-02 8.82731199e-01 8.79907534e-02 -7.42336571e-01
-7.60513306e-01 -5.40111303e-01 -1.12693179e+00 2.01762930e-01
2.21396819e-01 -3.11220497e-01 -1.10438037e+00 1.32627583e+00
-1.51662782e-01 -1.04762673e-01 1.17051013e-01 5.33374667e-01
6.98530734e-01 -4.71167527e-02 6.54384553e-01 -3.63752723e-01
1.35833991e+00 -7.75633752e-01 -7.20372617e-01 -2.04789639e-01
8.00898850e-01 -2.78428435e-01 7.91459918e-01 4.28250402e-01
-8.85145724e-01 -2.03669563e-01 -8.27041805e-01 3.11560124e-01
-6.30547926e-02 -1.27338558e-01 9.31378782e-01 3.68882477e-01
-6.88442171e-01 9.52161491e-01 -1.00250494e+00 -3.11571300e-01
8.37354302e-01 6.16279125e-01 -3.71147901e-01 1.48984879e-01
-9.05700624e-01 1.42301357e+00 4.61077511e-01 -1.34482950e-01
-7.17085779e-01 -9.13853049e-01 -4.06801075e-01 2.34362692e-01
-9.19248164e-02 -1.50529397e+00 1.11025870e+00 -7.45008707e-01
-8.55299532e-01 5.24359345e-01 -6.23378754e-01 -8.12279999e-01
5.24862587e-01 -1.68899149e-01 -3.66754055e-01 -2.15307251e-02
-3.42479080e-01 3.16946477e-01 2.99133986e-01 -1.18929112e+00
-4.52420592e-01 -5.06097376e-01 -2.11707056e-01 -1.05590127e-01
3.92046154e-01 -2.19037950e-01 7.52814651e-01 -3.82978499e-01
1.69761002e-01 -1.04088938e+00 -7.04881430e-01 4.70984839e-02
-6.40631437e-01 2.01908112e-01 6.52682185e-01 -6.97707891e-01
1.33467031e+00 -1.64674985e+00 -2.31134772e-01 1.28214642e-01
8.80763054e-01 2.64570341e-02 5.80410302e-01 3.74182492e-01
-5.05355835e-01 8.60581279e-01 -2.31211290e-01 1.44794658e-01
-3.57726306e-01 5.13211668e-01 -2.00472444e-01 -3.65012558e-03
6.05784833e-01 1.08919311e+00 -7.00701833e-01 -5.99583507e-01
4.70338494e-01 1.99372619e-01 -6.22101307e-01 5.36331296e-01
5.77863757e-05 6.16453409e-01 -5.10757148e-01 4.75618750e-01
1.41261473e-01 -6.77870631e-01 3.35927278e-01 1.43374220e-01
2.08265111e-01 1.14682227e-01 -4.45100546e-01 5.35994112e-01
-4.48936582e-01 6.51730299e-01 -7.51325309e-01 -9.87887561e-01
7.46915519e-01 4.37372983e-01 1.21848714e-02 5.52515462e-02
8.42019990e-02 -9.67630744e-03 6.30246937e-01 -5.23532331e-01
-4.26587373e-01 -9.95431602e-01 2.96320796e-01 3.39758277e-01
-4.41317201e-01 -1.65275648e-01 -4.40768749e-01 -1.80925466e-02
1.22293818e+00 -2.15142190e-01 1.24848151e+00 -4.02133703e-01
1.90374538e-01 5.33445358e-01 5.17449737e-01 9.02698636e-01
-3.91024724e-02 8.83477688e-01 5.82826257e-01 -1.15401220e+00
-7.63946056e-01 -1.03601813e+00 -3.88020158e-01 3.58446777e-01
-3.73823643e-01 -2.97561437e-01 -3.00564200e-01 -8.09703052e-01
1.57988414e-01 1.14840376e+00 -1.11060154e+00 -5.03248036e-01
-2.83411145e-01 -1.03940523e+00 4.28978145e-01 8.25729549e-01
-9.68756527e-02 -6.31207883e-01 -8.25481117e-01 2.87380606e-01
8.04210380e-02 -6.59707725e-01 2.56900549e-01 8.11925948e-01
-1.54346979e+00 -1.71610820e+00 1.35279059e-01 -7.61853904e-02
9.25994277e-01 -1.02497153e-01 1.51387775e+00 7.58691311e-01
-3.13355714e-01 -1.07750744e-01 -2.34237969e-01 -1.20597196e+00
-6.33744180e-01 -2.56669402e-01 1.14618406e-01 -6.21952295e-01
4.27556962e-01 -5.32409608e-01 -9.27626491e-01 4.03942019e-01
-6.34877145e-01 4.45755631e-01 7.62173653e-01 1.12282407e+00
4.15156901e-01 -2.16000259e-01 7.72348344e-01 -1.48168814e+00
5.18248320e-01 -7.72920132e-01 2.57624835e-01 3.21105510e-01
-1.57062805e+00 3.31129044e-01 9.63119328e-01 -1.96483448e-01
-6.46679461e-01 -3.97264101e-02 1.82022348e-01 -6.55801296e-02
-3.58100981e-01 4.44079459e-01 3.96504670e-01 1.64376244e-01
8.19438577e-01 -1.18074812e-01 -3.58562805e-02 -3.55831683e-01
-1.93895221e-01 5.77871621e-01 -2.74959515e-04 -3.44752759e-01
5.41218221e-01 1.75953254e-01 6.15698695e-01 -1.58596158e-01
-8.46234083e-01 5.29505275e-02 -8.03820908e-01 -6.63736369e-04
8.42741430e-01 -6.92103446e-01 -7.19473124e-01 -3.40592116e-01
-9.50979233e-01 -1.60960808e-01 -2.22141430e-01 5.98534882e-01
-5.20153582e-01 -7.51410276e-02 -2.05583870e-01 -7.74558604e-01
-5.18763304e-01 -9.82252657e-01 6.47175670e-01 6.45223781e-02
-1.25197303e+00 -1.41257989e+00 1.41991213e-01 5.14597297e-01
3.92012119e-01 6.88993871e-01 1.81942129e+00 -1.31399298e+00
-2.71301389e-01 -2.52045482e-01 -2.25417510e-01 -1.08416334e-01
3.57879728e-01 4.17271793e-01 -8.19528461e-01 3.33309203e-01
-3.63591194e-01 1.43103050e-02 4.01707411e-01 5.77992976e-01
1.39754725e+00 -7.70144105e-01 -6.17146611e-01 4.30761576e-01
1.27270257e+00 1.77044377e-01 4.37855184e-01 1.79352939e-01
6.20130241e-01 4.67488438e-01 4.86384302e-01 5.13435185e-01
3.51102650e-01 2.88636386e-01 2.00568378e-01 -2.37541348e-01
1.22542284e-01 -4.31456804e-01 -3.38281035e-01 3.97881567e-01
-5.51016569e-01 1.02761842e-01 -1.32502449e+00 7.36493841e-02
-1.96951056e+00 -6.06607735e-01 -8.40915918e-01 1.79004717e+00
7.04603612e-01 2.05612764e-01 3.03176674e-03 2.86321580e-01
2.91189134e-01 -6.26470685e-01 -5.32689929e-01 -9.83453095e-01
3.81104350e-02 2.12787002e-01 4.14871901e-01 4.13126349e-01
-5.29775620e-01 3.84481102e-01 8.54396343e+00 -1.69662517e-02
-7.28281200e-01 -1.53141141e-01 1.20930362e+00 6.12343997e-02
-2.86495388e-01 3.81054729e-01 -1.69621095e-01 3.36766750e-01
1.43198168e+00 -3.85245532e-01 6.97739702e-03 8.04402590e-01
9.52156842e-01 -4.08542007e-02 -1.61909354e+00 5.19712329e-01
-6.69974089e-01 -1.40021169e+00 2.63766021e-01 -8.94988850e-02
5.04872322e-01 -3.78534913e-01 -3.39307725e-01 1.19785130e-01
5.59000611e-01 -1.84031832e+00 3.01561236e-01 8.85151267e-01
5.98560810e-01 -4.15817589e-01 1.05202103e+00 3.30294043e-01
-4.03516233e-01 -2.59236842e-01 -1.64264247e-01 -8.40872109e-01
-2.81169154e-02 5.88170886e-01 -1.51873708e+00 3.38999510e-01
6.07675195e-01 5.60466111e-01 -5.60984671e-01 7.94678092e-01
-3.34179223e-01 1.03355646e+00 7.48530179e-02 1.81687828e-02
-1.26995772e-01 1.44664034e-01 1.28003374e-01 1.09711754e+00
1.94185469e-02 5.75778723e-01 -3.09800684e-01 6.85440660e-01
3.95011455e-01 -5.06436974e-02 -7.40637660e-01 -1.46735236e-02
1.28167897e-01 8.19151342e-01 -3.89505565e-01 -5.70722759e-01
-3.08819916e-02 4.86019105e-01 3.27494234e-01 1.12363100e-01
-7.30550766e-01 3.87345940e-01 4.40209329e-01 3.80332768e-01
-3.98120135e-01 1.63928419e-01 -1.16841936e+00 -1.01350558e+00
-5.26328385e-01 -9.46441650e-01 7.73745716e-01 -1.11654592e+00
-1.05424774e+00 6.23327553e-01 7.32766017e-02 -1.04850078e+00
-4.96667266e-01 -6.59695745e-01 -7.87234962e-01 1.01002038e+00
-1.10930824e+00 -1.04668081e+00 -4.60642606e-01 1.51152581e-01
4.61969346e-01 3.64668891e-02 1.38738608e+00 -5.19761026e-01
-6.35354877e-01 1.00488052e-01 -3.10756341e-02 -2.37439752e-01
2.68053949e-01 -1.47011590e+00 -6.03668913e-02 1.50189996e-01
-4.95104492e-01 9.32334661e-01 1.07076383e+00 -7.24166512e-01
-1.06254196e+00 -1.15594423e+00 8.92912447e-01 -8.98325801e-01
4.93441075e-01 4.33250606e-01 -8.09353828e-01 9.71686542e-01
-7.82875940e-02 -2.34329328e-01 1.37468696e+00 2.27099791e-01
-9.45300609e-02 2.48568356e-01 -1.35311508e+00 5.85381746e-01
8.80683720e-01 1.63692251e-01 -9.64607060e-01 4.22943562e-01
4.65739369e-01 1.07809193e-01 -1.01763129e+00 6.50169313e-01
8.25960934e-01 -9.71643984e-01 8.90406847e-01 -1.82490158e+00
9.02189970e-01 -6.64333627e-02 5.18539436e-02 -1.30753624e+00
-6.33085668e-01 -2.81499326e-01 -3.55707705e-01 6.24992073e-01
8.56995881e-01 -8.50727141e-01 6.11736178e-01 1.34268486e+00
2.31644601e-01 -1.23280489e+00 -4.55948055e-01 -3.81783336e-01
9.55545828e-02 -8.66637677e-02 6.35985732e-01 9.13899541e-01
3.25708181e-01 6.40688241e-01 -4.20081377e-01 3.05587649e-01
4.95919615e-01 8.04720595e-02 4.65120912e-01 -1.57800865e+00
-4.72069710e-01 -3.55727524e-01 -4.73264456e-01 6.96948618e-02
-1.79927945e-01 -6.55519605e-01 -3.54079604e-01 -1.69623971e+00
7.10423708e-01 -6.13134205e-01 -3.17885488e-01 6.21036172e-01
-8.24381411e-01 -1.53663695e-01 2.55211219e-02 4.15041000e-01
3.66720255e-03 3.31341289e-02 1.13918662e+00 1.44967899e-01
1.78095207e-01 2.77550042e-01 -1.12221837e+00 8.74920368e-01
1.07352591e+00 -8.28244627e-01 -4.16645467e-01 -3.14605944e-02
4.80070375e-02 3.36843580e-01 4.19296443e-01 -5.79848588e-01
-3.46695781e-01 -6.86522663e-01 8.79262984e-01 -1.59769341e-01
-1.17327049e-01 -1.00395238e+00 8.75705183e-01 1.23338771e+00
-6.39058948e-01 3.37956786e-01 2.13898286e-01 6.07079029e-01
4.61551696e-02 -1.14488296e-01 5.52036405e-01 -3.26057822e-01
-7.94071853e-02 1.38627350e-01 -3.90668422e-01 6.36514323e-03
1.06374252e+00 -1.04447961e-01 -2.41128191e-01 -4.86866891e-01
-8.13298643e-01 2.00889066e-01 3.31136227e-01 7.20061138e-02
7.21971631e-01 -8.61689866e-01 -7.86853671e-01 -1.78466454e-01
1.76536351e-01 -5.79615906e-02 7.56920204e-02 9.37648714e-01
-1.01316166e+00 7.85875499e-01 -1.64936513e-01 -3.31718981e-01
-1.38402092e+00 5.35054266e-01 6.55977666e-01 -3.57756317e-01
-4.52114731e-01 3.66622716e-01 4.17552054e-01 -2.73497701e-01
-2.25671202e-01 -4.51957613e-01 -2.04505414e-01 -5.55346131e-01
5.30369878e-01 5.09845674e-01 -2.13227719e-01 -1.04471803e-01
-7.66796827e-01 4.03704882e-01 1.50668800e-01 4.54294145e-01
1.52261257e+00 1.95465200e-02 6.85546473e-02 6.29518270e-01
7.07890749e-01 -5.60077250e-01 -8.61233890e-01 5.09286582e-01
1.31519765e-01 -4.10174012e-01 -2.37219438e-01 -1.50082111e+00
-5.89646161e-01 1.11584306e+00 4.76278782e-01 2.00116426e-01
9.91595566e-01 -2.25454137e-01 4.72992994e-02 4.55374956e-01
3.37744392e-02 -2.68545330e-01 -4.57403660e-01 -9.35967937e-02
9.29744363e-01 -1.63368976e+00 2.71784991e-01 -5.11937141e-01
-8.40206265e-01 1.36568844e+00 4.79126871e-01 -1.74094681e-02
6.29171669e-01 1.41998351e-01 2.03810036e-01 -2.23063275e-01
-1.37401605e+00 4.71280783e-01 4.64888781e-01 7.93650389e-01
6.20701909e-01 6.22660816e-01 -4.11045700e-01 1.09092641e+00
-1.98896825e-01 4.59077954e-01 6.34909809e-01 7.26608276e-01
-3.23553741e-01 -8.76519263e-01 -3.19905221e-01 1.10271788e+00
-6.73806131e-01 -4.43828702e-01 -6.23747170e-01 9.68057811e-01
-2.55487770e-01 9.93253767e-01 -3.75260413e-01 -4.72440779e-01
3.53840262e-01 4.55873251e-01 1.16016999e-01 -5.50398588e-01
-6.41246021e-01 -6.06370926e-01 4.61179793e-01 -3.05606723e-01
-1.01030162e-02 -4.90571618e-01 -1.48921001e+00 -4.72375304e-01
-2.96572093e-02 2.28048801e-01 3.25405002e-01 1.20801198e+00
6.99647784e-01 8.53690624e-01 5.66213727e-01 -3.49681862e-02
-5.62068284e-01 -1.08985054e+00 -2.98065633e-01 4.24078375e-01
5.08626103e-01 -5.81422269e-01 -7.12896645e-01 4.52929318e-01] | [8.322891235351562, 5.8629913330078125] |
7ad1ed59-28a0-46f7-ba52-687471a41ea8 | bi-vlgm-bi-level-class-severity-aware-vision | 2305.12231 | null | https://arxiv.org/abs/2305.12231v1 | https://arxiv.org/pdf/2305.12231v1.pdf | Bi-VLGM : Bi-Level Class-Severity-Aware Vision-Language Graph Matching for Text Guided Medical Image Segmentation | Medical reports with substantial information can be naturally complementary to medical images for computer vision tasks, and the modality gap between vision and language can be solved by vision-language matching (VLM). However, current vision-language models distort the intra-model relation and mainly include class information in prompt learning that is insufficient for segmentation task. In this paper, we introduce a Bi-level class-severity-aware Vision-Language Graph Matching (Bi-VLGM) for text guided medical image segmentation, composed of a word-level VLGM module and a sentence-level VLGM module, to exploit the class-severity-aware relation among visual-textual features. In word-level VLGM, to mitigate the distorted intra-modal relation during VLM, we reformulate VLM as graph matching problem and introduce a vision-language graph matching (VLGM) to exploit the high-order relation among visual-textual features. Then, we perform VLGM between the local features for each class region and class-aware prompts to bridge their gap. In sentence-level VLGM, to provide disease severity information for segmentation task, we introduce a severity-aware prompting to quantify the severity level of retinal lesion, and perform VLGM between the global features and the severity-aware prompts. By exploiting the relation between the local (global) and class (severity) features, the segmentation model can selectively learn the class-aware and severity-aware information to promote performance. Extensive experiments prove the effectiveness of our method and its superiority to existing methods. Source code is to be released. | ['Yuan Yixuan', 'Liu Jie', 'Chen Wenting'] | 2023-05-20 | null | null | null | null | ['graph-matching'] | ['graphs'] | [ 4.10613477e-01 1.94381565e-01 -4.86482143e-01 -5.61369538e-01
-9.93725181e-01 -2.73595184e-01 4.47878927e-01 4.66571897e-01
-2.67957121e-01 1.15901686e-01 4.07633483e-01 -3.48869830e-01
-2.46179730e-01 -7.36969352e-01 -3.74025017e-01 -5.30318797e-01
2.80822337e-01 1.48399323e-01 5.01280785e-01 4.00615099e-04
3.32769185e-01 1.28505051e-01 -1.36322081e+00 7.72397518e-01
1.08557093e+00 8.80721271e-01 6.95530951e-01 7.29018569e-01
-5.58731854e-01 1.04983485e+00 -3.48114848e-01 -1.65862605e-01
1.57467723e-01 -7.67994046e-01 -9.23636436e-01 5.91861904e-01
7.91684687e-01 -3.31642538e-01 -3.76488119e-01 1.37975240e+00
4.69950616e-01 -2.21054703e-01 5.79855263e-01 -1.25677872e+00
-7.61916280e-01 2.15689942e-01 -8.30411136e-01 4.46835190e-01
5.79798996e-01 5.16000748e-01 8.34204972e-01 -5.41275144e-01
5.18665612e-01 1.48266101e+00 2.74024099e-01 5.20835996e-01
-8.20976853e-01 -2.49829099e-01 5.16451299e-01 4.77289051e-01
-1.11661422e+00 2.84180045e-02 9.68687832e-01 -6.95875823e-01
7.74076283e-01 4.33995724e-01 6.94298863e-01 6.16997898e-01
4.28546339e-01 9.19710934e-01 1.23501384e+00 -3.72714341e-01
-1.47982806e-01 1.34544298e-02 4.04327095e-01 1.31189990e+00
-1.19954847e-01 9.33767296e-03 -5.47010660e-01 -1.02768943e-01
7.21820354e-01 1.04388878e-01 -4.32222277e-01 3.70766297e-02
-1.20733464e+00 8.34180236e-01 6.24921799e-01 2.99063861e-01
-3.41905028e-01 -2.48843998e-01 3.07785183e-01 1.47926822e-01
4.24081326e-01 3.63857523e-02 -2.34667897e-01 5.13951004e-01
-1.01236582e+00 1.15680126e-02 4.03951257e-01 9.57091510e-01
9.72247303e-01 -2.57533878e-01 -1.13803351e+00 8.92951787e-01
6.47267044e-01 5.88765502e-01 7.51488149e-01 -9.11633313e-01
4.94016349e-01 1.04163706e+00 -4.66165364e-01 -1.23752713e+00
-5.66186309e-01 -2.95212686e-01 -1.01185012e+00 -1.80186987e-01
7.43583441e-02 8.30223113e-02 -1.25380731e+00 1.63408852e+00
3.82316291e-01 1.88472539e-01 -1.29172415e-01 1.20223355e+00
1.53418469e+00 4.51172650e-01 2.97000557e-01 -5.74800014e-01
1.79108906e+00 -1.11489809e+00 -6.71237707e-01 -5.09105206e-01
8.24640453e-01 -8.32052529e-01 1.34805167e+00 -1.00491740e-01
-9.69045520e-01 -6.94966137e-01 -6.50792420e-01 -1.80021286e-01
-1.20585598e-01 1.11951865e-01 4.12367731e-01 1.52815506e-01
-1.22997355e+00 2.51288712e-03 -3.16309392e-01 -4.79730308e-01
4.39601719e-01 9.68722925e-02 -1.10364936e-01 -4.77051258e-01
-1.25260866e+00 6.99382126e-01 4.22276974e-01 -5.84570877e-02
-5.63974977e-01 -7.49429226e-01 -1.05037022e+00 -3.92564714e-01
5.25034249e-01 -1.19709194e+00 1.12596738e+00 -8.56160820e-01
-9.21400666e-01 1.32351458e+00 -3.71547371e-01 -1.52089149e-01
3.81778866e-01 4.77105707e-01 -2.77149588e-01 6.71037436e-01
4.18787867e-01 8.94363284e-01 1.00125921e+00 -1.16527402e+00
-8.70984256e-01 -4.65649068e-01 2.11495206e-01 4.88617092e-01
-1.98141783e-01 -4.80623059e-02 -8.28863740e-01 -6.91074371e-01
3.04973215e-01 -4.99318600e-01 -3.90046179e-01 6.62761182e-02
-4.63538826e-01 -4.56177056e-01 5.29929519e-01 -8.88801694e-01
1.30289304e+00 -2.14864159e+00 -1.06425330e-01 2.68389527e-02
6.98720396e-01 2.79455692e-01 -6.00886703e-01 6.89006895e-02
1.30782560e-01 -7.84887224e-02 -3.77232611e-01 -2.15365648e-01
-4.26159859e-01 3.60720903e-01 9.41903293e-02 2.95540959e-01
2.50822961e-01 1.28825629e+00 -1.16542459e+00 -1.21622968e+00
3.39909881e-01 2.23929256e-01 -5.42046607e-01 4.03880984e-01
-3.18218052e-01 5.69745421e-01 -7.54256904e-01 7.03899741e-01
6.50918663e-01 -4.50444549e-01 -1.79126501e-01 -6.93484902e-01
1.05658546e-01 -1.86382048e-02 -6.74418926e-01 1.70776427e+00
-3.83288294e-01 2.65640467e-01 5.72288036e-02 -1.06898665e+00
7.78603017e-01 1.08716488e-01 6.37069583e-01 -9.99983668e-01
-9.00388509e-02 -5.46011701e-02 -1.34445325e-01 -1.03534293e+00
5.65485656e-02 3.65270637e-02 1.89456746e-01 -3.42350677e-02
3.44972648e-02 -3.47249746e-01 2.52704620e-01 5.53487420e-01
9.97533798e-01 -3.33769619e-01 2.49773815e-01 -3.33968326e-02
8.91333401e-01 1.41469330e-01 4.39503670e-01 8.49791706e-01
-4.72975731e-01 7.81968832e-01 4.97358322e-01 -8.43124241e-02
-4.24680412e-01 -9.79827225e-01 9.53787789e-02 8.77616882e-01
4.13540900e-01 -5.13437212e-01 -8.44778419e-01 -9.33089137e-01
-1.13723755e-01 4.49759185e-01 -4.44013506e-01 -4.20117438e-01
-2.84586668e-01 -7.65052736e-01 2.25124314e-01 3.89111459e-01
7.06033945e-01 -1.04164183e+00 -2.82730103e-01 -9.21970513e-03
-3.58188689e-01 -1.31642985e+00 -9.68438148e-01 -3.78301054e-01
-7.84665108e-01 -1.28183973e+00 -8.70235145e-01 -1.11205602e+00
8.14982295e-01 5.36391556e-01 1.14436936e+00 3.59030902e-01
-5.12668490e-01 9.24223125e-01 -3.06857884e-01 -1.12243518e-01
-3.58567774e-01 -3.34895194e-01 -4.44812715e-01 2.17969656e-01
3.38620901e-01 -2.23876998e-01 -7.38113582e-01 -1.10078141e-01
-1.03000188e+00 4.72571313e-01 8.77067924e-01 8.04222226e-01
8.75849128e-01 -1.75828308e-01 1.98432535e-01 -8.91218543e-01
7.58917928e-01 -1.59806997e-01 -4.62119490e-01 5.69081843e-01
-7.64251173e-01 -1.65650293e-01 2.39657074e-01 -4.68056977e-01
-8.08146358e-01 8.04280415e-02 -1.30626157e-01 -4.77654696e-01
-3.34177554e-01 8.08902919e-01 -2.50851065e-01 -1.24255180e-01
2.85817713e-01 5.95203519e-01 -4.70637493e-02 -2.00145364e-01
5.57027459e-01 7.43931174e-01 7.63322949e-01 -2.64697880e-01
5.72427630e-01 2.57929742e-01 8.03031549e-02 -7.31833577e-01
-1.46926892e+00 -8.23516846e-01 -3.84949297e-01 -3.16990912e-01
1.22692513e+00 -8.43378246e-01 -5.87421775e-01 3.59982103e-01
-1.35274971e+00 -1.35815531e-01 -2.25869611e-01 3.93333107e-01
-5.78419089e-01 8.18597436e-01 -5.98678231e-01 -4.95081425e-01
-3.67440134e-01 -1.37188804e+00 1.34105623e+00 4.73819822e-01
1.96269661e-01 -1.19077551e+00 -2.43835017e-01 7.88704395e-01
7.61190709e-03 -1.62693672e-02 1.16497779e+00 -4.24697876e-01
-5.83718479e-01 4.61827358e-03 -9.11804080e-01 3.24992269e-01
2.28955150e-01 -3.79421175e-01 -7.34354138e-01 -1.38496980e-01
2.15825602e-01 -4.91295345e-02 1.06556678e+00 7.10715592e-01
1.31539536e+00 -3.59137654e-01 -2.46694237e-01 6.47883296e-01
1.32433200e+00 -4.39106598e-02 2.87865013e-01 2.46275831e-02
1.13024163e+00 7.48915195e-01 6.81828141e-01 1.69277042e-01
9.56111073e-01 4.23851818e-01 3.81991178e-01 -5.76640248e-01
-7.54041612e-01 -3.45490366e-01 2.24368304e-01 1.06683350e+00
3.17383528e-01 -1.19764805e-02 -8.98990810e-01 5.30520141e-01
-1.99797785e+00 -5.63732028e-01 -4.63326991e-01 1.82414854e+00
1.08520675e+00 -7.39967078e-02 1.36987995e-02 -3.25523555e-01
7.06652999e-01 2.97764450e-01 -4.77172285e-01 -2.21185729e-01
-1.15504220e-01 -3.27589780e-01 3.84871751e-01 8.47594619e-01
-1.06699860e+00 9.67065632e-01 5.44725037e+00 1.19479299e+00
-1.04598927e+00 2.77997494e-01 5.34563661e-01 3.29396695e-01
-5.49683213e-01 -4.39523682e-02 -6.93716645e-01 4.69299972e-01
3.90085936e-01 -1.31061956e-01 1.53652474e-01 3.91560137e-01
3.84644091e-01 -2.90334702e-01 -9.82927918e-01 1.29534721e+00
3.89987350e-01 -1.24850595e+00 5.26485324e-01 -1.01477899e-01
5.02364039e-01 -1.08890370e-01 4.61638719e-02 2.62508482e-01
-1.07220158e-01 -8.14898133e-01 1.53819114e-01 8.46778393e-01
8.02350223e-01 -3.32452714e-01 6.64775252e-01 4.41896826e-01
-1.42896795e+00 2.93339230e-02 -3.15991193e-01 2.77066648e-01
5.37119620e-02 7.56458938e-01 -7.51892447e-01 9.81979668e-01
5.45079887e-01 8.93459499e-01 -9.75744426e-01 1.05837679e+00
-3.57041031e-01 5.43685913e-01 2.23808885e-01 2.83677280e-01
4.01220977e-01 -2.58380711e-01 6.76563144e-01 1.36846018e+00
-1.39531046e-01 1.99637502e-01 8.08757424e-01 9.47086513e-01
3.10984135e-01 3.49896938e-01 -5.15457153e-01 2.66326219e-03
-9.64002311e-03 1.24019253e+00 -5.48310161e-01 -5.45403957e-01
-7.16551960e-01 1.01724303e+00 1.10074721e-01 6.12316012e-01
-5.04114449e-01 -1.63089857e-01 2.04080984e-01 1.52475551e-01
-2.49181896e-01 6.90208673e-02 -3.10535282e-01 -1.09811890e+00
4.19779532e-02 -8.28082085e-01 7.81791329e-01 -9.91808057e-01
-1.61027026e+00 4.78488356e-01 -5.24271429e-02 -1.30827355e+00
-1.03280984e-01 -5.18746436e-01 -4.42935139e-01 9.30857301e-01
-2.02428484e+00 -1.61162663e+00 -4.28484827e-01 1.13695097e+00
7.59676039e-01 -1.63317233e-01 4.96462345e-01 1.38227507e-01
-3.12829614e-01 5.37749887e-01 -7.16411114e-01 2.80291617e-01
7.71317184e-01 -1.14644718e+00 -3.45356017e-01 8.15966964e-01
1.33245379e-01 2.77639836e-01 3.74986231e-01 -8.24478924e-01
-1.20444822e+00 -1.51286781e+00 1.08175349e+00 -2.45728925e-01
5.40129364e-01 1.35817900e-01 -1.05281544e+00 2.70284265e-01
4.24885144e-03 1.66568100e-01 3.95711064e-01 -3.57741773e-01
-3.98257911e-01 -1.27192140e-01 -1.07029271e+00 6.88110471e-01
1.05063987e+00 -8.39631557e-01 -8.34472477e-01 8.27203572e-01
1.28458977e+00 -5.17814755e-01 -6.37745559e-01 7.39412189e-01
-4.20335941e-02 -7.39629269e-01 9.55423832e-01 -5.56361020e-01
4.17621672e-01 -4.07513916e-01 -2.31138662e-01 -9.11136270e-01
-3.81416708e-01 -3.05526882e-01 -2.89407466e-02 1.33297622e+00
1.44582242e-01 -5.14158845e-01 3.39834958e-01 2.40110785e-01
-2.38430977e-01 -8.88082504e-01 -7.40572333e-01 -5.04867673e-01
-8.97891074e-02 -4.35836881e-01 1.32239059e-01 1.00803900e+00
-2.86976874e-01 4.25241798e-01 -1.39662459e-01 5.01612842e-01
6.12864852e-01 4.36995894e-01 2.37703055e-01 -9.84866023e-01
-3.58329117e-01 -6.59741223e-01 -4.32777792e-01 -9.82612789e-01
2.05208063e-01 -1.29665017e+00 4.40343749e-03 -2.17644811e+00
7.18442619e-01 7.52923861e-02 -4.90203291e-01 6.60511315e-01
-6.38210177e-01 1.99121803e-01 3.26342374e-01 1.09306656e-01
-7.25283504e-01 2.82457262e-01 1.87763178e+00 -5.79946816e-01
-1.72982752e-01 -1.04641654e-01 -8.22438359e-01 8.55545878e-01
3.93294513e-01 -3.87604624e-01 -6.84633374e-01 -3.06436718e-01
-1.36530280e-01 4.09016609e-01 6.97180867e-01 -6.55091166e-01
5.44835329e-01 -5.41850567e-01 9.70456973e-02 -7.42891014e-01
8.90959054e-02 -4.49566066e-01 -7.19016910e-01 5.57675004e-01
-5.04788280e-01 -3.21113169e-01 6.40261024e-02 5.45217156e-01
-5.50768197e-01 -1.22131491e-02 5.97523689e-01 -3.25085461e-01
-9.09195542e-01 7.77570009e-01 -1.86170980e-01 3.96075070e-01
6.67806625e-01 -9.49976593e-02 -4.36502993e-01 -3.55305880e-01
-7.41210818e-01 7.03455448e-01 1.32130310e-01 5.23803413e-01
1.09224963e+00 -1.12363374e+00 -8.64064872e-01 1.75477117e-01
4.94255006e-01 1.11455537e-01 6.47326291e-01 1.24288380e+00
-1.18125223e-01 2.97975779e-01 1.92023128e-01 -1.05680776e+00
-1.38056540e+00 8.16455722e-01 4.42712963e-01 -3.49121779e-01
-5.79884171e-01 9.19414878e-01 7.75330186e-01 -3.22588354e-01
2.28566334e-01 -3.78708035e-01 -3.99572819e-01 2.39728764e-02
6.05799735e-01 -1.36326462e-01 -9.08336565e-02 -6.28468871e-01
-4.70744222e-01 1.08757472e+00 -3.89314853e-02 5.91539368e-02
7.25783467e-01 -4.92806822e-01 -4.41500753e-01 2.68894106e-01
1.09477162e+00 -2.51746804e-01 -8.89732838e-01 -5.13437569e-01
-1.32599309e-01 -2.60630220e-01 4.22964185e-01 -9.85984147e-01
-1.25290751e+00 1.03068471e+00 7.64744580e-01 -2.67569833e-02
1.64259684e+00 2.97164321e-01 8.70087981e-01 -1.09354332e-01
-1.21362181e-02 -7.74108291e-01 2.78602034e-01 3.74744415e-01
7.88132548e-01 -1.48736215e+00 -1.39621779e-01 -7.72691131e-01
-7.93118000e-01 8.58039081e-01 7.85829365e-01 1.09768234e-01
6.22294724e-01 1.43906204e-02 4.66234148e-01 -3.18816453e-01
-5.39450943e-01 -7.86040962e-01 1.10561550e+00 9.19200242e-01
3.04695189e-01 7.20571056e-02 -5.66266418e-01 7.53370404e-01
1.46815702e-01 -7.79146999e-02 1.68573156e-01 5.86587906e-01
-6.11406207e-01 -1.05196118e+00 -2.94141471e-01 4.80940729e-01
-9.56884176e-02 -4.17407691e-01 -4.27701801e-01 3.99985969e-01
4.66387808e-01 1.34057808e+00 -7.93112367e-02 -4.33703125e-01
2.75234163e-01 -1.29700154e-01 5.18894792e-01 -8.23585629e-01
-4.06655699e-01 2.84033358e-01 -1.86109498e-01 -8.42236519e-01
-5.68635941e-01 -2.12624505e-01 -1.52593493e+00 1.87241867e-01
6.18376061e-02 -2.61193246e-01 2.20108867e-01 1.27269089e+00
2.06636578e-01 7.60905385e-01 6.31302774e-01 -3.07094723e-01
-3.21728319e-01 -6.25293553e-01 -5.46439528e-01 5.12359500e-01
6.33320510e-01 -3.48104298e-01 -1.53474480e-01 1.82755753e-01] | [10.767301559448242, 1.4589191675186157] |
9c0a0fe6-f35b-4362-9895-4663133627e0 | wtr-a-test-collection-for-web-table-retrieval | 2105.02354 | null | https://arxiv.org/abs/2105.02354v1 | https://arxiv.org/pdf/2105.02354v1.pdf | WTR: A Test Collection for Web Table Retrieval | We describe the development, characteristics and availability of a test collection for the task of Web table retrieval, which uses a large-scale Web Table Corpora extracted from the Common Crawl. Since a Web table usually has rich context information such as the page title and surrounding paragraphs, we not only provide relevance judgments of query-table pairs, but also the relevance judgments of query-table context pairs with respect to a query, which are ignored by previous test collections. To facilitate future research with this benchmark, we provide details about how the dataset is pre-processed and also baseline results from both traditional and recently proposed table retrieval methods. Our experimental results show that proper usage of context labels can benefit previous table retrieval methods. | ['Brian D. Davison', 'Shuo Zhang', 'Zhiyu Chen'] | 2021-05-05 | null | null | null | null | ['table-retrieval'] | ['natural-language-processing'] | [ 7.03463936e-03 -2.07034677e-01 -5.55843353e-01 -2.14279518e-01
-1.73379672e+00 -1.04249310e+00 7.49343991e-01 8.01453471e-01
-2.75336146e-01 7.69034564e-01 4.99001563e-01 -1.87107801e-01
-1.53182879e-01 -9.45233881e-01 -5.50367236e-01 -9.92710367e-02
-5.68464436e-02 7.77723849e-01 8.98276567e-01 -4.71772999e-01
4.50913817e-01 -6.52306154e-02 -1.61302328e+00 8.56649339e-01
6.34239733e-01 1.31353986e+00 -6.71381652e-02 3.28121275e-01
-5.97774684e-01 8.72884870e-01 -8.19687128e-01 -8.61772656e-01
2.53243029e-01 -2.37326294e-01 -1.05321407e+00 -1.84235975e-01
9.58869159e-01 -2.38437921e-01 -3.28271240e-01 9.39151585e-01
3.33804190e-01 4.39330339e-02 7.29026556e-01 -1.14193487e+00
-4.46602076e-01 9.07753885e-01 -3.30959946e-01 5.03934801e-01
8.70613158e-01 -4.74002272e-01 1.57251573e+00 -8.32998335e-01
1.21361709e+00 1.11651158e+00 2.38950610e-01 4.30719033e-02
-9.64822650e-01 -4.50019985e-01 -5.65178059e-02 3.03313076e-01
-1.36134481e+00 -4.69667614e-01 5.20073295e-01 -9.80340987e-02
1.04914224e+00 5.11403859e-01 2.67569631e-01 1.04390156e+00
3.19232821e-01 7.70485103e-01 7.72095561e-01 -6.55405223e-01
-7.21779745e-03 4.27515149e-01 5.24187982e-01 2.98063546e-01
7.74215698e-01 -9.24477935e-01 -7.28983879e-01 -4.59312916e-01
3.32925409e-01 -5.19746780e-01 -5.76174865e-03 -7.29938447e-01
-1.36166537e+00 5.31509042e-01 2.35535040e-01 1.63579807e-01
-3.25497985e-01 -3.19693595e-01 6.99151754e-01 3.86457205e-01
1.37257189e-01 4.72135097e-01 -6.04063272e-01 -1.74923524e-01
-3.19039255e-01 5.86073101e-01 1.45934439e+00 1.65266395e+00
8.61611307e-01 -7.76194036e-01 -2.41317749e-01 9.04378951e-01
-2.03763018e-03 7.30849147e-01 2.55888462e-01 -1.02203059e+00
1.33540273e+00 8.50772083e-01 3.30364674e-01 -1.13682306e+00
2.03452837e-02 1.29936785e-01 -1.76030487e-01 -5.85633993e-01
5.18204629e-01 3.23363245e-01 -3.80539477e-01 1.19848526e+00
3.82243603e-01 -1.04691720e+00 1.79808110e-01 3.55349213e-01
1.03081000e+00 5.25785387e-01 -7.78115094e-02 -3.84188026e-01
1.92002988e+00 -8.02104652e-01 -1.19127476e+00 -3.61851275e-01
7.25767851e-01 -1.08136606e+00 1.46369112e+00 1.72547758e-01
-1.00028181e+00 -1.43070385e-01 -1.05202329e+00 -4.06804115e-01
-1.10847378e+00 -2.46203929e-01 5.14456153e-01 4.36627686e-01
-7.35509157e-01 1.08092867e-01 -4.08381015e-01 -9.68715668e-01
-9.53885689e-02 -7.10699707e-02 -4.00964022e-01 -2.85670608e-01
-1.36043394e+00 8.94763291e-01 4.69884813e-01 -4.55392480e-01
-2.52586484e-01 -2.99005955e-01 -7.89393902e-01 1.06146887e-01
1.23954868e+00 -2.96371996e-01 1.42752635e+00 2.68445779e-02
-4.95401651e-01 8.42499554e-01 -2.84361452e-01 -5.84666319e-02
1.06350921e-01 -2.80214578e-01 -6.65706933e-01 3.70205164e-01
6.35664165e-01 2.89273351e-01 1.43260911e-01 -1.18625951e+00
-8.67711544e-01 -3.37476462e-01 6.66937456e-02 3.79268587e-01
-3.60828698e-01 1.27678126e-01 -1.32649422e+00 -5.94998658e-01
6.53487816e-02 -7.05368817e-01 3.35705310e-01 -1.98659986e-01
-5.31190157e-01 -3.36232156e-01 5.83264172e-01 -8.33377779e-01
1.78515303e+00 -1.87680483e+00 -5.70203483e-01 3.97108376e-01
1.62942484e-02 -4.10000890e-01 -7.45792836e-02 1.30777979e+00
3.45369190e-01 5.81456304e-01 7.49149024e-02 4.07674551e-01
2.76762247e-01 3.01248878e-02 -4.13319826e-01 -1.46066874e-01
-2.46052593e-01 8.01551044e-01 -7.61438549e-01 -1.22928739e+00
-4.33037937e-01 5.50246872e-02 -3.08434248e-01 -5.87415025e-02
-5.18066108e-01 -5.58194160e-01 -7.97194183e-01 1.03982890e+00
1.09258436e-01 -5.37380755e-01 4.67279494e-01 -3.60691160e-01
2.53744125e-01 1.29549932e+00 -1.18525851e+00 1.47548628e+00
-7.01505840e-02 3.37905169e-01 -2.00202748e-01 3.58412229e-02
5.22044301e-01 3.12151194e-01 4.86829370e-01 -1.05851090e+00
-2.20205143e-01 3.04700077e-01 -1.81532592e-01 -5.11427224e-01
6.88042581e-01 8.51732135e-01 -4.23008740e-01 7.16885805e-01
-3.35788608e-01 -1.26800209e-01 1.32155395e+00 7.10961461e-01
1.29555106e+00 -1.34514878e-02 5.02472460e-01 -5.55440545e-01
6.33193672e-01 6.37153029e-01 2.02391759e-01 8.75985861e-01
6.29967973e-02 3.53394598e-01 7.66580224e-01 -2.65989572e-01
-1.03409886e+00 -1.11599505e+00 -3.03003281e-01 1.26930547e+00
5.05277216e-01 -1.24255860e+00 -7.53169000e-01 -1.22399342e+00
2.54007250e-01 5.68536401e-01 -7.72112787e-01 2.55744070e-01
-7.61820853e-01 -5.11867821e-01 5.84401898e-02 5.45509577e-01
3.51128757e-01 -1.18078053e+00 -1.74604356e-01 2.16213629e-01
-8.92584026e-01 -1.29947138e+00 -7.81611323e-01 3.15919101e-01
-6.43886566e-01 -1.45734811e+00 1.12661630e-01 -9.45207477e-01
2.20685005e-01 5.28274059e-01 1.91562676e+00 1.86545059e-01
3.68309319e-02 4.74296689e-01 -5.01310885e-01 -3.91681224e-01
-3.44491214e-01 5.81759870e-01 -2.73224771e-01 -7.96802998e-01
8.39681983e-01 -1.09570049e-01 -3.92246991e-01 7.27178931e-01
-1.09412277e+00 -3.42333764e-01 5.25828958e-01 6.73010707e-01
7.31860816e-01 3.10757726e-01 3.28627467e-01 -1.24982047e+00
8.59476030e-01 -4.37539667e-01 -7.65505493e-01 9.94064629e-01
-9.19542313e-01 3.44232529e-01 2.18053058e-01 -1.12967908e-01
-1.02728927e+00 -4.80346411e-01 1.42627656e-01 5.08640766e-01
3.04083228e-01 7.91538239e-01 -4.51496691e-01 3.98719192e-01
7.37865388e-01 -5.51424846e-02 -3.62978697e-01 -7.04142630e-01
1.44029096e-01 6.60781443e-01 3.15000236e-01 -8.92629564e-01
8.28552783e-01 1.64495841e-01 -3.02368551e-01 -2.71126002e-01
-7.93646216e-01 -6.34741962e-01 -7.46784329e-01 1.50855735e-01
5.90027690e-01 -8.01139295e-01 -5.25830567e-01 -2.49316558e-01
-7.55505979e-01 -3.48107517e-02 6.51743189e-02 1.16686441e-01
-2.04502985e-01 2.59440571e-01 -9.32885230e-01 -2.70182878e-01
-2.52336800e-01 -1.11963236e+00 1.13025057e+00 -1.96026444e-01
-4.72993582e-01 -8.76225054e-01 8.98267999e-02 4.37782764e-01
2.82546192e-01 -1.66862473e-01 1.49759424e+00 -1.22852015e+00
-8.67425442e-01 -5.37769794e-01 -1.70444787e-01 -3.53763551e-01
5.00135958e-01 5.03859892e-02 -4.65550125e-01 -6.41018376e-02
-3.43797863e-01 -5.39293766e-01 5.01031637e-01 -3.36783946e-01
6.91301465e-01 -7.11258888e-01 -6.45826876e-01 -7.32344836e-02
1.61351883e+00 5.25083303e-01 4.94155556e-01 9.39573586e-01
4.61479455e-01 8.17413449e-01 1.02356219e+00 2.76023746e-01
7.77267456e-01 7.90693283e-01 7.42263393e-03 4.99694943e-01
-1.48551494e-01 -5.75338483e-01 -4.07644967e-03 1.13997447e+00
8.37851346e-01 -6.86758161e-01 -8.47904861e-01 4.07102585e-01
-1.51817763e+00 -7.38290429e-01 1.47604883e-01 2.34320259e+00
1.23080087e+00 5.59790373e-01 1.57488525e-01 1.49710014e-01
5.22045195e-01 7.23854005e-02 -3.72354627e-01 -1.08514361e-01
-1.06948793e-01 -7.40348473e-02 4.60701257e-01 3.76695812e-01
-1.22889137e+00 8.87178004e-01 7.59780121e+00 7.97910273e-01
-3.47106695e-01 -4.36634511e-01 4.29528087e-01 8.96197334e-02
-4.78167713e-01 2.31782831e-02 -1.21235931e+00 2.01647207e-01
7.92753220e-01 -5.81778884e-01 4.39109951e-01 9.64810014e-01
-5.41217208e-01 -3.47916275e-01 -1.42985010e+00 6.52750194e-01
1.66463226e-01 -1.10925400e+00 5.61190188e-01 5.36529832e-02
3.59157085e-01 -2.70899951e-01 -2.00842261e-01 4.52633291e-01
2.94793546e-01 -2.69445628e-01 4.96390283e-01 -8.43090937e-02
7.22573578e-01 -5.22032380e-01 8.06907654e-01 -1.90152407e-01
-1.37311327e+00 3.85282397e-01 -2.40260363e-01 3.68995905e-01
-4.65568691e-01 1.86346635e-01 -8.01421404e-01 5.13890326e-01
1.15579689e+00 4.94522214e-01 -1.31471241e+00 6.82462871e-01
-1.42542079e-01 3.68937731e-01 -4.56279248e-01 -5.36206305e-01
-1.74377888e-01 -1.78318515e-01 1.86987132e-01 1.04411459e+00
-7.01097101e-02 -1.96911305e-01 1.26963466e-01 2.68129259e-01
-3.87067586e-01 6.72084808e-01 -7.93086112e-01 -1.40817165e-01
9.00288641e-01 1.27270639e+00 -1.09086585e+00 -8.09782684e-01
-7.09572852e-01 3.26324880e-01 3.54923904e-01 3.91809702e-01
-4.01617348e-01 -9.05767500e-01 5.62452257e-01 1.90867916e-01
4.99792665e-01 9.03053861e-03 -3.13963979e-01 -1.35443246e+00
8.14971805e-01 -1.21248186e+00 9.72419798e-01 -8.65412056e-01
-1.44284511e+00 7.99820781e-01 5.98950922e-01 -1.21703506e+00
-7.29588032e-01 -4.14686412e-01 8.42571817e-03 5.86230099e-01
-1.07310331e+00 -6.60265028e-01 -8.41569453e-02 4.68839139e-01
5.79599321e-01 8.52295198e-03 6.35310650e-01 3.87729824e-01
-4.32447255e-01 6.58654273e-01 2.66909301e-01 6.53874934e-01
1.33322716e+00 -1.53145146e+00 6.86183393e-01 8.61244678e-01
3.92826468e-01 1.11482489e+00 5.70226610e-01 -1.11019361e+00
-1.85502315e+00 -5.05664110e-01 1.28496397e+00 -1.12939751e+00
1.00839961e+00 -7.04311430e-01 -1.08375776e+00 8.75976682e-01
5.54955304e-01 -3.31974179e-01 6.56369627e-01 4.92378294e-01
-7.75548518e-01 -6.48989320e-01 -1.02002466e+00 9.16479588e-01
9.10366356e-01 -7.73529351e-01 -7.35410571e-01 4.85475719e-01
7.45114625e-01 -3.53314638e-01 -1.12154317e+00 3.82441640e-01
8.27441514e-01 -6.18330121e-01 1.02089989e+00 -2.67704755e-01
3.85435462e-01 -2.37976775e-01 -4.27179307e-01 -6.68437362e-01
-1.73053294e-01 -1.62186995e-01 -1.99741736e-01 1.50784600e+00
9.15674090e-01 -3.92141849e-01 6.78331912e-01 9.44477022e-01
3.45092624e-01 -4.18160260e-01 -4.84985769e-01 -7.71937728e-01
-7.27569759e-02 -1.19863763e-01 6.89204156e-01 7.51570880e-01
4.27029908e-01 7.31901526e-01 2.42555708e-01 -1.89273208e-01
5.04408419e-01 3.23949128e-01 8.25662732e-01 -1.14199471e+00
1.57997772e-01 -2.91832000e-01 1.49517938e-01 -8.21973443e-01
-3.53763461e-01 -7.80225694e-01 9.48723480e-02 -1.80950868e+00
3.58658880e-01 -1.92048490e-01 -2.37344012e-01 3.84342313e-01
-5.26159525e-01 1.22049145e-01 3.57972383e-02 6.36909544e-01
-1.18182933e+00 4.19242447e-03 1.12410688e+00 -1.97236568e-01
7.12656975e-02 -3.73864830e-01 -8.50169599e-01 1.18365668e-01
4.29455966e-01 -7.88771808e-01 -6.28505170e-01 -2.18185067e-01
5.80798090e-01 3.58065158e-01 -5.90970099e-01 -1.06970155e+00
2.99833000e-01 -2.75815070e-01 4.53668505e-01 -1.06378043e+00
1.32030621e-01 -9.74822462e-01 -1.75135657e-01 6.97941519e-03
-8.08069587e-01 8.21260631e-01 1.27380475e-01 4.60793346e-01
-4.75353152e-01 -1.20830029e-01 5.56254052e-02 -2.77633458e-01
-6.23506606e-01 1.21159121e-01 -6.93421185e-01 6.90130711e-01
4.21636313e-01 2.23007798e-01 -6.97766483e-01 -4.98820603e-01
-6.40536919e-02 4.24628735e-01 8.62605095e-01 6.72460258e-01
1.90974772e-01 -1.46876729e+00 -3.46537739e-01 -1.54902130e-01
9.07458544e-01 -3.28103989e-01 -4.50277686e-01 2.29623646e-01
-6.66568100e-01 8.27260077e-01 -7.88707882e-02 -3.04901272e-01
-1.14529037e+00 1.08904004e+00 -4.18372631e-01 -5.99844754e-01
-4.87899244e-01 8.82134438e-02 1.13977246e-01 -2.46098042e-01
5.39624095e-01 -4.79907632e-01 -3.47755343e-01 1.86269715e-01
8.35338354e-01 5.91202118e-02 6.44583941e-01 -4.94926311e-02
-5.10187864e-01 3.21794629e-01 -4.96842384e-01 -3.27333957e-01
7.03131735e-01 -4.65896875e-01 -3.15428913e-01 7.81964600e-01
1.25978124e+00 4.56215054e-01 -4.34156418e-01 -4.96582240e-01
8.47659290e-01 -5.41811585e-01 -5.84366739e-01 -1.25817716e+00
-6.55182123e-01 2.18999580e-01 7.18555078e-02 5.66273630e-01
9.66223538e-01 7.34400526e-02 8.37443531e-01 1.30933964e+00
6.19589627e-01 -1.15577328e+00 -6.42155204e-03 4.84076858e-01
8.27616751e-01 -1.45164776e+00 3.46355498e-01 -8.76583517e-01
-5.36840975e-01 9.92590070e-01 7.58131385e-01 3.11090171e-01
6.96247637e-01 4.67080712e-01 5.35158217e-01 -2.71106064e-01
-1.39777696e+00 -3.35631639e-01 5.04878640e-01 4.40724969e-01
6.72279716e-01 -6.41566575e-01 -5.83269536e-01 3.74304205e-01
-2.76645452e-01 -3.87321800e-01 4.62255955e-01 1.54873955e+00
-2.39188701e-01 -1.50236249e+00 -3.96089733e-01 8.03885937e-01
-7.04238057e-01 -3.93600047e-01 -8.03134620e-01 1.35291767e+00
-6.83687985e-01 1.03892148e+00 3.73954698e-02 -3.43163610e-01
5.15363872e-01 3.61527026e-01 2.53100187e-01 -5.81413090e-01
-8.35032761e-01 2.21081257e-01 6.95365012e-01 -6.45707130e-01
-6.90717101e-02 -6.39634252e-01 -9.21000540e-01 -1.91437334e-01
-1.80227593e-01 8.41215789e-01 4.92117971e-01 5.00523627e-01
2.61412174e-01 7.03866631e-02 3.15084994e-01 1.26347020e-01
-4.12223011e-01 -1.10986221e+00 -6.05689287e-01 7.00999796e-01
2.35853828e-02 -7.32251525e-01 -3.25130254e-01 1.99786946e-01] | [9.754591941833496, 7.930352210998535] |
8290ebf5-ada5-4ebc-8895-c8ed3e697603 | fooling-polarization-based-vision-using | 2303.1789 | null | https://arxiv.org/abs/2303.17890v1 | https://arxiv.org/pdf/2303.17890v1.pdf | Fooling Polarization-based Vision using Locally Controllable Polarizing Projection | Polarization is a fundamental property of light that encodes abundant information regarding surface shape, material, illumination and viewing geometry. The computer vision community has witnessed a blossom of polarization-based vision applications, such as reflection removal, shape-from-polarization, transparent object segmentation and color constancy, partially due to the emergence of single-chip mono/color polarization sensors that make polarization data acquisition easier than ever. However, is polarization-based vision vulnerable to adversarial attacks? If so, is that possible to realize these adversarial attacks in the physical world, without being perceived by human eyes? In this paper, we warn the community of the vulnerability of polarization-based vision, which can be more serious than RGB-based vision. By adapting a commercial LCD projector, we achieve locally controllable polarizing projection, which is successfully utilized to fool state-of-the-art polarization-based vision algorithms for glass segmentation and color constancy. Compared with existing physical attacks on RGB-based vision, which always suffer from the trade-off between attack efficacy and eye conceivability, the adversarial attackers based on polarizing projection are contact-free and visually imperceptible, since naked human eyes can rarely perceive the difference of viciously manipulated polarizing light and ordinary illumination. This poses unprecedented risks on polarization-based vision, both in the monochromatic and trichromatic domain, for which due attentions should be paid and counter measures be considered. | ['Yinqiang Zheng', 'Ko Nishino', 'Shohei Nobuhara', 'Zhihang Zhong', 'Zhuoxiao Li'] | 2023-03-31 | null | null | null | null | ['reflection-removal', 'color-constancy'] | ['computer-vision', 'computer-vision'] | [ 6.66549206e-01 3.23499382e-01 5.26474059e-01 2.20784381e-01
-2.66855001e-01 -1.09293973e+00 2.79885799e-01 -6.13680065e-01
-5.16199470e-01 4.61238861e-01 -2.46605054e-01 -7.50272930e-01
2.68258512e-01 -7.64275491e-01 -7.22613633e-01 -1.30713379e+00
5.87938488e-01 -1.91983148e-01 2.25976259e-01 -1.53888687e-01
4.98317540e-01 6.71779931e-01 -1.49734688e+00 -1.35007471e-01
7.49960721e-01 9.93260682e-01 -4.20296639e-01 7.87466347e-01
1.90784886e-01 1.06534123e-01 -4.90077078e-01 -7.26007104e-01
4.13067997e-01 -3.42544705e-01 -4.63182777e-01 6.10947683e-02
4.62042868e-01 -1.43652007e-01 -2.53875881e-01 1.48615217e+00
2.24576116e-01 -5.44563293e-01 5.00264943e-01 -1.20112205e+00
-1.00371301e+00 -2.37592593e-01 -7.56686687e-01 -1.93494380e-01
4.05284047e-01 8.67305577e-01 4.03148204e-01 -2.07696706e-01
4.27789927e-01 8.15708816e-01 3.49123925e-01 7.92551637e-01
-1.10943067e+00 -5.55436909e-01 -3.13492566e-01 9.51681510e-02
-1.01138234e+00 -3.73733401e-01 1.02688563e+00 -3.13051015e-01
6.46679103e-01 6.11820936e-01 7.78005123e-01 1.10665452e+00
4.55057651e-01 -2.60189269e-02 1.97071743e+00 -5.59199333e-01
3.09657659e-02 4.90565628e-01 -2.74817616e-01 7.09808230e-01
6.89011157e-01 5.39467692e-01 -4.12116051e-01 -1.23167731e-01
4.57686394e-01 -3.55615050e-01 -9.26400065e-01 -3.39732289e-01
-9.28120315e-01 8.94395858e-02 3.46724272e-01 -1.97721422e-01
-9.18351710e-02 -1.24968626e-01 -2.70581216e-01 1.83027953e-01
-8.45572799e-02 8.66772056e-01 1.00530185e-01 6.94309771e-02
-2.21080661e-01 -3.03368717e-01 7.05572486e-01 4.40371960e-01
4.05532181e-01 9.29044411e-02 3.84278029e-01 1.75714523e-01
6.94668353e-01 1.42895997e+00 -9.22751203e-02 -9.82527316e-01
1.89597845e-01 4.39582586e-01 3.60402524e-01 -9.38441515e-01
-2.40403250e-01 2.28229150e-01 -4.61776584e-01 1.00087833e+00
6.00149512e-01 -2.69749463e-01 -1.19752133e+00 1.48739159e+00
4.11374360e-01 -2.79342026e-01 4.20669854e-01 1.14048183e+00
5.60465097e-01 4.26066607e-01 -3.62136155e-01 -5.61144352e-01
1.40984607e+00 -2.96383739e-01 -3.82954895e-01 -1.14545263e-01
-9.21514332e-02 -1.09221041e+00 1.01126683e+00 7.14698434e-01
-8.20042372e-01 6.89373463e-02 -1.30828965e+00 2.18419373e-01
-1.13416523e-01 -5.00669718e-01 4.86461163e-01 1.75377750e+00
-7.22959876e-01 1.35670826e-01 -8.15053523e-01 -1.59876689e-01
2.94692010e-01 3.66888165e-01 -3.19846153e-01 -1.41267210e-01
-7.44692504e-01 7.18846023e-01 -2.21352369e-01 2.97151923e-01
-3.08456779e-01 -5.17220378e-01 -2.89563924e-01 -5.60849071e-01
2.29398146e-01 -5.01378179e-01 3.59978497e-01 -7.60936439e-01
-2.20378327e+00 1.22801745e+00 -7.45653287e-02 -1.57210380e-01
3.67047638e-01 1.00289628e-01 -8.12288105e-01 3.40548098e-01
-6.02417171e-01 1.23339131e-01 1.22920585e+00 -1.50257456e+00
-1.90406755e-01 -4.80384231e-01 1.24349847e-01 -3.86931896e-02
-1.06049448e-01 1.42071843e-01 -2.37070277e-01 4.44711238e-01
4.24665660e-01 -1.46756279e+00 4.89322878e-02 3.55192684e-02
-9.98091578e-01 5.44938087e-01 1.24507499e+00 -2.78210193e-01
2.60592043e-01 -2.40718770e+00 -3.84573609e-01 5.51425368e-02
1.07461274e-01 7.65084863e-01 -2.55643204e-02 1.60277292e-01
1.50613382e-01 1.78783327e-01 -1.60390973e-01 2.89540738e-01
-2.58067399e-01 -1.46953866e-01 -5.36710858e-01 9.85198438e-01
1.07688904e-02 7.50618517e-01 -5.02134860e-01 -2.78046429e-02
1.17275506e-01 7.94933021e-01 -2.19748124e-01 -6.47592843e-02
5.06420508e-02 8.09725463e-01 -3.03048342e-01 1.05297172e+00
1.20085311e+00 1.36501744e-01 2.51465619e-01 -4.42582428e-01
-4.55249399e-01 1.96882174e-03 -6.52210295e-01 7.72897482e-01
-1.32973492e-01 8.39775562e-01 1.48958981e-01 -3.89985412e-01
8.27899456e-01 1.71653852e-01 1.41894951e-01 -1.05768037e+00
8.56966972e-02 3.48792434e-01 9.85446270e-04 -6.13279283e-01
4.72076714e-01 -3.51680338e-01 -6.83223605e-02 4.33513135e-01
-6.87313259e-01 -4.97195542e-01 -4.49611962e-01 -1.33775428e-01
7.43472338e-01 1.51798412e-01 -3.92393112e-01 3.21079919e-04
3.35265964e-01 -5.18418401e-02 4.99613822e-01 4.24898773e-01
-2.44520262e-01 7.55476296e-01 5.50537527e-01 -3.61670762e-01
-8.75683069e-01 -1.45416653e+00 -1.88286602e-01 2.61602372e-01
9.16975558e-01 2.22215846e-01 -5.21999240e-01 -2.58898318e-01
-1.40158758e-01 5.35471261e-01 -6.74504414e-02 -2.14875773e-01
-3.66586715e-01 -9.56201375e-01 5.98029971e-01 -1.83990985e-01
6.50662482e-01 -7.03293622e-01 -1.08014715e+00 -3.01179677e-01
2.68967412e-02 -1.30862188e+00 1.45238116e-01 -2.85430759e-01
-4.35346246e-01 -1.57409334e+00 -4.70692098e-01 -3.91531363e-02
7.88376272e-01 3.95849735e-01 7.93381333e-01 -3.73272359e-01
-5.28125346e-01 9.60728526e-01 -4.18028198e-02 -6.84478581e-01
-5.77481747e-01 -6.67567611e-01 1.73609197e-01 3.29187840e-01
2.87752360e-01 -6.91584587e-01 -1.00717545e+00 4.46081072e-01
-5.42472899e-01 -1.14743672e-01 4.55276072e-01 2.59955138e-01
3.66518319e-01 -1.41515478e-01 -2.17142269e-01 -8.46885145e-01
2.14077502e-01 1.94712862e-01 -1.10024250e+00 3.21312487e-01
-6.86514437e-01 -2.48106614e-01 4.23309416e-01 -2.77230740e-01
-1.20892954e+00 -1.38100028e-01 1.74058080e-01 -1.00963823e-01
-3.50676864e-01 -6.11229129e-02 -3.66987258e-01 -8.44659150e-01
8.61088753e-01 2.77623773e-01 6.87558353e-02 8.32152441e-02
3.06363523e-01 6.46696627e-01 6.44168079e-01 -1.85240537e-01
1.01621342e+00 1.15339363e+00 5.64292252e-01 -1.42323244e+00
-3.65900546e-01 -2.81291902e-02 2.97248252e-02 -4.76651043e-01
9.66057420e-01 -5.81566334e-01 -1.50933850e+00 1.09404314e+00
-1.10776579e+00 3.45581658e-02 1.17067499e-02 3.01707327e-01
-2.32674867e-01 4.34999526e-01 -3.60882103e-01 -1.04941559e+00
-3.03475913e-02 -1.23722625e+00 5.87444365e-01 7.50670671e-01
3.39744896e-01 -6.02956295e-01 3.66119891e-02 7.97366977e-01
5.04711509e-01 6.83464050e-01 8.08463633e-01 3.67090255e-01
-1.06919336e+00 -2.62757152e-01 -1.52682498e-01 3.48003685e-01
8.25376585e-02 3.44390363e-01 -1.40478146e+00 -1.98126078e-01
2.67149299e-01 -2.92641848e-01 8.58562946e-01 1.87366784e-01
6.03962719e-01 -3.32794249e-01 -3.30045044e-01 1.05112660e+00
1.57904315e+00 4.31276828e-01 1.21334696e+00 3.48704726e-01
9.94532883e-01 6.70117974e-01 2.48400778e-01 -1.15456529e-01
2.23020986e-02 5.65690279e-01 9.63378072e-01 -3.50768715e-01
-4.36973833e-02 1.75064713e-01 3.73953909e-01 2.03898385e-01
-4.63881403e-01 -4.44801867e-01 -8.21590126e-01 4.50271405e-02
-1.03116584e+00 -7.62442768e-01 -6.33906066e-01 2.46990037e+00
6.27807021e-01 6.93317354e-02 -1.81675687e-01 4.83115204e-02
8.04494262e-01 3.80980857e-02 -6.00888968e-01 -6.02802038e-01
-4.05802518e-01 2.53352076e-02 9.72799242e-01 5.21237731e-01
-9.62468326e-01 6.43721819e-01 6.02898645e+00 -1.14816073e-02
-1.96390486e+00 -1.31592721e-01 4.09591764e-01 -1.69639319e-01
-7.83945501e-01 1.22989036e-01 -6.51957393e-01 4.65101451e-01
5.10453224e-01 1.78892866e-01 3.51311982e-01 2.06760988e-01
-1.79671749e-01 -3.70494068e-01 -5.71376860e-01 1.06979477e+00
9.27470177e-02 -9.63777065e-01 -2.25307480e-01 5.37562907e-01
4.75155473e-01 -1.02979891e-01 8.21656227e-01 -6.89019620e-01
-6.03364920e-03 -8.14546585e-01 3.81774336e-01 4.23178703e-01
9.33209956e-01 -5.60736597e-01 6.57604784e-02 -1.50744170e-01
-4.26239014e-01 -3.49111184e-02 -3.58476490e-01 2.37034783e-01
3.25290769e-01 6.06654465e-01 -6.18121624e-01 1.65818632e-01
7.22057343e-01 -8.99509937e-02 -1.09832659e-01 7.76177526e-01
-6.91067040e-01 5.99664330e-01 -4.18809414e-01 1.37879312e-01
3.77389938e-02 -4.45069700e-01 1.08912253e+00 5.28383434e-01
1.77684247e-01 2.07706437e-01 -4.64393139e-01 1.16371036e+00
2.01360986e-01 -6.64111972e-01 -7.33922541e-01 -1.61120608e-01
2.50481039e-01 1.37799776e+00 -7.32756317e-01 5.36426008e-01
-4.38500136e-01 9.92702901e-01 -5.22774100e-01 8.37759435e-01
-7.35022306e-01 -4.93315578e-01 8.80045533e-01 3.12462509e-01
2.13494971e-01 -4.83201563e-01 -3.78176063e-01 -9.98446286e-01
1.12230130e-01 -6.07074618e-01 -2.09464610e-01 -7.84116507e-01
-9.07916367e-01 3.63020033e-01 -5.45437396e-01 -1.09488904e+00
4.62794095e-01 -8.97003293e-01 -4.88323092e-01 1.02565622e+00
-1.72984028e+00 -1.19387257e+00 -1.77159190e-01 7.76221454e-01
-6.14980936e-01 1.05217718e-01 8.48561823e-01 -2.82662928e-01
-3.04465503e-01 4.67021614e-01 1.56500876e-01 -3.80069427e-02
8.11015368e-01 -8.10381174e-01 2.42549852e-01 1.27072108e+00
7.75681734e-02 6.23775601e-01 9.03994918e-01 -2.96807617e-01
-2.17457747e+00 -6.24771833e-01 2.54088074e-01 -5.59359729e-01
3.52913201e-01 -1.06268942e-01 -6.19605601e-01 1.57655254e-01
3.35083097e-01 1.81925446e-01 7.62936771e-01 -1.65231124e-01
-9.73981202e-01 -2.94567376e-01 -1.39867091e+00 7.13021100e-01
5.57299674e-01 -8.41359794e-01 -3.20162386e-01 2.25214139e-01
5.34645855e-01 -3.65939081e-01 -2.39641041e-01 3.46982807e-01
8.84961665e-01 -1.31951344e+00 9.87358689e-01 -1.22137144e-01
1.53009117e-01 -5.38257957e-01 -3.71998735e-02 -1.02248526e+00
1.03739463e-01 -1.06246901e+00 5.74182868e-01 1.11828256e+00
5.24414718e-01 -1.45052445e+00 8.33882391e-01 9.97309089e-01
-3.46435085e-02 -7.18556643e-02 -8.73926401e-01 -8.70300472e-01
-4.86162528e-02 -1.49471447e-01 3.70676547e-01 8.34490061e-01
-1.03415009e-02 -1.18799888e-01 -4.14766639e-01 9.10794735e-01
1.18860757e+00 3.44282717e-01 6.28493011e-01 -9.90364909e-01
-3.32782537e-01 -1.52995348e-01 -6.16937280e-01 -4.86294359e-01
-1.31133974e-01 -3.48694205e-01 3.96890799e-03 -9.55686510e-01
-6.45566806e-02 -3.08339745e-01 -1.02227010e-01 2.99691916e-01
1.48028955e-01 8.74125421e-01 3.71754616e-01 3.57267529e-01
-1.37986362e-01 -2.03009509e-02 1.32589447e+00 -2.81559438e-01
-5.48490584e-01 3.82040776e-02 -9.06925499e-01 6.64791226e-01
6.08901024e-01 -4.41180944e-01 -3.44441563e-01 -2.50161231e-01
8.37157309e-01 -7.05224872e-02 8.11641753e-01 -1.08736634e+00
3.59871715e-01 -3.02726001e-01 1.71521120e-02 6.91481456e-02
5.39721131e-01 -8.17100585e-01 2.80511320e-01 5.86239755e-01
3.32233042e-01 -7.36236989e-01 3.82220224e-02 6.15347981e-01
6.00872487e-02 8.34005028e-02 1.02826679e+00 -2.69528091e-01
-4.78202224e-01 8.61648936e-03 -4.32196647e-01 9.74765867e-02
1.07834220e+00 -5.66730022e-01 -1.37231731e+00 -1.78984657e-01
-1.63413957e-01 -5.43087721e-01 1.05133998e+00 2.71818414e-02
6.78766012e-01 -6.89534247e-01 -1.88944757e-01 3.72721046e-01
4.60102670e-02 -3.90139610e-01 5.04216909e-01 1.03252196e+00
-8.53251934e-01 3.15185905e-01 -3.89442533e-01 -6.77487493e-01
-1.35317147e+00 4.63754207e-01 3.92016232e-01 5.48287332e-01
-5.50950706e-01 9.19151366e-01 1.63127005e-01 1.10730216e-01
-3.84368092e-01 2.18582720e-01 8.91329199e-02 -4.77936476e-01
4.15506184e-01 2.78671056e-01 -1.01393171e-01 -4.49072689e-01
-5.86389005e-01 9.86953676e-01 1.59675062e-01 -1.78539261e-01
1.02918863e+00 -2.43970841e-01 -2.99620658e-01 9.63961035e-02
7.75771558e-01 7.43707061e-01 -1.38651097e+00 4.04705644e-01
-7.62819529e-01 -7.83840299e-01 -2.64529027e-02 -8.40849280e-01
-1.07462835e+00 1.06966627e+00 6.43841267e-01 6.68344676e-01
1.14741027e+00 -1.04230098e-01 6.85052156e-01 2.18978733e-01
4.77375507e-01 -9.25202131e-01 -5.80992782e-03 1.11476019e-01
4.25178677e-01 -1.13801062e+00 -7.69835785e-02 -7.54806399e-01
-5.80960453e-01 9.14215624e-01 3.82920653e-01 1.56045273e-01
5.94360769e-01 2.32870668e-01 6.47482038e-01 -2.09236979e-01
-1.78765759e-01 1.47583336e-01 8.35145041e-02 1.17739213e+00
3.90276052e-02 2.04294965e-01 -2.36942451e-02 -1.13516524e-02
-1.07000373e-01 -3.78033012e-01 9.89084005e-01 7.94439793e-01
-3.71271402e-01 -1.15159523e+00 -8.09274316e-01 -8.96134228e-02
-5.62061667e-01 1.36005357e-01 -5.47382295e-01 5.62833905e-01
6.65560812e-02 1.06256771e+00 -8.59859660e-02 -2.48794243e-01
3.29091787e-01 -6.54064119e-02 6.80330098e-01 -1.46932527e-01
4.55791876e-02 -1.48397148e-01 -3.29751819e-02 -7.43494689e-01
-6.43112659e-01 -4.68932211e-01 -8.65625679e-01 -3.06292415e-01
-2.04280540e-01 -3.90180022e-01 9.05352473e-01 7.84898818e-01
4.67750788e-01 -1.07295677e-01 7.52162576e-01 -5.80781817e-01
-2.06226781e-01 -7.60718808e-02 -9.22517955e-01 2.18084872e-01
4.35068011e-01 -4.00275618e-01 -6.61799133e-01 -2.50968844e-01] | [10.380586624145508, -2.677485227584839] |
16b11805-6f56-4b0c-8fbe-7eded3f7d154 | interpretability-of-multivariate-brain-maps | 1603.08704 | null | http://arxiv.org/abs/1603.08704v1 | http://arxiv.org/pdf/1603.08704v1.pdf | Interpretability of Multivariate Brain Maps in Brain Decoding: Definition and Quantification | Brain decoding is a popular multivariate approach for hypothesis testing in
neuroimaging. It is well known that the brain maps derived from weights of
linear classifiers are hard to interpret because of high correlations between
predictors, low signal to noise ratios, and the high dimensionality of
neuroimaging data. Therefore, improving the interpretability of brain decoding
approaches is of primary interest in many neuroimaging studies. Despite
extensive studies of this type, at present, there is no formal definition for
interpretability of multivariate brain maps. As a consequence, there is no
quantitative measure for evaluating the interpretability of different brain
decoding methods. In this paper, first, we present a theoretical definition of
interpretability in brain decoding; we show that the interpretability of
multivariate brain maps can be decomposed into their reproducibility and
representativeness. Second, as an application of the proposed theoretical
definition, we formalize a heuristic method for approximating the
interpretability of multivariate brain maps in a binary magnetoencephalography
(MEG) decoding scenario. Third, we propose to combine the approximated
interpretability and the performance of the brain decoding model into a new
multi-objective criterion for model selection. Our results for the MEG data
show that optimizing the hyper-parameters of the regularized linear classifier
based on the proposed criterion results in more informative multivariate brain
maps. More importantly, the presented definition provides the theoretical
background for quantitative evaluation of interpretability, and hence,
facilitates the development of more effective brain decoding algorithms in the
future. | ['Seyed Mostafa Kia'] | 2016-03-29 | null | null | null | null | ['brain-decoding', 'brain-decoding'] | ['medical', 'miscellaneous'] | [ 3.62688482e-01 1.02163658e-01 2.73874968e-01 -7.41973758e-01
-4.20025975e-01 -3.13706994e-01 4.04208660e-01 3.29778016e-01
-5.15252948e-01 6.74134851e-01 1.42813427e-02 -3.93071294e-01
-6.63265944e-01 -2.83888757e-01 -5.98091781e-01 -7.80963957e-01
-2.46905819e-01 4.52156991e-01 -1.58291221e-01 2.95618594e-01
3.13298523e-01 5.88855922e-01 -1.68033934e+00 -2.40141135e-02
1.24091125e+00 1.33237922e+00 4.86971170e-01 3.07627857e-01
2.02808350e-01 1.07366912e-01 -5.29942334e-01 -3.37537348e-01
-8.82907957e-02 -8.28086019e-01 -4.32592750e-01 5.37713990e-03
1.28082320e-01 1.54275879e-01 9.56786647e-02 1.31956267e+00
4.06297266e-01 1.30941004e-01 1.06458569e+00 -1.20119798e+00
-5.81419170e-01 6.67457521e-01 -1.34502739e-01 3.99390817e-01
4.24173884e-02 -5.60492277e-02 1.00241113e+00 -8.99090469e-01
1.18377365e-01 1.11669087e+00 2.14770555e-01 1.63400561e-01
-1.34689724e+00 -4.93270814e-01 1.42698243e-01 6.10414207e-01
-1.38293326e+00 -4.60285634e-01 5.62787235e-01 -6.55414045e-01
4.89342868e-01 5.77921808e-01 8.53970528e-01 8.00423086e-01
5.51920056e-01 2.94365108e-01 1.54372334e+00 -5.05050421e-01
4.62669879e-01 1.76324457e-01 6.00654185e-01 6.90682352e-01
6.74793303e-01 6.40715361e-02 -3.46938640e-01 -4.42028902e-02
4.52769727e-01 -2.83482611e-01 -3.86903822e-01 -1.73425555e-01
-1.14831483e+00 8.11477304e-01 4.00957346e-01 5.66849232e-01
-5.46217978e-01 6.40487224e-02 2.23813757e-01 1.18360622e-02
5.66962361e-01 4.73237902e-01 7.89003000e-02 1.19668409e-01
-9.76056755e-01 4.51602638e-02 4.67539608e-01 4.48030591e-01
5.21993160e-01 -8.19225088e-02 -1.41627699e-01 7.38591075e-01
5.08415163e-01 6.36444986e-01 5.35468221e-01 -5.90128839e-01
2.69033819e-01 3.35570872e-01 -2.50425994e-01 -1.28279829e+00
-7.66880810e-01 -5.19273400e-01 -9.36293125e-01 1.07850410e-01
4.42268461e-01 6.56964704e-02 -4.96841431e-01 1.80267572e+00
-6.38988540e-02 -6.25491515e-02 -1.98513821e-01 1.18014646e+00
4.76286739e-01 2.34694183e-01 2.00775921e-01 -3.41446519e-01
1.50029790e+00 -5.01071990e-01 -8.26966763e-01 -2.66448617e-01
2.49248073e-01 -2.92587221e-01 1.08628583e+00 4.39940780e-01
-9.19451356e-01 -3.85973305e-01 -1.11214757e+00 1.88125461e-01
-1.88699663e-01 2.09706426e-01 5.98425329e-01 7.08534896e-01
-1.04919791e+00 3.47798198e-01 -8.64092052e-01 -2.07119077e-01
2.90186614e-01 4.86818522e-01 -4.33830649e-01 1.41418427e-01
-1.08208466e+00 1.28326321e+00 6.23373389e-01 6.14707291e-01
-5.35416365e-01 -2.46545479e-01 -6.59970999e-01 2.64450073e-01
-7.21366704e-02 -5.22055686e-01 6.83438420e-01 -1.11787474e+00
-1.29487348e+00 7.63669968e-01 -3.81145239e-01 -3.31164360e-01
3.61299038e-01 2.78503835e-01 -3.98515224e-01 1.19135708e-01
-1.20004304e-01 4.18461174e-01 9.29417193e-01 -1.05939794e+00
-2.27670565e-01 -6.41957700e-01 -2.20853046e-01 4.64331582e-02
-2.96346366e-01 -5.68079054e-02 -1.10841587e-01 -4.48189884e-01
5.02827048e-01 -7.34525144e-01 -1.56419337e-01 -1.52727485e-01
-3.99090171e-01 -1.15984179e-01 8.39936361e-02 -8.34415674e-01
1.28160620e+00 -2.28847289e+00 6.28667831e-01 6.69854879e-01
5.97749233e-01 -2.67073989e-01 -6.66417703e-02 -2.93985605e-01
-3.33641946e-01 1.85443565e-01 -5.31496704e-01 -1.20048933e-01
8.80948529e-02 1.14677228e-01 2.97834948e-02 8.28792274e-01
2.78324544e-01 6.70139015e-01 -8.58830452e-01 -3.75898659e-01
1.86896533e-01 4.57860053e-01 -5.49114525e-01 9.14818421e-02
2.61903197e-01 5.93295932e-01 -4.58461493e-01 3.75587583e-01
6.01064920e-01 -2.25990430e-01 1.84442282e-01 -2.92177320e-01
-9.78500918e-02 2.63977312e-02 -8.34615648e-01 1.08606434e+00
-2.90438116e-01 9.12862003e-01 -1.65920779e-01 -1.39499640e+00
9.09614384e-01 1.81044444e-01 3.36052239e-01 -7.86912858e-01
3.93724173e-01 4.89484847e-01 5.35647810e-01 -5.77973247e-01
1.03963159e-01 -2.64037609e-01 1.08917885e-01 4.21386361e-01
4.02225405e-02 3.88185382e-02 1.84883222e-01 -4.70270574e-01
7.94551432e-01 -4.05604094e-01 4.71087933e-01 -6.06330037e-01
7.01251507e-01 -5.79900563e-01 2.44228318e-01 6.03814125e-01
-1.40682891e-01 5.00136554e-01 7.68729389e-01 -2.56559342e-01
-8.30165267e-01 -1.09532714e+00 -6.80366635e-01 7.24590778e-01
4.62996075e-03 -3.18438839e-03 -9.68739390e-01 -3.54289025e-01
-7.89998844e-02 9.09286022e-01 -6.17338598e-01 -2.00485542e-01
-4.43173259e-01 -1.18503606e+00 4.73850310e-01 2.63315231e-01
2.75876552e-01 -8.54799688e-01 -7.64070690e-01 -5.65281585e-02
-2.97368854e-01 -9.93900716e-01 -9.41250324e-02 4.13665712e-01
-1.01972723e+00 -1.12412024e+00 -5.19293070e-01 -5.59993267e-01
9.09802794e-01 -2.73462925e-02 9.05128837e-01 1.87166214e-01
4.09938283e-02 5.72789833e-02 -1.99612021e-01 -3.53699058e-01
-3.91897112e-01 -1.04921706e-01 2.31824473e-01 1.88841656e-01
2.21196324e-01 -4.88824874e-01 -4.24497426e-01 4.34234887e-01
-1.04577088e+00 1.91074327e-01 7.97848344e-01 8.58685195e-01
6.70861840e-01 -5.66594191e-02 6.98980272e-01 -4.04066384e-01
9.60552335e-01 -5.33235848e-01 -7.52403736e-01 4.00131166e-01
-7.75236607e-01 4.81246889e-01 6.45397604e-01 -4.43850219e-01
-8.20148706e-01 -2.61131465e-01 -1.10977851e-01 -1.49088120e-03
-8.57273489e-02 9.46269095e-01 -2.67300218e-01 -1.74354628e-01
7.12900698e-01 1.83340922e-01 1.81186963e-02 -3.88539284e-01
1.59265861e-01 6.10431850e-01 2.81814009e-01 -5.34359276e-01
3.25256288e-01 7.67432749e-02 2.03350455e-01 -9.16753173e-01
-3.70820969e-01 2.03271187e-03 -5.85455775e-01 -5.50605059e-01
9.26987290e-01 -3.63622516e-01 -8.50345612e-01 6.09642193e-02
-1.24095356e+00 4.08321470e-02 1.79970711e-01 8.84946823e-01
-6.18994176e-01 5.03495514e-01 -6.75784191e-03 -1.01782799e+00
-2.30334401e-01 -1.75798535e+00 8.38178813e-01 -8.29621702e-02
-3.67686629e-01 -1.06479156e+00 -3.60065699e-01 2.44058862e-01
2.72559136e-01 2.73950934e-01 1.37050569e+00 -6.88633859e-01
-3.44812572e-01 -2.34542906e-01 -4.16919827e-01 3.14352453e-01
-1.43759623e-01 -3.24844688e-01 -9.69101012e-01 4.67720181e-02
3.87115002e-01 1.93172142e-01 8.83106589e-01 5.56808233e-01
1.48364472e+00 -2.47395590e-01 -4.68023121e-02 4.39537913e-01
1.10224509e+00 2.65122801e-01 5.82727432e-01 1.69105008e-01
4.00291622e-01 8.25177550e-01 2.88903743e-01 3.18893909e-01
1.05611108e-01 6.40195370e-01 3.77534330e-01 1.70523614e-01
2.74141490e-01 2.53492892e-01 4.49128270e-01 1.00075948e+00
-4.10729855e-01 -2.72800744e-01 -9.32413876e-01 9.91935432e-02
-1.79617536e+00 -8.13081086e-01 -3.33203584e-01 2.47903085e+00
4.70537364e-01 5.45343384e-02 -1.62307635e-01 1.79325774e-01
9.28190947e-01 -1.53835952e-01 -5.33136547e-01 -4.51079160e-01
-3.70198458e-01 1.17840409e-01 3.75855505e-01 4.75641876e-01
-7.57250011e-01 2.78357297e-01 6.54533195e+00 5.91501355e-01
-1.21649206e+00 2.36350849e-01 7.63254285e-01 8.30153525e-02
-4.71663415e-01 -2.53953129e-01 -3.87983203e-01 5.98060012e-01
1.10831511e+00 -4.33686972e-01 6.62951112e-01 6.63899541e-01
5.54813087e-01 -2.53226697e-01 -1.37364197e+00 1.22889709e+00
2.21569017e-01 -9.14535880e-01 -7.22250268e-02 1.86364576e-01
2.41789013e-01 -2.21321076e-01 1.93934351e-01 -7.03640282e-02
-7.50795424e-01 -1.21974969e+00 1.13757491e+00 8.00376415e-01
5.66491663e-01 -5.35633862e-01 8.30600381e-01 2.71103978e-01
-7.26639628e-01 -9.56420898e-02 -4.42226082e-01 1.00262448e-01
1.22490115e-01 8.81534934e-01 -6.64246678e-01 1.91071853e-01
3.04983377e-01 4.13452595e-01 -7.59538770e-01 1.24406219e+00
-2.17735767e-01 4.80391860e-01 -3.27446073e-01 -3.58333617e-01
3.39213461e-02 -4.82150286e-01 5.89477837e-01 1.15960884e+00
5.83389044e-01 2.80237757e-02 -3.37687224e-01 1.20490086e+00
8.11269060e-02 2.88192391e-01 -3.63163859e-01 -2.08303168e-01
1.28862739e-01 1.09855056e+00 -1.06164980e+00 -1.11456998e-01
-1.72685936e-01 5.90369165e-01 3.54754925e-01 3.65640193e-01
-9.62506354e-01 8.17308761e-03 5.32248497e-01 1.39086004e-02
-2.07196817e-01 -2.32963994e-01 -8.72671485e-01 -1.26739645e+00
2.40211561e-01 -6.94148064e-01 1.57459214e-01 -5.75893223e-01
-1.16327333e+00 9.72130120e-01 4.71429497e-01 -9.37521040e-01
-3.23549271e-01 -9.05405104e-01 -3.35434079e-01 1.03405523e+00
-1.21164441e+00 -4.97756004e-01 -3.06010127e-01 1.74683630e-01
2.27574527e-01 -3.76608931e-02 9.56512868e-01 2.01797277e-01
-6.16092443e-01 4.56805408e-01 1.50263369e-01 -3.49771053e-01
2.43229508e-01 -1.07777262e+00 -6.92184791e-02 8.09687912e-01
-8.38752538e-02 6.49773061e-01 9.16256785e-01 -4.20602143e-01
-1.27518904e+00 -6.75629437e-01 8.58650923e-01 -1.16438545e-01
6.80151820e-01 -3.49067509e-01 -1.06254268e+00 3.00285876e-01
-1.03563264e-01 -3.05566043e-01 9.00164127e-01 1.04092270e-01
6.18254617e-02 -1.78155124e-01 -1.08464265e+00 6.64589047e-01
7.09437251e-01 -3.99797291e-01 -6.52662635e-01 3.71095687e-01
1.98396564e-01 9.74065438e-02 -8.95141482e-01 2.21424431e-01
8.57088685e-01 -9.49252546e-01 7.92257905e-01 -4.45423871e-01
2.27590904e-01 -1.83128342e-01 -1.78929046e-01 -1.48900068e+00
-3.36726457e-01 9.75404233e-02 1.45220291e-02 6.90992773e-01
5.33597529e-01 -9.12765861e-01 3.41661304e-01 8.38489234e-01
-1.18592106e-01 -8.11327577e-01 -1.16677415e+00 -6.79521203e-01
-1.07350826e-01 -7.35217035e-01 5.03049970e-01 5.52151144e-01
1.91927999e-01 3.01037818e-01 -6.24707453e-02 1.62944242e-01
8.76633942e-01 -1.08401932e-01 -2.64075901e-02 -1.37849545e+00
-2.22222373e-01 -9.82989490e-01 -8.62488031e-01 -6.68208539e-01
4.31809217e-01 -1.02471900e+00 2.18497172e-01 -1.27994144e+00
2.36774027e-01 -3.73201072e-01 -6.04595602e-01 1.71438903e-01
-2.46685952e-01 1.38673514e-01 2.18096003e-01 1.18665479e-01
-2.70809770e-01 4.58265662e-01 1.05802524e+00 -1.42085031e-01
-7.68247992e-02 -6.92552179e-02 -5.81682086e-01 6.57860696e-01
9.30729449e-01 -4.83537525e-01 -4.05728161e-01 -4.79095459e-01
2.59424448e-01 6.04306385e-02 4.75186348e-01 -9.00228381e-01
4.04213853e-02 -1.18018039e-01 5.15396476e-01 -2.85614163e-01
1.66097745e-01 -7.57620513e-01 2.70247340e-01 3.41205865e-01
-3.05621684e-01 1.86023057e-01 6.46991134e-02 4.41288590e-01
-1.84066907e-01 -4.77015465e-01 7.84009099e-01 2.39565149e-01
-5.82334161e-01 -6.91055655e-02 -5.73839784e-01 -5.47382347e-02
7.05814719e-01 -2.58798361e-01 5.11242338e-02 -1.65139049e-01
-7.63357103e-01 -6.93117306e-02 1.55296279e-02 2.04790652e-01
9.19742942e-01 -1.17157412e+00 -6.80565178e-01 3.48865658e-01
6.19369410e-02 -4.90537941e-01 8.23854282e-02 1.46778154e+00
-5.71186900e-01 5.87887526e-01 -4.38608855e-01 -7.10158229e-01
-1.03432488e+00 2.51269042e-01 3.77659649e-01 5.54492250e-02
-2.28975862e-01 3.01944673e-01 4.28750545e-01 1.44187376e-01
1.39156595e-01 -6.13893986e-01 -3.70687991e-01 1.42158329e-01
7.12508917e-01 4.14144903e-01 3.63813758e-01 -8.87250662e-01
-5.06763160e-01 2.58950740e-01 4.49782223e-01 -2.52649456e-01
1.40606833e+00 -1.17236070e-01 -4.32868540e-01 5.38313389e-01
1.07697344e+00 -3.90599847e-01 -7.21256375e-01 3.58864963e-01
2.09477544e-01 -3.35695505e-01 2.25196898e-01 -6.35194302e-01
-9.71892476e-01 9.36916471e-01 6.94258928e-01 3.36311221e-01
1.39315128e+00 -4.86528277e-02 8.72537121e-03 4.69734967e-01
3.61602634e-01 -7.94905603e-01 -4.72235501e-01 4.09637243e-01
1.25582385e+00 -1.11338508e+00 -6.52781501e-02 -3.05602342e-01
-4.20404404e-01 1.49873579e+00 2.52185225e-01 9.24548805e-02
6.69283390e-01 -2.81700697e-02 -3.30135584e-01 -1.99603170e-01
-1.96631461e-01 -9.44517627e-02 8.08234811e-01 4.23690557e-01
5.89739203e-01 4.39016342e-01 -9.66712236e-01 9.98021901e-01
-3.20523590e-01 -2.11101770e-01 1.45572469e-01 2.47012049e-01
-4.63130504e-01 -8.03074181e-01 -6.83133125e-01 7.12601721e-01
-1.95290938e-01 -1.81470171e-01 -1.26854375e-01 5.23367763e-01
4.65329811e-02 9.27072108e-01 -1.40814614e-02 -3.28767627e-01
2.70207673e-01 1.74714699e-01 6.04985654e-01 -2.58583665e-01
-2.60168314e-01 -1.22945964e-01 -1.09561607e-01 -3.16438138e-01
-3.51352006e-01 -6.41927123e-01 -1.10058260e+00 -1.94487229e-01
-4.50143486e-01 2.88415760e-01 9.55019295e-01 1.30673695e+00
1.39945924e-01 4.57870752e-01 2.95201153e-01 -9.44533050e-01
-4.47522104e-01 -9.41941619e-01 -5.90833008e-01 3.28984916e-01
1.01484142e-01 -1.02241290e+00 -5.19287527e-01 -1.44868299e-01] | [12.687114715576172, 3.397921562194824] |
9316ab3c-1f65-40b3-94ae-77595b56c883 | n-reference-transfer-learning-for-saliency | 2007.05104 | null | https://arxiv.org/abs/2007.05104v1 | https://arxiv.org/pdf/2007.05104v1.pdf | $n$-Reference Transfer Learning for Saliency Prediction | Benefiting from deep learning research and large-scale datasets, saliency prediction has achieved significant success in the past decade. However, it still remains challenging to predict saliency maps on images in new domains that lack sufficient data for data-hungry models. To solve this problem, we propose a few-shot transfer learning paradigm for saliency prediction, which enables efficient transfer of knowledge learned from the existing large-scale saliency datasets to a target domain with limited labeled examples. Specifically, very few target domain examples are used as the reference to train a model with a source domain dataset such that the training process can converge to a local minimum in favor of the target domain. Then, the learned model is further fine-tuned with the reference. The proposed framework is gradient-based and model-agnostic. We conduct comprehensive experiments and ablation study on various source domain and target domain pairs. The results show that the proposed framework achieves a significant performance improvement. The code is publicly available at \url{https://github.com/luoyan407/n-reference}. | ['Mohan S. Kankanhalli', 'Yongkang Wong', 'Yan Luo', 'Qi Zhao'] | 2020-07-09 | null | null | null | null | ['saliency-prediction-1'] | ['computer-vision'] | [ 2.14773536e-01 3.67088476e-03 -5.18732190e-01 -5.04493415e-01
-8.33039343e-01 2.38905824e-03 3.17005217e-01 -1.69513505e-02
-2.41265818e-01 8.43019247e-01 1.70676962e-01 1.01133041e-01
9.94725376e-02 -6.15074456e-01 -6.67476952e-01 -5.25064468e-01
2.94850886e-01 2.05925271e-01 7.52606630e-01 -2.38286152e-01
3.64361167e-01 -2.19161361e-02 -1.40734136e+00 1.22400679e-01
1.25711274e+00 1.09189522e+00 8.73978138e-01 9.05502141e-02
-3.37405242e-02 6.93871558e-01 -3.07767659e-01 -6.18187264e-02
1.87099084e-01 -4.40831423e-01 -7.51550078e-01 -6.46441523e-03
1.72868162e-01 -2.97231853e-01 -3.26279223e-01 1.22734952e+00
5.72784901e-01 1.65876314e-01 3.38248700e-01 -1.32485807e+00
-9.60441232e-01 2.89351314e-01 -7.89833188e-01 4.66199249e-01
-4.68081199e-02 2.35412627e-01 8.22248161e-01 -1.16592526e+00
5.47641575e-01 1.02847850e+00 3.24842453e-01 6.78610682e-01
-9.05128241e-01 -8.96922708e-01 1.87478647e-01 6.00878596e-01
-1.27995038e+00 -4.26889241e-01 1.16693592e+00 -8.77142251e-02
4.11816180e-01 -1.89610586e-01 5.92088699e-01 9.29945588e-01
-1.70661941e-01 1.14099622e+00 1.10406578e+00 -2.92830586e-01
3.57221127e-01 2.64438480e-01 -3.22448313e-02 5.46485782e-01
8.31441656e-02 6.28431961e-02 -7.06842721e-01 2.31201768e-01
7.40361810e-01 2.58125037e-01 -2.02240080e-01 -6.44597530e-01
-1.17469108e+00 7.87493289e-01 1.08942473e+00 2.06267357e-01
-3.89453799e-01 -3.85433994e-02 3.21875721e-01 1.31511251e-02
6.64941132e-01 3.02347064e-01 -3.76666784e-01 2.28080619e-02
-9.63783503e-01 9.63994935e-02 3.26603830e-01 9.76366043e-01
1.08201706e+00 1.78522274e-01 -1.41334698e-01 1.02949274e+00
1.68221474e-01 5.55314422e-01 6.87421978e-01 -6.50851429e-01
4.39267695e-01 7.00793684e-01 1.96956456e-01 -9.66500401e-01
-1.91126972e-01 -4.94967729e-01 -5.75749040e-01 1.40685976e-01
1.43637702e-01 -2.24271640e-01 -9.62508798e-01 1.78788459e+00
4.73183125e-01 7.22761035e-01 7.37027824e-02 1.32829905e+00
8.93883884e-01 7.40465999e-01 2.17690498e-01 1.46489206e-03
8.90667021e-01 -1.40549076e+00 -3.71071637e-01 -7.41840124e-01
2.82177866e-01 -6.20250762e-01 1.37548387e+00 -9.06978399e-02
-8.73003781e-01 -6.43107772e-01 -1.06167448e+00 -1.38902128e-01
-1.92545667e-01 1.68620795e-01 4.48772013e-01 4.87834495e-03
-1.00444627e+00 3.43343347e-01 -7.17074811e-01 -5.26951015e-01
9.78039503e-01 1.77122831e-01 2.49295700e-02 -3.06552976e-01
-1.27062297e+00 9.77829754e-01 5.00383973e-01 -4.33895066e-02
-1.23789084e+00 -7.71208167e-01 -8.75999570e-01 6.89541250e-02
4.32773918e-01 -2.49642313e-01 1.36329722e+00 -1.46791840e+00
-1.23031044e+00 8.41558218e-01 -3.02157402e-01 -4.42150384e-01
3.49776387e-01 -4.00139093e-01 -3.96114498e-01 4.20995317e-02
4.77600127e-01 9.57044661e-01 1.05301213e+00 -1.16333020e+00
-8.18559349e-01 -1.70419678e-01 -1.18863359e-02 4.90711242e-01
-5.67748010e-01 -6.39634877e-02 -4.77457136e-01 -5.40261328e-01
-1.30580023e-01 -7.59405196e-01 -1.42248809e-01 5.42612113e-02
-2.74936736e-01 -1.81048095e-01 9.48118687e-01 -4.49302554e-01
1.03517008e+00 -2.17720509e+00 -1.43854087e-02 -2.43789285e-01
1.32996693e-01 6.54924929e-01 -3.22256774e-01 6.82696179e-02
-1.16232693e-01 -3.02317619e-01 -2.60565788e-01 1.27580717e-01
-4.00664777e-01 -2.63445884e-01 -4.32309747e-01 2.39646539e-01
3.32013667e-01 1.03918934e+00 -1.17817736e+00 -5.26141107e-01
2.51669586e-01 1.94354698e-01 -2.91676760e-01 2.95940131e-01
-2.00054392e-01 4.02840316e-01 -7.24261701e-01 6.33668005e-01
7.06423283e-01 -5.97951174e-01 -1.82336971e-01 -1.90523624e-01
-9.12455171e-02 2.00127393e-01 -7.77958512e-01 1.97527575e+00
-3.97580266e-01 6.42441988e-01 -2.74532944e-01 -1.19809341e+00
1.16941059e+00 -8.65093917e-02 1.62392080e-01 -8.91143143e-01
1.57835677e-01 3.59920561e-01 6.64069504e-02 -3.16534311e-01
3.01078409e-01 -1.24320827e-01 1.20717712e-01 3.03779006e-01
8.89725685e-02 1.03235105e-02 2.44060205e-03 1.62054345e-01
8.83877397e-01 2.91605502e-01 3.18831623e-01 -3.04069936e-01
4.98162657e-01 4.09593016e-01 8.37595999e-01 3.14180434e-01
-6.03332222e-01 6.17206633e-01 1.46869168e-01 -3.66788596e-01
-1.04082596e+00 -1.11013269e+00 1.12456113e-01 1.23396516e+00
7.76566327e-01 -8.59592296e-03 -6.57520890e-01 -7.18435943e-01
-1.17035069e-01 7.03182518e-01 -6.72134519e-01 -6.42273664e-01
-2.79995888e-01 -3.18901718e-01 -3.44830193e-02 4.74643469e-01
1.00241435e+00 -1.46550429e+00 -8.15300226e-01 1.23557836e-01
-1.99650973e-01 -7.62309253e-01 -4.86053526e-01 -4.77526784e-02
-1.03598785e+00 -9.84834194e-01 -1.16270101e+00 -1.27774119e+00
7.95176744e-01 7.61049747e-01 9.16342854e-01 4.70408648e-02
-8.10460970e-02 -8.27951133e-02 -3.48870426e-01 -5.97946107e-01
3.12519707e-02 1.14721440e-01 -2.31889673e-02 8.20665993e-03
7.95368969e-01 -4.61962432e-01 -9.37939346e-01 4.89512503e-01
-6.45040751e-01 3.33243221e-01 7.57985055e-01 8.64018738e-01
6.70948446e-01 -3.81255299e-01 1.21790111e+00 -6.74377561e-01
5.39213479e-01 -7.47247994e-01 -5.74665606e-01 2.25487143e-01
-4.93179649e-01 -1.10759668e-01 6.07454598e-01 -4.27239120e-01
-1.22226357e+00 4.48240824e-02 2.45930284e-01 -6.69488311e-01
-1.37756526e-01 5.54977298e-01 -1.47188202e-01 1.86079368e-02
7.81903446e-01 4.64403450e-01 -1.20314196e-01 -4.06226695e-01
1.45183504e-01 6.35791540e-01 4.82878447e-01 -2.19761714e-01
9.08596754e-01 4.05112505e-01 -5.75112998e-01 -4.45257992e-01
-1.27640629e+00 -3.71339977e-01 -5.44446588e-01 -2.57726938e-01
6.22614503e-01 -1.10254419e+00 4.40098606e-02 4.66017634e-01
-8.56013536e-01 -4.98978287e-01 -2.47984469e-01 3.68914038e-01
-4.94876385e-01 3.92395593e-02 -1.07059263e-01 -4.45583403e-01
-4.99326974e-01 -9.33456004e-01 7.83428967e-01 8.56329858e-01
1.36950240e-02 -8.45265567e-01 8.47928822e-02 1.27227500e-01
4.73639667e-01 -6.45481572e-02 5.08176208e-01 -6.89422190e-01
-6.38206244e-01 -1.69443578e-01 -6.03810251e-01 2.35117272e-01
4.73902315e-01 -4.27602798e-01 -9.37988818e-01 -2.77508616e-01
-2.46814508e-02 -7.39095449e-01 8.34176242e-01 4.44386929e-01
1.12201929e+00 -4.72850576e-02 -5.72537541e-01 3.74581486e-01
1.39992440e+00 8.70011225e-02 3.98048341e-01 4.98859465e-01
6.10931396e-01 3.86209846e-01 1.21936882e+00 3.26283157e-01
4.76902455e-01 4.26019013e-01 4.36762094e-01 -2.02199385e-01
-1.91514820e-01 -6.20523810e-01 5.95785305e-02 4.60057616e-01
3.26388359e-01 3.51168722e-01 -1.08272946e+00 1.03208375e+00
-1.97528160e+00 -7.89860904e-01 4.25193936e-01 2.23591995e+00
9.80260313e-01 3.33487719e-01 8.07837173e-02 -3.55652571e-01
1.06363010e+00 2.84934819e-01 -1.21740246e+00 1.15145460e-01
1.13061935e-01 -2.32466240e-03 1.81310251e-01 1.42825440e-01
-1.17562926e+00 1.48303783e+00 5.21604776e+00 8.85199904e-01
-1.53848088e+00 3.16789985e-01 7.51778245e-01 -2.23706603e-01
-2.16847673e-01 1.03910286e-02 -6.04251981e-01 7.19269514e-01
6.53296113e-01 -7.61639178e-01 2.22846359e-01 1.32435882e+00
3.20778012e-01 -2.30447918e-01 -8.13092351e-01 1.03268135e+00
8.00795853e-02 -1.42615843e+00 -6.45698011e-02 -3.59925061e-01
9.23248589e-01 2.98918933e-01 3.29484284e-01 4.92855310e-01
2.02540949e-01 -7.35372543e-01 4.88884300e-01 2.11921439e-01
7.42392421e-01 -7.16284573e-01 4.87899572e-01 3.76103073e-01
-1.07398784e+00 -2.15245575e-01 -8.52882326e-01 -1.40340969e-01
5.80324233e-02 5.71060658e-01 -9.04198408e-01 1.72122851e-01
8.13805342e-01 1.23761272e+00 -7.04038262e-01 1.42308486e+00
-3.79583836e-01 6.86034262e-01 -2.32796390e-02 -1.21579982e-01
2.98263997e-01 -1.10993348e-01 4.47352052e-01 5.92667818e-01
3.37868184e-01 1.08113445e-01 2.20816001e-01 8.95047247e-01
-2.36796454e-01 1.59205526e-01 -6.19846046e-01 1.56631991e-01
7.53603041e-01 1.40973711e+00 -6.23127222e-01 -3.95743102e-01
-4.85455215e-01 9.27854359e-01 6.86625481e-01 4.58893478e-01
-8.96226466e-01 -4.29810137e-01 4.82589453e-01 4.91931476e-03
3.20964128e-01 1.93906218e-01 -3.65839958e-01 -1.26209593e+00
-6.23456109e-03 -6.52177393e-01 2.87874699e-01 -1.06124628e+00
-1.24459839e+00 5.70732534e-01 -2.19018295e-01 -1.56172287e+00
-4.49957252e-02 -1.22595191e-01 -9.66239989e-01 1.02556264e+00
-1.95960855e+00 -1.11621320e+00 -5.65583110e-01 7.58466363e-01
7.29906380e-01 -2.93108851e-01 5.14959395e-01 1.83987960e-01
-3.82937551e-01 5.15629113e-01 1.51877373e-01 6.93643391e-02
8.85452509e-01 -1.02071548e+00 3.29074919e-01 9.27110374e-01
-1.04323402e-01 3.58621716e-01 6.81952178e-01 -7.11012483e-01
-9.24233019e-01 -1.29868007e+00 5.95688641e-01 -1.30231172e-01
6.05444133e-01 -1.98037073e-01 -1.18758583e+00 5.64780951e-01
2.15700135e-01 2.39301965e-01 4.15221184e-01 -3.86878252e-02
-1.60331264e-01 -3.08289319e-01 -1.08057594e+00 6.73433244e-01
9.15574968e-01 -3.73777181e-01 -8.03823590e-01 3.18135053e-01
7.58626997e-01 -3.09397817e-01 -2.25568131e-01 2.68983573e-01
1.57872736e-01 -7.97467053e-01 7.71555245e-01 -5.63260019e-01
6.63088560e-01 -3.90017718e-01 1.50666520e-01 -1.56479847e+00
-3.78313154e-01 -1.99557960e-01 -7.07680965e-03 1.05593967e+00
4.75483775e-01 -4.74147886e-01 9.44427609e-01 5.58236480e-01
-2.76577383e-01 -8.17336380e-01 -7.69784272e-01 -7.01487482e-01
4.75809686e-02 2.01249681e-03 3.54960918e-01 8.47122133e-01
1.24371812e-01 4.57222015e-01 -4.38352168e-01 7.51817971e-02
6.56078219e-01 6.62078917e-01 7.21625984e-01 -1.08580875e+00
1.86494112e-01 -7.33294562e-02 -2.89569318e-01 -1.15439177e+00
2.26409242e-01 -9.11556125e-01 2.94838011e-01 -1.74189782e+00
4.39022094e-01 -5.07001281e-01 -6.54537797e-01 7.33826458e-01
-5.89798748e-01 1.82769105e-01 2.13930771e-01 4.00115967e-01
-9.52024519e-01 9.74882483e-01 1.35382092e+00 -2.04840481e-01
-2.74342328e-01 -3.88069563e-02 -8.81438673e-01 7.18319178e-01
1.15968752e+00 -4.07187551e-01 -6.53722644e-01 -5.33762455e-01
-2.95095861e-01 -1.70220375e-01 3.91358197e-01 -1.25480533e+00
2.79902399e-01 -3.19507658e-01 5.84928811e-01 -4.14175779e-01
2.67890215e-01 -6.01489365e-01 -4.22378540e-01 4.62087423e-01
-4.49757844e-01 -4.34112340e-01 3.01099122e-01 6.34989500e-01
-3.45790625e-01 -1.33059174e-01 1.01387119e+00 2.83059496e-02
-1.46651495e+00 6.01948678e-01 2.73292482e-01 3.12848628e-01
1.29361188e+00 -1.30600095e-01 -3.80779207e-01 -3.75510156e-01
-5.42886853e-01 5.46136975e-01 6.84813678e-01 8.66053820e-01
9.79745328e-01 -1.49053502e+00 -6.64267480e-01 -2.42096949e-02
3.92137319e-01 9.15634260e-02 4.38139498e-01 6.77144289e-01
-1.09550297e-01 2.23757491e-01 -6.63144350e-01 -6.16975307e-01
-9.69103277e-01 8.10627580e-01 2.82267332e-01 2.27897659e-01
-4.56599772e-01 9.56720054e-01 4.58354145e-01 -4.33594763e-01
-6.47165403e-02 1.44923449e-01 -2.73016244e-01 -2.19437122e-01
6.85182095e-01 9.56028700e-02 -3.38764817e-01 -5.74919641e-01
-4.65477586e-01 3.54612768e-01 -3.14757735e-01 1.32332519e-01
1.39981008e+00 -3.10248107e-01 2.82843292e-01 4.48799968e-01
1.07290781e+00 -4.87023234e-01 -1.73320377e+00 -7.31602311e-01
6.08754605e-02 -7.67287970e-01 1.29632195e-02 -8.56566966e-01
-1.15362477e+00 1.02945995e+00 8.43088090e-01 -3.70555639e-01
1.29308319e+00 1.39155596e-01 9.49515998e-01 2.74603665e-01
4.09125328e-01 -1.20846510e+00 4.16310370e-01 4.56534326e-01
9.26503241e-01 -1.67338943e+00 -2.43528649e-01 -2.04713285e-01
-1.06550467e+00 5.75996637e-01 1.16402149e+00 -4.57143426e-01
6.12150550e-01 -4.14303839e-01 1.45982563e-01 -7.02536106e-02
-5.51317930e-01 -3.29774767e-01 1.84043378e-01 7.69714832e-01
1.22516282e-01 -8.41289088e-02 -7.48674646e-02 6.13577664e-01
3.69902588e-02 5.04940748e-01 4.21420336e-01 9.52923059e-01
-8.95356119e-01 -8.27171504e-01 -6.48331121e-02 6.56586707e-01
-5.33915125e-02 -3.41873378e-01 -1.10898711e-01 5.31145275e-01
-1.89343542e-01 7.87338197e-01 -6.97956458e-02 -2.47988805e-01
1.49698079e-01 -3.67534570e-02 6.05797879e-02 -9.04080212e-01
1.83420535e-02 -7.32507780e-02 -3.79115790e-01 -5.08131921e-01
-4.44384038e-01 -5.20283818e-01 -1.41018689e+00 2.68073454e-02
-1.49881080e-01 2.37244204e-01 1.99595854e-01 8.03197682e-01
6.49271607e-01 3.10824841e-01 7.36133218e-01 -1.01606369e+00
-3.97110045e-01 -9.70818043e-01 -4.66805369e-01 4.37807173e-01
2.61064261e-01 -9.31761503e-01 -9.89722759e-02 5.71334772e-02] | [9.779558181762695, -0.31541258096694946] |
cdb6461e-1f86-484b-b12f-c6c9f997d0c7 | flownet3d-geometric-losses-for-deep-scene | 1912.01438 | null | https://arxiv.org/abs/1912.01438v3 | https://arxiv.org/pdf/1912.01438v3.pdf | FlowNet3D++: Geometric Losses For Deep Scene Flow Estimation | We present FlowNet3D++, a deep scene flow estimation network. Inspired by classical methods, FlowNet3D++ incorporates geometric constraints in the form of point-to-plane distance and angular alignment between individual vectors in the flow field, into FlowNet3D. We demonstrate that the addition of these geometric loss terms improves the previous state-of-art FlowNet3D accuracy from 57.85% to 63.43%. To further demonstrate the effectiveness of our geometric constraints, we propose a benchmark for flow estimation on the task of dynamic 3D reconstruction, thus providing a more holistic and practical measure of performance than the breakdown of individual metrics previously used to evaluate scene flow. This is made possible through the contribution of a novel pipeline to integrate point-based scene flow predictions into a global dense volume. FlowNet3D++ achieves up to a 15.0% reduction in reconstruction error over FlowNet3D, and up to a 35.2% improvement over KillingFusion alone. We will release our scene flow estimation code later. | ['Henry Howard-Jenkins', 'Shuda Li', 'Zirui Wang', 'Min Chen', 'Victor Adrian Prisacariu'] | 2019-12-03 | null | null | null | null | ['scene-flow-estimation'] | ['computer-vision'] | [-3.51546258e-01 -2.53259987e-01 2.30272561e-02 -2.40462348e-01
-2.35007107e-01 -6.53577745e-01 7.09521711e-01 8.81607085e-02
-3.17997932e-01 5.05631864e-01 6.71107531e-01 -3.05041730e-01
2.36433983e-01 -8.77134144e-01 -4.31882501e-01 -1.08176194e-01
-4.30051923e-01 2.55409449e-01 5.42592704e-01 -7.51551613e-02
3.62591773e-01 1.01082397e+00 -1.22188568e+00 6.12438954e-02
6.86703205e-01 1.06625497e+00 -2.13160992e-01 1.08484721e+00
-1.96583614e-01 1.14705062e+00 -3.90302360e-01 -1.92600280e-01
6.63536727e-01 -1.72670767e-01 -1.13727927e+00 -1.55849054e-01
1.15781403e+00 -1.00465357e+00 -8.07751060e-01 4.05731708e-01
4.62084949e-01 4.40051258e-01 3.66621733e-01 -1.22308874e+00
-4.13425900e-02 2.95156650e-02 -4.84621376e-01 6.24351919e-01
5.29486418e-01 6.05871558e-01 7.75791407e-01 -8.16112876e-01
1.14661860e+00 1.32827139e+00 7.88599074e-01 2.60546744e-01
-1.21229136e+00 -4.27357167e-01 -2.29742797e-03 8.91610533e-02
-1.06276739e+00 -5.45712054e-01 5.96852720e-01 -7.28493392e-01
1.52504575e+00 -5.75170759e-03 9.11352754e-01 6.55184984e-01
7.52758384e-02 7.15791345e-01 5.93861341e-01 5.34020811e-02
-5.03836870e-02 -3.54287982e-01 3.46631836e-03 9.45235729e-01
1.18483812e-01 3.33667547e-01 -6.02696538e-01 2.46712968e-01
1.07743192e+00 -4.66667801e-01 -3.39131922e-01 -6.19405508e-01
-1.29241943e+00 6.55902803e-01 9.55394983e-01 1.67141780e-02
-6.41416386e-02 5.86597800e-01 6.90676093e-01 -1.09017447e-01
6.53517604e-01 5.05111396e-01 -3.42766732e-01 -5.90520501e-01
-9.18107212e-01 6.25261605e-01 8.41444135e-01 8.71027768e-01
8.56237233e-01 2.19282851e-01 -1.95921898e-01 2.13161841e-01
3.26034963e-01 4.31026876e-01 -2.24049732e-01 -1.69423592e+00
6.93639278e-01 5.85924625e-01 2.17240918e-02 -1.35163569e+00
-6.19362175e-01 -4.10616606e-01 -5.53165197e-01 4.46688563e-01
7.42740214e-01 -6.05523326e-02 -6.62433147e-01 1.43556774e+00
5.83371341e-01 4.77883726e-01 -3.09298933e-01 1.16015697e+00
9.70104694e-01 5.85906088e-01 -1.29479244e-02 2.97922760e-01
7.18588054e-01 -1.28256702e+00 -2.63153851e-01 5.91578633e-02
9.93439436e-01 -7.89802372e-01 7.76047647e-01 1.20021932e-01
-1.32615626e+00 -5.17907023e-01 -8.64875257e-01 -5.27698696e-01
-6.74918070e-02 -3.99686903e-01 9.01651621e-01 4.63907033e-01
-1.26912880e+00 7.68311560e-01 -1.04080260e+00 -3.50524098e-01
8.35080147e-01 1.54887468e-01 -5.38837433e-01 -2.35869363e-01
-6.31051123e-01 9.69456792e-01 5.31864576e-02 -9.09467265e-02
-8.79923046e-01 -1.61524796e+00 -1.29055989e+00 6.00851476e-02
-1.73139963e-02 -1.25595415e+00 1.19358218e+00 -6.74893335e-02
-1.20411837e+00 8.29264104e-01 -2.60884047e-01 -4.93059069e-01
1.10229015e+00 -3.61987919e-01 1.78154826e-01 5.52906692e-01
1.74570844e-01 9.99966860e-01 4.16910313e-02 -9.13304806e-01
-4.95602697e-01 -7.16268122e-02 1.97784960e-01 1.84602678e-01
3.13073397e-02 -1.94614619e-01 -6.24638617e-01 -3.72643143e-01
-2.94172823e-01 -6.57385230e-01 -3.64822328e-01 7.44038999e-01
-3.96153659e-01 -3.65425367e-03 1.08306491e+00 -5.53803623e-01
1.01484001e+00 -1.85299158e+00 -2.67901868e-02 6.12000003e-02
7.27429628e-01 4.17517930e-01 -1.32991731e-01 1.33383855e-01
1.72458008e-01 1.33116946e-01 -4.36958492e-01 -6.68678463e-01
-1.94393709e-01 -2.60052662e-02 -1.55857712e-01 5.91405153e-01
3.56891155e-01 1.04977643e+00 -1.13446677e+00 -5.32270491e-01
1.12480247e+00 6.83697522e-01 -1.03934884e+00 -3.41106951e-02
1.95164774e-02 5.59346735e-01 -2.17124566e-01 5.08735955e-01
9.57484603e-01 -2.34531567e-01 -3.13745320e-01 -2.74574608e-01
-3.12301993e-01 5.89387715e-01 -1.08916843e+00 2.26725483e+00
-5.67725360e-01 1.05713773e+00 -4.50130403e-02 -4.69474196e-01
7.81139553e-01 -1.08796321e-01 1.08367419e+00 -6.66048050e-01
2.17397183e-01 -4.31923196e-02 -5.65820709e-02 -3.10461491e-01
6.93051159e-01 1.01461239e-01 1.51935309e-01 2.89797306e-01
3.26418638e-01 -5.91143429e-01 5.28944612e-01 6.15254104e-01
1.45838177e+00 4.44714606e-01 6.47494346e-02 -4.34279025e-01
5.92287540e-01 3.03982198e-01 4.64082479e-01 6.75361812e-01
-6.16121650e-01 8.00759256e-01 5.78332245e-01 -9.14390802e-01
-1.22469819e+00 -1.01874268e+00 -2.92508304e-01 3.20887446e-01
3.91830295e-01 -7.27148294e-01 -4.22517985e-01 -7.87885487e-01
4.44323868e-01 3.50498587e-01 -5.90008080e-01 1.52077720e-01
-9.08786297e-01 -3.66910189e-01 6.10487700e-01 5.75569689e-01
8.21570277e-01 -8.32033396e-01 -6.64534092e-01 3.00461978e-01
-3.82741615e-02 -1.57149827e+00 -3.78590524e-01 -3.84181738e-01
-8.70417118e-01 -1.31690657e+00 -5.08154809e-01 -2.88233638e-01
1.97598264e-01 3.08011234e-01 1.46686959e+00 3.77266675e-01
-5.22485197e-01 3.27541322e-01 -2.10288361e-01 1.54682267e-02
-2.79666513e-01 3.38831067e-01 -2.96410382e-01 -4.70542282e-01
4.33869381e-03 -7.14578867e-01 -1.09357131e+00 1.11816257e-01
-6.11096919e-01 1.67643771e-01 -1.94825679e-01 2.42151693e-01
1.36575907e-01 -4.34594005e-01 2.04419903e-02 -6.08352184e-01
8.24979097e-02 -3.81407708e-01 -6.52861655e-01 -3.70845109e-01
-2.32366949e-01 -4.79714051e-02 5.87962449e-01 2.33092949e-01
-1.03155279e+00 1.76667035e-01 -4.06513095e-01 -7.40151703e-01
-1.23420984e-01 4.31099720e-02 1.48456961e-01 -4.20810580e-01
7.05402613e-01 -2.81396896e-01 3.62129286e-02 -1.26144528e-01
5.07176697e-01 -1.86308429e-01 6.97139144e-01 -4.57566082e-01
7.70603955e-01 8.09322059e-01 5.51295340e-01 -7.09211707e-01
-8.79344761e-01 -7.04764545e-01 -9.66752529e-01 -5.99531174e-01
1.01089323e+00 -8.11901152e-01 -9.36099172e-01 5.72045624e-01
-1.28747272e+00 -7.76042640e-01 -3.82725477e-01 6.05363548e-01
-7.01776981e-01 4.44951266e-01 -7.61550426e-01 -3.01108420e-01
-2.88325936e-01 -1.27182806e+00 1.16928220e+00 -1.40547594e-02
-2.32824296e-01 -1.40465569e+00 3.06204230e-01 3.35294098e-01
5.93947053e-01 7.30335295e-01 2.45008275e-01 1.33571520e-01
-9.12786841e-01 1.80215806e-01 -6.17764711e-01 1.75729811e-01
-3.61937620e-02 3.22611541e-01 -9.75185037e-01 -2.03027409e-02
-4.89099205e-01 -1.28095567e-01 1.17227674e+00 5.81769466e-01
1.04346383e+00 1.94079056e-01 -1.82367057e-01 1.35155225e+00
1.53584409e+00 -1.66831255e-01 7.47091830e-01 3.01717281e-01
1.13222384e+00 3.85404825e-01 3.14415783e-01 4.53226984e-01
6.35429323e-01 6.98504627e-01 6.89917982e-01 -2.12852165e-01
-8.01406503e-01 -4.00650442e-01 -9.84044895e-02 4.14235443e-01
-5.15464880e-02 -3.05223197e-01 -1.03475034e+00 5.73525310e-01
-1.56711590e+00 -1.02513576e+00 -6.45416021e-01 1.98186183e+00
2.99150050e-01 2.46350199e-01 1.93602994e-01 -7.82890692e-02
2.85554320e-01 4.57186878e-01 -5.67516446e-01 -3.95176411e-01
-6.18066117e-02 3.42315167e-01 6.30897284e-01 1.13891542e+00
-1.22413158e+00 1.26225305e+00 6.64613199e+00 3.65804791e-01
-1.25313497e+00 -1.56247586e-01 5.23797393e-01 -3.28058571e-01
-2.23638564e-01 1.71106353e-01 -7.50291526e-01 2.58578062e-01
5.33482194e-01 -1.57847881e-01 1.60803184e-01 5.50137699e-01
5.05665600e-01 -3.61590683e-01 -1.01448035e+00 9.79846597e-01
-3.58375907e-02 -1.85116565e+00 4.45160605e-02 1.64923057e-01
7.25245893e-01 4.86380428e-01 -3.29880267e-01 1.28661573e-01
3.81880283e-01 -9.60844934e-01 5.36470056e-01 3.24693978e-01
7.84928262e-01 -6.13264382e-01 3.99100453e-01 2.37896591e-02
-1.34942627e+00 3.25286001e-01 -1.48513183e-01 -3.32666188e-01
8.89950633e-01 7.39858091e-01 -9.41560268e-01 6.13971651e-01
7.51742363e-01 1.40210056e+00 -4.43322837e-01 1.42020094e+00
-5.12137599e-02 4.34242785e-01 -5.17822862e-01 3.76042187e-01
4.37780082e-01 -6.81273192e-02 7.96841383e-01 1.52390110e+00
7.43569732e-02 -1.10759266e-01 1.57717690e-01 9.95983839e-01
-1.01574644e-01 -1.34328261e-01 -7.40554750e-01 6.07606709e-01
2.75778681e-01 1.27354777e+00 -7.03902900e-01 -4.64435339e-01
-2.51303852e-01 7.89523721e-01 4.06377494e-01 5.51412478e-02
-7.84238994e-01 -4.09831911e-01 1.25972176e+00 3.01125497e-01
1.94680542e-01 -6.63618565e-01 -4.97684211e-01 -1.29638672e+00
3.25006549e-03 -2.58905254e-02 1.35622889e-01 -8.04234922e-01
-1.01538146e+00 3.64791840e-01 -9.13639087e-03 -1.14069712e+00
-1.32689774e-01 -6.16412640e-01 -6.85138345e-01 8.84627938e-01
-1.81932843e+00 -8.65528286e-01 -8.74767900e-01 5.37485957e-01
3.41224134e-01 4.05315101e-01 4.18331981e-01 4.62329954e-01
-4.01563913e-01 2.37885162e-01 -6.47527039e-01 4.04035866e-01
5.56203187e-01 -1.29874837e+00 1.04433393e+00 1.07545793e+00
-1.23508401e-01 1.99060932e-01 4.90361601e-01 -4.66763496e-01
-1.17285895e+00 -1.28957093e+00 9.26606953e-01 -8.64834547e-01
5.78773618e-01 -3.75744283e-01 -6.80780232e-01 6.56362414e-01
-3.30222100e-02 6.58590078e-01 1.64177403e-01 -2.29555205e-01
-4.35306489e-01 -6.73173517e-02 -1.28599429e+00 3.89675736e-01
1.60832465e+00 -2.67516375e-01 -9.51845571e-02 2.06565291e-01
9.09459114e-01 -8.14134121e-01 -1.11059105e+00 5.21580756e-01
4.32015657e-01 -1.46751761e+00 1.27764893e+00 -4.63499635e-01
8.33941102e-01 -4.08899009e-01 2.72531249e-02 -1.09827197e+00
-2.18156785e-01 -6.78211093e-01 -2.95226842e-01 7.71456540e-01
6.49427325e-02 -3.74250710e-01 1.11658406e+00 5.37707090e-01
-5.55575788e-01 -5.34097195e-01 -8.34632754e-01 -4.89966959e-01
2.62599736e-01 -8.09328020e-01 3.45432311e-01 9.14608240e-01
-1.99141979e-01 5.50910868e-02 -1.58143431e-01 -1.67943880e-01
6.01515234e-01 -1.41898766e-01 1.23088551e+00 -9.18203831e-01
2.16157734e-02 -8.41915369e-01 -9.51186836e-01 -1.68421292e+00
1.54842645e-01 -1.07495439e+00 -2.95325696e-01 -1.82074714e+00
-3.12883258e-01 -6.01389110e-01 1.97689787e-01 1.66792810e-01
-6.97635487e-02 5.51120937e-01 5.87461948e-01 -7.92156383e-02
-6.44295037e-01 6.64081037e-01 1.75118744e+00 2.07082272e-01
-3.44120055e-01 -3.85146230e-01 -3.65390092e-01 6.13068879e-01
6.67645276e-01 -6.35394827e-02 -3.19373935e-01 -7.49951601e-01
-1.69196397e-01 2.01043203e-01 6.40641034e-01 -1.29889309e+00
1.48194253e-01 2.41481513e-02 5.02443492e-01 -7.18980074e-01
4.85107481e-01 -5.56629002e-01 -2.80762851e-01 5.52674413e-01
-1.40001401e-01 7.07323328e-02 4.78966147e-01 2.02518091e-01
-8.35337788e-02 3.66412401e-01 7.26072371e-01 -1.30319446e-01
-9.25795436e-01 6.53549552e-01 -8.21503028e-02 3.96001935e-01
9.60690439e-01 -2.21533075e-01 -6.00717902e-01 -3.40402693e-01
-4.38751370e-01 2.86085427e-01 5.04258573e-01 4.66810614e-01
5.82623661e-01 -1.15668380e+00 -7.22173929e-01 1.52757004e-01
-5.52430935e-02 4.17802393e-01 3.50364089e-01 8.35094213e-01
-1.49179387e+00 6.27454579e-01 -2.89055914e-01 -9.47713852e-01
-9.34252739e-01 2.23736897e-01 7.42462575e-01 -2.28353381e-01
-8.92561615e-01 8.29212546e-01 1.98991507e-01 -6.26018226e-01
2.47794692e-03 -6.49156153e-01 2.43638486e-01 -4.24463958e-01
4.19091254e-01 7.77709723e-01 1.40473977e-01 -7.01716959e-01
-5.69341063e-01 6.44066334e-01 2.87045568e-01 -6.22740239e-02
1.25024891e+00 -6.02752678e-02 6.55959696e-02 -9.36226323e-02
1.55226111e+00 -5.32083586e-02 -1.74393535e+00 6.71440214e-02
-4.51738894e-01 -9.38808024e-01 1.81195199e-01 -6.81622088e-01
-1.39467657e+00 9.77590024e-01 1.81166470e-01 -9.79808345e-02
7.78227329e-01 -1.37832761e-01 9.39723790e-01 -6.23777732e-02
2.41886258e-01 -3.37944269e-01 -4.61841412e-02 7.38683879e-01
6.62725389e-01 -1.20204926e+00 1.08633831e-01 -7.06973553e-01
-7.61796460e-02 1.11982155e+00 8.93082559e-01 -5.46786726e-01
8.01276565e-01 3.82442772e-01 4.04566638e-02 -2.55773842e-01
-6.13514125e-01 -1.28759965e-01 2.58124143e-01 5.90059519e-01
3.94652963e-01 -2.88952589e-01 1.99956924e-01 -6.08561933e-01
-5.75479925e-01 9.05742720e-02 5.71513116e-01 7.76316643e-01
-1.87630296e-01 -8.06094646e-01 -9.72156320e-03 1.29005253e-01
-1.41193092e-01 1.02599218e-01 -4.65679355e-02 1.02349269e+00
4.20639776e-02 8.55243742e-01 5.69753647e-01 -2.54182398e-01
5.64789891e-01 -5.52115738e-01 5.76466024e-01 -3.07198167e-01
-5.72095811e-01 -3.62887621e-01 1.11785553e-01 -1.16672146e+00
-5.42987227e-01 -6.45942152e-01 -1.26031172e+00 -1.00458932e+00
3.05083781e-01 -2.92394996e-01 6.87839806e-01 6.19649053e-01
5.35393596e-01 4.99395818e-01 4.57382619e-01 -1.24606824e+00
1.73669428e-01 -6.58410549e-01 -9.15063843e-02 4.85306352e-01
6.01180375e-01 -6.28216565e-01 -5.78003168e-01 -2.37798560e-02] | [8.678671836853027, -1.970813512802124] |
75635496-15e8-48aa-8de7-584f9c2089b3 | from-a-bird-s-eye-view-to-see-joint-camera | 2212.09298 | null | https://arxiv.org/abs/2212.09298v1 | https://arxiv.org/pdf/2212.09298v1.pdf | From a Bird's Eye View to See: Joint Camera and Subject Registration without the Camera Calibration | We tackle a new problem of multi-view camera and subject registration in the bird's eye view (BEV) without pre-given camera calibration. This is a very challenging problem since its only input is several RGB images from different first-person views (FPVs) for a multi-person scene, without the BEV image and the calibration of the FPVs, while the output is a unified plane with the localization and orientation of both the subjects and cameras in a BEV. We propose an end-to-end framework solving this problem, whose main idea can be divided into following parts: i) creating a view-transform subject detection module to transform the FPV to a virtual BEV including localization and orientation of each pedestrian, ii) deriving a geometric transformation based method to estimate camera localization and view direction, i.e., the camera registration in a unified BEV, iii) making use of spatial and appearance information to aggregate the subjects into the unified BEV. We collect a new large-scale synthetic dataset with rich annotations for evaluation. The experimental results show the remarkable effectiveness of our proposed method. | ['Song Wang', 'Feifan Wang', 'Wei Feng', 'Ruize Han', 'Zekun Qian'] | 2022-12-19 | null | null | null | null | ['camera-calibration', 'camera-localization'] | ['computer-vision', 'computer-vision'] | [-1.23905540e-01 -2.89394706e-01 6.15104795e-01 -4.31657553e-01
-6.33486152e-01 -7.66300797e-01 3.01566780e-01 -4.18669283e-01
-4.13885862e-01 3.90254110e-01 9.55348462e-02 2.79141366e-01
3.81672412e-01 -4.60166425e-01 -8.26123893e-01 -7.45298386e-01
7.37610996e-01 3.91979635e-01 5.70495307e-01 -1.83316041e-02
-8.17272440e-02 2.55445987e-01 -1.39020097e+00 -1.31073296e-01
5.76299310e-01 7.93190718e-01 2.65392214e-01 6.53018892e-01
6.30644858e-01 2.70437926e-01 -4.43779558e-01 -8.31118405e-01
4.98532414e-01 -3.14906061e-01 -4.85164046e-01 8.25545430e-01
8.94334137e-01 -4.86344665e-01 -1.58414822e-02 1.17410207e+00
5.93484700e-01 2.88268954e-01 2.81816155e-01 -1.52687645e+00
-2.78403878e-01 -2.18832687e-01 -9.36652124e-01 -9.01420116e-02
8.77610147e-01 1.88803121e-01 5.76267123e-01 -1.03600752e+00
7.23434806e-01 1.29407620e+00 6.32406235e-01 4.78435010e-01
-9.54492390e-01 -4.46306974e-01 3.32458138e-01 2.47676328e-01
-1.64472401e+00 -4.90617365e-01 8.13287437e-01 -6.23608172e-01
2.60026395e-01 1.90470487e-01 9.78664875e-01 1.04372787e+00
-2.21562460e-01 3.56558263e-01 1.13037753e+00 -3.50318074e-01
2.20404074e-01 4.89771307e-01 3.59724283e-01 6.74450934e-01
3.99559855e-01 -2.61472583e-01 -1.90325931e-01 -1.94505118e-02
1.00412536e+00 4.38295722e-01 -5.79188049e-01 -8.79443169e-01
-1.30466044e+00 4.26411599e-01 3.59076947e-01 -7.70220235e-02
-2.34218612e-01 -1.55109987e-01 9.17883515e-02 -2.05973968e-01
6.54725283e-02 -2.77959168e-01 -4.59443539e-01 2.42783368e-01
-4.95236576e-01 3.58520091e-01 3.93611014e-01 1.40292501e+00
7.17193663e-01 -2.83692002e-01 3.12888362e-02 6.87031865e-01
4.56445187e-01 8.78621578e-01 4.00476269e-02 -8.85991812e-01
8.39869440e-01 6.62379444e-01 6.94510460e-01 -1.17960584e+00
-4.16599244e-01 -2.20122933e-01 -7.49952674e-01 1.35505006e-01
7.63149977e-01 -2.98835188e-01 -3.81939650e-01 1.62949514e+00
8.26360464e-01 1.52346259e-02 -2.06208546e-02 1.26089573e+00
8.12758684e-01 4.93353218e-01 -3.53305489e-01 -2.21725598e-01
1.93467307e+00 -1.04514515e+00 -5.55685222e-01 -4.60242480e-01
1.35932371e-01 -7.49744177e-01 1.02064574e+00 3.48811418e-01
-1.18510675e+00 -8.26722145e-01 -8.75188351e-01 -1.72654614e-01
-1.74033090e-01 6.92726314e-01 1.78408757e-01 8.31988573e-01
-9.63000894e-01 -2.29505464e-01 -5.74828506e-01 -7.81401336e-01
5.08565791e-02 1.99293241e-01 -7.87610769e-01 -2.66724139e-01
-6.77522242e-01 7.02702701e-01 2.96643764e-01 4.50497806e-01
-8.58959556e-01 -3.39389205e-01 -9.85108078e-01 -1.37660235e-01
6.58020973e-01 -1.21130860e+00 8.56550157e-01 -8.41356874e-01
-1.24811971e+00 1.03266120e+00 -4.05041337e-01 3.51353198e-01
7.65094638e-01 -1.92737937e-01 -2.71956772e-01 4.60941754e-02
2.85597980e-01 5.16364463e-02 6.16164684e-01 -1.74970436e+00
-7.83861876e-01 -9.97580767e-01 1.30748197e-01 5.02903342e-01
-3.61995175e-02 2.09412053e-01 -1.11906421e+00 -3.08936834e-01
2.89224774e-01 -9.95832145e-01 -9.43525732e-02 3.70569676e-02
-6.46717548e-01 -2.76064477e-03 5.48126459e-01 -1.02943313e+00
6.10234857e-01 -1.92015481e+00 4.19028878e-01 8.93127099e-02
3.51526886e-01 -5.08495308e-02 1.30641952e-01 2.73176450e-02
-2.06533924e-01 -5.41982293e-01 7.96145946e-02 -6.50333107e-01
-2.83596754e-01 -1.92136794e-01 1.63196966e-01 8.42113853e-01
-3.91715139e-01 5.51440537e-01 -9.14559126e-01 -5.98820567e-01
5.75489342e-01 5.28483331e-01 -5.14938295e-01 6.28715456e-01
4.37538385e-01 6.97862923e-01 -2.89770991e-01 5.99053860e-01
1.14631581e+00 -1.30153894e-01 -1.07108252e-02 -5.21880448e-01
-8.51169080e-02 -4.81842488e-01 -1.88117945e+00 1.63524652e+00
-2.38862559e-01 1.41852766e-01 2.94167340e-01 -4.84200031e-01
6.77878439e-01 1.67804971e-01 4.71762717e-01 -2.59608209e-01
2.57841676e-01 -1.55714154e-01 -4.59477305e-01 -5.80769837e-01
3.90536338e-01 2.24648669e-01 -5.59124649e-02 1.74385637e-01
2.62811154e-01 1.09000891e-01 1.25636339e-01 1.12932473e-01
3.92705560e-01 2.33416364e-01 3.70821446e-01 7.90895298e-02
9.86941993e-01 -4.35478806e-01 7.45673478e-01 1.96125254e-01
-1.56506628e-01 9.63201046e-01 2.23840788e-01 -5.06727874e-01
-1.12784827e+00 -1.22231698e+00 2.19751477e-01 6.55819535e-01
5.60836434e-01 -3.75744939e-01 -1.21650922e+00 -5.96603811e-01
-3.11474204e-01 2.83303738e-01 -4.37053084e-01 3.06782871e-01
-4.75173622e-01 -7.55371213e-01 -1.28952250e-01 4.62518096e-01
9.41656053e-01 -3.66409093e-01 -6.14468575e-01 -3.79928619e-01
-7.23676085e-01 -1.72274685e+00 -9.37346399e-01 -4.73193556e-01
-4.54920501e-01 -1.46181619e+00 -1.05281103e+00 -7.71027684e-01
1.01982260e+00 8.20680857e-01 8.20113957e-01 -4.03926164e-01
6.09290823e-02 6.30516946e-01 -7.78557137e-02 -1.12299703e-01
2.26327434e-01 -4.10498589e-01 3.34364027e-01 6.20131731e-01
1.51052967e-01 -4.50987220e-01 -9.50789511e-01 8.04770708e-01
-2.98968613e-01 1.20418079e-01 3.34419489e-01 4.75891709e-01
7.15627909e-01 -1.93068951e-01 -2.33943313e-01 -6.17286086e-01
-4.40054515e-04 5.48701882e-02 -8.93219829e-01 5.83458364e-01
5.86571451e-03 -6.72021866e-01 4.31645870e-01 -1.59910217e-01
-1.07534873e+00 6.55873299e-01 -4.18200111e-03 -4.69930261e-01
-3.72074753e-01 -4.28643256e-01 -1.04941964e+00 -1.73007146e-01
3.80514115e-01 3.18863392e-01 -2.09683746e-01 -3.50497335e-01
5.88315725e-01 7.15011597e-01 7.62053251e-01 -3.17056626e-01
9.64187026e-01 6.94562674e-01 -1.81780770e-01 -8.96001875e-01
-7.07721472e-01 -7.54727602e-01 -1.33409524e+00 -4.98601735e-01
1.38682795e+00 -1.23679626e+00 -1.04659343e+00 6.06591284e-01
-1.55422258e+00 1.84277028e-01 6.37421431e-03 5.54484725e-01
-5.55690587e-01 6.63801074e-01 -1.64084703e-01 -7.61685252e-01
-1.45529270e-01 -1.41273725e+00 1.44730175e+00 4.08966243e-01
2.69864559e-01 -8.25949490e-01 -4.22354974e-02 8.21686685e-01
-2.68033832e-01 4.09040064e-01 2.11863756e-01 -8.84980038e-02
-7.46119499e-01 -3.13757330e-01 -2.75229782e-01 3.46983373e-01
-1.15996137e-01 -1.78481892e-01 -1.19628668e+00 -3.57611567e-01
2.98585206e-01 1.57906786e-01 2.69812226e-01 5.48641980e-01
6.41698778e-01 -6.22574650e-02 -5.39823353e-01 1.06677878e+00
1.61576056e+00 9.57271531e-02 4.74028051e-01 1.88521937e-01
1.12573111e+00 6.99617386e-01 6.34201050e-01 3.08690280e-01
8.78099442e-01 1.28920794e+00 4.24528748e-01 -3.37604642e-01
-1.00375742e-01 -2.25902572e-01 6.02344811e-01 7.50849068e-01
-5.15367031e-01 -1.14662588e-01 -6.34654760e-01 2.44670749e-01
-1.76183116e+00 -8.81622016e-01 -5.33492148e-01 2.53190708e+00
1.71217322e-02 -3.64299744e-01 4.95719910e-01 -5.98798990e-02
1.00003958e+00 -1.78312600e-01 -2.86708176e-01 3.54994118e-01
6.33564591e-02 -5.91838837e-01 5.12253940e-01 4.72978771e-01
-1.20434594e+00 7.25522637e-01 5.05940342e+00 2.48807311e-01
-5.89965761e-01 3.63880485e-01 4.09898460e-01 1.76259577e-01
3.38400751e-01 -1.10975355e-01 -1.15614283e+00 5.27838945e-01
2.20699564e-01 3.33269030e-01 3.76400560e-01 8.35039496e-01
2.52596706e-01 -1.86177358e-01 -1.06647182e+00 1.75222921e+00
6.12782478e-01 -6.19326770e-01 -1.82346985e-01 -1.78145356e-02
5.65976322e-01 -4.13133740e-01 -1.89047694e-01 -1.63255125e-01
7.66800642e-02 -3.28796327e-01 1.09328401e+00 6.70187473e-01
6.86914563e-01 -5.08303583e-01 7.67068386e-01 3.69196087e-01
-1.54967344e+00 -4.68053930e-02 -5.69579422e-01 1.78553507e-01
4.06690687e-01 2.32943535e-01 -3.14782262e-01 8.82263184e-01
9.06033576e-01 7.97347844e-01 -1.00801790e+00 9.86808717e-01
-2.16727868e-01 -2.53355354e-01 -3.56039545e-03 3.73420358e-01
-3.92238140e-01 -8.02443743e-01 5.41793942e-01 7.76252925e-01
4.01324928e-01 1.35316864e-01 3.18197995e-01 8.61483157e-01
3.43607306e-01 4.58095148e-02 -4.97469455e-01 9.15977716e-01
2.66742676e-01 1.51940620e+00 -7.78138816e-01 -2.87323415e-01
-6.76543534e-01 1.36666870e+00 2.67360955e-02 5.96826792e-01
-1.08023345e+00 -3.32116008e-01 5.19033074e-01 2.52192318e-01
3.24633896e-01 -7.77894333e-02 1.49478525e-01 -1.81618547e+00
5.67837119e-01 -6.49125099e-01 3.59055936e-01 -1.32881594e+00
-9.93750513e-01 7.65011072e-01 1.34719148e-01 -1.55440140e+00
2.59281807e-02 -6.50615752e-01 -4.06781644e-01 1.01175475e+00
-9.62923646e-01 -1.72352469e+00 -1.05865598e+00 1.10703731e+00
4.09438074e-01 -1.49338588e-01 5.65230787e-01 4.28549081e-01
-8.30829263e-01 5.45702398e-01 3.04143503e-02 2.32397527e-01
8.67880583e-01 -1.17588186e+00 1.80230737e-01 1.20707297e+00
-1.07674755e-01 6.09324574e-01 4.34385866e-01 -3.62307429e-01
-1.50163984e+00 -1.17845392e+00 5.65396488e-01 -1.23450220e+00
5.91594167e-02 -8.53607059e-01 -2.23094895e-01 1.07713175e+00
2.27899458e-02 1.97054476e-01 3.95253122e-01 -5.25032775e-03
-5.55121712e-02 -5.39967835e-01 -1.10483623e+00 4.17996973e-01
1.12627506e+00 -2.39975736e-01 -4.77756798e-01 4.43441123e-01
4.38291818e-01 -7.77710021e-01 -6.45691276e-01 -2.81449437e-01
5.88032722e-01 -1.13038158e+00 1.45502567e+00 -3.48102488e-02
5.50074317e-02 -8.83102894e-01 -2.64443845e-01 -1.24465585e+00
-1.69000700e-01 -2.91635096e-01 4.10445303e-01 1.56396472e+00
-2.47766569e-01 -5.99104702e-01 5.37571192e-01 6.74005926e-01
7.30623230e-02 -3.04669648e-01 -7.37240970e-01 -3.97946835e-01
-5.40998340e-01 -2.41957948e-01 4.16124254e-01 7.09251046e-01
-4.34068829e-01 5.72467148e-01 -6.65808737e-01 6.93615079e-01
1.12626767e+00 3.11461166e-02 1.45318544e+00 -1.18551326e+00
-2.70779461e-01 2.13195562e-01 -7.03519583e-01 -1.07505524e+00
-3.92615020e-01 -4.62874740e-01 -1.64428264e-01 -1.51016760e+00
7.22144961e-01 1.24632753e-01 3.66445512e-01 -1.41667128e-01
-3.51023078e-01 1.73252791e-01 3.43047768e-01 1.00906715e-01
-8.26024354e-01 4.04263794e-01 1.19264984e+00 5.80262765e-02
-5.01091965e-03 2.82610387e-01 -6.60922348e-01 1.00492942e+00
6.78467751e-02 -1.60507545e-01 -3.79212976e-01 -5.32359779e-01
-1.15236249e-02 3.26918870e-01 9.64793086e-01 -1.11526704e+00
2.75537252e-01 1.68413278e-02 8.72019053e-01 -7.60667801e-01
5.80278814e-01 -1.11906052e+00 4.08582449e-01 3.60401988e-01
1.57166228e-01 4.35182810e-01 -1.06472187e-01 7.76408076e-01
4.72902991e-02 3.71934846e-02 9.16602433e-01 -3.93546075e-01
-6.28336966e-01 4.10079241e-01 2.84182698e-01 -1.77586433e-02
1.50602365e+00 -3.10018510e-01 -2.14630902e-01 -2.74644762e-01
-8.32311213e-01 3.56003433e-01 8.68380785e-01 2.57008225e-01
7.70838439e-01 -1.40395343e+00 -5.80449820e-01 4.42920506e-01
2.27381632e-01 1.76713318e-01 6.29083037e-01 8.24958146e-01
-6.26276612e-01 2.92784423e-01 -3.10284227e-01 -9.56903398e-01
-1.81751323e+00 8.48319054e-01 7.55827487e-01 9.38300192e-02
-5.92274964e-01 6.70566916e-01 5.78766525e-01 -6.23887718e-01
1.62446246e-01 -2.82895356e-01 -3.64282697e-01 -3.58015522e-02
6.57770097e-01 4.89928335e-01 -1.11200206e-01 -1.24426794e+00
-5.21169901e-01 1.33442831e+00 3.24215859e-01 -2.05569386e-01
1.28515089e+00 -7.29785979e-01 -9.75367501e-02 2.71258980e-01
1.11689723e+00 1.54447570e-01 -1.24460590e+00 -2.67730474e-01
-7.24552751e-01 -9.46238101e-01 -3.87324095e-01 -4.63018477e-01
-1.24636340e+00 9.97749031e-01 1.04280519e+00 -2.91101456e-01
1.02855730e+00 -6.08792603e-02 5.79185188e-01 2.22240686e-01
7.93029130e-01 -9.39855635e-01 5.05351759e-02 -1.21438913e-02
7.78157651e-01 -1.27494109e+00 1.16284490e-01 -7.85395145e-01
-8.65369499e-01 1.06998515e+00 7.77197123e-01 -2.60999482e-02
3.98882687e-01 -2.97142506e-01 1.05851583e-01 -2.01762810e-01
-6.34309202e-02 -1.65064275e-01 2.39745319e-01 9.73774791e-01
-4.59406935e-02 -4.54798378e-02 2.13414326e-01 8.07102144e-01
-1.37373358e-01 -2.78391391e-01 5.64459622e-01 3.46498340e-01
-1.22756241e-02 -8.92121375e-01 -8.01467597e-01 -2.03935593e-01
-8.52318257e-02 3.85644078e-01 -2.79931217e-01 8.03308368e-01
5.84188342e-01 1.01417148e+00 -1.25086546e-01 -4.19240475e-01
1.02243900e+00 -1.76554188e-01 5.81672966e-01 -5.29952168e-01
-3.62692535e-01 2.39378139e-01 -3.44361663e-01 -7.07499981e-01
-5.41538596e-01 -9.34539616e-01 -5.57024360e-01 -1.69062555e-01
-3.01989377e-01 -8.73865783e-02 5.23766160e-01 7.49835610e-01
3.82742775e-03 3.47583890e-01 8.26708674e-01 -1.19226956e+00
-3.23130190e-01 -7.83650398e-01 -7.60356188e-01 8.60046685e-01
3.88928980e-01 -7.80668676e-01 -1.64704740e-01 4.81555283e-01] | [7.223688125610352, -1.009044885635376] |
252175bb-fbbb-4342-ae8b-a01e00b0cd15 | illiterate-dall-cdot-e-learns-to-compose | null | null | https://openreview.net/forum?id=h0OYV0We3oh | https://openreview.net/pdf?id=h0OYV0We3oh | Illiterate DALL$\cdot$E Learns to Compose | DALL$\cdot$E has shown an impressive ability of composition-based systematic generalization in image generation. This is possible because it utilizes the dataset of text-image pairs where the text provides the source of compositionality. Following this result, an important extending question is whether this compositionality can still be achieved even without conditioning on text. In this paper, we propose an architecture called $\textit{Slot2Seq}$ that achieves this text-free DALL$\cdot$E by learning compositional slot-based representations purely from images, an ability lacking in DALL$\cdot$E. Unlike existing object-centric representation models that decode pixels independently for each slot and each pixel location and compose them via mixture-based alpha composition, we propose to use the Image GPT decoder conditioned on the slots for a more flexible generation by capturing complex interaction among the pixels and the slots.
In experiments, we show that this simple architecture achieves zero-shot generation of novel images without text and better quality in generation than the models based on mixture decoders. | ['Sungjin Ahn', 'Fei Deng', 'Gautam Singh'] | 2021-09-29 | null | null | null | iclr-2022-4 | ['systematic-generalization'] | ['reasoning'] | [ 7.13097990e-01 6.45836234e-01 2.20900148e-01 -1.75255075e-01
-9.52389657e-01 -4.10482079e-01 8.33686709e-01 -3.79598856e-01
-2.82892048e-01 8.43183279e-01 4.33727205e-02 -2.33304635e-01
-2.86056716e-02 -9.98359382e-01 -1.09074128e+00 -8.89013529e-01
1.25511944e-01 6.63703084e-01 1.61708727e-01 -4.51235175e-01
-7.40835294e-02 -4.62655574e-02 -2.00358796e+00 6.92921400e-01
8.91762078e-01 8.70078206e-01 7.08359241e-01 8.00395250e-01
-5.21255374e-01 5.51828861e-01 -7.47179508e-01 -5.14520228e-01
6.12639844e-01 -1.11624444e+00 -3.50774199e-01 3.08350682e-01
5.75223565e-01 -2.72304773e-01 -1.95095405e-01 8.89551044e-01
5.73781252e-01 -9.77810007e-03 8.02757263e-01 -1.00799131e+00
-7.79011488e-01 1.04901934e+00 -4.14089262e-01 -2.80326396e-01
1.03156723e-01 3.87927979e-01 9.39313948e-01 -7.39562511e-01
9.98708248e-01 1.20792603e+00 3.48894715e-01 9.69788313e-01
-1.36523378e+00 -6.41763926e-01 2.22959638e-01 -9.66978744e-02
-1.29076362e+00 -4.72323745e-01 5.94408929e-01 -3.85576844e-01
9.14469421e-01 2.40881324e-01 5.84502518e-01 1.14777887e+00
-4.46854532e-02 1.00321209e+00 1.14568436e+00 -7.01543748e-01
2.66221315e-01 1.95838094e-01 -4.03558493e-01 8.91522348e-01
5.81249036e-02 7.74999186e-02 -6.03900552e-01 3.30753177e-01
8.63085687e-01 -2.27454379e-01 -1.55526519e-01 -2.62348622e-01
-1.30094910e+00 7.56657362e-01 3.52893203e-01 2.54404813e-01
-1.83445856e-01 6.72009647e-01 8.71929973e-02 3.21029097e-01
3.90516430e-01 4.82949436e-01 -3.03562909e-01 3.29615138e-02
-1.29996490e+00 3.66559774e-01 3.81252497e-01 1.31318974e+00
9.41389024e-01 4.88615364e-01 -3.42941135e-01 6.54650927e-01
9.75975916e-02 5.96856594e-01 7.22670734e-01 -9.61367130e-01
2.32935309e-01 4.68850285e-01 -4.59836498e-02 -2.68568993e-01
-1.46982878e-01 -4.72537637e-01 -8.22738588e-01 4.33656871e-01
2.54806846e-01 -2.71384865e-01 -1.49363708e+00 2.10921264e+00
5.45781292e-02 -7.77469501e-02 9.01115388e-02 5.65315604e-01
7.17434585e-01 7.29629815e-01 6.64680675e-02 -1.53165177e-01
1.31538057e+00 -1.14955044e+00 -5.09208441e-01 -8.87508541e-02
5.46447515e-01 -7.33643651e-01 1.12961900e+00 2.88661242e-01
-1.42578864e+00 -7.99538493e-01 -1.08257997e+00 -6.37990311e-02
-4.65904504e-01 1.91151202e-01 7.07740128e-01 1.00172126e+00
-1.49750769e+00 4.90248293e-01 -3.70427728e-01 -2.87476271e-01
3.54203373e-01 3.21190417e-01 -1.09688513e-01 -1.24900274e-01
-1.03674150e+00 6.28942966e-01 3.93770754e-01 -1.13590010e-01
-1.24480748e+00 -5.78361094e-01 -8.42333496e-01 -5.74729666e-02
2.97881693e-01 -1.26922631e+00 1.15143824e+00 -1.35112906e+00
-1.56262898e+00 7.85233259e-01 -2.84384191e-01 -6.55920804e-01
5.39370060e-01 2.98309058e-01 -3.16646546e-02 2.44574189e-01
1.50749668e-01 1.55887449e+00 1.14761174e+00 -1.47480071e+00
-5.77688992e-01 -2.45515525e-01 2.28277907e-01 2.31468171e-01
5.58175240e-03 -4.39660460e-01 -4.49978948e-01 -7.79733598e-01
6.47239238e-02 -1.01734269e+00 -3.46553087e-01 1.13040276e-01
-3.71333510e-01 5.43559939e-02 3.44099879e-01 -4.13688153e-01
7.76690423e-01 -2.09420347e+00 2.71973640e-01 -4.98714782e-02
8.28023553e-02 1.43131554e-01 -4.82343048e-01 5.24774492e-01
-3.07487771e-02 2.49696255e-01 -5.01894236e-01 -8.11645627e-01
2.50668943e-01 1.77245796e-01 -4.55793381e-01 -1.22827575e-01
2.97115207e-01 1.16327083e+00 -5.83646059e-01 -3.80383164e-01
9.38126445e-02 3.27714711e-01 -7.62753367e-01 -7.61828572e-02
-9.80235696e-01 2.18984544e-01 -6.75063729e-02 3.24253231e-01
5.71531773e-01 -1.51440904e-01 1.88414037e-01 1.37033552e-01
7.64651969e-02 4.92918082e-02 -1.02175915e+00 2.24342155e+00
-3.65278393e-01 4.82843280e-01 3.55279036e-02 -9.66872871e-01
8.86850297e-01 3.21082145e-01 3.15323234e-01 -9.29105341e-01
-3.88400853e-02 2.56716251e-01 3.84444967e-02 -7.77142122e-02
8.14753950e-01 -5.12012661e-01 -2.18052328e-01 6.60041749e-01
3.78736556e-01 -6.03794932e-01 5.28093815e-01 3.59019101e-01
9.60573852e-01 5.38347423e-01 -2.07901463e-01 -2.53974974e-01
1.79348215e-01 3.10442895e-02 1.94840789e-01 1.07913387e+00
2.82901287e-01 1.03187478e+00 3.51813108e-01 -9.16140005e-02
-1.31126761e+00 -1.27074659e+00 1.34655491e-01 8.90215695e-01
3.67910671e-03 -4.07473803e-01 -1.07324588e+00 -5.07021606e-01
-5.27261019e-01 7.88936496e-01 -7.99065650e-01 -1.11556634e-01
-3.87604177e-01 -8.58231723e-01 6.34053111e-01 2.40349904e-01
7.20308721e-01 -1.06818390e+00 -5.74009061e-01 1.77051619e-01
-2.64021099e-01 -8.56527388e-01 -4.16763335e-01 3.40215653e-01
-7.46111691e-01 -6.07940555e-01 -1.10499561e+00 -7.24543989e-01
7.32030392e-01 8.72388035e-02 1.15734124e+00 -6.84158430e-02
-5.36021054e-01 4.26634759e-01 -3.64314973e-01 -5.10373354e-01
-4.88859445e-01 9.42958742e-02 -4.66442734e-01 -9.64688659e-02
-1.40839800e-01 -6.83956563e-01 -6.12509012e-01 -3.86466365e-03
-1.43314302e+00 6.38845801e-01 7.70536721e-01 9.29846048e-01
5.52249670e-01 1.60320848e-01 5.72347224e-01 -9.60150659e-01
3.98153245e-01 -2.98726320e-01 -2.60102540e-01 1.72646105e-01
-4.51002479e-01 3.95538330e-01 6.65282190e-01 -3.56932759e-01
-1.21753252e+00 1.54034480e-01 -2.66001165e-01 -2.18536064e-01
-2.20221013e-01 1.32195100e-01 -1.80180997e-01 3.40489358e-01
8.50061059e-01 7.95072079e-01 5.15602082e-02 -2.59787798e-01
8.30062568e-01 1.97265729e-01 2.84697354e-01 -6.75555408e-01
6.06444776e-01 6.17933333e-01 -4.02222462e-02 -8.23199213e-01
-2.90737629e-01 4.44829883e-03 -2.78866202e-01 -1.01210214e-02
1.05724967e+00 -1.03957760e+00 -2.58533895e-01 4.13761258e-01
-1.06444752e+00 -6.54516995e-01 -8.65125954e-01 8.20411742e-02
-9.81046677e-01 3.05302739e-01 -5.19253433e-01 -8.65250468e-01
-2.15371862e-01 -1.32853866e+00 1.29768622e+00 3.68346088e-02
-3.55164930e-02 -7.33687222e-01 -2.13590249e-01 2.81192422e-01
4.91567165e-01 -3.80953588e-03 1.02364135e+00 -8.95538703e-02
-1.13713157e+00 1.65101245e-01 -2.03469113e-01 3.06352645e-01
-1.33120596e-01 -2.31627285e-01 -9.14198458e-01 -1.22185811e-01
-1.66348517e-01 -2.51686841e-01 1.31537187e+00 4.59275633e-01
8.85221064e-01 -1.81630313e-01 -1.45973146e-01 5.57365000e-01
1.48728406e+00 3.28916311e-01 1.02167583e+00 3.53720821e-02
4.18596208e-01 5.72180271e-01 -2.78544109e-02 3.41523707e-01
4.53004509e-01 6.82407618e-01 3.27233583e-01 -2.89270371e-01
-7.44268954e-01 -3.91705513e-01 5.11230767e-01 3.92309189e-01
8.79639536e-02 -5.81931293e-01 -4.44229633e-01 6.37666702e-01
-1.76527834e+00 -1.04588616e+00 7.42038414e-02 1.84592044e+00
9.02240038e-01 -2.13949066e-02 1.62920151e-02 3.55598032e-02
5.40727854e-01 9.87427533e-02 -3.21058154e-01 -3.12027901e-01
-5.44304550e-01 7.91616559e-01 4.55581725e-01 4.92748648e-01
-7.20209777e-01 1.08178532e+00 6.77130985e+00 1.27954078e+00
-9.29292202e-01 2.21597746e-01 7.45708466e-01 -3.16618353e-01
-8.58304083e-01 -2.34560091e-02 -8.23266149e-01 5.94763935e-01
9.51082468e-01 1.76120341e-01 4.91480649e-01 5.32783806e-01
-8.22826196e-03 -3.98032457e-01 -8.76815259e-01 8.75696659e-01
3.95867944e-01 -1.55910289e+00 4.80156451e-01 9.50191468e-02
1.06498468e+00 -3.03831846e-01 4.97840941e-01 2.88368642e-01
5.70216477e-01 -1.01082277e+00 1.09156084e+00 4.35604721e-01
9.87646401e-01 -4.23289895e-01 1.41655549e-01 4.04668570e-01
-8.66883457e-01 -2.82310247e-02 -2.81986922e-01 2.92507838e-02
2.79076606e-01 5.92955828e-01 -9.11998332e-01 6.17286801e-01
3.15124959e-01 8.35597813e-02 -5.57546198e-01 6.15391910e-01
-2.73446470e-01 2.51919895e-01 -2.32446313e-01 8.23038965e-02
2.91518837e-01 -5.30734695e-02 3.95993620e-01 1.25075376e+00
7.88558662e-01 -1.13178410e-01 -6.00070618e-02 1.21407461e+00
-1.72928683e-02 -1.47528708e-01 -6.44898891e-01 -3.42357419e-02
-2.24339925e-02 9.21899021e-01 -9.03331399e-01 -6.69282556e-01
3.13301710e-03 1.47659242e+00 -8.40228982e-03 3.65685791e-01
-8.70026290e-01 -1.57967508e-01 4.22992289e-01 2.07783818e-01
6.63457632e-01 -3.69873226e-01 -3.45502317e-01 -1.25400710e+00
-1.13342777e-01 -9.01186466e-01 7.19739273e-02 -1.10933363e+00
-8.93381059e-01 7.31240869e-01 1.45304516e-01 -8.93690825e-01
-4.71763194e-01 -5.88409662e-01 -2.04973251e-01 9.09040153e-01
-1.40725231e+00 -1.48903382e+00 -7.85744051e-04 7.41521478e-01
7.59474099e-01 -1.82077110e-01 8.98908198e-01 1.80735841e-01
-2.05911323e-01 6.37048364e-01 -4.92655188e-02 -2.59575963e-01
5.03257275e-01 -1.35518599e+00 4.22926515e-01 9.61737573e-01
4.62147295e-01 5.21808088e-01 7.26290286e-01 -5.49413383e-01
-1.12360263e+00 -1.00184453e+00 8.16741645e-01 -3.76283646e-01
-3.42309312e-03 -6.92419827e-01 -2.44372830e-01 6.07026398e-01
5.13767660e-01 -4.96252477e-01 6.15388095e-01 -3.72198552e-01
-4.33635920e-01 -7.50605240e-02 -1.00132310e+00 9.72115219e-01
1.44968522e+00 -4.00563121e-01 -3.73360217e-01 2.18817309e-01
9.41974223e-01 -1.53652549e-01 -3.09681565e-01 2.29006931e-01
2.89716572e-01 -1.25930393e+00 9.20822918e-01 -1.14040822e-01
7.52489626e-01 -4.16377246e-01 -3.33284169e-01 -1.21913362e+00
-1.56704411e-01 -6.92125559e-01 3.44956547e-01 9.55193579e-01
8.19190741e-01 -4.25800741e-01 9.33228791e-01 4.55571383e-01
-4.81660694e-01 -2.70855665e-01 -8.66524756e-01 -7.41406322e-01
3.22226018e-01 -6.40214384e-01 6.50548697e-01 3.44256520e-01
-2.05044463e-01 3.53174627e-01 -4.39578176e-01 -4.51195836e-01
5.43527007e-01 1.94360420e-01 8.76884460e-01 -6.13129437e-01
-7.52601862e-01 -6.94504619e-01 -2.46433467e-01 -1.39025211e+00
1.97539646e-02 -9.78771210e-01 3.91460568e-01 -1.66196167e+00
1.84230536e-01 -5.98881841e-01 1.28255233e-01 2.80883431e-01
-4.29966934e-02 6.59767985e-01 5.05663812e-01 1.99642982e-02
-6.22567117e-01 6.01405442e-01 1.52542877e+00 -2.73653448e-01
1.15549892e-01 -4.50715005e-01 -8.93539667e-01 3.91943514e-01
5.37759781e-01 -1.67552724e-01 -6.99449301e-01 -6.29887044e-01
2.56683707e-01 1.38280615e-01 2.38566086e-01 -1.11993313e+00
6.67831823e-02 1.61755979e-01 3.81164134e-01 -3.49114031e-01
7.57219672e-01 -5.79837799e-01 5.09240270e-01 3.04393828e-01
-4.80102867e-01 -1.59255207e-01 1.54480562e-01 5.14160395e-01
-1.01848923e-01 -3.84057075e-01 6.25047922e-01 -7.95595944e-01
-6.74353004e-01 1.43314615e-01 -5.89292765e-01 -2.98307270e-01
9.44915771e-01 -4.43805099e-01 -3.76339376e-01 -4.95046824e-01
-8.92427325e-01 -3.29915822e-01 7.22615421e-01 1.97730809e-01
6.49087846e-01 -1.16029501e+00 -6.22063398e-01 4.83870178e-01
2.14297436e-02 -1.33325858e-02 6.46288812e-01 4.99355167e-01
-2.89382368e-01 3.38713646e-01 -2.26965293e-01 -5.55734396e-01
-7.92138815e-01 6.17832422e-01 3.20292592e-01 -3.08057219e-01
-2.56398797e-01 1.11720002e+00 7.16112375e-01 -3.11782986e-01
-1.24790847e-01 -2.32506752e-01 3.45695287e-01 -5.84078254e-03
4.15681422e-01 -1.39856353e-01 -9.20316800e-02 -4.40071434e-01
1.91013873e-01 4.32415366e-01 1.68052390e-01 -8.47074628e-01
1.03142643e+00 -2.43708223e-01 -6.35861456e-02 3.27128261e-01
8.54294896e-01 -1.60061479e-01 -1.44232178e+00 1.85102537e-01
-4.35971111e-01 -2.74865776e-01 -4.29906398e-01 -1.04659557e+00
-9.52604592e-01 1.04705966e+00 5.35222769e-01 3.54271457e-02
1.19076264e+00 -5.53787164e-02 7.40910709e-01 1.01514705e-01
4.69830543e-01 -9.49437737e-01 3.91572714e-01 3.20626646e-01
7.92844832e-01 -8.86254787e-01 -4.73829627e-01 -3.29694539e-01
-5.90924561e-01 8.20191145e-01 4.77954835e-01 -1.71065465e-01
4.15526509e-01 2.29556292e-01 -1.42972633e-01 4.66914475e-02
-9.00266588e-01 -6.62440181e-01 -2.93373968e-02 8.99707735e-01
4.24040675e-01 6.89135492e-02 -3.66385132e-01 4.38260317e-01
-3.80888492e-01 -7.15511367e-02 4.56347704e-01 8.00873101e-01
-5.18192172e-01 -1.48779619e+00 -2.18323395e-01 2.66834378e-01
-3.30994099e-01 -5.08151889e-01 -2.51067370e-01 6.19617045e-01
6.68300152e-01 7.93773115e-01 2.24880949e-01 -3.01814616e-01
-1.11898102e-01 4.48267758e-01 9.48804796e-01 -7.66132772e-01
-5.40938318e-01 2.83577561e-01 1.81370899e-02 -2.64597744e-01
-4.99233603e-01 -5.52220166e-01 -1.18724000e+00 -1.16518907e-01
-2.51816828e-02 4.13443595e-02 8.42459917e-01 7.20713437e-01
5.08800566e-01 7.65613675e-01 1.94539323e-01 -8.56126308e-01
-1.24229871e-01 -8.16523075e-01 -7.07938671e-01 4.70905930e-01
8.97608474e-02 -3.07770491e-01 -1.07872047e-01 5.08213997e-01] | [11.31167221069336, -0.18442033231258392] |
0b4657d2-a494-4386-ba3a-c9ef20923765 | cook-gen-robust-generative-modeling-of | 2306.01805 | null | https://arxiv.org/abs/2306.01805v1 | https://arxiv.org/pdf/2306.01805v1.pdf | Cook-Gen: Robust Generative Modeling of Cooking Actions from Recipes | As people become more aware of their food choices, food computation models have become increasingly popular in assisting people in maintaining healthy eating habits. For example, food recommendation systems analyze recipe instructions to assess nutritional contents and provide recipe recommendations. The recent and remarkable successes of generative AI methods, such as auto-regressive large language models, can lead to robust methods for a more comprehensive understanding of recipes for healthy food recommendations beyond surface-level nutrition content assessments. In this study, we explore the use of generative AI methods to extend current food computation models, primarily involving the analysis of nutrition and ingredients, to also incorporate cooking actions (e.g., add salt, fry the meat, boil the vegetables, etc.). Cooking actions are notoriously hard to model using statistical learning methods due to irregular data patterns - significantly varying natural language descriptions for the same action (e.g., marinate the meat vs. marinate the meat and leave overnight) and infrequently occurring patterns (e.g., add salt occurs far more frequently than marinating the meat). The prototypical approach to handling irregular data patterns is to increase the volume of data that the model ingests by orders of magnitude. Unfortunately, in the cooking domain, these problems are further compounded with larger data volumes presenting a unique challenge that is not easily handled by simply scaling up. In this work, we propose novel aggregation-based generative AI methods, Cook-Gen, that reliably generate cooking actions from recipes, despite difficulties with irregular data patterns, while also outperforming Large Language Models and other strong baselines. | ['Amit Sheth', 'Vignesh Narayanan', 'Yuxin Zi', 'Renjith Prasad', 'Kanak Raj', 'Kaushik Roy', 'Revathy Venkataramanan'] | 2023-06-01 | null | null | null | null | ['food-recommendation'] | ['miscellaneous'] | [ 2.05775231e-01 -1.03056215e-01 -2.49258727e-01 -4.40607131e-01
-3.59797180e-01 -8.48942757e-01 5.20288169e-01 7.59591758e-01
-5.40564023e-02 3.06668669e-01 8.24491382e-01 -3.13662350e-01
1.23252824e-01 -1.23312271e+00 -7.28276610e-01 -6.09174728e-01
6.12514131e-02 5.56519628e-01 -4.41634774e-01 -6.41602099e-01
-1.16443805e-01 -2.25423664e-01 -1.66928756e+00 4.71339673e-01
8.55880797e-01 3.74065727e-01 1.24348849e-01 5.62085271e-01
-3.56020689e-01 9.15129304e-01 -8.73810723e-02 -2.68965214e-01
2.20857650e-01 -1.00359154e+00 -3.33587706e-01 2.89549261e-01
3.56073469e-01 -4.77172434e-01 3.22813064e-01 1.01256347e+00
4.29055467e-02 3.79949242e-01 6.85193777e-01 -9.58891332e-01
-1.21354413e+00 1.26794493e+00 -2.15982720e-01 -1.89622283e-01
6.87661231e-01 4.30215597e-01 1.11606145e+00 -5.60028970e-01
1.49709255e-01 1.35891259e+00 8.33130896e-01 5.44190347e-01
-1.64466512e+00 -4.49828535e-01 4.24237013e-01 -8.15267712e-02
-1.14126289e+00 -4.12501127e-01 7.41552830e-01 -4.05628115e-01
1.07742655e+00 4.99879301e-01 1.10183120e+00 9.94506121e-01
3.52089256e-01 8.05857658e-01 1.01904213e+00 -3.21615666e-01
4.69445318e-01 -1.38424829e-01 2.58769810e-01 5.29199719e-01
6.12486601e-01 2.12689638e-01 -1.93018705e-01 -3.73425446e-02
4.10267532e-01 5.60415149e-01 2.77770609e-01 -2.59794474e-01
-1.26466298e+00 1.25050640e+00 2.38199234e-01 1.59074932e-01
-8.59782338e-01 2.06240565e-01 2.31953189e-01 1.96633309e-01
5.75911343e-01 4.10294920e-01 -6.47453129e-01 1.39188126e-01
-1.07778823e+00 8.14608097e-01 1.11094666e+00 8.56983364e-01
6.87758207e-01 1.38229907e-01 2.74592340e-02 6.66328251e-01
6.06257975e-01 7.07436383e-01 2.51540899e-01 -7.62715042e-01
1.20768316e-01 4.75610197e-01 1.94646329e-01 -1.04915786e+00
-6.55634880e-01 1.85315475e-01 -1.07517278e+00 -9.86278206e-02
5.50565302e-01 -1.34230018e-01 -1.07343292e+00 1.73745918e+00
4.23838556e-01 -3.35116863e-01 -1.50266774e-02 8.15847993e-01
8.66008282e-01 7.30042398e-01 1.03846323e+00 -7.92172924e-02
1.70921743e+00 -6.53043926e-01 -7.41093755e-01 -2.57216483e-01
6.19130731e-01 -8.22435617e-01 1.05854881e+00 4.96850640e-01
-9.32389855e-01 -7.17037618e-01 -6.24770403e-01 -1.14419460e-01
-7.00552762e-01 -9.07770246e-02 1.11079407e+00 8.29054058e-01
-8.47700477e-01 6.04991138e-01 -1.10190713e+00 -5.27097225e-01
1.09368749e-02 1.18412003e-01 -3.34191392e-03 -5.00231743e-01
-1.11331499e+00 8.49516630e-01 7.22322285e-01 9.50653926e-02
-7.00681627e-01 -1.20201218e+00 -1.43728018e+00 1.17538041e-02
5.54662347e-01 -7.13579595e-01 1.26614356e+00 -9.75165486e-01
-1.42435730e+00 6.08624041e-01 1.03131413e-01 -3.03839862e-01
-2.94134654e-02 -3.32248181e-01 -6.36153281e-01 -5.65546870e-01
1.71163097e-01 8.47697735e-01 4.28092152e-01 -8.87468517e-01
-2.64430285e-01 -3.98175478e-01 6.87343031e-02 9.59344953e-02
1.13172978e-01 1.63513437e-01 4.88300055e-01 -7.37294912e-01
-6.59861639e-02 -9.33577359e-01 -7.03465402e-01 -1.46335661e-01
-4.78542864e-01 -3.87264282e-01 1.76725611e-01 -8.14780414e-01
1.26835954e+00 -2.02982879e+00 8.30840170e-02 2.48337369e-02
1.93142623e-01 -6.49769455e-02 -1.45525798e-01 6.38993263e-01
-1.09226005e-02 3.03951204e-01 -1.79190189e-01 -5.19895628e-02
5.78666031e-01 4.56179053e-01 7.47672841e-02 2.61100560e-01
2.51263976e-01 1.03619647e+00 -1.25621140e+00 -2.78035492e-01
4.57653642e-01 5.04660249e-01 -1.01497114e+00 5.09069636e-02
-9.96447444e-01 -3.14947069e-02 -3.52470726e-01 5.13801575e-01
4.90870923e-01 -1.92750320e-01 4.82418567e-01 -5.41428745e-01
-4.46125448e-01 4.75331277e-01 -1.41896653e+00 1.68987906e+00
-2.05983832e-01 -2.37118170e-01 -6.98970258e-02 -7.89591253e-01
6.99723244e-01 2.18840510e-01 7.40225077e-01 -5.31689763e-01
-7.13430345e-02 -1.36458175e-02 2.69946843e-01 -4.89856124e-01
7.41975248e-01 -5.94680727e-01 -3.65638644e-01 8.04765940e-01
-1.55039132e-01 -2.90570855e-01 6.95524931e-01 -4.29956764e-02
6.86182737e-01 6.70157611e-01 7.66773760e-01 -5.93585432e-01
1.15285151e-01 4.05985951e-01 4.95847851e-01 3.72277141e-01
3.15399803e-02 2.53661156e-01 -1.98394299e-01 -7.01175630e-01
-1.28969955e+00 -1.13463032e+00 1.91337094e-01 1.52117693e+00
-3.92680883e-01 -6.91102922e-01 -4.85024899e-01 -4.30663317e-01
2.80442655e-01 1.24808383e+00 -6.97387218e-01 -3.63877505e-01
-7.39677966e-01 -1.15976918e+00 1.11328930e-01 6.90758646e-01
3.73406082e-01 -1.18053329e+00 -7.63015568e-01 7.72999167e-01
-3.23256671e-01 -4.56274062e-01 -7.78089821e-01 2.94767618e-01
-7.52298713e-01 -9.35730338e-01 -4.23167437e-01 -4.53654855e-01
3.14716846e-01 1.82302326e-01 1.79183900e+00 9.09784585e-02
-3.95143479e-01 3.90502602e-01 -5.97813845e-01 -6.71802104e-01
-1.16808975e+00 -1.95372865e-01 -1.26149701e-02 -3.33895832e-01
1.19148016e+00 -3.44018102e-01 -8.45390916e-01 7.27380738e-02
-1.17043042e+00 2.26299420e-01 4.03826982e-01 4.19963747e-01
7.67705739e-01 1.77229583e-01 4.53959823e-01 -9.13824975e-01
5.31526804e-01 -1.07277107e+00 -6.23865612e-02 2.15330467e-01
-6.69785380e-01 3.75163019e-01 7.31074393e-01 -8.48403990e-01
-8.05759966e-01 1.79163486e-01 -2.32334614e-01 2.50889897e-01
-5.38063288e-01 6.77172184e-01 -2.11329036e-03 8.01188231e-01
8.55330825e-01 1.95459187e-01 -9.87879746e-03 -6.00499630e-01
8.93706083e-01 1.54619440e-01 2.47054219e-01 -6.43216133e-01
5.82545459e-01 -1.24027736e-01 -2.00837865e-01 -7.27295220e-01
-8.98510993e-01 -2.92697549e-01 -3.23510557e-01 6.11670911e-02
1.17482567e+00 -8.60838294e-01 -5.37088454e-01 1.54628232e-01
-5.38865447e-01 -6.36798263e-01 -5.46141326e-01 6.13709152e-01
-4.96040881e-01 1.39939025e-01 -7.09690630e-01 -6.58096731e-01
-4.50565487e-01 -8.05183530e-01 9.37395513e-01 2.46629678e-02
-8.24658692e-01 -1.11946058e+00 3.01602274e-01 1.43822134e-01
6.57766223e-01 5.27498484e-01 1.15475047e+00 -5.88899195e-01
-1.82175457e-01 3.12750190e-01 1.42722145e-01 1.75452545e-01
6.83387816e-01 1.20102298e-02 -2.95430988e-01 4.71363775e-03
-6.29300475e-02 -3.82813871e-01 7.14383423e-01 6.55935049e-01
5.32330215e-01 -5.97927332e-01 -3.46460342e-02 7.29323328e-02
1.44122291e+00 2.15743184e-01 4.63952094e-01 -1.25642776e-01
6.74796700e-01 6.34594321e-01 2.73305297e-01 3.58663082e-01
9.09219146e-01 5.45625508e-01 -7.84597360e-03 -2.76339889e-01
-5.02945483e-01 -4.77366328e-01 6.28001809e-01 7.49024928e-01
-1.30426481e-01 -2.35331699e-01 -6.07459426e-01 4.01773006e-01
-1.55787683e+00 -1.37779272e+00 -3.59181494e-01 2.14902544e+00
1.33057344e+00 -4.31835979e-01 4.68528360e-01 3.61808017e-02
9.81851444e-02 -5.35355788e-03 -5.74302375e-01 -4.87257749e-01
1.42804354e-01 2.73474723e-01 2.69228816e-01 3.84260774e-01
-1.17939591e+00 8.28031301e-01 6.47092438e+00 1.94943160e-01
-5.84076524e-01 -5.26053570e-02 6.21965110e-01 -4.72882874e-02
-6.59414172e-01 -3.35335910e-01 -8.68593156e-01 5.26187837e-01
1.19997132e+00 3.01093429e-01 1.03626156e+00 6.26349449e-01
3.06402564e-01 -1.54837042e-01 -1.55554426e+00 6.14538133e-01
2.17984140e-01 -9.77448881e-01 -4.67449687e-02 -1.26265630e-01
6.14319563e-01 -1.25324979e-01 -2.26018623e-01 6.02834940e-01
1.24693024e+00 -1.00617468e+00 1.04178202e+00 4.30604190e-01
3.36227417e-01 -5.47788739e-01 2.45138004e-01 2.88161397e-01
-1.42285407e+00 -6.05137981e-02 -5.24073839e-01 -2.16176569e-01
2.83958048e-01 6.57018840e-01 -4.19526815e-01 3.86298329e-01
6.94033682e-01 8.44410717e-01 -4.22407687e-01 5.35311937e-01
-6.33723363e-02 7.40714133e-01 -6.93493783e-01 -2.51794964e-01
2.10689217e-01 -5.83527863e-01 1.11709826e-01 1.25600362e+00
4.81801569e-01 3.89105856e-01 5.66286504e-01 1.43568981e+00
3.13050300e-01 4.04014558e-01 -5.51508963e-01 -3.79579514e-01
-1.68724507e-01 1.06911778e+00 -7.49535680e-01 -3.04536909e-01
-5.83509684e-01 7.83296049e-01 2.12318916e-02 2.84974486e-01
-7.55427122e-01 5.47758698e-01 7.68467546e-01 2.57130504e-01
2.26815268e-01 -2.97249407e-01 8.46326053e-02 -1.20296359e+00
-4.24157917e-01 -1.42978954e+00 4.95395899e-01 -5.76812625e-01
-1.57432735e+00 4.09856439e-02 3.20439696e-01 -6.25013709e-01
-6.61887646e-01 -3.59655291e-01 -3.38113934e-01 7.82242239e-01
-1.07916224e+00 -1.27671158e+00 -5.91541193e-02 5.12605309e-01
7.76038408e-01 3.93237770e-01 1.19698036e+00 3.76685917e-01
-2.76170611e-01 2.24786714e-01 -8.26324672e-02 4.34841886e-02
3.78444374e-01 -1.40532160e+00 4.93633628e-01 6.09697759e-01
4.53600198e-01 8.36883545e-01 9.79415834e-01 -1.01044178e+00
-1.74135530e+00 -1.14941108e+00 1.04984677e+00 -6.38358116e-01
4.56577361e-01 -5.08704543e-01 -7.07006991e-01 7.89788544e-01
1.54117271e-01 -5.44386983e-01 1.14702594e+00 3.56960028e-01
-4.82573897e-01 9.40672755e-02 -1.22304225e+00 6.46997690e-01
8.56264055e-01 -8.11353624e-02 -5.60970604e-01 4.21927273e-01
7.11003304e-01 -5.27357161e-02 -1.03579152e+00 5.48349991e-02
6.52150512e-01 -5.92863441e-01 1.33273065e+00 -9.31219399e-01
5.56038558e-01 -3.45955968e-01 -3.34976465e-01 -1.62012792e+00
-7.13564277e-01 -3.26050669e-01 -2.24995807e-01 1.03960180e+00
3.97310376e-01 -2.54024714e-01 5.07617712e-01 1.08509862e+00
-7.30435625e-02 -1.81297466e-01 -8.78414288e-02 -4.23024863e-01
2.10801557e-01 -2.10240617e-01 1.06344497e+00 1.02299201e+00
-3.15618664e-02 2.18162715e-01 -5.32088697e-01 -2.66303688e-01
7.63947308e-01 3.56122702e-01 5.50676525e-01 -1.18518698e+00
-5.81473529e-01 -5.46173394e-01 2.06975475e-01 -8.61471415e-01
-3.65601510e-01 -1.07770908e+00 4.13543046e-01 -1.63218701e+00
3.74614805e-01 -4.57215041e-01 -2.19470620e-01 7.74248481e-01
-4.51210976e-01 1.55319437e-01 3.12711984e-01 -2.55424887e-01
-2.67730892e-01 1.15592197e-01 1.17763650e+00 -3.99379849e-01
-5.69627643e-01 -1.01145476e-01 -1.12393200e+00 4.96803671e-01
1.02843118e+00 -6.17722273e-01 -5.89844048e-01 -3.45400542e-01
7.95490623e-01 -2.67035753e-01 4.71800894e-01 -6.18260205e-01
-1.54098779e-01 -5.01383364e-01 4.37116146e-01 -2.92328745e-01
-1.34957135e-01 -7.50545740e-01 9.18901324e-01 6.27255380e-01
-5.83141565e-01 1.92824528e-01 2.45474175e-01 5.03532410e-01
2.86864609e-01 -9.25824642e-02 2.75875300e-01 -5.29543161e-01
-5.03518462e-01 1.96588844e-01 -6.33371055e-01 -7.15791732e-02
6.12280607e-01 7.70005733e-02 -1.03917658e-01 -1.58164293e-01
-9.79411304e-01 -1.61680242e-03 3.84469897e-01 6.83850944e-01
1.95993662e-01 -1.46395159e+00 -1.13122857e+00 3.54058355e-01
7.98142925e-02 -1.88223019e-01 3.18100423e-01 5.81248701e-01
-3.09407592e-01 2.23011583e-01 -9.86941457e-02 -2.32195988e-01
-7.96694100e-01 1.18090749e+00 1.23692423e-01 -5.33013344e-01
-8.27765405e-01 2.51213938e-01 2.91891098e-01 -5.51670253e-01
-1.87602356e-01 -1.19640338e+00 -1.53159037e-01 -1.72308199e-02
7.55560815e-01 2.73597509e-01 -1.99626565e-01 -3.39603901e-01
-9.85800028e-02 2.75689840e-01 1.14085250e-01 5.94862759e-01
1.51245546e+00 -6.50904179e-02 -1.19430020e-01 5.09689271e-01
6.50377333e-01 -1.65124878e-01 -1.07688379e+00 -1.04947977e-01
-2.09423855e-01 4.46671359e-02 3.78713980e-02 -1.25169849e+00
-9.44193780e-01 4.57405388e-01 4.68316287e-01 6.43984079e-01
1.07410014e+00 4.65161130e-02 9.28382158e-01 1.45741582e-01
3.45809579e-01 -8.76511693e-01 -2.76255757e-01 -3.22513878e-02
7.65358567e-01 -1.34406734e+00 2.11408198e-01 -2.60745496e-01
-5.11690021e-01 8.95829380e-01 1.73987418e-01 -1.38727933e-01
7.91504443e-01 2.72718430e-01 -7.88574368e-02 -4.06604856e-01
-5.08139908e-01 -9.61674079e-02 4.60964113e-01 5.90881884e-01
7.58746028e-01 7.79342055e-01 -3.66026700e-01 6.68186724e-01
-2.42981836e-01 2.11954400e-01 1.34835005e-01 6.46306872e-01
-4.36775357e-01 -1.33924627e+00 -4.28643554e-01 6.72544241e-01
-3.11950624e-01 -6.43625081e-01 9.80932172e-03 7.36941397e-01
4.63196814e-01 1.23962629e+00 5.59079275e-02 1.38511555e-02
2.16493145e-01 3.60610813e-01 6.24465883e-01 -7.82517731e-01
-1.08410966e+00 3.09871793e-01 1.59301713e-01 -5.27089119e-01
-7.80124724e-01 -1.07586813e+00 -1.15832162e+00 -7.29603767e-01
1.65487036e-01 -2.68050823e-02 4.88423139e-01 7.87706792e-01
3.72434296e-02 4.34969932e-01 1.74598664e-01 -7.36977398e-01
-8.41761291e-01 -1.06509852e+00 -5.08077562e-01 8.64727497e-01
-2.09524520e-02 -4.73803848e-01 -1.28722206e-01 5.27944505e-01] | [11.521666526794434, 4.531466484069824] |
7dfa3827-5938-4dd3-89b2-5a6c0d700900 | fine-grained-temporal-relation-extraction | 1902.0139 | null | https://arxiv.org/abs/1902.01390v2 | https://arxiv.org/pdf/1902.01390v2.pdf | Fine-Grained Temporal Relation Extraction | We present a novel semantic framework for modeling temporal relations and event durations that maps pairs of events to real-valued scales. We use this framework to construct the largest temporal relations dataset to date, covering the entirety of the Universal Dependencies English Web Treebank. We use this dataset to train models for jointly predicting fine-grained temporal relations and event durations. We report strong results on our data and show the efficacy of a transfer-learning approach for predicting categorical relations. | ['Benjamin Van Durme', 'Siddharth Vashishtha', 'Aaron Steven White'] | 2019-02-04 | fine-grained-temporal-relation-extraction-1 | https://aclanthology.org/P19-1280 | https://aclanthology.org/P19-1280.pdf | acl-2019-7 | ['temporal-relation-extraction'] | ['natural-language-processing'] | [-2.71402299e-01 4.34392333e-01 -7.27705121e-01 -9.12190259e-01
-6.95153952e-01 -8.52088690e-01 9.01400864e-01 4.54536289e-01
-4.66124535e-01 9.61421907e-01 7.38404751e-01 -2.45216355e-01
-2.47662157e-01 -9.13215578e-01 -4.64628011e-01 -3.78327863e-03
-8.01604927e-01 6.50939584e-01 5.55714428e-01 -4.43790317e-01
-1.22172870e-01 3.48511070e-01 -1.11133397e+00 7.31486619e-01
3.32086772e-01 1.07486880e+00 -3.33228350e-01 4.98301774e-01
1.73147887e-01 1.41814721e+00 -4.58956093e-01 -5.97598851e-01
-2.33115822e-01 -2.42267966e-01 -1.27705193e+00 -5.96022487e-01
-3.43548834e-01 -9.04123783e-02 -9.30916071e-01 3.57064992e-01
5.16139530e-02 4.60276872e-01 7.30439544e-01 -1.48174083e+00
-7.90261626e-01 1.08094823e+00 6.72990382e-02 8.28078270e-01
7.47680426e-01 -4.50097740e-01 1.42249084e+00 -2.27171406e-01
9.16201651e-01 1.31926847e+00 7.64897883e-01 2.64444500e-01
-1.18952787e+00 -3.79408598e-01 2.04040200e-01 6.45471275e-01
-1.21507967e+00 -3.76356840e-01 4.65186357e-01 -2.40976259e-01
1.92108023e+00 7.92256650e-03 4.64384705e-01 1.28211737e+00
3.63117218e-01 4.44064856e-01 1.00323641e+00 -4.53608990e-01
1.35517508e-01 -5.94381213e-01 3.30172658e-01 2.48151422e-01
-4.04924601e-01 3.39216292e-01 -8.16916466e-01 -2.87341569e-02
7.42573798e-01 -4.06041682e-01 4.18285012e-01 3.56021792e-01
-1.36809981e+00 6.31648421e-01 4.38201815e-01 6.41602457e-01
-1.67116269e-01 4.85130072e-01 7.21516550e-01 4.88386303e-01
8.32009256e-01 3.04784983e-01 -9.95470047e-01 -4.82832104e-01
-3.63680333e-01 3.09429049e-01 9.67011690e-01 9.01202083e-01
2.04094097e-01 -5.12793839e-01 -2.21859500e-01 7.78055370e-01
-8.43218341e-02 -2.00800486e-02 3.59911054e-01 -1.10121703e+00
5.77147543e-01 3.96913022e-01 2.10555866e-01 -7.24824905e-01
-9.51411605e-01 3.52516890e-01 -1.16110690e-01 -5.39495289e-01
4.06856179e-01 -1.05973177e-01 -5.99686384e-01 2.17575049e+00
2.54700035e-01 5.08115530e-01 2.41404697e-01 4.60191339e-01
8.02950263e-01 6.58959329e-01 8.22854757e-01 -4.63227153e-01
1.42758024e+00 -3.97779912e-01 -1.02998209e+00 -1.19257785e-01
9.89235818e-01 -1.71057925e-01 7.27230489e-01 5.90437965e-04
-1.14720368e+00 -2.63367742e-01 -7.94068694e-01 -4.61060584e-01
-7.61014700e-01 -3.74393702e-01 1.27850294e+00 -3.58377099e-02
-8.88575256e-01 7.95924962e-01 -1.35146129e+00 -6.45114005e-01
1.33553430e-01 1.28417328e-01 -5.07627964e-01 5.19631684e-01
-1.93486059e+00 1.46334875e+00 9.42029119e-01 -3.08420174e-02
-4.79299635e-01 -7.17453480e-01 -1.04356456e+00 -5.01179658e-02
1.22660168e-01 -1.75065011e-01 1.52288139e+00 -1.52767718e-01
-9.24364746e-01 1.16268814e+00 -2.43413806e-01 -9.06834543e-01
1.42174698e-02 -1.42759830e-01 -1.23064804e+00 1.11361481e-01
2.55754024e-01 5.79120338e-01 -1.34656966e-01 -4.02136832e-01
-7.84009755e-01 -1.77198350e-01 2.88703263e-01 -7.60513321e-02
-2.68663079e-01 6.61504149e-01 -9.42413136e-02 -9.25332069e-01
4.86864038e-02 -8.04987669e-01 -4.04338986e-02 -4.48059648e-01
-1.53919756e-01 -1.01221204e+00 4.76737320e-01 -7.34564662e-01
1.54102659e+00 -2.04729199e+00 1.03271626e-01 -2.85480201e-01
-3.08451772e-01 -4.90705788e-01 -1.71168502e-02 6.53516054e-01
-4.86578047e-01 1.06199436e-01 -7.89491087e-02 -2.28973985e-01
2.26182938e-01 6.98548496e-01 -6.14129126e-01 2.81759977e-01
4.32927489e-01 1.41964877e+00 -1.13916051e+00 -8.46237242e-01
3.93805392e-02 1.16399989e-01 -1.38957217e-01 3.43026012e-01
-5.24849474e-01 4.06864166e-01 -3.47465277e-01 4.86910999e-01
-4.74976301e-02 -1.04921311e-01 5.17362416e-01 -1.71527863e-01
-7.39855990e-02 1.18974900e+00 -4.92331266e-01 1.94087493e+00
-4.27706271e-01 6.19380176e-01 -8.29317749e-01 -9.86228228e-01
8.04687679e-01 9.39323962e-01 7.22438931e-01 -1.03237045e+00
-5.16607016e-02 4.28246967e-02 -1.98601410e-01 -6.52744114e-01
2.20354304e-01 -4.19506609e-01 -9.42604780e-01 4.60069031e-01
3.53537470e-01 -2.63584614e-01 5.13302326e-01 1.53647497e-01
1.24626100e+00 5.06954789e-01 8.55434537e-01 -2.41793543e-01
-8.15246440e-03 1.35075971e-02 7.41851747e-01 2.31831178e-01
-3.41536582e-01 -4.60122041e-02 7.99884021e-01 -8.49249423e-01
-9.31371748e-01 -1.46601677e+00 -4.66035336e-01 1.37388456e+00
-2.08142027e-01 -6.53471649e-01 -1.24731576e-02 -9.22769785e-01
-9.74087790e-02 9.52100098e-01 -9.51450527e-01 6.01334423e-02
-9.14478004e-01 -5.02911150e-01 9.15754795e-01 1.22376895e+00
6.35821074e-02 -1.55617166e+00 -5.23503900e-01 4.10850644e-01
-6.78336024e-01 -1.55231094e+00 -1.50605127e-01 6.15926862e-01
-7.38664269e-01 -9.52550650e-01 2.71605879e-01 -8.96898627e-01
-3.18925589e-01 -8.05019021e-01 1.58227909e+00 -5.50303221e-01
2.94594839e-02 -5.68175549e-03 -5.14841735e-01 -2.92045951e-01
-2.69584626e-01 -7.39322836e-03 1.35231614e-01 -7.01655447e-01
6.83042049e-01 -9.33226764e-01 -8.23091790e-02 2.33854249e-01
-5.62008679e-01 -2.32411370e-01 -2.96843082e-01 4.09830630e-01
2.49956191e-01 9.94966775e-02 9.58927691e-01 -7.59826958e-01
4.49073941e-01 -8.72384250e-01 -4.20406133e-01 4.04393196e-01
-1.75535619e-01 7.58448467e-02 3.09305489e-01 -4.69950140e-01
-1.40824234e+00 -2.97642887e-01 -5.69002219e-02 4.08679098e-02
-4.28759933e-01 8.66658926e-01 1.53959349e-01 6.42947257e-01
7.21956193e-01 -3.48425776e-01 -8.70418012e-01 -1.91304341e-01
8.86640072e-01 2.54197419e-01 1.08077395e+00 -8.75224411e-01
2.28798881e-01 5.86874068e-01 1.79941446e-01 -1.55648187e-01
-1.21055269e+00 -2.07433224e-01 -1.00790739e+00 1.21148162e-01
1.24110186e+00 -1.03331256e+00 -7.65594721e-01 9.40397382e-02
-1.31382227e+00 -6.92044139e-01 -4.83603507e-01 6.11278415e-01
-9.07355368e-01 -2.01255232e-01 -1.28960013e+00 -5.31759024e-01
2.81540871e-01 -1.67306140e-01 9.90009010e-01 -8.44810754e-02
-9.67981279e-01 -1.50850511e+00 5.04481554e-01 -2.13218972e-01
-1.52577478e-02 6.75071895e-01 1.05317426e+00 -7.68900156e-01
-5.57149239e-02 7.32064471e-02 -1.22583456e-01 -4.17498738e-01
2.77099848e-01 -1.30249441e-01 -7.12829173e-01 4.20898110e-01
-2.38311201e-01 -8.11520040e-01 7.14818239e-01 3.52803200e-01
8.61046731e-01 -5.86544648e-02 -6.54783726e-01 2.98666298e-01
1.04706430e+00 6.47017002e-01 5.88760197e-01 3.34867358e-01
3.56728166e-01 7.73163021e-01 1.07628238e+00 5.59248745e-01
9.73441124e-01 8.45656395e-01 -1.30759075e-01 1.85762867e-01
-5.57261109e-02 -4.26329970e-01 2.62181833e-02 5.74675679e-01
-2.71142751e-01 -2.19188705e-01 -1.02083790e+00 1.03951252e+00
-2.03740191e+00 -1.24332547e+00 1.01211444e-02 1.43981612e+00
1.45037293e+00 2.73723602e-01 1.33461699e-01 -2.71042548e-02
4.53140140e-01 3.13924521e-01 -2.36060575e-01 -6.87233686e-01
-1.87198415e-01 6.10658884e-01 1.69865564e-01 4.93563920e-01
-1.60912979e+00 1.45649350e+00 8.07540512e+00 4.80041653e-01
-5.99797308e-01 2.42015719e-01 4.36587244e-01 -7.78195411e-02
-1.84408784e-01 1.36581331e-01 -5.23015380e-01 1.57437950e-01
1.68555284e+00 -3.39003950e-01 3.26633304e-01 3.22672397e-01
2.46717840e-01 4.70568500e-02 -1.62025177e+00 6.73484445e-01
-3.71385664e-01 -1.24311149e+00 -5.52992463e-01 -4.10544753e-01
5.46746254e-01 -5.49494885e-02 -4.42185163e-01 4.22475666e-01
1.01028824e+00 -1.09498096e+00 9.02652383e-01 5.51780641e-01
8.69885147e-01 -5.47075927e-01 2.77587682e-01 -1.97345376e-01
-1.70895839e+00 -1.58186182e-01 8.37235674e-02 -5.88486254e-01
8.46233249e-01 4.73719150e-01 -6.44904792e-01 5.74212432e-01
8.81691933e-01 1.36582124e+00 -3.77895087e-01 3.58247131e-01
-6.09972358e-01 6.44237101e-01 -3.75313133e-01 3.16472709e-01
-1.58319369e-01 1.46181792e-01 1.05662584e-01 1.29193592e+00
9.37022865e-02 6.35827005e-01 -2.90122125e-02 5.53102016e-01
-7.79460371e-02 -3.11971009e-01 -7.50460684e-01 -3.39431047e-01
9.15143609e-01 7.72091985e-01 -7.03193486e-01 -4.44496393e-01
-5.82828701e-01 7.17549026e-01 8.83805096e-01 2.76816458e-01
-1.29260838e+00 -3.32074761e-02 4.96035695e-01 -3.44499677e-01
1.66570961e-01 -4.65390265e-01 -5.15793741e-01 -1.13848126e+00
-1.64233092e-02 -2.22053930e-01 1.40864837e+00 -9.18098390e-01
-1.64044833e+00 5.82631350e-01 6.38421535e-01 -8.20902765e-01
-7.79131413e-01 -2.25803539e-01 -4.50029433e-01 5.68495512e-01
-1.07099640e+00 -1.29077244e+00 2.98318446e-01 7.85872221e-01
2.43986011e-01 3.40379268e-01 1.16491008e+00 2.39726335e-01
-1.37734964e-01 2.37829998e-01 -7.10164905e-01 3.37882429e-01
8.91835272e-01 -1.42690146e+00 8.05591941e-01 6.33079767e-01
3.54375660e-01 5.39290607e-01 6.14577949e-01 -8.57540309e-01
-7.07868218e-01 -9.28113282e-01 1.78544712e+00 -9.30265903e-01
1.38154793e+00 -4.48318750e-01 -7.46247292e-01 1.69045222e+00
2.41730377e-01 1.95369706e-01 6.07391059e-01 7.61210799e-01
-8.22686136e-01 7.80371344e-03 -8.34606826e-01 4.71667886e-01
1.60034215e+00 -1.07832062e+00 -1.43426549e+00 2.32901827e-01
1.23510849e+00 -4.69853938e-01 -1.59775651e+00 7.10154176e-01
4.73960876e-01 -4.76272106e-01 1.05817902e+00 -1.06010580e+00
6.49722517e-01 -5.19333780e-02 -3.41709346e-01 -1.19216299e+00
-2.85225660e-01 -3.10657531e-01 -4.69058961e-01 1.39640486e+00
7.73757994e-01 -6.11276686e-01 2.85899520e-01 8.94698322e-01
-1.61795989e-01 -4.42837149e-01 -1.21394980e+00 -8.53617847e-01
2.79729694e-01 -8.46016228e-01 5.90006351e-01 1.40458727e+00
9.82522666e-01 3.59017879e-01 -2.03107655e-01 4.56870943e-02
1.16658621e-01 4.04495031e-01 -1.75956145e-01 -1.11867023e+00
-2.41085291e-01 -1.62124738e-01 -4.56345290e-01 -4.69670534e-01
7.05844045e-01 -9.06256020e-01 -1.15626775e-01 -1.53727436e+00
-8.43687952e-02 -5.91991186e-01 -4.56198931e-01 1.00479245e+00
-4.93053049e-02 1.40554816e-01 -2.54775047e-01 -9.40886438e-02
-7.20121622e-01 2.41467640e-01 6.69173002e-01 1.42924607e-01
-1.72317788e-01 -3.80536199e-01 -2.20667303e-01 4.39026594e-01
6.53707504e-01 -5.28455138e-01 -6.16102993e-01 -3.72972041e-01
4.22218889e-01 7.67821789e-01 2.51167387e-01 -6.00333512e-01
1.77417308e-01 -5.56428075e-01 2.30270505e-01 -5.58211207e-01
3.79811049e-01 -5.17871499e-01 2.39540219e-01 1.02619603e-01
-9.51723337e-01 2.17789143e-01 4.77110535e-01 4.25129205e-01
-5.16181290e-01 5.32350600e-01 2.90050119e-01 1.40280172e-01
-1.12236392e+00 1.57590717e-01 -2.37343431e-01 4.62804049e-01
1.16622519e+00 4.66066211e-01 -6.18888795e-01 -2.38040194e-01
-1.53624737e+00 1.32084057e-01 -9.26467329e-02 7.59373367e-01
1.56939358e-01 -1.59769011e+00 -5.90886056e-01 -4.86744881e-01
2.59373635e-01 -4.79936391e-01 7.44476095e-02 6.94688499e-01
-7.78362304e-02 7.46148825e-01 -2.62284458e-01 -9.49008614e-02
-7.80502498e-01 9.95840132e-01 1.51431695e-01 -5.57169974e-01
-6.00475371e-01 8.87127221e-01 -1.34161875e-01 -2.92565376e-01
-1.73662510e-02 -7.40718901e-01 -3.02194208e-01 2.71726519e-01
3.59784186e-01 1.65648848e-01 -4.73031513e-02 -4.63606596e-01
-8.36021483e-01 1.32909477e-01 4.42152977e-01 -6.78064704e-01
1.43660128e+00 -1.13466315e-01 -4.25883889e-01 1.03712738e+00
1.18406534e+00 -3.79201949e-01 -1.04259479e+00 -3.06563616e-01
7.48415709e-01 2.67677177e-02 -5.15619993e-01 -8.42401147e-01
-4.39357370e-01 3.69342774e-01 -1.06234185e-01 6.15820050e-01
1.15954602e+00 7.11875856e-01 7.26147771e-01 1.29266515e-01
6.57287776e-01 -1.01459515e+00 -2.82761548e-02 8.72209251e-01
7.84674227e-01 -9.59462523e-01 -2.99574912e-01 -5.78325033e-01
-7.31705368e-01 9.65192437e-01 4.02080297e-01 -1.71191484e-01
9.43337023e-01 7.55254626e-01 -1.91525463e-02 -2.86706537e-01
-1.38597834e+00 -4.42724854e-01 3.49688828e-01 5.69634974e-01
9.30981934e-01 4.14219558e-01 -6.38729155e-01 7.37146556e-01
-4.37465459e-01 2.42637008e-01 2.62204688e-02 9.80783582e-01
5.51981181e-02 -1.24893177e+00 1.18893594e-01 1.00271024e-01
-5.26676595e-01 -2.21962601e-01 -2.15647265e-01 7.62989938e-01
8.74316543e-02 1.34866917e+00 5.05281389e-01 -2.52946109e-01
4.55846488e-01 4.61612493e-01 7.61661172e-01 -5.15664160e-01
-4.01436239e-01 -1.86702043e-01 9.02620733e-01 -1.02032793e+00
-8.05673480e-01 -9.03573275e-01 -1.68558514e+00 -1.38674080e-01
3.18213046e-01 8.50784481e-02 9.49501619e-02 1.28012359e+00
-1.61372535e-02 5.99532187e-01 4.11510587e-01 -4.13498461e-01
6.95093200e-02 -1.00026000e+00 -5.23444951e-01 6.94358766e-01
5.66153638e-02 -7.89232910e-01 -5.91732934e-02 4.62255925e-01] | [9.124157905578613, 9.17470932006836] |
47bdcc8e-1c88-42e7-82ba-db97a2e059f8 | huruf-an-application-for-arabic-handwritten | 2212.0861 | null | https://arxiv.org/abs/2212.08610v2 | https://arxiv.org/pdf/2212.08610v2.pdf | Huruf: An Application for Arabic Handwritten Character Recognition Using Deep Learning | Handwriting Recognition has been a field of great interest in the Artificial Intelligence domain. Due to its broad use cases in real life, research has been conducted widely on it. Prominent work has been done in this field focusing mainly on Latin characters. However, the domain of Arabic handwritten character recognition is still relatively unexplored. The inherent cursive nature of the Arabic characters and variations in writing styles across individuals makes the task even more challenging. We identified some probable reasons behind this and proposed a lightweight Convolutional Neural Network-based architecture for recognizing Arabic characters and digits. The proposed pipeline consists of a total of 18 layers containing four layers each for convolution, pooling, batch normalization, dropout, and finally one Global average pooling and a Dense layer. Furthermore, we thoroughly investigated the different choices of hyperparameters such as the choice of the optimizer, kernel initializer, activation function, etc. Evaluating the proposed architecture on the publicly available 'Arabic Handwritten Character Dataset (AHCD)' and 'Modified Arabic handwritten digits Database (MadBase)' datasets, the proposed model respectively achieved an accuracy of 96.93% and 99.35% which is comparable to the state-of-the-art and makes it a suitable solution for real-life end-level applications. | ['Md. Hasanul Kabir', 'Tasnim Ahmed', 'Sabbir Ahmed', 'Chowdhury Mohammad Abdullah', 'Fairuz Shaiara', 'Minhaz Kamal'] | 2022-12-16 | null | null | null | null | ['handwriting-recognition'] | ['computer-vision'] | [ 6.77236468e-02 -4.00537521e-01 2.91988969e-01 -6.21406972e-01
-2.44326312e-02 -5.98860681e-01 5.87885201e-01 8.31288770e-02
-7.77160406e-01 7.03132331e-01 -1.96064860e-01 -1.26089618e-01
-1.22382417e-01 -5.73374212e-01 -3.61634940e-01 -9.66965139e-01
1.01550318e-01 4.37718540e-01 3.49124312e-01 -2.22145066e-01
8.84295642e-01 1.08099806e+00 -1.29463935e+00 3.84779394e-01
7.45149732e-01 8.49044979e-01 -4.16206010e-03 9.02058423e-01
9.24543440e-02 9.87348139e-01 -8.15351069e-01 -6.87736094e-01
2.18898714e-01 -3.42613667e-01 -7.06942260e-01 4.53894585e-01
3.32256734e-01 -4.18916821e-01 -4.90937591e-01 8.10253799e-01
7.76575565e-01 3.20154339e-01 6.16064787e-01 -8.95975530e-01
-9.51563835e-01 5.04349768e-01 -6.65228009e-01 3.03525925e-01
-3.02170098e-01 1.00419469e-01 4.47715044e-01 -6.63381219e-01
5.02061903e-01 1.01997018e+00 6.35614634e-01 3.98072660e-01
-6.14100933e-01 -3.54284376e-01 -3.90624702e-02 3.35373104e-01
-1.26243293e+00 -2.01514900e-01 5.43445408e-01 -3.03301930e-01
1.20001936e+00 1.16156437e-01 9.99146402e-02 8.72736156e-01
2.45221451e-01 7.21285105e-01 1.30004644e+00 -6.44376874e-01
-4.74410132e-02 4.41705495e-01 5.86434066e-01 6.60315931e-01
2.74545133e-01 -4.47764963e-01 -3.00645888e-01 3.19332272e-01
7.20180869e-01 -2.04182550e-01 1.34801120e-01 3.33716393e-01
-1.12939179e+00 5.49166739e-01 4.08520937e-01 3.96294624e-01
-4.21949744e-01 -2.73724981e-02 5.20414472e-01 1.66835383e-01
-7.49459639e-02 1.86339319e-01 -3.16065907e-01 -2.41592497e-01
-7.92674184e-01 3.52036268e-01 8.42343032e-01 8.52882683e-01
1.90112785e-01 2.73900360e-01 -4.72044535e-02 1.08777094e+00
2.78237969e-01 1.49812162e-01 7.17941582e-01 -1.23919502e-01
6.67570710e-01 6.91239595e-01 1.45286828e-01 -1.12359846e+00
-3.80776018e-01 -3.33056957e-01 -1.02655220e+00 4.41805303e-01
6.90644264e-01 -4.49797094e-01 -1.18892455e+00 9.56225216e-01
-1.34701177e-01 -2.73490459e-01 1.55981898e-01 1.23492825e+00
6.79788411e-01 7.00251579e-01 -2.95669064e-02 3.50356400e-01
1.38492405e+00 -1.10809016e+00 -6.14579141e-01 -1.38683483e-01
-5.00953905e-02 -1.18821335e+00 8.38671684e-01 6.67074025e-01
-9.31293011e-01 -5.13150513e-01 -1.50106871e+00 -7.25385323e-02
-7.34435320e-01 8.74386072e-01 7.20879257e-01 9.86819685e-01
-7.02279925e-01 6.05637193e-01 -9.28520799e-01 -6.06402040e-01
5.17527401e-01 6.68186784e-01 -5.48968315e-01 3.04704183e-03
-5.49751461e-01 1.12653828e+00 7.30953336e-01 7.69424379e-01
-6.05186164e-01 1.22349717e-01 -1.35993525e-01 -1.34099647e-01
-8.69182721e-02 2.60046661e-01 9.15791214e-01 -1.18598890e+00
-1.80446625e+00 8.70891750e-01 3.19392771e-01 -3.69257659e-01
7.01121509e-01 -7.95968175e-02 -7.18757153e-01 -8.81033838e-02
-5.41575134e-01 4.46897030e-01 6.54282033e-01 -6.53578341e-01
-5.39614558e-01 -5.23417473e-01 -1.90721363e-01 2.99312919e-01
-3.95868570e-01 5.31369030e-01 -3.91018212e-01 -8.58875394e-01
1.59757927e-01 -7.31393516e-01 4.24440354e-02 -1.94421157e-01
-4.81311023e-01 1.50139658e-02 8.63552928e-01 -1.07192171e+00
1.12928224e+00 -2.09544063e+00 5.84294926e-03 2.83108771e-01
-3.93934131e-01 5.95529616e-01 7.53141269e-02 4.90729421e-01
1.19917132e-01 -2.23677635e-01 -2.99130976e-01 -7.63891786e-02
1.50796147e-02 -7.74967810e-03 -7.99141303e-02 7.27832019e-01
6.69887006e-01 3.76016617e-01 -2.45885164e-01 -1.93714932e-01
1.89111471e-01 7.46565938e-01 -1.21994711e-01 2.48651549e-01
1.06270926e-03 1.06300160e-01 -3.60565603e-01 1.12179053e+00
8.89258146e-01 1.48414463e-01 4.56035174e-02 -1.60696089e-01
-5.13639808e-01 -2.78809488e-01 -1.57595074e+00 1.13226497e+00
1.88274421e-02 1.01313448e+00 -1.50107592e-01 -9.32483912e-01
1.44738555e+00 1.93915963e-01 -2.59972483e-01 -4.02337998e-01
4.66654360e-01 5.24607480e-01 4.26348358e-01 -7.76607990e-01
5.93146682e-01 1.93773076e-01 3.38573724e-01 3.01383913e-01
2.40709379e-01 4.60718244e-01 3.76689285e-01 -4.21713263e-01
7.48039663e-01 3.24120075e-01 4.05208394e-02 -3.15762848e-01
9.30542111e-01 8.66280645e-02 1.97259203e-01 5.76961994e-01
-2.77983606e-01 8.06841671e-01 5.13716817e-01 -7.15501368e-01
-1.23454857e+00 -6.24533892e-01 -3.45700026e-01 9.01575029e-01
-1.48222506e-01 3.34092617e-01 -8.25024962e-01 -2.71859795e-01
-1.62078232e-01 2.28669643e-01 -5.97178876e-01 4.22842294e-01
-8.94515336e-01 -1.48095727e+00 1.20760310e+00 5.73486269e-01
1.34131277e+00 -1.53263831e+00 -8.84764433e-01 3.94314528e-01
4.19042528e-01 -9.47853804e-01 -3.79279703e-02 3.60575199e-01
-8.68903458e-01 -9.31115389e-01 -1.09415877e+00 -1.14055729e+00
8.81945431e-01 -4.43940371e-01 7.26522148e-01 -1.95494443e-02
-5.71582973e-01 -3.47176522e-01 -4.92697567e-01 -5.51941454e-01
-1.93562210e-01 4.03177232e-01 -2.47518420e-01 3.47592711e-01
5.45550764e-01 -1.51366174e-01 -3.62386614e-01 1.28577292e-01
-9.76205289e-01 -2.78687596e-01 9.62477863e-01 8.21078897e-01
-5.85841052e-02 9.12777036e-02 5.76164067e-01 -9.21225011e-01
9.65300024e-01 -2.77774990e-01 -8.00697148e-01 4.11153764e-01
-1.85254797e-01 1.37888128e-03 7.49648988e-01 -4.56077635e-01
-1.22872066e+00 1.29412021e-02 -3.06696087e-01 4.77894872e-01
-5.66351056e-01 4.52865332e-01 -1.34336591e-01 -2.97948152e-01
6.28033876e-01 1.97128192e-01 -1.24649089e-02 -5.99471688e-01
-2.81973273e-01 1.22691309e+00 6.20717824e-01 -5.02411962e-01
3.09175014e-01 8.98925662e-02 -7.67529057e-03 -1.04063046e+00
-2.29860786e-02 -1.01982072e-01 -8.63239229e-01 -6.97235242e-02
1.00251675e+00 -5.47108710e-01 -8.85850966e-01 1.41504860e+00
-1.17375851e+00 -5.72219603e-02 5.77744722e-01 3.15865517e-01
7.97294173e-03 2.77313292e-01 -7.65828192e-01 -8.68420482e-01
-4.33638632e-01 -1.35925591e+00 2.57292330e-01 7.00102210e-01
1.13619648e-01 -8.58243585e-01 -2.22272053e-01 8.16483423e-02
6.30591571e-01 4.40058678e-01 8.94454658e-01 -8.25580299e-01
-5.00265419e-01 -5.77616990e-01 -5.84637344e-01 6.74640954e-01
1.10681094e-01 3.51168692e-01 -1.03401434e+00 -2.33493611e-01
-4.29356307e-01 -4.81844842e-01 6.21390760e-01 6.10733926e-02
9.46514368e-01 -1.37536064e-01 2.34955072e-01 4.37116086e-01
1.67039680e+00 6.74898148e-01 9.62962449e-01 8.02612543e-01
5.84827185e-01 3.70162427e-01 2.62625515e-01 7.00705647e-01
-1.24731705e-01 4.17357624e-01 1.83866367e-01 3.34820412e-02
1.36676654e-01 4.01623309e-01 2.45841771e-01 6.51788175e-01
-5.83914459e-01 -3.87772977e-01 -1.01976490e+00 3.84982616e-01
-1.55802345e+00 -7.37216115e-01 -2.07580060e-01 1.98662353e+00
7.18257368e-01 6.72068894e-02 9.96656045e-02 5.32977879e-01
6.75684452e-01 -5.89885116e-02 -3.22398394e-01 -9.36754763e-01
-5.86038888e-01 4.07179326e-01 6.86459064e-01 2.28646591e-01
-1.49612045e+00 9.43953037e-01 5.20899057e+00 4.22091484e-01
-1.42537224e+00 -4.61586744e-01 7.42636621e-01 4.27924663e-01
6.02493227e-01 -3.29598159e-01 -1.06891060e+00 5.46612918e-01
7.59920418e-01 4.89324212e-01 4.65351254e-01 4.75864291e-01
1.64177455e-02 -1.31047830e-01 -6.32557690e-01 7.68176734e-01
3.07440490e-01 -1.15565157e+00 1.61376029e-01 -3.35960716e-01
6.98097646e-01 -3.17585468e-01 5.56370243e-02 -9.01781842e-02
-1.09492645e-01 -1.46738660e+00 4.62937534e-01 6.11673415e-01
4.10049140e-01 -1.08263218e+00 1.19694555e+00 1.32309109e-01
-5.87831378e-01 -2.84034848e-01 -5.93680680e-01 -1.59668565e-01
-3.70545179e-01 9.88189429e-02 -8.41884255e-01 2.24830091e-01
7.27061570e-01 3.81440341e-01 -8.89828622e-01 1.22437000e+00
3.44630629e-02 5.93722343e-01 -2.56646782e-01 -4.93699253e-01
7.23052680e-01 -2.74117053e-01 1.95361860e-03 1.66517913e+00
2.23491460e-01 1.29046485e-01 -4.34505135e-01 6.01377964e-01
-4.10111994e-03 3.03668320e-01 7.57065555e-03 -4.87384260e-01
1.65773720e-01 1.22794139e+00 -1.18589342e+00 -6.24876358e-02
-1.80588275e-01 1.19444156e+00 3.18658769e-01 2.51157731e-01
-6.70631826e-01 -1.15618157e+00 2.06347898e-01 -2.85861462e-01
3.53001982e-01 -2.83861578e-01 -4.43826199e-01 -8.81170034e-01
2.89764315e-01 -1.25165796e+00 2.19979405e-01 -4.07900095e-01
-1.07476819e+00 1.00341666e+00 -3.30808103e-01 -9.02892113e-01
5.00162840e-02 -1.41100562e+00 -6.07989192e-01 1.33912253e+00
-1.07401025e+00 -1.27456868e+00 -6.41609132e-01 3.18310827e-01
6.44228697e-01 -8.59631896e-01 7.87776530e-01 4.72997665e-01
-1.00402820e+00 7.98624456e-01 3.34297508e-01 6.35169625e-01
7.03541875e-01 -1.12732124e+00 2.22095445e-01 1.16777265e+00
-3.26186836e-01 6.58665776e-01 6.46648228e-01 -4.00103390e-01
-1.53849185e+00 -7.21820414e-01 7.59754956e-01 -4.52654622e-02
4.16959167e-01 -1.30167738e-01 -9.43597436e-01 5.73999107e-01
5.06366193e-01 -2.18624666e-01 4.13615584e-01 -4.52090561e-01
3.58758587e-03 -1.29952759e-01 -1.33120692e+00 5.49143136e-01
1.00066721e-01 5.52541837e-02 -3.42313170e-01 1.65069610e-01
-3.08998913e-01 -5.48807502e-01 -6.25015676e-01 1.29236266e-01
7.65038013e-01 -1.01140976e+00 6.96264327e-01 -7.30398059e-01
5.92255890e-01 -4.16579276e-01 -7.68284798e-02 -8.51426423e-01
-2.30563268e-01 -2.82324880e-01 1.66518569e-01 1.47726870e+00
3.77199948e-01 -4.26198959e-01 8.21974874e-01 7.41173804e-01
-8.65287036e-02 -9.02621210e-01 -5.18239558e-01 -3.08642209e-01
9.58213732e-02 2.16780275e-01 4.59541708e-01 8.78601789e-01
-4.25651938e-01 -1.04352660e-01 -4.39707547e-01 3.40363503e-01
3.92233580e-01 -2.53361493e-01 4.74697322e-01 -8.76073241e-01
-2.22991750e-01 -6.04553699e-01 -7.59656012e-01 -5.01237512e-01
-2.71989584e-01 -5.44356644e-01 -5.46270870e-02 -1.24998224e+00
-1.14245564e-01 -4.51440394e-01 -1.29206046e-01 4.40171003e-01
-1.07964098e-01 5.32230675e-01 1.88151523e-01 -3.83256972e-02
-1.36673413e-02 -2.83208322e-02 8.58663261e-01 -6.44625798e-02
-2.35608727e-01 8.41388553e-02 -3.39842141e-01 4.99286026e-01
1.10220313e+00 6.59463331e-02 -3.19690369e-02 -8.46094370e-01
-6.45731539e-02 -4.91516083e-01 1.44022927e-01 -1.06050539e+00
4.74325746e-01 -3.47297899e-02 9.50772107e-01 -6.34194136e-01
9.90930200e-02 -6.63004756e-01 -7.41909966e-02 3.14340413e-01
-2.94471860e-01 4.23040301e-01 4.64851111e-01 -9.24590752e-02
-3.29397351e-01 -7.77322233e-01 1.04374444e+00 -2.24082783e-01
-9.66451585e-01 -1.40199259e-01 -5.55842221e-01 -4.21295881e-01
9.81766045e-01 -5.59634089e-01 -3.65108848e-01 3.06339055e-01
-4.35657829e-01 -2.03198463e-01 1.42746046e-02 3.72461647e-01
6.33461177e-01 -9.57905769e-01 -1.06029546e+00 2.58576602e-01
-1.69635698e-01 -2.20503658e-01 2.37951905e-01 6.26594961e-01
-1.67744064e+00 3.42292368e-01 -1.02893937e+00 -1.49564222e-01
-1.37930071e+00 -2.86825895e-02 3.91788751e-01 5.03710881e-02
-4.30001199e-01 9.05694127e-01 -9.18755710e-01 -3.12880367e-01
6.32946074e-01 -1.03958167e-01 -5.69100261e-01 1.61961820e-02
6.00264966e-01 8.61043036e-01 3.77579957e-01 -6.39084876e-01
-3.86109471e-01 4.33771551e-01 -3.26505423e-01 -4.81690764e-02
1.45293927e+00 4.81661826e-01 -4.26576257e-01 3.09129834e-01
8.96173835e-01 -2.56158262e-01 -1.03936779e+00 2.71543741e-01
2.83337057e-01 -4.37589526e-01 -1.92993030e-01 -1.25302029e+00
-9.04502094e-01 9.60212469e-01 8.13351154e-01 -1.77290931e-01
1.13691139e+00 -9.33028996e-01 1.76092893e-01 9.29979742e-01
-6.57757148e-02 -1.39501119e+00 -2.44519278e-01 8.15004230e-01
1.00371742e+00 -1.25926685e+00 1.24641329e-01 -1.89686492e-02
-7.76756704e-01 1.78813887e+00 6.98269844e-01 -5.55132806e-01
4.49577361e-01 3.92894089e-01 4.53107089e-01 2.57957906e-01
-2.46276140e-01 3.10592085e-01 2.24097133e-01 5.25844216e-01
9.22895372e-01 -8.42505768e-02 -4.74698186e-01 7.08588839e-01
-3.99080552e-02 1.71738058e-01 7.56895185e-01 1.20319152e+00
-2.87903666e-01 -1.11814535e+00 -5.63178539e-01 4.85280663e-01
-8.54086936e-01 9.07281190e-02 -5.46646535e-01 8.92045975e-01
2.60534316e-01 6.66460097e-01 -2.73318253e-02 -3.11833918e-02
2.91745067e-01 1.47407562e-01 5.71084142e-01 -9.09129605e-02
-1.12900555e+00 -3.25148284e-01 -7.47953206e-02 2.11425513e-01
-3.13778877e-01 -5.73228717e-01 -1.04833996e+00 -3.93622190e-01
-1.11851685e-01 -1.98333606e-01 8.96627843e-01 9.42677677e-01
3.25172730e-02 5.39359629e-01 3.01995188e-01 -6.86692655e-01
-8.02586734e-01 -1.33272552e+00 -6.46069467e-01 2.49671340e-01
5.28406240e-02 -3.05354238e-01 8.39082003e-02 3.28447163e-01] | [11.831430435180664, 2.672022819519043] |
f6e90b6c-7aeb-47c5-bc82-13bf0e5634a6 | the-natural-language-pipeline-neural-text | null | null | https://aclanthology.org/2020.nl4xai-1.5 | https://aclanthology.org/2020.nl4xai-1.5.pdf | The Natural Language Pipeline, Neural Text Generation and Explainability | End-to-end encoder-decoder approaches to data-to-text generation are often black boxes whose predictions are difficult to explain. Breaking up the end-to-end model into sub-modules is a natural way to address this problem. The traditional pre-neural Natural Language Generation (NLG) pipeline provides a framework for breaking up the end-to-end encoder-decoder. We survey recent papers that integrate traditional NLG submodules in neural approaches and analyse their explainability. Our survey is a first step towards building explainable neural NLG models. | ['Claire Gardent', 'Albert Gatt', 'Juliette Faille'] | null | null | null | null | acl-nl4xai-inlg-2020-11 | ['data-to-text-generation'] | ['natural-language-processing'] | [ 3.66871327e-01 1.32422781e+00 -2.54753828e-01 -7.69077241e-01
-9.36461627e-01 -3.64611953e-01 9.13369358e-01 -1.08885385e-01
1.88821405e-01 8.64302397e-01 8.21257532e-01 -5.66328466e-01
4.53586638e-01 -7.28133619e-01 -9.44504678e-01 1.04230933e-01
3.43400389e-01 7.52495587e-01 -4.56510484e-01 -4.41212654e-01
2.49523178e-01 -2.34126434e-01 -1.16514695e+00 9.38628078e-01
7.40135133e-01 4.90453541e-01 2.68081546e-01 1.28641856e+00
-5.76581538e-01 1.06111419e+00 -5.22997558e-01 -7.01146662e-01
-1.49286002e-01 -1.17202866e+00 -1.13802576e+00 -1.63387790e-01
4.20018107e-01 -2.97474176e-01 -2.15658620e-01 5.87138116e-01
4.20384645e-01 -1.91805229e-01 7.19851136e-01 -1.16802895e+00
-1.45065022e+00 1.26035988e+00 1.14429526e-01 -2.79670745e-01
3.86806846e-01 5.33548221e-02 1.21694684e+00 -1.06782830e+00
8.85149837e-01 1.44389474e+00 6.08115613e-01 1.43440616e+00
-1.36664844e+00 -1.97409794e-01 3.13128710e-01 -1.00945406e-01
-8.11530471e-01 -7.42973089e-01 4.20951158e-01 -4.07821029e-01
1.74389851e+00 -9.34179034e-03 5.58003724e-01 1.34976733e+00
3.75480592e-01 1.05521953e+00 4.15770978e-01 -5.70263386e-01
4.18591350e-02 -2.37341866e-01 2.26497322e-01 7.21380949e-01
3.01415801e-01 4.24090713e-01 -9.49827015e-01 2.76690394e-01
6.60086334e-01 -4.97532666e-01 -7.97416344e-02 2.69676745e-01
-1.13239372e+00 1.06556785e+00 4.50572759e-01 9.95160714e-02
-1.73840627e-01 8.13342571e-01 3.31053913e-01 3.22813630e-01
5.93232512e-01 8.59060764e-01 -5.99687159e-01 -4.25489694e-01
-1.19792569e+00 7.02864408e-01 1.02459085e+00 1.33837199e+00
5.92456818e-01 5.27385712e-01 -5.07291675e-01 4.58545625e-01
5.49265921e-01 -1.15219742e-01 8.28819752e-01 -5.75820565e-01
6.75772548e-01 1.64589718e-01 -2.20145062e-01 -2.75765568e-01
-2.09728405e-01 -2.08267957e-01 -1.00637722e+00 1.08755566e-01
2.10521415e-01 -5.28233767e-01 -1.02895129e+00 1.70157599e+00
-2.69397646e-01 -1.65584847e-01 5.01927137e-01 7.41104007e-01
1.08787143e+00 1.00355220e+00 1.52222529e-01 1.76382348e-01
1.08909607e+00 -1.32474971e+00 -7.12549090e-01 -6.36233389e-01
7.19329894e-01 -5.69444895e-01 9.76493061e-01 1.25692621e-01
-1.53895092e+00 -7.38805950e-01 -7.98072278e-01 -8.34175646e-01
-3.15356493e-01 1.37409523e-01 7.48670459e-01 3.51676226e-01
-1.36677241e+00 7.72736847e-01 -7.68432856e-01 -3.21157239e-02
3.21365952e-01 2.62164503e-01 -2.09745407e-01 1.96216092e-01
-1.27138829e+00 1.06037164e+00 8.43711972e-01 -7.36486726e-03
-1.07085776e+00 -6.57823086e-01 -1.04881883e+00 3.69655117e-02
-4.44851778e-02 -1.42110348e+00 1.92542374e+00 -9.09277976e-01
-1.55651700e+00 6.60712540e-01 -7.19308674e-01 -9.08558547e-01
2.80481040e-01 -3.52642357e-01 -2.71656923e-02 -5.37363827e-01
3.05160861e-02 1.36275220e+00 5.93067586e-01 -9.54001963e-01
-3.75499606e-01 2.20493181e-03 -4.87837046e-02 1.46717072e-01
4.21110749e-01 -1.41634807e-01 -1.59602296e-02 -8.78116727e-01
-2.23520219e-01 -7.05753207e-01 -4.08243179e-01 -2.61080533e-01
-9.62377727e-01 -3.18708092e-01 1.44277573e-01 -6.67318940e-01
1.06130743e+00 -1.40309560e+00 3.28326344e-01 -5.53803205e-01
2.98299909e-01 -5.71686134e-04 -3.97064507e-01 7.96283126e-01
-4.90289509e-01 5.37882984e-01 -1.29824460e-01 -7.64945984e-01
3.76711547e-01 1.06859401e-01 -6.19292021e-01 -3.96759808e-01
7.24436700e-01 1.52916133e+00 -8.84797215e-01 -1.45994693e-01
7.08157197e-03 4.82266128e-01 -8.35638225e-01 5.74776590e-01
-1.04614580e+00 2.66428053e-01 -2.55040258e-01 3.02175701e-01
1.02670463e-02 -5.49177825e-02 -3.33364159e-01 3.95382136e-01
-1.03357680e-01 1.05344355e+00 -5.20670116e-01 2.04238772e+00
-6.18977666e-01 1.22302794e+00 -4.71913755e-01 -6.05838239e-01
9.78316844e-01 6.91012025e-01 -3.04106802e-01 -3.38956118e-01
-8.79364461e-02 2.95284927e-01 1.81651507e-02 -2.77893066e-01
8.25345635e-01 -5.48653543e-01 -1.91800550e-01 6.50447428e-01
5.24345994e-01 -3.63473117e-01 2.98780441e-01 3.09205621e-01
8.65156353e-01 5.07715285e-01 6.36620879e-01 -3.67734022e-02
9.05973986e-02 2.21185043e-01 2.08328757e-02 8.14301848e-01
4.07758117e-01 1.08743548e+00 5.11133373e-01 -7.49916494e-01
-1.44401813e+00 -8.84076715e-01 5.38400352e-01 9.58202183e-01
-6.47368550e-01 -8.67424786e-01 -1.13104558e+00 -6.26383960e-01
-2.78539985e-01 1.34483874e+00 -7.32036173e-01 -2.17047796e-01
-3.86316538e-01 -3.07185799e-01 1.00083697e+00 6.60526216e-01
5.54504395e-02 -1.28699076e+00 -2.59697616e-01 5.41831791e-01
-3.17597002e-01 -7.70880759e-01 -2.83714920e-01 4.96102750e-01
-9.74514604e-01 -3.95553261e-01 -6.66705608e-01 -8.35935652e-01
7.70047069e-01 -1.70255721e-01 1.66321230e+00 1.73874959e-01
-6.66266754e-02 -3.04764837e-01 -3.42939705e-01 -8.11206818e-01
-1.04806495e+00 5.38473368e-01 -3.75518739e-01 -6.44629776e-01
3.42367917e-01 -3.39124560e-01 -3.30224067e-01 -3.34848136e-01
-9.22774613e-01 1.04105937e+00 6.56590700e-01 7.36681044e-01
6.37887597e-01 -6.77540600e-01 7.52306163e-01 -1.01739597e+00
1.10765243e+00 -4.95819598e-01 -3.36606383e-01 1.56176671e-01
-7.53969133e-01 7.17676103e-01 9.68648076e-01 -5.76322339e-03
-8.90601218e-01 2.17000499e-01 -5.89687705e-01 2.01903075e-01
-3.90513182e-01 7.09361911e-01 -2.21106023e-01 7.34792531e-01
7.94883966e-01 2.98266172e-01 -1.52884349e-01 -5.14380395e-01
1.15162075e+00 5.10274172e-01 6.54986441e-01 -1.98626876e-01
7.48666704e-01 -2.91467667e-01 -3.18894625e-01 -3.67762536e-01
-1.07244205e+00 2.98895746e-01 -6.87208116e-01 7.14775473e-02
1.04140449e+00 -9.84948456e-01 -1.78249493e-01 -1.64258435e-01
-2.01918268e+00 -6.41124547e-01 -6.49285614e-01 6.21916912e-02
-9.65046763e-01 -1.02015801e-01 -5.95906436e-01 -6.40011191e-01
-7.85177827e-01 -9.18729186e-01 1.35099936e+00 1.86803430e-01
-9.66685712e-01 -1.26181686e+00 2.68646896e-01 2.36930817e-01
5.09035349e-01 9.66329426e-02 1.01788819e+00 -7.50361085e-01
-5.32036245e-01 -7.89905712e-02 -7.16945007e-02 1.35721907e-01
-1.92588836e-01 -1.90453887e-01 -1.04707372e+00 3.54218245e-01
-2.33468518e-01 -3.32439959e-01 1.07608294e+00 4.67070341e-01
1.08049393e+00 -8.13666761e-01 -1.19156227e-01 6.04784727e-01
1.34942782e+00 -1.62133038e-01 8.65419388e-01 -5.44139039e-05
9.06678438e-01 6.13928556e-01 2.20180243e-01 1.00984395e-01
6.85314775e-01 4.42092896e-01 4.31771845e-01 -2.56055772e-01
-5.13330102e-01 -9.91030991e-01 8.74612987e-01 7.77957916e-01
2.03945279e-01 -8.50095689e-01 -7.67401993e-01 6.47971392e-01
-1.82100546e+00 -1.08121705e+00 -5.40873468e-01 1.51187658e+00
1.12947261e+00 9.08992738e-02 -8.56077597e-02 -1.06302209e-01
3.44167829e-01 2.24237382e-01 -3.09710741e-01 -1.14880013e+00
2.78650448e-02 1.88690379e-01 1.47627369e-01 9.85263050e-01
-5.52986681e-01 1.41195309e+00 7.43730211e+00 6.70841396e-01
-8.20676863e-01 -5.95864691e-02 7.46028364e-01 -1.85611844e-01
-8.49664867e-01 3.48721087e-01 -1.11648607e+00 2.81658769e-01
1.56591535e+00 -3.59122455e-01 6.33999109e-01 1.03231871e+00
5.34408808e-01 2.90362358e-01 -1.77764845e+00 6.57461107e-01
1.16622776e-01 -1.79070139e+00 5.98437130e-01 -7.45511055e-02
6.77041650e-01 1.24081418e-01 -3.32994998e-01 3.32466066e-01
7.59637713e-01 -1.66814339e+00 1.07395589e+00 6.20019674e-01
8.73833597e-01 -5.20289838e-01 3.97248954e-01 4.90300149e-01
-7.89199173e-01 3.26063365e-01 -5.81500590e-01 -4.66061532e-01
7.96196401e-01 6.28689229e-01 -1.41190946e+00 3.81111205e-01
-1.72614753e-01 7.09633410e-01 -5.29107571e-01 5.38258970e-01
-8.07098925e-01 5.95547199e-01 3.77311051e-01 -2.64775664e-01
3.69965523e-01 2.39779890e-01 3.97690535e-01 1.57049108e+00
4.92178828e-01 -1.97655410e-01 -4.15163904e-01 1.51321042e+00
-3.37928832e-01 -2.93480784e-01 -9.45376158e-01 -3.87227565e-01
-6.58002123e-02 7.80773044e-01 -1.45073786e-01 -5.64852774e-01
-2.83742011e-01 1.36329377e+00 4.49876487e-01 1.55714035e-01
-6.88130200e-01 -4.53956991e-01 5.98959148e-01 1.35810405e-01
1.66432690e-02 -3.01766038e-01 -6.94376886e-01 -1.29186666e+00
-2.49828637e-01 -1.10989845e+00 -7.23613277e-02 -1.32132220e+00
-1.06206214e+00 8.54601085e-01 -1.87534779e-01 -6.84395492e-01
-1.32063472e+00 -5.94370723e-01 -6.64844394e-01 1.19145763e+00
-1.34068322e+00 -1.42069161e+00 -1.67441010e-01 1.28465891e-01
1.21459484e+00 -2.33056650e-01 1.22688079e+00 -2.39471734e-01
-2.22877339e-01 4.74461049e-01 -2.34658703e-01 1.26134411e-01
2.19194025e-01 -1.61624789e+00 1.61648870e+00 9.95665193e-01
5.73893964e-01 7.35025287e-01 7.74725020e-01 -8.54457796e-01
-1.22086477e+00 -1.34175313e+00 1.82439852e+00 -7.72250593e-01
3.33169937e-01 -7.50684202e-01 -5.42594194e-01 9.64928985e-01
7.84330845e-01 -5.70223212e-01 6.44985616e-01 1.67800963e-01
-1.98564678e-01 4.54610556e-01 -4.44577843e-01 8.47800136e-01
9.99454081e-01 -6.29928827e-01 -7.46308029e-01 3.19465965e-01
1.17058694e+00 -5.03907502e-01 -4.89071459e-01 -1.56335473e-01
6.11940742e-01 -7.65765369e-01 5.34768522e-01 -1.03126860e+00
1.47499740e+00 -1.86620504e-01 2.38534108e-01 -1.76511514e+00
-1.89429149e-01 -1.33896673e+00 -3.49760890e-01 1.17306697e+00
1.16972804e+00 -4.26372364e-02 8.98209035e-01 6.64610326e-01
-6.97106719e-01 -7.31975496e-01 -5.54045916e-01 -4.30236101e-01
4.03539598e-01 -9.07844663e-01 7.04743743e-01 4.45472330e-01
2.71879882e-01 1.02025974e+00 -5.59994102e-01 -5.96681774e-01
1.90900207e-01 -2.08413914e-01 9.68717217e-01 -1.02528572e+00
-5.05210161e-01 -5.42560697e-01 -6.19040690e-02 -1.53741634e+00
1.82394400e-01 -1.40659678e+00 4.99303043e-01 -2.32697201e+00
1.12593845e-01 3.10555279e-01 4.89189595e-01 6.89017355e-01
-2.17394963e-01 5.65693155e-02 2.89152563e-01 -4.55621928e-02
-3.25358212e-01 5.91857374e-01 1.15522504e+00 6.69677276e-03
8.11416190e-03 -1.50296077e-01 -1.26863456e+00 4.38528389e-01
8.40522349e-01 -6.53591573e-01 -5.71379840e-01 -1.26581728e+00
6.20705664e-01 2.45836124e-01 3.05167764e-01 -6.88262045e-01
6.79666102e-02 -1.24751143e-01 2.73839056e-01 -6.32788837e-01
1.10082723e-01 -1.37318209e-01 -2.47818306e-02 3.11096519e-01
-1.10002029e+00 3.96346390e-01 1.58406541e-01 2.77492523e-01
-4.10165191e-01 -3.09729815e-01 4.34106946e-01 -4.68453199e-01
-2.96236038e-01 1.08507268e-01 -7.79632807e-01 2.19773501e-01
5.11870384e-01 -2.21063510e-01 -4.55442101e-01 -1.04277694e+00
-6.77669764e-01 1.14878248e-02 1.96095675e-01 7.23692298e-01
7.49775708e-01 -1.21059823e+00 -1.12767768e+00 8.67209733e-02
1.55584095e-02 1.45458356e-01 -2.99806505e-01 1.57393724e-01
-6.47705078e-01 1.06406748e+00 -2.09199581e-02 4.30031605e-02
-9.72765326e-01 2.51617759e-01 5.56501627e-01 -5.26124120e-01
-3.98772925e-01 1.17759228e+00 1.66718945e-01 -3.33795011e-01
1.88725553e-02 -7.31950819e-01 1.22983105e-01 -2.90487915e-01
6.79337621e-01 -7.97397122e-02 -1.32042557e-01 -2.86369830e-01
8.51042420e-02 2.04099733e-02 -8.23853090e-02 -6.39574230e-01
1.34564424e+00 -8.89521092e-02 1.24191195e-01 5.69457114e-01
1.00080800e+00 -4.94990379e-01 -9.42863226e-01 2.32655242e-01
1.51394024e-01 2.35116392e-01 -1.85617119e-01 -1.07062435e+00
-6.70708716e-01 1.52518809e+00 -6.99082538e-02 2.44618297e-01
7.14666486e-01 2.28277892e-02 9.99286175e-01 4.55619782e-01
-9.00935307e-02 -8.44236851e-01 -8.97051245e-02 1.10693002e+00
1.29051626e+00 -9.49403703e-01 -2.58249074e-01 -9.94582027e-02
-8.46015275e-01 1.34961176e+00 6.40353501e-01 -1.08755440e-01
1.54325619e-01 2.90990114e-01 -5.98233417e-02 -1.07921176e-01
-1.46766675e+00 -1.84182659e-01 4.84475136e-01 7.20584333e-01
1.24203837e+00 -1.35803586e-02 -2.73146451e-01 9.49785173e-01
-9.58559632e-01 1.38652369e-01 6.44509792e-01 3.52874368e-01
-2.84543276e-01 -1.51800764e+00 5.30406237e-02 2.00739697e-01
-4.79222268e-01 -7.99319863e-01 -1.10866904e+00 5.98262787e-01
6.74857870e-02 1.04062009e+00 7.21003488e-03 -3.92549187e-01
1.05652265e-01 4.14900154e-01 2.84140170e-01 -1.08860004e+00
-9.03545499e-01 -1.91986039e-01 5.33021212e-01 -4.77145076e-01
9.34041589e-02 -2.84075767e-01 -1.49747300e+00 -2.73470342e-01
-1.25909954e-01 9.08842087e-02 8.36175144e-01 1.06186259e+00
6.53815150e-01 7.54624605e-01 1.26237884e-01 -8.56244564e-01
-2.82933295e-01 -1.09952664e+00 9.64620803e-03 -2.14323178e-02
5.60248256e-01 3.17002654e-01 7.21670836e-02 6.61382616e-01] | [11.676753997802734, 8.961003303527832] |
8bba6353-3e22-4133-b3f1-6171f340d23b | openshape-scaling-up-3d-shape-representation | 2305.10764 | null | https://arxiv.org/abs/2305.10764v2 | https://arxiv.org/pdf/2305.10764v2.pdf | OpenShape: Scaling Up 3D Shape Representation Towards Open-World Understanding | We introduce OpenShape, a method for learning multi-modal joint representations of text, image, and point clouds. We adopt the commonly used multi-modal contrastive learning framework for representation alignment, but with a specific focus on scaling up 3D representations to enable open-world 3D shape understanding. To achieve this, we scale up training data by ensembling multiple 3D datasets and propose several strategies to automatically filter and enrich noisy text descriptions. We also explore and compare strategies for scaling 3D backbone networks and introduce a novel hard negative mining module for more efficient training. We evaluate OpenShape on zero-shot 3D classification benchmarks and demonstrate its superior capabilities for open-world recognition. Specifically, OpenShape achieves a zero-shot accuracy of 46.8% on the 1,156-category Objaverse-LVIS benchmark, compared to less than 10% for existing methods. OpenShape also achieves an accuracy of 85.3% on ModelNet40, outperforming previous zero-shot baseline methods by 20% and performing on par with some fully-supervised methods. Furthermore, we show that our learned embeddings encode a wide range of visual and semantic concepts (e.g., subcategories, color, shape, style) and facilitate fine-grained text-3D and image-3D interactions. Due to their alignment with CLIP embeddings, our learned shape representations can also be integrated with off-the-shelf CLIP-based models for various applications, such as point cloud captioning and point cloud-conditioned image generation. | ['Hao Su', 'Fatih Porikli', 'Hong Cai', 'Shizhong Han', 'Xuanlin Li', 'Yinhao Zhu', 'Kaiming Kuang', 'Ruoxi Shi', 'Minghua Liu'] | 2023-05-18 | null | null | null | null | ['3d-shape-representation', '3d-classification', 'zero-shot-transfer-3d-point-cloud', 'open-set-learning'] | ['computer-vision', 'computer-vision', 'computer-vision', 'miscellaneous'] | [-1.40727907e-01 -1.27145022e-01 -1.75479919e-01 -3.05610240e-01
-1.10456848e+00 -8.11639369e-01 8.40888917e-01 1.72460496e-01
-9.76676568e-02 6.40937537e-02 3.92214835e-01 1.04588248e-01
2.97216207e-01 -7.59858549e-01 -9.12149668e-01 -3.27570856e-01
3.19433600e-01 9.07845914e-01 7.23697692e-02 -2.31151223e-01
1.92381755e-01 7.82332838e-01 -1.87469542e+00 3.41745287e-01
4.69380885e-01 1.14254642e+00 -1.50961857e-02 4.65234995e-01
-6.95725977e-01 -6.73884060e-03 -4.26984221e-01 -4.92973745e-01
3.87795299e-01 3.07859570e-01 -6.85885131e-01 2.99219817e-01
1.04731750e+00 -2.62472034e-01 -3.27712536e-01 7.72395849e-01
6.84019029e-01 8.28181207e-02 9.59209502e-01 -1.42487276e+00
-1.11605573e+00 6.56120405e-02 -7.82016635e-01 -2.62027979e-01
4.55404311e-01 1.86305940e-01 1.22050893e+00 -1.44281578e+00
8.90704632e-01 1.41163564e+00 6.50003612e-01 6.68038428e-01
-1.35201383e+00 -7.29935408e-01 2.59844154e-01 -3.46686542e-02
-1.55679739e+00 -3.19048345e-01 8.69673073e-01 -4.26826149e-01
1.33148718e+00 1.08691536e-01 6.76538169e-01 1.51675534e+00
-1.15109965e-01 9.73428667e-01 7.27290928e-01 -2.66713351e-01
5.35327233e-02 9.43499878e-02 -9.47838947e-02 6.73480690e-01
1.17277965e-01 -1.66217625e-01 -6.44873917e-01 -3.03036273e-01
8.59884322e-01 7.07505718e-02 7.30361119e-02 -9.95732665e-01
-1.41979432e+00 7.59712219e-01 6.74750805e-01 9.36659127e-02
-1.86441932e-02 2.87626475e-01 4.17845130e-01 6.63889349e-02
7.82689333e-01 4.07710969e-01 -4.07507271e-01 -1.18695153e-02
-8.11261594e-01 3.16159248e-01 6.24457300e-01 1.29475486e+00
7.69807935e-01 1.32942170e-01 -2.82417238e-01 1.13544571e+00
2.94103473e-01 9.29276943e-01 4.99957614e-02 -8.89885902e-01
5.86530149e-01 7.55942583e-01 -9.52708125e-02 -8.63589346e-01
-2.93935418e-01 -1.94805950e-01 -8.42949390e-01 3.09972078e-01
-1.11216240e-01 4.01221007e-01 -1.22652173e+00 1.43592298e+00
2.63076037e-01 2.36909851e-01 -3.87704894e-02 8.96595180e-01
1.19040680e+00 7.16422558e-01 7.46714398e-02 4.09910172e-01
1.13556266e+00 -9.44527686e-01 -1.91863671e-01 -2.74510592e-01
5.55943608e-01 -9.09694314e-01 1.29522908e+00 1.28252834e-01
-9.51538384e-01 -5.14470160e-01 -8.56717527e-01 -4.77969259e-01
-6.15261614e-01 -1.68120824e-02 5.22425056e-01 2.25300640e-01
-9.84127641e-01 4.68910247e-01 -6.03607833e-01 -5.96628666e-01
7.38841057e-01 1.46624073e-01 -6.91768765e-01 -3.18565935e-01
-6.58438087e-01 7.47423351e-01 8.62170756e-02 -6.24518991e-01
-8.26107085e-01 -9.75047112e-01 -1.10241580e+00 -1.56470671e-01
-5.07882982e-03 -9.58023131e-01 9.25872326e-01 -4.37727273e-01
-1.09307837e+00 1.53477657e+00 -3.35095860e-02 -1.59410909e-01
3.40377480e-01 -1.80996940e-01 -1.77490085e-01 2.47587070e-01
3.02638233e-01 1.39389312e+00 9.74483907e-01 -1.48130071e+00
-1.66321561e-01 -4.78022307e-01 4.31858562e-02 4.02217746e-01
-4.43915755e-01 -2.19550163e-01 -8.31473291e-01 -7.07878649e-01
2.14494377e-01 -8.46675992e-01 -6.61795363e-02 9.37757730e-01
-3.74479771e-01 -3.65729600e-01 1.10137129e+00 -2.26436064e-01
4.35953200e-01 -2.20458174e+00 2.46396303e-01 -6.42604828e-02
2.80792832e-01 4.70020175e-02 -6.23152494e-01 3.69957775e-01
-1.10513851e-01 2.76288420e-01 -1.57616362e-01 -8.54207039e-01
3.52292448e-01 3.31131428e-01 -4.28028405e-01 3.68669212e-01
4.94482338e-01 1.22436047e+00 -6.31992817e-01 -5.35901368e-01
7.42564678e-01 7.19828010e-01 -6.37915611e-01 1.74658716e-01
-5.37416101e-01 3.64061557e-02 -2.38487855e-01 1.03086972e+00
8.37981105e-01 -5.31372786e-01 -3.88430774e-01 -4.92680103e-01
1.74228773e-01 1.85191445e-02 -1.10391760e+00 2.32959700e+00
-6.65760338e-01 5.20629585e-01 -1.12920247e-01 -8.37532938e-01
1.17313147e+00 9.73332450e-02 6.21075273e-01 -6.23980105e-01
1.33629858e-01 -7.79454655e-04 -7.59967923e-01 -1.37657017e-01
5.71050227e-01 7.09238350e-02 -2.76553601e-01 5.58479488e-01
3.86166722e-01 -7.35340238e-01 -1.83479697e-01 4.34517890e-01
8.24568093e-01 4.65771109e-01 -4.78959130e-03 -9.69680026e-02
8.95083398e-02 3.70305367e-02 2.86748242e-02 5.71562827e-01
-4.58678082e-02 1.20825326e+00 2.06266299e-01 -5.37218571e-01
-1.31617820e+00 -1.37745655e+00 -2.78600305e-01 1.00177634e+00
3.96144360e-01 -5.35087049e-01 -1.64105386e-01 -7.99068153e-01
4.99996513e-01 6.81119382e-01 -6.67605102e-01 -6.01394065e-02
-1.09531149e-01 -3.29946786e-01 4.76724148e-01 7.66299605e-01
2.41805092e-01 -6.27350628e-01 2.03399789e-02 -8.31626132e-02
-2.19613761e-02 -1.36675787e+00 -4.15319830e-01 1.84301823e-01
-7.66202331e-01 -8.76431406e-01 -8.08926344e-01 -8.69186461e-01
6.27341986e-01 6.93754315e-01 1.47609842e+00 4.72853379e-03
-4.07579154e-01 8.69010985e-01 -4.65780973e-01 -3.48538518e-01
-2.28798226e-01 3.27047775e-03 9.63639617e-02 -2.88472265e-01
4.85828370e-01 -6.21272027e-01 -4.29466516e-01 2.37799466e-01
-8.63000870e-01 1.72039777e-01 3.85790497e-01 7.06529379e-01
9.66076076e-01 -7.17796564e-01 2.57915437e-01 -2.87816465e-01
2.27727562e-01 -4.64576989e-01 -3.70567143e-01 1.75953135e-01
-2.97839344e-01 3.01174130e-02 3.19826424e-01 -4.84728903e-01
-6.49228454e-01 1.33813724e-01 -2.92312384e-01 -1.27931118e+00
-4.04500127e-01 -1.05662316e-01 -1.03828333e-01 -2.49094069e-01
7.68822908e-01 1.78875133e-01 1.58887476e-01 -5.20466924e-01
8.94259036e-01 6.91970408e-01 4.18306619e-01 -8.28415930e-01
1.18131876e+00 8.25265646e-01 -1.43286064e-01 -8.96707177e-01
-1.07054305e+00 -6.91248119e-01 -7.02761710e-01 7.29889125e-02
9.34041619e-01 -1.29125476e+00 -3.87886852e-01 3.54493707e-01
-1.34509087e+00 -1.05293520e-01 -4.13160980e-01 1.11173153e-01
-7.63434350e-01 3.57542068e-01 -4.40867454e-01 -3.80191535e-01
-5.27802885e-01 -9.46219027e-01 1.95353091e+00 -1.98538527e-01
-8.14705417e-02 -9.27972674e-01 -5.64657077e-02 4.59029824e-01
1.90634608e-01 2.05629185e-01 9.07267034e-01 -6.31281853e-01
-7.31950045e-01 -5.11705838e-02 -6.66457653e-01 3.11926901e-01
-1.41953960e-01 -3.40414457e-02 -1.12688446e+00 -3.08168590e-01
-6.43695474e-01 -7.89253414e-01 7.98186004e-01 1.29955590e-01
1.37837577e+00 1.37969911e-01 -4.48812425e-01 9.47280765e-01
1.44800889e+00 -4.93257850e-01 4.09719408e-01 1.13126673e-01
9.86924231e-01 3.25675130e-01 5.75598121e-01 5.47025561e-01
3.48828644e-01 8.05218577e-01 8.45998347e-01 -1.60524055e-01
-3.98334265e-01 -4.96754676e-01 -2.28025746e-02 6.71120703e-01
2.55615324e-01 -2.49598086e-01 -9.65456843e-01 6.36186898e-01
-1.67770207e+00 -7.33580291e-01 1.00154206e-02 1.85294092e+00
5.53128064e-01 6.29228950e-02 -6.68028668e-02 -1.51702583e-01
6.02858305e-01 4.48603213e-01 -6.57019198e-01 -1.97311431e-01
-4.07979220e-01 4.27164704e-01 4.50398773e-01 2.05426320e-01
-1.23131740e+00 1.16736174e+00 5.96857119e+00 8.96254301e-01
-1.00552499e+00 5.10362536e-02 4.63015407e-01 -3.93833965e-01
-6.50105476e-01 -2.58505017e-01 -9.09160137e-01 -8.76882449e-02
4.48751092e-01 -6.92226738e-02 2.85096198e-01 1.04581082e+00
-2.22011656e-01 4.04923856e-01 -1.23729980e+00 1.56655705e+00
5.30296683e-01 -1.94257247e+00 5.01400054e-01 1.01308778e-01
9.47084725e-01 5.48207700e-01 -1.00629933e-01 3.63286495e-01
3.11229944e-01 -1.06447303e+00 8.07768404e-01 3.52298856e-01
1.29180169e+00 -6.96366966e-01 3.12203050e-01 1.53932631e-01
-1.34846509e+00 1.82042494e-01 -4.72561896e-01 2.67906159e-01
2.22464427e-01 4.70837355e-01 -5.75132549e-01 5.01255751e-01
9.40726519e-01 1.19104600e+00 -6.22434855e-01 9.66394305e-01
-2.71913037e-02 -1.85792334e-02 -5.86679876e-01 -9.61348191e-02
2.69615829e-01 1.13269553e-01 7.13720977e-01 1.10877693e+00
4.06810015e-01 -3.45914721e-01 2.84102350e-01 1.21468234e+00
-4.58260000e-01 1.94497555e-01 -1.06260371e+00 -9.22336653e-02
6.17866814e-01 1.18799341e+00 -5.16126573e-01 -3.32655370e-01
-6.07790828e-01 1.19374704e+00 4.71945435e-01 1.37398243e-01
-8.23568285e-01 -3.02052408e-01 1.09413517e+00 1.55646307e-02
5.39282501e-01 -4.10308152e-01 -4.65476185e-01 -1.30930007e+00
4.65324633e-02 -5.28876901e-01 1.08480774e-01 -1.29920077e+00
-1.77424800e+00 5.00364423e-01 -3.64432186e-02 -1.53700495e+00
6.39432147e-02 -8.56472433e-01 -3.88558924e-01 7.13215590e-01
-1.47809184e+00 -1.74294901e+00 -5.34736037e-01 7.37816155e-01
7.65238583e-01 -1.77653208e-01 1.13973248e+00 1.31136864e-01
-1.07448757e-01 4.43536580e-01 -2.23880652e-02 1.84439301e-01
8.32424402e-01 -1.11305594e+00 1.01300740e+00 3.16699415e-01
6.40944004e-01 1.60313606e-01 2.34823465e-01 -5.60932636e-01
-1.55918145e+00 -1.37441301e+00 5.06553113e-01 -7.87343144e-01
7.26863801e-01 -8.63453507e-01 -9.07166004e-01 5.18078208e-01
-3.00835073e-02 4.69960004e-01 5.69166422e-01 1.58540860e-01
-9.63826954e-01 -2.72175781e-02 -1.06256056e+00 6.24025464e-01
1.54436994e+00 -7.45633662e-01 -7.73757279e-01 5.39040267e-01
1.09649599e+00 -6.07421279e-01 -9.59095001e-01 3.83259892e-01
4.00741488e-01 -7.22072542e-01 1.34407353e+00 -7.04041839e-01
6.33668303e-01 -7.20593855e-02 -6.10305250e-01 -1.19375241e+00
-3.83378178e-01 -1.95811614e-01 -1.71848685e-01 1.12574446e+00
2.36575082e-01 -1.41895249e-01 1.01106918e+00 2.33674094e-01
-2.93440670e-01 -6.92007005e-01 -9.27073061e-01 -7.61422694e-01
2.84238845e-01 -9.00673926e-01 6.19926214e-01 1.02715945e+00
-4.16434646e-01 4.71208304e-01 -2.53208548e-01 1.25578821e-01
7.35809922e-01 4.46030498e-01 1.06743610e+00 -1.41625714e+00
7.22464174e-02 -4.28650677e-01 -7.30702877e-01 -1.25181031e+00
4.90038395e-01 -1.35380590e+00 -3.09304774e-01 -1.78005958e+00
2.35224456e-01 -5.25238991e-01 -6.17390573e-02 5.45858502e-01
2.98297465e-01 6.96337521e-01 3.96987140e-01 1.66675195e-01
-7.43389964e-01 9.10310507e-01 1.33931077e+00 -5.24513185e-01
2.29458027e-02 -4.15336788e-01 -6.95351362e-01 6.11777067e-01
4.88548070e-01 -1.63072467e-01 -2.35489577e-01 -7.63847709e-01
1.18908755e-01 -3.08514744e-01 8.10333133e-01 -9.51953411e-01
-5.14297187e-02 -1.16051711e-01 5.29120088e-01 -1.08214617e+00
8.36327553e-01 -8.32681537e-01 -1.32776305e-01 -1.84572026e-01
-1.69759110e-01 -2.20871270e-01 4.83152211e-01 6.67991340e-01
-2.33213548e-02 1.32572055e-01 7.83629298e-01 -1.50920838e-01
-8.96654665e-01 7.28024900e-01 2.37645060e-01 2.57601917e-01
1.00827026e+00 -3.43109578e-01 -5.12818694e-01 -2.06396103e-01
-6.38071597e-01 2.39849120e-01 8.80772889e-01 1.01307046e+00
9.80592489e-01 -1.72632694e+00 -6.19112909e-01 4.59675133e-01
8.53064656e-01 2.25442410e-01 3.25086266e-01 2.67422110e-01
-3.29475462e-01 2.92006671e-01 -2.09668845e-01 -1.18335450e+00
-1.11879742e+00 4.36258763e-01 2.33740449e-01 7.78606683e-02
-8.83151412e-01 8.00127923e-01 2.40189493e-01 -8.43996644e-01
3.89056683e-01 -2.06128165e-01 1.68704301e-01 -6.00189939e-02
2.92984843e-01 -1.21024407e-01 2.11476833e-01 -6.85505331e-01
-4.28052545e-01 1.10397875e+00 8.29262957e-02 1.11606710e-01
1.48014295e+00 -2.16987263e-02 1.81793079e-01 5.64993143e-01
1.50862336e+00 -2.60969877e-01 -1.26296151e+00 -4.22012001e-01
-4.94319916e-01 -5.06432116e-01 1.34153068e-01 -6.97487473e-01
-9.03482676e-01 1.07385838e+00 5.67902803e-01 -1.51485741e-01
6.60940826e-01 5.78038275e-01 7.39489794e-01 5.49088180e-01
4.75498259e-01 -9.95890856e-01 4.70470309e-01 6.57050908e-01
1.16921461e+00 -1.51286650e+00 -7.31287822e-02 -3.81734610e-01
-6.39050663e-01 1.01360059e+00 6.83683038e-01 -2.86905617e-01
6.62646115e-01 1.78720608e-01 3.32251117e-02 -3.80812585e-01
-9.56365883e-01 -4.31520700e-01 5.44955850e-01 9.58436310e-01
1.47162259e-01 -3.56354304e-02 4.67389643e-01 3.05046588e-01
-1.26513019e-01 -2.52932429e-01 1.60822332e-01 7.21333027e-01
-3.29198092e-01 -9.95645463e-01 -3.56599420e-01 5.41964233e-01
9.70529690e-02 -1.32518962e-01 -4.57502663e-01 8.99937391e-01
-3.48037668e-02 3.62080902e-01 5.48906565e-01 -3.80800903e-01
4.81455326e-01 1.37129471e-01 5.74866712e-01 -8.38975191e-01
-9.52315703e-02 4.54751449e-03 2.90949661e-02 -6.74470425e-01
-4.27027911e-01 -5.07735252e-01 -1.07440841e+00 -2.14968637e-01
-1.40937924e-01 -3.68694305e-01 8.98518860e-01 6.78998411e-01
7.14348853e-01 1.96814686e-01 5.13153255e-01 -1.39203143e+00
-3.06970954e-01 -7.21119523e-01 -4.28593040e-01 7.81941533e-01
4.91020493e-02 -9.64857101e-01 -4.88066792e-01 -1.90923825e-01] | [8.152865409851074, -3.3014943599700928] |
ab255369-11cf-4ad8-b4f1-8594411691c7 | hardware-trojan-attacks-on-neural-networks | 1806.05768 | null | http://arxiv.org/abs/1806.05768v1 | http://arxiv.org/pdf/1806.05768v1.pdf | Hardware Trojan Attacks on Neural Networks | With the rising popularity of machine learning and the ever increasing demand
for computational power, there is a growing need for hardware optimized
implementations of neural networks and other machine learning models. As the
technology evolves, it is also plausible that machine learning or artificial
intelligence will soon become consumer electronic products and military
equipment, in the form of well-trained models. Unfortunately, the modern
fabless business model of manufacturing hardware, while economic, leads to
deficiencies in security through the supply chain. In this paper, we illuminate
these security issues by introducing hardware Trojan attacks on neural
networks, expanding the current taxonomy of neural network security to
incorporate attacks of this nature. To aid in this, we develop a novel
framework for inserting malicious hardware Trojans in the implementation of a
neural network classifier. We evaluate the capabilities of the adversary in
this setting by implementing the attack algorithm on convolutional neural
networks while controlling a variety of parameters available to the adversary.
Our experimental results show that the proposed algorithm could effectively
classify a selected input trigger as a specified class on the MNIST dataset by
injecting hardware Trojans into $0.03\%$, on average, of neurons in the 5th
hidden layer of arbitrary 7-layer convolutional neural networks, while
undetectable under the test data. Finally, we discuss the potential defenses to
protect neural networks against hardware Trojan attacks. | ['Yingjie Lao', 'Joseph Clements'] | 2018-06-14 | null | null | null | null | ['neural-network-security'] | ['miscellaneous'] | [ 6.06006086e-01 -1.46486992e-02 -1.17462270e-01 -2.43967012e-01
5.73846698e-03 -8.93841267e-01 2.13992372e-01 -1.55799389e-01
-7.28859484e-01 3.74769896e-01 -7.42908716e-01 -1.06213856e+00
5.13911359e-02 -8.60351682e-01 -1.04424477e+00 -7.20688879e-01
-2.13168278e-01 -1.61239192e-01 2.99054533e-01 -5.52157052e-02
3.16661984e-01 1.00739491e+00 -1.48997712e+00 4.01727170e-01
2.45587274e-01 1.13503933e+00 -4.63210851e-01 8.71312797e-01
3.71057361e-01 5.96312821e-01 -1.05149853e+00 -6.87117994e-01
6.06253028e-01 1.34611400e-02 -5.64943433e-01 -3.84307593e-01
5.18409550e-01 -3.83676648e-01 -4.43913430e-01 1.43723130e+00
1.61778301e-01 -4.71815318e-01 7.37911761e-02 -1.65192783e+00
-2.23323539e-01 9.79985416e-01 -3.74624878e-01 2.12754339e-01
-3.40103030e-01 2.11245924e-01 5.91342747e-01 -1.08039364e-01
1.68345839e-01 1.04357755e+00 3.66046399e-01 8.28916430e-01
-8.68284523e-01 -1.30601597e+00 -4.38272618e-02 -1.56649783e-01
-1.22846913e+00 -3.47288221e-01 8.05136025e-01 -1.96968585e-01
1.05551684e+00 4.08179790e-01 2.51001596e-01 1.44632971e+00
8.08956683e-01 4.73735154e-01 7.79889643e-01 -4.67840999e-01
4.52214509e-01 1.61022872e-01 5.42182684e-01 8.01050246e-01
1.00161636e+00 5.07762551e-01 -1.26867086e-01 -5.77580214e-01
5.69901466e-01 4.45858799e-02 -2.80613787e-02 -1.05660498e-01
-6.26744747e-01 9.60742772e-01 4.27313179e-01 1.27669021e-01
1.07592627e-01 7.56382704e-01 7.61281431e-01 4.54757005e-01
-1.59739390e-01 6.49671555e-01 -8.63833845e-01 2.30551466e-01
-3.71253222e-01 -2.89843809e-02 8.24283600e-01 6.61002159e-01
2.84584016e-01 5.20298958e-01 9.20827031e-01 -2.06117500e-02
3.40755135e-01 3.25313270e-01 4.67519611e-01 -5.41633785e-01
3.01044703e-01 3.59150916e-01 -4.02208805e-01 -8.64598095e-01
-3.58677715e-01 -6.29859507e-01 -7.63163686e-01 6.37606025e-01
3.38613510e-01 -5.03367543e-01 -7.65792847e-01 1.70736742e+00
1.33617565e-01 3.81318212e-01 3.64910156e-01 5.33770204e-01
8.70647728e-02 4.44981307e-01 -1.00924529e-01 9.59355906e-02
1.49689710e+00 -5.61156988e-01 -3.38934332e-01 -3.51656675e-01
7.70776391e-01 -5.34560382e-01 5.14211476e-01 6.71366930e-01
-5.71904600e-01 -3.91683668e-01 -1.72188568e+00 4.52055842e-01
-6.25360012e-01 -1.62784949e-01 8.72178197e-01 1.74869537e+00
-6.99528456e-01 4.93689299e-01 -1.19919407e+00 -9.45815071e-03
5.77281237e-01 9.28508520e-01 -2.26918176e-01 1.66425064e-01
-1.18374968e+00 7.48443544e-01 7.45769382e-01 1.08857695e-02
-8.66449654e-01 -5.18543065e-01 -8.12795043e-01 7.43547529e-02
-6.03743792e-02 -2.89231539e-01 1.09623098e+00 -1.04484499e+00
-1.28420496e+00 4.22592103e-01 6.07768893e-01 -1.02880144e+00
1.72188476e-01 -1.38436496e-01 -7.09559798e-01 4.70245164e-03
-7.54993916e-01 4.46663797e-01 1.08745480e+00 -1.22732437e+00
-8.84692729e-01 -3.69274288e-01 1.90285921e-01 -5.95109642e-01
-6.28815353e-01 2.87489712e-01 4.35420185e-01 -2.36407191e-01
-1.40724316e-01 -1.18758857e+00 -3.01502198e-01 -1.62322342e-01
-5.29444575e-01 4.16297078e-01 1.41082036e+00 4.38052192e-02
8.00890803e-01 -2.37558150e+00 -3.61032039e-01 5.14847755e-01
9.42796916e-02 6.19964838e-01 -2.25515831e-02 -1.35780752e-01
-2.63565391e-01 2.58309305e-01 -2.54661776e-02 5.45252636e-02
-8.58597234e-02 4.51108247e-01 -8.76973569e-01 8.12178612e-01
2.24257484e-01 5.38096428e-01 -4.93749827e-01 1.66115597e-01
8.08184128e-03 5.09782910e-01 -6.53412521e-01 -2.45694593e-01
3.59381065e-02 -5.71501479e-02 -5.18294811e-01 5.75443387e-01
5.91393650e-01 3.47093344e-02 2.46333286e-01 6.23076335e-02
5.87551966e-02 1.27516434e-01 -8.65037620e-01 9.34149086e-01
-3.38847727e-01 7.92678952e-01 5.84883802e-02 -7.72734702e-01
6.80493832e-01 2.50148177e-01 -1.33693635e-01 -2.14324266e-01
8.73894393e-01 2.56069154e-01 8.01198304e-01 -6.12588041e-02
1.14195637e-01 1.06164530e-01 -3.08978498e-01 4.44213539e-01
-1.23008773e-01 1.31831616e-01 -4.91556853e-01 -1.27125278e-01
1.51594031e+00 -5.19916058e-01 7.20542967e-02 -2.44588673e-01
3.92646164e-01 -1.66879520e-02 2.66529053e-01 7.83553839e-01
-3.31551433e-01 -1.44167632e-01 4.58135068e-01 -7.29722321e-01
-9.18456256e-01 -8.45746219e-01 -2.25910693e-01 8.56783748e-01
-2.16987878e-01 -4.52274904e-02 -1.09372020e+00 -7.97044337e-01
-3.09030712e-02 6.73824608e-01 -5.00368237e-01 -6.64505064e-01
-6.13648713e-01 -6.94565713e-01 1.47633684e+00 4.55307633e-01
5.73457003e-01 -8.26015115e-01 -1.44823635e+00 -1.48120582e-01
8.78151536e-01 -1.21966755e+00 3.54037173e-02 8.95162463e-01
-1.05883634e+00 -1.00986803e+00 1.85130507e-01 -8.00897002e-01
8.24540615e-01 9.12845209e-02 4.99505460e-01 4.52364951e-01
-5.27543306e-01 1.92143153e-02 -1.16752170e-01 -7.77095973e-01
-7.02363551e-01 2.58336604e-01 4.55168694e-01 -5.41823208e-02
4.23938483e-01 -5.37488341e-01 -2.44709268e-01 1.09963313e-01
-1.16130149e+00 -3.15101862e-01 6.87172472e-01 7.62581289e-01
-1.03789993e-01 8.29645872e-01 3.24853003e-01 -1.13252091e+00
3.51370245e-01 -3.56221437e-01 -9.60422337e-01 -1.53458625e-01
-4.30386633e-01 2.49266759e-01 1.09877956e+00 -8.35369527e-01
-4.73537296e-01 4.23877984e-01 -3.58504117e-01 -5.54965436e-01
-2.79576838e-01 3.11269879e-01 -4.31263179e-01 -7.21731424e-01
8.86726797e-01 -2.74199247e-02 -1.04381263e-01 -5.03676012e-02
9.08447132e-02 4.48495626e-01 5.60849667e-01 -4.11108166e-01
1.20760679e+00 4.01181042e-01 1.62700132e-01 -9.17195082e-01
-2.31191620e-01 2.14153007e-01 -1.96781099e-01 5.84359355e-02
4.87916857e-01 -4.59013909e-01 -1.07752621e+00 7.28650689e-01
-1.22984922e+00 -2.34821364e-01 2.98652023e-01 3.19831401e-01
3.21809053e-02 7.91939795e-02 -7.71117568e-01 -8.88976038e-01
-2.81274319e-01 -1.63791597e+00 4.83027935e-01 1.39960393e-01
-1.23560019e-01 -7.43818581e-01 -4.13476497e-01 1.39243275e-01
3.46505940e-01 3.77615005e-01 1.34792912e+00 -1.10478640e+00
-5.87654650e-01 -7.30151117e-01 4.92284857e-02 5.34272790e-01
-1.49288461e-01 1.93708912e-01 -1.38672507e+00 -5.19852340e-01
3.65824521e-01 -2.81858057e-01 5.39880097e-01 3.67714502e-02
1.22512138e+00 -5.87899804e-01 -1.65418416e-01 7.57008851e-01
1.42721725e+00 7.01507092e-01 6.68816268e-01 5.12095869e-01
6.19215429e-01 4.57256228e-01 4.40187119e-02 1.65544301e-01
-5.15803397e-01 1.94564521e-01 1.04459107e+00 2.46996313e-01
6.86595440e-01 1.62053872e-02 4.96406674e-01 3.62188607e-01
3.99505138e-01 -1.24850422e-01 -9.72719073e-01 -6.44511404e-03
-1.26965737e+00 -6.71299160e-01 4.32010032e-02 2.04972553e+00
6.22803032e-01 8.93146217e-01 -2.82728195e-01 6.41392887e-01
5.67999065e-01 -9.18761715e-02 -6.58162773e-01 -9.44028139e-01
1.91914365e-01 6.40190065e-01 1.25588739e+00 2.30764806e-01
-1.15054893e+00 8.28768611e-01 6.34001446e+00 5.52221477e-01
-1.48345387e+00 -3.00875343e-02 6.19968057e-01 -4.68603186e-02
1.32560521e-01 -1.80792317e-01 -9.24921036e-01 2.32823923e-01
1.56655860e+00 9.44670588e-02 3.85608435e-01 1.29777539e+00
-3.17042440e-01 2.16559365e-01 -1.14006054e+00 6.72367036e-01
-9.51438174e-02 -1.29996955e+00 -1.24613643e-01 3.35517943e-01
3.35745811e-01 -5.99579141e-03 5.73305726e-01 2.22366869e-01
4.59195673e-01 -1.22421944e+00 7.90103376e-01 -4.89486754e-01
4.30034190e-01 -1.40873396e+00 8.90629292e-01 5.29318035e-01
-7.40645647e-01 -5.06047606e-01 -4.16269451e-01 -3.11361104e-01
-5.26171982e-01 4.99733053e-02 -8.18438053e-01 -3.48676480e-02
6.15338743e-01 -3.38372767e-01 -5.58560729e-01 6.23426557e-01
-1.87452719e-01 7.17754304e-01 -3.27352434e-01 -1.11413956e-01
3.41770977e-01 4.11005586e-01 -1.35020856e-02 1.03056645e+00
5.22271432e-02 -1.41312703e-02 -1.57620609e-01 4.16055590e-01
-3.27222258e-01 -5.30981779e-01 -9.73311484e-01 -2.63879240e-01
6.39456511e-01 1.18953061e+00 -1.08743763e+00 2.04019547e-01
-2.66339481e-01 6.78534627e-01 1.82709340e-02 7.48226373e-03
-9.98555064e-01 -6.09710872e-01 1.21838880e+00 -2.30357721e-01
-3.19756456e-02 -3.42993647e-01 -6.32703602e-01 -5.75158477e-01
-3.29165727e-01 -1.32580674e+00 2.62197912e-01 -1.84046030e-01
-8.16157460e-01 8.11466038e-01 -2.09111691e-01 -7.72862196e-01
-1.92576572e-01 -1.25486696e+00 -5.56258619e-01 3.83037388e-01
-1.00013757e+00 -7.68369079e-01 3.31261754e-01 3.63139987e-01
2.16076061e-01 -5.82545519e-01 9.15617824e-01 2.12856889e-01
-8.96611929e-01 1.14912677e+00 -3.03601027e-02 5.87217450e-01
1.63527727e-01 -4.73321438e-01 8.20686162e-01 1.33384693e+00
2.63344824e-01 1.09516311e+00 9.20192361e-01 -4.54650044e-01
-1.83344769e+00 -1.21470952e+00 2.88847595e-01 -3.12094808e-01
9.31256711e-01 -8.00365090e-01 -8.09237838e-01 8.69995475e-01
1.54258028e-01 7.02053457e-02 7.93893218e-01 -3.42045695e-01
-8.32864404e-01 1.11131325e-01 -1.45784652e+00 9.15235758e-01
5.94608843e-01 -5.48074007e-01 -2.52176791e-01 4.19879779e-02
9.66241002e-01 -3.82048547e-01 -5.46842337e-01 2.09571093e-01
7.15385854e-01 -6.48505509e-01 8.71555328e-01 -9.14404869e-01
1.79355219e-01 -1.84830979e-01 -4.10855681e-01 -8.19383323e-01
1.13515571e-01 -7.73930192e-01 -2.22512364e-01 7.40741134e-01
6.11295223e-01 -9.92767870e-01 1.30640197e+00 6.99301124e-01
1.00388564e-02 -5.64194441e-01 -1.14295852e+00 -7.93490589e-01
2.30386823e-01 -7.59926021e-01 6.23772323e-01 9.40200388e-01
-1.22484863e-02 3.78561802e-02 -1.61913395e-01 9.57267821e-01
6.92977309e-01 -6.33820295e-01 5.55799067e-01 -1.07211494e+00
-5.18359244e-01 -5.11399209e-01 -1.03982317e+00 -3.94592047e-01
4.51576531e-01 -6.78773344e-01 -1.00824401e-01 -2.56686807e-01
-3.35228652e-01 -2.20125273e-01 -4.60665286e-01 6.00088477e-01
4.75199372e-01 4.42361385e-01 1.87682644e-01 -2.18650624e-01
7.05427155e-02 -2.44905606e-01 3.67533088e-01 -2.16906980e-01
1.51837870e-01 1.28900737e-01 -5.74941337e-01 9.60591435e-01
1.03897834e+00 -8.11936200e-01 -6.47643268e-01 -3.74139309e-01
1.08800717e-01 -3.36755633e-01 3.29462677e-01 -1.29460979e+00
3.88994157e-01 4.74954098e-02 4.48727250e-01 -8.84475857e-02
1.23037748e-01 -1.37808955e+00 2.67542869e-01 1.28248906e+00
-4.83489871e-01 3.22688460e-01 5.15232563e-01 2.81692564e-01
3.10083896e-01 -4.06352073e-01 1.05551541e+00 2.45764613e-01
-5.16104221e-01 1.40632689e-01 -7.30778217e-01 -5.57674587e-01
1.13030612e+00 -2.22706243e-01 -7.79247642e-01 2.72792518e-01
-1.79031372e-01 -2.09293485e-01 5.40615618e-01 4.64565367e-01
7.21141100e-01 -8.77284527e-01 -9.03269202e-02 6.39401734e-01
-1.61169432e-02 -3.70470911e-01 -7.76169971e-02 2.02247366e-01
-8.30121934e-01 3.74119431e-01 -5.17844498e-01 -2.29244679e-01
-1.55852294e+00 8.26828122e-01 3.58766049e-01 8.12422261e-02
-2.82224745e-01 8.92106354e-01 7.75054768e-02 -9.32300538e-02
4.90098864e-01 -4.03182089e-01 1.09536320e-01 -5.31780541e-01
8.54814291e-01 -1.85780078e-02 3.04410487e-01 -2.34745592e-01
-4.11020458e-01 -9.41580907e-02 -4.46362048e-01 -1.19106092e-01
8.61514032e-01 6.00035787e-01 -3.85654420e-02 1.33229747e-01
1.38395727e+00 -1.18094243e-01 -9.08643723e-01 2.06194460e-01
3.36834133e-01 -2.37586096e-01 2.35041872e-01 -5.55804193e-01
-1.35456598e+00 8.24005246e-01 7.44338214e-01 3.71559590e-01
1.09319413e+00 -5.38901746e-01 7.83462822e-01 9.00786579e-01
8.46777976e-01 -4.13107276e-01 -1.61756381e-01 5.41899383e-01
2.97183115e-02 -6.50451541e-01 -2.26577058e-01 -3.67146134e-01
-1.82917863e-01 1.20093524e+00 5.66616118e-01 -5.97183764e-01
8.33323479e-01 1.15978360e+00 3.43282461e-01 1.16471096e-03
-8.41838896e-01 5.85183740e-01 -2.69852757e-01 6.87875092e-01
3.27267051e-02 -5.56883179e-02 2.28098571e-01 5.39069414e-01
-3.45348120e-01 -3.51804197e-01 7.69136488e-01 1.42150760e+00
-5.76556742e-01 -1.16378093e+00 -5.68150938e-01 3.33401173e-01
-9.28547978e-01 1.90416202e-02 -5.30722499e-01 8.88826251e-01
7.52060488e-02 8.34337711e-01 7.99662396e-02 -1.07879436e+00
7.45441169e-02 2.00998172e-01 3.06610107e-01 -5.32104790e-01
-9.46404040e-01 -3.41320217e-01 5.40694520e-02 -4.41352040e-01
2.89814979e-01 -3.39549214e-01 -1.33199596e+00 -5.46516836e-01
-4.13928926e-01 -6.87474534e-02 1.16449571e+00 9.31773841e-01
1.15085572e-01 6.01106048e-01 7.41340220e-01 -7.13847399e-01
-9.90449667e-01 -2.85074800e-01 -5.39643168e-01 -2.19151437e-01
3.59819949e-01 -4.64372337e-01 -5.64948857e-01 -1.46023005e-01] | [5.649139404296875, 7.620660781860352] |
efd3615d-73b8-49d5-a9ce-995d77b59cfa | a-preliminary-approach-for-learning | 2001.04432 | null | https://arxiv.org/abs/2001.04432v1 | https://arxiv.org/pdf/2001.04432v1.pdf | A Preliminary Approach for Learning Relational Policies for the Management of Critically Ill Children | The increased use of electronic health records has made possible the automated extraction of medical policies from patient records to aid in the development of clinical decision support systems. We adapted a boosted Statistical Relational Learning (SRL) framework to learn probabilistic rules from clinical hospital records for the management of physiologic parameters of children with severe cardiac or respiratory failure who were managed with extracorporeal membrane oxygenation. In this preliminary study, the results were promising. In particular, the algorithm returned logic rules for medical actions that are consistent with medical reasoning. | ['Neel Shah', 'Lakshmi Raman', 'Abdelaziz Farhat', 'Michael A. Skinner', 'Sriraam Natarajan'] | 2020-01-13 | null | null | null | null | ['respiratory-failure'] | ['medical'] | [ 6.69181123e-02 6.96077347e-01 -6.04178846e-01 -9.77226615e-01
-5.55140138e-01 -3.09404612e-01 1.14065461e-01 1.22658312e+00
-2.25531533e-01 9.33594644e-01 3.75957280e-01 -8.17283273e-01
-7.67451048e-01 -8.92648220e-01 -4.43959385e-01 -3.99335951e-01
-2.27081731e-01 1.00682092e+00 8.39749128e-02 4.07967985e-01
-8.57695267e-02 8.14790606e-01 -1.10228109e+00 8.71752203e-01
5.81735551e-01 7.08431184e-01 -4.33602482e-01 7.24671006e-01
-1.65999964e-01 1.82173336e+00 -1.06242642e-01 -3.14496130e-01
-2.16861367e-01 -1.77804098e-01 -6.48851156e-01 -4.62468892e-01
-1.79692343e-01 -3.09958249e-01 -3.88168871e-01 6.09782696e-01
3.31251770e-01 -7.66161904e-02 6.87244952e-01 -9.71525431e-01
-1.17519893e-01 9.90857005e-01 3.63505006e-01 -2.29307637e-02
5.08830547e-01 -1.44713700e-01 8.28971267e-01 -3.35870504e-01
4.84845549e-01 9.63111579e-01 5.30293107e-01 5.98365843e-01
-1.24641931e+00 -3.57520610e-01 -1.46144018e-01 2.84219831e-01
-1.52162445e+00 -4.28110331e-01 2.09680751e-01 -5.49933255e-01
1.11897981e+00 1.83662921e-01 5.48590720e-01 4.63129312e-01
5.53472042e-01 4.43573147e-01 8.94793630e-01 -5.35873771e-01
5.29645920e-01 4.33773756e-01 3.82712007e-01 8.65993798e-01
7.96637297e-01 3.07940096e-01 -4.65847999e-01 -8.35525870e-01
3.05469483e-01 1.17907241e-01 1.19122118e-01 -3.12491834e-01
-8.34823012e-01 6.78887665e-01 -7.92038813e-02 3.66641223e-01
-8.18625748e-01 -6.67475164e-03 3.38076323e-01 -1.47999063e-01
2.28815116e-02 2.60495424e-01 -8.36618185e-01 -1.93547517e-01
-5.04397511e-01 1.43274695e-01 1.08917987e+00 8.11460137e-01
8.47003981e-03 -5.03745914e-01 -9.24991965e-02 4.15335894e-01
6.97227478e-01 3.40356529e-01 1.17233694e-01 -1.00879145e+00
2.33831421e-01 7.11185038e-01 1.71363413e-01 -7.41089165e-01
-5.75913072e-01 4.00899023e-01 -5.81282675e-01 -1.98428378e-01
1.11542359e-01 -1.15975842e-01 -5.55064023e-01 1.21516669e+00
3.43184114e-01 1.08657040e-01 7.35903859e-01 1.81985691e-01
8.39990497e-01 1.31748512e-01 8.76333356e-01 -4.16964799e-01
1.11505091e+00 1.47208810e-01 -8.18979740e-01 4.27290499e-01
9.60594118e-01 -9.30847377e-02 -7.21848384e-02 7.28571892e-01
-9.92490292e-01 6.69630766e-02 -5.67436159e-01 4.87848252e-01
-2.07274333e-01 -2.55011052e-01 7.60881126e-01 6.44907534e-01
-7.58657694e-01 7.80625701e-01 -1.26184428e+00 -7.53074437e-02
6.15096927e-01 5.29315293e-01 -5.37098169e-01 -2.26588219e-01
-1.17723858e+00 1.19943190e+00 7.66130865e-01 -2.31775731e-01
-3.13166291e-01 -7.84128845e-01 -1.05459917e+00 9.16567519e-02
3.32147539e-01 -7.82496572e-01 1.23547089e+00 3.21833156e-02
-1.13693130e+00 9.01624382e-01 -2.13161498e-01 -6.49902105e-01
1.24281561e-02 -1.05463319e-01 -5.60621917e-01 3.71134013e-01
-3.46064687e-01 1.96713299e-01 3.59861064e-03 -6.47048354e-01
-8.42649162e-01 -5.29377937e-01 -4.55486804e-01 -2.73826152e-01
2.87710398e-01 3.23033154e-01 2.17651859e-01 -1.53627366e-01
2.19927609e-01 -5.90522349e-01 -6.15463853e-01 -5.45856595e-01
-4.96499017e-02 -3.64401460e-01 -6.63898215e-02 -5.06761432e-01
1.21968675e+00 -1.77346909e+00 -2.82501698e-01 5.71566164e-01
1.57040074e-01 1.18906856e-01 5.75120807e-01 2.38543972e-01
-4.12894398e-01 2.17781723e-01 -6.18291907e-02 4.07209992e-01
-1.83727920e-01 5.24939895e-01 -2.56841362e-01 9.07093808e-02
2.95654535e-01 8.45587432e-01 -9.57765937e-01 -1.03636479e+00
5.57742238e-01 4.85439479e-01 -4.59528714e-01 6.26778185e-01
-2.57549673e-01 3.41014087e-01 -7.31555760e-01 6.38403296e-01
2.88668573e-02 -2.28995219e-01 7.18539298e-01 1.02183342e-01
1.55491635e-01 6.43153608e-01 -7.33067393e-01 1.01900792e+00
1.24718182e-01 -1.66359380e-01 -1.86818108e-01 -6.25954866e-01
8.29670787e-01 8.71771336e-01 1.18398225e+00 -1.37506515e-01
2.86318183e-01 -1.60935298e-01 -1.94695249e-01 -1.12625873e+00
-4.16659534e-01 -7.75420547e-01 2.99023867e-01 2.49862060e-01
-2.67840058e-01 -4.06134695e-01 -2.48416841e-01 7.63429180e-02
1.39190614e+00 5.77222183e-02 1.09098506e+00 -1.00010596e-01
5.11639893e-01 2.58812010e-01 9.07542765e-01 5.74908018e-01
-1.64619014e-01 3.00390214e-01 4.52650368e-01 -5.78625619e-01
-3.83963943e-01 -1.13964891e+00 -5.67717850e-01 3.05619657e-01
-6.13598704e-01 -4.46395576e-01 -5.19180559e-02 -5.29850006e-01
3.10978085e-01 1.09067488e+00 -3.22227031e-01 -1.32142261e-01
-4.67309922e-01 -6.35494113e-01 4.65685159e-01 7.94937670e-01
-5.17701089e-01 -1.11702359e+00 -8.15573990e-01 6.78925455e-01
2.52016842e-01 -1.14571440e+00 3.27358067e-01 3.76147568e-01
-1.14586759e+00 -1.71189260e+00 1.77089468e-01 -3.77843648e-01
4.99411225e-01 -7.83877730e-01 8.66822183e-01 -7.67008886e-02
-6.28354192e-01 6.17941558e-01 -2.39352420e-01 -8.20692062e-01
-9.65758681e-01 -3.75121266e-01 2.88866669e-01 -5.28955519e-01
9.67550874e-01 -3.02755952e-01 -2.29072213e-01 -1.66504204e-01
-9.40158963e-01 -1.82623401e-01 3.84782076e-01 4.59873050e-01
7.68697977e-01 2.22537234e-01 5.53032279e-01 -1.37231731e+00
5.02392232e-01 -7.11944818e-01 -6.35049641e-01 6.21725082e-01
-1.35863018e+00 1.52743131e-01 3.37440968e-01 4.99056503e-02
-8.26912880e-01 2.53421307e-01 2.15988159e-02 -3.25419188e-01
-7.31934905e-01 7.15041697e-01 4.53180037e-02 5.50650954e-01
4.76140499e-01 -3.15343142e-01 -4.07981128e-02 -2.88434565e-01
-3.77562568e-02 8.60850394e-01 4.69527751e-01 -7.94247389e-01
1.77270442e-01 7.00126588e-02 3.68576914e-01 -1.99119180e-01
-6.39550269e-01 -3.63037944e-01 -5.12074471e-01 1.34000644e-01
1.09604609e+00 -5.63523293e-01 -1.17975664e+00 -9.18989852e-02
-7.91320801e-01 -1.18596114e-01 -4.14126575e-01 1.02032912e+00
-5.38019538e-01 7.87575468e-02 -5.09547889e-01 -1.00751305e+00
-3.39174181e-01 -7.99251139e-01 2.49535114e-01 1.06476014e-02
-7.01349080e-01 -1.15095568e+00 5.52424252e-01 2.72530317e-01
5.83598092e-02 4.90788341e-01 1.57554829e+00 -1.24506760e+00
-3.38966042e-01 -5.46346605e-01 1.17754489e-01 2.28744209e-01
5.92778683e-01 3.37403804e-01 -6.32522225e-01 1.91240475e-01
2.04407535e-02 -1.10783815e-01 1.85322072e-02 4.60991204e-01
1.18122399e+00 -3.55887085e-01 -4.51322615e-01 2.06750318e-01
1.20369148e+00 7.88592517e-01 2.39797726e-01 -7.45092556e-02
2.91344166e-01 9.98604596e-01 5.44426680e-01 6.42370880e-01
7.14106262e-01 1.29092231e-01 -1.87519804e-01 3.53665620e-01
4.86137450e-01 -2.67151505e-01 -1.26991451e-01 3.32578391e-01
-4.08141576e-02 1.29763752e-01 -1.62194705e+00 4.64282870e-01
-1.74854624e+00 -6.00363314e-01 8.05821344e-02 2.04573035e+00
1.13008332e+00 9.23606232e-02 -1.79257378e-01 -1.64559335e-02
5.77241778e-01 -5.52186847e-01 -4.20756251e-01 -5.87535381e-01
2.29605600e-01 5.82246244e-01 3.87657106e-01 4.87504482e-01
-7.14232862e-01 5.86389899e-01 7.69943428e+00 -2.98581719e-01
-5.95390320e-01 -2.84069240e-01 6.05586708e-01 -1.15575321e-01
-2.94250369e-01 -1.18219383e-01 -5.96820593e-01 1.41061679e-01
1.46062148e+00 -2.39667922e-01 1.73205197e-01 8.45256984e-01
3.13540280e-01 -1.80505812e-01 -1.44389153e+00 6.77176893e-01
-2.55806506e-01 -1.48718607e+00 -1.80384532e-01 -3.64934325e-01
3.25820535e-01 4.29312605e-03 -5.89719832e-01 -2.42988169e-02
7.23510861e-01 -1.24036777e+00 8.89662728e-02 9.39412892e-01
8.60607684e-01 -5.79423428e-01 9.84194100e-01 1.58839598e-01
-4.75564063e-01 -1.80312395e-01 2.64084674e-02 1.76143393e-01
4.54109721e-02 6.58240497e-01 -1.77141285e+00 4.58397359e-01
6.34214938e-01 5.32870352e-01 -6.27110805e-03 7.03844368e-01
-3.61880273e-01 8.39821815e-01 -1.91569373e-01 8.49103853e-02
-3.06916088e-01 4.94449399e-02 1.83924392e-01 9.66761827e-01
-2.23062128e-01 1.25519550e+00 -7.27749988e-02 5.80248654e-01
1.21357098e-01 3.10243696e-01 -9.45248544e-01 -2.06193104e-01
3.81057888e-01 7.59494007e-01 -3.20374548e-01 -5.18374681e-01
-3.52447867e-01 -3.80000800e-01 1.54388845e-01 1.05589800e-01
-4.51828063e-01 1.22702986e-01 4.29145753e-01 4.22214389e-01
1.47782594e-01 2.90850908e-01 -6.88967288e-01 -7.93613851e-01
-3.08383763e-01 -7.79704809e-01 1.00322306e+00 -6.02511764e-01
-1.07522905e+00 5.06806791e-01 4.81928974e-01 -6.99107528e-01
-7.30773509e-01 -3.91817182e-01 -1.84946984e-01 7.40609586e-01
-1.35871899e+00 -5.95678329e-01 2.07732812e-01 4.28763956e-01
-3.37888926e-01 -3.95698488e-01 1.30961275e+00 -7.97841400e-02
-4.44537938e-01 2.17614383e-01 -4.10004109e-01 1.15124986e-01
6.61448896e-01 -1.15658855e+00 -3.54244858e-01 3.32290649e-01
-2.82811314e-01 7.19944835e-01 6.36175990e-01 -1.05553484e+00
-1.21118164e+00 -1.12904906e+00 1.28006494e+00 -7.84229100e-01
4.39329356e-01 5.46908736e-01 -1.13994849e+00 7.27805853e-01
-2.40762442e-01 -2.69788690e-02 1.46431744e+00 -9.53702182e-02
-3.10134500e-01 -1.74982220e-01 -1.61899555e+00 -1.62531324e-02
3.44943315e-01 -4.51238066e-01 -1.25634408e+00 2.41258428e-01
4.94666040e-01 -2.66184628e-01 -1.64908969e+00 9.89515603e-01
3.65211129e-01 -2.20054641e-01 7.38888323e-01 -1.62245560e+00
2.85776347e-01 -2.78561741e-01 -2.36189410e-01 -6.23311639e-01
-2.23745480e-01 -7.06511557e-01 -4.26074624e-01 7.38574207e-01
4.48713899e-01 -8.46053183e-01 7.42568016e-01 1.92428911e+00
2.52957493e-01 -8.38020205e-01 -7.73884237e-01 7.50427917e-02
-1.21934585e-01 -4.89519984e-01 6.17026746e-01 9.31789696e-01
6.00663126e-01 -1.21277981e-01 2.19968900e-01 4.24424410e-01
5.53167284e-01 1.85743012e-02 -3.28524299e-02 -1.35190988e+00
-1.98863506e-01 -5.60828857e-03 -6.37295306e-01 2.12009817e-01
2.65715092e-01 -9.90677595e-01 5.33759296e-02 -1.71737206e+00
3.47602904e-01 -8.17420721e-01 -9.84927297e-01 9.99137461e-01
-2.00014889e-01 -6.35637462e-01 -3.82236511e-01 -2.56416589e-01
-2.96526164e-01 -5.28291240e-02 3.30498427e-01 1.85989141e-01
-3.87126535e-01 4.44505602e-01 -8.16164374e-01 7.15539753e-01
8.27934861e-01 -1.10807240e+00 -2.80578703e-01 1.06113128e-01
4.66773540e-01 8.77654910e-01 -1.33098468e-01 -4.54714596e-01
5.02538979e-01 -5.84239423e-01 2.52932757e-01 -7.22374380e-01
-2.96523452e-01 -1.10140741e+00 4.00082678e-01 7.72033572e-01
-7.53388643e-01 3.69346924e-02 2.58583724e-01 3.70185524e-01
-2.86577165e-01 -9.34559777e-02 6.96386158e-01 1.90439403e-01
-1.86769664e-01 1.19815581e-01 -8.77349138e-01 -1.32342175e-01
1.29687309e+00 2.36344114e-01 5.90047613e-03 -1.09945804e-01
-1.10191917e+00 2.05903560e-01 2.24539131e-01 2.57401288e-01
9.17776823e-01 -7.78964818e-01 -9.38337624e-01 1.96277693e-01
1.67731047e-01 2.72456735e-01 2.41642408e-02 7.90514529e-01
-8.15971017e-01 1.01437354e+00 2.86759511e-02 -3.52759361e-01
-1.32144368e+00 7.30047345e-01 2.73881197e-01 -1.97623879e-01
-6.59968913e-01 5.19385099e-01 -4.49507803e-01 -1.33853480e-01
5.05634069e-01 -6.06735587e-01 -3.09937745e-01 -3.70157093e-01
5.93720496e-01 1.93585172e-01 2.02592432e-01 -7.62710422e-02
-7.37813413e-01 -7.39944428e-02 -5.58915874e-03 7.34678805e-02
1.65857661e+00 3.99654657e-01 -4.33607489e-01 4.76956695e-01
5.85636020e-01 -1.10927597e-01 -4.49878514e-01 -1.88923612e-01
3.83867830e-01 -2.70547181e-01 2.83090416e-02 -1.01188385e+00
-6.51744246e-01 2.58697301e-01 3.44491214e-01 -1.10879451e-01
9.83276486e-01 2.18674138e-01 6.60898536e-02 8.13561141e-01
5.22591174e-01 -6.37272120e-01 -8.39733720e-01 1.71789363e-01
3.26894552e-01 -1.21875870e+00 3.93200248e-01 -4.16025221e-01
-7.57052124e-01 1.24874759e+00 2.26654291e-01 2.66893923e-01
1.01093650e+00 6.52357280e-01 2.20095828e-01 -3.60634208e-01
-1.36770654e+00 4.06766117e-01 2.07146376e-01 3.89383674e-01
6.03002667e-01 5.88898718e-01 -1.13889910e-01 7.54234672e-01
3.19190800e-01 6.46539032e-01 2.88127542e-01 1.23571026e+00
-2.76669323e-01 -1.37957561e+00 -4.26795334e-01 8.85324955e-01
-6.02143109e-01 -1.28331378e-01 -3.62721026e-01 3.76162201e-01
-6.00360855e-02 9.27051783e-01 -2.91107059e-01 -1.22119047e-01
5.21908700e-01 8.26669157e-01 5.28581262e-01 -8.90647173e-01
-6.28793478e-01 -3.47189844e-01 3.79257202e-01 -7.07985997e-01
-4.77252632e-01 -9.72999811e-01 -1.68711984e+00 2.23943725e-01
1.29174903e-01 4.50929582e-01 6.21190667e-01 7.91718841e-01
4.42019939e-01 5.81729949e-01 4.10440326e-01 6.13882601e-01
-5.05283594e-01 -6.71771824e-01 -4.09258127e-01 1.36332542e-01
2.68802941e-01 -2.56160259e-01 -9.95189101e-02 2.50154018e-01] | [8.419625282287598, 8.589703559875488] |
21fee6df-c079-4003-aaf6-c43b0ebe8b1c | temporal-stability-in-predictive-process | 1712.04165 | null | http://arxiv.org/abs/1712.04165v3 | http://arxiv.org/pdf/1712.04165v3.pdf | Temporal Stability in Predictive Process Monitoring | Predictive process monitoring is concerned with the analysis of events
produced during the execution of a business process in order to predict as
early as possible the final outcome of an ongoing case. Traditionally,
predictive process monitoring methods are optimized with respect to accuracy.
However, in environments where users make decisions and take actions in
response to the predictions they receive, it is equally important to optimize
the stability of the successive predictions made for each case. To this end,
this paper defines a notion of temporal stability for binary classification
tasks in predictive process monitoring and evaluates existing methods with
respect to both temporal stability and accuracy. We find that methods based on
XGBoost and LSTM neural networks exhibit the highest temporal stability. We
then show that temporal stability can be enhanced by hyperparameter-optimizing
random forests and XGBoost classifiers with respect to inter-run stability.
Finally, we show that time series smoothing techniques can further enhance
temporal stability at the expense of slightly lower accuracy. | ['Fabrizio Maria Maggi', 'Anna Leontjeva', 'Irene Teinemaa', 'Marlon Dumas'] | 2017-12-12 | null | null | null | null | ['predictive-process-monitoring'] | ['time-series'] | [ 5.20492852e-01 -9.27241985e-03 -2.55004108e-01 -4.34739769e-01
-3.61060500e-01 -3.00075024e-01 8.60350847e-01 8.59502554e-01
-2.89523870e-01 5.86829185e-01 2.15463303e-02 -5.54414392e-01
-5.96721649e-01 -1.07271552e+00 -2.04753861e-01 -4.22609031e-01
-3.02076638e-01 4.41073269e-01 1.22239411e-01 2.66808599e-01
3.57764989e-01 7.56139040e-01 -1.30383456e+00 4.54974920e-01
6.97314084e-01 1.55636191e+00 -2.59533882e-01 8.79269421e-01
-2.34629750e-01 1.19585574e+00 -4.92212683e-01 -1.86557814e-01
1.26422271e-01 -2.58875728e-01 -7.46410072e-01 2.76091974e-02
-3.14436644e-01 5.47737144e-02 1.30847305e-01 5.38749337e-01
-9.78157297e-03 3.96319449e-01 5.56013286e-01 -1.27134609e+00
-1.09712183e-01 5.01791239e-01 -2.04333752e-01 3.75866234e-01
2.87709743e-01 6.08058199e-02 8.14769626e-01 -3.21254075e-01
3.44314992e-01 8.06421161e-01 8.00675988e-01 7.87673593e-02
-1.49171460e+00 -2.74504453e-01 4.88354892e-01 1.51246876e-01
-7.26335645e-01 -5.18847883e-01 4.16207045e-01 -4.73346412e-01
1.16012132e+00 3.16079915e-01 5.64792216e-01 6.90072119e-01
7.35457838e-01 5.73092520e-01 7.87232101e-01 -5.06699324e-01
5.25277615e-01 5.08552901e-02 3.75534475e-01 3.30911905e-01
9.65449121e-03 3.55342217e-02 -6.02324963e-01 -3.78009290e-01
4.80014622e-01 5.06691217e-01 -9.27717835e-02 -1.31513521e-01
-1.01266921e+00 6.90354109e-01 -1.74879320e-02 4.28803205e-01
-9.04234886e-01 3.08756400e-02 4.32358027e-01 5.83604515e-01
6.47962034e-01 6.82074189e-01 -5.57155013e-01 -3.67673188e-01
-1.12288249e+00 1.51508510e-01 1.13391256e+00 7.97193348e-01
3.07490915e-01 1.33952364e-01 -4.73526776e-01 4.67129022e-01
-3.74768768e-03 -1.49939135e-01 6.81418300e-01 -9.42376673e-01
2.27727011e-01 5.18224657e-01 4.43284929e-01 -6.29766047e-01
-3.95285070e-01 -5.65128088e-01 -6.92640603e-01 2.64380842e-01
4.49081987e-01 -1.68541253e-01 -6.26974523e-01 1.36922109e+00
-7.67074432e-03 -3.04200407e-02 -1.01198405e-01 6.43127114e-02
-2.05648005e-01 6.66060448e-01 5.37479162e-01 -9.61003721e-01
1.04453647e+00 -7.79622376e-01 -8.74175072e-01 -7.65395686e-02
2.99592227e-01 -5.19164622e-01 6.38898730e-01 5.86030006e-01
-1.17648792e+00 -4.42884654e-01 -6.85562074e-01 5.81541419e-01
-1.40444189e-01 -1.20119773e-01 4.54942018e-01 5.65841377e-01
-7.40837634e-01 1.38818681e+00 -1.35953665e+00 -3.16030830e-01
2.59176195e-01 3.24123591e-01 1.30062565e-01 5.86875796e-01
-8.67947638e-01 8.83521318e-01 3.04119378e-01 8.90249237e-02
-4.26308125e-01 -6.01537645e-01 -2.48588562e-01 4.82747406e-01
3.43019426e-01 -4.19342756e-01 1.81449997e+00 -1.00709522e+00
-1.53379571e+00 -1.29969930e-02 -5.23662031e-01 -9.30608988e-01
8.96311939e-01 -3.19312155e-01 -6.09930336e-01 -1.90645859e-01
-2.25013182e-01 -2.06106156e-01 7.19094098e-01 -6.75284505e-01
-1.30189788e+00 -4.72405255e-01 -3.64413142e-01 -8.55140686e-02
-1.84246555e-01 1.15353122e-01 9.19870585e-02 -3.68031025e-01
3.02979946e-01 -7.25706697e-01 -5.80119848e-01 -5.95882773e-01
-9.29391459e-02 -1.25526816e-01 6.25543892e-01 -8.03407729e-01
1.36531544e+00 -1.71559882e+00 -4.43464905e-01 4.60747898e-01
7.55358040e-02 -1.01382382e-01 4.69144791e-01 4.37613994e-01
-1.05036601e-01 2.94755787e-01 1.63457274e-01 -2.16437221e-01
-8.84689763e-02 -1.48807928e-01 -5.95381320e-01 2.73893237e-01
2.70432442e-01 6.05602086e-01 -5.20689607e-01 -3.09687406e-01
1.98997006e-01 4.26450111e-02 1.77109659e-01 1.70349926e-01
-2.84186542e-01 3.24994683e-01 -5.06801248e-01 5.72707653e-01
-1.23026319e-01 -5.32092571e-01 3.17906260e-01 3.33577037e-01
-2.42119685e-01 3.76834393e-01 -1.03593731e+00 6.76979423e-01
-5.95068872e-01 5.45518637e-01 -2.47946039e-01 -8.88817549e-01
1.01097357e+00 6.48741364e-01 8.47050309e-01 -6.54139996e-01
-5.75758964e-02 7.44399205e-02 -1.66480601e-01 -4.33021113e-02
6.56728685e-01 -3.55287552e-01 2.50385016e-01 6.56098843e-01
-4.72602814e-01 4.43521410e-01 2.48812720e-01 -3.64474714e-01
1.29325426e+00 1.87397271e-01 6.37019515e-01 -3.84050719e-02
2.97646105e-01 6.88789263e-02 5.81220746e-01 8.08237076e-01
-2.72910058e-01 2.20028907e-02 6.09730065e-01 -6.13542259e-01
-8.81920218e-01 -8.79027605e-01 -3.48582119e-02 1.31473112e+00
-2.19306022e-01 -1.63374841e-01 -1.96154237e-01 -4.54869300e-01
-1.87108610e-02 1.20796478e+00 -5.83838880e-01 4.55194637e-02
-5.45624435e-01 -8.52564394e-01 6.23797104e-02 7.52108216e-01
2.11340532e-01 -1.01705122e+00 -9.40258861e-01 9.16930497e-01
3.40699732e-01 -7.79645622e-01 -2.04247814e-02 5.91189802e-01
-1.45198596e+00 -7.48907208e-01 -2.54397720e-01 -2.01936916e-01
2.32370332e-01 -1.46680743e-01 1.02011049e+00 -5.12575209e-01
3.57948989e-01 1.16900139e-01 -1.82639375e-01 -7.49581337e-01
-5.77746987e-01 2.38076016e-01 -8.36125314e-02 4.47710492e-02
4.22802567e-01 -7.03844845e-01 -1.86092511e-01 2.38449231e-01
-5.12686849e-01 1.12353712e-02 3.06555927e-01 4.71877247e-01
7.10223019e-01 3.62733454e-01 5.31590641e-01 -8.25078905e-01
9.88967955e-01 -2.82060385e-01 -5.78731954e-01 5.80726981e-01
-1.41180933e+00 2.64251560e-01 7.35248685e-01 -6.05104804e-01
-1.31150424e+00 3.63378637e-02 4.12175268e-01 -2.47509003e-01
-4.74145971e-02 6.73643172e-01 3.22175741e-01 4.10692096e-01
5.62871873e-01 3.44148010e-01 2.43522748e-02 -2.68917859e-01
-5.04656173e-02 3.33331883e-01 2.52523303e-01 -3.83732617e-01
2.41540417e-01 3.83500189e-01 7.85290971e-02 -3.44543934e-01
-4.85051632e-01 -2.38062546e-01 -3.97174358e-01 -4.95504379e-01
3.98357749e-01 -5.11635840e-01 -9.95542109e-01 1.24043159e-01
-8.13858926e-01 -2.99826473e-01 -5.01137316e-01 4.06893253e-01
-8.75716329e-01 -1.48249492e-01 -8.97201121e-01 -1.21562207e+00
-5.05228519e-01 -7.96102524e-01 5.89786053e-01 2.21023142e-01
-8.45631242e-01 -1.02249050e+00 -2.30550736e-01 -8.49327818e-02
5.27811527e-01 3.21901113e-01 8.35622966e-01 -1.16998231e+00
-4.94784534e-01 -6.86488211e-01 1.62466809e-01 3.99919786e-02
3.08396190e-01 3.05640131e-01 -7.38336802e-01 -1.29634053e-01
4.07900661e-01 5.45275748e-01 6.57532156e-01 6.90572560e-01
1.07238054e+00 -5.29593468e-01 -4.91738707e-01 1.91167474e-01
1.21380198e+00 9.63410258e-01 4.17636693e-01 7.67664969e-01
1.44377559e-01 8.46570671e-01 7.59465635e-01 6.50267720e-01
-1.48363322e-01 3.09110910e-01 -8.24085698e-02 2.73219705e-01
6.67594552e-01 -9.08100456e-02 3.64919513e-01 2.79334366e-01
-5.11160791e-01 -1.55410662e-01 -1.04804134e+00 4.52643663e-01
-2.15041494e+00 -1.35678184e+00 -2.90283054e-01 2.49263096e+00
4.41758424e-01 6.77512050e-01 4.26981717e-01 5.01426876e-01
8.43657792e-01 1.55982282e-02 -4.38448101e-01 -5.79745114e-01
1.77294105e-01 6.86039627e-02 7.76483119e-01 3.95019770e-01
-1.15784609e+00 3.45489174e-01 6.72351599e+00 4.76884723e-01
-1.18305659e+00 4.98239361e-02 1.08590388e+00 -5.15715599e-01
1.04085639e-01 1.02095269e-01 -7.95261264e-01 6.39926851e-01
1.55754101e+00 -5.15859365e-01 2.31863692e-01 1.06903636e+00
7.72608221e-01 -1.95720896e-01 -1.28721380e+00 3.92575204e-01
-6.33567333e-01 -1.36774135e+00 -3.81402224e-01 1.97966680e-01
6.37202263e-01 -2.94403970e-01 -1.77526519e-01 4.05152626e-02
5.73614419e-01 -1.07244563e+00 8.20698738e-01 1.00312519e+00
-3.92914712e-02 -9.16512787e-01 6.23641431e-01 3.52766037e-01
-1.15299261e+00 -4.65100050e-01 2.05310240e-01 -3.74388486e-01
4.64540094e-01 7.89102077e-01 -1.05045652e+00 4.54957902e-01
5.50587893e-01 4.59382921e-01 -2.09545925e-01 8.95584762e-01
-4.21612822e-02 8.35766196e-01 -2.51252115e-01 -1.17182493e-01
5.75458072e-02 -2.55400300e-01 3.07521969e-01 9.61509764e-01
5.16487896e-01 -3.43258083e-01 1.42223254e-01 6.81389391e-01
5.29416323e-01 1.67262390e-01 -2.67340004e-01 -9.06527862e-02
3.99716020e-01 7.58408010e-01 -9.04888630e-01 -5.29330373e-01
-1.44094124e-01 5.79533756e-01 3.38901803e-02 4.62159663e-01
-5.08096099e-01 -1.12949692e-01 4.61370170e-01 4.82534170e-01
3.12986702e-01 -2.89510101e-01 -1.06998014e+00 -6.22320712e-01
1.39211953e-01 -6.59759998e-01 5.41660726e-01 -4.04950827e-01
-1.19600391e+00 5.46706736e-01 -4.24667835e-01 -1.22128510e+00
-7.33270705e-01 -2.43442237e-01 -7.52793550e-01 9.62748230e-01
-1.00966370e+00 -7.54884183e-01 -8.37190077e-02 -3.35997134e-03
5.15286982e-01 -6.81417286e-02 7.20775723e-01 -3.71188968e-01
-5.71125805e-01 -1.82102904e-01 4.42631364e-01 -3.17593485e-01
4.38879550e-01 -1.25782287e+00 3.74577343e-01 7.98088789e-01
-2.16966867e-01 5.64715862e-01 9.90074992e-01 -9.03396189e-01
-8.68889570e-01 -1.17737246e+00 1.20935225e+00 -3.10803056e-01
8.38247478e-01 2.77753770e-01 -1.12331116e+00 9.27258313e-01
-1.02995515e-01 -5.05344033e-01 6.61372006e-01 4.97962356e-01
2.37773106e-01 -5.20494998e-01 -1.02147532e+00 4.53900844e-01
5.90355754e-01 -3.83203417e-01 -5.08839846e-01 8.74233767e-02
3.58313143e-01 3.19041759e-02 -1.41353559e+00 3.81623030e-01
6.33915901e-01 -8.80603135e-01 5.43728530e-01 -7.22900391e-01
3.76627803e-01 -8.02363232e-02 2.02347282e-02 -1.12119246e+00
-5.38081348e-01 -8.11632574e-01 -3.98370236e-01 1.02635050e+00
5.29515266e-01 -6.86749339e-01 8.94192100e-01 1.31019378e+00
3.27821434e-01 -6.71318650e-01 -7.51542330e-01 -8.66695344e-01
-3.26733530e-01 -6.88200772e-01 9.29867864e-01 7.76698351e-01
1.78270817e-01 2.55735517e-01 -2.73933768e-01 -6.00275472e-02
3.80177617e-01 4.15815890e-01 4.44191664e-01 -1.49108481e+00
-1.99390516e-01 -6.54858053e-01 -2.03284323e-01 -6.16062462e-01
-2.26101622e-01 -2.73147881e-01 5.24404533e-02 -1.50603426e+00
-1.43195063e-01 -4.18982863e-01 -7.62206733e-01 4.48203772e-01
-2.46611372e-01 -4.82558489e-01 9.35840905e-02 6.20154023e-01
-4.15659428e-01 2.98264921e-01 7.27580249e-01 1.01475738e-01
-8.11403632e-01 7.49659777e-01 -4.99597549e-01 5.23157775e-01
1.06264699e+00 -3.38170141e-01 -2.57977396e-01 7.89553151e-02
-8.72457214e-03 5.52590370e-01 1.69772521e-01 -9.80958998e-01
4.39603180e-01 -5.81946850e-01 6.81289732e-01 -4.05430794e-01
9.39464644e-02 -6.34168983e-01 7.03735590e-01 6.65515661e-01
-8.21295619e-01 3.69394004e-01 -7.43146092e-02 8.24872851e-01
-2.57045060e-01 -1.64039031e-01 6.13749266e-01 -3.54520045e-02
-4.91988987e-01 1.48950204e-01 -8.43422771e-01 -6.68953836e-01
1.12152088e+00 -4.64617461e-01 -9.88250375e-02 -6.36420131e-01
-1.00838065e+00 1.63274303e-01 2.30877787e-01 6.53070882e-02
2.63913553e-02 -8.95030856e-01 -4.51842219e-01 -5.52455150e-02
-1.78383902e-01 -3.10548097e-01 1.42264338e-02 1.03889775e+00
-1.61882803e-01 5.47404170e-01 -7.18408376e-02 -4.16534573e-01
-1.29310691e+00 5.38869739e-01 3.44221562e-01 -7.99363256e-01
-3.35370481e-01 3.52258831e-01 -4.35499161e-01 3.24493289e-01
1.75243497e-01 -4.98298436e-01 -2.27313891e-01 3.49237233e-01
5.55802822e-01 6.41624391e-01 4.03115660e-01 -3.61128896e-02
-9.86442938e-02 -2.07753927e-01 -1.57725200e-01 -2.95018703e-01
1.59439170e+00 2.89114043e-02 -7.61179551e-02 7.59981751e-01
5.31155288e-01 -1.79854825e-01 -1.40721393e+00 -2.27973297e-01
8.69681537e-01 -3.99026304e-01 7.82963037e-02 -7.58853793e-01
-7.71455705e-01 4.50177789e-01 4.58260149e-01 8.39529991e-01
1.28202033e+00 -3.53739738e-01 3.84376556e-01 3.01452756e-01
9.17688608e-02 -1.35701287e+00 -3.70802701e-01 3.72943461e-01
4.30759221e-01 -7.43954957e-01 9.19900183e-03 -2.32164130e-01
-7.06649542e-01 1.19713223e+00 2.23139882e-01 9.17673036e-02
6.49906158e-01 2.86522985e-01 -2.21872985e-01 1.47406340e-01
-1.24570465e+00 1.93764374e-01 -3.02627143e-02 3.85592669e-01
5.78231096e-01 4.02711362e-01 -1.89437285e-01 5.57808399e-01
-2.02600107e-01 2.64755726e-01 2.87439376e-01 1.17629766e+00
-5.11806726e-01 -9.66574073e-01 -4.93230999e-01 9.89555657e-01
-6.97503448e-01 1.11032166e-01 -1.84821233e-01 5.55864275e-01
-3.69959593e-01 1.17020380e+00 3.51044148e-01 -2.35597193e-01
5.35444081e-01 6.09056354e-01 2.20604632e-02 -2.36972108e-01
-7.82675266e-01 2.95271099e-01 5.31696022e-01 -4.14315641e-01
-2.34540954e-01 -1.12261438e+00 -9.36144292e-01 -4.10740644e-01
-2.09914669e-01 1.49840042e-01 6.97491527e-01 1.05256641e+00
4.09367234e-01 9.12684798e-01 6.07322395e-01 -3.62380445e-01
-8.61510515e-01 -1.06178343e+00 -5.35271227e-01 2.09807083e-01
2.69661820e-03 -2.99953729e-01 -2.46982779e-02 3.23649496e-01] | [8.531810760498047, 5.996852397918701] |
1b34a267-4023-4180-a23c-38bbbbc8ca47 | context-aware-proposal-network-for-temporal | 2206.09082 | null | https://arxiv.org/abs/2206.09082v1 | https://arxiv.org/pdf/2206.09082v1.pdf | Context-aware Proposal Network for Temporal Action Detection | This technical report presents our first place winning solution for temporal action detection task in CVPR-2022 AcitivityNet Challenge. The task aims to localize temporal boundaries of action instances with specific classes in long untrimmed videos. Recent mainstream attempts are based on dense boundary matchings and enumerate all possible combinations to produce proposals. We argue that the generated proposals contain rich contextual information, which may benefits detection confidence prediction. To this end, our method mainly consists of the following three steps: 1) action classification and feature extraction by Slowfast, CSN, TimeSformer, TSP, I3D-flow, VGGish-audio, TPN and ViViT; 2) proposal generation. Our proposed Context-aware Proposal Network (CPN) builds on top of BMN, GTAD and PRN to aggregate contextual information by randomly masking some proposal features. 3) action detection. The final detection prediction is calculated by assigning the proposals with corresponding video-level classifcation results. Finally, we ensemble the results under different feature combination settings and achieve 45.8% performance on the test set, which improves the champion result in CVPR-2021 ActivityNet Challenge by 1.1% in terms of average mAP. | ['Nong Sang', 'Yuanjie Shao', 'Changxin Gao', 'Shiwei Zhang', 'Huaxin Zhang', 'Xiang Wang'] | 2022-06-18 | null | null | null | null | ['action-classification'] | ['computer-vision'] | [ 5.09142160e-01 4.08182181e-02 -5.16459525e-01 -3.16474773e-02
-1.07760096e+00 -3.22739065e-01 7.71476269e-01 -1.99407876e-01
-4.57382351e-01 6.37306392e-01 5.62393486e-01 2.21542135e-01
-8.29901844e-02 -3.98048997e-01 -3.91420364e-01 -5.49025595e-01
-4.62746680e-01 1.69462770e-01 1.01584673e+00 -4.33658808e-03
3.10477704e-01 2.49024659e-01 -1.65763652e+00 1.06761777e+00
4.85675454e-01 1.61154795e+00 -7.41929188e-02 8.03686321e-01
3.31792384e-01 1.19159234e+00 -5.77762604e-01 -7.04087168e-02
4.28519875e-01 -4.76070017e-01 -8.01700294e-01 2.34219655e-02
6.06886804e-01 -4.02595997e-01 -4.33607668e-01 6.97952092e-01
5.15777230e-01 3.38708341e-01 6.49760187e-01 -1.53955424e+00
3.42102528e-01 8.19447339e-01 -6.72643423e-01 8.26831102e-01
5.16765654e-01 5.50730586e-01 1.19103467e+00 -1.16016531e+00
8.69998813e-01 1.19556618e+00 5.46264768e-01 4.91096765e-01
-9.79716539e-01 -5.94430923e-01 4.42779183e-01 9.10276473e-01
-1.51201117e+00 -5.42608619e-01 4.87022668e-01 -4.27806556e-01
1.11847711e+00 3.80860120e-01 8.53204906e-01 1.43775737e+00
3.48893516e-02 1.23618698e+00 7.43583143e-01 3.20201404e-02
2.96667010e-01 -2.97004104e-01 -1.77637979e-01 2.82243043e-01
-4.00786787e-01 6.81287944e-02 -8.77466798e-01 3.99071872e-02
5.42074680e-01 -2.31508031e-01 -3.74318212e-01 7.39881322e-02
-1.41071820e+00 4.72240269e-01 3.47361714e-01 3.69116902e-01
-6.73785865e-01 3.24456245e-01 5.89886487e-01 7.44784623e-02
3.99722338e-01 3.22976768e-01 -4.80546504e-01 -5.28251052e-01
-1.22022593e+00 3.28232110e-01 2.84252316e-01 8.92785013e-01
2.98253894e-01 -2.01598540e-01 -1.02644873e+00 7.03437150e-01
2.42834240e-02 -2.67277993e-02 3.61900687e-01 -1.25730217e+00
7.70293891e-01 5.08013010e-01 1.55948222e-01 -1.07308805e+00
-4.13551837e-01 -5.14440000e-01 -5.05251110e-01 -6.61417656e-03
1.41155288e-01 -1.20325133e-01 -8.36721241e-01 1.39979279e+00
3.62818480e-01 8.49311948e-01 -3.06138217e-01 1.06109297e+00
7.72722483e-01 5.96534133e-01 4.58153397e-01 -1.97796702e-01
1.20704722e+00 -1.26268291e+00 -4.67662454e-01 -2.09598303e-01
7.38652408e-01 -5.94597876e-01 4.60175246e-01 6.87035561e-01
-9.58830535e-01 -9.62646246e-01 -8.70981872e-01 4.91805524e-01
9.16064754e-02 4.97916043e-01 3.66383761e-01 1.96568817e-01
-9.71915603e-01 7.36410439e-01 -6.10880911e-01 -4.24582988e-01
7.40627050e-01 1.43293902e-01 -3.52257639e-01 5.73081113e-02
-1.12562966e+00 5.88670492e-01 7.75008738e-01 2.31088072e-01
-1.40573359e+00 -6.69800043e-01 -5.43125331e-01 -1.74221203e-01
8.97743464e-01 -3.56027991e-01 1.14190662e+00 -9.95967507e-01
-1.12098217e+00 4.78307933e-01 1.23085856e-01 -1.00067723e+00
9.03871715e-01 -3.06759030e-01 -6.43148899e-01 6.10919356e-01
2.52108663e-01 1.27897382e+00 9.48733032e-01 -6.05046511e-01
-1.48601615e+00 -9.84970760e-03 -5.68833016e-02 4.69570845e-01
1.09226964e-01 2.29592413e-01 -6.29905343e-01 -6.16075575e-01
1.86472520e-01 -6.90752506e-01 -2.42537230e-01 -7.95233324e-02
-3.84925425e-01 -5.06901681e-01 7.75128007e-01 -7.02914953e-01
1.49403489e+00 -2.22654724e+00 8.22802782e-02 1.55848667e-01
2.87939996e-01 1.48661479e-01 -3.83887529e-01 2.36037627e-01
-5.13635576e-02 -9.95657295e-02 3.35877776e-01 -2.04127327e-01
-5.07047027e-02 -2.46929124e-01 -3.21175694e-01 5.37818611e-01
3.30767035e-01 8.31755877e-01 -9.46152151e-01 -8.90740931e-01
4.37874734e-01 1.66387528e-01 -6.86100364e-01 -8.34904686e-02
-2.70161182e-01 4.03412491e-01 -4.90796655e-01 7.96218693e-01
2.53836602e-01 -5.01190312e-03 1.52094230e-01 -5.40042937e-01
-1.39524013e-01 4.35562938e-01 -1.52320170e+00 1.76324272e+00
1.65434107e-01 6.32425666e-01 -1.69070780e-01 -1.07426190e+00
7.25897551e-01 5.17425299e-01 9.25735295e-01 -6.04015470e-01
3.63859013e-02 8.13707560e-02 4.10893820e-02 -4.61391658e-01
4.90078658e-01 4.80477184e-01 -1.00870110e-01 -3.15703116e-02
2.90698141e-01 4.47236985e-01 4.15602893e-01 3.15158159e-01
1.66999114e+00 5.88074863e-01 1.88631877e-01 6.46981299e-02
6.84287488e-01 -1.44848404e-02 8.79263639e-01 9.85735238e-01
-8.20642412e-01 5.62845945e-01 6.83153391e-01 -5.27190983e-01
-6.19698107e-01 -9.27656889e-01 1.77627414e-01 1.28904784e+00
1.30260065e-01 -6.41207099e-01 -5.97765207e-01 -1.06644166e+00
-3.56020212e-01 5.96681654e-01 -6.90067887e-01 -8.72923136e-02
-8.01408887e-01 -3.24724793e-01 6.25019073e-01 6.33035064e-01
7.76226699e-01 -1.44052756e+00 -8.11932564e-01 4.29779053e-01
-6.89840317e-01 -1.33495247e+00 -5.02022505e-01 7.40680173e-02
-6.96917415e-01 -1.19798207e+00 -5.10800719e-01 -3.60265821e-01
5.84823824e-02 1.39370650e-01 9.48594153e-01 -1.89382404e-01
-4.56286460e-01 4.36655104e-01 -7.20499516e-01 1.20126978e-01
-1.42126799e-01 1.72524694e-02 6.76680207e-02 3.83782655e-01
4.95019823e-01 -5.46256423e-01 -1.03521216e+00 6.59670055e-01
-1.33653700e-01 3.08987801e-03 6.13252699e-01 3.87753397e-01
7.46204615e-01 8.76205415e-02 3.88026774e-01 -1.54640287e-01
1.04169838e-01 -4.64141041e-01 -2.38693953e-01 1.21127427e-01
-1.95245042e-01 -3.63600016e-01 5.34289852e-02 -5.75313687e-01
-8.69362473e-01 3.71088922e-01 6.57338202e-02 -6.66326463e-01
-3.78903359e-01 5.81652559e-02 5.10344282e-02 2.04299718e-01
9.23443019e-01 4.19319540e-01 -3.44780058e-01 -2.38213599e-01
2.74569184e-01 3.68383706e-01 6.26979113e-01 -3.25731516e-01
3.18828464e-01 3.55080903e-01 -1.93725273e-01 -5.25664032e-01
-7.75523961e-01 -7.20708668e-01 -6.16608262e-01 -8.94073904e-01
1.12664843e+00 -1.05776799e+00 -6.76385820e-01 2.20048040e-01
-1.07135773e+00 -2.82241136e-01 -3.59451681e-01 6.78285778e-01
-7.72027671e-01 3.40492696e-01 -4.28147167e-01 -8.72420311e-01
-3.46255660e-01 -8.95224452e-01 1.13070571e+00 -4.66819108e-02
-5.41035593e-01 -1.54755756e-01 -4.25375737e-02 5.38013995e-01
1.54920429e-01 2.26990506e-01 7.81909823e-02 -7.60226607e-01
-7.09663332e-01 -9.85921174e-02 -1.76287696e-01 3.01830024e-01
-2.02435598e-01 1.01994447e-01 -9.08979774e-01 -1.19937873e-02
-4.82972413e-01 -3.51379514e-01 1.27475750e+00 7.52071381e-01
1.25514460e+00 -2.39512593e-01 -6.67043805e-01 2.80339926e-01
9.70358729e-01 2.66818970e-01 7.72975385e-01 2.27521911e-01
2.98756480e-01 4.47164088e-01 1.21631908e+00 7.77419150e-01
-1.13887750e-01 1.02783060e+00 8.57740939e-01 2.22491130e-01
-3.49743158e-01 -1.69336274e-01 6.84178293e-01 7.57682025e-02
-4.27231997e-01 -2.86669523e-01 -7.12535203e-01 6.93432689e-01
-2.25674129e+00 -1.49091804e+00 -1.16037391e-01 1.91893399e+00
4.06363338e-01 5.65661907e-01 6.72134757e-01 2.94115186e-01
8.36807549e-01 3.61771852e-01 -3.18538636e-01 1.70613453e-01
-5.32916859e-02 -2.35474911e-02 3.40767980e-01 6.78272918e-02
-1.67577887e+00 9.80398238e-01 5.75439548e+00 1.29131985e+00
-7.47885942e-01 2.14171350e-01 6.74381554e-01 -4.08637226e-01
5.11295199e-01 -2.20336080e-01 -9.93391693e-01 5.10929406e-01
7.56703913e-01 2.38655880e-01 -1.38963638e-02 9.07148600e-01
5.79880714e-01 -3.21914792e-01 -1.10766602e+00 1.00464332e+00
-8.17107856e-02 -1.50989163e+00 -4.25589010e-02 -8.95220190e-02
4.57720101e-01 1.83675289e-01 -2.14372292e-01 6.62050366e-01
1.01995006e-01 -7.80373156e-01 9.96034026e-01 5.05168140e-01
5.41449904e-01 -6.13302588e-01 5.27323306e-01 2.74178207e-01
-1.60702372e+00 -5.88645816e-01 -1.50569439e-01 1.17498703e-01
3.89359206e-01 3.19940776e-01 -9.40845370e-01 4.96944875e-01
8.30310166e-01 1.19885278e+00 -4.32104915e-01 1.52318525e+00
-3.34144235e-01 8.98729563e-01 -3.51436138e-01 1.13046840e-01
4.71389472e-01 3.17838758e-01 7.24856675e-01 1.31778932e+00
3.71009797e-01 1.04425870e-01 5.13456821e-01 5.40908575e-01
2.22104371e-01 6.79049641e-02 -2.44093090e-01 2.88526237e-01
3.80129814e-01 1.34279537e+00 -8.92045498e-01 -5.36391437e-01
-1.54444009e-01 9.64118421e-01 -6.98583797e-02 1.66007325e-01
-1.43048620e+00 3.19438912e-02 5.43666244e-01 3.04820299e-01
6.30716085e-01 1.78011522e-01 2.59835631e-01 -8.94314468e-01
7.16543384e-03 -8.35695803e-01 9.01024282e-01 -7.55307019e-01
-8.62472236e-01 5.79458714e-01 1.11525908e-01 -1.78234541e+00
-1.48892328e-01 -2.68331677e-01 -6.57565951e-01 3.81863177e-01
-8.64425957e-01 -8.06734383e-01 -3.91840488e-01 7.33262599e-01
1.04905140e+00 -1.29902318e-01 3.95681173e-01 5.80824375e-01
-5.82999706e-01 5.00546396e-01 -6.93640769e-01 2.11709753e-01
5.89964092e-01 -8.85945618e-01 2.12023228e-01 9.81698871e-01
3.26837331e-01 -2.70559609e-01 6.56172335e-01 -8.42500389e-01
-7.26827383e-01 -1.40503931e+00 7.87756443e-01 -4.55468923e-01
7.14725554e-01 -1.58702150e-01 -3.84166092e-01 4.52043921e-01
-6.46750778e-02 3.81971836e-01 1.77128702e-01 -2.60938108e-01
-1.32795483e-01 -2.30434403e-01 -9.64956403e-01 5.74395239e-01
1.49633205e+00 -9.78562385e-02 -4.54637080e-01 4.53321695e-01
6.14415646e-01 -2.40595698e-01 -7.98716128e-01 4.90525156e-01
5.36334693e-01 -1.09132123e+00 9.79619920e-01 -6.24014378e-01
4.94975805e-01 -5.04240930e-01 -3.08918148e-01 -6.80892050e-01
-4.80373412e-01 -7.39196062e-01 -2.90699512e-01 9.97108936e-01
5.45685410e-01 1.24258600e-01 9.43171978e-01 5.85279725e-02
-3.07564855e-01 -6.23839617e-01 -1.40610123e+00 -8.39277148e-01
-7.31222570e-01 -1.05912352e+00 1.00767024e-01 5.02122343e-01
7.91825801e-02 2.17076033e-01 -6.97350681e-01 -1.46706060e-01
5.13345718e-01 -2.19757542e-01 6.42675579e-01 -9.07146275e-01
-4.69752520e-01 -5.82308590e-01 -8.02579522e-01 -1.26165211e+00
-2.95133114e-01 -6.00113809e-01 2.14541003e-01 -1.34217656e+00
1.95072621e-01 -2.40477487e-01 -6.06561899e-01 7.16100216e-01
9.38826427e-02 4.66963619e-01 2.75467902e-01 2.06121936e-01
-1.40208697e+00 3.57560337e-01 1.10601830e+00 -2.86354959e-01
-3.62346858e-01 2.86933303e-01 -1.25520676e-01 6.31375611e-01
5.81531286e-01 -5.11386514e-01 -4.52870369e-01 1.73853800e-01
-9.02153850e-02 2.06900269e-01 5.69041431e-01 -1.57811379e+00
1.28952205e-01 -3.85702699e-01 4.28773195e-01 -1.10863090e+00
5.83516061e-01 -6.64295614e-01 1.79048210e-01 6.73793197e-01
-4.59244967e-01 -3.59426558e-01 -2.07017232e-02 7.59060264e-01
-1.52991116e-01 1.32297993e-01 4.97198939e-01 -1.12356514e-01
-1.39988017e+00 4.94817317e-01 -5.87166190e-01 1.00980371e-01
1.35642362e+00 -5.46491027e-01 -3.07978094e-01 -3.03917319e-01
-1.26202250e+00 2.47691810e-01 -3.18456799e-01 5.89844942e-01
5.60698032e-01 -1.44367635e+00 -9.08289671e-01 -2.67485350e-01
1.88227370e-01 -4.02783185e-01 5.31694174e-01 1.36259329e+00
-1.24906287e-01 3.54211032e-01 -2.19775677e-01 -8.87635529e-01
-1.34537983e+00 4.52935576e-01 4.68667984e-01 -4.06129658e-01
-9.08880770e-01 1.17298198e+00 1.20435104e-01 3.32391620e-01
5.58836162e-01 -3.20213646e-01 -6.20518327e-01 4.06194359e-01
7.72561908e-01 6.52273595e-01 -1.86614007e-01 -6.90854847e-01
-7.83511460e-01 3.05115104e-01 9.57792774e-02 -1.36059240e-01
1.12390018e+00 1.55776113e-01 3.96019101e-01 -9.60876793e-02
8.03843439e-01 -5.08411169e-01 -1.58404744e+00 -1.98217362e-01
7.04781413e-02 -4.18619424e-01 -1.43823130e-02 -8.16971421e-01
-1.17378533e+00 6.29921913e-01 6.59961998e-01 3.15421931e-02
1.23892069e+00 1.72165394e-01 4.40189809e-01 1.84895843e-02
4.14038122e-01 -1.47322929e+00 4.75293517e-01 2.05656514e-01
1.06265354e+00 -1.02109981e+00 -1.21778883e-02 -4.00321513e-01
-7.71694541e-01 1.00584471e+00 9.46096361e-01 -1.49071395e-01
5.23349643e-01 -1.15317471e-01 -3.94607842e-01 -1.40081897e-01
-1.26615894e+00 -3.46478730e-01 5.29048562e-01 6.13908231e-01
-6.35344163e-02 6.34360611e-02 -4.72416490e-01 5.97024977e-01
2.95725435e-01 9.21073183e-02 3.33582044e-01 6.53035104e-01
-7.23003745e-01 -5.80517828e-01 -1.21477239e-01 5.90631306e-01
-4.22090530e-01 4.78123426e-02 -3.25501651e-01 6.39794290e-01
5.04754066e-01 1.02661538e+00 3.53663862e-02 -8.58543277e-01
3.66093606e-01 -6.97176680e-02 2.74828821e-01 -4.18582439e-01
-6.82596028e-01 3.12877774e-01 5.55354416e-01 -1.31018090e+00
-6.06580257e-01 -9.43692982e-01 -9.86577153e-01 4.97396700e-02
-7.04779550e-02 -2.60098856e-02 6.79346547e-02 9.23844576e-01
5.05547523e-01 6.38286650e-01 3.70529920e-01 -9.72337723e-01
-3.87784302e-01 -1.11897886e+00 -3.68873686e-01 5.86422756e-02
-1.93954892e-02 -8.59525025e-01 -3.70367944e-01 6.04743324e-02] | [8.310942649841309, 0.43152111768722534] |
50f77769-9f75-437d-919d-723b20cc9f0e | few-shot-learning-of-accurate-folding | 2208.09652 | null | https://arxiv.org/abs/2208.09652v1 | https://arxiv.org/pdf/2208.09652v1.pdf | Few-Shot Learning of Accurate Folding Landscape for Protein Structure Prediction | Data-driven predictive methods which can efficiently and accurately transform protein sequences into biologically active structures are highly valuable for scientific research and therapeutical development. Determining accurate folding landscape using co-evolutionary information is fundamental to the success of modern protein structure prediction methods. As the state of the art, AlphaFold2 has dramatically raised the accuracy without performing explicit co-evolutionary analysis. Nevertheless, its performance still shows strong dependence on available sequence homologs. We investigated the cause of such dependence and presented EvoGen, a meta generative model, to remedy the underperformance of AlphaFold2 for poor MSA targets. EvoGen allows us to manipulate the folding landscape either by denoising the searched MSA or by generating virtual MSA, and helps AlphaFold2 fold accurately in low-data regime or even achieve encouraging performance with single-sequence predictions. Being able to make accurate predictions with few-shot MSA not only generalizes AlphaFold2 better for orphan sequences, but also democratizes its use for high-throughput applications. Besides, EvoGen combined with AlphaFold2 yields a probabilistic structure generation method which could explore alternative conformations of protein sequences, and the task-aware differentiable algorithm for sequence generation will benefit other related tasks including protein design. | ['Yi Qin Gao', 'YuAn Liu', 'Lijiang Yang', 'Boxin Xue', 'Yi Isaac Yang', 'Diqing Chen', 'Fan Yu', 'Ningxi Ni', 'Jialiang Yu', 'Zidong Wang', 'Min Wang', 'Haotian Chu', 'Mengyun Chen', 'Sirui Liu', 'Jun Zhang'] | 2022-08-20 | null | null | null | null | ['protein-design'] | ['medical'] | [ 4.32902247e-01 -3.80804054e-02 -2.55712494e-02 -2.96005160e-01
-8.44510913e-01 -6.17997766e-01 3.10324520e-01 -1.27264503e-02
-6.24326766e-02 1.41242445e+00 4.27769274e-02 -4.71573710e-01
9.28751454e-02 -6.63170934e-01 -1.00996125e+00 -1.26960611e+00
2.29398489e-01 8.50033581e-01 2.33365506e-01 -5.50499141e-01
3.35921675e-01 4.17844474e-01 -1.58905387e+00 5.51947832e-01
1.62046230e+00 3.01388830e-01 8.41633439e-01 3.14293861e-01
-3.59870464e-01 1.24637693e-01 -3.09505910e-01 -3.92999828e-01
3.20112891e-02 -7.62090027e-01 -7.03414202e-01 -3.23611140e-01
-9.39237997e-02 3.61549020e-01 2.86798149e-01 9.26838756e-01
5.67437589e-01 -9.54511613e-02 8.61204386e-01 -6.80118263e-01
-7.72439778e-01 2.86722481e-01 -4.84171242e-01 -3.64331715e-02
2.86605597e-01 8.26183081e-01 8.50069344e-01 -8.08089316e-01
1.04740810e+00 1.12115955e+00 5.49748659e-01 6.80020809e-01
-1.71979094e+00 -3.73124689e-01 -2.48231277e-01 4.12246346e-01
-1.03919041e+00 -2.06959277e-01 6.04122639e-01 -4.15869027e-01
1.30812788e+00 2.99120873e-01 7.06587851e-01 1.47252536e+00
6.32886469e-01 6.43410563e-01 1.14023340e+00 -7.90816694e-02
4.01237935e-01 -4.29623187e-01 -1.72061697e-02 7.00402558e-01
-2.79186573e-03 1.19470313e-01 -5.97421467e-01 -8.04650486e-01
3.16348612e-01 2.39780955e-02 -3.19982260e-01 -3.21694672e-01
-1.24727058e+00 7.50888646e-01 2.96899587e-01 9.93756130e-02
-8.10445130e-01 -1.42505139e-01 1.82662234e-01 1.64685845e-01
2.88739413e-01 9.23459649e-01 -8.02916229e-01 -4.37452704e-01
-9.01271164e-01 5.80946386e-01 4.38177705e-01 6.26414478e-01
8.47554564e-01 -2.07234602e-02 9.64935124e-03 6.99368894e-01
1.01650454e-01 3.08413386e-01 6.44179702e-01 -6.31878078e-01
1.38081601e-02 7.54399717e-01 2.08024204e-01 -6.47717953e-01
-3.40266019e-01 -2.71670550e-01 -6.69480443e-01 8.58412683e-02
2.63557822e-01 1.51393443e-01 -1.13508976e+00 1.74776363e+00
4.91778165e-01 -1.94531560e-01 7.17077330e-02 8.93729746e-01
3.12767982e-01 9.32824969e-01 2.97273874e-01 -6.62064850e-01
1.29776728e+00 -6.97092593e-01 -4.41334367e-01 1.41826794e-01
5.48857152e-01 -7.22159564e-01 1.19222140e+00 6.40630901e-01
-7.88986921e-01 -3.30930591e-01 -1.11720562e+00 1.07000194e-01
-2.88802460e-02 -1.54399529e-01 7.87540674e-01 4.53739762e-01
-6.99678361e-01 1.04136622e+00 -1.08529437e+00 -2.85723507e-01
4.39953744e-01 3.91538322e-01 -4.04716104e-01 1.04482979e-01
-1.11114454e+00 1.10172284e+00 5.76027155e-01 -2.96782672e-01
-8.35270286e-01 -8.93427312e-01 -4.12560105e-01 -4.72575352e-02
1.66476667e-01 -1.06223607e+00 8.26742709e-01 -1.00811350e+00
-1.62645173e+00 4.70848024e-01 -5.21800816e-01 -6.10404372e-01
2.93095112e-01 -1.21667266e-01 -1.01502173e-01 -2.15963647e-01
-7.75975138e-02 6.60655260e-01 6.85115933e-01 -8.34802747e-01
-9.15083736e-02 -6.40610516e-01 -7.03044534e-01 1.47308543e-01
2.85394728e-01 -2.12722689e-01 1.94657743e-01 -6.19079232e-01
2.61107534e-01 -1.05996406e+00 -7.35691130e-01 -2.66771406e-01
-3.68805826e-01 -1.29957736e-01 6.73017144e-01 -6.04929388e-01
8.39548647e-01 -1.54790294e+00 6.60560787e-01 1.26805842e-01
1.08563907e-01 5.09234071e-01 -1.43321291e-01 9.10607696e-01
-2.18479976e-01 -2.04588454e-02 -7.05007255e-01 5.49743533e-01
-4.85552579e-01 7.45890737e-02 -2.53249764e-01 3.70066732e-01
4.09723222e-01 1.16214919e+00 -7.61538982e-01 -1.31861925e-01
1.95434559e-02 5.14353216e-01 -7.69923806e-01 2.35488653e-01
-9.02984619e-01 7.76605904e-01 -6.30550802e-01 6.08380198e-01
6.77500188e-01 -5.21588743e-01 5.84036589e-01 -3.00304949e-01
-2.95444671e-02 2.29995757e-01 -4.51354951e-01 1.67772555e+00
-8.38495120e-02 6.22149594e-02 -2.95862406e-01 -8.44169497e-01
1.29002368e+00 -3.01499460e-02 5.96988559e-01 -4.37238723e-01
-2.53033131e-01 2.63698995e-01 2.98683852e-01 -3.46648484e-01
2.08635345e-01 -5.18032610e-01 4.31149632e-01 2.35803828e-01
3.70696671e-02 1.26897484e-01 -3.49775031e-02 -5.19428626e-02
1.12074423e+00 7.65651345e-01 5.80758691e-01 -4.52485770e-01
4.77299601e-01 5.72426677e-01 7.56014824e-01 1.11763917e-01
3.17559317e-02 5.49082458e-01 2.80918151e-01 -6.60119176e-01
-1.36022592e+00 -1.02096963e+00 -5.90082966e-02 1.09540403e+00
1.70394510e-01 -4.36743021e-01 -7.69403338e-01 -6.33651078e-01
-5.29248640e-02 9.72356439e-01 -3.14618945e-01 -5.42250216e-01
-5.94686747e-01 -1.52067029e+00 4.01555538e-01 -7.46798515e-03
2.81340852e-02 -1.34700406e+00 -3.68803889e-01 7.55033255e-01
-4.62947637e-01 -2.93534726e-01 -4.71184105e-01 4.81314480e-01
-9.42843616e-01 -1.15729666e+00 -8.39177549e-01 -5.15978456e-01
3.69337738e-01 5.38620129e-02 1.06831980e+00 -1.23107381e-01
-4.90450472e-01 -6.07020378e-01 -3.37475479e-01 -2.64322102e-01
-9.65755939e-01 2.42537409e-01 3.24184120e-01 -1.35387197e-01
4.21790183e-01 -1.15119708e+00 -7.42375612e-01 4.15216804e-01
-7.37149000e-01 4.42669094e-01 7.43667722e-01 1.19412160e+00
9.95630443e-01 -5.55499971e-01 9.28399861e-01 -9.62163508e-01
5.35407424e-01 -4.14653897e-01 -5.41340828e-01 4.45436835e-01
-1.03882086e+00 9.32780266e-01 6.75521493e-01 -4.01097000e-01
-1.19446182e+00 4.23125386e-01 -6.35419548e-01 -2.49806732e-01
4.39210758e-02 4.18583781e-01 -3.32567126e-01 1.48512498e-01
1.02241385e+00 9.62917686e-01 4.39535528e-01 -7.66981661e-01
3.00435960e-01 3.60345989e-01 3.68648767e-01 -7.46164382e-01
4.01056468e-01 1.07001252e-01 8.58441517e-02 -8.17085207e-01
-3.93182307e-01 -2.43061915e-01 -5.41000247e-01 3.21757436e-01
6.86907291e-01 -5.05993187e-01 -8.81220043e-01 2.34988779e-01
-1.19824517e+00 2.53558122e-02 1.63857549e-01 6.97877482e-02
-1.02114379e+00 6.33409023e-01 -4.27348435e-01 -6.27562225e-01
-7.49712110e-01 -1.54991293e+00 9.73947763e-01 -6.30643070e-02
-3.66440564e-01 -3.39868635e-01 4.55131948e-01 4.79911804e-01
2.20735848e-01 4.29224104e-01 1.48947120e+00 -6.50144696e-01
-7.12482631e-01 4.54698652e-01 2.03319341e-01 2.12527458e-02
2.28723913e-01 1.86148837e-01 -5.92189670e-01 -4.04803939e-02
-2.43291318e-01 -3.07122827e-01 9.60619867e-01 3.95774215e-01
9.85672176e-01 -4.50550526e-01 -6.59734070e-01 7.15350986e-01
1.24425781e+00 3.41954142e-01 8.67873669e-01 4.58531648e-01
5.09653151e-01 4.98969823e-01 8.27705681e-01 2.03523383e-01
-3.64896283e-02 9.50011134e-01 5.34865141e-01 1.41518846e-01
2.15056077e-01 -4.60022867e-01 4.67041194e-01 6.74768806e-01
-3.85792345e-01 -2.32029900e-01 -9.37932074e-01 2.46809229e-01
-2.05272579e+00 -1.18132186e+00 -3.28179032e-01 1.91638935e+00
1.41855514e+00 -3.34669024e-01 1.81505665e-01 -4.00253743e-01
5.97034037e-01 -1.72687903e-01 -1.15807950e+00 -2.94811249e-01
-5.16827404e-01 2.86994308e-01 3.24602783e-01 1.53718784e-01
-5.81522524e-01 1.03581190e+00 6.77538300e+00 1.14719331e+00
-1.07937169e+00 -1.00814313e-01 7.48939991e-01 -1.54067129e-02
-4.39568639e-01 1.76488310e-01 -7.83699751e-01 8.06078553e-01
1.08686316e+00 -3.41832876e-01 3.17441553e-01 1.10108769e+00
3.74978632e-01 1.42509550e-01 -8.55134368e-01 7.69054174e-01
-4.94782418e-01 -2.02263403e+00 3.77623498e-01 3.16561669e-01
6.53987408e-01 2.00771496e-01 -9.36782584e-02 -3.09086759e-02
3.52058649e-01 -1.21672487e+00 1.79106548e-01 6.52518511e-01
7.04389453e-01 -1.14525723e+00 5.44263601e-01 7.62745321e-01
-5.96465051e-01 4.06159341e-01 -5.48369229e-01 3.44180226e-01
2.28775501e-01 6.23926222e-01 -1.15541339e+00 4.64903146e-01
3.11098784e-01 4.41783369e-01 -2.88412213e-01 6.48785293e-01
1.33323237e-01 7.16826797e-01 -2.26892591e-01 -2.05560133e-01
9.24424734e-03 -7.71931171e-01 7.54522324e-01 9.12336946e-01
4.13352281e-01 1.91324711e-01 1.96238771e-01 1.06984437e+00
2.09547192e-01 2.91286886e-01 -3.90308887e-01 -2.56570578e-01
3.53979766e-01 8.43207061e-01 -3.73758614e-01 -1.67220086e-01
2.38592699e-01 1.16355550e+00 5.11982739e-01 1.76056698e-01
-9.77466822e-01 -2.12243497e-01 1.11535144e+00 4.18975860e-01
2.45452061e-01 -1.37398005e-01 -3.17204706e-02 -1.07389617e+00
-2.06376418e-01 -1.35469615e+00 -4.16834392e-02 -6.30618989e-01
-1.31717539e+00 7.10097253e-01 -3.66113514e-01 -9.79842901e-01
-3.85474145e-01 -4.37747478e-01 -4.12236661e-01 9.90420520e-01
-1.05580640e+00 -9.95771348e-01 2.81646341e-01 3.39519270e-02
7.55506635e-01 -2.78376579e-01 8.72539341e-01 -1.31969139e-01
-3.41781169e-01 4.42755252e-01 6.58225834e-01 -9.41302836e-01
8.00597847e-01 -1.02936172e+00 8.40909719e-01 4.01069939e-01
-3.91817577e-02 8.01480114e-01 1.25788844e+00 -1.08730567e+00
-1.52420115e+00 -1.15718901e+00 5.64648986e-01 -5.41740835e-01
3.14616323e-01 -1.11449137e-01 -1.48512912e+00 4.32775654e-02
-2.24107787e-01 -5.77232778e-01 5.82254291e-01 -1.76546238e-02
-8.25279802e-02 3.23199004e-01 -1.24824798e+00 6.34648621e-01
1.30382383e+00 -6.03812188e-02 -5.56120694e-01 4.54698086e-01
9.55682695e-01 -1.20338559e-01 -7.50597596e-01 5.64093649e-01
4.66110706e-01 -1.07360756e+00 1.01360548e+00 -1.14175844e+00
4.55679595e-01 -4.48871464e-01 -1.01833366e-01 -1.35118997e+00
-5.22528708e-01 -7.98586905e-01 -3.68203670e-01 8.02930534e-01
8.47842455e-01 -5.80890238e-01 1.20883644e+00 1.67715624e-01
-2.55758584e-01 -1.13408923e+00 -8.73764396e-01 -8.36599052e-01
1.28005639e-01 2.98686028e-02 7.48579800e-01 6.93059444e-01
-6.68646321e-02 3.70257705e-01 -5.10604024e-01 -1.59875721e-01
4.19743061e-01 4.13703471e-01 7.58680761e-01 -1.08748174e+00
-5.80705702e-01 -2.21056759e-01 -2.86986351e-01 -1.05106366e+00
-4.99674026e-03 -1.01146436e+00 9.76467654e-02 -1.14568162e+00
4.87636656e-01 -1.23176783e-01 -1.50690796e-02 3.71773541e-01
-5.66949308e-01 -2.51190923e-03 -1.98225662e-01 5.34753978e-01
-1.09258652e-01 1.07122779e+00 1.33107650e+00 -5.77804781e-02
-4.17309552e-01 7.30364323e-02 -6.49374247e-01 3.66916507e-01
7.67154932e-01 -5.32623053e-01 -2.69630432e-01 4.25705045e-01
2.01654792e-01 8.61627534e-02 1.16673820e-01 -6.22425437e-01
-2.89098859e-01 -6.55834496e-01 3.09273541e-01 -7.81265199e-01
3.05394828e-01 -3.81661594e-01 8.76117945e-01 6.37416720e-01
-5.84733449e-02 2.34718204e-01 -1.27181187e-02 8.21734607e-01
-2.75580753e-02 -1.77138578e-02 1.00311911e+00 -2.36201957e-01
-3.32218647e-01 3.02302420e-01 -5.43595910e-01 -2.47071698e-01
9.83850539e-01 -4.74988312e-01 -4.34253551e-02 -9.93557274e-03
-7.63144851e-01 -1.15716740e-01 9.23733652e-01 3.13563198e-01
5.45256019e-01 -9.34837759e-01 -7.37881899e-01 2.16275558e-01
1.23003267e-01 -2.63039529e-01 1.96400151e-01 6.69298768e-01
-7.47090638e-01 4.78152633e-01 -4.43865597e-01 -8.91931176e-01
-1.41106272e+00 6.53753102e-01 2.70139724e-01 -3.46366167e-01
-4.85194683e-01 5.99926770e-01 3.39367837e-01 -5.42002320e-01
-5.20768702e-01 1.29814580e-01 2.12196290e-01 -3.04119587e-01
4.60296333e-01 1.54907808e-01 2.29570821e-01 -4.23612297e-01
-3.54142576e-01 2.48832077e-01 -4.24317032e-01 2.95143306e-01
1.69955373e+00 2.10744500e-01 -8.29828829e-02 1.43196866e-01
8.12638164e-01 -3.34564537e-01 -1.48044658e+00 1.27667457e-01
2.45466039e-01 -2.57794589e-01 -4.10480738e-01 -1.07642591e+00
-4.22020137e-01 5.88168561e-01 5.02979755e-01 -4.78702068e-01
8.77626002e-01 -1.30995344e-02 9.91989493e-01 6.36394024e-01
7.92611003e-01 -6.15575850e-01 -1.62352115e-01 3.44970256e-01
9.00426924e-01 -1.24552166e+00 -4.69566369e-03 -2.36266330e-01
-8.89209092e-01 1.05765104e+00 5.24623692e-01 1.90438524e-01
2.27973443e-02 8.89685601e-02 -8.94804597e-02 -1.65045455e-01
-1.07915449e+00 1.65587962e-01 3.49844843e-01 9.80387151e-01
5.34949958e-01 -7.32908677e-03 -5.87330103e-01 6.87186837e-01
-1.89028025e-01 -7.33304173e-02 8.14698264e-02 7.38584936e-01
-8.04729760e-01 -1.64068937e+00 -2.69088775e-01 4.68772769e-01
-3.58303726e-01 -3.47674608e-01 -7.27101326e-01 4.17541862e-01
-7.96978474e-02 3.59261692e-01 -6.02717936e-01 -2.58059353e-01
8.44416320e-02 5.26096761e-01 6.06667161e-01 -3.83841395e-01
-6.62361801e-01 -4.05266881e-02 -3.14848460e-02 -4.81670231e-01
-5.57787232e-02 -6.39619589e-01 -1.43838656e+00 -4.31278259e-01
-4.58442122e-01 4.18990016e-01 6.47455752e-01 7.63242066e-01
1.14039266e+00 1.96385071e-01 5.07802904e-01 -8.06888521e-01
-8.27509582e-01 -9.10595357e-01 -2.02771604e-01 2.92155564e-01
2.74582785e-02 -6.61338091e-01 1.48959443e-01 3.00902426e-01] | [4.714763641357422, 5.582594871520996] |
cbf4da58-a708-4218-af88-d33a3af9e495 | tukey-inspired-video-object-segmentation | 1811.07958 | null | http://arxiv.org/abs/1811.07958v2 | http://arxiv.org/pdf/1811.07958v2.pdf | Tukey-Inspired Video Object Segmentation | We investigate the problem of strictly unsupervised video object
segmentation, i.e., the separation of a primary object from background in video
without a user-provided object mask or any training on an annotated dataset. We
find foreground objects in low-level vision data using a John Tukey-inspired
measure of "outlierness". This Tukey-inspired measure also estimates the
reliability of each data source as video characteristics change (e.g., a camera
starts moving). The proposed method achieves state-of-the-art results for
strictly unsupervised video object segmentation on the challenging DAVIS
dataset. Finally, we use a variant of the Tukey-inspired measure to combine the
output of multiple segmentation methods, including those using supervision
during training, runtime, or both. This collectively more robust method of
segmentation improves the Jaccard measure of its constituent methods by as much
as 28%. | ['Jason J. Corso', 'Brent A. Griffin'] | 2018-11-19 | null | null | null | null | ['unsupervised-video-object-segmentation'] | ['computer-vision'] | [ 5.60685575e-01 -1.30391598e-01 7.36777857e-02 -2.77577370e-01
-8.94828141e-01 -9.17752683e-01 5.01636505e-01 2.80027032e-01
-5.73774219e-01 3.19343388e-01 -3.41090947e-01 -1.24950241e-02
1.37089705e-02 -3.45926285e-01 -1.08746886e+00 -9.70065892e-01
-2.83125583e-02 6.22360706e-01 8.53526413e-01 4.47001129e-01
5.13332963e-01 5.07321179e-01 -1.63853621e+00 1.92571908e-01
7.56631315e-01 1.07188332e+00 1.68175831e-01 9.72268045e-01
5.59397675e-02 7.19525993e-01 -8.54840100e-01 -2.59989768e-01
5.92216194e-01 -3.48319054e-01 -7.39532113e-01 7.35717535e-01
1.06832016e+00 -2.07765326e-01 -1.62234098e-01 1.34802568e+00
-7.13982359e-02 2.30228513e-01 6.44538462e-01 -1.24964917e+00
-1.07541546e-01 4.67621893e-01 -6.58652365e-01 5.14590561e-01
1.32011876e-01 3.49030316e-01 7.32904851e-01 -8.39066446e-01
7.35646427e-01 7.71273613e-01 5.68687320e-01 1.27089098e-01
-1.35226297e+00 -9.81382281e-02 1.91462159e-01 1.48715347e-01
-1.22801352e+00 -4.17639762e-01 6.75438523e-01 -8.58787119e-01
4.90118921e-01 3.64540786e-01 4.36883837e-01 6.80484533e-01
-2.19698831e-01 8.49498928e-01 1.01555574e+00 -2.25473329e-01
4.77876842e-01 8.63894895e-02 3.44912231e-01 6.33398473e-01
5.42052805e-01 -2.08562136e-01 -2.49363691e-01 -6.33678064e-02
5.18291771e-01 1.73382945e-02 -1.35823250e-01 -5.18646657e-01
-1.28956592e+00 3.48953933e-01 -4.34146002e-02 2.12050632e-01
-3.11806738e-01 2.79662460e-02 4.34702545e-01 2.99528036e-02
3.30258280e-01 4.44396764e-01 -5.60109794e-01 -4.79060039e-02
-1.55668318e+00 3.78424488e-02 7.34265864e-01 1.22023106e+00
7.63440788e-01 -1.57557260e-02 -2.40161270e-01 3.26199442e-01
7.88709223e-02 2.55302697e-01 1.36330783e-01 -1.30865765e+00
9.05541331e-02 5.64687312e-01 1.61092415e-01 -9.87953067e-01
-1.69050291e-01 -2.08418190e-01 -4.06875312e-01 3.36387426e-01
9.84909296e-01 9.64283571e-03 -1.11989427e+00 1.40232754e+00
6.59452260e-01 5.36152065e-01 -1.40446186e-01 8.37697029e-01
7.07810223e-01 3.89503837e-01 -1.88065529e-01 -5.34838617e-01
1.10931742e+00 -1.17358780e+00 -5.73307455e-01 -5.32445684e-02
2.27067217e-01 -7.89538205e-01 5.31814456e-01 6.72873616e-01
-9.92759228e-01 -7.04854012e-01 -1.04034698e+00 2.57035226e-01
-3.15754861e-01 6.42517954e-02 2.74560601e-01 7.20795810e-01
-7.99967766e-01 8.36380184e-01 -9.36080396e-01 -3.34482849e-01
5.53757608e-01 3.35106581e-01 -3.49480122e-01 1.38983792e-02
-3.98659796e-01 3.29911500e-01 5.06350636e-01 4.75043198e-03
-1.22255111e+00 -6.52336657e-01 -6.41257644e-01 -2.90550053e-01
9.16392744e-01 -2.63639599e-01 9.19108391e-01 -1.38495672e+00
-1.10032368e+00 1.12498331e+00 -9.71715748e-02 -5.88282049e-01
9.90804791e-01 -3.93265814e-01 -6.12903610e-02 5.83578348e-01
2.01992333e-01 5.00822783e-01 1.35543680e+00 -1.55103207e+00
-7.81852484e-01 -4.29095030e-01 -1.40198097e-01 -1.53067514e-01
8.26051235e-02 1.37003779e-01 -9.14329469e-01 -6.91940546e-01
1.58132538e-01 -1.04245269e+00 -2.24669859e-01 -1.31509259e-01
-4.81391340e-01 -6.29463494e-02 1.02472639e+00 -8.47456574e-01
1.12809610e+00 -2.28835702e+00 9.40774530e-02 4.02614385e-01
3.56719494e-01 2.68169433e-01 4.33940347e-03 -1.07823476e-01
7.63163567e-02 2.14150414e-01 -5.56030452e-01 -2.06273764e-01
-3.35905284e-01 2.08210662e-01 9.22910199e-02 8.34750295e-01
2.78203547e-01 3.87455851e-01 -1.00300121e+00 -7.62204766e-01
1.71862468e-01 -2.06202921e-02 -4.43752408e-01 2.61716455e-01
-3.84631127e-01 5.23671627e-01 5.27436547e-02 7.53497660e-01
8.49009573e-01 -6.38548508e-02 1.18074819e-01 -3.48296314e-01
-8.85929465e-02 -3.13361108e-01 -1.54872596e+00 1.72532415e+00
4.53766465e-01 7.81117320e-01 -3.17884497e-02 -9.59446669e-01
6.59923911e-01 8.10672790e-02 8.39542329e-01 4.04606424e-02
1.86384663e-01 1.34386793e-01 6.22159503e-02 -5.27029514e-01
5.88965952e-01 4.72879291e-01 2.42164224e-01 2.34147578e-01
2.84880430e-01 3.41166705e-02 7.24345207e-01 3.03899527e-01
1.47070885e+00 1.70423165e-01 7.46936202e-02 -4.69960332e-01
5.20847678e-01 1.38053134e-01 7.12327719e-01 1.08772779e+00
-5.67072630e-01 9.01010573e-01 5.47093391e-01 -1.17075831e-01
-1.07908058e+00 -9.84565258e-01 -1.04145490e-01 9.53332722e-01
3.27188104e-01 -2.88357556e-01 -1.15056384e+00 -1.01363313e+00
1.19447783e-01 3.65619302e-01 -6.21541679e-01 7.52837509e-02
-4.16630328e-01 -5.22871733e-01 4.97792304e-01 3.83728772e-01
3.64199579e-01 -8.00142586e-01 -7.10418463e-01 1.08080499e-01
-9.54672694e-02 -1.55134797e+00 -5.06354034e-01 3.72628748e-01
-1.01833117e+00 -1.39393151e+00 -4.85281795e-01 -6.46052480e-01
8.32428515e-01 2.90627718e-01 1.16478086e+00 2.22390682e-01
-3.40562254e-01 5.86810827e-01 -4.35300916e-01 2.61049494e-02
-4.50653434e-01 -3.32030147e-01 1.67213246e-01 3.10337842e-01
3.75135005e-01 -1.08872809e-01 -5.31330824e-01 6.08712554e-01
-1.04583275e+00 -4.65562612e-01 3.07377070e-01 4.09811318e-01
7.78766632e-01 5.47830045e-01 4.86005396e-02 -8.69023800e-01
1.00285513e-02 -2.55147517e-01 -8.07979405e-01 2.48909369e-01
-3.56484234e-01 -2.24196166e-01 4.15670544e-01 -6.14838481e-01
-8.11461508e-01 3.27698559e-01 4.31150436e-01 -8.16322565e-01
-5.75407267e-01 -3.33651453e-02 -2.65578359e-01 -1.02718629e-01
4.90313292e-01 -5.49995191e-02 -2.47136533e-01 -3.75886828e-01
2.68776953e-01 4.13386613e-01 9.51459289e-01 -4.56874967e-01
8.44240963e-01 6.85636103e-01 -7.23161176e-02 -9.25159812e-01
-6.74502432e-01 -9.03628469e-01 -1.06674325e+00 -4.55268562e-01
1.13211298e+00 -7.06254125e-01 -4.15151834e-01 7.01995492e-01
-1.04997230e+00 -3.46516192e-01 -3.08630049e-01 4.46415752e-01
-5.54741025e-01 6.16462767e-01 -4.22183394e-01 -8.50245893e-01
2.20669985e-01 -1.15546560e+00 9.87227440e-01 3.37426662e-01
-1.15018031e-02 -6.79131269e-01 -2.47786671e-01 6.63566172e-01
-3.54957670e-01 4.48291481e-01 3.43049586e-01 -1.04371953e+00
-8.00627708e-01 -1.77503601e-01 -1.48052976e-01 6.92239881e-01
1.58212215e-01 6.48113251e-01 -1.00910854e+00 -1.86662599e-01
1.68123513e-01 2.02349707e-01 9.28979516e-01 5.22498965e-01
1.20306849e+00 -2.21331701e-01 -2.20896646e-01 4.48309749e-01
1.48906755e+00 4.10987288e-01 6.04237497e-01 3.44455957e-01
9.64636922e-01 6.25384331e-01 8.74554515e-01 4.34322208e-01
-9.27097425e-02 3.78777206e-01 4.47264284e-01 5.85840158e-02
3.93343344e-02 2.26518378e-01 4.93260324e-01 3.83343756e-01
-2.32089892e-01 -8.17998797e-02 -9.07461286e-01 5.42491674e-01
-1.91091168e+00 -9.64230359e-01 -5.67024171e-01 2.36777163e+00
5.84911585e-01 4.64516908e-01 3.87025505e-01 1.99774653e-01
9.82639790e-01 -7.81711265e-02 -6.91196740e-01 -5.37632557e-04
-9.49495286e-02 -9.43128988e-02 9.41283226e-01 2.56261438e-01
-1.61209893e+00 1.00710022e+00 6.22001791e+00 7.76612043e-01
-5.94944477e-01 5.58841750e-02 6.37924194e-01 1.87564958e-02
2.01726034e-01 1.08405547e-02 -5.62723756e-01 6.31263793e-01
5.80785632e-01 2.16602504e-01 3.12139332e-01 9.74897206e-01
1.61312178e-01 -6.31109595e-01 -1.38090527e+00 8.50486815e-01
3.72268766e-01 -1.02139199e+00 -2.91195899e-01 5.54945171e-02
9.53773737e-01 -5.48268333e-02 -2.58983225e-01 -1.86225310e-01
-4.16983530e-04 -5.59100866e-01 8.73145938e-01 5.64925015e-01
3.20815831e-01 -5.35504997e-01 6.34713650e-01 1.34131357e-01
-1.05864990e+00 1.27154753e-01 -9.68295857e-02 3.34089756e-01
5.58679141e-02 7.80869424e-01 -6.18611872e-01 4.71895576e-01
8.65707457e-01 7.41806448e-01 -1.05222273e+00 1.37303078e+00
-6.72547193e-03 7.93985486e-01 -4.54977334e-01 4.61517692e-01
1.56090200e-01 -5.20360708e-01 1.02310145e+00 1.33261728e+00
-7.24080876e-02 -8.29765797e-02 4.77862448e-01 6.06911421e-01
1.87362619e-02 -1.16214946e-01 -4.32787925e-01 -1.00469030e-01
1.21079475e-01 1.37435341e+00 -1.45164895e+00 -6.59807265e-01
-2.58180946e-01 1.12454891e+00 -1.03065461e-01 4.39820349e-01
-1.03800464e+00 -1.55636847e-01 5.99029183e-01 -2.12277379e-02
8.09217274e-01 -3.43864471e-01 -2.78752536e-01 -1.00420356e+00
1.44768223e-01 -1.07875788e+00 3.75630438e-01 -6.16741180e-01
-1.19835794e+00 2.60518372e-01 -8.68041888e-02 -1.24093866e+00
4.51692641e-02 -6.58619642e-01 -4.65775758e-01 1.78238243e-01
-9.47441757e-01 -5.61635017e-01 -4.10940498e-01 4.54904169e-01
6.07398272e-01 -1.35469377e-01 7.48911276e-02 3.10189337e-01
-6.88042879e-01 2.48322710e-01 1.88575923e-01 3.97523731e-01
7.67466664e-01 -1.55328786e+00 6.71492741e-02 1.50279820e+00
4.34625238e-01 4.11490232e-01 9.11147833e-01 -8.54065537e-01
-1.27933896e+00 -1.16311049e+00 2.01965004e-01 -7.44680703e-01
7.33826399e-01 -2.45369896e-01 -9.28271115e-01 6.01125419e-01
6.03996264e-03 9.64745134e-02 5.81876457e-01 -2.34559417e-01
-2.17535660e-01 -2.20849305e-01 -1.21148705e+00 4.51751500e-01
1.06118977e+00 -2.45478958e-01 -7.08484471e-01 4.22524422e-01
7.77778029e-01 -3.24360937e-01 -8.98600399e-01 4.44013238e-01
3.41161102e-01 -1.13711941e+00 8.69398713e-01 -6.85206294e-01
2.66127437e-01 -8.15833747e-01 -2.81829923e-01 -6.50664210e-01
-7.31092840e-02 -7.85170972e-01 -3.41185421e-01 1.34718692e+00
1.00498855e-01 -5.67744710e-02 7.70250440e-01 6.56073928e-01
-1.13666974e-01 -3.20363283e-01 -7.63681471e-01 -9.44993198e-01
-4.95589256e-01 -5.54032147e-01 -2.07537152e-02 8.64609718e-01
-5.30351222e-01 -1.72400624e-01 -1.46750733e-01 4.81352180e-01
1.06915367e+00 -2.82492228e-02 1.07467258e+00 -1.20732713e+00
-1.52329624e-01 -4.96523708e-01 -8.26381862e-01 -9.58610415e-01
1.23868123e-01 -5.51073015e-01 4.58086342e-01 -1.24589944e+00
2.86782414e-01 -1.48835674e-01 -3.50224316e-01 5.46736680e-02
-3.85196000e-01 3.98127049e-01 2.22191185e-01 2.45043159e-01
-9.85457182e-01 1.30467480e-02 9.13656294e-01 -2.38137573e-01
-2.43092775e-01 2.13186555e-02 -2.81883776e-01 1.05277431e+00
6.47352099e-01 -7.50293434e-01 3.50710587e-03 -1.31706014e-01
-1.99549690e-01 -3.34991574e-01 5.15044212e-01 -1.19354224e+00
2.21045509e-01 -3.67544666e-02 4.40164000e-01 -5.98947942e-01
-1.55501381e-01 -1.16467237e+00 5.81530221e-02 3.51978600e-01
-5.52641042e-02 3.51867415e-02 7.38906339e-02 9.76295531e-01
-1.30951494e-01 -5.12786210e-01 8.17711711e-01 -2.20614195e-01
-1.06221652e+00 2.43390411e-01 -4.90737945e-01 3.91727120e-01
1.31466377e+00 -5.60527146e-01 -1.28438026e-01 7.21444860e-02
-8.67366314e-01 1.34709954e-01 6.47001684e-01 2.36426637e-01
3.33767354e-01 -8.64891529e-01 -4.94124830e-01 -2.67042220e-02
-6.31459802e-03 2.65561163e-01 -5.36749922e-02 1.24624395e+00
-6.26113713e-01 -2.89909273e-01 -6.01084456e-02 -1.20689416e+00
-1.65086377e+00 6.66806221e-01 2.06903994e-01 8.60082172e-03
-5.54201722e-01 7.74308145e-01 -2.43325159e-02 -2.96034329e-02
5.00923574e-01 -5.32925427e-01 3.98331136e-02 3.94014537e-01
4.01464313e-01 7.30300486e-01 -6.46723807e-03 -6.78456426e-01
-6.07755601e-01 4.80561525e-01 5.99948987e-02 -6.69910833e-02
1.02862692e+00 -6.54373690e-02 -2.15491772e-01 7.77357519e-01
1.01187122e+00 6.67233393e-02 -1.61269987e+00 -1.40168220e-01
3.84877056e-01 -7.14952409e-01 -1.73541695e-01 -5.02213836e-01
-1.03871465e+00 4.10220236e-01 6.26422048e-01 4.30613846e-01
1.05110085e+00 8.88578966e-02 3.86305630e-01 3.41823071e-01
2.22913012e-01 -1.52246952e+00 2.58421332e-01 2.99941272e-01
3.51749629e-01 -1.47074711e+00 3.30586404e-01 -4.09492880e-01
-7.37375200e-01 1.06197166e+00 4.80162442e-01 -2.23007426e-01
5.13381660e-01 2.68413186e-01 1.17494375e-01 -7.03261644e-02
-2.51491189e-01 -5.00097454e-01 5.00428438e-01 5.44077575e-01
-1.09047495e-01 -2.28611380e-02 1.30231827e-02 1.28833562e-01
1.28720373e-01 -1.98358282e-01 5.75529099e-01 9.19327199e-01
-5.29733181e-01 -4.68074590e-01 -5.08601189e-01 5.47882497e-01
-5.68041623e-01 4.60438095e-02 -5.23680329e-01 6.67400002e-01
3.58619630e-01 1.15811884e+00 3.07406306e-01 -2.78072119e-01
1.58910722e-01 4.30556424e-02 5.24011254e-01 -4.92314309e-01
-6.89602733e-01 4.29531068e-01 -1.33552596e-01 -7.36321151e-01
-1.09202504e+00 -1.08758008e+00 -1.19225430e+00 -5.76436985e-03
-5.07178128e-01 4.48464509e-03 6.24619603e-01 1.05256796e+00
2.82254275e-02 5.25469065e-01 5.21152854e-01 -9.26191032e-01
-7.87694529e-02 -7.71818876e-01 -5.60622871e-01 7.15952039e-01
3.67833197e-01 -4.89501953e-01 -6.28391922e-01 6.24590099e-01] | [9.0302734375, -0.3196549713611603] |
2ae7a165-a289-4021-915a-0a7a1e88ce9b | catching-image-retrieval-generalization | 2306.13357 | null | https://arxiv.org/abs/2306.13357v1 | https://arxiv.org/pdf/2306.13357v1.pdf | Catching Image Retrieval Generalization | The concepts of overfitting and generalization are vital for evaluating machine learning models. In this work, we show that the popular Recall@K metric depends on the number of classes in the dataset, which limits its ability to estimate generalization. To fix this issue, we propose a new metric, which measures retrieval performance, and, unlike Recall@K, estimates generalization. We apply the proposed metric to popular image retrieval methods and provide new insights about deep metric learning generalization. | ['Ivan Karpukhin', 'Maksim Zhdanov'] | 2023-06-23 | null | null | null | null | ['metric-learning', 'metric-learning', 'retrieval'] | ['computer-vision', 'methodology', 'methodology'] | [-2.02670872e-01 -5.73318303e-01 -4.49303716e-01 -5.68980217e-01
-7.84347713e-01 -6.94140315e-01 5.19654512e-01 4.59584177e-01
-5.82860649e-01 5.11895776e-01 -3.89792398e-02 -5.46234176e-02
-4.73260075e-01 -7.76584685e-01 -3.21670741e-01 -5.14079750e-01
1.85310245e-01 -7.08878273e-03 1.80981494e-02 -1.39153033e-01
5.17071545e-01 5.73036790e-01 -1.45584106e+00 1.73462868e-01
6.95568502e-01 1.44090438e+00 -1.19366720e-01 3.44160646e-01
3.45171466e-02 5.04968107e-01 -8.07907701e-01 -4.45634961e-01
2.77967811e-01 -3.10225427e-01 -7.85770893e-01 -4.66699570e-01
7.00928032e-01 -4.35762703e-01 -6.48564517e-01 1.00746596e+00
3.54918092e-01 2.04804346e-01 1.12847805e+00 -1.49739695e+00
-9.87623096e-01 5.70872724e-01 -2.13983208e-01 5.30033231e-01
1.48720503e-01 -1.83695123e-01 1.33976912e+00 -9.58590746e-01
2.96001285e-01 9.44103718e-01 9.03614998e-01 5.85726857e-01
-9.76492226e-01 -7.36783803e-01 -2.35455390e-02 2.83203095e-01
-1.81668735e+00 -3.19780856e-02 6.20510817e-01 -4.48640794e-01
4.89331603e-01 1.75336838e-01 1.75382346e-01 9.16348875e-01
1.07588232e-01 9.09100890e-01 9.60744798e-01 -2.79303819e-01
2.40418896e-01 1.49533838e-01 4.49814528e-01 3.80451858e-01
3.51913035e-01 1.03839032e-01 -3.66654903e-01 -1.26868546e-01
6.49097025e-01 3.72246593e-01 -4.96718884e-01 -3.03492814e-01
-8.36992800e-01 9.97713864e-01 8.19630563e-01 3.99782598e-01
1.92772761e-01 4.03775811e-01 4.51606840e-01 5.37107766e-01
3.33180368e-01 6.58893287e-01 -4.44901288e-01 -4.59074043e-02
-8.86942029e-01 2.86123097e-01 5.49106896e-01 8.01284790e-01
8.74633729e-01 -4.15313125e-01 -2.69729257e-01 1.07731533e+00
2.43096612e-02 4.75212842e-01 6.59829080e-01 -9.49859619e-01
1.50017033e-03 6.80202723e-01 -6.60084635e-02 -1.16292417e+00
-6.41299129e-01 -6.36649013e-01 -7.34175444e-01 -2.83391297e-01
3.09267133e-01 5.28254271e-01 -6.89977229e-01 2.05909967e+00
-1.16048001e-01 -1.53811455e-01 -1.57203779e-01 9.60506916e-01
7.57834196e-01 3.10009003e-01 3.16767059e-02 -1.26761869e-01
6.44811034e-01 -7.75585234e-01 -4.98413414e-01 1.37798414e-01
9.93572176e-01 -6.41699314e-01 1.50139534e+00 3.39904487e-01
-7.50719905e-01 -4.27401364e-01 -1.24891436e+00 -2.00318292e-01
-7.09308863e-01 3.36633213e-02 6.75388098e-01 5.18825531e-01
-9.22093809e-01 9.00843859e-01 -6.14377677e-01 -4.87850547e-01
3.53180170e-01 1.90661773e-01 -1.70983255e-01 -2.95776248e-01
-1.13252711e+00 8.48264992e-01 3.95440757e-01 -2.32390285e-01
-5.15930474e-01 -5.47290266e-01 -5.47457576e-01 4.22915280e-01
3.51449773e-02 -6.16926849e-01 1.27930892e+00 -4.22200114e-01
-1.04201508e+00 8.25768650e-01 2.38684654e-01 -2.79907078e-01
3.13627899e-01 -2.55188406e-01 -4.51314151e-01 7.66795576e-02
-2.44392514e-01 5.00968814e-01 6.05371833e-01 -1.05595815e+00
-3.62327993e-01 -5.94724238e-01 2.89157659e-01 -2.22925648e-01
-9.03221726e-01 -3.74129355e-01 -2.29182303e-01 -7.77068436e-01
3.08560401e-01 -7.71348596e-01 3.50668803e-02 1.15507677e-01
-1.32946506e-01 -5.77882469e-01 4.47982222e-01 -1.08748123e-01
1.62254560e+00 -2.25536537e+00 -2.53576376e-02 4.02947575e-01
4.27639902e-01 2.16485977e-01 -3.32372189e-01 5.23910224e-01
1.80668727e-01 1.91354752e-01 -2.46278644e-01 -8.15631971e-02
3.63651514e-01 2.75998324e-01 -5.31484783e-01 4.01454568e-01
9.79639813e-02 1.04530454e+00 -8.93086851e-01 -3.70581269e-01
-6.78935871e-02 3.74904215e-01 -5.12441456e-01 -1.00798383e-01
5.57136768e-03 -1.76949486e-01 -5.07884920e-01 4.67621088e-01
8.93398225e-01 -5.19984126e-01 -9.82462689e-02 -2.77256936e-01
3.71840447e-01 1.48315385e-01 -8.52293611e-01 1.69370615e+00
-5.07961750e-01 7.92471707e-01 -7.27801204e-01 -1.13056529e+00
9.83760238e-01 -2.51207858e-01 3.30603868e-01 -9.69554484e-01
2.48128399e-02 4.92817253e-01 -3.09259742e-01 -3.12890917e-01
6.50021195e-01 3.00099999e-02 -2.58379519e-01 6.25149906e-01
4.16723378e-02 5.13787474e-03 1.00811854e-01 2.71055996e-01
1.11087263e+00 -2.82239318e-01 1.90965608e-01 -5.50667882e-01
4.29117739e-01 -3.43761861e-01 3.58353406e-01 1.01932919e+00
-2.52770573e-01 6.80993736e-01 4.77381915e-01 -5.97760201e-01
-9.21595633e-01 -1.30376899e+00 -5.11882603e-01 1.16351485e+00
3.84575903e-01 -7.59436071e-01 -6.66523278e-01 -7.75813222e-01
3.60529631e-01 2.16936052e-01 -8.01713347e-01 -8.01108479e-01
-2.88673431e-01 -6.22672796e-01 6.57894492e-01 8.33372712e-01
2.13767901e-01 -4.65598971e-01 -3.40894341e-01 -2.22682342e-01
-1.98929444e-01 -9.37744796e-01 -5.88376164e-01 -2.88090371e-02
-8.95581663e-01 -1.13145232e+00 -8.08344901e-01 -5.68049550e-01
3.82786453e-01 4.08788890e-01 1.21421409e+00 3.71676028e-01
-2.37951860e-01 6.17347300e-01 -5.18329561e-01 -2.43152380e-01
-1.57214463e-01 5.10890245e-01 1.99798629e-01 -3.66942793e-01
7.38296509e-01 -4.48788702e-01 -9.17336643e-01 4.42828119e-01
-1.28916848e+00 -6.91733837e-01 5.99275172e-01 8.06357384e-01
5.14336288e-01 -1.81964859e-01 7.85794675e-01 -5.38030326e-01
9.45097566e-01 -2.28971079e-01 -4.53396708e-01 5.88091731e-01
-9.74429190e-01 3.48262578e-01 5.57195842e-01 -6.52412593e-01
-9.06643793e-02 -4.75038320e-01 1.71507552e-01 -5.12960076e-01
4.64019537e-01 7.62093425e-01 9.44619402e-02 -2.35815257e-01
7.91882634e-01 1.20143361e-01 -2.58311301e-01 -6.93098485e-01
4.00992751e-01 7.94895470e-01 4.46969628e-01 -4.98435229e-01
3.93091083e-01 2.92525917e-01 1.29265234e-01 -6.08870089e-01
-1.23161888e+00 -4.43736494e-01 -4.60734189e-01 1.39476210e-01
2.36057043e-01 -5.18376231e-01 -6.93697155e-01 3.02957207e-01
-9.77338195e-01 -1.49666145e-01 -3.57887596e-01 5.72572827e-01
-6.27606988e-01 3.25177431e-01 -5.55367589e-01 -5.15377879e-01
-2.52468646e-01 -9.16912973e-01 9.46344733e-01 7.79099166e-02
-1.56384751e-01 -9.64701116e-01 3.21121186e-01 -1.13396652e-01
7.38768041e-01 7.68548716e-03 1.03622091e+00 -7.74446905e-01
-3.76009017e-01 -4.37269121e-01 -6.41017973e-01 6.74631059e-01
1.08345538e-01 2.35265121e-03 -9.22453463e-01 -4.57289964e-01
-1.17927752e-02 -3.79432201e-01 1.40712643e+00 2.80570507e-01
1.71167517e+00 -1.75000370e-01 -2.30169818e-01 6.76250875e-01
1.44688320e+00 -2.30458155e-01 5.23585975e-01 4.22889560e-01
3.25409293e-01 2.11856797e-01 4.67899889e-01 4.92981255e-01
3.33365619e-01 7.39702046e-01 7.05412105e-02 2.51687467e-01
-2.77829319e-02 -2.48391598e-01 -6.93501532e-02 8.22709739e-01
2.82305360e-01 -3.32266629e-01 -9.17568445e-01 2.09324598e-01
-1.69194376e+00 -6.73045456e-01 3.86433691e-01 2.41130304e+00
7.25276887e-01 -7.48221800e-02 2.96607316e-02 2.30809316e-01
5.27911842e-01 -2.74346527e-02 -5.39281011e-01 -4.49681759e-01
-2.32180819e-01 2.63696134e-01 3.76225561e-01 4.50057536e-01
-1.00013542e+00 9.01976943e-01 7.50693989e+00 9.74945366e-01
-1.29224563e+00 -5.74174114e-02 4.36432362e-01 -2.66566247e-01
-4.32154626e-01 -2.69683719e-01 -5.64304888e-01 3.53694707e-01
6.13624752e-01 -4.61214185e-01 3.44865590e-01 7.77479053e-01
-4.12543833e-01 2.90496737e-01 -1.62685335e+00 1.26727569e+00
2.05927864e-01 -8.50301683e-01 4.93307710e-01 6.14475692e-03
5.83433449e-01 3.02404091e-02 5.67516565e-01 5.23792863e-01
-4.54861671e-02 -1.16128218e+00 1.33002207e-01 5.73064864e-01
6.46407664e-01 -8.25070798e-01 6.50174260e-01 1.05918847e-01
-8.54196608e-01 -2.95635968e-01 -8.21981549e-01 7.91297182e-02
-3.28222871e-01 7.03131795e-01 -3.94244730e-01 2.21964002e-01
6.62766993e-01 7.28162050e-01 -8.10724437e-01 1.28322136e+00
-4.30032834e-02 2.45852128e-01 -4.04800087e-01 -1.14239469e-01
2.70165145e-01 5.77289090e-02 1.65499762e-01 1.27697766e+00
5.06909549e-01 -1.32865477e-02 3.40435766e-02 8.45218360e-01
-3.94459635e-01 1.67483687e-01 -5.16589880e-01 -1.66404903e-01
5.52628160e-01 9.46038604e-01 -4.00422186e-01 -1.38599470e-01
-3.41913342e-01 8.48121226e-01 5.50481439e-01 5.41662991e-01
-7.20750391e-01 -5.53544819e-01 7.53153861e-01 3.94895338e-02
-6.39855117e-03 -2.87762731e-01 -1.17141619e-01 -1.26371193e+00
4.12379473e-01 -5.49879372e-01 6.94707692e-01 -5.14326930e-01
-1.66954660e+00 3.82248968e-01 -3.67322192e-03 -1.37756515e+00
-9.10077319e-02 -9.16786432e-01 -1.96591765e-01 4.39307928e-01
-1.50718141e+00 -8.71138573e-01 -4.97231662e-01 4.26357239e-01
-7.63107883e-03 -1.03838868e-01 6.54710650e-01 5.36681831e-01
-4.99427706e-01 1.29309905e+00 5.22689283e-01 1.60211012e-01
9.54516470e-01 -1.07852399e+00 1.65934995e-01 2.78081477e-01
3.62618059e-01 8.06112230e-01 4.97930199e-01 8.79968777e-02
-1.09912753e+00 -8.65994096e-01 6.72595263e-01 -6.00128233e-01
7.54040182e-01 -5.13414666e-02 -1.00972271e+00 4.26545739e-01
-4.38492686e-01 1.22317173e-01 9.63396668e-01 3.17574233e-01
-9.34859395e-01 -4.42523718e-01 -1.08628273e+00 4.13906395e-01
1.10035610e+00 -8.56328130e-01 -3.69384855e-01 3.33737791e-01
8.88399243e-01 2.05928963e-02 -1.04136908e+00 8.17474544e-01
9.91172135e-01 -8.73728752e-01 9.28504944e-01 -9.72132683e-01
3.12869579e-01 2.48610601e-01 -6.88935041e-01 -1.25751936e+00
-3.09468418e-01 8.07272457e-03 -3.90337437e-01 9.73128080e-01
3.77597213e-01 -6.21237099e-01 6.94734275e-01 6.85141563e-01
1.71441957e-01 -1.01681530e+00 -9.19409156e-01 -1.36962533e+00
7.80964553e-01 -4.20781642e-01 8.60099792e-01 9.72456932e-01
2.37968341e-01 1.56440780e-01 -3.83492447e-02 -1.36456609e-01
4.14954484e-01 2.90285707e-01 5.36680162e-01 -1.49402964e+00
-1.37325570e-01 -9.49886918e-01 -9.50524211e-01 -1.35920835e+00
1.96195260e-01 -1.06060553e+00 -4.12541449e-01 -1.14107418e+00
5.88027120e-01 -5.31700552e-01 -1.09026325e+00 1.90208063e-01
-8.22829381e-02 5.36147237e-01 4.07026429e-03 4.97739255e-01
-8.37229490e-01 7.80801833e-01 1.06088305e+00 -3.07160646e-01
-3.06508075e-02 -1.44082159e-01 -7.78981328e-01 5.17828345e-01
9.23570514e-01 -4.23534870e-01 -3.03825349e-01 -7.93595970e-01
5.31429946e-01 -5.42542815e-01 4.29232329e-01 -9.56249833e-01
1.33421466e-01 -1.47505011e-02 3.69737029e-01 -3.38961244e-01
2.70127445e-01 -6.51100814e-01 -6.80176258e-01 4.65405643e-01
-6.91393971e-01 2.54073739e-01 1.36811407e-02 5.06141603e-01
-3.99389058e-01 -3.03803593e-01 7.30872035e-01 2.35667750e-01
-4.86347109e-01 5.33870578e-01 1.82609722e-01 3.26278716e-01
6.19407177e-01 -6.99634757e-03 -5.62113762e-01 -4.14558858e-01
-5.94981015e-01 3.05142462e-01 6.52257323e-01 5.57395518e-01
7.52438188e-01 -1.79303038e+00 -6.02886796e-01 -4.60825935e-02
7.30510414e-01 -4.39521790e-01 2.95987427e-02 6.88869655e-01
-4.75821882e-01 5.40546656e-01 7.79041201e-02 -8.13980758e-01
-7.81939745e-01 9.35679197e-01 3.93898308e-01 -5.40400296e-02
-3.19196463e-01 6.57155573e-01 3.38308603e-01 -2.32142881e-01
4.01820332e-01 -4.34370250e-01 8.54546204e-03 -1.21875986e-01
6.77311659e-01 5.55881917e-01 1.74255297e-01 -2.39018813e-01
-3.70294034e-01 7.18158484e-01 -3.08717072e-01 -1.17583778e-02
1.12095642e+00 -1.52717769e-01 -4.12121937e-02 6.23216927e-01
1.84620214e+00 -5.67125380e-01 -7.32024312e-01 -6.91243589e-01
1.60813063e-01 -6.77775383e-01 6.52915239e-02 -6.14205599e-01
-1.02890670e+00 1.12768102e+00 8.74728084e-01 2.71886557e-01
1.24135101e+00 2.29216024e-01 9.87462282e-01 9.75369632e-01
3.41081291e-01 -1.26452982e+00 2.36390457e-01 5.80769300e-01
8.06475461e-01 -1.46180511e+00 1.23799466e-01 -2.75871740e-03
-1.49124205e-01 1.04746890e+00 5.05799830e-01 -1.20079480e-01
8.58890295e-01 -9.74248499e-02 6.78006932e-02 -2.39572003e-01
-4.17971909e-01 -5.80752343e-02 5.50937116e-01 3.13313693e-01
5.28547645e-01 -1.03077814e-01 -6.60816908e-01 5.16387641e-01
-3.75515133e-01 4.23320420e-02 1.41377389e-01 6.01206660e-01
-5.92618346e-01 -9.61252391e-01 2.40427241e-01 5.45930862e-01
-3.37503195e-01 -1.11065634e-01 -7.45940745e-01 8.07817340e-01
-2.92312115e-01 7.07834601e-01 1.25721753e-01 -5.62042296e-01
2.65969127e-01 -4.45473231e-02 7.86557078e-01 -2.31033936e-01
-7.06831887e-02 -6.55925930e-01 -6.19171083e-01 -6.37466669e-01
-3.55560005e-01 -1.64404288e-01 -7.31553078e-01 -4.30244684e-01
-5.46499431e-01 3.04170758e-01 5.80284953e-01 8.42404723e-01
2.61071682e-01 -8.50041062e-02 9.73509252e-01 -3.57833833e-01
-1.15806937e+00 -1.04092991e+00 -7.50926077e-01 7.53170907e-01
4.98195350e-01 -7.54575312e-01 -6.64638340e-01 -5.94309866e-01] | [9.428749084472656, 3.0729188919067383] |
450a414f-6748-4bc0-b60c-68313a0a1779 | blind-image-deblurring-with-local-maximum | null | null | http://openaccess.thecvf.com/content_CVPR_2019/html/Chen_Blind_Image_Deblurring_With_Local_Maximum_Gradient_Prior_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Chen_Blind_Image_Deblurring_With_Local_Maximum_Gradient_Prior_CVPR_2019_paper.pdf | Blind Image Deblurring With Local Maximum Gradient Prior | Blind image deblurring aims to recover sharp image from a blurred one while the blur kernel is unknown. To solve this ill-posed problem, a great amount of image priors have been explored and employed in this area. In this paper, we present a blind deblurring method based on Local Maximum Gradient (LMG) prior. Our work is inspired by the simple and intuitive observation that the maximum value of a local patch gradient will diminish after the blur process, which is proved to be true both mathematically and empirically. This inherent property of blur process helps us to establish a new energy function. By introducing an liner operator to compute the Local Maximum Gradient, together with an effective optimization scheme, our method can handle various specific scenarios. Extensive experimental results illustrate that our method is able to achieve favorable performance against state-of-the-art algorithms on both synthetic and real-world images.
| [' Guixu Zhang', ' Tingting Wang', ' Faming Fang', 'Liang Chen'] | 2019-06-01 | null | null | null | cvpr-2019-6 | ['blind-image-deblurring'] | ['computer-vision'] | [ 3.11840832e-01 -4.76423711e-01 1.94303662e-01 -1.45482898e-01
-1.94197774e-01 -3.63395065e-01 6.11809373e-01 -4.40785229e-01
-1.89332515e-01 8.61782730e-01 2.37890735e-01 5.79334646e-02
-3.08683753e-01 -4.82955933e-01 -4.29088742e-01 -8.82543623e-01
4.09726836e-02 -3.28593016e-01 9.50525627e-02 -1.15340687e-01
5.71146667e-01 3.35610449e-01 -1.22220349e+00 -3.49332541e-01
1.30924869e+00 7.60187089e-01 6.61282897e-01 4.29688096e-01
1.76796362e-01 8.44680190e-01 -3.91012967e-01 -7.63990507e-02
3.28574240e-01 -7.10517764e-01 -7.20298648e-01 3.49315375e-01
2.62914896e-01 -3.54972869e-01 -4.06749219e-01 1.53477764e+00
4.63561267e-01 1.82238907e-01 6.47478402e-01 -7.86018431e-01
-1.02717733e+00 1.39063969e-01 -8.83291960e-01 4.03185576e-01
2.95178890e-01 9.27502364e-02 4.62312847e-01 -9.48651195e-01
3.05714518e-01 8.64573896e-01 6.83968544e-01 2.36555427e-01
-1.10068595e+00 -3.31431329e-01 -3.05984676e-01 4.35996890e-01
-1.44320977e+00 -4.71895486e-01 1.02540839e+00 -3.10618013e-01
1.52605206e-01 3.69131565e-01 3.67312580e-01 6.60169661e-01
2.98098803e-01 5.38143992e-01 1.67361403e+00 -4.84665036e-01
1.70748904e-01 1.38972942e-02 2.39612564e-01 5.69810271e-01
4.39711034e-01 1.41273022e-01 -3.15092295e-01 -6.98210374e-02
1.03320909e+00 8.64831805e-02 -1.20686436e+00 -3.63037854e-01
-1.21622491e+00 4.45510119e-01 6.00023687e-01 5.60135663e-01
-4.35592115e-01 4.70118076e-02 -2.45841354e-01 1.21405534e-01
5.47734320e-01 4.11254972e-01 -1.20869176e-02 -7.97644481e-02
-1.15991342e+00 2.89285742e-02 6.99484408e-01 4.74322259e-01
8.15016448e-01 -1.32599607e-01 -2.08938017e-01 9.44583356e-01
3.32324058e-01 5.59276998e-01 5.93855202e-01 -8.60948980e-01
7.83850849e-02 1.63402304e-01 5.69347978e-01 -1.11562431e+00
5.01521677e-02 -5.24912775e-01 -1.23623812e+00 3.45700383e-01
5.23155093e-01 -1.29727826e-01 -7.74286509e-01 1.44787729e+00
1.48776874e-01 6.16415203e-01 -8.72481540e-02 1.44692731e+00
3.38148892e-01 6.06593609e-01 -4.11189944e-01 -5.47012568e-01
1.26704967e+00 -9.26778674e-01 -1.02856994e+00 -2.82269418e-01
-2.94289231e-01 -1.03202140e+00 8.28224242e-01 3.53318751e-01
-1.05249012e+00 -5.85732460e-01 -1.09909773e+00 6.52936697e-02
3.37567031e-02 7.01754093e-02 5.36680281e-01 6.50861681e-01
-1.07588375e+00 6.23693705e-01 -6.68299198e-01 -3.00382435e-01
1.56490624e-01 -2.32642628e-02 -2.71670312e-01 -2.49293849e-01
-1.08574355e+00 1.11186099e+00 1.96724981e-01 5.13702035e-01
-5.85215867e-01 -3.61926585e-01 -6.27918363e-01 1.07005969e-01
2.77918577e-01 -9.15804982e-01 1.00312388e+00 -8.67340922e-01
-1.69789779e+00 6.54894769e-01 -3.87452751e-01 -3.31074774e-01
6.73686206e-01 -4.42108124e-01 -2.46874154e-01 2.76667923e-01
-1.30369142e-01 3.92897660e-03 1.43060708e+00 -1.44011462e+00
-5.13101481e-02 -2.98282087e-01 -2.90232897e-01 2.39564076e-01
-3.28414768e-01 -5.26058637e-02 -2.88938344e-01 -9.23582792e-01
3.24640483e-01 -6.54158533e-01 -1.64892003e-01 -1.80495501e-01
-2.76497275e-01 2.77932346e-01 8.22925866e-01 -8.59856963e-01
1.32024896e+00 -2.09739304e+00 2.80916423e-01 -2.60444265e-02
3.74240160e-01 3.37531656e-01 2.29169339e-01 3.11735481e-01
-2.00786740e-01 -2.26760328e-01 -7.21931338e-01 -1.16067752e-01
-2.73328781e-01 -1.02253422e-01 -4.02338713e-01 9.86006916e-01
-7.82760158e-02 8.07345033e-01 -9.95952487e-01 -2.52733618e-01
3.95035118e-01 6.08125031e-01 -2.81652808e-01 5.47265112e-01
2.78667182e-01 6.27654672e-01 -3.31484795e-01 2.88524806e-01
1.33408582e+00 -2.91589290e-01 -1.71207905e-01 -2.95185477e-01
-3.65288377e-01 -3.87035429e-01 -1.11247492e+00 1.56652248e+00
-3.07447493e-01 7.28975713e-01 3.70266020e-01 -8.91639650e-01
8.19263577e-01 1.74573392e-01 2.20153347e-01 -2.24997684e-01
1.77824944e-01 3.96823913e-01 -2.80393094e-01 -8.47084224e-01
5.15364110e-01 -4.44197178e-01 4.93722349e-01 3.22878003e-01
-3.58409166e-01 -4.20666218e-01 -1.18672758e-01 -6.49752170e-02
9.04126465e-01 9.32898819e-02 3.79793763e-01 -5.57559133e-01
9.64168847e-01 -4.04491901e-01 2.18220860e-01 7.58934796e-01
-2.89783627e-01 9.58559275e-01 -1.45356096e-02 -2.08101362e-01
-7.43933022e-01 -8.87592077e-01 -2.97792733e-01 2.09666774e-01
8.37235272e-01 1.65839434e-01 -1.05011022e+00 -2.79844910e-01
-2.26501673e-01 4.36952680e-01 -4.84090328e-01 -4.47999388e-02
-5.19669354e-01 -9.92484868e-01 6.26084805e-02 2.34973598e-02
1.12250173e+00 -6.82778895e-01 -6.04374468e-01 5.62956333e-02
-3.10064882e-01 -9.22407866e-01 -7.79309571e-01 -3.56196582e-01
-9.06291187e-01 -1.08006191e+00 -1.53790712e+00 -7.74650037e-01
8.24732602e-01 9.64636862e-01 6.86653078e-01 8.20865035e-02
-1.76254749e-01 1.93597555e-01 -2.79339492e-01 2.20898300e-01
-4.01432842e-01 -3.72798264e-01 -1.42416671e-01 4.58747476e-01
2.08672159e-03 -6.14834726e-01 -8.94331157e-01 4.82091516e-01
-1.12996840e+00 1.77503869e-01 7.74148047e-01 9.50018883e-01
2.23283246e-02 3.95993263e-01 4.19951469e-01 -3.39326590e-01
1.04777050e+00 -2.74334460e-01 -7.23688662e-01 2.77654827e-01
-7.90725470e-01 2.55420536e-01 4.96599555e-01 -3.99658620e-01
-1.47743201e+00 -1.82493702e-01 1.65300190e-01 -4.22614843e-01
-1.19011804e-01 4.72230345e-01 2.73921508e-02 -3.74438167e-01
6.18818343e-01 5.49218297e-01 1.45965904e-01 -7.57149935e-01
4.85283285e-01 9.79525566e-01 9.55629051e-01 -3.00476104e-01
1.04741657e+00 6.45963728e-01 3.70679013e-02 -1.00327814e+00
-5.83636165e-01 -6.68238342e-01 -3.03263664e-01 -1.68279693e-01
7.21915305e-01 -7.03383327e-01 -7.15350151e-01 1.09430265e+00
-1.18818867e+00 -1.21159174e-01 1.48248181e-01 5.76024354e-01
-3.66814256e-01 1.08052778e+00 -7.29881227e-01 -8.97231042e-01
-3.38608295e-01 -1.18109071e+00 6.99521899e-01 4.41698998e-01
2.42938846e-01 -1.07242119e+00 1.39520347e-01 2.81742603e-01
8.74776781e-01 2.46802289e-02 3.22178870e-01 2.77223974e-01
-7.35081911e-01 -3.91667411e-02 -6.51839018e-01 5.30641258e-01
5.65714538e-01 -4.74411875e-01 -9.76445079e-01 -4.63974863e-01
7.62210608e-01 2.66364634e-01 1.04400337e+00 5.41899443e-01
9.25875485e-01 -2.95913249e-01 -2.47926891e-01 7.43762255e-01
1.67971146e+00 -2.10075691e-01 8.88711751e-01 2.57636160e-01
5.70587754e-01 2.14997590e-01 4.29508120e-01 3.80352825e-01
-6.56227171e-02 7.60197461e-01 4.37160999e-01 -1.16209477e-01
-3.73614907e-01 1.07903153e-01 2.70223975e-01 6.23275697e-01
-2.66241074e-01 -2.18920782e-01 -5.87119758e-01 5.53326607e-01
-1.82535779e+00 -9.96272683e-01 -4.68122303e-01 2.37541723e+00
1.09996629e+00 -1.89641923e-01 -4.34130967e-01 1.08883940e-02
1.01344824e+00 4.10537630e-01 -1.88314304e-01 2.03541160e-01
-2.14962929e-01 9.27206576e-02 5.42407751e-01 7.87208438e-01
-1.11045301e+00 7.43058562e-01 6.32040167e+00 8.13165307e-01
-1.24461722e+00 5.44373281e-02 3.97016227e-01 5.63268125e-01
-1.33643582e-01 8.80372524e-02 -1.16217598e-01 7.56778717e-01
4.29124773e-01 -3.76758337e-01 8.02757859e-01 3.62938732e-01
4.77068663e-01 -4.45115417e-01 -5.62607825e-01 1.38018894e+00
1.44724384e-01 -1.01849699e+00 -3.96684378e-01 -5.24063595e-02
8.36388767e-01 -2.42533609e-01 1.75985783e-01 -4.04858530e-01
-8.37056115e-02 -9.24492180e-01 4.59628761e-01 8.00207198e-01
7.21618712e-01 -3.16932023e-01 8.18177938e-01 3.41432303e-01
-7.74888694e-01 7.49473125e-02 -1.89762414e-01 -2.39458576e-01
3.55660021e-01 1.14029050e+00 -5.48432231e-01 8.12896550e-01
5.01830101e-01 8.53012025e-01 -5.13333738e-01 1.55894959e+00
-5.22660673e-01 5.08839369e-01 5.33733191e-03 2.34504238e-01
1.73825063e-02 -7.10905790e-01 8.95975888e-01 1.17829072e+00
5.04504204e-01 1.69921398e-01 -4.30459827e-01 1.11876392e+00
1.10876270e-01 -1.70049742e-01 -3.14754874e-01 2.79660463e-01
9.74725634e-02 1.23595989e+00 -5.72114050e-01 -2.91587114e-01
-2.01608002e-01 1.56336963e+00 -1.37766391e-01 6.92463338e-01
-7.72657812e-01 -4.88590181e-01 4.79041308e-01 -2.77730916e-02
1.67342469e-01 -3.67038459e-01 -3.77396703e-01 -1.56191778e+00
9.65216979e-02 -6.39922023e-01 -1.37604162e-01 -1.13882875e+00
-1.33849776e+00 5.67558169e-01 -1.87492203e-02 -1.39708972e+00
-4.95929085e-02 -4.78012681e-01 -6.41775608e-01 1.27079487e+00
-1.95291233e+00 -7.80523896e-01 -8.89351487e-01 4.89362210e-01
4.56783384e-01 2.02547535e-01 3.54233801e-01 2.04487145e-01
-3.88500452e-01 -7.11645037e-02 3.33939880e-01 -4.88371924e-02
9.39418793e-01 -1.22513151e+00 2.15057060e-01 1.31878400e+00
-2.61602968e-01 8.85426819e-01 1.21449888e+00 -5.25319755e-01
-1.34588933e+00 -6.05895162e-01 5.09738088e-01 -1.22976065e-01
7.13801146e-01 2.56942719e-01 -1.04634058e+00 2.71473229e-01
7.02355146e-01 8.31872150e-02 3.39918002e-03 -5.72542548e-01
-3.84191982e-02 -9.14529338e-02 -1.15403521e+00 4.77906197e-01
7.30160058e-01 -3.31727535e-01 -9.28062618e-01 1.06066681e-01
3.36282104e-01 -4.43632334e-01 -4.32883054e-01 4.16735083e-01
3.03577632e-01 -1.28529918e+00 9.85943139e-01 1.54690936e-01
4.04808074e-01 -7.34551370e-01 1.99062556e-01 -1.58612633e+00
-4.26564127e-01 -1.08172417e+00 -3.44625324e-01 9.82174873e-01
-1.58495069e-01 -9.67101574e-01 2.55094558e-01 3.88994604e-01
-7.87404403e-02 -4.36059833e-01 -5.47835052e-01 -8.01266789e-01
-2.84698188e-01 9.33984518e-02 2.68619776e-01 1.02511466e+00
-2.74626259e-02 3.31991315e-02 -6.44390106e-01 4.31523860e-01
9.95691419e-01 2.61618286e-01 5.53131163e-01 -9.84753728e-01
-3.41958940e-01 -4.91236955e-01 -2.70838708e-01 -1.57379651e+00
-1.17638029e-01 -3.71319890e-01 2.19737291e-01 -1.50889468e+00
4.09484684e-01 -2.61814207e-01 -1.94350868e-01 -1.92568116e-02
-6.73371375e-01 3.05626959e-01 3.74323153e-03 5.23190081e-01
-1.19086139e-01 6.05982125e-01 1.43639100e+00 -6.71101063e-02
-1.47624910e-01 1.07207462e-01 -6.20482504e-01 6.99453890e-01
6.62504077e-01 -1.57322884e-01 -2.44481191e-01 -4.74541962e-01
-1.01724789e-01 -1.14530232e-03 5.90875864e-01 -9.93164718e-01
3.31078440e-01 -1.42883927e-01 1.18652113e-01 -2.31770024e-01
3.43027741e-01 -8.73361230e-01 1.72828481e-01 3.16017538e-01
4.64491732e-02 -2.94824034e-01 -9.70674604e-02 7.05258548e-01
-3.93576711e-01 -5.05476594e-01 1.06858087e+00 -1.35835245e-01
-6.79207861e-01 -5.06839715e-02 -6.98141083e-02 -2.01576531e-01
8.28835130e-01 -1.65280685e-01 -4.71335769e-01 -6.05077744e-01
-3.31941932e-01 -1.32127389e-01 8.52260411e-01 1.64278150e-01
7.19761252e-01 -1.02290118e+00 -7.20859051e-01 3.13381702e-01
-2.66692191e-01 -3.01102817e-01 2.84035861e-01 1.19807196e+00
-8.38735044e-01 1.36749730e-01 -2.96636075e-02 -4.48951542e-01
-1.06548023e+00 7.55223691e-01 4.65905666e-01 7.97632784e-02
-5.51236033e-01 7.07873166e-01 1.65012628e-01 2.59544611e-01
-1.07924849e-01 -3.27363759e-01 -5.92504106e-02 -3.84658009e-01
7.35252976e-01 5.78781188e-01 -2.06039131e-01 -7.07342744e-01
-2.08181188e-01 9.15634930e-01 2.22778335e-01 -2.16739073e-01
1.21798849e+00 -5.83851099e-01 -5.78475773e-01 -1.56316124e-02
1.19007051e+00 4.27058190e-01 -1.51569104e+00 -1.67264864e-01
-5.83000407e-02 -9.28278685e-01 4.27695602e-01 -5.86769164e-01
-1.01072037e+00 6.08914375e-01 7.28423715e-01 5.81867099e-01
1.41753316e+00 -2.40032762e-01 9.26111758e-01 -1.40746012e-01
3.16913337e-01 -6.15247488e-01 1.04095545e-02 1.43261209e-01
9.86168683e-01 -1.27512765e+00 1.90572977e-01 -6.27512932e-01
-2.08115965e-01 1.13584042e+00 6.55547678e-02 -1.11799680e-01
5.23559093e-01 -3.20260599e-02 5.41279837e-02 9.69931483e-02
2.39305217e-02 -2.36782894e-01 3.22955638e-01 2.32407153e-01
3.09955627e-01 -2.30287567e-01 -6.33603513e-01 1.36177525e-01
2.37113729e-01 3.97992492e-01 6.69541836e-01 7.33407617e-01
-6.41317666e-01 -9.78310525e-01 -9.06963885e-01 -1.67987049e-01
-5.56189120e-01 -2.67792583e-01 4.52833343e-03 3.88619274e-01
-9.84808952e-02 1.18247056e+00 -3.17730039e-01 7.09508583e-02
-9.21955183e-02 -2.54646510e-01 5.85768044e-01 -1.25016958e-01
5.12982868e-02 2.81245470e-01 -4.83827174e-01 -2.21484497e-01
-7.75968969e-01 -4.29189891e-01 -8.57940018e-01 -1.00304507e-01
-5.93629658e-01 2.41762295e-01 5.77158034e-01 8.19329679e-01
2.06357986e-01 1.98035523e-01 7.95425653e-01 -9.92365658e-01
-6.49295747e-01 -1.11448216e+00 -7.63812721e-01 5.10277927e-01
8.71128261e-01 -5.32112837e-01 -8.35892618e-01 3.79971683e-01] | [11.595423698425293, -2.7497875690460205] |
22f575ea-9f08-4062-9691-f6323a14f66e | dense-regression-network-for-video-grounding | 2004.03545 | null | https://arxiv.org/abs/2004.03545v1 | https://arxiv.org/pdf/2004.03545v1.pdf | Dense Regression Network for Video Grounding | We address the problem of video grounding from natural language queries. The key challenge in this task is that one training video might only contain a few annotated starting/ending frames that can be used as positive examples for model training. Most conventional approaches directly train a binary classifier using such imbalance data, thus achieving inferior results. The key idea of this paper is to use the distances between the frame within the ground truth and the starting (ending) frame as dense supervisions to improve the video grounding accuracy. Specifically, we design a novel dense regression network (DRN) to regress the distances from each frame to the starting (ending) frame of the video segment described by the query. We also propose a simple but effective IoU regression head module to explicitly consider the localization quality of the grounding results (i.e., the IoU between the predicted location and the ground truth). Experimental results show that our approach significantly outperforms state-of-the-arts on three datasets (i.e., Charades-STA, ActivityNet-Captions, and TACoS). | ['Peihao Chen', 'Haoming Xu', 'Runhao Zeng', 'Mingkui Tan', 'Wenbing Huang', 'Chuang Gan'] | 2020-04-07 | dense-regression-network-for-video-grounding-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Zeng_Dense_Regression_Network_for_Video_Grounding_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Zeng_Dense_Regression_Network_for_Video_Grounding_CVPR_2020_paper.pdf | cvpr-2020-6 | ['video-grounding', 'natural-language-moment-retrieval'] | ['computer-vision', 'computer-vision'] | [ 5.61501123e-02 9.03912783e-02 -5.83269596e-01 -4.62412477e-01
-9.30936515e-01 -3.90520602e-01 4.21468019e-01 -7.71564171e-02
-4.41212147e-01 6.28717124e-01 5.95108047e-02 -3.87539677e-02
5.03111124e-01 -6.74895525e-01 -1.36758780e+00 -5.61507046e-01
6.23579882e-03 3.04796010e-01 5.72054446e-01 8.30008741e-03
-8.09095684e-04 1.70763299e-01 -1.28751111e+00 3.90681922e-01
6.22689068e-01 1.40514779e+00 2.52130240e-01 4.32873875e-01
1.85866579e-01 1.20959651e+00 -7.10468113e-01 -3.66809607e-01
2.46940434e-01 -5.14460564e-01 -6.42624855e-01 2.49793172e-01
7.72912741e-01 -6.81965768e-01 -6.83515251e-01 8.93841147e-01
1.30725786e-01 1.81138948e-01 2.85537213e-01 -1.59180021e+00
-3.30877364e-01 3.76559943e-01 -6.55745029e-01 4.22300518e-01
6.15300655e-01 4.16457281e-02 1.00597727e+00 -8.66348326e-01
8.19502532e-01 1.00532413e+00 6.41967654e-01 4.22203094e-01
-8.43152225e-01 -6.97699130e-01 2.80349523e-01 5.48976123e-01
-1.60536206e+00 -5.51711679e-01 6.36805594e-01 -4.91574049e-01
6.56690955e-01 -2.28669997e-02 5.68851948e-01 1.14381814e+00
-9.01119262e-02 8.69793832e-01 6.78862572e-01 -3.93958569e-01
2.77213335e-01 -8.29869807e-02 -1.77442163e-01 9.48378265e-01
1.25316642e-02 6.68287426e-02 -7.80869246e-01 7.20804483e-02
7.32077718e-01 1.13541193e-01 -4.75219756e-01 -4.97302532e-01
-1.47027159e+00 6.51144385e-01 6.57901287e-01 1.79412112e-01
-3.81732643e-01 3.82980555e-01 9.91598442e-02 -7.66865611e-02
5.35341084e-01 1.29904196e-01 -2.66466498e-01 -9.31743309e-02
-1.28729010e+00 1.53162524e-01 5.87129354e-01 1.11794376e+00
1.00609064e+00 -1.56957388e-01 -3.10488731e-01 3.53154778e-01
2.61537492e-01 3.14514577e-01 2.59580612e-01 -1.00947297e+00
8.64979267e-01 5.30941248e-01 4.35285598e-01 -1.13604724e+00
-1.83871314e-02 -2.97699660e-01 -3.41136664e-01 -2.16375425e-01
7.52043009e-01 -2.73466527e-01 -7.83818722e-01 1.61323154e+00
3.88239175e-01 8.04927051e-01 -1.72869667e-01 1.30444360e+00
7.58824408e-01 9.44880545e-01 -6.16788893e-05 -4.77540046e-02
8.59843612e-01 -1.29669952e+00 -5.43859780e-01 -5.10615826e-01
8.23377252e-01 -4.80083823e-01 9.10039723e-01 1.55213084e-02
-1.00276113e+00 -6.02085292e-01 -1.06469393e+00 1.12223275e-01
-1.13591421e-02 4.00567412e-01 3.39677095e-01 -1.54990796e-02
-8.89897168e-01 4.97243911e-01 -9.36806679e-01 -3.23642105e-01
3.02322537e-01 9.28931832e-02 -3.55822384e-01 -2.78905243e-01
-1.03240585e+00 6.05733275e-01 5.17052710e-01 1.38506874e-01
-1.11167228e+00 -6.28198266e-01 -9.22504783e-01 -1.28044277e-01
8.13428164e-01 -4.36008811e-01 1.14362490e+00 -1.34862721e+00
-9.76960659e-01 9.41182792e-01 -1.52449623e-01 -7.82390177e-01
6.95939064e-01 -3.32917511e-01 -1.24173544e-01 3.27169776e-01
4.36559886e-01 1.13424373e+00 6.77197516e-01 -1.21603203e+00
-9.82747376e-01 -1.38898417e-01 3.37632984e-01 1.47636771e-01
-2.96220258e-02 -8.97107050e-02 -9.21237230e-01 -5.29561400e-01
1.88104033e-01 -1.03748417e+00 -2.94405278e-02 1.96878731e-01
-5.57182252e-01 -1.49281323e-01 8.03292155e-01 -8.35077465e-01
1.22154403e+00 -2.16563892e+00 -1.98099241e-02 -8.18183925e-03
2.09023450e-02 2.13096291e-01 8.09572730e-03 9.18422416e-02
1.48883928e-02 8.26489646e-03 2.57579327e-01 -2.04931542e-01
-2.82986879e-01 2.65387177e-01 -4.55797255e-01 6.90069735e-01
1.35146514e-01 9.24667656e-01 -1.12496638e+00 -6.53058946e-01
1.85347140e-01 3.01469803e-01 -6.41847372e-01 6.17573261e-01
-4.79778498e-01 5.44001758e-01 -3.55607837e-01 6.38802528e-01
2.10869595e-01 -5.45011818e-01 -1.75481513e-01 -4.80466604e-01
-3.83285875e-03 2.41070524e-01 -1.03716481e+00 1.87849486e+00
-2.12629005e-01 8.43152821e-01 -4.74844277e-01 -8.57233524e-01
6.92544341e-01 2.73381919e-01 6.94624722e-01 -6.62236273e-01
-4.88793328e-02 -1.10100731e-01 -1.86412558e-01 -4.67122883e-01
3.60522002e-01 3.03788632e-01 9.99485329e-02 1.48740426e-01
1.08544834e-01 3.29943091e-01 2.05710769e-01 9.39304158e-02
1.06161773e+00 4.65420753e-01 2.21890315e-01 2.19024554e-01
2.21206427e-01 2.02633351e-01 8.57752323e-01 7.47162342e-01
-3.52169693e-01 9.61723328e-01 7.08225846e-01 -6.26415849e-01
-1.17486906e+00 -6.85716689e-01 3.04478526e-01 1.12575531e+00
5.85401297e-01 -5.23787856e-01 -1.02494252e+00 -8.50799263e-01
-3.57549608e-01 5.83652079e-01 -7.04896331e-01 -9.12292302e-03
-6.43527806e-01 -1.29445463e-01 4.33103800e-01 7.04532564e-01
7.94783533e-01 -9.08089638e-01 -6.44499779e-01 4.20923010e-02
-7.95336246e-01 -1.62653029e+00 -7.09794641e-01 -1.41989559e-01
-6.75859392e-01 -1.06030142e+00 -5.96962273e-01 -5.74356675e-01
4.68696773e-01 4.73549575e-01 1.18329847e+00 2.73402452e-01
1.26456484e-01 1.25003159e-01 -4.90019351e-01 -4.79537882e-02
-2.07068488e-01 -2.65814923e-03 -1.76388934e-01 2.74933130e-01
3.15320104e-01 4.80749644e-02 -7.58234739e-01 7.65739739e-01
-6.65427864e-01 4.04639959e-01 2.51686543e-01 6.48051679e-01
9.22373056e-01 -2.84277827e-01 3.04518104e-01 -5.20864964e-01
-2.01153889e-01 -6.06245577e-01 -5.71402133e-01 3.50354940e-01
-2.41810948e-01 -2.13029757e-01 4.32697862e-01 -3.76440942e-01
-6.01584554e-01 2.80494541e-01 5.14133982e-02 -8.76225710e-01
-1.67341679e-01 3.79040122e-01 -2.00141147e-01 1.50344506e-01
6.76445186e-01 4.74405661e-02 -1.88610479e-01 -1.12070307e-01
1.65968284e-01 5.17734408e-01 7.03244388e-01 -3.12090784e-01
6.36156142e-01 5.82774103e-01 -2.33926043e-01 -6.53020382e-01
-1.39952838e+00 -6.68004632e-01 -5.40179968e-01 -4.55069661e-01
1.00733936e+00 -1.28002572e+00 -3.69667888e-01 1.20979361e-01
-1.27462435e+00 -4.61982936e-01 -1.39039129e-01 5.17709434e-01
-7.66105235e-01 3.43999833e-01 -3.66814733e-01 -4.78363901e-01
-1.78768244e-02 -1.15057003e+00 1.42535043e+00 2.46483058e-01
-2.56765395e-01 -7.14739740e-01 -1.38415202e-01 5.06973803e-01
-2.07795098e-01 5.34939945e-01 3.62058342e-01 -5.60917079e-01
-9.73735332e-01 -4.88534570e-01 -3.49051654e-01 1.65390298e-01
-1.58834115e-01 9.77973044e-02 -9.32244599e-01 -1.74274996e-01
-2.09630966e-01 -3.10219795e-01 6.53471529e-01 4.72913712e-01
1.16177952e+00 -3.62928241e-01 -4.67727900e-01 8.23314726e-01
1.25013924e+00 5.29172383e-02 8.70089650e-01 4.52281743e-01
8.61921847e-01 2.68605173e-01 1.29886448e+00 3.53684932e-01
4.15941834e-01 1.13520837e+00 6.91980839e-01 -2.09414572e-01
-1.29091859e-01 -6.53737009e-01 5.41143894e-01 2.57071197e-01
-4.36029732e-02 -5.43573201e-01 -8.73577476e-01 5.61892092e-01
-2.19028211e+00 -1.10523498e+00 5.91732040e-02 2.29456878e+00
5.51735759e-01 7.66813979e-02 1.56281069e-01 -1.21035628e-01
9.11341608e-01 3.51459235e-01 -3.93961340e-01 3.61637294e-01
2.69504525e-02 -3.98435295e-01 5.59810936e-01 3.16652477e-01
-1.31044364e+00 1.06110454e+00 5.75316954e+00 6.01782501e-01
-1.04957771e+00 1.45621285e-01 9.39950705e-01 -2.11300388e-01
2.49540091e-01 -2.86141690e-02 -9.01941776e-01 7.11032391e-01
9.93447661e-01 1.39659926e-01 3.70791405e-01 1.00704026e+00
5.12375891e-01 -3.49410981e-01 -1.55480623e+00 1.11201668e+00
2.60325342e-01 -1.44743025e+00 -1.54196873e-01 -6.82200566e-02
8.46843660e-01 1.37122199e-01 -2.54895151e-01 1.66543365e-01
-5.39550185e-02 -7.93107688e-01 1.09580469e+00 3.65669757e-01
9.20906901e-01 -4.08957005e-01 8.10166955e-01 3.91723573e-01
-1.01677620e+00 1.51890546e-01 -3.09204191e-01 8.02460760e-02
1.95347056e-01 3.52023453e-01 -9.69832838e-01 2.87962258e-01
8.95136476e-01 9.44474816e-01 -5.52078068e-01 1.18382597e+00
-3.49811137e-01 6.67895913e-01 -3.54834497e-01 2.01277509e-01
4.70668495e-01 -4.77720127e-02 3.35192591e-01 9.78687763e-01
4.43976790e-01 -1.40264407e-01 4.72127736e-01 7.45078206e-01
-1.75312132e-01 -2.16802675e-02 -7.07886279e-01 1.20024793e-01
4.03887272e-01 9.56597745e-01 -7.36133397e-01 -3.79087597e-01
-6.07986867e-01 9.10183728e-01 3.55979860e-01 5.69262326e-01
-1.36484897e+00 2.55350545e-02 5.79292893e-01 5.82383037e-01
4.70235586e-01 -1.45403864e-02 1.19964249e-01 -1.29940546e+00
2.99150556e-01 -7.08787978e-01 4.01652306e-01 -1.24917459e+00
-7.90313780e-01 5.15681267e-01 -1.36489244e-02 -1.54048681e+00
-5.47319114e-01 -2.76158601e-01 -4.90096688e-01 4.81174201e-01
-1.34461272e+00 -1.06858385e+00 -8.15876186e-01 5.18639326e-01
5.13016701e-01 2.82563083e-03 3.36389244e-01 4.36149418e-01
-5.64644456e-01 5.07519245e-01 -1.97752014e-01 5.48721194e-01
7.67766893e-01 -9.20824230e-01 2.97654510e-01 1.01825655e+00
5.20742238e-01 4.78118435e-02 7.05309868e-01 -6.13873303e-01
-1.09050274e+00 -1.39852047e+00 8.33714664e-01 -5.89216471e-01
6.51444495e-01 -2.72196203e-01 -6.88694060e-01 1.03372717e+00
-3.11362259e-02 4.44211453e-01 3.54315698e-01 -2.52207398e-01
-1.93513528e-01 -1.88812703e-01 -7.86031961e-01 6.12676501e-01
9.77947593e-01 -3.92766923e-01 -3.02543342e-01 4.91568685e-01
7.21718550e-01 -7.59663403e-01 -3.91087681e-01 2.93229252e-01
4.10234958e-01 -1.22072661e+00 9.80068147e-01 -5.66788077e-01
6.70650184e-01 -5.21315455e-01 -3.42206657e-01 -1.06961846e+00
1.62220702e-01 -5.74863553e-01 -3.93783599e-01 1.03928363e+00
2.43940771e-01 1.72650084e-01 1.12185299e+00 6.13043904e-01
-7.98433796e-02 -8.30069721e-01 -1.04076266e+00 -6.22144222e-01
-5.76405942e-01 -4.47115511e-01 4.62410748e-01 8.14678967e-01
-1.66032642e-01 3.18510354e-01 -6.22634113e-01 2.30568498e-01
3.49531591e-01 5.22987638e-03 1.02989733e+00 -8.64203572e-01
-1.85714379e-01 1.86052009e-01 -8.37018669e-01 -1.52262282e+00
2.34709188e-01 -5.10404229e-01 3.79457265e-01 -1.51690614e+00
1.64445564e-01 -3.75503242e-01 -1.22197822e-01 4.40865457e-01
-6.64693564e-02 3.67448628e-01 2.46111959e-01 4.18379903e-01
-1.15712047e+00 3.06522191e-01 1.07214582e+00 -1.11009561e-01
-1.43392339e-01 6.28535822e-02 -2.18774244e-01 8.40260446e-01
5.30848742e-01 -6.32284641e-01 -3.04628938e-01 -5.09379506e-01
1.55512378e-01 3.51719618e-01 6.74071014e-01 -1.24199522e+00
3.18348020e-01 -1.62159070e-01 4.40628529e-01 -7.67123818e-01
3.51133853e-01 -1.00947964e+00 1.29060820e-01 7.14390427e-02
-4.49586630e-01 1.08535521e-01 -8.35711583e-02 7.90435910e-01
-2.82980889e-01 -2.61020392e-01 6.39491737e-01 4.78393249e-02
-9.08017218e-01 5.02260447e-01 1.09066755e-01 3.05519342e-01
1.39619052e+00 -3.31334978e-01 -3.02907795e-01 -6.85713112e-01
-5.62379539e-01 2.13539362e-01 7.10471392e-01 4.21763748e-01
5.77651262e-01 -1.36089945e+00 -5.61158121e-01 -6.13261946e-02
2.83610731e-01 1.56352401e-01 -6.59245849e-02 8.54537666e-01
-6.98564172e-01 4.72362161e-01 4.29794528e-02 -9.72587466e-01
-1.10517144e+00 6.56506956e-01 5.73884070e-01 -1.62838414e-01
-5.97860038e-01 7.51502037e-01 4.03993785e-01 2.24808782e-01
5.86118639e-01 -5.81169724e-01 1.64965346e-01 -5.67032658e-02
5.97981453e-01 2.72656709e-01 1.78653989e-02 -9.13501561e-01
-3.13524008e-01 1.95092231e-01 1.48278531e-02 8.78723059e-03
1.08256280e+00 -1.15242921e-01 1.96298286e-01 3.72803777e-01
1.28921032e+00 -3.56578618e-01 -1.94710195e+00 -4.36605841e-01
-3.04144733e-02 -7.02788174e-01 3.54835689e-02 -4.67377156e-01
-1.17508101e+00 7.88363218e-01 6.23221040e-01 -9.61170197e-02
8.09835553e-01 2.49241978e-01 7.96088338e-01 3.48939300e-01
5.15095294e-01 -1.08923912e+00 2.93770075e-01 1.34820104e-01
6.92698717e-01 -1.49820292e+00 -1.98704258e-01 -3.65201712e-01
-6.69903696e-01 8.77266645e-01 8.73472154e-01 -8.08660164e-02
2.16619104e-01 -3.83969918e-02 7.04063708e-03 1.91946304e-03
-5.38924217e-01 -1.40239686e-01 3.21429372e-01 5.18729448e-01
3.16871673e-01 -2.30497569e-01 3.37360084e-01 3.12677622e-01
1.27468675e-01 1.67220935e-01 4.83339071e-01 7.65346169e-01
-5.19044518e-01 -5.29071510e-01 -3.73079330e-01 3.66306335e-01
-4.60678935e-01 1.73899978e-02 -3.69093150e-01 8.01439524e-01
6.80233166e-02 9.80568826e-01 2.62341112e-01 -4.43991095e-01
2.51008809e-01 -1.82010099e-01 2.82899350e-01 -6.87081695e-01
-2.79461682e-01 1.61482379e-01 1.31034916e-02 -9.64947760e-01
-5.97415030e-01 -5.46665013e-01 -1.26370215e+00 -3.29917401e-01
-4.78753358e-01 1.35077670e-01 3.46715361e-01 9.64649558e-01
2.74700016e-01 3.68326455e-01 5.39171338e-01 -1.01767886e+00
-2.48176396e-01 -7.91370273e-01 -2.44574681e-01 5.51571250e-01
3.60041291e-01 -6.30043805e-01 -3.04250956e-01 4.96843606e-01] | [9.887646675109863, 0.6968645453453064] |
8cd88620-06cb-430e-8e45-e15cc2bbe6f7 | tilt-then-blur-or-blur-then-tilt-clarifying | 2207.06377 | null | https://arxiv.org/abs/2207.06377v3 | https://arxiv.org/pdf/2207.06377v3.pdf | Tilt-then-Blur or Blur-then-Tilt? Clarifying the Atmospheric Turbulence Model | Imaging at a long distance often requires advanced image restoration algorithms to compensate for the distortions caused by atmospheric turbulence. However, unlike many standard restoration problems such as deconvolution, the forward image formation model of the atmospheric turbulence does not have a simple expression. Thanks to the Zernike representation of the phase, one can show that the forward model is a combination of tilt (pixel shifting due to the linear phase terms) and blur (image smoothing due to the high order aberrations). Confusions then arise between the ordering of the two operators. Should the model be tilt-then-blur, or blur-then-tilt? Some papers in the literature say that the model is tilt-then-blur, whereas more papers say that it is blur-then-tilt. This paper clarifies the differences between the two and discusses why the tilt-then-blur is the correct model. Recommendations are given to the research community. | ['Stanley H. Chan'] | 2022-07-13 | null | null | null | null | ['image-smoothing'] | ['computer-vision'] | [ 4.60508943e-01 -3.40519637e-01 6.63833082e-01 -2.75556147e-01
2.23451644e-01 -5.05943894e-01 6.02320969e-01 -3.56519759e-01
-2.60160565e-01 5.96020699e-01 3.32285821e-01 -6.76218033e-01
-4.27487969e-01 -3.44215363e-01 -3.63955468e-01 -1.12856102e+00
2.27604926e-01 7.23169744e-02 1.80065870e-01 -9.60506219e-03
5.65511346e-01 3.19261044e-01 -1.32098520e+00 2.59624403e-02
9.73692834e-01 5.98609924e-01 4.62794572e-01 8.96864057e-01
-2.61197478e-01 8.66757870e-01 -5.75264752e-01 -1.37619540e-01
4.59453672e-01 -7.50160515e-01 -7.19974160e-01 4.37985629e-01
4.79753822e-01 -4.57916141e-01 -2.57458031e-01 1.34360826e+00
1.06628798e-01 1.36856034e-01 4.77493376e-01 -6.94916785e-01
-7.69570529e-01 -2.29909867e-02 -6.32506192e-01 3.88847888e-01
2.68516004e-01 1.37458250e-01 3.44198346e-01 -8.11539412e-01
4.50744450e-01 1.09101808e+00 8.68309379e-01 6.24280758e-02
-1.15530622e+00 -1.77301019e-01 -2.56369025e-01 1.61544248e-01
-1.16344988e+00 -3.46865088e-01 3.52764308e-01 -7.51626015e-01
6.26111686e-01 6.56246603e-01 6.64179802e-01 1.60155132e-01
4.84787941e-01 -2.75765620e-02 1.82641530e+00 -6.22587621e-01
-1.21534750e-01 1.53978899e-01 3.27927977e-01 3.11316997e-01
7.52535522e-01 3.73395920e-01 -3.28675270e-01 -6.45081624e-02
9.55821097e-01 -1.50523484e-01 -1.02631426e+00 -1.20014377e-01
-9.63076711e-01 3.25304151e-01 2.23002702e-01 4.21392083e-01
-4.12644923e-01 1.40865240e-03 -9.49291363e-02 3.85311216e-01
4.35148329e-01 8.75501633e-01 -2.13694558e-01 -1.66008621e-01
-1.17223239e+00 4.03254896e-01 7.74654329e-01 4.59979981e-01
6.56363428e-01 6.51498437e-02 1.78725809e-01 4.41563576e-01
5.53823411e-01 5.72319567e-01 3.36453527e-01 -1.14795554e+00
-1.41647294e-01 -7.37647265e-02 5.81595957e-01 -8.00078332e-01
-1.32559001e-01 -5.73540151e-01 -7.96060622e-01 6.87008798e-01
7.97625422e-01 -1.83876365e-01 -8.83506715e-01 1.12794614e+00
-4.24044840e-02 3.60104620e-01 1.28124818e-01 1.05208194e+00
4.67054635e-01 5.50152540e-01 -5.57522595e-01 -7.23634481e-01
1.25557172e+00 -9.23333764e-01 -1.18683255e+00 -3.07038337e-01
2.07290664e-01 -1.45077229e+00 5.84957123e-01 4.38560039e-01
-1.12824845e+00 -4.19043958e-01 -1.18609917e+00 1.53929740e-03
1.03603922e-01 -3.38673830e-01 2.76664019e-01 5.59134841e-01
-1.17196965e+00 8.84245336e-01 -6.02229655e-01 -4.46507961e-01
-3.02861154e-01 -1.23008832e-01 -3.17827687e-02 -2.20437422e-01
-6.29166961e-01 1.23815846e+00 -2.78886080e-01 4.34078664e-01
1.80922467e-02 -7.11299896e-01 -3.93204719e-01 -9.96037126e-02
-1.09802425e-01 -7.79822767e-01 1.23691189e+00 -9.76462305e-01
-1.43799686e+00 6.67884946e-01 -6.04519904e-01 -3.52066606e-01
6.82778358e-01 -5.60548127e-01 -2.30615914e-01 2.07519099e-01
-1.71205968e-01 -3.76847148e-01 9.64034677e-01 -1.70479572e+00
-4.58864242e-01 -1.17673531e-01 -2.21076086e-02 4.97278154e-01
5.84936798e-01 1.17207274e-01 -8.67741480e-02 -3.92736495e-01
6.95114374e-01 -1.04221654e+00 -2.15174004e-01 -1.31441146e-01
-3.08111995e-01 4.96880203e-01 1.03653312e+00 -7.82903314e-01
1.14984620e+00 -2.37733293e+00 -4.04143125e-01 -1.09276034e-01
2.12808475e-01 2.36353114e-01 2.31851980e-01 5.66622913e-01
-4.78457510e-01 -1.18107416e-01 -1.86569124e-01 -1.01900697e-01
-3.52236331e-01 2.07316920e-01 -5.78479826e-01 7.33416736e-01
-2.37535596e-01 2.93550491e-01 -7.66869962e-01 1.21750541e-01
3.45563680e-01 5.15095830e-01 -1.85074002e-01 6.66485429e-02
4.22958195e-01 6.65359139e-01 4.54265475e-02 -3.14079635e-02
1.28858507e+00 2.08576247e-02 9.61766392e-02 -4.91211593e-01
-8.39582205e-01 1.70385003e-01 -1.11501014e+00 7.67858863e-01
-6.47102445e-02 1.06137562e+00 4.89726841e-01 -4.72439855e-01
5.47297895e-01 4.94015336e-01 1.84503675e-01 -2.96091825e-01
-2.21109197e-01 6.42644346e-01 4.02294427e-01 -7.22940743e-01
6.34054363e-01 -6.58628702e-01 8.79622817e-01 8.64258781e-02
-5.74905574e-01 -5.40344059e-01 1.61874011e-01 -6.58269180e-03
9.79914010e-01 1.01073407e-01 1.61327153e-01 -6.53970420e-01
3.65775526e-01 8.58915001e-02 2.78158128e-01 8.30535829e-01
-1.73517466e-01 8.19912553e-01 2.16394171e-01 -4.69124943e-01
-8.70857418e-01 -8.15752208e-01 -3.58554661e-01 -5.24237230e-02
5.16258895e-01 -3.20268005e-01 -8.59508932e-01 2.08655238e-01
-2.65842468e-01 7.45887578e-01 -1.46531001e-01 8.99923295e-02
-4.16334748e-01 -9.74853337e-01 1.38365507e-01 -9.05852318e-02
6.90486968e-01 -4.79835868e-01 -1.00626826e+00 -8.89667645e-02
-4.11850959e-01 -1.16055059e+00 -3.57982010e-01 7.32280361e-03
-9.97260451e-01 -1.19094920e+00 -6.94418073e-01 -3.80368531e-01
7.71352649e-01 8.20740759e-01 8.92650843e-01 4.41372931e-01
-4.41663377e-02 2.96986938e-01 -1.63628772e-01 -2.37540513e-01
-3.84360611e-01 -1.20198166e+00 1.01119667e-01 -3.92373763e-02
1.99361876e-01 -4.53585267e-01 -7.25871146e-01 1.40390113e-01
-7.94610441e-01 1.59991160e-01 4.64367926e-01 7.05828846e-01
2.35763371e-01 5.90301692e-01 -3.93326044e-01 -6.74238563e-01
5.47740042e-01 2.72115827e-01 -6.20052218e-01 -2.20946476e-01
-9.17960107e-01 -1.37513012e-01 2.07778081e-01 -1.49328396e-01
-1.26221311e+00 -3.40104789e-01 1.23791926e-01 -3.83111686e-01
-3.27760279e-01 4.99075800e-01 1.41222730e-01 -3.72620821e-01
3.65999877e-01 2.08176360e-01 2.26272922e-02 -6.65043116e-01
-9.86762345e-03 5.02728701e-01 5.46383262e-01 5.93674220e-02
1.16004181e+00 7.09711373e-01 2.70387173e-01 -1.44743156e+00
-7.92145669e-01 -4.32391286e-01 -4.61450249e-01 -3.15982193e-01
1.11429870e+00 -7.14497328e-01 -3.16400051e-01 8.92459512e-01
-1.24197829e+00 -3.11043888e-01 -5.06900251e-01 7.82338738e-01
-2.25570768e-01 5.47339380e-01 -7.96759844e-01 -1.06573236e+00
1.28951976e-02 -1.25097346e+00 3.45832914e-01 5.38113654e-01
-5.21279164e-02 -1.00941133e+00 -6.49929494e-02 3.70136887e-01
8.03827226e-01 -8.67870227e-02 7.88496256e-01 7.74356648e-02
-6.57786012e-01 3.53622511e-02 -3.45405787e-01 4.38957930e-01
3.41298580e-01 2.57796705e-01 -9.60048258e-01 -1.35678038e-01
9.48341966e-01 6.58769488e-01 7.12338090e-01 9.57076430e-01
5.09108782e-01 -4.50756520e-01 1.71085056e-02 8.16915989e-01
1.82529402e+00 3.95737857e-01 1.08615339e+00 6.29629433e-01
5.69827497e-01 5.83146393e-01 3.42378736e-01 5.29969893e-02
1.04549542e-01 6.30840421e-01 2.37728834e-01 -2.82648683e-01
-4.00489628e-01 3.70900840e-01 3.23199749e-01 9.18044567e-01
-2.84417182e-01 -2.47035008e-02 -6.48136556e-01 3.37174296e-01
-1.38384748e+00 -1.13757706e+00 -1.24124265e+00 2.44592357e+00
5.33259273e-01 7.86283612e-02 -5.61676979e-01 2.17555210e-01
5.42746782e-01 1.32563129e-01 1.91142350e-01 -6.17489517e-01
-2.85302192e-01 -4.56121713e-02 8.11885178e-01 1.13479733e+00
-8.79123032e-01 3.45874101e-01 7.25278521e+00 2.63007820e-01
-1.30944836e+00 -8.72449502e-02 1.74192339e-01 4.06861931e-01
-2.04479381e-01 5.68558753e-01 -3.83013159e-01 4.79715228e-01
6.40275180e-01 -1.36065744e-02 3.89261514e-01 -1.76532883e-02
6.89845085e-01 -4.86930341e-01 -5.17322183e-01 9.27938938e-01
-1.36381537e-01 -9.35419917e-01 -1.14712454e-01 3.47513795e-01
6.72825098e-01 -6.93847388e-02 -4.84903380e-02 -4.69970196e-01
2.49189306e-02 -7.41164565e-01 6.77249968e-01 9.03863251e-01
4.94172454e-01 -1.46926701e-01 8.17301989e-01 2.65107572e-01
-8.84374499e-01 2.11725309e-01 -3.08255106e-01 -5.78145027e-01
5.76848149e-01 1.16147828e+00 -4.43280786e-01 7.17769384e-01
6.47745073e-01 4.49380040e-01 -3.22040468e-01 1.47206485e+00
-2.53111571e-01 4.48355585e-01 -1.59329951e-01 6.05582297e-01
-1.09682409e-02 -1.08473313e+00 9.26473916e-01 9.97406185e-01
8.47126067e-01 2.32521161e-01 -2.84578562e-01 5.50197601e-01
7.65610099e-01 -3.94684136e-01 -3.14384013e-01 -3.54658179e-02
1.54944345e-01 1.07672274e+00 -6.68851674e-01 -3.19191307e-01
-5.42648852e-01 1.28431821e+00 -7.59626627e-01 7.59020150e-01
-2.88149416e-01 -2.73662210e-01 8.89723241e-01 4.82754320e-01
2.52761871e-01 -3.13673019e-01 -4.36111033e-01 -1.08064616e+00
-2.10322738e-01 -6.89643800e-01 -4.17684883e-01 -1.37781549e+00
-1.07462072e+00 4.87526715e-01 -5.96606433e-02 -1.16731656e+00
1.06201582e-01 -8.62139404e-01 -7.54047215e-01 1.38710308e+00
-1.62398875e+00 -5.11209726e-01 -3.80074710e-01 2.86029100e-01
3.57944429e-01 5.54514349e-01 6.34237945e-01 2.01351956e-01
-1.68464571e-01 -5.22813201e-01 4.65565085e-01 -2.79371649e-01
9.20412302e-01 -1.61070955e+00 -1.59384497e-03 1.30895436e+00
-2.02421486e-01 8.87510002e-01 1.77445018e+00 -5.33096075e-01
-1.14598596e+00 -5.03001451e-01 1.21041942e+00 -3.57241452e-01
5.67526937e-01 3.89111340e-01 -1.06601930e+00 5.62793374e-01
7.71601439e-01 -1.13621518e-01 4.24837589e-01 -4.66437757e-01
-1.02606334e-03 -2.28752308e-02 -9.03959692e-01 3.47397923e-01
4.26435113e-01 -3.60597968e-01 -6.86703205e-01 2.64394283e-01
3.42659235e-01 -6.49880052e-01 -4.41080451e-01 2.31838703e-01
6.08019948e-01 -1.47691417e+00 7.92283177e-01 6.78164586e-02
4.35002893e-01 -7.50358462e-01 -9.82815102e-02 -1.42904794e+00
-6.23322427e-01 -6.40850186e-01 2.96071589e-01 1.00048196e+00
-2.11489368e-02 -8.91094625e-01 1.37729123e-01 3.89548898e-01
-3.10378134e-01 -2.06966460e-01 -6.77109182e-01 -7.63330579e-01
-1.85943916e-01 7.96690732e-02 2.81604707e-01 9.96411145e-01
-1.80916905e-01 4.55388337e-01 -4.08201069e-01 5.76438844e-01
8.77690136e-01 1.94778427e-01 5.08394003e-01 -1.37941551e+00
-4.31286246e-01 -2.48010218e-01 -1.27423763e-01 -1.27901149e+00
-6.49012446e-01 -9.45259109e-02 3.05361569e-01 -1.80159020e+00
-1.46424696e-01 -1.37828588e-01 1.19256653e-01 -4.35865670e-01
-1.61979496e-01 1.96358219e-01 3.00291181e-01 7.08795667e-01
2.76654750e-01 -1.18837655e-01 1.48620963e+00 3.69855613e-01
-8.27005506e-02 2.15053290e-01 -7.58486509e-01 1.01062262e+00
4.71216649e-01 -1.67936608e-01 -2.87197471e-01 -6.52055383e-01
2.72776008e-01 6.23675063e-02 5.25949180e-01 -9.95267987e-01
1.99098974e-01 -1.37361512e-01 1.61438361e-01 -3.22129726e-01
4.90571111e-01 -9.01692212e-01 5.69995999e-01 5.76425970e-01
5.26238531e-02 1.76733539e-01 2.95879524e-02 3.26356620e-01
-2.23475754e-01 -5.98976851e-01 9.72688913e-01 -4.46101010e-01
-4.50168550e-01 -3.18292469e-01 -8.35038781e-01 -3.86423677e-01
5.73199034e-01 -3.85404587e-01 -5.59724212e-01 -6.44197464e-01
-5.92471004e-01 -4.80174035e-01 9.10677850e-01 -2.17814550e-01
3.49498808e-01 -7.88662374e-01 -5.78448415e-01 1.80797830e-01
-5.86163402e-01 -2.06132382e-01 3.40423375e-01 1.54592943e+00
-1.15225828e+00 2.49452949e-01 2.29024395e-01 -3.93973261e-01
-1.53299069e+00 2.74547487e-01 7.33691812e-01 -1.27199590e-01
-7.15318203e-01 9.54291344e-01 5.02712607e-01 1.63024724e-01
-3.64127070e-01 -1.44796759e-01 -2.37643737e-02 -2.98217207e-01
8.69877160e-01 5.15737355e-01 -1.00267194e-01 -8.79382968e-01
-2.94224262e-01 8.30427408e-01 2.58950263e-01 -2.36816496e-01
9.86012459e-01 -7.64036298e-01 -7.24702716e-01 5.75825155e-01
6.42280281e-01 5.06506741e-01 -1.17575490e+00 8.65692347e-02
-3.39543641e-01 -7.44922519e-01 2.59324968e-01 -6.91161931e-01
-7.67190158e-01 9.48316753e-01 6.94220483e-01 6.78945363e-01
1.26426041e+00 -5.49145818e-01 3.11485589e-01 -1.11507498e-01
-1.31712914e-01 -6.82757556e-01 -3.78910094e-01 7.02630639e-01
9.10904825e-01 -5.80573320e-01 3.56476903e-01 -7.47087479e-01
-3.96228522e-01 1.18347514e+00 1.07938424e-01 -2.25429758e-01
8.53296220e-01 4.84296739e-01 5.26181996e-01 -1.17023207e-01
-4.59461272e-01 -2.11109102e-01 1.87007189e-01 5.49790502e-01
6.95923448e-01 -1.38530418e-01 -5.77543378e-01 -1.35519281e-01
-3.29516798e-01 1.65129110e-01 1.08642542e+00 7.40308940e-01
-6.74225748e-01 -1.02317715e+00 -9.57530916e-01 3.24975967e-01
-3.79697978e-01 -2.38716707e-01 -5.39412126e-02 4.70941484e-01
3.83579820e-01 1.29623783e+00 9.45600942e-02 -8.35332088e-03
3.04855913e-01 -3.63592654e-02 5.36692441e-01 -3.19690615e-01
-2.59100080e-01 6.37665451e-01 1.19616101e-02 -2.34036878e-01
-7.13298023e-01 -7.18211472e-01 -9.45418417e-01 -4.28365052e-01
-5.09378195e-01 3.58854294e-01 5.59247732e-01 1.19064867e+00
-2.07213592e-03 5.15723765e-01 2.44891465e-01 -8.63408744e-01
-5.10483325e-01 -1.03178072e+00 -1.08986211e+00 3.79109740e-01
8.46538961e-01 -3.73171717e-01 -8.60697746e-01 2.65338898e-01] | [11.69257640838623, -2.746903896331787] |
6e661c1a-1579-4291-a5e6-a7abbd4e1cfd | when-regression-meets-manifold-learning-for | 1805.064 | null | http://arxiv.org/abs/1805.06400v1 | http://arxiv.org/pdf/1805.06400v1.pdf | When Regression Meets Manifold Learning for Object Recognition and Pose Estimation | In this work, we propose a method for object recognition and pose estimation
from depth images using convolutional neural networks. Previous methods
addressing this problem rely on manifold learning to learn low dimensional
viewpoint descriptors and employ them in a nearest neighbor search on an
estimated descriptor space. In comparison we create an efficient multi-task
learning framework combining manifold descriptor learning and pose regression.
By combining the strengths of manifold learning using triplet loss and pose
regression, we could either estimate the pose directly reducing the complexity
compared to NN search, or use learned descriptor for the NN descriptor
matching. By in depth experimental evaluation of the novel loss function we
observed that the view descriptors learned by the network are much more
discriminative resulting in almost 30% increase regarding relative pose
accuracy compared to related works. On the other hand, regarding directly
regressed poses we obtained important improvement compared to simple pose
regression. By leveraging the advantages of both manifold learning and
regression tasks, we are able to improve the current state-of-the-art for
object recognition and pose retrieval that we demonstrate through in depth
experimental evaluation. | ['Sergey Zakharov', 'Mai Bui', 'Slobodan Ilic', 'Nassir Navab', 'Shadi Albarqouni'] | 2018-05-16 | null | null | null | null | ['pose-retrieval'] | ['computer-vision'] | [ 3.82787101e-02 4.85112071e-02 -2.51821756e-01 -4.27054733e-01
-1.39254045e+00 -5.61468422e-01 8.24893951e-01 2.08174974e-01
-6.31305218e-01 3.64360064e-01 6.28311038e-02 3.36390734e-01
-5.42982519e-01 -5.61439693e-01 -9.46409106e-01 -6.16032124e-01
5.86402901e-02 8.37056458e-01 -7.87203163e-02 1.32176206e-01
5.37119806e-01 9.21410382e-01 -1.83579695e+00 -9.06243771e-02
1.37380466e-01 1.37722719e+00 -2.16720119e-01 3.48680884e-01
1.66660592e-01 4.60346222e-01 -3.73228759e-01 -2.07412258e-01
5.12458503e-01 1.24785960e-01 -7.09461331e-01 -6.56262785e-02
1.40808403e+00 -5.53479493e-01 -3.63086849e-01 5.67921281e-01
6.88106537e-01 2.16238543e-01 9.42853391e-01 -9.98432100e-01
-4.39488739e-01 -1.19146489e-01 -3.25075537e-01 -9.89317968e-02
5.48577189e-01 -3.53490204e-01 1.04464555e+00 -1.21750438e+00
9.19618189e-01 1.35644448e+00 7.09514678e-01 3.94671559e-01
-1.26556933e+00 -3.88194740e-01 4.55049947e-02 4.00679529e-01
-1.57645488e+00 -5.39862752e-01 9.46483612e-01 -3.52273703e-01
1.20457494e+00 7.09918812e-02 4.09780979e-01 9.53589082e-01
7.22835585e-02 6.65058494e-01 7.83039033e-01 -7.00901508e-01
8.35664198e-03 1.32670254e-01 -1.43297404e-01 8.61726463e-01
3.68880145e-02 4.04089332e-01 -5.80767334e-01 -7.98171088e-02
7.13755906e-01 1.75913051e-01 -2.89062671e-02 -1.32828426e+00
-1.21347344e+00 8.85331035e-01 7.48085380e-01 1.85867414e-01
-2.14117810e-01 4.78501946e-01 4.54925060e-01 2.29665786e-01
5.41685402e-01 3.48041713e-01 -3.50911885e-01 -1.64980769e-01
-9.64510143e-01 1.98355332e-01 7.60002136e-01 9.38364685e-01
9.86867785e-01 -3.35799009e-01 -9.45396125e-02 6.76263690e-01
4.81600612e-01 6.57976985e-01 1.39935791e-01 -1.07896757e+00
3.85480851e-01 4.95758474e-01 -1.26449183e-01 -1.09482968e+00
-4.17071104e-01 -2.60907412e-01 -2.65036315e-01 4.37730968e-01
5.38943887e-01 3.32958579e-01 -8.34774494e-01 1.60973632e+00
3.12035829e-01 -3.25567126e-02 -1.74071431e-01 8.40877593e-01
5.47469020e-01 4.52645076e-03 -1.82127893e-01 1.84657902e-01
1.10959208e+00 -8.64066482e-01 -4.54678506e-01 2.51732320e-01
6.67716980e-01 -9.26584363e-01 5.66035211e-01 4.72337693e-01
-9.15922046e-01 -5.93578994e-01 -1.31103730e+00 -3.06861013e-01
-5.05075514e-01 2.75562048e-01 5.54658949e-01 4.88893121e-01
-1.14296234e+00 1.05359328e+00 -9.99664664e-01 -4.60072070e-01
3.95132482e-01 7.26407170e-01 -6.87157393e-01 -3.67010795e-02
-6.56934619e-01 1.25610816e+00 1.51889116e-01 -1.44707290e-02
-9.09590364e-01 -6.38133168e-01 -1.04834068e+00 -2.03860551e-01
2.49822289e-01 -7.44591594e-01 8.98475647e-01 -6.33942723e-01
-1.43513262e+00 1.05607796e+00 -7.82913342e-02 -3.01404655e-01
6.41821682e-01 -5.91238320e-01 2.42519826e-02 4.51120615e-01
-4.99811955e-03 9.37070966e-01 1.04516494e+00 -1.10868573e+00
-2.72027969e-01 -8.64034891e-01 3.57765779e-02 9.39809904e-02
-5.18778503e-01 -3.03462982e-01 -4.28934574e-01 -2.20419466e-01
3.19999903e-01 -1.04827940e+00 2.88911741e-02 5.38609505e-01
-1.51190702e-02 -4.04710233e-01 1.03295195e+00 -4.03266966e-01
6.46992147e-01 -1.93867934e+00 6.40795410e-01 1.99208289e-01
2.10183933e-01 1.51053414e-01 -3.01783621e-01 5.26962399e-01
-8.09957907e-02 -3.31592947e-01 3.98593277e-01 -8.07887256e-01
1.11200459e-01 8.80361423e-02 -1.90444842e-01 8.49573374e-01
3.60791773e-01 9.24040735e-01 -6.28210545e-01 -4.26808566e-01
5.71602345e-01 1.00092280e+00 -5.15586972e-01 1.46079332e-01
-2.56955400e-02 2.31032133e-01 -3.82020205e-01 7.18348682e-01
5.55265903e-01 2.23376211e-02 -1.56153306e-01 -5.62750459e-01
1.37814969e-01 2.71179676e-02 -1.25733054e+00 2.38513446e+00
-7.79852808e-01 8.01595032e-01 -2.63963103e-01 -1.17424405e+00
1.12781382e+00 1.34433940e-01 6.13995910e-01 -7.16190279e-01
9.02819633e-02 4.10656840e-01 -6.28749907e-01 -2.05345690e-01
4.68395740e-01 3.11998539e-02 7.98822790e-02 1.25703886e-01
5.45341015e-01 -9.00318101e-02 -3.04562032e-01 -1.56014681e-01
7.93436468e-01 9.05057847e-01 1.51686549e-01 -6.56290278e-02
7.48737097e-01 -1.60539955e-01 -8.16853866e-02 4.58002836e-01
-7.46579692e-02 6.92424059e-01 1.64238870e-01 -6.47798359e-01
-1.06131101e+00 -1.24173236e+00 -4.42378640e-01 7.99304605e-01
3.10729314e-02 -4.44614053e-01 -4.92741615e-01 -6.35341585e-01
2.98744142e-01 2.49103114e-01 -6.76647544e-01 -2.02519357e-01
-7.78230071e-01 2.98001450e-02 4.78481948e-01 6.02652073e-01
2.02593818e-01 -6.59563601e-01 -4.29163545e-01 -2.58849468e-03
3.06669354e-01 -1.21606016e+00 -3.64384621e-01 2.25646332e-01
-1.08026004e+00 -9.95773554e-01 -8.42109621e-01 -7.46197104e-01
5.68585753e-01 1.83488697e-01 1.07643986e+00 -1.68878794e-01
-6.13307118e-01 1.10192704e+00 -2.18310982e-01 -2.97521353e-01
-5.58192767e-02 2.99962193e-01 1.04920894e-01 -1.66736960e-01
5.32509208e-01 -7.54031062e-01 -7.94193685e-01 2.20573083e-01
-7.38723397e-01 -7.14083910e-01 7.20766306e-01 7.92088568e-01
6.49618924e-01 -8.40239465e-01 2.51822621e-01 -3.55757654e-01
3.04258000e-02 -3.09854206e-02 -8.21396232e-01 2.17897236e-01
-8.22479188e-01 6.44673049e-01 1.71916813e-01 -3.63108039e-01
-4.38413650e-01 5.08374751e-01 9.73736420e-02 -9.01641428e-01
-2.66512930e-01 6.84998110e-02 8.92984346e-02 -7.07835138e-01
5.97274780e-01 4.79933582e-02 4.82107103e-01 -5.70676506e-01
6.10415757e-01 3.90813857e-01 1.72114685e-01 -5.36579907e-01
1.03310156e+00 6.90829754e-01 7.38729656e-01 -6.45176053e-01
-8.23966384e-01 -7.32762873e-01 -1.21025991e+00 -2.39465609e-01
1.01954949e+00 -1.04589272e+00 -9.79840040e-01 5.75981662e-02
-1.15068269e+00 2.85559684e-01 -1.80891573e-01 7.01451123e-01
-9.63254809e-01 5.92474103e-01 -1.92346752e-01 -7.45546877e-01
-2.90782124e-01 -1.05015278e+00 1.91375220e+00 -3.09322238e-01
-1.08797409e-01 -1.22963643e+00 4.07189220e-01 3.85636568e-01
4.67878580e-01 4.46934998e-01 5.54819524e-01 -7.28507280e-01
-1.01172030e+00 -5.48740983e-01 -1.76008433e-01 2.74952948e-01
-6.18864372e-02 -2.47566402e-01 -1.15573883e+00 -6.18970037e-01
-1.23486459e-01 -5.95272124e-01 9.63421106e-01 1.41219109e-01
7.44383156e-01 2.84072757e-01 -3.56908441e-01 7.84041226e-01
1.79884744e+00 -3.12463403e-01 6.62810326e-01 6.52455747e-01
7.82730460e-01 7.10962534e-01 8.38377953e-01 2.56340027e-01
3.43946278e-01 1.28965116e+00 5.88164330e-01 4.54182327e-02
-2.47762784e-01 -2.44918391e-01 3.63507837e-01 4.62136000e-01
-1.56339109e-01 3.52497905e-01 -7.53153503e-01 3.06751370e-01
-1.76696503e+00 -8.03722620e-01 4.31622058e-01 2.43366194e+00
2.68952817e-01 -1.11167654e-01 1.75336689e-01 -2.60189157e-02
3.61468107e-01 2.01283488e-02 -3.55250508e-01 -2.21493408e-01
1.46220997e-01 3.74229401e-01 5.07696211e-01 4.67693627e-01
-1.29462075e+00 7.40726173e-01 5.96174574e+00 7.99585998e-01
-1.12619805e+00 -9.59657058e-02 4.28087078e-02 -3.45056564e-01
-5.72836585e-02 5.93805015e-02 -1.09264362e+00 -2.17134401e-01
9.47621465e-01 3.87421578e-01 2.91147441e-01 9.59076643e-01
-1.55621901e-01 2.66407263e-02 -1.65192056e+00 1.34440064e+00
6.46059692e-01 -1.13294446e+00 2.06823438e-01 4.06563669e-01
3.77739459e-01 9.81466025e-02 3.28977518e-02 3.58446240e-01
-5.06835282e-01 -1.06929767e+00 6.59566462e-01 6.94525659e-01
6.35660112e-01 -6.58390403e-01 6.23467326e-01 1.63746640e-01
-1.13348210e+00 1.09948210e-01 -4.84627277e-01 1.99922305e-02
-7.43956193e-02 1.64162740e-01 -7.13000596e-01 9.11523163e-01
4.08135474e-01 8.78487349e-01 -7.47051299e-01 1.02554202e+00
8.40373859e-02 -1.51913300e-01 -4.72542733e-01 -7.44559616e-02
2.24621683e-01 -2.70555854e-01 6.92299008e-01 9.00768161e-01
3.14414054e-01 -6.81987405e-01 1.84556201e-01 8.75226378e-01
1.07604668e-01 2.54086375e-01 -8.64591658e-01 2.26888657e-01
1.07155800e-01 1.34185398e+00 -4.09942120e-01 -4.85805143e-03
-3.61089766e-01 9.98380244e-01 6.34997666e-01 3.95690985e-02
-6.59318447e-01 -3.66097599e-01 5.63782156e-01 4.95538674e-02
8.26151431e-01 -5.23433924e-01 1.41438276e-01 -1.09467030e+00
3.80177170e-01 -4.22638744e-01 1.18773878e-01 -4.26363051e-01
-1.07422900e+00 4.65294987e-01 2.03447968e-01 -1.61918759e+00
-5.96440494e-01 -9.80929375e-01 -1.09046079e-01 7.10149169e-01
-1.63225067e+00 -1.53820968e+00 -2.11872712e-01 4.31282699e-01
3.52091730e-01 -1.90959424e-01 1.20699859e+00 4.41205591e-01
-4.67381850e-02 6.55603528e-01 1.87794521e-01 -1.29958261e-02
9.95103300e-01 -1.31347644e+00 6.60755336e-02 1.81076318e-01
5.20276248e-01 6.80250764e-01 3.73156279e-01 -1.19233809e-01
-1.80196261e+00 -6.78826094e-01 5.81213236e-01 -1.00900722e+00
5.00299394e-01 -4.76101130e-01 -5.02094209e-01 4.53211218e-01
1.08264843e-02 2.23391384e-01 5.35081446e-01 1.28108695e-01
-7.28166521e-01 -1.19173512e-01 -9.73925054e-01 1.53201774e-01
9.53628957e-01 -7.97746062e-01 -5.89724481e-01 3.78161520e-01
4.63518530e-01 -3.96785527e-01 -1.31670523e+00 5.23342907e-01
8.81047070e-01 -9.18000400e-01 1.25972080e+00 -5.14492333e-01
1.85661599e-01 -1.19335778e-01 -4.69253182e-01 -8.94313335e-01
9.22947377e-03 -1.75056577e-01 -1.79178998e-01 1.20804298e+00
3.22331995e-01 -4.08904195e-01 1.00913393e+00 4.26995516e-01
7.26832822e-02 -8.56927991e-01 -1.22420704e+00 -1.05224121e+00
3.07989955e-01 -1.99299619e-01 4.15475629e-02 3.28373373e-01
-3.12429458e-01 4.46413875e-01 -3.46979350e-01 -2.99942400e-02
1.09073782e+00 1.86509192e-01 9.38487530e-01 -1.35274577e+00
-2.78677881e-01 -2.41715550e-01 -1.17689323e+00 -1.16572654e+00
4.64909017e-01 -1.01639438e+00 -1.58956096e-01 -1.29006386e+00
2.17979148e-01 6.80129901e-02 -3.48438501e-01 1.90182254e-01
2.13080764e-01 4.63367045e-01 3.21784765e-01 2.35518947e-01
-8.34575713e-01 6.65795028e-01 1.09673345e+00 -1.71655402e-01
1.48614138e-01 -1.15314692e-01 -1.18474223e-01 3.33815515e-01
3.47678214e-01 -4.63264167e-01 -1.73923433e-01 -4.37425941e-01
2.40897294e-02 -9.96816233e-02 6.14272118e-01 -1.05805945e+00
2.41742879e-01 5.45235455e-01 6.70493245e-01 -7.39604354e-01
9.47087109e-01 -1.05399084e+00 -2.02193514e-01 4.34962362e-01
-3.38995665e-01 -4.79818285e-02 2.64990479e-01 7.11203814e-01
-3.28996956e-01 -1.21576905e-01 4.19366896e-01 3.92223448e-02
-7.74049997e-01 3.32857937e-01 1.87778041e-01 -2.90021211e-01
1.03208017e+00 -5.22943437e-01 -1.53842140e-02 -5.24421036e-01
-8.86354208e-01 -9.81195569e-02 5.00508368e-01 5.97581089e-01
7.91065335e-01 -1.59423459e+00 -4.63848174e-01 1.97516188e-01
3.60802233e-01 -3.51176709e-01 -1.45895794e-01 9.65837002e-01
-4.80819851e-01 7.64113247e-01 -3.24038088e-01 -1.04994678e+00
-1.37842059e+00 7.15293050e-01 3.84171903e-01 -1.24064006e-01
-3.91043335e-01 5.70351601e-01 -3.01119089e-01 -7.64448464e-01
6.56301916e-01 -1.46287009e-02 -8.94102082e-02 2.92401373e-01
1.86677247e-01 6.35865331e-01 3.72266382e-01 -7.29397595e-01
-5.76125622e-01 1.46079731e+00 -1.19197182e-01 -1.05174966e-01
1.55296874e+00 -8.89750570e-02 -5.91359511e-02 4.07755882e-01
2.02195024e+00 -1.34265214e-01 -1.15937889e+00 -3.05912077e-01
2.90907681e-01 -5.12212336e-01 1.74222067e-01 -3.43422741e-01
-7.24430740e-01 9.41060543e-01 1.12010312e+00 -2.04719335e-01
7.24716663e-01 1.81653947e-01 4.89981622e-01 8.34578753e-01
5.94287276e-01 -9.70302105e-01 4.24139082e-01 3.43554527e-01
1.01189029e+00 -1.49149024e+00 2.76760370e-01 -3.66048187e-01
-2.54571903e-02 1.65829039e+00 3.58169615e-01 -6.31782413e-01
6.94518626e-01 -3.76854353e-02 3.06578159e-01 -4.72174436e-01
-4.55659986e-01 -2.28548631e-01 8.09375763e-01 6.57649219e-01
4.22200948e-01 -4.12041396e-01 7.15182424e-02 -3.50231200e-01
1.09151624e-01 -3.30351502e-01 -1.08040072e-01 8.11684370e-01
-3.17726374e-01 -1.57331300e+00 -3.77713680e-01 1.05715983e-01
-2.67570585e-01 1.01486847e-01 -5.41815937e-01 1.02386403e+00
-2.69303948e-01 5.94718695e-01 -1.16624609e-02 -3.10625285e-01
5.94431281e-01 1.39894113e-01 1.09876966e+00 -4.75602269e-01
-3.04107726e-01 1.20902963e-01 -9.91805419e-02 -1.07970405e+00
-7.30429769e-01 -8.40650499e-01 -6.95598841e-01 3.28055829e-01
-5.84891200e-01 -2.03751370e-01 1.13886797e+00 8.90675724e-01
3.37627798e-01 2.38483757e-01 4.85997409e-01 -1.49446106e+00
-1.24290049e+00 -8.91126871e-01 -3.82192940e-01 4.34756011e-01
4.36378628e-01 -1.12662601e+00 -4.70174909e-01 -3.33801389e-01] | [7.855602264404297, -2.241067886352539] |
4c0c9c97-17f0-483a-a8cc-7ca3fa9a6c75 | bella-black-box-model-explanations-by-local | 2305.11311 | null | https://arxiv.org/abs/2305.11311v1 | https://arxiv.org/pdf/2305.11311v1.pdf | BELLA: Black box model Explanations by Local Linear Approximations | In recent years, understanding the decision-making process of black-box models has become not only a legal requirement but also an additional way to assess their performance. However, the state of the art post-hoc interpretation approaches rely on synthetic data generation. This introduces uncertainty and can hurt the reliability of the interpretations. Furthermore, they tend to produce explanations that apply to only very few data points. This makes the explanations brittle and limited in scope. Finally, they provide scores that have no direct verifiable meaning. In this paper, we present BELLA, a deterministic model-agnostic post-hoc approach for explaining the individual predictions of regression black-box models. BELLA provides explanations in the form of a linear model trained in the feature space. Thus, its coefficients can be used directly to compute the predicted value from the feature values. Furthermore, BELLA maximizes the size of the neighborhood to which the linear model applies, so that the explanations are accurate, simple, general, and robust. BELLA can produce both factual and counterfactual explanations. Our user study confirms the importance of the desiderata we optimize, and our experiments show that BELLA outperforms the state-of-the-art approaches on these desiderata. | ['Fabian Suchanek', 'Albert Bifet', 'Nedeljko Radulovic'] | 2023-05-18 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [-8.22514370e-02 7.36784995e-01 -5.36706269e-01 -6.61953986e-01
-7.14210272e-01 -6.55037165e-01 8.59292686e-01 3.15028727e-01
-5.19146360e-02 1.09763539e+00 7.48672336e-02 -8.14693868e-01
-2.77209610e-01 -6.84991181e-01 -8.04503202e-01 -4.70486253e-01
9.06798914e-02 5.47460973e-01 1.09771788e-01 -1.07180446e-01
3.32549661e-01 1.53893039e-01 -1.59768260e+00 3.98724914e-01
1.03059232e+00 9.50756252e-01 -1.38488904e-01 4.22959030e-01
1.16331764e-02 4.54329044e-01 -6.72732174e-01 -7.40936160e-01
1.56602666e-01 -5.50101280e-01 -6.39371514e-01 -1.18476309e-01
6.17433563e-02 -4.71203417e-01 1.60719901e-01 7.61895001e-01
-5.28763160e-02 9.81029049e-02 7.38847554e-01 -1.64139891e+00
-7.89207280e-01 9.55348969e-01 -2.21406177e-01 -1.53424680e-01
4.31204915e-01 2.51009226e-01 1.26519728e+00 -8.83234203e-01
4.30799663e-01 1.53601873e+00 5.47588110e-01 6.64563596e-01
-1.49077165e+00 -6.44543767e-01 5.24589419e-01 1.41828150e-01
-1.09361100e+00 -2.84256279e-01 3.86243790e-01 -3.45247120e-01
8.99225414e-01 6.32730126e-01 6.18633032e-01 1.12948370e+00
2.97071457e-01 4.85134631e-01 1.21275496e+00 -5.74914753e-01
6.23878896e-01 3.18200350e-01 -5.20399259e-03 4.53928232e-01
8.82897317e-01 5.58832824e-01 -5.22331417e-01 -3.60055864e-01
6.21609747e-01 -2.43334342e-02 -1.12951860e-01 -3.38562787e-01
-1.06765974e+00 1.05292964e+00 4.65977371e-01 2.87790541e-02
-3.49334836e-01 2.67921478e-01 -5.28557263e-02 2.54070133e-01
3.85763973e-01 8.79420161e-01 -6.66467011e-01 -7.10552335e-02
-7.93867409e-01 7.73648679e-01 6.32070899e-01 5.89502633e-01
5.26871741e-01 -1.53806210e-01 -1.84836373e-01 3.94461870e-01
6.26672924e-01 4.21662867e-01 5.24996817e-01 -1.02368569e+00
2.99823910e-01 6.68447256e-01 5.61822951e-01 -9.35142398e-01
-3.75308275e-01 -3.68993342e-01 -3.89652282e-01 5.59807658e-01
7.09883749e-01 -2.74564654e-01 -8.02141249e-01 1.69678771e+00
2.22738773e-01 -2.98552901e-01 1.30180478e-01 9.94893372e-01
3.31735313e-01 4.87149924e-01 5.16683646e-02 -2.17072085e-01
9.37828660e-01 -7.48881221e-01 -6.62344813e-01 -5.10707378e-01
7.64973164e-01 -6.02792203e-01 1.30991912e+00 3.57264876e-01
-1.05536127e+00 -3.84497494e-01 -1.08069408e+00 2.79149562e-01
-1.82960764e-01 -1.30482644e-01 8.51367891e-01 6.15034878e-01
-6.05000257e-01 8.27874005e-01 -8.64441216e-01 -2.19186265e-02
2.06172869e-01 3.32688332e-01 -1.33231238e-01 1.33799225e-01
-1.16151786e+00 1.10057890e+00 4.59154934e-01 -8.17905143e-02
-6.29507780e-01 -6.58116221e-01 -8.41215611e-01 4.15958434e-01
5.71883261e-01 -7.27080226e-01 1.59092522e+00 -1.02770495e+00
-1.12133813e+00 3.26359868e-01 -3.22368383e-01 -6.89356744e-01
8.60682786e-01 -8.80353302e-02 -3.58694494e-01 -3.58713180e-01
2.09598899e-01 6.64001822e-01 5.26996493e-01 -1.53110194e+00
-7.13813663e-01 -9.89674777e-02 2.87811100e-01 -4.00431901e-02
2.67912865e-01 -3.09580624e-01 1.89911395e-01 -5.00468135e-01
2.34743685e-01 -8.77074003e-01 -5.14191270e-01 3.77480648e-02
-6.54089272e-01 -6.65495321e-02 3.83721948e-01 -2.44911894e-01
1.51331413e+00 -1.97055721e+00 -4.15148079e-01 4.16563600e-01
-6.10681018e-03 -1.42818525e-01 2.36152142e-01 3.49818110e-01
-2.22113162e-01 8.47859800e-01 -2.90901393e-01 -7.73896351e-02
2.29114577e-01 2.80898750e-01 -6.08919501e-01 4.65343073e-02
1.95291147e-01 8.83228421e-01 -9.19640243e-01 -3.63102376e-01
2.37113357e-01 -1.28591016e-01 -7.15089619e-01 2.33881652e-01
-4.83337224e-01 2.27474108e-01 -3.89292300e-01 2.61406004e-01
3.98195595e-01 -3.89003485e-01 2.80286103e-01 4.05606002e-01
-7.46984407e-02 5.75505912e-01 -1.18401062e+00 1.07768047e+00
-3.94086987e-01 6.11106455e-01 -7.22078919e-01 -4.91862714e-01
8.82961810e-01 1.53333768e-01 -9.55646783e-02 -2.99932361e-01
-9.94003098e-03 4.36141551e-01 4.30149943e-01 -3.58834684e-01
2.69550949e-01 -4.41511095e-01 -1.31230474e-01 6.61174476e-01
-4.82297093e-01 -3.57110441e-01 1.70891821e-01 8.78107026e-02
8.48699152e-01 2.26768047e-01 8.41090381e-01 -3.02708074e-02
1.37642965e-01 1.88297331e-01 7.33102441e-01 1.01072598e+00
2.61592507e-01 8.03286493e-01 8.16184163e-01 -6.14148378e-01
-9.38395262e-01 -1.00308311e+00 -1.64531097e-01 4.92089391e-01
5.24326786e-02 -3.84971112e-01 -6.27092898e-01 -1.00530934e+00
2.76725322e-01 1.67267299e+00 -9.31316018e-01 -3.02404940e-01
-4.21744287e-02 -5.49362123e-01 8.90262350e-02 5.94458878e-01
2.00524494e-01 -7.44411707e-01 -8.89087439e-01 8.91364217e-02
-2.75791466e-01 -6.42831445e-01 -8.48341212e-02 4.75560129e-02
-1.03046596e+00 -1.17461491e+00 -1.43127829e-01 7.07461089e-02
9.83624995e-01 2.48051822e-01 1.12054348e+00 3.56614679e-01
4.51436341e-01 -2.57354200e-01 -2.87612379e-01 -8.09961915e-01
-6.81585550e-01 -3.01706851e-01 1.83900003e-03 -1.76075324e-01
4.40030158e-01 -5.07209301e-01 -5.26272833e-01 5.70112526e-01
-6.07831120e-01 2.35042378e-01 6.07818365e-01 9.47933853e-01
4.61809069e-01 -1.39952913e-01 7.28003144e-01 -1.23507440e+00
8.32396388e-01 -4.40288126e-01 -7.26744294e-01 4.46280122e-01
-1.18098593e+00 5.25862217e-01 7.75443017e-01 -3.73827308e-01
-1.22786129e+00 -1.66683033e-01 2.75140375e-01 -1.94764808e-01
-1.90444589e-01 7.83198059e-01 -5.78431711e-02 5.25219858e-01
1.14262426e+00 -3.45667481e-01 -2.07161143e-01 -3.49132478e-01
4.48781073e-01 4.95451659e-01 4.05304611e-01 -5.13653934e-01
9.12395775e-01 2.76323169e-01 -8.38157460e-02 -1.24118939e-01
-1.07394934e+00 1.97561726e-01 -4.10952061e-01 -9.56182033e-02
3.67752582e-01 -4.60386842e-01 -6.53661728e-01 -3.22351515e-01
-1.12845635e+00 -3.44078541e-01 -5.42363942e-01 5.55592835e-01
-6.69901133e-01 -3.58623941e-03 1.27311289e-01 -1.21653378e+00
1.61907211e-01 -1.03888261e+00 6.64941609e-01 1.96568593e-01
-1.14841402e+00 -9.89574969e-01 -1.88259453e-01 2.53319055e-01
1.44310996e-01 4.59322423e-01 1.16106629e+00 -9.20010746e-01
-5.68116724e-01 -4.56896633e-01 -2.34588478e-02 -1.85759347e-02
-5.45312651e-03 3.50710183e-01 -1.16455877e+00 5.52923270e-02
-1.14548236e-01 -1.01416521e-01 6.24647498e-01 5.00735343e-01
1.27220333e+00 -7.25355804e-01 -3.23588729e-01 1.22069232e-01
9.27483976e-01 2.69756168e-01 3.92647058e-01 3.68022442e-01
7.61411786e-02 8.09175014e-01 1.09504330e+00 5.48552454e-01
3.88352782e-01 6.63581967e-01 5.88355482e-01 -5.91653511e-02
3.48648727e-01 -6.63774490e-01 3.03195834e-01 -1.20848529e-02
-1.15082096e-02 -2.76645243e-01 -9.18491781e-01 5.66478491e-01
-2.36474895e+00 -9.67811644e-01 -2.56714851e-01 2.62414551e+00
5.96521795e-01 5.12447536e-01 -8.86513516e-02 2.02063918e-01
3.62592548e-01 -1.91842332e-01 -6.85844839e-01 -8.02760303e-01
-4.05163281e-02 -1.69454902e-01 3.15678716e-01 5.83426833e-01
-6.78383887e-01 6.11549973e-01 6.96061707e+00 3.99869800e-01
-7.70698905e-01 -8.73636529e-02 8.46575499e-01 -2.92274922e-01
-8.91741633e-01 4.45984304e-01 -3.01474094e-01 5.22194624e-01
1.05835640e+00 -6.83823347e-01 2.43926704e-01 1.11707580e+00
6.57680571e-01 -4.26684976e-01 -1.66398704e+00 6.06964409e-01
-3.68242949e-01 -1.43596959e+00 1.55070677e-01 1.55222595e-01
8.20006728e-01 -5.75293005e-01 1.39531597e-01 2.17218548e-01
6.36121273e-01 -1.27205050e+00 1.10723782e+00 3.43040258e-01
6.10976517e-01 -7.82977104e-01 1.12496018e+00 5.21404266e-01
-4.25379902e-01 -3.56354266e-01 -3.33575219e-01 -5.15371144e-01
1.01622105e-01 6.03694916e-01 -1.13889647e+00 3.36617261e-01
4.08520550e-01 2.13863954e-01 -3.81090909e-01 8.75015736e-01
-8.78732324e-01 7.62822807e-01 -2.02401921e-01 -1.07448936e-01
1.44835725e-01 1.25769004e-01 2.32243672e-01 7.60949135e-01
4.61519569e-01 2.40367010e-01 -8.16312432e-03 1.34042943e+00
8.32580924e-02 -1.48879707e-01 -6.70014083e-01 1.88573003e-01
7.13628948e-01 8.31240594e-01 -5.54314435e-01 -3.86217505e-01
-2.64035463e-01 3.09509337e-01 1.22238629e-01 3.95722777e-01
-8.47170591e-01 2.23946832e-02 7.52188325e-01 3.04503381e-01
-1.81841552e-01 1.69816092e-01 -9.27644730e-01 -1.06830800e+00
1.32355213e-01 -9.40227985e-01 4.42902863e-01 -9.59381819e-01
-1.05576396e+00 3.10595036e-01 3.92290205e-01 -1.30579638e+00
-9.59157705e-01 -4.07297492e-01 -8.59265566e-01 9.11496103e-01
-1.18579185e+00 -7.24094093e-01 -1.05228230e-01 -9.31494161e-02
3.87271404e-01 2.15603188e-01 9.57743287e-01 -4.15562689e-01
-4.87908781e-01 5.69066226e-01 -1.06179774e-01 -2.60095686e-01
6.00724876e-01 -1.32430756e+00 7.99587548e-01 9.02328491e-01
1.43607855e-01 9.90765810e-01 1.24365711e+00 -6.30816400e-01
-6.81270540e-01 -7.58569479e-01 1.10858905e+00 -6.62232161e-01
4.90580916e-01 -1.06361695e-01 -9.47533727e-01 8.14381540e-01
-1.74491465e-01 -3.06739777e-01 7.62824893e-01 5.70555627e-01
-1.71480983e-01 2.64122561e-02 -1.28868508e+00 9.09619391e-01
6.91130936e-01 -3.63052404e-03 -8.68993819e-01 9.04035270e-02
6.36789560e-01 -3.31415951e-01 -4.42336917e-01 5.44230007e-02
7.86224961e-01 -1.38045895e+00 6.46513581e-01 -1.00194526e+00
7.22553730e-01 -2.76563495e-01 2.87387776e-03 -1.59423029e+00
-1.69071361e-01 -6.57343030e-01 -8.54701549e-03 1.04332054e+00
1.20559585e+00 -8.09288085e-01 5.82137465e-01 1.44321370e+00
1.23468004e-01 -9.31677997e-01 -7.68999279e-01 -7.56652951e-01
-3.72205637e-02 -6.72361374e-01 1.13911688e+00 8.22301328e-01
3.20759833e-01 1.02397487e-01 -2.26617992e-01 1.36052519e-01
4.78503376e-01 3.33241105e-01 9.58569050e-01 -1.34120548e+00
-3.91424924e-01 -2.95844465e-01 -1.51099533e-01 -7.64512956e-01
3.45707461e-02 -5.58300018e-01 -8.04294422e-02 -1.49119937e+00
1.72142714e-01 -4.96596217e-01 -1.68208122e-01 6.91384614e-01
-3.86903495e-01 -3.20115387e-01 3.32425773e-01 2.17669457e-01
-1.43012270e-01 2.90070176e-01 1.03729439e+00 1.45591348e-01
-3.40327680e-01 2.74350673e-01 -1.15485716e+00 1.10633016e+00
9.70695198e-01 -7.32110202e-01 -5.67425966e-01 -1.90601930e-01
5.37364364e-01 7.10158199e-02 5.76728880e-01 -5.33691287e-01
-1.76163390e-01 -7.84286261e-01 5.12045383e-01 -2.65916020e-01
2.52382487e-01 -7.92995393e-01 2.46076405e-01 4.04447049e-01
-7.92435288e-01 4.20141183e-02 1.16007231e-01 5.07200480e-01
-1.09223060e-01 -4.21173573e-01 5.96840143e-01 -9.70697924e-02
-2.60879010e-01 -2.34103292e-01 -2.11176693e-01 -7.69375116e-02
1.03792238e+00 -3.16323847e-01 -3.27365071e-01 -7.59746909e-01
-5.10019183e-01 3.94455671e-01 5.80725312e-01 3.52726817e-01
6.30865037e-01 -1.19529808e+00 -6.12316072e-01 1.65972114e-01
2.67251968e-01 8.36297274e-02 -4.30872850e-02 5.34206688e-01
-1.85272411e-01 5.61292171e-01 5.48997987e-03 -3.67567092e-01
-9.92643416e-01 5.93745172e-01 2.54269630e-01 -3.26489508e-01
-3.30594897e-01 4.11162376e-01 3.62432122e-01 -3.83440614e-01
-1.12702623e-01 -7.07853734e-01 1.79647058e-02 -1.38487071e-01
5.82768738e-01 3.75236690e-01 -2.11966738e-01 -1.21852659e-01
-2.57056057e-01 -9.20605566e-03 2.42355783e-02 -4.15878057e-01
1.14274871e+00 -1.33886591e-01 3.35408390e-01 7.82157183e-01
3.84331465e-01 -8.99605602e-02 -1.31066310e+00 1.65244907e-01
2.37512350e-01 -9.38497961e-01 -1.85766086e-01 -1.29505408e+00
-4.70860779e-01 8.54175210e-01 -3.35749090e-02 4.95567709e-01
7.37087071e-01 -8.40231404e-02 1.25142425e-01 4.10258025e-01
2.55970240e-01 -9.13393736e-01 -3.96751851e-01 -7.25832433e-02
1.20558774e+00 -1.45234752e+00 1.00504056e-01 -3.80088359e-01
-9.06052768e-01 1.10131955e+00 7.22473025e-01 2.66716957e-01
1.10984117e-01 3.02725617e-04 3.02041680e-01 1.18757728e-02
-1.34222317e+00 2.03395158e-01 3.10988188e-01 5.36025226e-01
5.43665349e-01 3.34360093e-01 -3.57115000e-01 9.87458467e-01
-6.73136890e-01 -1.68403864e-01 7.91641891e-01 4.74031836e-01
-3.30234706e-01 -1.04252803e+00 -6.08763576e-01 5.57485819e-01
-3.54858696e-01 7.56209418e-02 -7.42539346e-01 1.22170818e+00
-1.89846218e-01 1.45180309e+00 -2.01253563e-01 -2.93737113e-01
5.23690164e-01 1.32001087e-01 -7.31554329e-02 -5.71331441e-01
-4.16595608e-01 -1.38939053e-01 2.71087766e-01 -6.22142017e-01
3.27285826e-02 -7.22332001e-01 -1.49388945e+00 -5.47768176e-01
-5.23623526e-01 4.36714917e-01 6.16909206e-01 1.07580006e+00
3.17379206e-01 3.99691552e-01 6.52946234e-01 -5.36721587e-01
-1.05668390e+00 -8.59636009e-01 -2.08846688e-01 3.87520313e-01
2.80094087e-01 -8.33865821e-01 -4.78492737e-01 -2.51348823e-01] | [8.752216339111328, 5.706368923187256] |
f471ab67-543f-42d1-9170-a28c0a5d7a3b | topic-guided-coherence-modeling-for-sentence | null | null | https://aclanthology.org/D19-1232 | https://aclanthology.org/D19-1232.pdf | Topic-Guided Coherence Modeling for Sentence Ordering by Preserving Global and Local Information | We propose a novel topic-guided coherence modeling (TGCM) for sentence ordering. Our attention based pointer decoder directly utilize sentence vectors in a permutation-invariant manner, without being compressed into a single fixed-length vector as the paragraph representation. Thus, TGCM can improve global dependencies among sentences and preserve relatively informative paragraph-level semantics. Moreover, to predict the next sentence, we capture topic-enhanced sentence-pair interactions between the current predicted sentence and each next-sentence candidate. With the coherent topical context matching, we promote local dependencies that help identify the tight semantic connections for sentence ordering. The experimental results show that TGCM outperforms state-of-the-art models from various perspectives. | ['Kyong-Ho Lee', 'Cheolheon Shin', 'Byungkook Oh', 'Seungmin Seo', 'Eunju Jo'] | 2019-11-01 | null | null | null | ijcnlp-2019-11 | ['sentence-ordering'] | ['natural-language-processing'] | [ 1.70277208e-01 4.30267351e-03 -6.29361808e-01 -6.98175132e-01
-8.51632059e-01 -5.39571159e-02 2.96063572e-01 5.63914120e-01
-3.45354944e-01 7.15451181e-01 1.17350304e+00 -1.81009218e-01
-6.63599297e-02 -4.54821199e-01 -6.22545958e-01 -3.78847659e-01
-4.39252146e-02 2.39083216e-01 4.58034664e-01 -2.34601706e-01
6.67343557e-01 -3.97370905e-01 -1.10186601e+00 7.57723927e-01
1.24545789e+00 4.53703612e-01 1.02279007e+00 6.38074815e-01
-4.29496765e-01 6.99182212e-01 -6.30006969e-01 -2.55028218e-01
-5.94845355e-01 -5.50837278e-01 -9.72873151e-01 -2.24326421e-02
5.96401870e-01 -1.52171165e-01 -4.39541519e-01 1.11361718e+00
1.42512977e-01 2.39958629e-01 3.10741156e-01 -5.02982140e-01
-7.94950664e-01 1.24657726e+00 -5.15707970e-01 5.14344215e-01
5.55732608e-01 -1.99972481e-01 1.63791037e+00 -8.58195961e-01
7.05734909e-01 1.37655258e+00 1.03922956e-01 2.49879673e-01
-1.14237928e+00 -3.48087370e-01 8.80690098e-01 5.82099617e-01
-1.09673524e+00 -3.41335714e-01 1.01494789e+00 -5.45959547e-02
1.60303605e+00 5.18873513e-01 6.88815176e-01 1.11285245e+00
8.61295998e-01 1.07874477e+00 3.93925190e-01 -4.18363929e-01
4.53066677e-02 -2.11318314e-01 9.61668611e-01 3.55716169e-01
7.91283026e-02 -7.11667359e-01 -9.60554302e-01 3.55836675e-02
2.36331880e-01 -8.30871239e-02 -2.90511668e-01 6.44421354e-02
-1.30774498e+00 7.86460340e-01 2.42283270e-01 5.61173081e-01
-3.33211929e-01 1.52902260e-01 7.26378977e-01 7.45507479e-02
7.48324275e-01 3.49502683e-01 -5.46861708e-01 -1.38876095e-01
-1.08267379e+00 2.32058633e-02 5.24257481e-01 1.24314821e+00
7.26016879e-01 -3.82374674e-01 -9.66394722e-01 7.05270946e-01
7.12388515e-01 2.82931596e-01 5.90090513e-01 -6.77896678e-01
1.11100483e+00 3.48982006e-01 -4.01942343e-01 -1.41110265e+00
-1.44147307e-01 -7.32957304e-01 -9.10497963e-01 -1.26203048e+00
-6.09435916e-01 6.95449561e-02 -3.69207054e-01 1.73062789e+00
-1.50310129e-01 2.79081553e-01 -6.84319362e-02 6.36425555e-01
7.95299590e-01 1.09549975e+00 1.85431018e-01 -6.59077346e-01
1.39772415e+00 -1.32603002e+00 -9.07834888e-01 -8.12381446e-01
6.55888140e-01 -6.60911322e-01 1.21968591e+00 -2.91741304e-02
-1.17456496e+00 -6.33938789e-01 -1.01903415e+00 -3.15894544e-01
3.12125683e-01 -3.73491347e-02 5.57094693e-01 -6.11011218e-03
-1.08339381e+00 4.01070714e-01 -9.30947125e-01 -2.46047512e-01
2.42879272e-01 9.56180170e-02 6.64735064e-02 -1.65767409e-02
-1.51043844e+00 7.58035600e-01 6.35031044e-01 -2.14248037e-04
-4.58963543e-01 -5.55702507e-01 -1.03462565e+00 7.07164884e-01
8.87030065e-02 -9.12761271e-01 1.10249364e+00 -1.36153385e-01
-1.32335317e+00 5.53741217e-01 -1.17766452e+00 -6.35685444e-01
-3.69113028e-01 -3.34603280e-01 -2.37449810e-01 2.21069366e-01
5.08248329e-01 6.93443596e-01 4.50929046e-01 -7.06089020e-01
-4.98176366e-01 -5.97964823e-02 -1.55982837e-01 5.70074022e-01
-7.81746507e-01 1.20732881e-01 -8.68566394e-01 -6.66196644e-01
3.03247839e-01 -3.49992901e-01 -3.27087998e-01 -8.10674191e-01
-1.01748204e+00 -4.58198547e-01 4.86534834e-01 -6.25535369e-01
2.06603146e+00 -1.97931480e+00 4.17235523e-01 -2.85324484e-01
6.83465833e-03 -2.61007100e-01 -2.53204077e-01 8.26578856e-01
1.59939498e-01 2.27275431e-01 -3.38145822e-01 -8.25933814e-01
-7.22838193e-02 1.25774577e-01 -4.98580456e-01 -1.83567498e-02
-1.49710821e-02 1.04454839e+00 -1.06307125e+00 -1.03976834e+00
1.02669924e-01 -8.11290753e-04 -7.40775943e-01 1.00362718e-01
-4.41422105e-01 4.10534382e-01 -7.21652687e-01 5.09106666e-02
5.99753797e-01 -5.74196160e-01 5.06494224e-01 -6.20134212e-02
1.12120762e-01 1.18870986e+00 -3.11465412e-01 2.29226470e+00
-5.55672526e-01 7.67384648e-01 -4.26635057e-01 -7.99008906e-01
9.28545356e-01 1.38943180e-01 2.28156656e-01 -7.41949439e-01
-2.13135943e-01 -1.87404230e-01 -1.85888946e-01 -4.45279598e-01
9.38632011e-01 2.18358397e-01 -4.24913287e-01 4.03156966e-01
5.75113557e-02 2.74808019e-01 4.32181150e-01 7.96723843e-01
8.49736214e-01 -3.15272629e-01 3.17915082e-01 -5.45319080e-01
6.94074035e-01 -2.61002064e-01 6.47571445e-01 7.43434966e-01
1.42169625e-01 5.42336285e-01 5.90438843e-01 9.62379426e-02
-6.44696891e-01 -1.03230107e+00 -1.09975435e-01 1.06584644e+00
4.09631193e-01 -9.71197069e-01 -6.26248419e-01 -5.45408309e-01
-2.99482375e-01 1.28932416e+00 -2.79564202e-01 -3.36867690e-01
-8.46539974e-01 -5.10080636e-01 -3.69060226e-02 5.42668760e-01
4.26541030e-01 -9.37996984e-01 -9.28689167e-02 4.69287872e-01
-7.29234755e-01 -1.05045068e+00 -1.17575216e+00 -1.31372809e-01
-1.08106923e+00 -6.02951348e-01 -3.94918710e-01 -1.10228968e+00
6.63159072e-01 5.88108242e-01 1.31307030e+00 -6.91952482e-02
4.06072974e-01 -1.75347388e-01 -3.88401866e-01 2.68981993e-01
3.33328760e-04 4.72257763e-01 -2.54666686e-01 -3.15538466e-01
5.82902670e-01 -5.58715463e-01 -7.20236659e-01 -1.32344797e-01
-4.08581227e-01 3.90196562e-01 4.46151942e-01 1.01157784e+00
4.38271821e-01 -2.50336468e-01 6.03903413e-01 -8.73103976e-01
1.04249549e+00 -5.29479682e-01 -1.33275092e-01 5.11332512e-01
-4.35016155e-01 8.22331682e-02 6.38348699e-01 -5.62722310e-02
-1.24715614e+00 -4.91279811e-01 -1.92306280e-01 1.56035116e-02
-9.38255899e-03 9.70342994e-01 -2.61559963e-01 8.29537392e-01
-4.54272404e-02 7.03434289e-01 -4.93265420e-01 -5.59324801e-01
4.08709735e-01 6.12525403e-01 4.50830221e-01 -3.92081648e-01
9.20493379e-02 2.37112008e-02 -2.88941115e-01 -7.80072212e-01
-1.34380734e+00 -6.96266294e-01 -7.82029271e-01 7.79572055e-02
9.37431335e-01 -9.37160194e-01 -4.52074528e-01 1.76151879e-02
-1.96898282e+00 2.32330382e-01 7.18375444e-02 4.86702681e-01
-1.88136026e-01 9.18562770e-01 -9.38183248e-01 -5.56857288e-01
-6.04488730e-01 -1.02645314e+00 1.24478972e+00 1.66986942e-01
-5.70676088e-01 -1.27040160e+00 1.58404246e-01 2.65887380e-01
6.25514090e-02 -5.40018678e-01 1.10438645e+00 -5.59436083e-01
-7.67278731e-01 -1.45989126e-02 -3.02966759e-02 5.08300923e-02
1.93906844e-01 -2.45095015e-01 -5.10007083e-01 -2.29961827e-01
1.72635302e-01 1.75090328e-01 1.39483476e+00 6.81910813e-01
1.25141907e+00 -5.06397009e-01 -6.32570088e-01 3.82848263e-01
9.76400077e-01 -7.88507164e-02 4.66432840e-01 2.45270595e-01
6.75249398e-01 4.86021966e-01 7.97506034e-01 3.28544557e-01
7.29223609e-01 7.05652714e-01 8.71130973e-02 2.73530632e-01
1.69324223e-02 -6.40448034e-01 5.40405273e-01 1.86762369e+00
6.20969117e-01 -5.83285332e-01 -5.55476367e-01 6.73360884e-01
-2.24961162e+00 -1.11336863e+00 -4.37505305e-01 1.93747556e+00
1.10273242e+00 4.74977762e-01 -5.54894269e-01 -3.55326235e-01
7.34931171e-01 7.30814695e-01 -1.13643959e-01 -5.16046047e-01
-1.85771957e-01 -1.12448156e-01 4.57488336e-02 9.36005712e-01
-1.06303251e+00 1.24476683e+00 6.50005341e+00 1.03540766e+00
-5.80893695e-01 1.33787677e-01 7.24233806e-01 -1.29308715e-01
-9.15815055e-01 1.64824396e-01 -1.28353143e+00 9.77754653e-01
7.74089992e-01 -6.45786643e-01 -1.62448317e-01 6.75302505e-01
5.58329701e-01 -1.76363319e-01 -1.01923621e+00 5.50150156e-01
4.39514071e-01 -1.74945831e+00 3.64036590e-01 -3.54814649e-01
9.20419335e-01 -1.71791807e-01 -7.58076608e-02 2.17770949e-01
-2.09359139e-01 -5.43406308e-01 4.40577507e-01 7.57706344e-01
2.92016715e-01 -7.60653496e-01 8.02768946e-01 5.08969009e-01
-1.36916864e+00 1.84943393e-01 -7.35653579e-01 -1.53203443e-01
5.86839020e-01 8.90824020e-01 -6.17009819e-01 6.77175760e-01
3.14726323e-01 1.39607179e+00 -5.96035540e-01 8.51361752e-01
-5.00005662e-01 8.44819605e-01 6.34511113e-02 -4.92263466e-01
3.63425165e-01 -6.87157884e-02 8.55322361e-01 1.54270124e+00
1.56688035e-01 -1.02174833e-01 2.79610932e-01 7.39492774e-01
-1.79431159e-02 2.80812949e-01 -2.69403845e-01 2.65895367e-01
8.73460293e-01 8.94936562e-01 -4.53833163e-01 -4.19003457e-01
-4.02306795e-01 1.30961716e+00 4.37151670e-01 2.40256444e-01
-5.92745423e-01 -3.59016478e-01 6.04884148e-01 -3.62859786e-01
4.01844531e-01 -3.33207607e-01 -6.63058579e-01 -1.50325489e+00
3.78021181e-01 -2.26376817e-01 2.18446016e-01 -6.22312069e-01
-1.36129713e+00 5.47694206e-01 1.01662628e-01 -1.22072625e+00
-3.27953756e-01 9.99261439e-02 -1.16715384e+00 8.83904457e-01
-1.37483764e+00 -1.00369871e+00 2.21519247e-01 1.18741155e-01
1.23653984e+00 -5.57741523e-03 7.62871087e-01 -5.91236465e-02
-6.85041487e-01 5.57608902e-01 1.25122994e-01 -1.30056351e-01
4.58441764e-01 -1.10524690e+00 8.16404939e-01 1.14177358e+00
2.54962862e-01 1.17449796e+00 6.24460459e-01 -9.57974017e-01
-1.13966930e+00 -1.13725674e+00 1.94300163e+00 -2.11307287e-01
6.40521049e-01 -5.79604983e-01 -1.03711259e+00 6.07864499e-01
8.22456896e-01 -9.72811878e-01 7.05901682e-01 6.50067627e-01
-5.85986897e-02 -9.77650657e-02 -1.67870820e-01 8.25907469e-01
1.11492968e+00 -7.61645317e-01 -1.07019567e+00 5.48271060e-01
1.54088652e+00 -2.97556162e-01 -6.29064023e-01 1.73188835e-01
1.24143191e-01 -7.64070690e-01 8.71889114e-01 -5.85025787e-01
9.62110221e-01 -6.51420876e-02 -9.03931707e-02 -1.25725198e+00
-7.42878377e-01 -4.09014463e-01 -4.56956983e-01 1.38722861e+00
7.11273193e-01 -3.45865130e-01 7.30263174e-01 1.57297209e-01
-5.96033812e-01 -9.84156668e-01 -9.09209549e-01 -6.28581047e-01
-9.09021795e-02 -3.54874253e-01 4.99852747e-01 6.78782105e-01
7.58424819e-01 1.04362261e+00 -4.92598295e-01 6.29278868e-02
4.65262711e-01 5.14868319e-01 1.23398267e-01 -7.05156982e-01
-4.66472089e-01 -5.18388093e-01 -9.54780728e-02 -2.15345335e+00
5.36618471e-01 -1.07578492e+00 2.74277985e-01 -1.86733007e+00
8.94861579e-01 1.79563865e-01 -5.41203082e-01 -7.39306444e-03
-8.72784138e-01 -4.58343506e-01 1.71338573e-01 3.28928351e-01
-1.23065233e+00 1.05275679e+00 1.38522172e+00 -3.76179129e-01
-2.43601073e-02 -3.97528671e-02 -8.65176380e-01 3.68021727e-01
7.27327585e-01 -5.38775504e-01 -6.33172095e-01 -9.29732859e-01
2.29639813e-01 2.24566370e-01 -5.83372861e-02 -6.13542855e-01
7.47965872e-01 -1.51523888e-01 8.87391940e-02 -1.32865834e+00
2.43358046e-01 -1.72789022e-01 -4.96692389e-01 4.23869938e-01
-9.65408981e-01 -4.78889756e-02 -2.36053392e-02 9.04505730e-01
-5.07861018e-01 -3.50632012e-01 1.23394459e-01 9.02703255e-02
-6.39864445e-01 1.79740876e-01 -4.47666883e-01 4.65683453e-02
6.97886646e-01 7.15190694e-02 -3.95727158e-01 -4.37475741e-01
-5.29981494e-01 6.83026671e-01 1.42991409e-01 5.97307503e-01
8.24664593e-01 -1.21865427e+00 -8.31086755e-01 9.43199173e-02
2.14959797e-03 2.45319568e-02 6.13510787e-01 5.85932612e-01
1.09151900e-01 1.14050102e+00 3.77796054e-01 -8.68607759e-01
-1.41336524e+00 3.94030839e-01 -2.79091924e-01 -6.37327790e-01
-6.50764704e-01 1.12774181e+00 4.73273963e-01 -1.19608387e-01
1.01371743e-01 -5.01665235e-01 -4.11549479e-01 -2.27903441e-01
6.26432478e-01 -6.21459410e-02 -1.31968617e-01 -2.09447697e-01
-5.08020759e-01 3.82620513e-01 -5.68593562e-01 -7.78065249e-02
1.10450375e+00 -6.02305055e-01 -6.31655216e-01 5.53733587e-01
1.34833682e+00 -1.31073929e-02 -8.89350891e-01 -2.42212847e-01
3.03683162e-01 -5.50059438e-01 1.99056491e-02 -3.09233844e-01
-6.48819745e-01 9.76801097e-01 -4.57113713e-01 4.64253910e-02
1.02242577e+00 2.79661715e-01 1.03302586e+00 3.83422375e-01
1.86263219e-01 -9.01357591e-01 2.14385137e-01 1.00987136e+00
9.56440270e-01 -8.03378642e-01 1.36871740e-01 -7.66899228e-01
-7.95490801e-01 1.02749217e+00 7.95007110e-01 3.71421501e-02
5.24881065e-01 -1.36499807e-01 -5.66205621e-01 -4.16463912e-02
-1.44190717e+00 9.57256556e-02 5.50151587e-01 1.73470646e-01
8.57802689e-01 1.60425156e-02 -7.52925813e-01 8.52669299e-01
-2.00785622e-01 -4.89836335e-01 1.98543459e-01 5.41065395e-01
-8.79541993e-01 -1.20957887e+00 2.13098794e-01 6.07465327e-01
-3.43652554e-02 -8.94232631e-01 -2.20925346e-01 1.30865157e-01
-2.98978955e-01 1.06908631e+00 4.79304492e-01 -4.38762516e-01
-5.49185264e-04 -5.37351258e-02 1.33209825e-01 -1.04719210e+00
-4.58808780e-01 3.42827827e-01 3.12633842e-01 -3.84833306e-01
-1.25924632e-01 -8.73848319e-01 -1.20377767e+00 -1.59470156e-01
-3.32219750e-01 4.33689147e-01 1.31570816e-01 1.11796367e+00
7.84933209e-01 8.62748086e-01 9.32497740e-01 -4.76140499e-01
-3.80375445e-01 -1.23504114e+00 -2.23224252e-01 7.58241909e-03
1.46654069e-01 -5.25031276e-02 -9.66036767e-02 -1.39937297e-01] | [12.23622989654541, 9.33604907989502] |
6380a1ea-f571-4481-bcc4-d4f7a532d6d8 | shading-annotations-in-the-wild | 1705.01156 | null | http://arxiv.org/abs/1705.01156v1 | http://arxiv.org/pdf/1705.01156v1.pdf | Shading Annotations in the Wild | Understanding shading effects in images is critical for a variety of vision
and graphics problems, including intrinsic image decomposition, shadow removal,
image relighting, and inverse rendering. As is the case with other vision
tasks, machine learning is a promising approach to understanding shading - but
there is little ground truth shading data available for real-world images. We
introduce Shading Annotations in the Wild (SAW), a new large-scale, public
dataset of shading annotations in indoor scenes, comprised of multiple forms of
shading judgments obtained via crowdsourcing, along with shading annotations
automatically generated from RGB-D imagery. We use this data to train a
convolutional neural network to predict per-pixel shading information in an
image. We demonstrate the value of our data and network in an application to
intrinsic images, where we can reduce decomposition artifacts produced by
existing algorithms. Our database is available at
http://opensurfaces.cs.cornell.edu/saw/. | ['Balazs Kovacs', 'Sean Bell', 'Kavita Bala', 'Noah Snavely'] | 2017-05-02 | shading-annotations-in-the-wild-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Kovacs_Shading_Annotations_in_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Kovacs_Shading_Annotations_in_CVPR_2017_paper.pdf | cvpr-2017-7 | ['shadow-removal', 'intrinsic-image-decomposition', 'image-relighting'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 7.49115288e-01 5.88197000e-02 5.65200031e-01 -7.28616834e-01
-7.00129628e-01 -6.94307029e-01 3.54540050e-01 -1.13525592e-01
-3.58968787e-02 4.62804198e-01 5.16439319e-01 -4.28366005e-01
6.09076917e-01 -8.07552457e-01 -6.62543714e-01 -4.62952971e-01
3.54536682e-01 3.01278889e-01 3.25556457e-01 -3.74651134e-01
2.70085961e-01 6.78733766e-01 -1.74796057e+00 6.64172232e-01
9.06384587e-01 8.59322131e-01 4.05960292e-01 9.16852057e-01
-1.95642754e-01 8.48376930e-01 -2.63368011e-01 8.06025788e-02
5.21155059e-01 -1.46852449e-01 -5.77028275e-01 3.87582064e-01
8.85291338e-01 -6.36624873e-01 -9.86151248e-02 6.36762023e-01
5.91597438e-01 -4.95427288e-03 3.44565958e-01 -1.13348818e+00
-5.95810294e-01 -2.64095426e-01 -7.04550982e-01 -2.43867978e-01
5.72649121e-01 4.00088102e-01 8.39842319e-01 -1.04203892e+00
5.11107326e-01 1.15483129e+00 8.22084606e-01 3.73271704e-01
-1.33282089e+00 -1.90789998e-01 -2.26396516e-01 -3.88411433e-01
-9.97064710e-01 -5.59321940e-01 9.89333332e-01 -6.78977549e-01
7.97615469e-01 4.85081166e-01 7.56181240e-01 7.17306435e-01
-1.53125942e-01 7.45412171e-01 1.52582002e+00 -5.73705435e-01
3.27994138e-01 9.16046742e-03 -1.15807496e-01 9.09109950e-01
-2.85413414e-01 1.84861273e-01 -8.17217290e-01 -2.76850075e-01
9.00680482e-01 -1.79738015e-01 -4.84018594e-01 -5.26747942e-01
-9.54566061e-01 6.22754574e-01 6.93415940e-01 -4.69500661e-01
-7.81466737e-02 2.30789870e-01 -4.55880500e-02 -1.51383653e-01
7.73309708e-01 3.04752439e-01 -7.48869956e-01 1.55637115e-01
-6.07614040e-01 3.43889028e-01 8.48340333e-01 7.58134544e-01
1.18586707e+00 9.40306783e-02 2.67995059e-01 7.96757519e-01
2.47921959e-01 7.25176692e-01 -1.50495544e-01 -1.69889331e+00
1.60976201e-01 5.07064998e-01 3.98983687e-01 -8.47577929e-01
-3.62318814e-01 2.71617740e-01 -4.65540260e-01 9.58397150e-01
4.60691273e-01 -1.47112831e-01 -8.69540513e-01 1.28696620e+00
5.59573233e-01 -4.17958125e-02 -2.58341283e-01 1.08439076e+00
1.16133690e+00 4.93522853e-01 -4.16548431e-01 3.75043333e-01
1.01789093e+00 -9.15006638e-01 -1.90303728e-01 -5.93174577e-01
3.99184793e-01 -9.97955799e-01 1.52219689e+00 6.88708186e-01
-9.40393686e-01 -3.63378435e-01 -7.59235084e-01 -8.46943796e-01
-3.62819701e-01 1.23211004e-01 1.13046169e+00 6.66806936e-01
-1.34116006e+00 4.78549272e-01 -6.22121871e-01 -3.62087786e-01
6.84031248e-01 1.26612961e-01 -2.17703804e-01 -2.95182765e-01
-4.69874740e-01 5.41572273e-01 -4.52072144e-01 1.63745522e-01
-9.00907636e-01 -9.05046880e-01 -9.47138727e-01 -2.61177272e-01
1.88280284e-01 -5.13777077e-01 1.41291761e+00 -1.23527241e+00
-1.24594915e+00 1.20293748e+00 -4.15743142e-01 5.09981625e-02
5.28325379e-01 -2.77155161e-01 2.69958496e-01 -2.00252056e-01
1.40770227e-01 8.13546121e-01 6.04473889e-01 -1.91059458e+00
-3.31923068e-01 -5.27599871e-01 2.27409482e-01 3.92008245e-01
1.19976372e-01 -1.85192645e-01 -3.80979955e-01 -2.06985474e-01
2.15825617e-01 -8.30909193e-01 -3.55042070e-01 6.33315980e-01
-5.30117691e-01 2.58023620e-01 8.49970460e-01 -9.27150846e-01
1.77842826e-01 -1.95689940e+00 -2.20491752e-01 3.25143076e-02
1.52297154e-01 -6.38259277e-02 -1.61603972e-01 4.73067909e-01
8.61659795e-02 -1.22263961e-01 -5.22132874e-01 -5.71989477e-01
-9.75642279e-02 2.65848249e-01 -6.64572954e-01 3.94266069e-01
5.57318442e-02 6.45740688e-01 -8.38500321e-01 -2.22226709e-01
6.37338400e-01 7.99730778e-01 -5.35659134e-01 3.86180758e-01
-7.20731735e-01 9.09141719e-01 -9.13571641e-02 7.83958018e-01
8.84911299e-01 -1.45607293e-01 7.82373399e-02 -5.47422469e-02
-3.35311323e-01 2.05607399e-01 -9.66614842e-01 1.87520838e+00
-7.08140671e-01 1.08072674e+00 2.17134237e-01 -1.58677101e-01
8.47422898e-01 -1.54361591e-01 4.68887478e-01 -7.02486217e-01
-3.16163078e-02 4.69140820e-02 -5.60326397e-01 -4.47225839e-01
6.61984801e-01 2.28434145e-01 4.10838008e-01 5.83991230e-01
-5.90865672e-01 -1.00481713e+00 -1.54601887e-01 1.83841020e-01
1.16778195e+00 5.96665502e-01 -3.94499376e-02 -3.44052017e-01
1.76354378e-01 3.53642046e-01 4.55677718e-01 5.34336448e-01
1.65153369e-01 1.26644790e+00 4.82905023e-02 -7.79451847e-01
-1.27662718e+00 -1.09066486e+00 -1.36466518e-01 1.13895941e+00
1.58214837e-01 -9.27299187e-02 -9.04839933e-01 1.66717134e-02
2.40523726e-01 5.68541884e-01 -6.46340728e-01 6.04617000e-01
-4.05872464e-01 -4.81333435e-01 5.09428643e-02 5.08754492e-01
5.28535306e-01 -1.17117155e+00 -9.77958620e-01 -3.07303131e-01
-5.17249964e-02 -1.04768515e+00 -3.68579835e-01 2.34544188e-01
-6.89841986e-01 -1.33091891e+00 -3.44724387e-01 -6.32749021e-01
8.01879883e-01 8.71754348e-01 1.76295364e+00 6.37201726e-01
-8.88332307e-01 6.86976492e-01 -7.46727958e-02 -6.17364228e-01
-2.31141865e-01 -4.81443852e-01 -5.75809360e-01 -2.62139469e-01
1.04234613e-01 -5.47860205e-01 -8.54759455e-01 3.09432775e-01
-1.00657427e+00 5.93947768e-01 -1.65915191e-01 5.42854548e-01
6.03888512e-01 -3.44392687e-01 -3.16888630e-01 -1.19297659e+00
3.51390183e-01 -9.61320009e-03 -1.02633333e+00 6.98932335e-02
-1.17802270e-01 -3.42770427e-01 3.14544141e-01 2.61039525e-01
-1.55373764e+00 5.57159126e-01 -4.05843509e-03 -5.41273179e-03
-5.78129590e-01 -5.63361906e-02 -2.03853279e-01 -4.08287853e-01
9.27692771e-01 -1.35856941e-01 -2.60443687e-01 -3.72317910e-01
6.68509245e-01 4.68200713e-01 6.53352976e-01 -5.63361526e-01
6.43210709e-01 1.10284090e+00 -4.14597094e-02 -1.26028800e+00
-1.16903281e+00 -4.20476496e-01 -7.21895874e-01 -2.78923392e-01
9.59733844e-01 -1.02558815e+00 -7.93415725e-01 6.06828213e-01
-1.35494065e+00 -1.21845841e+00 -1.77819803e-01 -2.55613476e-01
-6.38671160e-01 2.85731465e-01 -4.78359759e-01 -9.37547684e-01
-8.26073363e-02 -1.07538569e+00 1.65996933e+00 4.06624734e-01
-4.09342051e-02 -9.18739378e-01 2.40980417e-01 8.10185134e-01
2.80530542e-01 4.58795875e-01 6.54447317e-01 3.99052590e-01
-9.94247019e-01 2.64072269e-01 -4.36371654e-01 2.99352616e-01
1.98857933e-01 2.12689236e-01 -1.68767202e+00 9.85470414e-02
-6.88556358e-02 -7.41609097e-01 9.19490933e-01 4.35048282e-01
1.49243665e+00 1.13359191e-01 9.76784974e-02 7.34629273e-01
1.70221841e+00 -1.59617439e-01 8.07538867e-01 1.07372686e-01
1.04419494e+00 8.16668034e-01 3.07559311e-01 6.40775740e-01
7.13692009e-01 4.47497159e-01 6.38138354e-01 -7.27017343e-01
-4.51176375e-01 -2.05367774e-01 -1.48612916e-01 2.64387429e-01
-1.15253106e-01 -3.02601606e-01 -1.31859183e+00 3.97563964e-01
-1.70001996e+00 -6.22312784e-01 -6.60546124e-01 2.17008924e+00
7.73867965e-01 -3.14155310e-01 -3.66929978e-01 -6.85367286e-02
2.29699984e-01 2.27868557e-01 -6.14537835e-01 -4.01812583e-01
-3.73868674e-01 2.72068828e-01 6.31289363e-01 1.06049979e+00
-1.07055628e+00 1.05800915e+00 6.20171261e+00 6.88114539e-02
-8.36358190e-01 2.22272128e-02 1.02198315e+00 1.05322830e-01
-8.33566129e-01 2.59150922e-01 -2.52428949e-01 6.04768060e-02
2.31114611e-01 4.80683506e-01 9.04793322e-01 7.49032438e-01
5.17532229e-01 -8.95097613e-01 -9.93099868e-01 1.03022671e+00
2.42217481e-01 -1.31274271e+00 -6.25900686e-01 -3.75407785e-02
1.11395299e+00 3.98828059e-01 -7.96478316e-02 -4.22809154e-01
8.84887159e-01 -1.06703484e+00 4.31110501e-01 5.47069907e-01
7.08968222e-01 -2.14264914e-01 1.75257877e-01 2.77579576e-01
-9.12592828e-01 2.35907823e-01 -4.52805638e-01 -3.31119627e-01
1.51473239e-01 1.02607310e+00 -9.46433365e-01 -3.95095535e-02
7.80030966e-01 5.01328647e-01 -7.83733785e-01 8.11407924e-01
-5.86659074e-01 2.60658711e-01 -3.90801609e-01 1.76148713e-01
-3.36447477e-01 -5.67011714e-01 -6.01210557e-02 8.71831954e-01
-1.12036653e-01 3.63099098e-01 2.74038136e-01 9.07074630e-01
-1.62014320e-01 -2.03046411e-01 -9.70482767e-01 4.39079374e-01
2.51673460e-01 1.40706849e+00 -7.91520476e-01 -1.11785166e-01
-3.27946633e-01 1.23501062e+00 4.56514865e-01 5.93460739e-01
-4.40676749e-01 -1.45992101e-03 6.40720069e-01 1.35692060e-01
-1.31864429e-01 -5.45905769e-01 -9.53094959e-01 -1.05784273e+00
-1.51308766e-02 -6.00951374e-01 -1.95768595e-01 -1.82249534e+00
-1.15914142e+00 3.36017221e-01 -3.69706482e-01 -8.27556074e-01
5.56834377e-02 -7.20653415e-01 -6.47995472e-01 9.14491415e-01
-1.65732741e+00 -1.29037762e+00 -1.16003883e+00 4.61036026e-01
6.87316716e-01 3.27579647e-01 9.43209827e-01 -7.15433657e-02
-7.13954866e-03 -3.61576855e-01 1.40029490e-01 4.30382267e-02
6.28249586e-01 -1.56249714e+00 7.94296503e-01 6.96155488e-01
3.23098004e-01 8.61504823e-02 8.56779873e-01 -5.85273981e-01
-1.71429574e+00 -9.20034289e-01 3.66059631e-01 -7.60002911e-01
2.70725816e-01 -5.26697278e-01 -7.76245415e-01 6.32049739e-01
3.08896303e-01 1.94277734e-01 5.79617143e-01 -2.06711069e-02
-4.07656610e-01 -4.98986095e-02 -1.18750942e+00 7.79519022e-01
1.26733625e+00 -7.26943195e-01 2.15931498e-02 8.94829333e-01
4.91613239e-01 -8.63123775e-01 -5.04062951e-01 1.02105089e-01
5.44557810e-01 -1.69652617e+00 1.16172326e+00 -7.07104057e-02
8.74841571e-01 -4.89228815e-01 -4.38678950e-01 -1.32287848e+00
1.35217041e-01 -3.06850940e-01 4.44585621e-01 9.87944841e-01
2.40939975e-01 -4.47941184e-01 1.14338529e+00 1.30091858e+00
-2.13369414e-01 -3.43953818e-01 -4.93840724e-01 -1.21938683e-01
-1.41090512e-01 -4.95203853e-01 5.17599642e-01 7.62799084e-01
-6.90746009e-01 1.18741676e-01 -2.31780604e-01 3.31938803e-01
1.00585830e+00 5.96578181e-01 1.42491746e+00 -1.23733747e+00
-3.99386674e-01 1.78133082e-02 6.57636151e-02 -9.62979436e-01
1.52343020e-01 -5.54238617e-01 5.61001241e-01 -1.95235789e+00
9.07185972e-02 -7.79637396e-01 5.71809947e-01 6.92404151e-01
-1.05098143e-01 6.19340420e-01 1.46539286e-01 7.23241195e-02
-2.69658625e-01 5.17794132e-01 1.17976856e+00 5.47720045e-02
-9.31706578e-02 -3.30220670e-01 -5.05572319e-01 1.06916988e+00
1.09896255e+00 -1.17074341e-01 -4.13191974e-01 -1.13094068e+00
1.68861866e-01 -5.62152155e-02 6.43897116e-01 -8.09561610e-01
4.50666621e-02 -3.60741794e-01 6.74103260e-01 -6.70909822e-01
7.94554472e-01 -7.19971657e-01 1.29363939e-01 7.27416277e-02
-2.24020869e-01 -1.59413069e-01 1.22686855e-01 2.91264981e-01
2.85143524e-01 -4.10797037e-02 6.72483087e-01 -4.69251186e-01
-8.81229758e-01 1.29905224e-01 5.68904467e-02 9.09986496e-02
5.81241786e-01 -1.56516075e-01 -6.82411909e-01 -4.63493615e-01
-1.85438737e-01 8.36329237e-02 1.05620313e+00 1.18176855e-01
8.56141269e-01 -1.00728846e+00 -4.80713606e-01 4.34272677e-01
1.95496991e-01 5.32364905e-01 1.50462776e-01 2.42316678e-01
-1.14448833e+00 -1.66231200e-01 -4.46146429e-02 -7.49824286e-01
-1.36146653e+00 2.14243848e-02 3.32480371e-01 4.58578885e-01
-6.68578207e-01 9.67077971e-01 5.90578794e-01 -8.23174655e-01
1.24885179e-02 -3.13494235e-01 5.25966167e-01 -6.87947989e-01
4.99907494e-01 1.65711761e-01 2.72325784e-01 -4.45989668e-01
-1.25067443e-01 5.56051970e-01 6.43948853e-01 -3.35034937e-01
1.51784825e+00 -5.95425314e-04 -4.65206236e-01 5.19853592e-01
9.11367774e-01 1.85909227e-01 -1.79937899e+00 -8.94113034e-02
-1.83315352e-01 -1.09669876e+00 1.92373261e-01 -8.47863317e-01
-1.05256176e+00 1.04565644e+00 5.83971381e-01 3.38878706e-02
1.05967736e+00 -1.42050639e-01 5.69040120e-01 4.45514470e-01
6.81098223e-01 -1.07000852e+00 1.42593592e-01 4.37224299e-01
1.12269843e+00 -1.79213226e+00 2.35496864e-01 -8.05416882e-01
-6.89431429e-01 1.06682408e+00 6.36806488e-01 -2.30028674e-01
6.20068669e-01 6.85805798e-01 5.25051177e-01 -2.41835058e-01
-4.26581055e-01 -1.38903171e-01 2.30573323e-02 9.79379892e-01
6.49473965e-01 1.37224123e-01 3.23475033e-01 -1.70966625e-01
-3.14142019e-01 6.20038472e-02 8.20710599e-01 9.85651612e-01
-5.87444425e-01 -1.02529490e+00 -6.81161404e-01 3.50629836e-01
2.06417769e-01 -4.02160496e-01 -7.46875644e-01 2.52418041e-01
9.17344168e-02 1.10572505e+00 -7.19997287e-02 2.53438037e-02
1.25230774e-01 -3.43538135e-01 5.77306151e-01 -8.22763205e-01
-4.22880083e-01 1.47869941e-02 2.80337751e-01 -8.73179555e-01
-4.67810690e-01 -7.51578391e-01 -1.14001608e+00 -2.18542576e-01
7.22233206e-02 -5.22758842e-01 1.02833915e+00 6.54006183e-01
2.29158670e-01 3.14581394e-01 7.44112611e-01 -1.56167960e+00
2.38096178e-01 -5.17515421e-01 -5.26316941e-01 5.88123679e-01
4.05056566e-01 -3.04094732e-01 -3.26022446e-01 4.96080309e-01] | [9.725348472595215, -3.018434762954712] |
8c1de538-65ee-440e-b933-2aa7fda6679e | weakly-supervised-forced-alignment-of | 2306.00996 | null | https://arxiv.org/abs/2306.00996v1 | https://arxiv.org/pdf/2306.00996v1.pdf | Weakly-supervised forced alignment of disfluent speech using phoneme-level modeling | The study of speech disorders can benefit greatly from time-aligned data. However, audio-text mismatches in disfluent speech cause rapid performance degradation for modern speech aligners, hindering the use of automatic approaches. In this work, we propose a simple and effective modification of alignment graph construction of CTC-based models using Weighted Finite State Transducers. The proposed weakly-supervised approach alleviates the need for verbatim transcription of speech disfluencies for forced alignment. During the graph construction, we allow the modeling of common speech disfluencies, i.e. repetitions and omissions. Further, we show that by assessing the degree of audio-text mismatch through the use of Oracle Error Rate, our method can be effectively used in the wild. Our evaluation on a corrupted version of the TIMIT test set and the UCLASS dataset shows significant improvements, particularly for recall, achieving a 23-25% relative improvement over our baselines. | ['Vassilis Katsouros', 'Athanasios Katsamanis', 'Georgios Paraskevopoulos', 'Theodoros Kouzelis'] | 2023-05-30 | null | null | null | null | ['graph-construction'] | ['graphs'] | [ 3.32956225e-01 2.50534505e-01 9.80334803e-02 -4.40503985e-01
-1.41894150e+00 -7.01848269e-01 1.78005114e-01 2.73833066e-01
-4.03847694e-01 4.24985588e-01 5.62382758e-01 -4.37342286e-01
1.46116465e-01 1.15241837e-02 -3.49391282e-01 -4.56632644e-01
-8.04251134e-02 5.72249293e-01 8.76526311e-02 -5.37870117e-02
-2.83308208e-01 -2.54756995e-02 -1.10975647e+00 3.16329986e-01
9.25156593e-01 4.44158375e-01 4.01972800e-01 7.67629921e-01
1.62025616e-01 5.73087215e-01 -8.84460926e-01 -3.37031662e-01
-4.97718602e-02 -5.80945492e-01 -8.05370271e-01 1.68327227e-01
3.61978501e-01 -1.61331773e-01 -2.96224385e-01 9.14062679e-01
8.69640946e-01 -4.63967174e-02 8.49242434e-02 -8.19643676e-01
2.33288288e-01 8.09809208e-01 2.69068200e-02 5.47271967e-01
5.74386775e-01 -2.42071804e-02 1.08709085e+00 -7.00900257e-01
4.69068110e-01 1.04580641e+00 5.32010794e-01 4.95067626e-01
-1.25060928e+00 -5.55418015e-01 1.45881712e-01 1.41000941e-01
-1.24216509e+00 -1.00337434e+00 4.55860913e-01 -3.13586324e-01
1.63306117e+00 5.59381545e-01 5.08291245e-01 1.07819450e+00
-4.29918915e-01 5.86694360e-01 8.14820230e-01 -7.89522350e-01
8.85423571e-02 -4.21796709e-01 2.60365605e-01 4.13979232e-01
-2.36104965e-01 4.01993059e-02 -7.35647082e-01 -2.84761876e-01
1.63374066e-01 -6.47189081e-01 -4.70463812e-01 2.93351382e-01
-1.04521954e+00 3.85391831e-01 -2.38794401e-01 7.16509521e-01
-4.99989651e-02 -1.64613709e-01 5.77549160e-01 4.58913177e-01
8.53217840e-01 3.13390762e-01 -3.61950278e-01 -7.25135863e-01
-1.21191931e+00 -2.66149715e-02 5.38361609e-01 9.80781794e-01
1.14051722e-01 1.69850305e-01 -2.68403739e-01 1.19961441e+00
4.17439826e-02 5.47772527e-01 6.21274948e-01 -7.26939082e-01
9.72892821e-01 2.27609240e-02 -6.10656850e-02 -5.19438624e-01
-4.03877735e-01 -4.76015031e-01 -4.03708488e-01 -4.99536604e-01
3.63751054e-01 -2.32952356e-01 -1.08136272e+00 1.98088408e+00
1.62687182e-01 3.38758409e-01 -5.19257085e-03 5.32313824e-01
2.77684152e-01 5.10354161e-01 -3.48918810e-02 -7.40092397e-01
1.07794666e+00 -9.11293328e-01 -1.28584754e+00 -4.26861495e-01
9.55737472e-01 -1.00627041e+00 9.32749867e-01 4.80657995e-01
-1.35924089e+00 -2.91574389e-01 -8.98195386e-01 1.92171678e-01
3.09223086e-01 7.84283131e-02 1.06592663e-01 7.39603281e-01
-1.05269706e+00 5.47262549e-01 -1.21281707e+00 -1.94626704e-01
-1.82401925e-01 4.73946184e-01 -5.31602859e-01 -7.46240541e-02
-1.16390622e+00 9.03259277e-01 2.34150216e-01 1.95921049e-01
-4.93780434e-01 -4.68916565e-01 -9.52822626e-01 -8.34683925e-02
1.83852971e-01 -9.76881161e-02 1.65308821e+00 -6.48077428e-01
-1.37492955e+00 7.74915040e-01 -5.67662835e-01 -5.49307644e-01
4.20832306e-01 -3.98866743e-01 -7.79170930e-01 1.13782249e-01
-8.67320746e-02 1.82317272e-01 7.17271030e-01 -4.55932081e-01
-5.48660278e-01 -9.73677188e-02 -4.53670233e-01 3.54594022e-01
-3.59026939e-01 4.64466214e-01 -5.67066729e-01 -9.12451923e-01
2.43784413e-01 -1.24596536e+00 -1.07882865e-01 -8.38157594e-01
-5.83137572e-01 -1.10930070e-01 4.60021645e-01 -1.19070172e+00
1.67438996e+00 -2.35266018e+00 1.74518764e-01 1.22940324e-01
1.36613315e-02 6.16294086e-01 -1.86525643e-01 6.20937169e-01
-2.39424229e-01 2.36350432e-01 -2.95826733e-01 -7.89190114e-01
-1.85899720e-01 2.83122033e-01 -1.61860734e-01 4.45203125e-01
7.68131241e-02 5.11378586e-01 -8.88618708e-01 -2.78759211e-01
2.05753803e-01 2.77322561e-01 -6.34653091e-01 4.68970060e-01
3.85206938e-02 5.85893989e-01 1.38564467e-01 3.70730102e-01
3.52615207e-01 3.07834864e-01 5.46636581e-01 1.08057104e-01
-6.89441189e-02 1.33426964e+00 -8.25887561e-01 1.83261931e+00
-6.98532343e-01 6.70879543e-01 8.88353214e-02 -7.75745451e-01
4.96278137e-01 1.02009535e+00 1.95082337e-01 -6.39763772e-01
-3.62958782e-03 4.97005850e-01 4.67352092e-01 -4.84512538e-01
1.23621531e-01 -1.14178747e-01 -6.17161989e-02 3.30258667e-01
2.04421088e-01 -1.88098773e-01 6.29615113e-02 6.07142188e-02
1.45750415e+00 -3.34823757e-01 1.53558820e-01 -2.93080304e-02
1.34282693e-01 -2.55210757e-01 6.32215977e-01 5.06951451e-01
-1.37905821e-01 9.90465879e-01 3.35700393e-01 9.67788100e-02
-8.38251114e-01 -9.92032349e-01 -7.47337043e-02 7.93374062e-01
-5.51185071e-01 -8.66435111e-01 -1.15547121e+00 -4.42694575e-01
-4.81454700e-01 8.23247790e-01 4.52742577e-02 -1.53806269e-01
-1.13729966e+00 -4.77085233e-01 9.43405926e-01 4.95744556e-01
-1.32371426e-01 -1.02341342e+00 -7.80146243e-03 5.60505152e-01
-8.39971125e-01 -1.57647443e+00 -8.93927872e-01 3.26097012e-01
-7.95278192e-01 -6.45556629e-01 -5.38092196e-01 -9.23386753e-01
2.72536963e-01 9.46708843e-02 8.55916560e-01 1.29771143e-01
4.35423292e-02 6.79568127e-02 -5.03025234e-01 -7.47520328e-02
-9.14956927e-01 2.26106077e-01 3.83780271e-01 -1.88102767e-01
7.45615512e-02 -7.36362755e-01 -2.63876617e-01 2.76814908e-01
-4.76576298e-01 -4.74157790e-03 1.61085188e-01 8.87614191e-01
5.66329062e-01 -2.86045611e-01 5.11969447e-01 -7.91701794e-01
7.35139549e-01 -9.36255231e-02 -3.47168446e-01 9.71984342e-02
-4.92960066e-01 5.61569706e-02 4.52424943e-01 -5.21522701e-01
-8.38636696e-01 5.91776781e-02 -5.17325461e-01 -2.60029048e-01
-9.68988314e-02 5.17333210e-01 -3.18249404e-01 3.75918895e-01
4.31547463e-01 -7.77413547e-02 -1.91939622e-01 -7.17186034e-01
1.22055635e-01 1.12667048e+00 7.40854025e-01 -2.41235733e-01
4.20265853e-01 -1.21524468e-01 -6.78238630e-01 -9.75080073e-01
-4.55357045e-01 -7.52622247e-01 -4.36692595e-01 6.64560646e-02
6.72509909e-01 -8.78179431e-01 -4.77636466e-03 6.18666172e-01
-1.34822965e+00 -3.21468145e-01 6.56141862e-02 7.27335095e-01
-4.92053419e-01 5.70105433e-01 -7.09457576e-01 -7.63241589e-01
-3.33998173e-01 -1.22077763e+00 9.81964529e-01 -5.11482537e-01
-7.20485628e-01 -6.06131494e-01 3.53682339e-01 5.60806096e-01
2.08881393e-01 -2.18772382e-01 8.31345737e-01 -1.04972267e+00
-5.55542782e-02 -2.72367239e-01 5.04426122e-01 6.18333697e-01
4.42698896e-01 -2.73142755e-01 -1.23250151e+00 -4.22922671e-01
7.75762349e-02 -2.11383373e-01 4.42766637e-01 2.17993364e-01
7.36873448e-01 -4.13766444e-01 -2.62186378e-01 2.48729303e-01
6.52692139e-01 3.59975755e-01 6.12708688e-01 -2.86371440e-01
6.56695664e-01 6.22030258e-01 6.19465053e-01 1.83247373e-01
1.00617312e-01 1.17512167e+00 -8.71618986e-02 1.95537843e-02
-4.15542930e-01 -2.15716869e-01 4.93909180e-01 1.76904750e+00
1.33948594e-01 -6.12287045e-01 -1.04576790e+00 8.93811345e-01
-1.63524842e+00 -7.41296649e-01 -1.93457171e-01 2.52890897e+00
1.09449458e+00 2.94096589e-01 1.09048583e-01 6.16491914e-01
1.01287520e+00 5.63922897e-02 -4.11917046e-02 -5.42026937e-01
1.14859462e-01 5.42699218e-01 1.37385294e-01 9.73980069e-01
-8.38835061e-01 1.09227252e+00 6.32362318e+00 8.85875285e-01
-1.07028759e+00 4.59675908e-01 2.24787846e-01 -4.67902660e-01
-2.30471075e-01 -6.04520552e-02 -6.23086452e-01 5.47810674e-01
1.55377746e+00 1.64664596e-01 5.76124310e-01 1.02204651e-01
5.60362875e-01 6.96506277e-02 -1.15435266e+00 9.74117458e-01
-1.08635575e-01 -7.20684469e-01 -5.11375010e-01 -1.29205346e-01
4.06084150e-01 3.60541582e-01 -9.40844566e-02 -5.37151843e-02
-3.35692912e-02 -8.93649161e-01 8.87025476e-01 -6.14408441e-02
9.85434592e-01 -6.31781399e-01 5.89844525e-01 2.23426998e-01
-9.84224796e-01 2.25636616e-01 9.39114243e-02 9.25044790e-02
4.78170544e-01 7.63557911e-01 -1.26361489e+00 4.10405725e-01
3.44813228e-01 2.90813774e-01 -1.82018727e-01 1.02100408e+00
-4.58852053e-01 1.36413121e+00 -5.63041031e-01 2.34166771e-01
2.02893298e-02 2.47332260e-01 8.23713124e-01 1.46726441e+00
3.88104737e-01 -1.16576202e-01 2.90439073e-02 1.45518959e-01
1.68312844e-02 2.32800335e-01 -5.23128688e-01 -1.91479459e-01
8.59537721e-01 8.03432405e-01 -3.34707618e-01 -4.41201776e-02
-2.92925179e-01 1.10060024e+00 5.30617893e-01 1.91721141e-01
-6.91558063e-01 -2.95821995e-01 6.43175483e-01 2.18174588e-02
1.77755341e-01 -3.95463973e-01 1.21412529e-02 -9.48215365e-01
5.08171320e-01 -1.33673728e+00 2.06240043e-01 -3.99181396e-01
-7.91056573e-01 1.02856541e+00 -7.66138062e-02 -1.23465812e+00
-9.29419518e-01 -6.44179508e-02 -4.72914994e-01 9.27648306e-01
-1.00982332e+00 -7.97373891e-01 2.75032759e-01 3.37491810e-01
6.21653974e-01 8.99788290e-02 1.02636898e+00 6.16201282e-01
-5.79666674e-01 8.88155222e-01 -2.63414830e-01 -5.91972359e-02
7.88908124e-01 -1.12079978e+00 1.01936674e+00 1.08806777e+00
7.13955343e-01 4.01520252e-01 9.00707603e-01 -6.54812694e-01
-6.25878274e-01 -1.03640151e+00 1.65344524e+00 -2.75795490e-01
5.63819170e-01 -7.66600668e-01 -1.01895845e+00 6.17563009e-01
1.72293454e-01 -1.77302793e-01 6.26670659e-01 2.44983867e-01
-2.00318143e-01 -1.95096037e-03 -8.38716686e-01 5.47178030e-01
1.35758662e+00 -9.46740568e-01 -6.38755262e-01 5.64667344e-01
9.73999500e-01 -6.30831361e-01 -5.37885785e-01 3.46369058e-01
3.25018227e-01 -6.93207443e-01 4.31284785e-01 -4.27662492e-01
-2.14670047e-01 -5.92093207e-02 -1.79676175e-01 -1.66381383e+00
-7.15307938e-03 -1.17867780e+00 1.71964586e-01 1.47227418e+00
7.06269741e-01 -4.30965632e-01 2.84581989e-01 3.53406727e-01
-6.61433756e-01 -2.53571510e-01 -1.47693980e+00 -9.09027398e-01
-1.52366251e-01 -9.00294185e-01 2.60089099e-01 6.79565430e-01
6.07935786e-01 1.48634464e-01 -3.64186496e-01 4.04717147e-01
3.94848511e-02 -4.98079389e-01 1.05448037e-01 -7.44460881e-01
-6.36189699e-01 1.93409920e-02 -4.54860598e-01 -6.85336471e-01
2.29078069e-01 -8.22291255e-01 4.30638939e-01 -1.13396740e+00
-2.03481913e-01 -3.72367591e-01 -1.48438707e-01 6.41942799e-01
-1.92818046e-01 1.22279637e-01 1.65584564e-01 -3.93949486e-02
-3.46281439e-01 2.95120031e-01 6.23169899e-01 -5.97335063e-02
-3.44077796e-01 2.07669988e-01 -1.05829209e-01 6.68778956e-01
9.68288600e-01 -8.48378718e-01 -4.93514448e-01 -6.53528810e-01
-9.47757140e-02 3.70031893e-01 -1.54339865e-01 -1.11329293e+00
5.38838543e-02 3.21976632e-01 -5.52270055e-01 -3.12050879e-01
3.91309947e-01 -4.45744365e-01 2.26673767e-01 3.93387526e-01
-3.17082524e-01 2.59455919e-01 4.70546126e-01 3.13515097e-01
-3.94926846e-01 -3.75438422e-01 5.12108386e-01 1.75949752e-01
1.22551084e-01 -7.85984397e-02 -7.58667171e-01 3.51245850e-01
4.04988587e-01 2.16491804e-01 -7.91763663e-02 -5.86636841e-01
-9.75745678e-01 -6.95803091e-02 4.93908748e-02 4.83533353e-01
2.24891022e-01 -1.05510604e+00 -6.76556826e-01 3.20197463e-01
9.38554779e-02 -2.75370479e-01 1.08560726e-01 9.23285246e-01
-1.63495436e-01 5.37820339e-01 3.84921193e-01 -4.94643688e-01
-1.74714077e+00 9.44372192e-02 4.53697890e-01 -2.49335438e-01
-5.93266666e-01 8.36433887e-01 -1.54571697e-01 -1.86778545e-01
5.18164217e-01 -7.37793446e-01 1.57776862e-01 -1.55016422e-01
4.07696128e-01 3.17728192e-01 9.43127215e-01 -6.65130377e-01
-5.67159772e-01 1.60452917e-01 -1.68371156e-01 -6.32839322e-01
1.08354914e+00 -2.95126796e-01 1.06381558e-01 3.87421846e-01
1.07150674e+00 4.66359943e-01 -7.46939540e-01 -4.47599217e-03
3.21308047e-01 1.39000919e-03 1.07685260e-01 -8.02194357e-01
-7.27041781e-01 9.32191551e-01 5.22519886e-01 3.00848484e-01
1.04324317e+00 3.32979299e-02 1.01831353e+00 1.35666832e-01
3.28332067e-01 -1.04018497e+00 -3.84374797e-01 5.31107426e-01
7.87626565e-01 -7.94238150e-01 -7.10552156e-01 -6.33975029e-01
-4.92356420e-01 6.62056863e-01 1.85555294e-01 3.01063597e-01
3.04585785e-01 5.28517365e-01 3.95354986e-01 1.49099559e-01
-8.24822843e-01 -2.35697493e-01 3.65502775e-01 5.93424499e-01
6.75994694e-01 3.31134945e-01 -5.35627365e-01 5.35438478e-01
-5.71839690e-01 -4.26591814e-01 3.83133203e-01 7.64951050e-01
-2.41925448e-01 -1.52498043e+00 -1.59535632e-01 2.20694810e-01
-7.27687478e-01 -6.02281809e-01 -3.86142761e-01 3.59212935e-01
-6.71039447e-02 1.53285670e+00 1.83858931e-01 -3.94298762e-01
4.40376848e-01 4.25482452e-01 5.10467887e-01 -8.91165495e-01
-7.04625964e-01 5.35482228e-01 7.71640837e-01 -5.60598254e-01
-1.78916365e-01 -8.71759176e-01 -1.25770926e+00 2.00033367e-01
-6.32110298e-01 2.45272487e-01 4.99324918e-01 1.19684935e+00
2.70281911e-01 5.08572102e-01 5.65013587e-01 -3.82242292e-01
-5.70889592e-01 -1.40396535e+00 -3.72435063e-01 3.29893142e-01
6.71358109e-01 -3.50571215e-01 -5.62610805e-01 3.80267762e-02] | [14.433728218078613, 6.847911834716797] |
f7854d31-1c47-4832-889a-7e565c10ecb6 | autotransfer-automl-with-knowledge-transfer | 2303.07669 | null | https://arxiv.org/abs/2303.07669v1 | https://arxiv.org/pdf/2303.07669v1.pdf | AutoTransfer: AutoML with Knowledge Transfer -- An Application to Graph Neural Networks | AutoML has demonstrated remarkable success in finding an effective neural architecture for a given machine learning task defined by a specific dataset and an evaluation metric. However, most present AutoML techniques consider each task independently from scratch, which requires exploring many architectures, leading to high computational cost. Here we propose AutoTransfer, an AutoML solution that improves search efficiency by transferring the prior architectural design knowledge to the novel task of interest. Our key innovation includes a task-model bank that captures the model performance over a diverse set of GNN architectures and tasks, and a computationally efficient task embedding that can accurately measure the similarity among different tasks. Based on the task-model bank and the task embeddings, we estimate the design priors of desirable models of the novel task, by aggregating a similarity-weighted sum of the top-K design distributions on tasks that are similar to the task of interest. The computed design priors can be used with any AutoML search algorithm. We evaluate AutoTransfer on six datasets in the graph machine learning domain. Experiments demonstrate that (i) our proposed task embedding can be computed efficiently, and that tasks with similar embeddings have similar best-performing architectures; (ii) AutoTransfer significantly improves search efficiency with the transferred design priors, reducing the number of explored architectures by an order of magnitude. Finally, we release GNN-Bank-101, a large-scale dataset of detailed GNN training information of 120,000 task-model combinations to facilitate and inspire future research. | ['Jure Leskovec', 'Jiaju Liu', 'Jiaxuan You', 'Kaidi Cao'] | 2023-03-14 | null | null | null | null | ['automl'] | ['methodology'] | [ 9.31500420e-02 -2.68464983e-01 -4.09187116e-02 -1.38680443e-01
-6.28861785e-01 -7.75549769e-01 5.02295911e-01 -2.59477887e-02
-4.88788337e-01 2.45043203e-01 1.05353639e-01 -2.67929167e-01
-6.72692060e-01 -4.83188719e-01 -6.57023311e-01 -4.84892666e-01
1.73247695e-01 6.90576494e-01 1.62994564e-02 -1.80078715e-01
3.59966904e-01 2.70280659e-01 -1.22867179e+00 -3.24038938e-02
9.35657024e-01 1.39051354e+00 7.94628799e-01 4.04337734e-01
-1.00911096e-01 2.49576449e-01 -6.03241920e-01 -3.93110037e-01
5.36129117e-01 -2.08625287e-01 -8.51068139e-01 -3.76640707e-01
6.30792797e-01 2.96151966e-01 -4.16700989e-01 9.93736804e-01
6.59153342e-01 2.82927990e-01 8.97034407e-01 -1.28029478e+00
-9.50259030e-01 4.92019087e-01 -1.09267287e-01 2.50261545e-01
-3.36230576e-01 2.63936669e-01 1.39335263e+00 -8.88555050e-01
4.58771586e-01 1.05122435e+00 5.93272030e-01 5.50508499e-01
-1.70218194e+00 -7.87669897e-01 2.30279371e-01 2.42899686e-01
-1.34485483e+00 -1.96621031e-01 1.07653213e+00 -4.49794590e-01
1.33315551e+00 -1.20626390e-01 7.18490839e-01 1.20559871e+00
2.35768735e-01 4.26288992e-01 6.42693818e-01 -2.54600346e-01
2.62216926e-01 -1.26387447e-01 1.64887711e-01 6.88967526e-01
4.52567160e-01 -1.25571266e-01 -6.82954431e-01 -1.76200762e-01
5.13105750e-01 -4.54483069e-02 -3.08454186e-01 -7.45412230e-01
-1.11846423e+00 8.78488362e-01 8.10938954e-01 1.97362617e-01
-1.92694739e-01 5.20184219e-01 5.08626938e-01 5.29554427e-01
2.33936071e-01 1.18206608e+00 -5.97623050e-01 6.35650009e-02
-5.49273133e-01 6.95238858e-02 5.99417865e-01 1.19187319e+00
1.04221904e+00 2.76708126e-01 -7.38504231e-02 9.82673287e-01
2.06006989e-01 3.60535532e-01 5.20724714e-01 -7.37173259e-01
6.54669642e-01 9.04964149e-01 -4.54348266e-01 -9.69000220e-01
-5.62500954e-01 -8.71579349e-01 -7.80872166e-01 -1.58502936e-01
2.60459244e-01 -1.08993784e-01 -7.76641250e-01 2.13547111e+00
-8.24828073e-02 -3.03300053e-01 -1.06108487e-01 6.84199512e-01
4.39917505e-01 4.71871793e-01 7.69970715e-02 3.27352226e-01
1.44416094e+00 -1.23546064e+00 -3.27387273e-01 -8.17865670e-01
1.14652181e+00 -6.19760334e-01 1.29009378e+00 4.14722294e-01
-7.48161256e-01 -5.67534924e-01 -1.14355159e+00 -1.99590340e-01
-4.80624735e-01 2.11010769e-01 6.04790390e-01 5.60120642e-01
-1.41689873e+00 5.08965373e-01 -3.76558036e-01 -3.53560179e-01
3.42566788e-01 6.18882418e-01 -1.40592128e-01 -2.32425570e-01
-1.16166592e+00 9.76103783e-01 7.31099427e-01 -1.48601821e-02
-1.18596613e+00 -9.15499806e-01 -7.96063960e-01 3.27180654e-01
5.50630927e-01 -9.08468425e-01 8.67073536e-01 -4.92092043e-01
-1.13668609e+00 5.26317239e-01 1.17206879e-01 -4.14036214e-01
-9.70836654e-02 -3.72451507e-02 -3.31007034e-01 -1.73013270e-01
-6.64862692e-02 7.78676927e-01 9.69093084e-01 -8.49682212e-01
-2.14784309e-01 -2.41299897e-01 2.85403733e-03 2.75146961e-01
-1.05594206e+00 -3.42330664e-01 -7.09443808e-01 -7.35291839e-01
-1.15613542e-01 -1.09427643e+00 -1.70883924e-01 -3.12950946e-02
-3.28179806e-01 -5.73588669e-01 4.76743132e-01 -2.71808147e-01
1.36483872e+00 -2.18449926e+00 5.38073778e-01 1.87594518e-01
5.72847843e-01 2.37886578e-01 -6.91086411e-01 4.34821695e-01
-3.44261043e-02 2.97569484e-01 1.36163887e-02 -4.08049017e-01
6.79986849e-02 6.39862940e-02 -1.98317673e-02 5.75575903e-02
2.04229876e-01 1.11299539e+00 -7.62918055e-01 -3.61803025e-01
-4.49736863e-02 8.98013040e-02 -5.99467397e-01 9.62201804e-02
-3.09375405e-01 -1.43783182e-01 -5.21937788e-01 3.25712800e-01
1.85590029e-01 -7.30099380e-01 3.81673634e-01 -5.41883349e-01
4.64649439e-01 3.40966284e-01 -6.91365659e-01 1.99442542e+00
-8.10598791e-01 8.21708143e-01 -1.63388237e-01 -1.02355301e+00
1.38778543e+00 -1.78776264e-01 2.02773437e-01 -8.46473217e-01
-3.96075770e-02 3.15167040e-01 2.23274484e-01 -7.47131854e-02
2.88649023e-01 2.64332175e-01 -3.16804737e-01 7.12784052e-01
5.30796826e-01 -7.07179978e-02 8.90415981e-02 7.00585842e-02
1.41056836e+00 -3.77826869e-01 2.17006847e-01 -7.52120793e-01
1.15944423e-01 -8.11766014e-02 3.08319986e-01 7.12093055e-01
-9.97034311e-02 2.92266846e-01 5.74481428e-01 -7.15426743e-01
-1.09866798e+00 -9.26245093e-01 9.69553515e-02 1.37141907e+00
-6.01165146e-02 -5.05967796e-01 -5.97162127e-01 -8.40009809e-01
1.06440730e-01 6.65531456e-01 -7.47155249e-01 -6.06667995e-01
-6.53321683e-01 -3.68244529e-01 5.27266145e-01 4.99573678e-01
3.10780287e-01 -1.15358031e+00 -3.43376428e-01 2.72603408e-02
-1.98546033e-02 -9.71057594e-01 -9.50698495e-01 6.31697953e-01
-9.53656554e-01 -1.02421916e+00 -6.99664414e-01 -1.23974025e+00
8.09659123e-01 3.07002753e-01 1.34308314e+00 9.00008082e-02
-3.09971303e-01 1.52657330e-01 -3.70845236e-02 -6.80201426e-02
-9.05444622e-02 6.73388064e-01 1.71049207e-01 -2.23527282e-01
2.67153591e-01 -4.81466681e-01 -5.88146508e-01 6.20528400e-01
-6.92356825e-01 6.23151846e-02 7.93792546e-01 8.91085744e-01
4.14173037e-01 1.16387635e-01 6.57342494e-01 -6.18552029e-01
9.98810947e-01 -3.28256100e-01 -6.74966574e-01 5.37923813e-01
-1.10813391e+00 6.64510489e-01 1.01529419e+00 -6.44333363e-01
-5.73429823e-01 -1.86707050e-01 4.75749373e-01 -7.02435911e-01
3.30615073e-01 6.84206069e-01 -1.10683762e-01 -2.10500568e-01
1.02701735e+00 3.59228373e-01 1.77170500e-01 -6.16667986e-01
3.52282226e-01 2.04178244e-01 2.51502901e-01 -8.36195588e-01
7.19020665e-01 -1.31522670e-01 1.73240900e-01 -6.26836598e-01
-7.72723854e-01 -3.01892936e-01 -3.27566057e-01 5.51965758e-02
5.98462522e-01 -7.90211678e-01 -6.80334687e-01 4.03263539e-01
-1.12842166e+00 -7.42387950e-01 -4.19395752e-02 4.31995779e-01
-4.45935756e-01 1.39258519e-01 -2.94021249e-01 -4.12309766e-01
-5.21098614e-01 -1.41128540e+00 8.56409192e-01 3.89286615e-02
-2.22603843e-01 -1.24652803e+00 -3.01671978e-02 2.26273969e-01
6.04795098e-01 -2.61694044e-01 1.70156479e+00 -8.55623245e-01
-6.28644526e-01 4.75208573e-02 -5.02397597e-01 4.08168644e-01
3.15642953e-02 -5.59436142e-01 -5.84809601e-01 -5.75613081e-01
-4.21102270e-02 -5.61210632e-01 1.04341900e+00 3.97106290e-01
1.37079489e+00 -1.57805145e-01 -3.85172427e-01 7.93173015e-01
1.58867395e+00 1.01948723e-01 3.42680007e-01 3.60076964e-01
8.92176569e-01 5.68961859e-01 2.99905062e-01 1.01693071e-01
1.33444741e-01 9.28398967e-01 5.32707810e-01 2.42734835e-01
-1.81078702e-01 -3.14979285e-01 2.97355741e-01 1.11684716e+00
2.68220097e-01 -3.10693860e-01 -1.24714696e+00 7.33642399e-01
-1.78269124e+00 -2.77440488e-01 3.86900008e-01 2.01431441e+00
7.35838532e-01 3.37797642e-01 -8.93030390e-02 -2.89136827e-01
6.67612016e-01 2.63653487e-01 -9.49761093e-01 -3.82207513e-01
1.53096363e-01 1.53152540e-01 4.96097744e-01 1.18708521e-01
-7.44244695e-01 8.55101585e-01 6.06078339e+00 1.32355857e+00
-8.79712641e-01 3.27036306e-02 2.91515380e-01 -2.53413051e-01
-5.38719833e-01 -8.13797712e-02 -9.76009250e-01 4.12262887e-01
8.81511033e-01 -5.91425002e-01 7.99413919e-01 1.12130249e+00
-2.66593605e-01 5.22780418e-01 -1.51088583e+00 1.12457883e+00
2.36451894e-01 -1.51819134e+00 3.70890558e-01 4.08236712e-01
7.06496358e-01 2.21419677e-01 4.46229666e-01 6.08899772e-01
3.88012499e-01 -1.06662083e+00 6.14438653e-01 3.52954924e-01
8.65695417e-01 -6.53132796e-01 4.39857364e-01 3.13923955e-01
-1.41426945e+00 -4.00013685e-01 -6.00524485e-01 1.29232958e-01
-3.09106082e-01 3.64537209e-01 -9.59084451e-01 3.77362579e-01
5.88251710e-01 6.17738545e-01 -9.02530670e-01 1.00308800e+00
-1.48567528e-01 2.79859275e-01 -2.51688778e-01 -5.36576509e-01
3.73787940e-01 -1.33408487e-01 3.77109915e-01 8.79634142e-01
4.74375248e-01 -5.05691290e-01 4.51562554e-02 1.16080081e+00
-4.43381101e-01 -7.10177002e-03 -8.90749753e-01 -2.82338351e-01
8.14455390e-01 1.21739495e+00 -5.16483188e-01 1.92858651e-01
-1.61070615e-01 6.21759832e-01 8.42681289e-01 5.05485237e-01
-6.61728263e-01 -5.54810107e-01 7.74016082e-01 -1.15271084e-01
2.69971341e-01 -4.75607663e-01 -4.63088214e-01 -6.27064526e-01
7.05004781e-02 -7.91152000e-01 3.26666236e-01 -7.13590801e-01
-1.46448672e+00 8.25618804e-01 -1.09516636e-01 -1.03240609e+00
-2.31424019e-01 -9.61490631e-01 -5.97031951e-01 9.81758893e-01
-1.23542929e+00 -1.03048396e+00 -3.68669569e-01 3.89137179e-01
7.40451038e-01 -7.69230425e-01 7.58796096e-01 1.81854606e-01
-7.16111481e-01 8.77310991e-01 9.88275707e-02 -2.07802374e-02
7.36277223e-01 -1.14037085e+00 8.20292652e-01 5.36086559e-01
3.05862010e-01 7.65615523e-01 2.62399882e-01 -5.82536280e-01
-1.56521928e+00 -1.31363153e+00 6.69531643e-01 -4.96311933e-01
8.26118350e-01 -8.04600596e-01 -6.88416898e-01 4.79625553e-01
-2.07951702e-02 -2.21779868e-01 5.42230189e-01 4.84201133e-01
-5.02345920e-01 -3.11911374e-01 -5.69836140e-01 7.07577705e-01
1.54860806e+00 -6.93857491e-01 -3.28949958e-01 4.31586206e-01
1.17335522e+00 6.01225495e-02 -7.78775394e-01 1.15243033e-01
3.12965840e-01 -4.09534454e-01 1.00785387e+00 -7.06358433e-01
3.04004550e-01 4.41024527e-02 -1.97670937e-01 -1.63861489e+00
-6.40898407e-01 -5.67314804e-01 -1.68933034e-01 9.12595391e-01
8.15983772e-01 -7.57404506e-01 8.52154672e-01 4.37142432e-01
-3.81454468e-01 -1.11837041e+00 -8.51654589e-01 -1.21823454e+00
-1.26502430e-02 -4.40268815e-01 5.89986503e-01 7.43378103e-01
-3.61087471e-01 7.20735729e-01 -3.35972220e-01 -2.32791275e-01
8.23097229e-01 5.19911433e-03 6.25936925e-01 -1.51293826e+00
-4.23930377e-01 -6.62883401e-01 -2.26846173e-01 -1.24474585e+00
4.58416045e-01 -1.32987368e+00 -1.68385983e-01 -1.38845491e+00
1.19534910e-01 -7.46705055e-01 -4.69525367e-01 6.42244101e-01
8.87308791e-02 -6.00304902e-02 1.82634369e-01 4.70966578e-01
-6.60518050e-01 7.49017298e-01 1.21984613e+00 -3.21730196e-01
-2.01176926e-02 -3.34784240e-01 -9.18417692e-01 5.32336414e-01
8.20800483e-01 -6.41284168e-01 -7.95910180e-01 -9.85672891e-01
5.81629395e-01 -4.94257361e-01 1.10800438e-01 -1.13641357e+00
4.24754351e-01 8.65074620e-03 2.53834188e-01 -2.87097514e-01
4.03266966e-01 -8.90587389e-01 1.54412881e-01 5.44727802e-01
-5.14115334e-01 2.14511961e-01 2.41951376e-01 9.32704091e-01
1.20619416e-01 -4.36234951e-01 5.46644509e-01 1.99047714e-01
-8.74607980e-01 4.04378504e-01 1.22959293e-01 4.05691326e-01
9.01133657e-01 -2.23642826e-01 -6.89558983e-01 -1.32830143e-01
-4.42871600e-01 5.32527089e-01 4.39016372e-01 5.81803441e-01
7.15256393e-01 -1.55306172e+00 -4.59399641e-01 3.26576144e-01
4.92515683e-01 -4.41647284e-02 1.37783423e-01 4.89463151e-01
-3.65673810e-01 6.61268353e-01 -3.29526961e-01 -5.55685163e-01
-8.65065634e-01 6.96008265e-01 2.20954999e-01 -7.16698408e-01
-3.06791633e-01 9.53307390e-01 5.95418274e-01 -5.16703606e-01
2.65845507e-01 -1.56539842e-01 -4.69416529e-02 -9.32267867e-03
5.90678006e-02 3.21767718e-01 2.29995519e-01 -1.05994225e-01
-3.37302685e-01 6.33279324e-01 -3.85192245e-01 2.27893874e-01
1.23850179e+00 1.91040471e-01 1.42897397e-01 1.13368921e-01
1.58370137e+00 -7.28050232e-01 -1.15370560e+00 -5.98301351e-01
1.40876532e-01 -2.27854148e-01 1.51115537e-01 -6.82237446e-01
-1.12682784e+00 9.11296368e-01 3.80798548e-01 1.59108832e-01
1.20082414e+00 1.56874448e-01 6.82334960e-01 8.24869871e-01
6.40813291e-01 -1.13608873e+00 7.35679924e-01 7.02798069e-01
1.04925120e+00 -8.38519633e-01 -1.58421934e-01 -1.54753089e-01
-5.27016819e-01 9.58279908e-01 9.08516943e-01 -8.34970102e-02
5.87018847e-01 -6.76916838e-02 -3.33848625e-01 -5.38662672e-01
-1.10506523e+00 1.91964693e-02 6.19502902e-01 7.51695991e-01
1.68199763e-02 -7.90993497e-02 8.27970058e-02 6.89354956e-01
-2.05721974e-01 -4.30952698e-01 7.82855451e-02 4.97855246e-01
-3.76550078e-01 -1.05940557e+00 1.69458389e-01 8.80098403e-01
1.74454898e-01 -3.06082934e-01 -4.81306911e-01 6.04243100e-01
-3.02957922e-01 6.18150055e-01 6.57844618e-02 -7.75005996e-01
3.19717675e-01 2.45268330e-01 2.61448354e-01 -7.75402904e-01
-5.72105289e-01 -2.86284000e-01 1.61637515e-01 -5.38850248e-01
4.62377369e-01 -8.00116956e-02 -8.22593451e-01 -2.11306944e-01
-4.69165325e-01 1.06642973e-02 7.98034549e-01 6.42708778e-01
6.97409451e-01 6.54643536e-01 5.11184633e-01 -8.54052901e-01
-1.04607069e+00 -9.14421320e-01 -6.01529717e-01 2.56420285e-01
-4.86482531e-02 -9.76394296e-01 -3.94583255e-01 -3.93367767e-01] | [8.7780179977417, 3.3779296875] |
cea4521c-1456-482c-9449-aa5275df9475 | decoupled-multi-task-learning-with-cyclical | 2203.14448 | null | https://arxiv.org/abs/2203.14448v1 | https://arxiv.org/pdf/2203.14448v1.pdf | Decoupled Multi-task Learning with Cyclical Self-Regulation for Face Parsing | This paper probes intrinsic factors behind typical failure cases (e.g. spatial inconsistency and boundary confusion) produced by the existing state-of-the-art method in face parsing. To tackle these problems, we propose a novel Decoupled Multi-task Learning with Cyclical Self-Regulation (DML-CSR) for face parsing. Specifically, DML-CSR designs a multi-task model which comprises face parsing, binary edge, and category edge detection. These tasks only share low-level encoder weights without high-level interactions between each other, enabling to decouple auxiliary modules from the whole network at the inference stage. To address spatial inconsistency, we develop a dynamic dual graph convolutional network to capture global contextual information without using any extra pooling operation. To handle boundary confusion in both single and multiple face scenarios, we exploit binary and category edge detection to jointly obtain generic geometric structure and fine-grained semantic clues of human faces. Besides, to prevent noisy labels from degrading model generalization during training, cyclical self-regulation is proposed to self-ensemble several model instances to get a new model and the resulting model then is used to self-distill subsequent models, through alternating iterations. Experiments show that our method achieves the new state-of-the-art performance on the Helen, CelebAMask-HQ, and Lapa datasets. The source code is available at https://github.com/deepinsight/insightface/tree/master/parsing/dml_csr. | ['Stefanos Zafeiriou', 'Ying Li', 'Zheng Zhu', 'Jiankang Deng', 'Qingping Zheng'] | 2022-03-28 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Zheng_Decoupled_Multi-Task_Learning_With_Cyclical_Self-Regulation_for_Face_Parsing_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Zheng_Decoupled_Multi-Task_Learning_With_Cyclical_Self-Regulation_for_Face_Parsing_CVPR_2022_paper.pdf | cvpr-2022-1 | ['face-parsing'] | ['computer-vision'] | [ 2.60408700e-01 3.11763167e-01 1.46450745e-02 -6.79246604e-01
-8.24617982e-01 -5.13399243e-01 2.56910712e-01 -2.81441718e-01
2.49778386e-02 2.87656724e-01 -1.45397484e-01 -1.86984271e-01
8.99450779e-02 -6.24094784e-01 -9.68125582e-01 -5.76155126e-01
1.39742494e-01 2.66801268e-01 1.39610648e-01 -3.67095545e-02
-3.62164155e-02 4.95931238e-01 -1.50062621e+00 6.31625473e-01
8.91395450e-01 1.13940835e+00 2.92502791e-01 5.77190399e-01
-2.56580800e-01 5.98698378e-01 -2.25135699e-01 -6.58233345e-01
1.60683855e-01 -3.92520815e-01 -7.42157400e-01 2.62536079e-01
8.66838574e-01 -3.05607587e-01 -8.40539932e-02 1.01177394e+00
6.65361047e-01 -1.72889784e-01 3.47506762e-01 -1.10498190e+00
-7.98190653e-01 3.17593932e-01 -1.01375782e+00 1.15306996e-01
1.87475890e-01 1.22502178e-01 7.93901443e-01 -1.03192461e+00
4.86800760e-01 1.61547685e+00 6.92993045e-01 9.34459090e-01
-1.32149923e+00 -1.00268877e+00 5.52124798e-01 1.85835641e-02
-1.27940464e+00 -7.49835670e-01 1.03918588e+00 -2.72536278e-01
6.45949423e-01 -7.57561252e-02 2.15165585e-01 1.34545171e+00
-4.23360034e-04 6.52853727e-01 1.05631876e+00 -1.43419266e-01
4.04753909e-03 -8.34886730e-02 7.78817311e-02 1.33714628e+00
1.56948686e-01 -1.29348829e-01 -6.00006878e-01 4.40132730e-02
9.28996027e-01 -1.01515442e-01 -4.34910767e-02 -1.52922735e-01
-3.78175348e-01 6.00796700e-01 6.47144139e-01 2.57857561e-01
-5.09647019e-02 1.23513758e-01 1.69926390e-01 9.72885564e-02
6.82135105e-01 -1.28614515e-01 -6.23590887e-01 4.28576231e-01
-7.57207394e-01 5.81070744e-02 4.28405166e-01 9.63460028e-01
9.17115688e-01 -1.08836323e-01 -1.96598977e-01 1.08442616e+00
5.33367574e-01 2.46792912e-01 2.68842787e-01 -9.05033529e-01
5.94881475e-01 7.33636081e-01 -4.91018206e-01 -1.04969156e+00
-6.66979134e-01 -5.84209859e-01 -7.73809910e-01 5.98557256e-02
4.94703680e-01 -2.17981815e-01 -1.28930378e+00 2.12931252e+00
5.77232838e-01 3.80706340e-01 -1.67006671e-01 8.34693730e-01
1.15627909e+00 3.13328236e-01 3.22601825e-01 9.28963646e-02
1.68308187e+00 -1.11721575e+00 -3.44963372e-01 -7.10565269e-01
6.42540872e-01 -5.59557378e-01 8.80438626e-01 1.26613542e-01
-1.03671849e+00 -9.05568063e-01 -7.92060256e-01 -3.23889852e-01
-2.54085600e-01 4.26266074e-01 7.53507137e-01 5.52017808e-01
-1.23094153e+00 5.31680882e-01 -8.16035867e-01 -7.28473067e-02
1.05040467e+00 5.08758962e-01 -5.46059787e-01 -3.56959581e-01
-8.79178882e-01 3.84360671e-01 -3.32352072e-02 5.05817592e-01
-7.88547099e-01 -6.61617875e-01 -1.02123737e+00 3.48274000e-02
6.03604257e-01 -8.18464875e-01 1.14135265e+00 -1.09137249e+00
-1.36226976e+00 1.03299916e+00 -4.64330077e-01 2.48839602e-01
3.52574378e-01 -9.52351019e-02 -3.12359422e-01 1.17321670e-01
2.65733600e-01 9.27052259e-01 1.04585600e+00 -1.53879678e+00
-4.47499841e-01 -8.11546087e-01 -4.79923375e-02 1.56229371e-02
-7.58203194e-02 1.77867729e-02 -8.94716680e-01 -6.34312272e-01
3.46551239e-01 -8.35297525e-01 -1.81964729e-02 -4.04265374e-02
-5.42820573e-01 -3.81299108e-01 7.68207848e-01 -7.45281637e-01
9.52456653e-01 -2.15088010e+00 2.84165472e-01 -7.20618591e-02
2.98585802e-01 2.08321348e-01 -6.05743229e-01 -2.63783842e-01
-2.20680237e-01 2.96266139e-01 -3.74881595e-01 -1.00453269e+00
-3.72700453e-01 1.14160389e-01 1.12587966e-01 3.08780104e-01
7.13609993e-01 9.88158047e-01 -6.70411468e-01 -5.29933631e-01
-2.68477350e-02 7.11686373e-01 -7.90602446e-01 1.58964902e-01
-2.54851133e-01 7.73749411e-01 -5.13120651e-01 1.06869459e+00
1.08755720e+00 -4.92578596e-01 2.48448774e-01 -3.04069817e-01
2.13119656e-01 5.67521304e-02 -1.09576774e+00 1.82594907e+00
-6.45569026e-01 1.31878003e-01 5.70240021e-01 -9.30355549e-01
7.43979216e-01 2.33237650e-02 1.34780124e-01 -8.63453090e-01
8.47136006e-02 2.86662858e-02 -3.83031517e-02 -4.41973835e-01
-2.80950703e-02 9.90199521e-02 1.63689665e-02 1.02432758e-01
3.19104910e-01 3.52168173e-01 -8.29127990e-03 7.81873986e-03
9.51743305e-01 4.34204191e-01 -4.38111834e-02 -1.80216789e-01
5.98958552e-01 -4.60734129e-01 1.04769599e+00 4.25134003e-01
-2.42873743e-01 7.65281796e-01 7.69723892e-01 -4.24468607e-01
-5.24975598e-01 -9.81627226e-01 -7.21065402e-02 1.47750890e+00
1.02780648e-01 -3.21841627e-01 -9.51471150e-01 -1.00769424e+00
1.62805513e-01 2.00785577e-01 -8.69076967e-01 -4.15790677e-02
-7.99429417e-01 -9.40299809e-01 4.75159407e-01 5.78684926e-01
7.21913815e-01 -1.07509899e+00 -1.03071876e-01 1.39227688e-01
-9.79179889e-03 -1.33519816e+00 -5.37894905e-01 -2.20568478e-03
-7.21718907e-01 -1.18487203e+00 -3.82376313e-01 -9.23903286e-01
1.04055822e+00 2.63295263e-01 1.11103582e+00 3.89617562e-01
-4.64476943e-01 3.89051810e-02 -1.61570892e-01 9.56084505e-02
-1.36591151e-01 7.75341392e-02 -3.51321161e-01 2.25796595e-01
4.41422500e-02 -5.74322701e-01 -6.84560061e-01 3.85100007e-01
-7.65500426e-01 2.06062123e-01 7.40605175e-01 8.88182819e-01
6.35006130e-01 -2.37320855e-01 7.47598469e-01 -1.10424662e+00
1.24560453e-01 -5.18181324e-01 -5.83837748e-01 4.29781049e-01
-2.27527916e-01 5.40733635e-02 5.89867413e-01 -2.49588922e-01
-1.39532232e+00 2.79032409e-01 -2.48775080e-01 -4.62414980e-01
-3.53877544e-01 5.92519976e-02 -7.38088787e-01 -1.89333752e-01
2.09229454e-01 -2.17018183e-02 -1.77441970e-01 -6.40823066e-01
3.67192924e-01 5.51449716e-01 5.11434376e-01 -6.40979350e-01
5.56444883e-01 4.39473718e-01 -4.72710021e-02 -7.57407010e-01
-1.09522712e+00 -9.11515281e-02 -6.97492480e-01 -1.30682454e-01
1.03086972e+00 -1.10935390e+00 -6.97326899e-01 7.76804745e-01
-1.36794019e+00 -6.03715122e-01 3.31942946e-01 -2.08108738e-01
-2.10724831e-01 2.16999248e-01 -7.79049337e-01 -5.61548650e-01
-2.50324667e-01 -1.29352427e+00 1.54794371e+00 5.96653998e-01
2.16161624e-01 -9.14559186e-01 -4.24579471e-01 9.09050465e-01
1.16756536e-01 1.59901068e-01 7.90661931e-01 -4.18876082e-01
-6.28457427e-01 3.44861895e-01 -5.00208199e-01 2.30464742e-01
2.76424646e-01 -1.21740378e-01 -1.35226846e+00 -3.71474981e-01
-8.51888731e-02 -4.27685082e-01 1.11855793e+00 2.86200613e-01
1.74371254e+00 -1.63223177e-01 -4.61384833e-01 1.01082456e+00
1.20827007e+00 7.54281087e-03 4.64799523e-01 -8.80230069e-02
1.21026003e+00 7.62339056e-01 2.20145375e-01 1.15805127e-01
6.53159142e-01 5.81757128e-01 3.60598117e-01 -2.60269910e-01
-4.56729680e-01 -2.97618419e-01 1.87747359e-01 3.67082626e-01
1.91337541e-01 -1.55292392e-01 -8.08979809e-01 3.48066926e-01
-1.66404724e+00 -5.27134895e-01 7.07418174e-02 1.76049101e+00
6.42607927e-01 1.29218251e-01 -1.08825043e-01 -4.49109793e-01
1.00583529e+00 2.72228837e-01 -7.87706017e-01 -3.00937772e-01
-7.12938234e-02 2.17846677e-01 3.51615727e-01 4.90232825e-01
-1.24123549e+00 1.43014991e+00 4.78143167e+00 7.88132310e-01
-1.16077518e+00 3.72809798e-01 1.28639793e+00 4.35810387e-02
-2.31197730e-01 -2.18047723e-01 -1.13035166e+00 5.17969429e-01
5.72961211e-01 5.46719790e-01 4.53240156e-01 8.04313958e-01
-5.12895882e-02 4.52582305e-03 -9.49120164e-01 9.04116392e-01
2.26819798e-01 -1.07788956e+00 -9.78292152e-03 2.83977855e-03
6.22013032e-01 -9.28140953e-02 1.40911326e-01 2.94863462e-01
1.33943394e-01 -1.06324923e+00 5.16599357e-01 2.10684225e-01
9.68175173e-01 -5.74085057e-01 4.38789368e-01 -3.72233465e-02
-1.52702868e+00 -2.13453442e-01 -1.88186198e-01 2.68187970e-01
1.21017024e-02 5.79930305e-01 -3.67449731e-01 6.24247134e-01
8.02251041e-01 5.23718834e-01 -9.29080784e-01 4.27595139e-01
-3.21839184e-01 3.87422860e-01 -3.01408261e-01 6.03459120e-01
3.43886688e-02 -2.78970134e-02 3.41190070e-01 1.07176518e+00
2.21487194e-01 5.07328175e-02 1.75273702e-01 9.79700506e-01
-4.73312616e-01 -2.04629675e-01 -3.36918473e-01 2.28991106e-01
4.89016891e-01 1.60180819e+00 -9.64195967e-01 -1.35547012e-01
-6.56879723e-01 1.14254165e+00 8.57934475e-01 2.66740084e-01
-8.38888288e-01 5.05491868e-02 8.11573982e-01 8.76697898e-02
5.44055164e-01 -9.07583814e-03 -4.41703409e-01 -1.09231794e+00
2.69424826e-01 -7.21274078e-01 5.12274563e-01 -4.84177947e-01
-1.29955137e+00 7.33785808e-01 -2.47918382e-01 -6.36870682e-01
-4.49666381e-02 -7.94682860e-01 -8.50389779e-01 7.57561922e-01
-1.73207247e+00 -1.63497400e+00 -4.06147540e-01 4.81356382e-01
6.43426418e-01 -1.20193221e-01 5.93700111e-01 3.77683699e-01
-1.14175630e+00 9.56444502e-01 -5.25090694e-01 3.73926461e-01
7.88270712e-01 -1.14796710e+00 6.43074512e-01 8.73037815e-01
-1.11011326e-01 5.54237843e-01 8.88985991e-02 -8.55530083e-01
-1.33084869e+00 -1.35385072e+00 5.72195768e-01 -2.64037043e-01
3.10058951e-01 -7.05847204e-01 -1.12115455e+00 6.63349271e-01
-2.74787366e-01 5.64846694e-01 5.33689976e-01 2.58216709e-01
-6.73814535e-01 -3.18943024e-01 -1.23257673e+00 5.25157154e-01
1.53993702e+00 -5.28562903e-01 -2.19227418e-01 1.68271229e-01
7.01806247e-01 -4.26258445e-01 -6.13949776e-01 6.46959186e-01
3.91951114e-01 -1.13011384e+00 9.07067358e-01 -3.76647085e-01
4.92163956e-01 -1.02182366e-01 5.01724742e-02 -1.10787928e+00
-4.28098619e-01 -5.65140784e-01 -1.10197857e-01 1.55973494e+00
4.76603419e-01 -6.84645295e-01 9.83848393e-01 4.79575366e-01
-3.25123399e-01 -1.03053224e+00 -1.10550225e+00 -4.74380106e-01
1.36595681e-01 -2.18461081e-01 5.73211193e-01 7.96192348e-01
-5.11401117e-01 3.94476533e-01 -8.77646282e-02 4.96079803e-01
7.86578000e-01 1.78577080e-01 5.76989233e-01 -1.21641028e+00
-2.67457724e-01 -4.45719212e-01 -3.85688506e-02 -9.34756219e-01
5.36125958e-01 -9.97519016e-01 -8.77293274e-02 -1.39084804e+00
1.33339435e-01 -5.09902477e-01 -1.37415156e-01 8.38721454e-01
-4.74208117e-01 3.90719652e-01 1.23540439e-01 -1.46845251e-01
-7.18042552e-01 4.56477106e-01 1.46082067e+00 1.05457798e-01
-1.73987314e-01 -1.79850653e-01 -1.19046259e+00 7.67226458e-01
7.41153240e-01 -3.41754347e-01 -2.99718767e-01 -8.29565823e-01
-7.27505460e-02 3.05316634e-02 5.62344551e-01 -8.52324426e-01
2.36889794e-01 1.48869216e-01 8.03499699e-01 -1.46821827e-01
3.26516002e-01 -5.36163867e-01 -1.16977267e-01 2.21397996e-01
-7.42568225e-02 1.80970058e-02 4.76834446e-01 5.04540443e-01
1.10076023e-02 6.84217960e-02 1.00784910e+00 -2.07169279e-01
-8.17263782e-01 6.26249015e-01 2.26001114e-01 5.45351692e-02
8.38928998e-01 -3.85795161e-02 -5.57159603e-01 1.15448147e-01
-8.72374296e-01 5.55829227e-01 3.44763547e-01 6.77677333e-01
5.33970058e-01 -1.08288777e+00 -7.11808562e-01 6.23673975e-01
-9.44785366e-04 4.55741435e-01 8.48183692e-01 5.33959150e-01
-1.69741720e-01 1.15875624e-01 -1.33060515e-01 -6.15523398e-01
-1.33558357e+00 2.69795358e-01 5.01589537e-01 -1.88029408e-01
-4.49737847e-01 1.31223452e+00 5.52182615e-01 -5.98725855e-01
1.04715072e-01 -1.48499966e-01 -6.64393380e-02 1.47598594e-01
3.76958400e-01 7.31283054e-02 1.66288555e-01 -7.61038005e-01
-6.03053272e-01 8.74622643e-01 -3.52722794e-01 3.60317320e-01
1.31960714e+00 -1.99439272e-01 -1.87530473e-01 -1.91479474e-02
1.33876765e+00 -1.51170924e-01 -1.59135401e+00 -1.62915081e-01
-1.15920231e-01 -2.96717912e-01 2.59234756e-02 -8.80865216e-01
-1.46988451e+00 8.12951863e-01 6.54995680e-01 -2.84338057e-01
1.27249682e+00 3.62802595e-01 7.16951847e-01 7.06952289e-02
1.65185973e-01 -9.67644930e-01 1.87510833e-01 3.40858430e-01
9.54542458e-01 -1.56330574e+00 -1.94718674e-01 -9.10372555e-01
-4.98264223e-01 8.76326203e-01 1.20390749e+00 2.15886831e-02
6.68218493e-01 2.99211770e-01 -4.58490150e-03 -2.93554485e-01
-7.50211298e-01 -3.02401751e-01 2.91751713e-01 4.94587451e-01
4.90197122e-01 1.02878129e-02 1.85753778e-01 9.50236082e-01
5.51473275e-02 -1.45344257e-01 -1.21194599e-02 7.66393960e-01
-1.89911202e-01 -1.18722403e+00 -2.17042625e-01 3.62383425e-01
-5.24486542e-01 -3.55172865e-02 -3.79920363e-01 6.41914189e-01
4.78148520e-01 9.92086112e-01 2.88142502e-01 -3.67779493e-01
3.06758195e-01 2.08465368e-01 5.75419664e-01 -7.50351429e-01
-4.63183165e-01 1.90668002e-01 -8.02656729e-03 -8.33541393e-01
-1.34401396e-01 -5.34677386e-01 -1.30391073e+00 -4.89872992e-02
-7.39915967e-02 -2.05710262e-01 5.47500908e-01 9.01170254e-01
7.71046519e-01 7.98731565e-01 6.47412896e-01 -1.05051756e+00
-1.70265406e-01 -1.01213598e+00 -2.00474948e-01 3.34710121e-01
3.76969725e-01 -8.75871778e-01 -3.57858211e-01 -1.61291748e-01] | [13.435489654541016, 0.6584419012069702] |
95539aee-4b61-4470-a238-a784b46aa0b3 | don-t-take-it-literally-an-edit-invariant-1 | null | null | https://openreview.net/forum?id=kQWGURqrAW_ | https://openreview.net/pdf?id=kQWGURqrAW_ | Don't Take It Literally: An Edit-Invariant Sequence Loss for Text Generation | Neural text generation models are typically trained by maximizing log-likelihood with the sequence cross entropy (CE) loss, which encourages an exact token-by-token match between a target sequence with a generated sequence. Such training objective is sub-optimal when the target sequence is not perfect, e.g., when the target sequence is corrupted with noises, or when only weak sequence supervision is available. To address the challenge, we propose a novel Edit-Invariant Sequence Loss (EISL), which computes the matching loss of a target $n$-gram with all $n$-grams in the generated sequence. EISL is designed to be robust to various noises and edits in the target sequences. Moreover, the EISL computation is essentially an approximate convolution operation with target $n$-grams as kernels, which is easy to implement and efficient to compute with existing libraries. To demonstrate the effectiveness of EISL, we conduct experiments on a wide range of tasks, including machine translation with noisy target sequences, unsupervised text style transfer with only weak training signals, and non-autoregressive generation with non-predefined generation order. Experimental results show our method significantly outperforms the common CE loss and other strong baselines on all the tasks. EISL has a simple API that can be used as a drop-in replacement of the CE loss: https://anonymous.4open.science/r/EISLLoss. | ['Anonymous'] | 2022-01-16 | null | null | null | acl-arr-january-2022-1 | ['text-style-transfoer'] | ['natural-language-processing'] | [ 5.63180685e-01 -1.06831044e-01 -4.00110707e-03 -4.13256198e-01
-1.11002493e+00 -6.16670012e-01 6.66834474e-01 -5.86972199e-02
-5.69137156e-01 9.91656244e-01 1.48499176e-01 -4.39760059e-01
5.77042460e-01 -8.54581654e-01 -1.14064932e+00 -6.16720676e-01
1.48778677e-01 2.95936555e-01 -1.52696148e-01 -3.33763987e-01
1.28704250e-01 -6.81409016e-02 -1.14882612e+00 2.72262037e-01
1.07453191e+00 9.14063573e-01 5.21320581e-01 7.28648067e-01
-1.15097649e-01 5.09961128e-01 -6.32264912e-01 -8.03353846e-01
4.56785291e-01 -7.67108619e-01 -6.31118894e-01 -3.81269276e-01
1.83692217e-01 -3.07785541e-01 -2.73437828e-01 1.32045877e+00
7.98825741e-01 1.56838804e-01 6.31213188e-01 -1.06372464e+00
-7.76264131e-01 8.12146246e-01 -4.18817878e-01 -9.16496739e-02
3.73631150e-01 5.21672308e-01 9.47966695e-01 -1.18678331e+00
4.78607655e-01 1.21669352e+00 7.69298553e-01 5.69869578e-01
-1.17518330e+00 -8.98105979e-01 -7.01812580e-02 -4.58557680e-02
-1.23384786e+00 -3.77226025e-01 4.22000051e-01 -3.65656823e-01
1.06012797e+00 2.11852700e-01 1.05409771e-01 1.32984304e+00
1.92956328e-01 7.58315384e-01 8.04220855e-01 -5.32059848e-01
1.23015836e-01 6.54309392e-02 -1.89796016e-01 6.88908398e-01
-1.91054512e-02 2.00828344e-01 -5.50318599e-01 -2.64153391e-01
5.05162060e-01 -2.85201699e-01 -3.12863350e-01 1.47960871e-01
-1.38702142e+00 7.99623787e-01 1.03785954e-01 1.29204512e-01
-1.89860508e-01 3.36151093e-01 6.69351876e-01 5.66550314e-01
6.17993951e-01 1.94401965e-01 -4.21723664e-01 -4.55937803e-01
-7.43997931e-01 3.80948961e-01 7.66385376e-01 1.30530345e+00
8.86987448e-01 2.24207252e-01 -3.24079782e-01 9.83118773e-01
1.60262808e-02 5.08495212e-01 8.75455439e-01 -6.17931187e-01
7.72221148e-01 1.66992605e-01 2.23258182e-01 -5.92333794e-01
2.60908753e-01 -2.84247756e-01 -1.09884179e+00 -1.21877313e-01
3.57602239e-01 -4.58574831e-01 -7.11310446e-01 2.10743880e+00
2.04161093e-01 2.40076885e-01 -2.54349802e-02 6.59393907e-01
4.85124588e-01 8.84649456e-01 -1.42787844e-01 -2.09838882e-01
1.11096597e+00 -1.03497660e+00 -5.83900690e-01 -3.71118605e-01
7.85588086e-01 -1.00616670e+00 1.49964952e+00 1.99391861e-02
-1.19625390e+00 -4.33675557e-01 -9.62533474e-01 -8.56572390e-02
-3.45507562e-01 1.33355096e-01 2.60980099e-01 5.34145892e-01
-9.72189188e-01 8.43438447e-01 -7.05017328e-01 -6.21336140e-02
1.20611154e-01 1.08102180e-01 -2.56952733e-01 -1.28248811e-01
-1.60002887e+00 8.02498043e-01 6.68484271e-01 2.78123608e-03
-8.21985543e-01 -9.82240260e-01 -9.85309303e-01 7.82230645e-02
1.19948089e-01 -7.76416540e-01 1.49453127e+00 -1.22472084e+00
-1.72540963e+00 6.12005889e-01 -3.07139069e-01 -5.56436479e-01
8.83576214e-01 -3.16249043e-01 -2.28210762e-01 -2.22098142e-01
1.27003044e-01 2.85848618e-01 8.31664503e-01 -8.02554905e-01
-3.82909000e-01 8.15829709e-02 -3.20414364e-01 2.54112989e-01
-3.11508805e-01 1.57710627e-01 -3.04031491e-01 -1.12417114e+00
-6.50049508e-01 -9.34419990e-01 -2.52022505e-01 -2.30551288e-01
-4.13532645e-01 -2.68068582e-01 3.85684878e-01 -9.16190624e-01
1.20434809e+00 -2.03745627e+00 -1.59209654e-01 1.05970010e-01
-3.70398939e-01 4.87652540e-01 -4.15953368e-01 7.88443685e-01
-1.09948389e-01 1.03613049e-01 -7.02541351e-01 -3.66264224e-01
2.93963403e-01 -1.95817292e-01 -5.56698799e-01 9.76662561e-02
2.50769168e-01 9.66358304e-01 -1.08173740e+00 -2.62328953e-01
-2.33318016e-01 3.97456020e-01 -4.09679383e-01 3.86257738e-01
-4.62831736e-01 1.35749996e-01 -1.51500300e-01 2.62026608e-01
4.85200047e-01 -3.43400359e-01 3.62355486e-02 2.51904666e-01
3.42645682e-02 7.41893768e-01 -9.76565599e-01 1.76664746e+00
-5.08002341e-01 4.26406235e-01 -8.61128196e-02 -9.41003978e-01
9.34891999e-01 4.07984167e-01 -7.84632564e-02 -3.13288659e-01
-8.00567865e-02 5.21608949e-01 -9.17164609e-02 -1.52254224e-01
5.33961713e-01 -2.69021869e-01 -1.31388947e-01 7.51017094e-01
4.45243344e-02 -7.14035705e-02 3.93986136e-01 2.41137818e-01
1.18398893e+00 2.13527277e-01 2.92096645e-01 -9.87626314e-02
2.97545910e-01 -3.22139114e-01 6.45819485e-01 9.21058178e-01
1.54756486e-01 6.47279322e-01 3.31656098e-01 2.42480636e-02
-1.26864195e+00 -1.07251871e+00 2.84511656e-01 1.10608125e+00
-1.69528514e-01 -2.97423095e-01 -1.00417840e+00 -7.32303143e-01
-1.89365119e-01 9.31302786e-01 -2.62882859e-01 -3.25393766e-01
-7.39058912e-01 -8.31543744e-01 8.32419097e-01 5.03786564e-01
5.57442963e-01 -1.24102700e+00 -7.38638267e-02 3.83406192e-01
-5.67035794e-01 -8.07347596e-01 -1.44228876e+00 -9.56578180e-02
-6.15376592e-01 -6.25134528e-01 -8.48222613e-01 -1.02435064e+00
6.19061589e-01 5.16033405e-03 1.07142282e+00 -9.01129693e-02
-1.31616443e-01 -6.37649074e-02 -3.35338086e-01 -3.76319140e-01
-6.52116716e-01 2.59374321e-01 -9.39458143e-03 -1.20164466e-03
2.97229290e-01 -5.73872983e-01 -6.53896391e-01 1.68130368e-01
-1.09295785e+00 8.56236592e-02 5.71114302e-01 1.12664437e+00
4.50720847e-01 -2.45991692e-01 7.92137921e-01 -7.69998491e-01
9.42252100e-01 -4.79224890e-01 -5.00177741e-01 2.86068797e-01
-4.98231679e-01 1.99501470e-01 1.01856995e+00 -5.43797314e-01
-1.06338227e+00 -1.93421870e-01 -3.56544822e-01 -3.56168091e-01
9.32405815e-02 7.59567320e-01 -2.82151490e-01 4.26220238e-01
6.43122435e-01 7.74957597e-01 6.34863675e-02 -4.60986167e-01
3.32496941e-01 8.29681635e-01 4.75905895e-01 -5.70698440e-01
8.44366133e-01 1.23168686e-02 -3.71748656e-01 -5.83782911e-01
-6.53063059e-01 -2.50993520e-01 -1.02391385e-01 1.63643375e-01
3.89085293e-01 -9.44615722e-01 -3.68761778e-01 8.06898236e-01
-1.43504488e+00 -6.99262142e-01 -1.57436326e-01 5.63332200e-01
-6.51329875e-01 6.70265377e-01 -8.91635358e-01 -6.08833194e-01
-9.91376758e-01 -9.74036574e-01 7.82011151e-01 -1.05427444e-01
-3.57563913e-01 -8.02244484e-01 1.23336978e-01 1.04062548e-02
5.02466559e-01 7.85604715e-02 1.00168896e+00 -6.65628910e-01
-4.57658947e-01 -2.93111324e-01 8.84186383e-03 8.31746697e-01
1.76786244e-01 -8.80894288e-02 -5.97048342e-01 -4.83428270e-01
-3.07457130e-02 -4.42052305e-01 8.30175042e-01 9.38521177e-02
1.17659640e+00 -7.26770043e-01 -3.80721241e-02 5.69326460e-01
1.18780887e+00 1.67767093e-01 7.27648675e-01 1.17245831e-01
4.61595178e-01 4.92341310e-01 6.17676020e-01 5.33780694e-01
2.16536015e-01 5.67217231e-01 4.32378165e-02 7.60193169e-02
1.26547381e-01 -5.81108034e-01 8.88808370e-01 1.26506817e+00
1.91933170e-01 -4.05101895e-01 -6.72341645e-01 5.19146264e-01
-1.85942447e+00 -1.14751112e+00 1.17660604e-01 2.41802597e+00
1.46553123e+00 5.83996326e-02 -4.41618860e-02 -2.57869065e-01
9.81218040e-01 9.54209715e-02 -6.52711451e-01 -4.23509240e-01
-1.86423212e-01 5.51620424e-01 5.07547736e-01 6.81679547e-01
-9.78558958e-01 9.76992071e-01 5.69994354e+00 1.35057592e+00
-1.12728763e+00 1.23428799e-01 6.07189953e-01 -2.04084709e-01
-4.33515817e-01 -6.98269531e-02 -7.58695781e-01 1.00671768e+00
1.04563355e+00 -5.52324653e-01 5.82944989e-01 7.71867335e-01
2.73457199e-01 3.80152464e-01 -1.17477000e+00 8.05920362e-01
-7.42788017e-02 -1.23880601e+00 1.07114352e-01 -2.03486681e-01
6.67183816e-01 1.35195449e-01 1.44730434e-01 3.13294202e-01
7.37109303e-01 -1.04047525e+00 7.65559018e-01 2.49539405e-01
1.08822680e+00 -8.14230800e-01 5.82187593e-01 5.03912628e-01
-1.03665638e+00 3.13143373e-01 -3.56739700e-01 1.13872727e-02
3.79620552e-01 7.54593670e-01 -8.85892630e-01 4.65559632e-01
3.82539868e-01 4.62127268e-01 1.33780930e-02 7.19207168e-01
-2.77066410e-01 7.01069534e-01 -2.81679839e-01 -2.68111676e-01
2.79980004e-01 -5.31693161e-01 5.53778827e-01 1.67256904e+00
7.88122296e-01 -1.19939841e-01 -8.66051763e-02 8.87885988e-01
-5.05050778e-01 2.88461447e-01 -5.64445138e-01 -2.06169784e-01
5.60749590e-01 1.03158510e+00 -1.49926022e-01 -4.18458939e-01
-2.99336106e-01 1.44538248e+00 4.24814641e-01 4.33349550e-01
-8.90978098e-01 -1.08206725e+00 8.99636030e-01 2.23891297e-03
4.23933893e-01 -1.04883261e-01 -6.66823909e-02 -1.19080961e+00
4.30135459e-01 -1.16449940e+00 1.34890810e-01 -6.64721668e-01
-1.51861644e+00 6.60145283e-01 -3.10845017e-01 -1.15691042e+00
-6.16707087e-01 -3.59493703e-01 -7.14350581e-01 1.34141195e+00
-1.34943330e+00 -7.94919431e-01 7.70914182e-02 4.89783645e-01
5.87000549e-01 -1.00162134e-01 7.00788260e-01 4.17120844e-01
-4.65394884e-01 1.10636640e+00 4.70037401e-01 3.21725309e-01
9.85910714e-01 -1.20000470e+00 9.57999706e-01 1.07419336e+00
-2.39163786e-01 8.48950684e-01 6.74957514e-01 -7.17512071e-01
-1.15886903e+00 -1.55142200e+00 1.32568574e+00 -2.34178558e-01
5.53272069e-01 -7.02028215e-01 -9.00681734e-01 8.21456790e-01
3.84879887e-01 -2.82066613e-01 7.29846954e-01 -4.42200333e-01
-4.60449219e-01 6.40539452e-02 -9.06747878e-01 8.40434372e-01
1.15304852e+00 -5.44969797e-01 -3.25981200e-01 5.27795255e-01
1.07276511e+00 -3.76460254e-01 -7.28544593e-01 3.12770963e-01
4.74252075e-01 -6.08996272e-01 7.55739331e-01 -6.54027760e-01
7.51220524e-01 -2.05327496e-01 -1.52712062e-01 -1.70540118e+00
-8.19215551e-02 -1.13913989e+00 3.15162316e-02 1.32241356e+00
7.07234442e-01 -8.03427637e-01 5.73352814e-01 4.01193142e-01
-3.70703518e-01 -5.97515762e-01 -7.93632686e-01 -1.03400123e+00
3.12948674e-01 -2.83395022e-01 8.60789418e-01 8.98499727e-01
1.66268006e-01 2.83030152e-01 -6.81093693e-01 -2.26532042e-01
4.25545990e-01 -9.50589310e-03 7.70111918e-01 -4.21650589e-01
-6.27129853e-01 -4.75581884e-01 9.41891670e-02 -1.38811171e+00
3.14541936e-01 -1.16727340e+00 4.64397401e-01 -1.24097574e+00
3.29529233e-02 -2.58684337e-01 -1.71131402e-01 3.41135800e-01
-6.06843770e-01 8.15935656e-02 1.31416380e-01 4.10527065e-02
-2.41106212e-01 9.44747388e-01 9.25828636e-01 -8.12147409e-02
-6.36907592e-02 -7.38795660e-03 -6.16067946e-01 5.30589461e-01
1.02258611e+00 -4.68764931e-01 -2.88828164e-01 -4.95695859e-01
1.28993347e-01 -6.55693039e-02 -9.19063482e-03 -5.20556033e-01
-5.75976595e-02 -9.09655020e-02 -1.00942232e-01 -3.81673098e-01
1.30765229e-01 -2.73225814e-01 2.59884298e-01 4.58413929e-01
-6.59585893e-01 1.19411603e-01 4.97811334e-03 4.88139868e-01
-1.50091588e-01 -4.01833594e-01 7.22687721e-01 -2.75127977e-01
-1.75489113e-01 4.10962373e-01 -2.45049879e-01 3.75763953e-01
7.74421930e-01 1.70304235e-02 -3.22113156e-01 -5.50059259e-01
-9.15138647e-02 1.39313996e-01 4.27423656e-01 4.19112951e-01
5.29143035e-01 -1.51096821e+00 -1.10791290e+00 1.30754009e-01
1.33307129e-01 3.49547074e-04 6.61921427e-02 5.10411441e-01
-5.40112078e-01 2.86045969e-01 2.13267490e-01 -4.03933227e-02
-1.17045343e+00 4.47254509e-01 1.73900574e-01 -5.32425165e-01
-2.95925975e-01 8.78579021e-01 3.42749476e-01 -7.79071987e-01
2.11411878e-01 -1.08091086e-01 5.59387207e-01 -3.69255692e-01
8.18690300e-01 3.43955606e-01 1.73620909e-01 -4.71060276e-01
4.59731033e-04 5.81182204e-02 -2.70059764e-01 -3.18067580e-01
1.01838756e+00 -1.23283751e-02 -1.49482936e-01 4.21216846e-01
1.38535142e+00 -1.56760186e-01 -1.25651908e+00 -5.58124542e-01
-1.65577739e-01 -3.83194178e-01 -6.01958394e-01 -6.92104757e-01
-7.30019331e-01 8.73477340e-01 1.91632390e-01 -1.24491602e-01
9.15040553e-01 -3.94397765e-01 1.25690591e+00 2.98067510e-01
2.09393531e-01 -1.08003271e+00 1.39574304e-01 7.81563520e-01
1.07555580e+00 -9.18040454e-01 -5.42884052e-01 -2.26457745e-01
-6.44521356e-01 7.54862010e-01 5.49085915e-01 1.90524515e-02
4.67534959e-01 2.06562877e-01 -6.12894353e-03 5.18414080e-01
-8.98390472e-01 1.33724198e-01 -2.17770282e-02 4.56575632e-01
6.29354119e-01 4.87356223e-02 -5.54438710e-01 5.38860500e-01
-5.83043098e-01 9.81524289e-02 3.88122231e-01 8.17288280e-01
-2.66742319e-01 -1.39418900e+00 -1.44230708e-01 6.30978227e-01
-7.20356524e-01 -7.94670999e-01 -2.75612563e-01 3.00193310e-01
-2.99725324e-01 7.77130485e-01 -4.35809083e-02 -2.90647775e-01
2.02693462e-01 4.12936538e-01 1.92202926e-01 -5.31629860e-01
-7.29443789e-01 -5.32715879e-02 5.41846789e-02 -3.19568574e-01
1.20098330e-01 -6.05008185e-01 -1.07874382e+00 -4.93071020e-01
-2.88373560e-01 1.78352386e-01 4.90842700e-01 6.67430401e-01
6.06198192e-01 1.39448658e-01 7.22182333e-01 -5.76087356e-01
-1.17273676e+00 -1.22217965e+00 -4.93914634e-01 6.66102350e-01
2.34707654e-01 -9.92261693e-02 -3.95480067e-01 3.48881155e-01] | [11.71264362335205, 9.623723983764648] |
c21a461d-cc18-499a-845d-1a621e7d3596 | an-enhanced-deep-convolutional-encoder | null | null | https://link.springer.com/chapter/10.1007/978-3-319-60663-7_18 | https://link.springer.com/chapter/10.1007/978-3-319-60663-7_18 | An Enhanced Deep Convolutional Encoder-Decoder Network for Road Segmentation on Aerial Imagery | Object classification from images is among the many practical examples where deep learning algorithms have successfully been applied. In this paper, we present an improved deep convolutional encoder-decoder network (DCED) for segmenting road objects from aerial images. Several aspects of the proposed method are enhanced, incl. incorporation of ELU (exponential linear unit)—as opposed to ReLU (rectified linear unit) that typically outperforms ELU in most object classification cases; amplification of datasets by adding incrementally-rotated images with eight different angles in the training corpus (this eliminates the limitation that the number of training aerial images is usually limited), thus the number of training datasets is increased by eight times; and lastly, adoption of landscape metrics to further improve the overall quality of results by removing false road objects. The most recent DCED approach for object segmentation, namely SegNet, is used as one of the benchmarks in evaluating our method. The experiments were conducted on a well-known aerial imagery, Massachusetts roads dataset (Mass. Roads), which is publicly available. The results showed that our method outperforms all of the baselines in terms of precision, recall, and F1 scores. | ['Teerapong Panboonyuen'] | 2017-06-20 | null | null | null | null | ['road-segementation'] | ['computer-vision'] | [ 4.46523994e-01 1.36653148e-02 -1.68277584e-02 -5.56759298e-01
-5.41586161e-01 -5.06838679e-01 5.51862895e-01 -3.54070544e-01
-6.76145852e-01 8.17162156e-01 -1.44271329e-01 -3.91627073e-01
-1.10453553e-01 -9.73615468e-01 -1.00784183e+00 -5.70796490e-01
-2.76296794e-01 4.80198003e-02 5.47935545e-01 -1.63997710e-01
1.74866885e-01 6.14165664e-01 -1.72812915e+00 2.13878974e-02
1.01053143e+00 1.11719275e+00 3.04300934e-01 6.29500866e-01
5.45787960e-02 6.58717096e-01 -5.38069427e-01 -3.57289881e-01
7.18038380e-01 -2.44219825e-01 -1.14649105e+00 4.31357652e-01
8.56061816e-01 -8.39689195e-01 -2.58764386e-01 9.39788342e-01
3.98907036e-01 2.41652638e-01 6.24088883e-01 -9.47140813e-01
-6.53135955e-01 6.01333916e-01 -8.48836303e-01 7.18820989e-02
-1.80675447e-01 6.62105009e-02 8.97866488e-01 -6.84700131e-01
5.61768711e-01 1.11546421e+00 7.57167459e-01 9.94873717e-02
-9.96757388e-01 -5.23935795e-01 3.19123626e-01 1.81587506e-02
-1.60095918e+00 -6.10387027e-02 3.26394707e-01 -4.57687527e-01
9.19767737e-01 1.22913763e-01 5.03630579e-01 4.52745080e-01
-5.91945648e-02 9.14750397e-01 9.07982707e-01 -2.96808034e-01
-5.19682467e-02 1.24095388e-01 3.21487844e-01 7.30582893e-01
3.05346191e-01 3.69339809e-02 3.92079324e-01 2.87511826e-01
8.46233249e-01 -9.41488817e-02 -1.71352625e-01 -3.13205212e-01
-9.10933018e-01 8.15169573e-01 1.00503409e+00 7.67241642e-02
-3.80938143e-01 7.40406394e-05 5.15862405e-01 6.48136660e-02
5.18600643e-01 3.17643046e-01 -5.71040273e-01 2.83406556e-01
-1.01793242e+00 2.63019860e-01 2.91588873e-01 1.17430520e+00
7.71179974e-01 1.92820057e-01 4.36908938e-03 8.72556686e-01
1.26975447e-01 4.65739012e-01 3.09257895e-01 -8.72169733e-01
5.79766870e-01 9.14428949e-01 1.39408976e-01 -8.72293711e-01
-3.95408541e-01 -4.41259980e-01 -7.24773884e-01 2.99897403e-01
3.93268973e-01 -3.91601086e-01 -1.45645320e+00 1.28368616e+00
6.83018267e-02 -1.21351834e-02 2.40624219e-01 1.11460423e+00
9.89151597e-01 7.68563271e-01 1.33647680e-01 3.20033967e-01
1.18503058e+00 -1.01907051e+00 -4.34628040e-01 -3.96289021e-01
5.69564223e-01 -6.31503224e-01 1.05137002e+00 2.85509378e-01
-6.98631525e-01 -8.65238905e-01 -1.12429070e+00 -6.75025880e-02
-7.95545876e-01 8.85219991e-01 6.80101216e-01 5.37690699e-01
-9.13524866e-01 5.99623382e-01 -6.38221920e-01 -5.08685946e-01
7.82766640e-01 5.68629980e-01 -3.34422410e-01 -2.14973569e-01
-1.16632318e+00 7.89461136e-01 6.32806838e-01 4.39828753e-01
-7.68814027e-01 -2.42288902e-01 -8.89942706e-01 -1.79050192e-01
5.43748379e-01 -1.20512471e-01 1.04803073e+00 -1.26206160e+00
-1.23478341e+00 8.93606484e-01 1.92399889e-01 -6.88628972e-01
5.23069084e-01 -5.19044042e-01 -2.88562328e-01 1.12166926e-01
1.00788131e-01 1.34449708e+00 5.48074305e-01 -1.13233566e+00
-8.47293615e-01 -2.37507463e-01 3.73239011e-01 2.85161108e-01
5.83756045e-02 -1.74540728e-01 -3.92116785e-01 -5.12072563e-01
6.79487288e-02 -1.07236230e+00 -3.18949789e-01 -8.39735791e-02
-3.71063352e-01 -7.42935836e-02 1.07806623e+00 -8.64050686e-01
1.03976250e+00 -2.14288092e+00 -1.06475048e-01 1.61528453e-01
-1.34397402e-01 6.99708343e-01 -2.08797157e-01 2.17409149e-01
-1.49279326e-01 2.10543633e-01 -7.29219973e-01 1.43474862e-01
-2.77614504e-01 2.70482033e-01 -1.56095013e-01 2.75795728e-01
5.12583196e-01 7.60053158e-01 -6.36497378e-01 -4.54356998e-01
4.38268691e-01 3.49191964e-01 -2.31764063e-01 -6.53464198e-02
-6.87493458e-02 8.85063931e-02 -1.78991184e-01 5.92562795e-01
9.13302004e-01 1.19387604e-01 -3.19302887e-01 -2.59988308e-01
-4.21199918e-01 1.08769245e-01 -1.40055871e+00 1.42892063e+00
-2.75967538e-01 8.92365992e-01 -2.90995628e-01 -9.01833415e-01
1.11461532e+00 7.29236230e-02 2.30606049e-01 -4.39686030e-01
5.42561561e-02 2.98416410e-02 5.14490418e-02 -5.55123270e-01
8.61570835e-01 5.06285548e-01 8.92576501e-02 5.46505162e-03
6.27553388e-02 -1.94337890e-01 4.00928289e-01 2.90586203e-02
5.49015641e-01 4.97264624e-01 1.89173982e-01 -2.77707368e-01
4.68842536e-01 4.61569071e-01 4.43006903e-01 6.35873497e-01
-3.09000939e-01 7.80925751e-01 4.08676356e-01 -4.77574587e-01
-1.00681674e+00 -7.16019094e-01 -4.88422722e-01 8.21739972e-01
3.33103150e-01 -4.11605202e-02 -1.01087797e+00 -8.30739498e-01
-2.20756829e-01 5.72346985e-01 -5.74363351e-01 3.36202800e-01
-5.43134630e-01 -1.07141757e+00 7.89815009e-01 8.38527560e-01
1.19515252e+00 -1.07873058e+00 -8.33627105e-01 2.24177577e-02
-1.17044166e-01 -1.30104411e+00 4.45966907e-02 2.34939978e-01
-1.01210988e+00 -1.26718700e+00 -8.12816083e-01 -8.04749846e-01
7.46747375e-01 4.52738911e-01 8.26030076e-01 -1.17540017e-01
-5.87423563e-01 -2.91930679e-02 -4.77301210e-01 -3.85464489e-01
3.91765237e-02 3.73758554e-01 -5.56970298e-01 -2.46243939e-01
3.71028930e-01 1.33401543e-01 -6.84139848e-01 4.71388042e-01
-1.06892967e+00 1.63167581e-01 8.74954879e-01 7.13811636e-01
5.00602961e-01 1.89697370e-01 6.05848074e-01 -9.10009742e-01
1.96683630e-01 -3.31797987e-01 -8.43006909e-01 1.00080647e-01
-2.70684302e-01 -2.71524549e-01 4.26948577e-01 -5.84834069e-02
-1.02384210e+00 3.33348960e-01 -3.04322630e-01 1.15712024e-01
-5.62463105e-01 4.21885967e-01 -8.57867151e-02 -1.51042059e-01
5.18196523e-01 -7.62304589e-02 -1.72576293e-01 -2.88813144e-01
4.25154269e-01 9.27907050e-01 5.03088951e-01 -5.75578883e-02
3.61811846e-01 2.27625027e-01 -6.92141280e-02 -1.07463992e+00
-7.23793685e-01 -4.30151433e-01 -1.05707097e+00 -2.98577815e-01
9.39303935e-01 -1.10724902e+00 -9.11173448e-02 6.26873910e-01
-9.12937760e-01 -4.42183077e-01 -2.11239845e-01 6.45174861e-01
-2.32890889e-01 1.36464775e-01 -3.86126071e-01 -6.96147919e-01
-2.67068624e-01 -1.41629076e+00 1.00531518e+00 4.03861046e-01
3.94912250e-02 -6.38625503e-01 -4.70061928e-01 3.02043647e-01
3.33621114e-01 5.24124146e-01 6.88275158e-01 -4.85561460e-01
-5.87774277e-01 -2.28941932e-01 -7.11300313e-01 8.69167626e-01
1.74159706e-01 3.13870668e-01 -9.15880442e-01 -5.18097579e-02
-5.35305381e-01 -3.15859735e-01 1.14127982e+00 5.72003365e-01
1.31510973e+00 -6.56764442e-03 -4.08479184e-01 4.40004468e-01
1.63120615e+00 3.72802854e-01 1.11539650e+00 5.43972969e-01
6.84265375e-01 6.95525348e-01 9.64994371e-01 1.39674172e-01
3.45923752e-01 4.89692986e-01 6.81906164e-01 -5.78881085e-01
-2.35102490e-01 9.00904015e-02 1.92641437e-01 2.26744041e-01
-3.41331363e-01 -4.18099701e-01 -9.27004099e-01 7.66186416e-01
-1.76227510e+00 -7.42476463e-01 -5.81816316e-01 2.04854751e+00
2.75975913e-01 2.73099653e-02 2.03369990e-01 3.07020158e-01
7.40072548e-01 4.79449742e-02 -5.23219943e-01 -2.36876145e-01
-1.64376885e-01 6.89709745e-03 1.12719679e+00 3.13410401e-01
-1.80864286e+00 1.17628515e+00 6.54336643e+00 5.05080998e-01
-1.17451119e+00 -3.31736684e-01 6.48770988e-01 3.36144507e-01
5.45221806e-01 -1.83108360e-01 -8.62136364e-01 2.65163511e-01
6.72130287e-01 5.33795059e-01 1.02705069e-01 9.34714854e-01
9.09374729e-02 -5.63754618e-01 -6.08320594e-01 5.02783120e-01
-1.12599619e-02 -8.77894759e-01 -4.98912483e-02 -4.42570895e-02
7.95857966e-01 3.51237684e-01 2.19575688e-02 2.67645299e-01
3.72857600e-01 -8.85291696e-01 6.58350170e-01 3.24243098e-01
7.81179488e-01 -8.81563365e-01 1.25818729e+00 -4.26223641e-03
-1.23119283e+00 -1.57178625e-01 -4.95353132e-01 -8.62555392e-03
-3.75286490e-02 3.55900407e-01 -7.91459858e-01 7.92335391e-01
1.04884398e+00 7.65037179e-01 -9.09910917e-01 1.09966063e+00
-1.61397338e-01 7.94605553e-01 -3.52432698e-01 1.80353805e-01
7.91120410e-01 -3.69036376e-01 3.29240441e-01 1.36648285e+00
2.24058360e-01 1.47004083e-01 1.76070616e-01 7.15655148e-01
-1.91033911e-02 2.85252798e-02 -8.43644202e-01 2.70145208e-01
2.12995693e-01 1.20980048e+00 -9.84234512e-01 -2.87572861e-01
-4.72152591e-01 8.51982474e-01 -1.10077344e-01 3.72877359e-01
-1.02256119e+00 -7.96839297e-01 4.16148990e-01 1.98370308e-01
6.39391840e-01 -2.11741298e-01 -1.14770718e-01 -6.66252613e-01
-8.92620385e-02 -6.73898041e-01 2.21078947e-01 -9.13728356e-01
-8.43053281e-01 8.11794877e-01 4.27367568e-01 -1.31662083e+00
1.15757257e-01 -7.10402429e-01 -4.34972525e-01 7.53705084e-01
-1.73652768e+00 -1.24132371e+00 -6.11879945e-01 2.72363067e-01
7.99768984e-01 -1.59371838e-01 4.98856008e-01 5.05864918e-01
-8.62788379e-01 3.25097531e-01 8.33403766e-02 5.29453039e-01
5.21428466e-01 -1.18695045e+00 5.51894486e-01 1.12678766e+00
-1.11591667e-01 1.86499968e-01 3.06597620e-01 -4.18050230e-01
-7.30700314e-01 -1.61034918e+00 5.23388922e-01 -1.37207052e-02
2.73334414e-01 -1.46514893e-01 -8.43780756e-01 8.44190478e-01
1.41291976e-01 -8.40465277e-02 4.24119204e-01 -3.61385077e-01
2.43804410e-01 -1.41786888e-01 -1.28038204e+00 5.56380510e-01
8.62471819e-01 -1.78611241e-02 -3.61347139e-01 1.47565633e-01
6.29146814e-01 -4.29838926e-01 -6.49353921e-01 6.52617216e-01
4.04266447e-01 -8.93364191e-01 7.61719882e-01 -1.65547833e-01
5.70496321e-01 -7.58385718e-01 -3.07637781e-01 -1.20414782e+00
-1.33864507e-01 4.04037647e-02 3.59160423e-01 1.33533990e+00
5.25915444e-01 -5.49121857e-01 4.80274022e-01 3.18982184e-01
-3.39610249e-01 -7.83278584e-01 -4.36139286e-01 -7.16422379e-01
-7.29513988e-02 -1.74930692e-01 7.72073448e-01 6.24920547e-01
-7.95356095e-01 1.49857491e-01 -2.82359898e-01 2.96370357e-01
2.77733475e-01 -1.39815928e-02 8.57531726e-01 -1.04737580e+00
1.87766179e-01 -1.41978487e-01 -6.37464702e-01 -1.25631821e+00
-1.69338584e-01 -6.76871300e-01 8.87209252e-02 -1.86994576e+00
-2.32984215e-01 -4.74259079e-01 1.29954532e-01 6.03201210e-01
-2.46764794e-02 4.98511314e-01 7.33048767e-02 -4.10597697e-02
-3.45920682e-01 3.87427330e-01 1.17908466e+00 -1.59500614e-01
-2.50199795e-01 1.75545305e-01 -5.74274063e-01 8.88112307e-01
9.26459014e-01 -2.86515266e-01 -2.72217989e-01 -6.26484334e-01
-2.54882723e-01 -4.83422697e-01 4.41035837e-01 -1.11542535e+00
-1.26927853e-01 1.74482763e-01 5.43982148e-01 -9.94101226e-01
1.82912499e-02 -9.14951146e-01 -7.22506270e-02 3.14449042e-01
-2.68103778e-01 3.98508534e-02 5.22796273e-01 3.38310748e-01
-3.84724319e-01 -4.72927898e-01 9.56346869e-01 -3.35500836e-02
-1.30228877e+00 1.02466561e-01 -4.31986898e-01 -1.87092707e-01
1.34521902e+00 -5.51383078e-01 -2.56930828e-01 -9.77710448e-03
-3.92862141e-01 2.90807068e-01 1.97915778e-01 3.82177383e-01
4.44003761e-01 -9.60090518e-01 -7.09990919e-01 1.38323233e-01
7.98803717e-02 4.55363452e-01 7.83214197e-02 7.22030222e-01
-1.05340886e+00 4.97595936e-01 -3.15563023e-01 -6.80343568e-01
-1.29609656e+00 1.95346758e-01 4.34248716e-01 -1.20386250e-01
-7.43242085e-01 5.88363588e-01 4.23890688e-02 -6.46787465e-01
3.67232651e-01 -6.30266190e-01 -4.67636317e-01 5.77111617e-02
1.91543102e-01 5.89495063e-01 1.58418238e-01 -7.66762972e-01
-2.44803488e-01 6.96208119e-01 -1.36862382e-01 1.58102334e-01
1.51275837e+00 -2.08780974e-01 7.96256363e-02 2.30732501e-01
9.61149156e-01 -5.86195946e-01 -1.48383605e+00 -1.22970335e-01
-1.31318808e-01 -4.64671969e-01 2.68640697e-01 -1.03076231e+00
-1.15965176e+00 1.01169956e+00 9.89232242e-01 3.62734571e-02
1.25708091e+00 -3.73969793e-01 6.94820464e-01 5.21186531e-01
1.36854962e-01 -1.10676467e+00 -3.88775527e-01 5.83113909e-01
8.39613855e-01 -1.50509083e+00 2.28624105e-01 -4.05642927e-01
-9.09539521e-01 1.12798488e+00 7.73952186e-01 -3.46763223e-01
4.95439678e-01 2.29561150e-01 -2.10940242e-02 -1.52188629e-01
-5.95488138e-02 -6.52167141e-01 3.11530501e-01 5.27100027e-01
4.36873972e-01 3.81404944e-02 -2.28567362e-01 3.15023273e-01
-1.23153301e-02 6.55411258e-02 5.80210567e-01 9.01912749e-01
-5.73050559e-01 -6.98812306e-01 -2.59921342e-01 6.06693983e-01
-5.28947234e-01 2.61440873e-02 -4.85928416e-01 1.25735652e+00
4.14850175e-01 7.65476286e-01 2.58354604e-01 -2.38312319e-01
5.18357217e-01 -2.77306318e-01 2.53739893e-01 -4.65402931e-01
-5.89684367e-01 -2.35352263e-01 4.71789762e-02 -2.31105402e-01
-6.86584592e-01 -4.40734386e-01 -1.12661898e+00 -1.32767737e-01
-7.36048281e-01 -2.39645064e-01 7.85606086e-01 8.35261464e-01
2.41041049e-01 6.56356573e-01 5.42386830e-01 -9.89100277e-01
-1.23429880e-01 -1.05278802e+00 -5.13921022e-01 6.76140189e-02
1.55391186e-01 -7.07960010e-01 -1.49124950e-01 3.09770465e-01] | [9.157733917236328, -1.0376887321472168] |
e8e51306-aff3-4e31-8784-dfb2b99f9f4f | information-plane-analysis-of-deep-neural-1 | null | null | https://openreview.net/forum?id=B1l0wp4tvr | https://openreview.net/pdf?id=B1l0wp4tvr | Information Plane Analysis of Deep Neural Networks via Matrix--Based Renyi's Entropy and Tensor Kernels | Analyzing deep neural networks (DNNs) via information plane (IP) theory has gained tremendous attention recently as a tool to gain insight into, among others, their generalization ability. However, it is by no means obvious how to estimate mutual information (MI) between each hidden layer and the input/desired output, to construct the IP. For instance, hidden layers with many neurons require MI estimators with robustness towards the high dimensionality associated with such layers. MI estimators should also be able to naturally handle convolutional layers, while at the same time being computationally tractable to scale to large networks. None of the existing IP methods to date have been able to study truly deep Convolutional Neural Networks (CNNs), such as the e.g.\ VGG-16. In this paper, we propose an IP analysis using the new matrix--based R\'enyi's entropy coupled with tensor kernels over convolutional layers, leveraging the power of kernel methods to represent properties of the probability distribution independently of the dimensionality of the data. The obtained results shed new light on the previous literature concerning small-scale DNNs, however using a completely new approach. Importantly, the new framework enables us to provide the first comprehensive IP analysis of contemporary large-scale DNNs and CNNs, investigating the different training phases and providing new insights into the training dynamics of large-scale neural networks. | ['Robert Jenssen', 'Jose Principe', 'Shujian Yu', 'Michael Kampffmeyer', 'Sigurd Løkse', 'Kristoffer Wickstrøm'] | 2019-09-25 | null | null | null | null | ['information-plane'] | ['methodology'] | [-1.20235030e-02 1.92163572e-01 2.11380683e-02 -7.70130306e-02
9.89772156e-02 -6.31550968e-01 5.07116318e-01 -3.70822139e-02
-5.59202611e-01 5.42601347e-01 -2.94818670e-01 -3.95616770e-01
-6.72982872e-01 -7.76364684e-01 -7.11786628e-01 -9.94312108e-01
-5.11472166e-01 1.77555770e-01 2.63531506e-01 -8.63950048e-03
-2.26230198e-03 9.02385116e-01 -1.53359759e+00 -5.53123578e-02
5.55973530e-01 1.47706866e+00 1.49325624e-01 6.75837219e-01
9.17442217e-02 7.84415543e-01 -3.31550747e-01 -5.39266407e-01
2.00951055e-01 -6.47596270e-02 -8.69803667e-01 -1.89929798e-01
4.65397835e-01 -6.47155344e-02 -6.10439539e-01 1.20375109e+00
1.55842990e-01 4.21472862e-02 7.97106564e-01 -1.21231127e+00
-4.68837082e-01 7.07824886e-01 -8.33435636e-03 5.01203656e-01
-4.97401804e-01 4.61878106e-02 1.34169328e+00 -6.09625161e-01
5.85340977e-01 9.20486212e-01 6.93291605e-01 5.01211107e-01
-1.44636500e+00 -4.76093888e-01 -4.71875630e-03 2.24883929e-01
-1.23868012e+00 -1.74145967e-01 8.35639477e-01 -6.91909552e-01
7.77786255e-01 -4.80495691e-02 7.11635113e-01 1.11600184e+00
5.63466065e-02 8.23491514e-01 9.69095647e-01 -1.98445633e-01
3.06999415e-01 1.89742982e-01 3.49818826e-01 6.45132065e-01
4.78034765e-01 -4.47099507e-02 -2.48647168e-01 1.34847209e-01
7.74145722e-01 5.58198197e-03 -2.40360007e-01 -5.83197653e-01
-1.27945268e+00 8.50222588e-01 7.33600020e-01 6.72353804e-01
-2.10966423e-01 3.46120477e-01 5.01234412e-01 3.71900409e-01
4.16839540e-01 2.81120121e-01 -7.61671722e-01 -5.23292385e-02
-7.62614071e-01 1.80117086e-01 9.44930613e-01 6.92316473e-01
9.63563561e-01 -2.59760842e-02 3.70086312e-01 5.03497899e-01
3.26514632e-01 1.88628823e-01 2.45160967e-01 -9.21792626e-01
7.20357478e-01 7.59185493e-01 -4.43394065e-01 -1.10542393e+00
-6.94640100e-01 -7.75141537e-01 -1.41672552e+00 2.68905610e-01
6.79551125e-01 -2.69304365e-01 -2.54331440e-01 1.95276368e+00
1.14665926e-02 -1.15902163e-01 2.49036402e-01 5.38343728e-01
4.98494476e-01 2.61901289e-01 -2.19497681e-01 1.31340131e-01
1.28057241e+00 -3.54248166e-01 -3.38425815e-01 5.33342399e-02
9.38989162e-01 -2.32687101e-01 6.11267745e-01 3.16488206e-01
-8.39374363e-01 -3.26660037e-01 -1.11537099e+00 7.99840018e-02
-6.17468894e-01 3.34167689e-01 6.93689108e-01 5.99576116e-01
-1.31970990e+00 1.07706046e+00 -9.76974607e-01 -3.07661235e-01
8.63184273e-01 6.07651651e-01 -6.51403666e-01 1.65912047e-01
-1.27455509e+00 7.28955388e-01 6.36471927e-01 4.17073876e-01
-6.42405748e-01 -7.02785194e-01 -6.11807108e-01 2.22026765e-01
4.38637808e-02 -4.69020009e-01 8.82254779e-01 -9.16394711e-01
-1.43695319e+00 3.99538219e-01 2.61351824e-01 -4.56977904e-01
3.68055165e-01 6.10194467e-02 -9.39113125e-02 3.18793148e-01
-3.50707114e-01 6.49998844e-01 6.61814690e-01 -8.36738110e-01
-2.26818085e-01 -5.87789893e-01 1.37053773e-01 -2.89313763e-01
-7.36336291e-01 -2.91013300e-01 1.88533917e-01 -4.44089323e-01
2.37234473e-01 -9.15895820e-01 -9.70226899e-02 2.39981964e-01
-5.75744808e-01 -2.10115939e-01 6.56203508e-01 -2.71151274e-01
1.02129161e+00 -2.22799206e+00 6.20870054e-01 2.24895999e-01
7.00426280e-01 4.53697145e-01 -7.03003407e-02 4.66623724e-01
-3.43709797e-01 1.91522896e-01 -2.62308031e-01 -2.83171088e-01
1.24050692e-01 3.37832302e-01 -1.69022843e-01 6.48249805e-01
4.60010022e-01 9.07775640e-01 -6.05904400e-01 -2.02990606e-01
1.19052276e-01 6.87801838e-01 -5.75072527e-01 -2.47779518e-01
-1.30234227e-01 2.39316106e-01 -3.52090269e-01 5.90152517e-02
6.30363047e-01 -4.57710057e-01 1.28622591e-01 -3.59435141e-01
-1.64601862e-01 3.33199017e-02 -9.96135116e-01 1.27794707e+00
-3.37736994e-01 1.21495700e+00 7.73483515e-02 -1.60509849e+00
6.53286815e-01 3.69298458e-01 4.80028003e-01 -2.22198576e-01
3.28733504e-01 2.79740274e-01 3.63986135e-01 -3.45194012e-01
7.62750348e-03 -6.68653548e-02 2.41706237e-01 2.02217683e-01
5.22625327e-01 4.51835662e-01 2.25320131e-01 1.51164085e-01
1.24469531e+00 -5.15765607e-01 -2.80316584e-02 -5.37774265e-01
5.45118928e-01 -4.57897633e-01 1.05827883e-01 5.18979669e-01
-1.35464355e-01 2.30127975e-01 1.14036644e+00 -3.20254326e-01
-1.10886705e+00 -1.14912796e+00 -5.87246776e-01 6.11952007e-01
-4.36328858e-01 -8.33113343e-02 -8.50701928e-01 -5.30400395e-01
-7.52303924e-04 8.47291760e-03 -7.38939583e-01 -1.93490565e-01
-2.25396961e-01 -8.23557794e-01 8.69230151e-01 5.17144918e-01
7.35602677e-01 -8.71701419e-01 -4.03844923e-01 9.54868793e-02
2.51441419e-01 -1.48747027e+00 2.34986022e-01 6.01965070e-01
-1.30075192e+00 -1.17791593e+00 -8.39328647e-01 -6.18016303e-01
6.62722886e-01 -2.31378272e-01 8.28703284e-01 -2.39022300e-01
-3.02399158e-01 2.31600046e-01 -3.63866761e-02 -1.04262210e-01
-2.62870193e-01 6.36075795e-01 2.21081257e-01 2.23660707e-01
1.31101459e-01 -9.54016447e-01 -6.02547884e-01 3.39806974e-01
-1.36139143e+00 -1.26617849e-01 8.16102266e-01 6.60944641e-01
6.31480664e-02 3.76773655e-01 4.12930161e-01 -5.45592606e-01
4.51209038e-01 -6.00815177e-01 -7.60165334e-01 9.62281972e-02
-4.05116916e-01 4.17242348e-01 1.00403476e+00 -5.39489806e-01
-6.20636404e-01 -1.53538987e-01 -2.55670249e-01 -2.61911213e-01
-2.33306661e-01 5.58813989e-01 -8.19884017e-02 -4.01500106e-01
4.56990272e-01 2.28467360e-01 2.48273034e-02 -6.44092917e-01
3.49338382e-01 3.83388698e-01 2.59074569e-01 -3.92784059e-01
8.44809890e-01 5.21317959e-01 5.51096797e-01 -1.01913452e+00
-8.84160340e-01 -3.61031801e-01 -1.04257047e+00 -1.43906504e-01
9.11092579e-01 -5.21746576e-01 -1.17898953e+00 7.85032213e-01
-1.11447072e+00 -3.99488449e-01 -2.30731875e-01 7.38711476e-01
-4.93091136e-01 2.30664715e-01 -8.92704964e-01 -7.36716688e-01
2.21674633e-03 -1.08569884e+00 5.83391368e-01 1.48126513e-01
2.92308807e-01 -1.44585049e+00 4.82942536e-02 -5.28676137e-02
5.74980497e-01 1.17171384e-01 1.07140410e+00 -6.97211683e-01
-7.03164220e-01 -1.59374714e-01 -5.72980046e-01 9.18123424e-01
-1.98168263e-01 1.35602862e-01 -1.04132724e+00 -5.49624488e-02
3.60644832e-02 -1.11118250e-01 1.06289029e+00 2.91745394e-01
1.08776617e+00 -4.84960407e-01 5.47906570e-03 7.54228771e-01
1.50731051e+00 -3.87229890e-01 4.73857045e-01 1.79975361e-01
7.77287245e-01 5.81725121e-01 -3.35081577e-01 3.15682024e-01
1.96152061e-01 5.19986451e-01 6.43854260e-01 2.17024460e-01
1.35958001e-01 1.28198728e-01 2.32389808e-01 1.06811297e+00
-3.44858199e-01 -1.16558336e-01 -8.75192583e-01 3.64972532e-01
-1.62780368e+00 -8.79353285e-01 -1.62540853e-01 1.98011470e+00
6.32796407e-01 2.62473494e-01 -2.07868405e-02 4.55919415e-01
3.73148918e-01 2.98835218e-01 -5.68330824e-01 -9.95851159e-02
-2.54795283e-01 1.86976522e-01 6.63116872e-01 1.45482719e-01
-1.18590045e+00 5.83335221e-01 5.75403976e+00 6.60293221e-01
-1.11009991e+00 -4.72022332e-02 4.25962776e-01 1.38530925e-01
-2.52900682e-02 -1.25892356e-01 -9.29172039e-01 1.89749628e-01
1.04990673e+00 2.50675261e-01 4.51768696e-01 8.46752763e-01
-2.47189611e-01 1.05128512e-01 -1.12245154e+00 9.07631576e-01
-3.57427567e-01 -1.44104540e+00 2.72473712e-02 4.72942740e-01
6.30467236e-01 5.79342484e-01 3.66311491e-01 9.37667787e-02
-3.57624441e-02 -8.96116078e-01 4.58795786e-01 7.10105300e-01
4.55984205e-01 -6.84171915e-01 8.27418029e-01 5.12904048e-01
-1.10816884e+00 -1.22411147e-01 -6.27964258e-01 -1.40030906e-01
-3.32659334e-01 1.02597308e+00 -4.84353215e-01 3.01251024e-01
5.28726101e-01 9.58528697e-01 -5.97611487e-01 8.29391778e-01
2.97344774e-01 4.67867434e-01 -5.02777040e-01 -2.33327061e-01
4.08041805e-01 -2.64968485e-01 5.78086436e-01 9.77216721e-01
2.42201880e-01 -3.89130861e-01 -5.80283463e-01 1.11953664e+00
-3.07758451e-01 -1.99433416e-02 -8.06188583e-01 -4.86718506e-01
-1.21291868e-01 1.35274160e+00 -8.00519407e-01 2.04627216e-02
-4.08751011e-01 6.85565770e-01 7.22644448e-01 4.56464559e-01
-4.23616558e-01 -4.37806517e-01 1.17656648e+00 -6.09699488e-02
6.02527320e-01 -7.13505328e-01 -1.42736807e-01 -1.28355098e+00
2.53374517e-01 -3.10824007e-01 -1.12151101e-01 -1.07477963e-01
-1.37903607e+00 6.48094654e-01 2.47802623e-02 -1.02247775e+00
6.64626155e-03 -1.41221619e+00 -3.14730734e-01 6.85448229e-01
-1.56670678e+00 -5.73745847e-01 1.27423331e-01 5.99739254e-01
-1.69578716e-01 -1.66481212e-01 7.49921799e-01 4.88864958e-01
-9.17592704e-01 4.83212292e-01 5.69822252e-01 5.60666442e-01
-7.41633624e-02 -1.31272292e+00 2.50628144e-01 4.89635915e-01
3.04256380e-01 7.42301285e-01 4.26254779e-01 -8.83357897e-02
-1.50265348e+00 -8.46641004e-01 6.80607319e-01 -4.01703089e-01
1.12960279e+00 -6.91974044e-01 -9.44854856e-01 5.33399403e-01
-2.04098865e-01 5.56266725e-01 5.13468087e-01 7.28599578e-02
-4.96936411e-01 -3.91052693e-01 -7.68762767e-01 4.70490485e-01
1.00823247e+00 -7.37921476e-01 -1.69346139e-01 2.08493710e-01
5.96694469e-01 9.59496945e-02 -1.25063491e+00 4.30341601e-01
6.24383092e-01 -1.18757248e+00 1.03509653e+00 -6.31909490e-01
4.31416452e-01 6.94840923e-02 -2.46764198e-01 -1.12540603e+00
-1.02687076e-01 -4.04031456e-01 -2.17008367e-01 1.02443850e+00
5.47373593e-01 -9.41854298e-01 7.42420018e-01 4.88588363e-01
9.90567729e-02 -9.90163624e-01 -1.21407413e+00 -1.00132811e+00
3.02532434e-01 -7.65865743e-01 1.58480048e-01 6.29790902e-01
-1.63298417e-02 4.29249369e-02 2.40553729e-02 2.10315704e-01
7.27970541e-01 -4.79169309e-01 1.97965592e-01 -1.62559259e+00
-3.19503844e-01 -9.83121336e-01 -1.07133305e+00 -8.47042441e-01
4.05744374e-01 -1.05606771e+00 -3.41956645e-01 -1.10721362e+00
4.87400964e-02 -4.66726989e-01 -4.35777426e-01 1.75276250e-01
3.45462918e-01 2.76717633e-01 1.56021982e-01 2.32485682e-01
-4.46155101e-01 5.18113136e-01 1.15683222e+00 8.03681910e-02
1.63961202e-01 2.59354144e-01 -4.00445879e-01 9.04977798e-01
9.16391134e-01 -4.09373045e-01 -3.96759659e-01 -2.76602536e-01
7.68527448e-01 -3.35319638e-01 7.89052725e-01 -1.34192669e+00
2.67199606e-01 3.42359185e-01 3.12780410e-01 -2.98378557e-01
1.44843563e-01 -9.33369279e-01 -8.44565704e-02 3.21536064e-01
-4.59203333e-01 -1.09596156e-01 9.54010338e-02 6.82693839e-01
-2.90705174e-01 -3.33225250e-01 7.56303549e-01 -9.84581932e-02
-3.39305311e-01 6.03114903e-01 -4.75649416e-01 1.76237300e-01
8.00097346e-01 1.87724784e-01 -3.89010698e-01 2.96208058e-02
-8.03829253e-01 -2.93864369e-01 1.53074756e-01 4.12404723e-02
4.28847671e-01 -1.27423978e+00 -2.47975945e-01 3.33542854e-01
1.81172937e-02 6.83564842e-02 3.11531067e-01 1.18474150e+00
-5.79042077e-01 7.33509541e-01 -2.13175833e-01 -7.02246904e-01
-8.06414664e-01 4.49228883e-01 5.39866686e-01 -4.05430555e-01
-5.10738969e-01 7.81253517e-01 2.31959417e-01 -4.57474798e-01
3.17330331e-01 -7.63532639e-01 -3.06847006e-01 2.67304093e-01
3.61233324e-01 2.33636558e-01 1.15809053e-01 -5.86390078e-01
-2.94877350e-01 5.94192266e-01 2.80851759e-02 -1.71572398e-02
1.43491316e+00 -3.54080684e-02 -2.43852094e-01 7.48061597e-01
1.98107862e+00 -6.43121362e-01 -1.28260720e+00 -4.13705528e-01
2.34055042e-01 1.05472259e-01 2.13520795e-01 -2.03940704e-01
-1.51709080e+00 1.36241174e+00 4.13290888e-01 5.56974292e-01
9.90401804e-01 3.99551332e-01 5.81282675e-01 8.05105150e-01
1.46143645e-01 -8.45876515e-01 9.60429907e-02 6.94103062e-01
5.57644248e-01 -1.06716609e+00 -4.60868031e-01 -2.69423667e-02
-2.16688409e-01 1.59165311e+00 1.16694517e-01 -2.57193834e-01
1.26133263e+00 2.24994883e-01 -3.46293837e-01 -3.03951532e-01
-7.03294635e-01 -2.15609923e-01 3.89083236e-01 4.71401483e-01
2.47435853e-01 1.14813343e-01 2.68369436e-01 2.55054981e-01
-1.41174838e-01 -1.29321337e-01 2.64907867e-01 4.58223164e-01
-2.46274531e-01 -9.90940928e-01 2.14029476e-01 5.68021238e-01
-6.16683424e-01 -2.00592667e-01 -1.31692007e-01 8.36357653e-01
8.61102063e-03 3.70293260e-01 2.66923346e-02 -5.31592190e-01
1.27545044e-01 1.35386556e-01 5.58393836e-01 -3.03842127e-01
-2.26285592e-01 -6.33853793e-01 -3.09904784e-01 -5.01007438e-01
-6.19082987e-01 -7.63860762e-01 -8.63925278e-01 -3.57515037e-01
-4.00474608e-01 -1.62444785e-01 9.13874626e-01 1.17694986e+00
1.11807577e-01 4.57917809e-01 5.78242779e-01 -9.79944885e-01
-6.09647334e-01 -8.65540862e-01 -9.75453496e-01 1.40957579e-01
6.35760546e-01 -6.61534607e-01 -7.84425318e-01 -3.83473605e-01] | [7.9372735023498535, 3.5939571857452393] |
09595f24-20fd-4b13-bc97-a104b80ddb18 | improving-accented-speech-recognition-with | 2303.07924 | null | https://arxiv.org/abs/2303.07924v1 | https://arxiv.org/pdf/2303.07924v1.pdf | Improving Accented Speech Recognition with Multi-Domain Training | Thanks to the rise of self-supervised learning, automatic speech recognition (ASR) systems now achieve near-human performance on a wide variety of datasets. However, they still lack generalization capability and are not robust to domain shifts like accent variations. In this work, we use speech audio representing four different French accents to create fine-tuning datasets that improve the robustness of pre-trained ASR models. By incorporating various accents in the training set, we obtain both in-domain and out-of-domain improvements. Our numerical experiments show that we can reduce error rates by up to 25% (relative) on African and Belgian accents compared to single-domain training while keeping a good performance on standard French. | ['Yannick Estève', 'Lucas Maison'] | 2023-03-14 | null | null | null | null | ['accented-speech-recognition'] | ['speech'] | [ 8.46379697e-02 -1.27173185e-01 -1.36307508e-01 -6.36453629e-01
-1.13030720e+00 -8.39938402e-01 6.35099649e-01 8.08544531e-02
-6.80771470e-01 9.38747704e-01 3.57250035e-01 -3.95805299e-01
4.83823717e-02 -2.83975214e-01 -4.90841717e-01 -4.11528915e-01
1.02893971e-01 4.48051184e-01 2.26940081e-01 -6.12480104e-01
-1.30995855e-01 5.24824798e-01 -1.24034154e+00 2.75488496e-01
9.75094020e-01 6.82529867e-01 -6.69079088e-03 6.38732910e-01
5.74482307e-02 5.07751644e-01 -1.00619447e+00 -3.73575836e-01
1.33884460e-01 -8.72073174e-02 -7.97099233e-01 3.06027174e-01
3.91991675e-01 4.76267748e-02 -5.33754587e-01 9.52797174e-01
6.86036110e-01 3.71974349e-01 3.61328036e-01 -6.38097107e-01
-6.97871327e-01 7.72613406e-01 -1.28819451e-01 3.51288944e-01
6.66501343e-01 -1.23024443e-02 8.34937513e-01 -7.88567781e-01
4.50704455e-01 1.20855141e+00 6.34921670e-01 7.04659939e-01
-1.34746706e+00 -6.53619766e-01 9.35741961e-02 -1.28008937e-02
-1.41377163e+00 -9.64420438e-01 8.01801264e-01 5.60296774e-02
1.11398017e+00 3.36233348e-01 4.85234894e-02 1.21979523e+00
-3.28267515e-01 6.57281876e-01 1.30888462e+00 -6.20953262e-01
2.69081324e-01 2.72770464e-01 1.88009236e-02 2.21942663e-01
-4.01564658e-01 2.64998615e-01 -6.18246675e-01 1.24682924e-02
5.18527806e-01 -6.01694822e-01 -2.24475011e-01 6.58294782e-02
-1.27540767e+00 7.32812703e-01 2.38971397e-01 6.25391722e-01
-3.40045571e-01 -5.28178215e-01 4.64985847e-01 6.30931795e-01
6.36172056e-01 4.56912816e-01 -1.00063717e+00 -4.39019948e-01
-9.56240296e-01 2.18253620e-02 7.50093818e-01 8.90026808e-01
6.70542359e-01 5.58771431e-01 1.82848901e-01 1.63349831e+00
-1.25962019e-01 7.46652186e-01 8.59791994e-01 -7.27066636e-01
6.30715728e-01 2.79750556e-01 -1.34127801e-02 -4.27755207e-01
-4.29770797e-01 -4.56780404e-01 -8.32545161e-01 -1.47942483e-01
5.56542099e-01 -4.84497011e-01 -1.34692025e+00 1.82543445e+00
2.30447035e-02 -7.92382061e-02 3.54460657e-01 5.77861965e-01
3.91169399e-01 7.40830064e-01 1.69043839e-01 -5.30113339e-01
1.07980025e+00 -8.15406621e-01 -8.85599434e-01 -6.17250025e-01
6.58423722e-01 -1.06086648e+00 1.22624946e+00 3.98975581e-01
-1.11324847e+00 -8.57444346e-01 -8.34312379e-01 4.59334433e-01
-4.03739452e-01 4.29897159e-02 1.48991272e-01 1.08243728e+00
-1.03014958e+00 2.53195256e-01 -5.33845246e-01 -3.69798511e-01
1.55766224e-02 4.00718033e-01 -6.75998569e-01 -1.07030168e-01
-1.44024003e+00 9.54738259e-01 5.18420160e-01 -3.05790335e-01
-5.52728176e-01 -9.05044258e-01 -8.22195530e-01 -2.06975490e-01
3.52390736e-01 -9.40997899e-03 1.42426836e+00 -1.14151669e+00
-1.78668964e+00 8.40307653e-01 -2.94052601e-01 -6.63617432e-01
2.99533874e-01 -3.78390729e-01 -1.20962608e+00 -1.48777932e-01
-2.28884175e-01 6.01016641e-01 8.34233344e-01 -9.96401429e-01
-6.90038145e-01 -2.81697839e-01 -2.59046942e-01 2.39637747e-01
-3.35480422e-01 4.16468680e-01 -6.71110898e-02 -1.03118789e+00
-9.68090072e-02 -1.15736651e+00 -6.60507083e-02 -9.85959232e-01
-1.44974247e-01 -1.54659018e-01 8.40595663e-01 -9.08458889e-01
1.32740068e+00 -2.51370478e+00 -8.22168216e-02 1.52682245e-01
-4.03836727e-01 9.23555136e-01 -1.29425198e-01 2.84835786e-01
-2.90678382e-01 7.39144534e-02 -2.33641908e-01 -9.22099128e-02
-7.34526590e-02 4.31612402e-01 -2.50661552e-01 1.71532586e-01
3.85503560e-01 4.44790274e-01 -7.43127584e-01 -1.31765306e-01
4.05271322e-01 4.20357406e-01 -4.90422577e-01 2.50881732e-01
-1.01507485e-01 4.86071795e-01 5.84498756e-02 4.19774741e-01
5.12388706e-01 4.44191545e-01 1.83452144e-01 2.93650001e-01
5.48586808e-02 7.20220447e-01 -1.27711642e+00 1.48429990e+00
-7.33052969e-01 5.27165055e-01 1.75781593e-01 -9.24192250e-01
1.32143724e+00 6.02808833e-01 1.04903355e-01 -7.87463605e-01
-1.29825830e-01 4.33742434e-01 3.40585113e-01 9.51410383e-02
5.68379819e-01 -4.31649655e-01 -1.35071397e-01 1.71797350e-01
2.95289010e-01 -1.66336060e-01 1.85700268e-01 -2.17872545e-01
9.51963365e-01 -5.33225000e-01 3.17519993e-01 -3.80477339e-01
8.35241139e-01 -1.88303322e-01 4.95188445e-01 4.70868230e-01
-4.61636543e-01 6.89461291e-01 -8.89255702e-02 -3.45555514e-01
-1.04290199e+00 -1.14482188e+00 -2.68714696e-01 1.42737532e+00
-4.46962535e-01 -3.37573856e-01 -9.71517742e-01 -6.49047554e-01
-8.44005346e-02 9.59459603e-01 -7.00118765e-02 -1.31488144e-01
-9.43225563e-01 -7.95788169e-01 8.29280317e-01 5.84596574e-01
5.52597880e-01 -1.03493643e+00 3.06637675e-01 3.83092135e-01
-1.19062282e-01 -1.56615782e+00 -4.86018836e-01 4.44550067e-01
-6.15149558e-01 -5.30136287e-01 -8.46892595e-01 -9.50075090e-01
1.81178451e-01 -9.09170806e-02 1.11324549e+00 -3.10598522e-01
1.82590663e-01 2.47372791e-01 -4.70144272e-01 -3.05573434e-01
-1.15182650e+00 4.61469322e-01 6.36002719e-01 1.24393646e-02
4.84060794e-01 -3.30383569e-01 9.30919424e-02 5.92428803e-01
-6.56882823e-01 -6.62980914e-01 4.47881073e-01 9.66405213e-01
2.81757087e-01 3.87143567e-02 1.08659911e+00 -8.17373574e-01
6.87955081e-01 -1.71024472e-01 -4.73967463e-01 1.44665718e-01
-5.87419987e-01 2.58615166e-02 7.79146850e-01 -5.07470489e-01
-1.30258667e+00 1.45491228e-01 -5.66353023e-01 -2.46141911e-01
-7.96746194e-01 3.18529129e-01 -3.24156374e-01 1.16503388e-01
8.76872778e-01 1.96568668e-01 -6.91295192e-02 -5.55004001e-01
5.40770710e-01 1.13350964e+00 6.23155713e-01 -4.83852983e-01
8.04118872e-01 -2.49438077e-01 -7.26862669e-01 -1.27145553e+00
-7.50263035e-01 -4.90610123e-01 -8.43878448e-01 9.77533683e-02
4.67431247e-01 -1.04201043e+00 3.47903036e-02 5.88327885e-01
-8.53080690e-01 -4.26027626e-01 -2.84542024e-01 5.42204976e-01
-4.91814762e-01 2.08239734e-01 -6.00340486e-01 -7.61616528e-01
-8.29362050e-02 -9.73205149e-01 7.85427511e-01 -4.30426747e-02
-4.48096305e-01 -1.18092227e+00 1.75105035e-01 5.00562966e-01
5.36631346e-01 -4.83042598e-01 6.94366932e-01 -1.16415870e+00
5.60541488e-02 -9.99414325e-02 3.02367687e-01 9.12879586e-01
5.77296138e-01 -1.18959337e-01 -1.24420428e+00 -4.79106367e-01
-1.85902834e-01 -1.49929598e-01 5.64195931e-01 1.46172687e-01
6.81224406e-01 -2.47132882e-01 1.64875947e-02 3.52625936e-01
8.08613122e-01 4.83760297e-01 7.27066576e-01 3.97845328e-01
3.93402368e-01 4.55738187e-01 6.85096502e-01 1.85986489e-01
-5.15831001e-02 7.88970768e-01 -4.02051717e-01 -7.87464976e-02
-3.88868719e-01 -1.73411921e-01 5.02824545e-01 1.15477943e+00
1.50046915e-01 -2.72779256e-01 -1.09153688e+00 8.51113558e-01
-1.27803516e+00 -8.62489462e-01 3.18735749e-01 2.33479333e+00
1.29914904e+00 3.40443701e-01 5.61672628e-01 5.40459335e-01
9.19580877e-01 2.39659622e-01 -1.36510700e-01 -6.93421245e-01
-3.99655461e-01 4.10009265e-01 3.74203026e-01 7.58725107e-01
-1.26574183e+00 1.41663682e+00 7.48787594e+00 8.40270996e-01
-1.27767587e+00 2.27289852e-02 6.36152387e-01 1.43540382e-01
2.50980966e-02 -4.13645327e-01 -8.77439857e-01 2.82561362e-01
1.57631099e+00 -1.19702592e-01 4.67572212e-01 7.69829750e-01
6.74791932e-02 3.58388364e-01 -8.37014854e-01 9.57446039e-01
-4.41463143e-02 -8.80110562e-01 -1.44659489e-01 -2.77960896e-01
8.69020045e-01 3.15422565e-01 2.50519626e-03 4.78226691e-01
4.37775970e-01 -1.00389457e+00 4.71053988e-01 -1.98892400e-01
8.51618350e-01 -1.06403112e+00 6.94170892e-01 2.91660160e-01
-8.07985008e-01 -2.25530211e-02 -3.08065027e-01 -6.57474399e-02
1.14542834e-01 4.56393182e-01 -1.26910222e+00 3.89569551e-01
5.08532345e-01 2.71595716e-01 -5.31052351e-01 7.65392721e-01
-2.53268182e-01 1.13654482e+00 -4.10395801e-01 1.61276445e-01
1.66618273e-01 1.14012942e-01 5.19349456e-01 1.51547515e+00
3.62077206e-02 -1.04981605e-02 4.25076783e-02 1.69766814e-01
-2.32461676e-01 2.40559280e-01 -4.20783699e-01 -5.01716211e-02
7.57138908e-01 6.68242157e-01 -2.74543673e-01 -2.91870832e-01
-3.48327726e-01 9.22931910e-01 3.05831939e-01 4.91451651e-01
-5.97707331e-01 -4.87275094e-01 1.04685616e+00 1.53904721e-01
3.34456682e-01 -4.45234656e-01 -2.09174931e-01 -1.03569162e+00
-1.17721818e-01 -1.48410702e+00 2.73103088e-01 -4.18840379e-01
-1.20876324e+00 9.11680877e-01 -2.13342220e-01 -8.08392584e-01
-6.46288097e-01 -6.36474490e-01 -1.98426709e-01 8.75702322e-01
-1.46471965e+00 -9.55193996e-01 1.80046648e-01 6.32905245e-01
7.45186627e-01 -7.30781555e-01 1.10198998e+00 5.68806112e-01
-3.50783587e-01 1.05358601e+00 3.03799182e-01 3.79630864e-01
1.08999991e+00 -1.33083093e+00 7.93421924e-01 9.34494972e-01
6.59664214e-01 5.22798002e-01 6.69026554e-01 -4.29717511e-01
-8.73986483e-01 -9.40081358e-01 1.15376115e+00 -4.04125243e-01
8.33058000e-01 -3.35893333e-01 -1.16369689e+00 8.76679063e-01
3.19178581e-01 -1.60662718e-02 6.28023028e-01 5.98437846e-01
-5.91987789e-01 -2.36325115e-01 -1.02706599e+00 4.74127203e-01
8.18551838e-01 -9.96476352e-01 -9.35943186e-01 1.47121146e-01
7.63502479e-01 -3.78580540e-01 -1.02697194e+00 5.01908004e-01
1.65416196e-01 -8.05529356e-01 8.91047478e-01 -7.84492314e-01
-2.81227469e-01 -1.07771263e-01 -3.51098239e-01 -1.93930662e+00
-2.66267568e-01 -1.00206423e+00 2.60376513e-01 1.60154343e+00
6.69091582e-01 -6.20304525e-01 6.11779928e-01 3.75648767e-01
-2.58281678e-01 -5.52780852e-02 -1.06474018e+00 -1.10269368e+00
2.70300299e-01 -6.54198587e-01 5.60516059e-01 1.07451773e+00
4.56319153e-02 5.60174286e-01 -2.27936879e-01 2.74705321e-01
1.35486469e-01 -5.06653547e-01 6.49447680e-01 -1.01856077e+00
-9.46321934e-02 -2.79060274e-01 -5.85122645e-01 -9.42096293e-01
4.22943026e-01 -6.49003088e-01 1.57136127e-01 -9.01461661e-01
-4.37851876e-01 -5.77169120e-01 -5.12932897e-01 5.19280672e-01
-3.26071441e-01 1.67061299e-01 2.77400672e-01 -1.70488551e-01
-3.72772217e-01 3.52903962e-01 6.93107843e-01 -8.97587091e-02
-3.77225876e-01 2.58274794e-01 -5.09486616e-01 6.99903309e-01
1.04198658e+00 -3.13591093e-01 -2.13866755e-01 -4.58680809e-01
-5.35077870e-01 -8.90627876e-02 -2.77046829e-01 -1.22058451e+00
-2.41215993e-02 2.55226661e-02 1.95572689e-01 -2.60071993e-01
3.56336683e-01 -5.21121383e-01 -2.24388584e-01 1.39744431e-01
-4.60515767e-01 5.73102348e-02 6.70383453e-01 2.61759639e-01
-6.07430398e-01 6.68001501e-03 1.20663011e+00 1.59752250e-01
-8.66207600e-01 -1.66716829e-01 -4.64806110e-01 4.10924435e-01
6.51028275e-01 1.68105900e-01 -1.24285772e-01 -4.54584330e-01
-7.78494060e-01 -2.44580641e-01 3.62365961e-01 7.37843096e-01
2.52810985e-01 -1.17108071e+00 -8.82724285e-01 6.04129076e-01
9.41365808e-02 -3.77284527e-01 2.06224412e-01 3.17723900e-01
-1.95138931e-01 7.42842436e-01 -5.80565408e-02 -5.14852643e-01
-1.48157239e+00 4.93888855e-01 3.70985478e-01 -1.63115755e-01
-1.27771288e-01 9.67149377e-01 -4.32835370e-02 -9.95501459e-01
4.24472541e-01 -1.77610025e-01 3.23214903e-02 -7.59668127e-02
5.58482349e-01 3.48637015e-01 5.04790485e-01 -8.97253215e-01
-5.96894741e-01 3.27532619e-01 -2.55568892e-01 -3.35963875e-01
1.04352951e+00 -1.38722703e-01 4.55101907e-01 4.34305370e-01
1.03312385e+00 4.85868394e-01 -9.93148088e-01 -3.60097408e-01
4.36160743e-01 -3.28917980e-01 1.30432174e-01 -1.21952140e+00
-7.89345443e-01 7.68556118e-01 6.85092092e-01 3.95439297e-01
1.17211068e+00 -2.65619550e-02 8.41982186e-01 5.27618825e-01
2.36507863e-01 -1.33726156e+00 -1.96409553e-01 8.34142685e-01
6.89863026e-01 -1.24956739e+00 -3.69667888e-01 -3.32606316e-01
-9.39526379e-01 8.58640194e-01 3.44163090e-01 2.74058525e-02
6.14232779e-01 4.38111067e-01 6.39633060e-01 4.92040873e-01
-7.00488150e-01 -2.76704758e-01 2.93181568e-01 9.48436737e-01
6.51674092e-01 2.61518925e-01 -6.08329661e-02 4.43478853e-01
-5.16238272e-01 -2.70110786e-01 3.48582774e-01 6.13678932e-01
-4.89565164e-01 -1.40842235e+00 -5.28339207e-01 -3.68485041e-02
-6.82908237e-01 -1.63451746e-01 -4.66830105e-01 8.82237017e-01
-2.74838865e-01 1.27026391e+00 2.14403011e-02 -5.39422572e-01
6.72443628e-01 5.96689641e-01 2.26185694e-01 -6.44722581e-01
-6.91749394e-01 2.34980598e-01 5.20763695e-01 -1.59468442e-01
-2.91403174e-01 -8.14223886e-01 -1.10427237e+00 -2.72403628e-01
-4.09323931e-01 2.44104341e-01 6.17384493e-01 1.08571339e+00
3.09526592e-01 5.04166245e-01 7.99229085e-01 -3.27629000e-01
-8.88716698e-01 -1.30179918e+00 -6.11446142e-01 5.75503290e-01
4.57378149e-01 -4.87777025e-01 -4.79757607e-01 1.52840927e-01] | [14.365167617797852, 6.653918743133545] |
b37f08d5-d22c-4d29-8f56-f21c3c2e88c8 | mvco-dot-multi-view-contrastive-domain | 2304.07465 | null | https://arxiv.org/abs/2304.07465v1 | https://arxiv.org/pdf/2304.07465v1.pdf | MvCo-DoT:Multi-View Contrastive Domain Transfer Network for Medical Report Generation | In clinical scenarios, multiple medical images with different views are usually generated at the same time, and they have high semantic consistency. However, the existing medical report generation methods cannot exploit the rich multi-view mutual information of medical images. Therefore, in this work, we propose the first multi-view medical report generation model, called MvCo-DoT. Specifically, MvCo-DoT first propose a multi-view contrastive learning (MvCo) strategy to help the deep reinforcement learning based model utilize the consistency of multi-view inputs for better model learning. Then, to close the performance gaps of using multi-view and single-view inputs, a domain transfer network is further proposed to ensure MvCo-DoT achieve almost the same performance as multi-view inputs using only single-view inputs.Extensive experiments on the IU X-Ray public dataset show that MvCo-DoT outperforms the SOTA medical report generation baselines in all metrics. | ['Thomas Lukasiewicz', 'Junyang Chen', 'Wenting Xu', 'Zhenghua Xu', 'Xiangtao Wang', 'Ruizhi Wang'] | 2023-04-15 | null | null | null | null | ['medical-report-generation'] | ['medical'] | [-1.32439816e-02 3.53651911e-01 -3.62387121e-01 -4.87834156e-01
-1.35249054e+00 -3.70814204e-01 5.26825368e-01 -1.11998789e-01
1.23779275e-01 8.56803000e-01 7.07835317e-01 -1.47746474e-01
-7.71932900e-02 -6.34257495e-01 -6.77954257e-01 -5.56789994e-01
3.35213721e-01 4.79837120e-01 -7.34062940e-02 2.52989143e-01
-1.89634904e-01 -2.55263209e-01 -8.27243865e-01 8.41201007e-01
6.90400302e-01 8.32456410e-01 4.44860667e-01 6.96416497e-01
1.81796223e-01 1.18199933e+00 -4.23863471e-01 -4.97166842e-01
-6.77564144e-02 -9.42007065e-01 -9.06178355e-01 3.79945725e-01
4.62009102e-01 -8.02367091e-01 -4.21658605e-01 8.99084389e-01
7.86113679e-01 -1.52698904e-01 6.43505514e-01 -8.89943242e-01
-1.02097654e+00 6.60586298e-01 -7.60582447e-01 2.78449982e-01
4.12477404e-01 2.27482989e-01 8.77977729e-01 -8.45035851e-01
1.11779535e+00 1.03858888e+00 4.14122850e-01 7.79092431e-01
-1.10832191e+00 -7.43039429e-01 2.48626634e-01 4.62806188e-02
-9.23924685e-01 -4.18456942e-02 6.97155833e-01 -3.04456800e-01
6.55834973e-01 2.40562364e-01 5.05031049e-01 1.42416000e+00
5.15890777e-01 1.09289193e+00 1.33782411e+00 1.00397319e-01
-5.03177233e-02 -2.13761926e-02 -2.94030309e-01 8.43153000e-01
1.70903921e-01 9.79011804e-02 -4.97255117e-01 -2.11044550e-01
8.51550281e-01 2.48026162e-01 -4.30222839e-01 -3.44235539e-01
-1.69109881e+00 8.75721633e-01 6.38898790e-01 3.84019136e-01
-5.01858771e-01 4.60442491e-02 4.89337832e-01 2.16613904e-01
7.59522140e-01 3.05919379e-01 -2.89616525e-01 1.83676392e-01
-7.42551684e-01 3.46524477e-01 3.09141099e-01 1.07281625e+00
2.87695706e-01 -3.14398885e-01 -6.45739496e-01 7.95381725e-01
2.49931470e-01 5.88040650e-01 5.01114666e-01 -8.37198138e-01
1.07820392e+00 5.49708545e-01 -5.98652586e-02 -8.91584277e-01
-4.72788632e-01 -5.18660486e-01 -1.20047259e+00 -6.27051815e-02
-3.86041403e-02 -2.33029664e-01 -1.21581376e+00 1.63330233e+00
3.32002044e-01 1.18050352e-02 3.08920294e-01 9.60989535e-01
1.30871749e+00 4.62229520e-01 5.34372590e-03 -3.69442433e-01
1.40652311e+00 -1.18327892e+00 -9.31442797e-01 -1.47614991e-02
6.94172204e-01 -6.55663967e-01 7.42683887e-01 2.17499718e-01
-1.27275217e+00 -5.48944175e-01 -9.63735580e-01 1.61393151e-01
1.29001424e-01 2.50434995e-01 5.97236395e-01 2.14536209e-02
-7.29555011e-01 2.66325265e-01 -9.84732389e-01 5.90073504e-02
6.80120349e-01 2.88064927e-02 -5.90355217e-01 -7.85952270e-01
-1.03316033e+00 6.71693206e-01 3.45733881e-01 -1.99904516e-01
-1.02914453e+00 -1.09105957e+00 -8.72424901e-01 -2.85470486e-01
4.67557311e-01 -1.27426088e+00 1.37719536e+00 -7.37226486e-01
-1.14823067e+00 1.15876162e+00 -1.13912217e-01 -9.09831375e-02
7.81916916e-01 -1.86044827e-01 -5.42493463e-01 5.40410399e-01
5.06879151e-01 8.04025412e-01 4.97879833e-01 -1.56943679e+00
-4.75415766e-01 -3.13552141e-01 1.50109261e-01 3.47200006e-01
1.09284669e-01 -4.13307667e-01 -6.86907470e-01 -9.33330476e-01
9.81399044e-02 -1.02424407e+00 -2.65696108e-01 -1.67021543e-01
-7.00021029e-01 -2.49295548e-01 3.98705840e-01 -7.85673738e-01
1.06327534e+00 -2.00198102e+00 1.66396111e-01 -1.38190135e-01
5.58904409e-01 -2.36684997e-02 -2.38493145e-01 1.89700112e-01
-4.05991614e-01 -1.67257085e-01 -3.45620602e-01 -3.45513433e-01
-4.90491420e-01 2.78794110e-01 -8.37859213e-02 7.82451555e-02
3.09125930e-01 1.13297284e+00 -1.15041983e+00 -7.21790373e-01
-2.12756637e-02 2.42654204e-01 -7.33877718e-01 4.53513563e-01
-1.54466018e-01 9.35104311e-01 -6.07133925e-01 4.96558815e-01
7.65938342e-01 -1.01234698e+00 2.90701121e-01 -5.36108315e-01
5.31506181e-01 2.17369586e-01 -6.45274937e-01 2.28935504e+00
-7.98826933e-01 1.30990073e-01 -3.28293204e-01 -6.23414159e-01
4.96286690e-01 7.08111584e-01 9.01401103e-01 -9.68187511e-01
-2.47832507e-01 1.19352594e-01 -8.06112513e-02 -5.23672223e-01
2.74089068e-01 -4.61014360e-01 9.35756490e-02 7.96720505e-01
1.58752039e-01 7.51050562e-02 -8.04174244e-02 6.64235950e-01
1.17168415e+00 5.40089756e-02 3.07444632e-01 2.57321030e-01
3.65575582e-01 -1.66417256e-01 7.01931596e-01 5.63394547e-01
-6.00327104e-02 1.21066833e+00 4.54250604e-01 -4.01397914e-01
-8.78419101e-01 -1.32176793e+00 1.95979793e-02 4.27466869e-01
2.84965187e-01 -3.60005349e-01 -5.84513545e-01 -1.44057715e+00
-2.29476005e-01 5.21947861e-01 -7.91673541e-01 -1.68616548e-01
-5.30970812e-01 -6.96244419e-01 1.21951215e-01 8.09295893e-01
5.65922201e-01 -1.00544930e+00 -4.70979840e-01 2.60969877e-01
-7.04915524e-01 -1.35625482e+00 -9.68816876e-01 -2.22859874e-01
-9.70016301e-01 -1.11574543e+00 -1.35563481e+00 -4.57319289e-01
8.53472114e-01 3.24128032e-01 1.53980732e+00 6.51605576e-02
-1.43164158e-01 5.23409963e-01 -5.41163921e-01 -2.66498715e-01
-6.14356220e-01 -2.31563393e-02 -4.19960052e-01 -1.48410067e-01
3.94295827e-02 -3.39860022e-01 -9.86130655e-01 6.25958517e-02
-1.09564292e+00 5.93234122e-01 7.86001325e-01 1.31419969e+00
9.36408460e-01 -4.60782737e-01 6.39762759e-01 -1.34976554e+00
4.88942593e-01 -6.16791427e-01 3.28288116e-02 4.48317289e-01
-8.01341236e-01 7.70133957e-02 3.31668824e-01 -1.70782909e-01
-1.28011990e+00 -1.93800241e-01 -5.99263124e-02 -6.82539225e-01
1.44807901e-02 6.14113331e-01 -1.83619596e-02 6.34118557e-01
3.54710102e-01 3.82349551e-01 -2.44676191e-02 -1.89438432e-01
3.14888388e-01 3.66688788e-01 3.18909883e-01 -5.93176708e-02
3.79846424e-01 7.17927098e-01 -2.45403916e-01 2.52071381e-01
-1.41532815e+00 -4.25488979e-01 -3.52308542e-01 -1.15624405e-01
1.22244310e+00 -1.19317067e+00 -3.73685211e-01 2.87421405e-01
-1.32589328e+00 -1.81635786e-02 -6.86330572e-02 7.28772461e-01
-8.27883601e-01 1.87451333e-01 -8.05107355e-01 -1.25299722e-01
-6.60451949e-01 -1.44708252e+00 1.22740412e+00 1.21265613e-01
-1.37380734e-01 -1.02052212e+00 1.73861697e-01 6.03801072e-01
-5.29992655e-02 2.89528310e-01 8.58929932e-01 -3.57739091e-01
-5.61391652e-01 2.91420341e-01 -4.31307584e-01 3.34168792e-01
5.96172631e-01 -6.11182094e-01 -8.53734970e-01 -4.46005970e-01
-6.32729009e-03 -5.19938111e-01 1.01634896e+00 5.79178512e-01
1.38900542e+00 -1.30065933e-01 -4.72751349e-01 6.89830720e-01
1.44579387e+00 2.29867071e-01 4.77634162e-01 -3.57021787e-03
8.35414112e-01 2.96189457e-01 8.33220124e-01 4.52994734e-01
8.89564574e-01 4.49860811e-01 5.51307619e-01 -4.94049728e-01
-3.15316200e-01 -4.41719770e-01 -2.06924397e-02 1.29151976e+00
4.50843126e-02 -2.21504539e-01 -6.77771986e-01 5.98287463e-01
-1.75945663e+00 -9.49412942e-01 1.71575084e-01 1.73193944e+00
9.33058739e-01 1.16379885e-03 3.31884250e-02 -4.25041795e-01
4.33617026e-01 2.69210368e-01 -6.97121084e-01 3.08537632e-02
1.49379775e-01 -2.09946614e-02 2.90846944e-01 1.76192597e-01
-1.14955473e+00 2.69987524e-01 6.42651558e+00 5.02488077e-01
-1.03816330e+00 4.12776917e-01 9.70486581e-01 -1.43717110e-01
-7.48588979e-01 -5.13224900e-01 -3.92718494e-01 4.52697963e-01
5.40309250e-01 1.41156584e-01 -2.42064372e-01 8.07121217e-01
-1.22827649e-01 -3.91801260e-02 -1.24735129e+00 1.28672945e+00
3.59835267e-01 -1.78089797e+00 3.33688706e-01 2.84835190e-01
1.25219858e+00 2.32617091e-03 2.46669173e-01 1.33195922e-01
6.63428187e-01 -8.93820047e-01 1.85005993e-01 7.07422376e-01
9.20369208e-01 -7.59409964e-01 8.39229703e-01 9.03808400e-02
-9.82185602e-01 1.78888708e-01 -5.70910200e-02 8.03274691e-01
3.59474748e-01 6.24529481e-01 -7.82144189e-01 1.31100452e+00
5.19751072e-01 1.06787324e+00 -2.91486919e-01 7.10424185e-01
-4.48218659e-02 4.69323605e-01 2.91354120e-01 4.41638678e-01
3.02594155e-01 7.63513893e-02 2.09468111e-01 9.50715303e-01
3.12216640e-01 2.15926424e-01 3.08013022e-01 9.86443281e-01
-2.26853266e-01 -2.14391991e-01 -6.00122571e-01 1.94079325e-01
3.29635739e-02 1.14892662e+00 -5.52219748e-01 -6.86568022e-01
-8.16078961e-01 1.20415580e+00 2.26927385e-01 2.23854795e-01
-8.54275107e-01 1.93061143e-01 3.64980489e-01 1.82110327e-03
2.78815687e-01 2.29001373e-01 -2.17721358e-01 -1.29237819e+00
1.98451072e-01 -1.07825708e+00 7.43926942e-01 -9.70968783e-01
-1.64383519e+00 7.98299432e-01 3.61192115e-02 -1.77144194e+00
-5.17583668e-01 -3.05358291e-01 -3.48470360e-01 8.21818948e-01
-1.64104223e+00 -1.44477248e+00 -3.38710576e-01 5.72204113e-01
6.77957416e-01 -3.25925231e-01 9.28700387e-01 3.21119577e-01
-5.94390780e-02 6.76513195e-01 -1.24408051e-01 2.68626928e-01
8.90722156e-01 -1.41827655e+00 3.16769332e-01 4.30000871e-01
8.88916701e-02 4.49601024e-01 1.52845189e-01 -7.52031505e-01
-9.65712547e-01 -1.10070562e+00 3.80617976e-01 -6.13921583e-01
1.55337229e-01 1.97995335e-01 -9.25225735e-01 7.12166309e-01
5.70550144e-01 3.94545615e-01 9.70734298e-01 -9.11512449e-02
-4.57710743e-01 -9.47894230e-02 -1.08279204e+00 3.92073542e-01
9.35346246e-01 -5.85956216e-01 -3.55375588e-01 4.41181302e-01
1.02202594e+00 -7.31136203e-01 -1.36733174e+00 5.68519413e-01
3.36233586e-01 -9.09083247e-01 8.84824634e-01 -7.88675249e-01
1.27348101e+00 -1.26567215e-01 -7.62882009e-02 -1.60502386e+00
-2.24734262e-01 -7.34820440e-02 -1.60999596e-01 1.03561342e+00
4.48612541e-01 -3.27582091e-01 6.78979158e-01 1.20198213e-01
-1.95494801e-01 -1.04171371e+00 -8.15772235e-01 -3.63126338e-01
2.87611246e-01 -1.35318622e-01 5.07000983e-01 1.26607859e+00
-2.00910047e-01 4.02088761e-01 -5.25728643e-01 4.04078327e-02
5.69911838e-01 5.20849109e-01 5.35840094e-01 -6.73048258e-01
-8.32063138e-01 8.29444453e-02 -3.27045284e-03 -9.01271701e-01
-2.58355021e-01 -1.11970091e+00 -5.43081127e-02 -1.98437190e+00
9.89255071e-01 -3.19451094e-01 -5.30082405e-01 2.84089386e-01
-8.11593294e-01 1.99484527e-01 1.86952516e-01 1.93964422e-01
-7.70430326e-01 5.43286800e-01 2.13055253e+00 -3.70151728e-01
9.65582803e-02 -9.89425555e-03 -8.24646950e-01 5.42963624e-01
5.00802636e-01 -6.14716291e-01 -6.13434255e-01 -6.13320589e-01
1.87795997e-01 7.62451172e-01 2.46153742e-01 -8.53122294e-01
-1.06314227e-01 -2.09712982e-01 5.86466074e-01 -9.71627355e-01
1.23558685e-01 -5.85265338e-01 1.22884475e-02 5.84996164e-01
-5.57941258e-01 3.03884089e-01 1.74533632e-02 8.42650294e-01
-5.39762914e-01 1.47129908e-01 6.98476851e-01 -6.95356667e-01
-3.05584431e-01 6.06773317e-01 -3.01729534e-02 2.21193328e-01
9.62768376e-01 2.50057608e-01 -3.65453392e-01 -3.77441883e-01
-9.61279273e-01 3.89247924e-01 -3.08922008e-02 6.76679850e-01
9.22215819e-01 -1.56016695e+00 -9.83113766e-01 -1.46847993e-01
5.48760235e-01 2.86527932e-01 8.46349239e-01 9.48749661e-01
-3.20568949e-01 3.65645081e-01 -1.18208066e-01 -7.29368210e-01
-1.36167693e+00 7.38847136e-01 2.70625412e-01 -1.11392665e+00
-9.20338988e-01 6.58540606e-01 8.48223567e-01 -5.66380680e-01
-2.42126241e-01 -2.44913042e-01 9.96303558e-03 -2.38789991e-01
7.74439037e-01 -1.16177551e-01 8.74073654e-02 -2.82407016e-01
-1.53080404e-01 4.27020758e-01 -6.85406506e-01 -2.64976859e-01
1.31332362e+00 -1.91874549e-01 2.23712355e-01 4.11698461e-01
1.22205245e+00 -1.99264601e-01 -1.09567595e+00 -4.34930414e-01
-3.55358332e-01 -2.79253513e-01 2.14165691e-02 -1.10119295e+00
-1.36468697e+00 8.66386890e-01 7.76503563e-01 -4.27205414e-01
1.16167688e+00 2.75387824e-01 9.97563720e-01 5.82725219e-02
2.22484633e-01 -8.97300720e-01 8.51288497e-01 -6.07591383e-02
1.08753920e+00 -1.57217216e+00 1.74874350e-01 -4.61734891e-01
-1.35085285e+00 7.25006402e-01 9.09017146e-01 5.73597662e-02
6.57576740e-01 1.08049110e-01 2.69847602e-01 -4.02203411e-01
-1.08871496e+00 -5.13709001e-02 3.26606423e-01 7.13736415e-01
6.15374565e-01 2.30720654e-01 -1.84620783e-01 4.26688224e-01
2.54501611e-01 2.26430535e-01 4.26226109e-01 8.07904840e-01
2.29983687e-01 -1.04402649e+00 -2.08056659e-01 6.82946563e-01
-7.52179742e-01 -1.15086988e-01 -9.09172371e-02 6.00279868e-01
9.59574059e-02 6.80341005e-01 -7.38365129e-02 -5.25266826e-01
2.79140443e-01 -3.65265906e-01 5.98480642e-01 -7.39473283e-01
-5.03123581e-01 1.54104054e-01 -1.02594547e-01 -5.69521546e-01
-7.61320889e-01 -5.36270320e-01 -1.29149103e+00 9.53667518e-03
-3.42898853e-02 -1.64822266e-01 3.61565024e-01 9.71183538e-01
4.90826339e-01 1.07264233e+00 7.63069928e-01 -1.16797611e-01
-4.86890346e-01 -9.67631042e-01 -4.57242787e-01 7.64159203e-01
3.98611009e-01 -6.09362602e-01 3.06736022e-01 2.60856390e-01] | [15.032275199890137, -1.4103494882583618] |
7cb80578-1224-4360-a0c2-ea05a9ceddd2 | modelling-arbitrary-complex-dielectric | 2109.01928 | null | https://arxiv.org/abs/2109.01928v1 | https://arxiv.org/pdf/2109.01928v1.pdf | Modelling Arbitrary Complex Dielectric Properties -- an automated implementation for gprMax | There is a need to accurately simulate materials with complex electromagnetic properties when modelling Ground Penetrating Radar (GPR), as many objects encountered with GPR contain water, e.g. soils, curing concrete, and water-filled pipes. One of widely-used open-source software that simulates electromagnetic wave propagation is gprMax. It uses Yee's algorithm to solve Maxwell's equations with the Finite-Difference Time-Domain (FDTD) method. A significant drawback of the FDTD method is the limited ability to model materials with dispersive properties, currently narrowed to specific set of relaxation mechanisms, namely multi-Debye, Drude and Lorentz media. Consequently, modelling any arbitrary complex material should be done by approximating it as a combination of these functions. This paper describes work carried out as part of the Google Summer of Code (GSoC) programme 2021 to develop a new module within gprMax that can be used to simulate complex dispersive materials using multi-Debye expansions in an automatic manner. The module is capable of modelling Havriliak-Negami, Cole-Cole, Cole-Davidson, Jonscher, Complex-Refractive Index Models, and indeed any arbitrary dispersive material with real and imaginary permittivity specified by the user. | ['Antonios Giannopoulos', 'Craig Warren', 'Iraklis Giannakis', 'Sylwia Majchrowska'] | 2021-09-04 | null | null | null | null | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [-8.16528425e-02 -2.90747136e-01 1.00196254e+00 -6.87974393e-02
-4.11705494e-01 -1.97111726e-01 1.93220913e-01 -7.71566257e-02
-3.67539585e-01 8.61630261e-01 -1.20431393e-01 -7.82974839e-01
-3.01727027e-01 -1.30752480e+00 1.09108403e-01 -8.72570932e-01
-2.21285135e-01 7.03645766e-01 4.51717347e-01 -5.00568628e-01
1.33153647e-01 9.83823061e-01 -1.54591298e+00 1.07043058e-01
9.62681651e-01 8.52401137e-01 2.10330769e-01 7.60568738e-01
1.39409572e-01 7.39372432e-01 -2.98993468e-01 -1.08542331e-01
-3.61439586e-01 -2.06073105e-01 -7.92385161e-01 -5.23784876e-01
-7.35179603e-01 -4.15473133e-01 2.60023266e-01 6.99401259e-01
8.09525132e-01 2.03530863e-01 1.10331583e+00 -7.34943092e-01
-3.13750990e-02 2.36130878e-01 -4.98009324e-01 2.13753507e-01
8.44949603e-01 -4.08131838e-01 5.91063052e-02 -6.15356207e-01
3.15401375e-01 1.05732608e+00 1.15392900e+00 2.86323458e-01
-8.57361376e-01 -3.69192094e-01 -6.79127157e-01 1.71156392e-01
-1.28410542e+00 1.74533144e-01 4.34228003e-01 -3.62568796e-01
1.25578976e+00 7.96890557e-01 8.44344437e-01 7.88750529e-01
5.44413567e-01 8.21752101e-02 1.23103833e+00 -7.83030212e-01
4.85220104e-01 1.77056953e-01 4.72812325e-01 1.92417800e-01
4.68432009e-01 3.57173204e-01 4.86866474e-01 -7.78337777e-01
4.37136441e-01 -3.76329184e-01 -3.14083129e-01 -7.49593079e-02
-5.28482080e-01 7.43573308e-01 -1.06453143e-01 8.54929328e-01
-8.25576961e-01 -9.58472714e-02 2.88751960e-01 3.98833379e-02
8.03445339e-01 2.23489270e-01 -2.85826594e-01 -2.88249195e-01
-8.14027846e-01 7.69658625e-01 1.28937304e+00 5.11814356e-01
5.06205201e-01 1.25821188e-01 3.55647415e-01 7.38049984e-01
6.32452309e-01 8.37971568e-01 1.03538297e-01 -3.21303457e-01
-2.78589547e-01 -1.18590653e-01 4.24080402e-01 -8.11248004e-01
-7.33406305e-01 -5.00654459e-01 -8.30244422e-01 2.35578224e-01
4.06647585e-02 -4.13207024e-01 -6.59596980e-01 8.27851951e-01
6.12880707e-01 1.80999994e-01 6.13684237e-01 6.54856861e-01
1.23584545e+00 9.19553578e-01 1.97324902e-01 -4.41039801e-01
1.32398355e+00 -1.32319987e-01 -4.25126225e-01 6.13539293e-03
1.08754194e+00 -1.20822501e+00 -3.58708506e-03 5.59730887e-01
-1.10824203e+00 9.87333059e-02 -7.29144633e-01 5.77934682e-01
-2.05865905e-01 -2.55752236e-01 8.05452228e-01 1.14331102e+00
-9.70670283e-01 5.80862880e-01 -7.68816113e-01 -2.10105553e-01
-2.25825608e-01 -4.94618230e-02 -2.60032825e-02 -3.13195378e-01
-1.35624444e+00 1.18935263e+00 -1.97039880e-02 2.89933175e-01
-7.31225386e-02 -1.02869368e+00 -7.31162667e-01 -1.99842706e-01
-4.77867991e-01 -7.13226736e-01 1.24589860e+00 -1.69188157e-01
-1.37953925e+00 7.28647768e-01 3.25015038e-01 -1.66765571e-01
4.51759398e-01 7.30794370e-02 -9.49226201e-01 2.89192885e-01
-9.58841294e-02 -3.72933060e-01 3.19321156e-01 -1.53666449e+00
-1.77333817e-01 2.70312652e-02 -7.38610029e-02 1.63632967e-02
4.80965018e-01 4.49222088e-01 5.63084662e-01 -3.14365804e-01
3.99277180e-01 -5.90451300e-01 -4.64342833e-01 -7.19742477e-01
-2.36329988e-01 1.95827007e-01 7.41813302e-01 -9.92527187e-01
1.00314140e+00 -1.81208336e+00 -5.02364993e-01 7.08669901e-01
-1.88411340e-01 3.95863712e-01 3.66220742e-01 1.11423647e+00
-3.09348464e-01 -4.65990245e-01 -6.22914016e-01 2.60389894e-01
-1.45461008e-01 2.45389771e-02 -4.04902287e-02 7.48099446e-01
-5.32791317e-01 7.60728717e-02 -7.27816463e-01 -1.75778553e-01
1.79139376e-01 1.07026517e+00 -4.29730326e-01 -2.29403466e-01
2.48390049e-01 4.24307913e-01 -8.14015210e-01 6.07576966e-01
1.79928303e+00 3.52029949e-01 -1.24103256e-01 -1.92342892e-01
-7.56543756e-01 -6.67064562e-02 -1.41874766e+00 6.35279179e-01
-7.95578301e-01 7.17919916e-02 5.10668933e-01 -1.25529683e+00
1.10303152e+00 6.46129012e-01 6.87037826e-01 -8.53602767e-01
1.19530156e-01 8.10548007e-01 -7.09450319e-02 -7.75775790e-01
4.61269587e-01 -7.25927830e-01 4.66565192e-01 5.13552248e-01
-4.13679302e-01 -6.56985223e-01 6.56400993e-02 -1.00247808e-01
1.21950245e+00 -8.90824106e-03 -2.79805139e-02 -8.50334346e-01
5.70948839e-01 5.07054068e-02 -4.01455387e-02 4.17612106e-01
5.68336844e-01 3.72396082e-01 -2.50268038e-02 -5.13397098e-01
-7.61674106e-01 -7.91331470e-01 -9.47912753e-01 3.49689037e-01
2.34985258e-03 1.07423410e-01 -5.53113222e-01 5.00304341e-01
-3.30821872e-01 1.10132647e+00 -9.67330635e-02 -7.12786093e-02
-4.64844733e-01 -1.60828567e+00 4.99974519e-01 -3.22738849e-02
4.33391422e-01 -1.33017516e+00 -1.03195679e+00 6.10240996e-01
-1.05258718e-01 -8.46900642e-01 8.03593636e-01 2.78422356e-01
-9.77049112e-01 -1.00789094e+00 -1.06373739e+00 -6.20953977e-01
4.81576622e-01 -9.24636945e-02 1.24354398e+00 1.83581561e-01
-6.68969631e-01 9.03878272e-01 -9.45308626e-01 -3.24337631e-01
-7.57636487e-01 -4.82214183e-01 -3.91633242e-01 -4.25888449e-01
1.58885509e-01 -8.96454334e-01 -6.20880127e-01 3.93748164e-01
-1.05796278e+00 -1.75071403e-01 2.39213347e-01 2.08703578e-01
1.03727328e-02 4.37941164e-01 3.84473681e-01 -7.25153327e-01
8.17466855e-01 -5.82679570e-01 -3.47034574e-01 3.52896843e-03
-6.97861761e-02 -3.83958876e-01 3.38504642e-01 -1.95505381e-01
-1.31470406e+00 -6.90089345e-01 -9.56942081e-01 6.55999064e-01
-3.61586899e-01 1.00970709e+00 -4.82439660e-02 -6.34565949e-01
7.20022917e-01 1.62471414e-01 -3.77024442e-01 -5.87655425e-01
-4.46811825e-01 7.29448915e-01 2.21389160e-01 -8.43027890e-01
5.22137582e-01 2.66283393e-01 3.06883425e-01 -1.55804574e+00
-9.58983824e-02 -6.10110521e-01 1.49742305e-01 -3.23859513e-01
3.43207359e-01 -4.16444659e-01 -6.21555328e-01 1.08790207e+00
-1.29190361e+00 -3.57864171e-01 -9.35342759e-02 9.89364564e-01
-2.80198991e-01 6.30078912e-01 -6.32396936e-01 -1.19543576e+00
-6.99754715e-01 -7.09957242e-01 4.55500066e-01 1.68921977e-01
-2.01064661e-01 -1.35555875e+00 4.46506977e-01 -1.20063625e-01
8.97624195e-01 7.02618480e-01 9.07302916e-01 -3.95849571e-02
1.05039172e-01 -5.86549580e-01 -6.97429329e-02 2.62617648e-01
-3.60867172e-01 1.40898228e-01 -8.03339660e-01 -3.15830380e-01
5.69855750e-01 1.27139747e-01 5.52793860e-01 8.50633979e-01
6.32548988e-01 -8.48819390e-02 -5.13014078e-01 7.12089717e-01
1.92787576e+00 4.85939533e-01 1.43979049e+00 5.48368990e-01
-5.67232771e-03 2.89725363e-01 6.24412179e-01 9.73514915e-01
1.10016659e-01 1.65747851e-01 4.26259011e-01 -1.17235914e-01
6.71728969e-01 7.34475493e-01 -2.17817441e-01 3.87126952e-01
-1.02102995e+00 -2.70667017e-01 -1.03728795e+00 1.79953307e-01
-1.06155980e+00 -1.20906305e+00 -1.02624261e+00 2.15823436e+00
5.68078101e-01 -3.62650268e-02 -2.04058543e-01 7.39265323e-01
4.39507008e-01 -4.55182910e-01 3.00079823e-01 -7.84355879e-01
4.75961342e-02 9.24487054e-01 5.29209971e-01 7.15274811e-01
-5.80928743e-01 -5.61990477e-02 5.80800056e+00 5.33357084e-01
-1.20531392e+00 6.10891357e-02 -3.54719236e-02 7.50916660e-01
-6.27829194e-01 2.44202688e-02 -4.24125731e-01 3.84782881e-01
1.16684318e+00 -1.10706300e-01 -9.99863166e-03 4.70052451e-01
2.22797900e-01 -8.89297307e-01 -2.19152331e-01 6.20372593e-01
-3.96072805e-01 -8.39457929e-01 -5.41639805e-01 -1.08991247e-02
3.92363161e-01 -1.39698818e-01 -3.39554459e-01 2.36874849e-01
-4.63353321e-02 -7.10521460e-01 4.10945505e-01 8.85733426e-01
6.02294087e-01 -9.32473779e-01 9.69436586e-01 4.83919770e-01
-1.08154881e+00 3.73768032e-01 -5.33455312e-01 -1.16701961e-01
6.45469964e-01 1.34804475e+00 -2.76420802e-01 9.06272948e-01
7.96466112e-01 8.33418369e-02 2.01895878e-01 1.73639524e+00
4.83188301e-01 5.49409688e-01 -7.83686340e-01 3.16414982e-02
3.51894200e-01 -5.64640999e-01 5.15531838e-01 1.32277417e+00
1.14234471e+00 7.12202847e-01 -4.32726502e-01 5.49956083e-01
1.17948425e+00 2.56693572e-01 -1.78980753e-01 3.09944153e-01
1.16754144e-01 1.11337411e+00 -8.61845791e-01 3.11314836e-02
-2.16174081e-01 4.90499794e-01 -9.28599179e-01 5.57971954e-01
-1.01625502e+00 -7.07636416e-01 1.87512591e-01 7.73472428e-01
2.04543203e-01 -1.55660570e-01 1.37041986e-01 -4.96570706e-01
-2.67180830e-01 -4.24361825e-01 1.28898844e-02 -1.01057577e+00
-1.13740742e+00 9.20209706e-01 6.25108361e-01 -1.09859335e+00
-1.18368007e-01 -7.94827998e-01 -9.73584771e-01 1.44458723e+00
-1.39446998e+00 -1.08492148e+00 -3.39554638e-01 5.47762156e-01
-6.57504618e-01 2.20724419e-01 1.31345880e+00 6.74750030e-01
2.67356277e-01 -1.43041283e-01 5.47951221e-01 -4.81939256e-01
1.06593803e-01 -8.19902778e-01 1.82670094e-02 1.21315949e-01
-1.10387003e+00 2.11696520e-01 1.55648756e+00 -6.41917646e-01
-9.20037150e-01 -4.46294427e-01 9.36980069e-01 6.01518691e-01
4.84946311e-01 2.06889689e-01 -9.48458254e-01 1.70566976e-01
-9.15697739e-02 7.37292096e-02 7.37922013e-01 -3.29891890e-02
2.46827483e-01 2.44370788e-01 -1.68831968e+00 -1.31042272e-01
2.27693602e-01 1.02699891e-01 -2.66985923e-01 3.10532093e-01
-1.30070537e-01 -7.05738962e-01 -1.05910671e+00 6.27902269e-01
4.36400294e-01 -1.41564620e+00 1.17144156e+00 1.87611923e-01
1.24007881e-01 1.51808962e-01 -7.22471508e-04 -1.37884665e+00
-1.92372680e-01 -3.67985487e-01 2.95372665e-01 9.05660093e-01
7.51770213e-02 -1.24441433e+00 3.13943386e-01 5.40786207e-01
-3.84993792e-01 -6.89054072e-01 -1.00978315e+00 -5.40613115e-01
2.17011333e-01 -6.07860923e-01 5.08577764e-01 7.39711821e-01
3.11189163e-02 -2.17316806e-01 -1.25987023e-01 3.86825621e-01
7.52162576e-01 -7.77116492e-02 6.87113917e-03 -1.76927948e+00
-2.85543799e-01 1.36122346e-01 -4.57518160e-01 -5.11186600e-01
-2.15258703e-01 -3.47879440e-01 -1.40889257e-01 -1.93249273e+00
-3.62651229e-01 -1.17842984e+00 3.90545338e-01 -1.38090700e-01
5.62500954e-01 1.56551853e-01 -6.26681268e-01 -1.75710395e-01
2.97975540e-01 6.11428916e-02 1.28854859e+00 3.56143385e-01
-1.77215666e-01 6.72114253e-01 -2.44083896e-01 7.10542262e-01
7.25276947e-01 -5.89369357e-01 -1.09770790e-01 -3.18464227e-02
7.65606225e-01 3.57544065e-01 7.44667768e-01 -1.37353492e+00
-9.98671800e-02 -3.07890307e-03 2.14864120e-01 -9.15882885e-01
3.07655156e-01 -8.95346820e-01 8.41112733e-01 6.48223400e-01
6.98111653e-01 -1.20848857e-01 4.49346244e-01 -1.53839827e-01
6.72786031e-03 -1.01602089e+00 1.08595538e+00 -4.26302254e-01
-3.94286066e-01 -5.98529056e-02 -1.12787235e+00 -1.11907475e-01
1.00407434e+00 -5.52272081e-01 -3.01164061e-01 -2.36291617e-01
-8.59484792e-01 -4.01656985e-01 2.49778405e-01 -6.38497353e-01
5.61987996e-01 -9.52739537e-01 -8.47119451e-01 8.50883424e-02
-3.93623143e-01 2.82545690e-03 1.11592615e+00 9.42601562e-01
-1.44120002e+00 2.74244487e-01 -3.34680140e-01 -2.50624895e-01
-1.13193429e+00 -3.15525569e-02 8.45683813e-01 -6.31158054e-01
-8.38839471e-01 7.66617239e-01 -2.48382077e-01 -3.76542330e-01
-4.93266702e-01 -2.59299070e-01 -4.75966662e-01 7.89422169e-03
5.73111355e-01 6.67770326e-01 6.09677076e-01 -6.49575889e-01
-4.86263096e-01 6.42578185e-01 3.98090214e-01 -3.85069549e-02
1.85912192e+00 1.85633764e-01 -5.07964909e-01 -7.53525570e-02
9.27551806e-01 1.57249302e-01 -4.08223838e-01 2.92167455e-01
-5.61825216e-01 -2.66771135e-03 3.58101070e-01 -8.20516646e-01
-6.08283222e-01 4.65559155e-01 3.47404182e-01 7.43947566e-01
1.07583714e+00 -2.97705412e-01 6.99618280e-01 1.58525974e-01
5.89807212e-01 -8.48339021e-01 -6.72645926e-01 6.76486194e-01
8.19217324e-01 -4.70861226e-01 2.38360703e-01 -7.63455570e-01
-8.85172337e-02 1.42582929e+00 -5.96015081e-02 -1.25516742e-01
1.48777056e+00 8.47959459e-01 -1.23390824e-01 -3.27872962e-01
-1.62427604e-01 3.95882837e-02 -1.98891714e-01 1.10006595e+00
9.20926034e-01 1.43558145e-01 -7.67060816e-01 7.04731762e-01
-2.09579378e-01 3.59018117e-01 7.56757796e-01 1.43000710e+00
-6.31185770e-01 -1.08565736e+00 -1.02447021e+00 5.35186172e-01
-5.83696842e-01 -7.43677318e-02 4.98558015e-01 7.38491356e-01
5.38411066e-02 8.70187223e-01 -3.08842398e-02 3.53922725e-01
3.80641907e-01 -2.25227654e-01 7.43709743e-01 -3.87471169e-01
-3.88287514e-01 -2.03290991e-02 3.89643848e-01 2.00011104e-01
-7.01672316e-01 -6.47157609e-01 -1.34326935e+00 -7.43864298e-01
-3.43772918e-01 8.33412826e-01 8.37674439e-01 1.00517368e+00
-5.40800631e-01 5.64040244e-01 3.06745887e-01 -1.02507710e+00
-3.65033746e-01 -1.00410259e+00 -1.53218627e+00 -1.57254115e-01
1.93012378e-03 -7.95469105e-01 -5.66930830e-01 -3.68200898e-01] | [6.795579433441162, 1.6407396793365479] |
027d8f73-2015-45ab-b058-dec2ddadf45a | a-geometrically-aware-auto-encoder-for-multi | 2302.01616 | null | https://arxiv.org/abs/2302.01616v3 | https://arxiv.org/pdf/2302.01616v3.pdf | A geometrically aware auto-encoder for multi-texture synthesis | We propose an auto-encoder architecture for multi-texture synthesis. The approach relies on both a compact encoder accounting for second order neural statistics and a generator incorporating adaptive periodic content. Images are embedded in a compact and geometrically consistent latent space, where the texture representation and its spatial organisation are disentangled. Texture synthesis and interpolation tasks can be performed directly from these latent codes. Our experiments demonstrate that our model outperforms state-of-the-art feed-forward methods in terms of visual quality and various texture related metrics. | ['Sidonie Lefebvre', 'Yann Gousseau', 'Pierrick Chatillon'] | 2023-02-03 | null | null | null | null | ['texture-synthesis'] | ['computer-vision'] | [ 1.69275299e-01 1.30170763e-01 -2.28887454e-01 -2.21146286e-01
-7.65800118e-01 -2.52675325e-01 9.93595839e-01 -4.41293776e-01
2.03265011e-01 7.42965579e-01 6.29564404e-01 3.31421345e-01
7.11542964e-02 -8.17402601e-01 -8.67612958e-01 -1.02330303e+00
-3.32108960e-02 4.17670876e-01 1.13626756e-01 -2.17743292e-02
3.35063666e-01 2.36531317e-01 -1.62656569e+00 8.86342525e-01
5.55221498e-01 1.26420343e+00 4.13483173e-01 8.76654208e-01
2.19782010e-01 9.07827973e-01 3.79413180e-02 -1.96591809e-01
8.61501172e-02 -6.19771123e-01 -4.76267993e-01 2.87142932e-01
5.79899251e-01 -4.49566513e-01 -6.75108612e-01 7.10964084e-01
1.96426421e-01 -4.07653190e-02 1.00090218e+00 -5.07834733e-01
-1.49894714e+00 1.20072752e-01 -3.55293274e-01 -1.14362046e-01
9.55427065e-02 4.20881584e-02 1.14041126e+00 -1.22186899e+00
8.44735563e-01 1.43426025e+00 5.38177073e-01 4.38188076e-01
-1.95011771e+00 -3.16018313e-02 -3.10983360e-01 -8.14883485e-02
-1.22473860e+00 -9.61900413e-01 8.74854386e-01 -6.30900562e-01
8.01132083e-01 2.67250955e-01 7.57140219e-01 1.33166945e+00
8.48985016e-01 5.15243113e-01 1.40322959e+00 -6.18783593e-01
3.16111982e-01 -8.94807279e-02 -7.86301911e-01 9.93935406e-01
-8.51359442e-02 5.54229856e-01 -1.05643475e+00 7.77962059e-02
1.69287491e+00 -3.79389822e-01 1.44192809e-02 -6.79124653e-01
-1.30013752e+00 7.20216155e-01 2.40592480e-01 1.07573844e-01
-4.18294549e-01 6.50052607e-01 1.42236203e-01 3.63260478e-01
1.05209613e+00 4.16516572e-01 -2.46579684e-02 7.34310150e-02
-1.15072417e+00 2.53795147e-01 5.95298350e-01 8.62502635e-01
5.34991801e-01 4.12555426e-01 -3.65822732e-01 8.36840510e-01
5.05236626e-01 6.31916285e-01 2.59182990e-01 -1.19007134e+00
1.29320011e-01 -1.03018899e-02 -2.96756122e-02 -1.09710097e+00
4.03463334e-01 -5.36803938e-02 -1.14998639e+00 5.36675394e-01
7.75023410e-03 4.35541898e-01 -9.64847684e-01 1.43302786e+00
-2.11249992e-01 2.49975190e-01 -2.75847286e-01 7.49688506e-01
5.48486590e-01 6.45947397e-01 -2.18248233e-01 1.44829109e-01
1.24913013e+00 -9.83074605e-01 -9.62173998e-01 -2.96055842e-02
-2.25460276e-01 -9.87456739e-01 8.67749035e-01 3.05659473e-01
-1.49086511e+00 -6.46281898e-01 -1.17957938e+00 -4.71927494e-01
2.38047857e-02 5.10530174e-01 6.90635324e-01 1.99906826e-01
-1.32043397e+00 9.10920203e-01 -1.02547824e+00 3.77480537e-02
5.74087679e-01 1.66672841e-02 -5.50148249e-01 1.88848209e-02
-7.18379021e-01 7.28273928e-01 -4.00203653e-02 1.45296797e-01
-9.23324466e-01 -4.61072832e-01 -1.07103789e+00 -9.96680185e-02
-3.33220333e-01 -1.14675415e+00 8.46708477e-01 -7.30276644e-01
-2.06768703e+00 9.38495398e-01 -4.01477635e-01 -2.36471564e-01
4.37071621e-01 -1.95080161e-01 -3.87542367e-01 3.58332694e-01
3.52275893e-02 6.58383906e-01 1.40784132e+00 -1.25750268e+00
-1.14094801e-01 -3.49343084e-02 -5.31983078e-01 2.36190051e-01
-5.95452636e-02 -3.04407716e-01 -1.73877493e-01 -1.25579250e+00
2.35138625e-01 -9.33793128e-01 -1.58235624e-01 6.40101254e-01
-2.58984923e-01 5.61862767e-01 8.80228341e-01 -7.92428732e-01
1.06123805e+00 -2.01962447e+00 6.82960391e-01 4.13199849e-02
4.22368765e-01 -4.80724066e-01 -1.61450893e-01 4.13596004e-01
-1.98532175e-02 1.14755176e-01 3.91292535e-02 -7.94358909e-01
2.72617545e-02 2.79289037e-01 -6.26185477e-01 5.74130476e-01
4.11698043e-01 1.16087782e+00 -6.04948699e-01 -3.05499911e-01
5.10485053e-01 7.00189948e-01 -9.65475619e-01 3.47413987e-01
-5.08039653e-01 8.55577886e-01 -2.19506070e-01 5.41295290e-01
4.45268244e-01 -7.03082442e-01 5.49041927e-02 -4.46185350e-01
-1.55941844e-01 2.84326315e-01 -7.53299832e-01 2.03960276e+00
-6.78221464e-01 9.28136945e-01 -1.82691067e-01 -5.75366259e-01
8.60426664e-01 6.48929179e-01 2.17451915e-01 -6.94749832e-01
-1.31931171e-01 4.54611778e-01 -4.10139680e-01 -3.78628552e-01
5.94604433e-01 -1.79378614e-01 8.49501044e-02 4.16778862e-01
3.97381425e-01 -3.69703740e-01 -2.30979189e-01 -1.47410914e-01
7.70966887e-01 5.85986674e-01 1.18117198e-01 -6.12609327e-01
1.70557782e-01 -7.40466356e-01 1.63620245e-02 4.96834189e-01
4.37973410e-01 1.05209374e+00 4.15347934e-01 -7.91323543e-01
-1.86902201e+00 -1.40752983e+00 -2.90466845e-01 8.41639936e-01
-2.13144273e-01 -4.06360805e-01 -4.42025095e-01 8.36737752e-02
-7.77791366e-02 2.61736304e-01 -1.13310480e+00 1.07273482e-01
-4.60297495e-01 -2.91197687e-01 3.01857114e-01 3.69888753e-01
5.59599519e-01 -1.01844835e+00 -3.88266385e-01 1.85683906e-01
-2.68379837e-01 -1.01173043e+00 -5.67464173e-01 5.35534732e-02
-9.58105087e-01 -4.85598743e-01 -9.02339697e-01 -7.44876146e-01
6.22075558e-01 7.91009283e-04 1.21358609e+00 -2.27024347e-01
-4.05327320e-01 -4.16796803e-02 7.69698992e-02 2.17377558e-01
-5.26350200e-01 -1.76157311e-01 -1.17585152e-01 3.51142943e-01
-3.21304709e-01 -1.01788259e+00 -8.00829649e-01 2.90855527e-01
-8.48719299e-01 7.26368725e-01 6.39277995e-01 1.16894126e+00
8.98622334e-01 -2.89496869e-01 1.29178643e-01 -6.91987872e-01
3.88305455e-01 -2.19228312e-01 -4.79288638e-01 3.15262973e-01
-3.23112965e-01 7.34959424e-01 3.25911760e-01 -3.07906300e-01
-1.12969100e+00 6.45605773e-02 3.12501495e-03 -6.33329690e-01
-3.15364897e-02 2.02438384e-01 1.07328542e-01 -4.23918545e-01
7.00405300e-01 3.99141222e-01 1.39722839e-01 -5.56849539e-01
6.52135730e-01 1.27527967e-01 4.82324183e-01 -8.43892574e-01
4.53933597e-01 7.14957952e-01 2.41346091e-01 -9.16834414e-01
-6.79878175e-01 1.43582731e-01 -8.76094997e-01 -1.08679302e-01
1.03291738e+00 -1.07839227e+00 -4.70941097e-01 6.26310527e-01
-1.26196754e+00 -5.46547472e-01 -5.99209845e-01 1.92207679e-01
-1.26080811e+00 -4.56899917e-03 -1.03843296e+00 -5.25461555e-01
1.93208158e-02 -1.10871029e+00 1.69468570e+00 -2.83642858e-01
-2.50911534e-01 -1.30795896e+00 2.90067464e-01 -2.09636033e-01
8.11491132e-01 4.53987777e-01 9.52079833e-01 4.82938856e-01
-1.06208646e+00 1.78023204e-01 -2.53050447e-01 4.77348752e-02
2.40873188e-01 -3.98212932e-02 -1.17466104e+00 -2.41243050e-01
7.85318464e-02 -5.93390167e-01 1.26014042e+00 6.62027657e-01
1.32429433e+00 -8.81267726e-01 -1.14099264e-01 1.04982460e+00
1.56108916e+00 -4.61774439e-01 7.40119159e-01 2.92372750e-03
7.30787456e-01 4.43452597e-01 -4.09430742e-01 3.15216571e-01
2.00510323e-01 7.43793607e-01 1.06537268e-01 -1.47533447e-01
-6.84450507e-01 -4.38904315e-01 1.55828238e-01 1.12940598e+00
-5.08454144e-01 -1.24593571e-01 -4.80032921e-01 4.16400582e-01
-1.82697284e+00 -1.13422751e+00 1.14572056e-01 1.87790644e+00
9.50069189e-01 -1.62053276e-02 -3.27608198e-01 -2.91085280e-02
3.43790621e-01 6.23360336e-01 -4.18109536e-01 -4.76013809e-01
-5.36731780e-01 5.30813098e-01 3.64163578e-01 6.93169594e-01
-1.13048732e+00 8.81559193e-01 8.14594460e+00 9.97589111e-01
-1.06593013e+00 6.68330640e-02 9.69296992e-01 -4.80033830e-03
-6.81196809e-01 -3.42395762e-03 -2.90409267e-01 3.73342752e-01
9.06297982e-01 2.79096942e-02 6.71707749e-01 4.62574899e-01
-3.13304588e-02 -8.20853487e-02 -1.16228521e+00 9.42384720e-01
1.62646160e-01 -2.14291000e+00 2.94959545e-01 3.92319828e-01
1.14054036e+00 -1.32024720e-01 7.95807779e-01 -4.19783086e-01
3.27779055e-01 -1.35118961e+00 1.12523592e+00 1.14022005e+00
1.67014337e+00 -2.60099560e-01 1.03495404e-01 -2.14684859e-01
-1.36477900e+00 3.26042771e-01 -2.97380865e-01 2.09406137e-01
1.29659206e-01 5.91730416e-01 6.32004589e-02 4.44391549e-01
5.85507035e-01 1.04526079e+00 -3.09895694e-01 5.85635066e-01
-2.34678760e-01 3.09245735e-01 -2.55707907e-03 4.47002888e-01
-1.91507023e-02 -1.30194932e-01 3.54453087e-01 1.08276951e+00
5.06339788e-01 9.06157643e-02 -1.74242035e-01 1.36497974e+00
1.87670097e-01 -2.77554095e-01 -8.08009565e-01 -1.17879406e-01
1.28990561e-01 9.61576045e-01 -6.59460187e-01 -2.12510481e-01
-2.14070708e-01 1.39992070e+00 3.53364795e-01 6.25931680e-01
-3.57153833e-01 5.06697595e-03 5.98306835e-01 2.11885586e-01
6.58861876e-01 -5.61700642e-01 -6.42909467e-01 -1.46626759e+00
-4.42603230e-02 -6.93229556e-01 -3.48757923e-01 -9.96164203e-01
-1.41692328e+00 8.43313456e-01 -3.34496766e-01 -1.28157640e+00
-3.96757901e-01 -6.39720082e-01 -1.42114788e-01 1.19357181e+00
-1.14971101e+00 -1.53124607e+00 6.64985850e-02 4.81391221e-01
3.96623969e-01 -3.64292592e-01 1.41313505e+00 9.70729347e-03
2.41795182e-02 4.88121241e-01 3.58827710e-01 -2.29549438e-01
4.22283709e-01 -1.15808403e+00 7.69091487e-01 5.28914332e-01
3.14140975e-01 4.84999090e-01 7.24828660e-01 -6.28525198e-01
-1.30404663e+00 -9.01114345e-01 1.02435994e+00 -4.63370174e-01
6.91363692e-01 -8.18731606e-01 -7.76217818e-01 5.89842975e-01
4.30123389e-01 4.90016609e-01 5.65335691e-01 -4.01339792e-02
-5.46779394e-01 2.08841458e-01 -8.42624485e-01 8.10815871e-01
9.21287000e-01 -1.15629125e+00 -9.59713459e-02 2.15421036e-01
5.24384499e-01 -5.95635355e-01 -9.67956781e-01 9.34742168e-02
1.02018034e+00 -1.04276133e+00 9.50942397e-01 -4.64714110e-01
1.26545453e+00 9.37951133e-02 -4.31082249e-01 -1.24741852e+00
-8.42258275e-01 -8.15262616e-01 -5.77202260e-01 5.51116049e-01
1.84710369e-01 -1.94411471e-01 8.31318736e-01 2.06449002e-01
2.57040076e-02 -9.26242948e-01 -9.28199232e-01 -6.03251338e-01
-4.21739481e-02 -5.09762503e-02 3.28489572e-01 6.03303730e-01
-3.07841212e-01 2.85401791e-01 -1.00526357e+00 -1.99894100e-01
7.71987021e-01 2.55223781e-01 2.33186722e-01 -9.75730360e-01
-5.71587324e-01 -4.98773634e-01 -3.51730645e-01 -1.32996643e+00
1.50746657e-02 -8.06653857e-01 1.04908720e-01 -1.24328303e+00
1.94105327e-01 -3.61692160e-01 1.94459911e-02 1.71621650e-01
3.49632442e-01 6.15582585e-01 -2.07962841e-01 4.99522716e-01
-4.68654007e-01 1.16201317e+00 1.63985562e+00 2.03671649e-01
2.91762143e-01 -4.83643800e-01 -2.73299396e-01 5.82166970e-01
3.01068306e-01 -2.30803207e-01 -3.10925215e-01 -6.85863435e-01
4.12850231e-01 3.96287680e-01 6.82067156e-01 -9.11403716e-01
5.07175587e-02 -3.60912830e-01 7.74831772e-01 -1.25217244e-01
6.91354930e-01 -3.94850314e-01 3.99870813e-01 5.96637763e-02
-6.59859121e-01 1.06807701e-01 1.92298889e-01 7.69071400e-01
-3.66951019e-01 5.58392406e-01 8.27060401e-01 -8.67106840e-02
-3.74067456e-01 5.18643558e-01 -4.90924358e-01 -1.75566822e-01
4.88989234e-01 -3.14103216e-01 -3.68001461e-01 -4.22768027e-01
-9.52922940e-01 -6.56897485e-01 7.41157472e-01 3.89671028e-01
8.99607360e-01 -2.00316024e+00 -8.02342176e-01 8.36011589e-01
1.73047222e-02 -3.23515743e-01 2.88440853e-01 4.56289560e-01
-7.96866596e-01 5.85415721e-01 -7.28687525e-01 -5.89784801e-01
-7.10941434e-01 2.15591848e-01 3.74509364e-01 -2.88942873e-01
-8.95628810e-01 9.01110888e-01 4.34650362e-01 7.04819383e-03
-1.54328272e-01 -3.35402280e-01 2.65412331e-01 -3.98952991e-01
4.20157075e-01 -5.02845943e-02 -1.07587926e-01 -7.22384810e-01
9.91449654e-02 7.34802842e-01 2.73452640e-01 -7.18136966e-01
1.46543372e+00 -1.62587374e-01 -3.33031088e-01 7.87488639e-01
1.26302028e+00 3.82311232e-02 -1.82241714e+00 -3.71762097e-01
-3.11747819e-01 -8.16838861e-01 1.35308117e-01 -5.60386777e-01
-8.85366082e-01 9.95253205e-01 4.47586000e-01 4.77945544e-02
9.93563116e-01 -1.12617603e-02 5.82562268e-01 -7.60231018e-02
4.61537570e-01 -7.76256859e-01 5.86228669e-01 6.26266479e-01
1.50244403e+00 -1.03986442e+00 5.77719091e-03 -3.60419869e-01
-4.27547455e-01 1.15664935e+00 8.41880441e-02 -8.36582005e-01
1.04860306e+00 3.93608779e-01 -3.10494572e-01 -3.15052062e-01
-1.33641005e+00 3.13471973e-01 9.95548129e-01 4.55891937e-01
7.48399258e-01 1.38593018e-01 1.83587208e-01 1.00458711e-02
-3.42380553e-01 -1.11459881e-01 1.37717217e-01 4.75403041e-01
-1.79459706e-01 -1.11770761e+00 3.14424490e-03 3.69118750e-01
-4.07095522e-01 -3.55546504e-01 7.74023309e-02 3.20445180e-01
2.07756888e-02 2.77037203e-01 2.87916958e-01 -3.65127504e-01
-1.88387811e-01 -1.10340700e-01 1.04619968e+00 -5.04073977e-01
1.22520998e-01 3.85247469e-01 -2.75315102e-02 -9.13523912e-01
-3.16616595e-01 -5.97656190e-01 -3.35973620e-01 -3.80849630e-01
-8.99665803e-03 -2.62084156e-01 4.74498749e-01 6.41710281e-01
4.21918750e-01 6.90021634e-01 4.98850465e-01 -1.26862490e+00
-1.80825949e-01 -7.80025065e-01 -7.38823056e-01 1.12655856e-01
7.56700993e-01 -7.01480269e-01 -1.94796041e-01 6.44903421e-01] | [11.485966682434082, -0.504715085029602] |
f336c09b-ae2d-4146-99f1-eaee0ff13437 | dbpedia-abstracts-a-large-scale-open | null | null | https://aclanthology.org/L16-1532 | https://aclanthology.org/L16-1532.pdf | DBpedia Abstracts: A Large-Scale, Open, Multilingual NLP Training Corpus | The ever increasing importance of machine learning in Natural Language Processing is accompanied by an equally increasing need in large-scale training and evaluation corpora. Due to its size, its openness and relative quality, the Wikipedia has already been a source of such data, but on a limited scale. This paper introduces the DBpedia Abstract Corpus, a large-scale, open corpus of annotated Wikipedia texts in six languages, featuring over 11 million texts and over 97 million entity links. The properties of the Wikipedia texts are being described, as well as the corpus creation process, its format and interesting use-cases, like Named Entity Linking training and evaluation. | ['Martin Br{\\"u}mmer', 'Sebastian Hellmann', 'Milan Dojchinovski'] | 2016-05-01 | dbpedia-abstracts-a-large-scale-open-1 | https://aclanthology.org/L16-1532 | https://aclanthology.org/L16-1532.pdf | lrec-2016-5 | ['multilingual-nlp'] | ['natural-language-processing'] | [-5.81127644e-01 6.21569395e-01 -5.86331964e-01 -1.65739655e-01
-5.92833042e-01 -9.81024027e-01 8.95704389e-01 8.75102401e-01
-9.53318000e-01 1.33367777e+00 5.43167710e-01 -2.07800359e-01
-1.43483102e-01 -8.14870119e-01 -6.84521556e-01 1.40406698e-01
-4.59983796e-01 1.13040960e+00 5.39068341e-01 -5.46535492e-01
3.27222124e-02 -4.03707176e-02 -1.30727148e+00 2.07366198e-01
1.30511928e+00 4.62314457e-01 1.32682800e-01 3.99538159e-01
-8.47238243e-01 7.71930695e-01 -7.42163360e-01 -1.31574249e+00
-1.67647004e-01 -2.85910964e-02 -1.30296326e+00 -3.97139966e-01
5.53084612e-01 2.32836217e-01 -2.16899604e-01 8.48083019e-01
6.47979856e-01 -3.62099707e-02 4.20292199e-01 -1.09672582e+00
-6.55068398e-01 1.08109951e+00 -1.57403320e-01 2.08941042e-01
6.34987891e-01 -3.48512471e-01 1.29320073e+00 -6.00706816e-01
1.51160550e+00 8.34361553e-01 6.89818263e-01 4.59931284e-01
-8.95711064e-01 -2.45860726e-01 -5.21343648e-01 -4.17912230e-02
-1.19954443e+00 -3.34491879e-01 2.43844315e-01 -5.33736169e-01
1.51530266e+00 -8.80109444e-02 4.41122234e-01 1.02010775e+00
-2.66299605e-01 4.14986879e-01 6.54968619e-01 -7.64398813e-01
-2.33595580e-01 5.16557515e-01 2.38675356e-01 5.70421934e-01
8.27313244e-01 -4.38631922e-01 -4.65859286e-02 -2.07335770e-01
2.19002619e-01 -6.96216822e-01 -1.03788242e-01 -4.32554752e-01
-1.33316445e+00 6.49269104e-01 2.38861322e-01 5.20669460e-01
-3.38002652e-01 -2.68262088e-01 1.05766165e+00 3.40500295e-01
3.35495889e-01 9.06413436e-01 -9.63619351e-01 -4.04145479e-01
-5.48099995e-01 3.45637202e-01 1.44744480e+00 1.26846719e+00
7.10046649e-01 -4.51460779e-01 3.17108929e-01 1.15395784e+00
1.76936939e-01 4.21852678e-01 4.31312263e-01 -6.83921099e-01
1.25729418e+00 9.48547542e-01 3.12628448e-01 -9.56564546e-01
-5.62420189e-01 -1.80440888e-01 -3.20273370e-01 -2.47228533e-01
7.37237215e-01 -3.26747864e-01 -2.42041066e-01 1.71038103e+00
4.84809965e-01 -5.63506365e-01 4.72471118e-01 2.52383679e-01
1.33643973e+00 5.26512504e-01 4.80748236e-01 -2.83920854e-01
1.47057211e+00 -6.29555225e-01 -8.35831821e-01 -1.52359188e-01
8.73557687e-01 -8.08270216e-01 7.82220125e-01 5.40303402e-02
-1.15183949e+00 -1.29060060e-01 -6.88455462e-01 -3.51286411e-01
-1.13128328e+00 -1.39389157e-01 6.92762971e-01 3.82789016e-01
-5.97503960e-01 3.58515680e-01 -2.58863091e-01 -9.54879642e-01
2.90855348e-01 1.89960133e-02 -9.15046275e-01 -1.54569805e-01
-1.68240607e+00 1.49740350e+00 1.30217946e+00 -4.24684644e-01
1.08322456e-01 -7.04340160e-01 -9.06535327e-01 -1.90048337e-01
6.35851502e-01 -4.55613106e-01 9.66506004e-01 -4.76770282e-01
-6.15790188e-01 1.23372412e+00 3.95962059e-01 -5.72038472e-01
3.49401355e-01 -2.51941442e-01 -1.00222421e+00 1.14510417e-01
5.10166109e-01 5.30287623e-01 -5.42283133e-02 -8.47788572e-01
-8.67126584e-01 -2.68930554e-01 2.88491786e-01 1.57224834e-01
-5.43559968e-01 4.37157810e-01 -5.50070465e-01 -5.39361000e-01
-4.16570842e-01 -4.36776400e-01 -8.13659206e-02 -5.47056317e-01
-3.13239664e-01 -5.51119983e-01 2.71083832e-01 -9.69750404e-01
1.70108593e+00 -1.87902343e+00 -4.47156802e-02 5.63378353e-03
3.24905694e-01 3.13247830e-01 1.04934834e-01 9.84470606e-01
8.97789523e-02 4.57861036e-01 -1.50286466e-01 2.06372052e-01
2.64892280e-01 3.79468918e-01 -1.53838163e-02 -4.34290711e-03
1.96039438e-01 8.22696328e-01 -1.27742815e+00 -1.02122009e+00
-2.16060668e-01 1.87421948e-01 -2.48070732e-01 -2.26810113e-01
-4.33647633e-01 -6.85788170e-02 -4.71010745e-01 5.45052648e-01
1.24763802e-01 -1.17256679e-01 3.57634842e-01 -2.51284838e-01
-4.50920433e-01 6.62205815e-01 -1.16085207e+00 1.59084487e+00
-4.01821077e-01 7.63732731e-01 -5.42641282e-02 -6.31009996e-01
7.63370514e-01 6.25343025e-01 3.61046284e-01 -7.36403942e-01
-2.27619216e-01 5.58565080e-01 -8.99102539e-04 -1.00642991e+00
8.14172328e-01 1.04093656e-01 -5.50497413e-01 3.44531238e-01
3.75066787e-01 1.10882118e-01 1.24980307e+00 5.70513070e-01
9.80917573e-01 1.98662490e-01 7.34586298e-01 -2.39999264e-01
4.81784970e-01 6.39971018e-01 5.63344955e-01 1.80598542e-01
6.76350296e-02 5.16088679e-02 5.02184629e-01 -5.03101945e-01
-1.61275578e+00 -6.70269966e-01 -7.32083917e-01 8.35544050e-01
-2.61592865e-01 -6.04295731e-01 -5.86311221e-01 -7.94175684e-01
2.80194551e-01 6.90073609e-01 -4.89272296e-01 4.19804573e-01
-8.24310899e-01 -4.17012542e-01 1.02119756e+00 2.97973543e-01
4.49580848e-01 -1.41907144e+00 -1.66409999e-01 4.60995018e-01
-4.82503951e-01 -1.28846443e+00 6.18651696e-02 1.16413459e-01
-5.83968282e-01 -1.34349251e+00 -3.67040128e-01 -1.03993130e+00
3.67666066e-01 -4.97186899e-01 1.84580684e+00 -1.72754228e-01
-2.99705446e-01 2.06195474e-01 -4.57626790e-01 -5.44473886e-01
-7.35609055e-01 3.71798277e-01 -1.02582589e-01 -8.82851422e-01
5.55090070e-01 -2.46998608e-01 1.58764333e-01 -4.33676913e-02
-8.52274776e-01 -4.56813216e-01 5.29789448e-01 7.09865749e-01
2.24679604e-01 7.71135837e-02 8.94711614e-01 -1.38114786e+00
5.62722802e-01 -1.00481129e+00 -3.22891891e-01 6.63846135e-01
-5.82792580e-01 6.04413077e-02 5.39143324e-01 -5.25866970e-02
-1.15729845e+00 -3.66141915e-01 -1.63475364e-01 7.14496613e-01
-3.65078241e-01 1.08808947e+00 -3.12767595e-01 4.25319254e-01
9.17319059e-01 -3.56533229e-01 -2.76680470e-01 -7.05691576e-01
6.08922124e-01 6.90293908e-01 5.11799932e-01 -7.48027742e-01
7.34741747e-01 -2.08788201e-01 -4.71706420e-01 -1.07369375e+00
-8.57234359e-01 -4.47810113e-01 -9.33944464e-01 -1.93783082e-02
8.78277481e-01 -8.58684003e-01 -3.25272202e-01 -5.43596335e-02
-1.17929196e+00 -4.17575873e-02 -4.03684139e-01 3.89689177e-01
-5.24915904e-02 2.10603878e-01 -4.33617353e-01 -3.21792871e-01
-3.62422585e-01 -3.35404843e-01 3.47535670e-01 5.92380464e-02
-4.33516383e-01 -1.38530302e+00 4.96999919e-01 3.94382149e-01
9.60362405e-02 6.88085318e-01 1.18129492e+00 -1.20800447e+00
1.38870468e-02 -5.66410661e-01 -8.53517801e-02 2.15202160e-02
-1.61000699e-01 3.14919412e-01 -4.58287507e-01 1.80192571e-02
-8.59549403e-01 -6.24764204e-01 2.63626456e-01 -1.16505623e-01
2.39931792e-01 -5.57497859e-01 -3.89612168e-01 -3.99674512e-02
1.61998940e+00 -4.87469099e-02 6.55053258e-01 1.04664135e+00
5.89726210e-01 8.03772926e-01 6.06547177e-01 2.11911723e-01
7.09285796e-01 4.95067358e-01 2.27075338e-01 1.60714343e-01
-1.11953005e-01 -3.91191900e-01 -7.77133852e-02 1.16138923e+00
-2.31437892e-01 -3.00167173e-01 -1.41933072e+00 7.10008383e-01
-1.56274271e+00 -1.47355485e+00 -3.69233191e-01 2.15868282e+00
1.35089254e+00 3.68121713e-01 2.41941541e-01 -1.22634083e-01
7.89905608e-01 4.25297059e-02 -9.29949954e-02 -2.52572298e-01
-6.12119436e-01 -1.29736029e-02 4.02220309e-01 1.88461959e-01
-1.15479171e+00 7.78330684e-01 6.47913074e+00 6.26112342e-01
-5.10812640e-01 1.58693790e-01 -2.11462006e-02 2.87919760e-01
-2.14468807e-01 8.03294964e-03 -1.09298635e+00 5.89502215e-01
1.09201121e+00 -6.22043967e-01 1.41874313e-01 7.61425495e-01
-2.57581115e-01 9.96690765e-02 -8.42457771e-01 3.50773931e-01
-4.30005901e-02 -1.53852463e+00 -1.04069151e-02 5.78045547e-02
5.89381337e-01 4.53936309e-01 -6.24241114e-01 5.29545903e-01
4.86499816e-01 -5.93762517e-01 7.59389758e-01 2.51541525e-01
8.44008803e-01 -7.36857653e-01 1.03704643e+00 3.35129231e-01
-9.95061755e-01 -8.32536742e-02 -4.12790239e-01 1.72142982e-01
1.77803755e-01 7.84243047e-01 -5.28414309e-01 6.74581885e-01
7.97500730e-01 6.69892192e-01 -7.84457564e-01 1.27405369e+00
-5.53348362e-01 5.31966388e-01 -3.47847641e-01 -4.08619791e-01
1.61725104e-01 4.20505144e-02 5.58157861e-01 1.70571613e+00
-9.67182498e-03 -1.93333589e-02 1.58995509e-01 3.31446320e-01
-5.90532660e-01 6.46092057e-01 -9.01074648e-01 -4.94645983e-01
9.98882949e-01 1.47018409e+00 -4.42326248e-01 -4.21133906e-01
-7.99126208e-01 2.73556352e-01 8.41829717e-01 1.37344584e-01
-7.25785136e-01 -1.08994675e+00 3.06667387e-01 2.42454797e-01
1.40755787e-01 -1.55335769e-01 2.69286752e-01 -1.02278256e+00
3.43545258e-01 -8.01111281e-01 9.10674691e-01 -5.85408866e-01
-1.53576028e+00 7.44314671e-01 5.17901815e-02 -1.09019887e+00
-2.25318462e-01 -4.76922512e-01 -2.50063956e-01 6.55894458e-01
-1.39257717e+00 -1.06767941e+00 1.16842389e-02 3.39414418e-01
1.59766525e-01 -3.64972830e-01 8.92491281e-01 9.43830609e-01
-4.93473232e-01 4.05259788e-01 4.50544924e-01 7.35966206e-01
1.07984066e+00 -1.51617455e+00 4.12902206e-01 3.76649320e-01
1.67423546e-01 8.28379035e-01 7.21272767e-01 -7.95936465e-01
-1.12846005e+00 -9.90337968e-01 1.79918420e+00 -7.71930695e-01
1.41372967e+00 -2.85407871e-01 -1.13751602e+00 7.50028610e-01
4.63789761e-01 -1.22281022e-01 7.89250195e-01 4.64370221e-01
-4.96791363e-01 -1.06663890e-02 -1.19475293e+00 2.76108086e-01
8.67013872e-01 -3.35099339e-01 -1.34614778e+00 5.53029537e-01
5.63799500e-01 -4.78362113e-01 -1.42898893e+00 2.56695658e-01
3.05639505e-01 -3.96225095e-01 8.97898793e-01 -9.57054377e-01
7.00111270e-01 4.23748326e-03 1.07286796e-01 -1.17598844e+00
-1.81040958e-01 -4.59805369e-01 -8.88415053e-02 1.98475254e+00
1.21301830e+00 -6.68578506e-01 4.54383343e-01 7.87327647e-01
-2.91233003e-01 -3.30411226e-01 -7.57122040e-01 -8.53170395e-01
2.52771139e-01 -2.04325199e-01 4.85760421e-01 1.51184797e+00
7.12508202e-01 7.70974755e-01 1.48708254e-01 -1.51724398e-01
5.16001821e-01 -1.12000905e-01 5.86326003e-01 -1.72559929e+00
-1.28764892e-02 -4.29820001e-01 -3.77334148e-01 -3.02624106e-01
7.00883865e-02 -1.07604825e+00 -2.15706691e-01 -1.97788930e+00
2.11202949e-01 -7.36442685e-01 2.39435613e-01 4.77849782e-01
-1.78827107e-01 4.72072475e-02 -2.60713995e-01 3.94537240e-01
-9.11021769e-01 -6.77012354e-02 8.82962704e-01 1.04115695e-01
-7.54087418e-02 -5.40466189e-01 -3.35532904e-01 7.60510445e-01
7.15792656e-01 -4.17980731e-01 8.94196611e-03 -5.06508291e-01
9.40151334e-01 -3.04238379e-01 -2.07518816e-01 -8.18688691e-01
1.80947542e-01 9.40449387e-02 1.84221834e-01 -4.28860039e-01
-2.15193689e-01 -6.50094867e-01 1.63416907e-01 2.43262753e-01
-5.05688071e-01 3.22410107e-01 2.92046607e-01 2.15611547e-01
-5.38661361e-01 -5.91903090e-01 5.37579775e-01 -3.17911983e-01
-1.14088547e+00 7.03552887e-02 -1.43885940e-01 1.11116326e+00
1.09174180e+00 -9.69278961e-02 -8.27364028e-01 1.66103542e-01
-6.29923105e-01 3.58722478e-01 4.34938043e-01 4.90405262e-01
-1.65077657e-01 -1.31216669e+00 -9.23843503e-01 -3.53952050e-01
5.12196422e-01 -1.24176413e-01 -2.05941632e-01 4.96899903e-01
-6.56214535e-01 4.22750086e-01 -4.33093905e-01 4.67878580e-02
-9.40081120e-01 6.46307230e-01 -2.06615061e-01 -5.58000863e-01
-7.56876230e-01 4.09532905e-01 -6.95717096e-01 -8.03968966e-01
2.42893636e-01 3.85769933e-01 -7.77279019e-01 4.58055258e-01
4.39761698e-01 5.18160343e-01 1.06015667e-01 -7.09226310e-01
-2.66396701e-01 1.52488858e-01 9.44913030e-02 1.43287718e-01
1.68018258e+00 -1.59733325e-01 -5.54759860e-01 5.42543232e-01
1.07942879e+00 3.23182195e-01 -3.68519425e-01 -2.64777482e-01
8.10032964e-01 -1.49599230e-03 -4.05816615e-01 -1.13169336e+00
-6.27341270e-01 5.88789105e-01 -2.33869076e-01 6.84488773e-01
3.68981689e-01 4.59520459e-01 6.90966249e-01 7.34128475e-01
4.89129305e-01 -1.50686884e+00 -4.10813630e-01 9.84611750e-01
7.29537249e-01 -1.38064194e+00 2.02752888e-01 -3.59130681e-01
-6.50004923e-01 1.09367073e+00 5.30209541e-01 2.98829168e-01
4.63753104e-01 3.68680269e-01 -2.63432879e-02 -4.25204933e-01
-5.64401150e-01 -2.10995629e-01 1.03745408e-01 7.74117947e-01
9.09270167e-01 -2.71694511e-01 -8.21279466e-01 4.62013662e-01
-2.43878216e-01 -1.48789555e-01 3.76862824e-01 8.75650764e-01
-5.10765493e-01 -1.25744510e+00 2.25720163e-02 4.96266395e-01
-8.01987290e-01 -2.76088864e-01 -4.46809739e-01 1.54156733e+00
1.75305098e-01 6.03866100e-01 1.17373094e-01 2.74479270e-01
7.37035871e-01 1.91702485e-01 3.10108840e-01 -7.64936209e-01
-8.25526416e-01 -5.15186489e-01 1.10899866e+00 -1.85468674e-01
-6.53910577e-01 -7.77303636e-01 -1.56032896e+00 -3.76793891e-01
-3.61216545e-01 7.47647583e-01 7.56920278e-01 7.48431087e-01
2.96676099e-01 1.14380978e-01 9.58269835e-02 -1.57678097e-01
-2.30146393e-01 -1.02483964e+00 -5.42827606e-01 6.15609884e-01
-3.34782511e-01 -6.04370475e-01 -1.83237076e-01 1.71408236e-01] | [9.475006103515625, 8.887776374816895] |
1056c8fc-44cd-46c3-9057-0b916f3c3db5 | reporting-existing-datasets-for-automatic | 2306.12292 | null | https://arxiv.org/abs/2306.12292v1 | https://arxiv.org/pdf/2306.12292v1.pdf | Reporting existing datasets for automatic epilepsy diagnosis and seizure detection | More than 50 million individuals are affected by epilepsy, a chronic neurological disorder characterized by unprovoked, recurring seizures and psychological symptoms. Researchers are working to automatically detect or predict epileptic episodes through Electroencephalography (EEG) signal analysis, and machine, and deep learning methods. Good quality, open-source, and free EEG data acts as a catalyst in this ongoing battle to manage this disease. This article presents 40+ publicly available EEG datasets for adult and pediatric human populations from 2001-2023. A comparative analysis and discussion on open and private EEG datasets have been done based on objective parameters in this domain. Bonn and CHB-MIT remain the benchmark datasets used for the automatic detection of epileptic and seizure EEG signals. Meta-data has also been released for large EEG data like CHB-MIT. This article will be updated every year to report the progress and changing trends in the development of EEG datasets in this field. | ['Nidhi Goel', 'Sakshi Tiwari', 'Palak Handa'] | 2023-06-21 | null | null | null | null | ['seizure-detection'] | ['medical'] | [-3.17575008e-01 -2.50064820e-01 3.68268073e-01 -4.80275661e-01
-7.39375174e-01 -9.34751853e-02 1.93848744e-01 1.41821176e-01
-3.66269112e-01 1.23974609e+00 3.64113480e-01 1.38434440e-01
-3.47856969e-01 -4.74897653e-01 -2.19816461e-01 -6.99730217e-01
-8.83559942e-01 5.35383165e-01 -4.32056576e-01 4.28941138e-02
1.42334461e-01 3.47252816e-01 -1.20916867e+00 3.94401670e-01
8.99629056e-01 8.98006082e-01 2.86123514e-01 4.46701139e-01
5.46700239e-01 3.54820460e-01 -8.48705411e-01 -2.15895265e-01
-2.00723961e-01 -4.32274282e-01 -6.50809467e-01 -3.35573375e-01
-4.52141821e-01 -1.62282974e-01 -3.20193768e-01 7.87554979e-01
1.34270668e+00 -4.25791413e-01 5.79079270e-01 -1.16631973e+00
-3.22046459e-01 5.75800121e-01 -4.36288714e-01 7.23601580e-01
7.04463601e-01 8.33901539e-02 4.11529869e-01 -5.10250092e-01
3.35574687e-01 2.67043501e-01 7.86258817e-01 5.59299707e-01
-1.00779438e+00 -1.23492861e+00 -3.25183779e-01 7.11521685e-01
-1.65632534e+00 -1.11968055e-01 4.69996154e-01 -5.96935868e-01
1.35754800e+00 3.49569350e-01 1.61665249e+00 1.72699142e+00
7.73627877e-01 3.53036076e-01 1.26007032e+00 -6.10457286e-02
3.13956410e-01 -1.62339911e-01 2.73741633e-01 -1.31480321e-01
2.78115720e-01 1.86992720e-01 -1.10777557e+00 -5.64386964e-01
3.66333663e-01 -1.98947355e-01 -8.93254101e-01 4.53585714e-01
-1.23494565e+00 4.10370767e-01 -3.99195626e-02 7.48526573e-01
-8.95690441e-01 -1.88167974e-01 3.13598007e-01 5.52588880e-01
6.84050858e-01 5.08115292e-01 -7.28025317e-01 -9.15262938e-01
-1.17579806e+00 2.14547515e-01 6.97982550e-01 6.24926269e-01
3.23166311e-01 -1.29865538e-02 1.04464710e-01 9.13617790e-01
-4.04832698e-02 4.09875304e-01 9.44602430e-01 -2.98806578e-01
8.37895274e-02 4.26167905e-01 -1.27532616e-01 -8.09507072e-01
-9.21884656e-01 -6.50468171e-01 -1.04235756e+00 -3.59335989e-01
-4.35754210e-01 -6.25254273e-01 -6.65341139e-01 1.36507297e+00
-4.42189634e-01 3.77453774e-01 -1.14558652e-01 6.76425397e-01
8.44893932e-01 2.54959732e-01 -1.38912052e-01 -1.79611772e-01
1.28002632e+00 -8.66638348e-02 -8.19612861e-01 -2.79391254e-03
3.12070221e-01 -2.06648052e-01 3.45876902e-01 1.14359462e+00
-1.12022197e+00 1.11050859e-01 -9.23345804e-01 4.67817664e-01
-4.87972319e-01 -9.89992470e-02 6.85282409e-01 7.70772815e-01
-1.36397219e+00 4.46603298e-01 -1.00161278e+00 -4.36226338e-01
9.06531751e-01 8.28675568e-01 -6.90257668e-01 3.29802990e-01
-1.38348520e+00 1.15906107e+00 3.78579527e-01 -1.93197340e-01
-9.17075098e-01 -8.78696382e-01 -3.11232090e-01 -4.46618833e-02
-6.03215694e-01 -4.06134546e-01 7.23187804e-01 -5.25224149e-01
-9.98749733e-01 1.06497848e+00 2.43785456e-01 -7.25431800e-01
2.50542581e-01 -8.36222619e-02 -8.51820529e-01 9.54167992e-02
-4.13260572e-02 7.15613544e-01 2.91699767e-01 -3.48288834e-01
-7.25099862e-01 -7.79540539e-01 -6.71340227e-01 -8.76084790e-02
-2.80393690e-01 6.41618907e-01 1.10115804e-01 -8.15141141e-01
-1.93554983e-01 -6.53017998e-01 1.02748983e-01 -9.07791853e-01
-2.51539052e-01 -2.87908226e-01 2.77132988e-01 -9.28818047e-01
1.02205455e+00 -2.06907606e+00 3.78316734e-03 5.04253097e-02
4.28872019e-01 -2.09981650e-01 1.09877706e-01 3.62645358e-01
-6.37992144e-01 -1.51063457e-01 -1.13748387e-01 -4.97197546e-02
-4.48414907e-02 -2.56032884e-01 -1.30481705e-01 6.17442727e-01
5.15568443e-03 8.95600021e-01 -6.66965842e-01 2.38167465e-01
1.11578725e-01 5.91724515e-01 -1.66294247e-01 1.97325811e-01
6.29027605e-01 8.35270047e-01 -1.09980136e-01 6.53703868e-01
5.23773491e-01 -2.16273323e-01 -1.59712523e-01 2.69647002e-01
-1.31106675e-01 3.96354884e-01 -4.88086849e-01 1.81633759e+00
1.46548878e-02 1.03597724e+00 -9.82483476e-02 -9.75853086e-01
8.17154408e-01 8.70166659e-01 8.59470308e-01 -6.68349028e-01
4.27872121e-01 3.94002974e-01 3.10976326e-01 -4.93230224e-01
-3.54326665e-01 7.67526627e-02 1.34264022e-01 4.03074473e-01
4.11759645e-01 -5.09676300e-02 1.50212482e-01 -1.97913244e-01
1.59748554e+00 -3.21623087e-01 3.93863142e-01 -6.74352407e-01
2.00523496e-01 -3.33697230e-01 6.87941134e-01 2.78407991e-01
-1.09763213e-01 4.07503009e-01 3.22619468e-01 -6.56482697e-01
-5.51251590e-01 -7.78586745e-01 -9.12623167e-01 3.19578886e-01
-5.73391080e-01 -5.35581470e-01 -9.56699789e-01 8.59558806e-02
-2.34963477e-01 5.47858894e-01 -6.08621180e-01 -2.48388499e-01
-1.12313114e-01 -1.53462338e+00 5.93560100e-01 5.13870478e-01
5.51030397e-01 -1.33027506e+00 -1.00476253e+00 5.29594541e-01
-2.68782705e-01 -7.71222830e-01 2.70123988e-01 6.52523160e-01
-4.25385088e-01 -1.15894055e+00 -1.19928193e+00 -6.93387032e-01
2.87858516e-01 -7.87650704e-01 9.49585438e-01 -3.47928733e-01
-8.22282434e-01 2.75356442e-01 -2.64421552e-01 -9.35592711e-01
2.14684486e-01 7.71075636e-02 4.65195626e-01 -2.38541409e-01
1.10423422e+00 -1.26598239e+00 -7.59319425e-01 1.83303859e-02
-2.37118080e-01 -2.02283766e-02 5.46657801e-01 4.92193788e-01
3.34898770e-01 -5.42286560e-02 1.10616970e+00 -1.51402518e-01
8.89726281e-01 -7.56485701e-01 -4.93656099e-01 -8.62970427e-02
-5.36417305e-01 -6.54653609e-01 5.44429850e-03 -2.98476160e-01
-3.67337078e-01 -1.94583386e-01 -3.39834899e-01 -1.07083701e-01
-7.24834859e-01 3.45516443e-01 3.69995348e-02 -1.06707975e-01
4.08994406e-01 3.69436830e-01 -6.73619151e-01 -4.80521142e-01
-6.15312278e-01 1.24601758e+00 5.96097708e-01 -7.06235543e-02
6.83518276e-02 1.06959984e-01 -3.33842337e-01 -8.54761302e-01
3.57467472e-03 -2.28448331e-01 -4.39117849e-01 -1.14583150e-01
1.04066372e+00 -1.14126348e+00 -5.17297566e-01 9.90091920e-01
-1.10608232e+00 -2.73228735e-01 1.90003529e-01 9.67576623e-01
-5.07505178e-01 -2.29731426e-01 -6.09747231e-01 -5.91203392e-01
-8.72409105e-01 -1.27190661e+00 8.74961972e-01 -5.65408617e-02
-6.45966709e-01 -7.50034273e-01 4.98983949e-01 -5.60251735e-02
4.20276105e-01 4.55865800e-01 7.32701361e-01 -9.44406331e-01
-1.54555827e-01 -2.87719488e-01 7.21098781e-02 3.32200348e-01
2.85626531e-01 -4.66129720e-01 -1.12149739e+00 -3.96585763e-01
2.48796806e-01 -3.25740069e-01 4.04978633e-01 7.37200260e-01
1.32238472e+00 2.49990895e-01 -3.68485630e-01 8.17737579e-01
1.04782784e+00 9.35463667e-01 6.97142839e-01 4.00457591e-01
-3.14635225e-02 3.01477164e-01 -3.10930252e-01 7.26987123e-01
2.66674370e-01 2.30324581e-01 -5.01860045e-02 2.40080148e-01
2.41307750e-01 5.84649205e-01 3.13830897e-02 8.56399715e-01
-3.19609404e-01 -1.20470203e-01 -1.51217258e+00 7.21675634e-01
-1.20000732e+00 -7.36227274e-01 -1.16740607e-01 1.94810367e+00
9.32239532e-01 3.73832136e-02 2.52365936e-02 4.46753025e-01
5.75857580e-01 -5.46947360e-01 -4.00909573e-01 -8.04533288e-02
-3.72777432e-01 7.90793657e-01 1.50760651e-01 -3.73062462e-01
-7.39828169e-01 3.15327615e-01 7.82229185e+00 4.70023543e-01
-1.49059045e+00 3.92314166e-01 6.83078945e-01 -5.36287725e-01
2.59211957e-01 -6.20639563e-01 -3.39582056e-01 1.02468598e+00
1.45052624e+00 -6.10891640e-01 7.35311568e-01 2.66982943e-01
2.46824354e-01 -2.76643008e-01 -1.10102749e+00 1.70342755e+00
1.72947332e-01 -1.22745204e+00 -5.78152180e-01 3.02002132e-01
5.48881054e-01 9.09419358e-01 -9.83445272e-02 4.57300618e-02
-2.08672792e-01 -1.32083035e+00 4.04974431e-01 7.71733344e-01
1.10500598e+00 -9.88614321e-01 8.65540504e-01 2.40814805e-01
-6.94841862e-01 -2.17142060e-01 -4.33697402e-02 -1.74697101e-01
-3.45692858e-02 7.48387694e-01 -3.67023647e-01 4.16898102e-01
1.39520133e+00 9.52305436e-01 -6.12360358e-01 1.41560459e+00
-2.23335773e-01 6.97868288e-01 -2.06913754e-01 2.64414679e-02
-1.47144660e-01 -1.62522122e-02 6.43815696e-01 9.82426941e-01
5.44783652e-01 4.21469450e-01 -4.63611692e-01 8.53026748e-01
3.14362044e-03 -4.99659739e-02 -4.66597110e-01 -1.45610422e-01
1.87027961e-01 1.21809840e+00 -5.63403368e-01 -3.04893814e-02
-5.09154618e-01 8.22255850e-01 -2.38621552e-02 1.52867854e-01
-7.70214736e-01 -7.06113756e-01 5.26740253e-01 -3.42002250e-02
-4.04401153e-01 3.22780699e-01 -2.51756012e-01 -1.21869373e+00
-1.03714041e-01 -1.10362160e+00 1.29970118e-01 -1.10874867e+00
-1.16580856e+00 1.05805254e+00 7.94480741e-02 -8.29816103e-01
-3.74909759e-01 -5.90687215e-01 -7.18389571e-01 1.03052127e+00
-1.05516613e+00 -4.78994638e-01 -2.60847688e-01 7.39717185e-01
3.32429975e-01 -5.42694032e-01 1.30500925e+00 5.94395995e-01
-6.75887227e-01 2.67899871e-01 3.75151411e-02 2.14337349e-01
6.53659940e-01 -1.14567947e+00 2.93508738e-01 2.79243231e-01
-1.19098239e-01 2.97888190e-01 4.64784503e-01 -7.06417322e-01
-9.09301043e-01 -7.89348602e-01 8.31846356e-01 -3.18892807e-01
7.68786371e-01 -5.68132043e-01 -6.60952091e-01 7.24244952e-01
7.57640719e-01 -4.80238229e-01 1.09516108e+00 7.54620209e-02
5.66419482e-01 -1.33276403e-01 -1.24592721e+00 2.09081806e-02
9.09460306e-01 -3.11384648e-01 -1.02208638e+00 5.11647284e-01
-7.09734336e-02 -2.27875650e-01 -1.13249934e+00 5.84581316e-01
4.87833619e-01 -1.04348576e+00 3.95750821e-01 -3.27951878e-01
7.31251761e-02 5.61939180e-01 3.01307529e-01 -1.87142146e+00
-2.90775001e-01 -1.06393230e+00 5.88301532e-02 7.15311170e-01
1.92309335e-01 -1.05202019e+00 6.31725729e-01 5.92668235e-01
-2.77211875e-01 -1.07544041e+00 -1.13170111e+00 -5.55844247e-01
1.82962552e-01 -6.15343630e-01 1.12994957e+00 7.60787368e-01
6.30552948e-01 -2.34123189e-02 -8.14179797e-03 -5.51014058e-02
2.89764166e-01 -2.37878203e-01 1.35436073e-01 -1.43329775e+00
1.87525302e-01 -4.53996778e-01 -1.11889970e+00 4.40602601e-02
1.05197378e-01 -8.06509554e-01 -4.51124310e-01 -1.52311897e+00
4.68951136e-01 9.38848127e-03 -7.53754139e-01 7.50406981e-01
4.19704258e-01 3.44577551e-01 -5.40326834e-01 -1.94510251e-01
-9.91016701e-02 1.71002328e-01 5.78529119e-01 -3.36576179e-02
-6.22425340e-02 -9.20350999e-02 -5.48098028e-01 8.30953896e-01
1.20624435e+00 -4.97738391e-01 -2.16165274e-01 -3.73730510e-01
2.28517264e-01 1.09343849e-01 2.88979769e-01 -1.53089917e+00
2.14710787e-01 2.53869504e-01 1.00661469e+00 -1.08599508e+00
4.46638703e-01 -5.31563103e-01 5.26657462e-01 4.93310243e-01
2.25661471e-02 4.89805013e-01 3.29724580e-01 1.10107362e-02
-1.91935331e-01 2.68160135e-01 6.32539928e-01 -7.97951743e-02
-3.47804517e-01 4.54000533e-01 -7.79060543e-01 2.94416323e-02
1.21543896e+00 -1.15893051e-01 -1.02964021e-01 -2.71819562e-01
-1.14963484e+00 2.88933218e-01 4.00001295e-02 4.15522248e-01
7.10862994e-01 -1.15257096e+00 -1.06779933e+00 6.58682644e-01
9.05541033e-02 -5.59499145e-01 2.09842682e-01 1.44387245e+00
-4.85259682e-01 7.41994143e-01 -8.13439608e-01 -3.70348305e-01
-1.09112048e+00 7.06903711e-02 5.51420391e-01 8.35839957e-02
-1.00827146e+00 1.04252172e+00 -2.50850201e-01 1.74946770e-01
4.66057181e-01 -3.03177387e-01 -3.30313027e-01 1.12055853e-01
9.58504498e-01 4.45558816e-01 6.93424761e-01 -3.99870813e-01
-6.26292884e-01 -3.93074267e-02 2.12218910e-01 -7.59994537e-02
1.92944300e+00 2.19610736e-01 -5.81820548e-01 3.99101734e-01
1.00219512e+00 -4.57173198e-01 -3.50451738e-01 7.28379250e-01
-1.52207203e-02 -8.97789523e-02 4.03000176e-01 -1.38912487e+00
-1.14131629e+00 7.42371917e-01 1.03387630e+00 2.10941523e-01
1.56289732e+00 -2.66640205e-02 6.96669698e-01 2.64342576e-01
1.06776679e+00 -9.47550535e-01 -5.55853724e-01 1.27325043e-01
1.05473888e+00 -6.86653972e-01 -1.89502537e-01 3.10071081e-01
-1.87244236e-01 9.58161056e-01 3.61691803e-01 -2.49835819e-01
8.96711946e-01 7.39364862e-01 -2.28961989e-01 -4.71603483e-01
-9.61340904e-01 3.60891342e-01 1.35760441e-01 8.70742857e-01
5.77238202e-01 2.66040266e-01 -5.71577251e-01 1.28906059e+00
-6.43141150e-01 3.61390859e-01 3.06909978e-01 8.42865884e-01
-8.47630319e-04 -1.17096078e+00 -3.43653440e-01 1.21688366e+00
-8.53560448e-01 -3.74587119e-01 -5.84472001e-01 5.53920329e-01
4.69740629e-01 9.37166810e-01 1.31175742e-01 -2.80150026e-01
1.88280214e-02 2.92153060e-01 6.61806583e-01 -3.78990263e-01
-7.58197010e-01 9.56703052e-02 -2.30557397e-02 -5.21562934e-01
-2.18530729e-01 -8.28078330e-01 -1.14232886e+00 6.47812113e-02
-1.79994673e-01 4.54499602e-01 8.02511513e-01 6.75504982e-01
5.88014662e-01 6.46197498e-01 3.20436686e-01 -7.16933310e-01
-8.54842924e-03 -1.26936555e+00 -1.26229048e+00 -1.54240131e-01
2.78715074e-01 -6.78560555e-01 -4.66644615e-01 -2.30791613e-01] | [13.219476699829102, 3.506120443344116] |
242b44d2-1011-4775-9140-ecabfa14ceec | from-argument-search-to-argumentative | null | null | https://aclanthology.org/2021.sigdial-1.39 | https://aclanthology.org/2021.sigdial-1.39.pdf | From Argument Search to Argumentative Dialogue: A Topic-independent Approach to Argument Acquisition for Dialogue Systems | Despite the remarkable progress in the field of computational argumentation, dialogue systems concerned with argumentative tasks often rely on structured knowledge about arguments and their relations. Since the manual acquisition of these argument structures is highly time-consuming, the corresponding systems are inflexible regarding the topics they can discuss. To address this issue, we propose a combination of argumentative dialogue systems with argument search technology that enables a system to discuss any topic on which the search engine is able to find suitable arguments. Our approach utilizes supervised learning-based relation classification to map the retrieved arguments into a general tree structure for use in dialogue systems. We evaluate the approach with a state of the art search engine and a recently introduced dialogue model in an extensive user study with respect to the dialogue coherence. The results vary between the investigated topics (and hence depend on the quality of the underlying data) but are in some instances surprisingly close to the results achieved with a manually annotated argument structure. | ['Stefan Ultes', 'Wolfgang Minker', 'Johannes Daxenberger', 'Isabel Feustel', 'Carolin Schindler', 'Niklas Rach'] | null | null | null | null | sigdial-acl-2021-7 | ['relation-classification'] | ['natural-language-processing'] | [ 1.95442036e-01 1.00207746e+00 -1.22101650e-01 -2.31953621e-01
-7.56505311e-01 -9.12352681e-01 1.25230658e+00 1.00233698e+00
-5.81466258e-01 1.15442109e+00 3.84459406e-01 -8.25371861e-01
-4.36918646e-01 -9.79578733e-01 -4.89214137e-02 -2.20104277e-01
2.33625561e-01 9.80674386e-01 5.89323580e-01 -9.19196963e-01
5.07598042e-01 6.41124472e-02 -1.62292969e+00 5.58419466e-01
1.18828118e+00 7.68212914e-01 1.31244259e-02 5.70727706e-01
-7.54996002e-01 8.05052876e-01 -9.38535571e-01 -6.64114356e-01
-2.53569275e-01 -7.11102545e-01 -1.71687174e+00 -2.21143231e-01
-1.20628871e-01 2.95541257e-01 5.92950106e-01 9.17435229e-01
1.95434168e-01 -2.80948907e-01 6.64821863e-01 -9.28584158e-01
1.84030354e-01 9.70776081e-01 2.02698544e-01 1.22982621e-01
8.02355707e-01 -5.06571233e-01 1.24288106e+00 -4.06214148e-01
1.05428112e+00 1.20625103e+00 4.55975413e-01 3.83160383e-01
-1.15803289e+00 2.86538620e-02 1.13719031e-01 1.37901098e-01
-5.00598192e-01 -3.86656910e-01 7.21530914e-01 -4.48437214e-01
8.46530795e-01 5.81741929e-01 7.96295702e-01 8.75848532e-01
-2.50964373e-01 4.86256242e-01 1.34923685e+00 -9.52256620e-01
4.05860126e-01 5.92621505e-01 5.45686185e-01 5.19298077e-01
2.54996449e-01 -4.08471286e-01 -3.72695267e-01 -5.72316945e-01
2.68746376e-01 -9.76263165e-01 -2.83196151e-01 -4.82781291e-01
-1.01728261e+00 1.11237860e+00 7.31141269e-02 8.31437707e-01
-4.30808812e-01 -7.62543142e-01 9.01193500e-01 6.80664599e-01
7.02757001e-01 7.96671689e-01 -7.02360034e-01 -3.67794544e-01
-5.51991820e-01 5.47457993e-01 1.80865645e+00 3.17047596e-01
4.48422462e-01 -6.73072100e-01 -1.07262731e-01 9.67153072e-01
3.03683579e-01 -6.48933947e-02 3.25573802e-01 -7.19471216e-01
6.19342685e-01 1.09367681e+00 3.81447643e-01 -7.82342792e-01
-3.57173055e-01 -1.15587458e-01 -2.68175304e-01 3.51302087e-01
9.19130862e-01 -2.67335087e-01 1.20637100e-02 1.31788623e+00
7.24548161e-01 -9.26537097e-01 4.88513440e-01 6.06419563e-01
9.34698343e-01 2.38047749e-01 2.26044327e-01 -6.73746765e-01
1.60321820e+00 -6.83279693e-01 -9.11934197e-01 1.70925811e-01
9.88032341e-01 -1.05677247e+00 9.98157561e-01 4.79488730e-01
-1.05945289e+00 -2.07852677e-01 -9.05718267e-01 8.84906109e-03
-4.40584868e-01 4.66126949e-02 6.72468066e-01 6.79411232e-01
-7.21379280e-01 6.39085412e-01 -2.71985292e-01 -5.23013711e-01
-3.50853577e-02 2.75348067e-01 -2.46243984e-01 6.11121118e-01
-1.48262727e+00 1.31231797e+00 6.68424249e-01 -6.58294037e-02
1.20381802e-01 -1.42931536e-01 -6.46112978e-01 5.22050895e-02
6.16099894e-01 -5.88557661e-01 1.51420033e+00 -9.07525778e-01
-1.74318266e+00 9.97134686e-01 2.08280474e-01 -7.11218596e-01
9.77803648e-01 -3.58446121e-01 -7.42168501e-02 1.72860771e-01
2.42612407e-01 1.14736326e-01 4.31322217e-01 -1.02893531e+00
-6.88138604e-01 -1.76444903e-01 4.76245046e-01 3.98679137e-01
-5.97394444e-02 2.25164890e-01 6.80664182e-02 -2.56018579e-01
-6.29775152e-02 -6.93845451e-01 -7.98960626e-02 -2.03465462e-01
-5.94034374e-01 -8.84947419e-01 7.78559744e-01 -4.67789978e-01
1.30542684e+00 -1.53337157e+00 2.89799213e-01 2.14219630e-01
-8.36270582e-03 5.39450109e-01 4.20340896e-01 8.50009739e-01
9.02769715e-03 1.98096022e-01 -7.64233246e-02 1.22823201e-01
1.38602465e-01 3.03733021e-01 -3.80904734e-01 6.28965348e-02
-6.54274002e-02 4.90156978e-01 -9.38316405e-01 -8.41816008e-01
2.51717269e-01 1.44659534e-01 -2.83644468e-01 3.98487002e-01
-6.93475127e-01 3.76359850e-01 -9.68441129e-01 2.31092662e-01
6.72032610e-02 -1.55321836e-01 5.93043387e-01 -6.04042523e-02
-2.83320606e-01 1.07199013e+00 -9.80246723e-01 1.43449247e+00
-6.12471998e-01 4.48503613e-01 2.12416023e-01 -1.07662201e+00
9.96346295e-01 7.00727046e-01 2.37770781e-01 -4.67320204e-01
2.64513314e-01 5.06413400e-01 2.66456246e-01 -5.20699739e-01
4.74319130e-01 6.90841749e-02 8.21902081e-02 9.25499260e-01
-9.30094644e-02 -3.16418231e-01 5.61255872e-01 2.15845495e-01
8.95958006e-01 3.14077348e-01 6.70697987e-01 -6.23334885e-01
9.47658300e-01 4.87639725e-01 -1.79509312e-01 5.05707264e-01
2.89069027e-01 -5.80069125e-02 8.73750389e-01 -6.03892028e-01
-9.31693912e-01 -5.14189363e-01 -3.34518611e-01 9.58628476e-01
1.06436862e-02 -6.80867493e-01 -1.09077811e+00 -1.10865521e+00
-3.97122979e-01 7.14440346e-01 -5.71115971e-01 4.05793399e-01
-5.52334666e-01 -4.29521739e-01 2.91743070e-01 -2.98583329e-01
5.71822047e-01 -1.44986880e+00 -1.16770852e+00 4.28256959e-01
-4.18876082e-01 -7.82148719e-01 4.93440211e-01 2.64025033e-01
-7.84111083e-01 -1.65242791e+00 -2.80994654e-01 -3.92672420e-01
3.78702641e-01 -3.36906075e-01 1.36582017e+00 5.73205113e-01
7.78045729e-02 2.08488107e-01 -6.61412179e-01 -5.56801677e-01
-1.16426313e+00 6.16624236e-01 -5.64052343e-01 -4.60397035e-01
5.81108071e-02 -3.87970775e-01 -2.86540002e-01 2.19979197e-01
-7.16514707e-01 4.71225120e-02 1.87752068e-01 1.02190018e+00
-1.51840031e-01 -1.14536405e-01 6.67497039e-01 -1.49029303e+00
1.51226354e+00 -2.99666643e-01 -5.12450337e-01 5.12342632e-01
-9.11593795e-01 5.19826233e-01 3.87387216e-01 -2.24786490e-01
-1.27166629e+00 -3.67754281e-01 -4.19808477e-01 8.20095718e-01
-2.56407768e-01 7.04017997e-01 4.81453761e-02 1.17365167e-01
1.10870922e+00 -3.24914545e-01 1.10320382e-01 -5.28451800e-01
4.18054819e-01 6.15548015e-01 1.63168073e-01 -9.63414490e-01
4.88509804e-01 2.31049191e-02 -1.74720943e-01 -6.68499768e-01
-1.02260232e+00 -3.63731265e-01 -5.45669734e-01 -1.48239270e-01
4.56822753e-01 -1.83778107e-01 -7.53237307e-01 -1.35428563e-01
-1.46665752e+00 -9.56470370e-02 -4.69948500e-01 2.30274826e-01
-7.19001234e-01 6.04900599e-01 -4.71390098e-01 -1.03821337e+00
-6.09409869e-01 -9.10511613e-01 8.88775885e-01 -1.32362274e-02
-1.03405428e+00 -1.17805266e+00 3.21854770e-01 5.11313558e-01
2.85025269e-01 1.66809857e-01 1.34192741e+00 -1.30199254e+00
1.64426062e-02 -2.03617230e-01 -1.27729028e-01 -3.58093455e-02
2.17955738e-01 -9.74628106e-02 -7.55368769e-01 2.14140356e-01
8.21811110e-02 -4.16485488e-01 4.24425930e-01 -1.75093234e-01
3.86721581e-01 -4.24983799e-01 -3.44480157e-01 -5.63363791e-01
8.62924218e-01 -2.74286475e-02 4.77341503e-01 9.32300925e-01
-1.12709388e-01 1.24593532e+00 9.95533764e-01 2.16981962e-01
1.75262615e-01 9.97274399e-01 2.98362076e-01 1.85484358e-03
2.19908535e-01 -7.40166605e-02 -9.43703055e-02 3.46017748e-01
-2.39881039e-01 -2.51685232e-01 -8.05860817e-01 6.04984045e-01
-2.09589863e+00 -8.90212059e-01 -4.27498758e-01 2.06398392e+00
1.18353045e+00 4.62277561e-01 2.94980943e-01 5.98552823e-01
5.18783867e-01 4.06447127e-02 7.44717047e-02 -8.16197813e-01
-3.92400771e-02 2.99961597e-01 -2.33906493e-01 8.41266453e-01
-9.49383914e-01 7.26046443e-01 6.02086449e+00 5.76972425e-01
-6.44030809e-01 -1.27385795e-01 4.40586865e-01 6.27298594e-01
-2.63230205e-01 3.30550432e-01 -5.22717595e-01 2.12257937e-01
9.71131980e-01 -1.97309181e-01 1.62911080e-02 6.60750926e-01
8.34228098e-02 -5.78509808e-01 -1.10811114e+00 3.60610873e-01
-4.00973558e-01 -1.45558965e+00 -1.07657053e-01 8.65219757e-02
9.13649574e-02 -4.31578308e-01 -4.56344694e-01 9.99140441e-02
3.43813121e-01 -7.70357788e-01 4.28062648e-01 1.64952517e-01
2.24919513e-01 -5.00974298e-01 8.91327322e-01 6.54913306e-01
-6.04298770e-01 1.16583832e-01 1.24535918e-01 -2.45205998e-01
1.77756578e-01 4.16016072e-01 -1.16845345e+00 7.05122828e-01
3.49294841e-01 1.33862242e-01 -3.72861981e-01 7.88626194e-01
-6.11769319e-01 5.95628202e-01 -4.81845796e-01 -4.78530467e-01
1.96119919e-01 -2.44426996e-01 6.84668779e-01 1.07728291e+00
5.45544438e-02 3.09585463e-02 1.51602626e-01 4.34678018e-01
2.93234378e-01 8.15645576e-01 -5.02286553e-01 1.89298823e-01
1.70950368e-01 1.30723727e+00 -9.39024270e-01 -5.09166121e-01
-9.29916799e-02 5.50386310e-01 4.01201189e-01 -1.04699641e-01
-2.91233122e-01 -1.07728563e-01 4.13570367e-02 1.64950833e-01
3.76514606e-02 5.81700094e-02 -1.09485887e-01 -8.97110403e-01
3.95250708e-01 -1.03633976e+00 5.53902447e-01 -4.40842390e-01
-1.08847487e+00 1.14873993e+00 3.06491554e-01 -8.77562165e-01
-9.90161121e-01 -4.95758384e-01 -5.80600500e-01 9.06037688e-01
-1.45595169e+00 -8.90005529e-01 1.05996378e-01 2.30535716e-01
5.57233453e-01 -3.52703892e-02 1.29215443e+00 -1.64416417e-01
1.19454138e-01 -1.08429946e-01 -4.10865217e-01 -5.63778961e-03
6.07221544e-01 -1.32933736e+00 1.71210229e-01 1.54179230e-01
3.88692230e-01 5.93268096e-01 1.20189238e+00 -4.54468817e-01
-7.55582750e-01 -1.98410109e-01 1.40306377e+00 -3.56752008e-01
9.10334110e-01 -8.66876841e-02 -1.07663250e+00 -6.19599223e-02
7.84375250e-01 -3.51162940e-01 6.75532818e-01 7.11077332e-01
-3.25919747e-01 4.88982052e-01 -1.02902079e+00 4.13363904e-01
7.89499938e-01 -4.28802490e-01 -1.21987104e+00 4.66636777e-01
3.16481888e-01 -4.55679923e-01 -8.09099972e-01 2.74502724e-01
4.44759637e-01 -1.19033909e+00 6.72832787e-01 -7.56632864e-01
3.48747581e-01 -1.25778466e-01 2.75719970e-01 -1.16552365e+00
5.88460684e-01 -8.01337659e-01 5.85276857e-02 1.24945354e+00
9.16699767e-01 -7.07138002e-01 8.43107164e-01 7.97489703e-01
2.85554886e-01 -7.18247533e-01 -9.70075190e-01 -3.03752065e-01
7.73622319e-02 -8.98389071e-02 2.64351845e-01 9.52606618e-01
7.69896805e-01 1.06314218e+00 1.31553158e-01 -3.59110594e-01
2.29487687e-01 4.86717820e-01 7.72042274e-01 -2.01808310e+00
-2.71968693e-01 -5.47384799e-01 6.38862401e-02 -6.64652050e-01
3.04926068e-01 -5.76972246e-01 -1.10008359e-01 -1.68451869e+00
-2.56696343e-01 -6.47334933e-01 2.37386122e-01 3.22886914e-01
-6.44756034e-02 -3.65611047e-01 -5.74106053e-02 2.14111671e-01
-5.67431748e-01 2.68273532e-01 9.67459559e-01 5.48914634e-02
-4.73454237e-01 3.29304010e-01 -6.26981318e-01 8.94493520e-01
7.94388056e-01 -3.71009469e-01 -5.09081602e-01 2.55980730e-01
6.43623054e-01 3.62261266e-01 6.00852780e-02 -4.84445721e-01
2.75403470e-01 5.82312495e-02 -2.51038313e-01 -4.20144588e-01
1.59968466e-01 -8.23899746e-01 -9.55296028e-03 4.76092219e-01
-8.21044028e-01 4.44875509e-02 9.25013646e-02 1.95375264e-01
-4.73665118e-01 -7.14025915e-01 3.23522121e-01 -9.80038717e-02
-1.50358230e-01 -4.40674663e-01 -4.69659984e-01 1.78851664e-01
6.53984368e-01 -1.07366428e-01 -4.54846591e-01 -5.32883465e-01
-8.65212560e-01 5.63668236e-02 3.95820260e-01 1.56903222e-01
1.78241000e-01 -6.93287194e-01 -7.34660506e-01 -3.86756480e-01
1.19709365e-01 -2.82733049e-02 -3.73329997e-01 6.43390059e-01
-4.52148974e-01 5.14559567e-01 -1.10430315e-01 -4.08816099e-01
-1.73609364e+00 3.27687651e-01 1.34217486e-01 -9.56263781e-01
-5.58402061e-01 2.06072375e-01 -3.37077469e-01 -3.86444569e-01
4.47249599e-02 -2.23921403e-01 -7.97028065e-01 4.28073704e-01
3.15527290e-01 5.88889010e-02 1.94535881e-01 -4.71979827e-01
-3.34768696e-03 2.21397072e-01 -6.08173236e-02 -4.41988796e-01
1.21682405e+00 -9.28234085e-02 -3.64478230e-01 3.96917999e-01
4.79304314e-01 2.56343067e-01 -6.85299993e-01 -3.54506314e-01
6.42261565e-01 -1.30440488e-01 -3.31862867e-01 -9.86032963e-01
-1.71779245e-01 5.22887230e-01 1.01637259e-01 1.29330528e+00
6.63166583e-01 1.44187301e-01 2.57616788e-01 8.18788707e-01
4.10803914e-01 -1.09142482e+00 -3.31014335e-01 6.18890285e-01
9.66468751e-01 -1.19128919e+00 3.42220575e-01 -1.03152883e+00
-4.96289074e-01 1.54609847e+00 1.29973069e-01 1.72618508e-01
4.88189787e-01 1.44250378e-01 3.66077840e-01 -6.89530969e-01
-9.19958711e-01 -1.95958540e-01 1.58991873e-01 4.44936275e-01
9.00786817e-01 -3.08484107e-01 -1.29172945e+00 1.89420566e-01
-5.18047333e-01 -2.45543882e-01 3.94054025e-01 9.28512156e-01
-4.82491791e-01 -1.89508379e+00 -2.87448734e-01 1.19635217e-01
-6.17407203e-01 -1.30465496e-02 -1.04481208e+00 1.08417904e+00
-3.84897232e-01 1.18643379e+00 -4.55113083e-01 3.65914851e-01
4.92759168e-01 1.98902294e-01 4.57621366e-01 -8.64102244e-01
-1.31039882e+00 -9.08950716e-02 1.32591081e+00 -1.52329266e-01
-1.00334871e+00 -7.44823158e-01 -1.14103782e+00 1.05126336e-01
-7.35189915e-01 1.16288781e+00 7.67507553e-01 1.39266980e+00
8.19853023e-02 1.96411207e-01 2.81319678e-01 -3.45845729e-01
-5.23220301e-01 -1.09873950e+00 -1.46902919e-01 5.17878056e-01
-4.32961360e-02 -6.75319850e-01 9.54586361e-03 -1.73279956e-01] | [9.692276954650879, 9.544038772583008] |
abcf98cf-1d35-4660-9f4a-6d11d8597836 | learning-disentangled-representations-in-1 | 2307.03077 | null | https://arxiv.org/abs/2307.03077v1 | https://arxiv.org/pdf/2307.03077v1.pdf | Learning Disentangled Representations in Signed Directed Graphs without Social Assumptions | Signed graphs are complex systems that represent trust relationships or preferences in various domains. Learning node representations in such graphs is crucial for many mining tasks. Although real-world signed relationships can be influenced by multiple latent factors, most existing methods often oversimplify the modeling of signed relationships by relying on social theories and treating them as simplistic factors. This limits their expressiveness and their ability to capture the diverse factors that shape these relationships. In this paper, we propose DINES, a novel method for learning disentangled node representations in signed directed graphs without social assumptions. We adopt a disentangled framework that separates each embedding into distinct factors, allowing for capturing multiple latent factors. We also explore lightweight graph convolutions that focus solely on sign and direction, without depending on social theories. Additionally, we propose a decoder that effectively classifies an edge's sign by considering correlations between the factors. To further enhance disentanglement, we jointly train a self-supervised factor discriminator with our encoder and decoder. Throughout extensive experiments on real-world signed directed graphs, we show that DINES effectively learns disentangled node representations, and significantly outperforms its competitors in the sign prediction task. | ['Jinhong Jung', 'Geonwoo Ko'] | 2023-07-06 | null | null | null | null | ['disentanglement'] | ['methodology'] | [ 1.44777343e-01 3.71875942e-01 -6.16954565e-01 -7.34826803e-01
1.45967588e-01 -7.80228853e-01 8.13028991e-01 -9.70488414e-02
4.16069254e-02 4.99402523e-01 6.10032320e-01 -4.18510318e-01
-3.07969242e-01 -8.56737375e-01 -6.91300392e-01 -5.06549478e-01
-5.15272796e-01 4.80677158e-01 -5.92156872e-02 -3.15482438e-01
-1.22458719e-01 3.42530429e-01 -1.04169321e+00 1.18766353e-01
6.87756479e-01 6.78125560e-01 -5.71381390e-01 8.28880608e-01
1.95838079e-01 8.66577387e-01 -4.69582200e-01 -1.05449462e+00
3.03338349e-01 -3.84971291e-01 -1.72301859e-01 -9.45471302e-02
6.80507123e-01 -5.65522492e-01 -8.41544509e-01 8.66912186e-01
1.85711041e-01 -5.59161782e-01 1.02491736e+00 -1.83263242e+00
-1.07871127e+00 1.20978248e+00 -7.02828348e-01 -1.65937111e-01
1.14382692e-01 -2.33921871e-01 1.84104955e+00 -5.91925740e-01
5.42170227e-01 1.22232974e+00 6.35238886e-01 5.01121879e-01
-1.33261192e+00 -9.22564685e-01 4.37487036e-01 3.55032608e-02
-8.91960979e-01 -3.50249767e-01 1.17847109e+00 -3.15882236e-01
4.78741318e-01 2.00575739e-01 7.42809534e-01 1.43117499e+00
-1.03234109e-02 1.05511487e+00 8.92226517e-01 -5.04805297e-02
-9.96624026e-03 -2.32468545e-01 3.39144588e-01 1.10711014e+00
8.90988886e-01 9.07237679e-02 -7.90516913e-01 -4.46928710e-01
7.21021116e-01 1.68685973e-01 -4.88705188e-01 -1.03261483e+00
-1.38299572e+00 1.11746657e+00 9.17225122e-01 7.51090748e-03
-6.62164614e-02 6.25675619e-01 3.11991781e-01 5.58923483e-01
3.72236937e-01 1.11663789e-01 -4.59838182e-01 3.04183841e-01
-6.55018091e-01 -9.02060699e-03 1.21976900e+00 9.00099158e-01
6.81540191e-01 -7.29546845e-02 8.87257084e-02 4.61609393e-01
8.15832913e-01 6.49105310e-01 -4.71360944e-02 -7.51653135e-01
3.20586801e-01 7.27633655e-01 -3.22676659e-01 -1.45179892e+00
-2.88799644e-01 -5.24847388e-01 -1.04163837e+00 1.15025602e-01
3.43089998e-01 -4.68591675e-02 -9.83985126e-01 2.04216218e+00
-1.13150813e-01 1.27891600e-01 -2.76475847e-01 8.98346603e-01
7.11761832e-01 1.41818240e-01 -3.07400018e-01 1.55807748e-01
1.17815936e+00 -8.39803815e-01 -8.05204809e-01 -3.27244967e-01
3.15284848e-01 -2.81260759e-01 5.64038634e-01 2.67638326e-01
-8.76226723e-01 1.03787690e-01 -1.48320043e+00 -3.03969830e-01
-2.78348237e-01 6.98019639e-02 1.40374959e+00 6.15357459e-01
-1.05411863e+00 6.41457677e-01 -9.16712701e-01 -3.31809968e-02
5.79469502e-01 5.29789031e-01 -6.14548743e-01 -8.61573070e-02
-1.34693205e+00 6.72441304e-01 -4.02516991e-01 2.57945955e-01
-5.15377998e-01 -4.93896604e-01 -1.08807909e+00 2.89343268e-01
2.84530520e-01 -8.37758660e-01 8.49915385e-01 -8.30196798e-01
-1.17867887e+00 8.37522864e-01 -5.80793060e-02 -1.68890908e-01
6.06106579e-01 1.67526588e-01 -3.57951850e-01 1.30322710e-01
-3.77717800e-02 2.91874498e-01 1.10650647e+00 -1.42630398e+00
2.68427908e-01 -4.69893336e-01 4.92376208e-01 -1.60868838e-01
-5.88198066e-01 -2.20273331e-01 -3.83825570e-01 -8.01068366e-01
3.07142943e-01 -1.12025380e+00 -1.05124570e-01 8.37387443e-01
-5.30770898e-01 1.37998968e-01 5.49494565e-01 -3.81627530e-01
1.04522765e+00 -1.95938802e+00 4.70921189e-01 6.65911973e-01
1.24684632e+00 -4.65710498e-02 -2.20561638e-01 4.53757554e-01
-8.22739974e-02 2.83626616e-01 -1.69198573e-01 -4.88015264e-01
5.11113226e-01 5.43913305e-01 -1.32140547e-01 6.15083337e-01
2.57820308e-01 1.17708695e+00 -1.01974344e+00 -2.86641240e-01
-2.59966642e-01 8.44332993e-01 -6.62215769e-01 -2.28037164e-02
-5.00344597e-02 -7.09510669e-02 -4.71574217e-01 4.04801458e-01
9.11694348e-01 -6.42057955e-01 7.38356411e-01 -5.79840660e-01
5.80597162e-01 4.70097810e-01 -1.06702626e+00 1.46715200e+00
-4.21794772e-01 7.33098328e-01 3.54437470e-01 -7.29321361e-01
9.69660163e-01 1.45214945e-01 4.66833115e-01 -1.14570461e-01
2.64547199e-01 3.15892547e-01 1.95025057e-01 -5.82258590e-02
1.62111491e-01 -1.14317402e-01 -9.81766209e-02 9.10988808e-01
3.68548214e-01 1.87186077e-01 2.23630413e-01 9.51176047e-01
1.38909602e+00 -1.38371721e-01 2.15251178e-01 -1.30378634e-01
1.47660002e-01 -7.88804829e-01 5.63161492e-01 2.99825758e-01
-2.75033176e-01 5.36052048e-01 1.22266877e+00 -2.00006485e-01
-6.70151591e-01 -1.37375379e+00 2.10291296e-01 1.03449178e+00
6.42869473e-02 -6.60608649e-01 -3.02887615e-03 -1.10708058e+00
5.18987596e-01 2.49024719e-01 -1.01551807e+00 -4.60060120e-01
-3.82249445e-01 -6.27532542e-01 7.39225030e-01 6.82825983e-01
4.89254035e-02 -4.11189854e-01 1.72722369e-01 -2.69749761e-01
1.65287629e-01 -1.01146400e+00 -7.28030026e-01 4.60336238e-01
-4.84916419e-01 -1.55888319e+00 -5.95966697e-01 -6.20289207e-01
1.15164793e+00 5.45884788e-01 1.17627990e+00 4.11733508e-01
1.24756517e-02 3.65749180e-01 -2.26624727e-01 -4.87495400e-02
-3.12430710e-01 -2.21097153e-02 -7.65433386e-02 2.05389559e-01
3.71971816e-01 -9.71531451e-01 -6.72580123e-01 3.70937735e-01
-8.73942971e-01 1.57174647e-01 6.38986588e-01 9.03435111e-01
-2.85867155e-01 -5.72561264e-01 4.01232660e-01 -1.24574208e+00
7.27645159e-01 -5.42692900e-01 -2.25221470e-01 3.54780555e-01
-6.02652252e-01 6.08996689e-01 4.85850781e-01 -4.03559506e-01
-6.11994147e-01 -2.69399017e-01 3.52804482e-01 -4.46330369e-01
5.47258317e-01 5.22822618e-01 -3.98965955e-01 -2.39794910e-01
5.26806831e-01 -2.96014160e-01 1.97157815e-01 -1.45755991e-01
7.45067537e-01 3.33306313e-01 1.15618184e-01 -5.63920379e-01
1.26870906e+00 5.48654556e-01 1.82639167e-01 -1.63782492e-01
-7.68918812e-01 -6.85148686e-02 -6.76203728e-01 -4.11465839e-02
3.99169117e-01 -9.44548905e-01 -1.02492118e+00 2.38430187e-01
-1.00702488e+00 1.05943363e-02 1.73706822e-02 4.04930294e-01
-2.15019256e-01 4.33613360e-01 -8.85913491e-01 -4.28361267e-01
3.01157427e-03 -8.30436945e-01 1.05438352e+00 -1.67175546e-01
-3.02580893e-01 -1.32280409e+00 1.05376258e-01 3.46923798e-01
5.02483070e-01 2.39089936e-01 8.58517051e-01 -6.61856353e-01
-5.60854077e-01 -4.35682684e-01 -5.42338550e-01 2.06684127e-01
1.83861285e-01 2.00912967e-01 -8.39605093e-01 -1.90728843e-01
-6.45702660e-01 -3.86073947e-01 1.38957536e+00 -2.06289575e-01
6.25422299e-01 -3.78741205e-01 -3.61083120e-01 6.58296704e-01
7.83936501e-01 -5.78627408e-01 1.51143134e-01 -3.47242028e-01
1.10074031e+00 4.67225492e-01 -2.84511596e-01 3.26614261e-01
8.71719658e-01 3.05060267e-01 6.14865601e-01 -1.88705847e-01
-1.75520167e-01 -7.35628128e-01 4.16830271e-01 1.09087932e+00
-1.73895761e-01 -4.45356905e-01 -4.96892989e-01 2.08803609e-01
-1.89416742e+00 -8.54452670e-01 -2.43001238e-01 1.78163755e+00
7.68295884e-01 2.88021296e-01 -1.33794084e-01 3.06800574e-01
6.06013596e-01 8.05816114e-01 -6.89972758e-01 -2.25455612e-01
-3.68980110e-01 2.10648209e-01 7.85105526e-01 7.63781607e-01
-9.05877650e-01 4.63504285e-01 5.83662415e+00 1.31400868e-01
-9.70014572e-01 -2.61744201e-01 2.88931459e-01 2.97758095e-02
-1.24347806e+00 9.21612307e-02 -2.37638667e-01 9.05712321e-02
3.44423801e-01 -2.44962990e-01 5.68808615e-01 4.80384260e-01
-1.45577401e-01 5.95996082e-01 -1.37179124e+00 6.31950736e-01
3.71485174e-01 -1.01976252e+00 1.46164894e-01 1.41215712e-01
7.06737638e-01 2.45392416e-03 1.40920311e-01 6.35509714e-02
8.58471334e-01 -1.19859040e+00 5.53851306e-01 2.70714760e-01
6.62042618e-01 -2.77754426e-01 4.58998799e-01 -5.59153687e-03
-1.10126662e+00 1.81871414e-01 -8.75004679e-02 -2.28644103e-01
5.65929525e-03 7.58276105e-01 -6.54058337e-01 5.76602042e-01
2.12726071e-02 1.22898471e+00 -5.91631651e-01 5.16508043e-01
-8.68616283e-01 6.34534717e-01 -3.64083141e-01 -2.60197252e-01
-7.91969970e-02 -2.26702109e-01 2.73449957e-01 7.74892271e-01
-1.17740128e-02 -2.68176168e-01 -3.71858627e-02 9.02745306e-01
-5.18834889e-01 -2.71551967e-01 -7.12698340e-01 -4.83891875e-01
2.90872246e-01 1.27087820e+00 -6.80227280e-01 -9.58961472e-02
-6.54541731e-01 1.18592346e+00 5.63184321e-01 6.13807499e-01
-8.67624104e-01 -1.37180939e-01 9.80982482e-01 6.18472099e-02
4.41981822e-01 -5.72641134e-01 -1.64684162e-01 -1.87193370e+00
2.38867596e-01 -6.69282556e-01 2.78914273e-01 -4.97902066e-01
-1.67891800e+00 5.92540383e-01 -2.85064697e-01 -1.02841282e+00
-5.72604649e-02 -9.36765313e-01 -6.40091240e-01 5.82973778e-01
-1.66269279e+00 -1.65773213e+00 -2.51411408e-01 6.50753975e-01
-3.60664755e-01 1.27043039e-01 8.91743362e-01 3.81378621e-01
-3.02972198e-01 7.49517322e-01 -1.21630713e-01 4.36839640e-01
8.36757481e-01 -1.42301571e+00 5.68265557e-01 5.85972786e-01
5.41246951e-01 9.52142954e-01 4.58869547e-01 -5.86968362e-01
-1.69518006e+00 -6.45218015e-01 1.07804716e+00 -1.30254418e-01
1.25450444e+00 -9.68340218e-01 -5.83352685e-01 1.02809405e+00
1.06884360e-01 5.94369292e-01 1.08006978e+00 5.56760550e-01
-1.32371831e+00 -3.30588520e-02 -8.85243595e-01 7.24286735e-01
1.58602405e+00 -8.25874448e-01 -2.35467196e-01 1.10883079e-01
6.22590363e-01 -3.92107479e-03 -6.71414733e-01 4.83070940e-01
8.95269275e-01 -1.08407950e+00 8.23466361e-01 -7.85692990e-01
4.77681696e-01 -1.14854895e-01 -1.01527594e-01 -1.54881489e+00
-4.41432476e-01 -7.87924051e-01 -6.35953248e-01 9.05899286e-01
6.29026175e-01 -8.87212217e-01 1.09793282e+00 6.30162954e-01
4.05436575e-01 -7.99997330e-01 -5.33229649e-01 -2.44918883e-01
-2.03878313e-01 -2.79996544e-01 7.02527165e-01 1.32716095e+00
2.44395941e-01 5.16685009e-01 -6.13865137e-01 2.48794362e-01
8.35165977e-01 2.53708035e-01 5.95157385e-01 -1.35786748e+00
-6.71472907e-01 -5.79792023e-01 -7.87669063e-01 -1.36237466e+00
4.76521164e-01 -1.27916956e+00 -4.79660213e-01 -1.49213815e+00
2.15086162e-01 -3.27047765e-01 -4.09838676e-01 5.52836239e-01
-2.54816413e-01 3.37324470e-01 2.89726287e-01 -1.02391466e-01
-4.65207487e-01 7.72612154e-01 1.47680914e+00 -4.95965093e-01
4.18586701e-01 2.34605204e-02 -1.10850394e+00 6.77137434e-01
5.87555528e-01 -3.25519621e-01 -6.10783815e-01 -6.05989814e-01
7.97147214e-01 1.25687569e-02 4.91659194e-01 -2.04934776e-01
2.57927954e-01 1.09577395e-01 1.90175757e-01 -1.43235130e-02
4.10378605e-01 -8.93629193e-01 -1.94916725e-01 3.47674876e-01
-5.91123939e-01 -2.16814607e-01 -4.58980471e-01 8.08113456e-01
-1.60877779e-01 2.08858013e-01 1.91493824e-01 1.84911773e-01
1.44227417e-02 6.34713233e-01 1.01375893e-01 -9.23678055e-02
7.29642987e-01 9.51041952e-02 -5.13627768e-01 -9.04693604e-01
-8.73877943e-01 4.80079532e-01 5.24140418e-01 4.97954905e-01
6.77962899e-01 -1.48873925e+00 -6.40695751e-01 6.20397687e-01
2.45336279e-01 -2.86936313e-01 -1.81032151e-01 7.20377386e-01
-4.93419021e-01 -6.39821738e-02 -5.52093908e-02 -3.44374701e-02
-1.24719346e+00 3.38692933e-01 2.26133034e-01 -3.39961141e-01
-1.95441231e-01 7.86601007e-01 1.77888304e-01 -6.91773593e-01
3.08325499e-01 -4.99904931e-01 -1.78275816e-02 1.90159962e-01
8.17631185e-02 1.77804306e-01 -2.19031394e-01 -6.24340832e-01
-3.22805822e-01 4.14824456e-01 -5.26535101e-02 1.07722424e-01
1.35989869e+00 2.18894705e-01 -2.31610492e-01 4.41400498e-01
1.45890498e+00 5.00106096e-01 -1.24366713e+00 -3.93024832e-01
-1.18606664e-01 -5.19104242e-01 9.46161430e-03 -6.80462897e-01
-1.57739782e+00 9.25699949e-01 -3.06105524e-01 2.41059616e-01
7.74685740e-01 2.18314007e-02 6.30365312e-01 3.46513182e-01
2.86891550e-01 -1.81288138e-01 1.91161722e-01 2.93802708e-01
8.70942831e-01 -1.28869891e+00 1.97126701e-01 -8.91930044e-01
-6.75029159e-01 1.33643973e+00 3.80847901e-01 -4.98566747e-01
1.16218829e+00 5.19081652e-01 2.98031475e-02 -3.35937381e-01
-8.97402704e-01 3.46303023e-02 5.79287946e-01 4.51769084e-01
6.05482101e-01 4.22601581e-01 -3.96957248e-01 7.19758451e-01
-2.28864253e-01 -2.17628166e-01 5.74243844e-01 7.23056376e-01
-6.77062497e-02 -1.32123208e+00 1.11581787e-01 6.49947584e-01
-2.26340666e-01 -1.26323178e-01 -8.97763669e-01 5.74579358e-01
-1.33385986e-01 6.28675759e-01 -1.19504213e-01 -5.64523280e-01
1.25334272e-02 -3.40823263e-01 6.46057785e-01 -4.28123087e-01
-2.95307666e-01 -2.48233557e-01 2.46939033e-01 -4.62730914e-01
-4.28703994e-01 -4.58275586e-01 -1.26773620e+00 -6.22427106e-01
-3.55066210e-01 1.47782192e-01 5.58583975e-01 5.58913648e-01
3.54519546e-01 4.28638846e-01 7.48440921e-01 -6.72032177e-01
-7.34469473e-01 -6.59873307e-01 -7.53211200e-01 5.54175019e-01
5.83131015e-01 -8.61256540e-01 -7.37376273e-01 -3.27330440e-01] | [7.25329065322876, 6.226616859436035] |
2b767930-6ef4-4aba-a43a-6f4d41833b9d | self-learning-locally-optimal-hypertuning | 2210.10783 | null | https://arxiv.org/abs/2210.10783v1 | https://arxiv.org/pdf/2210.10783v1.pdf | Self-learning locally-optimal hypertuning using maximum entropy, and comparison of machine learning approaches for estimating fatigue life in composite materials | Applications of Structural Health Monitoring (SHM) combined with Machine Learning (ML) techniques enhance real-time performance tracking and increase structural integrity awareness of civil, aerospace and automotive infrastructures. This SHM-ML synergy has gained popularity in the last years thanks to the anticipation of maintenance provided by arising ML algorithms and their ability of handling large quantities of data and considering their influence in the problem. In this paper we develop a novel ML nearest-neighbors-alike algorithm based on the principle of maximum entropy to predict fatigue damage (Palmgren-Miner index) in composite materials by processing the signals of Lamb Waves -- a non-destructive SHM technique -- with other meaningful features such as layup parameters and stiffness matrices calculated from the Classical Laminate Theory (CLT). The full data analysis cycle is applied to a dataset of delamination experiments in composites. The predictions achieve a good level of accuracy, similar to other ML algorithms, e.g. Neural Networks or Gradient-Boosted Trees, and computation times are of the same order of magnitude. The key advantages of our proposal are: (1) The automatic determination of all the parameters involved in the prediction, so no hyperparameters have to be set beforehand, which saves time devoted to hypertuning the model and also represents an advantage for autonomous, self-supervised SHM. (2) No training is required, which, in an \textit{online learning} context where streams of data are fed continuously to the model, avoids repeated training -- essential for reliable real-time, continuous monitoring. | ['Francisco Javier Montans', 'Ricardo Callado', 'Miguel Angel Sanz', 'Luis Saucedo-Mora', 'Miguel Diaz-Lago', 'Ismael Ben-Yelun'] | 2022-10-19 | null | null | null | null | ['self-learning'] | ['natural-language-processing'] | [ 2.62323231e-01 9.50100049e-02 1.41750559e-01 -1.51254401e-01
-7.37065196e-01 -2.48418283e-02 2.95591056e-01 6.15749300e-01
-4.76866364e-01 8.69150639e-01 -4.03361320e-01 -2.27900311e-01
-8.09865832e-01 -8.92053425e-01 -4.60243881e-01 -1.05044031e+00
-4.73339856e-01 7.89408326e-01 5.87610960e-01 -4.48231071e-01
3.60731095e-01 8.79640222e-01 -2.02989793e+00 3.82573187e-01
7.60029495e-01 1.28835583e+00 3.29965889e-01 5.36317348e-01
5.13972998e-01 5.65792263e-01 -3.52070957e-01 -4.17090319e-02
-9.60666984e-02 -1.12130821e-01 -7.26416051e-01 1.62423819e-01
-2.42823437e-01 5.71837127e-02 3.14708203e-01 4.25875217e-01
5.02915442e-01 1.32236332e-01 8.60032260e-01 -8.49432349e-01
4.17171240e-01 3.21703404e-01 -2.00072050e-01 -5.27619012e-02
3.90130103e-01 9.29124430e-02 7.23737419e-01 -9.25194085e-01
4.55441356e-01 8.61768007e-01 9.88128960e-01 7.45242089e-02
-1.30002034e+00 -7.89782405e-03 -3.47478032e-01 6.10307932e-01
-1.06413245e+00 -2.02165931e-01 9.02595580e-01 -7.23402500e-01
8.46330523e-01 4.97259349e-01 8.04517925e-01 7.36932516e-01
5.20726144e-01 3.52382332e-01 1.21422660e+00 -7.00574398e-01
5.37704349e-01 1.05803236e-01 -3.56541239e-02 6.77028477e-01
2.37062111e-01 3.60307097e-01 -3.57901961e-01 -6.07360564e-02
4.38475877e-01 -2.10831031e-01 -3.94661166e-02 -3.25272948e-01
-9.03240919e-01 7.08397269e-01 1.41134828e-01 5.74131906e-01
-6.60849988e-01 -4.46084253e-02 6.30410850e-01 5.77536047e-01
6.30096138e-01 4.23989534e-01 -5.73972702e-01 -1.57530323e-01
-1.10660517e+00 1.59329265e-01 8.08682621e-01 3.33343050e-03
8.78557682e-01 2.11919531e-01 5.39421916e-01 8.17378700e-01
3.04141164e-01 4.52117234e-01 4.71829444e-01 -6.77369297e-01
1.57184452e-01 6.71714246e-01 -1.64832681e-01 -1.09655011e+00
-7.70413578e-01 -7.56136715e-01 -9.79178131e-01 7.95954168e-01
2.39166737e-01 -1.33420542e-01 -3.48862261e-01 1.09928358e+00
6.12025023e-01 -1.67152718e-01 -8.62270966e-02 6.44356787e-01
2.64579337e-02 6.14504337e-01 -2.80189663e-01 -6.26711369e-01
1.00433671e+00 -2.39642695e-01 -5.07901430e-01 -9.77552757e-02
7.96306551e-01 -6.90642655e-01 7.65078545e-01 1.06412148e+00
-9.29934919e-01 -5.47314942e-01 -1.24708855e+00 7.33448148e-01
-4.53092605e-01 3.49202454e-02 2.54901558e-01 6.25745177e-01
-7.89557695e-01 1.23233199e+00 -8.42318058e-01 -4.87795323e-02
-2.23613754e-02 6.01279140e-01 -4.04462606e-01 7.19428286e-02
-1.10622990e+00 1.10663629e+00 5.94195604e-01 5.24828374e-01
-8.66517484e-01 -5.98120868e-01 -4.62739229e-01 -1.44881979e-01
3.62350762e-01 -2.67452091e-01 7.73142219e-01 -6.66128874e-01
-1.60366464e+00 5.92869401e-01 3.04011822e-01 -3.29950005e-01
6.26511216e-01 -3.17995608e-01 -4.52877641e-01 3.45228910e-01
-4.04393613e-01 -7.71455169e-02 9.81460929e-01 -1.53031480e+00
-2.24093989e-01 -2.47236282e-01 -3.54855657e-01 -4.26157027e-01
-3.05088371e-01 -3.67462486e-01 4.26680446e-01 -5.50508380e-01
3.47471744e-01 -8.61251891e-01 -4.02782172e-01 -5.04394650e-01
-3.19184601e-01 1.46578655e-01 9.04515803e-01 -8.56923223e-01
1.26269972e+00 -1.83000994e+00 2.73258328e-01 7.60507762e-01
-9.58527401e-02 2.73984224e-01 2.77773350e-01 1.17576361e+00
-3.92650992e-01 -5.64307451e-01 -7.06809700e-01 -2.13751961e-02
-3.31252098e-01 3.09218347e-01 5.60209341e-02 5.37706673e-01
-1.42827742e-02 2.29304627e-01 -5.91210127e-01 -5.97276747e-01
5.41095972e-01 2.22417742e-01 -3.49509895e-01 1.24359868e-01
-2.35896260e-01 4.64012176e-01 -1.75920546e-01 3.81399900e-01
3.57473582e-01 2.20299587e-01 1.34562686e-01 -2.96026051e-01
-2.68530428e-01 -3.01230282e-01 -1.39507782e+00 1.20665061e+00
-7.95399606e-01 3.41708690e-01 9.58776250e-02 -1.51724851e+00
1.31249595e+00 6.50199831e-01 8.89247477e-01 -4.13607001e-01
2.29523867e-01 6.42755866e-01 -1.53281584e-01 -8.40591311e-01
-8.31583142e-02 -1.90000638e-01 8.49616230e-02 3.08890849e-01
6.75697699e-02 -1.29796311e-01 2.72380114e-01 -3.42960656e-01
9.21537101e-01 6.91872835e-03 1.67889550e-01 -5.72778344e-01
1.11268973e+00 -1.48908868e-01 3.87265891e-01 2.54597336e-01
5.70798099e-01 1.35499492e-01 4.16714340e-01 -5.13043702e-01
-9.81250763e-01 -7.09945261e-01 -7.75607154e-02 6.50453985e-01
-4.26512331e-01 -4.38225567e-02 -8.23978841e-01 -2.44637281e-01
-4.23188172e-02 5.66322863e-01 -4.67175841e-01 -2.85123557e-01
-8.01877677e-01 -9.76805210e-01 -6.72370270e-02 1.50727659e-01
-7.97628518e-03 -1.02351606e+00 -8.60343933e-01 5.12977123e-01
6.91550970e-02 -6.27580881e-01 6.16251588e-01 4.17977601e-01
-1.54403770e+00 -1.13820338e+00 -2.39946097e-01 -6.04154944e-01
5.46109319e-01 -4.79931504e-01 9.80756819e-01 3.20133358e-01
-6.80504978e-01 4.09580231e-01 -3.87058288e-01 -1.80752054e-01
-9.38090146e-01 5.03847376e-02 3.98326293e-02 2.55942315e-01
-2.75314033e-01 -1.02589405e+00 -4.80692685e-01 4.05011237e-01
-8.79718363e-01 -3.92256886e-01 9.05282021e-01 8.46420348e-01
4.90997136e-01 4.60197806e-01 9.18983221e-01 -9.30205822e-01
3.65542769e-01 -3.10371578e-01 -7.07956612e-01 1.60437316e-01
-1.14896822e+00 -8.24813247e-02 6.66339159e-01 -2.58892447e-01
-9.59374726e-01 3.37804691e-03 -5.45130849e-01 7.77655020e-02
-3.03879410e-01 7.93124318e-01 -2.49776006e-01 -2.19139844e-01
6.79897606e-01 2.76001897e-02 3.54799300e-01 -7.37989426e-01
1.35187395e-02 5.54693282e-01 4.67704505e-01 -7.16090143e-01
8.69506955e-01 1.78229108e-01 5.65450668e-01 -1.13885200e+00
-4.43175733e-01 -4.13736969e-01 -8.49077404e-01 -8.91228497e-01
4.47857499e-01 -3.87553692e-01 -1.00518680e+00 7.14665353e-01
-8.97919357e-01 -1.32413298e-01 -4.80287373e-01 6.16433918e-01
-7.39326298e-01 6.59914970e-01 -8.41066599e-01 -1.32579291e+00
-3.54489237e-01 -9.47888315e-01 6.94393873e-01 -2.13761359e-01
-3.50044332e-02 -1.02110898e+00 2.11574566e-02 5.40506542e-01
3.16322267e-01 7.12934434e-01 1.17276990e+00 -4.66755569e-01
-1.91665024e-01 -7.28471994e-01 4.99421358e-01 8.38778734e-01
-1.52289912e-01 3.93342339e-02 -9.97000694e-01 -5.44762373e-01
5.08405626e-01 -1.13116242e-01 6.10508800e-01 2.87589908e-01
8.93536866e-01 -2.48772293e-01 -1.65763363e-01 -1.08383670e-01
1.75806344e+00 1.71695098e-01 5.94990253e-01 3.54144692e-01
1.80742458e-01 1.05998683e+00 9.11874592e-01 7.17686713e-01
-2.98310399e-01 6.77723765e-01 8.14401031e-01 -1.26009583e-01
1.05394490e-01 2.89820373e-01 3.94983530e-01 1.13892531e+00
-4.00746524e-01 1.05644673e-01 -9.58606720e-01 4.70202267e-01
-1.64599264e+00 -9.09362972e-01 -5.74574292e-01 2.59625244e+00
7.31074989e-01 6.71146691e-01 1.23773880e-01 1.04548907e+00
5.33920169e-01 -1.75916865e-01 -8.16867575e-02 -3.98978353e-01
9.94084254e-02 2.87242562e-01 2.37310603e-01 6.79984868e-01
-8.32051575e-01 1.16584636e-01 5.24894953e+00 9.51666713e-01
-9.97701049e-01 9.10260156e-02 3.41391832e-01 2.52214760e-01
4.14521694e-02 2.43331157e-02 -3.70097280e-01 4.91805464e-01
1.28421533e+00 4.31719095e-01 1.83305800e-01 7.31781185e-01
4.98773903e-01 -4.71040249e-01 -8.91473591e-01 6.11183107e-01
-1.26715243e-01 -1.14489126e+00 -4.92469728e-01 2.06335410e-01
3.84291440e-01 -2.18600303e-01 -2.51867414e-01 -1.92718580e-01
-3.98433298e-01 -4.93528962e-01 7.08261609e-01 8.79880130e-01
4.33609575e-01 -1.00818276e+00 1.10660398e+00 5.34558237e-01
-8.52246761e-01 -6.37400031e-01 -7.91319460e-02 -2.95366626e-02
5.11249423e-01 1.50076628e+00 -7.64431655e-01 1.00275993e+00
5.39283693e-01 2.46235430e-01 -3.73904169e-01 1.02282691e+00
1.48889320e-02 9.00033355e-01 -5.21766961e-01 1.02388330e-01
9.11789201e-03 -3.50497186e-01 7.05641150e-01 1.02180624e+00
3.35846514e-01 -3.05815399e-01 3.54450531e-02 2.95779943e-01
7.84895182e-01 1.65126666e-01 -3.14657182e-01 4.68287140e-01
2.65025616e-01 1.15484548e+00 -8.51453841e-01 4.68171760e-02
7.96827674e-03 2.01494455e-01 -4.20514524e-01 -9.81366038e-02
-5.27226806e-01 -2.34494522e-01 1.34137675e-01 6.72071993e-01
2.80460119e-01 -4.27705884e-01 -2.03849822e-01 -4.21872586e-01
6.54189810e-02 -6.97057009e-01 3.77824932e-01 -3.68327588e-01
-1.25120497e+00 6.32821321e-01 2.41318092e-01 -1.50808513e+00
-6.81993425e-01 -7.94021249e-01 -5.24875343e-01 3.62385929e-01
-1.12152064e+00 -1.03379214e+00 -4.90032090e-03 3.50505501e-01
2.34544531e-01 -9.28606912e-02 7.85241842e-01 4.41285610e-01
-4.92038965e-01 6.55507147e-02 4.04609561e-01 -2.65608639e-01
3.95432562e-01 -1.12394667e+00 -2.43677422e-01 6.54719472e-01
-2.08382830e-01 -8.95312205e-02 1.19070947e+00 -6.33225858e-01
-1.21408820e+00 -5.23535967e-01 7.47870326e-01 -2.53953785e-02
7.00921357e-01 -2.12714121e-01 -1.02644801e+00 -8.30833837e-02
-7.42308423e-02 -2.87515640e-01 6.72618151e-01 1.47062674e-01
2.97067672e-01 -5.98881304e-01 -9.59499538e-01 -6.70610890e-02
3.92261386e-01 -2.66928643e-01 -5.47907233e-01 5.73131502e-01
2.76320964e-01 4.47488911e-02 -1.50640976e+00 5.80833912e-01
4.38406765e-01 -1.20283687e+00 1.13888907e+00 -2.24610791e-01
2.15110302e-01 -5.31150438e-02 -3.00886482e-02 -1.02800071e+00
-1.29750326e-01 -5.08462906e-01 -2.24649623e-01 1.27305496e+00
3.69039297e-01 -9.05346811e-01 7.04471648e-01 1.75334916e-01
-3.90621573e-01 -1.18594837e+00 -1.33310401e+00 -9.11334217e-01
-2.93349087e-01 -5.46535671e-01 1.94388121e-01 5.79844058e-01
-1.23018645e-01 3.94425429e-02 -3.32286835e-01 2.71702439e-01
8.12970579e-01 1.35863975e-01 2.71685392e-01 -1.86258268e+00
-6.29434824e-01 -1.92383185e-01 -5.65511823e-01 -1.19971819e-01
-1.42477229e-01 -4.86769527e-01 5.66047207e-02 -1.13879156e+00
-5.60956120e-01 -6.60057485e-01 -2.97152936e-01 3.18620950e-01
3.81922632e-01 1.54414494e-02 -2.34068140e-01 2.17750579e-01
1.32500995e-02 5.09245098e-01 8.66947234e-01 1.24460720e-01
-1.66307632e-02 4.34930086e-01 2.16207370e-01 6.69893086e-01
6.12496257e-01 -4.99613762e-01 -3.81656259e-01 1.34559229e-01
4.51718003e-01 4.67577755e-01 5.39105713e-01 -1.54633033e+00
1.23975329e-01 1.02051578e-01 2.38301039e-01 -6.49167776e-01
4.19444025e-01 -1.18861043e+00 4.20541078e-01 9.46885824e-01
8.68795663e-02 1.09947942e-01 -7.78660327e-02 5.36627829e-01
-5.19861579e-01 -8.48835528e-01 9.08106565e-01 1.53958639e-02
-5.21315932e-01 -1.92249954e-01 -7.01588154e-01 -4.96537566e-01
1.00637078e+00 -4.95041668e-01 -1.33261187e-02 -9.85510051e-02
-9.71621513e-01 -3.03681999e-01 3.19759727e-01 -4.77895364e-02
4.93081748e-01 -8.32663178e-01 -7.81565070e-01 3.12581837e-01
-1.67582706e-01 -9.68640596e-02 6.94539666e-01 1.27818155e+00
-7.51477718e-01 1.10482275e-01 -1.69925302e-01 -7.18022585e-01
-1.09881151e+00 6.17134869e-01 2.47793451e-01 -4.81260657e-01
-6.07652485e-01 5.39417803e-01 -5.40737987e-01 -1.47800073e-01
-1.91532433e-01 -6.74789622e-02 -2.71998584e-01 4.75168139e-01
1.29248023e-01 9.04093266e-01 9.14133906e-01 -4.08767700e-01
-2.87770599e-01 6.46330893e-01 2.77976930e-01 5.11836959e-03
1.88681543e+00 3.19501981e-02 -5.06804585e-01 7.56681144e-01
1.06533229e+00 -8.67687538e-02 -1.01577020e+00 1.73926786e-01
5.33042610e-01 -1.32448629e-01 2.38951921e-01 -7.19861329e-01
-9.93612289e-01 8.93934548e-01 8.93775940e-01 5.10590971e-01
1.43855584e+00 -1.85659453e-01 7.31453180e-01 4.18558031e-01
5.09382069e-01 -1.52077103e+00 5.66315949e-02 1.41035346e-02
9.73948777e-01 -7.67512739e-01 3.41721654e-01 -4.18773770e-01
-1.70357838e-01 1.52350163e+00 1.67130604e-01 -8.80531669e-02
8.73487592e-01 4.36068892e-01 -4.98771444e-02 -2.56258875e-01
-8.09219778e-01 8.70512351e-02 1.37207717e-01 4.73099828e-01
6.44684434e-02 -1.42300084e-01 -5.06787777e-01 1.23543881e-01
7.62891248e-02 -2.04653651e-01 3.32056671e-01 1.17727983e+00
-7.83963919e-01 -1.47341669e+00 -7.00835407e-01 4.75575864e-01
-1.66220948e-01 4.88151491e-01 1.86571822e-01 8.58811259e-01
1.80591643e-01 9.97628868e-01 -5.08650661e-01 -4.51160282e-01
4.65438783e-01 5.17657548e-02 3.35997730e-01 -1.67056382e-01
-7.11358905e-01 7.51875415e-02 2.27906615e-01 -5.27078032e-01
-4.06007797e-01 -1.02000630e+00 -1.00615954e+00 -8.15447494e-02
-6.57931864e-01 4.74482358e-01 9.30263579e-01 1.11660087e+00
1.51236949e-03 5.34840763e-01 1.09980476e+00 -1.18718660e+00
-6.90263987e-01 -9.05448556e-01 -8.00052822e-01 1.01039514e-01
6.42002523e-02 -8.66266787e-01 -5.10022283e-01 1.09192893e-01] | [6.678933143615723, 2.6414856910705566] |
80ea58ca-ce7d-4332-adaa-97992f86e53b | stand-alone-inter-frame-attention-in-video-1 | 2206.06931 | null | https://arxiv.org/abs/2206.06931v1 | https://arxiv.org/pdf/2206.06931v1.pdf | Stand-Alone Inter-Frame Attention in Video Models | Motion, as the uniqueness of a video, has been critical to the development of video understanding models. Modern deep learning models leverage motion by either executing spatio-temporal 3D convolutions, factorizing 3D convolutions into spatial and temporal convolutions separately, or computing self-attention along temporal dimension. The implicit assumption behind such successes is that the feature maps across consecutive frames can be nicely aggregated. Nevertheless, the assumption may not always hold especially for the regions with large deformation. In this paper, we present a new recipe of inter-frame attention block, namely Stand-alone Inter-Frame Attention (SIFA), that novelly delves into the deformation across frames to estimate local self-attention on each spatial location. Technically, SIFA remoulds the deformable design via re-scaling the offset predictions by the difference between two frames. Taking each spatial location in the current frame as the query, the locally deformable neighbors in the next frame are regarded as the keys/values. Then, SIFA measures the similarity between query and keys as stand-alone attention to weighted average the values for temporal aggregation. We further plug SIFA block into ConvNets and Vision Transformer, respectively, to devise SIFA-Net and SIFA-Transformer. Extensive experiments conducted on four video datasets demonstrate the superiority of SIFA-Net and SIFA-Transformer as stronger backbones. More remarkably, SIFA-Transformer achieves an accuracy of 83.1% on Kinetics-400 dataset. Source code is available at \url{https://github.com/FuchenUSTC/SIFA}. | ['Tao Mei', 'Jiebo Luo', 'Ting Yao', 'Yingwei Pan', 'Zhaofan Qiu', 'Fuchen Long'] | 2022-06-14 | stand-alone-inter-frame-attention-in-video | http://openaccess.thecvf.com//content/CVPR2022/html/Long_Stand-Alone_Inter-Frame_Attention_in_Video_Models_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Long_Stand-Alone_Inter-Frame_Attention_in_Video_Models_CVPR_2022_paper.pdf | cvpr-2022-1 | ['action-classification'] | ['computer-vision'] | [-0.26087043 -0.29416034 -0.07403517 -0.30848968 -0.40505713 -0.56211925
0.4872163 -0.46275002 -0.4020021 0.4341501 0.29445583 -0.0210753
0.07518449 -0.72683495 -1.0305592 -0.7891805 -0.02675223 -0.17317224
0.61379445 -0.06677594 0.00643632 0.41282228 -1.1873306 0.42888266
0.6583695 1.3065356 0.32144117 0.6472825 0.07347704 0.8247248
-0.18007864 -0.26252073 0.325673 -0.22344747 -0.8374543 -0.10205564
0.6496453 -0.85569125 -0.8369936 0.93859255 0.21031529 0.2997878
0.30958977 -1.1622795 -0.89662737 0.4957584 -0.65979844 0.71355677
0.06655709 0.38669333 0.9109769 -1.0875996 0.5841213 1.1330429
0.6787606 0.55953395 -0.84482193 -0.54554856 0.5527702 0.30832312
-1.3380432 -0.40054595 0.744647 -0.62061286 0.9762843 0.140991
0.89728487 0.9586606 0.19026889 1.0050925 0.5519397 0.24381989
-0.09969383 -0.5339298 0.03904637 0.7898402 -0.02372075 0.03466413
-0.6271838 0.31160134 1.2716058 0.30587885 -0.50830275 -0.1160052
-1.4576752 0.45520407 0.70717776 0.4003419 -0.4250533 0.481004
0.23588766 0.08971984 0.49140006 -0.1294947 -0.46058202 -0.14244796
-0.9549792 0.27289957 0.19639915 0.8577422 0.81990063 0.0335619
-0.31533456 0.44300514 0.15949722 0.29330382 0.48147556 -1.1955911
0.34090593 0.52551925 0.05413993 -1.1319999 -0.2701127 -0.2411634
-0.87774897 -0.03245857 0.5526557 -0.06752557 -0.9235847 1.9419385
0.38080707 0.7829971 -0.27653468 1.2084343 0.8887534 0.63463694
0.11992324 0.01129789 1.116132 -1.2323186 -0.40375987 0.05367051
0.47431695 -0.47458473 1.0006217 -0.04187757 -1.2768018 -0.91087705
-0.7887992 -0.37371758 -0.19803134 0.14043179 0.583723 -0.01772357
-1.1747767 0.77690524 -1.1857691 -0.11354457 0.57485574 0.36496934
-0.3973248 0.20330714 -1.2265574 0.42105103 0.16874817 0.33585063
-0.92314696 -0.963314 -0.7334186 0.13300534 0.2559936 -0.9637584
1.1054641 -1.1724039 -1.1254097 0.66245604 -0.44883233 -0.47107452
0.6289281 -0.4796379 -0.25791195 0.12509729 0.20305873 0.8984945
0.9286595 -0.8310497 -0.64303565 -0.17981894 0.38215655 0.07149628
-0.27564925 -0.19104998 -0.8341024 -0.8075211 0.14273033 -0.75189835
0.07240329 0.34214005 -0.2586727 -0.30762953 1.0436519 -0.6573089
1.564601 -2.2294471 0.36761644 -0.16491526 0.5837656 0.4217698
-0.178476 -0.02420318 -0.14456907 0.06762596 -0.13312136 -0.20309353
-0.20153515 0.09705036 -0.3897712 0.39800036 0.49079508 1.3417867
-0.97411424 -0.32406566 0.38595465 0.6466231 -0.6866529 0.02043617
-0.14992125 0.37033015 -0.62442034 0.59638494 0.70611006 -0.49653107
-0.1574511 -0.5834638 -0.23062956 0.11060981 -0.8788502 1.833966
-0.07376205 0.68813604 -0.13768291 -0.96127087 0.49756378 0.13576308
0.9004673 -0.54590386 0.0970468 -0.0438533 -0.12896256 -0.68684083
0.33833706 0.2468341 0.21977673 0.2767013 0.1259983 0.4912728
0.11197905 0.23949693 1.0348969 0.5466575 -0.09047122 -0.18536527
0.6775258 -0.256584 0.6693047 0.42280698 -0.48841965 0.90042233
0.34190416 -0.9685272 -0.96248245 -1.0762979 0.16439532 1.0682116
0.39644927 -0.33830762 -0.88587433 -0.76291895 0.13702375 0.01868143
-0.86124676 -0.1843579 -0.8076613 -0.50953686 0.35337326 0.9879953
0.78742445 -0.986536 -0.63677454 0.24406987 -0.41852334 -1.0879416
-1.0205352 -0.3128504 -0.5860038 -0.91983086 -0.8186921 -0.6870847
0.4932744 0.64065284 0.9267841 0.31262645 0.07961596 0.1674278
-0.288494 0.13509865 0.2541206 -0.03044708 0.07669317 0.3276259
0.44053987 -0.6721196 -1.138542 0.4194902 -0.86055714 0.2602811
0.3383302 0.7118579 0.63566303 -0.2788012 0.31411612 -0.5364394
0.10831785 -0.5448876 -0.36837488 0.14461546 0.07225325 -0.06060629
0.42522478 -0.510356 -0.7918545 0.15470424 -0.22681598 -1.0261855
-0.03555605 0.40009627 -0.14303681 -0.05773328 0.21995313 0.3450514
-0.04581438 -0.2518706 0.29972047 0.19441386 0.7639656 -0.42139325
0.71485955 0.8302183 -0.36685047 -0.61045086 -0.90172607 -0.3141068
-0.78194153 -0.44910878 1.2474498 -0.9945313 -0.8028645 0.7234762
-1.1417245 -0.61525923 -0.11265157 0.35591945 -0.46909568 0.33545715
-0.606343 -0.36832225 -0.2673833 -1.3023562 1.0680441 0.34824297
-0.24049121 -1.0698793 -0.10970256 0.26954997 0.3944642 0.16056931
0.60961235 -0.2555789 -0.9850928 0.0742128 -0.4570381 0.18471046
0.17443912 0.24224149 -1.0173571 -0.17966415 -0.30170342 0.0134274
1.090812 0.82349485 1.4056218 -0.06271198 -0.19416532 1.0282186
1.18248 0.15193965 0.7308196 0.30403942 1.1249007 0.15572242
0.34664848 0.31151697 0.60426784 0.8323094 0.36535895 -0.01918769
-0.28034103 -0.21366024 0.45644003 0.7883633 -0.56119794 -0.1115403
-0.6666597 0.49665684 -1.9835327 -1.3034104 0.02395836 2.04151
0.46759203 0.09999418 0.20533493 -0.22822404 0.61578715 0.45608982
-0.73370224 0.1907368 -0.24706666 -0.0144885 0.33555555 0.58272123
-1.366805 1.0218965 5.089051 0.74516577 -1.3841217 0.1827915
0.8105709 -0.28323928 -0.27446607 -0.07326739 -0.6893264 0.7399767
0.5482341 -0.04310405 0.4163789 0.522822 0.5179822 0.14333098
-0.93760216 0.77420884 -0.19671233 -1.6499497 0.22924095 -0.06060341
0.72469485 0.19662225 0.20418803 0.0795773 0.06632875 -0.8660312
1.0550015 0.67637914 0.7530587 -0.42502546 0.5377465 0.07114153
-1.6105276 -0.00695723 -0.3214813 -0.0683223 0.26850206 0.3278832
-0.01910168 0.47913286 1.1291208 1.2427351 -0.31679112 0.8222884
0.12958205 0.4372627 -0.35015485 0.3831633 0.4500725 -0.23908737
0.42496967 1.0917906 0.39705962 0.21640769 0.07485922 0.8560957
-0.07694959 -0.35177955 -0.30763525 0.0434855 0.48436937 1.1599557
-0.60313654 -0.48633894 -0.71500236 1.037324 0.45573545 0.62647915
-1.2493256 -0.08852123 1.0958722 0.32681808 0.66740984 -0.35920727
-0.06335661 -1.3120543 0.18270919 -0.4599367 0.31835696 -0.82371557
-1.1021706 0.64575636 -0.19749548 -1.3793644 0.09295002 -0.5363364
-0.86395335 0.7111319 -1.3613164 -1.2518059 -0.61520964 0.8077744
0.7142195 0.11352558 0.27966726 0.410584 -0.7743189 0.687507
-0.20214692 0.37525848 0.56212205 -0.98576474 0.68241984 0.93722856
0.08763057 0.61466324 0.211302 -0.54146254 -1.269299 -1.3459282
0.6598309 -0.5760126 0.8564118 -0.05901741 -1.2037739 0.74450296
0.03013813 0.52321386 0.21276374 -0.52920496 -0.36787203 -0.12009539
-0.65887785 0.5150609 1.4662522 -0.5956579 -0.22125547 0.12198329
0.8761136 -0.5966034 -0.9536379 0.35540015 0.7239818 -1.2352566
1.2294062 -0.6785822 0.8018686 -0.6464753 -0.13165933 -0.86301345
-0.62578833 -0.6157914 -0.41162512 0.9940243 0.16921909 -0.43265817
0.76307327 0.6188123 -0.2880899 -1.079846 -0.88492984 -0.4774096
0.14577255 -0.3927234 0.64810413 0.93203294 -0.37576547 -0.05764205
-0.39554763 0.23497453 0.2967597 0.03233772 0.63596886 -0.83496225
-0.3184312 -0.6412803 -0.56380624 -1.5003215 0.11907959 -0.72623885
-0.27532294 -1.239255 0.19651024 -0.19634742 -0.533147 0.57879
-0.41715407 0.4119495 0.24769069 0.32654044 -0.6049261 0.43984815
1.4814818 -0.05300407 -0.22792347 -0.17691663 -0.41692814 0.7599449
0.64008677 0.02805524 -0.4326857 -0.82977694 -0.07644299 -0.11524212
0.74327403 -1.01415 0.27112883 -0.25565395 0.65570676 -0.59794456
0.3218596 -0.6200412 0.3826824 0.29382166 -0.22467381 0.32134742
0.11865729 0.5869122 -0.25490955 0.31913626 0.58414966 -0.12798753
-1.1682127 0.94544363 -0.1298506 0.06246376 1.0931828 -0.37868023
-0.3370432 -0.30333394 -0.87565583 0.17943633 0.5225152 0.4697224
0.7031463 -1.4472439 -0.41164568 0.33625153 -0.18519247 0.18018897
0.7791677 1.1897367 -0.5114986 0.36443514 -0.18922958 -0.6785181
-1.0123981 0.58372617 0.5414067 0.0879349 -0.7534616 1.1615727
0.7165982 0.2499671 0.04688212 -0.6327587 -0.05444765 -0.08441114
0.6755151 0.07222401 -0.16725753 -0.90971744 -0.37437755 0.8524928
-0.19921084 0.2255628 1.3473281 -0.2623681 -0.01717194 0.1989591
1.3372577 -0.27091503 -1.9607307 -0.30252495 -0.36419845 -0.59947526
0.04371794 -0.23147583 -1.7254504 0.86756 0.29568875 0.06312787
1.2526908 0.07372298 1.0207064 -0.09769467 0.14939277 -0.6436991
0.1398599 0.48000494 0.8012544 -1.0761077 -0.24956577 -0.28952187
-0.6426743 0.9959038 0.83644384 -0.27910784 0.7901709 0.19096768
-0.03900542 -0.14154257 -0.7937615 -0.17525809 0.48476034 0.31979203
0.48309526 -0.11667235 0.06216976 0.63323265 0.05629984 0.20884252
0.13780637 0.7918533 -0.28218892 -0.6195058 0.06752123 0.2872465
-0.29293808 -0.0358194 -0.15586258 0.5974337 0.2871346 0.39978313
0.32732937 -0.7247792 0.14321804 -0.12700523 0.3025276 -0.07424499
-0.42434272 0.20999661 -0.39789113 -0.98069793 -0.6058309 -0.6624805
-1.4251761 -0.5766248 0.03859846 -0.27856594 -0.03377705 0.93884945
0.6001705 0.5880038 0.46668342 -1.0742412 -0.13725536 -0.6904157
-0.08122817 0.51937413 0.54753953 -0.89496255 -0.2531262 0.36690518] | [8.998528480529785, 0.23565573990345] |
7f0a3139-e1b2-4e8d-8caa-d381616c56e1 | complementary-network-with-adaptive-receptive | 2001.03893 | null | https://arxiv.org/abs/2001.03893v1 | https://arxiv.org/pdf/2001.03893v1.pdf | Complementary Network with Adaptive Receptive Fields for Melanoma Segmentation | Automatic melanoma segmentation in dermoscopic images is essential in computer-aided diagnosis of skin cancer. Existing methods may suffer from the hole and shrink problems with limited segmentation performance. To tackle these issues, we propose a novel complementary network with adaptive receptive filed learning. Instead of regarding the segmentation task independently, we introduce a foreground network to detect melanoma lesions and a background network to mask non-melanoma regions. Moreover, we propose adaptive atrous convolution (AAC) and knowledge aggregation module (KAM) to fill holes and alleviate the shrink problems. AAC explicitly controls the receptive field at multiple scales and KAM convolves shallow feature maps by dilated convolutions with adaptive receptive fields, which are adjusted according to deep feature maps. In addition, a novel mutual loss is proposed to utilize the dependency between the foreground and background networks, thereby enabling the reciprocally influence within these two networks. Consequently, this mutual training strategy enables the semi-supervised learning and improve the boundary-sensitivity. Training with Skin Imaging Collaboration (ISIC) 2018 skin lesion segmentation dataset, our method achieves a dice co-efficient of 86.4% and shows better performance compared with state-of-the-art melanoma segmentation methods. | ['Zhen Chen', 'Yixuan Yuan', 'Xiaoqing Guo'] | 2020-01-12 | null | null | null | null | ['skin-lesion-segmentation'] | ['medical'] | [ 6.65469408e-01 2.27971762e-01 -2.28496134e-01 -2.13998452e-01
-4.24937963e-01 -2.85305887e-01 2.55897284e-01 2.59163906e-03
-7.79179335e-01 5.94018459e-01 -6.91175312e-02 -2.92628825e-01
-1.52580654e-02 -8.25632155e-01 -2.78087705e-01 -1.07957625e+00
4.59359437e-01 -1.46862999e-01 6.08261347e-01 7.30093494e-02
1.54621512e-01 5.57775438e-01 -9.85395551e-01 3.60069275e-01
1.45181274e+00 1.02883422e+00 2.07167163e-01 5.92212260e-01
-4.11613792e-01 6.18217170e-01 -2.44390473e-01 -4.49194193e-01
1.65337324e-01 -3.43018293e-01 -8.56552362e-01 3.05670530e-01
3.39812964e-01 -2.26474598e-01 -1.27681091e-01 1.24746239e+00
4.83456373e-01 -2.66437232e-01 7.46780455e-01 -5.52328169e-01
-4.83117163e-01 3.13632429e-01 -1.07975304e+00 2.51111448e-01
-3.73071223e-01 2.12780878e-01 4.92861658e-01 -4.19543028e-01
4.73735213e-01 7.16602623e-01 8.75618756e-01 8.40286613e-01
-1.05337358e+00 -3.91880751e-01 1.56871885e-01 -1.18912421e-02
-1.23567176e+00 -1.40199542e-01 4.76003200e-01 -3.60813379e-01
4.96175617e-01 4.80416179e-01 7.63856411e-01 7.24656284e-01
1.77382201e-01 7.88902760e-01 1.11769319e+00 -4.90789771e-01
4.81719524e-02 2.82877564e-01 6.89064339e-02 8.82662594e-01
2.08825357e-02 -2.41414890e-01 7.72888819e-03 1.87070280e-01
1.18621874e+00 9.73554477e-02 -2.93403625e-01 -7.49638081e-02
-9.32495058e-01 6.15078926e-01 8.48863244e-01 3.89702797e-01
-3.62571895e-01 2.96631275e-04 3.39071065e-01 -1.53950825e-01
3.92046750e-01 2.66636848e-01 -1.13418169e-01 3.97454709e-01
-9.05988932e-01 -3.43646973e-01 4.01860327e-01 2.84321308e-01
5.52605748e-01 -3.09133738e-01 -4.65403467e-01 9.56936777e-01
2.46576354e-01 6.51687160e-02 4.65447605e-01 -6.00220740e-01
1.47133201e-01 1.04760158e+00 -2.22431019e-01 -6.99299991e-01
-5.63297033e-01 -5.23841679e-01 -1.24195731e+00 3.07436198e-01
7.08671570e-01 -1.74847797e-01 -1.43830359e+00 1.23816586e+00
7.94665396e-01 1.66458428e-01 -1.45762697e-01 9.97610629e-01
8.07343185e-01 1.88612476e-01 3.59421968e-01 -1.87413573e-01
1.37083900e+00 -1.19185781e+00 -7.45325029e-01 5.66555709e-02
7.53985107e-01 -7.22854137e-01 8.12954366e-01 2.02156439e-01
-1.11548400e+00 -3.18840683e-01 -8.67305219e-01 -2.16719043e-02
-2.30528906e-01 4.70052004e-01 8.05850387e-01 6.93672001e-01
-1.03183877e+00 3.32066745e-01 -1.08213639e+00 -3.86141717e-01
8.28108191e-01 5.75225472e-01 -1.90832555e-01 1.50338084e-01
-9.34895217e-01 7.02539265e-01 2.53377378e-01 3.98819715e-01
-3.91413063e-01 -7.82758296e-01 -6.43418908e-01 -3.67578685e-01
2.67293215e-01 -7.00969815e-01 7.21107662e-01 -1.36745882e+00
-1.59708250e+00 9.45579708e-01 -1.84046924e-01 -4.36099738e-01
7.29351759e-01 1.50408983e-01 -2.88949788e-01 5.29471159e-01
-1.31764486e-01 8.57838631e-01 9.13318217e-01 -1.03169239e+00
-6.84589922e-01 -3.31129849e-01 7.96750002e-03 4.63925272e-01
-6.73277915e-01 -1.21550053e-01 -6.61193311e-01 -6.12930357e-01
9.27644894e-02 -5.16240597e-01 -6.22650146e-01 4.41272557e-01
-7.25610733e-01 6.84006959e-02 6.55645311e-01 -9.74485397e-01
1.57101154e+00 -2.16475391e+00 1.13973834e-01 5.51332116e-01
4.72122252e-01 5.00860572e-01 -1.27175346e-01 -3.73690784e-01
9.09463167e-02 1.71486840e-01 -5.66193938e-01 -1.82166636e-01
-5.36975205e-01 -2.24924926e-02 3.72496724e-01 5.97934783e-01
5.49717605e-01 8.25483322e-01 -6.66896880e-01 -9.48367715e-01
4.47134495e-01 5.83903253e-01 -3.17605168e-01 1.95648447e-02
-8.07948261e-02 5.60433269e-01 -6.00134075e-01 9.11396563e-01
9.67242718e-01 -3.69082600e-01 2.02357948e-01 -3.41294289e-01
-1.25227317e-01 -3.25702816e-01 -1.02639866e+00 1.49569809e+00
-2.76388913e-01 2.12385982e-01 6.11651361e-01 -7.15455055e-01
6.50324702e-01 1.69026464e-01 4.44970846e-01 -5.58857501e-01
2.31796935e-01 2.13240981e-01 1.97580144e-01 -7.03440428e-01
7.83405006e-02 -8.74342099e-02 4.91411716e-01 -6.71022162e-02
-1.88752338e-01 1.02466501e-01 6.30004033e-02 -1.30330278e-02
1.09280193e+00 1.25268213e-02 2.11457953e-01 -2.03654230e-01
9.35839236e-01 -7.73580447e-02 5.97978592e-01 3.78940552e-01
-3.64884377e-01 5.56293190e-01 6.38524294e-01 -3.54678839e-01
-6.73268020e-01 -1.00282216e+00 -5.11866987e-01 6.60005689e-01
4.24048185e-01 6.42577559e-02 -1.12317789e+00 -1.02705419e+00
-1.09716564e-01 1.18556134e-01 -1.03174365e+00 1.15886718e-01
-4.91311818e-01 -1.10044718e+00 5.80053389e-01 5.73891521e-01
8.17487419e-01 -8.72362792e-01 -4.15086359e-01 -2.87227463e-02
5.46823964e-02 -8.00764322e-01 -6.34077191e-01 1.28841326e-01
-7.50894189e-01 -1.34239817e+00 -1.03520525e+00 -9.69994426e-01
1.28366375e+00 7.09419101e-02 5.49249589e-01 1.80020496e-01
-7.85818577e-01 -1.92844141e-02 -1.12092525e-01 -2.08362073e-01
-3.78482968e-01 2.22067893e-01 -4.65056568e-01 2.73011118e-01
2.77721882e-01 -2.39574775e-01 -9.29220021e-01 3.26244116e-01
-1.11012435e+00 3.29077780e-01 1.04071689e+00 1.12322140e+00
8.57586503e-01 3.75723615e-02 3.71785820e-01 -1.01632810e+00
3.43003213e-01 -2.79430062e-01 -4.74191725e-01 4.86259729e-01
-3.41848969e-01 -3.20015132e-01 4.40700620e-01 -4.30798322e-01
-1.30966699e+00 3.20470780e-01 -1.51839063e-01 -1.97348207e-01
-2.37154052e-01 1.15872599e-01 -1.20545411e-02 -5.37066281e-01
6.58278942e-01 1.66341923e-02 2.90499061e-01 -4.00117710e-02
9.56228822e-02 6.94423318e-01 3.21481049e-01 5.06161079e-02
5.46072125e-01 6.45563424e-01 1.30830452e-01 -7.27630496e-01
-7.84435868e-01 -5.45042336e-01 -8.11041296e-01 -2.26875752e-01
1.23563135e+00 -7.80563533e-01 -4.89464581e-01 9.41157103e-01
-9.01250899e-01 -5.45646012e-01 -2.85236031e-01 3.13417792e-01
-6.35062605e-02 4.83158916e-01 -9.19667840e-01 -6.97351933e-01
-5.31185746e-01 -1.11620867e+00 8.87827218e-01 8.26918244e-01
4.19079466e-03 -1.23853946e+00 -2.67183661e-01 4.53619570e-01
4.91731763e-01 4.92161274e-01 7.23608971e-01 -3.40752095e-01
-4.40435320e-01 -1.41015589e-01 -6.64655805e-01 6.33886278e-01
3.17341596e-01 1.65376157e-01 -9.32601810e-01 -1.48400530e-01
-3.17807108e-01 -1.23863131e-01 1.34805918e+00 8.25932622e-01
1.63151681e+00 -7.85891786e-02 -5.26986301e-01 1.00744689e+00
1.44390821e+00 2.99553704e-02 7.63819098e-01 1.59142852e-01
9.15091336e-01 7.35098600e-01 4.17545468e-01 1.59117401e-01
2.20221788e-01 2.01247394e-01 4.80683655e-01 -8.88810039e-01
-3.91777605e-01 1.03824504e-01 -1.33711040e-01 4.39762384e-01
-2.87650555e-01 -8.03955644e-03 -6.70500278e-01 6.30026639e-01
-1.69351530e+00 -5.27915776e-01 -2.39161447e-01 2.05188560e+00
1.11074471e+00 6.90562129e-02 -2.84710433e-02 -1.89526677e-01
1.02668428e+00 4.73579131e-02 -7.67555714e-01 -2.05689475e-01
-1.48773566e-01 4.70518678e-01 6.33331656e-01 5.98925650e-01
-1.41821814e+00 9.29446816e-01 5.79470396e+00 1.32475793e+00
-1.21589768e+00 8.49179476e-02 9.78280842e-01 2.24766042e-02
-3.40761900e-01 -2.51904130e-01 -7.59596705e-01 4.21271712e-01
2.55190492e-01 6.39473259e-01 1.74819395e-01 3.69683892e-01
1.83005899e-03 -3.63872558e-01 -5.22339344e-01 6.81464791e-01
-1.02884822e-01 -1.38721192e+00 -4.71194051e-02 1.91902742e-01
8.31358790e-01 -2.91032523e-01 1.54984102e-01 -2.95212954e-01
-3.47919911e-02 -1.18371093e+00 1.51159558e-02 9.40395534e-01
1.14520848e+00 -6.71618044e-01 9.66169357e-01 4.34655100e-02
-9.99304831e-01 -3.74794491e-02 -2.97244638e-01 3.69373918e-01
-1.26048282e-01 7.97720730e-01 -8.08956265e-01 3.57913852e-01
3.22133213e-01 6.51131630e-01 -6.67155564e-01 1.19946063e+00
-1.74914762e-01 5.63399255e-01 -2.29005963e-01 -1.42901674e-01
3.52968752e-01 -2.31991738e-01 4.56640303e-01 1.25923908e+00
-8.27345029e-02 -1.19640701e-01 -1.33340478e-01 9.12416995e-01
5.62839443e-03 2.35853568e-01 3.03836502e-02 7.61157945e-02
2.27177247e-01 1.74358571e+00 -9.17281091e-01 -5.26600927e-02
-2.03168154e-01 1.10156691e+00 1.47259116e-01 4.90911037e-01
-8.11203063e-01 -6.11874759e-01 3.63246948e-01 3.14302295e-01
5.57865798e-02 2.88983196e-01 -5.94875813e-01 -9.17014301e-01
-1.13522047e-02 -4.95869190e-01 2.94629991e-01 -1.34680420e-01
-1.35010898e+00 3.62125248e-01 -5.09471536e-01 -1.02314782e+00
4.97183114e-01 -7.29073644e-01 -9.26283896e-01 1.00720859e+00
-1.90630066e+00 -1.39819860e+00 -5.94840527e-01 5.51662266e-01
2.29955211e-01 1.32911913e-02 7.09722936e-01 1.31738842e-01
-8.76668096e-01 8.24446857e-01 -5.63853746e-03 2.99540669e-01
8.09581935e-01 -1.44635296e+00 -1.69998221e-02 7.11053073e-01
-6.40868902e-01 5.14488161e-01 -1.13688335e-01 -6.97955191e-01
-7.98375189e-01 -1.26778531e+00 4.10817772e-01 -6.46402687e-02
6.89062715e-01 -3.28954197e-02 -8.89144719e-01 7.57627636e-02
1.86123237e-01 2.71883190e-01 9.25073147e-01 -1.75613627e-01
5.65723106e-02 -2.48617545e-01 -1.42817163e+00 6.43873334e-01
6.90041065e-01 -2.24801362e-01 -4.72863354e-02 4.06107754e-01
4.97174770e-01 -6.69822991e-01 -9.43802893e-01 6.71758711e-01
5.65491676e-01 -9.36956644e-01 8.38245153e-01 -2.30798572e-01
4.53529209e-01 -3.08478415e-01 3.27706635e-01 -9.23049629e-01
-3.34269047e-01 -3.49278539e-01 5.54176457e-02 1.13675225e+00
3.85258406e-01 -7.15934455e-01 9.75682259e-01 3.23462844e-01
-2.36567587e-01 -1.37032187e+00 -9.37864184e-01 -6.44556582e-02
1.52938202e-01 9.74986106e-02 2.41542593e-01 8.01310182e-01
-2.70656288e-01 -1.73487365e-01 2.27992740e-02 1.20475814e-01
6.53659463e-01 -3.17354500e-01 3.16071868e-01 -8.65960836e-01
-1.56295493e-01 -7.42552042e-01 -2.95295388e-01 -8.10485721e-01
-1.80714563e-01 -8.20624292e-01 -1.31328970e-01 -1.53067815e+00
3.34882438e-01 -6.27745092e-01 -5.12011349e-01 6.49217129e-01
-4.92093563e-01 4.73375827e-01 -2.45382950e-01 1.64794505e-01
-5.15647948e-01 1.30702466e-01 1.91164517e+00 -2.28828683e-01
-4.05617416e-01 1.39587015e-01 -7.55078912e-01 9.40236628e-01
8.03060949e-01 4.18805033e-02 -1.10261515e-01 -2.78721273e-01
-1.60797924e-01 -2.23102018e-01 6.33873284e-01 -9.43315029e-01
4.66192812e-01 -1.30607054e-01 8.80975544e-01 -4.54450071e-01
6.29960224e-02 -5.06128550e-01 -3.44705790e-01 6.20129883e-01
-3.46309930e-01 -7.05125511e-01 2.23246217e-01 5.79466105e-01
-1.60292715e-01 -2.27212191e-01 1.11430144e+00 -2.65846133e-01
-5.29288471e-01 4.09618974e-01 -2.83239335e-01 -3.63412648e-01
1.37150049e+00 -5.16064703e-01 -3.39438289e-01 3.03463459e-01
-8.67941320e-01 4.84444559e-01 4.33070630e-01 -1.09700672e-01
5.89857101e-01 -9.81441796e-01 -5.48893631e-01 3.03089738e-01
-6.17971830e-02 3.39174211e-01 8.10224414e-01 1.45368838e+00
-8.19812953e-01 -2.90449616e-02 -1.14604324e-01 -5.76343715e-01
-1.27907884e+00 5.07634059e-02 7.95246243e-01 -6.05095685e-01
-3.13100547e-01 1.32583129e+00 3.46003741e-01 -4.19862956e-01
2.96404958e-01 -2.97644973e-01 -4.46054608e-01 -7.40890205e-02
5.06802142e-01 3.18250120e-01 -1.21278010e-01 -3.59894544e-01
-2.66289622e-01 6.87527835e-01 -4.63686109e-01 4.22539026e-01
8.39882195e-01 -8.90255868e-02 -4.82434571e-01 -5.31193316e-02
7.73431361e-01 -2.33737621e-02 -1.48423588e+00 -1.50112942e-01
-4.59881425e-01 -3.22201073e-01 2.96838284e-01 -1.06279409e+00
-1.12916768e+00 9.28608239e-01 8.93845260e-01 1.32308885e-01
1.48378623e+00 -2.54296362e-01 8.94649327e-01 -1.07487574e-01
-2.60495275e-01 -1.28360081e+00 5.56910932e-02 1.14013314e-01
3.84640515e-01 -1.28255832e+00 -7.98869412e-03 -8.27778935e-01
-7.70874321e-01 1.28890574e+00 9.01675642e-01 -1.37563378e-01
5.30367494e-01 4.91231889e-01 2.44589552e-01 1.76225393e-03
-1.34709105e-01 -3.51409316e-01 5.40417433e-01 6.22006118e-01
3.44341278e-01 1.37476340e-01 -5.12437344e-01 4.95386839e-01
4.83528525e-01 -4.57761772e-02 1.30131096e-01 4.81893271e-01
-4.58541095e-01 -1.10542345e+00 -1.67145580e-01 8.19456875e-01
-6.18592739e-01 -2.30259672e-01 -3.42657387e-01 7.89421141e-01
6.10885382e-01 5.89264810e-01 1.90187305e-01 -2.35485807e-01
-1.38948202e-01 -4.43336993e-01 4.62970197e-01 -5.35009742e-01
-9.11633968e-01 4.27015185e-01 -2.14568093e-01 -3.92198294e-01
-5.84466159e-01 -4.51570034e-01 -1.34208834e+00 8.71084034e-02
-5.89080155e-01 -1.74539924e-01 4.85008597e-01 9.09352243e-01
9.31617711e-03 8.29088986e-01 5.41195035e-01 -3.33188474e-01
-3.61161798e-01 -9.09025490e-01 -7.18255460e-01 5.01861162e-02
2.89700091e-01 -2.68008649e-01 -2.30754495e-01 -2.52745226e-02] | [15.607069969177246, -2.9114456176757812] |
7ca3aed5-6d38-48d3-8c27-3b47217a45d2 | analysis-of-sparse-subspace-clustering | 2204.00723 | null | https://arxiv.org/abs/2204.00723v1 | https://arxiv.org/pdf/2204.00723v1.pdf | Analysis of Sparse Subspace Clustering: Experiments and Random Projection | Clustering can be defined as the process of assembling objects into a number of groups whose elements are similar to each other in some manner. As a technique that is used in many domains, such as face clustering, plant categorization, image segmentation, document classification, clustering is considered one of the most important unsupervised learning problems. Scientists have surveyed this problem for years and developed different techniques that can solve it, such as k-means clustering. We analyze one of these techniques: a powerful clustering algorithm called Sparse Subspace Clustering. We demonstrate several experiments using this method and then introduce a new approach that can reduce the computational time required to perform sparse subspace clustering. | ['Enrico Au-Yeung', 'Mehmet F. Demirel'] | 2022-04-01 | null | null | null | null | ['face-clustering'] | ['computer-vision'] | [ 1.83637947e-01 -3.56107146e-01 -1.38056234e-01 -3.84318650e-01
-1.71972886e-01 -7.98214138e-01 3.78905624e-01 2.16666698e-01
-6.64187968e-02 9.55519602e-02 1.72093660e-01 -1.44235522e-01
-3.58784556e-01 -5.98809004e-01 -1.97300762e-01 -1.02407205e+00
-1.08501658e-01 6.16763294e-01 2.11568922e-01 2.22859859e-01
5.78594744e-01 7.20721722e-01 -1.87541437e+00 2.50490665e-01
9.14769948e-01 5.38454831e-01 2.42427498e-01 2.92418182e-01
-6.86713815e-01 5.12714326e-01 -5.61454415e-01 1.42174438e-01
9.47761536e-02 -6.56072617e-01 -8.85618567e-01 7.35143900e-01
-1.59614071e-01 3.39996457e-01 1.09907016e-01 1.32690120e+00
9.78039950e-02 5.97670913e-01 1.06218541e+00 -1.34587908e+00
-3.23533088e-01 6.68811262e-01 -1.04503417e+00 -2.54362911e-01
3.08463454e-01 -6.15854025e-01 3.27925920e-01 -6.92272842e-01
5.92958987e-01 1.53331983e+00 4.31814343e-01 4.01954532e-01
-1.40446115e+00 -4.98020947e-01 -1.96944624e-02 4.06233847e-01
-1.66043448e+00 -2.26405397e-01 8.81255150e-01 -5.86659133e-01
2.53785700e-01 3.85414720e-01 1.97740182e-01 1.59575269e-01
-2.12650612e-01 8.16982090e-01 1.15587318e+00 -7.41846263e-01
5.67006290e-01 1.39761493e-01 5.23389041e-01 2.89560378e-01
5.16703986e-02 -6.97005212e-01 8.88197869e-03 -4.75445747e-01
4.92744803e-01 2.64773160e-01 -2.43025154e-01 -7.68695652e-01
-1.12198210e+00 1.01510203e+00 1.89377964e-01 9.00864780e-01
-1.85334668e-01 -2.40474790e-01 2.67868996e-01 1.24227647e-02
1.21503480e-01 2.72647679e-01 8.48479196e-02 2.71556377e-01
-1.10130990e+00 -1.80001125e-01 9.16084230e-01 9.21504915e-01
9.71232533e-01 -1.92263305e-01 2.68481910e-01 1.14157832e+00
2.84881890e-01 1.86431766e-01 6.88829064e-01 -1.19888473e+00
-3.04744065e-01 8.37536037e-01 -1.44968808e-01 -1.17925024e+00
-2.53428578e-01 1.00393370e-01 -1.18815351e+00 7.27402326e-03
3.03212792e-01 8.37927777e-03 -8.87119234e-01 1.34941959e+00
5.48898757e-01 4.33034956e-01 -7.04549327e-02 6.66690409e-01
7.96384633e-01 9.73382950e-01 -2.27262247e-02 -5.99212646e-01
1.07389545e+00 -8.84626389e-01 -8.44385326e-01 3.74079108e-01
4.01604950e-01 -1.00388956e+00 6.35191679e-01 6.52342558e-01
-7.68879950e-01 -5.64147294e-01 -5.98550797e-01 4.44219261e-01
-4.86987948e-01 4.46134284e-02 8.90006542e-01 8.60886335e-01
-1.14415920e+00 4.97002691e-01 -6.82558954e-01 -8.78466666e-01
1.07993707e-02 5.71886897e-01 -5.03878951e-01 -2.59119987e-01
-3.21685135e-01 2.41867423e-01 7.14548230e-01 -1.44029140e-01
-5.93792856e-01 -1.08310670e-01 -5.80073535e-01 9.82220024e-02
4.63417470e-01 -2.99342155e-01 8.15679550e-01 -1.06408107e+00
-1.19529915e+00 9.46497440e-01 -6.68984652e-01 -1.10950388e-01
-1.76275373e-01 1.32251948e-01 -4.70852047e-01 1.91511095e-01
1.25839666e-01 3.60179305e-01 1.07470882e+00 -1.58951652e+00
-4.62153971e-01 -6.68979228e-01 -6.50448859e-01 1.74245089e-01
-4.74033922e-01 4.73120809e-01 -6.69396520e-01 -4.97522026e-01
8.83775115e-01 -1.13794303e+00 -7.88302064e-01 -5.84148824e-01
-6.70099676e-01 -6.50385916e-01 1.47477281e+00 -3.67335618e-01
1.14128220e+00 -2.46305037e+00 6.22773468e-01 6.99816525e-01
3.03480029e-01 1.25090361e-01 1.71441674e-01 6.50400937e-01
-3.55737478e-01 1.13715008e-01 -6.01109803e-01 -2.31698185e-01
-2.54013538e-01 1.29095644e-01 -1.35689571e-01 5.38770020e-01
-3.69187862e-01 3.08094382e-01 -7.54843950e-01 -6.03285909e-01
4.52399731e-01 2.47773975e-01 -2.49002218e-01 1.26315221e-01
1.01068176e-01 6.94002569e-01 -3.16253394e-01 4.97275651e-01
6.92393363e-01 -1.97467387e-01 5.04590094e-01 -1.63639545e-01
-1.15620695e-01 -6.09470189e-01 -1.87482429e+00 1.39999247e+00
3.55532736e-01 4.34397668e-01 3.25846791e-01 -1.65236866e+00
8.65004122e-01 3.04141164e-01 1.02548933e+00 3.54656249e-01
2.08131298e-01 5.19767739e-02 -7.00520054e-02 -4.57873195e-01
2.90618718e-01 -6.61942586e-02 2.38480978e-02 5.60430229e-01
1.25526264e-02 -1.81911364e-01 4.91133332e-01 3.83491278e-01
8.26658070e-01 -4.99932170e-01 4.64746624e-01 -5.57006001e-01
9.14043427e-01 2.84900725e-01 5.07737160e-01 4.32558209e-01
2.34067552e-02 5.67487478e-01 1.52695939e-01 -1.02972791e-01
-7.51818001e-01 -9.26047564e-01 5.68020483e-03 8.94283056e-01
2.49703333e-01 -4.33828026e-01 -9.19503570e-01 -2.61259466e-01
-5.45592904e-02 2.45435789e-01 -2.89224058e-01 2.53045764e-02
-3.92865986e-01 -7.09273100e-01 7.63535202e-02 3.46972942e-01
5.67892492e-01 -9.06859398e-01 -3.96966264e-02 3.34621519e-02
-5.79686500e-02 -7.58131385e-01 -2.34121397e-01 2.45715350e-01
-1.11215258e+00 -1.05999291e+00 -8.40530217e-01 -1.31605601e+00
1.10313678e+00 8.71380270e-01 8.14518392e-01 3.00086260e-01
-4.10253197e-01 7.48516738e-01 -7.05025613e-01 -2.98719436e-01
-3.04219872e-01 2.15720814e-02 3.56028527e-01 3.11305523e-01
7.54349589e-01 -5.93691707e-01 -8.73934552e-02 3.60428214e-01
-1.13610244e+00 -3.65180492e-01 3.64237458e-01 4.61046278e-01
8.17516148e-01 8.97033811e-01 2.73263425e-01 -1.17736590e+00
7.06316710e-01 -5.19401193e-01 -4.54074949e-01 3.11365753e-01
-2.04246879e-01 -2.61937857e-01 6.76691055e-01 -5.07514119e-01
-8.84854376e-01 7.69734502e-01 2.15097338e-01 -5.76929152e-01
-7.65986621e-01 5.59660375e-01 -6.46491766e-01 -2.22663671e-01
5.32361925e-01 5.23882687e-01 4.47745174e-02 -6.67043626e-01
4.94569749e-01 9.99446094e-01 6.33527040e-01 -4.88883555e-01
1.05417407e+00 6.45678222e-01 9.81783196e-02 -1.49037993e+00
-3.89356166e-01 -1.32165778e+00 -1.04130888e+00 -3.36501956e-01
1.07184899e+00 -4.79270458e-01 -6.80285811e-01 3.77124578e-01
-8.13721359e-01 2.39277378e-01 5.34695275e-02 5.08984149e-01
-4.21064079e-01 7.48949945e-01 -3.08452010e-01 -8.27780247e-01
9.84066278e-02 -9.67236459e-01 6.72476232e-01 4.89763290e-01
-2.85352439e-01 -9.55063462e-01 7.12955147e-02 2.41170391e-01
3.69512178e-02 3.36754173e-01 1.09067881e+00 -7.14268386e-01
-2.94425189e-01 -1.41589940e-02 9.99320969e-02 2.44914889e-01
5.41297257e-01 2.58765340e-01 -7.03958929e-01 -3.93566012e-01
3.60480994e-01 1.17719583e-01 7.78868794e-01 5.42713821e-01
1.51631546e+00 -3.06098186e-03 -9.11346376e-01 4.95861173e-01
1.39325762e+00 6.99811935e-01 5.71156800e-01 -1.10169098e-01
9.53761876e-01 9.67476487e-01 4.92147058e-01 1.42830953e-01
-1.56502470e-01 4.18089718e-01 -3.55704203e-02 -2.07149044e-01
3.10259491e-01 2.98973024e-01 3.71629633e-02 1.13807178e+00
-1.59593552e-01 1.02099642e-01 -9.27659273e-01 4.56948042e-01
-2.13603973e+00 -1.29000294e+00 -3.80830795e-01 2.04625249e+00
4.25399482e-01 -4.90647078e-01 4.00269747e-01 6.45196855e-01
1.21392119e+00 -1.67155445e-01 -3.26773256e-01 -3.49848419e-01
-9.15292948e-02 -1.23099172e-02 1.88462034e-01 2.10048690e-01
-1.33837700e+00 1.14947736e+00 7.19503689e+00 8.49302649e-01
-1.00992453e+00 -2.61646152e-01 5.17420232e-01 6.40638769e-01
-2.75954120e-02 9.02540237e-02 -4.79773879e-01 4.44663823e-01
5.45963883e-01 -3.42209846e-01 5.36917686e-01 1.09881413e+00
1.89333379e-01 -3.73052359e-01 -9.91474211e-01 1.28772616e+00
2.70424902e-01 -1.06376386e+00 1.28465056e-01 -2.84145563e-03
1.02991855e+00 -6.61624849e-01 -2.70110313e-02 -9.48530510e-02
4.77841377e-01 -1.15335488e+00 2.68723350e-02 3.70165855e-02
3.63170147e-01 -9.30705965e-01 3.48568916e-01 6.13944411e-01
-1.46567035e+00 -1.61444560e-01 -4.46961671e-01 5.27635850e-02
1.17257133e-01 8.60205650e-01 -5.87949574e-01 5.99048078e-01
8.73662889e-01 8.56548965e-01 -4.35917139e-01 1.49345183e+00
4.64913063e-02 7.66571045e-01 -4.04629946e-01 9.49169621e-02
1.10743977e-01 -1.01262391e+00 4.76378739e-01 9.36573327e-01
2.88130581e-01 5.08450449e-01 5.63834965e-01 5.88355899e-01
1.67668596e-01 4.16351408e-01 -8.63240659e-01 -1.38846800e-01
6.54542685e-01 1.40189838e+00 -1.64603245e+00 -5.25434017e-01
-2.93160498e-01 9.36110795e-01 -2.06510484e-01 2.44242668e-01
-4.56166595e-01 -4.95812505e-01 4.37791795e-01 8.93536862e-03
2.46647060e-01 -6.86092794e-01 -1.64513066e-01 -8.70184302e-01
-5.88026047e-01 -8.71333480e-01 4.74105835e-01 -4.51694846e-01
-1.12217426e+00 3.02645534e-01 1.74145803e-01 -1.23623395e+00
-1.04341261e-01 -4.74257469e-01 -6.90893173e-01 5.20321071e-01
-6.99276388e-01 -7.34003067e-01 -3.76772940e-01 1.16043699e+00
5.60891509e-01 -3.38539124e-01 8.43638837e-01 2.47208849e-01
-5.73891997e-01 -1.34326324e-01 6.67070687e-01 1.62295565e-01
8.36477816e-01 -1.33465290e+00 -3.66974890e-01 9.00372207e-01
5.14585197e-01 9.93906558e-01 5.88026881e-01 -5.04848540e-01
-1.23820376e+00 -8.39603722e-01 6.32566571e-01 -9.23960134e-02
3.45659018e-01 -2.02833906e-01 -1.00794399e+00 4.86067325e-01
1.55017242e-01 -4.13216174e-01 1.25400162e+00 1.53679416e-01
-3.14972252e-02 4.01321650e-02 -1.18909943e+00 5.62867880e-01
7.23367035e-01 -2.47983202e-01 -5.49090922e-01 5.09826243e-01
3.51484865e-01 9.59045738e-02 -7.90073216e-01 9.36955512e-02
2.46953666e-02 -1.10973978e+00 8.82846892e-01 -2.88597047e-01
-9.56098437e-02 -9.57527816e-01 -1.24826767e-01 -1.28683591e+00
-7.63895035e-01 -6.16564810e-01 7.49352276e-02 1.41985881e+00
-1.54059663e-01 -4.30374622e-01 1.08209336e+00 2.23835260e-01
1.11997925e-01 -8.12953934e-02 -5.19323051e-01 -8.39748263e-01
-2.20480561e-01 1.20525829e-01 5.61926126e-01 1.44741702e+00
2.16894120e-01 3.99098337e-01 -7.48885621e-04 1.38726756e-02
9.23773110e-01 4.45568144e-01 7.92888880e-01 -1.80711854e+00
7.97936395e-02 -5.46047866e-01 -5.46089292e-01 -7.08489776e-01
4.12883013e-01 -7.64594555e-01 2.62014922e-02 -1.65597248e+00
3.14888299e-01 -4.62933391e-01 -1.16862524e-02 2.57513911e-01
-4.70601879e-02 1.61443323e-01 1.48025453e-01 6.53110206e-01
-5.57496369e-01 7.95291215e-02 7.55366385e-01 -1.36331767e-01
-5.00932634e-01 -2.49701290e-04 -7.63377726e-01 9.20915246e-01
7.68194914e-01 -4.03457522e-01 -5.37312925e-01 2.27383478e-03
-5.51116109e-01 -2.73042887e-01 -3.23705375e-01 -1.28161418e+00
5.22892892e-01 -3.03768516e-01 4.08034295e-01 -7.74404943e-01
6.35232925e-02 -1.24269724e+00 5.72894514e-01 4.84198272e-01
9.51962397e-02 -1.22718677e-01 -9.17186439e-02 5.60190380e-01
-5.87277472e-01 -4.32180524e-01 9.39636827e-01 -3.48661214e-01
-1.11731851e+00 -7.70919141e-04 -7.35846162e-01 -4.94503379e-01
1.54963756e+00 -3.57846767e-01 2.20294923e-01 -5.00796258e-01
-9.51079369e-01 2.28974715e-01 6.82062209e-01 1.37690678e-01
5.28224587e-01 -1.21857655e+00 -3.98313820e-01 -4.66159284e-02
-1.11318879e-01 7.18513504e-02 2.78772153e-02 7.69188166e-01
-5.69671631e-01 4.71686870e-01 -1.61201820e-01 -1.06577766e+00
-1.84772766e+00 1.13164818e+00 -9.17332247e-02 2.04646289e-02
-4.19613242e-01 6.14651561e-01 4.13537413e-01 -2.92096943e-01
3.11350018e-01 3.33443344e-01 -7.16691732e-01 9.41066816e-02
6.68767750e-01 6.99974179e-01 -2.93492317e-01 -9.24968183e-01
-3.82687956e-01 1.00750399e+00 1.69030413e-01 1.01485766e-01
1.19998777e+00 -7.86969289e-02 -8.69004846e-01 6.15506113e-01
9.70214248e-01 -1.18934251e-01 -4.33186918e-01 -7.81948641e-02
4.40232217e-01 -3.34631294e-01 -2.26479098e-01 -1.27959251e-01
-1.00171208e+00 5.96581578e-01 5.60768247e-01 6.28296018e-01
1.36521494e+00 1.49038211e-01 3.68786246e-01 6.36046469e-01
4.69419509e-01 -1.23711956e+00 9.56727564e-03 3.78387839e-01
4.48799998e-01 -1.11713982e+00 4.68953438e-02 -1.02879655e+00
-3.90206546e-01 8.67721260e-01 3.43858212e-01 -2.84783006e-01
1.03291142e+00 2.62789398e-01 -3.50626707e-02 -9.58793089e-02
-6.13177828e-02 -3.37013721e-01 1.62080377e-01 8.63650918e-01
5.31709254e-01 1.52083173e-01 -5.32044172e-01 2.98491597e-01
1.04816057e-01 -3.01195145e-01 3.41001332e-01 1.11423695e+00
-8.85227144e-01 -1.35301113e+00 -1.14294600e+00 6.02992296e-01
-2.00943038e-01 3.46903890e-01 -7.80289888e-01 2.66774893e-01
1.07866645e-01 1.26214290e+00 -5.79431094e-02 -3.77778143e-01
-4.37317714e-02 1.87788546e-01 3.50133449e-01 -9.02744353e-01
-4.95213956e-01 4.20124620e-01 -5.38312376e-01 -3.98701608e-01
-1.07073367e+00 -6.11260414e-01 -1.36357808e+00 -3.92882317e-01
-3.67337734e-01 7.98925221e-01 5.93402684e-01 8.60627532e-01
7.99635947e-02 2.63123512e-01 9.00446951e-01 -9.99783576e-01
6.05801716e-02 -6.54119432e-01 -1.07493925e+00 6.10271275e-01
-2.47093260e-01 -4.67529088e-01 -2.91194856e-01 6.95561767e-01] | [7.693382740020752, 4.49768590927124] |
b7502cb4-ec0c-4a5c-9ad3-9885fe7d8cee | a-cross-study-analysis-of-drug-response | 2104.08961 | null | https://arxiv.org/abs/2104.08961v2 | https://arxiv.org/pdf/2104.08961v2.pdf | A cross-study analysis of drug response prediction in cancer cell lines | To enable personalized cancer treatment, machine learning models have been developed to predict drug response as a function of tumor and drug features. However, most algorithm development efforts have relied on cross validation within a single study to assess model accuracy. While an essential first step, cross validation within a biological data set typically provides an overly optimistic estimate of the prediction performance on independent test sets. To provide a more rigorous assessment of model generalizability between different studies, we use machine learning to analyze five publicly available cell line-based data sets: NCI60, CTRP, GDSC, CCLE and gCSI. Based on observed experimental variability across studies, we explore estimates of prediction upper bounds. We report performance results of a variety of machine learning models, with a multitasking deep neural network achieving the best cross-study generalizability. By multiple measures, models trained on CTRP yield the most accurate predictions on the remaining testing data, and gCSI is the most predictable among the cell line data sets included in this study. With these experiments and further simulations on partial data, two lessons emerge: (1) differences in viability assays can limit model generalizability across studies, and (2) drug diversity, more than tumor diversity, is crucial for raising model generalizability in preclinical screening. | ['Rick Stevens', 'Yitan Zhu', 'George Zaki', 'Hyunseung Yoo', 'Justin M. Wozniak', 'Eric Stahlberg', 'Maulik Shukla', 'Alexander Partin', 'Sergei Maslov', 'Pinyi Lu', 'Stewart He', 'Jason Gans', 'Ya Ju Fan', 'Yvonne Evrard', 'Veronika Dubinkina', 'Xiaotian Duan', 'James Doroshow', 'Judith Cohn', 'Austin Clyde', 'Cristina Garcia-Cardona', 'Thomas Brettin', 'Prasanna Balaprakash', 'Jonathan Allen', 'Fangfang Xia'] | 2021-04-18 | null | null | null | null | ['drug-response-prediction'] | ['medical'] | [ 2.59036005e-01 -5.23219109e-01 -5.80438256e-01 -1.73388720e-01
-1.16128194e+00 -6.62714481e-01 5.11910677e-01 7.39084482e-01
-5.92492342e-01 1.13139081e+00 1.40113205e-01 -8.94127548e-01
-2.94799179e-01 -4.42699373e-01 -6.22367263e-01 -7.06849217e-01
-8.32520425e-04 4.46670294e-01 2.36957315e-02 7.60231763e-02
2.75433362e-01 5.48911810e-01 -5.46841145e-01 4.12041545e-01
8.38125110e-01 7.24820018e-01 -1.12579569e-01 5.95257163e-01
3.01120073e-01 5.27418852e-01 -3.75949204e-01 -6.79833069e-03
-4.16636690e-02 -2.48240650e-01 -8.02937388e-01 -2.49988705e-01
1.98676199e-01 -6.72383755e-02 -1.63840294e-01 3.39231312e-01
6.82620883e-01 -3.37797463e-01 9.73715782e-01 -1.34419012e+00
-3.77885401e-01 3.86423290e-01 -3.77718925e-01 1.79614738e-01
1.09328791e-01 6.13448322e-01 9.09660816e-01 -5.99435687e-01
5.16227424e-01 7.45145023e-01 9.54163671e-01 4.03559655e-01
-1.63772845e+00 -8.43183875e-01 -1.73264995e-01 -2.74277627e-01
-1.46597397e+00 -4.17805910e-01 9.26956981e-02 -7.43771076e-01
9.43107247e-01 3.60978007e-01 5.91520667e-01 1.22081351e+00
8.81793261e-01 3.69236410e-01 1.42749727e+00 -9.09924209e-02
3.16765577e-01 2.20293224e-01 1.47652656e-01 4.79897439e-01
4.34941411e-01 1.44071519e-01 -3.58493000e-01 -5.37947416e-01
5.44603825e-01 1.31416926e-02 -5.06784260e-01 -9.86713618e-02
-1.08918095e+00 9.18541133e-01 1.15308046e-01 2.76809216e-01
3.33776101e-02 8.86326805e-02 5.59545994e-01 2.59279311e-01
4.09577131e-01 6.32538378e-01 -8.07519913e-01 1.16898596e-01
-8.93261611e-01 2.54899800e-01 8.50910783e-01 2.15863034e-01
2.58147836e-01 -2.00101778e-01 -1.27558023e-01 8.15094650e-01
1.96764290e-01 9.55213308e-02 8.17684591e-01 -5.18280327e-01
5.71460426e-02 3.81581247e-01 1.30617633e-01 -6.23155653e-01
-9.16409254e-01 -7.74254024e-01 -5.97111225e-01 2.03228801e-01
7.82977223e-01 -2.84033060e-01 -8.16274285e-01 1.80462205e+00
-8.88238251e-02 -4.79054861e-02 5.42383306e-02 3.57496947e-01
7.59191632e-01 2.45927721e-01 5.87440789e-01 -2.90757895e-01
1.16155219e+00 -4.71055031e-01 -2.03331977e-01 -2.24012882e-02
1.50978923e+00 -7.07510471e-01 8.62990975e-01 4.23715979e-01
-6.25618875e-01 -2.04171315e-02 -1.02412069e+00 2.77430713e-01
-5.11017442e-01 -4.99359854e-02 7.40122795e-01 6.01848066e-01
-8.41137826e-01 6.26087189e-01 -7.16954350e-01 -6.27333045e-01
6.83408976e-01 5.87796271e-01 -6.13879025e-01 -1.16600782e-01
-1.02120757e+00 1.08101618e+00 3.40687156e-01 -2.96992123e-01
-9.53119397e-01 -1.20606732e+00 -4.18085247e-01 3.83198857e-02
-1.34987563e-01 -8.88495803e-01 9.19988036e-01 -5.47138095e-01
-1.15689659e+00 7.43648410e-01 -3.99721637e-02 -6.20433629e-01
3.97995114e-01 4.90218103e-01 -3.11678320e-01 -2.97565430e-01
-6.27102256e-02 6.17379546e-01 -6.33705407e-02 -1.06091189e+00
-6.63726807e-01 -4.50888216e-01 -4.83143121e-01 1.74363241e-01
-1.44314468e-01 -5.65414783e-04 9.54673216e-02 -2.67173141e-01
-2.18490511e-01 -1.25563955e+00 -1.80102944e-01 -1.18007243e-01
-2.65877247e-01 -4.90375198e-02 5.45601368e-01 -4.72999871e-01
1.20252705e+00 -1.71735919e+00 -2.64708817e-01 6.60146028e-02
3.84002805e-01 7.67261386e-02 -2.93472558e-01 5.87900996e-01
-2.81910181e-01 5.46297789e-01 8.40081722e-02 8.91733244e-02
-5.20376086e-01 -2.06244498e-01 -3.43616009e-02 7.24347234e-01
3.61287534e-01 1.01850247e+00 -6.41763449e-01 -5.07911444e-01
-1.91207528e-01 2.77380645e-01 -4.38396096e-01 -3.51062596e-01
-1.25219986e-01 6.05045378e-01 -4.05033588e-01 6.75983787e-01
3.15995067e-01 -5.46664834e-01 2.14920878e-01 -1.34472787e-01
8.36409256e-03 6.92079514e-02 -3.46759856e-01 1.14779067e+00
-3.74890774e-01 5.39960861e-01 -3.69449347e-01 -8.55656922e-01
5.96254945e-01 2.90368974e-01 9.20118451e-01 -5.16652107e-01
1.65999353e-01 4.31763887e-01 7.69078255e-01 -3.05868499e-02
-1.71744555e-01 -4.19876695e-01 1.17547527e-01 2.61487246e-01
-1.63787425e-01 -2.03228761e-02 3.07670683e-02 -1.69585440e-02
1.51159334e+00 -4.03425992e-01 3.92210811e-01 -6.25567138e-01
2.08637834e-01 3.82498801e-01 7.56339669e-01 5.01903296e-01
-4.33087617e-01 4.64745045e-01 7.38090217e-01 -4.54542935e-01
-9.35217261e-01 -8.16400170e-01 -8.49419832e-01 5.99214971e-01
-3.44784200e-01 -1.42853275e-01 -2.80754745e-01 -5.65709472e-01
3.24717671e-01 6.67542815e-01 -9.74552035e-01 -2.18530953e-01
2.22573616e-03 -1.48074770e+00 8.18359256e-01 3.42400104e-01
1.36813773e-02 -3.05634230e-01 -2.02794433e-01 3.47118914e-01
1.65684968e-01 -9.24302459e-01 -2.11981535e-01 7.23414719e-01
-7.32975960e-01 -1.48509455e+00 -6.20491624e-01 -6.14817977e-01
3.80736947e-01 2.01536804e-01 9.87374246e-01 3.04445922e-01
-1.54339030e-01 1.10727310e-01 -3.55630890e-02 -8.25498581e-01
-6.82147682e-01 5.95559254e-02 8.26886147e-02 -4.94063109e-01
4.28110838e-01 -2.52644062e-01 -4.82328713e-01 5.25342166e-01
-5.19327879e-01 -1.16631985e-01 6.54540300e-01 1.26611280e+00
8.33763421e-01 -3.20898928e-02 9.30375576e-01 -1.12059212e+00
6.17695451e-01 -7.84975767e-01 -3.23617578e-01 3.07708085e-01
-8.65777731e-01 -1.63744181e-01 7.82199144e-01 -4.98801917e-01
-6.84710205e-01 -2.34701261e-02 2.54004076e-02 -3.75502780e-02
-4.04955000e-02 9.74652529e-01 2.82124221e-01 -4.32704687e-01
9.73587990e-01 1.40726805e-01 5.17489731e-01 -4.39827517e-02
-4.99728799e-01 4.94991302e-01 -8.27650875e-02 -5.44286549e-01
3.20161372e-01 2.89579093e-01 4.56288397e-01 -7.14121878e-01
-6.41645074e-01 -1.16101801e-01 -3.08504134e-01 1.36563867e-01
6.38330162e-01 -1.07504988e+00 -9.05728579e-01 5.18078208e-01
-5.85523427e-01 -1.06673861e+00 2.74136543e-01 7.41236985e-01
-4.52168494e-01 2.63274871e-02 -7.89431512e-01 -2.54811049e-01
-2.46406466e-01 -1.54517651e+00 9.36048150e-01 6.50630286e-03
-6.08157575e-01 -1.46656406e+00 3.51848155e-01 3.49216282e-01
3.59244376e-01 5.61153948e-01 1.44003117e+00 -1.43257022e+00
-1.77551746e-01 -4.25698221e-01 -1.62791550e-01 -3.24536204e-01
4.51022267e-01 4.01038587e-01 -9.00000811e-01 -6.19670153e-01
-3.21966052e-01 -6.63383186e-01 7.68702984e-01 8.20028841e-01
1.19084239e+00 6.67267069e-02 -1.09500837e+00 5.47505140e-01
1.45832586e+00 5.62812328e-01 4.70732629e-01 3.23268592e-01
2.76806533e-01 5.71523905e-02 2.70958126e-01 1.29470676e-01
3.04711759e-01 5.78256309e-01 8.13251883e-02 -4.39830720e-01
7.56160095e-02 -1.62616268e-01 -6.70193732e-02 1.74806565e-01
3.90899837e-01 -4.72639382e-01 -1.14688849e+00 1.84781387e-01
-1.27053237e+00 -7.03854918e-01 -2.13314638e-01 2.56594515e+00
9.72196162e-01 3.69632810e-01 3.25824320e-01 -7.04932734e-02
2.79582202e-01 -4.75746125e-01 -7.55645096e-01 -2.45188400e-01
-3.80506277e-01 -5.04426844e-02 7.09015727e-01 4.72217858e-01
-8.60223055e-01 4.64708209e-01 7.88462925e+00 8.98519158e-01
-1.55966043e+00 -1.71722621e-01 1.60314727e+00 -1.55195743e-01
-1.25913605e-01 2.04303488e-02 -8.51932943e-01 4.59984720e-01
1.13179255e+00 -5.18880129e-01 -7.97944367e-02 1.90263197e-01
5.70801139e-01 -3.28817725e-01 -1.53770447e+00 4.75870281e-01
-2.74930030e-01 -1.55495620e+00 -1.97176367e-01 5.96106410e-01
7.72915184e-01 3.59532505e-01 2.61444658e-01 3.19038689e-01
6.26199663e-01 -1.53453732e+00 2.16268413e-02 3.35236818e-01
8.84984136e-01 -4.90526468e-01 8.05784106e-01 3.31243813e-01
-4.15525645e-01 -1.58484042e-01 -7.78691322e-02 5.25998324e-02
-3.45872432e-01 4.89980221e-01 -1.37557232e+00 2.86860883e-01
3.34735513e-01 4.28227663e-01 -7.95617700e-01 1.10696399e+00
6.10758424e-01 8.83507490e-01 -3.45495433e-01 -1.24289259e-01
1.76698864e-01 1.44775778e-01 -5.03869578e-02 1.12935579e+00
3.06174457e-01 1.14555448e-01 1.85870111e-01 4.67474967e-01
1.98616534e-01 1.65230155e-01 -4.48455542e-01 -2.02923268e-01
5.45849800e-01 1.15797162e+00 -7.91543424e-01 -4.81410325e-02
-6.24837399e-01 2.68920809e-01 3.76191407e-01 2.93045044e-01
-8.67153108e-01 7.12344050e-02 6.15614414e-01 2.22285792e-01
-2.17312813e-01 5.86690642e-02 -6.52119100e-01 -7.81920254e-01
-7.18426347e-01 -1.00037551e+00 7.33680725e-01 -4.94813889e-01
-1.27839458e+00 2.79051542e-01 -1.49285480e-01 -1.15557861e+00
4.07398976e-02 -8.78448427e-01 -6.61632121e-01 1.09276581e+00
-1.21339393e+00 -8.51594388e-01 -1.71587709e-02 4.17387933e-02
3.04354668e-01 -1.69396162e-01 1.08694744e+00 7.99460039e-02
-9.81345475e-01 8.35327983e-01 3.70397449e-01 -1.74397275e-01
9.85019743e-01 -9.29356217e-01 -4.78902236e-02 1.91898048e-01
-2.51364857e-01 5.71002722e-01 8.09413552e-01 -6.90277576e-01
-1.22465587e+00 -1.09431255e+00 3.37671757e-01 -7.07521439e-01
8.11986327e-01 2.37040192e-01 -9.38461900e-01 7.79946268e-01
-2.52806768e-03 -9.20080841e-02 1.39309788e+00 3.53216618e-01
-1.68684840e-01 6.36064913e-03 -1.14490664e+00 5.97392321e-01
3.79181802e-01 -1.70768842e-01 2.84577668e-01 6.30331159e-01
2.18292087e-01 -5.64605117e-01 -1.60505056e+00 4.56635952e-01
7.06421137e-01 -6.84242725e-01 7.24180639e-01 -9.36189830e-01
5.27222931e-01 -2.55063534e-01 -1.67594310e-02 -1.62494612e+00
-6.62925243e-01 -3.73597890e-02 5.67768574e-01 9.36981618e-01
1.19924474e+00 -7.41891325e-01 9.94446039e-01 9.80949879e-01
7.27814063e-02 -1.43814874e+00 -7.97451735e-01 -8.00316632e-01
9.14164484e-01 -1.64714783e-01 3.75551939e-01 9.82498527e-01
3.36598903e-01 3.91771615e-01 2.10876744e-02 -1.02265559e-01
2.38055706e-01 -2.52812445e-01 7.27660060e-01 -1.19660950e+00
-3.91034991e-01 -6.90802336e-01 -4.21946228e-01 -3.02614808e-01
1.92621797e-01 -9.45236206e-01 -3.43927503e-01 -1.25006235e+00
6.47665441e-01 -7.50778079e-01 -4.93202925e-01 7.17550874e-01
-3.47613513e-01 1.64491743e-01 -9.03111100e-02 1.10249907e-01
-2.37186432e-01 -3.14847231e-02 1.13640428e+00 -2.24400848e-01
-2.18376517e-01 -4.55672257e-02 -9.74102259e-01 4.03210461e-01
9.00307834e-01 -3.30374539e-01 -4.21618253e-01 -2.19963878e-01
-2.38986854e-02 3.10399950e-01 2.30104893e-01 -1.04655886e+00
5.10368794e-02 -6.59082711e-01 1.01900578e+00 -1.28822878e-01
2.17559442e-01 -5.34060657e-01 6.43305302e-01 9.51363981e-01
-4.93539214e-01 -7.25557953e-02 5.79336643e-01 4.95555609e-01
1.44417077e-01 8.50973204e-02 9.09929931e-01 -2.51891855e-02
-1.10014543e-01 5.82133830e-01 -6.85931683e-01 -1.48920760e-01
9.36384201e-01 -3.44458818e-01 -5.35510063e-01 -2.83882916e-01
-5.50844550e-01 3.30611259e-01 7.41283715e-01 -4.55007441e-02
3.79469097e-02 -1.13951540e+00 -9.15562391e-01 -2.63374567e-01
4.46374416e-01 -4.61040080e-01 1.79040283e-01 1.13705325e+00
-4.84907418e-01 8.10834587e-01 -1.35263532e-01 -5.36619723e-01
-1.15381670e+00 5.46281457e-01 6.59735441e-01 -5.97279549e-01
1.43599331e-01 6.07414365e-01 3.62029433e-01 -8.79771411e-02
-3.04033458e-01 -2.51414895e-01 -5.22988215e-02 -2.08692387e-01
1.54897213e-01 1.80258170e-01 2.22596049e-01 -4.77277100e-01
-3.61808568e-01 1.22675136e-01 -4.92433310e-01 2.97499746e-01
1.10843205e+00 3.90926898e-01 2.36964032e-01 4.69483793e-01
1.30484784e+00 -1.85741723e-01 -1.11374736e+00 1.51119500e-01
-1.82485104e-01 -1.73335493e-01 2.06972182e-01 -1.13041878e+00
-6.50405884e-01 4.51261222e-01 5.55473447e-01 8.41687545e-02
1.01040316e+00 -3.34822685e-01 3.85140777e-01 2.10192531e-01
2.29713321e-01 -8.91163588e-01 -2.86089301e-01 2.03815028e-01
7.41108179e-01 -1.42100835e+00 3.70978296e-01 -3.99292409e-01
-6.78637028e-01 1.04570937e+00 6.88855648e-01 3.01727295e-01
6.27045989e-01 2.58686543e-01 -3.09469569e-02 -8.85365456e-02
-1.29282403e+00 3.61352146e-01 -1.19644359e-01 5.36419332e-01
9.00038779e-01 -5.11349272e-03 -7.10490167e-01 6.67332649e-01
5.96923009e-02 4.48065579e-01 7.28304863e-01 5.76742768e-01
-2.85652936e-01 -1.13347733e+00 -1.77501947e-01 1.03815973e+00
-6.18435621e-01 -1.09160557e-01 -4.73329186e-01 1.15425098e+00
-4.01366621e-01 9.30847228e-01 -5.29269055e-02 -3.83989662e-01
4.74225245e-02 8.00478831e-02 3.99500310e-01 -3.93449008e-01
-4.67803180e-01 4.03865017e-02 2.84864843e-01 -9.74154919e-02
5.49081117e-02 -6.80105031e-01 -1.16438079e+00 -4.09903705e-01
-5.50994992e-01 7.10751414e-02 4.18526798e-01 1.05803156e+00
4.89919394e-01 3.63160551e-01 6.63295031e-01 -3.36793840e-01
-5.95696330e-01 -8.69772553e-01 -5.43581724e-01 2.71895617e-01
3.26343179e-02 -5.70102632e-01 -2.77076811e-01 -4.35750745e-02] | [15.181254386901855, -2.9655954837799072] |
d15b2701-c4c5-4b87-86b1-2b540bab94f7 | machine-learning-for-classification-of-1 | 2209.02249 | null | https://arxiv.org/abs/2209.02249v1 | https://arxiv.org/pdf/2209.02249v1.pdf | Machine Learning For Classification Of Antithetical Emotional States | Emotion Classification through EEG signals has achieved many advancements. However, the problems like lack of data and learning the important features and patterns have always been areas with scope for improvement both computationally and in prediction accuracy. This works analyses the baseline machine learning classifiers' performance on DEAP Dataset along with a tabular learning approach that provided state-of-the-art comparable results leveraging the performance boost due to its deep learning architecture without deploying heavy neural networks. | ['Yusuf Uzzaman Khan', 'Rajat Maheshwari', 'Jeevanshi Sharma'] | 2022-09-06 | null | null | null | null | ['emotion-classification', 'emotion-classification'] | ['computer-vision', 'natural-language-processing'] | [-1.38709515e-01 -2.38651466e-02 5.73981255e-02 -7.24147022e-01
-6.31943107e-01 -2.57897042e-02 3.28283608e-01 2.77408034e-01
-6.08083308e-01 9.52426076e-01 -1.05656952e-01 1.94251373e-01
-3.79634947e-01 -4.61131036e-01 -3.92694503e-01 -7.06059098e-01
-6.42261028e-01 7.51084164e-02 -2.49252930e-01 -2.70891994e-01
3.02260846e-01 3.83236170e-01 -1.68604922e+00 6.77457929e-01
5.40732980e-01 1.71163642e+00 -2.57781684e-01 3.16505194e-01
3.50152180e-02 8.41632128e-01 -7.05196738e-01 -4.30972099e-01
2.64396686e-02 -3.67825404e-02 -5.12921572e-01 -4.91066933e-01
5.59739359e-02 2.75896758e-01 -3.22740763e-01 5.62668025e-01
8.07021081e-01 -3.76472995e-02 5.81090808e-01 -1.55073690e+00
-3.95375609e-01 2.74126202e-01 -3.73768687e-01 6.16173804e-01
4.46687609e-01 -3.28335702e-01 9.16814804e-01 -1.04072821e+00
2.70109683e-01 4.69926089e-01 1.02341747e+00 5.19381106e-01
-8.79115939e-01 -1.09545910e+00 -6.97243586e-02 8.27451169e-01
-1.25202322e+00 -2.56550521e-01 8.34713459e-01 -3.97719502e-01
1.66893101e+00 2.30728850e-01 8.66208136e-01 1.42201352e+00
5.16993344e-01 7.04170942e-01 1.51026785e+00 -2.51210928e-01
3.56537521e-01 5.81806004e-01 6.67513788e-01 4.34462428e-01
-7.54021704e-02 8.24178904e-02 -9.59550738e-01 1.60119738e-02
9.93984416e-02 -1.45617321e-01 -2.48610955e-02 -1.33447861e-02
-6.19321644e-01 7.04989016e-01 3.41631591e-01 4.74768430e-01
-7.59918571e-01 -1.17141090e-01 8.88160288e-01 5.41343927e-01
7.35216200e-01 7.16250002e-01 -9.61806655e-01 -7.77422726e-01
-1.04340136e+00 8.57079774e-02 6.87420607e-01 4.85336542e-01
3.78988683e-01 3.11202347e-01 1.56079158e-01 9.34736133e-01
-3.01110864e-01 7.71812648e-02 7.20528543e-01 -3.82872701e-01
1.93641335e-01 8.62156451e-01 -1.75845653e-01 -1.10307479e+00
-9.26731169e-01 -7.03890145e-01 -9.12250400e-01 3.02796185e-01
-1.57735422e-02 -6.48864269e-01 -6.15129411e-01 1.46935833e+00
-3.13274622e-01 6.33828267e-02 1.70977473e-01 6.72398210e-01
8.02085161e-01 4.40778822e-01 1.63617924e-01 -1.56405464e-01
1.14982438e+00 -8.77450287e-01 -9.86139834e-01 -2.96798170e-01
4.91754770e-01 -2.61521995e-01 8.10382366e-01 1.07458222e+00
-9.25460339e-01 -4.25899804e-01 -1.19043005e+00 1.38180375e-01
-8.12437892e-01 2.79883325e-01 1.30702651e+00 9.38220203e-01
-1.05203021e+00 5.93769073e-01 -7.16283143e-01 -3.00884813e-01
9.31429386e-01 7.76123464e-01 -6.78237438e-01 4.11557078e-01
-1.15272653e+00 1.27793455e+00 6.13087535e-01 7.23502785e-02
-4.55214471e-01 -8.33908677e-01 -4.84625220e-01 3.90300959e-01
-6.52881861e-02 -4.67597395e-02 7.84118831e-01 -1.17952561e+00
-1.50455141e+00 6.81395829e-01 -5.06914826e-03 -6.23600841e-01
2.82210141e-01 -2.61885971e-01 -9.98827040e-01 -6.42139688e-02
-5.93989611e-01 7.19783008e-01 6.54667497e-01 -7.05745757e-01
-6.74548149e-01 -5.76450408e-01 -3.30859244e-01 -7.48074278e-02
-7.42610812e-01 2.14255288e-01 2.20452979e-01 -5.33377886e-01
-1.14227213e-01 -7.42358983e-01 1.54790267e-01 -2.99991369e-01
1.60518274e-01 -4.04789835e-01 8.80778015e-01 -7.42773652e-01
1.27257597e+00 -2.03022122e+00 -4.97661680e-02 2.21098557e-01
-1.36612639e-01 1.98642254e-01 4.00142893e-02 3.35711509e-01
-7.62699425e-01 -1.77529141e-01 7.40089938e-02 -9.00666416e-02
5.63093498e-02 3.22886296e-02 -2.99577445e-01 2.13849902e-01
3.19494069e-01 8.06573093e-01 -5.05146146e-01 1.10863894e-02
2.78563052e-01 6.78762019e-01 -3.42140555e-01 1.68784648e-01
4.82796490e-01 4.05118577e-02 4.90604946e-03 6.03091657e-01
4.88052934e-01 5.72590306e-02 4.54419032e-02 -2.93442935e-01
-8.92065316e-02 2.06976846e-01 -1.13502550e+00 1.53722441e+00
-3.87880147e-01 9.30370510e-01 -2.11754665e-01 -1.55448616e+00
1.02305675e+00 5.21295547e-01 6.90228045e-01 -9.57843304e-01
2.46197730e-01 2.46822521e-01 1.82832271e-01 -6.51834607e-01
7.01737648e-04 -1.82573915e-01 1.67291655e-04 1.58514515e-01
4.87309277e-01 3.81167591e-01 -2.88517684e-01 -3.14053446e-01
9.21093464e-01 1.47492215e-01 3.42090219e-01 -4.67911452e-01
3.78448725e-01 -3.00938822e-02 4.69307363e-01 4.37640131e-01
-3.64476800e-01 2.01481029e-01 4.61319298e-01 -8.39523494e-01
-7.00111151e-01 -6.39279008e-01 -5.05518854e-01 1.24996543e+00
-3.97902280e-01 -5.24271131e-01 -6.82649195e-01 -4.69966680e-01
-3.00428808e-01 7.13661909e-01 -1.13934767e+00 -2.90769696e-01
-1.04657277e-01 -1.25295234e+00 5.69664001e-01 8.80306423e-01
6.17432117e-01 -1.30793226e+00 -1.25229454e+00 2.34075159e-01
1.41439542e-01 -1.03185439e+00 5.13979197e-01 8.82123232e-01
-9.19915915e-01 -9.06782091e-01 -3.30347389e-01 -5.30477703e-01
2.53465354e-01 -2.42018059e-01 9.91311491e-01 -2.19990656e-01
-6.19457603e-01 2.05212668e-01 -5.08740842e-01 -1.08939064e+00
5.01378775e-01 2.94286609e-01 1.54386431e-01 -1.74227580e-01
9.49684381e-01 -8.20761263e-01 -5.43157041e-01 -2.45042108e-02
-5.42146325e-01 -1.92486390e-01 5.95722914e-01 9.66092646e-01
2.16921940e-01 4.21831489e-01 1.11764002e+00 -4.36250985e-01
8.18161190e-01 -6.22633338e-01 -6.48919865e-02 1.01138003e-01
-9.50174630e-01 -8.81531388e-02 4.11389410e-01 -8.81903768e-02
-8.54273558e-01 -4.61090654e-02 -3.44093412e-01 8.63518640e-02
-4.02816772e-01 6.67816103e-01 2.41552368e-01 -2.47682735e-01
5.98188937e-01 1.60836011e-01 -2.86258191e-01 -3.99510175e-01
-2.24316731e-01 9.63337064e-01 2.68824250e-01 -3.54430646e-01
1.47766499e-02 1.09450579e-01 -1.38102770e-01 -6.40752137e-01
-8.40349078e-01 -3.18922222e-01 -7.93251514e-01 -1.94826007e-01
6.99710011e-01 -8.88767064e-01 -8.19591224e-01 5.75795114e-01
-8.15055251e-01 -4.33395691e-02 -7.43032470e-02 6.74281716e-01
-3.33656937e-01 -2.72736490e-01 -4.99224007e-01 -9.63736117e-01
-7.84430861e-01 -7.95604229e-01 4.68205690e-01 1.41627461e-01
-4.70077515e-01 -8.04697335e-01 4.60982462e-03 6.99803233e-02
6.78891063e-01 2.76893109e-01 9.05624211e-01 -9.31579828e-01
2.07539022e-01 -6.22448921e-01 -1.42890424e-01 3.98398697e-01
-5.53148240e-02 -3.60517442e-01 -1.61070859e+00 -1.58725247e-01
8.29220414e-02 -7.79837847e-01 4.58573401e-01 2.34879687e-01
1.67016387e+00 -7.36468136e-02 -1.64743796e-01 6.43818617e-01
1.43972206e+00 4.53784257e-01 6.57325923e-01 7.13587582e-01
2.65804715e-02 6.13367319e-01 2.29342043e-01 6.24702930e-01
5.44450916e-02 5.45989811e-01 3.45217019e-01 -1.27818093e-01
3.06179136e-01 3.02471578e-01 9.24566239e-02 7.13279009e-01
-2.75463402e-01 1.95431739e-01 -9.85173523e-01 3.75876725e-01
-1.75133705e+00 -1.00701523e+00 3.67928222e-02 1.64841628e+00
4.61730421e-01 1.07752994e-01 2.35270381e-01 6.41297758e-01
2.03161225e-01 -1.15922883e-01 -6.22749805e-01 -9.37576115e-01
-1.77558124e-01 6.40301704e-01 -4.57332432e-02 -2.18057692e-01
-1.12070918e+00 5.70042789e-01 7.43823195e+00 6.51145935e-01
-1.57830143e+00 1.29419014e-01 6.81422114e-01 -3.44845116e-01
4.90057826e-01 -8.02489340e-01 -5.21397173e-01 4.79406059e-01
1.46813715e+00 1.27374843e-01 5.74773490e-01 9.08012211e-01
1.78598426e-02 -1.81791946e-01 -9.14363325e-01 1.49967611e+00
2.91846633e-01 -1.09440255e+00 -2.85688967e-01 -2.08341584e-01
5.29390812e-01 2.26760268e-01 8.60824287e-02 8.00806522e-01
-4.39319730e-01 -1.38383532e+00 5.75959563e-01 4.74987000e-01
5.29799581e-01 -1.14107382e+00 1.26702046e+00 1.98066264e-01
-5.90333045e-01 -4.46852803e-01 -3.16676229e-01 -7.03194082e-01
-3.72990608e-01 3.99919778e-01 -3.96469682e-01 5.26642561e-01
1.27315164e+00 8.04989398e-01 -6.85137987e-01 1.03275990e+00
-8.93260315e-02 8.23150218e-01 -2.47362509e-01 -2.02425316e-01
2.01199010e-01 -1.16735503e-01 -8.76733586e-02 1.61902332e+00
3.09799433e-01 3.07730913e-01 -2.51694292e-01 4.38524961e-01
4.60075885e-02 3.32939714e-01 -3.70879233e-01 2.36847755e-02
1.70755342e-01 1.50591743e+00 -5.90667725e-01 -1.48515850e-01
-4.76330459e-01 8.27714980e-01 6.58540249e-01 7.75527582e-02
-7.80624568e-01 -6.16120517e-01 6.29659176e-01 -1.20673567e-01
2.12131724e-01 6.21182136e-02 -8.98577511e-01 -1.03725898e+00
1.05333865e-01 -8.35794270e-01 4.93005723e-01 -8.26416731e-01
-1.34469879e+00 8.90312791e-01 -1.23976789e-01 -7.39725232e-01
-9.87980366e-02 -1.02058554e+00 -6.65327728e-01 6.89853132e-01
-1.45844293e+00 -9.91539240e-01 -4.74315852e-01 6.68573022e-01
3.13338786e-01 -4.48563367e-01 1.42036581e+00 5.44210196e-01
-7.06473231e-01 8.55164707e-01 1.27270505e-01 2.45858654e-02
6.00703657e-01 -1.07858133e+00 -4.19931322e-01 2.66685665e-01
4.06205766e-02 3.62023562e-01 5.60486495e-01 7.61073083e-02
-1.13639104e+00 -4.98032123e-01 8.44964445e-01 -3.33889872e-01
5.59869111e-01 -7.30278909e-01 -8.43343616e-01 6.08944416e-01
6.56427801e-01 -1.08745351e-01 1.11425090e+00 5.84809065e-01
-1.36964038e-01 -4.45291907e-01 -1.25081921e+00 8.21772590e-02
5.80680847e-01 -5.17814815e-01 -7.88310289e-01 1.24653049e-01
-2.30515078e-01 -2.45148122e-01 -9.73931730e-01 5.96955955e-01
9.07661915e-01 -1.04198062e+00 7.46402383e-01 -8.02795649e-01
4.63218689e-01 4.13171917e-01 -1.12469979e-01 -1.60269427e+00
-2.67189980e-01 -2.40514174e-01 -2.32152522e-01 1.10331035e+00
5.77159762e-01 -5.33033669e-01 7.56027281e-01 1.00847065e+00
-1.40287906e-01 -1.44722855e+00 -9.74064827e-01 -5.77552199e-01
1.30043939e-01 -9.34747815e-01 4.79844540e-01 1.15414691e+00
6.97554946e-01 5.17220676e-01 -4.01421577e-01 -2.68050551e-01
8.61714557e-02 -9.27784443e-02 1.47613838e-01 -1.38360000e+00
1.64624006e-01 -5.15269220e-01 -7.94570744e-01 1.20345883e-01
3.13865423e-01 -9.03084457e-01 -3.24299216e-01 -1.40183806e+00
3.70696694e-01 -1.47026077e-01 -9.48391080e-01 8.59229982e-01
-2.77534015e-02 3.96900296e-01 2.34011728e-02 -2.97620893e-01
-5.33490896e-01 5.61026394e-01 3.81555527e-01 -1.66978136e-01
-1.64496288e-01 -2.76640773e-01 -7.33665466e-01 7.21728146e-01
1.16537392e+00 -5.12254477e-01 -4.36133355e-01 -3.46947670e-01
3.42874378e-01 -2.07939088e-01 1.80349126e-01 -1.60968673e+00
1.40008777e-01 4.41705555e-01 9.44598675e-01 -4.48028117e-01
5.70535600e-01 -1.06887078e+00 1.47332549e-01 2.80325025e-01
-4.88679379e-01 3.36096048e-01 8.05045187e-01 1.62246883e-01
-2.65794724e-01 -8.37267861e-02 7.04411983e-01 1.97337344e-01
-9.84520197e-01 1.74112096e-01 -4.63528842e-01 -7.85246640e-02
1.21722019e+00 -5.01461804e-01 -6.36776239e-02 -1.70296475e-01
-1.05551505e+00 -8.76477733e-02 -2.61691093e-01 6.36946559e-01
5.14694929e-01 -1.05965459e+00 -5.64897537e-01 3.44514132e-01
1.46525502e-01 -8.46077204e-01 2.99220532e-01 8.64839017e-01
-3.44370566e-02 7.69995332e-01 -1.03148532e+00 -3.24306458e-01
-1.19084525e+00 3.70916784e-01 5.30356526e-01 -2.35074937e-01
-6.45556152e-01 8.43716204e-01 -4.42457765e-01 -2.99748898e-01
4.99830425e-01 -2.72234017e-03 -5.55725157e-01 4.47962046e-01
6.71914518e-01 6.09295011e-01 6.45336509e-01 -2.09471732e-01
-6.22329712e-01 1.44351527e-01 -3.90183404e-02 1.44095913e-01
1.92320549e+00 1.93096429e-01 6.37804298e-03 5.98199248e-01
1.24610162e+00 -5.93096316e-01 -9.08209860e-01 2.95218080e-01
3.85469258e-01 -1.80115402e-01 5.84280729e-01 -1.65535617e+00
-1.16873729e+00 1.24486792e+00 1.31439996e+00 -1.14709809e-01
1.35521662e+00 -4.09746796e-01 5.03945231e-01 6.80214167e-01
6.22887015e-01 -1.39571357e+00 -4.23578173e-02 6.21136069e-01
8.21705163e-01 -1.39322650e+00 -7.97642246e-02 7.02491403e-02
-7.98780978e-01 1.29762280e+00 6.34324253e-01 -3.22767347e-01
1.20869589e+00 5.80608845e-01 1.25786006e-01 -5.68771362e-01
-9.13909853e-01 3.23641926e-01 4.02154565e-01 6.77773654e-01
6.44204080e-01 3.81335169e-02 -5.26197314e-01 1.38834798e+00
-3.51441175e-01 4.35524106e-01 1.51829049e-01 7.68726289e-01
-3.86765935e-02 -8.89143527e-01 1.83545068e-01 9.05802071e-01
-9.49502707e-01 -2.22360298e-01 -3.25371534e-01 8.91893446e-01
2.46235043e-01 9.60375071e-01 5.50735518e-02 -5.69691658e-01
4.99476522e-01 6.04264200e-01 4.28601980e-01 -3.01596761e-01
-1.05370963e+00 -4.79304701e-01 1.01844713e-01 -6.19078040e-01
-5.10714114e-01 -6.29024506e-01 -1.20310843e+00 2.10006293e-02
-2.43871972e-01 2.77136087e-01 1.01863301e+00 9.89065170e-01
6.06950819e-01 7.47897089e-01 4.06249553e-01 -7.46044636e-01
-2.16091260e-01 -1.27638066e+00 -8.85509133e-01 2.88902700e-01
3.15155648e-02 -7.38972723e-01 -1.18926629e-01 -1.46667227e-01] | [13.183027267456055, 3.490743637084961] |
18b6e27e-c929-4af5-aade-dab5e115adcc | online-decision-transformer | 2202.05607 | null | https://arxiv.org/abs/2202.05607v2 | https://arxiv.org/pdf/2202.05607v2.pdf | Online Decision Transformer | Recent work has shown that offline reinforcement learning (RL) can be formulated as a sequence modeling problem (Chen et al., 2021; Janner et al., 2021) and solved via approaches similar to large-scale language modeling. However, any practical instantiation of RL also involves an online component, where policies pretrained on passive offline datasets are finetuned via taskspecific interactions with the environment. We propose Online Decision Transformers (ODT), an RL algorithm based on sequence modeling that blends offline pretraining with online finetuning in a unified framework. Our framework uses sequence-level entropy regularizers in conjunction with autoregressive modeling objectives for sample-efficient exploration and finetuning. Empirically, we show that ODT is competitive with the state-of-the-art in absolute performance on the D4RL benchmark but shows much more significant gains during the finetuning procedure. | ['Aditya Grover', 'Amy Zhang', 'Qinqing Zheng'] | 2022-02-11 | null | null | null | null | ['d4rl'] | ['robots'] | [ 4.59040664e-02 3.40970516e-01 -7.19738841e-01 -2.10810974e-01
-9.97182906e-01 -7.52496123e-01 8.85575414e-01 2.50501223e-02
-7.17318833e-01 1.06192982e+00 2.81644434e-01 -6.99354827e-01
6.35056570e-02 -4.35582608e-01 -9.70135570e-01 -3.54548842e-01
-1.77994668e-01 7.62964487e-01 5.70699126e-02 -2.77361721e-01
2.54463136e-01 3.42316806e-01 -1.35550976e+00 2.96086729e-01
8.52681339e-01 8.67643118e-01 4.57322836e-01 9.93225932e-01
-2.84761667e-01 1.22414863e+00 -2.67493546e-01 -1.15555644e-01
2.88993835e-01 -3.97421211e-01 -9.31739509e-01 -1.05046161e-01
-9.60918814e-02 -3.47895056e-01 -4.54375625e-01 7.46393025e-01
3.97954971e-01 5.04937768e-01 2.34472394e-01 -9.09719169e-01
-3.79182488e-01 1.08185434e+00 -3.21685880e-01 1.25261815e-02
1.04974717e-01 6.08873069e-01 9.25988674e-01 -5.23504138e-01
4.55972731e-01 1.35509264e+00 4.32687819e-01 8.36020529e-01
-1.55418289e+00 -5.46202242e-01 5.25491238e-01 9.60138738e-02
-9.64144707e-01 -4.88257140e-01 5.04733741e-01 -1.71325982e-01
1.53360057e+00 9.22544897e-02 6.80190384e-01 1.32037520e+00
2.34294772e-01 1.29548609e+00 1.20125389e+00 -6.15248203e-01
5.71315229e-01 6.71221018e-02 -1.13976046e-01 8.08789015e-01
-1.11862630e-01 7.77236164e-01 -5.86455047e-01 -2.77970314e-01
7.25537419e-01 -3.47692341e-01 2.15646863e-01 -7.06879854e-01
-1.07659781e+00 9.80985045e-01 2.21442245e-02 -1.77119533e-03
-2.76808947e-01 5.45703351e-01 7.75298715e-01 5.08377731e-01
2.58128822e-01 7.37142742e-01 -9.23391402e-01 -7.94216812e-01
-1.13159215e+00 4.52526897e-01 9.22583222e-01 9.60522115e-01
6.57231450e-01 4.07262027e-01 -2.77445614e-01 6.99970186e-01
2.42273971e-01 1.89994797e-01 9.79993880e-01 -1.36809540e+00
3.80509913e-01 8.61305594e-02 2.44803041e-01 7.57819414e-02
-3.65360200e-01 -5.63641727e-01 -2.44999364e-01 1.36732534e-01
1.65161848e-01 -4.51282114e-01 -8.91969204e-01 2.17470074e+00
3.03815044e-02 3.89631003e-01 2.33930841e-01 5.25779128e-01
-5.21563552e-02 6.17577374e-01 4.09775704e-01 -5.20855010e-01
7.44273782e-01 -1.45850730e+00 -5.79379976e-01 -5.93672276e-01
8.66325915e-01 -1.91584975e-01 1.35825431e+00 6.91667378e-01
-1.40460944e+00 -5.38315594e-01 -1.11001980e+00 -2.93436125e-02
-9.40961614e-02 -3.33010443e-02 9.12098050e-01 6.96202457e-01
-1.28713965e+00 8.12827528e-01 -1.12970638e+00 -6.45187572e-02
2.43774623e-01 4.10470188e-01 9.57885087e-02 2.66162485e-01
-1.25641870e+00 1.14570594e+00 8.06565762e-01 -3.41638654e-01
-1.63779533e+00 -8.87693465e-01 -8.52114141e-01 -2.71086451e-02
9.11366105e-01 -5.88205397e-01 1.99459267e+00 -9.65992749e-01
-2.24820590e+00 5.23504674e-01 -4.51035686e-02 -1.14298117e+00
6.52533352e-01 -4.41861957e-01 -2.47122005e-01 -2.29208738e-01
-3.45233858e-01 6.02249384e-01 8.65017176e-01 -1.21776056e+00
-5.91545582e-01 -2.23459154e-01 9.14172381e-02 2.54623622e-01
-9.07188579e-02 -1.54447049e-01 -4.02577966e-01 -6.94944263e-01
-7.17948318e-01 -9.75608289e-01 -6.02539241e-01 -5.51746249e-01
-2.10273996e-01 -2.35913351e-01 5.35864830e-01 -7.10399747e-01
1.59999537e+00 -1.72878528e+00 2.29427025e-01 1.99567392e-01
3.43804136e-02 3.95742118e-01 -5.03157914e-01 5.40404081e-01
5.14831766e-02 1.35907352e-01 -1.70536637e-01 -5.49736977e-01
4.27302957e-01 4.37114805e-01 -5.66353858e-01 1.30501479e-01
-1.47527829e-01 1.22072613e+00 -1.02958417e+00 -2.90775210e-01
2.07989872e-01 -6.69452474e-02 -9.17433023e-01 2.42788821e-01
-1.00421381e+00 3.49846095e-01 -3.98661584e-01 3.69256288e-01
5.18247895e-02 -1.87637940e-01 7.82712162e-01 3.72930467e-01
-1.36231065e-01 4.68047053e-01 -9.21029270e-01 2.12272763e+00
-9.17591512e-01 3.93501997e-01 7.58757219e-02 -1.01622832e+00
6.88152611e-01 2.27448642e-01 3.76384169e-01 -6.81880951e-01
-1.24195486e-01 2.16006935e-01 -1.40221179e-01 -2.34795749e-01
8.00167322e-01 -1.50337577e-01 -1.59644663e-01 6.45114005e-01
3.01789492e-01 -1.18098140e-01 2.79861361e-01 2.02626467e-01
1.14869463e+00 7.06022024e-01 5.39095044e-01 -3.01730722e-01
3.39172900e-01 -1.40951816e-02 4.83621240e-01 1.24059689e+00
-1.32575095e-01 -2.41385594e-01 3.93814415e-01 -2.72477061e-01
-1.26574743e+00 -9.54384804e-01 2.23308131e-01 1.55536187e+00
-3.53852808e-01 -7.65497863e-01 -8.17251563e-01 -9.48112845e-01
2.32793674e-01 1.22333729e+00 -7.18907118e-01 -3.71895701e-01
-6.88095450e-01 -4.39439446e-01 6.84598565e-01 5.82277596e-01
4.64433461e-01 -1.25007868e+00 -7.52377868e-01 6.07105792e-01
7.82969818e-02 -8.75053227e-01 -5.53377569e-01 7.41681397e-01
-1.14620972e+00 -5.32920837e-01 -3.56235594e-01 -4.22635436e-01
9.50505733e-02 -3.16277713e-01 1.42604542e+00 -2.22126648e-01
-1.66558549e-01 3.88471663e-01 -1.01293720e-01 -2.50443697e-01
-7.56789148e-01 2.96660721e-01 2.45559752e-01 -5.42276144e-01
1.18540652e-01 -6.32007182e-01 -1.68732747e-01 -7.13225082e-02
-7.86706984e-01 6.85963631e-02 6.32189035e-01 1.18320990e+00
6.47928059e-01 -7.04661459e-02 6.38708830e-01 -9.48194683e-01
8.78632605e-01 -3.57901901e-01 -8.77944052e-01 4.66305345e-01
-1.06716681e+00 8.71993661e-01 8.14391911e-01 -6.15043759e-01
-1.18608153e+00 -1.07608557e-01 -1.57811508e-01 -6.78420186e-01
1.07109912e-01 4.55856085e-01 -4.43355516e-02 -6.81150034e-02
6.19330883e-01 6.94832742e-01 1.04644388e-01 -5.50965548e-01
6.01984799e-01 2.74114490e-01 2.35630393e-01 -1.07577300e+00
6.64583683e-01 -6.19412884e-02 -1.85634360e-01 -4.33825582e-01
-8.82501960e-01 2.03227289e-02 -2.51210093e-01 2.72672065e-02
4.59977686e-01 -9.50108945e-01 -9.26430941e-01 6.79783076e-02
-5.34642816e-01 -1.53036904e+00 -7.11769044e-01 2.44559109e-01
-1.43463039e+00 2.06655264e-01 -7.26391375e-01 -1.07356870e+00
-2.62834489e-01 -1.15721214e+00 8.00922573e-01 8.21318775e-02
-1.45781547e-01 -1.19212115e+00 4.02031183e-01 3.27774175e-02
5.01046538e-01 -2.31999084e-01 1.06219554e+00 -7.16986001e-01
-4.70770985e-01 3.08679432e-01 3.73184919e-01 4.60319966e-01
-3.45058978e-01 -3.49901140e-01 -7.85039186e-01 -5.81406236e-01
-5.63060157e-02 -1.04530871e+00 9.35645998e-01 3.11689228e-01
1.62773955e+00 -5.56174159e-01 -1.27910629e-01 7.02663541e-01
1.43571103e+00 4.15015757e-01 4.07178491e-01 5.76233685e-01
2.87110448e-01 1.11027032e-01 8.41854572e-01 8.23793769e-01
3.08978945e-01 6.08778179e-01 1.35380685e-01 3.99520546e-01
1.21458188e-01 -9.61863220e-01 8.00338089e-01 5.64417839e-01
1.45525247e-01 -8.18617046e-02 -8.59409690e-01 3.18826258e-01
-2.06411266e+00 -1.01707590e+00 8.55066955e-01 2.20117164e+00
1.34035158e+00 3.23954672e-01 4.04551029e-01 -2.16321900e-01
8.15154389e-02 2.57637113e-01 -1.15588260e+00 -8.40020955e-01
2.14599371e-02 5.02395570e-01 7.24340141e-01 8.75881672e-01
-8.07732403e-01 1.36013627e+00 7.08809853e+00 1.22783184e+00
-8.89176071e-01 1.13594495e-01 8.32755864e-01 -2.94417173e-01
-4.67952132e-01 1.25448611e-02 -1.00556374e+00 3.61616701e-01
1.45952165e+00 -3.34303170e-01 1.33441460e+00 9.56009150e-01
1.32803813e-01 -4.80731055e-02 -1.12187052e+00 7.77667761e-01
-2.39528015e-01 -1.47261477e+00 -6.58710897e-02 1.05174959e-01
9.45119083e-01 3.58079135e-01 1.10814184e-01 1.12244189e+00
1.06496811e+00 -1.05754542e+00 8.79250467e-01 4.44237381e-01
6.52584970e-01 -1.09704673e+00 1.97864026e-02 9.05280471e-01
-8.76418293e-01 -6.11120045e-01 -2.09402338e-01 4.94855493e-02
-4.12867814e-02 1.92764789e-01 -8.07934761e-01 2.87329346e-01
1.96894243e-01 5.27926683e-01 -3.20034236e-01 5.72068572e-01
-3.11622709e-01 8.09965193e-01 -1.64319053e-01 -8.24391022e-02
5.12706697e-01 -8.85727182e-02 4.34039593e-01 1.14109719e+00
5.65064093e-03 -2.10777923e-01 4.60330456e-01 7.37477124e-01
-1.38479993e-01 -1.89110823e-02 -4.03201252e-01 -5.23104250e-01
4.03702438e-01 8.43310773e-01 -1.58145279e-01 -4.42505598e-01
-8.35696757e-02 8.25949669e-01 6.78996742e-01 4.71535981e-01
-7.38436997e-01 -3.92800458e-02 7.36603975e-01 -2.93565482e-01
5.24926424e-01 -4.87234384e-01 -2.47146457e-01 -1.25095034e+00
-3.49801451e-01 -1.45467627e+00 3.35942835e-01 -3.46770644e-01
-8.83286953e-01 3.76262307e-01 -7.22836331e-02 -6.77009761e-01
-7.70742893e-01 -5.65927804e-01 -2.02952191e-01 6.36037409e-01
-1.51427817e+00 -7.56831825e-01 4.06650245e-01 4.96422201e-01
9.45758760e-01 -1.82666376e-01 9.02084410e-01 -2.63710618e-01
-4.65461761e-01 8.69444311e-01 4.88259017e-01 -3.42052281e-01
2.53984809e-01 -1.47274876e+00 5.22284150e-01 5.69539607e-01
8.83795321e-02 5.25841773e-01 7.57414341e-01 -6.98923588e-01
-1.83341503e+00 -1.04428196e+00 5.73415697e-01 -3.01458180e-01
9.33282256e-01 -5.88720262e-01 -5.62985778e-01 1.10367703e+00
2.82996058e-01 -2.60906488e-01 3.98008168e-01 2.40839705e-01
-2.08623081e-01 1.60711855e-01 -8.95790398e-01 7.31069922e-01
1.20472324e+00 -6.15076780e-01 -4.50796396e-01 3.42210531e-01
9.72502768e-01 -5.91645837e-01 -9.50872898e-01 2.34038115e-01
4.34251457e-01 -6.24300241e-01 9.72565532e-01 -1.09561276e+00
1.23206936e-01 1.94746032e-01 -1.31499603e-01 -1.41299284e+00
-1.18220717e-01 -1.35312939e+00 -9.01994884e-01 7.66327798e-01
5.05995333e-01 -4.38243628e-01 8.31193089e-01 4.75763470e-01
-6.32350072e-02 -9.37890232e-01 -6.95164263e-01 -1.22115481e+00
5.11199594e-01 -6.29671574e-01 6.24983072e-01 5.52551091e-01
2.45384306e-01 1.30056560e-01 -6.44173861e-01 -3.70039046e-01
5.79866350e-01 1.01774447e-01 6.24241292e-01 -7.81922817e-01
-9.77290094e-01 -5.14907598e-01 3.63002837e-01 -1.44648886e+00
6.42534256e-01 -9.72843051e-01 2.35619694e-01 -1.00600994e+00
1.19457446e-01 -4.92773712e-01 -5.12326837e-01 4.65408355e-01
6.47018403e-02 -5.74065506e-01 3.00066859e-01 -1.23675615e-01
-8.33203614e-01 8.99824798e-01 1.21856427e+00 1.31815001e-01
-3.97959650e-01 -7.84344822e-02 -5.82328260e-01 3.89625669e-01
1.02805746e+00 -1.59451738e-01 -8.18317831e-01 -2.34355450e-01
2.81666607e-01 4.85937119e-01 -6.82413802e-02 -8.81372988e-01
1.40995145e-01 -4.50662762e-01 8.84643272e-02 -2.99197644e-01
3.44390899e-01 -3.88742864e-01 -1.52449787e-01 7.33897746e-01
-1.11448276e+00 1.29693076e-01 5.70397437e-01 7.61928856e-01
1.14991836e-01 -1.41519547e-01 7.96395361e-01 -2.44431481e-01
-7.83469975e-01 3.56330693e-01 -6.25177085e-01 4.51964170e-01
9.10081148e-01 1.23928361e-01 -4.66196463e-02 -3.79588723e-01
-1.01806068e+00 4.98901188e-01 4.05472815e-01 3.74187738e-01
3.97013396e-01 -1.09213376e+00 -4.23527449e-01 2.28095025e-01
-1.25695005e-01 -4.30873096e-01 1.04531378e-01 5.92076898e-01
-6.86187968e-02 8.82631898e-01 3.08075454e-02 7.97800347e-03
-6.70412958e-01 9.16698337e-01 4.59682077e-01 -1.12757468e+00
-3.38767976e-01 8.76258790e-01 4.71283793e-02 -7.55703032e-01
5.23226976e-01 -3.85802209e-01 1.70904219e-01 -3.17111939e-01
4.68592882e-01 2.73851782e-01 -1.09286889e-01 2.57855415e-01
-1.54346317e-01 -1.62855268e-01 -2.10323587e-01 -5.71435571e-01
1.22015727e+00 -7.24175870e-02 3.25152338e-01 6.28252685e-01
9.91195261e-01 -3.34964186e-01 -1.62605250e+00 -5.51237583e-01
3.71559352e-01 -2.06763849e-01 3.06155890e-01 -1.35590410e+00
-6.45625353e-01 7.16036916e-01 3.78570974e-01 -1.64975047e-01
9.48444963e-01 -3.43580782e-01 6.96572423e-01 7.90573597e-01
6.51644528e-01 -1.60129631e+00 1.13826081e-01 8.60718250e-01
7.07446635e-01 -1.02455568e+00 -3.76686603e-01 3.51498097e-01
-8.24434519e-01 1.03948212e+00 6.19072020e-01 -1.45852849e-01
4.81754959e-01 5.50039470e-01 -3.89423639e-01 3.50762963e-01
-1.63094175e+00 1.91956398e-03 -6.36690035e-02 4.07534152e-01
1.41931027e-01 2.66906768e-01 -6.84944391e-02 7.23972023e-01
-8.69468451e-02 4.23072278e-01 1.32403061e-01 1.17434013e+00
-6.28694892e-01 -1.52148473e+00 -8.65522772e-03 4.31954384e-01
-2.27747455e-01 -3.92120689e-01 -1.69814341e-02 7.27814078e-01
-3.86344373e-01 6.36129141e-01 -1.11998938e-01 -5.08384764e-01
-5.96905388e-02 3.70295554e-01 7.61325717e-01 -5.90375304e-01
-9.05492127e-01 6.39019832e-02 3.96483332e-01 -1.07050931e+00
1.84857979e-01 -8.30349863e-01 -1.18905663e+00 -2.45174840e-01
4.14982699e-02 3.68059367e-01 6.02569103e-01 1.07580638e+00
4.43921477e-01 5.28398752e-01 6.47087455e-01 -6.13254130e-01
-1.33290040e+00 -8.09934378e-01 -6.17990017e-01 -3.33801806e-01
4.42992359e-01 -5.80214202e-01 -1.51224658e-01 -2.37463564e-01] | [4.122016906738281, 1.8962018489837646] |
5b164e94-7892-494b-9811-a3118ffbd879 | dada-dialect-adaptation-via-dynamic | 2305.13406 | null | https://arxiv.org/abs/2305.13406v1 | https://arxiv.org/pdf/2305.13406v1.pdf | DADA: Dialect Adaptation via Dynamic Aggregation of Linguistic Rules | Existing large language models (LLMs) that mainly focus on Standard American English (SAE) often lead to significantly worse performance when being applied to other English dialects. While existing mitigations tackle discrepancies for individual target dialects, they assume access to high-accuracy dialect identification systems. The boundaries between dialects are inherently flexible, making it difficult to categorize language into discrete predefined categories. In this paper, we propose DADA (Dialect Adaptation via Dynamic Aggregation), a modular approach to imbue SAE-trained models with multi-dialectal robustness by composing adapters which handle specific linguistic features. The compositional architecture of DADA allows for both targeted adaptation to specific dialect variants and simultaneous adaptation to various dialects. We show that DADA is effective for both single task and instruction finetuned language models, offering an extensible and interpretable framework for adapting existing LLMs to different English dialects. | ['Diyi Yang', 'William Held', 'Yanchen Liu'] | 2023-05-22 | null | null | null | null | ['dialect-identification'] | ['natural-language-processing'] | [-1.08129509e-01 -2.77240962e-01 -3.77792031e-01 -7.00036824e-01
-1.07254076e+00 -1.22327435e+00 6.06934965e-01 3.79504226e-02
-3.47791314e-01 4.83237743e-01 3.39380831e-01 -7.78206050e-01
-6.09332435e-02 -6.50966406e-01 -6.04304969e-01 -1.19372986e-01
2.81889826e-01 6.75732434e-01 1.31058425e-01 -8.48021150e-01
-1.75108369e-02 6.71475172e-01 -1.43694425e+00 2.67391354e-01
1.33042848e+00 3.28507572e-01 2.59363204e-01 7.03855634e-01
-5.02574027e-01 6.08168066e-01 -6.65751398e-01 -5.57193160e-01
1.08476155e-01 -1.73144400e-01 -9.39676583e-01 -2.21413314e-01
9.98198628e-01 -1.13962933e-01 3.06277007e-01 8.58726203e-01
5.91227472e-01 -3.92404273e-02 4.90466535e-01 -9.03686225e-01
-7.87405789e-01 1.19338596e+00 9.65644792e-03 1.44771384e-02
3.54896784e-01 1.91561133e-01 9.86137867e-01 -8.52119386e-01
3.61477494e-01 1.47203481e+00 1.02434194e+00 7.90973186e-01
-1.63435781e+00 -6.35440886e-01 5.20906746e-01 -1.03094973e-01
-1.36544228e+00 -8.04251254e-01 4.14914936e-01 -4.84130383e-01
1.27764750e+00 5.38821995e-01 1.04466036e-01 1.08835614e+00
1.14723116e-01 6.79563105e-01 1.22542667e+00 -6.75348520e-01
-2.02612326e-01 5.64245462e-01 4.80466247e-01 4.53092247e-01
2.89631151e-02 -1.59625247e-01 -3.20470214e-01 -2.77322512e-02
3.15979809e-01 -5.18422127e-01 -7.03173652e-02 -4.42236543e-01
-1.28234661e+00 6.54832244e-01 -1.10077836e-01 3.92272294e-01
-3.14315297e-02 -3.79620969e-01 6.61086917e-01 8.17374945e-01
1.58155933e-01 7.51788497e-01 -9.25425887e-01 -1.76045686e-01
-5.78072727e-01 1.82558939e-01 6.85061991e-01 1.19290042e+00
8.58545005e-01 4.37672406e-01 -2.29048710e-02 1.50215948e+00
2.58412063e-02 5.88526726e-01 8.52112710e-01 -7.70247459e-01
4.62249190e-01 6.61956191e-01 -2.12994024e-01 -3.21190387e-01
-6.06845856e-01 -4.45025355e-01 -5.88447332e-01 3.38093549e-01
6.14000261e-01 5.60400113e-02 -5.58722675e-01 2.24016905e+00
2.06130370e-01 -5.65597236e-01 2.06612319e-01 2.59190530e-01
5.32542586e-01 3.87107700e-01 2.08470568e-01 3.37759465e-01
1.15050793e+00 -9.97965157e-01 -4.96083558e-01 -4.39274490e-01
8.47415686e-01 -1.06474161e+00 1.99483812e+00 2.67737269e-01
-1.11685658e+00 -9.04685736e-01 -1.05694091e+00 -3.54169071e-01
-7.94189155e-01 5.77509170e-03 4.05578852e-01 1.25247645e+00
-1.33685017e+00 6.38609901e-02 -4.05009806e-01 -4.70414758e-01
-3.69092911e-01 6.70870125e-01 -3.77532214e-01 1.74053639e-01
-1.24556994e+00 1.14429867e+00 3.08758169e-01 -4.39846337e-01
-6.83357179e-01 -9.23345983e-01 -1.03686249e+00 -1.31877899e-01
7.39079667e-03 -4.73465949e-01 1.32204378e+00 -1.19080770e+00
-1.80432940e+00 1.17846453e+00 -1.89117208e-01 -3.23492378e-01
3.02868485e-01 -2.53636330e-01 -8.54868948e-01 -6.56042218e-01
7.79957771e-02 3.43784332e-01 6.39624357e-01 -1.08787489e+00
-7.38470137e-01 -1.59016281e-01 1.44391105e-01 2.57882476e-01
-5.80173075e-01 3.95327479e-01 -2.82905281e-01 -8.73424768e-01
-4.00584280e-01 -8.45642328e-01 -5.16897365e-02 -7.09588945e-01
-3.92955631e-01 -2.38217652e-01 3.65031213e-01 -6.47373497e-01
1.74854422e+00 -1.92436385e+00 2.86162347e-01 2.28221938e-01
-1.56410672e-02 4.29869294e-01 -4.18845385e-01 2.68078178e-01
-1.47104084e-01 -7.20450841e-03 -2.58164436e-01 -4.80855733e-01
3.81908655e-01 2.31032029e-01 -3.66456181e-01 5.69942258e-02
1.34861112e-01 7.25278974e-01 -6.04097366e-01 -1.63994744e-01
2.35950798e-01 2.47137621e-01 -7.70686209e-01 4.06598598e-01
1.84916165e-02 2.03290835e-01 7.66914338e-02 5.79392791e-01
7.29079247e-01 5.05320847e-01 3.23687345e-01 1.03319816e-01
-4.30244744e-01 6.76738203e-01 -1.27516210e+00 1.56860495e+00
-1.20288134e+00 2.66686410e-01 4.03518319e-01 -5.27049661e-01
1.04398549e+00 8.38072679e-04 -6.38665780e-02 -4.08262998e-01
-2.70280004e-01 6.54634416e-01 3.36787701e-02 -9.49073434e-02
7.94530034e-01 -5.65050021e-02 -7.51437187e-01 5.44995129e-01
2.61099458e-01 -3.13696474e-01 1.35076761e-01 -1.13896005e-01
5.97980142e-01 2.57762343e-01 5.23333013e-01 -8.38901103e-01
1.03074348e+00 -1.58111244e-01 3.82438689e-01 9.54203963e-01
-1.16477177e-01 4.89459574e-01 -1.64931312e-01 -3.93230289e-01
-1.00097811e+00 -1.61861885e+00 -3.54258746e-01 1.82349253e+00
-2.77703822e-01 -5.45899749e-01 -9.87638891e-01 -7.55883396e-01
-1.85948405e-02 1.11913347e+00 -2.50535667e-01 -4.20949906e-01
-9.35991168e-01 -6.70226097e-01 1.01563656e+00 4.94608074e-01
2.61618048e-01 -8.22312295e-01 -1.07153524e-02 3.56893212e-01
-2.97978491e-01 -9.16448891e-01 -1.08113050e+00 3.48362446e-01
-4.73556846e-01 -7.67289281e-01 -3.95213038e-01 -1.10674548e+00
2.79586375e-01 -5.30135445e-03 1.62523341e+00 -1.96945027e-01
2.61933237e-01 5.60501575e-01 9.19749886e-02 -3.00134212e-01
-1.29127443e+00 7.89736748e-01 5.28938532e-01 1.16882600e-01
7.40418553e-01 -3.76968235e-01 1.11301750e-01 5.36820114e-01
-7.28218377e-01 -8.66474584e-02 2.67326772e-01 7.44940221e-01
3.82315665e-01 -5.40224016e-01 7.68513620e-01 -1.26569605e+00
8.01358879e-01 -4.48841095e-01 -5.93407035e-01 6.20159864e-01
-9.59553361e-01 3.19449902e-01 1.33732963e+00 -5.48614025e-01
-1.25039256e+00 -5.28578460e-02 -7.15566993e-01 3.94843578e-01
-5.75671256e-01 1.84054315e-01 -6.47973359e-01 -7.68062696e-02
9.21793163e-01 3.51322860e-01 -1.23350061e-01 -7.51276374e-01
6.26433730e-01 1.00484371e+00 8.32866371e-01 -1.15491617e+00
7.01446831e-01 -8.92712176e-02 -7.14629173e-01 -5.87461829e-01
-3.34944576e-01 -3.05615276e-01 -1.00541496e+00 1.09048627e-01
5.85146546e-01 -1.02938449e+00 -2.54864484e-01 8.55646193e-01
-6.80051446e-01 -7.23877907e-01 -2.86330938e-01 1.13793232e-01
-6.25999391e-01 2.43557781e-01 -6.28216326e-01 -2.96936244e-01
-3.23740304e-01 -1.39392734e+00 9.14419115e-01 -6.45933226e-02
-8.32973063e-01 -1.38992810e+00 1.03815749e-01 1.84121475e-01
8.69048893e-01 -4.68584597e-01 1.36691821e+00 -8.97909284e-01
-1.00332938e-01 -9.39163342e-02 1.99624851e-01 4.64996755e-01
4.91167367e-01 -2.01823059e-02 -1.05385756e+00 -4.28831965e-01
-2.32130989e-01 -4.43819046e-01 3.88205826e-01 5.93569241e-02
7.36818910e-01 -3.11039507e-01 4.52423654e-02 9.68970418e-01
1.14913356e+00 -4.62814197e-02 1.25043809e-01 7.24698663e-01
9.92703319e-01 6.76269948e-01 2.22216889e-01 -1.94489658e-01
8.18417668e-01 1.01074624e+00 -2.08495632e-01 -8.02269727e-02
-4.29623902e-01 -2.82727331e-01 9.76562560e-01 1.29294348e+00
3.69640797e-01 1.15133099e-01 -9.68414605e-01 7.81432867e-01
-1.33601749e+00 -5.95126688e-01 -6.74145147e-02 2.58524346e+00
1.42858005e+00 1.62256926e-01 3.16116989e-01 -1.63247719e-01
5.67529917e-01 -6.64306134e-02 -4.04153138e-01 -1.02933407e+00
-6.02120996e-01 1.33612528e-01 4.91628110e-01 1.15258181e+00
-8.48871052e-01 1.33386242e+00 7.08019066e+00 9.84842479e-01
-1.13435125e+00 1.53258637e-01 3.90736967e-01 1.53229132e-01
-7.61031449e-01 -3.93951461e-02 -1.31132698e+00 9.10591781e-02
1.31716800e+00 -2.79965639e-01 6.35293484e-01 7.00953364e-01
-2.63962839e-02 4.46833998e-01 -1.10781455e+00 6.38321400e-01
7.81070674e-04 -9.34425533e-01 4.39783216e-01 -4.44408089e-01
7.28959858e-01 1.56913936e-01 1.31069750e-01 6.62737489e-01
8.28923583e-01 -9.57506716e-01 1.09538543e+00 8.61785486e-02
1.10088265e+00 -8.38744402e-01 2.32355356e-01 3.15175563e-01
-1.22838199e+00 -1.56498849e-01 -2.00981125e-01 1.09338835e-01
-2.89939176e-02 -9.46463346e-02 -6.97760403e-01 3.31110299e-01
6.40164375e-01 3.49540710e-01 -1.00960422e+00 4.59325492e-01
1.64625153e-01 7.85384536e-01 -3.66356641e-01 3.42517257e-01
6.53380156e-02 -2.21229479e-01 6.17514789e-01 1.66174054e+00
2.39845708e-01 -8.65111470e-01 2.17162356e-01 5.97433448e-01
1.09197445e-01 5.26071966e-01 -6.25941217e-01 2.99153417e-01
6.55335784e-01 9.64113951e-01 -8.71641487e-02 -4.59659994e-01
-6.16969049e-01 9.60826457e-01 8.00413489e-01 4.07267392e-01
-4.63577241e-01 -4.44344372e-01 1.22502077e+00 9.78593826e-02
-8.46725851e-02 -3.17584991e-01 -5.41735113e-01 -1.38391924e+00
-1.96466729e-01 -1.57935059e+00 7.08452046e-01 -2.48607695e-01
-1.31763899e+00 7.22855747e-01 3.94409858e-02 -9.97923255e-01
-5.41728914e-01 -7.90844381e-01 -4.12113041e-01 1.14748025e+00
-1.34110832e+00 -1.50818956e+00 2.35540897e-01 8.39782655e-01
7.97154844e-01 -6.67424262e-01 1.37239265e+00 4.35275376e-01
-5.41400194e-01 1.45902264e+00 4.92226213e-01 -6.22115955e-02
1.18531156e+00 -1.70736456e+00 8.43245149e-01 9.48525310e-01
4.49606702e-02 1.03440058e+00 5.98804057e-01 -2.84757525e-01
-1.13444257e+00 -1.16786826e+00 1.41917503e+00 -9.20528233e-01
6.54403329e-01 -7.55079448e-01 -1.04856765e+00 1.10888851e+00
2.58381486e-01 -6.74179375e-01 8.73840809e-01 6.23260915e-01
-7.78524935e-01 -4.61371571e-01 -9.12929296e-01 1.01260877e+00
1.08067238e+00 -9.87920046e-01 -5.02296031e-01 -5.45361601e-02
6.28396332e-01 -2.76882261e-01 -1.00277472e+00 1.82380348e-01
4.41566885e-01 -1.08812726e+00 1.14739215e+00 -7.62571633e-01
-2.12387636e-01 -3.00154030e-01 -4.08485979e-01 -1.68585467e+00
-4.29515481e-01 -7.87976384e-01 3.40714395e-01 1.67938554e+00
6.99899256e-01 -8.30555677e-01 -3.42231058e-02 4.91125196e-01
-4.31000501e-01 -4.77408618e-02 -7.43380308e-01 -9.82608557e-01
8.98038983e-01 -6.92760408e-01 1.07635140e+00 1.25933015e+00
-2.79242657e-02 4.26210731e-01 -3.43460560e-01 1.43832266e-01
2.61318535e-01 3.95353474e-02 9.56336498e-01 -9.34920788e-01
-5.39315403e-01 -1.01724768e+00 1.07071754e-02 -1.01518798e+00
5.00675797e-01 -1.23285866e+00 -5.33931702e-03 -8.13648820e-01
-2.83547193e-01 -7.85675764e-01 -3.09285372e-01 4.34738398e-01
-5.38354635e-01 1.14568003e-01 1.00368418e-01 -4.33129705e-02
-2.58808672e-01 1.25301793e-01 5.66276133e-01 -2.38111287e-01
-3.91129524e-01 2.85889983e-01 -9.01984334e-01 5.91483116e-01
9.27439272e-01 -1.66547850e-01 -3.50109339e-01 -7.09241569e-01
8.34832340e-02 -6.38884306e-01 -2.57473201e-01 -9.44079936e-01
-9.45775583e-03 -1.09306842e-01 -1.61157004e-04 2.59624690e-01
-1.56468391e-01 -5.79439759e-01 5.74119352e-02 1.81323946e-01
-6.00271106e-01 5.18441796e-01 7.32664645e-01 -1.73559651e-01
-2.02636212e-01 -2.23795176e-01 9.70082700e-01 1.13975294e-02
-1.00419533e+00 -4.80047092e-02 -7.94429064e-01 2.67065793e-01
5.30030251e-01 -1.52159467e-01 -4.30483781e-02 -1.78742915e-01
-5.05189598e-01 8.33146200e-02 7.47944176e-01 9.53759372e-01
-2.40724138e-03 -1.42066109e+00 -9.88046229e-01 8.02968800e-01
3.96639645e-01 -3.08455050e-01 1.40984476e-01 3.64486605e-01
-2.84803063e-01 4.34154630e-01 -3.38443577e-01 -3.16431850e-01
-1.42666280e+00 6.13972366e-01 8.13257098e-01 -3.75837028e-01
-2.00383037e-01 8.32382679e-01 3.64124060e-01 -1.66765130e+00
1.57255694e-01 -3.22791606e-01 -2.00299695e-01 -7.85605088e-02
5.48240364e-01 1.47381574e-01 2.89999992e-01 -1.04469132e+00
-3.30033928e-01 7.49694407e-01 -3.00436050e-01 1.35167971e-01
7.67148793e-01 -7.20206022e-01 -1.39107153e-01 8.50213408e-01
9.67561901e-01 1.01756990e+00 -9.30117846e-01 -5.10966837e-01
3.10886502e-01 5.23331016e-02 -4.19581264e-01 -1.06546748e+00
-4.71138746e-01 7.24899113e-01 6.51856601e-01 1.12872265e-01
1.16198742e+00 -2.10846215e-01 6.81978405e-01 3.03358674e-01
3.56975406e-01 -1.21152556e+00 -6.87016606e-01 1.01881039e+00
7.17294157e-01 -1.23768389e+00 -4.85124886e-01 -8.55271295e-02
-6.76670015e-01 1.26051247e+00 8.48681331e-01 4.28238243e-01
4.98548359e-01 4.00159687e-01 8.63569677e-01 4.46697384e-01
-6.01563215e-01 -1.06995128e-01 5.50495327e-01 8.11003983e-01
8.63775671e-01 3.11375856e-01 -6.86362311e-02 8.21902812e-01
-6.26613379e-01 -5.17435372e-01 3.89570266e-01 5.31929612e-01
-2.58772939e-01 -1.73487175e+00 -6.33169591e-01 1.47286072e-01
-4.09875333e-01 -4.48146909e-01 -3.95431936e-01 7.24954963e-01
2.59125471e-01 9.00347710e-01 6.18118271e-02 -4.12885308e-01
5.77885985e-01 5.89192808e-01 2.98990846e-01 -6.83075607e-01
-1.20034206e+00 -2.61302084e-01 3.51545542e-01 -3.63084704e-01
1.56793460e-01 -7.04301059e-01 -1.09759033e+00 -4.43734348e-01
1.43114805e-01 -1.06987525e-02 2.12766200e-01 8.97791505e-01
3.61721247e-01 2.67624319e-01 5.33932567e-01 -3.87777030e-01
-7.75456071e-01 -8.96086037e-01 -4.47942853e-01 5.25893450e-01
5.38170457e-01 -2.44067132e-01 -1.06608286e-01 5.97251058e-02] | [11.093363761901855, 10.031608581542969] |
ab29a8cd-cac1-4978-b9ac-75baaa02b863 | autonomous-crater-detection-on-asteroids | 2204.00477 | null | https://arxiv.org/abs/2204.00477v1 | https://arxiv.org/pdf/2204.00477v1.pdf | Autonomous crater detection on asteroids using a fully-convolutional neural network | This paper shows the application of autonomous Crater Detection using the U-Net, a Fully-Convolutional Neural Network, on Ceres. The U-Net is trained on optical images of the Moon Global Morphology Mosaic based on data collected by the LRO and manual crater catalogues. The Moon-trained network will be tested on Dawn optical images of Ceres: this task is accomplished by means of a Transfer Learning (TL) approach. The trained model has been fine-tuned using 100, 500 and 1000 additional images of Ceres. The test performance was measured on 350 never before seen images, reaching a testing accuracy of 96.24%, 96.95% and 97.19%, respectively. This means that despite the intrinsic differences between the Moon and Ceres, TL works with encouraging results. The output of the U-Net contains predicted craters: it will be post-processed applying global thresholding for image binarization and a template matching algorithm to extract craters positions and radii in the pixel space. Post-processed craters will be counted and compared to the ground truth data in order to compute image segmentation metrics: precision, recall and F1 score. These indices will be computed, and their effect will be discussed for tasks such as automated crater cataloguing and optical navigation. | ['Fabio Curti', 'Dario Spiller', 'Francesco Latorre'] | 2022-04-01 | null | null | null | null | ['template-matching'] | ['computer-vision'] | [ 1.26726389e-01 1.64605938e-02 2.64981866e-01 -3.46362710e-01
-7.18260586e-01 -5.50248682e-01 7.71016419e-01 -4.51364011e-01
-7.58643985e-01 3.94862205e-01 -3.86639774e-01 -2.26024240e-01
1.53038368e-01 -9.94547069e-01 -7.12336779e-01 -8.74974787e-01
-2.62273282e-01 8.93696964e-01 2.23551735e-01 -3.90457004e-01
4.09249157e-01 6.89301491e-01 -1.58332133e+00 8.13350081e-02
3.46877873e-01 1.26579142e+00 3.81236255e-01 9.32838559e-01
4.78908181e-01 2.84957021e-01 -3.90065283e-01 1.93599388e-01
7.20349610e-01 -2.91698158e-01 -3.64116907e-01 -5.38866706e-02
5.02446949e-01 -6.73223957e-02 -2.31653363e-01 8.94782543e-01
1.90547332e-01 7.44285807e-02 8.72132719e-01 -5.11874259e-01
-3.75212401e-01 4.05841887e-01 -5.48839331e-01 5.16426086e-01
1.92758784e-01 2.22982407e-01 9.18096423e-01 -9.26348984e-01
6.97504938e-01 9.11560655e-01 1.03675294e+00 1.08732499e-01
-7.98768580e-01 -8.23676109e-01 -8.50437939e-01 1.92005128e-01
-1.30514419e+00 -2.84115583e-01 3.45497459e-01 -6.46168947e-01
1.08546746e+00 1.88811511e-01 9.56788182e-01 2.38628477e-01
3.89075160e-01 5.16832054e-01 1.15404344e+00 -6.63773537e-01
1.82402030e-01 -7.78782442e-02 -1.37184575e-01 7.42844760e-01
2.65369236e-01 5.46791911e-01 -3.15966159e-01 2.98712134e-01
8.88925731e-01 -3.77440780e-01 -2.39178985e-01 -2.17999518e-01
-1.15220118e+00 1.21570313e+00 7.49619663e-01 5.82251787e-01
-4.73915398e-01 -8.55758507e-03 2.74833798e-01 3.63664687e-01
3.05509359e-01 6.39647067e-01 -2.85431296e-01 2.80961335e-01
-1.42103899e+00 3.79146934e-01 6.36727273e-01 5.87888360e-01
1.11061537e+00 1.56905562e-01 3.90302956e-01 6.20799661e-01
5.66093028e-01 7.51631856e-01 6.75421655e-01 -5.45524538e-01
-8.68572965e-02 6.68568134e-01 -7.46250898e-02 -8.77406895e-01
-8.56934369e-01 -2.45648354e-01 -5.89407861e-01 6.63019896e-01
3.03828716e-01 -4.11728062e-02 -1.49387848e+00 8.58623624e-01
1.34304509e-01 -1.66425064e-01 1.43375009e-01 1.29909706e+00
9.29791152e-01 5.66536784e-01 -3.50677431e-01 3.43619168e-01
1.43138540e+00 -6.98386192e-01 -1.46848008e-01 -6.25117362e-01
9.99381617e-02 -9.23493564e-01 8.06030273e-01 3.39519024e-01
-6.75293863e-01 -5.78463495e-01 -1.55992782e+00 1.39550924e-01
-5.96652508e-01 5.41206837e-01 7.92927921e-01 6.72259212e-01
-1.02280998e+00 5.15386283e-01 -8.81516993e-01 -5.45280516e-01
2.65138268e-01 3.23856533e-01 -2.72071719e-01 2.08381265e-01
-1.05112338e+00 1.05584919e+00 6.68240428e-01 3.23227108e-01
-1.46455359e+00 -9.26924720e-02 -1.22469771e+00 -1.42254189e-01
-2.53017604e-01 -6.19428694e-01 1.26790655e+00 -8.80876839e-01
-1.13179672e+00 1.26913404e+00 3.79374474e-01 -1.04275060e+00
4.45754260e-01 1.01310320e-01 -3.00720125e-01 1.70245335e-01
3.20871353e-01 6.73203468e-01 7.64721155e-01 -1.22809803e+00
-1.09257388e+00 -4.95368958e-01 -3.67910951e-01 2.33886048e-01
4.80781883e-01 -1.26472831e-01 -7.57002175e-01 -2.89780408e-01
4.88488853e-01 -1.06698418e+00 -1.25084326e-01 -7.04800308e-01
-2.19424248e-01 3.00680175e-02 8.64843071e-01 -1.00035322e+00
4.05879498e-01 -1.87468767e+00 -2.38315463e-01 4.76922423e-01
-2.43046999e-01 -5.89467436e-02 1.67242110e-01 -5.11127822e-02
1.58714145e-01 -5.21890223e-01 -7.84810066e-01 -3.19414437e-02
-3.12323779e-01 6.62730411e-02 1.13981448e-01 8.81598592e-01
-2.30038360e-01 6.94490969e-01 -5.65018535e-01 -2.97938436e-01
5.69267511e-01 2.18731046e-01 -1.53408751e-01 -2.81782970e-02
-2.86806077e-01 3.96762967e-01 -1.08805336e-01 1.10888028e+00
7.64116585e-01 3.18893313e-01 -2.01273605e-01 -6.05766848e-02
-4.17413026e-01 -1.65461935e-02 -7.43507266e-01 1.53115106e+00
-5.63062429e-01 1.21485353e+00 2.89188623e-01 -7.57524192e-01
1.31153190e+00 -7.69436359e-02 1.57652766e-01 -1.02603340e+00
5.36458075e-01 3.28026414e-01 8.45194086e-02 -5.22154927e-01
1.04301333e+00 -4.14979637e-01 -2.25036755e-01 3.49586010e-01
7.24736229e-02 -6.29407108e-01 3.07965606e-01 1.66652910e-02
7.68017590e-01 1.47002101e-01 2.32733693e-02 -4.82354850e-01
4.85338628e-01 7.24655092e-01 1.81264594e-01 8.38036597e-01
1.04579039e-01 7.53774703e-01 -4.25763205e-02 -9.46322262e-01
-1.35199571e+00 -7.05983758e-01 -4.92457539e-01 8.00021648e-01
1.27183720e-01 6.84185699e-02 -8.17133307e-01 -3.27847153e-01
8.08948725e-02 6.62537396e-01 -9.91685450e-01 8.74637254e-03
-3.22489858e-01 -1.21660161e+00 5.38689613e-01 3.19485307e-01
5.11546075e-01 -1.49749482e+00 -1.16393578e+00 3.74229737e-02
9.86272916e-02 -8.82871211e-01 2.02985749e-01 4.06884223e-01
-7.05391228e-01 -1.61556661e+00 -4.68661100e-01 -9.61956143e-01
5.47162354e-01 -4.76349294e-02 1.02915442e+00 -1.76351070e-01
-6.35876536e-01 2.46682480e-01 -3.82849723e-01 -4.39440072e-01
-4.34463501e-01 4.78116749e-03 -3.10353767e-02 -1.72315702e-01
6.36271834e-01 -2.46273682e-01 -6.49036467e-01 3.90768677e-01
-7.04498649e-01 8.04226175e-02 8.76837194e-01 8.68765175e-01
7.25988090e-01 -1.24355733e-01 -1.46092065e-02 -5.66993415e-01
3.34596694e-01 -1.11686438e-01 -1.10743546e+00 -1.73292682e-02
-7.36904979e-01 -2.72156507e-01 1.53028756e-01 1.86709002e-01
-9.07972157e-01 5.81933439e-01 -1.23809464e-01 -2.89424121e-01
-1.49101034e-01 3.51850510e-01 3.45859557e-01 -5.63109398e-01
9.28158164e-01 4.51416075e-01 -5.91714978e-02 -3.01068217e-01
2.48595446e-01 9.24908638e-01 1.17274117e+00 3.60516794e-02
8.44171822e-01 8.33721995e-01 -3.60133559e-01 -1.06780791e+00
-5.75880051e-01 -8.19876611e-01 -6.00225866e-01 -4.26429987e-01
1.16969752e+00 -1.04882407e+00 -4.13281500e-01 5.43567181e-01
-7.89730787e-01 -4.77647096e-01 -3.56863886e-02 5.62655509e-01
-5.98958671e-01 5.70684485e-02 -4.86762762e-01 -8.42225492e-01
-8.47165108e-01 -1.13786042e+00 1.02768147e+00 5.61251462e-01
2.76410341e-01 -6.75660789e-01 4.47933853e-01 4.24117327e-01
2.81944454e-01 4.86507058e-01 2.68897027e-01 -5.43513119e-01
-4.75627601e-01 -6.62437201e-01 -3.59526664e-01 2.95179754e-01
-2.43549809e-01 -2.52096117e-01 -1.07869995e+00 -2.44613588e-01
-1.35387909e-02 -2.72145957e-01 1.45422769e+00 5.78623235e-01
4.89795506e-01 4.36578572e-01 -3.38939160e-01 9.84688938e-01
1.57915401e+00 4.75378364e-01 7.08118081e-01 1.25771701e+00
3.01706821e-01 1.01587832e-01 7.82598794e-01 3.80228966e-01
2.09299177e-01 3.28107029e-01 7.93978214e-01 -3.40534411e-02
-2.89546046e-02 2.01993629e-01 1.01491302e-01 1.52438834e-01
-5.65104127e-01 5.58910549e-01 -1.24172366e+00 7.72199333e-01
-1.45483983e+00 -7.00078309e-01 -2.46293157e-01 2.20315218e+00
3.80845606e-01 2.46533647e-01 -2.24838272e-01 6.81984872e-02
6.31853759e-01 -8.42730179e-02 -2.65765905e-01 -3.93254817e-01
-3.80776495e-01 3.13310444e-01 1.22310519e+00 5.37586808e-01
-1.46466458e+00 1.16159964e+00 6.52150583e+00 5.08717060e-01
-1.17808223e+00 3.80151980e-02 6.10371411e-01 2.37948358e-01
2.03125149e-01 1.11040734e-01 -7.99217105e-01 1.96193784e-01
7.47794509e-01 2.42358729e-01 6.29522264e-01 1.08558571e+00
-5.66204370e-04 -7.13872433e-01 -5.62949836e-01 9.58878517e-01
5.09536624e-01 -1.49703145e+00 -3.99579197e-01 -8.54213312e-02
8.53017032e-01 9.00863469e-01 -4.77290720e-01 4.27662253e-01
1.21980004e-01 -9.50591564e-01 8.99813235e-01 5.52467704e-01
8.89855266e-01 -9.54696596e-01 1.12203932e+00 6.25126660e-02
-1.03592634e+00 6.82923123e-02 -6.95811093e-01 -5.23555987e-02
9.05175805e-02 4.32085663e-01 -1.37549031e+00 4.33897913e-01
1.22499359e+00 3.02688509e-01 -6.79466426e-01 1.22121668e+00
-3.10280889e-01 2.69996345e-01 -7.01121032e-01 -3.67580913e-02
6.10543072e-01 -5.27189493e-01 4.29703623e-01 1.54782450e+00
3.54967594e-01 -2.30858922e-01 -2.22701564e-01 7.38325059e-01
6.34047017e-02 -8.41057580e-03 -5.81803679e-01 1.19570501e-01
-3.41655649e-02 1.83843648e+00 -1.15356112e+00 -4.14441437e-01
-3.28957662e-02 6.54568374e-01 -1.98562101e-01 -1.92170948e-01
-6.60831034e-01 -6.11630857e-01 1.21813994e-02 7.86657706e-02
5.19415855e-01 -1.08749427e-01 -5.27636886e-01 -6.48867011e-01
-4.62272078e-01 -5.05790532e-01 5.28852344e-01 -1.09406626e+00
-7.78745115e-01 9.20143664e-01 -2.93633491e-01 -1.06224334e+00
-1.05300762e-01 -9.19263899e-01 -9.49239075e-01 7.03389704e-01
-1.40187263e+00 -1.27981079e+00 -5.73511302e-01 3.06358159e-01
4.69486713e-01 -6.89684331e-01 5.80619335e-01 1.01331048e-01
4.82842214e-02 1.91773936e-01 3.01065505e-01 2.65349239e-01
3.51714343e-01 -1.41623390e+00 4.45162356e-01 7.28217125e-01
2.58468479e-01 4.84829172e-02 9.25257146e-01 -6.93780124e-01
-1.26978552e+00 -1.00771105e+00 6.09347045e-01 -3.20252091e-01
6.73887908e-01 -2.13213950e-01 -6.55993044e-01 8.16872597e-01
3.02973837e-01 -1.98219538e-01 3.23801249e-01 -1.68549344e-01
1.01244554e-01 5.00600711e-02 -1.25574517e+00 -1.02336913e-01
2.00739622e-01 -3.03343832e-01 -8.98756742e-01 5.81772983e-01
-1.65375203e-01 -7.64286637e-01 -6.02042198e-01 6.22022867e-01
6.74865544e-01 -1.23153639e+00 6.91545904e-01 9.30648148e-02
3.38751853e-01 -4.71970439e-01 -4.02710736e-02 -1.22693002e+00
-2.91080866e-02 -2.28354290e-01 5.49628019e-01 4.46081221e-01
6.61046088e-01 -2.16848314e-01 1.04596138e+00 -2.91973174e-01
-5.52322328e-01 -2.70034373e-01 -1.15692091e+00 -5.37559509e-01
-2.02219099e-01 -7.04710066e-01 1.86884880e-01 8.63673151e-01
-3.02279800e-01 2.33555734e-01 -1.99965119e-01 2.77596563e-01
6.20661497e-01 5.70242226e-01 9.10395563e-01 -1.25795496e+00
-9.89614129e-02 -3.43178153e-01 -4.47575569e-01 -3.73157263e-01
-3.07443470e-01 -9.36025977e-01 4.72568750e-01 -1.61144781e+00
9.98501386e-03 -5.23159325e-01 1.56332254e-01 5.22768915e-01
5.41810811e-01 1.15302181e+00 -1.45927053e-02 4.80239391e-01
-2.90726006e-01 3.85678142e-01 9.07033503e-01 -1.11844815e-01
-1.90265402e-01 5.76617569e-02 1.05273291e-01 9.62285399e-01
9.51897919e-01 -3.35766882e-01 5.76367259e-01 -1.32610962e-01
2.75272816e-01 -2.00012811e-02 3.96300077e-01 -1.50461018e+00
2.03016877e-01 4.48601991e-01 8.62405300e-01 -1.20767438e+00
3.90014410e-01 -6.75471485e-01 8.12407210e-02 7.70268559e-01
4.53453690e-01 -1.67696849e-01 3.89384657e-01 4.18852717e-01
-2.36614674e-01 -4.89356697e-01 1.06892002e+00 -5.04040539e-01
-1.12054610e+00 -4.76244427e-02 -4.15962458e-01 -7.44627059e-01
1.03829551e+00 -4.08811331e-01 1.72897175e-01 -1.28681123e-01
-9.04152870e-01 1.91757247e-01 6.06159389e-01 4.48856115e-01
5.54230690e-01 -9.22594845e-01 -7.42477179e-01 5.49929738e-01
2.86165804e-01 2.43898451e-01 1.91459939e-01 8.67374659e-01
-1.35103703e+00 5.39908409e-01 -4.12218183e-01 -9.15556133e-01
-1.07421684e+00 1.80561736e-01 6.91143751e-01 3.43745400e-04
-7.65461147e-01 8.07989717e-01 -9.00339335e-02 -7.43856847e-01
-2.76405543e-01 -1.66209951e-01 -5.20788670e-01 1.51215225e-01
4.41979200e-01 5.13508124e-03 4.58815098e-01 -8.71402264e-01
-2.47363552e-01 6.98878706e-01 1.83349550e-01 -8.14566389e-02
1.56478274e+00 1.41449332e-01 -2.09825650e-01 1.41848803e-01
6.71514928e-01 -1.45524114e-01 -1.24449217e+00 -2.45273840e-02
1.38261661e-01 -2.26916909e-01 3.68780017e-01 -9.26156938e-01
-1.32895947e+00 4.61137414e-01 1.04618669e+00 1.49104491e-01
9.52447414e-01 3.73378515e-01 6.13065243e-01 5.27990162e-01
2.02622011e-01 -1.20913231e+00 -5.92144728e-01 7.04575598e-01
8.74095261e-01 -1.43694746e+00 1.47808969e-01 4.54167336e-01
-5.06794095e-01 1.53803957e+00 4.45340067e-01 -6.61256671e-01
2.66371071e-01 3.29065561e-01 3.95556182e-01 -6.96834922e-01
9.27293673e-03 -3.76312941e-01 2.34619156e-01 4.50786710e-01
1.58617049e-01 2.36177713e-01 -4.46323097e-01 4.99880582e-01
-8.89210939e-01 -1.89787254e-01 5.34167349e-01 6.85562372e-01
-1.02356040e+00 -3.89510304e-01 -1.06646216e+00 5.28026819e-01
-2.23416582e-01 -1.00646853e-01 -4.78292227e-01 1.07264638e+00
3.96642268e-01 4.60901767e-01 4.83498514e-01 -4.34054554e-01
1.94297824e-02 -9.56509113e-02 3.27805996e-01 -2.57152706e-01
-6.25354290e-01 2.32837275e-01 5.61393574e-02 2.28991569e-03
-4.52600598e-01 -7.89716601e-01 -1.31400549e+00 1.99049458e-01
-4.31624800e-01 3.39242816e-01 1.23544502e+00 9.11173820e-01
-3.57443869e-01 3.73630911e-01 4.91384089e-01 -1.57086480e+00
-1.13169802e-02 -1.48264563e+00 -7.63663232e-01 7.20846727e-02
1.85291171e-01 -5.92461228e-01 -5.91713250e-01 1.12593569e-01] | [8.616206169128418, -1.5213983058929443] |
a76d9632-ceee-4a5a-9584-b2ef7b22d59e | masked-vision-language-transformers-for-scene | 2211.04785 | null | https://arxiv.org/abs/2211.04785v1 | https://arxiv.org/pdf/2211.04785v1.pdf | Masked Vision-Language Transformers for Scene Text Recognition | Scene text recognition (STR) enables computers to recognize and read the text in various real-world scenes. Recent STR models benefit from taking linguistic information in addition to visual cues into consideration. We propose a novel Masked Vision-Language Transformers (MVLT) to capture both the explicit and the implicit linguistic information. Our encoder is a Vision Transformer, and our decoder is a multi-modal Transformer. MVLT is trained in two stages: in the first stage, we design a STR-tailored pretraining method based on a masking strategy; in the second stage, we fine-tune our model and adopt an iterative correction method to improve the performance. MVLT attains superior results compared to state-of-the-art STR models on several benchmarks. Our code and model are available at https://github.com/onealwj/MVLT. | ['Jian Zhang', 'Weigang Qi', 'Shengming Zhang', 'Ying Peng', 'Jie Wu'] | 2022-11-09 | null | null | null | null | ['scene-text-recognition'] | ['computer-vision'] | [ 3.97141516e-01 -4.59636748e-01 -6.42244592e-02 -4.11310852e-01
-5.53856730e-01 -2.31142014e-01 7.43772328e-01 -6.03261851e-02
-4.75291669e-01 1.45247459e-01 5.18854082e-01 -3.97099733e-01
5.29382825e-01 -5.89155078e-01 -7.23702371e-01 -5.07589579e-01
8.30256522e-01 2.02865079e-01 6.39393747e-01 4.35241871e-02
3.97649765e-01 8.02240595e-02 -1.17469323e+00 7.91108549e-01
8.77260387e-01 1.00588572e+00 6.73054278e-01 6.05334520e-01
-5.28276265e-01 1.34667742e+00 -5.05516715e-02 -4.62961107e-01
1.29600957e-01 -4.42191511e-01 -6.42296553e-01 4.28198427e-01
2.95913517e-01 -2.40993395e-01 -5.49214423e-01 1.05371249e+00
4.23683643e-01 -1.60914771e-02 4.57026005e-01 -7.60773301e-01
-9.46979642e-01 5.83550096e-01 -8.02369475e-01 1.04092002e-01
2.73188204e-01 3.27524424e-01 8.52584422e-01 -1.47486639e+00
1.40209377e-01 1.26335609e+00 4.32459444e-01 4.65290874e-01
-1.12036169e+00 -3.65185231e-01 3.79120827e-01 3.60881358e-01
-1.54889715e+00 -9.14739311e-01 8.21419001e-01 -5.63522160e-01
9.15589750e-01 5.21350764e-02 2.94238597e-01 1.02812958e+00
2.92604774e-01 1.15063322e+00 1.12700784e+00 -6.03752196e-01
-1.44546507e-02 1.79822952e-01 8.64555687e-02 9.02378738e-01
-1.09003350e-01 -1.41009197e-01 -6.79740250e-01 3.46338660e-01
7.28564262e-01 2.17649534e-01 -4.07730639e-01 -1.04692735e-01
-1.25947320e+00 6.42078280e-01 3.84421915e-01 3.80630046e-01
-2.73627013e-01 9.58473533e-02 3.35003585e-01 1.44100934e-01
3.88872027e-01 -1.34068608e-01 -6.38904050e-02 6.54735863e-02
-1.12979281e+00 -4.29652095e-01 5.00928283e-01 7.84386694e-01
8.18147957e-01 9.35732797e-02 -4.95729983e-01 1.00535703e+00
5.51439226e-01 6.21631384e-01 7.61371732e-01 -3.59424114e-01
8.55580211e-01 8.17353845e-01 -1.17094301e-01 -8.67309213e-01
-1.81217566e-02 -3.19032103e-01 -9.55186069e-01 -7.87216872e-02
4.21508402e-03 1.88227713e-01 -1.25408638e+00 1.22159648e+00
-3.32138464e-02 2.19153851e-01 4.18668985e-02 1.00304377e+00
1.07768381e+00 7.49756634e-01 -1.39467210e-01 1.16341440e-02
1.43526495e+00 -1.40254378e+00 -7.17839062e-01 -6.34704471e-01
4.97773886e-01 -8.89119148e-01 1.43241704e+00 2.80893683e-01
-1.05521345e+00 -5.01957774e-01 -9.60692644e-01 -4.65418667e-01
-1.27812400e-01 8.68433475e-01 -9.15931724e-03 3.83867741e-01
-1.05784237e+00 -2.63381898e-02 -9.61655974e-01 -3.54572266e-01
4.25875336e-01 2.00902864e-01 -1.70873269e-01 -7.58295581e-02
-8.98248911e-01 6.28066003e-01 2.76404649e-01 3.62676561e-01
-9.81264055e-01 1.29603576e-02 -9.78141367e-01 2.81295534e-02
4.96152401e-01 -6.57455504e-01 1.49174523e+00 -1.27255976e+00
-1.70346367e+00 9.55723464e-01 -6.95555687e-01 -3.53837401e-01
6.37580693e-01 -1.95237160e-01 -2.73855627e-01 2.43999809e-01
3.40801850e-02 4.27151829e-01 1.18275034e+00 -1.13946843e+00
-4.42734301e-01 -2.49538705e-01 -8.27500671e-02 4.18255150e-01
-4.17545170e-01 1.97083175e-01 -9.14490938e-01 -7.40331411e-01
2.32807305e-02 -4.03456986e-01 -7.65026286e-02 4.33460437e-03
-5.75631082e-01 7.48311803e-02 7.91366994e-01 -8.15936446e-01
1.33773172e+00 -2.28599167e+00 9.13785100e-02 -2.98856318e-01
3.98118198e-01 5.01201093e-01 -8.58448073e-02 5.01658440e-01
2.10689440e-01 -1.11667983e-01 -2.67623425e-01 -7.99085021e-01
1.28007708e-02 -7.58664459e-02 -4.42951143e-01 2.68075287e-01
4.92480509e-02 1.04395127e+00 -5.35216868e-01 -6.56385660e-01
3.86480719e-01 3.83303106e-01 -2.95712680e-01 2.50151366e-01
-4.80478168e-01 3.51767361e-01 -5.02652049e-01 5.88946998e-01
6.42380178e-01 -7.08903432e-01 -1.05117850e-01 -1.74427181e-01
-3.01938593e-01 4.82447654e-01 -8.61816347e-01 1.56580091e+00
-3.19348484e-01 6.54945493e-01 -1.19744223e-02 -8.57557714e-01
8.90701532e-01 2.80563802e-01 -1.22297801e-01 -9.40675318e-01
4.22170848e-01 5.36362873e-03 -4.09563899e-01 -5.91891825e-01
4.52894449e-01 -4.25310843e-02 2.46141091e-01 4.58560407e-01
-1.74994752e-01 2.57299066e-01 -4.80107293e-02 2.47035697e-01
9.22722161e-01 1.15836905e-02 2.77036607e-01 8.65196362e-02
9.73831892e-01 -1.27971247e-01 5.37509263e-01 6.62897944e-01
-1.96131505e-02 7.90348053e-01 2.47201487e-01 -2.83691347e-01
-6.96331561e-01 -1.01273298e+00 1.78082526e-01 1.00537336e+00
2.46352196e-01 -4.40613538e-01 -6.69705570e-01 -5.78339338e-01
-2.34293804e-01 7.57618606e-01 -4.23543513e-01 -1.72908515e-01
-2.68907964e-01 -4.62819606e-01 4.92863238e-01 7.49115765e-01
1.14895391e+00 -9.08761501e-01 -5.06560862e-01 -4.07145321e-02
-3.39006662e-01 -1.35462749e+00 -1.02372944e+00 1.23073934e-02
-6.28211379e-01 -6.08071089e-01 -7.06376672e-01 -9.99274015e-01
8.16931665e-01 5.66233337e-01 7.04265594e-01 5.68063296e-02
4.61752042e-02 1.34787574e-01 -4.40191567e-01 -1.33239329e-01
-2.44633943e-01 -3.89011502e-02 -2.71475762e-01 5.52754521e-01
6.26508772e-01 -2.79794097e-01 -5.36407411e-01 3.37896109e-01
-7.92128563e-01 7.69096673e-01 8.03197742e-01 8.19393933e-01
5.79991400e-01 -6.67064041e-02 3.40879982e-04 -7.69984484e-01
4.54671949e-01 -7.56530762e-02 -7.07335174e-01 4.97876763e-01
-4.48996753e-01 7.76024535e-03 6.96351230e-01 -4.77905273e-01
-1.22327650e+00 2.43631065e-01 -9.29993242e-02 -6.60775304e-01
-9.18126330e-02 7.14962840e-01 -2.65077591e-01 -7.92738795e-02
3.95182550e-01 7.56442428e-01 -3.61477882e-01 -6.75655961e-01
1.75336212e-01 1.07341075e+00 3.96807611e-01 -2.80144066e-01
6.12497389e-01 3.34155500e-01 -6.09122932e-01 -8.51563573e-01
-9.38800037e-01 -4.03929800e-01 -7.09167421e-01 -1.01954594e-01
8.98629308e-01 -1.13769090e+00 -4.13473547e-01 7.77897179e-01
-1.11836565e+00 -7.69148827e-01 6.34085611e-02 4.58514154e-01
-3.75961572e-01 5.94503105e-01 -7.03068554e-01 -7.18259633e-01
-4.25848663e-01 -1.17470312e+00 1.21951962e+00 1.72511071e-01
3.75374317e-01 -8.80085945e-01 -1.10564247e-01 5.86757958e-01
3.04814458e-01 -3.56344730e-01 6.41888142e-01 -5.35413623e-01
-8.55637014e-01 1.67680144e-01 -6.62757695e-01 3.19247752e-01
3.14897060e-01 -1.80808455e-01 -1.00665247e+00 -2.78681725e-01
2.61615962e-01 -2.74353385e-01 1.33642137e+00 2.22323671e-01
1.17132103e+00 -4.30951834e-01 -2.25852951e-01 6.94711626e-01
1.38103545e+00 1.21291921e-01 6.13031745e-01 2.17414558e-01
1.18879640e+00 1.51244149e-01 2.95182198e-01 3.46820623e-01
8.28317583e-01 7.18985915e-01 1.33612439e-01 -1.59005895e-02
-3.72720033e-01 -5.11549771e-01 7.33306348e-01 1.18207359e+00
3.47296238e-01 -3.91050488e-01 -1.26785886e+00 4.71193641e-01
-2.16449738e+00 -8.25313985e-01 -1.49150893e-01 2.12619042e+00
9.12374973e-01 3.11951369e-01 -1.24397472e-01 -2.54756515e-03
8.83286476e-01 3.58789712e-01 -5.39326370e-01 -3.13286871e-01
-1.84102133e-01 -1.53573468e-01 4.02338266e-01 6.66607082e-01
-1.08196974e+00 1.33349729e+00 5.40663099e+00 7.37037361e-01
-1.55779791e+00 2.43859380e-01 4.51055825e-01 6.29588664e-02
-2.58961618e-01 5.28528877e-02 -8.31959426e-01 4.70469892e-01
6.98782504e-01 -8.65864828e-02 4.95877117e-01 4.19820875e-01
3.95032614e-01 -6.40567243e-02 -8.88564348e-01 1.02915406e+00
4.05855089e-01 -1.31139946e+00 2.37902224e-01 -1.78567663e-01
4.34521347e-01 3.11963558e-01 3.89341749e-02 5.54992668e-02
1.39554292e-01 -9.78248358e-01 1.03107500e+00 6.60177946e-01
8.45573843e-01 -1.58845603e-01 5.38865924e-01 6.51430011e-01
-1.46094096e+00 2.08248235e-02 -2.90689707e-01 -8.80219266e-02
-4.82898988e-02 6.28115475e-01 -5.71692586e-01 4.48129714e-01
4.73815560e-01 1.02368379e+00 -8.15654635e-01 9.97514307e-01
-5.16235054e-01 8.11089575e-01 -4.31791097e-02 6.82116486e-03
1.45882651e-01 -9.15952846e-02 4.37346727e-01 1.35861158e+00
3.53148580e-02 -6.36593476e-02 3.37528437e-01 9.24944699e-01
-2.04097003e-01 2.08481267e-01 -4.24860448e-01 -2.33156949e-01
2.96848357e-01 9.97287273e-01 -6.11986339e-01 -2.35985190e-01
-8.87043357e-01 1.32751763e+00 4.26275939e-01 4.73362595e-01
-7.69013643e-01 -3.51843029e-01 -2.54557212e-03 8.07465240e-02
6.84787869e-01 -3.67380053e-01 -4.35330451e-01 -1.74674630e+00
2.98950851e-01 -8.80446017e-01 2.50955015e-01 -1.08377671e+00
-1.03043163e+00 6.36409521e-01 -4.74937826e-01 -1.14160526e+00
1.22062869e-01 -6.77631617e-01 -5.59790790e-01 9.05722141e-01
-1.66612589e+00 -1.50161326e+00 -4.86384422e-01 7.85702765e-01
9.09891546e-01 -1.30538046e-01 5.13057768e-01 2.46337041e-01
-8.27194929e-01 6.79741621e-01 7.48331696e-02 4.26903784e-01
6.95820808e-01 -9.93720591e-01 5.10335505e-01 1.28146982e+00
1.90654173e-01 4.47684586e-01 2.70234704e-01 -6.54663801e-01
-1.44727492e+00 -1.25954616e+00 1.14670956e+00 -2.88226634e-01
6.75415218e-01 -6.92790151e-01 -1.02810490e+00 1.00090456e+00
3.75250012e-01 -1.58057779e-01 3.30722868e-01 -3.71689439e-01
-5.18954933e-01 -1.97242349e-01 -7.62986600e-01 6.98403776e-01
8.20659816e-01 -9.29938257e-01 -7.68192112e-01 5.88113181e-02
6.96281195e-01 -5.25583744e-01 -1.89600781e-01 1.14840984e-01
3.11081678e-01 -9.13617969e-01 5.54202855e-01 5.01913503e-02
5.57276189e-01 -5.73497891e-01 -2.79168487e-01 -8.30762088e-01
-3.18549275e-01 -5.91551840e-01 -6.21151291e-02 1.23098540e+00
3.35980535e-01 -7.15410054e-01 4.66394991e-01 2.03299731e-01
-1.94297835e-01 -6.40821278e-01 -7.70865738e-01 -6.97067916e-01
-1.89493477e-01 -3.71764600e-01 2.70228982e-01 7.78374195e-01
-1.15001753e-01 6.94637537e-01 -5.11788011e-01 2.43375763e-01
4.95937943e-01 2.78470546e-01 5.96175551e-01 -8.55196953e-01
-4.29192066e-01 -5.09463370e-01 -1.91642940e-01 -1.63497281e+00
6.98602870e-02 -8.37476969e-01 1.75135821e-01 -1.78838313e+00
5.36622286e-01 9.41257030e-02 -2.13873938e-01 8.23760629e-01
-2.24711016e-01 1.40478253e-01 2.74424165e-01 4.03977871e-01
-8.03777754e-01 9.06425536e-01 1.14459872e+00 -2.67642111e-01
-6.93258569e-02 -8.37264806e-02 -7.33524024e-01 7.76445568e-01
8.53917122e-01 -2.26749226e-01 -4.74618226e-01 -1.11204052e+00
7.30255106e-03 -1.85584202e-02 5.31580627e-01 -9.55160260e-01
6.28525019e-01 -2.83188730e-01 2.86320806e-01 -7.43452787e-01
3.79621446e-01 -6.27201915e-01 -2.48408601e-01 3.82073045e-01
-4.26249176e-01 5.40880486e-02 1.79734975e-01 4.44510311e-01
-2.87158608e-01 -1.64547443e-01 7.83752441e-01 -6.45792903e-03
-7.91565895e-01 1.98124364e-01 -5.31001508e-01 -1.84303988e-02
6.28228843e-01 -1.69620007e-01 -5.88376641e-01 -3.93875986e-01
-4.89289820e-01 4.68895137e-01 6.70298278e-01 3.98813218e-01
9.66032088e-01 -1.15437973e+00 -6.64606214e-01 3.16311866e-01
1.99020907e-01 -1.48063436e-01 -3.40464455e-03 1.11379182e+00
-5.15561879e-01 4.52563822e-01 1.85492098e-01 -5.87823331e-01
-1.26733351e+00 6.62671745e-01 5.35907090e-01 -1.55468494e-01
-7.80525506e-01 8.00955474e-01 5.88333070e-01 -1.04662225e-01
4.56285566e-01 -3.18867743e-01 -1.53216094e-01 -4.26579326e-01
7.36895502e-01 -1.31555483e-01 -2.38394681e-02 -7.96551526e-01
-4.72807914e-01 7.76136279e-01 -3.26766014e-01 -2.71553844e-01
9.44127977e-01 -5.18272996e-01 -9.55958217e-02 7.63148010e-01
8.64986718e-01 8.39430094e-02 -1.17204678e+00 -7.87263632e-01
3.58383507e-02 -5.15323997e-01 4.17729914e-01 -8.63588333e-01
-1.06044543e+00 1.25928867e+00 3.51156473e-01 -1.14087984e-01
1.46734130e+00 -2.45716199e-01 8.28367233e-01 3.80610615e-01
1.46320283e-01 -8.41590405e-01 3.00876349e-01 7.93337762e-01
8.76185238e-01 -1.31182241e+00 -1.61094859e-01 -3.63201529e-01
-1.03589344e+00 9.45643663e-01 5.76587856e-01 5.88156655e-02
5.26998103e-01 3.01766306e-01 3.24568957e-01 -8.98283124e-02
-1.02986455e+00 -3.68484229e-01 4.25561517e-01 1.43337905e-01
5.32834768e-01 -1.26551047e-01 -8.51061791e-02 6.64020836e-01
1.40004843e-01 3.25644016e-01 3.77593160e-01 8.67296636e-01
-5.43095648e-01 -1.02405918e+00 -2.46792078e-01 3.45337301e-01
-2.85122961e-01 -4.95102495e-01 -6.41322792e-01 1.34054035e-01
-1.55190423e-01 1.11009216e+00 -1.06042042e-01 -6.09611273e-01
1.82640463e-01 2.13944614e-02 3.36982518e-01 -8.00198019e-01
-5.36599994e-01 4.48469937e-01 -6.66281804e-02 -4.37093675e-01
-2.96854436e-01 -5.93428850e-01 -1.25923359e+00 -2.92677194e-01
-2.95695722e-01 -1.71064615e-01 3.90078634e-01 1.04999900e+00
3.73765349e-01 4.36850101e-01 6.81192696e-01 -6.26519322e-01
-2.78069139e-01 -7.43603706e-01 -1.93210840e-01 -1.41009409e-02
6.00469112e-01 -3.45365852e-01 -2.06066117e-01 3.94089967e-01] | [11.827999114990234, 2.1226558685302734] |
ac17a083-e965-4daf-af8c-cba25d3a4eda | multimodal-learning-for-non-small-cell-lung | 2211.0328 | null | https://arxiv.org/abs/2211.03280v1 | https://arxiv.org/pdf/2211.03280v1.pdf | Multimodal Learning for Non-small Cell Lung Cancer Prognosis | This paper focuses on the task of survival time analysis for lung cancer. Although much progress has been made in this problem in recent years, the performance of existing methods is still far from satisfactory. Traditional and some deep learning-based survival time analyses for lung cancer are mostly based on textual clinical information such as staging, age, histology, etc. Unlike existing methods that predicting on the single modality, we observe that a human clinician usually takes multimodal data such as text clinical data and visual scans to estimate survival time. Motivated by this, in this work, we contribute a smart cross-modality network for survival analysis network named Lite-ProSENet that simulates a human's manner of decision making. Extensive experiments were conducted using data from 422 NSCLC patients from The Cancer Imaging Archive (TCIA). The results show that our Lite-ProSENet outperforms favorably again all comparison methods and achieves the new state of the art with the 89.3% on concordance. The code will be made publicly available. | ['Steven Weidong Su', 'Sai Ho Ling', 'Fan Yang', 'Xiaoshui Huang', 'Yaxiong Wang', 'Yujiao Wu'] | 2022-11-07 | null | null | null | null | ['survival-analysis'] | ['miscellaneous'] | [-1.18014418e-01 -2.59895593e-01 -7.82051325e-01 -3.89747739e-01
-1.14755857e+00 -3.49395812e-01 4.44799125e-01 3.57127279e-01
-5.87639630e-01 8.81780088e-01 4.18770611e-01 -9.35072601e-01
-9.25075635e-02 -8.11932504e-01 1.38866939e-02 -9.88355041e-01
-8.34028646e-02 7.84901738e-01 2.75232404e-01 1.49357975e-01
-3.53509516e-01 6.01495981e-01 -7.38866568e-01 3.81372243e-01
4.30814296e-01 1.00798941e+00 -1.17966123e-01 8.64578366e-01
-1.04396619e-01 1.02321362e+00 -3.29825236e-03 -2.07512274e-01
-1.72062472e-01 -3.18988770e-01 -1.02669847e+00 -4.93699908e-01
2.10749149e-01 -4.73505765e-01 -7.04031587e-01 4.19990093e-01
8.81834328e-01 -3.83994490e-01 9.14571464e-01 -1.24004698e+00
-1.68715313e-01 4.79009092e-01 -4.31603849e-01 1.27199680e-01
-9.24625099e-02 1.93897903e-01 8.78794134e-01 -7.63262272e-01
6.21035397e-01 6.88179731e-01 1.22427118e+00 6.43389285e-01
-6.89570129e-01 -3.90114605e-01 -4.45829153e-01 2.81133592e-01
-1.29523301e+00 -3.06192134e-02 2.89982229e-01 -4.64375794e-01
8.13199222e-01 4.39674079e-01 7.44262099e-01 1.27352023e+00
6.90429270e-01 7.30535150e-01 1.06721008e+00 -3.28446925e-01
6.39299005e-02 2.08474100e-01 1.39755592e-01 1.08534038e+00
8.02725852e-02 3.39521825e-01 -2.09838018e-01 -3.55011135e-01
5.88439882e-01 4.21118528e-01 -2.26756170e-01 -3.09390157e-01
-1.54202163e+00 9.11176085e-01 7.58821070e-01 4.94041502e-01
-2.00701609e-01 2.77709931e-01 8.03081036e-01 1.75386459e-01
4.72594857e-01 -3.49511683e-01 -4.02554661e-01 1.41161054e-01
-1.23065436e+00 -8.71651620e-02 8.64656329e-01 3.46134007e-01
-1.86913580e-01 -3.46479535e-01 -5.96885681e-01 4.81689513e-01
3.35404932e-01 2.14399353e-01 8.04031014e-01 -4.63327259e-01
1.16725475e-01 6.18956327e-01 -2.19903916e-01 -4.52565253e-01
-1.02320182e+00 -5.11832893e-01 -1.52390492e+00 2.56707996e-01
6.31182611e-01 -1.23501867e-01 -1.18592811e+00 1.42761517e+00
1.53876945e-01 -1.73535943e-01 1.35304704e-01 8.35423172e-01
1.43716896e+00 3.10106456e-01 4.50195789e-01 -2.71707743e-01
1.50023520e+00 -9.14792836e-01 -6.67947888e-01 1.11623637e-01
1.25205648e+00 -3.11016649e-01 7.39023566e-01 9.24805701e-02
-8.59658957e-01 5.80470404e-03 -5.19026458e-01 -1.20753184e-01
-4.00645196e-01 4.03576165e-01 6.96093500e-01 4.92423534e-01
-1.31216168e+00 5.21672189e-01 -1.08982396e+00 -9.29356337e-01
6.43623292e-01 4.27674592e-01 -7.03284681e-01 -2.18200553e-02
-1.17442465e+00 1.00769866e+00 3.35727453e-01 -1.42416283e-01
-1.22157669e+00 -7.96000361e-01 -2.88569033e-01 -6.15530349e-02
3.71559918e-01 -1.44512856e+00 1.62745273e+00 -6.04934990e-01
-1.03712368e+00 1.09569490e+00 -1.54923692e-01 -4.32716101e-01
1.05678344e+00 3.28946739e-01 -1.81284457e-01 5.79233579e-02
-1.73296109e-01 5.99332571e-01 1.62740350e-01 -8.71312380e-01
-7.08242774e-01 -4.72113997e-01 -5.20653009e-01 1.05131604e-01
-5.76537967e-01 1.17518619e-01 -5.76944411e-01 -6.06763422e-01
-2.55220711e-01 -1.13943863e+00 -4.50193524e-01 7.33898342e-01
-6.61256909e-01 -5.12468815e-01 8.47691238e-01 -8.18339288e-01
1.32017541e+00 -1.77230668e+00 5.55346999e-03 9.46605951e-02
6.46838665e-01 -9.40401703e-02 2.51926899e-01 5.17263710e-01
-3.11814338e-01 2.67840177e-01 -3.56529742e-01 -6.29815459e-01
-3.09783369e-01 1.30614758e-01 3.40498328e-01 7.91075826e-01
-5.00601292e-01 1.20391548e+00 -6.12064004e-01 -1.27526379e+00
4.85104769e-02 5.87017775e-01 2.34610285e-03 2.26348624e-01
7.52951950e-02 4.29114550e-01 -3.95805359e-01 9.29302216e-01
1.91648647e-01 -7.62276113e-01 2.48426467e-01 -1.55489817e-01
-6.25707582e-02 -8.94579068e-02 -2.16334268e-01 1.52378929e+00
-2.34788984e-01 6.59351051e-01 -1.90872043e-01 -7.25936353e-01
2.63410568e-01 9.28084016e-01 8.57207835e-01 -2.32251048e-01
4.73862886e-01 1.74509227e-01 2.03333385e-02 -5.33221900e-01
-2.26515129e-01 -4.76601452e-01 2.31258094e-01 4.04132962e-01
-3.30895811e-01 -4.60641086e-02 -1.28339484e-01 2.58847445e-01
1.55737686e+00 -4.43404943e-01 7.68645942e-01 2.50242110e-02
4.62534696e-01 2.31658459e-01 2.49345988e-01 5.38982570e-01
-5.01647413e-01 7.25509524e-01 5.18741369e-01 -5.37718773e-01
-8.74669313e-01 -1.06748283e+00 -1.94679484e-01 8.05051088e-01
-4.81796414e-01 -2.04741240e-01 -3.88021111e-01 -1.19544733e+00
-2.94033196e-02 4.24079955e-01 -9.98418152e-01 1.26487553e-01
-2.94652432e-01 -9.68685567e-01 7.73193777e-01 9.95868802e-01
4.15619642e-01 -9.28055644e-01 -4.06073332e-01 1.05878055e-01
-1.74967915e-01 -7.07411706e-01 -2.51606911e-01 3.64752114e-01
-1.31422150e+00 -1.12984383e+00 -1.17609978e+00 -7.57726312e-01
4.64523822e-01 -5.49307577e-02 1.31999719e+00 3.54554266e-01
-5.15114605e-01 2.11001307e-01 -9.85027775e-02 -3.89818698e-01
-5.03720224e-01 3.42397988e-01 -3.99942636e-01 -4.33582336e-01
2.28775173e-01 -3.22678506e-01 -9.60245728e-01 1.73459515e-01
-7.55408168e-01 3.73952329e-01 8.77332985e-01 9.12300646e-01
7.84912229e-01 9.52307880e-02 3.63229811e-01 -1.04551446e+00
3.64555389e-01 -7.79418528e-01 -2.24861056e-01 2.06588924e-01
-8.51886153e-01 -2.77603358e-01 6.26448572e-01 -6.25850782e-02
-8.76120746e-01 3.49958241e-01 -1.43109053e-01 -1.34902269e-01
-3.64742666e-01 8.49790692e-01 4.27898884e-01 -7.46906623e-02
6.25333011e-01 6.01338521e-02 5.40752560e-02 -2.97244042e-01
-1.67825237e-01 6.13293767e-01 4.77509618e-01 8.01065788e-02
7.07437813e-01 7.61340857e-01 4.98182684e-01 -2.40961343e-01
-9.51964319e-01 -5.23833990e-01 -5.51368475e-01 -3.61960024e-01
9.59785581e-01 -8.09938908e-01 -9.06341434e-01 6.46941900e-01
-7.08726227e-01 -4.20327634e-01 2.20647026e-02 4.00630683e-01
-4.20231581e-01 2.69471169e-01 -8.55451941e-01 -3.53740036e-01
-7.53802717e-01 -8.99865746e-01 1.01180720e+00 9.16852057e-02
-1.57361731e-01 -1.59830606e+00 4.29813206e-01 2.61393100e-01
6.94812179e-01 5.79422534e-01 1.11168396e+00 -8.30443084e-01
-2.27640510e-01 -5.78564405e-01 -3.91213179e-01 -1.54502362e-01
8.02991912e-02 1.85718104e-01 -8.67505312e-01 -5.07742584e-01
-4.13368970e-01 -5.15917480e-01 9.65532243e-01 6.55593038e-01
1.70009530e+00 8.83965939e-02 -1.15938377e+00 5.39849877e-01
1.57576036e+00 1.18768038e-02 4.65977281e-01 2.99161315e-01
7.95917869e-01 5.45671940e-01 2.34630972e-01 4.17067826e-01
5.67954898e-01 3.39237183e-01 7.52403677e-01 -4.91642565e-01
-2.36193359e-01 3.22125703e-02 -8.28939304e-02 3.88515562e-01
-2.28258863e-01 -8.47334862e-01 -1.53282416e+00 6.15197897e-01
-1.77992475e+00 -7.32739389e-01 -5.59935331e-01 1.86783206e+00
8.47374737e-01 -5.59031516e-02 4.73978743e-02 2.09623083e-01
4.53255862e-01 -3.28853019e-02 -4.90930736e-01 9.68877450e-02
1.61320418e-01 -1.33014321e-01 5.92199326e-01 1.68432835e-02
-1.18286216e+00 4.18714762e-01 6.81295061e+00 8.59817088e-01
-1.23293757e+00 3.41960996e-01 1.08291495e+00 -1.28273204e-01
-7.65745249e-03 -2.60115802e-01 -4.68263030e-01 1.48319364e-01
1.21904850e+00 -1.30029261e-01 -1.06264941e-01 6.65894151e-01
3.03813189e-01 -2.51652539e-01 -1.35333896e+00 9.20897543e-01
-2.75871426e-01 -1.30625033e+00 -3.55166703e-01 2.74989128e-01
3.86921138e-01 3.46497536e-01 1.69275105e-01 3.10162574e-01
4.30337191e-01 -1.52064002e+00 1.57246843e-01 8.21513653e-01
1.21236956e+00 -5.76955438e-01 1.22605789e+00 4.95837510e-01
-1.16901624e+00 3.56158577e-02 2.37188593e-01 5.60913861e-01
4.49659862e-02 6.33428633e-01 -1.29175901e+00 7.50616372e-01
7.53681362e-01 8.28906536e-01 -8.42552304e-01 1.13113534e+00
1.55066103e-01 9.96432126e-01 -2.61637181e-01 -6.55438527e-02
1.31840155e-01 5.75072944e-01 7.70053044e-02 1.20104599e+00
5.94775617e-01 1.21176094e-01 -1.52588887e-02 3.12550843e-01
-1.31485939e-01 7.47839287e-02 -5.87173879e-01 -2.07564503e-01
9.73573998e-02 1.54441512e+00 -7.83798814e-01 -4.07118708e-01
-6.99121416e-01 8.18502128e-01 1.77416936e-01 -1.74198728e-02
-9.73384976e-01 6.08295761e-02 -1.53838739e-01 3.87463748e-01
-1.40511096e-01 1.33497939e-01 -3.51690948e-01 -7.06501901e-01
-5.17373502e-01 -6.15770042e-01 1.08128881e+00 -9.04775202e-01
-1.62958622e+00 6.27441704e-01 -9.53386351e-02 -1.32345760e+00
-7.09349811e-02 -8.12606573e-01 -8.15840364e-01 7.37719297e-01
-1.42764568e+00 -1.36071754e+00 -6.51158094e-01 6.72816575e-01
3.29841316e-01 -2.97462285e-01 1.14930224e+00 1.38366222e-01
-6.19503260e-01 6.99865580e-01 2.64887094e-01 2.36261427e-01
1.02172685e+00 -1.42587984e+00 -4.42361757e-02 1.09758854e-01
-5.34626901e-01 -8.98880810e-02 3.57487857e-01 -6.69946790e-01
-1.22239733e+00 -1.17776275e+00 1.17217851e+00 -6.57402813e-01
6.98104858e-01 2.77308553e-01 -8.20155382e-01 5.77400565e-01
4.20860410e-01 3.99392933e-01 9.79687631e-01 -2.06333771e-01
5.27081750e-02 -2.19614953e-02 -1.15030408e+00 4.87314373e-01
8.12223434e-01 -2.64120519e-01 -6.44400492e-02 6.57476485e-01
2.91909784e-01 -3.82535845e-01 -1.09918582e+00 7.23111033e-01
5.15645087e-01 -9.17891622e-01 9.60123062e-01 -6.23311639e-01
4.31137711e-01 -1.92049220e-01 8.30609053e-02 -1.08475411e+00
-5.29738009e-01 1.92972228e-01 -1.19352806e-03 7.14441836e-01
4.84397322e-01 -4.24605906e-01 1.09291565e+00 4.67977613e-01
-1.56506255e-01 -1.22034645e+00 -1.04564166e+00 -5.02162337e-01
4.29209650e-01 -1.68622479e-01 8.10608491e-02 8.97513032e-01
-3.44878793e-01 1.82371736e-02 -1.41621783e-01 5.67379817e-02
5.70515633e-01 -1.03526734e-01 2.58858353e-01 -1.49211848e+00
-7.54645392e-02 -4.29719716e-01 -1.35667682e-01 -1.31051660e-01
1.09725565e-01 -1.10331845e+00 -2.01232731e-01 -2.01763034e+00
1.03090119e+00 -2.81523734e-01 -6.67561829e-01 9.46267784e-01
7.24839270e-02 2.51801938e-01 -3.56886894e-01 3.40338349e-01
-5.71230650e-01 3.13452005e-01 1.26471436e+00 -1.45445272e-01
3.03948879e-01 1.94118470e-02 -4.06240463e-01 7.62461722e-01
9.06371772e-01 -6.69647336e-01 -9.25913155e-02 -9.07695517e-02
1.19698912e-01 8.47496688e-01 7.34181762e-01 -1.07837367e+00
4.26032215e-01 -3.13372284e-01 6.64509714e-01 -1.04481351e+00
1.37164474e-01 -1.06393528e+00 3.90720695e-01 9.07196462e-01
-2.74846464e-01 1.58164099e-01 9.90028232e-02 6.83462799e-01
-3.00584823e-01 7.58128539e-02 8.34952474e-01 -1.43278137e-01
-3.51978034e-01 8.51102889e-01 -4.83789057e-01 -2.00068176e-01
1.15002060e+00 -8.39865580e-02 -5.67882836e-01 -4.19768393e-01
-8.46277297e-01 2.70886540e-01 4.04833794e-01 -1.55951396e-01
4.54525083e-01 -1.50370598e+00 -9.47832942e-01 -4.07933056e-01
1.99398696e-01 -1.11682668e-01 4.24941450e-01 1.40671182e+00
-7.12479293e-01 6.14726305e-01 8.93834047e-03 -5.71291387e-01
-1.65898931e+00 4.83144462e-01 9.16001499e-01 -8.44017386e-01
-6.16774499e-01 9.41668749e-01 2.76763916e-01 -2.94817001e-01
4.64888632e-01 -1.24648921e-01 -3.67336214e-01 2.75736786e-02
1.58945039e-01 2.79101968e-01 2.02627152e-01 -3.20387453e-01
-6.19641960e-01 3.45437020e-01 -1.81752846e-01 -1.03325114e-01
1.21246660e+00 9.90666300e-02 -1.82271078e-01 5.28059781e-01
1.31069624e+00 -4.05527562e-01 -6.46006942e-01 -2.44671971e-01
-9.63085443e-02 -7.68466815e-02 4.46834952e-01 -1.37283063e+00
-1.24016190e+00 8.59361887e-01 1.08543909e+00 1.22310601e-01
1.34104037e+00 1.84278548e-01 8.37174714e-01 3.91650081e-01
-3.86998914e-02 -5.29463470e-01 4.53637503e-02 4.18682188e-01
7.41139472e-01 -1.69900525e+00 -2.54038274e-02 -1.55656129e-01
-5.98424613e-01 1.31350482e+00 5.35449564e-01 3.49324286e-01
8.96371245e-01 3.15140724e-01 3.46388280e-01 -3.12338620e-01
-1.14981794e+00 4.62634154e-02 3.27584922e-01 2.38290295e-01
1.06513417e+00 2.70568967e-01 -2.02685043e-01 5.54544985e-01
6.73854537e-03 3.90875846e-01 2.12631151e-01 9.50615764e-01
-2.69847542e-01 -1.07905221e+00 -3.47552687e-01 6.80096567e-01
-8.17951024e-01 -1.59197479e-01 -5.98441243e-01 1.15911901e+00
-2.02881485e-01 5.17387450e-01 -1.87148213e-01 -1.50752097e-01
3.14207524e-02 1.28829658e-01 1.70197412e-01 -3.84424597e-01
-6.45426214e-01 -2.37283513e-01 2.19825998e-01 -3.05222392e-01
-2.75521994e-01 -6.17518485e-01 -1.46429384e+00 -5.73448062e-01
-1.40485829e-02 -1.38267353e-01 5.88569462e-01 6.58887446e-01
-2.11237878e-01 8.62341166e-01 3.81609172e-01 -3.95753026e-01
-4.03928667e-01 -8.32414865e-01 -4.42906231e-01 -4.65154052e-02
4.74070370e-01 -4.35639381e-01 -4.78579491e-01 1.08114704e-02] | [15.28002643585205, -2.4676408767700195] |
76503bfa-1bc8-405f-98c3-188411214d50 | a-deep-relevance-matching-model-for-ad-hoc | 1711.08611 | null | http://arxiv.org/abs/1711.08611v1 | http://arxiv.org/pdf/1711.08611v1.pdf | A Deep Relevance Matching Model for Ad-hoc Retrieval | In recent years, deep neural networks have led to exciting breakthroughs in
speech recognition, computer vision, and natural language processing (NLP)
tasks. However, there have been few positive results of deep models on ad-hoc
retrieval tasks. This is partially due to the fact that many important
characteristics of the ad-hoc retrieval task have not been well addressed in
deep models yet. Typically, the ad-hoc retrieval task is formalized as a
matching problem between two pieces of text in existing work using deep models,
and treated equivalent to many NLP tasks such as paraphrase identification,
question answering and automatic conversation. However, we argue that the
ad-hoc retrieval task is mainly about relevance matching while most NLP
matching tasks concern semantic matching, and there are some fundamental
differences between these two matching tasks. Successful relevance matching
requires proper handling of the exact matching signals, query term importance,
and diverse matching requirements. In this paper, we propose a novel deep
relevance matching model (DRMM) for ad-hoc retrieval. Specifically, our model
employs a joint deep architecture at the query term level for relevance
matching. By using matching histogram mapping, a feed forward matching network,
and a term gating network, we can effectively deal with the three relevance
matching factors mentioned above. Experimental results on two representative
benchmark collections show that our model can significantly outperform some
well-known retrieval models as well as state-of-the-art deep matching models. | ['W. Bruce Croft', 'Yixing Fan', 'Jiafeng Guo', 'Qingyao Ai'] | 2017-11-23 | null | null | null | null | ['ad-hoc-information-retrieval'] | ['natural-language-processing'] | [ 2.21606672e-01 -3.45055014e-01 -2.39945009e-01 -4.16263312e-01
-1.32547045e+00 -3.69544655e-01 9.37884450e-01 4.60516483e-01
-5.19366622e-01 1.15442641e-01 4.03258175e-01 -8.74181166e-02
-5.80771923e-01 -8.55288208e-01 -4.79895890e-01 -2.41656378e-01
3.36843967e-01 9.15495157e-01 2.17377052e-01 -6.71694100e-01
4.39846933e-01 2.34505072e-01 -1.59087694e+00 3.75995129e-01
6.91320240e-01 1.35781538e+00 1.75880477e-01 3.19003999e-01
-4.24322963e-01 3.27852160e-01 -7.71178663e-01 -4.34474319e-01
1.45294204e-01 -2.10411102e-01 -1.27881920e+00 -4.49040085e-01
6.46685660e-01 -4.75247502e-01 -7.12174356e-01 1.03316700e+00
7.89236605e-01 4.18613911e-01 6.37392461e-01 -1.07776511e+00
-1.23512733e+00 4.35271174e-01 -2.39301801e-01 3.92251521e-01
7.33897924e-01 -1.93089679e-01 1.62768805e+00 -1.11381602e+00
2.98649997e-01 1.43281567e+00 4.53993738e-01 3.82666886e-01
-7.64284968e-01 -5.76586127e-01 1.32644102e-02 7.16386437e-01
-1.34870172e+00 -2.71852851e-01 7.21227288e-01 -1.77556351e-02
1.14529371e+00 4.07701224e-01 5.15816450e-01 1.01320434e+00
2.23021299e-01 1.07507277e+00 4.93383825e-01 -5.59193254e-01
4.29398641e-02 -1.59052432e-01 5.24147451e-01 2.47506693e-01
-1.15193509e-01 3.86635289e-02 -3.99692893e-01 -3.63194197e-01
2.46680290e-01 3.10214788e-01 -3.49846661e-01 -1.85695395e-01
-1.08186710e+00 9.98620570e-01 6.44224644e-01 6.12114251e-01
-2.92970031e-01 9.10184383e-02 5.56102097e-01 7.14089990e-01
2.99115300e-01 6.76770806e-01 -2.02530488e-01 3.61312687e-01
-9.66715753e-01 7.25438952e-01 7.35269010e-01 1.12851679e+00
7.47023761e-01 -5.94401717e-01 -7.04000294e-01 1.14268863e+00
5.51932991e-01 5.22154391e-01 9.41138148e-01 -6.49804592e-01
4.98128921e-01 7.02969372e-01 -2.10724667e-01 -1.10853004e+00
-3.01756680e-01 -3.64424229e-01 -7.54320204e-01 -5.97330034e-01
4.10817079e-02 5.60039878e-01 -8.75811219e-01 1.71434247e+00
-1.07803605e-01 -2.46805608e-01 -5.67078441e-02 1.15188825e+00
1.36222827e+00 7.15937912e-01 1.87460519e-02 1.05603516e-01
1.60457742e+00 -1.10273337e+00 -8.35427046e-01 -5.26387274e-01
3.37766379e-01 -1.03009748e+00 1.09424949e+00 -2.14004830e-01
-1.26406622e+00 -6.46504045e-01 -9.21545982e-01 -6.59467459e-01
-6.47222102e-01 -3.42165291e-01 7.32323050e-01 1.17301054e-01
-1.23472869e+00 3.30339640e-01 1.88024584e-02 -5.94081879e-01
5.70649803e-02 2.91685432e-01 -1.38019383e-01 -4.63365495e-01
-2.04098606e+00 9.82660234e-01 6.92692325e-02 9.10205394e-02
-6.62196517e-01 -4.99983311e-01 -8.92030060e-01 6.33753836e-01
2.62731075e-01 -1.16334760e+00 1.56884897e+00 -7.10656166e-01
-1.01964819e+00 1.17875028e+00 -2.28988856e-01 -4.36378330e-01
-3.29046957e-02 -2.32261628e-01 -5.38177192e-01 1.74880221e-01
8.05143863e-02 7.44606018e-01 7.50447452e-01 -7.19813287e-01
-3.38751167e-01 -4.57426071e-01 3.09825689e-01 4.35084552e-01
-3.28592420e-01 3.44837427e-01 -8.97067666e-01 -6.25887752e-01
9.24257040e-02 -6.87493026e-01 1.02643475e-01 5.99110611e-02
2.64526587e-02 -9.27121699e-01 5.31041503e-01 -5.27224600e-01
1.32128763e+00 -2.06155729e+00 8.04543495e-02 2.06384435e-02
9.89931449e-02 4.68725175e-01 -6.78098142e-01 8.58430564e-01
2.33509956e-04 9.81800556e-02 -1.70912132e-01 -3.61237466e-01
4.76129442e-01 2.34122779e-02 -7.56714165e-01 5.86049445e-02
2.64397021e-02 1.47624278e+00 -9.45230544e-01 -4.74446923e-01
-2.88050137e-02 3.77313107e-01 -2.63388664e-01 4.98851091e-01
-1.95800602e-01 -3.67436409e-01 -4.56305474e-01 7.45848596e-01
5.59657037e-01 -4.17647511e-01 -1.01475425e-01 -2.57115453e-01
4.50614691e-01 7.25909889e-01 -4.79643941e-01 2.10053730e+00
-5.35353005e-01 7.99137235e-01 2.70229131e-02 -1.18195689e+00
9.00464416e-01 3.40434402e-01 3.62663388e-01 -1.37120199e+00
1.02685407e-01 2.55660534e-01 -2.41898343e-01 -3.90033364e-01
1.01240456e+00 -1.72951490e-01 -2.24021643e-01 6.46318316e-01
4.37932089e-02 -3.86438131e-01 1.05970040e-01 3.63488972e-01
1.11625910e+00 -4.89733011e-01 6.50630966e-02 -1.38102043e-02
6.36876285e-01 -5.60008921e-02 1.09910376e-01 1.06765795e+00
-3.75517815e-01 8.45812917e-01 -5.37668094e-02 -2.71590501e-01
-5.88074625e-01 -9.46006894e-01 -1.08437628e-01 1.33942449e+00
6.27177596e-01 -3.71009499e-01 -4.62271512e-01 -4.20924157e-01
8.76671448e-02 1.55117854e-01 -5.05031049e-01 -7.36310482e-01
-6.54821992e-01 -5.43392122e-01 4.52271551e-01 4.21599627e-01
4.59296048e-01 -1.44233704e+00 2.76390295e-02 2.28198573e-01
-4.45698768e-01 -1.04885113e+00 -7.69578695e-01 -1.12947254e-02
-5.42973280e-01 -8.84345233e-01 -1.04905224e+00 -1.21771264e+00
1.57007560e-01 8.58742774e-01 1.59814548e+00 4.75507170e-01
-3.88023287e-01 5.89906216e-01 -5.38677573e-01 -1.99332774e-01
-1.46657631e-01 2.91582942e-01 -2.61086911e-01 -2.95096099e-01
9.41214919e-01 -3.13239604e-01 -9.71699774e-01 3.39659899e-01
-1.32711482e+00 -5.42273045e-01 7.68763661e-01 1.04460621e+00
5.15778303e-01 -3.81202251e-01 1.00090837e+00 -3.12425196e-01
1.34085345e+00 -5.40705681e-01 -2.65305370e-01 8.14735591e-01
-7.30600238e-01 1.72672838e-01 3.44043493e-01 -3.05833519e-01
-6.58474505e-01 -6.72724843e-01 -3.67601246e-01 -4.71834928e-01
-2.04822291e-02 7.64268517e-01 -9.91702378e-02 1.48557788e-02
4.08547193e-01 3.73527139e-01 1.48393195e-02 -4.68473554e-01
3.43174726e-01 9.26000178e-01 3.32976699e-01 -5.84977746e-01
6.20918751e-01 2.52492040e-01 -3.42712492e-01 -4.88178074e-01
-1.05342114e+00 -1.05530703e+00 -1.91904202e-01 2.51809567e-01
6.11427486e-01 -8.05651903e-01 -6.50413215e-01 2.33556509e-01
-1.40503705e+00 1.89756900e-01 -3.35343897e-01 1.54564038e-01
-3.44212413e-01 7.61131942e-01 -5.65189123e-01 -3.45352978e-01
-1.04227364e+00 -1.13710856e+00 1.52065885e+00 1.42823488e-01
-2.64167219e-01 -1.01806855e+00 2.85139889e-01 6.26750052e-01
8.69116187e-01 -6.90467596e-01 1.16503775e+00 -8.83403838e-01
-6.69348717e-01 -6.18059695e-01 -6.06329024e-01 1.60591871e-01
-9.84616429e-02 -6.24314666e-01 -1.04859531e+00 -2.39491388e-01
-1.97891388e-02 -6.39588296e-01 1.37819767e+00 1.83143958e-01
1.22828794e+00 -4.57027480e-02 -2.56128550e-01 3.26010734e-01
1.01753914e+00 -4.10006847e-03 6.75679564e-01 3.58384907e-01
2.84253240e-01 8.08024585e-01 5.73068559e-01 9.06655043e-02
6.72118843e-01 1.04025602e+00 4.55658078e-01 4.20189686e-02
-2.24179864e-01 -1.08176962e-01 2.92017516e-02 9.45361257e-01
7.43550658e-01 -5.60458124e-01 -7.82261729e-01 7.64754236e-01
-2.08704662e+00 -1.12321901e+00 3.50679219e-01 2.10951519e+00
9.10238922e-01 -2.80490398e-01 -2.46454254e-01 -1.17848620e-01
6.79083824e-01 3.96832973e-01 -5.56106925e-01 -3.69209677e-01
-1.67240351e-01 3.88648123e-01 -1.77009791e-01 3.54026943e-01
-1.11302924e+00 8.03160250e-01 6.36016083e+00 1.11930621e+00
-8.45593333e-01 -4.15534973e-02 2.01534376e-01 9.83044282e-02
-7.43831933e-01 -3.31734605e-02 -7.15371668e-01 1.60705402e-01
5.63463271e-01 -3.85348111e-01 3.64284098e-01 5.75192928e-01
-4.02216285e-01 2.06370592e-01 -1.39543962e+00 1.25395226e+00
4.76045340e-01 -1.23197198e+00 5.34868121e-01 -1.48202553e-01
2.82158285e-01 9.47707146e-02 1.96819752e-01 9.61452723e-01
-1.37273088e-01 -1.10058320e+00 2.94568449e-01 4.80889350e-01
4.36896443e-01 -4.79950488e-01 8.52560222e-01 2.22347856e-01
-1.17372704e+00 -1.40528828e-02 -6.88603997e-01 1.64528012e-01
9.60882530e-02 5.36608756e-01 -4.73663867e-01 6.94595933e-01
6.28893077e-01 6.48917079e-01 -5.24580896e-01 1.20640862e+00
1.36990449e-04 3.85880210e-02 -1.59522563e-01 -9.69375148e-02
3.90477508e-01 7.56709650e-02 4.01450753e-01 1.21645355e+00
2.02008754e-01 -2.47096628e-01 -7.06441998e-02 9.59064603e-01
-4.39479738e-01 1.74219280e-01 -4.95410919e-01 -9.51827541e-02
5.94256103e-01 1.24244416e+00 -2.98634563e-02 -3.65713000e-01
-4.57364082e-01 8.38948786e-01 4.23171520e-01 4.24610525e-01
-4.73481029e-01 -5.35821736e-01 6.79518104e-01 -7.50148967e-02
-4.74999808e-02 1.51918203e-01 9.84559208e-02 -1.20935333e+00
3.06849986e-01 -9.54743803e-01 5.97537160e-01 -6.72837794e-01
-1.80445564e+00 5.41176677e-01 1.52722178e-02 -1.18393350e+00
-4.63762224e-01 -3.26506257e-01 -6.07644856e-01 1.10207486e+00
-2.00489497e+00 -1.08455491e+00 -2.94391870e-01 5.71395099e-01
7.01806247e-01 -1.27234414e-01 9.47710633e-01 5.49852848e-01
-2.56498009e-01 7.93616056e-01 4.16034972e-03 9.58992168e-02
9.49213922e-01 -9.69393849e-01 6.75935864e-01 4.01631922e-01
2.34386146e-01 9.53362405e-01 1.82014138e-01 -2.00253636e-01
-1.47933638e+00 -7.56108880e-01 1.42105865e+00 -3.12468439e-01
6.97649240e-01 -2.69788176e-01 -1.19136691e+00 2.00167730e-01
5.46041012e-01 -9.07560661e-02 6.81006908e-01 1.78435668e-01
-6.50093555e-01 -1.50482520e-01 -7.64154851e-01 5.06862998e-01
1.00902176e+00 -1.00471401e+00 -1.14624619e+00 5.35507202e-01
1.07803929e+00 -1.70051157e-01 -6.22468531e-01 5.82678556e-01
7.30603695e-01 -6.19429529e-01 1.43756390e+00 -6.41861200e-01
2.96381980e-01 1.71555921e-01 -3.82640600e-01 -1.13084924e+00
-3.01035851e-01 -2.76665509e-01 -1.97010219e-01 9.59974229e-01
2.29240924e-01 -3.68245274e-01 3.66668254e-01 7.36821055e-01
-2.57996231e-01 -8.02519917e-01 -9.11464155e-01 -7.70766854e-01
2.87886828e-01 -2.59250641e-01 9.54582214e-01 8.69071245e-01
5.54959103e-03 7.47180998e-01 -1.00235738e-01 -2.63742059e-01
3.17896158e-01 7.23042250e-01 4.22375768e-01 -1.34969676e+00
-3.23992342e-01 -9.77835774e-01 -1.74016535e-01 -1.58238518e+00
6.55418396e-01 -1.13424742e+00 3.83524597e-02 -1.99044383e+00
4.84391689e-01 -3.75372499e-01 -4.75544065e-01 2.56420285e-01
-4.39134121e-01 4.16107895e-03 6.30442426e-02 4.80687559e-01
-9.05115128e-01 9.26851988e-01 1.18066537e+00 -7.34086812e-01
1.68095574e-01 -8.90708994e-03 -8.29047918e-01 2.30411783e-01
6.00661337e-01 -2.75815010e-01 -4.18179482e-01 -9.08637106e-01
6.38820827e-01 1.48258120e-01 4.63751882e-01 -5.24594247e-01
6.22049809e-01 1.06768303e-01 -5.05427271e-02 -6.63734972e-01
5.31710267e-01 -9.12006497e-01 -5.04051626e-01 1.06923394e-01
-8.03682208e-01 1.87845036e-01 1.14943326e-01 5.76645911e-01
-8.39826405e-01 -5.00508666e-01 3.54881823e-01 -1.09251007e-01
-8.37161124e-01 4.51562464e-01 -1.74708471e-01 3.88572723e-01
3.59032929e-01 9.12171304e-02 -6.63078666e-01 -6.91179335e-01
-3.63415569e-01 5.31801939e-01 1.98480617e-02 9.54008877e-01
9.03993368e-01 -1.39187348e+00 -6.17593229e-01 1.39020368e-01
6.38558686e-01 -1.79780796e-02 2.88106054e-01 4.68713462e-01
2.04108879e-02 9.59335804e-01 2.34883025e-01 -4.10208255e-01
-1.01399755e+00 6.78926766e-01 2.83492953e-01 -7.47253299e-01
-2.56380111e-01 8.47836733e-01 4.99163032e-01 -7.00507224e-01
6.04338527e-01 -1.87806651e-01 -3.56270403e-01 1.67777732e-01
7.58328736e-01 4.49626334e-02 4.82852340e-01 -4.45778877e-01
-5.42425156e-01 7.79226065e-01 -5.91190100e-01 2.86422744e-02
8.44472528e-01 -2.62591392e-01 -1.91836208e-01 -4.50023152e-02
1.63837337e+00 -6.18067920e-01 -1.26224697e-01 -8.84053111e-01
9.97867882e-02 -1.65327951e-01 1.31637916e-01 -8.20762157e-01
-9.33852255e-01 1.20992374e+00 4.47760671e-01 2.56240517e-01
1.17122257e+00 1.97372094e-01 1.34432399e+00 8.47360671e-01
1.75141990e-01 -1.07685912e+00 2.18614116e-01 1.00856626e+00
1.26838458e+00 -1.47697496e+00 -7.08200559e-02 7.67921135e-02
-1.45126984e-01 8.83906484e-01 5.87902367e-01 6.79374188e-02
6.19595826e-01 -3.35461676e-01 4.13241051e-02 -5.51621079e-01
-9.07748640e-01 -5.12558401e-01 9.06057596e-01 2.85744280e-01
4.88757432e-01 -2.77146429e-01 -4.18127090e-01 5.30222297e-01
7.63673931e-02 -1.47963896e-01 -1.97642237e-01 7.91574121e-01
-6.48925662e-01 -1.23354459e+00 -1.36522189e-01 3.10625702e-01
-3.23541909e-01 -5.48696578e-01 -7.51254618e-01 5.50239027e-01
-5.63931346e-01 1.06723130e+00 2.45476797e-01 -3.16009521e-01
3.66893053e-01 1.73873767e-01 3.29031855e-01 -6.90548897e-01
-1.01047814e+00 -2.27024034e-01 -1.74441829e-01 -5.48791051e-01
-3.48598748e-01 2.27317624e-02 -9.04617071e-01 -2.04966620e-01
-5.18875062e-01 2.92578429e-01 5.65076768e-01 1.11090302e+00
5.46920061e-01 3.15030307e-01 5.45058191e-01 -5.42131364e-01
-9.51737463e-01 -1.02080870e+00 -3.69622529e-01 6.67450488e-01
2.53371507e-01 -4.65861559e-01 -4.70569134e-01 -6.39563501e-01] | [11.410110473632812, 7.735055446624756] |
aebd3a2a-1507-4698-be9b-25804db680a2 | clip2tv-an-empirical-study-on-transformer | 2111.0561 | null | https://arxiv.org/abs/2111.05610v2 | https://arxiv.org/pdf/2111.05610v2.pdf | CLIP2TV: Align, Match and Distill for Video-Text Retrieval | Modern video-text retrieval frameworks basically consist of three parts: video encoder, text encoder and the similarity head. With the success on both visual and textual representation learning, transformer based encoders and fusion methods have also been adopted in the field of video-text retrieval. In this report, we present CLIP2TV, aiming at exploring where the critical elements lie in transformer based methods. To achieve this, We first revisit some recent works on multi-modal learning, then introduce some techniques into video-text retrieval, finally evaluate them through extensive experiments in different configurations. Notably, CLIP2TV achieves 52.9@R1 on MSR-VTT dataset, outperforming the previous SOTA result by 4.1%. | ['Lili Zhao', 'Dedan Chang', 'Sheng Chen', 'Weiqi Sun', 'Jingyu Liu', 'Zijian Gao'] | 2021-11-10 | null | null | null | null | ['video-text-retrieval'] | ['computer-vision'] | [ 2.92991728e-01 -8.01440358e-01 -3.13462615e-01 -6.31396621e-02
-1.42905390e+00 -4.32723582e-01 9.89613175e-01 1.54547185e-01
-4.35643494e-01 3.77627134e-01 4.90900815e-01 1.47684380e-01
-4.07197297e-01 -2.46770546e-01 -4.45281416e-01 -5.67629039e-01
6.03904016e-02 4.49999511e-01 3.01595032e-01 -2.60924101e-01
5.08140743e-01 -2.02595785e-01 -1.59981024e+00 9.00704026e-01
3.81238103e-01 1.25877237e+00 7.33602494e-02 7.32709408e-01
-2.66615838e-01 1.21556211e+00 -5.23301721e-01 -6.12311244e-01
1.83712672e-02 -2.08594739e-01 -6.79368913e-01 -3.97693515e-02
7.07833827e-01 -5.56376398e-01 -9.04987514e-01 7.07527876e-01
7.03458011e-01 2.21164837e-01 9.46270823e-01 -1.02838361e+00
-7.33259976e-01 7.88682401e-01 -9.41789567e-01 3.11052978e-01
8.41311991e-01 -5.24306834e-01 1.40708601e+00 -1.37771702e+00
6.31508589e-01 1.47072744e+00 3.84579629e-01 1.30487457e-01
-5.59544802e-01 -4.34865862e-01 -1.38747141e-01 6.96603835e-01
-1.65792572e+00 -5.82796335e-01 5.41141272e-01 -3.13463271e-01
1.02440834e+00 3.31736267e-01 4.33816135e-01 1.12774086e+00
3.07411194e-01 1.48911524e+00 8.75578880e-01 -5.59390426e-01
-3.69759411e-01 2.34223623e-02 3.27634066e-02 5.21884203e-01
-2.76929862e-03 -2.38260031e-01 -9.11589324e-01 -6.16538199e-03
3.87933612e-01 1.52397692e-01 -2.36012027e-01 -3.22968930e-01
-1.19451702e+00 7.26694047e-01 1.48822561e-01 6.32215261e-01
-1.65896952e-01 3.04369003e-01 8.74401271e-01 4.52262372e-01
3.87248039e-01 -1.53584227e-01 -2.02916171e-02 -2.52264738e-01
-1.39017689e+00 6.12016097e-02 3.74738485e-01 1.06618154e+00
4.03979778e-01 -1.33479480e-02 -5.52738011e-01 1.17052042e+00
4.98430341e-01 6.93137527e-01 8.05049896e-01 -6.75357819e-01
7.13776588e-01 2.91086674e-01 -3.48019987e-01 -1.11338353e+00
1.54594585e-01 -3.06655616e-02 -6.36296153e-01 -6.32388294e-01
-2.64214516e-01 1.70990020e-01 -9.43090916e-01 8.08507740e-01
-2.70795822e-01 2.06637438e-02 9.13087502e-02 6.59087658e-01
1.20201719e+00 9.31782901e-01 -5.02676368e-02 -3.00669581e-01
1.35096335e+00 -1.10931885e+00 -1.02727389e+00 1.73225701e-02
6.14618361e-01 -1.25217831e+00 5.44222116e-01 2.81167448e-01
-1.28466225e+00 -4.04429704e-01 -8.91125977e-01 -2.33627573e-01
-3.84662926e-01 4.39744443e-01 1.51791111e-01 3.70027632e-01
-1.25713313e+00 2.81664878e-01 -5.05780399e-01 -5.05365133e-01
3.57526600e-01 1.92405984e-01 -3.60512614e-01 -4.12534505e-01
-1.27117419e+00 7.99959004e-01 4.47027206e-01 -1.48335785e-01
-9.03912544e-01 -4.06160891e-01 -6.54741168e-01 6.22445196e-02
4.51992750e-01 -6.54535890e-01 1.11096120e+00 -7.33668864e-01
-1.17948294e+00 8.54634941e-01 -3.00628155e-01 -4.90840852e-01
4.82552350e-01 -4.63914931e-01 -5.71039081e-01 7.26216972e-01
1.23505108e-02 6.02425277e-01 1.29156148e+00 -1.12776530e+00
-7.41912842e-01 -2.16439500e-01 -4.92309183e-02 4.05009717e-01
-6.26702487e-01 3.80621612e-01 -1.04122639e+00 -8.15788090e-01
-1.63637698e-01 -7.70205498e-01 3.02314401e-01 -1.65372923e-01
-1.96035787e-01 -4.75963801e-01 1.07547247e+00 -6.12865865e-01
1.51705885e+00 -2.17816544e+00 2.92409927e-01 1.86356921e-02
3.52595985e-01 4.25188124e-01 -1.48337886e-01 1.00502264e+00
-9.41435546e-02 3.19590978e-02 2.85928428e-01 -5.33457458e-01
7.46186748e-02 -1.06813826e-01 -5.58526635e-01 4.91518229e-01
-1.45854622e-01 1.08469272e+00 -5.16700208e-01 -1.19152319e+00
6.03302717e-01 6.97835207e-01 -2.67190069e-01 -1.35237545e-01
8.75111744e-02 -2.71768659e-01 -6.33485556e-01 9.43933427e-01
3.02451819e-01 -3.54164869e-01 5.54732122e-02 -4.49044883e-01
-9.45603251e-02 -2.91277692e-02 -9.60256636e-01 1.78364646e+00
-1.27249330e-01 1.09247315e+00 -1.42720729e-01 -1.13255298e+00
5.50199330e-01 6.05678558e-01 9.20979977e-01 -9.90935326e-01
3.33949476e-01 1.42291397e-01 -6.25468493e-01 -5.36628425e-01
1.10170484e+00 -7.70011311e-03 4.18287404e-02 2.78921008e-01
2.99144953e-01 1.82646900e-01 4.30610120e-01 6.57572746e-01
9.61353421e-01 9.18781832e-02 1.92094848e-01 2.08630979e-01
7.91964352e-01 -5.20055443e-02 -1.69620454e-01 6.77415848e-01
-1.73027828e-01 7.38925755e-01 1.13559350e-01 -6.79888204e-02
-1.03880560e+00 -8.09925675e-01 -1.94478169e-01 1.26274562e+00
2.75743425e-01 -1.01599920e+00 -2.66892761e-01 -4.36899245e-01
-1.01756819e-01 2.90154368e-01 -4.60943729e-01 -1.59406126e-01
-5.94008148e-01 -4.37893540e-01 6.64776146e-01 4.90177572e-01
3.82371336e-01 -6.41372085e-01 -1.95548683e-01 -1.07099958e-01
-5.21528065e-01 -1.31894886e+00 -6.39340580e-01 -2.53041148e-01
-6.06612265e-01 -8.45036387e-01 -1.15151596e+00 -8.22240591e-01
9.57046673e-02 9.82774913e-01 9.01224613e-01 1.55449301e-01
-3.56272608e-01 1.02263606e+00 -9.72623050e-01 -9.11746323e-02
-8.86908695e-02 -2.51521114e-02 -1.27733976e-01 1.03148423e-01
3.30573976e-01 -2.07543388e-01 -5.36795974e-01 2.53067523e-01
-1.17074859e+00 -1.30533472e-01 6.03264034e-01 7.16705620e-01
5.67358017e-01 -3.57588828e-02 3.32175754e-02 -3.74434501e-01
5.14892280e-01 -5.03497481e-01 -2.35744789e-01 7.31232703e-01
-5.46963811e-01 -1.50511205e-01 1.76371187e-01 -2.56193310e-01
-8.49891722e-01 -2.22030014e-01 1.02599621e-01 -1.04203141e+00
4.34019148e-01 8.30213964e-01 4.33608741e-01 -1.09042376e-01
3.43192607e-01 6.51596308e-01 -3.06775749e-01 -3.04111481e-01
3.00671786e-01 8.84425282e-01 6.45880923e-02 -3.96516442e-01
8.43625665e-01 4.77555931e-01 -1.71425954e-01 -1.01770902e+00
-7.50346541e-01 -8.47913623e-01 -4.92100269e-01 -5.26405334e-01
7.24856377e-01 -1.30383193e+00 -5.05957723e-01 2.51112223e-01
-7.88669586e-01 3.46023858e-01 -1.63161457e-02 4.68164444e-01
-5.35411060e-01 8.11994374e-01 -6.44327223e-01 -6.47328436e-01
-6.46887779e-01 -1.21186483e+00 1.64492464e+00 -2.05343664e-01
1.47383645e-01 -1.00303388e+00 1.64532229e-01 7.39666343e-01
4.65896100e-01 -4.05919552e-01 5.35097301e-01 -7.28248477e-01
-4.78841275e-01 -4.56756443e-01 -3.81900609e-01 6.58906102e-02
-4.33944874e-02 1.59856141e-01 -8.46051276e-01 -7.13554621e-01
-2.92711318e-01 -7.05045879e-01 1.19003105e+00 4.35822368e-01
9.97539222e-01 6.61563501e-02 -5.30211329e-01 3.43060106e-01
1.50934601e+00 2.72776961e-01 6.86713755e-01 3.91881585e-01
7.27262855e-01 4.13693875e-01 7.68278599e-01 5.23657918e-01
4.41564143e-01 9.53676820e-01 1.58190265e-01 3.58823180e-01
-2.45063365e-01 -2.48936400e-01 5.54400265e-01 1.26575363e+00
4.96351905e-02 -7.54275084e-01 -7.48404384e-01 3.63986403e-01
-1.99215364e+00 -1.16277933e+00 1.24997012e-01 1.83705842e+00
6.32760346e-01 -1.74277022e-01 6.47463053e-02 2.04737738e-01
7.40372479e-01 5.95273733e-01 -9.67679098e-02 8.07473361e-02
-3.11537921e-01 5.79248704e-02 3.14846396e-01 2.84488350e-01
-1.28102922e+00 1.04429400e+00 6.93813992e+00 1.38328493e+00
-1.05159354e+00 -1.15300253e-01 1.32376343e-01 -4.02060211e-01
-1.85476318e-01 -1.68404788e-01 -7.61669755e-01 1.68252245e-01
9.62646902e-01 -5.73415160e-01 1.39534742e-01 4.18560743e-01
-1.83644295e-01 2.95198001e-02 -9.64231133e-01 1.48714709e+00
6.59560084e-01 -1.26619470e+00 6.52423024e-01 -2.37672210e-01
5.20378649e-01 2.58579478e-02 3.50479186e-01 3.64610940e-01
-1.46801472e-01 -9.48513865e-01 7.77839184e-01 3.92444640e-01
1.01831675e+00 -5.41581452e-01 6.78447902e-01 -1.79518223e-01
-1.55774498e+00 4.04917374e-02 -3.02315891e-01 5.96505046e-01
1.80932522e-01 2.50958472e-01 -5.24694085e-01 1.06656992e+00
8.55037689e-01 1.45273483e+00 -6.11723423e-01 1.05005133e+00
3.64455700e-01 3.25146943e-01 -1.73567876e-01 -5.14479987e-02
3.78916413e-01 8.72826874e-02 6.20655417e-01 1.45099902e+00
4.42570984e-01 -1.10753007e-01 1.94362357e-01 8.71637464e-02
-1.50249556e-01 5.49324334e-01 -5.47517955e-01 -3.46712321e-01
2.80839831e-01 1.20137668e+00 -4.48193461e-01 -5.43391883e-01
-6.10346437e-01 9.85295475e-01 -3.71422991e-02 3.79072130e-01
-8.36693883e-01 -2.69153744e-01 1.88251719e-01 -2.96063125e-01
6.66270196e-01 -7.51895905e-02 4.67755765e-01 -1.45882475e+00
9.91061702e-02 -1.06725681e+00 5.98804176e-01 -1.05122948e+00
-1.03127706e+00 5.38224816e-01 2.60892004e-01 -1.63440621e+00
-3.41797233e-01 -5.00724971e-01 4.07626629e-02 3.39496344e-01
-1.57019615e+00 -1.29027224e+00 -8.33085403e-02 9.89423573e-01
1.00595212e+00 -4.47828859e-01 5.77346921e-01 7.73275435e-01
-4.88521993e-01 6.97857916e-01 6.34861529e-01 1.48373291e-01
1.04566622e+00 -9.17720914e-01 -1.93461195e-01 5.07256687e-01
6.12347603e-01 4.04257953e-01 5.12641370e-01 -5.18694937e-01
-1.91494298e+00 -7.00142920e-01 6.60753846e-01 -2.97764361e-01
9.96812582e-01 -9.73942783e-03 -6.52346849e-01 7.74541318e-01
7.23690569e-01 -2.06804514e-01 5.49426496e-01 -2.86455512e-01
-5.54057479e-01 -2.01965630e-01 -7.57466078e-01 6.06227934e-01
6.60041630e-01 -8.13951135e-01 -7.49116123e-01 4.05586332e-01
5.06981015e-01 -3.87661904e-01 -9.85163510e-01 4.07965332e-01
6.94197476e-01 -8.13065708e-01 1.29264343e+00 -2.03045771e-01
6.57217979e-01 -6.42137676e-02 -5.86632669e-01 -8.96000445e-01
-2.37814829e-01 -3.76138389e-01 -2.75292099e-01 1.28629422e+00
9.81969852e-03 9.02061071e-03 3.95293832e-01 -1.69585094e-01
2.47002766e-03 -5.10013282e-01 -1.03090870e+00 -4.60766554e-01
1.03175513e-01 -5.37792802e-01 -7.37763271e-02 8.28612268e-01
1.49000481e-01 5.10496736e-01 -8.63522112e-01 -4.13373798e-01
5.55615306e-01 5.75527698e-02 3.68674874e-01 -9.05108988e-01
1.58631317e-02 -5.49340069e-01 -5.16006649e-01 -1.32485080e+00
9.23599005e-02 -1.03010929e+00 -1.12832494e-01 -1.37960875e+00
6.64477408e-01 2.19598457e-01 -3.34810436e-01 3.10659319e-01
1.04134493e-02 4.31911439e-01 3.60470176e-01 6.25507474e-01
-1.33663213e+00 7.16879845e-01 1.07336438e+00 -4.71604258e-01
3.78525347e-01 -3.92448634e-01 -3.80808294e-01 3.69096845e-01
3.98754954e-01 -1.48357421e-01 -3.98444355e-01 -6.90042138e-01
1.95015252e-01 4.01924282e-01 1.12370595e-01 -8.59643757e-01
6.08978868e-01 2.65014499e-01 3.13469827e-01 -1.14279175e+00
6.95094466e-01 -1.01246524e+00 -1.47331404e-02 1.72894597e-01
-6.17098451e-01 3.03051531e-01 1.75039113e-01 5.80153108e-01
-6.89922929e-01 -2.22134352e-01 4.24989492e-01 5.45047373e-02
-7.57716417e-01 3.70991588e-01 -6.34850919e-01 2.11950373e-02
8.94800127e-01 -1.20553493e-01 -3.64426613e-01 -6.54417217e-01
-3.39124113e-01 4.97245133e-01 1.35245457e-01 7.00734138e-01
1.07895243e+00 -1.42437267e+00 -8.90420258e-01 -1.26515344e-01
3.87277693e-01 -7.69810677e-01 3.69165927e-01 9.33147848e-01
-3.09321046e-01 1.05727696e+00 2.37039421e-02 -7.06649661e-01
-1.72830093e+00 5.82427382e-01 1.84533924e-01 -2.98244506e-01
-6.10753655e-01 5.31791449e-01 2.97755674e-02 3.84238720e-01
3.73513311e-01 2.84884393e-01 -6.49900615e-01 5.46418428e-01
8.03695798e-01 3.81375104e-01 1.20305322e-01 -1.05988503e+00
-3.94217581e-01 1.03807938e+00 -6.38530910e-01 -1.66354984e-01
1.09170580e+00 -5.64707875e-01 -1.38646722e-01 5.94515324e-01
1.57620096e+00 -2.39764139e-01 -3.89863342e-01 -6.17657542e-01
9.49810147e-02 -5.82447350e-01 3.53364319e-01 -4.36223418e-01
-1.22723758e+00 8.51902664e-01 7.43737042e-01 1.72705144e-01
1.16735923e+00 1.39479563e-01 9.62511659e-01 5.23574650e-01
3.94917101e-01 -1.04785907e+00 4.79448378e-01 5.94202220e-01
7.25377083e-01 -1.27598119e+00 3.25200051e-01 -7.73990229e-02
-6.20495439e-01 1.16205347e+00 1.00037217e-01 2.27112360e-02
6.50095344e-01 2.62876488e-02 -1.33930594e-01 -3.95176977e-01
-1.11420274e+00 -2.31518209e-01 7.93321669e-01 1.41777068e-01
7.51046121e-01 -4.77217704e-01 -2.62949824e-01 -1.29614934e-01
2.17510223e-01 5.03935292e-02 1.92231864e-01 1.11828351e+00
-5.19921422e-01 -1.13700294e+00 -2.99036354e-01 4.59939450e-01
-8.26851606e-01 -4.27538633e-01 -3.87749344e-01 7.08379328e-01
-6.28631771e-01 9.95777667e-01 -9.20198336e-02 -6.32474959e-01
2.82842338e-01 -1.77949890e-02 5.21975875e-01 -2.56430387e-01
-5.47230959e-01 5.95657468e-01 2.03369204e-02 -5.31109691e-01
-9.73329067e-01 -8.12289894e-01 -8.86122942e-01 -4.61392224e-01
-5.02977788e-01 2.16198638e-01 3.25485170e-01 9.96628404e-01
1.21653184e-01 5.01754642e-01 7.10052669e-01 -7.56902874e-01
-2.35527635e-01 -9.26863313e-01 -5.24891615e-01 1.47990763e-01
3.67750585e-01 -5.69550991e-01 -4.37149346e-01 5.63888364e-02] | [10.449665069580078, 0.8480726480484009] |
13229c80-631b-4e97-9a4a-70384f7b679a | learning-uncertainty-with-artificial-neural | 2105.05559 | null | https://arxiv.org/abs/2105.05559v1 | https://arxiv.org/pdf/2105.05559v1.pdf | Learning Uncertainty with Artificial Neural Networks for Improved Remaining Time Prediction of Business Processes | Artificial neural networks will always make a prediction, even when completely uncertain and regardless of the consequences. This obliviousness of uncertainty is a major obstacle towards their adoption in practice. Techniques exist, however, to estimate the two major types of uncertainty: model uncertainty and observation noise in the data. Bayesian neural networks are theoretically well-founded models that can learn the model uncertainty of their predictions. Minor modifications to these models and their loss functions allow learning the observation noise for individual samples as well. This paper is the first to apply these techniques to predictive process monitoring. We found that they contribute towards more accurate predictions and work quickly. However, their main benefit resides with the uncertainty estimates themselves that allow the separation of higher-quality from lower-quality predictions and the building of confidence intervals. This leads to many interesting applications, enables an earlier adoption of prediction systems with smaller datasets and fosters a better cooperation with humans. | ['Jochen De Weerdt', 'Hans Weytjens'] | 2021-05-12 | null | null | null | null | ['predictive-process-monitoring'] | ['time-series'] | [ 8.77951756e-02 4.87299800e-01 -1.34626508e-01 -7.68230021e-01
-2.78721571e-01 -1.46590695e-01 6.75103366e-01 3.18231761e-01
-3.23905975e-01 1.13641441e+00 -2.19365433e-01 -4.17590350e-01
-6.18294597e-01 -1.09898162e+00 -6.66538179e-01 -6.62438095e-01
-2.66943395e-01 8.11938107e-01 2.37502694e-01 2.96811432e-01
2.70963639e-01 5.84219873e-01 -1.78010595e+00 -5.10883965e-02
7.92258561e-01 1.35273588e+00 4.77449931e-02 4.98709977e-01
-2.72509605e-01 8.59710693e-01 -5.16368389e-01 -3.63081068e-01
-1.63356010e-02 1.00445665e-01 -3.73100579e-01 -3.47617865e-01
-1.38576731e-01 -4.06365812e-01 8.00866187e-02 8.93566787e-01
1.04118846e-01 1.03972852e-01 1.04830563e+00 -1.29675877e+00
-3.55802745e-01 9.63641763e-01 -1.02228001e-01 -3.49274635e-01
1.46975666e-01 -4.81399260e-02 5.56580544e-01 -2.76321173e-01
-1.21772187e-02 1.24190176e+00 1.08651114e+00 4.10647780e-01
-1.21232188e+00 -5.60908616e-01 -1.78867392e-02 2.51021147e-01
-1.26660454e+00 -2.88031012e-01 4.43325967e-01 -5.09206831e-01
8.38852763e-01 2.85675704e-01 4.88378316e-01 1.10081887e+00
5.40872931e-01 4.83797848e-01 1.06931078e+00 -4.95502025e-01
8.67215693e-01 5.64145446e-01 2.10560843e-01 8.75934735e-02
7.09630907e-01 6.57735705e-01 -2.34838367e-01 -1.37248933e-01
7.37534583e-01 3.02359134e-01 -3.02597821e-01 -4.57040608e-01
-7.45881259e-01 6.89225256e-01 7.94099942e-02 4.02155668e-01
-4.80484247e-01 2.47722462e-01 2.29981706e-01 2.56894410e-01
3.33508015e-01 4.00413424e-01 -6.33231759e-01 -2.68048704e-01
-9.53197062e-01 2.63349324e-01 1.31728435e+00 7.59881079e-01
5.73074937e-01 1.17441259e-01 -7.54135773e-02 4.66955394e-01
5.38608730e-01 4.50017601e-01 4.34294909e-01 -1.03406799e+00
1.39128953e-01 3.45194429e-01 5.24403691e-01 -1.01922357e+00
-4.27210033e-01 -2.66161680e-01 -1.00849950e+00 5.88687062e-01
5.85346997e-01 -4.25161242e-01 -9.13229048e-01 1.50008607e+00
3.38532068e-02 1.01299681e-01 1.29694313e-01 3.35563749e-01
6.66613132e-02 5.85079014e-01 1.39524668e-01 -2.73086488e-01
8.68182957e-01 -2.63141245e-01 -8.97391140e-01 -8.16046000e-02
2.42877737e-01 -3.56621623e-01 4.68067050e-01 8.54935348e-01
-7.31839001e-01 -6.11104250e-01 -9.90867615e-01 6.53260529e-01
-6.59001946e-01 -3.07895154e-01 7.17550218e-01 9.55849826e-01
-6.35599375e-01 1.34891331e+00 -1.09633315e+00 -3.40396054e-02
2.25416645e-01 4.75583255e-01 -6.33641034e-02 7.52002969e-02
-1.28408086e+00 1.28896439e+00 8.37817729e-01 3.54620367e-01
-5.79380691e-01 -5.74219048e-01 -7.26180553e-01 1.21222891e-01
4.19099897e-01 -2.48835281e-01 1.38097203e+00 -9.30036545e-01
-1.56446540e+00 5.55931628e-02 1.26722351e-01 -7.91796505e-01
8.45417023e-01 -2.14239076e-01 -4.99895096e-01 -4.92840379e-01
-4.97988254e-01 3.43504339e-01 7.58362830e-01 -1.19201446e+00
-7.82254696e-01 -3.34418684e-01 -4.39780205e-01 -2.70983577e-01
1.21532110e-02 -4.55093384e-01 1.78849012e-01 -2.25661114e-01
1.00837126e-01 -6.85281098e-01 -4.53435659e-01 -3.73280942e-02
-4.05206650e-01 -2.48347700e-01 5.62342405e-01 -4.88893181e-01
1.15247560e+00 -1.84765470e+00 -3.22517872e-01 5.26665211e-01
6.98500425e-02 1.20217696e-01 4.65731651e-01 4.84640211e-01
8.77761394e-02 2.43518934e-01 -3.58252198e-01 -3.25515032e-01
2.69375056e-01 6.19468868e-01 -3.16079974e-01 1.79633021e-01
1.52663603e-01 4.83763576e-01 -6.51442468e-01 -2.20340058e-01
5.04543662e-01 5.24257183e-01 1.95615396e-01 2.44657978e-01
-2.40696281e-01 8.74207318e-02 -4.31717426e-01 3.69374812e-01
5.71527302e-01 1.88690331e-02 1.15202419e-01 9.36404243e-02
-1.97036326e-01 -1.10989496e-01 -1.54954624e+00 8.46347034e-01
-3.27032179e-01 5.47370851e-01 -1.85564354e-01 -1.07025790e+00
1.13550186e+00 5.00986338e-01 2.54859149e-01 -1.21015713e-01
3.54697019e-01 2.56987482e-01 -4.23117913e-02 -2.70905375e-01
3.83714020e-01 -3.17427248e-01 1.55269742e-01 3.14357907e-01
3.17836627e-02 -3.98217291e-01 1.29246250e-01 -3.75650913e-01
6.32226467e-01 2.16351017e-01 5.66145837e-01 -2.78798819e-01
1.89570993e-01 -3.92199188e-01 6.40972793e-01 1.12555575e+00
-2.19863102e-01 4.16952550e-01 4.72120583e-01 -7.48058319e-01
-9.02139068e-01 -1.12894225e+00 -4.35817808e-01 5.54823339e-01
-2.34872490e-01 1.41455844e-01 -4.72004503e-01 -4.79909360e-01
4.33595449e-01 1.35791731e+00 -7.02989638e-01 -1.86326012e-01
1.42204002e-01 -7.03361750e-01 1.79218665e-01 7.03568161e-01
1.21674642e-01 -9.47263479e-01 -6.04992568e-01 3.90265763e-01
2.81444579e-01 -6.52733386e-01 4.29525554e-01 7.38896728e-01
-1.17522526e+00 -1.03574848e+00 -4.96213078e-01 8.71157572e-02
3.50837797e-01 -5.70095122e-01 1.09663320e+00 -2.46710867e-01
7.11188838e-02 3.32170606e-01 -5.27560189e-02 -1.20365214e+00
-6.80734038e-01 -3.77946466e-01 3.37092280e-01 -2.78119475e-01
6.52854979e-01 -7.12515116e-01 -6.40375465e-02 3.79541218e-01
-7.99080253e-01 -4.42968816e-01 6.17076516e-01 5.95033884e-01
6.42375946e-01 5.97631872e-01 6.05805457e-01 -8.75245690e-01
9.18316901e-01 -4.06169146e-01 -9.90303695e-01 4.35028404e-01
-1.41395414e+00 3.40080768e-01 4.60709900e-01 -3.35554630e-01
-1.26882279e+00 8.51302892e-02 4.70213778e-02 -1.64288893e-01
-6.30137920e-01 4.65734065e-01 8.92880931e-02 1.12900235e-01
7.61750281e-01 3.48634720e-02 5.10162748e-02 -4.89691854e-01
1.48637041e-01 6.99142814e-01 4.55412477e-01 -5.39806962e-01
4.48367625e-01 8.00386220e-02 2.08328336e-01 -7.83361018e-01
-5.52989841e-01 -1.91654563e-01 -6.05175078e-01 -2.38914475e-01
4.90547180e-01 -5.42790651e-01 -7.62211442e-01 3.96605521e-01
-9.38842535e-01 -1.72463506e-01 -6.54748559e-01 6.57816291e-01
-6.44067168e-01 3.03131402e-01 -4.70838517e-01 -1.60772550e+00
-2.30266616e-01 -9.07481134e-01 5.09772360e-01 3.69163841e-01
-4.38511848e-01 -1.06812298e+00 -4.20744047e-02 -1.64166644e-01
5.58574319e-01 1.91295236e-01 6.57444835e-01 -1.00989544e+00
-3.02386820e-01 -7.03272045e-01 -7.63775781e-02 7.36551344e-01
2.07428820e-02 4.40577179e-01 -1.07727277e+00 1.11934640e-01
1.84946477e-01 -1.84083104e-01 7.32040048e-01 7.65602469e-01
1.09647787e+00 -2.41957292e-01 -4.62781817e-01 6.98629767e-02
1.32038701e+00 4.88262832e-01 7.32425213e-01 2.28555888e-01
2.03591436e-02 7.34302461e-01 3.66422653e-01 6.06245220e-01
2.63822488e-02 2.60009140e-01 5.35037816e-01 3.94794226e-01
5.09583294e-01 -1.15639642e-01 1.86486408e-01 4.77924347e-01
-4.12431091e-01 -2.53802806e-01 -1.04687059e+00 2.95190305e-01
-1.93041050e+00 -1.02284908e+00 -2.12196812e-01 2.58469129e+00
8.07169020e-01 5.33832788e-01 -1.79609254e-01 3.89404625e-01
6.41006827e-01 -3.54913890e-01 -5.75764298e-01 -5.07429719e-01
2.30425790e-01 -1.30550638e-01 6.50127530e-01 4.87042129e-01
-1.13110876e+00 2.43095785e-01 7.49914169e+00 8.68322551e-01
-7.65735447e-01 -3.26852560e-01 8.47825944e-01 9.79360715e-02
-1.32720664e-01 -3.18156958e-01 -8.89132679e-01 7.14766622e-01
1.50559664e+00 -5.09584993e-02 2.08754256e-01 1.16869283e+00
2.35361561e-01 -5.53606868e-01 -1.29473877e+00 6.05477273e-01
-4.85589564e-01 -1.14330518e+00 -1.43822461e-01 -5.62364887e-03
6.70513988e-01 -1.22097462e-01 -1.01997808e-01 3.69742692e-01
7.39243984e-01 -1.15734184e+00 6.77021027e-01 1.17743468e+00
2.03464672e-01 -1.11790264e+00 1.37911451e+00 5.69456458e-01
-5.45903325e-01 -3.29446614e-01 -7.43837297e-01 -4.15776074e-01
1.98479861e-01 1.25503004e+00 -9.15921926e-01 4.87289220e-01
6.26380920e-01 3.09480757e-01 -1.40616983e-01 1.11908424e+00
-2.87422687e-01 6.11404657e-01 -6.90877140e-01 -3.80697280e-01
-3.92973647e-02 -2.07724288e-01 2.63851881e-01 9.47455466e-01
5.44554174e-01 -3.50102186e-01 -6.78286105e-02 1.07170403e+00
5.09572923e-01 -2.45824426e-01 -5.97748578e-01 -7.19738286e-03
5.91359496e-01 7.17607081e-01 -4.80925858e-01 -1.90228179e-01
-2.01973915e-01 4.32959795e-01 1.23253159e-01 3.14958513e-01
-5.00566125e-01 -3.75711769e-01 4.60269839e-01 -1.30110323e-01
1.22208640e-01 -2.98009068e-02 -5.72358429e-01 -7.03170419e-01
-1.70456842e-01 -5.49373209e-01 2.76321232e-01 -5.96639812e-01
-1.70139360e+00 4.89113510e-01 2.67154157e-01 -9.22280252e-01
-8.67761016e-01 -1.00873983e+00 -5.15199304e-01 9.71293271e-01
-1.24878109e+00 -7.30242372e-01 2.64908485e-02 1.42725900e-01
8.72209072e-02 -7.15658069e-02 1.09855747e+00 -3.06743443e-01
-4.29744899e-01 9.19154510e-02 6.17335975e-01 -2.16928124e-01
4.33769256e-01 -1.48997378e+00 -7.20200986e-02 5.19272864e-01
-6.01755120e-02 4.11195368e-01 1.13041615e+00 -7.36137748e-01
-7.58648932e-01 -8.36502314e-01 7.94471920e-01 -6.61952198e-01
7.88661361e-01 2.26320233e-02 -9.69170928e-01 6.27123356e-01
-1.84785366e-01 -3.64792198e-01 6.86567247e-01 4.81825441e-01
-8.24063122e-02 -3.19236845e-01 -1.38667643e+00 3.28975767e-01
3.46409470e-01 -2.47258469e-01 -8.69668543e-01 -7.82478824e-02
5.39953053e-01 -7.01423560e-05 -1.13259852e+00 5.29756665e-01
8.86517406e-01 -1.38936770e+00 8.42469931e-01 -4.22708482e-01
2.61005640e-01 6.46223426e-02 -1.96170017e-01 -1.22667027e+00
-9.66350287e-02 -2.09301323e-01 -4.59930927e-01 1.20783734e+00
7.15192437e-01 -9.01644766e-01 8.49510372e-01 1.50916195e+00
2.00593323e-01 -7.06596613e-01 -1.05734873e+00 -1.03066325e+00
1.16898678e-01 -1.03140593e+00 7.68531740e-01 6.29285038e-01
-5.39832935e-03 -3.40983391e-01 -4.90048438e-01 2.17994437e-01
9.57960963e-01 -1.15369670e-01 3.13210458e-01 -1.97878635e+00
-2.55104303e-01 -5.34588873e-01 -6.45356655e-01 -4.59770530e-01
-9.59363282e-02 7.69200549e-02 3.88631076e-01 -1.19709468e+00
-1.97119758e-01 -5.49210608e-01 -6.20274842e-01 2.59432793e-01
2.74735074e-02 -2.80523479e-01 -1.41625687e-01 2.94892471e-02
-2.54725754e-01 4.06527102e-01 7.24171519e-01 6.05938211e-03
-2.55428225e-01 6.94797099e-01 -4.35786813e-01 9.83086467e-01
1.10699475e+00 -5.72276533e-01 -6.07898355e-01 4.13690247e-02
2.55272001e-01 -4.82033230e-02 2.36374125e-01 -1.31824374e+00
2.63213575e-01 -3.93432289e-01 7.68522501e-01 -6.08258128e-01
2.66095787e-01 -1.30806637e+00 6.62811041e-01 5.94955206e-01
-4.46316272e-01 -4.56810832e-01 1.78252861e-01 9.95042384e-01
-2.06673458e-01 -8.51114929e-01 6.14210904e-01 -3.36605251e-01
-5.64582646e-01 6.66992888e-02 -5.65328419e-01 -6.13060832e-01
9.98596013e-01 -3.87261719e-01 2.56327808e-01 -6.71706557e-01
-1.05949473e+00 7.90229067e-02 1.66317627e-01 1.88191578e-01
2.98761398e-01 -9.28852856e-01 -4.68517870e-01 6.42655939e-02
-1.41489789e-01 7.53362030e-02 2.01134011e-01 3.64093810e-01
-1.79734856e-01 6.84183538e-01 -1.05062887e-01 -5.45857012e-01
-7.23410487e-01 6.07509851e-01 4.96378064e-01 -3.91411960e-01
-2.03942418e-01 7.54194438e-01 -2.85897791e-01 -4.55744922e-01
6.73954666e-01 -5.86356878e-01 -2.02786237e-01 -6.02544807e-02
8.08612645e-01 7.03755558e-01 -4.34016362e-02 -4.63395640e-02
1.67285409e-02 7.37336054e-02 1.82531342e-01 -1.78506244e-02
1.36150348e+00 -2.17101261e-01 -2.51483973e-02 1.02337921e+00
4.64074582e-01 -4.66233552e-01 -1.52377820e+00 -2.35695429e-02
4.61538494e-01 -3.50130171e-01 2.49496982e-01 -1.14857173e+00
-7.36296773e-01 9.47397828e-01 7.26392508e-01 6.24376774e-01
9.47448373e-01 -3.08266610e-01 -2.19380967e-02 8.10924113e-01
5.82223415e-01 -1.18022847e+00 -6.08609796e-01 3.04013342e-01
8.22883964e-01 -1.38653624e+00 9.24005210e-02 -1.62238821e-01
-3.94252867e-01 1.41863620e+00 3.78324330e-01 2.76430458e-01
8.81826162e-01 4.30324048e-01 -3.22569385e-02 2.11231828e-01
-7.60388076e-01 5.32909222e-02 2.86783516e-01 1.25999498e+00
3.39558423e-01 3.97313088e-01 -1.00735694e-01 1.00285172e+00
4.94178161e-02 5.05298316e-01 4.52751577e-01 7.28606105e-01
-7.01148927e-01 -1.07712293e+00 -6.29468977e-01 6.79596364e-01
-4.04604733e-01 1.69256896e-01 -4.84080873e-02 8.85006607e-01
1.01645477e-01 1.08071554e+00 -2.43965853e-02 -2.62477160e-01
3.66099924e-01 4.10281569e-01 1.49724931e-01 -1.68261737e-01
-1.31943598e-01 -3.41475099e-01 2.46252865e-01 -5.72731793e-01
-1.58500955e-01 -7.11621642e-01 -9.61727560e-01 -4.68172997e-01
-4.75588679e-01 4.45987612e-01 1.01191556e+00 9.07464087e-01
2.59154648e-01 5.12848020e-01 3.15823317e-01 -8.56168807e-01
-1.24198914e+00 -1.18150711e+00 -1.07599056e+00 -3.41610648e-02
1.67628065e-01 -7.13969827e-01 -5.84029913e-01 -1.06207060e-03] | [7.400041103363037, 3.8447766304016113] |
5d616c95-1802-4990-a8d5-b224f32e28d0 | frob-few-shot-robust-model-for-classification | null | null | https://openreview.net/forum?id=mZsZy481_F | https://openreview.net/pdf?id=mZsZy481_F | FROB: Few-shot ROBust Model for Classification with Out-of-Distribution Detection | Nowadays, classification and Out-of-Distribution (OoD) detection in the few-shot setting remain challenging aims mainly due to rarity and the limited samples in the few-shot setting, and because of adversarial attacks. Accomplishing these aims is important for critical systems in safety, security, and defence. In parallel, OoD detection is challenging since deep neural network classifiers set high confidence to OoD samples away from the training data. To address such limitations, we propose the Few-shot ROBust (FROB) model for classification and few-shot OoD detection. We devise a methodology for improved robustness and reliable confidence prediction for few-shot OoD detection. We generate the support boundary of the normal class distribution and combine it with few-shot Outlier Exposure (OE). We propose a self-supervised learning few-shot confidence boundary methodology based on generative and discriminative models, including classification. The main contribution of FROB is the combination of the generated boundary in a self-supervised learning manner and the imposition of low confidence at this learned boundary. FROB implicitly generates strong adversarial samples on the boundary and forces samples from OoD, including our boundary, to be less confident by the classifier. FROB achieves generalization to unseen anomalies and adversarial attacks, with applicability to unknown, in the wild, test sets that do not correlate to the training datasets. To improve robustness, FROB redesigns and streamlines OE to work even for zero-shots. By including our learned boundary, FROB effectively reduces the threshold linked to the model’s few-shot robustness, and maintains the OoD performance approximately constant and independent of the number of few-shot samples. The few-shot robustness analysis evaluation of FROB on different image sets and on One-Class Classification (OCC) data shows that FROB achieves competitive state-of-the-art performance and outperforms benchmarks in terms of robustness to the outlier OoD few-shot sample population and variability. | ['Sotirios A. Tsaftaris', 'Mehrdad Yaghoobi', 'Nikolaos Dionelis'] | 2021-09-29 | null | null | null | null | ['one-class-classification'] | ['miscellaneous'] | [ 7.65542760e-02 3.74837592e-02 -8.48003477e-02 -3.35930586e-02
-9.99899268e-01 -3.87898684e-01 6.37550294e-01 2.82610089e-01
-6.03494793e-02 4.34807986e-01 -2.78104872e-01 6.85163736e-02
-1.23855129e-01 -8.39844227e-01 -8.88781190e-01 -7.41104901e-01
-2.01672554e-01 2.40085527e-01 6.05278790e-01 -1.85084686e-01
1.07896276e-01 5.00888050e-01 -1.67519939e+00 2.21103266e-01
7.24156916e-01 1.15822721e+00 -5.53957760e-01 7.32331514e-01
2.80110955e-01 5.11808157e-01 -1.18366432e+00 -2.74384230e-01
5.89680672e-01 -4.64601099e-01 5.12892269e-02 -2.21903734e-02
5.73593199e-01 -4.39378589e-01 -3.51831079e-01 1.14064121e+00
7.58637130e-01 2.25539654e-01 1.06594753e+00 -1.70844281e+00
-5.75468004e-01 7.70858750e-02 -4.79035497e-01 3.79521906e-01
1.33795232e-01 4.93530154e-01 5.19266546e-01 -8.35948110e-01
4.64511484e-01 9.64287281e-01 8.49615932e-01 7.56025493e-01
-1.18624723e+00 -6.48974836e-01 -1.69760972e-01 -1.45866305e-01
-1.40658975e+00 -3.15172017e-01 6.51788652e-01 -3.58608127e-01
7.48115063e-01 2.07456902e-01 2.14459851e-01 1.64910388e+00
5.51316798e-01 3.84654880e-01 8.19700003e-01 -4.59278136e-01
7.66945481e-01 2.60442019e-01 1.40135869e-01 3.11720848e-01
5.09504735e-01 6.16125703e-01 -5.89609504e-01 -3.60621810e-01
1.99886903e-01 2.92470038e-01 -2.27021933e-01 -2.43973523e-01
-3.34855497e-01 8.86629045e-01 1.74176276e-01 2.16660663e-01
-8.47581476e-02 1.00699916e-01 5.28615892e-01 4.41544265e-01
5.96936226e-01 3.62869859e-01 -1.41311154e-01 -3.27262506e-02
-9.82701063e-01 8.49920809e-02 9.44129586e-01 9.38429594e-01
5.26606083e-01 4.91226226e-01 -4.99228626e-01 8.04287255e-01
-1.28072482e-02 6.72337115e-01 5.26880801e-01 -3.94159317e-01
2.04146996e-01 3.48651946e-01 -1.46204636e-01 -9.52421427e-01
-2.17449591e-02 -5.71537971e-01 -8.82809162e-01 7.25095451e-01
4.72813129e-01 -1.09502748e-01 -1.27014208e+00 1.60106158e+00
4.02704269e-01 5.83300769e-01 1.12625428e-01 6.22873664e-01
5.43072581e-01 5.56311071e-01 -8.84168670e-02 -1.00295283e-01
1.03852892e+00 -4.41129386e-01 -5.08908331e-01 -1.62180841e-01
6.73972964e-01 -3.04634005e-01 1.02412212e+00 4.83264387e-01
-4.66693133e-01 -5.98252237e-01 -1.41595674e+00 6.92604125e-01
-6.70164824e-01 -4.38361198e-01 3.82344276e-02 1.20724535e+00
-4.56695229e-01 7.99639642e-01 -5.06793439e-01 -1.76549762e-01
6.36195779e-01 1.41249999e-01 -3.88689190e-01 -1.51478767e-01
-1.37140822e+00 6.04168475e-01 4.63218004e-01 -2.23127067e-01
-1.25840902e+00 -8.83304298e-01 -1.08179331e+00 -1.51560735e-02
5.58569252e-01 -1.38491467e-01 7.05464065e-01 -6.83592141e-01
-1.14280283e+00 7.95657039e-01 3.70608360e-01 -9.36911345e-01
7.96067417e-01 -2.54894167e-01 -7.07473218e-01 1.81641415e-01
1.28969774e-01 9.17083398e-02 1.65354550e+00 -1.18388307e+00
-3.36574614e-01 -2.23516077e-01 -4.40041512e-01 -5.32886326e-01
-4.77223754e-01 -2.30529115e-01 -1.59763247e-01 -8.26334655e-01
-1.91630989e-01 -7.61267304e-01 2.84289289e-02 -1.27470210e-01
-5.45543849e-01 -3.56916711e-02 1.25207019e+00 -1.64987370e-01
1.18670201e+00 -2.62210560e+00 -6.36157990e-01 3.53368908e-01
1.19327463e-01 5.69558382e-01 4.65314984e-02 3.60186607e-01
-1.92995057e-01 -5.36860228e-02 -4.12772179e-01 -3.35850775e-01
2.04151586e-01 2.03151703e-01 -8.60953331e-01 7.23656476e-01
6.95026219e-01 6.80509567e-01 -8.77034783e-01 -1.46182477e-01
2.84309834e-01 2.59421319e-01 -5.02821088e-01 4.55823302e-01
-8.70497376e-02 9.96988360e-03 -8.29918161e-02 9.61673081e-01
7.76512802e-01 3.44433755e-01 -5.45734465e-01 8.55282135e-03
3.26645702e-01 -5.23424447e-01 -1.44828475e+00 1.00247312e+00
-1.92083791e-01 3.51137310e-01 -3.91877800e-01 -8.79173994e-01
1.33792245e+00 2.61767119e-01 1.35347202e-01 -5.32312989e-01
3.21248144e-01 4.37403977e-01 -2.24129066e-01 -4.18691784e-01
1.52670011e-01 -2.87395328e-01 -3.63661051e-01 2.89648145e-01
5.26364863e-01 -3.29492301e-01 2.74134576e-02 2.41146281e-01
1.36768079e+00 -3.38331535e-02 3.89274418e-01 -4.28052880e-02
2.62156188e-01 -4.28599179e-01 7.21155524e-01 1.27259612e+00
-5.07629216e-01 8.84611845e-01 6.50663018e-01 -2.57189572e-01
-1.01487577e+00 -1.33997118e+00 -3.03982019e-01 6.23736084e-01
1.42884508e-01 -2.25245371e-01 -7.57369220e-01 -1.03841925e+00
3.16067547e-01 1.10267365e+00 -8.41335475e-01 -8.78382921e-01
-2.05123857e-01 -8.92817259e-01 8.62901926e-01 4.06382173e-01
2.82817632e-01 -9.24840391e-01 -5.49756646e-01 2.37086818e-01
5.99321783e-01 -1.14237344e+00 -2.90128589e-01 6.27323806e-01
-4.51992989e-01 -1.12624550e+00 -5.00544608e-01 -2.20582604e-01
5.24638057e-01 -7.97414258e-02 8.16364646e-01 -1.84482671e-02
-7.06527472e-01 5.21328866e-01 -5.77843070e-01 -8.59491467e-01
-8.30571115e-01 -3.89470845e-01 3.26308757e-01 2.81325966e-01
4.43389654e-01 -5.41417837e-01 -1.94819167e-01 4.10558522e-01
-1.21652007e+00 -1.06392336e+00 3.37171346e-01 1.12099743e+00
5.08628964e-01 3.00424695e-01 9.67033148e-01 -9.35645700e-01
4.63321209e-01 -8.24377418e-01 -4.97577697e-01 -1.12019137e-01
-5.75250268e-01 -8.09298679e-02 9.12605166e-01 -7.99443662e-01
-7.13241518e-01 -3.82782221e-01 -1.03877328e-01 -1.10907578e+00
-6.11338854e-01 -2.18168497e-01 -2.56077409e-01 -7.53344893e-02
1.26124477e+00 -5.01539111e-02 4.91237151e-05 -2.22405955e-01
1.92498431e-01 6.66470110e-01 5.62332809e-01 -4.11224961e-01
1.30582273e+00 5.60871601e-01 8.22157413e-02 -1.05960631e+00
-1.09353101e+00 -4.65698481e-01 -4.53454614e-01 -2.86352485e-01
5.69007635e-01 -6.93826318e-01 -3.17779809e-01 6.80659652e-01
-7.38405287e-01 -2.96117991e-01 -8.40330958e-01 2.80311316e-01
-5.01493216e-01 3.66101652e-01 -3.01900446e-01 -1.19760263e+00
-3.00234944e-01 -8.84186387e-01 1.17548871e+00 1.26640096e-01
-1.65698752e-01 -7.59135842e-01 7.60635808e-02 -1.96111724e-01
1.40960991e-01 9.10159588e-01 6.87206209e-01 -1.26577556e+00
-2.00879678e-01 -9.48590338e-01 1.39204279e-01 8.98070991e-01
-5.32839112e-02 7.35282302e-02 -1.42750204e+00 -4.71523166e-01
4.31560427e-01 -5.40752888e-01 9.98134375e-01 1.92263216e-01
1.01172364e+00 -6.87963590e-02 -7.85416961e-02 7.06604004e-01
1.45442104e+00 4.21545319e-02 7.08963096e-01 1.85709506e-01
3.53457123e-01 4.29029018e-01 9.38426435e-01 5.64974189e-01
-7.02489614e-01 5.04562378e-01 7.17948437e-01 6.06519915e-02
-1.66845962e-01 -2.43498772e-01 5.83664477e-01 1.21504683e-02
2.89077580e-01 -4.17535514e-01 -7.54044354e-01 4.53421205e-01
-1.56350422e+00 -1.30753160e+00 2.40075588e-03 2.52250934e+00
4.40924168e-01 8.30840826e-01 1.72744244e-01 5.81836760e-01
8.81663442e-01 1.77881807e-01 -6.08798146e-01 -4.99442816e-01
-1.80411786e-01 3.17241341e-01 4.29771990e-01 1.17550828e-01
-1.26303279e+00 6.87844634e-01 5.46543026e+00 1.22557414e+00
-9.46020305e-01 1.65248469e-01 7.50685990e-01 -9.64045748e-02
1.84764951e-01 -2.45283082e-01 -1.12705648e+00 7.44488895e-01
9.75441396e-01 4.07403708e-02 3.62792574e-02 1.08815598e+00
-9.88915786e-02 -4.45361696e-02 -1.02496433e+00 7.29702830e-01
5.94591498e-01 -1.06296790e+00 -1.84985489e-01 1.04122259e-01
8.63199472e-01 -1.36378661e-01 1.75473168e-01 6.64659798e-01
1.93439107e-02 -9.40727353e-01 8.03604186e-01 4.14086789e-01
9.29323614e-01 -1.01684391e+00 1.02014148e+00 6.38119221e-01
-8.36800933e-01 -4.59527314e-01 -4.72204357e-01 2.92911768e-01
-1.17635585e-01 9.35222089e-01 -8.27353120e-01 5.43726623e-01
7.09077120e-01 5.09361744e-01 -4.74245846e-01 8.31859827e-01
-2.53668845e-01 7.60070503e-01 -3.41458917e-01 1.64414123e-01
2.12161347e-01 1.03208303e-01 1.01630604e+00 1.13579690e+00
2.29505762e-01 -2.25164533e-01 1.59219071e-01 9.59451675e-01
-6.72258511e-02 -2.55572200e-01 -9.68532979e-01 1.34458989e-01
5.54100275e-01 9.59724367e-01 -7.18044281e-01 -2.18626946e-01
-1.12897314e-01 8.32484305e-01 6.68380335e-02 3.03449899e-01
-1.04653203e+00 -7.87873089e-01 4.92317349e-01 1.03233136e-01
4.18165714e-01 1.77072212e-01 -9.09954831e-02 -1.03547621e+00
-1.09057240e-02 -9.24690425e-01 5.62771559e-01 -2.00235054e-01
-1.77401268e+00 6.30553901e-01 2.54785996e-02 -1.77906275e+00
-3.01571697e-01 -5.86319208e-01 -1.06461358e+00 4.79347527e-01
-1.20118594e+00 -9.55337286e-01 -3.86423707e-01 4.73524332e-01
5.47640383e-01 -4.15600151e-01 8.28733861e-01 8.89772847e-02
-5.91702640e-01 9.98158753e-01 2.20306925e-02 2.33657971e-01
1.00182998e+00 -1.17057979e+00 3.93302590e-01 1.20433724e+00
7.28093237e-02 1.96908444e-01 7.84755588e-01 -8.76027226e-01
-9.32611644e-01 -1.37137651e+00 1.14742868e-01 -6.11365616e-01
8.14360201e-01 -6.39340520e-01 -1.07261527e+00 1.76163182e-01
-4.45254356e-01 5.89138865e-01 6.09909058e-01 -3.33796799e-01
-5.73424459e-01 -1.61661506e-01 -1.57090402e+00 4.01815236e-01
6.73133373e-01 -4.77633893e-01 -8.12145948e-01 2.54371464e-01
7.83300459e-01 -2.04123572e-01 -6.59260333e-01 4.55095708e-01
1.08766302e-01 -1.16790009e+00 9.95262027e-01 -4.99316245e-01
4.79783528e-02 -3.30682874e-01 -2.61419028e-01 -1.20348227e+00
1.59947291e-01 -7.33823419e-01 -6.10591233e-01 1.39658213e+00
4.57948148e-02 -8.24301243e-01 5.88007450e-01 2.09052056e-01
-3.25884908e-01 -7.92710185e-01 -1.18916476e+00 -1.42979801e+00
-2.19804183e-01 -6.92986786e-01 3.59737873e-01 8.19376111e-01
-3.97098899e-01 -1.70393422e-01 -6.21619821e-01 5.45043588e-01
1.00165367e+00 -2.20678434e-01 1.05396056e+00 -1.15201819e+00
-5.71208894e-01 6.08040243e-02 -8.82050455e-01 -1.85287327e-01
-9.48069338e-03 -7.22475648e-01 1.93754658e-01 -6.94640815e-01
-1.84260815e-01 -2.36583799e-01 -3.85898143e-01 3.96254271e-01
-1.96033910e-01 7.30783343e-01 1.54449448e-01 6.38098344e-02
-4.22165990e-01 6.17695987e-01 5.17434776e-01 -1.25187114e-01
-2.86670506e-01 1.67898297e-01 -1.84228644e-01 7.13684857e-01
5.97476065e-01 -9.71642792e-01 -2.08111167e-01 5.09287417e-01
-2.58513242e-01 -3.26594412e-01 6.24013066e-01 -1.46676171e+00
-8.32606703e-02 1.39022052e-01 4.62973535e-01 -5.17694175e-01
3.76132607e-01 -7.44905055e-01 -3.63768935e-01 7.16821194e-01
6.44321069e-02 -5.95768511e-01 1.98902413e-01 1.04087687e+00
-1.22637078e-01 -5.98233283e-01 1.15623379e+00 2.45960176e-01
-4.92170393e-01 4.26082134e-01 -1.72463566e-01 5.43253422e-01
1.50041676e+00 -5.52491665e-01 -3.70459557e-01 -2.03061044e-01
-5.89118958e-01 -5.82763366e-02 4.19120371e-01 4.00330812e-01
7.24535525e-01 -1.21312582e+00 -5.57619035e-01 8.37938964e-01
4.69166517e-01 -1.11015983e-01 3.29036117e-01 6.95667684e-01
-1.40550435e-01 -2.58973181e-01 -7.73648322e-02 -7.36539304e-01
-9.81158495e-01 9.00348783e-01 3.72660458e-01 4.89992164e-02
-6.62227035e-01 5.83052516e-01 -3.56181078e-02 -1.03528686e-01
4.30348188e-01 -3.11191636e-03 9.98576656e-02 8.85103717e-02
6.63746536e-01 4.21094626e-01 1.33291036e-01 -3.45720410e-01
-1.85464740e-01 2.92192042e-01 7.32282251e-02 2.88165718e-01
1.31088185e+00 3.10959756e-01 3.88394713e-01 7.41558492e-01
1.10063839e+00 2.13492393e-01 -1.45766783e+00 6.93364814e-02
-3.66414785e-02 -6.54355168e-01 -1.91496164e-01 -5.40858507e-01
-6.25814080e-01 8.74472022e-01 7.21244872e-01 5.09823680e-01
9.05407071e-01 -7.93293417e-02 6.80192769e-01 1.85735747e-01
1.58891380e-01 -1.16759121e+00 6.30216479e-01 4.76854444e-01
5.69419384e-01 -1.24456215e+00 -1.74631551e-01 -1.95143059e-01
-6.05805874e-01 1.04114950e+00 6.59592807e-01 -5.60456812e-01
7.02268839e-01 4.63834971e-01 -6.37536719e-02 -8.63388181e-02
-5.61257422e-01 5.49581125e-02 4.30089116e-01 9.57176268e-01
-3.46664757e-01 -3.44307661e-01 2.09270328e-01 7.79004633e-01
1.85757801e-01 -3.74421746e-01 4.36490893e-01 8.79192770e-01
-5.89734733e-01 -6.95758522e-01 -7.43566692e-01 5.91628194e-01
-3.18016320e-01 5.77180348e-02 -1.37723207e-01 8.46344292e-01
3.78434926e-01 9.25638556e-01 1.15656443e-01 -4.00328219e-01
4.52906370e-01 2.98693657e-01 -6.22905418e-02 -7.10142851e-01
-5.17885029e-01 1.26307048e-02 -2.06117019e-01 -5.96219778e-01
2.63989985e-01 -4.14638966e-01 -1.09083152e+00 7.32086301e-02
-6.81680679e-01 6.31576627e-02 4.69298482e-01 9.01312768e-01
2.39898458e-01 6.33423865e-01 9.59963977e-01 -7.12167025e-01
-1.24597454e+00 -8.39145780e-01 -8.29755723e-01 7.87519336e-01
4.61322904e-01 -7.92262316e-01 -1.19899404e+00 -3.38167399e-01] | [7.822861194610596, 2.427135705947876] |
01acbf8d-405e-467b-a550-5a723a2bc1a0 | coco-a-coupled-contrastive-framework-for | 2306.04979 | null | https://arxiv.org/abs/2306.04979v2 | https://arxiv.org/pdf/2306.04979v2.pdf | CoCo: A Coupled Contrastive Framework for Unsupervised Domain Adaptive Graph Classification | Although graph neural networks (GNNs) have achieved impressive achievements in graph classification, they often need abundant task-specific labels, which could be extensively costly to acquire. A credible solution is to explore additional labeled graphs to enhance unsupervised learning on the target domain. However, how to apply GNNs to domain adaptation remains unsolved owing to the insufficient exploration of graph topology and the significant domain discrepancy. In this paper, we propose Coupled Contrastive Graph Representation Learning (CoCo), which extracts the topological information from coupled learning branches and reduces the domain discrepancy with coupled contrastive learning. CoCo contains a graph convolutional network branch and a hierarchical graph kernel network branch, which explore graph topology in implicit and explicit manners. Besides, we incorporate coupled branches into a holistic multi-view contrastive learning framework, which not only incorporates graph representations learned from complementary views for enhanced understanding, but also encourages the similarity between cross-domain example pairs with the same semantics for domain alignment. Extensive experiments on popular datasets show that our CoCo outperforms these competing baselines in different settings generally. | ['Xiao Luo', 'Xian-Sheng Hua', 'Chong Chen', 'Zeyu Ma', 'Long Lan', 'Mengzhu Wang', 'Li Shen', 'Nan Yin'] | 2023-06-08 | null | null | null | null | ['graph-classification', 'graph-representation-learning'] | ['graphs', 'methodology'] | [ 1.73448861e-01 2.57199287e-01 -3.36496234e-01 -3.86085808e-01
-2.02914655e-01 -7.81594455e-01 4.51858729e-01 1.61568522e-01
1.02172956e-01 4.83011186e-01 2.33916163e-01 -2.08020538e-01
-2.20216468e-01 -9.65946853e-01 -5.87392211e-01 -5.13934433e-01
8.16354677e-02 7.38115251e-01 2.25095078e-01 -4.66892272e-01
3.92992347e-02 3.08222294e-01 -7.61030972e-01 2.06176508e-02
1.19582629e+00 6.87007785e-01 1.24568582e-01 1.48940608e-01
-4.44278181e-01 8.02923620e-01 -1.47437841e-01 -4.24242049e-01
1.59362644e-01 -5.30391097e-01 -8.53338182e-01 2.11297333e-01
7.23215997e-01 -1.14181593e-01 -5.68375409e-01 1.26252615e+00
3.32257420e-01 1.78046718e-01 6.49787188e-01 -1.41418695e+00
-1.19162512e+00 5.58864594e-01 -7.66398370e-01 2.78259397e-01
2.67802715e-01 5.92227243e-02 1.18862379e+00 -6.45049453e-01
9.52114940e-01 1.21458721e+00 5.51935554e-01 3.72306764e-01
-1.41665256e+00 -7.55635440e-01 6.05364561e-01 2.16257215e-01
-1.03713322e+00 6.83422163e-02 1.43971634e+00 -4.11742389e-01
7.32497096e-01 -3.73939663e-01 8.06026638e-01 1.24625492e+00
-3.00190926e-01 6.26760721e-01 1.06812203e+00 -1.83260411e-01
-5.15409037e-02 -2.76097339e-02 1.66102722e-01 9.87030029e-01
2.87084997e-01 2.95772590e-02 -3.95871788e-01 9.00383517e-02
1.14710248e+00 5.56679852e-02 -3.00438315e-01 -1.17401779e+00
-1.16577029e+00 8.83435547e-01 1.03488624e+00 8.30219686e-02
-7.97954425e-02 -1.58950776e-01 6.60550892e-01 4.78158176e-01
6.60921335e-01 5.19129932e-01 -4.35896993e-01 3.57147545e-01
-2.28451386e-01 -3.49473916e-02 5.75092256e-01 1.31542850e+00
8.86662364e-01 6.93415701e-02 1.85865164e-01 9.16586399e-01
2.45004773e-01 2.29307801e-01 7.26902634e-02 -4.18899238e-01
9.80904341e-01 1.33371818e+00 -6.18484259e-01 -1.36941993e+00
-3.88349384e-01 -6.34217143e-01 -1.17244291e+00 -8.37273449e-02
3.91157627e-01 1.26449883e-01 -8.87532771e-01 1.81594050e+00
3.56809109e-01 2.29118869e-01 -3.34749930e-02 9.01366889e-01
1.11513674e+00 3.16808403e-01 1.39383644e-01 2.17739210e-01
1.19387388e+00 -1.24222660e+00 -4.28755015e-01 -4.95777160e-01
8.63599062e-01 -3.63800555e-01 1.16678631e+00 1.61241367e-01
-6.76761568e-01 -5.24642706e-01 -1.01537359e+00 -2.50982553e-01
-5.25069594e-01 -1.56821832e-01 7.69884884e-01 1.35385081e-01
-9.62114692e-01 4.76344526e-01 -4.76012975e-01 -6.06527150e-01
7.71078944e-01 1.59036666e-01 -7.37812281e-01 -4.01272207e-01
-1.20863008e+00 6.39223993e-01 7.18835115e-01 -9.33200270e-02
-5.93041778e-01 -4.36530590e-01 -1.25053942e+00 1.11325219e-01
7.61647046e-01 -6.87911272e-01 5.62435269e-01 -9.75962102e-01
-1.27932084e+00 9.67805266e-01 4.16942567e-01 -9.84484404e-02
3.46638024e-01 2.64567882e-02 -4.19279575e-01 3.17153186e-01
2.54281223e-01 7.07636476e-01 7.24815309e-01 -1.12026823e+00
-4.94560972e-02 -5.45107305e-01 3.04664195e-01 5.17050564e-01
-4.33200866e-01 -5.57790935e-01 -6.20108366e-01 -9.65759993e-01
4.30186331e-01 -9.92689312e-01 -3.39421928e-01 7.91704059e-02
-3.62827957e-01 -3.54112536e-01 8.75517249e-01 -7.52594411e-01
1.05553973e+00 -2.13955379e+00 5.33410251e-01 1.84829623e-01
6.42635643e-01 5.00878096e-02 -5.15372396e-01 3.57151628e-01
-3.40929151e-01 -1.25925448e-02 -3.32642525e-01 2.54255593e-01
-2.13904247e-01 1.80271819e-01 -1.90076098e-01 3.52226675e-01
3.77929568e-01 1.23271704e+00 -1.21529925e+00 -6.27931893e-01
2.17081830e-01 1.98390603e-01 -6.40839040e-01 2.69602567e-01
-3.64952296e-01 7.69378364e-01 -6.49161637e-01 5.25355518e-01
9.44288313e-01 -9.83049333e-01 6.89860106e-01 -4.09084648e-01
5.86760044e-01 2.10494667e-01 -9.84929144e-01 2.10759330e+00
-1.89973027e-01 3.63647133e-01 -1.25882223e-01 -1.53326857e+00
1.23011780e+00 -1.17693752e-01 1.86258033e-01 -9.57978308e-01
-1.37539878e-01 1.21256515e-01 -1.18967786e-01 -3.24610621e-01
1.15324460e-01 8.71721562e-03 -6.43443763e-02 2.89721191e-01
3.56763154e-01 -9.31866243e-02 1.15089297e-01 6.34047270e-01
1.03312683e+00 5.06107986e-01 4.38122541e-01 -3.24398726e-01
6.76907361e-01 4.09680568e-02 5.76762676e-01 1.44038573e-01
-2.64327347e-01 4.53700542e-01 8.03888500e-01 -5.71565211e-01
-9.21968997e-01 -1.24029803e+00 3.22899848e-01 1.04273093e+00
7.00479925e-01 -3.85445565e-01 -3.68372411e-01 -1.40216553e+00
-4.82326746e-02 1.74704880e-01 -5.59746265e-01 -3.52968514e-01
-8.37859392e-01 -3.28250766e-01 2.28418916e-01 8.41340661e-01
7.48886108e-01 -9.80781972e-01 2.32209489e-01 1.34691820e-01
-2.81592667e-01 -1.35377312e+00 -7.99275100e-01 7.94365723e-03
-1.08919573e+00 -1.35333467e+00 -6.78802550e-01 -1.23436844e+00
9.99049366e-01 6.44173026e-01 1.42085385e+00 1.66770443e-01
8.06621090e-02 3.47691059e-01 -3.94880086e-01 1.59680977e-01
-1.27800047e-01 3.55777413e-01 -2.77434587e-01 -2.67565697e-01
3.80869836e-01 -9.56048548e-01 -4.69443619e-01 4.17364240e-01
-5.81523061e-01 3.01937640e-01 6.21882796e-01 1.00861382e+00
5.47785878e-01 -2.41523191e-01 7.76355088e-01 -1.32555270e+00
6.91376150e-01 -5.35035968e-01 -6.55666053e-01 4.20333296e-01
-8.53038728e-01 3.06427300e-01 8.72668386e-01 -2.78558850e-01
-9.54470217e-01 -1.07092895e-01 3.42611879e-01 -6.96611583e-01
-1.15936399e-01 6.88144863e-01 -4.25451159e-01 -2.46378705e-01
6.21126950e-01 2.13998988e-01 2.16958448e-01 -4.10459012e-01
6.92800701e-01 -2.07671449e-02 4.07594115e-01 -6.20501518e-01
9.24942017e-01 2.64813632e-01 1.08392842e-01 -5.20209670e-01
-8.53913784e-01 -5.30919731e-01 -9.18811262e-01 -6.24882169e-02
6.80031300e-01 -1.12415433e+00 -2.77669519e-01 3.73514026e-01
-9.48763132e-01 -4.13271397e-01 8.63851011e-02 2.86950111e-01
-4.73352045e-01 8.21128786e-01 -3.83387476e-01 -2.73791291e-02
-4.29523401e-02 -8.80530834e-01 9.43068445e-01 2.05927327e-01
1.43061414e-01 -1.51687121e+00 1.15588814e-01 4.63922650e-01
1.41548350e-01 4.86866742e-01 1.19922304e+00 -8.07821274e-01
-7.15282381e-01 1.61776409e-01 -7.81330168e-01 2.39903063e-01
3.43724340e-01 -3.71260464e-01 -4.85318691e-01 -4.98114705e-01
-5.39716721e-01 -5.21013558e-01 8.31860483e-01 2.08619222e-01
1.24466968e+00 -1.75423637e-01 -5.45617163e-01 8.91456485e-01
1.41745138e+00 -2.52009749e-01 3.06094170e-01 2.83205599e-01
1.38210857e+00 7.69722402e-01 4.75547910e-01 -1.35923207e-01
5.93285680e-01 5.30199230e-01 5.83719909e-01 -2.58958966e-01
-2.55310982e-01 -5.18594980e-01 1.22584254e-01 1.06351399e+00
-3.26280631e-02 -2.46416569e-01 -9.58375990e-01 5.92723727e-01
-1.78027999e+00 -6.52561665e-01 -1.36593834e-01 1.75798225e+00
4.30026889e-01 2.41187394e-01 1.64844126e-01 -5.15938818e-01
1.03239250e+00 5.05051017e-01 -8.57085049e-01 -1.49334343e-02
-2.28626102e-01 9.78250057e-02 3.09800446e-01 2.27219239e-01
-1.17752635e+00 1.05705452e+00 5.14719296e+00 6.37608469e-01
-9.28184032e-01 -1.22117452e-01 4.62540805e-01 4.36059296e-01
-5.07635593e-01 1.84809536e-01 -2.49091536e-01 2.44359791e-01
3.71700048e-01 -1.45191699e-01 5.55295646e-01 1.17736757e+00
-4.36538130e-01 4.98838663e-01 -1.08789980e+00 1.12409365e+00
5.84335141e-02 -1.41471601e+00 3.56083751e-01 2.92744428e-01
8.20090055e-01 1.09028667e-01 -3.67688127e-02 5.67645192e-01
5.89390635e-01 -9.44837391e-01 7.76908696e-02 1.66299015e-01
8.34128976e-01 -6.70288682e-01 5.38104415e-01 1.78543732e-01
-1.66484928e+00 1.14301451e-01 -3.78813624e-01 1.53739035e-01
-3.38287875e-02 3.45763654e-01 -7.11084306e-01 1.03701127e+00
5.67284048e-01 1.47735858e+00 -8.96968782e-01 5.74053168e-01
-4.73080993e-01 3.87715071e-01 2.65021808e-02 1.54327586e-01
3.76115501e-01 -7.80615091e-01 5.10444283e-01 8.57525885e-01
1.77295171e-02 -2.14636028e-02 6.00722909e-01 1.00304127e+00
-4.32877153e-01 1.74625650e-01 -1.05007076e+00 -4.73495126e-01
5.06648123e-01 1.42514920e+00 -8.56852949e-01 -2.88343996e-01
-6.90220177e-01 1.07852566e+00 1.07864714e+00 5.20444751e-01
-6.47680640e-01 -3.37748438e-01 2.83321559e-01 3.87451723e-02
1.57550558e-01 -3.21830332e-01 -1.90214545e-01 -1.50454831e+00
9.76150036e-02 -9.55030084e-01 8.27419221e-01 -6.06926739e-01
-1.84387863e+00 5.32848895e-01 -3.15758027e-02 -1.47245347e+00
1.42489970e-01 -7.41769373e-01 -5.47715247e-01 8.58827651e-01
-1.66740000e+00 -1.53241134e+00 -7.66146362e-01 7.12334871e-01
3.72031331e-01 -3.53085518e-01 4.75983769e-01 2.38141492e-01
-5.29444218e-01 5.94206035e-01 -1.04333028e-01 5.05965352e-01
9.10664916e-01 -1.46073544e+00 8.12546372e-01 5.07544994e-01
2.16800123e-01 6.73228323e-01 6.37611523e-02 -8.74065518e-01
-1.19287789e+00 -1.28574717e+00 3.25306475e-01 -4.60775882e-01
7.58559465e-01 -4.18430090e-01 -1.40876722e+00 8.72371078e-01
2.08200246e-01 4.09037292e-01 5.45086980e-01 4.36160892e-01
-8.27841580e-01 4.85395528e-02 -8.23068619e-01 5.82345605e-01
1.62617493e+00 -7.64741540e-01 -6.11841798e-01 4.16548043e-01
9.18361783e-01 -3.88018906e-01 -9.38091874e-01 5.44341385e-01
2.07818463e-01 -8.56788397e-01 1.00479972e+00 -9.41460133e-01
4.15642291e-01 -2.21664310e-01 1.08888701e-01 -1.46614575e+00
-5.70013940e-01 -2.37593397e-01 -6.68682531e-02 1.22472501e+00
2.40713239e-01 -8.62645388e-01 8.93288434e-01 -2.04183925e-02
-1.06379591e-01 -6.63778126e-01 -4.66784030e-01 -8.92768502e-01
1.77879155e-01 2.39092633e-01 5.19594789e-01 1.60222363e+00
-2.66105402e-02 9.56591725e-01 -1.46806344e-01 2.39772767e-01
7.25268543e-01 3.78766567e-01 9.28887367e-01 -1.58164632e+00
-1.65245026e-01 -4.33084995e-01 -4.92007107e-01 -1.21778917e+00
4.41529363e-01 -1.44943094e+00 -4.20183808e-01 -1.68599808e+00
2.87829131e-01 -3.72175157e-01 -4.31308895e-01 3.23104888e-01
-4.07360762e-01 -1.32622018e-01 -9.75140035e-02 1.62859797e-01
-7.38837898e-01 7.96695530e-01 1.68898749e+00 -4.07340229e-01
-1.14162460e-01 -4.40380484e-01 -8.24378014e-01 7.48721421e-01
6.70194209e-01 -3.00082058e-01 -8.83483112e-01 -4.66147602e-01
3.87568831e-01 -2.35261419e-03 4.42855984e-01 -7.68461943e-01
1.64705813e-01 -2.40612954e-01 3.11781049e-01 -5.06117344e-01
-1.65276065e-01 -7.85966933e-01 -1.39425978e-01 2.19257027e-01
-2.03042820e-01 8.98358598e-02 9.98813510e-02 1.21537936e+00
-2.71307290e-01 2.11770564e-01 8.44571054e-01 -3.19339454e-01
-1.01434588e+00 7.78494418e-01 4.39798862e-01 4.71841186e-01
8.47691059e-01 -2.96144575e-01 -5.67995131e-01 -2.68620849e-01
-8.64407837e-01 5.48708856e-01 7.14450657e-01 6.16098285e-01
5.47300220e-01 -1.66781497e+00 -5.25499105e-01 2.15037912e-01
5.57701707e-01 3.20896298e-01 3.85191828e-01 6.66647077e-01
-5.48981667e-01 1.37016863e-01 -4.81472909e-01 -7.37756729e-01
-1.01318228e+00 9.41623032e-01 1.96067959e-01 -6.40187263e-01
-8.80224228e-01 9.66339052e-01 7.13849664e-01 -1.14472735e+00
-1.14058994e-01 7.78617803e-03 -3.79707307e-01 5.01596294e-02
-6.68021366e-02 2.19135135e-01 -1.97628543e-01 -4.69942689e-01
-4.23947215e-01 7.78905869e-01 -3.66278470e-01 3.85012686e-01
1.30588305e+00 -1.80712178e-01 -2.04968959e-01 6.04734644e-02
1.14287341e+00 -2.51707643e-01 -1.33127916e+00 -6.26464546e-01
1.36085898e-01 -3.62934679e-01 -3.10624033e-01 -6.35236561e-01
-1.21850300e+00 1.07240665e+00 2.87885368e-01 1.33958846e-01
1.08480537e+00 1.04514413e-01 7.09155977e-01 3.57234299e-01
3.41801047e-01 -9.49212492e-01 7.62482822e-01 5.21892667e-01
9.61268842e-01 -1.49610841e+00 1.09351464e-01 -8.08457434e-01
-7.11152852e-01 1.20690048e+00 1.27002168e+00 -3.81595790e-01
5.67025959e-01 -4.26734775e-01 -1.41341716e-01 -6.16016626e-01
-4.09795374e-01 -1.85683623e-01 5.23021579e-01 9.25856233e-01
3.08873385e-01 -5.01508079e-02 -7.14082569e-02 3.18613261e-01
1.34410322e-01 -4.05449837e-01 1.67949244e-01 5.91010451e-01
-1.47532389e-01 -1.19532776e+00 1.98704004e-01 4.33904201e-01
2.12503374e-01 -1.74301028e-01 -8.04112494e-01 1.02666450e+00
-1.27591789e-01 5.25390208e-01 -1.33760870e-01 -3.40938687e-01
3.78159910e-01 -7.55070373e-02 5.66667795e-01 -8.09237480e-01
-1.67653918e-01 2.98248208e-03 5.69838323e-02 -5.28766036e-01
-3.79464775e-01 -3.21587473e-01 -1.17956698e+00 -1.57505780e-01
-3.55541259e-01 1.84866171e-02 -2.53134761e-02 8.57662141e-01
5.18946469e-01 7.15697467e-01 4.49981689e-01 -4.49842572e-01
-2.91051358e-01 -8.59476984e-01 -6.74391150e-01 7.18692183e-01
1.85007587e-01 -1.00434959e+00 -2.10688561e-01 -9.95573029e-02] | [7.292663097381592, 6.243254661560059] |
932e38b6-fbc8-4ac0-ae56-fbda8f5a79db | mmcr4nlp-multilingual-multiway-corpora | 1710.01025 | null | http://arxiv.org/abs/1710.01025v3 | http://arxiv.org/pdf/1710.01025v3.pdf | MMCR4NLP: Multilingual Multiway Corpora Repository for Natural Language Processing | Multilinguality is gradually becoming ubiquitous in the sense that more and
more researchers have successfully shown that using additional languages help
improve the results in many Natural Language Processing tasks. Multilingual
Multiway Corpora (MMC) contain the same sentence in multiple languages. Such
corpora have been primarily used for Multi-Source and Pivot Language Machine
Translation but are also useful for developing multilingual sequence taggers by
transfer learning. While these corpora are available, they are not organized
for multilingual experiments and researchers need to write boilerplate code
every time they want to use said corpora. Moreover, because there is no
official MMC collection it becomes difficult to compare against existing
approaches. As such we present our work on creating a unified and
systematically organized repository of MMC spanning a large number of
languages. We also provide training, development and test splits for corpora
where official splits are unavailable. We hope that this will help speed up the
pace of multilingual NLP research and ensure that NLP researchers obtain
results that are more trustable since they can be compared easily. We indicate
corpora sources, extraction procedures if any and relevant statistics. We also
make our collection public for research purposes. | ['Sadao Kurohashi', 'Raj Dabre'] | 2017-10-03 | null | null | null | null | ['multilingual-nlp'] | ['natural-language-processing'] | [-1.69916913e-01 -4.21023965e-01 -5.76121390e-01 -4.11512077e-01
-1.43816626e+00 -1.14680481e+00 5.57861567e-01 5.66314101e-01
-8.02968264e-01 1.28808129e+00 3.51598084e-01 -9.89246309e-01
2.38231644e-01 -3.58526260e-01 -6.71346307e-01 -2.64630079e-01
2.83424109e-01 7.79755294e-01 1.74462467e-01 -4.43188608e-01
2.62315750e-01 2.63800830e-01 -9.10893798e-01 3.09636414e-01
1.16263616e+00 -1.51511058e-01 5.74601352e-01 3.88558269e-01
-2.47981906e-01 3.28441560e-01 -4.53932971e-01 -8.21970105e-01
2.69796252e-01 -4.40894306e-01 -1.21786106e+00 -4.06537205e-01
3.21280718e-01 2.59057581e-01 4.76794541e-01 1.13993466e+00
3.57968330e-01 -2.24663809e-01 3.48530442e-01 -9.22924221e-01
-4.14581686e-01 9.76931751e-01 -4.87194151e-01 3.13322157e-01
4.58823919e-01 2.07979418e-02 1.16280901e+00 -7.24132359e-01
1.26540506e+00 1.04587162e+00 4.00663912e-01 3.35003108e-01
-1.09555233e+00 -6.30654335e-01 -8.45665485e-02 1.19946167e-01
-1.10677791e+00 -5.28421640e-01 4.73787755e-01 -4.47681576e-01
1.24467516e+00 2.41760254e-01 3.08786064e-01 1.08315790e+00
2.35793799e-01 5.52109420e-01 1.26683128e+00 -9.86134768e-01
-3.66009891e-01 4.94951725e-01 -6.86148480e-02 6.32193804e-01
3.66453707e-01 -4.16010141e-01 -3.16617072e-01 -3.14773284e-02
4.21087921e-01 -5.70597470e-01 -1.95789009e-01 -1.25786066e-01
-1.61430824e+00 1.01880145e+00 -1.69246510e-01 1.17218518e+00
-1.17047064e-01 -3.03920984e-01 6.29484534e-01 6.92838490e-01
5.94275594e-01 6.32769465e-01 -8.78907561e-01 -4.96587604e-01
-8.66241038e-01 1.53588101e-01 9.52189386e-01 9.10555840e-01
7.59953797e-01 -2.10955307e-01 6.50613666e-01 1.21796405e+00
4.02498767e-02 6.54903889e-01 6.22554719e-01 -7.67200291e-01
9.27701652e-01 4.27294731e-01 9.31997076e-02 -9.08265591e-01
-2.88188368e-01 -1.17648788e-01 -2.95760006e-01 -1.92199871e-01
6.36799872e-01 -1.93006545e-01 -3.57981741e-01 1.64176798e+00
8.35527331e-02 -7.94818342e-01 3.69761050e-01 6.38892055e-01
3.41001093e-01 8.39050710e-01 1.24752112e-01 -2.57862866e-01
1.44836318e+00 -7.62082040e-01 -6.12932920e-01 -2.35602394e-01
1.10686088e+00 -1.60573506e+00 1.14110482e+00 3.30766529e-01
-1.15328753e+00 -2.77737617e-01 -9.71495509e-01 -3.10412884e-01
-5.77779889e-01 -5.03934808e-02 6.66338325e-01 6.54763758e-01
-1.07159209e+00 3.22640568e-01 -8.33591700e-01 -8.35750997e-01
-1.65388420e-01 2.82767508e-02 -7.56061256e-01 -3.49230587e-01
-1.28202927e+00 1.49348700e+00 3.90616715e-01 -1.32967368e-01
-2.13438332e-01 -2.19539940e-01 -8.46620083e-01 -5.49291193e-01
7.83793926e-02 -1.61695763e-01 1.22758126e+00 -9.37522590e-01
-9.47433650e-01 1.25632989e+00 -2.77961105e-01 -2.47383997e-01
4.27417129e-01 -7.61701167e-02 -6.01051807e-01 -6.45994097e-02
6.74138904e-01 6.65960431e-01 4.75695208e-02 -8.66261005e-01
-8.84602189e-01 -2.66697526e-01 -2.40500942e-01 1.86498329e-01
-2.51995534e-01 8.24786723e-01 -4.62573916e-01 -6.83737218e-01
-6.58042952e-02 -1.09025109e+00 -1.68909758e-01 -6.30227268e-01
-2.28791311e-02 -2.32867017e-01 5.38008869e-01 -1.22616816e+00
1.04657340e+00 -1.64761662e+00 1.07176103e-01 -1.65537462e-01
-2.24336803e-01 2.98127800e-01 -2.71375924e-01 8.58367205e-01
2.09294215e-01 5.72042823e-01 -6.73115999e-02 -1.36041254e-01
-7.34654814e-02 3.84397417e-01 -1.49246559e-01 4.55549061e-01
1.73695937e-01 7.47791529e-01 -9.66327310e-01 -6.65394962e-01
3.81121337e-02 2.80817866e-01 -3.26967388e-01 -2.72183388e-01
-5.33454269e-02 5.88350892e-01 -2.45847002e-01 5.71612954e-01
2.85786718e-01 1.34619594e-01 5.16148269e-01 4.26499903e-01
-6.83141768e-01 8.23634267e-01 -7.49708056e-01 1.81607735e+00
-8.21710289e-01 8.02103162e-01 -6.20382950e-02 -7.86677897e-01
8.89821112e-01 5.80771625e-01 2.02654898e-01 -5.33995032e-01
-6.40433803e-02 8.69455993e-01 3.91344905e-01 -2.12456807e-01
8.54619205e-01 -2.92402744e-01 -3.48337740e-01 6.30645514e-01
5.61244674e-02 -2.73142964e-01 8.28912079e-01 1.74200013e-01
6.62046075e-01 1.15495406e-01 5.04902065e-01 -6.55174673e-01
5.55130661e-01 5.93809724e-01 8.61483216e-01 2.60224164e-01
-1.28065795e-01 2.39806667e-01 3.64504308e-01 -3.79614413e-01
-1.39341402e+00 -8.21671426e-01 -4.89562273e-01 1.05853033e+00
-4.56474692e-01 -3.54276806e-01 -4.78171259e-01 -4.63865608e-01
-4.10790741e-01 6.65394843e-01 -6.59811422e-02 5.29187977e-01
-8.66200924e-01 -5.61764181e-01 6.39386714e-01 1.83352724e-01
7.84516260e-02 -1.01508594e+00 -1.40591457e-01 3.98644537e-01
-6.96055412e-01 -1.25015128e+00 -4.96049494e-01 3.82446237e-02
-6.99800491e-01 -1.17462206e+00 -7.91413784e-01 -1.31050825e+00
3.53952885e-01 1.61329448e-01 9.70303893e-01 -6.07356690e-02
9.58756264e-03 5.76397330e-02 -5.68919122e-01 -4.46629345e-01
-1.02643418e+00 5.03784835e-01 1.77138239e-01 -6.21548057e-01
6.50300503e-01 -3.26501906e-01 1.52267903e-01 1.00367814e-01
-8.03518772e-01 1.27053380e-01 4.60230172e-01 5.58022499e-01
2.82364726e-01 -3.66533935e-01 6.64378166e-01 -1.13300943e+00
7.74429321e-01 -3.51276845e-01 -6.55349076e-01 4.58914787e-01
-6.42149031e-01 1.25966191e-01 5.65482497e-01 2.39142571e-02
-9.39001143e-01 -2.99181968e-01 -2.96404123e-01 2.88080126e-01
-2.33850315e-01 7.81257749e-01 -2.63870806e-02 1.41309753e-01
5.96162200e-01 6.94587901e-02 -2.78524846e-01 -6.55102611e-01
2.12231934e-01 9.38253284e-01 3.64991605e-01 -7.70428538e-01
7.51261890e-01 -1.96556777e-01 -5.58087528e-01 -8.28112364e-01
-3.33424658e-01 -7.66024888e-01 -7.87682414e-01 -3.39995362e-02
9.06599760e-01 -1.01379836e+00 -6.98503032e-02 2.46977821e-01
-1.36085820e+00 -3.00886244e-01 3.53342026e-01 8.09885383e-01
-2.54808187e-01 2.56582499e-01 -8.43725085e-01 -2.81499147e-01
-1.39323696e-01 -1.48145628e+00 8.08137774e-01 -8.68878514e-02
-5.32356143e-01 -1.43796158e+00 4.55576509e-01 6.14783406e-01
2.18042165e-01 1.46221757e-01 1.17777514e+00 -6.43585443e-01
-2.12697342e-01 -7.79151171e-02 8.81652310e-02 1.39225066e-01
2.55267859e-01 2.50029624e-01 -4.02075678e-01 -4.08632129e-01
-3.88624728e-01 -4.83500123e-01 2.85574585e-01 -2.36404035e-03
1.01672828e-01 -4.05108094e-01 -2.16296822e-01 2.10912861e-02
1.44425094e+00 2.08795235e-01 3.96421254e-01 6.63748264e-01
5.09880602e-01 8.71859431e-01 4.24831182e-01 -2.25132406e-01
6.83333695e-01 6.59139156e-01 -5.11867166e-01 -3.15944962e-02
1.11467861e-01 -2.75591016e-01 5.76680005e-01 1.77924025e+00
-5.85167259e-02 -5.87978326e-02 -1.43114769e+00 8.55961084e-01
-1.61648178e+00 -8.34846556e-01 -3.79058450e-01 2.19772410e+00
1.32387924e+00 -4.26962972e-02 6.14359863e-02 -1.50557026e-01
7.15719104e-01 -8.30742493e-02 9.25770774e-02 -7.86397696e-01
-2.24208906e-01 3.52738202e-01 5.63640356e-01 9.40295517e-01
-8.65399897e-01 1.47447896e+00 5.99625444e+00 6.09867454e-01
-1.19944239e+00 3.53179216e-01 3.84648710e-01 2.24536791e-01
-5.63127041e-01 3.02756429e-01 -9.55894351e-01 3.00043613e-01
1.15523207e+00 -5.55088580e-01 3.78243089e-01 5.96003354e-01
3.52290094e-01 -2.61661202e-01 -9.48645830e-01 7.15740323e-01
-1.31152302e-01 -1.22074139e+00 -3.30161452e-01 1.67207643e-01
8.22529316e-01 6.04655445e-01 -2.32602671e-01 3.61371160e-01
8.20481241e-01 -5.88809550e-01 6.25943363e-01 -6.90942034e-02
8.38647187e-01 -7.21321106e-01 7.35284328e-01 4.86501306e-01
-1.02379131e+00 5.55956006e-01 -5.51320255e-01 -8.56498480e-02
2.90892065e-01 4.95320857e-01 -1.02662015e+00 7.48219192e-01
4.45352137e-01 5.01835823e-01 -6.96172237e-01 8.97432923e-01
-4.24057275e-01 7.50943661e-01 -1.45870969e-01 -2.42031485e-01
4.54360753e-01 -2.83933550e-01 5.29976010e-01 1.39551151e+00
2.63427079e-01 -5.26477396e-01 4.22631472e-01 2.74008125e-01
-1.54346049e-01 9.83771443e-01 -9.05920982e-01 -4.06875998e-01
4.08219099e-01 1.16920912e+00 -1.00646698e+00 -9.52486917e-02
-7.75109708e-01 8.67889106e-01 5.39392233e-01 2.03021631e-01
-3.40990722e-01 -5.98306298e-01 6.33211255e-01 -9.82315168e-02
-6.24256348e-03 -6.38760269e-01 -5.94045110e-02 -1.48399127e+00
3.94641422e-02 -1.29703283e+00 3.81756753e-01 -5.28160870e-01
-1.17225552e+00 8.13407660e-01 -2.20892690e-02 -1.00351298e+00
-7.30105162e-01 -5.44402421e-01 1.20234042e-02 1.17682064e+00
-1.31004345e+00 -1.11387539e+00 5.82275987e-01 2.52226949e-01
5.73764980e-01 -3.07269901e-01 1.05893993e+00 4.44797039e-01
-4.26255465e-01 5.34004569e-01 2.93166906e-01 5.43013871e-01
1.17494702e+00 -9.78810310e-01 3.62789303e-01 1.08679044e+00
7.32948303e-01 1.03727543e+00 6.78660572e-01 -7.86178827e-01
-1.14676011e+00 -8.74472022e-01 1.75970542e+00 -5.19369602e-01
1.20828426e+00 -4.70913738e-01 -6.88450098e-01 9.28586602e-01
6.76035523e-01 -5.95838547e-01 9.00221884e-01 6.66405857e-01
-3.61047417e-01 4.71971706e-02 -7.83466637e-01 6.06314898e-01
5.28904378e-01 -6.20948970e-01 -7.86690831e-01 6.20626152e-01
6.21876299e-01 -1.44986570e-01 -9.37633097e-01 -2.23458931e-02
3.40980649e-01 -5.08920550e-01 2.36611813e-01 -6.31109536e-01
5.14379621e-01 -2.80399680e-01 -1.89722374e-01 -1.48059630e+00
-3.42923142e-02 -5.33829391e-01 1.08406627e+00 1.44346809e+00
1.09523129e+00 -9.21125650e-01 2.60327995e-01 3.11331213e-01
-2.19225466e-01 -1.97328225e-01 -8.82521808e-01 -9.93951261e-01
7.32112825e-01 -5.98469853e-01 3.13023001e-01 1.51327491e+00
4.81943876e-01 6.62933409e-01 -1.69468805e-01 -2.20096529e-01
3.16697419e-01 1.14272438e-01 7.32793570e-01 -1.08091009e+00
-2.45024979e-01 -3.85123163e-01 -2.56384999e-01 -4.54627365e-01
2.00762883e-01 -1.37960982e+00 1.16126209e-01 -1.55485177e+00
3.39731842e-01 -6.30841732e-01 6.80801570e-02 7.84681141e-01
-1.59197092e-01 3.84272069e-01 2.15786174e-01 4.83372927e-01
-2.00724155e-01 -3.01773939e-02 1.07130206e+00 5.49573638e-02
-1.73599392e-01 -4.51166749e-01 -8.60449255e-01 4.93225992e-01
9.68095303e-01 -6.94782376e-01 -7.53526688e-02 -7.23169506e-01
2.40726173e-01 -1.69133414e-02 -2.13838845e-01 -8.46277893e-01
7.28974165e-03 -4.00033236e-01 1.18249446e-01 -3.54276150e-01
-2.61289537e-01 -4.56089795e-01 1.44612446e-01 3.87138784e-01
-1.73884600e-01 5.81913471e-01 3.90136272e-01 -1.79966003e-01
-3.82177114e-01 -4.50475901e-01 5.85295975e-01 -2.75498956e-01
-6.58027053e-01 3.94989401e-02 -5.61724961e-01 2.11696491e-01
8.63485694e-01 3.06549609e-01 -3.41582179e-01 -3.91359568e-01
-2.52864361e-01 1.96757257e-01 8.69457901e-01 6.51600957e-01
-8.91534612e-02 -1.05342054e+00 -1.07265449e+00 -2.27638874e-02
2.13715836e-01 -5.04511476e-01 -3.70441228e-01 8.03723991e-01
-9.26207662e-01 9.17247713e-01 -3.97583097e-01 -2.69534677e-01
-1.24314129e+00 3.96404237e-01 -2.81061888e-01 -5.62288344e-01
-5.04757762e-01 4.59739625e-01 -2.72811025e-01 -9.11532462e-01
-2.62715220e-01 3.23401019e-02 -9.62220877e-03 -3.62202083e-03
3.78390849e-01 -8.72952417e-02 1.71052098e-01 -1.00356507e+00
-3.03303659e-01 6.18253760e-02 -4.43232894e-01 -7.13360906e-01
1.51718450e+00 -1.05046049e-01 -4.46871638e-01 8.54272246e-01
1.21353233e+00 6.50900364e-01 -4.29483265e-01 -3.93399410e-02
4.84721154e-01 -2.66750902e-01 -3.48483413e-01 -7.89609790e-01
-5.04037082e-01 7.70624936e-01 9.81205925e-02 -1.65410861e-02
7.13428438e-01 5.51109500e-02 8.50814402e-01 4.29573506e-01
7.99173832e-01 -1.25865185e+00 -5.91193318e-01 8.07219982e-01
7.14660406e-01 -1.15160477e+00 -7.60674179e-02 -2.89282829e-01
-5.24944067e-01 9.33341146e-01 1.62168875e-01 3.32904398e-01
1.91719994e-01 3.70991111e-01 5.91351092e-01 2.01544538e-01
-6.47009254e-01 -2.09395155e-01 -7.21395612e-02 6.00215018e-01
1.06213367e+00 1.72245070e-01 -1.20343626e+00 1.70131326e-01
-6.40951514e-01 -2.99061775e-01 6.47535026e-01 8.87948275e-01
-3.37756753e-01 -2.28556466e+00 -3.71692121e-01 9.77543071e-02
-8.08552742e-01 -5.32723546e-01 -4.13368136e-01 1.25808704e+00
-1.23499922e-01 1.03121698e+00 -2.65225381e-01 -8.05908591e-02
-9.47744846e-02 4.02894884e-01 6.76751316e-01 -7.60778785e-01
-4.50949907e-01 2.91497767e-01 5.46409369e-01 3.02389059e-02
-4.86614645e-01 -1.00879180e+00 -9.81852472e-01 -6.08346403e-01
-1.40995964e-01 7.92994678e-01 9.44598496e-01 9.94105101e-01
-6.24067597e-02 -1.23244390e-01 2.80419111e-01 -3.84144306e-01
-7.80671760e-02 -1.08533728e+00 -2.19768912e-01 1.23859309e-01
-2.81135082e-01 -2.26758122e-01 4.06855978e-02 4.28027302e-01] | [10.62663745880127, 10.113409996032715] |
ccc42f46-a67c-4a68-be5f-557d660d2d9a | improving-partition-block-based-acoustic-echo | 2008.03944 | null | https://arxiv.org/abs/2008.03944v1 | https://arxiv.org/pdf/2008.03944v1.pdf | improving partition-block-based acoustic echo canceler in under-modeling scenarios | Recently, a partitioned-block-based frequency-domain Kalman filter (PFKF) has been proposed for acoustic echo cancellation. Compared with the normal frequency-domain Kalman filter, the PFKF utilizes the partitioned-block structure, resulting in both fast convergence and low time-latency. We present an analysis of the steady-state behavior of the PFKF and found that it suffers from a biased steady-state solution when the filter is of deficient length. Accordingly, we propose an effective modification that has the benefit of the guaranteed optimal steady-state behavior. Simulations are conducted to validate the improved performance of the proposed method. | ['Jing Lu', 'Wenzhi Fan'] | 2020-08-10 | null | null | null | null | ['acoustic-echo-cancellation', 'acoustic-echo-cancellation'] | ['medical', 'speech'] | [-1.29772887e-01 -3.66947711e-01 3.87602955e-01 -1.41668737e-01
-7.02189803e-01 -2.94423789e-01 1.75805658e-01 -4.60266650e-01
-2.50765234e-01 7.14241147e-01 2.67330140e-01 -5.77458203e-01
-4.45952684e-01 -1.64900944e-01 -3.48960817e-01 -1.04058194e+00
-3.63182634e-01 -4.84903097e-01 2.33675644e-01 2.42813662e-01
1.59885988e-01 3.41370434e-01 -1.40425467e+00 -4.99074638e-01
9.12650585e-01 8.04639101e-01 2.55267829e-01 1.04076111e+00
3.68930936e-01 1.88414976e-01 -7.36762941e-01 2.52196342e-01
3.65749747e-01 -4.95547503e-01 -1.50283173e-01 -1.97817191e-01
3.48807126e-01 -6.47159517e-01 -6.46381319e-01 1.05892527e+00
8.69393408e-01 4.52135742e-01 3.14287543e-01 -8.36311102e-01
4.13395315e-01 1.83377758e-01 -3.44017625e-01 3.57899785e-01
3.84756118e-01 -2.30033308e-01 4.20147657e-01 -9.28884506e-01
4.92070839e-02 1.25184715e+00 1.08114254e+00 3.46477419e-01
-1.01051867e+00 -1.17866564e+00 -1.91858307e-01 -2.37766907e-01
-1.51470137e+00 -7.83770740e-01 5.95664084e-01 -2.57295907e-01
9.27187979e-01 2.11988747e-01 8.50894272e-01 2.04482555e-01
7.90826738e-01 3.68384212e-01 1.25922036e+00 -4.94049937e-01
1.39346093e-01 -4.74406064e-01 5.02596498e-01 4.79920417e-01
4.78179991e-01 8.68582249e-01 -6.83866680e-01 -6.57662988e-01
7.98862934e-01 -3.81413221e-01 -6.33223295e-01 -9.99276042e-02
-7.97471821e-01 3.12241405e-01 -8.04497078e-02 1.35160252e-01
-3.03863823e-01 3.47788334e-01 2.25675806e-01 4.95822132e-01
4.99404222e-01 1.29949957e-01 -2.04577491e-01 -3.80731076e-01
-1.19702244e+00 6.05062068e-01 1.01358664e+00 6.96431577e-01
4.21455324e-01 8.21656346e-01 3.56555551e-01 3.81029606e-01
8.75313997e-01 1.24443626e+00 2.30148360e-01 -7.35909402e-01
9.04202610e-02 -3.07547837e-01 5.62197983e-01 -7.52519608e-01
-5.92511535e-01 -7.89614737e-01 -7.94920146e-01 1.08465150e-01
3.68268847e-01 -7.07565188e-01 -9.42337990e-01 1.39015937e+00
3.87878656e-01 8.16917360e-01 4.58563656e-01 9.05578017e-01
5.94788432e-01 1.17565167e+00 -2.55437583e-01 -8.76033425e-01
8.09903264e-01 -4.71282005e-01 -1.61747861e+00 -1.62049472e-01
2.17408285e-01 -1.06299019e+00 2.24470869e-01 4.84481424e-01
-9.72668231e-01 -6.51668191e-01 -1.36688590e+00 6.73142433e-01
3.25711608e-01 -1.37124747e-01 1.85962155e-01 1.14439619e+00
-1.13745236e+00 3.90425920e-01 -9.32518542e-01 -3.79696414e-02
-7.85557866e-01 3.66174579e-01 7.30926218e-03 3.69700551e-01
-1.25119960e+00 7.16488540e-01 2.24758372e-01 6.69895947e-01
-7.32778549e-01 -7.33742893e-01 -5.88349462e-01 4.17067632e-02
7.28656277e-02 -2.93019950e-01 1.69593108e+00 -4.28807139e-01
-1.91283846e+00 -3.07506114e-01 -6.38753951e-01 -3.76810312e-01
5.28803505e-02 -6.50342584e-01 -9.61855650e-01 1.45924278e-02
-4.55453306e-01 -3.43709826e-01 1.28355765e+00 -9.88325059e-01
-8.77565920e-01 1.04228444e-01 -4.84755546e-01 1.72332108e-01
5.89848869e-02 -3.17359567e-01 -1.55624310e-02 -6.74193501e-01
7.24804640e-01 -1.11615860e+00 -3.58367980e-01 -5.74414551e-01
1.49235234e-01 2.51365662e-01 8.82276595e-01 -6.32363915e-01
2.04880404e+00 -2.52322698e+00 -3.23751807e-01 4.50916559e-01
-1.99258879e-01 5.55692017e-01 3.60530704e-01 9.66403484e-01
2.50023068e-03 -4.07099903e-01 -3.47253717e-02 1.22567743e-01
-3.67874622e-01 4.89595383e-02 -4.22155768e-01 9.49063003e-01
-3.27335924e-01 -9.18750744e-03 -9.59929109e-01 -6.47643879e-02
3.71517718e-01 3.01835328e-01 -5.43740094e-01 3.84948254e-01
7.29119182e-01 4.69510019e-01 -4.71776575e-01 4.57446903e-01
1.12200511e+00 4.02476430e-01 2.36162990e-01 -4.26273316e-01
-7.27860987e-01 3.96562070e-01 -1.80452871e+00 1.29516184e+00
-2.71501034e-01 7.26399660e-01 9.41039741e-01 -4.35997009e-01
1.02116299e+00 8.00092936e-01 4.14158136e-01 -3.19421947e-01
2.97589228e-02 4.31774467e-01 3.20836216e-01 -2.75582403e-01
7.05447435e-01 -4.10930723e-01 1.10625871e-01 3.18667471e-01
-1.25630811e-01 -2.66738027e-01 -9.42213610e-02 1.74149498e-01
1.03913677e+00 -1.72212794e-01 5.21461248e-01 -8.73885870e-01
6.38099015e-01 -3.41633409e-01 8.40294063e-01 7.81692386e-01
-4.01371151e-01 -1.33659421e-02 -3.93096983e-01 8.47748294e-03
-5.04049897e-01 -9.37945426e-01 -3.85046870e-01 3.80645722e-01
4.05756414e-01 -4.91145581e-01 -4.19632316e-01 2.09153682e-01
1.67785764e-01 3.28593343e-01 3.08665913e-02 -1.03228033e-01
-7.04327166e-01 -5.43426931e-01 6.63791478e-01 2.08036810e-01
3.62255514e-01 -2.30395243e-01 -7.49863505e-01 6.81121767e-01
1.40955867e-02 -7.78783262e-01 -4.00863647e-01 1.98617041e-01
-1.18185031e+00 -6.93858266e-01 -5.01986086e-01 -6.19453728e-01
5.32215476e-01 6.69441700e-01 5.28810561e-01 -7.04646185e-02
4.07941341e-01 5.98898292e-01 -3.44322443e-01 -5.24415672e-01
-3.81604195e-01 -4.90870655e-01 5.86118042e-01 4.75351028e-02
-1.06628694e-01 -2.29367614e-01 -7.18629658e-01 5.18476129e-01
-7.07034290e-01 -2.76421875e-01 5.14552593e-01 1.14072454e+00
4.72796410e-02 2.88226873e-01 8.03795993e-01 -3.79842073e-01
9.54962015e-01 2.45528994e-03 -1.12130725e+00 -1.41634181e-01
-9.24645901e-01 -1.70336708e-01 3.74653459e-01 -3.00809503e-01
-1.64737380e+00 1.71182349e-01 -3.60215306e-01 -1.39303312e-01
3.36632371e-01 7.51420081e-01 6.82937875e-02 -4.11518812e-01
2.88279921e-01 4.08438772e-01 3.78649116e-01 -5.80641568e-01
1.68429583e-01 8.63504767e-01 7.39835441e-01 -1.69780076e-01
1.04933751e+00 1.97610199e-01 -8.26506969e-03 -1.26187015e+00
-1.05682172e-01 -9.40596163e-01 -1.55856058e-01 -5.48893809e-01
1.81435108e-01 -1.28963017e+00 -7.17958093e-01 9.59514260e-01
-8.11803579e-01 1.09007902e-01 2.26142481e-01 1.33026481e+00
-2.17539623e-01 6.33579731e-01 -7.58371949e-01 -1.54460669e+00
-4.29867297e-01 -7.95566022e-01 5.47135770e-01 3.66217971e-01
-1.27086103e-01 -7.66976833e-01 4.53993291e-01 -4.90752071e-01
4.87961531e-01 -1.47827953e-01 1.43426150e-01 -7.06763864e-02
-4.20388371e-01 -2.48284370e-01 1.40527442e-01 2.62542903e-01
4.97219041e-02 9.52183381e-02 -8.85948598e-01 -9.44243014e-01
3.30146134e-01 3.71458769e-01 4.35295999e-01 7.57285416e-01
-2.21733022e-02 -1.82525426e-01 -4.93592530e-01 6.87949240e-01
1.40365016e+00 7.83482671e-01 4.03685629e-01 -1.02035655e-02
-3.05905547e-02 1.53764188e-01 1.17544055e+00 5.88600695e-01
-1.89090908e-01 3.09818178e-01 -3.19416053e-04 -5.27582839e-02
1.15994535e-01 -1.21984147e-01 4.49333370e-01 1.51630318e+00
2.52218604e-01 -2.71872312e-01 -5.40220976e-01 6.18892431e-01
-1.76573169e+00 -6.89339697e-01 -4.30304408e-01 2.25017285e+00
4.87873316e-01 1.22152613e-02 -2.69856513e-01 4.10259545e-01
5.18456936e-01 2.96215892e-01 -1.74322844e-01 -6.43810809e-01
3.54718328e-01 1.16390607e-03 9.94832993e-01 8.59420419e-01
-1.04458320e+00 4.72630024e-01 8.20065975e+00 7.90584326e-01
-1.08789504e+00 -8.67688805e-02 -6.43337607e-01 3.47932458e-01
1.11940823e-01 1.95790529e-01 -1.05139780e+00 2.84286141e-01
1.35896921e+00 -6.76490963e-01 -1.47107337e-02 3.77552539e-01
6.20033205e-01 -3.56709957e-01 -2.78807580e-01 8.11185181e-01
-3.70663762e-01 -6.68619275e-01 -5.40586293e-01 -8.45582411e-03
5.05638063e-01 -4.42688435e-01 -2.70327747e-01 9.14280564e-02
-1.44073471e-01 -2.25081876e-01 4.99462157e-01 5.40732741e-01
5.40221453e-01 -8.43229353e-01 6.59184754e-01 4.59355175e-01
-1.63639271e+00 -2.08287328e-01 -4.18026716e-01 -6.67243421e-01
6.72325790e-01 8.58124375e-01 -8.01697195e-01 7.68736839e-01
5.09253144e-01 4.75173205e-01 3.71109605e-01 1.58474040e+00
6.99626431e-02 1.14924538e+00 -5.25878966e-01 -3.27909477e-02
3.00281942e-01 -2.84892201e-01 9.87396121e-01 1.19237471e+00
7.50311255e-01 6.54688120e-01 1.71670213e-01 -4.39212769e-02
7.98985064e-01 -1.78425938e-01 -4.38818723e-01 1.22632019e-01
7.29030788e-01 6.99767411e-01 -1.81510717e-01 -3.41773897e-01
-5.64162850e-01 7.61536956e-02 -5.64865053e-01 5.82652748e-01
-4.00500119e-01 -7.18250751e-01 9.24963355e-01 -5.93272150e-02
4.55971599e-01 -6.94295228e-01 2.89035905e-02 -6.89046323e-01
-4.51930970e-01 -7.00351775e-01 1.15517698e-01 -5.02526045e-01
-8.20441782e-01 5.44138908e-01 3.08122516e-01 -1.84384251e+00
-5.20017922e-01 -1.29131898e-01 -2.73820758e-01 1.10456944e+00
-1.35878003e+00 -4.80432183e-01 5.01818508e-02 4.69784200e-01
2.86226839e-01 -2.91388165e-02 7.40186989e-01 5.65451264e-01
-1.90249875e-01 4.81914818e-01 7.95010746e-01 -2.25685671e-01
7.75724113e-01 -8.35730493e-01 3.37901503e-01 1.43572092e+00
-5.66512764e-01 1.09852469e+00 1.44755399e+00 -1.04616582e+00
-1.66543138e+00 -6.91217124e-01 8.28153431e-01 3.60411406e-01
5.68864107e-01 -5.70002757e-02 -9.87566352e-01 2.78273970e-01
2.57901609e-01 -2.99657524e-01 4.49567944e-01 -1.14549853e-01
2.18485937e-01 -4.66348112e-01 -8.63040507e-01 1.92369297e-01
5.23277223e-01 -3.95560056e-01 -8.61356914e-01 -2.71523505e-01
4.21794504e-01 -8.13285828e-01 -8.29116821e-01 6.85325682e-01
1.07191491e+00 -9.58736062e-01 6.82727337e-01 2.25934237e-01
-7.81252861e-01 -8.57952178e-01 9.24609974e-02 -1.51856411e+00
-6.08223140e-01 -1.13814068e+00 -1.88962430e-01 8.98704112e-01
2.07404241e-01 -1.03588355e+00 4.11520332e-01 1.44526243e-01
-5.07555485e-01 -1.39410660e-01 -1.18327200e+00 -1.08212316e+00
-5.34564793e-01 -4.64405656e-01 8.54064450e-02 1.66847929e-01
2.40958884e-01 1.86071217e-01 -9.11033273e-01 8.62813056e-01
7.69621432e-01 1.07392304e-01 7.79775381e-01 -9.55333650e-01
-1.97848782e-01 2.39834443e-01 -2.11650938e-01 -1.78437507e+00
-1.18920423e-01 2.07279354e-01 6.49192691e-01 -1.27462721e+00
-4.97406185e-01 -2.84513652e-01 -3.49358946e-01 -2.75689691e-01
-3.17740291e-01 -7.67401233e-02 -3.43343727e-02 1.05835572e-02
1.30585432e-01 3.40376645e-01 1.04621017e+00 2.45343775e-01
-4.10533965e-01 5.46037138e-01 1.76035836e-01 5.16065717e-01
3.60386342e-01 -6.17010832e-01 -6.52743876e-01 -1.61806002e-01
-4.67000976e-02 7.90226758e-01 -1.75154179e-01 -1.25909209e+00
5.34012198e-01 1.29732206e-01 -2.65761949e-02 -1.26436257e+00
5.41778743e-01 -1.10334957e+00 7.75572956e-01 1.11470664e+00
1.83808312e-01 1.99023739e-01 3.90832126e-01 1.03223515e+00
-6.09447002e-01 -1.10844195e-01 8.07205915e-01 3.59707028e-01
-7.73382425e-01 -2.34635696e-01 -1.14097881e+00 -4.54796702e-01
6.76196754e-01 -4.31196302e-01 8.85632914e-03 -6.93499148e-01
-4.96209651e-01 6.86540082e-02 -1.74988061e-02 1.27905561e-03
7.19834745e-01 -1.02789807e+00 -5.50207734e-01 6.21957898e-01
-4.44882721e-01 -5.84842622e-01 5.57146668e-01 8.30495834e-01
-6.12157524e-01 7.29511023e-01 1.06458664e-01 -7.69547343e-01
-1.43641996e+00 1.15820542e-01 3.83533180e-01 1.81226619e-02
-6.50502503e-01 6.62204325e-01 1.01111308e-01 -6.51627481e-02
3.11709702e-01 -4.70536977e-01 7.32704177e-02 -1.87855601e-01
7.27903008e-01 6.23151362e-01 1.38820991e-01 -5.40916443e-01
-4.47423339e-01 8.46559584e-01 4.23877537e-01 -4.57228214e-01
7.11266756e-01 -6.36917055e-01 -6.98173344e-02 4.80248034e-01
7.98421502e-01 4.25784558e-01 -1.23515320e+00 -1.93228975e-01
-3.34839284e-01 -8.70634913e-01 3.48134428e-01 -4.80790377e-01
-7.38874018e-01 4.30428416e-01 9.68694806e-01 5.30787259e-02
1.35149872e+00 -9.50480461e-01 8.50512087e-01 1.97188511e-01
6.03251576e-01 -7.98446238e-01 -4.90750343e-01 8.02510321e-01
4.13775563e-01 -5.03104210e-01 4.56184775e-01 -4.59335268e-01
-1.49492715e-02 1.03690016e+00 3.81385118e-01 -2.45945439e-01
1.38089383e+00 7.31573045e-01 6.44989371e-01 3.23709100e-01
-8.74534309e-01 1.22601785e-01 3.45039636e-01 4.43785936e-01
4.42152888e-01 3.37745249e-02 -1.03000236e+00 2.90684164e-01
2.99320947e-02 -5.28074577e-02 5.87841094e-01 1.32512188e+00
-8.38426769e-01 -9.16381240e-01 -9.36989844e-01 3.27446163e-01
-6.00285292e-01 7.36506134e-02 2.99110413e-01 7.19164252e-01
-6.70837402e-01 1.37721539e+00 -5.86111918e-02 -4.95026320e-01
5.06146014e-01 5.49893454e-02 1.31564453e-01 -1.01545483e-01
-6.87983871e-01 1.08144331e+00 2.52591759e-01 -5.35182357e-01
-7.31459022e-01 -5.24743915e-01 -1.22620678e+00 -1.99771687e-01
-9.59087253e-01 9.14281547e-01 6.66458547e-01 7.57126987e-01
2.65448004e-01 5.71860313e-01 7.09926903e-01 -6.47805214e-01
-7.63778806e-01 -1.04909301e+00 -1.00844765e+00 -4.03403461e-01
9.40159321e-01 -6.83444500e-01 -7.66862929e-01 -2.75111616e-01] | [15.123194694519043, 5.745194435119629] |
98b43b5c-dae9-44e1-b3e5-9173dd83f28a | inferring-disease-correlation-from-healthcare | 1510.03051 | null | http://arxiv.org/abs/1510.03051v1 | http://arxiv.org/pdf/1510.03051v1.pdf | Inferring disease correlation from healthcare data | Electronic Health Records maintained in health care settings are a potential
source of substantial clinical knowledge. The massive volume of data,
unstructured nature of records and obligatory requirement of domain
acquaintance together pose a challenge in knowledge extraction from it. The aim
of this study is to overcome this challenge with a methodical analysis,
abstraction and summarization of such data. This is an attempt to explain
clinical observations through bio-medical and genomic data. Discharge summaries
of obesity patients were processed to extract coherent patterns. This was
supported by Machine Learning and Natural Language Processing based
technologies and concept mapping tool along with biomedical, clinical and
genomic knowledge bases. Semantic relations between diseases were extracted and
filtered through Chi square test to remove spurious relations. The remaining
relations were validated against biomedical literature and gene interaction
networks. A collection of binary relations of diseases was derived from the
data. One set implied co-morbidity while the other set contained diseases which
are risk factors of others. Validation against bio-medical literature increased
the prospect of correlation between diseases. Gene interaction network revealed
that the diseases are related and their corresponding genes are in close
proximity. Conclusion: This study focuses on deducing meaningful relations
between diseases from discharge summaries. For analytical purpose, the scope
has been limited to a few common, well-researched diseases. It can be extended
to incorporate relatively unknown, complex diseases and discover new traits to
help in clinical assessments. | ['Anand Ashish', 'Priyadarshini Gargi'] | 2015-10-11 | null | null | null | null | ['clinical-knowledge'] | ['miscellaneous'] | [ 4.13075894e-01 4.19639081e-01 -4.63561416e-01 -3.41193497e-01
-3.54745120e-01 -3.36431265e-01 1.75840124e-01 1.12930501e+00
-5.45502082e-02 1.22814286e+00 6.29679084e-01 -3.14749777e-01
-1.26025403e+00 -8.95917892e-01 -1.41193494e-01 -4.95781392e-01
-5.34895539e-01 6.33458674e-01 -3.76682073e-01 -1.81680262e-01
1.11782707e-01 4.03935462e-01 -1.35175002e+00 5.84291935e-01
1.08644152e+00 4.34618145e-01 8.21585651e-04 5.79751492e-01
-2.39340320e-01 6.58572614e-01 -4.73904371e-01 -4.52455074e-01
-1.50146082e-01 -5.67394733e-01 -7.56737292e-01 -1.75093174e-01
-3.62137884e-01 2.33038202e-01 2.34141052e-02 9.05624509e-01
5.83498180e-01 -2.22777933e-01 7.21551001e-01 -1.19324267e+00
-7.72742987e-01 7.16065645e-01 -6.22137859e-02 2.16959015e-01
6.47482395e-01 -3.12347800e-01 8.56606066e-01 -3.09791833e-01
8.85387123e-01 1.04904866e+00 8.89036298e-01 2.64875948e-01
-9.86762106e-01 -3.91854107e-01 -2.97588617e-01 -7.30852364e-03
-1.45766973e+00 -1.38081148e-01 2.28267476e-01 -6.27238810e-01
1.30625916e+00 6.91983998e-01 5.99518538e-01 4.76481110e-01
3.94288838e-01 -3.24290127e-01 7.04423964e-01 -6.05607748e-01
-3.71603481e-02 4.20977384e-01 4.25808251e-01 6.42110944e-01
1.00626910e+00 -5.94123080e-02 -4.39716130e-01 -7.49490619e-01
2.32814789e-01 5.07193387e-01 -2.68787593e-01 3.76291841e-01
-1.07259619e+00 6.39049947e-01 1.54663369e-01 8.04722548e-01
-6.15428865e-01 -5.05531430e-01 5.47242820e-01 2.31419429e-01
2.93588549e-01 5.63244998e-01 -8.34070861e-01 2.62461871e-01
-6.18648350e-01 2.43916158e-02 1.08624864e+00 1.05755508e+00
3.58283609e-01 -5.79547703e-01 1.88100301e-02 6.39294922e-01
4.11511630e-01 4.32786122e-02 8.58672857e-01 -3.77573818e-01
2.43136480e-01 1.30060196e+00 -2.85694897e-01 -1.51611531e+00
-6.46632254e-01 -2.30510801e-01 -1.00830257e+00 -6.65335715e-01
1.07534371e-01 -3.57916415e-01 -7.04582036e-01 1.33141935e+00
4.15346622e-01 9.75744724e-02 5.04136562e-01 5.12891114e-01
1.10861170e+00 1.51829183e-01 5.73428750e-01 -5.94687521e-01
2.04822397e+00 -6.71669543e-02 -1.13299239e+00 5.81804156e-01
9.69540775e-01 -6.38289928e-01 1.18852101e-01 4.13797617e-01
-7.05986440e-01 -9.39769298e-02 -6.63246512e-01 9.72748920e-02
-9.18066800e-01 1.03135422e-01 9.36388671e-01 5.26163757e-01
-6.78926528e-01 4.54495192e-01 -6.40241623e-01 -1.00487089e+00
6.67477250e-01 5.52231193e-01 -5.38646340e-01 7.74582922e-02
-1.46812975e+00 1.00854301e+00 7.57510006e-01 -1.49465948e-01
2.09788308e-02 -1.03270674e+00 -6.51688755e-01 -7.81888068e-02
3.33783090e-01 -1.23027265e+00 3.56842428e-01 -3.81460011e-01
-6.25706732e-01 9.26894128e-01 -2.53996581e-01 -5.19414365e-01
-2.91560013e-02 4.05891873e-02 -1.10004830e+00 2.85312533e-01
2.45571166e-01 4.44570705e-02 -2.73188710e-01 -6.88673496e-01
-6.54284835e-01 -8.69717956e-01 -7.16276288e-01 6.20510690e-02
-3.43329787e-01 1.19662158e-01 1.81256562e-01 -6.17651403e-01
1.75903440e-01 -5.08124232e-01 -4.77754116e-01 -3.85780245e-01
-5.11625111e-01 -1.57359779e-01 3.76760334e-01 -6.97338462e-01
1.54505146e+00 -1.57655311e+00 -2.08839059e-01 3.49979669e-01
5.29783487e-01 1.68008700e-01 5.64280152e-01 1.03716409e+00
-3.72382760e-01 5.76483786e-01 -8.38190839e-02 6.34248495e-01
-4.07771021e-01 3.48934442e-01 2.61201244e-02 3.97046536e-01
6.24667346e-01 9.49409068e-01 -7.87804663e-01 -8.31473649e-01
6.46699443e-02 5.73218048e-01 -5.81291080e-01 -2.92103775e-02
-1.30511820e-01 3.32570016e-01 -8.47985625e-01 6.82098925e-01
4.03248012e-01 -2.91468710e-01 5.25244117e-01 -4.27046299e-01
5.47965169e-02 3.10675591e-01 -7.79689074e-01 1.51181865e+00
3.03677112e-01 2.28615943e-02 -3.23644221e-01 -1.37661111e+00
1.14835358e+00 7.99019277e-01 1.03096926e+00 -1.52862072e-01
2.02368826e-01 -4.90685254e-02 2.41842598e-01 -1.33369207e+00
2.28905100e-02 -3.76858979e-01 1.45890832e-01 -3.50164510e-02
-1.20145604e-01 3.99567902e-01 8.83962810e-02 2.13598832e-01
1.15775156e+00 -2.52465010e-01 1.14214289e+00 -4.13881183e-01
6.07176542e-01 6.11642420e-01 6.22459769e-01 3.38759273e-01
1.38047099e-01 9.59176570e-02 5.53124070e-01 -4.01495606e-01
-7.15475440e-01 -9.21562970e-01 -6.40627682e-01 2.06181735e-01
-2.73749799e-01 -6.98961258e-01 -2.18271360e-01 -1.89027488e-01
1.62485793e-01 3.84530723e-01 -6.47819638e-01 -1.37715742e-01
-1.39116379e-03 -1.30053866e+00 7.43192375e-01 -1.12348478e-02
1.55995354e-01 -6.24915600e-01 -5.56514919e-01 4.64246124e-01
4.77821231e-02 -7.87847579e-01 2.45577380e-01 1.05962075e-01
-1.18332064e+00 -1.54273307e+00 -1.24146670e-01 -7.29757369e-01
7.51011789e-01 -4.28388029e-01 9.43100631e-01 2.65578657e-01
-9.29241538e-01 -5.95292225e-02 -3.98523897e-01 -9.75616097e-01
-4.73070323e-01 -2.13030353e-01 5.13168871e-02 -3.23701143e-01
1.08502793e+00 -6.60992444e-01 -5.69127917e-01 -8.56188387e-02
-1.01728904e+00 -2.74957091e-01 7.69912720e-01 6.52306616e-01
4.83361721e-01 5.50168037e-01 1.08407366e+00 -1.22672915e+00
8.23201239e-01 -1.22604263e+00 -9.83384475e-02 2.74456412e-01
-9.07431245e-01 -6.89875111e-02 2.80730486e-01 -1.04603566e-01
-9.60826933e-01 6.57384889e-03 1.51753768e-01 3.57140630e-01
-6.32188797e-01 1.02074993e+00 -1.06990710e-01 7.38221943e-01
6.54391646e-01 -1.61482040e-02 3.20632368e-01 -4.63631451e-01
-7.33785331e-02 9.44019258e-01 3.46051425e-01 -2.68502653e-01
-1.07158925e-02 3.06863815e-01 4.15618956e-01 -6.82445943e-01
-5.33771992e-01 -6.47943556e-01 -7.34394252e-01 3.16371083e-01
1.01425099e+00 -7.84585893e-01 -9.64747548e-01 -4.41500157e-01
-8.18153918e-01 6.19463265e-01 2.08014734e-02 9.17997062e-01
1.71456896e-02 8.81119296e-02 -3.68529379e-01 -7.44835317e-01
-6.43075466e-01 -4.63540733e-01 6.92379296e-01 1.58161938e-01
-8.68539453e-01 -1.25739527e+00 2.68911928e-01 2.93378681e-01
6.80076629e-02 7.69671679e-01 1.53739405e+00 -1.28310621e+00
9.74285416e-03 -3.95365983e-01 -9.52589735e-02 -3.47822607e-01
9.43479121e-01 1.53787911e-01 -4.83657509e-01 3.79353672e-01
-2.01880075e-02 3.52854311e-01 3.89749497e-01 3.94429743e-01
6.76980257e-01 -8.01153660e-01 -9.06405330e-01 2.30493188e-01
1.59928143e+00 5.95561862e-01 6.16619110e-01 -1.38780503e-02
5.63076079e-01 9.46600437e-01 3.78554374e-01 4.33123678e-01
3.58865708e-01 2.65994638e-01 -1.60823330e-01 2.13943087e-02
1.41346246e-01 5.66443987e-02 -3.75513524e-01 7.16251731e-01
-2.57201225e-01 -2.44417191e-01 -1.18784058e+00 6.87373817e-01
-1.69395137e+00 -8.61731827e-01 -8.20333421e-01 1.99419463e+00
1.16593361e+00 -2.77189493e-01 1.70203317e-02 1.64252341e-01
7.02171504e-01 -9.60215449e-01 -2.51173675e-01 -5.91333389e-01
-2.09570646e-01 2.85666406e-01 3.46622467e-01 2.15980574e-01
-6.27686739e-01 4.60583478e-01 6.34014797e+00 2.13463306e-01
-6.68645859e-01 -2.15755939e-01 5.76001823e-01 8.54297876e-02
-1.95227429e-01 -1.48022696e-01 -7.15124726e-01 4.80581433e-01
1.29614878e+00 -4.70208943e-01 -1.87005848e-01 3.68132055e-01
4.64806378e-01 -2.70576358e-01 -1.00266612e+00 5.69194555e-01
-8.92022848e-02 -1.41817832e+00 1.18012443e-01 2.90580541e-01
6.43997729e-01 -2.41956681e-01 -6.14150584e-01 -3.25675011e-01
3.48990947e-01 -1.14269698e+00 -4.49003041e-01 1.14066648e+00
6.36121690e-01 -5.25628805e-01 1.06407118e+00 6.51195496e-02
-8.48046720e-01 2.94011235e-02 -3.60747308e-01 -2.78352320e-01
2.59177163e-02 1.09037578e+00 -1.48859656e+00 1.19849455e+00
6.39219761e-01 9.57902551e-01 -3.62911344e-01 9.21171010e-01
2.18160763e-01 5.82366109e-01 1.88078992e-02 -1.22996420e-02
-1.59332842e-01 -2.60190666e-01 2.25287184e-01 1.30749655e+00
5.24179459e-01 8.77854586e-01 -1.42258808e-01 8.53052020e-01
3.63093555e-01 5.99233627e-01 -9.56198633e-01 -4.09618497e-01
4.66445804e-01 1.11282492e+00 -8.43153775e-01 -6.33067846e-01
-4.23069894e-01 4.43285465e-01 -1.42404497e-01 1.22430556e-01
-3.04802775e-01 -4.61348057e-01 7.53854275e-01 2.28084117e-01
-1.86993778e-01 3.93095762e-01 -4.91963208e-01 -9.06331062e-01
-4.55583006e-01 -7.25599229e-01 1.04213893e+00 -4.78051662e-01
-1.55926514e+00 5.71719944e-01 1.14338405e-01 -1.05152953e+00
-1.78678066e-01 -3.43384713e-01 -1.09364122e-01 1.09610522e+00
-1.06188726e+00 -8.16021681e-01 -2.11067975e-01 6.72694087e-01
-5.95388003e-02 -3.01937997e-01 1.40788555e+00 4.84140247e-01
-5.21149874e-01 1.06296554e-01 -6.72287494e-02 2.39118878e-02
6.98703051e-01 -9.39628065e-01 -3.38947803e-01 1.06094450e-01
-5.66087723e-01 1.06849635e+00 5.92243493e-01 -1.29011142e+00
-1.16053832e+00 -1.14205658e+00 1.61925447e+00 -4.68085021e-01
6.19331002e-01 6.59901127e-02 -1.06756818e+00 4.61620003e-01
1.89445645e-01 -2.28839681e-01 1.62494910e+00 -1.75243020e-01
3.71680595e-02 7.19570071e-02 -1.50503004e+00 2.97745913e-01
9.60943043e-01 -1.16120242e-01 -1.06787431e+00 4.43208158e-01
6.73707187e-01 2.02189833e-01 -1.62425458e+00 3.64210427e-01
4.64757293e-01 -4.19591039e-01 1.08549201e+00 -1.31486976e+00
5.44805110e-01 -2.92356312e-01 3.91173027e-02 -9.24697042e-01
-2.14002416e-01 -4.45352912e-01 2.63013363e-01 1.38281751e+00
6.21465683e-01 -7.60439813e-01 5.16897738e-01 9.19109523e-01
7.16844499e-02 -7.62579501e-01 -4.77359593e-01 -3.28596681e-01
-3.50064933e-01 -8.31768513e-02 6.23090863e-01 1.51556933e+00
7.10243702e-01 4.63192999e-01 -6.91125765e-02 2.80186981e-01
3.74586284e-01 5.38091511e-02 2.75207698e-01 -1.76478934e+00
1.81662142e-02 -2.02943996e-01 -7.99716949e-01 2.83115089e-01
-3.62803429e-01 -1.06372166e+00 -5.97031951e-01 -1.94509661e+00
2.54401267e-01 -5.41408062e-01 -2.32647777e-01 4.63577092e-01
-4.94103096e-02 -8.22541043e-02 -5.64269543e-01 9.56850052e-02
2.23112665e-02 -5.99901862e-02 9.32112455e-01 1.41141668e-01
-5.64432204e-01 -1.77368745e-01 -1.24211347e+00 5.48584044e-01
8.99187863e-01 -8.39334130e-01 -5.70802450e-01 2.11435646e-01
5.16946733e-01 1.33632332e-01 1.88604027e-01 -4.89754826e-01
3.96595567e-01 -2.33274460e-01 4.25772250e-01 -5.88001907e-01
-1.51084155e-01 -1.08378983e+00 8.81739438e-01 8.83529603e-01
-4.57816660e-01 1.23847991e-01 2.48511493e-01 5.72409451e-01
-2.17385694e-01 -1.22993886e-01 1.01931803e-02 -2.18271613e-01
-2.65302360e-01 -1.87341183e-01 -4.65382576e-01 -1.47633940e-01
1.21972132e+00 -4.70454663e-01 -3.39374393e-01 2.45071724e-02
-1.15449357e+00 1.19027264e-01 -1.27756566e-01 2.05990225e-01
4.84589189e-01 -1.14724445e+00 -8.74813437e-01 -4.41308431e-02
1.28484175e-01 -7.02926293e-02 2.45388314e-01 1.19332957e+00
-6.33828640e-01 1.02306795e+00 -2.50332057e-01 -2.25631818e-01
-1.64228594e+00 9.40208852e-01 -1.91061068e-02 -1.63600117e-01
-6.59865141e-01 4.19591755e-01 -1.30268484e-01 -1.53357416e-01
1.31993787e-02 -5.94544768e-01 -6.80151343e-01 5.07674396e-01
5.56872785e-01 3.21334898e-01 1.68077990e-01 -6.13676310e-01
-8.15813482e-01 5.18315315e-01 3.45330946e-02 4.14232820e-01
1.77782273e+00 -3.24344754e-01 -8.31998825e-01 3.46160382e-01
1.15550780e+00 1.40919805e-01 8.28359723e-02 -1.00428112e-01
5.15718877e-01 -1.11506104e-01 -4.47742641e-01 -1.14008760e+00
-7.66519010e-01 2.60205537e-01 3.20337623e-01 3.82100075e-01
1.29173076e+00 2.53968060e-01 1.12930633e-01 3.20203841e-01
-2.25595862e-01 -6.70036972e-01 -8.23046088e-01 -4.70167911e-03
6.33569241e-01 -1.07237971e+00 3.20277989e-01 -9.39199269e-01
-4.81507242e-01 1.20519280e+00 9.33893621e-02 1.89340740e-01
1.03043902e+00 4.07759130e-01 -1.31411344e-01 -7.84139991e-01
-1.07406378e+00 -1.60195003e-03 4.50626969e-01 6.69656754e-01
9.71086800e-01 8.64649266e-02 -1.13172114e+00 1.10642695e+00
-2.64893919e-01 5.75901389e-01 2.04666555e-01 8.23294163e-01
-2.48944744e-01 -1.06699193e+00 -4.92013693e-01 9.79109704e-01
-1.05562270e+00 -4.80590492e-01 -6.81342959e-01 7.79010832e-01
6.35396004e-01 1.14057350e+00 1.24365985e-02 -1.93070054e-01
1.66352764e-01 2.93656677e-01 5.21997698e-02 -6.57548666e-01
-7.55150199e-01 1.67964429e-01 4.15945709e-01 -1.94326848e-01
-8.81165445e-01 -5.40568709e-01 -1.59450722e+00 -3.80769931e-02
-3.29205632e-01 4.80381131e-01 5.20011783e-01 9.79804456e-01
8.41592550e-01 9.56057429e-01 -8.81671384e-02 3.67613226e-01
1.71603084e-01 -8.68744135e-01 -5.29280424e-01 4.35716331e-01
7.79343098e-02 -2.76983172e-01 1.38639644e-01 7.31667697e-01] | [8.439491271972656, 8.604217529296875] |
445eb151-4e7b-4fb6-a6d3-9413ef7c3518 | task-adaptive-spatial-temporal-video-sampler | 2207.09759 | null | https://arxiv.org/abs/2207.09759v3 | https://arxiv.org/pdf/2207.09759v3.pdf | Task-adaptive Spatial-Temporal Video Sampler for Few-shot Action Recognition | A primary challenge faced in few-shot action recognition is inadequate video data for training. To address this issue, current methods in this field mainly focus on devising algorithms at the feature level while little attention is paid to processing input video data. Moreover, existing frame sampling strategies may omit critical action information in temporal and spatial dimensions, which further impacts video utilization efficiency. In this paper, we propose a novel video frame sampler for few-shot action recognition to address this issue, where task-specific spatial-temporal frame sampling is achieved via a temporal selector (TS) and a spatial amplifier (SA). Specifically, our sampler first scans the whole video at a small computational cost to obtain a global perception of video frames. The TS plays its role in selecting top-T frames that contribute most significantly and subsequently. The SA emphasizes the discriminative information of each frame by amplifying critical regions with the guidance of saliency maps. We further adopt task-adaptive learning to dynamically adjust the sampling strategy according to the episode task at hand. Both the implementations of TS and SA are differentiable for end-to-end optimization, facilitating seamless integration of our proposed sampler with most few-shot action recognition methods. Extensive experiments show a significant boost in the performances on various benchmarks including long-term videos.The code is available at https://github.com/R00Kie-Liu/Sampler | ['Weiyao Lin', 'John See', 'Weixian Lv', 'Huabin Liu'] | 2022-07-20 | null | null | null | null | ['few-shot-action-recognition'] | ['computer-vision'] | [ 4.48940426e-01 -4.19816464e-01 -4.74264294e-01 -1.43805668e-01
-6.59822643e-01 -3.21457535e-02 3.10195386e-01 -2.80787706e-01
-4.39467311e-01 4.20731664e-01 3.35645586e-01 4.18982133e-02
-2.03328785e-02 -4.80051905e-01 -5.77346027e-01 -8.15560520e-01
7.90368319e-02 -1.85035631e-01 6.64090097e-01 1.26469005e-02
5.07051408e-01 1.90334484e-01 -1.43137622e+00 3.12854469e-01
8.02958965e-01 1.18152654e+00 6.87317848e-01 5.97099006e-01
-4.80172224e-02 1.02169347e+00 -4.38314557e-01 -2.92299818e-02
2.65748262e-01 -7.00462222e-01 -4.74119961e-01 3.81067038e-01
2.49380305e-01 -6.07448995e-01 -4.39726055e-01 9.70366001e-01
5.48534214e-01 4.66859818e-01 2.55558670e-01 -1.08560288e+00
-2.39048034e-01 4.35612410e-01 -8.18749905e-01 8.80969226e-01
2.25313112e-01 4.83472377e-01 8.21787715e-01 -1.10139525e+00
3.34752500e-01 9.78964984e-01 2.28062555e-01 4.59658802e-01
-8.14930916e-01 -5.59450567e-01 4.78953987e-01 6.47851467e-01
-1.26117754e+00 -7.14131892e-01 1.00078142e+00 -2.91788965e-01
7.74909854e-01 9.34573486e-02 7.85846472e-01 8.16529572e-01
3.52821164e-02 1.19048750e+00 5.17122924e-01 -2.14508846e-01
3.14032704e-01 -3.68173569e-01 -1.31803274e-01 5.87690532e-01
-1.92514092e-01 -9.38553512e-02 -8.44140947e-01 1.87617302e-01
1.06702769e+00 3.24755728e-01 -5.02164602e-01 -3.04468751e-01
-1.19620645e+00 5.30012488e-01 2.59968609e-01 2.43226483e-01
-7.03109324e-01 2.22678825e-01 5.31739235e-01 1.59759253e-01
3.51193607e-01 9.41578597e-02 -2.77062625e-01 -5.15198171e-01
-1.08715379e+00 1.58580095e-02 1.93384737e-01 8.77336979e-01
7.25650251e-01 3.17636222e-01 -5.61851084e-01 8.74386907e-01
3.82501520e-02 2.31240675e-01 6.18281424e-01 -1.08902955e+00
6.57778740e-01 5.04099369e-01 6.56698570e-02 -1.00302076e+00
2.09383611e-02 -2.40119308e-01 -5.91685832e-01 -3.61008197e-02
3.39195162e-01 -7.81654716e-02 -7.83520281e-01 1.41613030e+00
3.85670900e-01 7.41260171e-01 -2.72606492e-01 1.29048574e+00
4.14821357e-01 7.22255170e-01 1.51337892e-01 -5.70541084e-01
1.44967210e+00 -1.22008753e+00 -7.01781809e-01 -4.22631681e-01
3.21095288e-01 -6.10866606e-01 1.27406418e+00 1.68451428e-01
-1.13345885e+00 -6.23664975e-01 -1.02291393e+00 9.84973237e-02
1.44861162e-01 3.78668964e-01 4.02921617e-01 2.30369866e-01
-6.59064949e-01 3.05182546e-01 -9.79823530e-01 -1.71946540e-01
7.61445284e-01 2.01979920e-01 3.75749879e-02 -4.41558957e-02
-1.08437598e+00 4.21576649e-01 3.72998565e-01 1.65009633e-01
-1.00260067e+00 -5.45798600e-01 -7.62250543e-01 1.56578004e-01
9.03189003e-01 -3.99680346e-01 1.31622505e+00 -1.27205408e+00
-1.66060901e+00 2.79043406e-01 -4.72236961e-01 -5.13796151e-01
4.77617472e-01 -3.42256933e-01 -2.08221048e-01 6.53758049e-01
1.25618935e-01 5.58904171e-01 1.16291368e+00 -6.79489017e-01
-9.63387072e-01 -2.47396901e-01 2.72013992e-01 5.82484007e-01
-6.17947400e-01 2.16110215e-01 -8.27179372e-01 -9.38190341e-01
1.13992980e-02 -5.45717657e-01 -1.31407499e-01 1.16386198e-01
1.76566560e-02 -1.50376648e-01 8.68141830e-01 -4.70789492e-01
1.51326430e+00 -2.28824115e+00 2.36787468e-01 -2.59093821e-01
6.65274188e-02 5.72089076e-01 -2.32202597e-02 1.41286716e-01
1.06593654e-01 -3.30361009e-01 -7.75477216e-02 -2.34095812e-01
-4.56066102e-01 -5.53728268e-02 -3.22027653e-01 3.80846947e-01
3.50768238e-01 7.94750392e-01 -1.06457114e+00 -7.20115364e-01
5.57028532e-01 3.53589684e-01 -5.70637703e-01 2.80991048e-01
-1.85757145e-01 3.68041068e-01 -8.72820854e-01 8.10892642e-01
3.83393437e-01 -3.35220218e-01 -4.79091890e-02 -2.17843682e-01
-3.07738155e-01 2.25109980e-01 -1.11809599e+00 1.75604117e+00
-3.05815935e-01 4.82926726e-01 -7.56084323e-02 -1.01542485e+00
6.81633234e-01 2.18776062e-01 6.19502485e-01 -8.44304204e-01
1.51487127e-01 -4.17554118e-02 -1.62059069e-01 -5.58508813e-01
4.43262368e-01 1.85505785e-02 2.50889122e-01 2.92866349e-01
-1.45351887e-01 4.50804919e-01 3.08785319e-01 8.40860531e-02
9.82514322e-01 3.07912797e-01 4.70431477e-01 5.68092242e-02
6.66413426e-01 -8.14086571e-02 9.15787160e-01 4.67826962e-01
-6.86425447e-01 6.99152410e-01 3.84017289e-01 -3.14807564e-01
-7.17083752e-01 -7.90949166e-01 2.06314638e-01 1.38526261e+00
5.36481798e-01 -4.68532234e-01 -6.50267005e-01 -5.22283733e-01
-3.27780515e-01 5.91825485e-01 -5.25180936e-01 -3.00423205e-01
-8.46066892e-01 -3.86653513e-01 1.55444786e-01 5.72303653e-01
7.60823429e-01 -1.35756195e+00 -1.19902754e+00 2.58490920e-01
-1.95755675e-01 -1.01340032e+00 -1.00321114e+00 -2.03525484e-01
-9.53062892e-01 -1.07826602e+00 -9.10120189e-01 -7.50961423e-01
5.88811934e-01 7.95058429e-01 5.86122870e-01 -9.13791284e-02
-1.59197196e-01 1.43620357e-01 -5.74084342e-01 -5.28548844e-02
1.57401517e-01 4.95735183e-02 -1.94746912e-01 3.59150559e-01
4.50426370e-01 -4.53843862e-01 -9.97092307e-01 6.04488194e-01
-8.69908333e-01 2.78198332e-01 5.85918725e-01 8.77097785e-01
6.11949325e-01 1.11334138e-02 5.37044108e-01 -4.23935205e-01
2.31343567e-01 -4.00764942e-01 -4.31082428e-01 1.19280405e-01
-2.31214035e-02 -2.55893171e-01 7.89706230e-01 -5.26272476e-01
-1.01077330e+00 9.49494839e-02 1.41648769e-01 -8.94317091e-01
6.16469570e-02 3.35305452e-01 -2.75413841e-01 2.15819269e-01
2.60261595e-01 5.98179936e-01 8.44894722e-02 -2.16951504e-01
-5.39535284e-03 5.22880316e-01 3.17853808e-01 -2.97310919e-01
3.84026170e-01 4.14132357e-01 -3.30147743e-01 -8.64931226e-01
-7.14270949e-01 -6.04094148e-01 -4.94278729e-01 -5.56519449e-01
7.40462661e-01 -9.86871898e-01 -5.03790855e-01 5.61447263e-01
-7.78578281e-01 -3.71096581e-01 -1.21913284e-01 5.89618981e-01
-6.29064679e-01 3.46569538e-01 -4.53146785e-01 -7.88060546e-01
-2.79295683e-01 -1.26791704e+00 1.14711189e+00 4.61513728e-01
4.82350811e-02 -5.51567376e-01 -3.02620649e-01 2.97959149e-01
3.11350852e-01 -1.24985367e-01 4.67958719e-01 -1.04788765e-01
-8.74936581e-01 2.77413148e-02 -1.98495895e-01 1.87233537e-01
3.69615287e-01 -1.51995108e-01 -6.69331670e-01 -3.47858816e-01
1.44973829e-01 -2.34195381e-01 9.32575345e-01 6.66534543e-01
1.40075994e+00 -1.52095288e-01 -2.25742951e-01 6.43728077e-01
1.32938302e+00 5.47739387e-01 7.06161618e-01 3.14008713e-01
8.65466118e-01 2.43398398e-01 1.12157679e+00 6.48321509e-01
1.69336051e-01 9.87279177e-01 3.00546497e-01 1.65339038e-01
-8.23876932e-02 -3.02528918e-01 5.84249437e-01 5.85685432e-01
-1.87727869e-01 -1.11363709e-01 -6.27620280e-01 5.24777353e-01
-2.08064270e+00 -1.35945177e+00 4.08003509e-01 2.24092960e+00
7.93894172e-01 3.21007520e-01 3.37978035e-01 1.22094482e-01
8.28026474e-01 6.85540676e-01 -8.05224776e-01 1.28800467e-01
1.29941076e-01 -1.25766098e-01 2.64780760e-01 2.11686254e-01
-1.15750372e+00 9.84199107e-01 5.19709110e+00 1.16395974e+00
-1.28540838e+00 -2.53627226e-02 7.72615850e-01 -6.09987438e-01
1.08800553e-01 3.49505804e-02 -7.62442589e-01 7.92220592e-01
5.78519166e-01 -1.60476297e-01 3.96045744e-01 8.01902473e-01
8.14555705e-01 -3.05641502e-01 -7.59610951e-01 1.06683373e+00
5.53720966e-02 -1.31317711e+00 2.58307420e-02 -2.61397332e-01
4.79336888e-01 -2.23082200e-01 -9.29284934e-03 2.26337671e-01
-4.34429049e-01 -5.69706202e-01 7.77853191e-01 4.93371040e-01
6.12442791e-01 -7.57793307e-01 2.86839277e-01 2.61728436e-01
-1.42014468e+00 -3.92072558e-01 -3.93193811e-01 -1.26092300e-01
2.71229506e-01 3.83437783e-01 -4.78085876e-01 2.36092269e-01
5.25838792e-01 1.07018590e+00 -4.11443412e-01 1.17870200e+00
-1.49939284e-01 7.30086625e-01 -6.75619245e-02 -8.34396631e-02
3.41312170e-01 -1.61158577e-01 7.41429806e-01 9.33716893e-01
2.87648052e-01 3.97459149e-01 3.02028060e-01 4.71961379e-01
1.94489747e-01 8.85415599e-02 -3.12384337e-01 1.06453318e-02
5.56358099e-01 1.03863215e+00 -8.54556501e-01 -4.60926592e-01
-5.84293664e-01 1.03408313e+00 1.94169119e-01 3.71162862e-01
-1.03404748e+00 -3.59696507e-01 6.30245924e-01 2.03019187e-01
7.84449756e-01 -1.98872462e-01 -1.36593029e-01 -1.18912113e+00
2.36565977e-01 -9.51036096e-01 4.48805511e-01 -6.80912137e-01
-6.93940997e-01 3.22783947e-01 -3.00065447e-02 -1.62024796e+00
-8.09202045e-02 -2.20645413e-01 -7.56550491e-01 5.88704646e-01
-1.33745301e+00 -6.96472168e-01 -4.29788530e-01 5.34422219e-01
1.38153899e+00 -1.86294094e-01 1.79817751e-01 3.58000398e-01
-9.37176049e-01 4.27645773e-01 -2.21478045e-01 -4.67947539e-04
5.81956029e-01 -8.68162036e-01 2.32596859e-01 1.10284638e+00
-6.46279901e-02 4.38717037e-01 6.00964546e-01 -6.34558499e-01
-1.54452837e+00 -1.07242846e+00 5.07802844e-01 -3.57271619e-02
4.73065615e-01 -7.59679005e-02 -7.92286634e-01 3.16254586e-01
5.22448905e-02 1.36689663e-01 3.43031764e-01 -3.69673610e-01
1.16926670e-01 -3.15309644e-01 -7.19184995e-01 7.50334799e-01
1.17106926e+00 -4.56366509e-01 -4.40621674e-01 5.23373857e-02
5.54715633e-01 -3.39642137e-01 -3.62366468e-01 2.06840649e-01
5.71794987e-01 -1.05385292e+00 8.65067840e-01 -2.60319471e-01
6.57251477e-01 -5.82531631e-01 8.03663454e-04 -1.00279522e+00
-3.54439318e-01 -5.68938255e-01 -3.99887025e-01 9.97066259e-01
-8.41885284e-02 -2.64672130e-01 8.13147187e-01 3.82175505e-01
-1.45534709e-01 -1.06463313e+00 -8.13800037e-01 -5.24490178e-01
-5.83072960e-01 -3.24131429e-01 4.15082842e-01 5.99719405e-01
-3.73697430e-02 1.59164011e-01 -6.24003530e-01 4.01157960e-02
4.53307301e-01 2.56369889e-01 6.50338769e-01 -4.80881602e-01
-4.65776831e-01 -5.22415996e-01 -4.29052502e-01 -1.53204036e+00
-1.74467653e-01 -3.81432623e-01 1.80589542e-01 -1.20487463e+00
2.73528069e-01 3.31367068e-02 -5.47981262e-01 3.70927691e-01
-5.30618489e-01 1.78846791e-01 2.52264172e-01 3.98896992e-01
-9.27315176e-01 8.16404045e-01 1.39058089e+00 -7.38233849e-02
-3.55449796e-01 1.15060806e-01 -5.54448545e-01 6.37846351e-01
7.66394675e-01 -2.46738896e-01 -8.00908387e-01 -4.75561142e-01
-3.43372852e-01 3.38074505e-01 4.12776828e-01 -1.10222232e+00
3.51389527e-01 -4.89646703e-01 3.55825603e-01 -5.76700389e-01
3.34457278e-01 -5.43486774e-01 -1.42527819e-01 4.07073945e-01
-4.04001832e-01 -5.86090498e-02 6.04118481e-02 6.48183942e-01
-3.02533656e-01 -1.65239230e-01 8.38358939e-01 -8.38438869e-02
-1.19210660e+00 5.03571510e-01 -3.23177993e-01 6.08364530e-02
1.24254251e+00 -4.77227718e-01 -5.30455746e-02 -3.23857456e-01
-4.66848314e-01 4.29119825e-01 4.59446400e-01 4.89073336e-01
9.08013165e-01 -1.33788848e+00 -4.49398369e-01 2.63787359e-01
6.15702234e-02 -1.32134661e-01 5.50596595e-01 1.05890191e+00
-3.12632322e-01 2.56867349e-01 -3.28439325e-01 -5.50007164e-01
-1.27548897e+00 4.32549626e-01 3.32764387e-01 1.58482060e-01
-7.76912034e-01 8.47262084e-01 3.50918919e-01 6.10424459e-01
4.14950371e-01 -2.49156237e-01 -3.86947483e-01 1.85150117e-01
9.04994547e-01 4.31056350e-01 -3.78101438e-01 -6.42733991e-01
-4.46598172e-01 4.70960140e-01 -1.87344000e-01 4.19931673e-02
1.18375158e+00 -2.74405509e-01 3.28443915e-01 5.47682285e-01
1.06526887e+00 -3.17817718e-01 -1.91181159e+00 -2.60711074e-01
-2.71153808e-01 -1.08212245e+00 2.28031889e-01 -3.68354678e-01
-1.25574434e+00 7.21847832e-01 4.55675066e-01 -6.95058927e-02
1.53133440e+00 -2.66234338e-01 9.72012520e-01 7.87700489e-02
5.00340462e-01 -1.23591852e+00 4.49958414e-01 3.88493955e-01
6.57701552e-01 -1.05865455e+00 7.00969202e-03 -4.08092350e-01
-8.75691473e-01 9.97327447e-01 8.33083808e-01 -1.91651806e-01
3.99161428e-01 -2.68282313e-02 -7.50179440e-02 -7.65348598e-02
-8.47891808e-01 -2.74673611e-01 1.92143559e-01 2.36867636e-01
2.32831731e-01 -2.44455829e-01 -4.20236498e-01 4.49299991e-01
3.44609588e-01 3.41170460e-01 2.69309521e-01 1.01927376e+00
-7.00171769e-01 -8.09623599e-01 -1.06800564e-01 5.08198082e-01
-3.68139327e-01 -5.95354438e-02 3.94891687e-02 3.02325606e-01
-1.60323195e-02 8.07613254e-01 1.44742981e-01 -3.40734154e-01
2.50938296e-01 -1.68291867e-01 3.18425208e-01 -4.44363207e-01
-2.99945682e-01 4.06574398e-01 -9.78343487e-02 -8.20394695e-01
-5.56716263e-01 -9.18498755e-01 -1.27936041e+00 2.50130203e-02
-2.76831478e-01 5.12249134e-02 1.54920429e-01 9.27959442e-01
4.59019035e-01 6.30943298e-01 8.42597544e-01 -1.02072251e+00
-4.71677244e-01 -8.23410332e-01 -4.30743188e-01 2.51462787e-01
3.46041322e-01 -8.12322557e-01 -2.34591886e-01 1.12916648e-01] | [8.794281959533691, 0.3370566964149475] |
5e154e7b-6fa6-49cb-8fa6-908224e3f105 | equivariance-with-learned-canonicalization | 2211.06489 | null | https://arxiv.org/abs/2211.06489v3 | https://arxiv.org/pdf/2211.06489v3.pdf | Equivariance with Learned Canonicalization Functions | Symmetry-based neural networks often constrain the architecture in order to achieve invariance or equivariance to a group of transformations. In this paper, we propose an alternative that avoids this architectural constraint by learning to produce canonical representations of the data. These canonicalization functions can readily be plugged into non-equivariant backbone architectures. We offer explicit ways to implement them for some groups of interest. We show that this approach enjoys universality while providing interpretable insights. Our main hypothesis, supported by our empirical results, is that learning a small neural network to perform canonicalization is better than using predefined heuristics. Our experiments show that learning the canonicalization function is competitive with existing techniques for learning equivariant functions across many tasks, including image classification, $N$-body dynamics prediction, point cloud classification and part segmentation, while being faster across the board. | ['Siamak Ravanbakhsh', 'Yoshua Bengio', 'Yan Zhang', 'Arnab Kumar Mondal', 'Sékou-Oumar Kaba'] | 2022-11-11 | null | null | null | null | ['point-cloud-classification'] | ['computer-vision'] | [ 1.11415461e-01 3.78461868e-01 -3.61632377e-01 -5.48065424e-01
-3.40160340e-01 -9.66963112e-01 6.96169853e-01 -4.21289176e-01
-3.56437296e-01 4.38214213e-01 1.58374518e-01 -4.22685236e-01
-2.72243649e-01 -5.62389493e-01 -1.23781157e+00 -6.38393939e-01
-8.83249417e-02 7.69436955e-01 -2.41827834e-02 -2.94499129e-01
2.75067091e-01 8.96031499e-01 -1.33956695e+00 1.21207677e-01
5.08014262e-01 6.78894639e-01 -5.77067100e-02 7.26742566e-01
1.19170755e-01 2.37897187e-01 8.13520700e-02 -2.27249518e-01
7.27009416e-01 -3.05642784e-01 -1.01287508e+00 6.85944706e-02
8.45464170e-01 3.38325500e-02 -1.41951948e-01 8.22829425e-01
9.06470194e-02 3.40200007e-01 8.35475445e-01 -1.22270429e+00
-4.70684439e-01 6.98110461e-01 -2.60649353e-01 -5.27603962e-02
-2.55553633e-01 7.07795694e-02 1.54377735e+00 -6.69593573e-01
1.03882384e+00 1.34825218e+00 8.28764975e-01 7.49740183e-01
-1.77181244e+00 -4.66205627e-01 3.85944784e-01 -4.02171090e-02
-1.21716082e+00 -5.92011809e-01 8.09036434e-01 -4.83515590e-01
1.05497825e+00 3.19327205e-01 8.14796507e-01 7.33931124e-01
4.40254003e-01 5.40173650e-01 6.40570402e-01 -3.09634864e-01
2.87737310e-01 -3.99938881e-01 2.02662200e-01 1.00868309e+00
3.80656540e-01 -1.37770638e-01 -4.12197083e-01 -1.33561298e-01
9.92622793e-01 -1.37065053e-01 -7.09171817e-02 -1.12652087e+00
-1.09800208e+00 8.00008118e-01 6.56743824e-01 3.98265421e-02
-9.30090919e-02 6.04037464e-01 3.85076791e-01 4.33649331e-01
1.17826328e-01 1.08529115e+00 -8.96831214e-01 2.73299605e-01
-5.00421286e-01 5.20439923e-01 8.47076774e-01 9.06807065e-01
9.25369263e-01 2.66312808e-01 5.87874949e-01 5.93213558e-01
1.92282557e-01 9.15234387e-02 4.64002192e-01 -1.52313125e+00
1.50748730e-01 4.32489395e-01 -2.20617235e-01 -8.44550788e-01
-7.93467045e-01 -3.50823522e-01 -8.41127634e-01 4.39317882e-01
4.75357145e-01 -4.81416397e-02 -9.65050757e-01 2.27604198e+00
2.23039076e-01 -1.29143342e-01 -1.20277375e-01 7.23741591e-01
1.79809898e-01 4.49537665e-01 -1.54447734e-01 8.88463333e-02
9.25512493e-01 -6.96953237e-01 -9.00002792e-02 -9.61591229e-02
7.36289740e-01 -4.45226938e-01 1.22062337e+00 4.59617317e-01
-1.17833102e+00 -3.76128584e-01 -1.15994108e+00 -2.19417199e-01
-1.59335718e-01 -8.17876309e-02 1.09721279e+00 3.62376511e-01
-1.37512982e+00 1.09870648e+00 -1.36190248e+00 -6.01660728e-01
3.64138782e-01 8.44458222e-01 -5.68571806e-01 5.07789910e-01
-4.81805682e-01 7.42394626e-01 4.45944011e-01 -2.15679839e-01
-8.01194370e-01 -8.31470132e-01 -7.48627782e-01 -8.69018286e-02
5.48301153e-02 -1.21613884e+00 1.37788844e+00 -1.36537182e+00
-1.53766418e+00 8.23737144e-01 -5.65687716e-02 -5.61892450e-01
4.23302531e-01 -9.43718012e-03 2.52671599e-01 -4.47129495e-02
-2.23714858e-01 1.11418235e+00 9.14628327e-01 -1.00114322e+00
-3.46892416e-01 -4.80881989e-01 3.64033133e-01 3.09197485e-01
-2.06520751e-01 -2.19451100e-01 -3.50335449e-01 -4.95626807e-01
7.04604328e-01 -1.44438303e+00 -5.37153482e-01 3.32541257e-01
-5.72756290e-01 -2.76656389e-01 6.38824165e-01 -3.64010334e-01
5.20782709e-01 -1.76784337e+00 6.51794612e-01 5.23584008e-01
4.15930688e-01 -3.60321164e-01 -1.69521302e-01 4.22267094e-02
-5.42618632e-01 2.26323262e-01 -1.90370485e-01 -1.33492872e-01
5.76438643e-02 6.08001292e-01 -3.32154095e-01 7.66073108e-01
2.09009990e-01 9.52219307e-01 -3.40635419e-01 -2.16627359e-01
1.21459421e-02 2.70089924e-01 -1.20129883e+00 -2.10117236e-01
-4.46696490e-01 2.88364381e-01 -3.24121714e-01 3.34674895e-01
3.75690401e-01 -1.83287606e-01 5.36112130e-01 -2.82687843e-01
-5.43663837e-02 4.02415305e-01 -1.17955506e+00 1.75721288e+00
-2.53153890e-01 6.44585967e-01 2.67931491e-01 -1.40194118e+00
5.88649154e-01 -6.50559589e-02 6.03178501e-01 -1.23136573e-01
1.86788008e-01 7.12248161e-02 4.10971075e-01 3.62161770e-02
2.80193269e-01 -1.70249164e-01 -9.81461704e-02 4.37222749e-01
3.57572585e-01 -1.53734714e-01 8.30103010e-02 3.57378684e-02
9.10529137e-01 5.65891027e-01 2.42822260e-01 -8.81950319e-01
2.72937655e-01 1.98407248e-01 7.16565549e-01 6.15748227e-01
1.83610111e-01 5.52538931e-01 5.98851621e-01 -9.68726873e-01
-1.44466078e+00 -1.29856884e+00 -2.05045100e-02 1.28792357e+00
-2.01125458e-01 -5.12691498e-01 -9.93815303e-01 -4.22935843e-01
-8.49335566e-02 4.72112387e-01 -7.78897107e-01 -2.02431485e-01
-1.05685377e+00 -7.03603268e-01 4.41126019e-01 7.22912490e-01
1.43067166e-01 -8.49082887e-01 -5.85074961e-01 1.59753710e-02
1.68319970e-01 -7.86106944e-01 -5.19895077e-01 6.27045989e-01
-1.37783110e+00 -1.11889565e+00 -5.91157600e-02 -6.23162091e-01
9.94013846e-01 2.11415961e-01 1.04938769e+00 -4.31451425e-02
-2.36378700e-01 4.28553045e-01 2.08720252e-01 -3.02550644e-01
-3.59134167e-01 6.09083235e-01 4.67686385e-01 -3.91524732e-01
-2.76106615e-02 -1.01969814e+00 -3.36988837e-01 3.02060544e-01
-8.48976254e-01 1.12073332e-01 4.43540066e-01 7.68939257e-01
8.68297935e-01 -2.74351835e-01 -1.06603302e-01 -7.88098991e-01
3.50660622e-01 -1.37664098e-02 -7.72383392e-01 -1.01256937e-01
-5.38349390e-01 7.09629714e-01 9.92447138e-01 -3.23919296e-01
-6.37383521e-01 5.50922930e-01 -1.01866961e-01 -4.68316227e-01
-1.80775061e-01 3.33586097e-01 -3.18601072e-01 -4.26805288e-01
8.82389963e-01 -5.44744581e-02 3.16518188e-01 -5.44737458e-01
6.65602207e-01 -3.89329970e-01 5.88065803e-01 -1.05957317e+00
1.07422411e+00 6.33170247e-01 4.42316562e-01 -8.92198145e-01
-6.34763658e-01 -3.17358114e-02 -1.05953646e+00 1.24853104e-01
8.63205016e-01 -6.88433468e-01 -7.46521950e-01 1.04490601e-01
-1.16451943e+00 -3.33138317e-01 -3.50542128e-01 2.40207180e-01
-1.01704001e+00 1.88940868e-01 -3.71675462e-01 -1.79740757e-01
-1.80216819e-01 -1.16782296e+00 6.85421467e-01 -2.69601829e-02
-4.95853812e-01 -9.05700624e-01 2.15478525e-01 -1.59897760e-01
1.97359860e-01 1.81910947e-01 1.22444379e+00 -6.97359920e-01
-8.14517021e-01 1.14861853e-01 2.16312990e-01 1.74493909e-01
-1.77750453e-01 2.99760371e-01 -6.45823538e-01 -3.45053792e-01
-2.59661436e-01 -1.98872149e-01 9.88185585e-01 5.73582768e-01
1.61907029e+00 -4.94851530e-01 -2.45355144e-01 1.25991166e+00
1.10506094e+00 -1.41710550e-01 3.88610125e-01 4.59796488e-01
9.52363729e-01 5.91321766e-01 -1.17860407e-01 -7.40019828e-02
2.72115886e-01 7.02129006e-01 5.02572238e-01 3.59129682e-02
1.21746689e-01 -2.10715219e-01 4.89504635e-01 7.98926353e-01
-4.83244300e-01 2.91339040e-01 -9.81527746e-01 2.44220510e-01
-1.90772641e+00 -8.69850278e-01 1.13132678e-01 1.83401060e+00
5.44254601e-01 1.89852715e-01 3.93147975e-01 -2.70266384e-01
3.22100192e-01 1.27732232e-01 -7.55768538e-01 -5.69913805e-01
-5.90893216e-02 4.47161853e-01 8.59001279e-01 5.38241148e-01
-1.16772854e+00 1.15336168e+00 7.67800474e+00 3.98813128e-01
-1.22480023e+00 -5.70575856e-02 5.40697753e-01 -6.10598363e-04
-4.16494101e-01 1.05012774e-01 -7.84620941e-01 -2.30614290e-01
8.39098990e-01 -3.36672366e-02 6.91347539e-01 1.21764445e+00
9.60830748e-02 6.79509103e-01 -1.56401265e+00 7.05382884e-01
-4.27528918e-02 -1.79019046e+00 3.56509238e-01 1.13936640e-01
6.53021812e-01 3.56040508e-01 1.37035504e-01 -2.28826199e-02
7.03540027e-01 -9.93226349e-01 8.74458373e-01 4.21440572e-01
4.78569597e-01 -8.08592319e-01 -5.27268015e-02 2.62336254e-01
-1.03661883e+00 9.29144546e-02 -5.85480809e-01 -1.32972553e-01
-2.18362018e-01 -1.64698660e-01 -7.96507537e-01 2.50448436e-01
6.06146276e-01 5.75347185e-01 -5.77573538e-01 7.37105370e-01
-1.25757560e-01 5.39912522e-01 -5.90464830e-01 1.40261590e-01
3.60673994e-01 -5.33365488e-01 6.24559700e-01 9.02674079e-01
2.56732643e-01 9.66062304e-03 1.28968418e-01 1.04595125e+00
-1.49618343e-01 4.70351838e-02 -8.68622720e-01 -3.79095525e-02
-6.09489679e-02 1.30304062e+00 -9.39010561e-01 -7.79170543e-02
-1.47549003e-01 6.38087332e-01 4.42034453e-01 2.86979377e-01
-6.26347423e-01 -1.70318916e-01 1.18816137e+00 3.81785147e-02
4.28481162e-01 -6.29256606e-01 -4.56103176e-01 -1.38526928e+00
-6.75243586e-02 -8.92701089e-01 2.57775247e-01 -6.15760863e-01
-9.76713598e-01 3.56190950e-01 2.65170336e-01 -7.71240830e-01
-5.43955564e-01 -1.34235132e+00 -7.64418662e-01 4.08774346e-01
-7.45680034e-01 -1.20134664e+00 7.20100626e-02 5.61602175e-01
5.02553642e-01 -3.20799857e-01 8.08988094e-01 -1.13098927e-01
-6.42257631e-01 6.06598616e-01 1.08998250e-02 8.16774294e-02
3.74874800e-01 -1.53963184e+00 6.99862182e-01 9.08086300e-01
5.87123990e-01 1.03488803e+00 9.40379560e-01 -3.11454237e-01
-1.87013328e+00 -1.02138197e+00 3.28902930e-01 -6.73705637e-01
6.78840697e-01 -5.07232845e-01 -6.04800045e-01 1.21996987e+00
1.62714049e-01 -1.88589692e-01 4.85141069e-01 4.57410723e-01
-7.38660693e-01 -2.31657311e-01 -6.15786970e-01 1.03211927e+00
1.46298087e+00 -3.80925179e-01 -5.30867636e-01 3.20043296e-01
7.90202081e-01 -4.66202974e-01 -5.79730451e-01 3.42493564e-01
8.09468925e-01 -6.81381404e-01 1.09648788e+00 -1.28276360e+00
3.88423622e-01 -3.29420298e-01 -3.39055747e-01 -1.41111684e+00
-5.06662548e-01 -8.39970291e-01 9.02627110e-02 4.81608868e-01
6.42489135e-01 -6.83432579e-01 1.06525958e+00 6.47144079e-01
-4.47137833e-01 -5.06941259e-01 -8.52681816e-01 -7.24460423e-01
6.17013872e-01 -5.50066650e-01 5.52259505e-01 1.07788384e+00
-3.47442150e-01 3.97999555e-01 -1.13776088e-01 8.13845843e-02
6.36325240e-01 1.15418419e-01 1.06480670e+00 -1.37387669e+00
-5.02803266e-01 -8.42273355e-01 -6.21278465e-01 -1.12142253e+00
6.97888911e-01 -1.34109759e+00 -1.76842526e-01 -9.63137984e-01
8.36983249e-02 -2.39710584e-01 -1.32021412e-01 7.56539524e-01
5.07866681e-01 3.02932531e-01 1.37375474e-01 1.56254038e-01
-4.42755193e-01 2.48162538e-01 1.03743970e+00 -4.10911143e-02
-1.77346930e-01 -1.44871622e-01 -9.05439258e-01 1.24638271e+00
9.94106352e-01 -3.16384733e-01 -4.50570256e-01 -5.65033317e-01
4.95263696e-01 -4.90122467e-01 4.18226242e-01 -8.45832944e-01
6.51519969e-02 -4.39243555e-01 5.84425449e-01 -4.26466197e-01
2.55048960e-01 -6.25456870e-01 2.97957748e-01 4.80368137e-01
-4.88247216e-01 5.74085176e-01 1.57060027e-01 2.91038811e-01
2.23976269e-01 -1.47081539e-01 9.54392910e-01 -3.67330104e-01
-5.66095114e-01 4.87318695e-01 -3.20672154e-01 7.97747299e-02
6.13168240e-01 -1.39320657e-01 -4.95249331e-02 -2.87276238e-01
-9.04454291e-01 4.52621244e-02 7.61867702e-01 3.14653188e-01
2.64270008e-01 -1.23276627e+00 -3.10212433e-01 3.50765735e-01
-1.00715213e-01 -3.14239636e-02 -2.49898657e-01 6.21576965e-01
-9.16554332e-01 4.83104467e-01 -5.50953209e-01 -7.87714005e-01
-1.16916883e+00 2.65629023e-01 7.59530306e-01 -4.24940139e-02
-7.68033028e-01 5.70721388e-01 5.44751406e-01 -8.65227222e-01
4.33169678e-02 -5.99604905e-01 2.43396774e-01 -3.32181752e-01
5.43468446e-03 1.63123176e-01 -1.92753803e-02 -6.83243573e-01
-1.40789703e-01 8.31925213e-01 -1.49014518e-01 -2.19264179e-01
1.55111539e+00 2.66983926e-01 -4.26252007e-01 2.55947143e-01
1.21137595e+00 -2.50389755e-01 -1.28111994e+00 5.39761484e-02
-1.37514817e-02 -3.82570177e-02 -1.97706774e-01 -2.09364995e-01
-1.04004192e+00 7.42406428e-01 1.74789980e-01 3.53956297e-02
7.87698507e-01 1.37603089e-01 2.76670367e-01 1.36417115e+00
1.49248511e-01 -1.06800890e+00 -4.56222482e-02 8.11517537e-01
1.00339139e+00 -7.31368780e-01 5.10127246e-02 -2.72816658e-01
-2.19255298e-01 1.31794965e+00 8.20831120e-01 -7.19162762e-01
6.88428402e-01 1.34675920e-01 -1.40044481e-01 -2.71008044e-01
-9.02401567e-01 1.63159817e-01 4.19798911e-01 5.21286607e-01
2.98854411e-01 5.31209633e-02 -1.56305686e-01 1.84931055e-01
-8.35353553e-01 -7.30323076e-01 4.37982917e-01 5.90829253e-01
-4.58445847e-01 -1.34549046e+00 -2.31234729e-01 4.10117179e-01
-3.20235491e-01 1.81005687e-01 -7.35354602e-01 1.20999956e+00
3.63292359e-02 -6.82138205e-02 2.47150570e-01 -2.51935035e-01
2.73250043e-01 4.31334615e-01 8.00003648e-01 -4.49075967e-01
-3.08946759e-01 1.80492222e-01 1.27485218e-02 -9.76282775e-01
-4.05696720e-01 -9.36577618e-01 -1.16178119e+00 -3.23291063e-01
4.80520688e-02 -9.60670635e-02 6.64909124e-01 9.93484318e-01
3.71222526e-01 3.52475852e-01 3.61090809e-01 -1.13412488e+00
-8.03646863e-01 -4.89015102e-01 -2.93337315e-01 2.68073112e-01
2.89879888e-01 -6.25353932e-01 -2.72566020e-01 2.67271608e-01] | [8.84866714477539, 2.4843215942382812] |
c9b6f450-d517-43bc-b71d-679766c25318 | ingb-informed-nonlinear-granular-ball | 2307.01224 | null | https://arxiv.org/abs/2307.01224v1 | https://arxiv.org/pdf/2307.01224v1.pdf | INGB: Informed Nonlinear Granular Ball Oversampling Framework for Noisy Imbalanced Classification | In classification problems, the datasets are usually imbalanced, noisy or complex. Most sampling algorithms only make some improvements to the linear sampling mechanism of the synthetic minority oversampling technique (SMOTE). Nevertheless, linear oversampling has several unavoidable drawbacks. Linear oversampling is susceptible to overfitting, and the synthetic samples lack diversity and rarely account for the original distribution characteristics. An informed nonlinear oversampling framework with the granular ball (INGB) as a new direction of oversampling is proposed in this paper. It uses granular balls to simulate the spatial distribution characteristics of datasets, and informed entropy is utilized to further optimize the granular-ball space. Then, nonlinear oversampling is performed by following high-dimensional sparsity and the isotropic Gaussian distribution. Furthermore, INGB has good compatibility. Not only can it be combined with most SMOTE-based sampling algorithms to improve their performance, but it can also be easily extended to noisy imbalanced multi-classification problems. The mathematical model and theoretical proof of INGB are given in this work. Extensive experiments demonstrate that INGB outperforms the traditional linear sampling frameworks and algorithms in oversampling on complex datasets. | ['GuoYing Wang', 'Yabin Shao', 'Qun Liu', 'Hao Zhou', 'Min Li'] | 2023-07-03 | null | null | null | null | ['imbalanced-classification'] | ['miscellaneous'] | [ 9.84145552e-02 -1.72474101e-01 -7.63888538e-01 -4.54047531e-01
-6.72723055e-01 3.16536650e-02 2.17289820e-01 -2.02940544e-03
-3.31060812e-02 1.22893918e+00 1.34028688e-01 -3.96582820e-02
-1.98599190e-01 -1.14385617e+00 -5.91682851e-01 -8.69617641e-01
2.56045312e-01 6.57459199e-01 1.73530713e-01 -8.79550725e-02
2.91735142e-01 2.57794589e-01 -1.75784969e+00 5.10381758e-01
1.51426184e+00 1.10331464e+00 -8.04916397e-02 1.29073754e-01
-4.62077051e-01 7.08061874e-01 -8.74717593e-01 -1.17059521e-01
2.81178504e-01 -3.89561176e-01 -4.58181143e-01 1.47355601e-01
-8.23687948e-03 -3.13915521e-01 5.52892350e-02 1.13536024e+00
7.50283599e-01 1.37631493e-02 7.51350224e-01 -1.41416347e+00
-6.42283976e-01 7.63171136e-01 -9.83093023e-01 1.44843191e-01
1.75141275e-01 -1.37729675e-01 5.34616351e-01 -1.03258204e+00
1.81001067e-01 1.43723547e+00 9.05237019e-01 2.70114779e-01
-1.08749974e+00 -8.90364885e-01 1.67090729e-01 1.65507972e-01
-1.50025141e+00 -8.39825198e-02 8.37490261e-01 -2.56546825e-01
3.18839222e-01 6.51231587e-01 1.26706362e+00 7.15828180e-01
-1.53632760e-01 1.40113270e+00 1.34177387e+00 -3.37683171e-01
6.65004134e-01 3.24698836e-01 2.32539758e-01 1.08960904e-01
1.04865813e+00 -4.27646562e-02 -5.27185082e-01 -6.27414227e-01
4.38072681e-01 4.07260031e-01 -4.29256618e-01 -3.15513104e-01
-9.67645586e-01 1.02254021e+00 3.65889162e-01 1.01220019e-01
-2.87144840e-01 -3.00507307e-01 5.78377843e-01 7.62322918e-03
1.02845919e+00 1.84965834e-01 -8.06216672e-02 -5.28465360e-02
-1.33484638e+00 7.93865919e-01 6.40438259e-01 1.15594614e+00
7.05643237e-01 1.21112294e-01 -2.29850814e-01 9.37202692e-01
9.74615291e-03 6.25524282e-01 9.08498287e-01 -6.72851682e-01
7.38787949e-01 7.13750362e-01 2.46188149e-01 -1.13314450e+00
-2.17588991e-01 -5.80258369e-01 -1.36298299e+00 -9.67054516e-02
3.71561676e-01 1.29329130e-01 -4.83159900e-01 9.97332573e-01
7.14949667e-01 9.03948620e-02 -2.15761378e-01 7.21850395e-01
7.30090082e-01 6.88625395e-01 -3.07600319e-01 -5.56102753e-01
1.26388705e+00 -8.09311152e-01 -1.18192363e+00 -1.18983053e-02
3.84567469e-01 -6.12317860e-01 1.30355179e+00 6.84997499e-01
-1.14292574e+00 -4.89459366e-01 -1.24956512e+00 1.71563163e-01
-3.23076755e-01 -7.22476318e-02 6.02135122e-01 1.05147731e+00
-3.96233737e-01 4.76499468e-01 -5.48467815e-01 3.12190503e-01
9.06641245e-01 -8.30526799e-02 3.21920693e-01 -4.66345400e-01
-1.36597884e+00 2.77129233e-01 4.83537048e-01 2.15843931e-01
-2.80844897e-01 -7.58767188e-01 -7.34219611e-01 -2.69625515e-01
1.77833855e-01 -6.16014719e-01 7.66651571e-01 -7.24233389e-01
-1.37513649e+00 4.67851669e-01 -4.07370627e-01 -4.14585084e-01
8.74627888e-01 -1.42208636e-01 -1.64527223e-01 -5.90917803e-02
9.15026590e-02 -7.38993213e-02 6.69215739e-01 -1.09590769e+00
-2.82419503e-01 -6.32713497e-01 -4.56040263e-01 2.00451404e-01
-3.65662336e-01 -2.74824172e-01 2.12526783e-01 -1.07020032e+00
5.96980989e-01 -5.55354834e-01 -4.58493471e-01 -2.80112654e-01
-6.15949929e-01 4.81878631e-02 6.87400818e-01 -2.14527145e-01
1.53638256e+00 -1.91325247e+00 -3.57067704e-01 2.19176486e-01
2.42908329e-01 3.19666624e-01 3.40623856e-01 3.64945531e-01
9.58835781e-02 2.99849093e-01 -3.75532627e-01 -1.93020776e-01
-1.07728593e-01 2.71802507e-02 -4.57063437e-01 7.03542769e-01
-2.38932878e-01 6.83975935e-01 -1.04127860e+00 -5.68789840e-01
3.70747480e-03 2.18147114e-01 -6.99980676e-01 -8.42437148e-03
4.65701744e-02 2.47810464e-02 -5.46726108e-01 9.67822015e-01
1.47241008e+00 -2.70847976e-01 -2.13363364e-01 -2.40434140e-01
2.13371783e-01 1.57279953e-01 -1.74325895e+00 9.37280357e-01
-2.06742272e-01 7.73063824e-02 -1.14012904e-01 -1.25976861e+00
1.35771096e+00 8.74948502e-02 4.88045245e-01 -1.96723416e-01
-7.53751621e-02 6.72896326e-01 -5.30533314e-01 -6.15093052e-01
7.05803573e-01 -5.06701529e-01 1.29626378e-01 4.22679991e-01
-5.57451189e-01 -4.44302261e-01 5.40070944e-02 -1.69346899e-01
3.46139401e-01 -1.79170370e-01 6.75157130e-01 -5.80492496e-01
4.47703838e-01 5.62352724e-02 1.00920975e+00 7.72187591e-01
-2.19526127e-01 9.48235631e-01 3.79915059e-01 -6.54556036e-01
-1.01318622e+00 -6.51800096e-01 -6.43018782e-01 3.81092995e-01
5.01221180e-01 -2.56265879e-01 -6.81313336e-01 -4.26605791e-01
3.08208257e-01 5.49606621e-01 -6.17754221e-01 -1.53317362e-01
-3.01876664e-01 -1.55206728e+00 4.86199081e-01 3.30595046e-01
9.14031863e-01 -5.69739401e-01 -1.28471240e-01 2.72157997e-01
-5.15302122e-01 -7.02855945e-01 -3.96343112e-01 -2.55951464e-01
-1.24371755e+00 -1.02382600e+00 -8.57128203e-01 -3.48774552e-01
5.33803284e-01 4.98576641e-01 1.11092806e+00 1.92605192e-03
-1.86818078e-01 -5.27526796e-01 -5.01849890e-01 -8.49483967e-01
-6.43198937e-02 1.80229694e-01 1.37891322e-01 1.71060547e-01
6.82937622e-01 -5.16166687e-01 -9.44716632e-01 6.14413321e-01
-1.07836962e+00 -1.74402475e-01 3.00054431e-01 1.19955349e+00
9.08311784e-01 4.39945698e-01 9.33388293e-01 -1.15472317e+00
7.40770280e-01 -6.58180356e-01 -2.56613821e-01 -6.28296509e-02
-6.31007016e-01 -3.87844861e-01 7.00839639e-01 -6.69556618e-01
-8.28868508e-01 -5.71074843e-01 -1.52007103e-01 -3.14699292e-01
1.94210589e-01 2.89619416e-01 -4.92910832e-01 9.25545618e-02
7.19282448e-01 3.07491660e-01 1.94442838e-01 -5.63093781e-01
-9.41250250e-02 1.15411770e+00 -2.58245498e-01 -7.01177001e-01
3.72920930e-01 9.23171937e-01 -8.23627412e-02 -7.24795818e-01
-1.06576526e+00 -2.36992419e-01 -1.57777667e-01 -2.08688527e-02
1.43617034e-01 -9.97479200e-01 -6.36007488e-01 7.37940490e-01
-6.80796087e-01 -7.64821842e-02 -7.61732399e-01 6.88686728e-01
-6.74006402e-01 2.55858481e-01 -4.12977010e-01 -1.19842112e+00
-2.69937426e-01 -1.28426886e+00 1.11637866e+00 1.66705117e-01
1.01170801e-01 -8.08958471e-01 -1.90682456e-01 5.09276569e-01
5.06120622e-01 4.11994755e-01 5.62849224e-01 -6.00903273e-01
-4.09493297e-01 -5.15579820e-01 -1.01219364e-01 4.53016192e-01
1.21388778e-01 6.18068762e-02 -8.45841885e-01 -4.22074795e-01
6.45810425e-01 -4.69009519e-01 7.80475557e-01 7.38148332e-01
1.76477242e+00 -5.70386827e-01 -4.11114931e-01 8.66048932e-01
1.55527163e+00 -4.01570387e-02 6.87033057e-01 5.57837665e-01
3.33065152e-01 4.82294649e-01 1.10842812e+00 8.38541746e-01
3.60398799e-01 4.25633937e-01 3.33302140e-01 -1.14900090e-01
2.10601673e-01 -4.03606445e-01 -8.92556757e-02 9.02588487e-01
1.00583635e-01 4.10412140e-02 -4.77096260e-01 5.25249243e-01
-1.61328912e+00 -1.17393386e+00 -3.67448896e-01 2.32835698e+00
1.21254909e+00 9.96640623e-02 3.86516362e-01 7.54417419e-01
1.05703175e+00 4.92474914e-01 -7.35541224e-01 -1.31892219e-01
-3.95790368e-01 -1.97076976e-01 5.93069375e-01 3.54582191e-01
-8.95146728e-01 1.58147171e-01 6.86438322e+00 1.70499492e+00
-6.75048351e-01 9.12303850e-02 1.21912992e+00 -3.35149795e-01
-6.33874476e-01 -2.61638731e-01 -1.29057074e+00 1.12604749e+00
2.52895892e-01 -2.71827072e-01 2.94942886e-01 9.00685072e-01
2.21769437e-01 -2.75627315e-01 -6.13470674e-01 1.05017388e+00
-4.13148887e-02 -1.31542540e+00 2.13286579e-01 4.37415205e-02
1.07538056e+00 -3.87357742e-01 5.02637736e-02 3.35791081e-01
1.75067887e-01 -1.10870993e+00 6.10044241e-01 6.03319049e-01
7.75072992e-01 -9.61931825e-01 1.10406995e+00 9.00968969e-01
-9.68598962e-01 -5.99953122e-02 -8.56288314e-01 -2.91518331e-01
9.71568227e-02 1.73371446e+00 -3.33341777e-01 5.82520068e-01
6.63513303e-01 5.86907506e-01 -2.23459378e-01 1.16498601e+00
2.11620688e-01 6.23489141e-01 -5.80886662e-01 -4.93981481e-01
-8.40043053e-02 -4.38262641e-01 6.14976704e-01 8.13149810e-01
4.66905028e-01 3.03514972e-02 1.95344195e-01 7.67547071e-01
1.70976698e-01 3.96480024e-01 -6.43293798e-01 4.97853398e-01
9.98902440e-01 6.13350034e-01 -4.90242451e-01 -7.77215004e-01
3.58865187e-02 2.30472326e-01 5.81816547e-02 3.36084664e-01
-7.65589535e-01 -7.14553177e-01 4.09683764e-01 4.68525320e-01
1.91221878e-01 2.89424628e-01 -8.73259544e-01 -1.36490083e+00
3.40956837e-01 -1.30825889e+00 2.72185087e-01 -4.28687483e-01
-1.46643639e+00 5.68058550e-01 1.75782651e-01 -1.76828671e+00
2.02876270e-01 -1.43465683e-01 -3.05505127e-01 9.48482156e-01
-1.47177875e+00 -5.25962770e-01 -4.67790157e-01 4.09951806e-01
5.51443517e-01 -6.30982667e-02 4.63273138e-01 3.07918936e-01
-4.88768458e-01 6.63605392e-01 6.14459515e-01 -2.97813207e-01
4.02088881e-01 -9.63385344e-01 2.45757662e-02 3.33686173e-01
-5.37929356e-01 5.53866744e-01 7.58300662e-01 -6.56348586e-01
-1.00015175e+00 -1.29399860e+00 6.37219489e-01 6.25078976e-02
2.47562677e-01 -3.63737136e-01 -9.49778020e-01 3.18689644e-01
-3.18187505e-01 4.40374643e-01 9.81529117e-01 -7.09604323e-02
-7.88156614e-02 -4.73870516e-01 -1.76993024e+00 3.34410697e-01
8.94084394e-01 -4.18531001e-02 -4.56698447e-01 7.21112549e-01
6.18955195e-01 -5.90927064e-01 -1.00192595e+00 6.84045851e-01
5.29079795e-01 -1.41423476e+00 1.14904487e+00 -3.92908186e-01
3.28754336e-01 -3.77159953e-01 -3.11135173e-01 -1.34023821e+00
-5.55118583e-02 -5.54633915e-01 -3.04319412e-01 1.17896354e+00
-4.20759507e-02 -1.18667603e+00 1.02503085e+00 3.23897824e-02
3.99834186e-01 -1.41557205e+00 -9.47287738e-01 -1.09695685e+00
1.64640650e-01 -3.00048709e-01 1.42705953e+00 8.78420770e-01
1.28676310e-01 -2.46060520e-01 -4.30133611e-01 -1.72739223e-01
1.05592752e+00 3.55776072e-01 7.59778678e-01 -1.14810824e+00
-2.52714217e-01 -3.57138336e-01 -2.56843746e-01 -1.40264165e+00
-3.17207247e-01 -6.67187631e-01 -2.69753098e-01 -1.18170190e+00
2.19747171e-01 -9.05653954e-01 1.21269487e-01 -3.36576819e-01
-5.98190546e-01 4.73028421e-01 -1.91762015e-01 6.63957536e-01
-3.20868373e-01 9.44136024e-01 1.60536265e+00 -1.93805397e-01
-2.20032245e-01 5.40394545e-01 -8.02511632e-01 7.54874468e-01
8.94128621e-01 -4.62631136e-01 -4.22324866e-01 5.62550407e-03
2.65213281e-01 -2.88599776e-03 -5.92672601e-02 -1.01681328e+00
-9.41284746e-02 -3.70121211e-01 4.96644884e-01 -1.13678241e+00
2.76164353e-01 -6.58953011e-01 1.24764234e-01 4.59246784e-01
-2.49408752e-01 -1.59277335e-01 -1.75423309e-01 6.79111242e-01
-5.64007819e-01 -4.11306679e-01 9.71924603e-01 -2.99650133e-01
3.72657239e-01 4.49856371e-01 -1.17646225e-01 5.12991309e-01
1.05831265e+00 -5.69621503e-01 -1.95668072e-01 -2.52892435e-01
-4.20914024e-01 2.43756443e-01 5.40165961e-01 -3.09862256e-01
6.20158494e-01 -1.89241970e+00 -7.24049330e-01 5.43715239e-01
-1.11105971e-01 6.50246441e-01 3.89815837e-01 9.31759417e-01
-6.72498524e-01 1.19575799e-01 2.21233174e-01 -7.28772342e-01
-7.59235978e-01 6.02200985e-01 3.33263189e-01 -4.15346175e-01
-4.48526263e-01 7.65619755e-01 -1.41529948e-01 -7.46921003e-01
2.15169549e-01 -4.37722802e-01 -7.50827864e-02 2.42327407e-01
9.39413667e-01 9.07706976e-01 1.36682212e-01 -2.08678603e-01
-1.16475388e-01 4.43384439e-01 3.47552747e-02 2.10230172e-01
1.28926969e+00 -1.15273707e-01 -3.00349653e-01 8.59535277e-01
1.18944240e+00 1.86859012e-01 -8.31092179e-01 -3.88536751e-01
-5.13744831e-01 -1.12553394e+00 -1.59308985e-01 -2.09852457e-01
-1.11667347e+00 7.10571527e-01 6.17218427e-02 8.59358549e-01
1.12613893e+00 -5.30140460e-01 8.21533978e-01 2.83076265e-03
5.92572927e-01 -1.21336401e+00 -1.98047593e-01 1.49650455e-01
7.76624441e-01 -1.03362429e+00 6.70059323e-01 -6.31321549e-01
-4.65566874e-01 8.76214802e-01 5.01183033e-01 -4.63301331e-01
9.48433876e-01 4.41086143e-01 -2.13254809e-01 7.54773989e-02
-4.91008312e-01 2.79808104e-01 -1.24945894e-01 6.51850700e-01
3.28865528e-01 1.09284684e-01 -9.17249203e-01 7.05553532e-01
-3.80565912e-01 3.68260771e-01 4.94533062e-01 6.99896216e-01
-5.35150945e-01 -7.43987262e-01 -9.48808491e-01 8.66525888e-01
-4.36177105e-01 -1.20515907e-02 3.02193403e-01 7.12807357e-01
3.11798036e-01 7.11570859e-01 1.03345171e-01 -2.37845570e-01
3.49996865e-01 -1.77207664e-01 1.41937703e-01 -2.94725329e-01
-1.96431443e-01 -8.25537369e-02 -2.96211839e-01 -4.43179965e-01
-4.13470924e-01 -8.61642182e-01 -6.47590399e-01 -5.35291910e-01
-7.91562676e-01 5.20824850e-01 5.46323359e-01 6.24465287e-01
2.12227508e-01 4.06853855e-01 9.84374762e-01 -9.09182072e-01
-1.06091976e+00 -1.08761370e+00 -1.22031069e+00 3.04425687e-01
3.40792000e-01 -8.30486536e-01 -8.34266245e-01 -4.92919028e-01] | [8.819238662719727, 4.086338520050049] |
a2599866-35a3-40ed-9bbb-51fd5af27d95 | on-hyperbolic-embeddings-in-2d-object | 2203.08049 | null | https://arxiv.org/abs/2203.08049v3 | https://arxiv.org/pdf/2203.08049v3.pdf | On Hyperbolic Embeddings in 2D Object Detection | Object detection, for the most part, has been formulated in the euclidean space, where euclidean or spherical geodesic distances measure the similarity of an image region to an object class prototype. In this work, we study whether a hyperbolic geometry better matches the underlying structure of the object classification space. We incorporate a hyperbolic classifier in two-stage, keypoint-based, and transformer-based object detection architectures and evaluate them on large-scale, long-tailed, and zero-shot object detection benchmarks. In our extensive experimental evaluations, we observe categorical class hierarchies emerging in the structure of the classification space, resulting in lower classification errors and boosting the overall object detection performance. | ['Abhinav Valada', 'Alexander Braun', 'Christopher Lang'] | 2022-03-15 | null | null | null | null | ['zero-shot-object-detection'] | ['computer-vision'] | [-9.83506218e-02 -1.93170741e-01 -2.44090259e-01 -4.89401907e-01
-5.04732072e-01 -8.40763032e-01 7.48408258e-01 6.62713587e-01
-5.98606408e-01 -1.50168970e-01 -1.51063930e-02 -2.27679074e-01
-4.34433311e-01 -8.17905366e-01 -2.42433384e-01 -4.35723215e-01
-6.51299506e-02 2.63989449e-01 8.98116410e-01 7.30796084e-02
5.45616508e-01 7.21893907e-01 -1.39425218e+00 3.88693631e-01
2.16706634e-01 1.36636496e+00 -1.46682501e-01 8.27322721e-01
2.11851642e-01 4.15634602e-01 -3.13538909e-01 -5.41217089e-01
2.65146732e-01 -1.64740756e-01 -6.79004490e-01 -6.88666478e-02
7.43498266e-01 -6.10798784e-02 -3.36040139e-01 9.96547043e-01
1.80195138e-01 -5.50970621e-02 1.06554437e+00 -1.44183278e+00
-7.31401265e-01 3.02461028e-01 -4.60513592e-01 5.99765718e-01
2.02148899e-01 9.98917446e-02 1.34107888e+00 -1.30042410e+00
5.70688367e-01 1.17656076e+00 8.71803164e-01 2.22004950e-02
-1.33607686e+00 -3.58569533e-01 4.42081783e-03 3.37787926e-01
-1.83830893e+00 -5.37020452e-02 5.98624885e-01 -5.80409527e-01
7.17310011e-01 3.20127904e-01 7.28936851e-01 6.92114770e-01
4.59307969e-01 6.75235510e-01 8.20445657e-01 -2.92708158e-01
3.56172532e-01 2.43953705e-01 5.14232218e-01 6.86827481e-01
3.76449913e-01 1.43407136e-01 -1.44227684e-01 -2.91186988e-01
4.30299759e-01 1.12487346e-01 -4.11411971e-02 -8.82414520e-01
-9.58365023e-01 1.05220914e+00 1.01372969e+00 2.34491259e-01
-2.42290273e-01 4.18510847e-02 3.13977301e-01 2.54981574e-02
3.81822765e-01 6.40921772e-01 1.31127732e-02 7.72581324e-02
-7.58906662e-01 2.02527776e-01 5.78753173e-01 1.06365359e+00
5.84195137e-01 -5.16131103e-01 -4.62009043e-01 4.52451199e-01
2.94202834e-01 1.72856823e-01 2.97360331e-01 -7.69901812e-01
7.53425956e-02 9.62369680e-01 -6.08036332e-02 -1.10935402e+00
-6.18554533e-01 -8.83363724e-01 -2.88856119e-01 1.83046341e-01
4.10333186e-01 4.63313669e-01 -5.27321041e-01 1.31914842e+00
5.63164651e-01 -7.56272376e-02 -5.00822626e-02 8.85331869e-01
6.62770092e-01 4.34802711e-01 7.37469792e-02 1.93697155e-01
1.67672479e+00 -5.92126429e-01 -1.58864036e-02 1.06517963e-01
8.26201916e-01 -5.00678360e-01 1.03956985e+00 3.32714207e-02
-7.36279190e-01 -5.21852851e-01 -1.00519896e+00 -2.25605801e-01
-6.01733327e-01 3.56739163e-02 2.39109099e-01 8.11319232e-01
-6.78291440e-01 3.63797039e-01 -7.13425040e-01 -5.38948417e-01
6.79899693e-01 -1.15010180e-01 -1.81246489e-01 -1.55842423e-01
-5.81864655e-01 7.37201154e-01 5.36612272e-01 -3.52408856e-01
-6.34140670e-01 -8.48032475e-01 -7.17719674e-01 1.83911085e-01
3.37663367e-02 -3.17579478e-01 1.11680949e+00 -3.46069962e-01
-7.98839629e-01 9.66782093e-01 2.05655679e-01 -5.92757702e-01
6.59856081e-01 7.35490918e-02 -1.19642884e-01 6.51467741e-02
7.86646977e-02 9.06009316e-01 8.44725847e-01 -9.61533606e-01
-1.08343053e+00 -7.84741163e-01 9.73512605e-02 1.43335149e-01
-4.54645365e-01 2.14492083e-01 -1.18674517e-01 -5.83188415e-01
4.58641797e-01 -1.11209476e+00 3.19519192e-02 4.98081088e-01
-3.00982058e-01 -8.16934943e-01 1.04359031e+00 1.15351751e-01
9.45262551e-01 -2.60158348e+00 -1.69584900e-01 2.28260204e-01
5.03802478e-01 -8.27338174e-02 -9.50096250e-02 2.31721088e-01
1.76316395e-01 -3.00073046e-02 -3.02046333e-02 -1.36612847e-01
5.94293736e-02 -1.55844778e-01 -5.54440916e-01 9.44168031e-01
3.20740044e-01 9.92347479e-01 -7.69082129e-01 -5.83986521e-01
-9.61716920e-02 2.44632408e-01 -5.63137591e-01 5.23769520e-02
3.79103720e-02 -6.00029379e-02 -2.20793545e-01 5.43252170e-01
4.65782046e-01 -2.39837080e-01 -3.55750620e-01 -2.69447953e-01
-1.81084603e-01 2.06471812e-02 -1.10993958e+00 1.24199080e+00
-7.55281001e-02 1.02776539e+00 -4.01476115e-01 -6.64659619e-01
9.74554241e-01 -2.21718818e-01 4.07841921e-01 -8.34771991e-01
3.50198150e-02 7.40282908e-02 2.75810093e-01 -1.10999927e-01
7.55189300e-01 2.63209432e-01 6.47908524e-02 3.35033238e-01
-1.12700433e-01 -5.26444279e-02 2.09895939e-01 3.36767823e-01
1.00593221e+00 -2.99741656e-01 4.45206136e-01 -6.23063743e-01
1.54838473e-01 2.77233303e-01 1.61616117e-01 1.08963645e+00
-4.19489235e-01 7.76398063e-01 4.97066557e-01 -4.63603586e-01
-1.13090861e+00 -1.41709185e+00 -8.89456213e-01 1.44004667e+00
4.45074350e-01 -5.28483391e-01 -5.66865385e-01 -6.99521065e-01
2.09791198e-01 7.23835468e-01 -7.04331398e-01 -4.85476136e-01
-5.01821935e-01 -6.52649939e-01 6.67567313e-01 7.62495935e-01
2.24655271e-01 -7.06318319e-01 -1.13290930e+00 2.74921730e-02
2.69903660e-01 -1.14046955e+00 -7.38900840e-01 2.69393325e-01
-7.60119855e-01 -1.29887915e+00 -5.24889827e-01 -7.03678787e-01
4.22746718e-01 4.84262198e-01 9.35599625e-01 -1.07212596e-01
-1.00914299e+00 4.81196731e-01 -5.15983701e-01 -5.62147856e-01
-1.97304003e-02 1.39084771e-01 8.16956758e-02 1.29493117e-01
4.29737955e-01 1.25263363e-01 -8.47068250e-01 7.01152921e-01
-6.22442484e-01 -3.90114695e-01 3.96020234e-01 3.92487526e-01
4.23243403e-01 -4.14721146e-02 2.03176603e-01 -1.90161258e-01
2.83837855e-01 -4.38045979e-01 -6.63590670e-01 2.12085754e-01
-7.03704596e-01 -9.50110853e-02 1.79708883e-01 -6.19377792e-01
-5.14875948e-01 1.14094997e-02 2.63752311e-01 -3.77728939e-01
-8.58698506e-03 1.28151193e-01 2.69211948e-01 -1.62348315e-01
1.00062263e+00 2.81052262e-01 -2.24990323e-01 -3.25470209e-01
5.78091621e-01 6.99518204e-01 4.64258581e-01 -2.86878258e-01
7.60568917e-01 7.54780710e-01 1.00953281e-01 -7.69864202e-01
-9.77785707e-01 -9.61743474e-01 -8.03010225e-01 -2.98176873e-02
9.08398390e-01 -7.34327972e-01 -6.36254132e-01 1.57733127e-01
-1.07423663e+00 -4.07167152e-02 -6.78378105e-01 5.73111892e-01
-4.08352256e-01 1.38965081e-02 -3.44579369e-01 -8.21567953e-01
-1.49892017e-01 -9.39466178e-01 1.28101504e+00 1.92978621e-01
-1.96961269e-01 -6.98534310e-01 1.14076599e-01 2.08760099e-03
3.07691365e-01 -2.00748816e-01 1.00026309e+00 -9.80552793e-01
-5.67051589e-01 -5.93509674e-01 -7.08737075e-01 -1.14544705e-02
-2.55773902e-01 6.76471442e-02 -7.12056279e-01 -3.46502870e-01
-1.25191495e-01 -9.02608782e-02 9.03674185e-01 2.57952392e-01
1.06394684e+00 -2.32832864e-01 -4.06417966e-01 5.55488706e-01
1.28025424e+00 -6.82317987e-02 2.95617938e-01 2.51281351e-01
4.59183842e-01 6.54240668e-01 7.17900693e-01 4.76547211e-01
3.63070011e-01 8.67459059e-01 3.63183469e-01 5.50284684e-02
-1.30446911e-01 -1.27970308e-01 9.14802998e-02 6.16954491e-02
5.59537709e-01 -1.02389604e-01 -1.15320933e+00 4.33111489e-01
-1.75369096e+00 -8.82383466e-01 -4.37536597e-01 2.10385823e+00
5.33983350e-01 3.45516086e-01 6.08737409e-01 3.14977430e-02
6.91735327e-01 -2.46875197e-01 -5.27290702e-01 -9.20565799e-02
-6.77972883e-02 -2.45837793e-01 4.87349778e-01 1.19459435e-01
-1.31467319e+00 7.25820303e-01 6.59958744e+00 7.26643920e-01
-1.02998722e+00 3.29364650e-02 4.95069712e-01 -6.77967370e-02
4.88842964e-01 -1.58875495e-01 -1.04312241e+00 1.72404557e-01
5.20502865e-01 -3.45031917e-01 -1.56158388e-01 1.28382194e+00
-2.41887793e-01 1.12718873e-01 -1.73960173e+00 9.64034140e-01
1.04031526e-01 -1.28452122e+00 -7.11717550e-03 2.08917424e-01
3.88808697e-01 2.57663250e-01 3.19159091e-01 3.60598505e-01
8.73624906e-02 -8.60629857e-01 1.20617044e+00 2.07587436e-01
5.01001537e-01 -4.80472684e-01 4.71970409e-01 4.02161181e-01
-1.53145850e+00 -4.82268274e-01 -3.78788799e-01 1.86464429e-01
-4.01089966e-01 1.63029119e-01 -9.85091269e-01 9.05841142e-02
8.74507964e-01 4.40667957e-01 -1.20037544e+00 1.66236007e+00
3.71851325e-01 4.49076504e-01 -4.47518885e-01 -2.70903140e-01
4.88232464e-01 1.99346803e-02 7.66385436e-01 1.42502522e+00
8.55722725e-02 2.98514068e-01 4.98755723e-01 1.10744488e+00
2.83115078e-02 2.45493874e-01 -5.81626236e-01 2.41743803e-01
3.70563805e-01 1.38110924e+00 -1.10696685e+00 -2.03741312e-01
-3.39020640e-01 4.72674310e-01 2.20726699e-01 8.10527578e-02
-8.64764512e-01 -6.04677320e-01 6.12839103e-01 3.11203271e-01
2.81660855e-01 -3.13637555e-01 -4.19245631e-01 -7.27905929e-01
3.47315371e-02 -2.45784640e-01 7.59247243e-01 -6.79787457e-01
-1.13804626e+00 4.06765431e-01 6.65180236e-02 -1.14817488e+00
4.99037951e-02 -8.12121689e-01 -6.14594638e-01 3.39553893e-01
-1.04619098e+00 -9.63497877e-01 -5.46545029e-01 4.66164589e-01
4.48222667e-01 -3.85401770e-02 4.51595068e-01 1.54990656e-02
-4.02422309e-01 5.85106075e-01 2.42624134e-01 5.23750722e-01
4.37417954e-01 -1.38625968e+00 4.15500700e-01 5.61752796e-01
6.35581493e-01 5.28626621e-01 6.18537247e-01 -3.70270044e-01
-1.36798573e+00 -1.41699398e+00 3.67082626e-01 -8.80800247e-01
8.62609982e-01 -5.79550922e-01 -9.96186197e-01 4.76484835e-01
-5.00470400e-01 5.39973676e-01 5.21692336e-01 1.37395024e-01
-1.00489748e+00 -1.52248338e-01 -1.05692387e+00 4.80140358e-01
1.12913918e+00 -6.32052720e-01 -5.39948702e-01 4.88245606e-01
6.29750252e-01 -1.31511781e-03 -8.76512468e-01 4.42047745e-01
7.09331930e-01 -6.94516420e-01 9.95298386e-01 -9.72327828e-01
8.48778188e-02 -2.76022732e-01 -5.08204758e-01 -9.64762568e-01
-7.43120909e-01 -3.42320576e-02 2.38898173e-02 8.11594009e-01
2.68197179e-01 -4.24980849e-01 9.09637809e-01 2.12468028e-01
1.41143529e-02 -6.99000776e-01 -1.11978054e+00 -1.22420359e+00
2.77332634e-01 -3.68622124e-01 2.46312618e-01 4.68668222e-01
-1.01351336e-01 3.46828133e-01 5.31447470e-01 2.82248765e-01
9.03042197e-01 4.53592211e-01 6.02788925e-01 -1.52870786e+00
4.79329079e-02 -9.62793469e-01 -1.27275503e+00 -8.68523240e-01
-1.30831450e-01 -1.04555142e+00 1.05492622e-01 -1.19447601e+00
4.93487358e-01 -5.65113902e-01 -3.65489960e-01 1.03695944e-01
3.51287723e-02 7.18049109e-01 2.81399399e-01 5.32940567e-01
-1.03205931e+00 5.30738592e-01 5.57474494e-01 -2.77977198e-01
-3.46722938e-02 -7.24598765e-02 -4.28287715e-01 5.55135727e-01
5.01091182e-01 -6.78000689e-01 -2.19844189e-02 -2.39452839e-01
9.10342336e-02 -4.87301946e-01 8.08651030e-01 -1.20929730e+00
5.09572327e-01 -3.38349380e-02 3.16135973e-01 -4.24180746e-01
2.70648777e-01 -7.83733606e-01 -2.92639732e-01 7.45703638e-01
-5.73154032e-01 -3.02189030e-02 -1.35382563e-01 7.20757484e-01
7.59782493e-02 -1.06794983e-01 1.39734888e+00 3.60958636e-01
-6.38813853e-01 3.87716889e-01 4.02126461e-02 1.95774496e-01
1.36032915e+00 -4.83571172e-01 -4.33642179e-01 1.56065539e-01
-5.37371695e-01 4.99453917e-02 4.45418835e-01 6.99739814e-01
5.96477985e-01 -1.27528059e+00 -8.47783327e-01 -9.40309372e-03
8.00519943e-01 -3.47879827e-01 -4.02830929e-01 9.38656926e-01
-2.85900474e-01 5.66746473e-01 -6.61839545e-02 -1.26493108e+00
-1.35619640e+00 7.33123124e-01 5.83481848e-01 3.90072703e-01
-8.30604196e-01 7.79741943e-01 5.93114376e-01 -3.07202756e-01
4.01768863e-01 -4.23464924e-01 -4.79258709e-02 2.77953804e-01
4.75275248e-01 6.69161320e-01 1.34498281e-02 -6.58001840e-01
-5.13109326e-01 5.40448546e-01 -1.91337749e-01 -2.65382323e-02
8.96105528e-01 2.43610591e-02 1.80272877e-01 5.43308854e-01
1.40315628e+00 -5.26139200e-01 -1.11895239e+00 -5.51150978e-01
3.95586222e-01 -5.11727393e-01 9.02337208e-02 -3.41359198e-01
-5.06011188e-01 6.74153984e-01 1.09075332e+00 6.64887249e-01
4.58218426e-01 6.01552665e-01 2.10675135e-01 7.07755208e-01
3.72780770e-01 -9.35753226e-01 3.94411653e-01 4.36807603e-01
8.17593932e-01 -1.28584361e+00 2.84594819e-02 -4.56221581e-01
-2.36348897e-01 1.09361184e+00 5.81278563e-01 -2.17579171e-01
8.51610184e-01 2.80920602e-02 -4.69520569e-01 -4.07901645e-01
-5.54004848e-01 -2.85962552e-01 5.71256340e-01 4.41267192e-01
8.85973126e-02 4.21474129e-02 1.98261037e-01 3.38017255e-01
-3.08718592e-01 -4.79834259e-01 1.86319754e-01 8.02037239e-01
-9.28243339e-01 -1.28959715e-01 -3.58348042e-01 3.63253057e-01
-1.60125583e-01 1.43723702e-03 -5.43481410e-01 7.37675905e-01
-5.44622950e-02 9.16646481e-01 6.61780596e-01 -3.11417002e-02
4.88413274e-01 -1.81738377e-01 4.83862817e-01 -7.21477330e-01
-3.58181626e-01 -3.48471910e-01 -4.22812372e-01 -6.44411385e-01
3.52189600e-01 -7.33850479e-01 -1.17693233e+00 3.23994756e-02
-5.89864254e-01 6.71307296e-02 7.81973302e-01 7.37466812e-01
4.02242064e-01 -2.31409795e-03 6.41718090e-01 -4.80653346e-01
-1.13482106e+00 -7.44246781e-01 -4.79757637e-01 4.25482154e-01
2.78101027e-01 -5.87618113e-01 -1.49469897e-01 -2.89784104e-01] | [9.484905242919922, 1.5046013593673706] |
f3f96753-59d6-4c76-bae4-1e6c220999c0 | rethinking-the-trigger-injecting-position-in | 2304.02277 | null | https://arxiv.org/abs/2304.02277v2 | https://arxiv.org/pdf/2304.02277v2.pdf | Rethinking the Trigger-injecting Position in Graph Backdoor Attack | Backdoor attacks have been demonstrated as a security threat for machine learning models. Traditional backdoor attacks intend to inject backdoor functionality into the model such that the backdoored model will perform abnormally on inputs with predefined backdoor triggers and still retain state-of-the-art performance on the clean inputs. While there are already some works on backdoor attacks on Graph Neural Networks (GNNs), the backdoor trigger in the graph domain is mostly injected into random positions of the sample. There is no work analyzing and explaining the backdoor attack performance when injecting triggers into the most important or least important area in the sample, which we refer to as trigger-injecting strategies MIAS and LIAS, respectively. Our results show that, generally, LIAS performs better, and the differences between the LIAS and MIAS performance can be significant. Furthermore, we explain these two strategies' similar (better) attack performance through explanation techniques, which results in a further understanding of backdoor attacks in GNNs. | ['Stjepan Picek', 'Gorka Abad', 'Jing Xu'] | 2023-04-05 | null | null | null | null | ['backdoor-attack'] | ['adversarial'] | [ 1.37932360e-01 6.08197808e-01 -4.56716806e-01 1.25747621e-01
5.23447841e-02 -1.11442459e+00 6.05468929e-01 1.68551907e-01
8.93178731e-02 3.59149009e-01 -2.77251959e-01 -1.16488814e+00
-1.17527701e-01 -1.21874523e+00 -1.14475799e+00 -5.51297128e-01
-2.76317149e-01 -1.16652898e-01 6.33275330e-01 -4.16906595e-01
9.82849523e-02 7.97806323e-01 -1.08916128e+00 4.18483436e-01
2.20723107e-01 2.92171031e-01 -5.98322809e-01 8.01265180e-01
-5.31536080e-02 7.62196839e-01 -1.09540665e+00 -7.76898086e-01
6.22609138e-01 -4.85306621e-01 -6.11733258e-01 -8.49390149e-01
7.27155328e-01 -1.29666939e-01 -7.79582441e-01 1.34739780e+00
1.17171422e-01 -3.22085291e-01 9.90214571e-02 -1.58951080e+00
-4.01039362e-01 1.41733944e+00 -4.20649946e-01 4.13781792e-01
2.10179135e-01 1.76504001e-01 9.71699893e-01 -1.48712844e-01
5.04081368e-01 1.15047824e+00 6.08472288e-01 8.37596357e-01
-1.24149477e+00 -1.09244847e+00 3.81209910e-01 4.52858359e-02
-1.05648732e+00 5.35632782e-02 9.20688093e-01 7.32962275e-03
1.16722023e+00 8.19234192e-01 5.84894717e-01 1.57376552e+00
6.64650738e-01 4.94797498e-01 8.76237631e-01 -4.92334783e-01
2.15122342e-01 2.80990779e-01 7.32009709e-01 7.17746556e-01
1.03989196e+00 8.42077971e-01 -5.05475044e-01 -5.34725547e-01
5.91801941e-01 5.33104725e-02 -3.19481790e-01 -1.50700182e-01
-5.71453094e-01 8.97817552e-01 6.07039332e-01 3.15703958e-01
-8.03454686e-03 6.01447999e-01 5.00370860e-01 5.18375516e-01
-1.14374058e-02 7.38788664e-01 -4.69504982e-01 1.33708417e-01
-7.92386889e-01 2.57105857e-01 1.00672948e+00 6.63307011e-01
5.96044660e-01 3.81804079e-01 -1.63693070e-01 -2.31530473e-01
2.49942780e-01 2.56232798e-01 8.86339769e-02 -1.12986393e-01
6.45179927e-01 5.99700451e-01 -6.49237216e-01 -1.22575569e+00
-3.44989240e-01 -9.72569644e-01 -3.98099571e-01 3.00299287e-01
4.22062904e-01 -3.66327286e-01 -1.22963679e+00 1.86702204e+00
8.63247663e-02 6.21481776e-01 -9.67189111e-03 3.85679007e-01
7.32172489e-01 5.75210214e-01 -1.32461593e-01 2.33594522e-01
1.23981893e+00 -8.06471586e-01 -7.48092830e-01 -4.46254045e-01
7.59538114e-01 -1.68415889e-01 1.09926116e+00 4.79427874e-01
-6.58326685e-01 -1.52896345e-01 -1.61025965e+00 5.80700219e-01
-1.01346934e+00 -4.33524460e-01 1.13206053e+00 1.76577926e+00
-8.94398510e-01 7.36112714e-01 -9.21189249e-01 5.23192920e-02
2.81471193e-01 5.44146359e-01 -4.04721677e-01 2.96276391e-01
-1.50723326e+00 5.96583784e-01 6.63814306e-01 -1.63227007e-01
-1.52406919e+00 -8.78095269e-01 -7.87494183e-01 1.56100765e-01
5.84188819e-01 -3.36632311e-01 8.37192178e-01 -3.99445027e-01
-1.23274076e+00 4.84599173e-01 4.14874285e-01 -1.15075266e+00
3.29242706e-01 -2.27929905e-01 -6.43312216e-01 -4.55109552e-02
-4.99886930e-01 9.25636515e-02 8.35381150e-01 -1.30859792e+00
-5.56988502e-03 -3.00437182e-01 5.25246561e-01 -6.27324998e-01
-6.54602289e-01 3.11420532e-03 -1.23585999e-01 -6.58234477e-01
4.52135224e-03 -9.07199502e-01 -3.74296516e-01 -7.26226151e-01
-1.24070776e+00 1.97031915e-01 1.19115019e+00 -3.01367879e-01
1.62743008e+00 -1.99830937e+00 -2.97315478e-01 7.63408720e-01
5.09339631e-01 6.60895526e-01 -9.02757347e-02 7.75844276e-01
-7.02795148e-01 6.77022874e-01 1.32160172e-01 3.24143097e-02
2.40715556e-02 2.40852162e-01 -1.12481713e+00 4.01415199e-01
-7.45252222e-02 1.01732624e+00 -7.10842788e-01 2.39455566e-01
7.21773505e-02 2.23309949e-01 -6.90686941e-01 -4.28464785e-02
-3.88396829e-01 -1.85773924e-01 -3.29073042e-01 6.10453427e-01
6.53717935e-01 9.15296301e-02 9.77771580e-02 5.84709048e-02
4.63786513e-01 3.78465503e-01 -9.66285527e-01 7.89050817e-01
5.50158368e-03 6.35135174e-01 -3.37384313e-01 -3.84492993e-01
9.07730162e-01 2.94142395e-01 -2.22942501e-01 -2.60125965e-01
2.46394068e-01 3.57327983e-03 2.49820858e-01 1.26234367e-01
4.61729169e-01 1.20941296e-01 -7.10895360e-02 4.06482160e-01
-5.78754432e-02 3.54981244e-01 4.12317254e-02 5.17143905e-01
1.44930279e+00 -3.78937453e-01 1.87769771e-01 -1.62020415e-01
3.82045001e-01 -9.13800746e-02 2.01433480e-01 1.30968845e+00
7.44540989e-02 2.21724704e-01 1.21863723e+00 -3.60281616e-01
-3.90458614e-01 -1.11028075e+00 3.02306443e-01 6.88996553e-01
6.25407696e-02 -1.08053577e+00 -1.08977997e+00 -1.28733253e+00
7.59094581e-02 1.15801513e+00 -8.46297026e-01 -8.82730067e-01
-3.89928073e-01 -5.41139841e-01 1.44877517e+00 6.34008229e-01
3.55894357e-01 -8.63685668e-01 -4.44490582e-01 -1.26004070e-01
7.07276225e-01 -1.05254185e+00 -2.83394545e-01 5.81605911e-01
-9.97406244e-01 -1.26443124e+00 -6.72712326e-02 -1.79653212e-01
6.84520483e-01 8.53416324e-02 9.15563524e-01 4.36658859e-01
-1.92120567e-01 7.61701539e-02 -2.99867481e-01 -7.59840727e-01
-6.96189582e-01 1.88895494e-01 1.27750590e-01 -1.98052794e-01
3.02258402e-01 -5.74657619e-01 -5.48048913e-02 1.79562062e-01
-1.10086489e+00 -4.58939135e-01 3.57389331e-01 4.90566403e-01
2.01216489e-01 4.16868120e-01 5.47584444e-02 -1.54568148e+00
8.60455751e-01 -3.34391803e-01 -9.04046655e-01 9.04191881e-02
-8.54595423e-01 7.79565424e-02 9.56273496e-01 -6.41240537e-01
-2.11187944e-01 -2.45090559e-01 -1.68782279e-01 -8.54646802e-01
2.04327777e-02 4.44672048e-01 -3.49727035e-01 -4.76188630e-01
1.23025286e+00 5.30492961e-02 -3.63377005e-01 -2.51467139e-01
3.16673279e-01 -8.63214210e-02 3.08312535e-01 -4.91201967e-01
1.35556686e+00 3.49117815e-01 3.46948236e-01 -5.78242302e-01
-3.79387140e-01 1.04041785e-01 1.10861927e-01 8.87762681e-02
4.80563045e-01 -2.65454233e-01 -7.99922407e-01 4.21841264e-01
-1.04179132e+00 -2.35224888e-01 -3.39798361e-01 9.30081978e-02
2.49673072e-02 3.31576735e-01 -6.26693189e-01 -9.41552997e-01
-3.47971678e-01 -1.15062797e+00 5.61279535e-01 3.01719576e-01
-1.54920548e-01 -1.19534051e+00 1.46952076e-02 -2.39902347e-01
1.60454333e-01 5.33501267e-01 1.21143949e+00 -1.40961230e+00
-6.66884184e-01 -7.35380709e-01 3.60810518e-01 3.19236487e-01
-4.54438031e-02 2.31677189e-01 -1.28214312e+00 -4.16684777e-01
4.05057427e-03 1.73345879e-01 8.53459060e-01 2.13161126e-01
1.14381695e+00 -5.51295877e-01 -7.92670608e-01 8.44644070e-01
1.35217619e+00 2.69699126e-01 8.81699443e-01 4.06815857e-01
7.59733915e-01 2.51216441e-01 2.36321911e-01 -1.14758931e-01
-4.05677795e-01 3.96274120e-01 1.20240152e+00 -2.73694605e-01
1.89212650e-01 -8.51405144e-01 7.72041976e-01 4.70663086e-02
1.81294903e-01 -4.93218809e-01 -9.04673457e-01 1.73568159e-01
-1.36456597e+00 -8.77321959e-01 -3.00467461e-01 2.29615688e+00
3.94810706e-01 8.02418113e-01 1.23695597e-01 6.65693343e-01
5.75248182e-01 5.06692171e-01 -2.36784548e-01 -9.44902480e-01
-4.27062297e-03 5.23762286e-01 1.11509585e+00 4.74901229e-01
-9.10207987e-01 1.14520919e+00 7.20613670e+00 8.02735448e-01
-1.25935781e+00 -1.80531502e-01 3.10571879e-01 -1.80989802e-02
-6.96263909e-01 5.15465677e-01 -1.19626451e+00 7.97538534e-02
1.21254778e+00 -1.51694298e-01 4.09327567e-01 9.36036050e-01
-2.54926741e-01 6.06216013e-01 -1.41186774e+00 4.90257174e-01
-3.65379751e-02 -1.63867521e+00 4.52072114e-01 5.11753619e-01
4.16941315e-01 -1.60993695e-01 3.45401227e-01 4.04281020e-01
3.39026481e-01 -1.15808558e+00 6.47488475e-01 -1.72406808e-02
4.04935211e-01 -1.15924668e+00 6.35936499e-01 3.85803908e-01
-9.16422844e-01 -2.26066247e-01 -2.37170160e-01 1.50408730e-01
-3.96277338e-01 5.86087704e-01 -1.14539933e+00 8.27709377e-01
5.43294191e-01 1.33221045e-01 -9.34683383e-01 5.82840860e-01
-7.40083337e-01 1.23116660e+00 -3.00169915e-01 -2.39730090e-01
3.77675742e-01 8.49632621e-02 8.91784489e-01 9.66318250e-01
-1.86007738e-01 -4.56794500e-01 -6.38817027e-02 1.13111818e+00
-1.25538379e-01 -3.63512218e-01 -1.10606611e+00 -5.77033818e-01
4.04427618e-01 9.91837442e-01 -8.98158073e-01 3.48761387e-04
-4.92133573e-02 4.65625316e-01 -1.29180387e-01 3.22142988e-01
-1.11217666e+00 -7.02973604e-01 7.67911136e-01 3.68092477e-01
-1.89118460e-02 -9.28979963e-02 -2.65383452e-01 -7.78944850e-01
-8.01477730e-02 -1.23644066e+00 6.79800689e-01 -3.89617234e-01
-9.11920130e-01 7.31542468e-01 1.72246441e-01 -7.84174323e-01
-2.08441824e-01 -9.47445154e-01 -1.03368151e+00 6.89653754e-01
-1.10891092e+00 -9.62969899e-01 1.20613538e-01 7.39314556e-01
-1.03801809e-01 -3.97291511e-01 8.51171732e-01 7.40786418e-02
-7.36411512e-01 1.32970011e+00 -5.38491011e-01 3.54391068e-01
1.35901794e-01 -1.18386126e+00 1.15525317e+00 1.41379237e+00
6.90020382e-01 1.34818447e+00 9.87275660e-01 -9.73329008e-01
-1.57737434e+00 -1.08747053e+00 2.91786730e-01 -6.70325160e-01
6.75007224e-01 -6.57391548e-01 -9.17021930e-01 1.15609777e+00
1.24224618e-01 -4.92143929e-02 6.40087664e-01 2.96855032e-01
-8.59954834e-01 1.20108984e-01 -1.19794095e+00 1.03056383e+00
9.95301187e-01 -5.83786488e-01 -2.87500918e-01 3.61069053e-01
1.11719382e+00 -5.97017944e-01 -2.83245862e-01 3.00898522e-01
2.16574818e-01 -1.03379869e+00 1.03301024e+00 -1.05621731e+00
5.61865158e-02 -2.66674072e-01 -1.39010428e-02 -1.03721094e+00
5.65602556e-02 -1.05686116e+00 -6.67992830e-01 1.02789533e+00
7.29373634e-01 -1.34990525e+00 1.36578476e+00 7.83352181e-02
9.44215059e-02 -7.93508351e-01 -7.44120657e-01 -1.17976546e+00
1.19528957e-01 -7.52249181e-01 9.09271836e-01 1.02604866e+00
-1.64795429e-01 -1.82693955e-02 -3.15006107e-01 8.14741075e-01
5.11264205e-01 -3.35352629e-01 9.01293278e-01 -9.24308538e-01
-6.74562395e-01 -3.44015270e-01 -6.38496578e-01 -4.44480181e-01
1.38419792e-01 -1.35170913e+00 -6.96873844e-01 -9.73968744e-01
-2.99651563e-01 6.63447678e-02 -5.50407290e-01 8.44734013e-01
-9.89346802e-02 7.54059330e-02 1.98584557e-01 -3.38149250e-01
2.29689479e-01 4.35757823e-03 7.62339890e-01 -1.93185389e-01
-3.27585965e-01 2.74285555e-01 -8.89379621e-01 6.09561563e-01
7.11187720e-01 -9.94750440e-01 -8.91881764e-01 6.14846218e-03
5.00289619e-01 -3.08921218e-01 5.27694225e-01 -1.03778219e+00
2.15792432e-01 -1.25736371e-02 1.63991258e-01 -6.53647780e-01
1.63135454e-01 -8.83954227e-01 2.82116979e-01 9.90816772e-01
-2.69512266e-01 2.52821833e-01 6.57701671e-01 6.20226264e-01
1.46216303e-01 -3.29848081e-01 4.44287509e-01 7.75952041e-02
-2.37409025e-01 3.10162872e-01 -3.00663561e-01 -1.92225873e-01
1.23176205e+00 -5.16563714e-01 -7.11170554e-01 -3.35623652e-01
-5.00139356e-01 -1.26586705e-01 3.11921924e-01 6.00044310e-01
7.42070198e-01 -8.82048488e-01 -1.93944067e-01 7.50772953e-01
1.97938476e-02 -5.04311383e-01 -1.10839508e-01 2.11352140e-01
-5.84342420e-01 5.30928195e-01 -5.94056994e-02 -3.14670801e-01
-1.44579446e+00 9.66792881e-01 6.32051229e-01 -5.43182909e-01
-6.21789575e-01 1.02196658e+00 4.08332124e-02 -4.76198792e-01
2.95582026e-01 -4.25921857e-01 -5.94181642e-02 -3.46115500e-01
4.45326626e-01 4.01235193e-01 3.17691565e-01 1.16222098e-01
-6.16703451e-01 1.85247958e-01 -3.16028833e-01 7.98822120e-02
8.36566329e-01 1.00030208e+00 -1.68381080e-01 4.61537242e-02
8.16570342e-01 6.95448995e-01 -5.01341879e-01 2.69780725e-01
-1.03215193e-02 -3.91496986e-01 -9.66802090e-02 -7.15792656e-01
-1.25500762e+00 8.25816929e-01 5.02879262e-01 6.98715806e-01
8.41437221e-01 -1.61353841e-01 6.43192530e-01 5.66458344e-01
5.13862967e-01 -3.48687738e-01 -1.41420424e-01 5.53245425e-01
5.95402002e-01 -1.08804852e-01 5.17057627e-02 -8.40173483e-01
-1.01772927e-01 9.43704069e-01 7.20407903e-01 -4.02616948e-01
6.38497889e-01 5.91983557e-01 -7.77332634e-02 -3.26220274e-01
-7.61617184e-01 4.71542686e-01 6.34144694e-02 6.70267940e-01
-9.97254625e-02 -1.33652873e-02 -3.71797569e-02 8.50557148e-01
-6.74432814e-01 -4.86694455e-01 7.91245997e-01 7.87629664e-01
-1.03402831e-01 -1.28607130e+00 -5.79746068e-01 4.46207672e-01
-7.52792358e-01 -2.81966478e-01 -1.02372897e+00 1.23819768e+00
-1.98493898e-01 9.10796523e-01 -5.53658009e-01 -1.13905299e+00
5.77332079e-01 1.65244684e-01 3.65726024e-01 -8.52896273e-01
-1.24686909e+00 -5.74039102e-01 1.69707283e-01 -8.59686613e-01
6.54676557e-01 2.70785578e-02 -1.31817091e+00 -3.98650408e-01
-6.62174523e-01 3.04565310e-01 7.46309459e-01 5.32902122e-01
2.22126678e-01 9.74982679e-01 5.91358244e-01 -1.52172804e-01
-7.83720076e-01 -4.40718085e-01 -5.67466915e-01 -7.08563551e-02
2.25005463e-01 -4.20191735e-01 -7.30283380e-01 -7.08949566e-01] | [5.778233528137207, 7.636438369750977] |
8fa7db5e-1b65-458e-bcfb-f40452176451 | obtaining-referential-word-meanings-from | null | null | https://aclanthology.org/P17-1023 | https://aclanthology.org/P17-1023.pdf | Obtaining referential word meanings from visual and distributional information: Experiments on object naming | We investigate object naming, which is an important sub-task of referring expression generation on real-world images. As opposed to mutually exclusive labels used in object recognition, object names are more flexible, subject to communicative preferences and semantically related to each other. Therefore, we investigate models of referential word meaning that link visual to lexical information which we assume to be given through distributional word embeddings. We present a model that learns individual predictors for object names that link visual and distributional aspects of word meaning during training. We show that this is particularly beneficial for zero-shot learning, as compared to projecting visual objects directly into the distributional space. In a standard object naming task, we find that different ways of combining lexical and visual information achieve very similar performance, though experiments on model combination suggest that they capture complementary aspects of referential meaning. | ['Sina Zarrie{\\ss}', 'David Schlangen'] | 2017-07-01 | null | null | null | acl-2017-7 | ['referring-expression-generation'] | ['computer-vision'] | [ 1.41923875e-01 -1.01579968e-02 -4.07852262e-01 -7.22281098e-01
-3.88656974e-01 -6.40089214e-01 1.03329265e+00 5.81341758e-02
-6.96446776e-01 5.27995169e-01 8.46578002e-01 -9.23889950e-02
-9.00230184e-02 -8.52651119e-01 -4.18686002e-01 -5.31525791e-01
1.70263320e-01 4.16296124e-01 -1.75092518e-01 -9.18549001e-02
2.79730409e-01 6.58291459e-01 -1.62560809e+00 1.72766685e-01
5.41897044e-02 6.60427451e-01 2.01832891e-01 4.27016765e-01
-8.98634732e-01 9.26020324e-01 -6.64277256e-01 -3.29319298e-01
1.58244409e-02 -4.00646806e-01 -1.08727920e+00 2.53993213e-01
8.54178548e-01 6.93084225e-02 -1.04415976e-01 1.05398560e+00
4.72344965e-01 5.01683295e-01 1.07613075e+00 -1.38160372e+00
-1.72693515e+00 5.45456052e-01 -1.69073790e-01 1.70882672e-01
4.30434585e-01 8.27134680e-03 1.52747130e+00 -1.07286119e+00
9.58968282e-01 1.71628654e+00 3.36155176e-01 8.98233235e-01
-1.72503901e+00 -4.97629523e-01 3.56219381e-01 2.11833626e-01
-1.41311181e+00 -4.94887739e-01 7.25943089e-01 -7.02326357e-01
1.05207205e+00 1.45094976e-01 5.35509586e-01 1.32091975e+00
-2.38993630e-01 6.51430130e-01 9.62062240e-01 -5.16194403e-01
2.11334869e-01 2.29495168e-01 4.06735122e-01 3.52084875e-01
1.45456135e-01 -7.49322167e-03 -5.91259539e-01 -2.81296372e-01
8.26706648e-01 9.21713486e-02 -2.61160374e-01 -6.72794402e-01
-1.38170242e+00 1.16976714e+00 7.69581974e-01 6.44391835e-01
-3.61952692e-01 6.84025586e-01 1.50114134e-01 1.69826478e-01
4.20823306e-01 6.99968398e-01 -1.86201140e-01 -1.59647558e-02
-4.15843636e-01 1.75900295e-01 5.97158670e-01 1.18352258e+00
8.98877442e-01 4.02602926e-02 -5.35016537e-01 1.15555918e+00
4.27634686e-01 4.45083380e-01 6.20999694e-01 -1.01598859e+00
-2.13223532e-01 3.64811450e-01 2.34037936e-02 -6.96888208e-01
-3.78959417e-01 1.43591985e-01 -2.09610492e-01 2.17811465e-01
3.69530082e-01 2.08534658e-01 -1.05231452e+00 2.38683629e+00
3.10203303e-02 9.80956554e-02 -2.10343003e-02 9.53658402e-01
1.27251482e+00 4.68886286e-01 9.25242960e-01 3.37700769e-02
1.62240708e+00 -6.22219682e-01 -7.33120084e-01 -4.09299970e-01
6.82870269e-01 -6.25599980e-01 1.45637131e+00 -2.77696222e-01
-7.42929578e-01 -4.64159250e-01 -6.73463762e-01 -6.45529211e-01
-8.50790739e-01 -3.06575865e-01 9.13087666e-01 4.04677927e-01
-1.07504904e+00 2.74102509e-01 -2.93435186e-01 -8.07281017e-01
6.44999027e-01 6.70827627e-02 -5.47319829e-01 -3.17047955e-03
-9.18683529e-01 1.30140710e+00 3.31975371e-01 -5.23317277e-01
-6.37285650e-01 -6.11346841e-01 -1.20371497e+00 6.42148331e-02
1.70337677e-01 -8.57707918e-01 1.29713869e+00 -1.22869086e+00
-1.07652342e+00 1.55657446e+00 -4.29490685e-01 -1.07412957e-01
-2.33940169e-01 -1.28227025e-01 -2.04005718e-01 -1.64194703e-01
3.71220440e-01 1.31011927e+00 7.83354998e-01 -1.38565481e+00
-2.85508662e-01 -2.02146322e-01 3.45876247e-01 1.96374014e-01
-3.81307900e-01 2.63113379e-01 -6.88798130e-02 -7.33052909e-01
-1.04503781e-01 -6.76374853e-01 3.41322124e-02 6.19292796e-01
-2.82548606e-01 -7.39628553e-01 7.16372371e-01 -1.52304461e-02
8.07518005e-01 -2.19669771e+00 2.36603752e-01 -8.40880647e-02
2.69971907e-01 -1.53505906e-01 -5.55824935e-01 3.02512199e-01
-4.82735097e-01 3.28444660e-01 -1.13526784e-01 -2.19708294e-01
2.94175357e-01 6.87158644e-01 -4.49856311e-01 3.82338464e-01
4.13953066e-01 1.26858997e+00 -1.10527658e+00 -4.52746063e-01
1.82474360e-01 5.48228383e-01 -4.11723137e-01 2.35089913e-01
-3.13368440e-01 8.31279755e-02 -4.56284136e-01 5.47971249e-01
2.32533231e-01 -3.82299364e-01 -5.71051911e-02 -1.41502440e-01
2.45666131e-01 3.03532809e-01 -8.17771792e-01 1.67002153e+00
-7.39049137e-01 8.32377970e-01 -3.56623769e-01 -8.94219100e-01
8.54057789e-01 3.37989837e-01 8.25491101e-02 -6.47472262e-01
2.76242733e-01 -2.35248610e-01 -1.11216657e-01 -6.41012192e-01
4.48053837e-01 -9.14106786e-01 -2.87657231e-01 7.64727116e-01
4.32225615e-01 -2.40829393e-01 3.97478528e-02 2.55199134e-01
8.16120028e-01 2.48847231e-01 7.40994692e-01 -2.23625392e-01
1.71382576e-01 -3.10937446e-02 3.03709179e-01 5.65224290e-01
-1.79346919e-01 5.79883695e-01 4.24482316e-01 -3.99533004e-01
-1.05798006e+00 -1.31871235e+00 -3.58622193e-01 1.58894992e+00
2.15344578e-01 -2.23240748e-01 -1.85648248e-01 -4.69495803e-01
1.90200001e-01 1.46541953e+00 -8.13837826e-01 -2.12424174e-01
-2.40914837e-01 -3.31618488e-01 2.06799746e-01 8.26599240e-01
-3.96729559e-01 -1.54959691e+00 -8.22352886e-01 -1.77969821e-02
2.34457269e-01 -9.78260815e-01 -5.05286634e-01 3.46441120e-01
-3.96055013e-01 -9.81582284e-01 -8.94489348e-01 -8.92606974e-01
5.99071562e-01 3.79154116e-01 1.63280916e+00 9.94340423e-03
-4.26423788e-01 1.06587613e+00 -3.61110836e-01 -4.75472778e-01
-2.77710259e-01 -4.30157214e-01 -5.37483133e-02 -5.94509533e-03
9.33395386e-01 -4.99846131e-01 -1.14774644e-01 -1.12473302e-01
-9.01250422e-01 -1.64659515e-01 2.91729838e-01 8.82164121e-01
5.72210550e-01 -9.94401038e-01 3.39875221e-01 -9.55678284e-01
7.83676624e-01 -7.19074667e-01 -2.50375092e-01 2.51056224e-01
-2.47807994e-01 2.85837978e-01 1.08364642e-01 -8.38318765e-01
-9.25496578e-01 -1.27807915e-01 1.42900914e-01 -5.29677868e-01
-4.96000826e-01 1.14728190e-01 -1.66236877e-01 4.18507531e-02
6.93193257e-01 -8.70283917e-02 -3.47317569e-02 -4.73471731e-01
1.17774618e+00 3.88081372e-01 2.43429944e-01 -6.30826950e-01
5.87734938e-01 4.61180985e-01 -8.17223191e-02 -9.79993880e-01
-1.06819510e+00 -5.06881416e-01 -6.98722005e-01 1.28906712e-01
1.34926224e+00 -7.21388817e-01 -5.85891366e-01 -2.62968302e-01
-1.53046644e+00 -5.80632426e-02 -8.63271713e-01 4.68419373e-01
-8.83078516e-01 -5.00265546e-02 -2.32917145e-01 -6.42910600e-01
1.06048845e-01 -6.97340608e-01 1.36959660e+00 1.44129738e-01
-7.48122811e-01 -1.22874486e+00 1.11540094e-01 -1.63410157e-01
4.46456820e-01 3.49799246e-02 1.30971515e+00 -9.96373355e-01
-3.52297395e-01 2.40623996e-01 -5.85911155e-01 1.72227815e-01
4.06237274e-01 -3.06374609e-01 -1.18151438e+00 1.33448482e-01
-8.94763172e-02 -6.16607964e-01 9.69601393e-01 3.39934945e-01
9.41054404e-01 -2.62647986e-01 -3.72972101e-01 5.47282100e-01
1.55774403e+00 3.66035812e-02 2.92585403e-01 1.79455131e-01
9.47809696e-01 8.37194443e-01 3.57253700e-01 2.82974750e-01
3.48568469e-01 7.15436995e-01 3.08586806e-01 -8.09502378e-02
-4.78319883e-01 -1.58408508e-01 1.87388524e-01 1.68259695e-01
3.05251032e-01 -2.45069414e-01 -9.85250473e-01 7.75290251e-01
-1.45121527e+00 -1.21232140e+00 -1.93695109e-02 1.97048950e+00
1.01420367e+00 -4.06609505e-01 -2.74353568e-03 -5.68251908e-01
7.18622744e-01 5.47898829e-01 -4.27169234e-01 -5.87132454e-01
-2.66416192e-01 3.78531367e-01 2.76929021e-01 4.38916415e-01
-9.50043023e-01 1.23437047e+00 7.37441444e+00 5.48167586e-01
-9.70978916e-01 3.96923751e-01 2.52339691e-01 -2.29356900e-01
-8.03962827e-01 3.20128314e-02 -5.30573547e-01 1.17715821e-01
6.43365681e-01 -4.69761938e-01 2.54395545e-01 9.73191977e-01
-2.17285365e-01 1.27750233e-01 -1.81092298e+00 1.25577188e+00
4.85894263e-01 -1.18036699e+00 4.89234418e-01 -6.09607436e-03
4.22703028e-01 -1.46611437e-01 -1.66441570e-03 1.84527487e-01
5.48128664e-01 -1.37221825e+00 8.12796056e-01 7.43345559e-01
7.91788042e-01 -3.16314369e-01 3.32032621e-01 -1.70186516e-02
-9.35902596e-01 -2.63862722e-02 -6.20544314e-01 -1.57451048e-01
3.95531058e-01 7.19399899e-02 -5.56837380e-01 -1.84945807e-01
3.81800443e-01 7.41526067e-01 -6.22926056e-01 8.10010791e-01
-4.94834334e-01 9.25264210e-02 -4.74655889e-02 -1.29791290e-01
3.11148167e-01 8.59681889e-02 3.98589730e-01 1.18183434e+00
2.91346550e-01 5.90404160e-02 1.85261846e-01 1.55703878e+00
-1.68101907e-01 3.17939341e-01 -1.24384546e+00 -3.20376337e-01
5.09531617e-01 1.13181317e+00 -5.10501087e-01 -6.48235023e-01
-8.25022519e-01 7.95846939e-01 5.38548827e-01 4.39965814e-01
-6.13491416e-01 -5.87237589e-02 1.13916683e+00 -1.08396128e-01
2.72844434e-01 -2.18829527e-01 -1.90522224e-01 -1.00197148e+00
-3.39554459e-01 -2.95148343e-01 2.18320966e-01 -1.14196777e+00
-1.86933732e+00 3.06597054e-01 1.47299170e-01 -1.16025662e+00
-3.26426774e-01 -9.97305989e-01 -6.00414157e-01 1.06341171e+00
-1.43890750e+00 -1.33539534e+00 -1.19991720e-01 4.31035012e-01
6.34935379e-01 4.10935432e-02 1.39124608e+00 -2.01983765e-01
1.34363204e-01 3.08925778e-01 -3.99322063e-01 1.73568264e-01
8.03333163e-01 -1.37479115e+00 3.36098909e-01 3.62089604e-01
9.85881388e-01 9.72306013e-01 6.06608391e-01 -4.09564018e-01
-1.12215364e+00 -8.33962023e-01 1.20570612e+00 -8.70911658e-01
8.81013632e-01 -2.78945327e-01 -1.00465572e+00 9.65988636e-01
2.88768530e-01 2.44434580e-01 1.17984116e+00 4.40073550e-01
-8.85188282e-01 4.35861051e-01 -9.46363270e-01 5.38555562e-01
1.24059486e+00 -1.02235508e+00 -1.31405950e+00 4.11388665e-01
8.83199871e-01 1.99928492e-01 -6.63774252e-01 -2.73959875e-01
3.39092195e-01 -6.24280512e-01 1.08888113e+00 -1.16215134e+00
3.17315191e-01 -1.23620696e-01 -4.89920825e-01 -1.33532643e+00
-6.64719343e-01 1.18553080e-01 2.85526812e-01 1.36221719e+00
1.95554450e-01 -4.47573334e-01 4.74030554e-01 7.55127847e-01
2.08108619e-01 -2.64533371e-01 -9.08213258e-01 -9.11787271e-01
3.58485520e-01 -4.27255273e-01 5.24145722e-01 1.30184197e+00
-9.57541261e-03 9.72891927e-01 -1.48972958e-01 -3.26947629e-01
4.30583537e-01 2.73250490e-01 4.62424278e-01 -1.37340391e+00
-5.16614802e-02 -7.91325748e-01 -8.34755123e-01 -9.16788876e-01
8.40469241e-01 -1.28712606e+00 8.17364678e-02 -1.64244580e+00
5.45289576e-01 -2.97674745e-01 -4.62452948e-01 8.47601712e-01
-1.89718217e-01 3.16776723e-01 4.76385564e-01 -5.32768704e-02
-4.44149196e-01 5.31325638e-01 1.15049350e+00 -3.78209114e-01
1.62298545e-01 -6.70989752e-01 -8.71473551e-01 8.30977261e-01
3.50460321e-01 -5.31647205e-01 -5.06103039e-01 -8.54708910e-01
1.24704450e-01 -2.99840480e-01 6.82059884e-01 -3.21530700e-01
-1.33307397e-01 -5.11980653e-01 3.18342179e-01 -5.30922972e-02
6.18943155e-01 -8.68695378e-01 -2.01253131e-01 -5.51447012e-02
-7.21305966e-01 -1.97095927e-02 5.15807346e-02 4.42464083e-01
-3.49684060e-01 -4.37293917e-01 7.84385562e-01 -3.52505237e-01
-1.25306201e+00 1.08123668e-01 -2.91551352e-01 2.26826549e-01
1.00780773e+00 -4.91606385e-01 -2.33928561e-01 -4.47660059e-01
-9.19360876e-01 -2.86889244e-02 7.40026474e-01 8.83265138e-01
6.11166000e-01 -1.81522524e+00 -5.48400044e-01 -5.63263558e-02
7.24573076e-01 -5.62794983e-01 -1.79902151e-01 4.62633431e-01
1.34156242e-01 3.47528756e-01 -3.80693167e-01 -5.14174402e-01
-1.21712267e+00 1.08965552e+00 1.39522627e-01 6.14067912e-01
-6.31133914e-01 1.32843006e+00 8.12852621e-01 -3.52109790e-01
2.01400369e-01 -1.95658371e-01 -4.82428133e-01 4.28960562e-01
4.88779932e-01 -1.19225524e-01 -6.19877517e-01 -1.25698864e+00
-4.99842733e-01 8.88579965e-01 1.86294600e-01 -1.79884583e-01
1.28060091e+00 -1.77184716e-01 -2.96384335e-01 1.12002945e+00
1.45475590e+00 -2.34655514e-01 -7.28667021e-01 -5.29263437e-01
4.04979438e-01 -6.91804230e-01 -1.54178455e-01 -4.71350849e-01
-7.75730193e-01 1.09736764e+00 6.44793391e-01 1.14544302e-01
7.27796495e-01 6.94716096e-01 2.63160318e-01 3.86445761e-01
3.60806674e-01 -7.51201391e-01 1.79330572e-01 3.92990053e-01
9.98573601e-01 -1.21122682e+00 -3.04361656e-02 -2.52181828e-01
-7.09487557e-01 9.74225342e-01 5.31307161e-01 -1.37701914e-01
4.85147029e-01 -4.35702950e-02 2.75955230e-01 -4.10274655e-01
-8.66346061e-01 -6.89518690e-01 3.82931769e-01 8.95159423e-01
8.85056853e-01 1.52845129e-01 -2.82685697e-01 3.76644701e-01
-1.16622940e-01 -2.49685466e-01 3.64479095e-01 8.55955720e-01
-4.49467957e-01 -1.12455416e+00 -2.64824092e-01 4.08047229e-01
-1.69267222e-01 -3.67505997e-01 -5.84181607e-01 7.63825059e-01
2.23527178e-01 6.03914201e-01 5.76328993e-01 1.83199301e-01
2.50403911e-01 4.41457808e-01 6.38212621e-01 -1.23515224e+00
-2.24979356e-01 -2.17878595e-01 1.95730686e-01 -7.12447107e-01
-6.30409777e-01 -3.23560566e-01 -1.40192592e+00 8.14076066e-02
-1.12086803e-01 1.33166770e-02 4.67396706e-01 9.49490845e-01
1.17958963e-01 4.71774012e-01 3.88184078e-02 -9.03350651e-01
-4.36114758e-01 -8.95510912e-01 -7.12720454e-01 1.06158543e+00
3.11067730e-01 -1.06669116e+00 -2.69484460e-01 3.02249491e-01] | [10.51675033569336, 2.1669788360595703] |
a9651cde-9528-4e0d-a0a2-0f473673d082 | contrastive-regularized-u-net-for-video | null | null | https://ieeexplore.ieee.org/abstract/document/10098744/ | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10098744 | Contrastive-Regularized U-Net for Video Anomaly Detection | Video anomaly detection aims to identify anomalous segments in a video. It is typically trained with weakly supervised video-level labels. This paper focuses on two crucial factors affecting the performance of video anomaly detection models. First, we explore how to capture the local and global temporal dependencies more effectively. Previous architectures are effective at capturing either local and global information, but not both. We propose to employ a U-Net like structure to model both types of dependencies in a unified structure where the encoder learns global dependencies hierarchically on top of local ones; then the decoder propagates this global information back to the segment level for classification. Second, overfitting is a non-trivial issue for video anomaly detection due to limited training data. We propose weakly supervised contrastive regularization which adopts a feature-based approach to regularize the network. Contrastive regularization learns more generalizable features by enforcing inter-class separability and intra-class compactness. Extensive experiments on the UCF-Crime dataset shows that our approach outperforms several state-of-the-art methods. | ['Joon Huang Chuah', 'Maylor Karhang Leung', 'Hui-Fuang Ng', 'Hung-Khoon Tan', 'Yu Tong Cheng', 'Kian Yu Gan'] | 2023-04-11 | null | null | null | ieee-access-2023-4 | ['video-anomaly-detection', 'anomaly-detection-in-surveillance-videos', 'anomaly-detection', 'anomaly-detection-in-surveillance-videos'] | ['computer-vision', 'computer-vision', 'methodology', 'methodology'] | [ 8.97685811e-02 -2.08453819e-01 -4.10917848e-01 -5.44020712e-01
-4.53111202e-01 -2.43726298e-01 4.63297427e-01 2.26169854e-01
-4.14395213e-01 2.68198878e-01 2.84502059e-01 -1.04584016e-01
1.82580069e-01 -4.68375176e-01 -9.40497220e-01 -6.61587298e-01
-6.06574118e-01 -2.94110198e-02 4.93909150e-01 9.80853289e-02
1.78092629e-01 4.24284697e-01 -1.55290055e+00 6.81755662e-01
8.16345334e-01 1.16017389e+00 -4.24024582e-01 8.28034163e-01
-2.66794860e-02 1.56415653e+00 -2.49039367e-01 -1.76586658e-01
3.46671283e-01 -6.08952403e-01 -8.92477095e-01 3.76750886e-01
7.34208882e-01 -8.50294948e-01 -5.70773065e-01 1.13948214e+00
-1.06130995e-01 8.94559622e-02 7.56367207e-01 -1.44035459e+00
-5.21088600e-01 1.62413329e-01 -6.32846892e-01 9.68605578e-01
2.85229415e-01 1.71078146e-01 1.23857784e+00 -6.94951892e-01
5.23087800e-01 1.03321552e+00 5.66322207e-01 5.52317917e-01
-1.10862458e+00 -3.27968985e-01 8.71300101e-01 5.87832808e-01
-1.14855981e+00 -3.77415925e-01 7.25310981e-01 -7.97527134e-01
1.14771855e+00 1.47557184e-01 5.06560206e-01 1.41820848e+00
2.61676788e-01 1.14645553e+00 4.36293632e-01 -1.97060168e-01
5.71206212e-02 -3.55192602e-01 4.71119553e-01 9.06142414e-01
3.94170024e-02 -5.42677753e-02 -6.01262569e-01 -3.00196230e-01
6.99823022e-01 1.90468416e-01 -8.77970830e-02 -3.67927700e-01
-7.67620623e-01 8.83500099e-01 1.58675343e-01 4.63051528e-01
-3.34474057e-01 2.97720373e-01 9.50777829e-01 6.73226416e-01
7.91468561e-01 9.47650298e-02 -3.80183071e-01 -2.85861701e-01
-8.96992981e-01 9.40335393e-02 4.28777367e-01 7.18744814e-01
6.25149071e-01 2.28498027e-01 -3.44752342e-01 6.68229699e-01
3.54967117e-01 -2.45924205e-01 5.11449218e-01 -8.58132243e-01
4.59847689e-01 5.49897909e-01 -1.86390802e-01 -1.01244628e+00
-3.20186913e-01 -1.86395824e-01 -7.02441216e-01 1.46384880e-01
5.46673238e-01 -8.64446089e-02 -1.10391808e+00 1.83817089e+00
-1.26530472e-02 7.80794144e-01 -2.76507139e-01 8.18673551e-01
3.64834070e-01 6.02106631e-01 2.82570750e-01 -1.95329726e-01
8.39767218e-01 -1.16802001e+00 -7.65368462e-01 -2.94128358e-01
1.04788661e+00 -3.48145187e-01 8.45207095e-01 2.71585613e-01
-9.98658955e-01 -3.23648930e-01 -9.11115348e-01 3.56511213e-02
-2.69039750e-01 1.39597543e-02 3.18129212e-01 2.07850441e-01
-1.06500745e+00 6.32723093e-01 -1.28555441e+00 -3.79846960e-01
6.60680175e-01 2.29498252e-01 -4.17453974e-01 9.70470137e-04
-1.01259196e+00 5.81918240e-01 3.79146338e-01 2.64252666e-02
-9.77861702e-01 -3.11637819e-01 -1.12643480e+00 3.19081433e-02
5.26459992e-01 -3.32459331e-01 1.03519619e+00 -1.49918926e+00
-1.13121915e+00 8.82796168e-01 -4.18756515e-01 -5.90999782e-01
4.44911689e-01 -5.43490708e-01 -4.85489041e-01 3.30954015e-01
1.14374556e-01 3.08352381e-01 1.16421628e+00 -9.19392765e-01
-7.61328101e-01 -3.73481572e-01 6.29244223e-02 1.00804068e-01
-5.47670007e-01 2.89315939e-01 -6.83296442e-01 -8.96033347e-01
1.55989483e-01 -6.63530648e-01 -1.00682519e-01 -6.52360693e-02
-1.37289479e-01 -3.11548740e-01 1.14073491e+00 -7.92574525e-01
1.57036304e+00 -2.32809734e+00 1.58274040e-01 3.86713564e-01
2.15532944e-01 2.41677001e-01 -2.79219478e-01 3.34400646e-02
-4.35432822e-01 1.26604050e-01 -2.69818068e-01 -3.62957746e-01
-3.02771181e-01 5.10677516e-01 -1.79372743e-01 6.91977322e-01
6.14717245e-01 6.66035950e-01 -9.55945015e-01 -3.21070433e-01
-1.10005811e-01 9.76073518e-02 -9.25357819e-01 3.86258572e-01
-7.19230473e-02 6.28621161e-01 -3.93626779e-01 7.38604844e-01
3.66496503e-01 -1.14791170e-01 -5.13130277e-02 2.10968420e-01
-6.09404501e-03 3.02616209e-01 -8.95464063e-01 1.70618069e+00
1.18542559e-01 7.76207566e-01 -2.39177071e-03 -1.58706713e+00
4.34532553e-01 3.14222842e-01 7.39719927e-01 -6.35834992e-01
-3.62774655e-02 1.30891085e-01 1.13123888e-02 -9.35452104e-01
8.32328126e-02 1.77430958e-01 1.34795666e-01 2.42610946e-01
3.44424099e-01 8.69955540e-01 3.27968657e-01 3.23062062e-01
1.56624997e+00 2.62514204e-01 1.65427521e-01 -2.27701128e-01
6.69518590e-01 -3.23514372e-01 9.23944473e-01 9.06750500e-01
-5.07782757e-01 7.20831692e-01 9.86131907e-01 -8.17490280e-01
-8.27682555e-01 -8.58392298e-01 5.88662736e-02 1.46454382e+00
-6.50987327e-02 -7.34027267e-01 -7.51118124e-01 -1.27520287e+00
-1.84967339e-01 4.98567730e-01 -7.77657449e-01 -3.92163873e-01
-8.43711913e-01 -7.02978253e-01 5.22911668e-01 7.95041800e-01
3.47397357e-01 -1.08246875e+00 -4.10791725e-01 1.69131234e-01
-2.59569496e-01 -1.29928112e+00 -4.64039207e-01 2.49615878e-01
-9.70682979e-01 -1.04062617e+00 -2.55174726e-01 -6.33329630e-01
7.51911879e-01 -5.09793349e-02 1.12454784e+00 4.45957243e-01
-1.68031573e-01 5.21978676e-01 -5.51645637e-01 -8.83090720e-02
-3.14630836e-01 -3.23741212e-02 1.54784203e-01 4.27411526e-01
7.67339051e-01 -5.89588165e-01 -3.38035315e-01 2.06114739e-01
-9.89756465e-01 -5.06977201e-01 3.60115439e-01 8.55051994e-01
3.86203587e-01 -2.47752778e-02 3.18617612e-01 -9.40955579e-01
3.84126991e-01 -6.35888815e-01 -3.94365728e-01 6.27815500e-02
-1.26846731e-01 3.77497412e-02 5.91035903e-01 -2.19147459e-01
-9.36287940e-01 7.11854622e-02 -1.84003875e-01 -8.14473927e-01
-5.49443603e-01 4.19809610e-01 -5.94964065e-02 1.80558205e-01
4.50479269e-01 8.63963142e-02 -1.90464184e-01 -3.62563789e-01
-5.62464893e-02 3.01567852e-01 5.10881662e-01 -5.70936382e-01
6.08061016e-01 4.88326877e-01 -2.54525900e-01 -1.00931656e+00
-8.48870993e-01 -7.15864360e-01 -1.07458115e+00 -3.21353883e-01
1.14544499e+00 -1.02009463e+00 -1.98319107e-01 6.70390010e-01
-1.12580192e+00 -4.08257514e-01 -3.46500963e-01 3.98795098e-01
-4.81651157e-01 6.52949750e-01 -1.03504252e+00 -6.55909181e-01
1.83612734e-01 -1.11790586e+00 9.26900089e-01 -3.31153333e-01
-2.71152198e-01 -1.03603339e+00 1.70493096e-01 1.31649047e-01
9.52829421e-02 2.40053058e-01 7.48319983e-01 -9.44258988e-01
-4.72903103e-01 -2.26341397e-01 -1.70810938e-01 6.32366955e-01
-1.66846961e-02 1.16941445e-01 -1.03970647e+00 -3.60384047e-01
1.25700206e-01 -2.13554293e-01 1.28211212e+00 3.75043005e-01
1.61852288e+00 -5.23275256e-01 -1.94043517e-01 6.93131208e-01
1.05841792e+00 7.57814646e-02 7.18243837e-01 4.41845983e-01
1.07138956e+00 4.91593659e-01 4.47034329e-01 4.37999308e-01
2.24823743e-01 5.23505569e-01 6.07221484e-01 1.24437988e-01
2.95297146e-01 -3.73761021e-02 8.70312572e-01 7.17359424e-01
-2.34153584e-01 -2.51219928e-01 -9.41110075e-01 6.10852003e-01
-2.35909462e+00 -1.25677478e+00 -1.76413208e-01 2.11215615e+00
3.52779478e-01 3.18421513e-01 3.44644278e-01 9.96887535e-02
5.51160693e-01 2.67816931e-01 -3.19248080e-01 -5.48660517e-01
5.87555729e-02 -2.95489818e-01 3.67649257e-01 3.96007240e-01
-1.59261441e+00 8.94278228e-01 6.30435467e+00 5.67960680e-01
-1.02728987e+00 9.84268710e-02 7.31554508e-01 -3.10868442e-01
1.25237912e-01 -7.98839927e-02 -3.72953922e-01 6.53280914e-01
1.05478942e+00 4.64290679e-01 1.55184895e-01 7.47032642e-01
1.15784459e-01 2.46062763e-02 -1.41723311e+00 8.78125548e-01
2.42649138e-01 -1.04554236e+00 2.01865420e-01 7.78901428e-02
5.97371042e-01 2.74426907e-01 -1.21838339e-01 2.71930188e-01
-7.08981827e-02 -8.83305848e-01 6.43905878e-01 5.19946516e-01
3.62930804e-01 -7.68448412e-01 7.96382606e-01 3.23614180e-01
-1.19798386e+00 -3.15407455e-01 -2.31067717e-01 -2.81683028e-01
1.10087045e-01 3.47989529e-01 -1.21265724e-01 3.47147375e-01
1.10736561e+00 1.17840707e+00 -6.33835375e-01 8.55091810e-01
-2.16120750e-01 8.07534397e-01 -2.49501243e-01 7.54715919e-01
6.65110171e-01 -2.50715315e-01 8.04558694e-01 1.46405554e+00
5.20438254e-02 9.67729930e-03 4.46083397e-01 5.65031707e-01
-2.31032241e-02 -2.83172522e-02 -8.71445954e-01 8.00581276e-02
-1.61682323e-01 9.19896603e-01 -6.05402231e-01 -3.98483902e-01
-1.04549587e+00 1.30125678e+00 6.40753031e-01 4.70704913e-01
-9.06340361e-01 8.22199807e-02 7.45750785e-01 2.25411773e-01
3.97498608e-01 -2.00349823e-01 1.11160325e-02 -1.73823893e+00
2.57603467e-01 -9.57016230e-01 9.57398355e-01 -1.69482052e-01
-1.35283518e+00 5.26762664e-01 -1.17730290e-01 -1.33723176e+00
-3.99245501e-01 -7.25190580e-01 -9.90836918e-01 3.11712742e-01
-1.47726595e+00 -1.00127518e+00 -1.48087025e-01 8.76944423e-01
6.97436631e-01 -2.75049269e-01 6.51215613e-01 4.65988219e-01
-1.05543208e+00 7.17294216e-01 -9.27850604e-02 6.30274236e-01
7.26219594e-01 -1.26096320e+00 2.67130077e-01 1.55427754e+00
2.53648371e-01 3.05694699e-01 3.76171559e-01 -6.14582598e-01
-7.23102927e-01 -1.19137645e+00 7.69843042e-01 -5.43961227e-01
8.24151278e-01 -3.35402131e-01 -1.41168678e+00 1.20281374e+00
8.44951421e-02 5.98622501e-01 7.18328595e-01 1.99978217e-01
-5.66746593e-01 1.73970595e-01 -8.02215159e-01 3.64408612e-01
1.17413807e+00 -7.52102196e-01 -5.53816676e-01 1.55412152e-01
3.99140567e-01 -2.98082411e-01 -5.31462908e-01 3.78761113e-01
2.51417935e-01 -1.28673756e+00 6.30482197e-01 -1.06302452e+00
5.20340800e-01 -1.97725266e-01 -1.35851905e-01 -1.03163242e+00
-4.84857440e-01 -5.52680314e-01 -6.96783841e-01 1.03993011e+00
2.73230553e-01 -3.90796930e-01 8.21634471e-01 6.12704039e-01
-3.28069448e-01 -7.05133915e-01 -8.61867607e-01 -8.94439399e-01
-9.93998051e-02 -6.80761874e-01 1.81550793e-02 1.12937522e+00
1.95234641e-01 2.41212219e-01 -6.26531065e-01 4.19649035e-01
4.52468216e-01 -4.22728002e-01 4.39432681e-01 -1.17784584e+00
-9.64370072e-02 -4.45210069e-01 -9.38999534e-01 -1.10325646e+00
4.68790740e-01 -7.55591989e-01 -7.62522817e-02 -8.80026996e-01
2.96299100e-01 6.80954382e-02 -6.42937303e-01 5.95593333e-01
-2.10422933e-01 2.81656057e-01 -1.60048306e-01 3.11014533e-01
-1.06424677e+00 4.81942147e-01 6.23537004e-01 2.24814117e-02
-2.97355920e-01 -5.61650507e-02 -2.38593325e-01 1.28743708e+00
7.62917399e-01 -4.98373836e-01 -2.76410192e-01 -7.00144053e-01
-1.53418565e-02 -2.58359730e-01 3.69960874e-01 -1.10915530e+00
2.30993301e-01 5.84721789e-02 3.54052663e-01 -3.30784708e-01
6.28650235e-03 -8.88141394e-01 -7.11758554e-01 2.50969529e-01
-4.15458560e-01 2.00095296e-01 1.26494735e-01 9.39986765e-01
-6.26830757e-01 -2.18170524e-01 7.73698211e-01 -2.00229976e-02
-9.29315865e-01 6.60543978e-01 -7.77442157e-01 1.72923088e-01
1.09760571e+00 -1.06367044e-01 5.41662425e-03 -5.48924744e-01
-1.04000115e+00 3.24998260e-01 3.08119327e-01 5.81666946e-01
5.17734528e-01 -1.49542356e+00 -6.62678838e-01 5.00280261e-01
3.40462863e-01 -1.89426318e-01 2.13247612e-01 1.07253432e+00
-4.35962766e-01 2.14623570e-01 -2.69818276e-01 -8.91027510e-01
-1.37056994e+00 5.24210870e-01 4.12352920e-01 -3.87718648e-01
-7.07374632e-01 9.00738895e-01 3.78952235e-01 9.57144483e-04
5.32202661e-01 -3.94403696e-01 -2.67878562e-01 8.77336487e-02
6.32778406e-01 4.04498935e-01 -6.27476606e-04 -9.42802906e-01
-4.73733842e-01 3.92258614e-01 -5.24565220e-01 2.86523432e-01
1.22837520e+00 -1.64304629e-01 -2.33873114e-01 4.90769923e-01
1.41985798e+00 -2.89615780e-01 -1.58517802e+00 -3.09491575e-01
4.58046347e-01 -5.27238667e-01 4.78436686e-02 -1.73574880e-01
-1.21896744e+00 8.73339653e-01 4.62515146e-01 3.28147858e-01
1.22364724e+00 -1.25125926e-02 5.90234518e-01 4.12105292e-01
-2.15017751e-01 -1.26564705e+00 4.07412976e-01 8.13932419e-01
5.71889997e-01 -1.44393408e+00 -1.59294367e-01 -2.27717057e-01
-6.46837473e-01 1.11607563e+00 1.07044959e+00 -3.46883714e-01
6.34717464e-01 9.73348320e-02 -7.52380565e-02 -3.52669537e-01
-7.54338324e-01 -1.54706359e-01 5.07152855e-01 2.68017918e-01
5.67740202e-01 -4.39988047e-01 4.07174490e-02 4.21156436e-01
6.18788660e-01 -2.63405710e-01 4.43010181e-01 1.15367246e+00
-3.25384289e-01 -1.05507326e+00 -1.50142983e-01 7.08796203e-01
-1.09963393e+00 4.98771928e-02 -2.47130945e-01 5.65323770e-01
2.59505242e-01 7.36083329e-01 5.02915204e-01 -2.70192683e-01
1.82890236e-01 3.84307325e-01 2.34679192e-01 -6.09023750e-01
-2.95705378e-01 2.34627798e-01 -9.50448122e-03 -1.11451256e+00
-5.57332933e-01 -8.11301708e-01 -1.05749404e+00 -9.86140147e-02
-1.23521835e-01 -3.44500206e-02 -6.05532713e-02 1.17861176e+00
3.03572327e-01 5.34595788e-01 4.71492589e-01 -7.47373879e-01
-1.30770415e-01 -7.24006832e-01 -5.31264007e-01 8.17268789e-01
7.90762007e-01 -5.49721003e-01 -5.64389944e-01 2.00856164e-01] | [7.853390216827393, 1.5923407077789307] |
03337c27-4fe9-4556-99c2-7e0bef24c5c4 | glean-generative-latent-bank-for-image-super | 2207.14812 | null | https://arxiv.org/abs/2207.14812v1 | https://arxiv.org/pdf/2207.14812v1.pdf | GLEAN: Generative Latent Bank for Image Super-Resolution and Beyond | We show that pre-trained Generative Adversarial Networks (GANs) such as StyleGAN and BigGAN can be used as a latent bank to improve the performance of image super-resolution. While most existing perceptual-oriented approaches attempt to generate realistic outputs through learning with adversarial loss, our method, Generative LatEnt bANk (GLEAN), goes beyond existing practices by directly leveraging rich and diverse priors encapsulated in a pre-trained GAN. But unlike prevalent GAN inversion methods that require expensive image-specific optimization at runtime, our approach only needs a single forward pass for restoration. GLEAN can be easily incorporated in a simple encoder-bank-decoder architecture with multi-resolution skip connections. Employing priors from different generative models allows GLEAN to be applied to diverse categories (\eg~human faces, cats, buildings, and cars). We further present a lightweight version of GLEAN, named LightGLEAN, which retains only the critical components in GLEAN. Notably, LightGLEAN consists of only 21% of parameters and 35% of FLOPs while achieving comparable image quality. We extend our method to different tasks including image colorization and blind image restoration, and extensive experiments show that our proposed models perform favorably in comparison to existing methods. Codes and models are available at https://github.com/open-mmlab/mmediting. | ['Chen Change Loy', 'Jinwei Gu', 'Xintao Wang', 'Xiangyu Xu', 'Kelvin C. K. Chan'] | 2022-07-29 | null | null | null | null | ['colorization'] | ['computer-vision'] | [ 4.26362038e-01 2.42834315e-01 1.42422944e-01 -2.36563757e-01
-9.72563565e-01 -6.01816595e-01 6.65096581e-01 -9.46748614e-01
-8.84462241e-03 6.02958620e-01 2.66482353e-01 -2.56869256e-01
4.91417497e-01 -9.00933266e-01 -1.11817312e+00 -5.98936439e-01
3.77273023e-01 2.02776089e-01 -1.60783395e-01 -3.22489679e-01
-8.59697163e-02 3.12198818e-01 -1.16147387e+00 4.62225050e-01
1.00698340e+00 6.62397623e-01 7.98987225e-02 7.02316701e-01
1.40006989e-01 1.08659279e+00 -5.78232229e-01 -6.96076274e-01
5.90215147e-01 -8.23176622e-01 -6.67662919e-01 2.26577669e-01
5.30593455e-01 -8.28579009e-01 -3.95491004e-01 1.07467151e+00
6.28531754e-01 -8.02498311e-02 4.84820992e-01 -1.19030607e+00
-1.21467030e+00 5.87229848e-01 -6.85004771e-01 -2.77076155e-01
2.83044130e-01 5.42226911e-01 7.10658312e-01 -1.07839274e+00
4.76900965e-01 1.53616369e+00 7.11221576e-01 8.86648893e-01
-1.46937871e+00 -9.44122970e-01 1.01666868e-01 -5.90763660e-03
-1.20025325e+00 -8.03469002e-01 7.93118298e-01 -1.78856269e-01
6.19196236e-01 1.94707900e-01 4.31930274e-01 1.35819232e+00
-5.14342263e-02 7.76652694e-01 1.30380118e+00 -3.02328408e-01
1.47663010e-02 -1.45771459e-01 -6.47756815e-01 8.17966700e-01
1.04807466e-01 1.28604785e-01 -4.60058689e-01 1.49954870e-01
1.23289299e+00 -7.07900375e-02 -4.45979685e-01 -1.38670459e-01
-1.08333910e+00 8.73988450e-01 8.07548761e-01 -2.36290976e-01
-3.46587449e-01 6.11396670e-01 -4.81746756e-02 2.78093189e-01
5.74576497e-01 2.75650799e-01 3.96188162e-02 2.27500975e-01
-9.80725944e-01 6.62706196e-02 4.94262189e-01 1.05861020e+00
9.58122790e-01 5.95753789e-01 -1.79340586e-01 8.75973821e-01
2.33674973e-01 6.02364540e-01 2.34585285e-01 -1.31975746e+00
3.41529936e-01 3.01763386e-01 1.29204050e-01 -7.66325772e-01
2.26575762e-01 -4.18246239e-01 -1.02109087e+00 6.35963678e-01
5.88601902e-02 -3.45539808e-01 -1.39913249e+00 1.78008533e+00
1.71047568e-01 3.25968623e-01 -1.04038492e-02 8.26981246e-01
8.19269180e-01 7.17440486e-01 -9.30593908e-02 2.32723564e-01
1.23332715e+00 -1.37823391e+00 -5.07275343e-01 -6.04587972e-01
1.23116830e-02 -1.01783502e+00 1.11943853e+00 2.47572914e-01
-1.58998239e+00 -5.75031042e-01 -1.02206159e+00 -3.80971760e-01
-6.67052194e-02 1.94799025e-02 6.71183825e-01 6.58286154e-01
-1.56445956e+00 3.50802988e-01 -8.82705808e-01 -1.48494184e-01
7.03155935e-01 2.82977909e-01 -2.47782081e-01 -4.65440065e-01
-8.37628126e-01 6.39400125e-01 4.59811799e-02 2.29959592e-01
-1.33355618e+00 -7.77233481e-01 -1.09706008e+00 -8.93619433e-02
1.98766068e-01 -1.21552145e+00 1.13227618e+00 -1.34314978e+00
-1.87131464e+00 6.74806356e-01 -1.94647774e-01 -3.58112246e-01
6.32452428e-01 -3.52054030e-01 -2.03926802e-01 2.54476249e-01
-3.11193466e-02 1.03684342e+00 1.23392344e+00 -1.58548450e+00
-2.05969155e-01 1.99538410e-01 3.39671135e-01 1.43793240e-01
-5.59449233e-02 8.85129198e-02 -7.86907136e-01 -8.83160889e-01
-1.12540983e-01 -8.95776510e-01 -2.99545109e-01 1.16861716e-01
-5.65369904e-01 4.59367394e-01 7.20171094e-01 -9.47759628e-01
7.30611205e-01 -2.03145289e+00 3.29087675e-01 -6.23236746e-02
2.65310585e-01 2.62037516e-01 -4.79429603e-01 4.23742473e-01
-1.62527665e-01 2.27167368e-01 -5.65225005e-01 -7.38304257e-01
1.71547547e-01 2.31494129e-01 -4.43010688e-01 3.95567060e-01
3.68663698e-01 1.25482500e+00 -6.82817519e-01 -2.20320240e-01
4.21188205e-01 9.98049438e-01 -8.56198609e-01 3.89519602e-01
-2.15893209e-01 6.32530868e-01 1.38239339e-02 8.16508949e-01
8.86617482e-01 -4.20580685e-01 6.61795959e-02 -2.72572070e-01
2.63626605e-01 2.38075286e-01 -9.35302079e-01 1.75105238e+00
-6.77063227e-01 5.35788357e-01 2.63991922e-01 -7.63695896e-01
6.53216422e-01 2.37731516e-01 9.46120080e-03 -7.66043365e-01
-1.03239909e-01 8.42647702e-02 -2.87538230e-01 -9.69378874e-02
4.13583666e-01 -1.61486454e-02 7.53550306e-02 4.03824151e-01
1.40326217e-01 -2.50815898e-01 2.56034248e-02 3.85317326e-01
1.00540876e+00 5.16948164e-01 -3.02539114e-02 1.81514740e-01
2.72848666e-01 -3.22562903e-01 5.11021495e-01 6.08768761e-01
2.61745840e-01 1.07545269e+00 3.54504794e-01 -1.89136863e-01
-1.30803478e+00 -1.25915611e+00 2.56921202e-01 9.81856942e-01
-1.18190581e-02 -1.87229633e-01 -8.90853524e-01 -5.07695138e-01
-2.96244055e-01 6.52163863e-01 -5.67164361e-01 5.20577431e-02
-6.15772843e-01 -9.06837165e-01 5.12211800e-01 5.18658400e-01
8.53193521e-01 -1.22422338e+00 -1.45063758e-01 -2.95933373e-02
-2.12745026e-01 -1.11724186e+00 -6.10970080e-01 -1.95626527e-01
-7.57009983e-01 -8.23508143e-01 -9.28411484e-01 -6.90654397e-01
1.02167189e+00 2.14311779e-01 1.44386208e+00 1.46563739e-01
-1.68704882e-01 3.28862786e-01 -3.35313797e-01 -7.02157766e-02
-6.18757486e-01 -1.21865511e-01 -3.52133751e-01 5.29648587e-02
-1.83527082e-01 -8.60677838e-01 -1.13859725e+00 2.20811650e-01
-1.04200304e+00 5.38765669e-01 8.70245457e-01 9.94515121e-01
6.00656331e-01 -3.05387735e-01 3.92857403e-01 -1.22790527e+00
2.86852837e-01 -3.81066948e-01 -6.27066672e-01 3.25866342e-02
-5.53355098e-01 -4.06419933e-02 7.28170753e-01 -2.39822358e-01
-1.24003243e+00 1.55575166e-03 -3.64424586e-01 -4.70443338e-01
-4.41344976e-02 1.00272581e-01 -3.08039129e-01 -3.55299383e-01
4.67447460e-01 3.21164221e-01 3.56769818e-03 -4.47758436e-01
8.43614459e-01 2.37574860e-01 8.10955644e-01 -4.12161767e-01
1.30220473e+00 6.51228845e-01 -2.65305161e-01 -2.82483190e-01
-6.52962983e-01 1.51042596e-01 -2.44659528e-01 -4.35055979e-02
7.34458923e-01 -1.38307130e+00 -4.50150311e-01 6.31296694e-01
-9.31246996e-01 -7.84081757e-01 -2.93240368e-01 7.61900842e-02
-6.22704864e-01 2.99181193e-01 -8.91646981e-01 -4.58372414e-01
-6.17318511e-01 -1.14782143e+00 1.14140034e+00 2.07487866e-01
1.51588574e-01 -9.99565005e-01 -2.55449653e-01 5.59054255e-01
7.40599036e-01 5.12024581e-01 6.04481041e-01 2.02696234e-01
-1.06588674e+00 7.79875293e-02 -3.62417758e-01 6.31966770e-01
1.37918800e-01 -8.92689750e-02 -1.06757474e+00 -5.67331314e-01
-1.62357584e-01 -4.85750943e-01 1.15408659e+00 3.64564925e-01
1.25907445e+00 -6.54094100e-01 -5.73292412e-02 1.34674633e+00
1.69489133e+00 -1.36168543e-02 1.17711711e+00 2.69918293e-01
1.07567978e+00 5.65601774e-02 1.40386030e-01 2.16546312e-01
5.41781127e-01 4.81584311e-01 7.22966850e-01 -7.23797023e-01
-7.02432692e-01 -3.40185821e-01 7.25962460e-01 8.27935040e-01
-3.03013176e-01 -3.83541495e-01 -5.16852558e-01 5.57697892e-01
-1.49848580e+00 -1.00584280e+00 2.85260201e-01 1.88871944e+00
1.02337563e+00 -2.05508590e-01 -1.08532362e-01 -1.86136648e-01
5.02435565e-01 3.66811097e-01 -6.86863363e-01 -3.08524430e-01
-2.14694589e-01 7.21541822e-01 6.58720613e-01 7.66737759e-01
-8.70269299e-01 1.13551819e+00 6.64933443e+00 7.95028150e-01
-1.05322838e+00 2.30055690e-01 8.39990914e-01 -1.98117122e-01
-7.28558958e-01 9.56045762e-02 -4.40346152e-01 4.61972803e-01
7.42734909e-01 1.77042693e-01 8.07456195e-01 6.26907110e-01
6.95838481e-02 2.02944383e-01 -8.93705189e-01 8.86808634e-01
1.89848483e-01 -1.41047907e+00 3.87114644e-01 -3.52649726e-02
1.09682775e+00 6.39927760e-02 5.52614629e-01 1.77864328e-01
8.85839701e-01 -1.36247301e+00 7.27152765e-01 4.74696279e-01
1.24757516e+00 -7.48552620e-01 4.11401272e-01 -1.58103749e-01
-9.85385597e-01 1.81827873e-01 -3.15362900e-01 2.06134468e-01
2.34636009e-01 4.79952097e-01 -5.81396043e-01 6.25658214e-01
7.03238785e-01 7.98953831e-01 -6.53010547e-01 6.03961706e-01
-6.88227952e-01 7.91905940e-01 -1.01182405e-02 7.79982209e-01
8.61165300e-02 -2.04676494e-01 3.49429429e-01 9.84631956e-01
4.16723341e-01 -4.10931557e-02 -8.18845481e-02 1.10624826e+00
-3.97454262e-01 -3.54279906e-01 -4.20345545e-01 1.28073111e-01
3.43297809e-01 1.23270142e+00 -4.36048269e-01 -3.11206251e-01
-5.79258323e-01 1.52917039e+00 2.06720829e-01 7.92703331e-01
-1.07725108e+00 -3.16727340e-01 6.94445074e-01 1.09334365e-01
5.73688865e-01 -1.05453543e-01 -2.46434689e-01 -1.37795758e+00
-3.15914974e-02 -1.27842939e+00 9.18319169e-03 -1.09959531e+00
-1.28186727e+00 8.78552318e-01 -3.15362960e-01 -1.12143278e+00
-4.50376153e-01 -4.37469095e-01 -6.34110391e-01 1.10007000e+00
-1.87278235e+00 -1.73176444e+00 -5.84289014e-01 8.93849015e-01
5.96350193e-01 -3.88509855e-02 8.05164814e-01 2.95343757e-01
-6.68016434e-01 7.94658482e-01 6.62769526e-02 2.69573778e-01
8.58008862e-01 -1.32878184e+00 8.88995767e-01 1.35564959e+00
3.61171667e-03 5.52595854e-01 6.53168797e-01 -4.73118991e-01
-1.41942072e+00 -1.33269072e+00 2.45548189e-01 -3.09753507e-01
3.85833561e-01 -4.89800841e-01 -6.57241702e-01 1.06857586e+00
7.96045184e-01 6.32709116e-02 4.16292548e-01 -2.56274879e-01
-5.33466637e-01 -9.11349505e-02 -1.07655931e+00 7.84238040e-01
1.15032494e+00 -6.19112194e-01 -1.51379835e-02 3.02027136e-01
7.38705635e-01 -6.68183506e-01 -6.88053668e-01 2.84610093e-01
3.70598555e-01 -1.10848272e+00 1.28341019e+00 -1.09454006e-01
8.30828846e-01 -4.92972553e-01 -1.24670319e-01 -1.34896481e+00
-4.45421964e-01 -8.95327985e-01 -1.67728484e-01 1.19933581e+00
2.21222565e-01 -7.56650269e-01 6.12375975e-01 3.88787389e-01
-1.96103185e-01 -5.42035103e-01 -4.02016640e-01 -5.17021060e-01
7.51117617e-02 -2.03287214e-01 7.81139314e-01 7.96918690e-01
-7.69369602e-01 1.43103644e-01 -9.08925712e-01 3.10656339e-01
9.16469097e-01 1.68197080e-01 1.04990005e+00 -5.44165432e-01
-6.04137719e-01 -1.55362993e-01 -9.05831680e-02 -1.17816353e+00
-4.94604446e-02 -8.07808459e-01 2.16188598e-02 -1.75825894e+00
2.37842500e-01 -3.43383580e-01 -1.24867782e-01 7.61308312e-01
-2.78200060e-01 1.06098545e+00 3.26863289e-01 2.74683982e-01
-3.57716262e-01 5.74187577e-01 1.40748703e+00 -2.18841165e-01
1.35894060e-01 -3.50043148e-01 -1.03681087e+00 7.72206306e-01
7.26351321e-01 -2.41734624e-01 -4.74780649e-01 -9.13358390e-01
2.06956223e-01 -1.05198137e-01 7.19479561e-01 -1.01229227e+00
-6.06813096e-03 -1.81293990e-02 6.77680194e-01 -1.45460024e-01
5.43176413e-01 -5.52647531e-01 5.66792130e-01 2.35706493e-01
-5.24850041e-02 -8.11976753e-03 2.11821407e-01 3.40685099e-01
-2.43095785e-01 1.68322727e-01 8.99599373e-01 -2.81751931e-01
-6.84354901e-01 4.00975108e-01 -6.59339726e-02 -3.68865691e-02
7.02963471e-01 -1.41283303e-01 -5.31812489e-01 -6.97965741e-01
-5.39051771e-01 -4.10588868e-02 9.16163146e-01 2.23338857e-01
7.70215929e-01 -1.31119645e+00 -9.70030606e-01 2.27053702e-01
-2.56281137e-01 2.26888359e-01 3.91782939e-01 5.34650147e-01
-8.16768765e-01 -5.20062372e-02 -3.59368175e-01 -3.51607591e-01
-9.58797634e-01 4.89562869e-01 3.62084806e-01 -1.70937136e-01
-7.34955609e-01 1.02578270e+00 7.25878954e-01 -1.37137011e-01
-1.39581054e-01 -2.99644154e-02 2.45063946e-01 -4.24553841e-01
4.67169046e-01 -1.39353517e-02 -1.94977924e-01 -6.08944893e-01
-1.29377425e-01 4.73749071e-01 -6.65539578e-02 -1.97183549e-01
1.46019971e+00 -2.29174852e-01 -2.84576118e-01 -9.23211500e-02
8.86930704e-01 2.85171896e-01 -1.75376272e+00 -1.04070373e-01
-7.47313499e-01 -6.00878775e-01 5.50255366e-02 -9.67712820e-01
-1.65041697e+00 7.50393867e-01 4.89370912e-01 -3.13253164e-01
1.75444257e+00 -1.34068504e-01 8.10555696e-01 -1.73882559e-01
4.09007758e-01 -4.54803497e-01 2.40451038e-01 1.09332569e-01
1.12271738e+00 -1.18495727e+00 4.35263403e-02 -4.89840120e-01
-6.59889638e-01 7.41322815e-01 6.26307368e-01 -4.37203765e-01
2.18602940e-01 3.82771045e-01 2.37605274e-01 8.99002850e-02
-6.80533767e-01 -1.79752514e-01 2.53576636e-01 6.99586391e-01
4.44514364e-01 -2.36752499e-02 5.68794236e-02 1.71361119e-01
-2.78143674e-01 -4.26365957e-02 6.67261541e-01 7.40722656e-01
7.26498291e-02 -1.35044813e+00 -3.80899966e-01 1.35484636e-01
-5.72572589e-01 -5.34859061e-01 -6.77969381e-02 6.18376493e-01
1.45701498e-01 8.38640571e-01 -5.76603375e-02 -3.07678640e-01
-1.00009151e-01 -2.11198062e-01 4.66899991e-01 -5.45085847e-01
-4.29159135e-01 1.61562845e-01 -8.52197185e-02 -9.09866750e-01
-6.13511503e-01 -2.94053853e-01 -7.60461569e-01 -6.13291323e-01
-4.71537299e-02 -3.11298549e-01 4.93711114e-01 5.47228336e-01
5.51234484e-01 8.45310152e-01 6.11676455e-01 -9.81789649e-01
-9.98470858e-02 -9.05860126e-01 -2.26482272e-01 4.06171590e-01
4.71932232e-01 -3.16390961e-01 -3.03135186e-01 5.27341008e-01] | [11.552242279052734, -0.6509395837783813] |
ec12364b-f03f-4e9a-a76c-683f3f08a310 | a-simple-local-minimal-intensity-prior-and-an | 1906.06642 | null | http://arxiv.org/abs/1906.06642v5 | http://arxiv.org/pdf/1906.06642v5.pdf | A Simple Local Minimal Intensity Prior and An Improved Algorithm for Blind Image Deblurring | Blind image deblurring is a long standing challenging problem in image
processing and low-level vision. Recently, sophisticated priors such as dark
channel prior, extreme channel prior, and local maximum gradient prior, have
shown promising effectiveness. However, these methods are computationally
expensive. Meanwhile, since these priors involved subproblems cannot be solved
explicitly, approximate solution is commonly used, which limits the best
exploitation of their capability. To address these problems, this work firstly
proposes a simplified sparsity prior of local minimal pixels, namely patch-wise
minimal pixels (PMP). The PMP of clear images is much more sparse than that of
blurred ones, and hence is very effective in discriminating between clear and
blurred images. Then, a novel algorithm is designed to efficiently exploit the
sparsity of PMP in deblurring. The new algorithm flexibly imposes sparsity
inducing on the PMP under the MAP framework rather than directly uses the half
quadratic splitting algorithm. By this, it avoids non-rigorous approximation
solution in existing algorithms, while being much more computationally
efficient. Extensive experiments demonstrate that the proposed algorithm can
achieve better practical stability compared with state-of-the-arts. In terms of
deblurring quality, robustness and computational efficiency, the new algorithm
is superior to state-of-the-arts. Code for reproducing the results of the new
method is available at https://github.com/FWen/deblur-pmp.git. | [] | 2020-10-29 | null | null | null | null | ['blind-image-deblurring'] | ['computer-vision'] | [ 1.95051506e-01 -4.90656585e-01 -5.10097742e-02 1.34449348e-01
-4.90212739e-01 -2.54590183e-01 3.92436326e-01 -6.67043984e-01
-2.08563313e-01 8.08654606e-01 3.90624136e-01 -5.47553487e-02
-1.03377670e-01 -3.82368147e-01 -4.62827951e-01 -1.05009735e+00
2.51351625e-01 -3.62928122e-01 1.60073042e-01 1.28850937e-01
4.75846261e-01 2.28503853e-01 -1.38544810e+00 -1.84885979e-01
1.32590759e+00 9.11956668e-01 5.99872947e-01 2.54081041e-01
8.06030631e-02 4.47677761e-01 -8.65407139e-02 1.19540870e-01
4.49345350e-01 -6.08813405e-01 -4.52319890e-01 3.38077515e-01
3.90480697e-01 -5.22615850e-01 -6.44004405e-01 1.62238681e+00
6.25628531e-01 1.59811541e-01 4.59542632e-01 -7.97777414e-01
-7.97090948e-01 7.90459961e-02 -1.01949990e+00 3.97279024e-01
2.49235526e-01 2.35401213e-01 5.61037719e-01 -9.86730278e-01
2.32072264e-01 9.49134588e-01 5.95535636e-01 1.68613479e-01
-9.84989643e-01 -6.56035721e-01 -4.89520729e-02 4.13545102e-01
-1.63246191e+00 -6.08412683e-01 8.03123474e-01 -3.92689109e-01
3.15027148e-01 3.88455689e-01 4.45588917e-01 6.21096671e-01
1.47206649e-01 7.41365194e-01 1.51521313e+00 -3.39688510e-01
1.34176821e-01 -2.86795329e-02 1.66976422e-01 5.49684644e-01
4.49153721e-01 2.13400885e-01 -4.00729746e-01 -1.56626046e-01
1.00400472e+00 2.18391806e-01 -1.17504346e+00 -1.53545335e-01
-1.27383590e+00 4.08456832e-01 3.41207832e-01 3.20824474e-01
-4.10811394e-01 -7.50716105e-02 -9.83089581e-02 8.08996335e-02
5.51176965e-01 1.13667138e-01 -1.39059722e-01 6.55684015e-03
-1.17885923e+00 9.83676612e-02 4.91482824e-01 7.22234428e-01
7.96609998e-01 5.78275919e-02 -1.44388840e-01 1.02772450e+00
4.31726724e-01 6.88970089e-01 5.58716357e-01 -1.05327797e+00
1.93771437e-01 1.32046893e-01 4.89949763e-01 -1.15188956e+00
5.57881482e-02 -5.80364823e-01 -1.30704939e+00 5.63443005e-02
2.36347035e-01 -9.34183151e-02 -8.33510220e-01 1.27901161e+00
3.10699522e-01 5.26373804e-01 -5.07885329e-02 1.39037263e+00
5.60224831e-01 9.95835423e-01 -4.74482983e-01 -6.94863498e-01
1.32736588e+00 -9.69984889e-01 -9.76247489e-01 -3.97212148e-01
-1.45737687e-02 -1.05026031e+00 7.45158732e-01 5.41540086e-01
-1.09206319e+00 -5.73864043e-01 -9.89546776e-01 1.03132486e-01
1.47268951e-01 2.70192415e-01 4.19695526e-01 5.86852729e-01
-9.89470601e-01 4.32594419e-01 -5.99198401e-01 -2.16617823e-01
3.37046087e-01 4.37804349e-02 -9.30862129e-02 -5.01264036e-01
-1.12757587e+00 8.85445535e-01 2.75862217e-01 4.23231751e-01
-7.60313272e-01 -5.53563714e-01 -6.74173057e-01 -8.78457446e-03
3.27418536e-01 -6.62046790e-01 9.47603703e-01 -8.33510518e-01
-1.66631258e+00 4.76973295e-01 -4.75690544e-01 -4.55905721e-02
4.73024100e-01 -5.17603040e-01 -3.17087084e-01 3.26475024e-01
-7.46343955e-02 2.26036966e-01 1.35640740e+00 -1.34150076e+00
-3.50217253e-01 -1.72574610e-01 -2.85687327e-01 3.97175640e-01
-3.42165202e-01 5.12552597e-02 -6.37474537e-01 -7.68920183e-01
4.10213768e-01 -7.36562908e-01 -2.44020402e-01 1.44198444e-03
-2.52725273e-01 2.71712065e-01 7.49315143e-01 -9.98613358e-01
1.31320226e+00 -2.28783655e+00 4.61992733e-02 -1.93808734e-01
1.59801930e-01 5.24452806e-01 1.13421626e-01 3.09180260e-01
-2.83820033e-02 -2.41237864e-01 -6.45961761e-01 -3.90864275e-02
-2.86588222e-01 -5.92400730e-02 -3.19596261e-01 9.96866941e-01
-1.28208220e-01 5.37148237e-01 -7.29825079e-01 -2.98986018e-01
4.82921660e-01 5.67652941e-01 -3.40716422e-01 2.15348870e-01
1.72234863e-01 7.30848193e-01 -5.32041788e-01 5.99633157e-01
1.30381334e+00 -3.17209840e-01 -1.98205739e-01 -3.87078285e-01
-5.42588592e-01 -2.50997871e-01 -1.38187397e+00 1.56806540e+00
-3.53310376e-01 5.85026979e-01 5.14640212e-01 -9.91274357e-01
7.40296125e-01 3.57715845e-01 2.51935244e-01 -4.34689701e-01
1.07915603e-01 3.45303595e-01 -2.84721971e-01 -7.50693977e-01
4.19196904e-01 -7.34587982e-02 5.19767880e-01 2.12460175e-01
-4.62220788e-01 -1.75693095e-01 5.87433949e-02 -1.99590232e-02
7.41277754e-01 1.03071854e-01 4.03495103e-01 -4.06164557e-01
8.22520375e-01 -2.79722661e-01 6.94466829e-01 5.68764448e-01
-2.84514457e-01 9.35339749e-01 -7.97820911e-02 -3.06174308e-02
-7.83876777e-01 -8.81665766e-01 -3.70737195e-01 4.65549141e-01
8.52300346e-01 -8.39623362e-02 -8.03957045e-01 -9.65573415e-02
-2.87403613e-01 3.88083875e-01 -1.84253171e-01 1.29965752e-01
-5.83272219e-01 -1.09716904e+00 1.02732584e-01 -5.61303739e-03
1.10001051e+00 -5.35357058e-01 -4.11006272e-01 -2.64063086e-02
-5.52323222e-01 -1.08651578e+00 -7.65452087e-01 -4.91520733e-01
-9.77194250e-01 -9.52362895e-01 -1.42524314e+00 -8.32404137e-01
9.39455807e-01 1.01548707e+00 5.84934533e-01 5.96704967e-02
-1.90335169e-01 2.47180521e-01 -4.53706682e-01 2.01353624e-01
-3.86939794e-02 -5.57529867e-01 7.95508549e-03 4.33701128e-01
-5.13660237e-02 -6.99409068e-01 -1.06058717e+00 5.46706796e-01
-1.03036177e+00 2.54997909e-01 1.02462101e+00 9.60232079e-01
4.77155447e-01 4.40554023e-01 1.72850281e-01 -4.37303960e-01
5.49237669e-01 -3.63420248e-01 -6.91650569e-01 4.98072505e-02
-7.21702576e-01 -1.30454093e-01 5.97737968e-01 -3.84266526e-01
-1.39172018e+00 -1.05773233e-01 2.24414274e-01 -5.22453845e-01
-1.50713399e-01 4.26805645e-01 -1.36400729e-01 -3.83280784e-01
3.92024636e-01 8.10564339e-01 8.34010243e-02 -7.68568039e-01
2.92353272e-01 9.54806149e-01 6.66865647e-01 -3.10355872e-01
9.39267576e-01 5.91438770e-01 -1.63530469e-01 -1.12410235e+00
-4.94975239e-01 -7.50664294e-01 -1.88375100e-01 -1.61993191e-01
6.45234585e-01 -1.07342422e+00 -4.64940757e-01 1.00898969e+00
-1.07744503e+00 -7.79763907e-02 1.54157907e-01 8.29084635e-01
-3.34409535e-01 1.09932411e+00 -7.87929058e-01 -8.12714458e-01
-3.22032481e-01 -1.15866351e+00 6.25936627e-01 5.66084385e-01
2.98986167e-01 -7.87018776e-01 -1.07140847e-01 5.16606212e-01
7.10564733e-01 -8.44858959e-02 3.74798596e-01 1.85657665e-01
-7.80945122e-01 -5.07736839e-02 -6.23722911e-01 5.23215234e-01
2.26453304e-01 -3.38649094e-01 -9.41115975e-01 -5.91382205e-01
4.69964743e-01 9.71212238e-02 9.61916864e-01 7.13377953e-01
1.00896358e+00 -4.16109979e-01 -3.22809190e-01 8.22464108e-01
1.50685239e+00 2.74244808e-02 9.25559103e-01 2.66678691e-01
5.71898878e-01 4.02872056e-01 6.63850546e-01 5.56671083e-01
1.19252652e-01 6.27921343e-01 4.20688212e-01 -1.85013458e-01
-3.62811625e-01 1.72503546e-01 5.87953389e-01 9.46419537e-01
-2.35293224e-01 -1.29679337e-01 -6.10783100e-01 6.18213773e-01
-1.89317083e+00 -8.74828219e-01 -3.21810991e-01 2.30920744e+00
9.67591643e-01 -3.31189811e-01 -4.97011602e-01 1.15170717e-01
1.07066286e+00 4.90290165e-01 -4.19781387e-01 2.16677994e-01
-2.16334507e-01 -4.01781760e-02 7.04989910e-01 6.24137819e-01
-9.93080616e-01 6.75002396e-01 5.42148304e+00 1.15346050e+00
-1.00757539e+00 2.65721470e-01 4.72296506e-01 1.43739700e-01
-5.49084693e-02 2.95323133e-01 -5.60853720e-01 8.42837274e-01
4.07176942e-01 -1.60577953e-01 7.15005517e-01 3.96404237e-01
6.96007311e-01 -4.93306160e-01 -5.57320237e-01 1.40051019e+00
1.51716053e-01 -9.91896808e-01 -2.07705006e-01 1.71022654e-01
8.29979181e-01 -1.44728079e-01 1.87565044e-01 -2.59163499e-01
-6.06359690e-02 -8.67905557e-01 5.30813515e-01 6.40046895e-01
7.20802784e-01 -3.92825454e-01 7.56869435e-01 4.40822303e-01
-8.98367822e-01 5.44543862e-02 -4.87065732e-01 -2.01538965e-01
3.32865000e-01 1.22483313e+00 -5.22251576e-02 8.34673524e-01
7.18271077e-01 1.00284564e+00 -1.50207892e-01 1.44324434e+00
-4.16955501e-01 6.71530843e-01 -1.40929610e-01 4.66857821e-01
6.07742220e-02 -6.19208932e-01 8.79652858e-01 1.12344968e+00
6.30986273e-01 4.94114727e-01 8.53121355e-02 8.26263011e-01
2.40390792e-01 3.49516161e-02 -2.48911828e-01 1.85865790e-01
4.10645932e-01 1.15580881e+00 -4.12632614e-01 -2.55079836e-01
-4.96523976e-01 1.18556595e+00 -2.89851040e-01 7.30153441e-01
-6.91499531e-01 -4.04070705e-01 5.70018291e-01 1.69988684e-02
4.27365184e-01 -3.29720140e-01 -1.59028843e-01 -1.51694715e+00
1.69368580e-01 -9.45119917e-01 3.82650346e-02 -9.95434701e-01
-1.28784037e+00 3.48365068e-01 -1.59679428e-01 -1.59934711e+00
3.18141282e-01 -4.08726394e-01 -6.87913060e-01 1.20873141e+00
-1.95437205e+00 -9.19160068e-01 -7.09273219e-01 6.81332588e-01
5.87126851e-01 1.27624974e-01 2.66819656e-01 3.61973464e-01
-7.32320547e-01 1.76758364e-01 6.09118402e-01 -1.86915621e-01
8.91744375e-01 -7.00415432e-01 -3.38712595e-02 1.44399285e+00
-3.86758149e-01 7.29776144e-01 9.54185903e-01 -5.73396444e-01
-1.41692710e+00 -8.15708578e-01 3.26347381e-01 2.22239226e-01
5.17352283e-01 2.75086284e-01 -1.01465642e+00 1.80358842e-01
5.03147244e-01 1.23760991e-01 2.21821129e-01 -4.83479619e-01
-1.28573760e-01 -8.28594789e-02 -1.01032710e+00 4.61871415e-01
7.07942069e-01 -2.46218726e-01 -5.10143459e-01 2.71744132e-01
3.84991884e-01 -4.32588100e-01 -5.67839861e-01 4.27364051e-01
3.11888218e-01 -1.20302248e+00 1.02975249e+00 3.10114831e-01
3.76095533e-01 -8.67332458e-01 -5.76248355e-02 -1.34734988e+00
-5.49414456e-01 -9.61590171e-01 -3.56528789e-01 1.13859642e+00
-7.91318640e-02 -1.09760094e+00 3.84815872e-01 2.49153003e-01
-2.46681497e-01 -5.12133360e-01 -7.31426001e-01 -8.54176700e-01
-2.20672354e-01 1.93346012e-02 1.49146631e-01 1.05488706e+00
-1.33062780e-01 -6.38320968e-02 -8.64207566e-01 5.53100586e-01
9.83004510e-01 3.05083096e-01 5.43901682e-01 -8.65447879e-01
-4.25889641e-01 -4.26272660e-01 -1.95828915e-01 -1.64489377e+00
-2.48374581e-01 -5.21000624e-01 2.46171817e-01 -1.61727321e+00
3.86442810e-01 -3.17272812e-01 -1.21492147e-01 1.81949034e-01
-4.34269309e-01 4.43873852e-01 7.73670599e-02 8.64453554e-01
-2.01244831e-01 6.92675471e-01 1.56767023e+00 -1.47914842e-01
-1.41498625e-01 -4.62446734e-02 -5.94937623e-01 7.26841807e-01
7.35964477e-01 -2.38361835e-01 -2.37676293e-01 -5.23646653e-01
-1.59500912e-01 1.38939202e-01 3.89382958e-01 -9.63411570e-01
3.98222029e-01 -2.95326769e-01 2.60822296e-01 -3.93091619e-01
3.35899681e-01 -6.37977660e-01 2.57714748e-01 3.91481847e-01
1.16729468e-01 -4.83682871e-01 2.10494678e-02 7.13801146e-01
-4.66764092e-01 -3.89863461e-01 1.00699770e+00 -2.38181174e-01
-7.32516289e-01 2.58140981e-01 -3.14522386e-01 -6.18427284e-02
8.35995674e-01 -4.16675806e-01 -4.73700821e-01 -4.81146544e-01
-3.18579793e-01 1.11363508e-01 6.64325833e-01 6.28706589e-02
6.88912094e-01 -9.92803395e-01 -8.81784141e-01 1.13389142e-01
-2.78093755e-01 -5.25387786e-02 6.08181298e-01 1.47694528e+00
-6.57718122e-01 1.33588776e-01 -8.91764238e-02 -4.65488940e-01
-1.05587173e+00 5.13789892e-01 1.99345708e-01 4.26834747e-02
-6.84266090e-01 7.37440646e-01 5.61958492e-01 1.76609322e-01
-6.26414716e-02 -5.75330667e-02 -1.15937524e-01 -3.00173879e-01
7.74693310e-01 6.41010225e-01 -2.06256092e-01 -6.98049009e-01
-1.60340577e-01 9.58962262e-01 -8.37527663e-02 -5.69461733e-02
1.14495313e+00 -7.06165910e-01 -5.35242558e-01 -1.77294284e-01
1.05591905e+00 3.78907353e-01 -1.46283591e+00 -3.69696409e-01
-3.99672657e-01 -1.00380719e+00 5.02611458e-01 -5.20728648e-01
-1.20922983e+00 7.73621798e-01 6.84210598e-01 3.06431297e-02
1.52148211e+00 -3.27419311e-01 1.09170425e+00 -6.18978702e-02
3.11799049e-01 -6.92810357e-01 -3.26271690e-02 3.65883529e-01
8.58148813e-01 -1.22296607e+00 4.31139141e-01 -7.02438295e-01
-2.73352265e-01 9.37428296e-01 3.18855554e-01 -1.06466286e-01
5.64557195e-01 -4.64337654e-02 1.62167251e-02 1.11430094e-01
-1.63734615e-01 -2.19378486e-01 2.66719252e-01 3.98656994e-01
2.51940876e-01 -1.63539872e-01 -6.17042422e-01 2.65740752e-01
2.70349264e-01 1.72826469e-01 6.81286573e-01 7.13581026e-01
-6.49562955e-01 -7.36996949e-01 -8.82712126e-01 1.94212422e-01
-5.27618408e-01 -4.88894790e-01 2.30989084e-01 2.71035045e-01
7.96015337e-02 1.21446157e+00 -3.71692032e-01 -1.18864119e-01
-5.43891378e-02 -4.45505559e-01 3.64922285e-01 -1.68752700e-01
7.06832632e-02 3.03740889e-01 -2.61844218e-01 -4.16586637e-01
-6.45978928e-01 -4.89268094e-01 -8.70108545e-01 -2.13174820e-01
-6.16758823e-01 2.37902269e-01 3.72131050e-01 7.50536263e-01
3.65303010e-01 6.20047711e-02 7.53806472e-01 -1.09285343e+00
-5.59776664e-01 -9.95294750e-01 -7.81988084e-01 1.38059899e-01
5.24346352e-01 -5.60848176e-01 -8.18930864e-01 3.29822183e-01] | [11.49273681640625, -2.679994583129883] |
83bff61a-977f-469c-a521-4d3c047239da | prosocialdialog-a-prosocial-backbone-for | 2205.12688 | null | https://arxiv.org/abs/2205.12688v2 | https://arxiv.org/pdf/2205.12688v2.pdf | ProsocialDialog: A Prosocial Backbone for Conversational Agents | Most existing dialogue systems fail to respond properly to potentially unsafe user utterances by either ignoring or passively agreeing with them. To address this issue, we introduce ProsocialDialog, the first large-scale multi-turn dialogue dataset to teach conversational agents to respond to problematic content following social norms. Covering diverse unethical, problematic, biased, and toxic situations, ProsocialDialog contains responses that encourage prosocial behavior, grounded in commonsense social rules (i.e., rules-of-thumb, RoTs). Created via a human-AI collaborative framework, ProsocialDialog consists of 58K dialogues, with 331K utterances, 160K unique RoTs, and 497K dialogue safety labels accompanied by free-form rationales. With this dataset, we introduce a dialogue safety detection module, Canary, capable of generating RoTs given conversational context, and a socially-informed dialogue agent, Prost. Empirical results show that Prost generates more socially acceptable dialogues compared to other state-of-the-art language and dialogue models in both in-domain and out-of-domain settings. Additionally, Canary effectively guides conversational agents and off-the-shelf language models to generate significantly more prosocial responses. Our work highlights the promise and importance of creating and steering conversational AI to be socially responsible. | ['Maarten Sap', 'Yejin Choi', 'Gunhee Kim', 'Daniel Khashabi', 'Ximing Lu', 'Liwei Jiang', 'Youngjae Yu', 'Hyunwoo Kim'] | 2022-05-25 | null | null | null | null | ['dialogue-safety-prediction', 'rules-of-thumb-generation'] | ['natural-language-processing', 'natural-language-processing'] | [-3.53471376e-02 9.81118262e-01 6.79255798e-02 -5.31233490e-01
-4.69569117e-01 -8.18877161e-01 9.92098808e-01 -1.62578002e-01
-1.27900466e-01 1.13669443e+00 1.05581510e+00 -2.54618049e-01
1.46555752e-01 -5.20122051e-01 7.14823082e-02 -2.86809236e-01
3.86672944e-01 8.86129498e-01 -3.09428990e-01 -8.44178677e-01
2.29423434e-01 -6.48732409e-02 -9.80916023e-01 5.00673592e-01
1.08691323e+00 1.77397519e-01 -5.73896766e-01 8.37907076e-01
1.12531625e-01 1.43628168e+00 -1.00546134e+00 -9.18478787e-01
-8.04441273e-02 -8.53851080e-01 -1.40637517e+00 -4.27202582e-02
-1.24652367e-02 -8.95918667e-01 -1.68050319e-01 5.72375000e-01
7.71788001e-01 4.57740724e-01 9.34026718e-01 -1.47914791e+00
-8.12352836e-01 1.10779119e+00 3.84143107e-02 -3.22162449e-01
1.01277089e+00 7.85592675e-01 1.02999496e+00 -2.43967310e-01
8.59588325e-01 1.88003492e+00 5.18191397e-01 1.36267865e+00
-1.32851923e+00 -6.77479625e-01 -2.20333189e-01 -2.61314780e-01
-4.50362653e-01 -8.86201203e-01 6.72743797e-01 -5.82775474e-01
1.11443365e+00 5.21190047e-01 7.32970417e-01 1.90733981e+00
-1.13243245e-01 8.36172581e-01 1.18849874e+00 -3.47846784e-02
2.76795715e-01 4.52092648e-01 1.54057115e-01 3.80711943e-01
-4.50936794e-01 8.32471997e-02 -5.79095304e-01 -8.95306289e-01
8.10671300e-02 -7.53457844e-01 -5.14693111e-02 2.32210219e-01
-9.09487307e-01 1.31010938e+00 -1.15821123e-01 2.38230545e-02
-3.76605034e-01 -2.69189149e-01 8.84894311e-01 4.37378526e-01
7.74812877e-01 1.03523433e+00 -3.31532568e-01 -8.91611159e-01
-4.28512283e-02 8.00432444e-01 1.52401662e+00 8.45890403e-01
3.60689580e-01 4.70818952e-02 -5.31522393e-01 1.42492390e+00
3.29666018e-01 4.90265220e-01 1.49889469e-01 -1.66477191e+00
2.69129664e-01 5.36280811e-01 3.45349252e-01 -9.13776755e-01
-6.36741340e-01 3.66077453e-01 -4.63974655e-01 -3.17264259e-01
5.76258481e-01 -8.61523509e-01 8.32799897e-02 1.93514037e+00
4.91786569e-01 -5.06626070e-01 6.23449683e-01 6.99331403e-01
1.07614422e+00 8.36479306e-01 1.85163394e-01 -4.28923190e-01
1.07018781e+00 -7.37028360e-01 -7.49779105e-01 -1.39994577e-01
1.10982156e+00 -6.06188834e-01 1.19714975e+00 3.15908730e-01
-1.04604709e+00 4.43308353e-01 -1.63453773e-01 -1.20524943e-01
2.38543041e-02 -5.00609934e-01 5.48032880e-01 7.72125483e-01
-1.05040789e+00 1.65253356e-01 1.15447767e-01 -6.05432928e-01
1.56336930e-02 -5.75154312e-02 -1.63275927e-01 3.51963043e-01
-1.71612680e+00 1.22699094e+00 9.39816644e-04 -3.05095702e-01
-8.31376553e-01 -5.49814820e-01 -9.58253920e-01 -2.53700376e-01
4.71857041e-01 -5.29627800e-01 1.75407851e+00 -8.44533145e-01
-2.26409316e+00 9.93929207e-01 4.49678749e-01 -3.77573758e-01
8.02528322e-01 -2.30344206e-01 -1.54032946e-01 5.12525178e-02
2.70796776e-01 9.10290062e-01 4.71450359e-01 -1.18652487e+00
-1.85814679e-01 5.14210127e-02 4.09403414e-01 5.46073377e-01
-2.85118431e-01 5.74873745e-01 4.06421274e-01 -3.10506463e-01
-9.75022793e-01 -1.11244369e+00 -1.44752890e-01 -4.39528435e-01
-9.88431156e-01 -8.59885752e-01 5.60071766e-01 -5.26976407e-01
1.00186467e+00 -1.82223952e+00 1.79230422e-01 -4.11232971e-02
2.59628743e-01 4.68156338e-01 -2.34810174e-01 7.84020364e-01
2.68647969e-01 2.59163976e-01 -1.81572605e-02 -3.40157509e-01
4.01894212e-01 7.26611465e-02 -2.71285534e-01 2.00290948e-01
4.41086479e-03 5.92822313e-01 -1.38462961e+00 -6.12096608e-01
3.01237732e-01 1.12933040e-01 -1.02004707e+00 7.37461030e-01
-4.81019676e-01 6.52309239e-01 -5.42343080e-01 1.23362780e-01
1.86725602e-01 2.29452372e-01 2.63646603e-01 5.10614157e-01
2.40076147e-02 8.90317798e-01 -3.36875916e-01 1.19361699e+00
-5.19687951e-01 5.29717624e-01 3.34811091e-01 -2.01016709e-01
1.08075833e+00 4.68840897e-01 3.24522376e-01 -4.48277086e-01
3.65895808e-01 7.98898339e-02 1.67632118e-01 -6.89081609e-01
7.29023576e-01 -1.87904209e-01 -5.83141685e-01 1.29464400e+00
-1.86320975e-01 -6.35945976e-01 9.47849229e-02 7.44946837e-01
1.21450734e+00 -2.82833844e-01 4.05930072e-01 -1.83084652e-01
5.86171508e-01 3.59484479e-02 5.56657016e-01 7.51888454e-01
-6.92931473e-01 -4.08837385e-02 1.08588982e+00 -2.86245346e-01
-8.33832920e-01 -5.81032157e-01 2.31198907e-01 1.61418533e+00
-3.29739571e-01 -3.67964923e-01 -1.14105403e+00 -8.79625380e-01
-4.76528145e-03 1.36925197e+00 -3.59633029e-01 -2.30951920e-01
-4.01750386e-01 -3.29058945e-01 1.24031782e+00 -3.01215559e-01
5.63083053e-01 -1.56060147e+00 -2.38238513e-01 4.50042069e-01
-6.74892068e-01 -1.05423546e+00 -6.75708354e-01 -5.04700065e-01
7.21023977e-02 -1.24069107e+00 -2.42225364e-01 -2.52039641e-01
2.80391484e-01 7.97927827e-02 1.14012122e+00 6.30280003e-02
1.25730485e-01 6.02245390e-01 -5.69682479e-01 -1.76550686e-01
-1.50644052e+00 -1.26380116e-01 2.71152049e-01 -1.55218512e-01
4.09918070e-01 -4.15107965e-01 -3.65170896e-01 5.75697720e-01
-4.67353553e-01 3.63591135e-01 -2.93985516e-01 8.53168607e-01
-8.69265616e-01 -8.20083678e-01 1.13684797e+00 -1.19947135e+00
1.74103212e+00 -7.92697847e-01 5.80937639e-02 9.77020562e-02
-3.48609239e-01 -2.92526394e-01 1.01867771e+00 -1.85721114e-01
-1.49207604e+00 -4.16203529e-01 -1.51804328e-01 3.55416179e-01
-3.67388725e-01 3.87866497e-02 7.87256807e-02 4.34748024e-01
1.15737402e+00 6.45985678e-02 4.45491523e-01 -3.54590081e-02
8.06857109e-01 1.44913673e+00 9.02263001e-02 -1.10750902e+00
3.96999329e-01 -2.39822134e-01 -8.45695138e-01 -9.77329433e-01
-6.24454319e-01 -2.30255038e-01 -3.73185892e-03 -8.23927522e-01
6.90156221e-01 -7.26075888e-01 -1.43195689e+00 7.85525322e-01
-1.46803117e+00 -8.81322205e-01 -1.30871847e-01 -2.89122369e-02
-8.62362862e-01 4.52113748e-01 -1.10473609e+00 -1.35372353e+00
-7.03572631e-01 -1.02336228e+00 6.46142066e-01 1.50374519e-02
-1.38401389e+00 -9.19363856e-01 3.38774353e-01 1.07525420e+00
4.46319073e-01 1.26675949e-01 9.18529034e-01 -1.26644945e+00
3.01295936e-01 1.63447842e-01 -5.13671823e-02 3.70197833e-01
2.46522710e-01 2.25872591e-01 -8.35946202e-01 1.11452699e-01
-1.23500094e-01 -1.46129274e+00 -1.04118753e-02 -1.78350627e-01
4.64221776e-01 -1.26862705e+00 2.02502161e-01 -2.18739778e-01
8.09729248e-02 3.23316753e-01 4.61231440e-01 1.79932564e-02
3.15518379e-01 1.37715220e+00 6.74886823e-01 1.02742136e+00
9.79779541e-01 5.68991005e-01 1.05045252e-01 3.26586396e-01
3.24845731e-01 -4.18699622e-01 8.88122737e-01 5.54847956e-01
3.10200632e-01 -4.61070329e-01 -9.82151628e-01 4.67970073e-01
-1.88163912e+00 -1.24162579e+00 -9.63949934e-02 1.46158564e+00
1.74222493e+00 -1.96010262e-01 5.68653941e-01 -3.73280972e-01
7.47806191e-01 4.54819977e-01 -5.47065496e-01 -1.34665298e+00
7.51180351e-02 -5.46536326e-01 -2.17240542e-01 1.04263794e+00
-6.21679485e-01 1.29442382e+00 5.92777872e+00 5.21318436e-01
-5.86059749e-01 -1.50804827e-02 9.76279676e-01 -1.26049981e-01
-7.23259091e-01 -1.96374714e-01 -4.26078796e-01 5.91501355e-01
1.01706064e+00 -5.41230977e-01 7.46810496e-01 8.79226923e-01
7.31180668e-01 -9.52070504e-02 -1.23662603e+00 4.71800417e-01
1.32623672e-01 -1.04671729e+00 -1.61458805e-01 -8.05737004e-02
4.87898439e-01 -3.15627098e-01 -1.34476304e-01 7.17387497e-01
1.41403615e+00 -9.69460309e-01 4.30352032e-01 1.50583550e-01
5.48957765e-01 -5.91770530e-01 5.35077870e-01 7.50449955e-01
-8.28922242e-02 -5.20729199e-02 1.18020624e-01 -2.78985411e-01
1.90013722e-01 -3.22230756e-02 -1.46134615e+00 -1.58331200e-01
3.07642490e-01 8.93175781e-01 5.92219606e-02 5.04943058e-02
-2.73468524e-01 6.42469704e-01 -3.27671409e-01 -6.41579926e-01
4.02874559e-01 -3.67664099e-01 8.36815834e-01 1.37400365e+00
-3.91290516e-01 6.40463412e-01 1.34110078e-01 9.99255300e-01
-2.49419674e-01 1.47540167e-01 -1.01231766e+00 -1.91719085e-01
8.25057685e-01 1.08671772e+00 1.41265392e-01 -2.49542788e-01
2.25173622e-01 6.77771330e-01 2.85928518e-01 2.09849447e-01
-5.17240465e-01 7.27695078e-02 9.07179594e-01 -2.37087697e-01
-8.56352985e-01 3.30316812e-01 -1.99095219e-01 -8.74729633e-01
-5.46433628e-01 -1.67822194e+00 1.67289570e-01 -7.22601056e-01
-1.68140841e+00 4.24747825e-01 6.61609769e-02 -6.77257419e-01
-9.97146726e-01 -8.91489461e-02 -8.10251355e-01 6.25212610e-01
-8.25795591e-01 -9.82866347e-01 1.01675078e-01 4.83886510e-01
7.06135094e-01 -4.24539506e-01 1.09404588e+00 -2.27829710e-01
-7.19566941e-01 5.94684243e-01 -2.45916560e-01 2.26442620e-01
1.19229352e+00 -1.08394766e+00 3.37599993e-01 4.75932285e-02
-9.05485511e-01 6.93395019e-01 1.12030780e+00 -7.92445898e-01
-1.03418648e+00 -8.46955001e-01 1.01531291e+00 -6.44334555e-01
9.22183573e-01 -4.62159753e-01 -7.12264478e-01 5.39879322e-01
6.25315130e-01 -8.61390948e-01 1.20292711e+00 2.55685180e-01
-3.96923959e-01 5.03184378e-01 -1.67112088e+00 1.04595029e+00
1.07746053e+00 -6.23288929e-01 -8.13261211e-01 8.94987106e-01
8.30050588e-01 -3.77896279e-01 -7.26727188e-01 -3.29369247e-01
3.36169302e-01 -1.10507095e+00 4.95896101e-01 -9.73594904e-01
8.36288631e-01 6.10780835e-01 1.62804320e-01 -1.78363347e+00
1.11925870e-01 -1.59994042e+00 2.27291852e-01 1.48469305e+00
3.88971567e-01 -7.84419060e-01 5.85180938e-01 1.55475819e+00
-1.95116058e-01 -4.37476814e-01 -9.50487792e-01 -2.85204679e-01
3.79166007e-01 -2.22505093e-01 5.50098717e-01 1.31729102e+00
9.60581005e-01 8.07284355e-01 -8.85514617e-01 -5.55340171e-01
4.45473045e-01 -4.16109413e-01 1.27961802e+00 -9.99170303e-01
-3.11858617e-02 -4.75982875e-01 3.56672376e-01 -9.44538057e-01
8.97023916e-01 -5.53429425e-01 2.90379047e-01 -1.32701790e+00
7.26302713e-02 -6.41390502e-01 8.01808536e-01 7.14804947e-01
-1.74605474e-01 -4.40465450e-01 2.57990748e-01 -4.40586247e-02
-6.45161331e-01 1.02774251e+00 1.16751742e+00 -1.02407008e-01
-6.26726985e-01 -5.02632037e-02 -1.00307465e+00 8.10508549e-01
8.74859095e-01 -3.57843250e-01 -4.99223799e-01 8.73079002e-02
1.89329758e-01 5.02071381e-01 1.13230653e-01 -9.83444154e-02
9.43209678e-02 -7.52820134e-01 -6.26339614e-01 3.77802141e-02
4.84385788e-01 -3.24979663e-01 -2.63582051e-01 2.84203053e-01
-1.22385025e+00 -3.87618840e-01 -1.33252991e-02 3.46820503e-01
2.67515123e-01 -9.30799395e-02 8.11749995e-01 -1.28557727e-01
-6.10547466e-03 -2.55470276e-01 -1.07347214e+00 6.37826979e-01
1.18071675e+00 1.05550714e-01 -1.04150689e+00 -1.29381490e+00
-3.36128891e-01 8.19221079e-01 4.66403008e-01 6.50280714e-01
6.11552596e-01 -9.46532965e-01 -1.15130150e+00 -3.39506358e-01
2.29609445e-01 -3.19492072e-01 3.82547021e-01 2.34737679e-01
-3.92005742e-01 2.46826068e-01 -1.72898605e-01 -1.45214781e-01
-1.29531860e+00 -8.59821737e-02 5.38280487e-01 -1.11563079e-01
-4.98347133e-01 8.79812181e-01 5.68095446e-02 -1.17829168e+00
2.51501650e-01 1.76539153e-01 -4.66172159e-01 2.77196318e-01
5.20362496e-01 4.56055522e-01 -5.90379715e-01 -7.50221550e-01
-1.39246047e-01 -4.29645598e-01 -3.02017719e-01 -4.05139953e-01
1.06424677e+00 -7.87931159e-02 -3.90497833e-01 2.74918020e-01
8.91373158e-01 2.92857736e-02 -1.02536619e+00 -2.69931197e-01
-2.06036255e-01 -4.73945618e-01 -5.73455334e-01 -1.24911392e+00
-3.21830690e-01 3.25338840e-01 -3.44476491e-01 6.62993670e-01
3.31079781e-01 -1.91750973e-01 9.12630856e-01 7.14559078e-01
3.78347397e-01 -1.52505779e+00 3.93128067e-01 1.01305747e+00
1.44054341e+00 -1.26159334e+00 -2.82698601e-01 -3.51865619e-01
-1.59116423e+00 7.45714962e-01 1.15454721e+00 3.49607050e-01
1.67985469e-01 -1.67051956e-01 4.13708746e-01 -1.42204121e-01
-1.35284126e+00 4.51738745e-01 -4.45673615e-01 7.42787480e-01
5.22055149e-01 3.63034278e-01 -4.13384676e-01 5.19642711e-01
-6.92972302e-01 -3.38707685e-01 1.16390324e+00 5.16750991e-01
-2.26066068e-01 -9.18777466e-01 -2.28669316e-01 3.51352692e-01
-1.17111593e-01 -8.86848420e-02 -1.58130932e+00 3.96693051e-01
-5.83153129e-01 1.79041207e+00 -3.71705383e-01 -5.61962426e-01
4.16601032e-01 7.70906582e-02 -4.24990177e-01 -9.01594281e-01
-1.66484094e+00 -4.44045573e-01 1.30212855e+00 -4.27747995e-01
-2.15020180e-01 -7.16316223e-01 -1.09425616e+00 -1.02104723e+00
-1.24989800e-01 3.88254046e-01 1.57649085e-01 8.80206168e-01
4.71464306e-01 -1.36853129e-01 1.11288440e+00 -4.92791653e-01
-9.32680011e-01 -1.32345998e+00 -1.46988988e-01 7.57693529e-01
2.63561904e-01 -2.99792945e-01 -6.16126776e-01 -5.14298379e-01] | [12.834390640258789, 8.0730562210083] |
0f502e66-80c0-414e-8e15-52b88b9f0cfd | crowdsourcing-ground-truth-for-medical | 1701.02185 | null | http://arxiv.org/abs/1701.02185v2 | http://arxiv.org/pdf/1701.02185v2.pdf | Crowdsourcing Ground Truth for Medical Relation Extraction | Cognitive computing systems require human labeled data for evaluation, and
often for training. The standard practice used in gathering this data minimizes
disagreement between annotators, and we have found this results in data that
fails to account for the ambiguity inherent in language. We have proposed the
CrowdTruth method for collecting ground truth through crowdsourcing, that
reconsiders the role of people in machine learning based on the observation
that disagreement between annotators provides a useful signal for phenomena
such as ambiguity in the text. We report on using this method to build an
annotated data set for medical relation extraction for the $cause$ and $treat$
relations, and how this data performed in a supervised training experiment. We
demonstrate that by modeling ambiguity, labeled data gathered from crowd
workers can (1) reach the level of quality of domain experts for this task
while reducing the cost, and (2) provide better training data at scale than
distant supervision. We further propose and validate new weighted measures for
precision, recall, and F-measure, that account for ambiguity in both human and
machine performance on this task. | ['Lora Aroyo', 'Anca Dumitrache', 'Chris Welty'] | 2017-01-09 | null | null | null | null | ['medical-relation-extraction'] | ['medical'] | [ 2.25764643e-02 8.91888916e-01 2.49850467e-01 -8.81277561e-01
-1.11328804e+00 -6.92999542e-01 4.66178477e-01 7.97760248e-01
-9.14123654e-01 1.14257479e+00 4.53411132e-01 -2.54483193e-01
-9.78836939e-02 -4.93692905e-01 -5.09616375e-01 -2.86440134e-01
2.68400967e-01 1.16451621e+00 2.59461850e-01 -3.19694161e-01
-1.13417432e-02 6.52189031e-02 -9.38596785e-01 5.00828207e-01
8.34391534e-01 5.84502757e-01 6.44284242e-04 3.88993889e-01
7.67694414e-02 1.28823090e+00 -9.15634453e-01 -1.03593123e+00
2.57307559e-01 -2.60283887e-01 -1.29764950e+00 1.87432781e-01
1.90449879e-01 -5.71508221e-02 1.72009051e-01 8.51306200e-01
5.25902331e-01 -3.57108898e-02 6.68995202e-01 -1.13199186e+00
-6.01075232e-01 7.84467340e-01 -2.28118986e-01 1.16796985e-01
6.46897733e-01 -2.66750362e-02 1.12430406e+00 -7.18082666e-01
8.68013322e-01 1.19878769e+00 8.06376874e-01 7.37539172e-01
-1.03124774e+00 -3.19881797e-01 -1.53017789e-01 -2.62908250e-01
-1.27445996e+00 -4.67772514e-01 2.88296640e-01 -8.40215921e-01
7.78056681e-01 1.98013693e-01 2.35667929e-01 9.59705412e-01
-1.35112464e-01 4.19323176e-01 1.15721643e+00 -8.65491629e-01
3.64141971e-01 5.31305373e-01 4.40315306e-01 9.86316741e-01
5.86309850e-01 -2.67046630e-01 -7.28020251e-01 -5.68719745e-01
2.90134281e-01 -3.35775495e-01 -1.65912420e-01 7.90752098e-02
-1.09786332e+00 7.35642254e-01 4.17084903e-01 4.29826736e-01
-2.67483294e-01 -2.75428981e-01 3.97488922e-01 1.28909484e-01
7.53794134e-01 8.42477739e-01 -6.01635098e-01 9.42875743e-02
-5.56541800e-01 4.52757597e-01 1.14351976e+00 8.45141113e-01
6.39758646e-01 -7.47241557e-01 -4.14845973e-01 7.69438446e-01
2.08060727e-01 4.70536917e-01 3.93050224e-01 -1.26921380e+00
7.19223499e-01 1.06218708e+00 7.72496104e-01 -9.23948765e-01
-6.03053331e-01 8.32745209e-02 -3.57221156e-01 1.08774871e-01
8.73912930e-01 -3.84377241e-01 -4.86589551e-01 1.62656271e+00
3.94672453e-01 -6.57545090e-01 8.53003114e-02 9.64516878e-01
8.03685248e-01 -2.90429562e-01 3.64302397e-01 -2.17853338e-01
1.74428761e+00 -7.10323572e-01 -8.65547121e-01 -2.92285562e-01
1.12117624e+00 -7.04863608e-01 1.02463198e+00 2.15239882e-01
-1.06789041e+00 -3.68011057e-01 -8.20176840e-01 -3.61738324e-01
-2.17406407e-01 6.33628964e-02 4.33970064e-01 7.69723117e-01
-1.04412842e+00 6.50281012e-01 -5.66475630e-01 -4.98870522e-01
2.81163841e-01 3.75103772e-01 -5.64723969e-01 -1.71713829e-01
-1.30761850e+00 1.28417766e+00 1.15363561e-01 -2.23958455e-02
-3.14779073e-01 -3.19189541e-02 -8.43589246e-01 -1.41739592e-01
5.89810133e-01 -5.23535609e-01 1.48357928e+00 -8.85978281e-01
-6.76451802e-01 1.58188486e+00 -4.31672275e-01 -3.85032862e-01
6.89229727e-01 -1.11897737e-01 -1.55159742e-01 4.87285294e-02
7.53519535e-01 4.66531754e-01 1.33002847e-01 -1.51880598e+00
-6.83450818e-01 -7.71505713e-01 6.74405843e-02 1.73940629e-01
-8.07016194e-02 3.07674468e-01 -8.35220143e-02 -2.83978432e-01
-1.36069732e-03 -9.87226129e-01 -3.11655670e-01 -2.40648493e-01
-4.29787129e-01 -6.60882115e-01 9.09416676e-02 -8.88942659e-01
9.33403015e-01 -1.85129702e+00 -2.99719125e-01 4.00300860e-01
7.67393649e-01 2.54290760e-01 1.99677750e-01 3.15904856e-01
3.43990475e-01 5.38547993e-01 -3.13373357e-01 -6.48916245e-01
-1.20011434e-01 4.91570860e-01 2.04139506e-03 3.20613295e-01
2.49289617e-01 8.23107839e-01 -1.11342454e+00 -9.56004441e-01
-4.52739716e-01 -1.34840906e-01 -1.81240439e-01 2.32880220e-01
-1.67701513e-01 2.92280018e-01 -4.24123943e-01 4.95364755e-01
1.22668810e-01 -4.57567781e-01 4.56810832e-01 2.39753976e-01
1.72052026e-01 3.04032296e-01 -8.32662225e-01 1.34635985e+00
-8.86889733e-03 5.24409115e-01 4.56224442e-01 -6.65256917e-01
8.03290188e-01 5.78483701e-01 3.20673823e-01 -5.97170770e-01
1.82046801e-01 3.68015021e-01 5.82806095e-02 -9.03237641e-01
4.10434842e-01 -2.03585625e-01 -2.07181051e-01 7.11120069e-01
-1.54116735e-01 -1.79005966e-01 2.01704800e-01 3.16813618e-01
1.29053104e+00 -3.66516888e-01 3.18623930e-01 -4.51117873e-01
2.81329870e-01 4.59749430e-01 5.72717905e-01 9.04294908e-01
-5.16233921e-01 4.08526629e-01 5.75952768e-01 -7.05491006e-01
-1.01492894e+00 -6.27383411e-01 -3.48868361e-03 1.29069519e+00
-1.02566369e-02 -1.91110894e-01 -1.13373232e+00 -1.05606937e+00
-1.87218219e-01 6.11112356e-01 -7.35943675e-01 2.24334165e-01
-4.35132474e-01 -7.21927166e-01 6.71482146e-01 4.05837268e-01
3.03597987e-01 -1.09938717e+00 -5.81814528e-01 1.69448003e-01
-6.58093631e-01 -1.42843735e+00 -2.86849916e-01 1.10645577e-01
-4.98522997e-01 -1.51068175e+00 -3.57121289e-01 -6.96812510e-01
9.12257373e-01 3.97644080e-02 1.53531134e+00 5.32658994e-01
-1.62461177e-01 4.58186030e-01 -5.41598141e-01 -9.28863943e-01
-6.49853468e-01 7.16762394e-02 8.08457062e-02 -3.46307546e-01
9.21004534e-01 -7.12331012e-02 -2.76031733e-01 6.90908194e-01
-7.71640360e-01 -3.45921159e-01 2.91828036e-01 8.40667844e-01
8.35569426e-02 -2.37712577e-01 6.54818714e-01 -1.39482963e+00
1.17066848e+00 -2.57479131e-01 -1.93628222e-01 5.77513933e-01
-7.79622674e-01 1.50663793e-01 1.41091458e-02 3.15093920e-02
-9.47914422e-01 1.11299912e-02 2.01771706e-01 3.34310025e-01
-1.46036565e-01 3.95421535e-01 8.60439613e-02 1.37642249e-01
1.48567486e+00 -7.09387064e-01 1.84776470e-01 -2.90517598e-01
-1.73443239e-02 9.87210989e-01 2.66614884e-01 -8.89534831e-01
4.04198617e-01 5.13678133e-01 -3.06154609e-01 -4.56620902e-01
-1.48219061e+00 -6.89574182e-01 -7.20725358e-01 -3.60574722e-02
1.03348911e+00 -1.03455174e+00 -7.34788537e-01 -2.41255119e-01
-1.40930033e+00 -2.96563834e-01 -4.32531565e-01 4.48099107e-01
-1.61718160e-01 1.56475201e-01 -5.99628866e-01 -1.05145025e+00
-2.33860016e-01 -9.00695384e-01 1.26084054e+00 -5.17019071e-02
-7.99603283e-01 -1.01454329e+00 1.67184860e-01 1.10735905e+00
1.46366939e-01 1.59976974e-01 7.53399432e-01 -1.32674038e+00
-1.42414451e-01 -4.40289527e-01 -2.77002186e-01 1.38404369e-01
4.76213992e-02 -3.85143995e-01 -1.13235724e+00 6.70579821e-02
1.74078286e-01 -8.07672560e-01 4.14165497e-01 -1.14583053e-01
7.00937510e-01 -2.92978227e-01 -4.51097071e-01 -4.59574968e-01
8.83443177e-01 -6.38376251e-02 3.52726132e-01 2.68564671e-01
5.28496325e-01 1.31684875e+00 4.76928920e-01 2.10852087e-01
6.05349004e-01 5.75482130e-01 -1.54808253e-01 -2.07608044e-01
3.76421332e-01 7.02795479e-03 -1.81999490e-01 4.86657560e-01
-5.36845922e-01 -1.85969844e-01 -1.39494300e+00 6.68258548e-01
-2.19673014e+00 -7.27522612e-01 -1.14835404e-01 1.91648436e+00
1.12076426e+00 1.73306957e-01 1.72000214e-01 1.62919596e-01
7.57713258e-01 -4.82712328e-01 1.72232538e-02 -1.90222651e-01
9.54558142e-03 9.94859785e-02 5.40959656e-01 7.77232707e-01
-7.75395572e-01 7.54914045e-01 6.42250299e+00 3.67818415e-01
-2.66722053e-01 4.14298087e-01 8.18466187e-01 1.47845194e-01
-1.73558697e-01 -2.11085215e-01 -6.87524557e-01 1.48030475e-01
8.21677625e-01 7.42981397e-03 3.56502026e-01 7.62878478e-01
3.13339561e-01 -3.12185347e-01 -1.42587328e+00 5.97412109e-01
2.18686715e-01 -9.35399234e-01 -2.69522220e-01 4.84008789e-02
7.08067060e-01 -1.18869126e-01 -6.80058300e-01 3.53047132e-01
7.75159955e-01 -1.16343451e+00 6.34984910e-01 3.18756878e-01
4.43924576e-01 -3.00770402e-01 1.39082932e+00 7.16982126e-01
-4.16815072e-01 -6.57669362e-03 -2.34825164e-01 -3.02782327e-01
4.45237523e-03 8.79362643e-01 -1.44847059e+00 2.36813292e-01
5.91378987e-01 -9.31392536e-02 -7.54154205e-01 5.07433236e-01
-3.31948102e-01 3.14980447e-01 -1.11583710e-01 -3.01044971e-01
-2.83227228e-02 2.60924071e-01 2.32293054e-01 1.16287422e+00
-1.98981851e-01 3.61306041e-01 4.35417950e-01 6.21417344e-01
-4.08117622e-01 1.85685188e-01 -7.98847854e-01 1.15240999e-01
4.95553553e-01 9.59738851e-01 -7.94782281e-01 -3.62912983e-01
-1.78665102e-01 6.65500998e-01 7.50098705e-01 2.46476337e-01
-3.06181520e-01 -1.06396787e-01 4.91378307e-02 3.21158588e-01
-3.54560345e-01 -9.15400088e-02 -6.07166767e-01 -9.45676029e-01
5.43592691e-01 -1.00064361e+00 3.95503461e-01 -7.77086496e-01
-1.61025536e+00 7.85818577e-01 -1.29942611e-01 -6.82986796e-01
-4.28952545e-01 -8.95614147e-01 -7.13254735e-02 1.11022925e+00
-9.90646243e-01 -9.71190989e-01 -4.33949590e-01 4.23873782e-01
2.35009745e-01 -6.26385584e-02 1.01133025e+00 6.61344454e-02
-2.09972247e-01 4.02686626e-01 -4.88989145e-01 5.99237382e-01
1.00081468e+00 -1.40360498e+00 2.11101875e-01 4.06945825e-01
2.54564613e-01 8.62502277e-01 6.83199644e-01 -9.09815669e-01
-4.45939332e-01 -8.28894734e-01 1.60142541e+00 -1.37565064e+00
3.61706406e-01 -3.28251302e-01 -7.67502487e-01 6.54765248e-01
2.53766943e-02 -1.48909956e-01 1.01333582e+00 6.61270201e-01
-4.16506171e-01 2.27507025e-01 -1.53464019e+00 2.26569727e-01
1.11119163e+00 -8.17809761e-01 -9.83601868e-01 8.81038547e-01
7.04223990e-01 -3.83826822e-01 -4.87646103e-01 1.90063536e-01
2.73209393e-01 -9.28160369e-01 4.96517241e-01 -9.56385791e-01
4.94907916e-01 -2.16480359e-01 -2.36255050e-01 -1.16507030e+00
-1.05476484e-01 -3.68375212e-01 5.35618842e-01 1.08551300e+00
1.02412438e+00 -3.31304669e-01 7.03356147e-01 1.72320127e+00
2.28676498e-01 -4.72072333e-01 -9.26868200e-01 -5.36761105e-01
-9.37691107e-02 -3.41338992e-01 2.21243829e-01 1.16066611e+00
3.08404118e-01 6.25231683e-01 4.05709073e-03 1.65680662e-01
4.55200702e-01 -4.76533651e-01 6.17709875e-01 -1.59450305e+00
-1.87218159e-01 2.11169705e-01 -2.33163700e-01 -2.36294717e-01
2.40008339e-01 -5.61440051e-01 3.64892155e-01 -1.70046258e+00
2.73116767e-01 -5.76154411e-01 3.50929886e-01 5.32924712e-01
-4.99505818e-01 1.50595829e-01 4.90054339e-02 5.42355001e-01
-8.97199512e-01 -1.81476280e-01 1.09103870e+00 5.53921312e-02
-7.94454664e-02 -1.32261768e-01 -1.11137295e+00 9.46715236e-01
5.28894424e-01 -7.31563032e-01 -2.28200778e-01 -6.92926049e-01
5.75668633e-01 3.30403564e-03 4.14137512e-01 -7.06112146e-01
4.00242120e-01 -5.73801734e-02 3.73284787e-01 3.49168599e-01
1.04975045e-01 -9.60526168e-01 -3.79161328e-01 2.32191876e-01
-7.92153656e-01 2.07621038e-01 -2.40561143e-01 6.54250205e-01
-1.46372437e-01 -6.06084347e-01 2.85412490e-01 -5.89781702e-01
-3.52893695e-02 -3.22109371e-01 -1.87066257e-01 4.85615492e-01
7.96164334e-01 2.06339806e-01 -6.91171706e-01 -2.91893572e-01
-1.04081929e+00 4.21688080e-01 4.95358437e-01 -5.21030501e-02
-3.45727801e-03 -8.67043674e-01 -1.01172030e+00 -2.96039224e-01
2.15908930e-01 2.34104544e-01 -3.11751038e-01 5.46616673e-01
-5.38481891e-01 2.41499692e-01 2.01571465e-01 -4.59504843e-01
-1.32647622e+00 3.81729394e-01 3.14399391e-01 -5.78880966e-01
1.40584842e-03 8.34935367e-01 -2.71047533e-01 -4.56742615e-01
3.02047014e-01 -1.76873431e-01 -2.36940458e-01 7.97021091e-02
5.94002247e-01 2.40967214e-01 3.20752949e-01 -4.77314770e-01
-3.46397609e-01 2.09124926e-02 -3.83164398e-02 -6.33059740e-01
1.02646565e+00 -1.30563062e-02 -2.93625236e-01 5.26735425e-01
6.10600948e-01 3.82696867e-01 -5.17918289e-01 -3.73267502e-01
5.09525418e-01 -3.42109740e-01 -4.99093503e-01 -1.02753198e+00
-2.05574453e-01 6.21459246e-01 2.42511615e-01 8.48001063e-01
4.29441035e-01 3.61291140e-01 2.70360142e-01 9.22106624e-01
7.09919870e-01 -1.39656544e+00 3.64224017e-02 4.25575197e-01
7.91745603e-01 -1.85361922e+00 6.57553300e-02 -7.21810400e-01
-9.90279078e-01 7.57934451e-01 6.57876909e-01 2.52281934e-01
3.13503861e-01 2.76541859e-01 4.87928450e-01 -7.19984531e-01
-5.57427645e-01 -2.06791982e-01 1.29904687e-01 6.45692170e-01
8.24868441e-01 2.97631711e-01 -4.69741255e-01 8.05801690e-01
-2.50754535e-01 -1.58196909e-03 3.99699658e-01 1.00332177e+00
-4.30047631e-01 -1.08268726e+00 -5.59891164e-01 4.67984229e-01
-6.90010190e-01 2.62758974e-03 -1.08119822e+00 6.14644885e-01
5.36627948e-01 1.38209701e+00 -1.16709575e-01 -1.90451488e-01
5.54712832e-01 3.40204239e-01 3.42141688e-01 -9.99132216e-01
-8.83365273e-01 -5.46541095e-01 8.57551157e-01 -2.31916025e-01
-8.07049513e-01 -4.10643548e-01 -1.13472676e+00 9.54908431e-02
-3.66839916e-01 7.92898417e-01 4.49418396e-01 1.44533741e+00
1.77078173e-01 1.58421978e-01 2.64867514e-01 -3.68397117e-01
-5.40579498e-01 -1.20387471e+00 -4.14251894e-01 9.95077014e-01
9.67132971e-02 -4.94010568e-01 -5.64127445e-01 3.74881238e-01] | [9.686211585998535, 4.804385662078857] |
8845d842-f80f-4ebb-b9a9-831eaa44c5f5 | stem-unsupervised-structural-embedding-for | 2112.00712 | null | https://arxiv.org/abs/2112.00712v2 | https://arxiv.org/pdf/2112.00712v2.pdf | STEM: Unsupervised STructural EMbedding for Stance Detection | Stance detection is an important task, supporting many downstream tasks such as discourse parsing and modeling the propagation of fake news, rumors, and science denial. In this paper, we propose a novel framework for stance detection. Our framework is unsupervised and domain-independent. Given a claim and a multi-participant discussion - we construct the interaction network from which we derive topological embedding for each speaker. These speaker embedding enjoy the following property: speakers with the same stance tend to be represented by similar vectors, while antipodal vectors represent speakers with opposing stances. These embedding are then used to divide the speakers into stance-partitions. We evaluate our method on three different datasets from different platforms. Our method outperforms or is comparable with supervised models while providing confidence levels for its output. Furthermore, we demonstrate how the structural embedding relate to the valence expressed by the speakers. Finally, we discuss some limitations inherent to the framework. | ['Oren Tsur', 'Dan Vilenchik', 'Vladyslav Kozhukhov', 'Ron Korenblum Pick'] | 2021-12-01 | null | null | null | null | ['discourse-parsing'] | ['natural-language-processing'] | [ 4.28983122e-02 6.42833054e-01 -6.95006847e-01 -4.72286612e-01
-3.66852313e-01 -7.73616791e-01 1.11642838e+00 7.22245395e-01
-1.10195920e-01 5.24065673e-01 1.14049542e+00 -4.62198704e-01
1.74202070e-01 -8.68146956e-01 -4.63274181e-01 -4.02754009e-01
1.34672225e-01 4.26533312e-01 1.49848446e-01 -4.77724195e-01
6.38655901e-01 -4.38352786e-02 -1.12165034e+00 6.29772425e-01
6.81297243e-01 5.01543880e-01 -5.85205257e-01 3.49015057e-01
-4.04508173e-01 1.11086690e+00 -8.26470733e-01 -6.41202867e-01
-3.09884369e-01 -5.67816556e-01 -9.98148739e-01 -1.22487880e-01
2.21726254e-01 2.53986511e-02 -1.48477912e-01 9.64952052e-01
1.90736622e-01 -1.58950180e-01 1.03385961e+00 -1.20417058e+00
-7.89152384e-01 1.20141399e+00 -6.04734480e-01 2.66210467e-01
4.51819956e-01 -3.60502571e-01 1.39688885e+00 -8.74892354e-01
1.13962293e+00 1.48638439e+00 6.96053684e-01 4.68928307e-01
-1.29784775e+00 -4.26466644e-01 4.02141899e-01 2.96697736e-01
-5.80579281e-01 -3.99703145e-01 1.27918041e+00 -8.20716918e-01
2.53949583e-01 4.38639522e-01 5.78895092e-01 1.50449038e+00
2.19300278e-02 7.56946564e-01 1.29827893e+00 -2.63134837e-01
2.30188787e-01 5.23218334e-01 9.45773721e-01 4.57601368e-01
4.57875460e-01 -3.06291521e-01 -7.63865113e-01 -6.63185120e-01
4.72653210e-02 -3.66431117e-01 -3.21626574e-01 -2.06241384e-01
-1.32283616e+00 1.46164632e+00 5.22766709e-01 4.60503697e-01
-3.50450695e-01 -1.75725073e-01 5.27031720e-01 4.42918777e-01
9.61815059e-01 3.20551425e-01 1.42685935e-01 8.38597417e-02
-8.62678170e-01 4.29388821e-01 1.16121137e+00 2.74668217e-01
3.55649382e-01 -4.11158860e-01 -2.86383599e-01 6.65817857e-01
6.39068902e-01 1.23256944e-01 2.86708832e-01 -7.90911317e-01
5.37651598e-01 7.62706161e-01 7.60900155e-02 -1.67648494e+00
-3.85537684e-01 -2.78419703e-01 -6.72358155e-01 1.68910380e-02
5.14827728e-01 -2.99559366e-02 -4.04682070e-01 1.87789357e+00
5.46242833e-01 -3.68495621e-02 2.00785741e-01 7.88591027e-01
1.03826058e+00 5.12295365e-01 -1.24538586e-01 -1.76137805e-01
1.68401110e+00 -8.25870991e-01 -9.49433386e-01 -2.29077086e-01
4.99140233e-01 -8.14724326e-01 7.22747982e-01 -1.01834215e-01
-9.24013376e-01 8.87290835e-02 -1.03432965e+00 -1.57917827e-01
-2.17824399e-01 -3.67286682e-01 3.43892425e-01 7.12475955e-01
-6.18305266e-01 6.24867976e-01 -5.12402236e-01 -6.56090602e-02
5.05501926e-01 -2.34890401e-01 -1.18333874e-02 4.44427162e-01
-1.44861722e+00 9.15004790e-01 3.26195359e-03 -5.65831698e-02
-4.53913718e-01 -5.58970630e-01 -7.98535943e-01 -2.24159196e-01
4.21034135e-02 -5.77792823e-01 9.56900299e-01 -7.68709242e-01
-1.38433719e+00 1.23773527e+00 -4.28182095e-01 -7.26941049e-01
7.43437231e-01 -5.83830774e-02 -3.42311680e-01 7.68147856e-02
4.33288336e-01 2.20260784e-01 8.43666017e-01 -1.27866590e+00
5.37768891e-03 -4.02608305e-01 1.84928894e-01 1.62709743e-01
-3.76987636e-01 3.59196424e-01 2.21837714e-01 -5.58066726e-01
5.16945779e-01 -8.69433522e-01 8.08913633e-02 -2.25629926e-01
-9.32853222e-01 -4.73269910e-01 9.80102062e-01 -6.51828170e-01
1.27864420e+00 -2.02542114e+00 4.34483558e-01 2.13119835e-01
7.19889343e-01 -3.40023071e-01 3.12352836e-01 5.62949061e-01
-2.95019485e-02 5.73888242e-01 -3.42500299e-01 -2.08028957e-01
1.67449564e-01 -9.99232382e-02 -5.57783425e-01 7.58337021e-01
9.92070585e-02 8.87443542e-01 -8.17739844e-01 -6.11967742e-01
-3.47410798e-01 4.54265565e-01 -5.85739851e-01 -1.40072808e-01
-2.84915924e-01 4.92945641e-01 -3.58290553e-01 4.71520454e-01
5.06686985e-01 -4.17434037e-01 7.21827328e-01 -2.23650023e-01
-1.04276016e-01 1.03896594e+00 -7.85505056e-01 8.92293036e-01
5.89558035e-02 1.14402986e+00 2.52841502e-01 -1.01176059e+00
9.08031940e-01 2.87817031e-01 1.11892596e-01 -2.95407683e-01
3.17597777e-01 1.17158175e-01 2.21939728e-01 -3.80242258e-01
5.23992300e-01 -2.66608864e-01 -3.60587329e-01 1.19006801e+00
-4.61062670e-01 2.78509051e-01 -8.47365260e-02 6.68647707e-01
6.68542802e-01 -4.01718885e-01 2.54787207e-01 -5.29282749e-01
4.74290937e-01 -1.73383467e-02 4.93828565e-01 5.03325224e-01
-3.64145815e-01 1.24865711e-01 1.43700182e+00 -2.39572376e-01
-9.08635736e-01 -9.29643154e-01 -2.35150129e-01 1.00973725e+00
1.94149554e-01 -4.67841595e-01 -6.34904861e-01 -8.26283991e-01
1.56870514e-01 7.69226551e-01 -1.14144063e+00 2.39745513e-01
-7.68819988e-01 -7.35603988e-01 7.19402254e-01 2.58077234e-01
2.72936612e-01 -1.04262197e+00 -4.67184484e-01 5.19644395e-02
-6.97529852e-01 -8.35262239e-01 -1.66565537e-01 -2.34585434e-01
-6.49379492e-01 -1.22578382e+00 -4.70555425e-01 -7.70329595e-01
4.83810842e-01 1.20716222e-01 1.08460081e+00 1.13019891e-01
4.09938067e-01 -1.16009817e-01 -3.25468443e-02 -3.61067563e-01
-7.70323932e-01 3.75405438e-02 1.24860264e-01 8.78646299e-02
2.93955505e-01 -4.31717336e-01 -4.76684391e-01 1.39686748e-01
-6.75016820e-01 -3.85255851e-02 -1.33325040e-01 8.90886962e-01
7.71881416e-02 -4.72376198e-01 6.45543694e-01 -1.36565828e+00
1.28145552e+00 -8.48438621e-01 -2.48752348e-03 -7.75461942e-02
-5.56793511e-01 -6.81216409e-03 5.61816804e-02 -4.24252272e-01
-8.55575740e-01 -7.57934153e-01 6.44644871e-02 2.33289808e-01
1.67528033e-01 6.94000602e-01 8.06720406e-02 4.30877507e-01
8.71924341e-01 -2.95787960e-01 2.54383922e-01 -3.14644843e-01
6.71463728e-01 8.04197252e-01 2.50883512e-02 -3.97859067e-01
7.03249037e-01 8.13185334e-01 -4.32412356e-01 -1.09247458e+00
-1.12927067e+00 -2.89152116e-01 -4.44989741e-01 -1.65824220e-01
7.90853918e-01 -6.70603693e-01 -7.65170574e-01 1.83500886e-01
-1.49605811e+00 2.33811252e-02 -6.18835650e-02 2.48212397e-01
-1.56627268e-01 5.61018646e-01 -8.92924368e-01 -9.50094998e-01
-3.01157713e-01 -7.75418580e-01 4.99650538e-01 -2.56602354e-02
-8.21294367e-01 -1.30580854e+00 3.97272348e-01 7.07512736e-01
2.43328676e-01 5.62284291e-01 1.09868145e+00 -8.46228480e-01
-2.12301575e-02 1.71298176e-01 -1.68303698e-01 -6.95536882e-02
2.30287835e-02 1.08037286e-01 -9.95991826e-01 -4.23732959e-02
1.79111436e-01 -2.07961991e-01 1.09502375e+00 3.10310602e-01
4.87292945e-01 -7.12165594e-01 -4.57958609e-01 3.56830843e-02
8.53309989e-01 -5.54935396e-01 3.27882856e-01 6.61602497e-01
5.64673901e-01 1.27412140e+00 2.58974195e-01 5.27320020e-02
3.92719775e-01 4.10309017e-01 2.68691629e-01 2.05185056e-01
1.52637169e-01 -2.82572240e-01 5.77970088e-01 9.55339491e-01
1.01107284e-01 -2.25605175e-01 -9.50762212e-01 7.36460984e-01
-1.90326512e+00 -1.33274829e+00 -7.46295094e-01 1.75072527e+00
8.42305362e-01 5.39883733e-01 3.37948829e-01 3.55144739e-01
7.77552247e-01 7.28810370e-01 -2.48017311e-01 -6.91660941e-01
-3.58022690e-01 -1.27009615e-01 1.41352698e-01 1.00232208e+00
-1.00643528e+00 8.67133737e-01 6.55378389e+00 -8.77981633e-03
-1.05890036e+00 4.25183713e-01 4.97181654e-01 2.61192266e-02
-8.65014255e-01 1.19008936e-01 -4.26436871e-01 4.14808363e-01
9.59249258e-01 -4.33058619e-01 -2.24921390e-01 5.68373740e-01
2.33907446e-01 1.38428584e-01 -9.14903164e-01 3.22867900e-01
2.71969557e-01 -1.82233298e+00 -1.41490504e-01 2.39940807e-01
6.08777225e-01 1.26646170e-02 1.59924939e-01 6.57162890e-02
3.19708318e-01 -9.14620936e-01 8.32489610e-01 2.48015206e-02
3.52548182e-01 -5.35370052e-01 6.53894901e-01 3.55741799e-01
-5.36791325e-01 1.06723877e-02 -9.37257409e-02 -2.83531010e-01
3.29446018e-01 8.62061203e-01 -9.59211171e-01 3.39239180e-01
3.74892890e-01 7.19968855e-01 -2.27494568e-01 2.28787303e-01
-6.51435494e-01 9.94180322e-01 -4.06306013e-02 -2.11023539e-01
2.38799214e-01 -2.30944574e-01 8.45794499e-01 1.27085423e+00
-2.86719739e-01 -1.01891011e-01 -2.06316803e-02 9.90120590e-01
-4.04398739e-01 1.90779775e-01 -6.67893589e-01 -2.61465639e-01
5.38583994e-01 1.10002124e+00 -7.18264401e-01 -5.45724034e-01
-1.01482466e-01 8.33983064e-01 3.92215908e-01 2.53222704e-01
-8.36284161e-01 -8.18136856e-02 9.40290809e-01 2.89823920e-01
-5.79118244e-02 -5.03123328e-02 -6.70184612e-01 -1.20338392e+00
2.29571648e-02 -8.53278458e-01 2.75796920e-01 -2.08901882e-01
-1.39846087e+00 5.63410401e-01 -2.48455629e-01 -9.49312806e-01
-2.69099981e-01 -4.21419472e-01 -9.70379472e-01 7.61832952e-01
-1.49076426e+00 -9.32430267e-01 -1.95057578e-02 1.93376958e-01
1.39975697e-01 1.45531833e-01 8.17950070e-01 -9.39806551e-03
-6.13103807e-01 2.83428848e-01 -8.66285414e-02 5.84001303e-01
7.57290065e-01 -1.14107382e+00 4.53255624e-01 4.44422513e-01
1.74847767e-01 9.51539397e-01 1.05474114e+00 -7.33772695e-01
-6.85269833e-01 -7.97832727e-01 1.53077960e+00 -7.68395305e-01
1.14348280e+00 -5.75470746e-01 -1.09834111e+00 7.86561549e-01
5.18912077e-01 -4.53489065e-01 1.16037023e+00 7.46398032e-01
-9.63707864e-01 6.55827403e-01 -1.04301655e+00 6.56190455e-01
8.93612564e-01 -6.88174546e-01 -1.18597412e+00 5.17627120e-01
6.75077617e-01 -2.65994519e-01 -5.78896165e-01 -1.25057980e-01
6.04389668e-01 -1.08375335e+00 8.21176112e-01 -1.02750552e+00
9.52887356e-01 -1.65310651e-02 -2.57995594e-02 -1.29019642e+00
-4.49639529e-01 -3.37817073e-01 -3.07892293e-01 1.14326513e+00
7.38376617e-01 -7.51995265e-01 7.31012940e-01 5.40066324e-02
2.87566364e-01 -6.78783596e-01 -8.07331741e-01 -4.41349328e-01
5.01294851e-01 -1.34772167e-01 3.82144183e-01 1.54062784e+00
4.49133217e-01 9.07864928e-01 -2.74499506e-01 1.95747480e-01
8.96711946e-01 6.27619922e-01 5.13770878e-01 -1.75683606e+00
-1.03897572e-01 -7.19780564e-01 -2.25962847e-01 -8.12051415e-01
4.77175415e-01 -1.09954071e+00 -3.65444750e-01 -1.66140354e+00
4.03827965e-01 -3.03000242e-01 -1.21152960e-01 1.49663880e-01
-7.97884166e-02 2.63252348e-01 3.44891250e-02 5.90520501e-01
-2.44283214e-01 3.79033059e-01 8.50057364e-01 -4.12314236e-01
-3.16182017e-01 -1.96741685e-01 -1.11151695e+00 1.05058384e+00
1.01061344e+00 -6.90892994e-01 -1.75684225e-02 -1.76718876e-01
5.29285729e-01 -1.38876885e-01 4.74901289e-01 -2.31053293e-01
-6.13838285e-02 -1.49512380e-01 1.37266051e-02 -6.51476264e-01
2.77738154e-01 -9.37030017e-02 -2.76662737e-01 6.29641354e-01
-6.85163379e-01 -5.61955795e-02 -1.86721414e-01 7.68344522e-01
-1.82310566e-01 -5.09419702e-02 6.63951397e-01 1.01636469e-01
3.37358825e-02 -3.27454090e-01 -4.51307505e-01 4.28717822e-01
7.97856867e-01 1.43893167e-01 -8.14249456e-01 -6.56850278e-01
-8.01659763e-01 5.85082360e-02 3.89639854e-01 5.57091773e-01
4.83451635e-01 -1.23928428e+00 -1.17408359e+00 -2.33893678e-01
6.28034249e-02 -6.00360155e-01 -3.17921750e-02 1.15372550e+00
-1.80741057e-01 1.92364380e-01 4.57432903e-02 -5.94215035e-01
-1.56823039e+00 2.92323470e-01 -2.02990770e-02 -2.84000009e-01
-6.26348913e-01 6.51691854e-01 5.86020835e-02 -7.08434701e-01
-1.08464554e-01 -7.51647577e-02 -5.64412892e-01 7.77663052e-01
5.96855521e-01 3.50773066e-01 -2.48813346e-01 -1.00707030e+00
-5.80564022e-01 1.58862382e-01 -2.54644603e-01 -5.13986170e-01
1.27331364e+00 -1.36474939e-02 -3.77642512e-01 8.50706339e-01
1.18579853e+00 6.92459583e-01 -5.76562643e-01 -3.36491048e-01
3.84037316e-01 -1.45870298e-01 -1.15899667e-01 -4.09047663e-01
-4.81352180e-01 8.35931361e-01 -5.55475391e-02 7.33397961e-01
1.19248018e-01 3.16462934e-01 5.20902812e-01 1.33211687e-02
-1.85316995e-01 -1.04978454e+00 2.14622691e-01 5.89514792e-01
8.88483107e-01 -1.15792191e+00 9.94177163e-02 -7.90078282e-01
-8.12074840e-01 8.29692602e-01 2.24130470e-02 -1.36970386e-01
5.38486898e-01 -2.75319666e-02 4.08621281e-01 -6.70901895e-01
-7.82269180e-01 2.09601864e-01 2.01196715e-01 2.73235202e-01
9.43986297e-01 2.97286332e-01 -9.65534389e-01 4.10515577e-01
-6.03124619e-01 -5.93892455e-01 8.16538215e-01 7.06790745e-01
-5.07880092e-01 -1.01104319e+00 -5.23453116e-01 2.70628244e-01
-5.83313346e-01 1.95504585e-03 -1.07088947e+00 5.89485765e-01
-2.34418854e-01 1.05469251e+00 1.08734027e-01 -3.62500489e-01
4.81757075e-02 3.30890834e-01 1.16513193e-01 -6.34879172e-01
-7.01453686e-01 -1.66390568e-01 6.02584839e-01 -2.00061992e-01
-5.65591574e-01 -8.19627821e-01 -1.21404171e+00 -6.93262517e-01
-2.04594031e-01 4.87314433e-01 6.30197167e-01 7.79369950e-01
3.36441010e-01 4.68803823e-01 6.28749251e-01 -2.14984357e-01
-6.59630001e-01 -1.07568359e+00 -2.08509237e-01 5.85439086e-01
4.73715246e-01 -6.51782155e-01 -6.93374634e-01 -2.66469955e-01] | [8.832176208496094, 10.060871124267578] |
66d15c00-2db4-4bfc-aa38-4141cb174887 | temporal-segment-networks-towards-good | 1608.00859 | null | http://arxiv.org/abs/1608.00859v1 | http://arxiv.org/pdf/1608.00859v1.pdf | Temporal Segment Networks: Towards Good Practices for Deep Action Recognition | Deep convolutional networks have achieved great success for visual
recognition in still images. However, for action recognition in videos, the
advantage over traditional methods is not so evident. This paper aims to
discover the principles to design effective ConvNet architectures for action
recognition in videos and learn these models given limited training samples.
Our first contribution is temporal segment network (TSN), a novel framework for
video-based action recognition. which is based on the idea of long-range
temporal structure modeling. It combines a sparse temporal sampling strategy
and video-level supervision to enable efficient and effective learning using
the whole action video. The other contribution is our study on a series of good
practices in learning ConvNets on video data with the help of temporal segment
network. Our approach obtains the state-the-of-art performance on the datasets
of HMDB51 ( $ 69.4\% $) and UCF101 ($ 94.2\% $). We also visualize the learned
ConvNet models, which qualitatively demonstrates the effectiveness of temporal
segment network and the proposed good practices. | ['Luc van Gool', 'Xiaoou Tang', 'Yu Qiao', 'Limin Wang', 'Zhe Wang', 'Yuanjun Xiong', 'Dahua Lin'] | 2016-08-02 | null | null | null | null | ['multimodal-activity-recognition'] | ['computer-vision'] | [ 2.88714677e-01 -2.12602526e-01 -6.80733562e-01 -2.96539843e-01
-3.36280257e-01 3.55681293e-02 4.68454510e-01 -8.56291711e-01
-4.39177126e-01 6.93004012e-01 2.68285125e-01 -3.83752026e-02
-2.76918322e-01 -3.99817526e-01 -8.77221823e-01 -8.37092578e-01
-4.77499872e-01 1.23477001e-02 4.69437867e-01 -7.19966516e-02
1.88460469e-01 5.50244331e-01 -1.38996422e+00 8.58974755e-01
2.57909834e-01 1.41520619e+00 8.22618455e-02 6.93973422e-01
-7.07378564e-03 1.84331596e+00 -4.75218087e-01 -1.03801049e-01
3.66473049e-01 -4.52150583e-01 -1.04910624e+00 4.11973447e-01
5.97256124e-01 -7.86878705e-01 -1.01895666e+00 5.81852853e-01
2.16938019e-01 2.89972246e-01 4.47906137e-01 -1.29885721e+00
-4.18142766e-01 4.47757244e-01 -3.71130198e-01 6.64373517e-01
1.57651603e-01 5.16951859e-01 7.44012773e-01 -5.18773973e-01
7.87979543e-01 1.11679673e+00 6.73165739e-01 7.15586126e-01
-7.11448371e-01 -5.31259358e-01 3.56743515e-01 8.47047269e-01
-9.53485787e-01 -4.96287823e-01 5.90476096e-01 -3.36417675e-01
1.45662963e+00 -1.25748470e-01 1.11099589e+00 1.52734232e+00
1.64413512e-01 1.55426252e+00 7.35805094e-01 -2.01525360e-01
5.32730995e-03 -4.69680995e-01 8.20180625e-02 8.01218331e-01
-3.42361063e-01 3.14230621e-01 -8.89537930e-01 4.02324319e-01
1.10439527e+00 4.46408838e-01 -2.53389806e-01 -2.57641166e-01
-1.12097442e+00 5.35536468e-01 5.13093412e-01 5.69678009e-01
-4.83734697e-01 6.90575123e-01 7.19023228e-01 4.04838949e-01
2.21740797e-01 -5.20169809e-02 -4.95589703e-01 -7.18241394e-01
-1.09293032e+00 -1.34150833e-01 2.97101557e-01 9.54080880e-01
2.38954023e-01 6.01679087e-01 -3.64026368e-01 6.43447280e-01
1.35291025e-01 3.03151995e-01 5.61558783e-01 -1.34356809e+00
2.92560160e-01 5.99014580e-01 -7.19367876e-04 -7.41879642e-01
-1.58611208e-01 -2.12125853e-01 -9.67779875e-01 1.58439279e-01
3.76001298e-01 3.21197212e-02 -1.21300280e+00 1.40274775e+00
-1.18024878e-01 6.87745869e-01 1.94616124e-01 6.62215412e-01
8.64274263e-01 6.07440293e-01 2.52100259e-01 -2.61485815e-01
8.32147539e-01 -1.42809486e+00 -7.28929758e-01 -1.32906958e-02
7.81061471e-01 -1.99684441e-01 7.06993341e-01 4.82745737e-01
-1.04223657e+00 -9.60726440e-01 -8.93386066e-01 9.08189490e-02
-1.37739792e-01 4.77125436e-01 7.28465855e-01 2.34364480e-01
-1.16150701e+00 8.75198483e-01 -1.32863939e+00 -5.11855721e-01
9.56386983e-01 4.71467853e-01 -3.85756046e-01 -1.76914066e-01
-9.27026749e-01 5.90219915e-01 4.40935552e-01 1.88398004e-01
-1.64683676e+00 -5.61370492e-01 -7.25762546e-01 -4.80090976e-02
5.27736783e-01 -2.81044036e-01 1.30485296e+00 -1.61540532e+00
-1.41368318e+00 5.57880819e-01 -2.27191653e-02 -1.15943313e+00
6.23635292e-01 -3.63434255e-01 -4.67953235e-01 7.91502714e-01
-1.31935820e-01 1.00317264e+00 1.01282072e+00 -6.09007835e-01
-6.67279065e-01 -9.98732299e-02 2.32603237e-01 -6.02295324e-02
-3.83105636e-01 9.04596820e-02 -4.54871356e-01 -6.84776783e-01
-2.50802249e-01 -8.26631606e-01 -1.60959259e-01 1.64932325e-01
-4.35523465e-02 -4.50604796e-01 1.30309486e+00 -6.48276925e-01
1.11676145e+00 -2.28932524e+00 8.09305757e-02 -3.44673365e-01
9.56527665e-02 7.92312324e-01 -3.07114303e-01 4.43635672e-01
-2.28145152e-01 -1.32820755e-01 5.33400178e-02 -1.69038191e-01
-2.31255606e-01 4.58725750e-01 -2.40883082e-01 3.53403181e-01
1.47455469e-01 1.23056948e+00 -6.61869228e-01 -5.62087595e-01
5.13813674e-01 5.42670727e-01 -2.75660306e-01 1.34873584e-01
-3.58022809e-01 2.23128721e-01 -5.31723499e-01 7.89215803e-01
1.44679770e-01 -4.91893649e-01 2.41928309e-01 -3.72275323e-01
-9.05747786e-02 -1.73433758e-02 -8.91329050e-01 1.74529827e+00
1.67961553e-01 1.03815508e+00 -3.03425133e-01 -1.43968987e+00
6.01589739e-01 6.10525668e-01 1.09986973e+00 -8.51155102e-01
1.25876948e-01 -8.52708369e-02 -1.12481035e-01 -8.76223147e-01
2.41041511e-01 1.71858460e-01 4.35325265e-01 2.73981512e-01
4.64962602e-01 5.20019114e-01 4.62823540e-01 1.94724247e-01
1.40383112e+00 5.92235863e-01 3.10800970e-02 -4.66997027e-02
4.22592670e-01 9.32094529e-02 4.91210729e-01 6.65234864e-01
-5.71199059e-01 3.01892012e-01 5.40844619e-01 -1.11630893e+00
-8.86798739e-01 -5.36095977e-01 2.02307954e-01 9.42746520e-01
-2.01576486e-01 -3.90421182e-01 -6.89761877e-01 -9.81799304e-01
-2.86039621e-01 2.70917237e-01 -9.12732065e-01 -1.26106322e-01
-7.14221418e-01 -2.56305873e-01 7.14099944e-01 1.18921196e+00
1.30982018e+00 -1.56158280e+00 -7.99575388e-01 1.53441474e-01
-2.24755332e-01 -1.25542390e+00 -1.94534332e-01 -6.27583638e-02
-1.18653429e+00 -1.38350546e+00 -7.90989220e-01 -7.63461232e-01
3.69772881e-01 5.10083973e-01 9.51748729e-01 1.85384378e-01
-2.48581395e-01 7.20871806e-01 -6.02805018e-01 -4.85257208e-02
6.10696115e-02 -1.35547206e-01 -4.39079143e-02 2.14482844e-01
6.83599591e-01 -5.68544626e-01 -6.86692297e-01 4.57789987e-01
-9.49963152e-01 -7.83748701e-02 6.48149848e-01 8.22353065e-01
4.77568895e-01 1.43368924e-02 2.51516581e-01 -3.99282664e-01
-6.22586757e-02 -1.09725758e-01 -4.40050542e-01 3.38194191e-01
-1.46096349e-01 -1.68975160e-01 4.57322478e-01 -5.29386759e-01
-9.28793073e-01 1.77027985e-01 -5.76300137e-02 -1.01418984e+00
-2.66310960e-01 3.73298287e-01 3.04835886e-01 -2.11609960e-01
4.93883997e-01 6.43364608e-01 6.33779168e-02 -5.01631141e-01
-2.86649372e-02 3.73834312e-01 3.04945260e-01 -2.09441766e-01
1.08095050e-01 7.82040656e-01 -6.79818690e-02 -1.08607686e+00
-9.08059716e-01 -4.82535720e-01 -9.46226120e-01 -5.25360107e-01
1.18729460e+00 -1.05423450e+00 -7.68219709e-01 9.03576791e-01
-8.72137785e-01 -8.94298136e-01 -4.73348141e-01 6.27878487e-01
-9.25457656e-01 4.89151478e-01 -8.76777291e-01 -6.06425643e-01
-9.00767520e-02 -1.00350308e+00 8.34424555e-01 1.13829918e-01
2.07348347e-01 -8.63861799e-01 -8.20353925e-02 5.02146780e-01
3.17898184e-01 2.48654410e-01 4.05340523e-01 -4.11067635e-01
-1.08377945e+00 3.76626924e-02 -2.39981890e-01 8.99705052e-01
-1.43977836e-01 1.96187183e-01 -7.46614099e-01 -3.07565778e-01
-1.04875259e-01 -7.60785997e-01 1.27516115e+00 7.50720084e-01
1.43205070e+00 -1.33817643e-01 -3.05062622e-01 7.11130142e-01
1.33991539e+00 6.32153034e-01 1.23669946e+00 3.23391557e-01
7.38154709e-01 1.76496893e-01 6.65376902e-01 4.12893891e-01
-2.23761320e-01 6.73999727e-01 4.77271050e-01 3.64155811e-03
-3.06501150e-01 -1.07062489e-01 7.32398927e-01 5.57977438e-01
-7.56104827e-01 -1.76511422e-01 -7.35905886e-01 6.38824522e-01
-2.03062057e+00 -1.52632475e+00 5.23401648e-02 1.63678074e+00
3.96783799e-01 1.76545084e-01 3.71923029e-01 1.21155269e-01
3.06718618e-01 5.31383455e-01 -5.30744314e-01 -1.51050724e-02
-1.00408040e-01 2.87505358e-01 3.26143384e-01 -8.48371536e-02
-1.39489532e+00 1.04822206e+00 6.48889256e+00 1.01754367e+00
-1.12610161e+00 -1.49407210e-02 6.03795946e-01 -2.59389400e-01
5.39439738e-01 -3.17972422e-01 -7.49574900e-01 3.04181486e-01
1.06419337e+00 2.26333231e-01 1.18728630e-01 9.42397833e-01
4.11680132e-01 2.85122171e-02 -1.00110006e+00 1.15602171e+00
1.60704419e-01 -1.70291257e+00 8.95137936e-02 4.51084562e-02
9.04895604e-01 2.33026952e-01 -5.14294021e-03 4.61455405e-01
2.08694607e-01 -1.20504296e+00 4.86636370e-01 6.60162926e-01
7.20487356e-01 -6.18118167e-01 7.54020333e-01 9.02552903e-02
-1.33720756e+00 -4.85520691e-01 -4.21498448e-01 -6.56868145e-02
1.69754520e-01 -5.23665957e-02 -3.54324907e-01 4.12961632e-01
1.01069999e+00 1.71961308e+00 -4.39489663e-01 1.01634991e+00
-2.36290067e-01 8.81230175e-01 8.12844634e-02 7.96237402e-03
8.32712173e-01 1.56696886e-02 2.70234734e-01 1.18964469e+00
1.94739446e-01 2.67998219e-01 2.44309112e-01 3.73569995e-01
-1.26632014e-02 -1.70085698e-01 -7.81377077e-01 -4.45491970e-01
-2.61975735e-01 7.86197901e-01 -5.51431060e-01 -6.04813755e-01
-5.60996890e-01 8.26058805e-01 2.17260852e-01 4.73702967e-01
-9.20413673e-01 1.40703544e-01 6.00325406e-01 1.24091208e-01
9.52774167e-01 -3.35378855e-01 3.63004297e-01 -1.15903354e+00
6.60113096e-02 -1.07770622e+00 5.95903039e-01 -8.98827672e-01
-7.80604899e-01 5.00582039e-01 1.15979359e-01 -1.56336761e+00
-2.66973972e-01 -9.50438738e-01 -2.82398105e-01 4.23520952e-02
-1.24922717e+00 -1.22892690e+00 -3.16556364e-01 1.10618961e+00
1.02520418e+00 -4.55700815e-01 5.81316710e-01 3.48624051e-01
-5.21732569e-01 4.20067191e-01 8.08937550e-02 6.15654111e-01
1.79308698e-01 -8.18792522e-01 1.02690510e-01 8.84060025e-01
3.18900734e-01 1.29233062e-01 1.60664976e-01 -4.65116620e-01
-1.35057235e+00 -1.20678186e+00 5.43764114e-01 -3.04114640e-01
5.74278116e-01 9.22055468e-02 -6.48134053e-01 1.07360387e+00
4.82467830e-01 2.19618171e-01 3.94224137e-01 -2.94612348e-01
-2.77080625e-01 -3.06443691e-01 -8.30464602e-01 4.79317307e-01
1.42231905e+00 -4.78443086e-01 -4.53676373e-01 4.29768384e-01
5.79248369e-01 -9.75310579e-02 -8.34550917e-01 5.62852919e-01
6.72189236e-01 -1.19662380e+00 9.08995748e-01 -1.12395704e+00
7.70422876e-01 -1.19135097e-01 -2.67225742e-01 -7.92292356e-01
-3.28009725e-01 -6.05952203e-01 -4.62229401e-01 5.96317112e-01
4.07304950e-02 -1.21828802e-01 1.16444707e+00 1.80498362e-01
-2.76450813e-01 -8.56805682e-01 -1.01391685e+00 -1.06707036e+00
-4.07059699e-01 -5.39025187e-01 9.30324793e-02 7.31762767e-01
-3.74907196e-01 -6.28855526e-02 -8.53042603e-01 -4.09464806e-01
4.08365726e-01 -1.49745584e-01 6.89077556e-01 -6.42839670e-01
-3.65709096e-01 -2.81908661e-01 -9.23872709e-01 -1.45747137e+00
1.06572337e-01 -2.15912044e-01 -2.18297303e-01 -1.39857709e+00
2.63474226e-01 1.06660932e-01 -7.12837994e-01 7.10373640e-01
5.34107864e-01 4.12383169e-01 2.48718351e-01 1.88392982e-01
-1.16523623e+00 7.00530887e-01 1.40458953e+00 -4.31043535e-01
1.14700831e-01 3.21419835e-02 -1.03739016e-01 6.26618743e-01
7.61436462e-01 -2.63399303e-01 -6.92655146e-01 -5.40485203e-01
-3.03664237e-01 1.75118059e-01 6.39357328e-01 -1.42191780e+00
1.06632680e-01 -2.85794705e-01 4.19322610e-01 -7.44188547e-01
5.98970771e-01 -8.73736739e-01 1.51571691e-01 7.69270897e-01
-4.62725163e-01 -1.50853589e-01 1.74089283e-01 6.70209229e-01
-5.23828506e-01 6.28310964e-02 7.76337445e-01 -5.04851341e-01
-1.46218884e+00 5.75724840e-01 -4.16056454e-01 -8.46833661e-02
1.22750592e+00 -5.44330955e-01 -2.53125221e-01 -3.73284966e-01
-9.53544021e-01 2.17481833e-02 1.50801316e-01 4.28792864e-01
8.51046562e-01 -1.51253784e+00 -5.14302254e-01 1.27036333e-01
-1.31868739e-02 -5.23107290e-01 6.21244013e-01 1.14004886e+00
-5.64525604e-01 7.53070354e-01 -7.25804865e-01 -7.69155681e-01
-1.40197492e+00 6.26650631e-01 5.64265609e-01 -3.43138576e-01
-8.10400844e-01 9.92058158e-01 4.34358455e-02 2.50875205e-01
6.68358207e-01 -3.49862993e-01 -4.64047581e-01 -1.56796262e-01
7.81018853e-01 4.27721411e-01 -2.26526976e-01 -5.98429739e-01
-2.73227364e-01 4.55335677e-01 -1.81807667e-01 1.40339091e-01
1.63554335e+00 2.78831273e-01 9.91161764e-02 3.73309612e-01
1.28030968e+00 -9.65712488e-01 -1.78691745e+00 -1.63716912e-01
-1.62779287e-01 -6.27232850e-01 -7.42152557e-02 -6.37502134e-01
-1.61207485e+00 1.00457191e+00 7.39370167e-01 9.44009125e-02
1.23351443e+00 -1.85876504e-01 8.51302803e-01 6.43750191e-01
3.68333012e-01 -1.27507651e+00 7.96698153e-01 6.36144340e-01
7.64226198e-01 -1.20783925e+00 -1.71860546e-01 -1.36591969e-02
-6.79454625e-01 1.19187772e+00 7.40716457e-01 -2.86396414e-01
6.55670583e-01 -1.07533172e-01 1.54602164e-02 -5.52891254e-01
-1.05696058e+00 -1.76462248e-01 2.37135664e-01 7.24057972e-01
2.72592872e-01 -1.55643269e-01 -6.81081340e-02 -2.57541072e-02
5.55152118e-01 4.74020928e-01 2.28034541e-01 1.25663698e+00
-3.31476241e-01 -9.27195013e-01 2.77021646e-01 4.78324324e-01
-6.22936487e-01 4.29694913e-02 -3.42632681e-01 1.02492607e+00
3.74983400e-02 7.37037957e-01 -4.82124910e-02 -5.66214740e-01
1.40408829e-01 1.25543699e-01 5.75583458e-01 -1.87364191e-01
-4.09737498e-01 -2.88768727e-02 1.98280260e-01 -1.23237371e+00
-1.13631129e+00 -7.25754917e-01 -8.85286272e-01 -2.48914644e-01
1.89861313e-01 -6.78463802e-02 1.96920201e-01 1.06220794e+00
4.24701363e-01 5.79541385e-01 3.23596716e-01 -7.62860000e-01
-3.66766959e-01 -1.07460618e+00 -7.70795941e-01 4.22052175e-01
3.71013522e-01 -7.36594796e-01 -1.44341215e-01 4.43243951e-01] | [8.46163272857666, 0.6142162680625916] |
0f485ea0-484d-4f4b-89ed-49146d3f55f6 | a-strategy-oriented-bayesian-soft-actor | 2303.04193 | null | https://arxiv.org/abs/2303.04193v1 | https://arxiv.org/pdf/2303.04193v1.pdf | A Strategy-Oriented Bayesian Soft Actor-Critic Model | Adopting reasonable strategies is challenging but crucial for an intelligent agent with limited resources working in hazardous, unstructured, and dynamic environments to improve the system's utility, decrease the overall cost, and increase mission success probability. This paper proposes a novel hierarchical strategy decomposition approach based on the Bayesian chain rule to separate an intricate policy into several simple sub-policies and organize their relationships as Bayesian strategy networks (BSN). We integrate this approach into the state-of-the-art DRL method -- soft actor-critic (SAC) and build the corresponding Bayesian soft actor-critic (BSAC) model by organizing several sub-policies as a joint policy. We compare the proposed BSAC method with the SAC and other state-of-the-art approaches such as TD3, DDPG, and PPO on the standard continuous control benchmarks -- Hopper-v2, Walker2d-v2, and Humanoid-v2 -- in MuJoCo with the OpenAI Gym environment. The results demonstrate that the promising potential of the BSAC method significantly improves training efficiency. | ['Ramviyas Parasuraman', 'Qin Yang'] | 2023-03-07 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [-4.65547532e-01 5.57841539e-01 -3.58190387e-01 -1.29306326e-02
-3.50872785e-01 -1.10323861e-01 6.34371459e-01 -5.67690544e-02
-7.73642063e-01 1.29400563e+00 2.34607071e-01 -2.55495340e-01
-7.01336443e-01 -5.19123197e-01 -4.92213577e-01 -9.66321766e-01
-1.62982285e-01 8.37340772e-01 8.90168011e-01 -5.70649087e-01
1.52275085e-01 3.26357663e-01 -1.42070687e+00 -3.13292503e-01
9.22353506e-01 7.45327711e-01 2.06050113e-01 5.30577004e-01
4.52492684e-01 1.09285545e+00 -5.43445349e-01 6.86762854e-02
3.19914341e-01 -2.19170108e-01 -5.01471460e-01 -2.22134158e-01
-4.38115567e-01 -2.65951067e-01 -3.51459891e-01 7.17326343e-01
7.05851734e-01 8.70662570e-01 6.38518035e-01 -1.21391273e+00
3.14049155e-01 7.40050316e-01 -3.27914119e-01 -5.01169562e-02
2.31033251e-01 5.09995937e-01 3.76004666e-01 1.42030537e-01
4.83293742e-01 1.66030312e+00 4.82968509e-01 6.84609175e-01
-1.18550754e+00 -1.64631158e-01 7.17848480e-01 3.89781654e-01
-9.36750412e-01 -3.18132676e-02 6.03780150e-01 -2.42539003e-01
1.02122533e+00 -6.09498471e-03 1.04934943e+00 1.37421668e+00
2.70349562e-01 1.04096305e+00 1.48318386e+00 -2.73628324e-01
8.32711577e-01 -3.14830154e-01 1.16001680e-01 8.32653165e-01
4.24588650e-01 5.60845613e-01 -4.44402188e-01 -1.42858356e-01
4.64701742e-01 -3.41318309e-01 7.83899352e-02 -4.28060383e-01
-9.93910730e-01 6.65390551e-01 1.09245211e-01 8.56493413e-03
-7.79636025e-01 5.18972814e-01 2.28685126e-01 -2.05305242e-03
6.93250224e-02 2.62329608e-01 -3.39820355e-01 -5.59932649e-01
-4.72025126e-01 7.48241842e-01 1.05408728e+00 7.75297880e-01
4.82413359e-02 6.40825689e-01 -3.39264423e-01 5.92724562e-01
5.97192705e-01 2.01219961e-01 2.70224839e-01 -1.63047826e+00
1.90928891e-01 2.75945604e-01 6.60186827e-01 -6.67470276e-01
-6.56274140e-01 -3.97049308e-01 -3.84430766e-01 7.64775515e-01
3.58504146e-01 -3.31262320e-01 -6.84867799e-01 1.63456011e+00
5.90285838e-01 -1.74512058e-01 1.61944866e-01 7.03551948e-01
3.08723241e-01 6.67689681e-01 2.21166193e-01 -2.44862750e-01
9.02974665e-01 -1.05695915e+00 -8.98339212e-01 -3.44605595e-01
-5.46631180e-02 2.70367746e-04 8.05371523e-01 9.23905373e-01
-1.22304928e+00 -5.07834136e-01 -1.09346783e+00 6.54893339e-01
-1.65680841e-01 -1.90274697e-02 4.69438136e-01 5.03126383e-01
-6.81082964e-01 7.40443468e-01 -1.25224459e+00 -2.62045830e-01
1.66118145e-01 3.49769592e-01 1.93951190e-01 9.18324962e-02
-1.01737332e+00 1.42540419e+00 8.64854276e-01 -7.75267705e-02
-1.68474627e+00 -2.08964169e-01 -4.70309556e-01 -1.07780462e-02
9.41069543e-01 -5.50813437e-01 1.40991104e+00 -2.63133794e-01
-2.51282620e+00 -7.52169341e-02 5.60359895e-01 -6.05877936e-01
6.96860373e-01 -4.47946578e-01 1.96055640e-02 1.94126785e-01
-2.08008870e-01 6.80915296e-01 7.36790001e-01 -1.58382070e+00
-9.14412200e-01 -1.30080178e-01 3.47128689e-01 6.04221880e-01
1.18255481e-01 -2.59448409e-01 -1.77665129e-01 -3.56450051e-01
-2.49705926e-01 -1.05964363e+00 -6.12840474e-01 -5.21643400e-01
-2.24065736e-01 -7.43453443e-01 4.37020600e-01 -5.86097360e-01
1.36332059e+00 -1.91507888e+00 8.04389477e-01 1.78116828e-01
-1.91402793e-01 3.04994911e-01 3.89929354e-01 3.70639622e-01
5.12330532e-01 -2.70395756e-01 2.24917680e-02 -1.91799432e-01
4.06908542e-01 8.08934808e-01 8.05538967e-02 4.71569508e-01
-4.74711895e-01 2.09718555e-01 -1.23717415e+00 -6.56728745e-01
3.84765774e-01 7.66144395e-02 -5.67632735e-01 3.52156699e-01
-6.79818094e-01 5.07194757e-01 -8.71474922e-01 5.74701905e-01
-1.89846363e-02 4.15826529e-01 6.26018584e-01 4.32197563e-02
-2.19556823e-01 -9.38995108e-02 -1.36190593e+00 1.58449519e+00
-1.11985214e-01 1.23122051e-01 3.20089996e-01 -1.26959479e+00
8.13046455e-01 2.44591728e-01 7.25815594e-01 -4.92099524e-01
4.51048970e-01 1.87155411e-01 1.41490167e-02 -5.69625020e-01
4.77545112e-01 9.66284573e-02 -2.20996171e-01 -1.70049965e-01
3.48488718e-01 -2.46284932e-01 5.58048666e-01 1.87911555e-01
1.14715624e+00 1.06425846e+00 4.22995150e-01 -5.11985004e-01
3.63016844e-01 3.12473029e-02 1.00275230e+00 1.00014639e+00
-6.17495060e-01 -1.49197310e-01 7.45388567e-01 -1.82852149e-01
-5.88697195e-01 -9.80853438e-01 3.76295954e-01 1.04308176e+00
1.22131348e-01 -2.35630631e-01 -6.07902706e-01 -7.42123783e-01
1.80142075e-01 1.24502325e+00 -4.77248728e-01 -1.42752662e-01
-5.31389475e-01 -6.81005538e-01 5.28566778e-01 4.27516341e-01
6.97639823e-01 -9.06227589e-01 -1.18253636e+00 6.83910310e-01
-1.40138820e-01 -9.60740209e-01 2.03770339e-01 2.98072666e-01
-6.62980795e-01 -9.69180286e-01 -4.92110670e-01 -1.34159833e-01
2.31329158e-01 -2.27688521e-01 8.34251404e-01 -3.62150043e-01
1.76610380e-01 7.92605698e-01 -5.46503544e-01 -6.12061799e-01
-3.98089677e-01 -1.62603229e-01 4.62187886e-01 -1.52943626e-01
-3.41488600e-01 -4.57103491e-01 -5.25699556e-01 5.71751297e-01
-2.71120399e-01 -2.38128863e-02 3.71463805e-01 7.85866857e-01
5.20597994e-01 2.23846987e-01 4.38048601e-01 -3.57315004e-01
6.35740042e-01 -3.54348332e-01 -9.60555494e-01 2.72093743e-01
-8.12014222e-01 2.02548310e-01 4.57339793e-01 -5.95806360e-01
-1.63535202e+00 3.23278792e-02 4.73281508e-03 -2.95910269e-01
5.15802093e-02 4.84067947e-01 -3.18222791e-02 -1.62487235e-02
6.96727216e-01 5.82777485e-02 1.05386205e-01 -4.04750109e-01
3.17776322e-01 3.90037507e-01 4.56136793e-01 -1.09838402e+00
3.80421817e-01 2.13831410e-01 3.34123552e-01 -4.76007104e-01
-8.18106174e-01 -2.50613123e-01 -7.84846693e-02 -8.90493035e-01
1.05569124e+00 -7.04738915e-01 -1.38026035e+00 5.18619537e-01
-7.65255630e-01 -9.85233605e-01 -3.04747015e-01 6.82722211e-01
-9.01251316e-01 2.72373617e-01 -4.14590240e-01 -1.24018300e+00
2.27728188e-01 -1.26409566e+00 4.72480565e-01 6.50014818e-01
2.31643002e-02 -8.93397987e-01 3.44861776e-01 4.19139802e-01
5.47640860e-01 5.34919977e-01 6.45300984e-01 -6.04214787e-01
-1.74374834e-01 -1.68526918e-02 5.90686142e-01 4.19482887e-01
-4.24245805e-01 -2.45570280e-02 -3.49461019e-01 -4.28463072e-01
-8.97847563e-02 -7.08144426e-01 5.83941758e-01 4.76807833e-01
7.50041127e-01 -2.18176231e-01 -2.59760052e-01 9.76976976e-02
1.27899182e+00 4.93225306e-01 5.49714029e-01 7.27699697e-01
2.85136312e-01 5.79379916e-01 1.09852874e+00 7.06643462e-01
6.45891070e-01 7.93037236e-01 8.11110675e-01 4.52944726e-01
1.90502647e-02 -1.94075614e-01 8.07217300e-01 7.52383530e-01
-7.45517492e-01 -3.31981152e-01 -7.10555494e-01 2.77386308e-01
-2.56273866e+00 -1.14421368e+00 1.85199499e-01 1.92821109e+00
8.38038385e-01 5.98546207e-01 4.30241704e-01 4.65056635e-02
4.38143820e-01 1.79901868e-01 -5.31418383e-01 -3.07414263e-01
2.51614451e-01 -1.77799344e-01 6.86849713e-01 5.81933379e-01
-8.55282784e-01 8.25693488e-01 6.10204840e+00 1.18819499e+00
-5.84239244e-01 1.90741435e-01 1.58147037e-01 -1.64614499e-01
3.46595138e-01 1.87558457e-01 -1.01783454e+00 5.66453397e-01
1.15136349e+00 7.39036500e-02 8.75586152e-01 1.16470957e+00
4.35756028e-01 -7.70416439e-01 -7.25031555e-01 4.51037645e-01
-1.47149637e-01 -9.21885431e-01 -5.30891001e-01 -1.46258682e-01
6.14343166e-01 2.68021151e-02 -4.98597622e-01 8.87587130e-01
1.27217793e+00 -5.16704679e-01 1.23757088e+00 9.02855694e-01
-2.07379665e-02 -5.97472072e-01 5.70463359e-01 6.83951080e-01
-9.48557973e-01 -7.15482533e-01 -2.57266343e-01 1.22458175e-01
3.49996269e-01 -2.16821721e-03 -4.07259971e-01 9.49123025e-01
9.09391224e-01 4.86742973e-01 -1.85849130e-01 9.35524344e-01
-6.95231199e-01 7.67330468e-01 -5.27485073e-01 -4.22365963e-01
5.56247115e-01 -3.18544865e-01 9.31321859e-01 6.20089471e-01
-6.57408079e-03 3.33105564e-01 6.67690158e-01 4.36646879e-01
7.59597838e-01 -5.89777648e-01 -1.66666165e-01 1.50579199e-01
4.71679598e-01 1.12172019e+00 -7.17236578e-01 -4.67838734e-01
6.72536865e-02 7.30448887e-02 1.82956174e-01 2.93122381e-01
-1.04141748e+00 -2.17124358e-01 1.51612759e-01 -2.98871070e-01
2.46718869e-01 -5.06526172e-01 1.11835331e-01 -6.86906099e-01
-4.00845170e-01 -1.12513375e+00 4.08202082e-01 -8.24431777e-01
-8.67734611e-01 3.94371539e-01 8.59624386e-01 -1.26508009e+00
-2.49311462e-01 -4.80014890e-01 -2.67302722e-01 2.78602570e-01
-1.34172320e+00 -8.53738844e-01 -3.36277813e-01 6.14314079e-01
7.27862537e-01 -4.56390858e-01 6.28977597e-01 -3.91374864e-02
-8.24823081e-01 -6.00030608e-02 2.38329291e-01 -5.30319870e-01
3.45166326e-01 -1.28130293e+00 -5.51280260e-01 8.20958734e-01
-8.97763729e-01 4.90529612e-02 9.06316698e-01 -8.61213624e-01
-1.30976164e+00 -6.48349583e-01 -5.50589077e-02 -6.23992532e-02
5.10055721e-01 8.94004479e-02 -4.29046839e-01 4.93444353e-01
3.96015823e-01 -2.95542270e-01 2.18466315e-02 -1.55457199e-01
4.06038225e-01 -3.80321890e-01 -1.20492911e+00 7.14912772e-01
8.38511646e-01 2.86497921e-01 -8.52986813e-01 2.35127196e-01
6.34353459e-01 -5.00524342e-01 -9.27725315e-01 4.57443416e-01
6.30648255e-01 -8.36857080e-01 9.19462621e-01 -6.90122306e-01
-8.81380811e-02 -3.92457217e-01 -2.09565744e-01 -1.69737327e+00
-4.38022882e-01 -1.02761579e+00 -5.76603234e-01 8.24392080e-01
-1.37320030e-02 -6.21113956e-01 5.16396821e-01 4.68609720e-01
-6.28536761e-01 -6.40069783e-01 -1.26281476e+00 -1.15094531e+00
-3.01520705e-01 -1.46797001e-01 1.77172691e-01 3.57672900e-01
-8.00096095e-02 3.59749570e-02 -5.53343356e-01 9.39209238e-02
1.07492399e+00 -6.56308830e-01 7.33388186e-01 -1.09099257e+00
-7.03781962e-01 -3.27714801e-01 1.49705738e-01 -7.25048006e-01
1.67439327e-01 -2.67712206e-01 4.77289498e-01 -1.85499561e+00
-3.39404881e-01 -5.99016130e-01 -3.84466648e-01 6.38562560e-01
5.12984283e-02 -7.08567441e-01 4.05172825e-01 8.17664713e-02
-1.08067656e+00 1.05275881e+00 1.23433709e+00 -1.24455497e-01
-2.96280593e-01 1.72561347e-01 -1.54850110e-01 1.03816116e+00
8.99852037e-01 -7.92860448e-01 -5.09416997e-01 -1.87220186e-01
2.31455892e-01 6.06195152e-01 2.12847665e-01 -1.33000886e+00
3.86425614e-01 -7.88350999e-01 -2.04500154e-01 -4.66805339e-01
5.79257429e-01 -9.16137695e-01 3.27689573e-02 1.03117585e+00
-2.79348165e-01 4.26821923e-03 2.24677362e-02 9.71847832e-01
8.32301974e-02 -3.53119582e-01 7.56510258e-01 -2.70071119e-01
-7.33227313e-01 -3.04830164e-01 -1.00211394e+00 -8.50976035e-02
1.43054581e+00 -2.63431724e-02 -4.16138530e-01 -2.93890178e-01
-1.04845953e+00 8.04414630e-01 -1.07247263e-01 -7.56040728e-03
3.89197141e-01 -9.80442822e-01 -4.36851591e-01 -4.90494668e-01
-4.84061509e-01 -8.35836232e-02 1.43945962e-01 1.02520823e+00
-4.65853244e-01 1.40571460e-01 -7.64917195e-01 -3.67852986e-01
-8.39714170e-01 2.98228055e-01 3.92081738e-01 -8.18616688e-01
-4.99078572e-01 3.67709726e-01 -6.84269369e-01 -3.75492811e-01
7.47011185e-01 -1.39177218e-01 -5.94817817e-01 -8.13378915e-02
-1.43814221e-01 1.00595307e+00 -3.95257980e-01 8.38109702e-02
-2.59231567e-01 2.36315325e-01 3.31461340e-01 -5.29658735e-01
1.64153147e+00 6.19512498e-02 1.65612504e-01 3.77045214e-01
-1.17524996e-01 -4.36221927e-01 -1.79252088e+00 5.16652986e-02
6.80489093e-02 -2.65236765e-01 3.21409106e-01 -9.97366965e-01
-6.85166299e-01 3.33409607e-01 7.82189250e-01 2.02810943e-01
8.00975084e-01 -3.95837486e-01 5.88767268e-02 7.20093071e-01
9.73312199e-01 -1.78854907e+00 3.82815927e-01 6.44640923e-01
8.92083526e-01 -8.43070447e-01 4.99697924e-01 1.16308697e-01
-9.94956255e-01 1.00159490e+00 8.46528769e-01 -3.40286106e-01
5.33087850e-01 4.75204363e-02 -6.12812817e-01 -5.99670522e-02
-1.03652883e+00 -4.50212002e-01 -1.16236471e-01 6.08935714e-01
-4.23285306e-01 1.70818746e-01 -5.90665221e-01 7.58747339e-01
4.10598695e-01 1.70438781e-01 6.72118723e-01 1.50459814e+00
-6.19029343e-01 -1.07179165e+00 -3.60240549e-01 1.05588175e-01
-2.08424881e-01 6.70458972e-01 1.68413907e-01 1.12773466e+00
4.03200276e-02 9.58755851e-01 -3.83116037e-01 -4.18159515e-01
5.95401943e-01 -1.16623528e-01 6.78066432e-01 -3.63806099e-01
-7.15132833e-01 3.45696986e-01 6.87293291e-01 -7.94498980e-01
-7.65930057e-01 -8.02861571e-01 -1.17997301e+00 -1.95328772e-01
-3.06565445e-02 1.55867487e-01 7.44246602e-01 8.67773175e-01
1.63246423e-01 1.09196746e+00 5.01229286e-01 -1.25626266e+00
-1.23306715e+00 -1.12945580e+00 -2.99788177e-01 -1.61194429e-01
-1.26726389e-01 -1.38425946e+00 -3.00335020e-01 -3.38649929e-01] | [4.170470237731934, 2.077409267425537] |
d729acfe-9131-47e5-9d4d-10d9e7dceac8 | cascaded-interactional-targeting-network-for | null | null | http://openaccess.thecvf.com/content_cvpr_2016/html/Zhou_Cascaded_Interactional_Targeting_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Zhou_Cascaded_Interactional_Targeting_CVPR_2016_paper.pdf | Cascaded Interactional Targeting Network for Egocentric Video Analysis | Knowing how hands move and what object is being manipulated are two key sub-tasks for analyzing first-person (egocentric) action. However, lack of fully annotated hand data as well as imprecise foreground segmentation make either sub-task challenging. This work aims to explicitly address these two issues via introducing a cascaded interactional targeting (i.e., infer both hand and active object regions) deep neural network. Firstly, a novel EM-like learning framework is proposed to train the pixel-level deep convolutional neural network (DCNN) by seamlessly integrating weakly supervised data (i.e., massive bounding box annotations) with a small set of strongly supervised data (i.e., fully annotated hand segmentation maps) to achieve state-of-the-art hand segmentation performance. Secondly, the resulting high-quality hand segmentation maps are further paired with the corresponding motion maps and object feature maps, in order to explore the contextual information among object, motion and hand to generate interactional foreground regions (operated objects). The resulting interactional target maps (hand + active object) from our cascaded DCNN are further utilized to form discriminative action representation. Experiments show that our framework has achieved the state-of-the-art egocentric action recognition performance on the benchmark dataset Activities of Daily Living (ADL). | ['Yang Zhou', 'Richang Hong', 'Bingbing Ni', 'Qi Tian', 'Xiaokang Yang'] | 2016-06-01 | null | null | null | cvpr-2016-6 | ['foreground-segmentation', 'hand-segmentation'] | ['computer-vision', 'computer-vision'] | [ 4.14280713e-01 9.26983356e-02 -4.00980771e-01 -3.05806965e-01
-5.72278261e-01 -4.63854074e-01 5.71898639e-01 -7.28407443e-01
-4.48352516e-01 6.96356475e-01 4.41419333e-01 7.49353915e-02
1.64751619e-01 -5.79167843e-01 -8.49777818e-01 -8.91329050e-01
2.65024245e-01 4.35422301e-01 5.74927092e-01 1.12999314e-02
1.29514024e-01 5.98848224e-01 -1.57757020e+00 3.92362833e-01
7.19199955e-01 8.05819571e-01 2.93357760e-01 7.48641849e-01
1.34376198e-01 1.06863189e+00 -3.81643295e-01 9.06059891e-03
2.26963475e-01 -4.34415460e-01 -1.20337939e+00 3.60550791e-01
2.50546217e-01 -8.51651549e-01 -5.47426343e-01 7.56081343e-01
7.43684590e-01 3.79545540e-01 5.85924447e-01 -1.17969334e+00
-2.28302151e-01 2.89187789e-01 -7.18918920e-01 3.54223847e-01
3.05037618e-01 7.13532865e-01 5.82469642e-01 -7.18990266e-01
7.98453629e-01 1.27982116e+00 6.23159744e-02 6.56361341e-01
-9.29880381e-01 -5.12567282e-01 5.26053250e-01 3.69309366e-01
-1.33190989e+00 -3.57447475e-01 9.68603671e-01 -6.30451977e-01
9.84240592e-01 1.31308556e-01 8.84857178e-01 1.47067225e+00
2.07171310e-03 1.56293213e+00 7.24027932e-01 -2.65012652e-01
2.08467051e-01 -2.17597440e-01 -1.53974993e-02 8.17343533e-01
-1.29268467e-01 -1.85057342e-01 -3.66060555e-01 3.56147051e-01
1.22999179e+00 1.88535303e-01 -2.33490393e-01 -6.28727794e-01
-1.44827080e+00 2.57663101e-01 6.00375473e-01 3.92089665e-01
-7.59477794e-01 3.32885623e-01 3.76998454e-01 -4.48056817e-01
2.52208054e-01 -2.15921059e-01 -6.11232758e-01 -3.15180480e-01
-9.00111198e-01 3.05144757e-01 4.01389599e-01 9.84976053e-01
5.57449162e-01 -1.23860694e-01 -7.86282182e-01 4.39162582e-01
2.57143140e-01 3.78352761e-01 1.92195177e-01 -9.22085702e-01
8.78813446e-01 8.45908105e-01 3.28406870e-01 -6.66885018e-01
-5.91932416e-01 -2.88657367e-01 -4.81087774e-01 2.77440399e-01
8.38444233e-01 -1.00964852e-01 -1.20064342e+00 1.65994358e+00
6.73531473e-01 -4.70603667e-02 -1.54194653e-01 1.31297922e+00
6.14991248e-01 3.60562891e-01 4.32521820e-01 1.06065735e-01
1.39490354e+00 -1.43366480e+00 -6.24326348e-01 -3.80495220e-01
6.78406119e-01 -1.52171910e-01 1.20740545e+00 2.02827901e-01
-1.05580747e+00 -8.08937550e-01 -6.53228104e-01 -2.05295891e-01
-2.91856408e-01 4.06563908e-01 6.87962890e-01 4.51181114e-01
-5.82038283e-01 3.17154437e-01 -1.17061508e+00 -2.65207738e-01
9.43950117e-01 3.13261956e-01 -4.21627253e-01 6.91613108e-02
-9.52330291e-01 6.13901377e-01 6.28522336e-01 3.59788209e-01
-1.15019178e+00 -4.48475957e-01 -9.00632858e-01 -9.59267467e-02
7.88889050e-01 -6.81371272e-01 1.00796700e+00 -1.06870639e+00
-1.31927526e+00 8.36717606e-01 -1.93623647e-01 7.12494254e-02
9.92111742e-01 -4.21394944e-01 1.19198775e-02 2.77438045e-01
1.91432461e-01 9.26367998e-01 7.80584395e-01 -1.26135921e+00
-8.77397954e-01 -7.74509370e-01 1.58745512e-01 3.76054049e-01
-9.28815380e-02 3.49189527e-02 -8.51058125e-01 -7.45827675e-01
-3.19999573e-03 -8.69285405e-01 9.74223912e-02 -3.00607476e-02
-6.08691335e-01 -5.88164389e-01 1.05637634e+00 -1.16073215e+00
1.20335031e+00 -1.85456848e+00 6.61793888e-01 3.29284742e-02
1.33673802e-01 2.62945801e-01 1.30746858e-02 -1.88730776e-01
-1.29077792e-01 -3.79598558e-01 -3.13010037e-01 -2.49614194e-01
-1.14240611e-04 2.35186160e-01 1.30918816e-01 5.19590020e-01
2.61796057e-01 1.36944699e+00 -1.18795288e+00 -7.81458914e-01
6.26596034e-01 5.34219146e-01 -4.20124650e-01 3.03352088e-01
-4.19956565e-01 1.08875799e+00 -6.37917101e-01 9.62928534e-01
3.56847227e-01 -8.82672742e-02 -1.77100897e-02 -3.75177205e-01
1.21912904e-01 -7.83055127e-02 -1.28677261e+00 2.00586414e+00
-2.82565832e-01 2.75663972e-01 -6.95473105e-02 -7.83587575e-01
2.46901989e-01 4.05239940e-01 7.21006513e-01 -5.33531189e-01
4.02192473e-01 -1.42679393e-01 -1.00680320e-02 -7.50989616e-01
-5.12580909e-02 2.42708817e-01 1.62068121e-02 6.05079591e-01
1.92415714e-01 3.59944403e-01 2.51932610e-02 -1.45357355e-01
8.93969715e-01 9.41407442e-01 1.18239522e-01 -5.42854033e-02
6.68764472e-01 -6.01808541e-03 5.88428617e-01 4.75938320e-01
-5.96506059e-01 6.12971544e-01 4.06097889e-01 -5.00897348e-01
-1.08648789e+00 -8.63425434e-01 2.55141139e-01 1.36674190e+00
2.58624852e-01 1.05954118e-01 -1.21428192e+00 -1.17727602e+00
-2.30010360e-01 5.37121832e-01 -8.02139878e-01 6.19494058e-02
-8.86784196e-01 -3.74662817e-01 5.14800489e-01 1.44044125e+00
9.40659583e-01 -1.72470808e+00 -8.69066119e-01 1.13373555e-01
-3.51064742e-01 -1.01032436e+00 -6.44973576e-01 -5.23778945e-02
-4.81969625e-01 -1.24841738e+00 -1.17182648e+00 -7.01576889e-01
6.87899888e-01 3.78369167e-02 7.42436826e-01 -3.38457942e-01
-4.85505849e-01 5.54050803e-01 -2.24338606e-01 -9.05978829e-02
1.88165501e-01 4.23519611e-02 6.35950714e-02 2.98945308e-01
4.45225924e-01 -5.47570527e-01 -1.04285014e+00 4.24796849e-01
-6.50201857e-01 2.84599960e-01 8.29820156e-01 6.49142385e-01
7.26770401e-01 -8.31348822e-02 2.88299739e-01 -4.37561184e-01
1.64771035e-01 -7.48807639e-02 -2.29433939e-01 4.17006493e-01
9.60276201e-02 -2.31156293e-02 1.07769154e-01 -5.97596645e-01
-1.54444897e+00 6.91949427e-01 -4.15936522e-02 -4.97193754e-01
-5.55320561e-01 -1.73208535e-01 -7.65363216e-01 2.31886983e-01
4.71601665e-01 4.11161363e-01 -2.93988496e-01 -4.10684824e-01
6.02625370e-01 8.60703826e-01 8.35111916e-01 -6.27994895e-01
3.55443120e-01 7.48954177e-01 -2.48417795e-01 -4.74080086e-01
-9.44201052e-01 -5.89933395e-01 -1.48622167e+00 -5.57009041e-01
1.54595327e+00 -7.23800480e-01 -6.58372641e-01 7.56183922e-01
-1.20975554e+00 -8.79864335e-01 -1.78927079e-01 3.84202480e-01
-9.08529520e-01 2.98185050e-01 -3.20174754e-01 -9.64657903e-01
-3.10008019e-01 -1.25143218e+00 1.53646660e+00 1.87505186e-01
-3.19301009e-01 -6.87705100e-01 -2.03394979e-01 8.14786673e-01
-1.68997660e-01 4.68875557e-01 4.80888963e-01 -3.54757071e-01
-7.42860675e-01 -2.60673732e-01 -4.09276843e-01 3.30749333e-01
2.46782690e-01 -3.97040427e-01 -1.05242741e+00 -1.08902767e-01
-3.16212475e-01 -2.97209024e-01 6.66778028e-01 4.55522686e-01
1.40287971e+00 -2.16482192e-01 -5.71522415e-01 5.16885757e-01
6.14512444e-01 1.83988720e-01 7.12326586e-01 1.27344236e-01
1.29908073e+00 7.13886797e-01 7.86944509e-01 5.30581653e-01
3.14301401e-01 9.10627782e-01 3.12005192e-01 -1.13203920e-01
-2.75999606e-01 -1.31364599e-01 1.17713772e-01 1.21354563e-02
-6.94605350e-01 2.13234536e-02 -9.60995257e-01 5.81118345e-01
-2.16241384e+00 -1.06006598e+00 -8.83991942e-02 1.79332113e+00
6.66110218e-01 1.26438841e-01 6.09514654e-01 1.64791346e-01
7.56692290e-01 1.92956224e-01 -9.92133498e-01 3.70347023e-01
1.74435541e-01 1.05372861e-01 2.63609707e-01 2.01393425e-01
-1.56597245e+00 1.23954725e+00 4.69401026e+00 6.97809577e-01
-8.44597697e-01 2.68990189e-01 3.77979428e-01 -1.88900337e-01
3.90242785e-01 -3.75720441e-01 -6.54746592e-01 5.73944092e-01
2.02598572e-01 5.23722708e-01 4.13432926e-01 9.78980422e-01
3.95005971e-01 -1.32479459e-01 -1.18188834e+00 1.13075018e+00
7.26462901e-02 -7.28549600e-01 -3.87291275e-02 1.02725267e-01
7.73806810e-01 -3.60122621e-01 -2.68642068e-01 3.85735840e-01
1.07729807e-01 -9.24900234e-01 9.35608685e-01 7.76195943e-01
6.98790789e-01 -7.39824176e-01 6.39904022e-01 6.31804764e-01
-1.44916129e+00 -3.08191419e-01 1.53151423e-01 3.56573053e-02
3.43996793e-01 1.29824895e-02 -3.48869532e-01 4.24171090e-01
8.44225943e-01 7.10521042e-01 -5.46310544e-01 4.96615261e-01
-5.31325221e-01 2.73251265e-01 -2.08436340e-01 2.12545097e-01
3.85962725e-01 7.90570378e-02 4.14852202e-01 1.14069772e+00
-3.17256987e-01 4.00058836e-01 1.71383694e-01 1.20745420e+00
1.91231027e-01 -2.95498639e-01 -2.34803751e-01 8.64832837e-04
-2.22083032e-02 1.19080102e+00 -9.42778349e-01 -5.13640404e-01
-3.27233493e-01 1.60377443e+00 3.01471531e-01 5.52798808e-01
-9.18798506e-01 -3.05119008e-01 6.71364605e-01 1.60031542e-01
3.71142179e-01 -1.64095178e-01 -2.09520578e-01 -1.13132298e+00
3.05786759e-01 -5.77732146e-01 4.00861084e-01 -8.81176829e-01
-8.88107181e-01 1.43565074e-01 5.83360679e-02 -1.00013053e+00
-3.32372278e-01 -7.07107961e-01 -5.48498273e-01 9.17572737e-01
-9.36509609e-01 -1.59649432e+00 -9.99643743e-01 9.29675937e-01
7.79029012e-01 4.21872139e-02 3.47192824e-01 2.50555962e-01
-8.03270519e-01 4.50866550e-01 -5.29510856e-01 5.78893960e-01
3.88693094e-01 -1.23722982e+00 2.95967370e-01 7.30670094e-01
-7.78592750e-02 3.32237214e-01 1.47055224e-01 -8.95255983e-01
-1.23730350e+00 -1.26275766e+00 2.54342109e-01 -1.11094975e+00
1.81678310e-01 -3.72175664e-01 -5.88259876e-01 9.04633343e-01
-3.70529652e-01 2.25407317e-01 1.97185069e-01 -3.21593344e-01
8.97244513e-02 2.18383953e-01 -1.03195262e+00 7.31717467e-01
1.78626490e+00 -5.60966730e-01 -8.00352752e-01 4.82554555e-01
3.87278914e-01 -5.30537665e-01 -5.51373661e-01 4.46775943e-01
8.42796743e-01 -8.72975528e-01 1.17800319e+00 -8.77728939e-01
4.23683435e-01 -4.00559992e-01 2.23489366e-02 -6.77049279e-01
-2.04451442e-01 -2.65117705e-01 -6.00221515e-01 1.04760337e+00
-2.21856192e-01 -1.91727027e-01 1.11304295e+00 6.70068026e-01
-5.39023094e-02 -8.97163391e-01 -7.31983900e-01 -6.59313202e-01
-2.26700827e-01 -4.29761857e-01 3.64890277e-01 6.99041963e-01
-2.22908296e-02 5.21129072e-02 -1.53197676e-01 2.76311077e-02
6.54553354e-01 -9.43491608e-02 1.03568292e+00 -8.88310075e-01
-7.97482952e-02 -4.42335010e-01 -5.72749138e-01 -1.29708481e+00
3.64078552e-01 -7.06850052e-01 1.23983845e-01 -1.77277589e+00
4.11746204e-01 -1.98298201e-01 -3.14111978e-01 6.71304286e-01
-4.67824548e-01 3.46124291e-01 -1.73848984e-03 1.89965889e-01
-6.45999253e-01 4.27432507e-01 1.76165736e+00 -2.31101841e-01
-6.06936693e-01 1.87415197e-01 -2.09099382e-01 8.08711648e-01
5.28706849e-01 -8.20088834e-02 -3.33109915e-01 -1.34201959e-01
-4.47092533e-01 -1.72254667e-02 9.02219474e-01 -1.09204352e+00
1.88173160e-01 -2.52640426e-01 7.34432399e-01 -9.21101570e-01
3.29292059e-01 -7.75960028e-01 -1.46756113e-01 4.42759395e-01
-2.21468627e-01 -5.55678606e-01 -9.21834726e-03 6.49577439e-01
9.05570239e-02 1.46427512e-01 6.11171126e-01 -3.58923227e-01
-1.06817985e+00 4.18538272e-01 3.45300883e-02 -1.27675943e-03
1.40382016e+00 -5.99117637e-01 1.69360593e-01 -1.64640978e-01
-1.06653404e+00 3.07429761e-01 1.99998647e-01 5.56470573e-01
3.48343879e-01 -1.25789011e+00 -4.37113136e-01 2.33705059e-01
1.33078501e-01 5.03416359e-01 5.02044737e-01 1.06423664e+00
-3.37635934e-01 5.68234384e-01 -4.47052538e-01 -8.00854683e-01
-1.22891796e+00 6.61232293e-01 4.88908857e-01 -9.47682634e-02
-9.74219799e-01 9.44134057e-01 5.00124812e-01 -4.00898486e-01
4.85489696e-01 -3.84384602e-01 -4.38437387e-02 7.77687132e-02
5.75279236e-01 6.82910919e-01 -2.03598753e-01 -7.54861832e-01
-5.35796881e-01 6.24953568e-01 2.42114782e-01 -1.01111956e-01
1.17647576e+00 1.08020686e-01 5.44943325e-02 1.16657779e-01
1.01172006e+00 -4.87348199e-01 -1.94093680e+00 -1.28151089e-01
-2.39038631e-01 -5.19022405e-01 -2.02727132e-02 -8.99911702e-01
-1.16440940e+00 1.13031816e+00 7.25633383e-01 -5.14884949e-01
9.79444325e-01 4.23716724e-01 7.94533670e-01 4.16959614e-01
4.42705125e-01 -1.41275764e+00 4.44991380e-01 1.56990394e-01
8.85613859e-01 -1.25803876e+00 -1.13532647e-01 -3.32413018e-01
-8.52502108e-01 9.34179246e-01 9.90614891e-01 1.76366195e-01
3.11808467e-01 6.52866438e-02 -1.44234225e-01 -2.48707145e-01
1.67166114e-01 -5.05091429e-01 5.24623275e-01 6.26937211e-01
-4.16940451e-03 9.44046304e-02 4.63806093e-02 8.90730321e-01
1.97776347e-01 2.79974431e-01 -2.79764503e-01 1.33964884e+00
-2.48880297e-01 -6.42232716e-01 -3.74060661e-01 4.59746659e-01
-1.07515313e-01 3.47270012e-01 -5.80237150e-01 8.33067954e-01
4.90686119e-01 5.76981187e-01 1.07345268e-01 -3.16334844e-01
4.98623312e-01 2.83337235e-01 8.73639882e-01 -5.75273573e-01
-3.28700393e-01 2.86039268e-03 -2.65443385e-01 -1.03659081e+00
-5.96311092e-01 -6.48751855e-01 -1.44102633e+00 -2.59201061e-02
-1.59487650e-01 -6.22218966e-01 2.85779983e-01 1.41225159e+00
1.10654041e-01 9.04157698e-01 2.59426180e-02 -1.67861223e+00
-2.98812985e-01 -1.05617595e+00 -5.71651161e-01 5.89393079e-01
1.16230674e-01 -1.08177507e+00 -6.28523380e-02 3.29189718e-01] | [7.70041561126709, 0.1854279637336731] |
a0ee90fa-4e3e-4754-94f3-094a7f46c194 | the-foes-of-neural-networks-data-efficiency | null | null | https://openreview.net/forum?id=X6_vet6HWX | https://openreview.net/pdf?id=X6_vet6HWX | The Foes of Neural Network’s Data Efficiency Among Unnecessary Input Dimensions | Input dimensions are unnecessary for a given task when the target function can be expressed without such dimensions. Object's background in image recognition or redundant sentences in text classification are examples of unnecessary dimensions that are often present in datasets. Deep neural networks achieve remarkable generalization performance despite the presence of unnecessary dimensions but it is unclear whether these dimensions negatively affect neural networks or how. In this paper, we investigate the impact of unnecessary input dimensions on one of the central issues of machine learning: the number of training examples needed to achieve high generalization performance, which we refer to as the network's data efficiency. In a series of analyses with multi-layer perceptrons and deep convolutional neural networks, we show that the network's data efficiency depends on whether the unnecessary dimensions are \emph{task-unrelated} or \emph{task-related} (unnecessary due to redundancy). Namely, we demonstrate that increasing the number of \emph{task-unrelated} dimensions leads to an incorrect inductive bias and as a result degrade the data efficiency, while increasing the number of \emph{task-related} dimensions helps to alleviate the negative impact of the \emph{task-unrelated} dimensions. These results highlight the need for mechanisms that remove \emph{task-unrelated} dimensions, such as crops or foveation for image classification, to enable data efficiency gains.
| ['Xavier Boix', 'Tomotake Sasaki', 'Sanjana Srivastava', "Vanessa D'Amario"] | 2021-01-01 | null | null | null | null | ['foveation'] | ['computer-vision'] | [ 6.26219213e-01 7.91244060e-02 1.11937739e-01 -4.78599161e-01
-4.85967584e-02 -4.69693184e-01 3.10079962e-01 1.79997504e-01
-7.73351371e-01 3.53532016e-01 2.30137438e-01 -6.77476525e-01
-5.67457974e-01 -7.36604810e-01 -7.67286599e-01 -5.75782597e-01
3.52031112e-01 -1.67197242e-01 -1.54084802e-01 -3.89384687e-01
4.98125702e-01 6.04593873e-01 -1.91631043e+00 4.00777698e-01
6.33101285e-01 9.29842234e-01 4.05922949e-01 4.53029931e-01
-3.29360515e-01 7.44585812e-01 -9.22120035e-01 -1.90071583e-01
2.84805030e-01 -3.29843491e-01 -6.20117068e-01 -5.05736582e-02
3.46887946e-01 -3.57772410e-01 -6.82189735e-03 9.81060326e-01
6.14014924e-01 -4.85573076e-02 7.47126400e-01 -9.61196780e-01
-8.41826737e-01 5.11672735e-01 -3.04633498e-01 6.54953599e-01
-3.39844882e-01 1.16715774e-01 1.03131127e+00 -7.75574267e-01
2.97767252e-01 1.09963894e+00 7.03577340e-01 5.28904140e-01
-1.41819417e+00 -4.99888718e-01 3.11694205e-01 -4.83913153e-01
-9.75990117e-01 -5.97755015e-01 9.69195902e-01 -6.01891994e-01
1.06583357e+00 5.02999902e-01 3.39601815e-01 1.04370177e+00
2.39955738e-01 4.36921984e-01 9.16203678e-01 -5.52666783e-01
1.69188410e-01 5.19252419e-01 5.40178776e-01 1.53786898e-01
6.23160422e-01 7.05073178e-02 -3.12432349e-01 1.59440473e-01
6.53171241e-01 -1.40770942e-01 -2.41235182e-01 1.14448210e-02
-7.87909687e-01 8.95312726e-01 4.07156169e-01 6.22859538e-01
-3.34433883e-01 8.68236497e-02 4.29249167e-01 5.22690475e-01
4.08212632e-01 1.14614201e+00 -7.20776677e-01 4.36486863e-02
-4.13671434e-01 1.38078645e-01 4.64382380e-01 8.08191597e-01
6.45782471e-01 2.85672575e-01 -2.10055932e-01 1.11378264e+00
-1.80879578e-01 3.98927480e-01 6.57935560e-01 -8.20309818e-01
6.38731778e-01 9.99470353e-01 -1.86203703e-01 -1.19752264e+00
-8.27296138e-01 -9.41961706e-01 -7.96163142e-01 1.28220078e-02
5.73274672e-01 -2.80108720e-01 -5.68140864e-01 2.01436734e+00
-2.29664013e-01 -1.06223726e+00 2.08268166e-01 8.85858595e-01
7.15586245e-01 2.67519355e-01 2.50637293e-01 -1.55834824e-01
1.30579674e+00 -3.91010076e-01 -6.11183882e-01 -8.71808171e-01
1.17764056e+00 -7.35230744e-01 1.62140238e+00 3.68378237e-02
-8.80021691e-01 -5.74918568e-01 -1.03529370e+00 -1.75501049e-01
-6.80857718e-01 3.35164219e-01 5.31283915e-01 7.28532135e-01
-7.44731545e-01 7.06365645e-01 -4.92776819e-02 -8.20666403e-02
3.94105732e-01 5.38374960e-01 -1.90184832e-01 2.24645570e-01
-1.19512892e+00 9.72427130e-01 4.04248595e-01 -2.71117017e-02
1.96519103e-02 -6.92929089e-01 -6.54242098e-01 4.27046895e-01
8.31705555e-02 -4.67559129e-01 9.22187388e-01 -1.44310427e+00
-7.02165365e-01 6.46971047e-01 1.30628422e-01 -2.34959245e-01
1.41687915e-01 -8.67255628e-02 -2.99533218e-01 -2.06688523e-01
-2.13367231e-02 8.63039434e-01 8.26669395e-01 -1.06623411e+00
-3.79786700e-01 -5.89862823e-01 1.33148894e-01 2.59602308e-01
-1.21404266e+00 -2.61348665e-01 7.59742036e-02 -8.01019490e-01
3.14146072e-01 -9.80056643e-01 1.43455535e-01 -1.84853137e-01
-2.11600631e-01 -2.64798075e-01 7.81578481e-01 -4.20178801e-01
1.32409692e+00 -2.40779710e+00 -1.16899855e-01 -6.85121641e-02
3.14535648e-01 2.31565744e-01 -2.70775408e-01 1.12205707e-01
-2.99620271e-01 4.36680615e-01 -4.35770005e-02 7.30553344e-02
-1.05769873e-01 2.19502375e-01 -5.69101125e-02 1.08109295e-01
5.85618138e-01 7.04222023e-01 -4.85098839e-01 -5.59823066e-02
6.01265393e-02 4.15868014e-01 -6.41568303e-01 -3.37244660e-01
-1.75716057e-01 1.10950537e-01 -3.19920272e-01 1.44952491e-01
4.82051581e-01 -3.76362205e-01 2.55341772e-02 -3.46464694e-01
-2.04089716e-01 5.88960290e-01 -7.51219988e-01 9.77201581e-01
-4.70884293e-01 9.71072733e-01 -3.22937816e-01 -1.02838886e+00
1.03973961e+00 6.43136129e-02 1.97375402e-01 -1.27587342e+00
2.85354942e-01 2.75899023e-01 6.02308810e-01 -5.20435452e-01
6.78727090e-01 -3.02149504e-01 -5.45488624e-03 6.79109931e-01
-2.54372746e-01 2.26035714e-02 2.81173617e-01 -1.52330920e-01
1.00582981e+00 -4.12247211e-01 -1.09296292e-01 -6.44639969e-01
2.42098317e-01 -3.59466374e-02 4.86605793e-01 7.74481118e-01
-7.68377408e-02 2.54995942e-01 8.37388515e-01 -5.78097522e-01
-1.44335115e+00 -4.60403264e-01 -2.65510887e-01 1.46276963e+00
-2.64061153e-01 -2.63566464e-01 -5.36167026e-01 -4.66383666e-01
1.55085355e-01 1.04100204e+00 -8.23366642e-01 -5.73401809e-01
-4.02312636e-01 -9.35282588e-01 6.46263421e-01 7.40265429e-01
3.99035901e-01 -9.61616576e-01 -1.04514742e+00 -3.41887660e-02
3.05853412e-02 -8.78257692e-01 -1.33031547e-01 7.23821163e-01
-1.01796961e+00 -6.58621550e-01 -6.32724583e-01 -4.94659305e-01
1.04424918e+00 3.96312445e-01 1.06657565e+00 2.25371078e-01
-1.14738815e-01 7.87768140e-02 -2.33364403e-01 -9.37351882e-01
-4.32522178e-01 2.35117704e-01 -2.40797713e-01 -5.44052005e-01
5.87779403e-01 -4.88867581e-01 -5.40406406e-01 3.03647190e-01
-1.07621312e+00 2.22062379e-01 8.74854386e-01 8.47823977e-01
2.15289280e-01 6.37949556e-02 9.69877362e-01 -9.02093172e-01
1.06460083e+00 -2.01655328e-01 -2.90170789e-01 -5.40606529e-02
-6.82648659e-01 4.34501767e-01 7.46557415e-01 -6.19422376e-01
-7.41890669e-01 -2.78451085e-01 -4.85525131e-02 -1.54432043e-01
-2.30039749e-02 4.74116683e-01 3.51029448e-02 2.27300152e-01
1.16833270e+00 1.84426382e-01 5.58505245e-02 -3.15310776e-01
-5.17173484e-02 7.27291524e-01 -2.67842144e-01 -2.47109652e-01
1.64208323e-01 1.70190796e-01 1.35240098e-02 -9.84430254e-01
-9.52978432e-01 -1.33302629e-01 -3.66240203e-01 -1.06979012e-01
6.29950821e-01 -5.79122245e-01 -5.82106948e-01 1.19382434e-01
-1.01919901e+00 -2.63964444e-01 -3.48382860e-01 5.32801986e-01
-2.04292998e-01 -1.12341836e-01 -2.69577026e-01 -7.65606999e-01
-4.33345914e-01 -1.25423253e+00 7.58213460e-01 1.45217955e-01
-6.19910002e-01 -7.26628602e-01 -7.61231840e-01 2.74470508e-01
6.61267698e-01 -1.86511859e-01 1.64730811e+00 -9.22734141e-01
-2.75767036e-02 -2.18888879e-01 -5.13782144e-01 8.26297462e-01
3.02523226e-01 -3.20085883e-01 -1.09503889e+00 -5.86316362e-02
2.34668985e-01 -1.84335679e-01 8.79845023e-01 3.65350634e-01
1.21378922e+00 -6.06507242e-01 -1.73456836e-02 1.59866363e-01
1.10583651e+00 3.17145586e-01 3.26342225e-01 3.62959981e-01
3.62890929e-01 9.31118667e-01 1.60547122e-01 3.07696998e-01
5.87112084e-02 2.09249958e-01 2.24899158e-01 -1.97334170e-01
1.05159804e-02 2.27912813e-01 -4.91320491e-02 4.50668544e-01
3.65640163e-01 -2.11583212e-01 -9.64433432e-01 5.21696806e-01
-1.28877723e+00 -6.30370498e-01 -1.59985945e-01 2.16097140e+00
8.00304949e-01 4.66425717e-01 -1.90847024e-01 6.07061446e-01
5.75638950e-01 6.65105088e-03 -7.25047588e-01 -8.97439659e-01
-4.13225472e-01 -1.37018785e-01 4.91343975e-01 1.30130216e-01
-7.86691189e-01 5.41904628e-01 5.57153034e+00 3.85805130e-01
-1.32961023e+00 -2.50236332e-01 7.73316324e-01 -2.56196350e-01
-4.93072569e-01 -3.19027364e-01 -7.68172264e-01 4.03519511e-01
8.46683860e-01 -8.16761032e-02 2.39103675e-01 8.50614846e-01
1.58787087e-01 -2.21449763e-01 -1.39789927e+00 6.41692102e-01
1.55961709e-02 -1.19098699e+00 1.85210049e-01 1.77646995e-01
4.23851848e-01 -2.98296481e-01 3.99445891e-01 2.95334876e-01
-1.23759486e-01 -9.07548547e-01 5.93851507e-01 1.39918879e-01
4.98643547e-01 -6.52973115e-01 6.66117549e-01 5.11785805e-01
-4.25302416e-01 -4.94325697e-01 -4.94591951e-01 -4.68704998e-01
-5.08840919e-01 9.03088391e-01 -7.21979320e-01 -6.95381984e-02
7.53961504e-01 1.28220022e-01 -7.63593137e-01 3.62705410e-01
1.33683369e-01 2.41282687e-01 -2.37974092e-01 -3.17855239e-01
1.23728804e-01 1.44161135e-01 3.27245295e-01 1.03037488e+00
2.18434468e-01 7.20133260e-03 -6.66334510e-01 9.38781679e-01
-1.50819495e-01 9.67907086e-02 -8.60489309e-01 -4.72188622e-01
5.31072974e-01 7.81967342e-01 -7.43118286e-01 -4.17640582e-02
-3.75036031e-01 6.06813490e-01 3.58615667e-01 3.77504766e-01
-3.87641221e-01 -3.61157894e-01 4.01083499e-01 1.65604115e-01
8.53341147e-02 -1.77512281e-02 -1.01154947e+00 -5.47491729e-01
3.10502648e-01 -6.79621160e-01 8.62918571e-02 -7.32949913e-01
-1.17718041e+00 3.71542931e-01 -2.99858540e-01 -9.13750231e-01
-2.69509032e-02 -7.24654317e-01 -1.31957605e-01 1.03486395e+00
-1.24628866e+00 -6.45309687e-01 -2.17521191e-01 2.63935506e-01
4.67494607e-01 -1.59450278e-01 9.21732366e-01 3.21716636e-01
-4.86051202e-01 6.79828703e-01 1.93278909e-01 2.64606893e-01
6.17574632e-01 -9.10697281e-01 1.94726840e-01 4.64220017e-01
-1.34531006e-01 9.12153125e-01 7.24780083e-01 -5.05762458e-01
-1.24021900e+00 -9.44799185e-01 1.18256795e+00 -3.25288981e-01
1.52881205e-01 -3.49185318e-01 -7.81355977e-01 5.65342784e-01
-3.85975659e-01 -3.31247449e-01 9.16967094e-01 3.04879099e-01
-5.29743791e-01 -2.60698318e-01 -9.25670922e-01 9.84845221e-01
1.00596201e+00 -3.52300853e-01 -5.73085487e-01 2.91368723e-01
8.17014813e-01 9.51388329e-02 -7.16522574e-01 6.01930320e-01
5.68990648e-01 -1.11448455e+00 7.26577163e-01 -6.20573282e-01
7.73077190e-01 2.39351287e-01 -3.51624727e-01 -1.13213396e+00
-4.36168462e-01 6.82378784e-02 3.16723973e-01 9.57716644e-01
8.93471479e-01 -7.02854514e-01 5.22498846e-01 8.48698914e-01
-1.89415291e-01 -7.57674277e-01 -7.49312699e-01 -7.31258869e-01
4.66025770e-01 -3.57948542e-01 4.72566873e-01 1.01641572e+00
-3.07836771e-01 7.20117867e-01 1.67517424e-01 -2.06872046e-01
1.75647452e-01 -1.69276506e-01 3.95695657e-01 -1.33850229e+00
2.12953053e-02 -7.14139044e-01 -1.27878100e-01 -8.10025334e-01
-6.80497065e-02 -6.05423689e-01 -2.27632597e-01 -1.40878451e+00
-1.07333049e-01 -7.98498631e-01 -1.87517911e-01 5.57990074e-01
-8.66155699e-02 -3.39478441e-02 3.61391246e-01 3.00688535e-01
-3.36102806e-02 3.06345910e-01 1.20055473e+00 -6.31021261e-02
-4.23394114e-01 3.38358358e-02 -1.13528156e+00 7.55282462e-01
9.17458773e-01 -4.32415724e-01 -6.89703941e-01 -8.60423803e-01
6.50835931e-01 -3.54841173e-01 2.33437493e-01 -9.07964349e-01
4.40911204e-02 -3.98350619e-02 8.44031334e-01 -3.16043079e-01
5.34496844e-01 -9.95495498e-01 -1.80971637e-01 5.66746950e-01
-7.51589417e-01 3.23178530e-01 6.58811390e-01 1.50633812e-01
9.48231444e-02 -4.53178972e-01 5.80926597e-01 -1.19054213e-01
-3.05834442e-01 -5.75397015e-01 -4.54237163e-01 8.70198831e-02
8.29071939e-01 -3.33443135e-01 -7.14354932e-01 -9.58433822e-02
-4.91813600e-01 -1.29940480e-01 2.49523565e-01 7.04731226e-01
3.69214743e-01 -9.71293807e-01 -5.24514139e-01 3.55359226e-01
1.56584606e-01 -2.82773972e-01 2.39397943e-01 8.37389588e-01
-2.23710671e-01 7.14878082e-01 -2.13482291e-01 -5.20277321e-01
-1.17213750e+00 5.29581547e-01 3.38730246e-01 -5.84307406e-03
-2.33046830e-01 8.67387354e-01 4.16288376e-01 -2.56941348e-01
3.83286893e-01 -4.44406569e-01 -1.60413370e-01 3.27750117e-01
3.17291588e-01 1.98525921e-01 2.92256802e-01 -2.24799186e-01
-2.46449426e-01 1.47108749e-01 -4.25643861e-01 1.21413097e-01
1.26826334e+00 -6.71134517e-02 9.82920229e-02 6.71708524e-01
1.10480142e+00 -4.96865273e-01 -1.02289963e+00 -9.28668603e-02
-4.40633073e-02 -2.97844797e-01 1.68537408e-01 -9.24933136e-01
-7.45240867e-01 1.08599925e+00 5.72092116e-01 4.43892390e-01
1.26749313e+00 -2.04806298e-01 3.04414451e-01 7.06617355e-01
-1.97249800e-01 -1.48920202e+00 3.70510131e-01 5.34246802e-01
9.84841347e-01 -1.03644228e+00 3.27060819e-02 -2.93335348e-01
-6.20997131e-01 9.61823165e-01 7.82222331e-01 2.36902013e-01
4.49744523e-01 2.38803774e-01 -1.53863266e-01 -2.28894234e-01
-7.08289266e-01 -9.79244709e-02 6.79368004e-02 4.18466866e-01
6.59738243e-01 -1.12460323e-01 -5.99056780e-01 8.24933946e-01
-4.50235814e-01 -6.38220757e-02 3.96587670e-01 1.12944055e+00
-4.61092621e-01 -6.54653549e-01 -2.38275513e-01 1.13470352e+00
-3.90835732e-01 -3.28852743e-01 -5.26454747e-01 8.66500080e-01
2.69664973e-01 9.37784493e-01 4.41217840e-01 -2.88185179e-01
3.64070922e-01 3.15996021e-01 3.44093978e-01 -6.31524324e-01
-8.03427219e-01 -1.15061663e-01 1.88368112e-01 6.38918877e-02
-6.19985946e-02 -3.06087792e-01 -1.14969635e+00 -3.25339735e-01
-5.33283234e-01 -3.20269972e-01 8.65662992e-01 1.04081762e+00
5.61501324e-01 7.77834952e-01 3.75350893e-01 -2.56318241e-01
-7.56892920e-01 -1.04916513e+00 -3.22072178e-01 4.89163816e-01
3.01857799e-01 -4.99597669e-01 -4.96860981e-01 -9.41832457e-03] | [8.646174430847168, 3.6194229125976562] |
c0a30995-ba73-43e6-b26d-983e148d1a4e | task-planning-in-robotics-an-empirical | 1804.08229 | null | http://arxiv.org/abs/1804.08229v3 | http://arxiv.org/pdf/1804.08229v3.pdf | Task Planning in Robotics: an Empirical Comparison of PDDL-based and ASP-based Systems | Robots need task planning algorithms to sequence actions toward accomplishing
goals that are impossible through individual actions. Off-the-shelf task
planners can be used by intelligent robotics practitioners to solve a variety
of planning problems. However, many different planners exist, each with
different strengths and weaknesses, and there are no general rules for which
planner would be best to apply to a given problem.
In this article, we empirically compare the performance of state-of-the-art
planners that use either the Planning Domain Description Language (PDDL), or
Answer Set Programming (ASP) as the underlying action language. PDDL is
designed for task planning, and PDDL-based planners are widely used for a
variety of planning problems. ASP is designed for knowledge-intensive
reasoning, but can also be used for solving task planning problems. Given
domain encodings that are as similar as possible, we find that PDDL-based
planners perform better on problems with longer solutions, and ASP-based
planners are better on tasks with a large number of objects or in which complex
reasoning is required to reason about action preconditions and effects. The
resulting analysis can inform selection among general purpose planning systems
for particular robot task planning domains. | ['Piyush Khandelwal', 'Yuqian Jiang', 'Shiqi Zhang', 'Peter Stone'] | 2018-04-23 | null | null | null | null | ['robot-task-planning'] | ['robots'] | [ 2.76734203e-01 6.45879626e-01 -3.48685950e-01 -3.31459165e-01
-4.15413558e-01 -6.15593910e-01 5.17017245e-01 8.15534294e-02
-2.19443634e-01 8.24475765e-01 4.90974724e-01 -4.47837144e-01
-6.56957567e-01 -8.33685160e-01 -3.44762176e-01 -2.17602521e-01
-1.18762650e-01 1.10642278e+00 8.24711502e-01 -6.16558790e-01
5.61619878e-01 6.59529269e-01 -1.55001259e+00 2.72380888e-01
8.39079440e-01 5.27385831e-01 7.98017681e-01 4.67339933e-01
-1.70452908e-01 1.15373266e+00 -4.30488348e-01 4.48513180e-01
4.14762408e-01 -1.77921891e-01 -1.72284234e+00 9.71177667e-02
-4.15856242e-01 -3.23834211e-01 -2.60953397e-01 6.76465273e-01
1.99523196e-01 3.82875651e-01 6.88443840e-01 -1.67533493e+00
-2.97647685e-01 5.11933088e-01 2.69437879e-01 -1.64767280e-01
9.91289198e-01 5.62341154e-01 7.42479444e-01 -2.67812014e-01
6.37792587e-01 1.67473400e+00 1.89154714e-01 7.77357399e-01
-8.79143178e-01 -8.61303881e-02 2.32296139e-01 3.34562778e-01
-1.01246548e+00 -5.43481231e-01 3.36973816e-01 -4.56924200e-01
1.81309938e+00 -8.03942326e-03 4.66747105e-01 5.06209910e-01
6.05378926e-01 6.46739900e-01 7.47827411e-01 -4.50703949e-01
3.86321515e-01 -3.81389409e-01 4.81664054e-02 5.69868386e-01
2.55739510e-01 1.13426670e-01 -2.82149285e-01 -2.57758498e-01
9.07683372e-01 -3.17962706e-01 -7.22772032e-02 -6.03627443e-01
-1.47924387e+00 7.80566752e-01 1.22204058e-01 2.94008344e-01
-6.80700898e-01 1.90011889e-01 5.35964251e-01 4.26360846e-01
-2.59714693e-01 1.27564013e+00 -6.67619526e-01 -3.40943366e-01
-8.09346363e-02 9.64427888e-01 1.33099616e+00 1.39874566e+00
8.66493881e-01 -2.38743082e-01 -2.39479959e-01 6.43510878e-01
3.59274656e-01 1.22734219e-01 2.90705830e-01 -1.83338940e+00
5.43132424e-01 7.92348742e-01 7.28223860e-01 -7.14419246e-01
-9.76136863e-01 6.02803946e-01 1.57098994e-01 5.81964433e-01
4.51717973e-01 -2.30733365e-01 -7.97480702e-01 1.40745938e+00
2.54271030e-01 -7.05637097e-01 6.60231709e-01 8.53481531e-01
4.86966074e-01 8.05594325e-01 1.30145088e-01 -1.69931576e-01
1.37198341e+00 -1.22726214e+00 -6.49472654e-01 -8.81112337e-01
1.06970572e+00 -4.42995340e-01 1.04658496e+00 3.02547425e-01
-1.27064204e+00 -2.53737628e-01 -9.61946070e-01 -2.07516015e-01
-1.79551199e-01 -3.65221411e-01 7.65658617e-01 2.25117952e-01
-1.14250076e+00 4.49775904e-01 -8.82889271e-01 -9.06746805e-01
4.73573133e-02 5.35613120e-01 -2.25206435e-01 -5.65106928e-01
-9.19603944e-01 1.83924186e+00 9.39417899e-01 -3.30948234e-01
-1.14403152e+00 -1.59912601e-01 -1.10967243e+00 -2.02542804e-02
8.34400952e-01 -6.25315011e-01 1.90407956e+00 -3.56473029e-01
-1.71743524e+00 2.58855045e-01 1.12591162e-01 -3.66830707e-01
-6.13531694e-02 7.94078261e-02 -5.01647517e-02 1.08448513e-01
5.09439647e-01 8.09380531e-01 2.58727223e-01 -9.72437203e-01
-8.84199321e-01 -1.36357129e-01 8.39930654e-01 8.77214432e-01
2.89250344e-01 2.21813127e-01 -1.80247083e-01 1.78190023e-01
2.50873595e-01 -1.08274758e+00 -6.08521938e-01 -1.87071979e-01
-3.82385030e-02 -7.33157575e-01 5.75646400e-01 -2.72077680e-01
7.55086303e-01 -1.80420196e+00 3.28326106e-01 -1.26497030e-01
-3.17866325e-01 -5.72644582e-04 -3.35501909e-01 8.39394093e-01
2.75795847e-01 -1.19552128e-01 -1.71218961e-01 4.65061545e-01
3.93942475e-01 7.73280084e-01 -8.19001570e-02 1.94364399e-01
1.61080435e-01 6.42810225e-01 -1.23003733e+00 -6.16803288e-01
3.46147448e-01 -3.96623939e-01 -6.23263240e-01 2.63056964e-01
-9.21017408e-01 2.95225978e-01 -9.53868091e-01 5.18557191e-01
-5.01397029e-02 5.05408943e-02 5.74644685e-01 4.94118541e-01
-2.98463851e-01 8.50809097e-01 -1.10020363e+00 1.92722023e+00
-5.31404376e-01 4.10636872e-01 1.44901618e-01 -8.74136806e-01
9.64817584e-01 5.43255210e-01 3.91743422e-01 -5.64291954e-01
-1.09217748e-01 3.91375273e-01 3.33903313e-01 -8.90596986e-01
5.33328116e-01 -2.04834208e-01 -3.21649432e-01 6.67976320e-01
-1.81162551e-01 -9.12870944e-01 6.90187454e-01 -2.30744958e-01
1.86112046e+00 4.17698622e-01 8.03537130e-01 -4.37498957e-01
2.93051213e-01 1.12758672e+00 6.12289906e-01 7.33462453e-01
-2.48498082e-01 1.00883812e-01 5.13911963e-01 -7.21141160e-01
-8.79534364e-01 -4.09814388e-01 1.56693831e-01 1.27641308e+00
4.63105291e-01 -3.25361222e-01 -4.95292783e-01 -4.27856952e-01
-3.21507215e-01 1.27879715e+00 3.30370143e-02 2.18478218e-02
-7.95068741e-01 -1.19143553e-01 4.32012796e-01 4.74932224e-01
4.59875822e-01 -1.87919235e+00 -1.45977485e+00 6.03766739e-01
-4.71134454e-01 -1.12949431e+00 -7.09232837e-02 4.40033525e-01
-7.57942438e-01 -1.28997374e+00 -3.69976789e-01 -9.75476742e-01
6.15429163e-01 4.34029579e-01 1.06648040e+00 4.72279638e-02
3.73645067e-01 6.40906096e-01 -6.84046924e-01 -7.06142783e-01
-7.02259600e-01 1.30043961e-02 2.83619445e-02 -1.02112329e+00
2.26235434e-01 -2.34879538e-01 8.29351023e-02 6.11637294e-01
-6.32422268e-01 2.03175768e-01 7.49398351e-01 4.79281515e-01
1.30823374e-01 7.44104147e-01 4.30885136e-01 -4.78531450e-01
1.19637394e+00 -3.63547951e-01 -3.53177667e-01 3.58814567e-01
-5.05701959e-01 3.09454978e-01 4.45793778e-01 -1.00596189e-01
-1.28425908e+00 2.44084686e-01 1.09013483e-01 1.74478084e-01
-5.57042420e-01 8.11313450e-01 -3.65468532e-01 1.31143615e-01
9.42924440e-01 -6.09166697e-02 4.53648828e-02 -8.75232965e-02
9.73386467e-02 4.23040748e-01 2.92500496e-01 -1.10665059e+00
1.55558646e-01 1.99518595e-02 1.93444099e-02 -5.53353846e-01
-3.94281715e-01 -4.24220204e-01 -4.02182102e-01 -1.12677008e-01
9.97197688e-01 -4.80260342e-01 -6.60838664e-01 1.14565454e-01
-1.52394354e+00 -1.21176052e+00 -4.12758440e-01 2.71570951e-01
-1.47468853e+00 -1.48818530e-02 -2.76538819e-01 -7.39864826e-01
1.95011348e-01 -1.53258443e+00 9.15162086e-01 1.19949698e-01
-7.13461399e-01 -6.08713686e-01 -4.32066247e-02 -4.07181634e-03
2.26611912e-01 6.42565936e-02 1.25352836e+00 -7.34219790e-01
-4.27778959e-01 6.28072843e-02 -1.14974886e-01 -1.89497858e-01
2.56025732e-01 -5.18934548e-01 -8.88883844e-02 6.25850335e-02
-6.21850565e-02 -3.63867909e-01 -1.70791581e-01 5.03547907e-01
6.72257960e-01 -5.42898655e-01 -6.91167831e-01 -3.10564607e-01
1.00240314e+00 9.10649598e-01 9.98639524e-01 8.03373814e-01
1.64340481e-01 1.18039787e+00 1.43628323e+00 2.35126927e-01
7.35826969e-01 8.21895838e-01 5.97547531e-01 5.59098780e-01
7.80256167e-02 2.82629416e-03 4.92065698e-01 1.18261334e-02
-1.08304851e-01 -2.89962649e-01 -1.54386365e+00 8.07863832e-01
-2.41380191e+00 -9.06194985e-01 -6.48407564e-02 1.77370965e+00
7.18354464e-01 1.77422315e-01 2.60951340e-01 8.89932141e-02
5.15349448e-01 -8.82538036e-02 -6.14519775e-01 -6.45783842e-01
5.98659813e-01 -1.81802332e-01 5.38308382e-01 5.94103932e-01
-8.32465947e-01 1.23412669e+00 7.00839663e+00 3.79573256e-01
-4.35424745e-01 -1.46856010e-01 -3.50286156e-01 3.47263306e-01
7.43755996e-02 3.81487906e-01 -5.60945988e-01 -9.60701257e-02
8.16280901e-01 -4.09100771e-01 6.84964836e-01 1.22261035e+00
2.84923494e-01 -5.39092302e-01 -1.40459907e+00 4.66806799e-01
-3.56864244e-01 -1.22816420e+00 -2.89784670e-01 -6.30125925e-02
2.72241235e-01 -1.82895049e-01 -5.19215465e-01 5.51210582e-01
1.04382026e+00 -1.09762669e+00 8.30224276e-01 1.50000542e-01
1.13705710e-01 -5.41489244e-01 6.18440747e-01 9.14578855e-01
-1.04205668e+00 -5.48830211e-01 -5.78875422e-01 -7.26358056e-01
5.72927773e-01 -3.04975845e-02 -1.46795058e+00 5.11313498e-01
6.19462669e-01 4.81568426e-01 1.68562606e-02 1.02016234e+00
-5.20066679e-01 -2.08981216e-01 -3.37457687e-01 -3.32861364e-01
4.36437964e-01 3.61236632e-02 5.91228962e-01 7.96945691e-01
1.85244396e-01 5.80105186e-01 6.94273174e-01 5.89636445e-01
9.37500179e-01 -3.06278139e-01 -1.09571552e+00 -9.71579999e-02
7.61246204e-01 7.04286516e-01 -8.71492565e-01 -2.29982615e-01
-2.71008492e-01 4.58592176e-01 2.11746812e-01 3.07493001e-01
-6.16207480e-01 -3.23600322e-01 8.16866636e-01 8.65589827e-02
-1.26851484e-01 -6.45745158e-01 -2.99972683e-01 -3.79720956e-01
-1.98218569e-01 -1.24991477e+00 2.69681513e-01 -1.22318518e+00
-8.21988344e-01 4.95971799e-01 7.68133402e-01 -1.13597596e+00
-7.10863233e-01 -7.28749096e-01 -4.48897332e-01 6.30112827e-01
-1.32658517e+00 -7.87015855e-01 -2.63736069e-01 5.74850619e-01
1.00194490e+00 -6.76181018e-02 1.12952054e+00 -3.65339726e-01
-6.10683598e-02 -6.12466335e-01 -8.49728882e-01 -3.34290922e-01
3.42249334e-01 -9.60454345e-01 2.76898503e-01 5.69678843e-01
-6.74766064e-01 4.58987355e-01 9.97227907e-01 -8.83285403e-01
-1.61724234e+00 -8.19465280e-01 8.11414421e-01 -3.62845808e-01
2.43477657e-01 3.63037974e-01 -5.73766291e-01 1.11576879e+00
1.32205084e-01 -5.25448501e-01 1.54071115e-02 1.93359599e-01
2.10225791e-01 5.02067626e-01 -1.39373553e+00 8.62964451e-01
1.33149683e+00 -9.12769511e-02 -1.18941009e+00 7.50904739e-01
8.50137293e-01 -8.21341872e-01 -6.84202433e-01 3.04529220e-01
3.64361048e-01 -6.96941614e-01 8.11799049e-01 -6.58396900e-01
4.26823139e-01 -6.95187151e-01 -3.66474420e-01 -1.41195953e+00
-8.02528322e-01 -4.31044251e-01 2.49980703e-01 6.26572549e-01
6.11287296e-01 -7.47775793e-01 4.52382594e-01 1.30999494e+00
-8.00128102e-01 -4.31005239e-01 -5.50198436e-01 -1.15268838e+00
-1.00197218e-01 -5.79945266e-01 8.35316181e-01 5.49729049e-01
8.09284627e-01 5.69755614e-01 1.29717305e-01 2.87119418e-01
1.49507821e-01 -1.54164165e-01 8.73872340e-01 -1.22778845e+00
4.43672277e-02 -3.54559004e-01 -1.53068295e-02 -1.03691494e+00
4.65234578e-01 -6.69739664e-01 8.34123790e-01 -2.42123413e+00
-4.48140532e-01 -8.87841702e-01 5.14105380e-01 1.15209603e+00
4.63364691e-01 -8.54028106e-01 4.12533998e-01 4.32768464e-01
-5.44349551e-01 2.23474756e-01 1.52816629e+00 -1.29637659e-01
-6.17689550e-01 -3.22038718e-02 -7.56233633e-01 1.05618215e+00
1.03265297e+00 -4.18232888e-01 -7.56883919e-01 -6.82305336e-01
4.26497102e-01 4.83285248e-01 -4.34625447e-02 -1.17346668e+00
5.97471774e-01 -1.14327812e+00 -4.22386795e-01 -1.61988586e-01
3.79593134e-01 -1.10508513e+00 3.23435515e-01 5.75064540e-01
-4.23505098e-01 1.30029216e-01 2.43372664e-01 2.33244047e-01
1.62716396e-02 -9.09374058e-01 5.24716258e-01 -5.93358397e-01
-1.41346443e+00 -1.14994124e-01 -1.07343960e+00 -1.14752464e-01
1.46311367e+00 -4.12371516e-01 -4.58727777e-01 -3.93702298e-01
-5.14811039e-01 5.99371791e-01 6.58208013e-01 3.28865826e-01
6.15903080e-01 -1.07726920e+00 -4.17141110e-01 -2.97325611e-01
2.30565414e-01 4.90758806e-01 -2.36208647e-01 6.26633823e-01
-7.10541964e-01 7.23157167e-01 -7.02507257e-01 -2.16637209e-01
-8.47567976e-01 6.31167769e-01 1.92817003e-01 -4.38357025e-01
-7.41999745e-01 5.12070000e-01 2.42723987e-01 -8.51704061e-01
1.14680424e-01 -5.45158684e-01 -4.04236943e-01 -3.93569857e-01
4.39720988e-01 5.42838216e-01 -1.80430293e-01 -3.00416648e-01
-6.54476523e-01 2.06780687e-01 2.73094118e-01 -3.59017491e-01
1.27594137e+00 -3.85796539e-02 -5.79930395e-02 4.57824506e-02
1.63801849e-01 -8.45047116e-01 -1.22967577e+00 -4.68421951e-02
3.65473896e-01 -2.89713591e-01 -1.47686601e-01 -8.09268415e-01
-3.55771124e-01 3.36045057e-01 -2.62775451e-01 5.96141756e-01
9.75693882e-01 1.03873990e-01 3.97568613e-01 9.01663780e-01
1.23598826e+00 -1.33039486e+00 2.59374470e-01 1.15610635e+00
1.37654603e+00 -8.48483562e-01 2.62663245e-01 -7.00748503e-01
-1.07834160e+00 1.15066600e+00 7.67561674e-01 -1.08141020e-01
1.21818207e-01 4.46845025e-01 -3.69087309e-01 -5.20268083e-01
-8.99784625e-01 -4.48946625e-01 -2.76319861e-01 1.27136755e+00
-1.00767598e-01 3.34580280e-02 -4.36613709e-01 2.28200734e-01
-1.39804974e-01 1.84719235e-01 8.16495895e-01 1.73009086e+00
-9.31779325e-01 -1.16797578e+00 -6.92665040e-01 3.86747628e-01
2.19539434e-01 3.73077512e-01 -5.00574768e-01 9.70730662e-01
-8.73893648e-02 1.53103697e+00 -2.40435451e-03 -1.60267591e-01
5.89702368e-01 1.36129782e-01 8.64982545e-01 -1.28682137e+00
-3.90445054e-01 -4.59263384e-01 9.17159736e-01 -8.64289343e-01
-6.01946592e-01 -9.16522026e-01 -2.03330541e+00 -2.04612404e-01
6.81023300e-03 1.19866012e-03 5.17191410e-01 1.38859916e+00
2.41179273e-01 5.29760361e-01 -1.13327038e-02 -1.22187734e+00
-4.49585378e-01 -7.60031819e-01 -3.71256888e-01 -1.55769393e-01
-1.51463032e-01 -1.01810312e+00 2.17702761e-01 3.46876457e-02] | [4.4195942878723145, 0.9395314455032349] |
145c6d1f-c3d9-40fb-809e-625cd17862d5 | generating-reasonable-and-diversified-story | null | null | https://aclanthology.org/C18-1088 | https://aclanthology.org/C18-1088.pdf | Generating Reasonable and Diversified Story Ending Using Sequence to Sequence Model with Adversarial Training | Story generation is a challenging problem in artificial intelligence (AI) and has received a lot of interests in the natural language processing (NLP) community. Most previous work tried to solve this problem using Sequence to Sequence (Seq2Seq) model trained with Maximum Likelihood Estimation (MLE). However, the pure MLE training objective much limits the power of Seq2Seq model in generating high-quality storys. In this paper, we propose using adversarial training augmented Seq2Seq model to generate reasonable and diversified story endings given a story context. Our model includes a generator that defines the policy of generating a story ending, and a discriminator that labels story endings as human-generated or machine-generated. Carefully designed human and automatic evaluation metrics demonstrate that our adversarial training augmented Seq2Seq model can generate more reasonable and diversified story endings compared to purely MLE-trained Seq2Seq model. Moreover, our model achieves better performance on the task of Story Cloze Test with an accuracy of 62.6{\%} compared with state-of-the-art baseline methods. | ['Xiao Ding', 'Zhongyang Li', 'Ting Liu'] | 2018-08-01 | generating-reasonable-and-diversified-story-1 | https://aclanthology.org/C18-1088 | https://aclanthology.org/C18-1088.pdf | coling-2018-8 | ['cloze-test'] | ['natural-language-processing'] | [ 6.12696707e-01 4.48033571e-01 3.07456050e-02 -2.90781915e-01
-1.17460263e+00 -7.43090749e-01 9.91606593e-01 -2.86431491e-01
-2.53208548e-01 1.18412995e+00 8.15558612e-01 1.27313748e-01
5.19620359e-01 -8.85883510e-01 -8.36175621e-01 -4.31789279e-01
2.09918767e-01 7.54016519e-01 -2.69520562e-02 -5.67837417e-01
2.44350478e-01 -2.57835060e-01 -1.24589527e+00 7.56930590e-01
9.28146243e-01 3.68436903e-01 2.31463417e-01 9.99612510e-01
2.61493679e-02 1.26767933e+00 -1.17239344e+00 -6.50492132e-01
8.41597021e-02 -1.09961951e+00 -8.29173744e-01 -2.41047099e-01
3.24260704e-02 -4.21750247e-01 -4.69128340e-01 7.16823995e-01
9.41329479e-01 1.51786849e-01 1.00452387e+00 -1.24294519e+00
-9.56109643e-01 1.11731982e+00 -2.87953019e-01 -2.37699300e-02
8.51338327e-01 6.00507557e-01 1.01746070e+00 -6.87146544e-01
1.05784881e+00 1.11430180e+00 5.59183300e-01 9.93378520e-01
-1.02473688e+00 -6.00788116e-01 -2.03294814e-01 1.53288290e-01
-1.20533276e+00 -4.16917115e-01 6.26945615e-01 -5.18147171e-01
9.67113435e-01 3.11813056e-01 4.22107548e-01 1.92656648e+00
6.63988069e-02 1.22310460e+00 7.41071522e-01 -2.52503633e-01
3.79595071e-01 -6.54688627e-02 -4.25716966e-01 2.75750626e-02
-2.29584098e-01 1.79077938e-01 -6.81608260e-01 -2.70814169e-02
7.75330126e-01 -6.39503419e-01 -2.57016271e-01 3.50101411e-01
-1.57383168e+00 1.12211084e+00 1.00080386e-01 2.09427416e-01
-4.78191257e-01 2.83153385e-01 6.44661903e-01 1.20387219e-01
4.95764494e-01 1.05382705e+00 2.12689321e-02 -5.88586271e-01
-9.41748202e-01 9.25116658e-01 6.56845152e-01 1.20071912e+00
-2.41000295e-01 4.57694918e-01 -1.10208428e+00 9.88648653e-01
-3.36980104e-01 5.90135336e-01 6.16467953e-01 -8.54531944e-01
6.71831369e-01 1.44861639e-01 1.82007894e-01 -6.76949322e-01
5.75051494e-02 -4.25300330e-01 -9.75667953e-01 1.62948072e-01
2.47062951e-01 -4.96085793e-01 -5.89178622e-01 1.94754910e+00
-1.05728125e-02 1.84444845e-01 4.42824125e-01 8.70614886e-01
1.04529285e+00 1.06686664e+00 -7.03964457e-02 -1.09034337e-01
1.07914400e+00 -1.16540694e+00 -7.50106990e-01 -4.02194500e-01
5.52978039e-01 -7.04062343e-01 1.37826061e+00 3.54148418e-01
-1.32106090e+00 -2.94284731e-01 -1.00057864e+00 2.60271388e-03
7.42857084e-02 -1.32764205e-02 3.84547979e-01 5.45156717e-01
-8.60355020e-01 4.95585948e-01 -2.81571507e-01 -9.04309899e-02
4.89447832e-01 -2.02783406e-01 -8.68827626e-02 -1.08254485e-01
-1.52577281e+00 8.26864541e-01 8.09959769e-01 -4.07623082e-01
-1.25615966e+00 -6.95924461e-01 -1.05870795e+00 -1.52586609e-01
2.44199216e-01 -9.41272020e-01 1.57061636e+00 -9.39889729e-01
-1.66340828e+00 6.96402609e-01 -4.75742891e-02 -6.74256325e-01
9.44847345e-01 -2.83168405e-01 -2.25599602e-01 -6.50342405e-02
3.84517521e-01 1.07303333e+00 7.10783243e-01 -1.01145899e+00
-3.18013787e-01 4.54661727e-01 -1.01352058e-01 3.64219964e-01
1.12162428e-02 1.82150960e-01 -5.27993254e-02 -1.35992098e+00
-6.29159093e-01 -7.06223965e-01 -1.74991220e-01 -6.01901650e-01
-6.07389927e-01 -3.73085558e-01 2.63747960e-01 -8.44825387e-01
1.32929277e+00 -1.84818196e+00 1.24014251e-01 -5.40748417e-01
-2.51469463e-01 3.94355953e-01 -4.68195915e-01 6.55379176e-01
-3.82966734e-02 1.94615960e-01 -5.05591154e-01 -3.31866860e-01
1.36763766e-01 -1.04675680e-01 -6.95025325e-01 -1.51463911e-01
5.45870662e-01 1.20201302e+00 -1.35137320e+00 -4.45944697e-01
-1.83424920e-01 3.45981330e-01 -6.93609416e-01 4.59248662e-01
-6.91728771e-01 6.11076713e-01 -3.20939869e-01 2.56571472e-01
2.19621196e-01 4.51178811e-02 -3.05979967e-01 7.20347881e-01
3.30875874e-01 4.30083901e-01 -7.84241140e-01 1.87316382e+00
-2.71353543e-01 8.50546598e-01 -8.40966403e-01 -2.82231331e-01
9.40073192e-01 5.85912406e-01 -1.51603952e-01 -5.77081859e-01
1.70266688e-01 -1.40345283e-02 -2.52510250e-01 -5.82951903e-01
9.75871742e-01 -3.29888195e-01 -7.01663613e-01 7.18322694e-01
-7.75252208e-02 -5.84010124e-01 4.65224266e-01 3.66057605e-01
1.35704911e+00 3.89140755e-01 3.14824015e-01 2.38807425e-01
2.06043288e-01 2.33528912e-01 4.88717377e-01 9.76275086e-01
6.90746158e-02 1.28274953e+00 6.45681143e-01 -1.35456190e-01
-1.51156902e+00 -9.96729314e-01 3.05309922e-01 9.35064435e-01
-2.54665792e-01 -3.25815976e-01 -1.06194162e+00 -7.28078127e-01
-4.48615193e-01 1.63425136e+00 -5.19343555e-01 -2.91749686e-01
-6.66898310e-01 -5.90833306e-01 1.12571001e+00 4.62958038e-01
6.30044460e-01 -1.83290803e+00 -5.07044137e-01 3.67157400e-01
-7.58952200e-01 -1.06828237e+00 -8.15114498e-01 -5.34353912e-01
-3.29574525e-01 -5.57009876e-01 -9.84788835e-01 -6.17596328e-01
2.49724388e-01 -2.33902320e-01 1.43277168e+00 -4.77450281e-01
1.37982368e-02 -2.22275257e-01 -7.80962706e-01 -6.33171022e-01
-1.02656972e+00 1.41861498e-01 -1.86814472e-01 -2.12472752e-01
-8.70390609e-03 -4.24781770e-01 -2.44594485e-01 -5.28263003e-02
-1.14658046e+00 6.63051009e-01 4.62415129e-01 9.21342373e-01
3.38748306e-01 -6.80003390e-02 1.19581866e+00 -9.92182553e-01
1.03045619e+00 -6.76631629e-01 -4.17237766e-02 1.06916815e-01
-1.16722614e-01 6.34903684e-02 1.01927614e+00 -5.78890681e-01
-1.16087461e+00 -2.34199181e-01 -3.43773842e-01 -9.28653777e-02
-3.49420428e-01 3.60734075e-01 -3.11092526e-01 8.63303483e-01
9.50160861e-01 4.86549675e-01 -3.62930626e-01 2.70833448e-03
5.44644475e-01 6.50566936e-01 9.01836753e-01 -3.22586268e-01
6.57621145e-01 9.21923295e-02 -4.61851031e-01 -4.44658637e-01
-9.80787754e-01 8.77699479e-02 -1.40398353e-01 -3.84684592e-01
9.76953685e-01 -9.63161349e-01 -2.39280805e-01 6.60457075e-01
-1.57387149e+00 -6.41665876e-01 -6.01050556e-01 2.33607382e-01
-1.02840126e+00 -8.68455917e-02 -4.94985849e-01 -8.25100601e-01
-6.11857891e-01 -7.62341380e-01 1.14202476e+00 1.51359335e-01
-9.53097999e-01 -6.13682270e-01 3.75815511e-01 5.18101394e-01
2.35516936e-01 9.05890882e-01 6.12195611e-01 -7.37796307e-01
-3.83643121e-01 -3.34776074e-01 1.53133735e-01 1.37461647e-01
-2.46289089e-01 -4.45567071e-01 -9.09713387e-01 3.61848325e-02
-1.18965402e-01 -7.19557583e-01 7.72165000e-01 2.16061413e-01
7.65553117e-01 -6.60640240e-01 5.03464155e-02 3.88637066e-01
1.10554230e+00 2.24619448e-01 1.04140711e+00 2.93101162e-01
6.16988659e-01 4.53381479e-01 7.56827116e-01 6.95734441e-01
3.33911866e-01 5.47333479e-01 2.35526294e-01 2.80997276e-01
-2.32250497e-01 -8.42166722e-01 7.57856369e-01 3.85935277e-01
1.67463198e-01 -9.48897302e-01 -7.96086490e-01 7.46980429e-01
-1.85518658e+00 -1.50859678e+00 -1.58638582e-01 1.81976509e+00
1.32042205e+00 3.58749717e-01 1.54157683e-01 1.13188863e-01
8.52844775e-01 3.76858592e-01 -6.62244201e-01 -5.83361268e-01
-5.19110382e-01 1.15167191e-02 3.01339347e-02 2.67738938e-01
-9.99544203e-01 1.21176708e+00 6.42707396e+00 1.29872477e+00
-6.92273915e-01 5.59726171e-02 7.27521718e-01 -5.24704695e-01
-5.78148067e-01 -3.01910639e-01 -5.82688034e-01 6.50455594e-01
8.44539940e-01 -4.62325066e-01 4.59320664e-01 8.75182271e-01
4.23059702e-01 7.13386461e-02 -1.23660946e+00 8.72348905e-01
5.00487328e-01 -1.46856701e+00 4.26979363e-01 -3.02861571e-01
1.25136662e+00 -3.75999838e-01 1.21184727e-02 6.95781231e-01
6.10844970e-01 -1.49057007e+00 1.19499183e+00 5.60544014e-01
1.10819650e+00 -9.37586248e-01 7.62782037e-01 7.71168172e-01
-6.78883135e-01 2.56971002e-01 -7.58807063e-02 -2.63371348e-01
6.86802983e-01 4.97757286e-01 -1.33721364e+00 3.80891651e-01
1.55156687e-01 5.83028436e-01 -3.24210823e-01 7.97857404e-01
-7.88470507e-01 8.89046490e-01 1.14908658e-01 -3.31726968e-01
2.91671604e-01 1.93687513e-01 1.02235365e+00 1.35177791e+00
2.34528676e-01 3.28539401e-01 -1.56534743e-02 1.23466420e+00
-3.23090881e-01 -6.63738139e-03 -8.51858854e-01 -4.60317105e-01
3.19420248e-01 8.01867783e-01 -4.75415230e-01 -4.60631013e-01
2.38914594e-01 1.37951279e+00 1.19788408e-01 1.23980373e-01
-1.05374479e+00 -5.53482234e-01 3.40068102e-01 -1.29916407e-02
1.67574927e-01 9.37017724e-02 -6.75496459e-01 -9.97527003e-01
-6.34132102e-02 -1.17302740e+00 2.18394101e-01 -1.21154380e+00
-1.27817225e+00 9.26700056e-01 2.13139039e-02 -1.29566705e+00
-1.03611875e+00 1.84734628e-01 -9.56470549e-01 8.97603750e-01
-1.00936592e+00 -1.23083472e+00 2.01857015e-02 9.11024883e-02
1.18803155e+00 -5.17333329e-01 7.98835814e-01 -1.78667903e-01
-3.12672317e-01 8.90639126e-01 -2.25605536e-03 3.13129395e-01
6.90070808e-01 -1.33449137e+00 1.04808760e+00 9.63211715e-01
-1.95209496e-02 4.33251262e-02 1.16825688e+00 -9.62580681e-01
-7.08704352e-01 -1.50108504e+00 1.30678797e+00 -6.71884298e-01
3.38995576e-01 -5.19114673e-01 -7.16882408e-01 4.51463789e-01
6.48818195e-01 -8.44525814e-01 6.15444243e-01 -4.83041316e-01
-3.23167562e-01 6.49032712e-01 -1.23484540e+00 1.04495406e+00
1.21647573e+00 -1.59897596e-01 -7.44507432e-01 3.78935724e-01
1.18379915e+00 -4.82841462e-01 -4.32546347e-01 2.95747668e-01
2.51884282e-01 -7.70718873e-01 7.17227221e-01 -7.17439353e-01
1.48800242e+00 -7.08282143e-02 1.14099480e-01 -1.67482221e+00
-2.06128284e-01 -1.02220261e+00 1.22450598e-01 1.45224452e+00
5.71188986e-01 -2.45574459e-01 6.67035460e-01 2.87025839e-01
-3.10388535e-01 -6.82404101e-01 -5.79679608e-01 -8.64413917e-01
2.65404284e-01 -5.59314728e-01 8.28411996e-01 8.39036882e-01
1.17027111e-01 7.97949672e-01 -7.97000587e-01 -2.85802543e-01
3.37598592e-01 -9.21913013e-02 8.69573116e-01 -4.63146001e-01
-5.79699814e-01 -5.83154023e-01 -9.48514789e-02 -8.42523217e-01
3.14808279e-01 -1.06781507e+00 3.71664762e-01 -1.69780624e+00
3.21037084e-01 2.96505690e-02 4.36372221e-01 3.54726166e-01
-5.38203537e-01 2.88752347e-01 3.48489255e-01 -6.27585351e-02
-6.54434264e-01 9.50427830e-01 1.45651138e+00 -1.20477326e-01
-1.99046254e-01 -1.28156394e-01 -7.51227438e-01 4.98883605e-01
9.07977045e-01 -5.93198717e-01 -5.28931320e-01 -4.55873042e-01
4.56658453e-01 3.78344893e-01 2.76152015e-01 -8.68327618e-01
-3.89999636e-02 -3.03569138e-01 1.95320085e-01 -5.44323444e-01
2.54542053e-01 1.17779315e-01 2.30582818e-01 2.17442349e-01
-7.98515499e-01 -9.87557992e-02 5.95588759e-02 3.77096266e-01
-1.90053374e-01 -4.79750037e-01 5.27367830e-01 -2.84525812e-01
-3.92985374e-01 4.89352457e-02 -5.58297276e-01 6.67924285e-01
1.16409886e+00 -1.37168318e-01 -3.31011862e-01 -9.40321565e-01
-3.87373090e-01 3.65417391e-01 3.83074552e-01 6.64889991e-01
8.29692662e-01 -1.63569200e+00 -1.68052328e+00 -1.34410098e-01
1.94719881e-01 1.54656813e-01 2.22796977e-01 4.76585850e-02
-5.07484376e-01 2.55330503e-01 -8.34389776e-02 -1.65503234e-01
-1.10641396e+00 4.25760746e-01 -1.12980284e-01 -4.92307186e-01
-4.81247485e-01 1.07347345e+00 1.70242012e-01 -3.96284312e-01
6.20917976e-02 1.71995506e-01 -3.60650361e-01 -1.22278899e-01
8.97544205e-01 4.51557547e-01 -3.87571961e-01 -5.92250288e-01
1.02874413e-01 -8.14065859e-02 -1.07596628e-02 -7.02589631e-01
1.28347623e+00 3.55343968e-01 2.36441553e-01 5.18066764e-01
8.85128558e-01 -1.46824747e-01 -1.20652044e+00 6.28477195e-03
-1.19235992e-01 -3.23683739e-01 -5.12497604e-01 -1.12781179e+00
-3.96757305e-01 5.69707632e-01 -1.88635573e-01 6.39777631e-02
9.11486685e-01 1.80036500e-02 1.29619944e+00 6.54646754e-02
2.99180090e-01 -9.64340150e-01 4.22277212e-01 9.41596925e-01
1.51399755e+00 -1.06273258e+00 -5.23603678e-01 -2.09860027e-01
-1.45601177e+00 5.37762105e-01 5.61634183e-01 -3.17543179e-01
-2.14381441e-01 1.66034818e-01 -5.53639419e-02 2.91378379e-01
-1.05462277e+00 -1.10568879e-02 1.77972928e-01 7.77751625e-01
3.76124144e-01 2.60185808e-01 -3.90042722e-01 1.03930795e+00
-9.71817195e-01 1.34201318e-01 7.31070697e-01 6.26317382e-01
-2.68606842e-01 -1.00801623e+00 -3.67130011e-01 3.09436917e-01
-3.71822000e-01 -4.38309103e-01 -6.81045890e-01 4.30066913e-01
1.14339091e-01 1.08497632e+00 -5.37237003e-02 -3.90639096e-01
2.17673600e-01 1.92542136e-01 2.95195818e-01 -8.71090114e-01
-9.22391891e-01 -2.57521927e-01 4.74098772e-01 -1.82653204e-01
1.46285981e-01 -9.61143255e-01 -1.21291065e+00 -3.26917797e-01
-2.97004916e-02 2.31577009e-02 2.28812888e-01 9.39921081e-01
2.90795356e-01 8.30942154e-01 5.89023054e-01 -5.78634441e-01
-5.09137690e-01 -1.29006135e+00 -2.54272699e-01 6.96036756e-01
-1.77760865e-03 -6.36527091e-02 -8.44242349e-02 4.25863206e-01] | [11.731973648071289, 8.908620834350586] |
40b19155-c33c-49cd-9064-dcb261cd70ba | a-reinforcement-learning-approach-to-improve | 2106.00654 | null | https://arxiv.org/abs/2106.00654v1 | https://arxiv.org/pdf/2106.00654v1.pdf | A reinforcement learning approach to improve communication performance and energy utilization in fog-based IoT | Recent research has shown the potential of using available mobile fog devices (such as smartphones, drones, domestic and industrial robots) as relays to minimize communication outages between sensors and destination devices, where localized Internet-of-Things services (e.g., manufacturing process control, health and security monitoring) are delivered. However, these mobile relays deplete energy when they move and transmit to distant destinations. As such, power-control mechanisms and intelligent mobility of the relay devices are critical in improving communication performance and energy utilization. In this paper, we propose a Q-learning-based decentralized approach where each mobile fog relay agent (MFRA) is controlled by an autonomous agent which uses reinforcement learning to simultaneously improve communication performance and energy utilization. Each autonomous agent learns based on the feedback from the destination and its own energy levels whether to remain active and forward the message, or become passive for that transmission phase. We evaluate the approach by comparing with the centralized approach, and observe that with lesser number of MFRAs, our approach is able to ensure reliable delivery of data and reduce overall energy cost by 56.76\% -- 88.03\%. | ['Ivana Dusparic', 'Maxime Gueriau', 'Babatunji Omoniwa'] | 2021-06-01 | null | null | null | null | ['industrial-robots'] | ['robots'] | [-1.72460765e-01 5.21317482e-01 -3.06057215e-01 2.14930788e-01
-2.62380056e-02 -6.03241384e-01 2.66293705e-01 1.39325157e-01
-1.45264462e-01 1.26675093e+00 -4.23311353e-01 -1.66197211e-01
-2.36704245e-01 -1.18207836e+00 -3.16416055e-01 -1.15039635e+00
-3.42816889e-01 1.30741432e-01 3.22042406e-01 -1.80482522e-01
1.78556561e-01 3.19733292e-01 -1.31215310e+00 -6.17660642e-01
1.07243478e+00 1.19619524e+00 5.48395157e-01 6.42382503e-01
4.60095584e-01 6.37175322e-01 -8.78637433e-01 1.61177233e-01
-1.11020342e-01 -4.66862082e-01 -4.56151515e-01 -1.24771774e-01
-1.21772778e+00 -3.11279416e-01 -1.27604350e-01 7.60223329e-01
5.43320775e-01 1.75984219e-01 4.11273062e-01 -1.89314520e+00
-5.46207547e-01 4.27641839e-01 -5.30017078e-01 2.74191976e-01
2.08822757e-01 1.37724429e-01 2.04744324e-01 -4.54840064e-02
4.57982451e-01 5.73537707e-01 1.51421830e-01 7.06108391e-01
-2.52859056e-01 -4.77663547e-01 9.15793295e-04 2.77501881e-01
-9.45521057e-01 -5.38857996e-01 7.18794465e-01 2.37806335e-01
1.22101545e+00 7.83463269e-02 9.84244168e-01 3.59263420e-01
1.05928934e+00 3.34423184e-01 8.35669816e-01 -3.27760100e-01
7.87057757e-01 1.41090453e-01 -4.62106287e-01 6.49664998e-01
5.04452288e-01 -6.32216632e-02 -2.97782063e-01 2.68907603e-02
3.04514706e-01 -7.90160000e-02 -9.90318954e-02 -1.05758943e-01
-9.35562551e-01 4.45627600e-01 5.23773313e-01 4.48909223e-01
-9.41596150e-01 4.55936283e-01 -2.12232471e-02 2.33069822e-01
4.45742875e-01 2.52169758e-01 -4.60269839e-01 -4.65306133e-01
-3.34079385e-01 -4.62718874e-01 9.32802975e-01 1.08062541e+00
5.26103735e-01 2.27196530e-01 -2.17488166e-02 1.43370241e-01
6.26777172e-01 8.24105918e-01 3.89235795e-01 -1.26750827e+00
1.93259001e-01 5.62162101e-01 6.58236265e-01 -6.18621945e-01
-6.27429068e-01 -6.87300712e-02 -8.77763391e-01 2.75502712e-01
-3.34440500e-01 -9.71638739e-01 -5.70313573e-01 1.36600947e+00
5.28499305e-01 1.86481968e-01 6.75351262e-01 7.77597427e-01
2.04076841e-01 1.03275704e+00 2.10077390e-01 -8.98817122e-01
1.24448359e+00 -6.67194784e-01 -1.12624717e+00 1.98756531e-03
3.42148185e-01 -3.81330371e-01 4.47091669e-01 3.07992399e-01
-1.47381902e+00 4.42388467e-02 -1.45881462e+00 5.22494316e-01
-6.42497659e-01 -1.70923516e-01 3.70538294e-01 8.80537808e-01
-1.30641615e+00 5.84566176e-01 -1.12908149e+00 -5.36733687e-01
3.17390829e-01 8.93457294e-01 2.89875507e-01 5.81575036e-02
-1.11018515e+00 8.81524920e-01 2.11823508e-01 -2.21460730e-01
-9.04885769e-01 -2.34663546e-01 -4.83276814e-01 3.95474471e-02
1.41263738e-01 -9.57614958e-01 1.07413352e+00 -4.00526166e-01
-2.11923528e+00 -5.94071001e-02 3.35810408e-02 -5.32729924e-01
1.40729949e-01 3.36667687e-01 -5.69014430e-01 4.83076632e-01
-2.48325080e-01 5.21810114e-01 7.07175910e-01 -1.03557742e+00
-9.77251649e-01 -3.21392477e-01 2.42785886e-01 4.61880863e-01
-5.10283887e-01 -3.62241119e-01 2.41203979e-01 1.22390412e-01
-2.29313523e-01 -1.05968392e+00 -1.70140490e-01 -3.28481704e-01
-1.50052652e-01 -6.92960322e-01 1.57153368e+00 -1.59173459e-01
7.13990450e-01 -1.70267582e+00 3.14765461e-02 7.34144896e-02
-6.08229935e-02 1.22967009e-02 4.38735753e-01 5.22567213e-01
8.32510531e-01 2.49101102e-01 -4.38649505e-02 -3.68465453e-01
-2.09101960e-01 6.18856490e-01 3.88081491e-01 5.97682834e-01
-4.56000753e-02 7.14534819e-01 -1.10507894e+00 -2.58069396e-01
4.99072909e-01 6.22633636e-01 -3.55504341e-02 1.03563294e-01
-1.46922749e-02 5.68874776e-01 -7.92782009e-01 9.25030231e-01
5.78064501e-01 -1.76228583e-01 2.95272946e-01 1.56571761e-01
-1.05874829e-01 1.05676129e-01 -8.41327369e-01 1.41868222e+00
-7.75960445e-01 3.48511994e-01 6.72624469e-01 -9.55667615e-01
4.74737734e-01 7.29652107e-01 1.09278560e+00 -9.86807466e-01
4.33301687e-01 2.40696773e-01 -2.97701448e-01 -5.25300503e-01
5.39893210e-01 5.60284182e-02 -1.42549068e-01 4.98310059e-01
2.52926517e-02 2.42745969e-03 2.06798404e-01 8.65976363e-02
1.46348763e+00 -1.82401478e-01 5.61130382e-02 -3.54214340e-01
4.30034667e-01 -1.28949538e-01 4.71931666e-01 2.32236907e-01
-6.09090805e-01 -5.32429397e-01 -1.42717585e-02 1.82585955e-01
-6.06891572e-01 -9.33884442e-01 5.22267520e-01 7.46629417e-01
1.07896185e+00 1.27295434e-01 -8.04736793e-01 -4.85289454e-01
-2.54017115e-01 8.09296072e-01 -9.25643742e-02 -6.81235135e-01
-4.46583539e-01 -8.82996678e-01 -3.87866721e-02 1.15373038e-01
6.63917661e-01 -9.67639506e-01 -1.26085222e+00 5.73803365e-01
-1.12028293e-01 -9.49280500e-01 -1.49898514e-01 3.43387336e-01
-8.86464298e-01 -1.09339082e+00 -2.09857091e-01 -4.89292711e-01
7.99988687e-01 5.77774584e-01 9.10241246e-01 2.93603748e-01
2.44671270e-01 6.40814781e-01 -5.22480667e-01 -7.75428593e-01
-3.01036090e-01 1.26212299e-01 2.06352890e-01 -3.38562727e-01
-4.82108928e-02 -6.67443454e-01 -1.21970487e+00 5.17789245e-01
-6.31119490e-01 -4.73534077e-01 4.01235789e-01 2.20841199e-01
4.81710970e-01 5.78822315e-01 1.33937407e+00 -3.37658346e-01
9.37271714e-01 -1.10143185e+00 -6.31975889e-01 6.88380226e-02
-9.38821197e-01 -2.02967167e-01 7.71636665e-01 -9.40353237e-03
-8.24503839e-01 1.89506114e-01 1.76682666e-01 1.41441086e-02
2.22599134e-01 -1.07205883e-01 -3.31257939e-01 -1.06908016e-01
6.95824102e-02 9.35545377e-03 -1.87947117e-02 1.57329068e-01
2.67052740e-01 7.34079063e-01 3.01388741e-01 3.03017050e-02
5.52758873e-01 4.51560915e-01 2.94889688e-01 -5.20213068e-01
3.79216969e-01 -1.70036793e-01 1.63745001e-01 -6.98360920e-01
8.41288090e-01 -9.44347680e-01 -1.35582483e+00 2.24120647e-01
-8.84002388e-01 -2.90793270e-01 -2.51446158e-01 4.37391102e-01
-3.87266934e-01 -1.66262910e-01 -4.54680622e-01 -1.21082008e+00
-7.38612890e-01 -9.80419517e-01 8.37336600e-01 8.14227402e-01
1.69394791e-01 -8.94307971e-01 -3.63511443e-01 -6.81942934e-03
6.80095196e-01 4.14199859e-01 5.78651965e-01 1.25269906e-03
-9.11486566e-01 -1.45764723e-01 5.42695403e-01 9.92130488e-03
6.32310033e-01 -2.46763274e-01 -6.85889602e-01 -5.96219063e-01
2.46779211e-02 1.58084735e-01 1.24303348e-01 5.24568260e-01
6.97728932e-01 -4.67803180e-01 -8.56527746e-01 3.27556767e-03
1.55846071e+00 8.69413614e-01 5.74657738e-01 -6.25550523e-02
9.41056386e-02 2.19617903e-01 8.44896197e-01 6.47902191e-01
7.21555829e-01 1.69610277e-01 1.19815755e+00 2.20957085e-01
2.56261081e-01 5.55017591e-02 6.67139351e-01 8.50337744e-01
-1.57485485e-01 -1.17625368e+00 -1.73436001e-01 5.41146636e-01
-1.80471492e+00 -4.80738878e-01 1.68400966e-02 2.03655005e+00
2.18953207e-01 1.25929072e-01 1.69946328e-01 4.47683126e-01
7.18315303e-01 -2.61185825e-01 -8.84685636e-01 -5.24635792e-01
2.87967205e-01 -1.64846495e-01 9.64866877e-01 3.79592657e-01
-3.91651392e-01 5.33706188e-01 5.68860292e+00 4.02574986e-01
-9.82857406e-01 5.79141498e-01 4.83750761e-01 -2.82018542e-01
-9.53305066e-02 -3.12547982e-01 -1.50381908e-01 9.53336775e-01
1.55821133e+00 -4.10189778e-01 8.30884814e-01 6.11520350e-01
8.86576474e-01 -8.75054002e-01 -7.09557176e-01 6.60642445e-01
-4.36973423e-01 -1.16005671e+00 -6.64375246e-01 2.30065569e-01
9.32193875e-01 6.45075506e-03 -3.63561749e-01 -1.72599986e-01
5.70202827e-01 -4.62277412e-01 4.18743879e-01 5.90189636e-01
1.67528674e-01 -1.32583666e+00 8.49493980e-01 6.38070345e-01
-1.21934414e+00 -4.11013275e-01 -2.67853707e-01 -2.41875574e-01
5.36720574e-01 8.95507991e-01 -7.57298291e-01 7.92950213e-01
6.99828148e-01 2.15293080e-01 1.02821849e-01 6.34695292e-01
-1.82643726e-01 3.77668321e-01 -5.31026483e-01 -6.03623092e-01
-4.94540446e-02 -3.65156054e-01 5.12734175e-01 2.52227157e-01
6.93463683e-01 4.21913803e-01 1.32823780e-01 4.65767950e-01
-3.89842451e-01 -4.90168631e-01 -6.45296156e-01 3.44024822e-02
9.71315444e-01 1.46868122e+00 -1.24237919e+00 -3.48377705e-01
-5.61952740e-02 1.01259565e+00 -4.15462404e-01 1.76230848e-01
-8.17324758e-01 -6.28575802e-01 7.27632880e-01 3.09705567e-02
3.25696878e-02 -4.15273219e-01 -1.97727099e-01 -3.57080907e-01
5.26450835e-02 2.26726666e-01 -5.41952066e-02 -9.42520916e-01
-8.07068110e-01 3.35526049e-01 -3.07348073e-01 -1.15256107e+00
-4.03419495e-01 -4.61900346e-02 -7.76248276e-01 3.39292437e-01
-1.84124875e+00 -5.62143564e-01 -3.93533498e-01 8.17360342e-01
3.60775530e-01 1.27288058e-01 5.73908746e-01 3.41783732e-01
-4.47396576e-01 3.87450159e-02 3.24138522e-01 -6.33989275e-01
7.45912120e-02 -1.16830254e+00 -3.25380564e-01 7.40918159e-01
-6.46633565e-01 -6.18280359e-02 7.31769025e-01 -6.61227345e-01
-2.18663931e+00 -1.05926311e+00 3.41377109e-01 -5.95768094e-02
2.26458311e-01 -1.50367633e-01 -5.12410626e-02 9.01517868e-02
9.21615720e-01 7.47636780e-02 3.14556539e-01 -8.92130971e-01
1.18528056e+00 -3.51535439e-01 -1.98203325e+00 3.70159954e-01
8.98781300e-01 2.25746576e-02 1.68979630e-01 4.13918644e-01
7.57268429e-01 -2.32840583e-01 -1.02076447e+00 8.76903459e-02
1.14054531e-01 -5.38698256e-01 3.37460697e-01 1.06417015e-01
-3.12335223e-01 -4.95302767e-01 1.26456087e-02 -1.80756485e+00
1.42367110e-01 -1.08929801e+00 -7.60695457e-01 1.22882032e+00
3.00462842e-01 -9.39367652e-01 7.87086427e-01 2.73009926e-01
-3.03893834e-01 -7.57888556e-01 -1.39030290e+00 -5.93673944e-01
-3.12812924e-01 6.47983551e-02 7.49504685e-01 4.00375813e-01
5.35415471e-01 2.86915749e-01 -1.51568264e-01 6.34091258e-01
4.97078866e-01 -3.45083654e-01 2.87676454e-01 -1.12183440e+00
4.13308114e-01 2.95143694e-01 -1.81407765e-01 -6.57585800e-01
-1.31024078e-01 -3.67490619e-01 2.36904308e-01 -2.16356850e+00
-5.16467929e-01 -5.04826427e-01 -5.26261091e-01 3.59895617e-01
2.71636695e-01 3.02787185e-01 1.96373612e-01 1.06036685e-01
-8.84176135e-01 6.67492151e-01 1.26814914e+00 -5.12343496e-02
-4.59353715e-01 3.08803618e-01 -5.74907899e-01 3.80523115e-01
1.40301752e+00 -7.17313588e-01 -9.15868282e-01 -8.55141729e-02
1.48420155e-01 5.44754863e-01 1.03495792e-01 -1.20308638e+00
5.55341482e-01 -1.91793814e-01 2.34491885e-01 -3.24871063e-01
3.03586751e-01 -1.36436367e+00 1.96501285e-01 1.08337224e+00
5.37888229e-01 3.95027488e-01 -3.01762551e-01 9.10383582e-01
3.01129907e-01 1.52586862e-01 5.45691609e-01 -7.91496877e-03
-4.29806292e-01 3.55315208e-03 -1.18724084e+00 -5.11950374e-01
1.77843416e+00 -3.53862464e-01 -5.49172044e-01 -5.59574246e-01
-5.63878000e-01 7.77637959e-01 3.95537555e-01 5.07365644e-01
3.60500395e-01 -9.45773661e-01 -2.80700605e-02 -2.80945212e-01
-5.71947753e-01 -2.07485467e-01 4.95771840e-02 8.48477006e-01
-1.41474426e-01 2.75310129e-01 -2.75407672e-01 -4.91226643e-01
-8.62782300e-01 3.99000764e-01 4.33134288e-01 -1.37162998e-01
-1.56463683e-01 2.25800782e-01 -9.50999022e-01 2.54011095e-01
-8.01272765e-02 -5.85003853e-01 -1.34589933e-02 -7.34684318e-02
1.57013182e-02 1.03331506e+00 3.10892999e-01 -1.99595839e-01
-6.15990341e-01 2.57320523e-01 7.65693963e-01 1.18739493e-01
1.36455142e+00 -8.57876062e-01 -7.96735585e-02 4.71586175e-02
8.09139371e-01 -2.01416150e-01 -1.32756364e+00 5.34180522e-01
-1.39793068e-01 5.70834009e-03 3.08924943e-01 -1.02728641e+00
-1.40166819e+00 1.14767790e-01 1.10503161e+00 8.50186467e-01
1.71202421e+00 1.76792294e-02 9.23313260e-01 3.05340827e-01
1.03875327e+00 -1.47013843e+00 1.73163176e-01 -4.36942242e-02
1.11913815e-01 -9.64794934e-01 -5.25544398e-03 -4.78073686e-01
-2.89765775e-01 7.92543054e-01 4.52441752e-01 -1.61546707e-01
7.88935184e-01 4.25160408e-01 -4.44914520e-01 -2.12717026e-01
-7.09008217e-01 -7.34364912e-02 -8.30380797e-01 9.73275423e-01
-1.93785220e-01 3.82657021e-01 -4.38579321e-01 3.41854900e-01
8.93691927e-02 2.53934354e-01 8.29360485e-01 1.50945342e+00
-8.23223293e-01 -1.01838541e+00 -1.69971570e-01 4.41829443e-01
-4.80417192e-01 5.72985291e-01 3.41738649e-02 6.61529422e-01
4.30121362e-01 1.88727748e+00 3.07721078e-01 -2.70440608e-01
2.93137252e-01 -2.58529991e-01 2.50905514e-01 -1.37447044e-01
-3.29575270e-01 -2.33130693e-01 1.40877277e-01 -4.41685796e-01
-9.05627251e-01 -3.29739958e-01 -1.81569541e+00 -4.46087509e-01
-5.27577400e-01 5.01287937e-01 1.24672222e+00 1.10252678e+00
8.18944454e-01 1.03565788e+00 1.20675218e+00 -8.51679981e-01
2.06335615e-02 -7.34016895e-01 -7.79815614e-01 -7.38884211e-01
4.13979709e-01 -7.03371763e-01 -1.78479806e-01 -1.61223665e-01] | [5.858743667602539, 1.71694815158844] |
0e5bc697-e2f3-4a38-a8b6-43a7922a65ba | activeglae-a-benchmark-for-deep-active | 2306.10087 | null | https://arxiv.org/abs/2306.10087v1 | https://arxiv.org/pdf/2306.10087v1.pdf | ActiveGLAE: A Benchmark for Deep Active Learning with Transformers | Deep active learning (DAL) seeks to reduce annotation costs by enabling the model to actively query instance annotations from which it expects to learn the most. Despite extensive research, there is currently no standardized evaluation protocol for transformer-based language models in the field of DAL. Diverse experimental settings lead to difficulties in comparing research and deriving recommendations for practitioners. To tackle this challenge, we propose the ActiveGLAE benchmark, a comprehensive collection of data sets and evaluation guidelines for assessing DAL. Our benchmark aims to facilitate and streamline the evaluation process of novel DAL strategies. Additionally, we provide an extensive overview of current practice in DAL with transformer-based language models. We identify three key challenges - data set selection, model training, and DAL settings - that pose difficulties in comparing query strategies. We establish baseline results through an extensive set of experiments as a reference point for evaluating future work. Based on our findings, we provide guidelines for researchers and practitioners. | ['Bernhard Sick', 'Bernd Bischl', 'Moritz Wirth', 'Denis Huseljic', 'Matthias Aßenmacher', 'Lukas Rauch'] | 2023-06-16 | null | null | null | null | ['active-learning', 'active-learning'] | ['methodology', 'natural-language-processing'] | [ 1.56098604e-01 1.02433242e-01 -7.98148274e-01 -5.10861933e-01
-1.35388017e+00 -6.88578367e-01 6.77438736e-01 2.26581365e-01
-8.74346614e-01 5.80864191e-01 3.25097114e-01 -3.12016189e-01
-1.83974117e-01 -3.75955999e-01 -4.24226671e-01 -8.37147385e-02
7.79759511e-02 6.71690941e-01 2.59525567e-01 -6.99207112e-02
2.41531774e-01 4.49661374e-01 -1.25914657e+00 4.93417144e-01
8.64173591e-01 8.27237368e-01 6.11921214e-02 5.09723425e-01
-3.87858182e-01 1.35090578e+00 -7.17988610e-01 -6.32763326e-01
6.62311763e-02 -4.30360101e-02 -1.23324645e+00 -3.41355264e-01
7.20278323e-01 -3.96461397e-01 -3.42080593e-02 3.56337190e-01
8.78369451e-01 4.53652561e-01 3.27901304e-01 -1.38174570e+00
-6.74717605e-01 8.38772237e-01 2.17165679e-01 2.57986635e-01
2.44803548e-01 1.83135644e-01 1.36153328e+00 -1.30509830e+00
7.09760725e-01 9.22192335e-01 8.19913328e-01 7.94898987e-01
-1.19969952e+00 -6.98637307e-01 3.80557030e-01 2.71946549e-01
-1.43605494e+00 -1.03808331e+00 5.17413199e-01 -1.58494696e-01
1.49295902e+00 6.62776709e-01 5.36380827e-01 1.44932473e+00
-3.62885565e-01 1.27876759e+00 5.41274548e-01 -6.65553212e-01
1.09545477e-01 4.07150745e-01 3.45403135e-01 7.18608260e-01
1.36751160e-01 -6.42310679e-02 -9.80991960e-01 -6.26004577e-01
2.91742802e-01 -2.17455015e-01 -9.78393480e-02 -5.83286285e-01
-8.45240057e-01 6.34465873e-01 6.96240664e-02 2.12167919e-01
-2.76269227e-01 3.21466774e-02 3.57552022e-01 2.55477130e-01
8.71435285e-01 1.00459909e+00 -6.16859138e-01 -2.54535556e-01
-1.09683728e+00 4.98794794e-01 1.05351961e+00 1.05712914e+00
6.27204180e-01 -3.43795359e-01 -6.66504562e-01 1.10954106e+00
8.23437870e-01 4.44364287e-02 1.16511788e-02 -1.19556928e+00
5.60577810e-01 8.38447571e-01 2.38511026e-01 -6.68974757e-01
-1.82047801e-03 -5.40917397e-01 -1.65763482e-01 -2.23727673e-01
2.27292433e-01 -5.59532829e-02 -3.31647307e-01 1.32106757e+00
1.25192264e-02 -5.46742184e-03 -8.65278021e-02 4.69397515e-01
1.20372677e+00 4.04893279e-01 5.61650872e-01 -3.57835233e-01
7.39748418e-01 -1.16133666e+00 -6.90560043e-01 -4.04039472e-01
1.23196590e+00 -8.99587452e-01 1.48837113e+00 1.56306371e-01
-1.26297641e+00 -2.92996079e-01 -7.60968328e-01 -4.70585108e-01
-2.49788165e-01 4.47909571e-02 8.19412231e-01 5.23006201e-01
-1.19276893e+00 5.17799035e-02 -1.10844028e+00 -3.99043620e-01
5.23889005e-01 2.06619844e-01 -3.72385718e-02 1.40065342e-01
-1.23444450e+00 9.98488665e-01 3.17841500e-01 -5.29593974e-02
-9.12649751e-01 -1.22717142e+00 -5.89630663e-01 8.40115026e-02
5.89261413e-01 -6.06375813e-01 1.94617832e+00 -8.14181209e-01
-1.08681118e+00 1.05261850e+00 -3.76850724e-01 -7.01384902e-01
4.88542438e-01 -4.67151791e-01 -4.05520141e-01 -2.33036980e-01
-2.02817079e-02 8.12427819e-01 3.86923067e-02 -1.00243747e+00
-6.79286182e-01 -7.30916567e-04 5.76362908e-01 7.09143102e-01
-7.88520396e-01 4.72286433e-01 -9.24303234e-01 -4.86438215e-01
-4.81527984e-01 -7.52118051e-01 -2.62577593e-01 1.86707452e-01
-3.10005784e-01 -8.03452611e-01 6.47625804e-01 -1.02875769e-01
1.88626981e+00 -1.99622607e+00 -4.14318621e-01 -1.39984339e-01
5.33410788e-01 3.04707021e-01 -4.73656356e-02 7.22131610e-01
2.74948508e-01 4.72864747e-01 9.51478407e-02 -8.89802516e-01
1.56427711e-01 1.89016551e-01 -6.24046266e-01 -2.30113417e-02
2.72174478e-01 1.20402038e+00 -9.48806942e-01 -9.43408966e-01
-8.11069608e-02 4.76543456e-01 -6.33775890e-01 5.63954592e-01
-5.28771341e-01 2.25653931e-01 -4.03538406e-01 5.94441414e-01
-1.08556049e-02 -5.19708931e-01 5.40787429e-02 -1.27028853e-01
-2.73106545e-01 9.83466625e-01 -6.60592854e-01 1.90828538e+00
-5.89049160e-01 7.70516336e-01 -2.14078501e-01 -7.60917068e-01
7.49224782e-01 5.18433928e-01 6.90953374e-01 -8.54197085e-01
-4.59402442e-01 2.24225938e-01 -2.30403230e-01 -4.41763580e-01
7.21530259e-01 6.76556468e-01 7.04545751e-02 7.15722322e-01
8.79653394e-02 1.66978031e-01 3.38755250e-01 5.07101297e-01
1.24768794e+00 -1.42788375e-03 4.81698476e-02 -2.14152947e-01
2.50270039e-01 7.73568973e-02 4.75580782e-01 1.22676206e+00
-2.83183753e-01 6.43244013e-02 1.07508242e-01 -3.09075743e-01
-6.01579845e-01 -9.04097140e-01 8.68328288e-02 1.83407950e+00
-1.23272166e-01 -7.98243284e-01 -4.21024770e-01 -1.10216737e+00
-1.82470605e-01 1.08206737e+00 -5.10391414e-01 -1.02226444e-01
-5.77427864e-01 -7.81184435e-01 8.68057489e-01 5.37660182e-01
4.33771580e-01 -1.05993032e+00 -4.67522472e-01 8.13660249e-02
-3.10192704e-01 -8.07684958e-01 -4.41086471e-01 1.85466364e-01
-8.67213309e-01 -9.96957660e-01 -3.39611173e-01 -7.56558359e-01
2.75553763e-01 2.81134337e-01 1.97278380e+00 3.40245217e-01
-2.34115683e-02 9.41211939e-01 -2.82641411e-01 -9.00654972e-01
-3.82560968e-01 6.15580320e-01 -3.74702305e-01 -5.24906814e-01
7.97206640e-01 1.66262507e-01 -6.71590567e-01 4.57178444e-01
-6.71672761e-01 7.97811076e-02 4.54769671e-01 5.96020043e-01
6.53508961e-01 -4.43016380e-01 3.81101072e-01 -1.32296991e+00
1.01056027e+00 -4.75246131e-01 -5.02798915e-01 9.00008440e-01
-1.33706903e+00 -1.10142447e-01 -6.97636530e-02 -1.54697120e-01
-1.31117713e+00 -9.01021734e-02 -2.60621727e-01 7.94338249e-03
1.32999748e-01 7.10067809e-01 4.89939526e-02 4.71424684e-02
1.08598280e+00 -2.25681126e-01 -4.60283875e-01 -6.93915009e-01
3.22226167e-01 6.77317202e-01 1.13088265e-01 -6.74405158e-01
3.84648114e-01 1.62208125e-01 -7.51241982e-01 -5.69974720e-01
-1.40160322e+00 -7.77259827e-01 -5.27100384e-01 -5.96850477e-02
3.60938728e-01 -8.81754458e-01 -2.53863126e-01 8.15350637e-02
-1.02648747e+00 -6.51080489e-01 -5.41423857e-01 2.64108002e-01
-3.28438401e-01 9.73192081e-02 -5.44192433e-01 -8.88826311e-01
-7.90811956e-01 -1.22208071e+00 9.24366891e-01 -1.12233292e-02
-7.75866330e-01 -1.53980887e+00 4.48128492e-01 7.46149778e-01
6.06002212e-01 -4.00573075e-01 7.21647561e-01 -1.12719667e+00
-6.00535870e-01 6.45516382e-04 1.64989695e-01 1.73920617e-01
-1.34764850e-01 2.51986831e-01 -1.18262541e+00 -3.05586845e-01
-2.70540029e-01 -7.74235487e-01 6.00382805e-01 6.12312928e-02
1.45518398e+00 -2.98360705e-01 -4.79198575e-01 3.66445094e-01
1.02461970e+00 1.96991771e-01 3.94268900e-01 3.84219110e-01
4.97521251e-01 4.47017431e-01 8.42722952e-01 -2.94447429e-02
5.14757693e-01 8.81927669e-01 3.12311500e-02 -4.13666904e-01
-1.20887145e-01 -2.46244714e-01 2.49075949e-01 9.33329940e-01
1.78697240e-02 -7.96063781e-01 -1.29095984e+00 3.66109848e-01
-1.72990155e+00 -9.36999559e-01 2.06313133e-01 2.33647609e+00
1.22038078e+00 3.79851311e-01 -3.82966828e-03 -1.12742111e-01
1.36660874e-01 3.63785699e-02 -6.51666582e-01 -1.53842583e-01
-1.23843312e-01 3.62375140e-01 2.60609895e-01 5.42160869e-01
-1.04778790e+00 9.67225730e-01 7.61546469e+00 8.18071246e-01
-1.20823705e+00 3.24849635e-01 5.86192131e-01 -2.54485339e-01
-3.75820905e-01 2.39360407e-01 -1.18401742e+00 6.67091012e-02
1.04852712e+00 -5.52683175e-01 -5.21680340e-02 8.49844992e-01
1.75880715e-01 2.94735551e-01 -1.55895591e+00 7.97890365e-01
5.05340211e-02 -1.60129237e+00 3.90804820e-02 -4.91711348e-02
5.73019981e-01 4.82780367e-01 1.58179834e-01 7.46590555e-01
5.24448037e-01 -9.16127980e-01 5.85620701e-01 7.11235344e-01
6.49700582e-01 -2.12389275e-01 3.59270453e-01 4.61054325e-01
-8.72576952e-01 -4.43551242e-02 -8.98771808e-02 -5.86262755e-02
-2.14801386e-01 9.97721478e-02 -1.43089831e+00 2.56428629e-01
6.68370545e-01 7.92017162e-01 -8.23490560e-01 1.13243222e+00
-1.52149983e-02 1.21078730e+00 -1.13796212e-01 -7.25655481e-02
1.38041914e-01 3.56297344e-01 3.55589211e-01 1.53676736e+00
-1.46148711e-01 -4.06144232e-01 3.24146450e-01 1.01318955e+00
-4.22732085e-01 4.07669216e-01 -6.00578010e-01 -5.53431571e-01
1.13833559e+00 1.12643397e+00 -2.33509079e-01 -3.46873254e-01
-6.66132987e-01 5.89280248e-01 5.95956862e-01 4.00895655e-01
-5.19136727e-01 -4.55240272e-02 4.18019384e-01 2.27500245e-01
-5.62862873e-01 -1.77547455e-01 -1.55983776e-01 -1.13525534e+00
-2.31880680e-01 -1.17520213e+00 7.86652803e-01 -6.29301786e-01
-1.32462275e+00 3.05865884e-01 3.79765451e-01 -9.80048180e-01
-4.94667351e-01 -1.79569989e-01 -4.04402643e-01 8.95302534e-01
-1.18166339e+00 -1.15428436e+00 -3.18312794e-01 2.60636449e-01
8.00376892e-01 -4.05957639e-01 1.27856612e+00 5.58481097e-01
-6.95302129e-01 1.16194677e+00 3.72742042e-02 2.74590760e-01
9.60108519e-01 -1.29438567e+00 7.47186363e-01 6.70606017e-01
6.36349142e-01 1.13829494e+00 4.02072668e-01 -4.74077910e-01
-1.21213162e+00 -9.34172928e-01 1.25523651e+00 -1.06767511e+00
5.95305979e-01 -6.09904289e-01 -8.22221816e-01 1.04034770e+00
1.70148820e-01 -2.20379278e-01 1.12873864e+00 6.78923965e-01
-1.79207176e-01 -1.57362223e-01 -7.80466676e-01 5.16527236e-01
1.28928030e+00 -7.82247603e-01 -4.58987892e-01 4.52757537e-01
8.02058101e-01 -4.12904441e-01 -1.07560790e+00 6.51043355e-01
6.31421745e-01 -8.72095525e-01 1.11124229e+00 -9.37954664e-01
-6.37572855e-02 3.46476287e-01 7.75603428e-02 -8.89912248e-01
-3.66577566e-01 -5.70689797e-01 -4.67336118e-01 1.35173786e+00
9.05468702e-01 -4.38027740e-01 9.80241358e-01 9.91002619e-01
-3.40562850e-01 -1.12102425e+00 -4.29438442e-01 -3.78537774e-01
7.35955760e-02 -8.20358932e-01 4.22802716e-01 1.25086486e+00
-3.41824383e-01 5.29643297e-01 1.30638123e-01 -2.77113497e-01
4.93099958e-01 -3.14472288e-01 8.85506511e-01 -1.27694392e+00
-1.97486758e-01 -3.47934633e-01 2.31245205e-01 -1.38648462e+00
1.25846967e-01 -8.86098802e-01 -2.34402150e-01 -1.68716550e+00
2.76877046e-01 -1.07902765e+00 -6.32899165e-01 7.92851329e-01
-1.41444132e-01 2.16293827e-01 -6.91873357e-02 5.82519114e-01
-1.19926417e+00 2.20401093e-01 8.05958271e-01 -3.20774555e-01
-2.90425390e-01 1.54686689e-01 -9.29583311e-01 5.21252811e-01
5.38044572e-01 -3.79973024e-01 -1.12872052e+00 -9.85584855e-01
6.85094535e-01 -3.34359139e-01 1.51011139e-01 -5.79960763e-01
5.11996150e-01 -2.53892541e-01 1.12322778e-01 -6.35023892e-01
3.24659824e-01 -5.96442699e-01 3.34842503e-02 6.48180619e-02
-1.30857944e+00 9.13220644e-02 1.73055351e-01 1.59073949e-01
-5.51881678e-02 -2.78012216e-01 4.12279129e-01 -2.00207472e-01
-9.58432615e-01 5.12317836e-01 -2.88651824e-01 5.53175807e-01
5.41934371e-01 -1.46417487e-02 -5.16506195e-01 -3.59395325e-01
-6.61521554e-01 5.13124645e-01 2.68806875e-01 7.53963411e-01
3.36658627e-01 -1.17985904e+00 -7.12341487e-01 -4.75983787e-03
7.04195380e-01 1.40087754e-01 -1.08070172e-01 8.44279766e-01
-4.89209265e-01 7.64229298e-01 2.80510753e-01 -5.67184746e-01
-1.54097891e+00 3.54275592e-02 6.89933300e-01 -6.04424715e-01
-1.07073173e-01 1.21922076e+00 -9.12767574e-02 -6.71082497e-01
8.52691829e-01 -1.11201003e-01 -3.70059222e-01 3.67729273e-03
3.52045029e-01 4.23736453e-01 5.48678637e-01 -1.98034808e-01
-2.56446242e-01 -4.27970082e-01 -4.87167776e-01 -1.63082048e-01
1.10799491e+00 -2.15046555e-01 1.41433242e-03 8.47862720e-01
1.02846408e+00 1.82051420e-01 -8.56499553e-01 -6.67427480e-01
4.94640827e-01 -3.01671982e-01 4.46627349e-01 -1.26731074e+00
-6.06575727e-01 6.06925368e-01 8.10969234e-01 2.90471613e-01
8.77335966e-01 -5.17390147e-02 4.51344073e-01 9.43094790e-01
2.71439046e-01 -1.24719155e+00 8.66753012e-02 6.21277690e-01
8.70138347e-01 -1.17249560e+00 1.52315628e-02 -1.85743570e-01
-5.52748680e-01 6.15619004e-01 7.82473385e-01 6.72043502e-01
6.23445570e-01 3.16415429e-01 4.96517599e-01 -3.94363046e-01
-1.39777434e+00 1.80048928e-01 7.92702317e-01 3.00857306e-01
1.33750975e+00 -3.20184141e-01 -1.93014711e-01 2.78513104e-01
-7.07081854e-02 1.79761112e-01 -2.04932764e-01 1.14396513e+00
-2.33539581e-01 -1.45446813e+00 5.52902520e-02 6.00724816e-01
-3.94663751e-01 -2.74369985e-01 -9.30875063e-01 6.59167707e-01
-6.37264980e-04 1.11641467e+00 2.18787357e-01 -2.08001643e-01
4.27790940e-01 2.12410435e-01 3.95675540e-01 -1.13139880e+00
-8.68354797e-01 -2.65718102e-01 6.33785665e-01 -6.35131538e-01
-4.88219231e-01 -3.87360275e-01 -8.40180099e-01 -1.54919580e-01
-5.45939624e-01 5.10035336e-01 4.21150684e-01 7.46325254e-01
6.64351821e-01 1.05969653e-01 2.44672269e-01 -4.88879569e-02
-6.94502592e-01 -1.07919192e+00 -1.05037866e-02 3.83871108e-01
-1.62479803e-01 -5.25338173e-01 -1.14629343e-01 -2.77241822e-02] | [9.661767959594727, 4.461142539978027] |
3fd20a78-5ce2-4448-8acf-b3af8daac052 | feature-based-decipherment-for-machine | null | null | https://aclanthology.org/J18-3006 | https://aclanthology.org/J18-3006.pdf | Feature-Based Decipherment for Machine Translation | Orthographic similarities across languages provide a strong signal for unsupervised probabilistic transduction (decipherment) for closely related language pairs. The existing decipherment models, however, are not well suited for exploiting these orthographic similarities. We propose a log-linear model with latent variables that incorporates orthographic similarity features. Maximum likelihood training is computationally expensive for the proposed log-linear model. To address this challenge, we perform approximate inference via Markov chain Monte Carlo sampling and contrastive divergence. Our results show that the proposed log-linear model with contrastive divergence outperforms the existing generative decipherment models by exploiting the orthographic features. The model both scales to large vocabularies and preserves accuracy in low- and no-resource contexts. | ['Parker Riley', 'Iftekhar Naim', 'Daniel Gildea'] | 2018-09-01 | null | null | null | cl-2018-9 | ['decipherment'] | ['natural-language-processing'] | [-8.23771115e-03 -5.05713046e-01 -2.77326196e-01 -4.01531756e-01
-7.88359284e-01 -7.44153261e-01 9.38404083e-01 1.27137780e-01
-5.12126446e-01 7.93928266e-01 4.94346261e-01 -3.25460494e-01
-2.81414613e-02 -8.69180918e-01 -8.26095879e-01 -5.31560838e-01
1.74766287e-01 7.48552144e-01 -1.44943550e-01 6.03557862e-02
2.05047652e-01 1.71022236e-01 -1.42652369e+00 1.57961100e-01
1.01123202e+00 2.06831262e-01 5.18260419e-01 8.14153075e-01
-1.12391405e-01 1.01682216e-01 -2.34879300e-01 -6.14825070e-01
1.51156709e-01 -5.31243145e-01 -2.71141171e-01 -4.60603327e-01
4.25293982e-01 -4.13082004e-01 -2.94215202e-01 1.06987393e+00
5.25091588e-01 -2.01914117e-01 9.82720792e-01 -8.16885412e-01
-1.16020167e+00 9.33223069e-01 -7.04232305e-02 7.61713088e-02
4.05271351e-01 3.53392735e-02 1.56251562e+00 -1.09503901e+00
4.54378754e-01 1.65738320e+00 5.67670226e-01 1.17277615e-01
-1.46206546e+00 -9.27641213e-01 -1.88958123e-01 8.27963650e-02
-1.87372029e+00 -4.24445033e-01 6.22328281e-01 -4.61843997e-01
1.21013212e+00 -1.05131917e-01 6.85999155e-01 1.36729372e+00
3.20400655e-01 6.99894369e-01 1.49124789e+00 -7.15394199e-01
3.30863595e-01 1.47414701e-02 1.63212776e-01 7.94743836e-01
6.16745889e-01 4.29593861e-01 -1.05251765e+00 -4.09805804e-01
6.39814138e-01 2.75311060e-02 2.61617899e-01 -1.22313730e-01
-1.15185046e+00 9.45450723e-01 -1.05684303e-01 4.20246542e-01
-5.48158363e-02 6.05338335e-01 -1.62464585e-02 2.51539260e-01
2.72829086e-01 4.30606753e-01 -1.91401452e-01 -3.21246773e-01
-1.11346066e+00 5.81190782e-03 9.70350862e-01 1.08485806e+00
1.03906190e+00 1.13704719e-01 2.29259521e-01 6.64312422e-01
7.43969858e-01 8.16085339e-01 6.51154757e-01 -7.47299790e-01
2.27786869e-01 7.72075504e-02 -6.13903143e-02 -7.74913430e-01
1.58343926e-01 -2.04105988e-01 -6.67158723e-01 -4.44864631e-01
4.06183153e-01 2.19008699e-01 -6.36996865e-01 2.04187083e+00
-1.04084708e-01 4.46230054e-01 2.03223214e-01 4.46616024e-01
4.38154042e-01 7.24095702e-01 1.48444638e-01 -2.97219694e-01
1.59398282e+00 -3.70118916e-01 -8.18243265e-01 -4.75913197e-01
4.70767587e-01 -1.08103085e+00 1.22948980e+00 2.83649504e-01
-8.02904785e-01 -4.35170949e-01 -1.05858040e+00 -2.77503729e-01
-1.58099338e-01 2.38966435e-01 9.33941483e-01 1.09976625e+00
-9.87714529e-01 5.17717600e-01 -1.02981913e+00 -5.96952736e-01
9.73621011e-02 2.89995372e-01 -2.37072930e-01 -2.09782511e-01
-1.26723087e+00 6.30944014e-01 5.23684680e-01 -1.91440687e-01
-9.64324474e-01 -2.36452818e-01 -8.60038102e-01 1.42585799e-01
-3.43810134e-02 -9.17812109e-01 8.62360239e-01 -4.96025890e-01
-1.90179682e+00 7.64198661e-01 -4.45224375e-01 -2.73831248e-01
-5.58977649e-02 -2.94351399e-01 -1.85814857e-01 -8.97237435e-02
-1.23767145e-01 6.48551524e-01 9.48097587e-01 -1.02322948e+00
2.23091483e-01 -3.78913343e-01 -3.64966244e-01 1.68473482e-01
-5.11890471e-01 1.01032995e-01 -3.35300803e-01 -9.61905777e-01
1.29533321e-01 -8.87975752e-01 2.01558426e-01 -1.32679999e-01
-2.78566927e-01 -1.63137883e-01 1.47841692e-01 -4.96862561e-01
1.24149370e+00 -2.13239646e+00 3.13476175e-02 1.21608665e-02
1.19727366e-02 6.74885586e-02 -2.40860462e-01 7.15509355e-01
5.16621947e-01 -6.64366176e-03 -3.32410872e-01 -6.66901171e-01
3.81883204e-01 7.74648070e-01 -5.57645559e-01 4.52417731e-01
1.62552744e-01 1.01120889e+00 -9.15324450e-01 -5.35894156e-01
-7.38673955e-02 4.51513827e-01 -1.17686343e+00 2.86516011e-01
-3.62321764e-01 4.22335677e-02 -1.50861695e-01 5.17772555e-01
6.74690723e-01 -2.53930092e-01 7.26483345e-01 2.54843593e-01
5.97225018e-02 5.93315661e-01 -1.01739883e+00 1.88347018e+00
-5.39265096e-01 4.78174031e-01 -4.02746618e-01 -6.93085790e-01
1.05579066e+00 -2.39544902e-02 -4.38301742e-01 -1.86628267e-01
1.67765439e-01 4.12668645e-01 1.36208117e-01 -1.64535835e-01
7.77121663e-01 -6.52876318e-01 -4.59612012e-01 7.58829355e-01
4.03914839e-01 -3.88650388e-01 -1.11539952e-01 4.25505280e-01
7.81002402e-01 1.54390618e-01 7.28664339e-01 -2.16790915e-01
4.33822684e-02 -4.31605518e-01 4.93929923e-01 9.96577978e-01
-1.63022056e-02 3.89062345e-01 2.99482375e-01 1.25392243e-01
-9.69061434e-01 -1.66419530e+00 -1.15871452e-01 1.05878317e+00
5.74054457e-02 -8.05797637e-01 -5.00290573e-01 -5.05248131e-03
8.68995935e-02 1.04381883e+00 -2.09078461e-01 -2.76341408e-01
-2.65278786e-01 -8.78916144e-01 1.03283060e+00 4.32575256e-01
6.23238087e-02 -7.04344690e-01 5.58196865e-02 2.47831136e-01
-2.23834693e-01 -1.11071551e+00 -4.12242651e-01 2.20945850e-01
-1.09965491e+00 -3.85472715e-01 -3.47306937e-01 -7.01055646e-01
3.78994554e-01 5.07387333e-02 9.70792472e-01 -5.09983301e-01
-2.59291112e-01 3.57832193e-01 -1.84472635e-01 -7.93533176e-02
-7.58232772e-01 -6.16215467e-02 4.92126971e-01 -1.76121891e-01
9.37986791e-01 -1.00406933e+00 -2.96417505e-01 5.49784638e-02
-8.23610246e-01 -2.31619924e-01 9.00272191e-01 9.97018993e-01
7.35859871e-01 -4.35966551e-01 2.56525546e-01 -6.60533667e-01
5.72928846e-01 -5.04453838e-01 -7.35395730e-01 2.94851452e-01
-7.05490947e-01 6.39107585e-01 7.10798681e-01 -6.47495508e-01
-1.04015744e+00 2.52347123e-02 4.29942831e-02 -1.67742446e-01
-1.32871166e-01 4.63815093e-01 -2.16634080e-01 3.19610357e-01
5.65147340e-01 6.56459153e-01 -8.91073719e-02 -7.20199049e-01
8.70806634e-01 8.32738638e-01 5.39381921e-01 -8.62723470e-01
8.00994694e-01 4.56938118e-01 -9.48785469e-02 -9.05423582e-01
-4.87214804e-01 -3.70719314e-01 -8.22635949e-01 3.64701033e-01
9.11022067e-01 -1.17560422e+00 -5.27093351e-01 3.99775624e-01
-1.17044711e+00 3.85899417e-04 -5.50245941e-02 1.05092001e+00
-6.41592383e-01 8.28919530e-01 -7.61583030e-01 -8.46152484e-01
-6.83611035e-02 -9.45791483e-01 1.19771230e+00 -2.70352751e-01
-4.72700566e-01 -1.03767657e+00 6.63374901e-01 3.27848971e-01
2.02349529e-01 -4.74662989e-01 1.23968208e+00 -5.73133051e-01
-1.02403152e+00 -9.14134234e-02 -3.56759578e-02 2.20462173e-01
-5.02376556e-02 2.40866877e-02 -9.70857322e-01 -2.02005476e-01
-2.88362414e-01 -4.82113004e-01 1.07046020e+00 2.32329205e-01
4.53052759e-01 -4.37095195e-01 -2.14333281e-01 6.55206859e-01
1.31283343e+00 -5.01512766e-01 4.80739862e-01 -3.56243908e-01
4.25898820e-01 2.39085391e-01 1.90694675e-01 6.19014561e-01
5.63565731e-01 5.77993453e-01 -1.31485105e-01 6.01031840e-01
-1.67982310e-01 -1.14463532e+00 9.16602135e-01 1.82392895e+00
2.09687248e-01 -3.80806118e-01 -8.62735450e-01 6.56613648e-01
-1.50376689e+00 -1.04947281e+00 7.70879723e-03 2.13898468e+00
1.22859180e+00 -1.75438430e-02 -2.06222951e-01 -2.59723395e-01
5.16354561e-01 1.14351816e-01 -1.84711874e-01 -5.53424239e-01
-5.11229575e-01 4.91685271e-01 2.89387017e-01 8.32505286e-01
-5.47324002e-01 1.40401995e+00 7.37087870e+00 9.45272684e-01
-5.76974750e-01 3.59613925e-01 -1.90860443e-02 -9.18370038e-02
-8.15281153e-01 5.88458598e-01 -1.10474503e+00 4.19689268e-01
1.10463226e+00 -7.88030699e-02 7.84814537e-01 5.89971483e-01
3.12584266e-02 -2.36563116e-01 -1.23758888e+00 1.02518725e+00
4.08697963e-01 -9.61144745e-01 3.44345450e-01 4.58209887e-02
7.12309897e-01 1.12093948e-01 2.80924618e-01 1.69397593e-01
9.32446003e-01 -9.61957693e-01 6.28032565e-01 5.19642413e-01
8.55486155e-01 -3.67459834e-01 1.93817049e-01 5.22559226e-01
-9.10148799e-01 1.17085464e-01 -6.85541928e-01 -3.61866266e-01
2.60154307e-01 5.57721853e-01 -8.25008154e-01 6.12792224e-02
2.52020419e-01 5.03465772e-01 -4.45767581e-01 3.92359763e-01
-6.93065464e-01 1.06908655e+00 -5.61055422e-01 -5.11953294e-01
2.51751635e-02 -3.86026859e-01 5.71412921e-01 1.25903153e+00
7.13253915e-01 -7.05097914e-02 -1.91061005e-01 1.15362906e+00
-2.09198564e-01 2.94448107e-01 -6.13792777e-01 -5.44189751e-01
8.69140923e-01 7.03420877e-01 -5.43679535e-01 -3.91009390e-01
-3.19614053e-01 1.29535270e+00 5.05518258e-01 2.23166302e-01
-4.73061770e-01 -2.16085717e-01 6.83483481e-01 -1.66060865e-01
6.52995408e-01 -8.20063651e-01 -1.80974826e-01 -1.44771338e+00
-2.53280431e-01 -7.58890688e-01 2.35748500e-01 -7.33597457e-01
-1.52493715e+00 1.99334413e-01 2.42996424e-01 -7.62067616e-01
-5.65382421e-01 -7.00864971e-01 -2.33171195e-01 8.64020348e-01
-1.43018091e+00 -1.29966044e+00 2.19539031e-01 4.64024931e-01
7.83726051e-02 -2.61410087e-01 1.17328107e+00 1.02957293e-01
-2.25161970e-01 9.32612419e-01 6.98951185e-01 -2.23937035e-01
9.46511149e-01 -1.19282115e+00 4.77124691e-01 9.05011296e-01
7.88968027e-01 1.44369268e+00 7.09046543e-01 -7.39775419e-01
-1.62661541e+00 -8.72817695e-01 1.57779729e+00 -6.00424945e-01
7.66467035e-01 -9.75717008e-01 -6.40398264e-01 6.95687473e-01
8.55193585e-02 -3.27382267e-01 1.28592730e+00 3.40712339e-01
-1.14575338e+00 1.60390601e-01 -7.46658683e-01 6.89244092e-01
1.08520544e+00 -1.17649841e+00 -1.05915964e+00 2.57380903e-01
7.07975447e-01 3.67164195e-01 -8.69504392e-01 -2.80489832e-01
8.35097611e-01 -6.21383548e-01 8.38683248e-01 -5.03161609e-01
2.37241223e-01 -1.90285489e-01 -6.00200057e-01 -1.16446757e+00
-3.11624199e-01 -1.09574461e+00 -6.23176396e-02 9.66118634e-01
3.13361883e-01 -6.90497994e-01 3.16226125e-01 4.25501168e-02
2.77147502e-01 1.65512413e-02 -1.16530347e+00 -1.25571752e+00
4.29297447e-01 -7.62549222e-01 5.27249396e-01 8.38864744e-01
1.68459252e-01 5.23670554e-01 -5.85305035e-01 2.19448879e-01
8.65443766e-01 3.56097192e-01 7.94276178e-01 -1.09500694e+00
-1.01212692e+00 -1.87864050e-01 -4.02696818e-01 -1.59170043e+00
4.24695760e-01 -1.45591891e+00 3.38483676e-02 -9.96068895e-01
5.76250553e-01 -1.33896500e-01 -9.89637822e-02 1.03015251e-01
-3.60501260e-01 3.99130672e-01 -8.80428329e-02 5.09083450e-01
-4.83846337e-01 9.27008629e-01 7.13674664e-01 1.44514874e-01
-6.85636476e-02 -3.47423077e-01 -6.87724829e-01 7.36261606e-01
6.75662518e-01 -8.21189702e-01 -1.78296268e-01 -5.30279875e-01
6.35050297e-01 -1.73012808e-01 3.45604748e-01 -5.96017659e-01
2.92346537e-01 -3.70903276e-02 5.17580844e-02 -5.12294114e-01
5.55648625e-01 -2.31118947e-01 1.94002360e-01 5.07197082e-01
-3.96176934e-01 7.15382919e-02 -6.49177954e-02 1.02544355e+00
-1.34459749e-01 -2.21299738e-01 6.23646498e-01 -1.64049286e-02
-2.24327296e-01 5.80413900e-02 -9.02112365e-01 2.16348514e-01
5.30516505e-01 -2.07324877e-01 1.01107685e-02 -4.02301967e-01
-4.60505188e-01 -5.37605464e-01 5.24365246e-01 2.47422323e-01
6.64924264e-01 -1.31542599e+00 -7.80606627e-01 3.77088755e-01
4.10814375e-01 -7.81224608e-01 -1.29413724e-01 4.61976171e-01
-4.08519328e-01 6.44614279e-01 2.49480829e-02 -4.82234627e-01
-1.19752467e+00 4.10716742e-01 -2.69630849e-01 -3.34581882e-01
-1.18619464e-01 8.62870455e-01 1.02174871e-01 -6.73991203e-01
-1.91860393e-01 -4.75253969e-01 2.39158452e-01 -3.90452817e-02
8.82371068e-02 2.02949449e-01 -3.65208507e-01 -7.05702722e-01
-3.04526687e-01 5.66392362e-01 -1.02709778e-01 -6.24602556e-01
1.20573401e+00 -2.43878677e-01 -2.77234405e-01 7.09059775e-01
1.24282062e+00 3.30269277e-01 -8.10397744e-01 -6.99484050e-01
-8.27682763e-02 -5.33412755e-01 -1.30169541e-01 -4.47236329e-01
8.80427752e-03 1.30001605e+00 2.93523610e-01 -5.59935868e-01
7.42389202e-01 9.44627225e-02 7.87931442e-01 7.87054300e-01
5.58291376e-01 -9.42921042e-01 1.71637118e-01 7.52124071e-01
3.73018205e-01 -8.96375358e-01 2.00608388e-01 -3.29668045e-01
-3.98726821e-01 7.81276345e-01 4.45583165e-02 -1.26973122e-01
7.58985579e-01 3.37743908e-01 -2.79114246e-01 1.05855741e-01
-8.75351846e-01 -1.26598537e-01 1.70784369e-01 6.59167230e-01
4.12005961e-01 4.13644254e-01 -6.07915699e-01 7.42461979e-01
-6.49016976e-01 -1.71614796e-01 3.69481117e-01 6.13455772e-01
-4.28356051e-01 -1.36015058e+00 -3.58739704e-01 2.11700439e-01
-3.45158905e-01 -9.33674812e-01 -5.26650608e-01 4.07428920e-01
-8.82132724e-02 7.49300301e-01 2.15766445e-01 -2.39512712e-01
-4.88475531e-01 4.06021595e-01 7.99459517e-01 -6.82229221e-01
-2.82089829e-01 2.52128303e-01 -1.53942227e-01 -2.51692563e-01
-1.91443592e-01 -7.09626079e-01 -7.48221576e-01 -5.09488165e-01
-6.25244439e-01 -3.92594859e-02 6.14501238e-01 7.98869371e-01
5.78817785e-01 -2.37906709e-01 3.88774782e-01 -2.78488785e-01
-9.67821836e-01 -1.03626180e+00 -7.45402932e-01 2.06154361e-01
-2.04277992e-01 -4.20186192e-01 -4.88569409e-01 2.05796212e-01] | [11.166964530944824, 9.58345890045166] |
11ff9bb3-5044-45f0-b7d6-d8f8b37b4031 | sketch3t-test-time-training-for-zero-shot | 2203.14691 | null | https://arxiv.org/abs/2203.14691v1 | https://arxiv.org/pdf/2203.14691v1.pdf | Sketch3T: Test-Time Training for Zero-Shot SBIR | Zero-shot sketch-based image retrieval typically asks for a trained model to be applied as is to unseen categories. In this paper, we question to argue that this setup by definition is not compatible with the inherent abstract and subjective nature of sketches, i.e., the model might transfer well to new categories, but will not understand sketches existing in different test-time distribution as a result. We thus extend ZS-SBIR asking it to transfer to both categories and sketch distributions. Our key contribution is a test-time training paradigm that can adapt using just one sketch. Since there is no paired photo, we make use of a sketch raster-vector reconstruction module as a self-supervised auxiliary task. To maintain the fidelity of the trained cross-modal joint embedding during test-time update, we design a novel meta-learning based training paradigm to learn a separation between model updates incurred by this auxiliary task from those off the primary objective of discriminative learning. Extensive experiments show our model to outperform state of-the-arts, thanks to the proposed test-time adaption that not only transfers to new categories but also accommodates to new sketching styles. | ['Yi-Zhe Song', 'Tao Xiang', 'Pinaki Nath Chowdhury', 'Vaishnav Potlapalli', 'Ayan Kumar Bhunia', 'Aneeshan Sain'] | 2022-03-28 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Sain_Sketch3T_Test-Time_Training_for_Zero-Shot_SBIR_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Sain_Sketch3T_Test-Time_Training_for_Zero-Shot_SBIR_CVPR_2022_paper.pdf | cvpr-2022-1 | ['sketch-based-image-retrieval'] | ['computer-vision'] | [ 2.27042139e-01 -2.16628417e-01 -2.07544878e-01 -5.70405543e-01
-8.35037887e-01 -9.07395482e-01 8.51963103e-01 -3.28144729e-01
-2.25212216e-01 3.59468490e-01 -7.94038624e-02 2.62011252e-02
-1.82872042e-01 -8.15955997e-01 -7.41900086e-01 -4.29744661e-01
2.77083218e-01 8.45237792e-01 3.05078745e-01 -2.76928782e-01
3.02416205e-01 5.42951763e-01 -1.62529278e+00 4.57771689e-01
4.66924042e-01 9.73684788e-01 3.19720894e-01 6.87231779e-01
-3.78437877e-01 2.48510882e-01 -5.71841359e-01 -6.97270453e-01
4.51305926e-01 -3.43086511e-01 -4.91574883e-01 3.31809729e-01
9.95119691e-01 -5.82131982e-01 -6.27877533e-01 7.94998825e-01
4.91619021e-01 2.99070150e-01 9.66580808e-01 -1.32545269e+00
-9.82832193e-01 1.39370039e-01 -6.44257739e-02 -1.28184214e-01
1.36245981e-01 2.83098251e-01 1.01218700e+00 -1.26119328e+00
9.26025927e-01 1.32100594e+00 3.96168262e-01 1.10540414e+00
-1.40673268e+00 -6.71042681e-01 3.72033060e-01 3.02866876e-01
-1.33079040e+00 -4.56835657e-01 1.11667645e+00 -3.07005584e-01
3.10274661e-01 3.38276654e-01 5.91893554e-01 1.55550861e+00
-2.51154125e-01 9.92753267e-01 8.76196802e-01 -3.35983157e-01
5.94422281e-01 5.42319179e-01 -1.78362086e-01 5.26298344e-01
-7.62040764e-02 3.22494209e-02 -6.84812307e-01 -6.18271418e-02
8.34253490e-01 1.95122972e-01 -6.97583482e-02 -1.06867099e+00
-1.09053349e+00 7.81165421e-01 3.70238245e-01 3.83620024e-01
3.90596711e-03 2.87786573e-01 4.02221739e-01 7.02178478e-01
3.82352084e-01 5.16478896e-01 -4.34580654e-01 -2.22520828e-02
-1.15476227e+00 3.82072367e-02 7.59157717e-01 1.16814148e+00
9.11713481e-01 -2.88077556e-02 -3.81113976e-01 1.05509961e+00
-2.09062966e-03 4.51537728e-01 5.17642319e-01 -7.56544709e-01
2.30878636e-01 3.49973828e-01 -1.02452077e-01 -6.46429420e-01
3.07976067e-01 -3.70069593e-01 -8.08364451e-01 3.34833473e-01
1.02389142e-01 3.59828800e-01 -1.20559323e+00 1.91617620e+00
1.60771966e-01 1.62784114e-01 -2.09866092e-01 7.31032550e-01
4.52885896e-01 5.11251032e-01 -6.03583828e-02 8.92750323e-02
9.42578793e-01 -9.07162189e-01 -5.28591394e-01 -2.73067713e-01
6.43309876e-02 -7.50248909e-01 1.53259563e+00 4.93237048e-01
-1.02118635e+00 -9.48066771e-01 -1.26342285e+00 -1.66265652e-01
-7.84794033e-01 1.49963871e-01 5.56618154e-01 5.98436415e-01
-1.00906205e+00 7.28059113e-01 -4.72227007e-01 -8.29865158e-01
4.40211654e-01 1.80633843e-01 -3.69268328e-01 -3.60621989e-01
-9.01489377e-01 7.53100872e-01 1.53733030e-01 -1.97879940e-01
-1.00627697e+00 -7.38600016e-01 -6.61115766e-01 2.05268309e-01
3.50519091e-01 -7.50461638e-01 9.74972606e-01 -1.13683879e+00
-1.72159886e+00 7.99409389e-01 -8.12543258e-02 1.56197667e-01
6.72629535e-01 1.32998541e-01 -3.05989474e-01 1.72683880e-01
1.04669176e-01 9.07953680e-01 1.29802227e+00 -1.73822975e+00
-3.36784989e-01 -3.21398526e-01 4.70485128e-02 8.48532468e-02
-6.95937693e-01 -7.72357643e-01 -8.47009540e-01 -9.36778188e-01
6.05346635e-02 -8.92107725e-01 2.04261228e-01 7.93275476e-01
1.53893769e-01 -1.66941851e-01 9.53323901e-01 -2.86610633e-01
1.05211031e+00 -2.17919326e+00 3.89537781e-01 3.69324863e-01
-9.40709263e-02 2.19818413e-01 -6.02763355e-01 5.84603667e-01
4.35966030e-02 -3.42172533e-02 -3.54380906e-01 -7.20774055e-01
3.61010194e-01 5.83069623e-01 -6.82240248e-01 1.43275395e-01
3.91414493e-01 8.99175763e-01 -9.73845899e-01 -4.11413163e-01
3.51077765e-01 3.46697718e-01 -6.97311938e-01 5.40369451e-01
-2.88244784e-01 3.09036523e-01 -2.31077939e-01 8.02009344e-01
8.49752784e-01 -8.67814049e-02 1.95011437e-01 -3.58039021e-01
2.00446144e-01 -2.88767695e-01 -1.13903987e+00 2.33740163e+00
-6.08393431e-01 3.74897897e-01 -1.79983884e-01 -1.05736017e+00
9.49659288e-01 5.60004823e-02 1.95427492e-01 -7.50800550e-01
-3.14424574e-01 2.04352051e-01 -3.85947764e-01 -2.74175882e-01
3.41924280e-01 -3.91733646e-01 -6.67592660e-02 4.04934585e-01
6.85015559e-01 -4.64916587e-01 -7.62283802e-03 2.92717189e-01
9.55894053e-01 2.78109342e-01 -8.45657736e-02 -1.46551132e-01
4.12636578e-01 -3.82772803e-01 2.15652481e-01 1.05544782e+00
1.09121474e-02 8.62906039e-01 2.12600470e-01 -4.77944642e-01
-1.20953333e+00 -1.45529342e+00 -1.27538636e-01 1.34276307e+00
2.01695457e-01 -2.10666239e-01 -2.49564826e-01 -1.05382824e+00
2.20199332e-01 7.48051286e-01 -8.45967114e-01 -4.74741876e-01
-3.11700076e-01 -5.82478195e-02 1.74214318e-01 3.63437206e-01
2.54318565e-01 -8.49779606e-01 -2.64046788e-01 6.10482171e-02
2.69069135e-01 -7.24177480e-01 -7.84587920e-01 1.20145530e-01
-9.16550338e-01 -8.87914777e-01 -1.15873539e+00 -8.01283062e-01
7.07944632e-01 3.81265312e-01 1.16022921e+00 -8.20934772e-03
-2.54513681e-01 1.11546361e+00 -3.87472689e-01 -5.22761643e-02
-3.35507900e-01 1.33799821e-01 -1.44628793e-01 3.94925445e-01
1.24168389e-01 -8.82253289e-01 -7.02684939e-01 2.70459622e-01
-1.26920354e+00 -3.23845260e-02 9.08270895e-01 1.13883162e+00
3.46126646e-01 -5.30289114e-01 5.74132979e-01 -7.91312635e-01
3.64392340e-01 -2.94318169e-01 -2.76394397e-01 8.19168806e-01
-7.83027112e-01 3.74722958e-01 5.20839512e-01 -1.02188790e+00
-1.00750947e+00 4.84148599e-03 -2.40286980e-02 -9.37292218e-01
-5.90985501e-03 7.02109486e-02 -3.36580306e-01 -1.26760975e-01
5.68796813e-01 3.65550309e-01 4.27358113e-02 -4.92686719e-01
7.14449465e-01 5.86177111e-01 4.48608905e-01 -8.33918273e-01
1.25845659e+00 6.41037405e-01 4.86313278e-04 -7.90669203e-01
-6.27495348e-01 -5.86621523e-01 -7.33334541e-01 -3.07761878e-01
5.09297132e-01 -7.04074800e-01 -4.14578855e-01 1.61889344e-01
-1.07909524e+00 -4.35938030e-01 -6.34962797e-01 5.75639121e-03
-7.20670462e-01 4.40061837e-01 -2.80949771e-01 -6.79907024e-01
-2.72591650e-01 -8.35892141e-01 1.28399217e+00 -1.40018925e-01
-1.32558923e-02 -1.00626242e+00 2.83942133e-01 -2.19522700e-01
6.29529238e-01 -9.51725617e-02 1.09736109e+00 -4.47027534e-01
-7.12147832e-01 -3.46032470e-01 -2.23156750e-01 6.03468180e-01
1.38643265e-01 -2.64469475e-01 -1.20825112e+00 -6.30457580e-01
-8.03030282e-02 -6.49549305e-01 1.04636979e+00 -1.98057100e-01
1.39437044e+00 -2.80520350e-01 -1.01968855e-01 6.60819471e-01
1.53094220e+00 -4.40707579e-02 7.37440467e-01 -7.03354226e-03
7.10788608e-01 4.62690264e-01 5.41217089e-01 3.38380158e-01
1.17533751e-01 9.99002874e-01 1.91242248e-01 8.94650221e-02
-5.31582952e-01 -6.72909617e-01 3.48389268e-01 8.85994494e-01
2.95593858e-01 -2.11629316e-01 -4.70244378e-01 6.52726889e-01
-1.79423892e+00 -8.48126054e-01 7.26601422e-01 2.56493092e+00
9.18637156e-01 -1.34914383e-01 -1.18951045e-01 -1.31175101e-01
4.57930356e-01 2.01145858e-01 -7.19862401e-01 -2.11260185e-01
1.66732594e-02 6.23113453e-01 1.65577844e-01 3.99313033e-01
-8.26538980e-01 9.57183301e-01 5.49053192e+00 1.07620597e+00
-1.21696723e+00 2.28227869e-01 2.64341563e-01 -4.57822829e-02
-6.13185346e-01 1.85091510e-01 -2.90739387e-01 3.23650360e-01
4.60476041e-01 -2.91791908e-03 6.93091094e-01 8.80434096e-01
-6.35618627e-01 3.77247572e-01 -1.64140737e+00 1.14768457e+00
5.31926930e-01 -1.19303942e+00 6.89552069e-01 -7.53118470e-02
6.91941082e-01 -4.40853357e-01 4.74635690e-01 7.56734550e-01
-1.65690005e-01 -5.45175970e-01 8.96258235e-01 8.15209270e-01
1.26864636e+00 -4.80700225e-01 2.70346999e-01 1.18381761e-01
-1.10362148e+00 1.89499054e-02 -5.84706426e-01 3.72651726e-01
-2.09457651e-01 1.11766055e-01 -6.92683458e-01 4.02011991e-01
3.42051178e-01 7.16874003e-01 -1.01642966e+00 8.71962965e-01
-1.78031474e-02 1.64751306e-01 -3.08101345e-02 1.32943451e-01
3.56690511e-02 -8.37591439e-02 5.21576643e-01 1.08366787e+00
6.47583604e-01 -1.45171747e-01 4.23747487e-02 9.47377980e-01
-7.54303038e-02 -2.72745192e-01 -7.77196825e-01 -6.73596188e-02
3.26130778e-01 1.13368142e+00 -3.88566822e-01 -5.83569348e-01
-3.06912988e-01 1.70879984e+00 3.79601061e-01 5.64586461e-01
-4.92102891e-01 -3.30959201e-01 5.41493118e-01 1.35255054e-01
5.34033835e-01 -6.43045679e-02 -8.25518370e-02 -1.54435837e+00
3.91445793e-02 -6.65773809e-01 3.86880904e-01 -8.36377263e-01
-1.79434383e+00 3.82090867e-01 -3.19452994e-02 -1.44485915e+00
-2.28133544e-01 -6.80386126e-01 -6.39050424e-01 5.88675499e-01
-1.55079103e+00 -1.66816509e+00 -3.55707794e-01 6.41995192e-01
8.96234572e-01 -3.48870993e-01 1.01272523e+00 4.37510341e-01
-1.45025477e-02 1.04567409e+00 1.30589232e-01 -1.25559464e-01
1.33891773e+00 -1.33637691e+00 2.40023196e-01 6.36805177e-01
5.21787167e-01 5.93270063e-01 6.24034882e-01 -4.13938731e-01
-1.70208752e+00 -9.71069098e-01 5.41555762e-01 -6.52885199e-01
5.48925996e-01 -7.83906579e-01 -1.05421627e+00 5.10370731e-01
7.28465021e-02 1.28152654e-01 4.77307886e-01 2.51639634e-01
-8.23992431e-01 -5.37386775e-01 -1.10530531e+00 5.62192500e-01
1.23548877e+00 -9.91679370e-01 -5.46667337e-01 1.04421861e-01
3.92486185e-01 -7.27467909e-02 -6.24053061e-01 3.26742858e-01
8.90118420e-01 -7.22107351e-01 1.08748507e+00 -6.85395718e-01
2.83667237e-01 -6.33111298e-02 -4.07095164e-01 -1.23691499e+00
-4.33475167e-01 -2.89157838e-01 -2.37327576e-01 1.35005593e+00
5.93587197e-02 -2.89575994e-01 8.07062685e-01 5.75432897e-01
-5.89944376e-03 -5.05218804e-01 -1.06839871e+00 -9.44982469e-01
1.25672519e-01 -2.52426505e-01 4.76429880e-01 9.19727266e-01
-2.34745830e-01 3.93143624e-01 -4.25059617e-01 -9.90642793e-03
7.85458565e-01 1.67379647e-01 1.10827744e+00 -1.32017970e+00
-4.73982871e-01 -5.13219595e-01 -4.33590442e-01 -1.14823163e+00
2.53535956e-01 -7.98643053e-01 -6.76226243e-02 -1.22627389e+00
2.90386140e-01 -5.60331702e-01 -5.09190679e-01 2.85277665e-01
-1.17914006e-01 3.43058556e-01 4.68496919e-01 3.49027097e-01
-6.81663930e-01 9.19017136e-01 1.15366924e+00 -6.07634842e-01
-8.73612165e-02 2.40703486e-02 -4.60079461e-01 1.78737938e-01
3.42442274e-01 -2.96795547e-01 -6.86257422e-01 -4.14783508e-01
1.56545341e-01 -1.72531322e-01 5.54643571e-01 -9.71713543e-01
2.82888174e-01 -7.73397163e-02 5.14955163e-01 -3.72238934e-01
6.81453407e-01 -1.03682661e+00 6.24989979e-02 1.52842253e-01
-4.49443549e-01 -2.86281914e-01 2.02878624e-01 8.80451262e-01
-6.35806322e-02 -2.11277112e-01 7.92386353e-01 -1.25348911e-01
-8.73653114e-01 5.89423299e-01 7.43277892e-02 -3.01368773e-01
6.46572590e-01 -3.41755748e-01 -1.44312277e-01 -4.71924096e-01
-7.63383389e-01 -1.91462543e-02 7.63525248e-01 6.58574402e-01
8.96251857e-01 -1.65674269e+00 -4.26979393e-01 3.75339091e-01
7.44496703e-01 -3.75390619e-01 4.98280048e-01 2.65307128e-01
2.54089087e-02 7.94179812e-02 -3.39154959e-01 -6.10207617e-01
-8.65381002e-01 1.01695216e+00 -3.80375818e-03 -1.74047910e-02
-4.18387204e-01 6.91631913e-01 4.96980369e-01 -6.51915967e-01
3.87220293e-01 2.15821415e-02 2.54583120e-01 1.85766906e-01
3.17332774e-01 1.65459402e-02 -4.15997133e-02 -1.60051480e-01
-8.33783746e-02 7.23529100e-01 -1.27479494e-01 -5.11177957e-01
1.19200981e+00 -1.37060612e-01 1.50401905e-01 9.49507833e-01
1.35742533e+00 -2.60617137e-01 -1.42265677e+00 -2.96137154e-01
-1.85220167e-01 -8.11233103e-01 -2.31774181e-01 -1.05137718e+00
-8.58223081e-01 1.07830656e+00 9.70127642e-01 -1.77743033e-01
1.07252932e+00 1.39263123e-01 5.63839257e-01 6.31017506e-01
5.07553220e-01 -1.17435360e+00 6.85562909e-01 2.89857119e-01
1.26158845e+00 -1.37029946e+00 -4.73878346e-02 1.11050606e-01
-5.72920978e-01 1.24993598e+00 5.20253181e-01 -1.74791023e-01
5.39245069e-01 -3.28805774e-01 -1.86180472e-01 -9.93018895e-02
-7.94717193e-01 -1.99288517e-01 5.53180039e-01 7.03798115e-01
-7.05513209e-02 -5.94708845e-02 2.60917656e-02 2.81894952e-01
2.47088999e-01 3.70228179e-02 1.83283687e-01 8.44198763e-01
-1.59257203e-01 -1.51336336e+00 -3.14815231e-02 4.17985827e-01
3.78981858e-01 1.12390421e-01 -4.64110255e-01 7.95781672e-01
2.21833542e-01 4.87273067e-01 1.35869980e-01 -3.51179957e-01
4.73202139e-01 3.67328346e-01 8.80959272e-01 -7.80432761e-01
-1.25226855e-01 -1.15572333e-01 -4.81464386e-01 -5.08449793e-01
-1.84635997e-01 -4.29041147e-01 -3.52462649e-01 -9.86323580e-02
-1.99074119e-01 3.73216234e-02 6.90154254e-01 6.42677307e-01
3.67934883e-01 2.88084328e-01 8.12766790e-01 -1.12517798e+00
-8.28017175e-01 -8.95884275e-01 -6.49208307e-01 6.71077490e-01
3.80506366e-01 -7.66195118e-01 -4.72491622e-01 1.04876617e-02] | [11.598742485046387, 0.6689432263374329] |
2c05782f-9791-4a6d-a5a7-4edf964a4be2 | improving-visual-relationship-detection-using | 1809.00204 | null | http://arxiv.org/abs/1809.00204v1 | http://arxiv.org/pdf/1809.00204v1.pdf | Improving Visual Relationship Detection using Semantic Modeling of Scene Descriptions | Structured scene descriptions of images are useful for the automatic
processing and querying of large image databases. We show how the combination
of a semantic and a visual statistical model can improve on the task of mapping
images to their associated scene description. In this paper we consider scene
descriptions which are represented as a set of triples (subject, predicate,
object), where each triple consists of a pair of visual objects, which appear
in the image, and the relationship between them (e.g. man-riding-elephant,
man-wearing-hat). We combine a standard visual model for object detection,
based on convolutional neural networks, with a latent variable model for link
prediction. We apply multiple state-of-the-art link prediction methods and
compare their capability for visual relationship detection. One of the main
advantages of link prediction methods is that they can also generalize to
triples, which have never been observed in the training data. Our experimental
results on the recently published Stanford Visual Relationship dataset, a
challenging real world dataset, show that the integration of a semantic model
using link prediction methods can significantly improve the results for visual
relationship detection. Our combined approach achieves superior performance
compared to the state-of-the-art method from the Stanford computer vision
group. | ['Stephan Baier', 'Yunpu Ma', 'Volker Tresp'] | 2018-09-01 | null | null | null | null | ['visual-relationship-detection'] | ['computer-vision'] | [ 1.35034651e-01 2.78731193e-02 -3.18838447e-01 -4.96029139e-01
-2.80717701e-01 -3.04216534e-01 9.86410975e-01 6.48365438e-01
-9.54256654e-02 4.59053904e-01 -1.21597829e-03 -1.40301570e-01
-1.63444638e-01 -8.94751906e-01 -1.01300323e+00 -2.76883632e-01
-9.78674591e-02 7.26949513e-01 8.83463025e-01 -1.25702977e-01
2.38958020e-02 4.79483485e-01 -2.08413363e+00 7.25443423e-01
2.04980254e-01 1.07197320e+00 4.52564001e-01 4.68144357e-01
-3.28009784e-01 1.16829729e+00 -2.06472605e-01 -5.57548583e-01
-1.28824130e-01 -1.02099821e-01 -8.96497965e-01 1.93097398e-01
1.04587710e+00 -1.13757975e-01 -3.49293292e-01 1.06048715e+00
-1.75214037e-02 -4.70843539e-02 7.17578709e-01 -1.78015184e+00
-8.10699821e-01 4.70453650e-01 -4.92790163e-01 1.02194004e-01
6.56299353e-01 -4.54593599e-01 1.18314397e+00 -8.78907084e-01
1.16100311e+00 1.51382661e+00 5.50099313e-01 1.82721823e-01
-1.40309119e+00 -4.48867053e-01 4.58778581e-03 8.88351321e-01
-1.49848056e+00 -2.83068746e-01 7.23025560e-01 -8.39063227e-01
1.08681357e+00 1.90368265e-01 6.48287237e-01 6.39917672e-01
-6.88585266e-02 8.35640371e-01 7.19908357e-01 -5.90607047e-01
-7.16638938e-02 5.30376971e-01 1.25599205e-01 1.05638838e+00
2.15226978e-01 -1.66905731e-01 -6.68459415e-01 -3.51193286e-02
5.85625768e-01 -8.61520544e-02 -8.35202262e-02 -1.34089994e+00
-1.30581474e+00 6.53172791e-01 8.71086121e-01 3.10394645e-01
-6.34065196e-02 3.59473407e-01 3.80635530e-01 -4.43405136e-02
3.29228878e-01 1.12142076e-03 -2.20264986e-01 5.20856082e-01
-8.55045557e-01 2.64665574e-01 7.44621038e-01 1.23014593e+00
8.93351376e-01 -5.03071249e-01 -2.65789449e-01 6.35181963e-01
4.52999383e-01 1.65391341e-01 -8.61784667e-02 -8.78922522e-01
5.61268747e-01 8.52166057e-01 6.88622966e-02 -1.23007798e+00
-3.37296814e-01 1.44366613e-02 -4.56831843e-01 2.87327230e-01
3.09256285e-01 7.22513378e-01 -7.45641947e-01 1.49081290e+00
3.25046718e-01 1.70077547e-01 1.13738649e-01 7.90673971e-01
1.42797780e+00 4.73103911e-01 3.48224074e-01 -1.74648046e-01
1.66245115e+00 -1.04054308e+00 -7.32346117e-01 -1.56711087e-01
5.06069243e-01 -7.95887411e-01 6.80885911e-01 -9.97702777e-02
-9.67479289e-01 -6.91916704e-01 -1.01257241e+00 -2.87500590e-01
-9.10188735e-01 2.79883146e-01 6.73586905e-01 2.02866182e-01
-1.05894351e+00 3.59130710e-01 -5.04340172e-01 -1.05643260e+00
5.44117987e-01 2.69088387e-01 -8.28742325e-01 -8.41737986e-02
-9.97912169e-01 8.88870299e-01 8.12051415e-01 -2.32586667e-01
-8.58766377e-01 -3.67804736e-01 -1.05549788e+00 6.92649707e-02
5.61016023e-01 -9.46096778e-01 8.27881336e-01 -8.30447495e-01
-6.21535301e-01 1.62299192e+00 -1.77000716e-01 -4.47712541e-01
5.25583446e-01 5.69709502e-02 -4.14266974e-01 4.01828468e-01
4.32086766e-01 1.00378990e+00 6.10750854e-01 -1.83078051e+00
-7.81038344e-01 -1.67867124e-01 2.02467754e-01 1.20420173e-01
-2.03587994e-01 3.47653240e-01 -9.18481112e-01 -9.97495055e-02
1.28771976e-01 -7.87736058e-01 1.72701076e-01 5.11708319e-01
-6.96317255e-01 -3.24553370e-01 1.04181254e+00 -3.63135457e-01
7.56660342e-01 -2.08988309e+00 1.68173924e-01 1.44569620e-01
2.59879172e-01 2.47292802e-01 -8.84792767e-03 4.85141188e-01
-3.64957988e-01 6.72704577e-02 9.46177542e-03 -5.31015098e-01
-7.38099292e-02 4.08086270e-01 -2.12879002e-01 4.61951315e-01
-9.59581602e-03 9.42786992e-01 -9.76562500e-01 -1.09330761e+00
3.35641623e-01 4.22226369e-01 -9.76312384e-02 3.39994788e-01
-3.28353524e-01 1.92401916e-01 -1.34029299e-01 4.84802395e-01
4.08052117e-01 -4.77906048e-01 3.13852251e-01 -7.04136431e-01
-9.48446244e-02 -5.50408207e-04 -1.12418365e+00 1.60550606e+00
-7.22636357e-02 9.50044811e-01 -2.92681217e-01 -9.55946326e-01
1.03056920e+00 1.58242911e-01 3.63766134e-01 -5.95652282e-01
-1.14072166e-01 -6.77407458e-02 -4.05181050e-01 -7.88812160e-01
4.13353622e-01 4.30952795e-02 1.29773647e-01 1.63105177e-03
1.60897836e-01 -4.85166302e-03 5.84322751e-01 5.74113131e-01
7.13899910e-01 4.76451516e-01 5.32498002e-01 -1.51376203e-01
6.85922325e-01 2.76393354e-01 3.08144957e-01 7.93856502e-01
-8.36055055e-02 5.24548233e-01 8.53196740e-01 -4.97881442e-01
-1.15521431e+00 -1.30099928e+00 -1.47794396e-01 1.00250638e+00
5.86131752e-01 -5.56580961e-01 -3.31785351e-01 -6.15957975e-01
1.68469250e-01 5.21418333e-01 -6.53288960e-01 2.23918650e-02
-4.00621593e-01 -2.40609124e-01 1.24882944e-01 6.04422331e-01
5.10668457e-01 -9.88310575e-01 -6.51543319e-01 -2.15500921e-01
-1.23186536e-01 -1.68908226e+00 5.74827418e-02 2.33087945e-03
-4.88450319e-01 -1.21547651e+00 -5.68212159e-02 -1.02884614e+00
6.94721639e-01 2.46522382e-01 1.44488883e+00 3.21230263e-01
-4.65179503e-01 6.69271410e-01 -2.15554431e-01 -2.74053127e-01
-4.57256138e-01 -4.53026205e-01 -3.39900516e-02 1.88320711e-01
2.64047116e-01 -2.06127375e-01 -2.89797276e-01 2.18318373e-01
-8.08462858e-01 3.70170951e-01 2.97153234e-01 6.02161348e-01
6.89914644e-01 6.98614791e-02 -1.96203142e-01 -9.01746571e-01
-5.40834591e-02 -4.80644435e-01 -6.80661380e-01 7.49694347e-01
-3.82916898e-01 1.81608871e-01 1.62032500e-01 -2.27067128e-01
-9.36004579e-01 5.34622192e-01 4.71225232e-01 -6.67340398e-01
-2.41547808e-01 3.72510582e-01 -3.11287761e-01 -9.78332479e-03
4.88285840e-01 -1.48751158e-02 -1.17106304e-01 -4.39747244e-01
6.90581322e-01 2.52047658e-01 8.04452002e-01 -6.07366301e-02
9.63063955e-01 7.53423274e-01 5.71752250e-01 -6.40478253e-01
-9.24387872e-01 -9.37855184e-01 -1.20033777e+00 -4.60331947e-01
1.23718798e+00 -9.38285232e-01 -5.71509063e-01 4.98829549e-03
-1.41239071e+00 -2.97979470e-02 2.60006581e-02 3.64946276e-01
-7.37973630e-01 4.81199741e-01 -2.42907345e-01 -4.47278738e-01
8.99284631e-02 -8.56868386e-01 1.11474097e+00 1.73677634e-02
1.44075096e-01 -1.15913177e+00 -1.58365428e-01 5.54272652e-01
-4.94768955e-02 2.86138684e-01 1.23673332e+00 -6.54697239e-01
-8.78261566e-01 -6.08313642e-02 -7.37439871e-01 -9.38067853e-04
-1.01081483e-01 1.29887238e-01 -8.64611447e-01 3.78224067e-02
-7.31451988e-01 -3.01891923e-01 9.79234815e-01 4.64040376e-02
9.56812620e-01 -6.44923374e-03 -9.60767686e-01 3.80125612e-01
1.79771781e+00 -5.91744892e-02 6.31803453e-01 3.78740311e-01
1.05655265e+00 8.22501719e-01 7.69744217e-01 1.06976308e-01
6.06875420e-01 1.27489316e+00 9.09469903e-01 -3.42330039e-01
-5.36468506e-01 -3.32719237e-01 -2.66010035e-03 1.00516655e-01
-7.83874542e-02 -1.47215381e-01 -1.21230733e+00 6.86130822e-01
-2.41757178e+00 -1.12427068e+00 -7.67926753e-01 2.11780405e+00
5.50125718e-01 3.91324721e-02 1.29435599e-01 -1.57465041e-01
9.29258823e-01 -1.01728186e-01 -4.99685556e-02 -1.78425536e-01
-1.30491048e-01 -4.08864498e-01 5.36842644e-01 3.77591014e-01
-1.51496422e+00 1.06978345e+00 6.34977484e+00 5.88408053e-01
-7.18044758e-01 2.25067601e-01 2.36844078e-01 3.30873817e-01
-1.64319083e-01 1.56026691e-01 -9.90699828e-01 -9.21316445e-02
4.53789353e-01 -3.45628113e-02 1.34910032e-01 9.08529401e-01
-8.61978233e-02 -2.71767050e-01 -1.49378550e+00 1.09312248e+00
6.10115170e-01 -1.56660736e+00 2.90109813e-01 -1.57055750e-01
4.82766271e-01 -3.01660419e-01 -3.77901532e-02 -1.84797763e-03
8.77695978e-02 -9.33661103e-01 1.08436823e+00 8.14941764e-01
6.44130290e-01 -3.42046529e-01 7.17127204e-01 1.19626381e-01
-1.57903600e+00 -3.19961621e-03 -4.46606159e-01 1.10283993e-01
9.89544615e-02 2.79322803e-01 -1.11856699e+00 6.27627254e-01
8.95535409e-01 1.05994391e+00 -1.08317542e+00 1.16443014e+00
-4.28221643e-01 -3.49579751e-02 -8.20486099e-02 1.32024940e-02
-5.80773614e-02 7.15927258e-02 5.79001546e-01 1.17566037e+00
-1.70925520e-02 -3.20270509e-01 3.94282043e-01 1.17822289e+00
5.70079610e-02 1.73871785e-01 -1.00035751e+00 3.39102596e-01
3.00039202e-01 1.34506679e+00 -9.43277717e-01 -5.04927874e-01
-7.15459645e-01 8.13068569e-01 5.77720523e-01 2.09798366e-01
-8.60195220e-01 -7.67606273e-02 3.70894283e-01 3.63872737e-01
5.65026045e-01 -1.30759493e-01 1.49778426e-01 -9.32197690e-01
7.64331967e-02 -1.33373350e-01 5.27889609e-01 -1.34500670e+00
-1.24167645e+00 5.32313168e-01 5.99823236e-01 -1.34715927e+00
-2.04620764e-01 -7.29287505e-01 -3.54094595e-01 5.29782116e-01
-1.58791912e+00 -1.88777506e+00 -5.42125523e-01 7.39515781e-01
2.23639503e-01 -1.65321350e-01 7.55279303e-01 2.73935467e-01
-2.39582449e-01 1.20426238e-01 -2.66426831e-01 2.47783437e-01
5.57600081e-01 -1.30818856e+00 1.19202122e-01 7.62881517e-01
6.58909321e-01 3.11667114e-01 8.65606904e-01 -7.12371469e-01
-1.07694364e+00 -1.30142725e+00 1.20077825e+00 -6.70267403e-01
8.58817577e-01 -6.05611384e-01 -9.24096763e-01 8.44472706e-01
2.09107831e-01 4.15281147e-01 4.01540518e-01 5.81394918e-02
-7.21763432e-01 -9.17482749e-02 -7.92549610e-01 4.55424368e-01
1.25943887e+00 -6.60014212e-01 -7.35159397e-01 5.85672140e-01
6.41906500e-01 -1.94526270e-01 -7.48867333e-01 4.00917292e-01
4.38706338e-01 -1.13113570e+00 1.44028497e+00 -6.99054539e-01
5.26413262e-01 -6.01091325e-01 -3.08591515e-01 -7.66661763e-01
-2.09813103e-01 9.58292037e-02 -3.56679857e-02 1.44216430e+00
2.45563954e-01 7.31474981e-02 5.76385438e-01 3.37151289e-01
1.42056644e-01 -3.22053581e-01 -8.04205596e-01 -9.06194806e-01
-4.28808719e-01 -3.12096685e-01 1.35245278e-01 8.16150486e-01
-3.52361165e-02 5.08401215e-01 -4.17534232e-01 3.01017582e-01
8.80641758e-01 3.82514060e-01 7.99973249e-01 -1.43566918e+00
-2.26206079e-01 -2.04611346e-01 -1.15493464e+00 -6.20545626e-01
4.31245297e-01 -1.16724920e+00 -3.07946764e-02 -2.17561126e+00
6.55755699e-01 -2.68907696e-01 -9.35443267e-02 8.27769458e-01
-6.38666144e-03 5.44182837e-01 4.77878749e-01 2.99364418e-01
-1.11408806e+00 2.62815535e-01 7.66166806e-01 -3.89368862e-01
5.74459247e-02 -4.33968365e-01 -2.39099767e-02 8.38962078e-01
3.13923091e-01 -5.71076214e-01 -3.16597462e-01 -1.65481851e-01
4.87484694e-01 4.37894985e-02 9.17323291e-01 -9.88481343e-01
4.62087482e-01 5.59286959e-02 2.69563407e-01 -7.02634275e-01
5.25795817e-01 -1.08801973e+00 3.95918041e-01 5.74620306e-01
-3.51029485e-01 -1.63764402e-01 1.12214625e-01 7.72728443e-01
-3.60412419e-01 -2.11164206e-01 6.50570571e-01 -2.10231960e-01
-1.44200659e+00 7.39919916e-02 -5.38840368e-02 -5.29336989e-01
1.43624878e+00 -2.08890155e-01 -6.17942095e-01 -3.38525742e-01
-9.35485840e-01 2.89527535e-01 4.18895006e-01 7.15904713e-01
7.88250268e-01 -1.50852346e+00 -5.76305866e-01 -1.06607951e-01
8.92168403e-01 -5.71749151e-01 -4.92381491e-02 8.26439023e-01
-5.02350867e-01 3.83221179e-01 -4.87277925e-01 -9.80997980e-01
-1.92606628e+00 1.06318855e+00 2.79080302e-01 -2.66646147e-02
-6.38889730e-01 6.22078657e-01 3.22375506e-01 -1.91433802e-01
2.89913207e-01 -1.38409017e-02 -7.18022406e-01 2.79411256e-01
2.32198969e-01 2.70265173e-02 -1.48685396e-01 -1.15111017e+00
-7.18184054e-01 8.84946048e-01 1.99338123e-01 2.21837386e-01
1.15041208e+00 -1.86996847e-01 -5.23647308e-01 7.61067331e-01
1.28884852e+00 -2.63243049e-01 -7.98538327e-01 -4.09643322e-01
9.45422277e-02 -5.29140651e-01 2.13661138e-02 -4.96458173e-01
-9.28136110e-01 6.86460197e-01 7.17877269e-01 4.45560992e-01
8.34221303e-01 8.74472678e-01 2.72141490e-02 3.15444827e-01
4.61835772e-01 -6.78060770e-01 3.15968007e-01 1.52311817e-01
8.69447887e-01 -1.45761418e+00 3.54812562e-01 -1.15524340e+00
-6.16374075e-01 1.19631183e+00 4.85309660e-01 1.08775459e-01
6.36963487e-01 -1.03798106e-01 -1.53355703e-01 -6.40618563e-01
-8.52219105e-01 -8.56615305e-01 6.74983859e-01 7.48703420e-01
1.99191108e-01 7.28352591e-02 -8.32869783e-02 -3.20107751e-02
2.69510061e-01 -1.97219178e-01 4.16303009e-01 7.37587333e-01
-3.56406897e-01 -1.17779529e+00 -3.43608975e-01 2.33145788e-01
-5.58549501e-02 -1.16789527e-01 -5.24780869e-01 9.99319017e-01
5.69336891e-01 8.48920524e-01 4.70811129e-01 -1.43365175e-01
3.27099562e-01 -3.85545455e-02 5.99192321e-01 -7.69356370e-01
-2.39912763e-01 -2.19787747e-01 3.58177602e-01 -8.44621062e-01
-1.05538785e+00 -7.39615917e-01 -1.27100563e+00 1.07266329e-01
-8.10053274e-02 -2.75636643e-01 7.15018034e-01 9.97898281e-01
-1.01314196e-02 3.22752386e-01 9.17617828e-02 -5.76667607e-01
2.66696095e-01 -4.68617320e-01 -6.26305699e-01 9.85572994e-01
2.07013905e-01 -1.08693624e+00 -3.92222144e-02 7.04972029e-01] | [10.267148971557617, 1.6766589879989624] |
67092234-1f89-4039-b7b2-ba5d84196fb2 | matroids-hitting-sets-and-unsupervised | 1705.08992 | null | http://arxiv.org/abs/1705.08992v2 | http://arxiv.org/pdf/1705.08992v2.pdf | Matroids Hitting Sets and Unsupervised Dependency Grammar Induction | This paper formulates a novel problem on graphs: find the minimal subset of
edges in a fully connected graph, such that the resulting graph contains all
spanning trees for a set of specifed sub-graphs. This formulation is motivated
by an un-supervised grammar induction problem from computational linguistics.
We present a reduction to some known problems and algorithms from graph theory,
provide computational complexity results, and describe an approximation
algorithm. | ['Vahab Mirrokni', 'Leonid Peshkin', 'Virginia Savova', 'Nicholas Harvey', 'David Karger'] | 2017-05-24 | null | null | null | null | ['dependency-grammar-induction'] | ['natural-language-processing'] | [ 6.17351174e-01 8.97196651e-01 -5.23899913e-01 -3.97998005e-01
-2.47320473e-01 -7.11442471e-01 1.93613559e-01 3.50369304e-01
1.28637671e-01 7.40229487e-01 -2.48111412e-01 -6.92530751e-01
-5.62047362e-01 -1.15058970e+00 -5.98154187e-01 -4.12455380e-01
-7.01882958e-01 8.63673508e-01 2.75515258e-01 -4.48476933e-02
1.25661761e-01 4.62457031e-01 -1.39744091e+00 -2.26884007e-01
8.16047668e-01 2.24532261e-01 1.75341964e-01 7.76952446e-01
-3.04227203e-01 5.58258355e-01 -3.33962440e-01 -4.48706299e-01
1.07121408e-01 -7.05617428e-01 -1.38140583e+00 9.45892036e-01
3.86496395e-01 4.55743253e-01 -4.02848482e-01 1.41690576e+00
3.85623425e-02 1.90869113e-03 7.11601555e-01 -1.59497285e+00
-5.25386274e-01 1.09947193e+00 -4.45414960e-01 2.58932233e-01
6.93017304e-01 -6.39271557e-01 1.56496608e+00 -1.45741940e-01
8.80415618e-01 1.12222302e+00 3.55963677e-01 4.36067909e-01
-1.25620711e+00 -5.86791784e-02 3.33536565e-01 1.42619178e-01
-1.47992611e+00 3.77874188e-02 7.73610175e-01 -1.99838892e-01
1.03331089e+00 5.45558751e-01 5.70117950e-01 3.85735363e-01
7.25594396e-03 3.35556507e-01 1.13232481e+00 -1.09436452e+00
1.56708658e-01 -1.37307778e-01 7.00649202e-01 1.45140636e+00
9.79532361e-01 -8.49095825e-03 -2.03677952e-01 -9.93472040e-02
5.78594923e-01 -5.55830657e-01 -2.33997956e-01 -3.45197320e-01
-5.31897783e-01 9.53769326e-01 3.03608347e-02 6.53479815e-01
-6.98978454e-02 7.72551000e-02 9.57232341e-02 4.27404493e-01
1.95389479e-01 3.54587704e-01 -6.05702698e-01 5.85732579e-01
-5.02696514e-01 -1.59253091e-01 1.34160852e+00 1.34055817e+00
1.18767822e+00 -5.24500906e-02 6.67222023e-01 3.05697024e-01
3.71237904e-01 5.23094475e-01 -1.91129595e-01 -5.32396436e-01
3.01510304e-01 1.02664626e+00 -5.08525372e-01 -1.08076489e+00
-6.37313068e-01 -2.13555560e-01 -2.96032757e-01 -3.85374069e-01
3.42476696e-01 -2.27008313e-02 -8.84290397e-01 1.72338033e+00
4.03328449e-01 -1.68920070e-01 -1.49411410e-01 4.93648976e-01
9.17690575e-01 4.75054383e-01 -1.15221158e-01 -8.25519383e-01
1.07830369e+00 -8.05738568e-01 -8.35764825e-01 -3.10818493e-01
8.66345704e-01 -2.62193233e-01 7.87092745e-01 3.30212921e-01
-8.80773723e-01 2.89388355e-02 -9.07838702e-01 1.04747079e-01
-4.30134416e-01 -2.94936091e-01 9.85638916e-01 9.54962254e-01
-1.37979555e+00 5.64577520e-01 -4.60232973e-01 -7.46822715e-01
-3.52322251e-01 6.22459710e-01 -5.87695479e-01 -1.84823170e-01
-7.64040709e-01 6.36594772e-01 1.02210414e+00 -1.42762780e-01
-6.13687217e-01 1.57529861e-01 -1.22955978e+00 -1.79029986e-01
7.36021876e-01 -5.53319752e-01 1.03708577e+00 -1.11459982e+00
-8.14786434e-01 1.43624616e+00 -3.94868761e-01 -4.13604379e-01
-3.93959135e-01 5.53280473e-01 -7.16953754e-01 4.93193984e-01
2.06379935e-01 9.99441519e-02 5.12690485e-01 -1.33375609e+00
-6.01962686e-01 -7.28664219e-01 2.80736387e-01 1.11860037e-01
-2.14081883e-01 1.39131099e-01 -7.99846292e-01 -3.23597401e-01
4.55592841e-01 -1.04709458e+00 -4.36112732e-01 -9.39732969e-01
-8.62307608e-01 -5.30213416e-01 4.81443048e-01 -4.58464622e-01
1.65929508e+00 -1.73625278e+00 3.80631626e-01 7.12206185e-01
6.19090617e-01 -2.22338036e-01 -2.72516012e-01 6.72356129e-01
-4.37351942e-01 3.86025846e-01 -3.46173286e-01 2.95180500e-01
-2.87241247e-02 9.07718480e-01 6.82376772e-02 5.28078020e-01
-2.38352686e-01 9.10416007e-01 -9.44368184e-01 -9.14004624e-01
-2.87186890e-03 -3.92507821e-01 -3.04309905e-01 6.01841025e-02
-4.75610882e-01 -1.15930021e-01 -5.60987830e-01 6.77204967e-01
4.78473753e-01 -3.14074248e-01 9.05507386e-01 3.67056429e-01
5.08647859e-02 5.07528245e-01 -1.36800444e+00 1.25557542e+00
-6.85730800e-02 2.65875578e-01 2.66622275e-01 -1.36131883e+00
1.12868476e+00 4.24396582e-02 4.57198381e-01 -1.77655175e-01
2.87490219e-01 1.33560188e-02 2.73418762e-02 -6.97444260e-01
1.50165170e-01 -7.43978247e-02 -4.34297025e-01 6.38805509e-01
1.78446367e-01 -2.72330344e-01 7.94702649e-01 7.29441404e-01
1.29965436e+00 -1.93512231e-01 6.16356611e-01 -5.98967493e-01
3.59802097e-01 3.62620026e-01 4.41741884e-01 8.46127987e-01
-6.16021603e-02 1.99747309e-01 6.98376238e-01 -5.35221756e-01
-6.29290462e-01 -7.81315506e-01 9.86605417e-03 1.08564925e+00
1.49788633e-01 -8.68432283e-01 -1.07639909e+00 -8.15517783e-01
-2.54378170e-01 3.57658803e-01 -6.37955904e-01 2.07638398e-01
-5.22224963e-01 -6.08910680e-01 9.95135084e-02 1.73691228e-01
-3.75604816e-02 -9.74653125e-01 -1.44733474e-01 1.46684051e-01
-1.19451225e-01 -1.40483725e+00 -4.79463756e-01 3.25795889e-01
-1.17204690e+00 -1.57071471e+00 2.82880187e-01 -1.55789256e+00
1.23844075e+00 3.52559447e-01 1.30034602e+00 6.44677579e-01
-1.91013411e-01 5.24040222e-01 -5.49666047e-01 -2.83350825e-01
-5.25547504e-01 3.17298293e-01 -1.33758143e-01 -3.84224474e-01
4.15928781e-01 -6.79442704e-01 3.86880964e-01 8.15296322e-02
-8.13002288e-01 -8.22093934e-02 2.64485270e-01 3.53739977e-01
9.23397422e-01 6.68651283e-01 1.94275066e-01 -1.42012465e+00
4.52279001e-01 -2.80183852e-01 -7.50228763e-01 6.64136708e-01
-9.00165439e-01 3.89714062e-01 5.45751691e-01 2.23323822e-01
-4.48664308e-01 4.26611722e-01 6.81984574e-02 2.85450101e-01
-2.17124626e-01 8.75126064e-01 -5.94273806e-01 -3.26479375e-01
4.28180426e-01 2.18637824e-01 -2.09050000e-01 -3.85199487e-01
5.75510919e-01 3.13563347e-01 4.97113645e-01 -4.54388857e-01
7.80350804e-01 2.79323637e-01 5.88133335e-01 -9.47491884e-01
-8.33227515e-01 -5.06593585e-01 -8.45951080e-01 -2.34343618e-01
5.75105011e-01 -1.84388384e-01 -3.74201566e-01 -3.43691230e-01
-1.15617585e+00 -2.18983386e-02 -1.03478298e-01 3.24215561e-01
-6.30428612e-01 7.48765051e-01 -4.03044075e-01 -8.20476532e-01
-1.70533776e-01 -6.13306403e-01 6.92713320e-01 -3.78466886e-03
-2.19016284e-01 -1.28774810e+00 4.09517556e-01 2.41217613e-01
-4.86859679e-01 4.58012640e-01 1.26243424e+00 -9.45407867e-01
-4.36311990e-01 -1.88822642e-01 3.09111159e-02 -7.78852627e-02
2.37788379e-01 -8.42556581e-02 -2.89236993e-01 -2.49456048e-01
-3.55976641e-01 1.36026338e-01 6.83029830e-01 2.84476787e-01
7.63730466e-01 -4.42767471e-01 -7.90745854e-01 3.99486959e-01
1.71169138e+00 2.68404245e-01 1.86084569e-01 1.23532275e-02
7.59678662e-01 7.19731688e-01 2.41438985e-01 -1.75286666e-01
4.92492139e-01 3.20560038e-02 3.90549839e-01 1.22540340e-01
8.61151889e-02 -5.05856633e-01 -2.48296950e-02 1.27005494e+00
-2.46062413e-01 -5.40700257e-01 -8.98116350e-01 8.09939981e-01
-1.66951895e+00 -5.53242743e-01 -7.96988547e-01 1.98846686e+00
5.29329479e-01 3.47968237e-03 3.80472898e-01 5.01899898e-01
1.18152845e+00 4.26040925e-02 2.04550847e-02 -9.46039855e-01
-4.14823800e-01 7.01263785e-01 6.20696843e-01 8.76277804e-01
-9.91924822e-01 1.28093886e+00 8.47917652e+00 4.08151776e-01
-3.33014041e-01 1.07662352e-02 8.17449689e-02 5.22474587e-01
-6.75944448e-01 5.23716450e-01 -5.02361357e-01 -1.79171324e-01
1.14814043e+00 -6.78820789e-01 4.94133264e-01 7.47436643e-01
-1.26605064e-01 -2.12293029e-01 -9.46817935e-01 4.42001760e-01
2.52429277e-01 -9.22403216e-01 1.36059210e-01 2.63480425e-01
8.94767761e-01 -2.60328054e-01 -4.49421823e-01 -1.35255292e-01
7.27234185e-01 -9.31601942e-01 7.48666152e-02 -2.39208087e-01
7.33127415e-01 -7.44691193e-01 2.69909739e-01 3.30571026e-01
-1.51709807e+00 1.58948954e-02 -3.86661112e-01 -3.27871978e-01
1.55575916e-01 5.28361201e-01 -9.01324034e-01 1.05484712e+00
2.82937139e-01 3.87025982e-01 -4.00959164e-01 8.84610295e-01
-7.57311940e-01 7.66491771e-01 -2.09821388e-01 -2.96279341e-01
1.98683515e-01 -5.06772518e-01 5.38160503e-01 1.30855179e+00
-6.87974542e-02 6.56389117e-01 4.99062687e-01 3.93441528e-01
-1.87402502e-01 5.98166764e-01 -1.16336250e+00 -3.08796734e-01
4.61027831e-01 1.25004601e+00 -1.42182267e+00 -1.84959412e-01
-5.77480972e-01 7.23376751e-01 5.09752095e-01 1.15184203e-01
-4.31240171e-01 -6.58980608e-01 -3.21564153e-02 1.82422206e-01
4.27854985e-01 -4.47516501e-01 6.51226565e-03 -8.63756657e-01
-1.25591844e-01 -6.78577542e-01 1.06133556e+00 -5.36106944e-01
-9.65004981e-01 7.06799090e-01 1.95675492e-01 -4.17233616e-01
-1.61290079e-01 -6.21918976e-01 -4.61174637e-01 3.75680119e-01
-1.05016124e+00 -8.66410017e-01 1.50651723e-01 6.80142224e-01
7.27736875e-02 1.70468599e-01 9.24468637e-01 -9.87799615e-02
-4.12348926e-01 2.22793266e-01 -4.56961036e-01 1.42978802e-01
-2.53886163e-01 -1.61234796e+00 2.80304819e-01 1.46615529e+00
5.98520935e-01 4.15803552e-01 8.15532148e-01 -8.52080107e-01
-1.85925257e+00 -1.09199452e+00 1.58802509e+00 -1.97674066e-01
7.84392953e-01 -1.30264685e-01 -5.95521808e-01 1.26233232e+00
2.40757763e-01 6.42424300e-02 6.43605769e-01 3.78066689e-01
-3.29724997e-01 2.01658949e-01 -1.02846181e+00 3.05458426e-01
1.73250508e+00 -3.81604970e-01 -7.74673760e-01 7.07772076e-01
9.06419933e-01 -4.37393725e-01 -7.74714530e-01 2.47502103e-01
8.78354684e-02 -4.73883569e-01 4.84788895e-01 -1.01303005e+00
-1.11270361e-01 -2.22608402e-01 4.40967828e-02 -1.20425689e+00
-4.25760210e-01 -1.23548126e+00 1.20186031e-01 8.30231965e-01
6.33081913e-01 -7.11435556e-01 1.00304353e+00 2.59541392e-01
-1.11314520e-01 -4.84355211e-01 -9.41247582e-01 -1.09539139e+00
-1.74165413e-01 -3.71544987e-01 4.49258119e-01 1.03198946e+00
6.67584538e-01 8.96932125e-01 -1.29575685e-01 3.57636929e-01
9.27522957e-01 7.28624821e-01 4.03883785e-01 -1.58190918e+00
-1.98295712e-01 -1.08572029e-01 -6.12283528e-01 -7.59489417e-01
8.26569259e-01 -1.51558232e+00 -3.96539979e-02 -2.05883121e+00
3.59784395e-01 -2.47046083e-01 2.11971253e-01 5.62866211e-01
1.57148093e-01 -1.82505231e-02 -1.71841368e-01 -2.68217891e-01
-7.10799277e-01 -1.71029031e-01 1.17082453e+00 4.38956980e-04
-1.58870429e-01 -1.17780797e-01 -9.19016719e-01 7.03940749e-01
7.88729191e-01 -7.65083611e-01 -5.93672872e-01 -3.96577388e-01
3.00953627e-01 4.07927424e-01 -1.64992154e-01 -4.39041764e-01
1.14902951e-01 -4.12214994e-01 -6.20922804e-01 -4.87949461e-01
-5.12009323e-01 -6.55388832e-01 2.81378895e-01 6.02660537e-01
-2.93367714e-01 4.82646614e-01 -9.87197310e-02 4.75816816e-01
-9.95536074e-02 -7.17444658e-01 4.34405565e-01 -2.28521168e-01
-8.47501636e-01 4.29561526e-01 -2.28790015e-01 2.28811488e-01
1.11465251e+00 -3.61242712e-01 -1.99887186e-01 -2.51316607e-01
-1.14695144e+00 2.58967936e-01 2.28788421e-01 5.20296879e-02
5.34966946e-01 -9.15223897e-01 -6.58738017e-01 -8.04322883e-02
1.02343835e-01 -9.62054580e-02 -2.34885871e-01 3.54640752e-01
-5.24936974e-01 5.87108254e-01 1.78325817e-01 -2.09766269e-01
-1.62975025e+00 9.53570783e-01 6.84569776e-02 -2.77128249e-01
-5.73229909e-01 7.60791540e-01 1.86128523e-02 -4.01394308e-01
-5.10634370e-02 -1.70092747e-01 -2.91445524e-01 -5.24832904e-01
1.02773175e-01 3.17107588e-01 8.90646130e-02 -8.95085096e-01
-3.85422558e-01 5.26340961e-01 1.52317464e-01 -5.03701158e-02
1.11253262e+00 -1.86720103e-01 -7.71119058e-01 3.75531286e-01
1.16555369e+00 4.24755625e-02 -1.78683355e-01 -4.67216760e-01
4.87931609e-01 -1.90967306e-01 -1.61107272e-01 -2.72935539e-01
-1.10461807e+00 1.80529878e-01 -2.34232858e-01 9.24680233e-01
1.35882282e+00 7.06267476e-01 4.21169609e-01 5.09393275e-01
6.65033758e-01 -9.33633327e-01 -3.85015041e-01 5.20897925e-01
5.52617848e-01 -7.26169884e-01 3.86987403e-02 -1.42119002e+00
-2.86876500e-01 1.13449883e+00 3.27946991e-01 -1.84704378e-01
8.15768003e-01 4.54938024e-01 -4.20616865e-01 -6.04632318e-01
-8.60984921e-01 -8.61962259e-01 6.79160729e-02 9.88086700e-01
3.41883123e-01 4.04816478e-01 -9.05696690e-01 3.85839105e-01
-5.51489949e-01 -2.14655235e-01 7.69443393e-01 9.56727326e-01
-8.44606280e-01 -1.31453109e+00 -1.45090252e-01 2.81984508e-01
-3.65269929e-01 -9.14455801e-02 -1.04995883e+00 8.99445653e-01
1.82338938e-01 1.24508417e+00 1.40555687e-02 -3.65847737e-01
3.26858252e-01 -1.22683696e-01 1.18665636e+00 -1.09634364e+00
-1.24048382e-01 1.39427809e-02 6.68887079e-01 -1.28587857e-01
-6.79929614e-01 -6.35574341e-01 -1.43182516e+00 -3.17521065e-01
-7.79322684e-01 6.65770292e-01 3.91096652e-01 1.03027403e+00
5.52933011e-03 1.92200735e-01 7.77189136e-01 2.93721780e-02
-8.81642401e-02 -6.90294683e-01 -1.17170930e+00 2.19858170e-01
-1.70812935e-01 -2.66159505e-01 -4.70289558e-01 1.29190087e-01] | [7.066451549530029, 5.3870930671691895] |
fe49393e-479d-4341-8c7b-dd2d8c4fe62b | tree-structured-parzen-estimator | 2304.11127 | null | https://arxiv.org/abs/2304.11127v3 | https://arxiv.org/pdf/2304.11127v3.pdf | Tree-Structured Parzen Estimator: Understanding Its Algorithm Components and Their Roles for Better Empirical Performance | Recent advances in many domains require more and more complicated experiment design. Such complicated experiments often have many parameters, which necessitate parameter tuning. Tree-structured Parzen estimator (TPE), a Bayesian optimization method, is widely used in recent parameter tuning frameworks. Despite its popularity, the roles of each control parameter and the algorithm intuition have not been discussed so far. In this tutorial, we will identify the roles of each control parameter and their impacts on hyperparameter optimization using a diverse set of benchmarks. We compare our recommended setting drawn from the ablation study with baseline methods and demonstrate that our recommended setting improves the performance of TPE. Our TPE implementation is available at https://github.com/nabenabe0928/tpe/tree/single-opt. | ['Shuhei Watanabe'] | 2023-04-21 | null | null | null | null | ['hyperparameter-optimization'] | ['methodology'] | [-3.30536783e-01 -4.86388355e-01 -5.43719411e-01 -3.48104417e-01
-9.23297882e-01 -7.35177219e-01 4.53962952e-01 -2.00447038e-01
-4.50813591e-01 8.36154103e-01 2.24350333e-01 -3.87671083e-01
-3.07719171e-01 -4.30097193e-01 -5.66340566e-01 -8.09557915e-01
1.54123962e-01 3.96899015e-01 1.75428301e-01 1.00206152e-01
4.54413414e-01 5.92499003e-02 -1.14913917e+00 -1.52927071e-01
8.81528676e-01 7.30648994e-01 8.20947737e-02 5.17262638e-01
2.86621213e-01 -5.76324351e-02 -5.34737468e-01 -4.54210937e-01
3.93329740e-01 -1.90091163e-01 -4.41974640e-01 -1.87887341e-01
2.72781223e-01 -3.17307413e-01 -2.72285163e-01 1.02383065e+00
8.90513778e-01 1.79451346e-01 8.44560742e-01 -1.38874912e+00
-2.29612872e-01 9.52293575e-01 -8.43846142e-01 4.82864380e-01
-1.70510970e-02 4.39158440e-01 1.18424881e+00 -9.20345187e-01
2.21656442e-01 1.19426084e+00 7.58119106e-01 1.93016261e-01
-1.46621609e+00 -1.09813213e+00 2.08664134e-01 1.06857441e-01
-1.72285616e+00 -5.76566935e-01 5.91464400e-01 -3.61936718e-01
6.89893246e-01 3.17290537e-02 4.84263957e-01 1.30468941e+00
2.47170597e-01 8.87272537e-01 9.73072648e-01 -2.82193959e-01
4.34613883e-01 1.45475820e-01 3.82779598e-01 5.44111013e-01
6.67039394e-01 4.47130024e-01 -6.36160731e-01 -6.50975287e-01
8.27404976e-01 -2.65748501e-01 -3.13409954e-01 -5.04477501e-01
-9.47525859e-01 1.08522797e+00 3.34787853e-02 -1.06789216e-01
-1.24894902e-01 6.37423158e-01 1.82598084e-01 7.68252835e-02
3.61807793e-01 6.69183254e-01 -7.47230589e-01 -4.56814706e-01
-7.47198641e-01 5.40541828e-01 8.91333401e-01 1.10329044e+00
4.55323398e-01 -2.31644914e-01 -5.21306574e-01 1.17783856e+00
5.00661552e-01 5.07857919e-01 1.75285906e-01 -1.19619012e+00
3.23951602e-01 1.51431084e-01 5.91477215e-01 -7.49793231e-01
-4.80028808e-01 -6.52559161e-01 -6.34617448e-01 -2.58950472e-01
7.21121728e-01 -4.55286592e-01 -6.49828911e-01 1.89205670e+00
4.21616077e-01 1.51950344e-01 -4.85403717e-01 8.73592734e-01
3.99113178e-01 5.54017663e-01 2.35734746e-01 -1.19476356e-01
1.40359259e+00 -1.08445907e+00 -6.84342861e-01 -4.74048525e-01
5.74607134e-01 -8.01980019e-01 1.48691189e+00 5.09115398e-01
-1.03533471e+00 1.23855449e-01 -1.04578412e+00 2.05138564e-01
1.18211331e-02 1.91513509e-01 8.68580937e-01 8.97318721e-01
-6.15295112e-01 4.42304224e-01 -8.60857368e-01 -3.34816307e-01
5.40511429e-01 2.51370728e-01 2.16644138e-01 1.22471228e-02
-1.21279168e+00 6.92716300e-01 2.69250005e-01 -7.50020072e-02
-8.95703673e-01 -9.77537751e-01 -4.36768949e-01 2.55085886e-01
8.70603561e-01 -9.64050651e-01 1.79830813e+00 -8.01933110e-02
-1.93093419e+00 3.70649606e-01 -1.67202204e-01 -2.97723442e-01
4.25980479e-01 -5.65825760e-01 -1.88338757e-02 -2.06933305e-01
-2.53692329e-01 4.23474401e-01 7.05449879e-01 -1.00159955e+00
-4.87937272e-01 -2.15590194e-01 -1.67731985e-01 3.07590961e-01
-1.96179777e-01 1.16847359e-01 -8.88665080e-01 -7.69370556e-01
2.72368155e-02 -1.09439254e+00 -4.57945526e-01 -2.02714443e-01
-4.30788338e-01 -2.80512571e-01 1.40051246e-01 -2.86777522e-02
1.85136294e+00 -2.24365234e+00 -2.41883755e-01 3.12457979e-01
1.81175351e-01 -2.04859883e-01 -1.26371905e-01 4.56791461e-01
1.68611512e-01 4.94363636e-01 -2.90401459e-01 -1.04358055e-01
3.82203162e-01 -2.58119971e-01 -1.93595052e-01 4.48270321e-01
-3.09427679e-01 6.68585956e-01 -7.54503131e-01 -3.98800850e-01
8.14422518e-02 1.78317741e-01 -9.22520638e-01 1.10261157e-01
-3.25637877e-01 1.46126941e-01 -7.91482925e-01 6.00196958e-01
6.20220542e-01 -6.85327411e-01 2.30393559e-01 -2.89100796e-01
7.19465464e-02 4.03397262e-01 -1.39451623e+00 1.30688691e+00
-1.38529897e-01 3.60218644e-01 9.85883027e-02 -6.94034934e-01
4.90128428e-01 7.97339007e-02 3.05106044e-01 -2.78541386e-01
3.06603134e-01 1.32464677e-01 1.50403395e-01 -2.59527206e-01
3.30189854e-01 2.11639315e-01 -1.79437250e-01 6.06539726e-01
-2.15089887e-01 -4.13436651e-01 3.92901331e-01 8.60193148e-02
1.13806081e+00 -9.57597978e-03 6.32140875e-01 -5.48008144e-01
-7.75169358e-02 -1.02087380e-02 8.59310031e-01 1.28101325e+00
-4.51837927e-01 4.10474628e-01 5.30594587e-01 -1.54510647e-01
-9.86806095e-01 -9.77165341e-01 -6.69582605e-01 1.32677150e+00
1.51112169e-01 -7.55711615e-01 -5.82034707e-01 -3.21810782e-01
3.30725253e-01 9.13899779e-01 -3.81446540e-01 -8.64670500e-02
-3.23655367e-01 -1.36287141e+00 4.45034206e-01 2.34613702e-01
4.32213396e-01 -5.86320519e-01 -5.11952579e-01 1.89569354e-01
-1.39508396e-01 -9.83439147e-01 -8.81102741e-01 3.89663093e-02
-1.03195667e+00 -9.31996524e-01 -5.11886001e-01 -1.10857531e-01
2.48693153e-01 1.49873048e-01 1.30006397e+00 -2.19023898e-01
-1.82594731e-01 4.09643948e-01 -2.22890377e-01 -6.76350832e-01
-5.04948609e-02 4.36334848e-01 1.67044267e-01 -4.79121506e-01
4.28545862e-01 -6.57586932e-01 -8.79080176e-01 7.51886487e-01
-5.81922412e-01 -6.18859045e-02 5.89374363e-01 9.24430251e-01
5.63171089e-01 -7.33891875e-02 4.43890750e-01 -1.26729083e+00
7.92557895e-01 -6.13454819e-01 -1.19549406e+00 2.62677014e-01
-9.73270774e-01 1.84988618e-01 9.72293839e-02 -5.92147052e-01
-1.02403736e+00 -2.42261454e-01 1.62063032e-01 -2.59144217e-01
1.55641481e-01 7.73970783e-01 3.62938643e-02 2.29487032e-01
8.48147035e-01 -3.13862622e-01 -2.09602267e-01 -5.60138941e-01
2.34067634e-01 6.91678226e-01 1.28454059e-01 -1.04694402e+00
6.75521374e-01 8.54658708e-02 -2.97433436e-01 -4.61028188e-01
-1.03279924e+00 -3.03845495e-01 -5.03139868e-02 1.30262926e-01
2.48220325e-01 -8.94456983e-01 -6.15485132e-01 3.31284791e-01
-8.61177623e-01 -7.87632585e-01 1.00635551e-01 7.41502643e-01
-5.97277761e-01 2.05454066e-01 -5.53383052e-01 -7.31743515e-01
-3.51332098e-01 -1.28701580e+00 7.07298577e-01 4.75226253e-01
-4.06264544e-01 -8.39807391e-01 1.23417877e-01 3.13982010e-01
5.29223323e-01 -3.32700253e-01 7.88785577e-01 -6.05267167e-01
-6.57116592e-01 4.76876423e-02 -5.27939424e-02 -1.75798804e-01
-8.68086070e-02 3.23530763e-01 -7.28930175e-01 -4.15037155e-01
-6.05367310e-02 -1.37746722e-01 7.72060752e-01 1.07482684e+00
1.59686279e+00 -1.79823682e-01 -5.18452585e-01 9.97303367e-01
1.33603597e+00 -4.22038659e-02 3.59114796e-01 4.65593994e-01
1.97499707e-01 1.07448146e-01 7.39991367e-01 8.77510428e-01
2.68499672e-01 7.68970311e-01 5.26909046e-02 3.81224722e-01
3.92542422e-01 -1.55932978e-01 4.40188587e-01 4.86523151e-01
1.89338535e-01 -5.33362448e-01 -1.08425236e+00 2.33766153e-01
-1.93272901e+00 -5.71935058e-01 2.05794185e-01 2.34378099e+00
1.27402568e+00 2.75246084e-01 2.04384238e-01 -4.22247708e-01
7.61097729e-01 9.58150178e-02 -9.13908184e-01 -9.94576588e-02
2.27941722e-02 -5.22641651e-02 8.13486934e-01 5.07730842e-01
-1.03809714e+00 1.08939278e+00 7.34858036e+00 1.21154058e+00
-5.72376251e-01 3.54973525e-02 6.81720257e-01 -4.48782742e-01
-1.50375411e-01 2.56937802e-01 -1.18177092e+00 4.70278263e-01
1.00769067e+00 -6.36918724e-01 6.77133501e-01 8.13978374e-01
5.43594480e-01 -4.92310226e-01 -1.18850815e+00 1.05396068e+00
-4.20343488e-01 -1.09056687e+00 -3.37833583e-01 5.86556494e-02
7.21156597e-01 2.57414490e-01 2.13240102e-01 4.39700186e-01
8.86644006e-01 -1.08329558e+00 3.90255123e-01 2.14809865e-01
5.09850323e-01 -4.70841348e-01 5.98267436e-01 8.26859847e-02
-7.12772727e-01 -2.86333710e-01 -4.01932150e-01 1.19609140e-01
1.49324611e-01 9.98899460e-01 -7.20432043e-01 8.61152560e-02
9.65788662e-01 5.14622629e-01 -4.88268852e-01 1.61476672e+00
-3.83165598e-01 1.29947495e+00 -8.32519650e-01 -2.36853510e-01
-1.09556340e-01 -3.68783951e-01 8.14559937e-01 9.89831626e-01
3.13001573e-01 2.43974477e-01 -2.46717744e-02 1.05912971e+00
-1.24591231e-01 2.83005498e-02 -2.43451044e-01 -2.14513898e-01
1.42329478e+00 1.06002462e+00 -5.58462799e-01 -1.51756629e-01
-2.76680261e-01 1.51398212e-01 1.59461260e-01 7.32267737e-01
-1.07304144e+00 -1.97546393e-01 9.29751396e-01 -6.39547855e-02
3.47197592e-01 -1.97210342e-01 -4.94682610e-01 -1.14644623e+00
-2.00391978e-01 -1.19691718e+00 6.78512216e-01 -6.32834673e-01
-1.61916733e+00 -7.80793279e-02 5.00092149e-01 -9.96413827e-01
-7.28702918e-02 -5.48308730e-01 -4.30473715e-01 6.11283422e-01
-1.02405214e+00 -2.75169462e-01 -2.10821822e-01 2.25508496e-01
3.90344262e-01 -5.79167437e-03 5.30853391e-01 2.41119862e-01
-1.20372331e+00 1.06551480e+00 4.10416275e-01 -1.99068457e-01
1.12460589e+00 -9.79118705e-01 2.73262203e-01 4.73212153e-01
-2.14176327e-01 9.84864593e-01 1.17138529e+00 -3.91318709e-01
-1.40613139e+00 -5.25441051e-01 1.38902694e-01 -6.05490029e-01
8.12210262e-01 -2.51818150e-01 -7.12209284e-01 6.86066270e-01
9.50370505e-02 -3.05472404e-01 7.74691641e-01 7.23223686e-01
-2.93241590e-01 -1.48600861e-01 -8.98787141e-01 1.01297474e+00
1.06647658e+00 -9.29120034e-02 -2.05292106e-01 2.77318627e-01
5.25011539e-01 -5.36547840e-01 -1.04917741e+00 5.81190407e-01
6.85391665e-01 -8.19895923e-01 9.52753723e-01 -3.70361716e-01
1.03497609e-01 -1.75247550e-01 -1.09491907e-01 -1.37024975e+00
-3.74623954e-01 -8.75546455e-01 -2.19562382e-01 1.13796115e+00
6.59500897e-01 -8.45787048e-01 7.57101774e-01 9.25190151e-01
1.11391269e-01 -9.34944153e-01 -5.87693095e-01 -7.93108821e-01
2.95610905e-01 -4.43086058e-01 6.72924101e-01 7.68526495e-01
-1.77777648e-01 5.44152558e-01 -2.73353040e-01 2.00095862e-01
7.65845716e-01 1.67120360e-02 9.99440193e-01 -9.91998672e-01
-4.76576805e-01 -8.01875293e-01 3.03206831e-01 -1.36144757e+00
-2.44891882e-01 -5.26412606e-01 1.23430658e-02 -1.09106827e+00
5.52268684e-01 -5.68148732e-01 -3.23988736e-01 3.41321588e-01
-5.47595024e-01 -1.67392418e-01 -7.71652013e-02 2.70156443e-01
-5.67187965e-01 8.84622276e-01 1.01191068e+00 1.26610041e-01
-3.16815048e-01 2.06352621e-01 -9.14977551e-01 7.49702811e-01
1.16141021e+00 -8.43544662e-01 -4.31034476e-01 -3.91596615e-01
4.27617610e-01 -9.51301530e-02 6.15468211e-02 -6.17653668e-01
2.96828270e-01 -4.96505678e-01 1.74506247e-01 -4.51214403e-01
1.72679752e-01 -4.69905227e-01 1.46694690e-01 1.89486980e-01
-3.74295890e-01 1.80714250e-01 3.60667080e-01 6.75474763e-01
1.29607394e-01 -3.86308968e-01 8.50100636e-01 1.99238673e-01
-4.49070156e-01 3.86720151e-01 -3.65809351e-01 4.47202116e-01
7.08814025e-01 5.57891419e-03 -4.36585009e-01 -4.36937422e-01
-4.15235341e-01 7.04027653e-01 3.47398996e-01 9.92836505e-02
1.32993743e-01 -1.04811192e+00 -6.95700288e-01 -2.46318415e-01
8.95501971e-02 3.07960920e-02 2.14739755e-01 9.83922362e-01
-2.86477029e-01 2.89643675e-01 2.28638738e-01 -5.92098117e-01
-9.65772629e-01 1.50333732e-01 4.85409260e-01 -4.79623705e-01
-3.55777115e-01 9.00215745e-01 2.78227925e-01 -4.38889354e-01
4.29774523e-01 -2.29277566e-01 1.15906157e-01 -2.05955461e-01
2.56014258e-01 4.49022770e-01 -3.21322888e-01 3.76662791e-01
-2.91333735e-01 4.10737813e-01 -1.58235401e-01 -3.49178642e-01
1.34229589e+00 -2.53384650e-01 1.52950976e-02 4.69189256e-01
7.56167114e-01 9.58894193e-02 -1.38825846e+00 -3.41216981e-01
1.36478975e-01 -5.89423597e-01 3.79853100e-01 -8.37444127e-01
-8.60540748e-01 4.28133160e-01 5.81117988e-01 -1.79451615e-01
9.68269348e-01 -1.99101537e-01 4.31915939e-01 6.37705922e-01
4.26994413e-01 -1.19856966e+00 1.14195682e-02 5.99175453e-01
7.18111575e-01 -1.31008029e+00 4.93402660e-01 -5.25423884e-01
-5.74316919e-01 6.86298013e-01 7.17570126e-01 2.48790588e-02
1.03897011e+00 4.81635749e-01 -1.55536979e-01 -2.34844491e-01
-1.05200171e+00 1.08021781e-01 1.92403644e-01 1.14732623e-01
4.49749887e-01 -7.75806755e-02 -6.56654537e-01 9.14221466e-01
-4.44834709e-01 8.55639428e-02 3.39171857e-01 7.50714540e-01
-3.45055044e-01 -1.18685687e+00 -3.89820367e-01 7.31330395e-01
-5.38139045e-01 -2.88859069e-01 -5.65995760e-02 8.99044693e-01
-6.10570312e-01 9.47551370e-01 -9.41944495e-02 -1.76702932e-01
1.57307804e-01 5.12733683e-02 4.83122438e-01 -5.71100533e-01
-4.16994959e-01 3.73624980e-01 1.90686896e-01 -7.18501985e-01
-1.36418000e-01 -7.92579174e-01 -8.22544277e-01 -6.41837597e-01
-5.73572099e-01 3.49238545e-01 5.38913131e-01 6.28561020e-01
6.08550131e-01 2.09919885e-01 4.67180729e-01 -5.46582699e-01
-1.13436365e+00 -1.13066268e+00 -6.12934709e-01 -2.99949106e-02
-8.54683667e-02 -1.16702855e+00 -7.07337618e-01 -3.86127800e-01] | [8.102594375610352, 3.9026646614074707] |
16c0f66a-25db-4ebd-ae82-08c6782ddb75 | moccasin-efficient-tensor-rematerialization | 2304.14463 | null | https://arxiv.org/abs/2304.14463v2 | https://arxiv.org/pdf/2304.14463v2.pdf | Moccasin: Efficient Tensor Rematerialization for Neural Networks | The deployment and training of neural networks on edge computing devices pose many challenges. The low memory nature of edge devices is often one of the biggest limiting factors encountered in the deployment of large neural network models. Tensor rematerialization or recompute is a way to address high memory requirements for neural network training and inference. In this paper we consider the problem of execution time minimization of compute graphs subject to a memory budget. In particular, we develop a new constraint programming formulation called \textsc{Moccasin} with only $O(n)$ integer variables, where $n$ is the number of nodes in the compute graph. This is a significant improvement over the works in the recent literature that propose formulations with $O(n^2)$ Boolean variables. We present numerical studies that show that our approach is up to an order of magnitude faster than recent work especially for large-scale graphs. | ['Bistra Dilkina', 'Christopher Lott', 'Harris Teague', 'Haoming Li', 'Burak Bartan'] | 2023-04-27 | null | null | null | null | ['edge-computing'] | ['time-series'] | [-2.68578846e-02 1.13554932e-01 -5.05593009e-02 -2.33132094e-01
1.18717819e-01 -4.46735829e-01 -3.77865578e-03 9.50567424e-02
-6.60993636e-01 7.74078012e-01 -5.54796219e-01 -5.93171895e-01
-4.17339057e-01 -9.40463901e-01 -9.44659531e-01 -5.59047997e-01
-3.95736188e-01 5.43158948e-01 4.28570099e-02 -3.01908404e-02
1.42958239e-01 6.61910951e-01 -1.51622403e+00 -2.13326961e-01
4.18392211e-01 1.62167394e+00 -1.25995681e-01 7.44186640e-01
-2.25295424e-02 8.37194562e-01 -3.65235537e-01 -6.67705178e-01
4.79743659e-01 1.18477412e-01 -6.32374883e-01 -1.15417771e-01
5.48439145e-01 -4.19770144e-02 -3.30113471e-01 1.36643076e+00
4.29311603e-01 2.34727085e-01 1.15508363e-01 -1.46328449e+00
-3.07141244e-01 8.51597548e-01 -2.88040817e-01 4.17019576e-01
-4.46238756e-01 -3.80677283e-01 7.96540082e-01 -7.00926006e-01
6.29147947e-01 5.17896295e-01 6.72771394e-01 5.19623816e-01
-8.54107320e-01 -7.51336932e-01 2.66020715e-01 4.30801630e-01
-1.70622158e+00 -4.87459481e-01 7.03152418e-01 -1.12083301e-01
1.44602406e+00 5.46648264e-01 9.29135263e-01 3.91717374e-01
-4.17578146e-02 4.01281565e-01 6.85115457e-01 -4.50141460e-01
6.27565205e-01 9.57698822e-02 5.63968539e-01 1.02108514e+00
9.09373224e-01 -1.09356299e-01 -6.22090101e-01 3.51870172e-02
7.06619501e-01 -2.34582335e-01 -2.58249864e-02 8.62581953e-02
-5.13285398e-01 6.00930035e-01 4.00287598e-01 2.54019529e-01
-3.70825917e-01 8.13446283e-01 5.58610141e-01 3.50176960e-01
5.69658637e-01 2.08151847e-01 -5.33621967e-01 -3.31183851e-01
-1.04023182e+00 5.81008531e-02 1.04687512e+00 1.32732320e+00
5.22949398e-01 5.41887939e-01 5.20084143e-01 3.15331876e-01
1.32488266e-01 1.30713657e-01 -9.91120711e-02 -8.65487874e-01
5.85633218e-01 5.91654599e-01 -2.87935674e-01 -1.08730233e+00
-4.08665925e-01 -7.59566247e-01 -1.12919331e+00 1.29465491e-01
2.86662579e-01 -5.73500514e-01 -8.15271616e-01 1.48506808e+00
1.97595477e-01 3.14291835e-01 -2.17645362e-01 7.54479825e-01
6.62475765e-01 3.70408595e-01 -1.21597648e-01 -1.69958830e-01
1.12015510e+00 -1.00420439e+00 -6.47264481e-01 -3.84805650e-01
7.54434884e-01 -3.32942039e-01 3.84670824e-01 4.67388093e-01
-1.49753606e+00 6.16197623e-02 -1.27713442e+00 -9.50355735e-03
-5.76540649e-01 2.32391581e-01 1.46286345e+00 1.13539362e+00
-1.51732409e+00 6.08955145e-01 -8.88947189e-01 1.36221737e-01
4.16453809e-01 9.41023290e-01 -2.84619272e-01 -6.58922270e-02
-7.61796951e-01 7.24695563e-01 4.50943083e-01 5.96350431e-01
-4.97599155e-01 -6.64861262e-01 -6.36494577e-01 3.09437186e-01
5.23025930e-01 -5.99589109e-01 7.19508111e-01 -6.66404665e-01
-1.27357447e+00 5.97841203e-01 5.31184599e-02 -5.01047075e-01
2.37049624e-01 1.44837901e-01 -2.34168634e-01 -4.60910797e-03
-5.60124874e-01 3.37488025e-01 5.32830477e-01 -6.51675463e-01
-4.04900223e-01 -5.48208714e-01 2.93900669e-01 -6.44542575e-02
-7.09729612e-01 5.73207997e-02 -7.57525682e-01 -3.46250951e-01
1.83012038e-01 -1.18206191e+00 -4.07626122e-01 2.32301988e-02
-3.63341063e-01 -1.82616934e-01 5.51878452e-01 -3.86953801e-01
1.35391009e+00 -1.89076006e+00 3.86403471e-01 6.02912724e-01
6.55609310e-01 1.80908918e-01 3.43861312e-01 -1.30518749e-02
9.19228569e-02 3.07797909e-01 1.47962749e-01 -4.04942006e-01
-5.86546920e-02 2.69038171e-01 2.56269187e-01 3.30675989e-01
-1.77991942e-01 6.76901042e-01 -3.28031391e-01 -4.72594827e-01
-1.84870332e-01 5.70944309e-01 -6.55647814e-01 -3.33568186e-01
-1.59012586e-01 -2.89478391e-01 -4.33499366e-01 7.36549556e-01
7.84537375e-01 -5.49109459e-01 4.40505564e-01 -2.88386911e-01
-6.86834976e-02 5.33643737e-02 -1.48065948e+00 1.53328669e+00
-3.12690347e-01 8.38691115e-01 4.75328952e-01 -1.18708861e+00
5.28472722e-01 3.18447292e-01 4.58635807e-01 -3.10965031e-01
7.17693865e-01 2.16929778e-01 -5.70475459e-02 -2.46554062e-01
6.05657041e-01 -1.26685217e-01 1.78713188e-01 3.33647013e-01
8.80855322e-02 2.87354231e-01 4.23932105e-01 2.08527029e-01
1.29163158e+00 -5.41975439e-01 -3.83244157e-01 -4.65351164e-01
1.43859424e-02 -7.70259500e-02 6.43393457e-01 5.68105876e-01
-1.68869913e-01 3.80257256e-02 9.07543063e-01 -4.97380823e-01
-1.15066290e+00 -4.38038558e-01 5.99602284e-03 7.85031319e-01
-1.16484284e-01 -5.33051431e-01 -1.03553176e+00 -2.05845192e-01
-2.58712590e-01 3.33454102e-01 -6.37288928e-01 9.53652188e-02
-6.33325636e-01 -9.82393682e-01 5.70551217e-01 8.91555548e-01
5.74521184e-01 -5.18740296e-01 -8.47817361e-01 -2.05580965e-02
4.38224763e-01 -1.39769435e+00 -1.94326550e-01 5.57703674e-01
-1.28370655e+00 -8.22277129e-01 -4.13463116e-01 -8.68553281e-01
1.05493951e+00 -1.38040036e-01 9.98337865e-01 4.77773786e-01
-3.47810805e-01 -6.31208345e-02 -1.15964755e-01 -4.17777300e-01
3.98227841e-01 2.86433131e-01 1.27452180e-01 -3.09801310e-01
4.18744564e-01 -6.72924042e-01 -3.64970028e-01 -1.08122416e-01
-9.13220346e-01 2.23447308e-01 4.23124313e-01 6.66871727e-01
5.79226077e-01 7.32664347e-01 5.46179973e-02 -1.16347897e+00
5.98809302e-01 -3.46812397e-01 -1.12124932e+00 2.88679361e-01
-1.13209271e+00 3.34577918e-01 6.03714645e-01 -5.12387276e-01
-4.71072078e-01 2.03071505e-01 1.78078532e-01 -5.66302299e-01
5.28165638e-01 8.94406319e-01 -9.42891277e-03 -7.74591804e-01
3.28702837e-01 -2.46869430e-01 -3.92359495e-01 -2.66751915e-01
-6.49776757e-02 1.63229138e-01 3.34648266e-02 -7.72910297e-01
4.12417799e-01 2.70165175e-01 8.21456313e-01 -6.85502410e-01
-2.63335824e-01 2.64380518e-02 -3.03382069e-01 -4.42450523e-01
5.47330201e-01 -5.11116147e-01 -1.26760066e+00 2.38966554e-01
-9.31069434e-01 -2.54027873e-01 4.69262637e-02 4.68314916e-01
-2.92289145e-02 -6.90302923e-02 -7.88692594e-01 -1.01493180e+00
-5.48475921e-01 -1.15660632e+00 4.67993855e-01 4.19524640e-01
4.04268682e-01 -1.00746191e+00 -5.74737489e-01 2.22203568e-01
7.50654697e-01 3.03044587e-01 9.02807415e-01 -1.73774794e-01
-9.51898932e-01 -5.19861341e-01 -4.70703632e-01 1.82113588e-01
-6.55991435e-01 -7.76810795e-02 -6.16167903e-01 -3.08663756e-01
8.22970048e-02 1.85400154e-02 5.60475349e-01 3.46002907e-01
1.24939537e+00 -4.05141979e-01 -3.35397899e-01 8.95328522e-01
1.81681454e+00 2.24496335e-01 5.10341585e-01 3.26482654e-02
8.12941909e-01 2.60104209e-01 1.02196805e-01 4.11708802e-01
1.47594526e-01 3.49280685e-01 6.68613851e-01 2.01753467e-01
1.89302295e-01 1.88143060e-01 -1.28962770e-01 1.12281203e+00
-6.37832403e-01 -2.82350570e-01 -1.00343561e+00 5.38505375e-01
-1.73767245e+00 -5.56153655e-01 -1.82317436e-01 2.04634047e+00
4.45820957e-01 4.41208005e-01 -1.35921478e-01 3.84491891e-01
5.83518863e-01 -1.01774246e-01 -4.96763289e-01 -5.24298489e-01
1.31115735e-01 7.22233176e-01 1.22419119e+00 4.44334477e-01
-6.44130647e-01 8.51554334e-01 6.57745123e+00 6.02345109e-01
-1.20549631e+00 3.27477962e-01 5.73964715e-01 -6.04152501e-01
-3.97541784e-02 6.37285933e-02 -8.96490693e-01 3.26043665e-01
1.34782147e+00 -1.20788760e-01 8.03122580e-01 1.03978360e+00
-3.53525788e-01 -3.79261039e-02 -1.07577968e+00 1.19380486e+00
7.41426051e-02 -1.57214975e+00 -5.18751621e-01 1.79147258e-01
6.94762707e-01 2.53449857e-01 9.01725367e-02 -3.03916465e-02
1.91741690e-01 -1.08273828e+00 6.66825950e-01 2.87738979e-01
8.73080730e-01 -1.03145695e+00 7.12264776e-01 1.91818431e-01
-1.34964073e+00 7.33368248e-02 -5.84373593e-01 -3.09173703e-01
1.69402450e-01 8.58588219e-01 -4.88788962e-01 2.95388490e-01
9.59769845e-01 1.68994278e-01 -3.37175161e-01 8.90204430e-01
2.10631534e-01 4.56982166e-01 -8.17108512e-01 -6.21654093e-01
3.05191670e-02 -2.20002770e-01 2.87760884e-01 8.95420074e-01
3.59982193e-01 3.34199429e-01 -3.57661188e-01 7.52717376e-01
-4.84546363e-01 -8.80028009e-02 -4.22692567e-01 -4.46562052e-01
4.88686651e-01 1.22105622e+00 -1.02991295e+00 -3.15174311e-01
-2.57880628e-01 7.06196129e-01 6.39452159e-01 3.03140730e-01
-8.30036283e-01 -4.33507353e-01 4.50568438e-01 6.17977045e-03
3.41248214e-01 -7.18809962e-01 -6.06580615e-01 -1.12456751e+00
4.46703225e-01 -3.61597866e-01 1.63426891e-01 -4.87408161e-01
-8.09071362e-01 8.31700563e-01 5.68762235e-03 -6.30032837e-01
1.63673777e-02 -1.12108970e+00 -2.32505605e-01 7.15664208e-01
-1.25771868e+00 -8.31177115e-01 -2.72024781e-01 3.57742965e-01
1.43559778e-03 -2.48118252e-01 8.68397176e-01 8.64862144e-01
-9.49658394e-01 8.81881475e-01 -8.94626752e-02 1.79580320e-02
-3.53920430e-01 -9.03743565e-01 2.75084913e-01 1.02945542e+00
-4.12437394e-02 8.42442572e-01 7.81524181e-01 -7.19843626e-01
-2.04072309e+00 -7.18127966e-01 9.58946705e-01 1.21745594e-01
5.87549686e-01 -5.09970427e-01 -4.10649002e-01 9.59081709e-01
-8.50184113e-02 3.96625966e-01 6.62386537e-01 3.44374627e-01
-1.68944389e-01 -2.04281673e-01 -1.11012638e+00 5.13398945e-01
1.25626314e+00 -4.81320977e-01 4.54752862e-01 4.19317067e-01
5.30776322e-01 -7.70231843e-01 -1.05631769e+00 2.30769083e-01
4.65678245e-01 -5.87035775e-01 6.70141399e-01 -7.12886691e-01
9.25591290e-02 -7.17511401e-02 -2.90083319e-01 -7.79118419e-01
1.05357789e-01 -6.70551896e-01 -4.21650171e-01 7.02942908e-01
7.42086530e-01 -5.38621724e-01 1.52906108e+00 1.30445492e+00
-5.00858873e-02 -8.88525486e-01 -1.18769276e+00 -7.37623274e-01
-1.42559424e-01 -7.99534619e-01 5.17323971e-01 9.32687044e-01
-4.85002674e-04 1.35621116e-01 -2.18299881e-01 8.98733884e-02
7.24803329e-01 -3.58648092e-01 2.07009256e-01 -1.18122363e+00
-5.69978952e-01 -4.04396653e-01 -7.41729915e-01 -7.30452061e-01
-1.43048465e-02 -9.22437668e-01 -4.27642107e-01 -1.27040756e+00
2.62473430e-02 -8.47629428e-01 -2.28259012e-01 4.46600497e-01
4.60172027e-01 1.94075823e-01 1.55222267e-01 -2.21148968e-01
-5.15540898e-01 3.38201374e-02 8.63142312e-01 1.78446099e-02
1.63355708e-01 -8.30635726e-02 -6.00713253e-01 5.39975941e-01
8.85323167e-01 -6.35157526e-01 -4.38337237e-01 -1.04726458e+00
1.02123785e+00 9.99560654e-02 1.56777009e-01 -1.16866183e+00
8.51996899e-01 2.43648831e-02 2.27158964e-02 -2.88185298e-01
6.30394995e-01 -1.17389071e+00 4.98606771e-01 3.45941395e-01
-2.77133491e-02 7.39537954e-01 4.00777191e-01 3.50083262e-01
-3.42077529e-03 -6.65753663e-01 3.31926078e-01 -1.82104021e-01
-4.74959373e-01 4.44880128e-01 -1.07007078e-03 -2.68161833e-01
9.43681419e-01 -6.65322989e-02 -5.90791225e-01 -1.98586971e-01
-7.53191352e-01 6.69403300e-02 8.91690478e-02 -2.44231615e-02
5.38954914e-01 -1.16630471e+00 -1.45490289e-01 1.18787169e-01
-4.52809304e-01 2.45571151e-01 3.37951332e-01 8.32659960e-01
-1.09034276e+00 6.80553198e-01 -2.02613458e-01 -1.96781591e-01
-1.29237723e+00 5.25817096e-01 4.35811102e-01 -2.85356492e-01
-4.01409626e-01 1.36246634e+00 -7.19278991e-01 1.37816787e-01
4.92063761e-01 -2.84992665e-01 2.51634032e-01 -3.61254007e-01
3.26464027e-01 6.93885386e-01 5.25240123e-01 -2.59294868e-01
-3.63614768e-01 3.09630930e-01 -1.87012091e-01 -2.38015592e-01
1.55575383e+00 3.06438357e-01 -5.91801107e-01 9.60870758e-02
1.13892901e+00 -2.90072739e-01 -8.44688773e-01 -9.22780931e-02
-3.09272800e-02 -1.06784344e-01 7.27704227e-01 -4.88020509e-01
-1.77967608e+00 7.56957591e-01 6.45723283e-01 2.52396554e-01
1.36806166e+00 -1.98277652e-01 6.79493487e-01 6.45615399e-01
6.24767363e-01 -1.57367957e+00 -3.94855320e-01 5.13624191e-01
4.07540947e-01 -7.84729004e-01 1.76761568e-01 -7.13792562e-01
-2.90375687e-02 1.01744890e+00 7.28315234e-01 -1.84098914e-01
1.11187589e+00 8.76861632e-01 -4.27112937e-01 -5.87168992e-01
-7.62586474e-01 8.44288170e-02 5.97115196e-02 7.71207213e-02
2.96306044e-01 2.38777429e-01 -5.47635555e-01 5.70129395e-01
-4.52129334e-01 2.88118005e-01 3.99027586e-01 1.12904644e+00
-1.07702643e-01 -1.19022810e+00 2.29106285e-02 6.99680626e-01
-5.92507005e-01 -3.01640362e-01 -1.78628743e-01 5.56336343e-01
1.85614377e-01 6.89937294e-01 1.88474879e-01 -7.33649075e-01
1.73602886e-02 4.22097929e-02 6.77205384e-01 -2.30007276e-01
-6.81983232e-01 -3.34222198e-01 3.75035405e-01 -3.60302687e-01
-1.06951728e-01 -4.13645923e-01 -1.37470579e+00 -8.10552061e-01
-5.98426104e-01 -4.12165225e-02 1.43036568e+00 7.87287891e-01
5.00226736e-01 7.32173562e-01 2.40528770e-02 -6.08569205e-01
-1.42001852e-01 -7.23762870e-01 -6.92856133e-01 -7.60704651e-02
-1.64905190e-01 -5.99122405e-01 -3.41904074e-01 -2.44332612e-01] | [8.41599178314209, 3.2758729457855225] |
ad0be8c6-7a64-45ae-8f0e-2f3061dd82d8 | predicting-rare-events-by-shrinking-towards | 2305.187 | null | https://arxiv.org/abs/2305.18700v1 | https://arxiv.org/pdf/2305.18700v1.pdf | Predicting Rare Events by Shrinking Towards Proportional Odds | Training classifiers is difficult with severe class imbalance, but many rare events are the culmination of a sequence with much more common intermediate outcomes. For example, in online marketing a user first sees an ad, then may click on it, and finally may make a purchase; estimating the probability of purchases is difficult because of their rarity. We show both theoretically and through data experiments that the more abundant data in earlier steps may be leveraged to improve estimation of probabilities of rare events. We present PRESTO, a relaxation of the proportional odds model for ordinal regression. Instead of estimating weights for one separating hyperplane that is shifted by separate intercepts for each of the estimated Bayes decision boundaries between adjacent pairs of categorical responses, we estimate separate weights for each of these transitions. We impose an L1 penalty on the differences between weights for the same feature in adjacent weight vectors in order to shrink towards the proportional odds model. We prove that PRESTO consistently estimates the decision boundary weights under a sparsity assumption. Synthetic and real data experiments show that our method can estimate rare probabilities in this setting better than both logistic regression on the rare category, which fails to borrow strength from more abundant categories, and the proportional odds model, which is too inflexible. | ['Jacob Bien', 'Gregory Faletto'] | 2023-05-30 | null | null | null | null | ['marketing'] | ['miscellaneous'] | [ 4.30592775e-01 2.59448826e-01 -7.59820700e-01 -6.95637822e-01
-6.16074264e-01 -4.42306101e-01 3.33049238e-01 6.22891128e-01
-4.68945444e-01 8.16442788e-01 1.98390633e-01 -4.84028876e-01
-3.42034817e-01 -6.80474699e-01 -7.99316585e-01 -5.41351676e-01
-3.33661020e-01 5.77784598e-01 -1.05358675e-01 1.79831013e-01
1.73916817e-01 1.30832806e-01 -1.34002852e+00 3.60624164e-01
7.83623397e-01 9.69298899e-01 -3.11347932e-01 4.12461549e-01
8.96551162e-02 4.51156020e-01 -3.50558996e-01 -5.90906620e-01
4.27396417e-01 -2.01116890e-01 -4.36265975e-01 7.33411238e-02
7.90387809e-01 -4.29816991e-01 9.26084593e-02 1.07037938e+00
-5.46296500e-02 -1.02454266e-02 1.21129632e+00 -1.61731124e+00
-4.99593765e-01 9.65757906e-01 -1.16948652e+00 1.49089217e-01
1.84918568e-01 -6.57806471e-02 1.58966112e+00 -7.12017119e-01
3.42294484e-01 1.10393488e+00 8.30519199e-01 -4.13815640e-02
-1.55703783e+00 -8.95687222e-01 4.17407423e-01 -4.28355113e-02
-1.13608384e+00 -2.61569470e-01 3.52058738e-01 -6.07469440e-01
5.94573379e-01 2.94490576e-01 4.95387733e-01 1.07887328e+00
1.33606821e-01 6.51259899e-01 1.06686747e+00 -3.51283163e-01
2.43678719e-01 3.41901839e-01 6.48867011e-01 4.63288754e-01
7.64063239e-01 8.88741761e-02 -5.17146230e-01 -6.31150067e-01
3.70811790e-01 4.54498500e-01 -5.63297272e-02 -3.37706357e-01
-9.95467544e-01 1.14980900e+00 2.93012500e-01 -1.25979081e-01
-3.09270740e-01 -9.63181537e-03 3.12869325e-02 3.54964644e-01
2.91013390e-01 3.09514046e-01 -5.62676251e-01 -8.60006432e-04
-1.02414310e+00 3.67985189e-01 1.00288904e+00 7.53241897e-01
5.32076418e-01 -4.85514700e-01 -3.35210487e-02 8.54553163e-01
-1.04471073e-02 2.02304542e-01 2.79480487e-01 -7.44842768e-01
5.64533412e-01 4.62114871e-01 3.84415001e-01 -8.17980766e-01
-5.10754406e-01 -4.37360257e-01 -8.06827307e-01 9.64480489e-02
1.15287900e+00 -1.70549482e-01 -8.30751717e-01 1.79169881e+00
1.76003426e-01 -7.58939981e-02 -3.83273631e-01 7.18778312e-01
-2.01988384e-01 5.58445632e-01 4.42354798e-01 -3.85850281e-01
1.46124780e+00 -4.68620569e-01 -4.01091933e-01 -3.91273648e-01
6.43295288e-01 -5.91556966e-01 1.15783405e+00 7.03147173e-01
-7.93089449e-01 -7.53687173e-02 -1.06048012e+00 -6.05689874e-03
-9.59094539e-02 -7.91624784e-02 8.76685560e-01 5.66134393e-01
-2.32380822e-01 6.85633659e-01 -7.88464665e-01 -8.77589360e-02
5.48161149e-01 3.38899463e-01 -3.34717005e-01 -1.64885238e-01
-1.09173608e+00 7.51500905e-01 3.76020037e-02 -1.54480964e-01
-2.20354684e-02 -8.54471028e-01 -8.41114700e-01 3.58513534e-01
5.07748783e-01 -2.69550979e-01 1.02962244e+00 -1.02092040e+00
-7.17019200e-01 6.36060357e-01 -1.31474972e-01 -5.55259943e-01
4.96761084e-01 -2.70124316e-01 -2.11868286e-01 -3.98607105e-01
1.05136126e-01 6.03106201e-01 7.52436936e-01 -8.76757562e-01
-9.34755504e-01 -5.33961654e-01 -3.84643562e-02 5.92695735e-02
-4.37781304e-01 -3.86231057e-02 1.80628881e-01 -5.67825615e-01
1.63541347e-01 -9.60098624e-01 -2.96086073e-01 1.39401808e-01
-4.78744775e-01 -2.90236086e-01 1.66055277e-01 -4.86408949e-01
1.34547198e+00 -2.17188191e+00 -6.10217340e-02 6.03837550e-01
2.65222073e-01 -3.51381987e-01 2.26011068e-01 6.67180493e-02
-2.05066353e-01 2.06906110e-01 -4.19283569e-01 -9.86647326e-03
1.36190206e-01 -3.90339196e-02 -4.42706913e-01 5.91353953e-01
2.99910635e-01 2.98541725e-01 -6.61356390e-01 -1.83493614e-01
-3.25944930e-01 -4.06017415e-02 -8.99170578e-01 -2.48375937e-01
2.66085733e-02 -4.50462222e-01 -1.26092866e-01 4.13421333e-01
6.64211392e-01 -3.89049232e-01 5.09513497e-01 -1.83714628e-01
-8.25737789e-03 4.23663229e-01 -1.51139462e+00 7.65631914e-01
-4.07533824e-01 4.08928692e-01 -1.16726168e-01 -1.10442007e+00
5.38867772e-01 -6.43915161e-02 3.73119742e-01 -1.92196190e-01
-3.20981070e-02 1.98563859e-01 2.43540332e-01 -3.10424000e-01
3.99447411e-01 -5.25835574e-01 -4.86738205e-01 4.09544796e-01
-1.03351638e-01 1.58085898e-01 1.25107720e-01 4.61856052e-02
9.57280338e-01 -2.63990432e-01 5.88196635e-01 -2.85485089e-01
-1.75362393e-01 -3.86467273e-03 8.54461133e-01 9.87380683e-01
-1.53476000e-01 5.18849492e-01 1.13852823e+00 -3.87493789e-01
-1.15063250e+00 -1.32013750e+00 -6.80789948e-01 1.17604804e+00
-1.05490133e-01 -2.61553288e-01 -2.16998503e-01 -8.61595035e-01
7.37602293e-01 1.01569235e+00 -6.89639032e-01 -2.09947541e-01
-3.48741055e-01 -1.11178279e+00 1.49009421e-01 6.53172970e-01
-1.07911587e-01 -4.04852480e-01 -4.00466174e-01 1.19441688e-01
4.11451943e-02 -7.82316208e-01 -4.91436988e-01 6.69303179e-01
-7.44366407e-01 -1.11070621e+00 -4.76516813e-01 -5.54211080e-01
7.20734298e-01 -2.54636128e-02 9.85349298e-01 -1.15856864e-01
-1.78118333e-01 -1.89806506e-01 -7.48083517e-02 -4.18979079e-01
-6.63636029e-02 1.83983609e-01 1.44567683e-01 -7.71042630e-02
8.82056415e-01 -5.60234904e-01 -4.60941911e-01 3.46016556e-01
-7.29408979e-01 -1.95179179e-01 5.21196842e-01 1.09282863e+00
2.96478957e-01 1.02929726e-01 7.27690160e-01 -1.18228126e+00
3.84614259e-01 -1.15275276e+00 -6.62279785e-01 2.14447007e-01
-5.32230735e-01 1.63182318e-01 5.58078825e-01 -9.26179707e-01
-6.74755692e-01 -8.31222162e-02 1.44253135e-01 4.68379706e-02
-6.02759868e-02 5.22049606e-01 1.04208313e-01 5.29644608e-01
7.73940504e-01 -4.87267613e-01 -5.31031117e-02 -3.55525911e-01
1.09163560e-02 9.04643834e-01 2.19478682e-01 -6.28982604e-01
6.19152844e-01 2.18487293e-01 -3.96848917e-02 -5.68033993e-01
-1.09431732e+00 -3.35805386e-01 -3.54381055e-01 1.86546475e-01
5.13043225e-01 -7.99542189e-01 -8.19163740e-01 3.65405753e-02
-6.10877097e-01 -4.59182858e-01 -3.84016216e-01 8.11597824e-01
-1.85130849e-01 -5.90729620e-03 -5.45454025e-01 -8.33472908e-01
3.05718839e-01 -8.91357899e-01 8.76542926e-01 8.90767053e-02
-6.35725439e-01 -7.27059484e-01 -2.68708766e-01 1.32501855e-01
7.99634159e-02 5.94054386e-02 1.30061364e+00 -9.64495778e-01
-1.35724574e-01 -5.23525000e-01 -1.56219080e-01 2.60750055e-01
1.33749202e-01 1.24457732e-01 -5.34840882e-01 -1.14439756e-01
-1.97634008e-02 -4.22633797e-01 1.04174626e+00 4.78567719e-01
1.29432487e+00 -5.68398476e-01 -3.61582249e-01 2.65791416e-01
1.22622180e+00 6.64346367e-02 3.26942772e-01 1.23075500e-01
5.07607579e-01 8.44763219e-01 5.78100264e-01 5.08710861e-01
4.66695994e-01 7.13524044e-01 -2.49362197e-02 4.40372117e-02
4.54828292e-01 -2.24104062e-01 3.84303957e-01 3.10860783e-01
2.95143247e-01 -1.47219956e-01 -8.21286976e-01 5.28867781e-01
-1.62392712e+00 -8.31435978e-01 -1.74280256e-01 2.73660994e+00
1.17126775e+00 6.97636247e-01 3.82741302e-01 2.64182717e-01
6.73674405e-01 -1.29828572e-01 -5.56178689e-01 -5.67222655e-01
1.81324974e-01 1.23436302e-01 8.55434716e-01 6.67286634e-01
-9.51060712e-01 4.08351779e-01 6.71857595e+00 6.83224559e-01
-8.07047307e-01 -1.42683819e-01 1.01515090e+00 -5.44683635e-01
-4.00366575e-01 1.61428705e-01 -1.12904692e+00 6.66498840e-01
8.67961824e-01 -6.74085319e-02 2.80924350e-01 8.68621767e-01
-1.42805904e-01 -4.64465886e-01 -1.43162608e+00 4.66640741e-01
-1.97894141e-01 -8.03108156e-01 -6.66356683e-02 2.41952181e-01
7.50894189e-01 -4.16277260e-01 2.08181903e-01 3.59012276e-01
8.18804264e-01 -1.03987551e+00 7.26896584e-01 2.41248012e-01
6.17093027e-01 -6.92289054e-01 6.00302815e-01 3.82374734e-01
-5.43792844e-01 -4.78667110e-01 -4.03938174e-01 -4.14656192e-01
-3.35885026e-02 1.16606855e+00 -8.38549733e-01 -1.07761927e-01
5.06550133e-01 3.19092155e-01 -1.91459194e-01 8.60944271e-01
-4.38793823e-02 8.93777370e-01 -7.50903606e-01 2.07414702e-02
9.38414261e-02 -1.62641704e-01 2.74654657e-01 9.73519087e-01
2.52254516e-01 2.33781356e-02 1.75428167e-01 6.44577205e-01
-2.32329726e-01 1.14098988e-01 -4.97673392e-01 2.09393546e-01
6.10973239e-01 1.03506649e+00 -6.14413738e-01 -3.63309264e-01
-7.20790029e-01 3.76129717e-01 4.35511857e-01 3.20038527e-01
-8.06767106e-01 -2.49224603e-01 6.55935645e-01 4.92081732e-01
3.33872139e-01 -5.79907820e-02 -5.94085574e-01 -1.14339781e+00
8.94698650e-02 -9.09674823e-01 7.64321983e-01 -4.19747204e-01
-1.54180431e+00 -2.58418247e-02 2.60011703e-01 -1.02891517e+00
-3.49966705e-01 -5.52080810e-01 -4.78517532e-01 7.90422797e-01
-1.26160550e+00 -4.58228856e-01 7.39661679e-02 1.18043542e-01
3.46696109e-01 3.29024971e-01 5.09144127e-01 2.00244427e-01
-6.66093230e-01 8.25395525e-01 1.18935257e-01 -1.20584592e-01
9.97921407e-01 -1.32281661e+00 1.37628034e-01 4.65831101e-01
4.80536371e-03 7.14586139e-01 8.88598084e-01 -7.62435973e-01
-7.60999799e-01 -7.45119631e-01 1.02137470e+00 -2.05079138e-01
7.64912903e-01 -4.76264685e-01 -9.61422622e-01 1.01181364e+00
-3.39868575e-01 -7.28524029e-02 9.58776593e-01 8.38701367e-01
-6.20889425e-01 -2.50384420e-01 -1.28074384e+00 7.15982616e-01
7.69851387e-01 -1.83821738e-01 -6.21423006e-01 4.09596086e-01
3.14115345e-01 -7.93918560e-04 -9.33478653e-01 4.11586463e-01
1.04415274e+00 -7.46067703e-01 7.51679957e-01 -8.52508247e-01
7.46848583e-01 -5.16022369e-02 -3.10094744e-01 -1.21694183e+00
-2.19151080e-01 -2.60338247e-01 4.06462736e-02 1.17964816e+00
7.38477707e-01 -5.99701703e-01 7.70296156e-01 1.01290250e+00
3.50026250e-01 -8.41585517e-01 -8.12765777e-01 -6.72752559e-01
2.39249930e-01 -4.37322438e-01 5.45022368e-01 1.01500750e+00
3.17178875e-01 2.82240778e-01 -4.12972510e-01 1.60231758e-02
7.99651265e-01 1.08600691e-01 5.74392438e-01 -1.38303161e+00
-5.65266669e-01 -5.08224428e-01 -2.82492459e-01 -8.77294838e-01
-1.34075671e-01 -7.92751789e-01 5.06380945e-02 -8.43496442e-01
3.91209066e-01 -8.61091077e-01 -3.75907570e-01 5.80042362e-01
-5.18710971e-01 1.99226588e-01 4.59746365e-03 1.40153125e-01
-2.03423351e-01 8.80096853e-02 7.23993897e-01 -5.73594049e-02
-3.22604626e-01 2.49483317e-01 -1.06302500e+00 1.01920927e+00
5.86955130e-01 -8.84817839e-01 -1.92481712e-01 6.99326545e-02
5.81635118e-01 3.06382030e-01 2.87753373e-01 -6.07335806e-01
9.16963443e-02 -3.79724026e-01 7.62879014e-01 -3.71802628e-01
4.13966849e-02 -8.58464360e-01 6.27650842e-02 3.13071012e-01
-9.12155628e-01 -9.75432172e-02 -1.00821890e-01 5.27887046e-01
7.96847939e-02 -3.83985311e-01 7.13601649e-01 2.44255096e-01
7.69983139e-03 -5.17049432e-02 -2.62307912e-01 1.63037866e-01
8.05168331e-01 -1.67943597e-01 -2.06229851e-01 -2.72948086e-01
-7.88115203e-01 3.30693603e-01 3.26716185e-01 2.21606165e-01
1.02533057e-01 -1.30674756e+00 -6.90904021e-01 2.59531975e-01
7.10972846e-02 -3.31864417e-01 2.18856502e-02 9.45521533e-01
1.69981867e-01 -2.73168255e-02 -1.88940361e-01 -5.02412796e-01
-9.60544825e-01 3.92860502e-01 1.88228562e-02 -4.40416813e-01
-4.61561441e-01 7.50978053e-01 4.91005272e-01 -1.26108393e-01
4.00139868e-01 -5.80810070e-01 -1.27982736e-01 4.32512105e-01
4.97481734e-01 4.35786515e-01 -8.74288101e-03 -1.29261762e-01
-3.20376486e-01 1.92342252e-01 -6.38656497e-01 -9.26723182e-02
1.39596176e+00 1.33315831e-01 1.86221108e-01 6.37704968e-01
1.11348677e+00 1.93936467e-01 -1.42156160e+00 -4.62072529e-02
1.17146619e-01 -5.99879861e-01 -1.73539475e-01 -7.52385318e-01
-7.59873748e-01 4.86816704e-01 2.27021575e-01 4.82326061e-01
8.44258666e-01 -8.73533487e-02 5.98557055e-01 9.31168720e-02
1.76194623e-01 -1.02072835e+00 -3.01593214e-01 6.86509460e-02
5.76807737e-01 -1.19456244e+00 3.53421003e-01 -5.24660289e-01
-6.23130083e-01 8.84990394e-01 4.01116490e-01 -3.31813395e-01
6.78669631e-01 3.25547606e-01 -5.04945636e-01 5.20153642e-02
-8.82110596e-01 2.44211853e-01 2.47077405e-01 2.02041134e-01
4.41810638e-01 3.00090432e-01 -5.17863154e-01 8.95475328e-01
-2.98862129e-01 -1.32724941e-01 6.46651030e-01 6.85421884e-01
-3.86577129e-01 -9.09898996e-01 -4.06317264e-01 1.32530677e+00
-7.82129824e-01 -2.44800761e-01 4.94931825e-02 8.48385453e-01
1.98312268e-01 7.79952347e-01 6.03074789e-01 -2.74195731e-01
4.35924470e-01 3.39555919e-01 3.65082443e-01 -5.74148655e-01
-3.69168878e-01 -4.82333601e-02 2.78719485e-01 -2.46217713e-01
-3.76315080e-02 -1.12751567e+00 -8.39595079e-01 -4.78472948e-01
-4.78493899e-01 6.06142543e-02 4.86084551e-01 9.02410805e-01
-4.46852483e-02 2.05834046e-01 7.41433680e-01 -4.26807642e-01
-1.13190508e+00 -9.30611253e-01 -7.29159534e-01 5.39215446e-01
3.64944428e-01 -8.65931630e-01 -8.97180021e-01 -4.73455079e-02] | [8.42539119720459, 4.713194370269775] |
3ab87a7a-53ff-4f1a-8ed8-0600dcf36729 | emotion-recognition-in-conversation-using | 2207.07238 | null | https://arxiv.org/abs/2207.07238v1 | https://arxiv.org/pdf/2207.07238v1.pdf | Emotion Recognition in Conversation using Probabilistic Soft Logic | Creating agents that can both appropriately respond to conversations and understand complex human linguistic tendencies and social cues has been a long standing challenge in the NLP community. A recent pillar of research revolves around emotion recognition in conversation (ERC); a sub-field of emotion recognition that focuses on conversations or dialogues that contain two or more utterances. In this work, we explore an approach to ERC that exploits the use of neural embeddings along with complex structures in dialogues. We implement our approach in a framework called Probabilistic Soft Logic (PSL), a declarative templating language that uses first-order like logical rules, that when combined with data, define a particular class of graphical model. Additionally, PSL provides functionality for the incorporation of results from neural models into PSL models. This allows our model to take advantage of advanced neural methods, such as sentence embeddings, and logical reasoning over the structure of a dialogue. We compare our method with state-of-the-art purely neural ERC systems, and see almost a 20% improvement. With these results, we provide an extensive qualitative and quantitative analysis over the DailyDialog conversation dataset. | ['Lise Getoor', 'William Wang', 'Charles Dickens', 'Connor Pryor', 'Alon Albalak', 'Pegah Jandaghi', 'Eriq Augustine'] | 2022-07-14 | null | null | null | null | ['emotion-recognition-in-conversation'] | ['natural-language-processing'] | [-1.08437903e-01 6.71258628e-01 1.20954767e-01 -8.17111075e-01
-2.26790339e-01 -5.68184495e-01 1.14724946e+00 5.12798786e-01
-3.55332762e-01 5.49244702e-01 9.50604975e-01 -1.94315195e-01
1.48312366e-02 -8.37434053e-01 -4.04208601e-01 -2.35384881e-01
-1.78924665e-01 6.76444411e-01 -2.39575446e-01 -7.14547813e-01
-9.04784724e-02 4.09218818e-01 -1.29945493e+00 7.12716281e-01
4.48900759e-01 1.03895593e+00 -5.90567052e-01 9.45638955e-01
-5.46400487e-01 1.37247860e+00 -5.25889277e-01 -7.83068120e-01
-5.19814134e-01 -2.51646101e-01 -9.41883385e-01 -2.95794159e-01
8.03141817e-02 -6.21685646e-02 -1.71077520e-01 8.37237716e-01
4.29552138e-01 2.86620557e-01 8.74596894e-01 -1.35761750e+00
-6.22995079e-01 1.05918467e+00 3.18030179e-01 -1.96088955e-01
8.24006319e-01 1.14629060e-01 1.27755451e+00 -8.16921413e-01
6.60274029e-01 1.85911393e+00 9.07293260e-01 8.32787514e-01
-1.28873694e+00 -6.70617670e-02 2.20218971e-01 1.31262347e-01
-7.50687063e-01 -7.19928980e-01 9.40346479e-01 -5.33829510e-01
1.45984316e+00 3.03169280e-01 6.42126381e-01 1.40827429e+00
7.35432282e-03 1.15087712e+00 1.05947888e+00 -5.86259723e-01
4.73941177e-01 6.07972741e-01 6.38686478e-01 8.77230525e-01
-4.37621802e-01 -8.95231031e-03 -7.89000630e-01 -4.46920097e-01
-8.31566676e-02 -3.02508354e-01 -2.40409106e-01 -8.62502009e-02
-7.86519647e-01 1.20307207e+00 5.32211550e-02 4.13773060e-01
-4.37239796e-01 1.37447506e-01 7.54306972e-01 3.15918744e-01
6.03518784e-01 7.23423243e-01 -5.63891947e-01 -6.57448113e-01
-5.96802950e-01 4.50483471e-01 1.72624671e+00 4.38681126e-01
3.98573428e-01 -1.10300995e-01 -2.52191275e-01 1.05234206e+00
5.99871397e-01 6.81299195e-02 3.74393582e-01 -1.20920765e+00
1.71082005e-01 8.26828241e-01 -1.31247699e-01 -1.18858981e+00
-6.58937812e-01 1.93858027e-01 -4.80043173e-01 8.23562145e-02
2.74050921e-01 -5.52163422e-01 -2.69129604e-01 1.92769837e+00
1.83836043e-01 -1.99055985e-01 7.22234249e-01 3.40461791e-01
1.03252721e+00 9.08077419e-01 -1.67523976e-02 -4.12635803e-02
1.40705347e+00 -8.50414574e-01 -1.04407704e+00 -1.87825695e-01
6.89831972e-01 -2.85692185e-01 9.87530529e-01 5.66384315e-01
-1.06741965e+00 -1.71774384e-02 -8.93164098e-01 -2.69316912e-01
-1.01945078e+00 -2.35293433e-01 8.55592966e-01 6.55078351e-01
-1.34983552e+00 4.73331660e-01 -5.69642484e-01 -5.60531616e-01
1.04213394e-01 1.16199099e-01 -3.54831636e-01 2.52206832e-01
-1.61801147e+00 1.29573119e+00 3.07564110e-01 2.48707965e-01
-3.64962578e-01 -4.42839324e-01 -1.42090178e+00 2.07912117e-01
3.68316174e-01 -3.09564143e-01 1.37726510e+00 -7.47440934e-01
-2.23388577e+00 7.84515440e-01 -4.63548303e-02 -6.41475976e-01
2.52659291e-01 -3.01978678e-01 -3.18272650e-01 -1.42691778e-02
-4.66552943e-01 7.36975372e-01 2.99284577e-01 -1.18570864e+00
-1.90433130e-01 -2.88975775e-01 3.12929749e-01 1.48286507e-01
-3.41315061e-01 4.36051756e-01 -2.88485557e-01 -2.77946323e-01
-5.84952712e-01 -7.58450329e-01 -8.77005607e-02 -1.49931297e-01
-5.73339224e-01 -9.78513956e-01 5.51025510e-01 -4.66223359e-01
1.21093392e+00 -1.97842801e+00 4.57443953e-01 2.01111168e-01
2.23918363e-01 1.53486818e-01 -3.80023085e-02 6.94972575e-01
7.20428228e-02 2.64040291e-01 -2.31942162e-01 -9.04739618e-01
9.31782126e-01 3.52299035e-01 -4.34611499e-01 2.43230648e-02
5.49564958e-01 9.67550695e-01 -5.96748173e-01 -4.74114269e-01
2.51208454e-01 5.65872848e-01 -7.62246788e-01 4.78057593e-01
-6.51729107e-01 -1.96942776e-01 -1.46775633e-01 4.14902538e-01
1.31207794e-01 9.84096229e-02 4.89105612e-01 -4.63545062e-02
-1.06210448e-01 5.02844989e-01 -9.34043646e-01 1.35591078e+00
-6.66268528e-01 8.30233216e-01 1.67973682e-01 -8.80664647e-01
8.75161767e-01 5.08144677e-01 4.99608666e-02 -2.75736749e-01
3.50820303e-01 -2.00432137e-01 -2.66694278e-01 -7.45531440e-01
6.24681771e-01 -1.43945470e-01 -5.93988776e-01 6.40420794e-01
3.21700513e-01 -3.49641979e-01 1.47551581e-01 3.88427645e-01
9.87777710e-01 3.34879719e-02 2.73675144e-01 1.68376639e-02
5.61636269e-01 -2.13754058e-01 4.41116273e-01 7.14744270e-01
-3.07070643e-01 -3.95853780e-02 1.18885064e+00 -5.15201867e-01
-5.56755364e-01 -7.40890980e-01 3.90830450e-03 1.38596821e+00
-7.08499968e-01 -7.64373839e-01 -8.20872009e-01 -6.73784733e-01
-2.42738724e-02 9.03140664e-01 -9.19204414e-01 -6.11018762e-02
-3.72019440e-01 -6.34762228e-01 1.04969537e+00 5.12606263e-01
2.12559134e-01 -1.46668577e+00 -4.02014256e-01 3.36261392e-01
-6.67397156e-02 -1.22540498e+00 8.20989981e-02 3.58933777e-01
-1.60245419e-01 -8.56437206e-01 -1.33736074e-01 -5.37633002e-01
6.74913600e-02 -8.47627938e-01 1.31224537e+00 -2.12855443e-01
-1.33305699e-01 7.75035679e-01 -4.60489780e-01 -5.58848202e-01
-6.91258132e-01 -1.50621131e-01 2.25643978e-01 1.88301057e-01
6.59431398e-01 -4.40094620e-01 1.19080201e-01 -3.78753930e-01
-8.28405499e-01 -4.69441302e-02 -1.07906358e-02 8.28536749e-01
-1.97158292e-01 -1.73410937e-01 4.26193595e-01 -1.04787755e+00
1.41346037e+00 -5.39903581e-01 -3.48596811e-01 4.14446622e-01
-2.02449381e-01 4.17070508e-01 5.32924354e-01 -2.43362725e-01
-1.13522160e+00 -1.47890031e-01 -4.54397351e-01 -3.11273634e-02
-2.59790897e-01 9.62026119e-01 -1.29901022e-01 4.74040836e-01
5.61860442e-01 2.06694584e-02 2.07648128e-01 -2.34824941e-01
8.18552196e-01 9.57840741e-01 4.84525710e-01 -8.04116428e-01
6.21660566e-03 1.49453878e-01 -4.53817278e-01 -9.68686461e-01
-7.49826550e-01 -1.19487487e-01 -2.56846219e-01 -3.92321020e-01
1.04911256e+00 -4.70474660e-01 -1.28554332e+00 4.09147024e-01
-1.45193505e+00 -7.32509911e-01 -1.61399096e-01 3.43851596e-01
-6.24973953e-01 2.13659495e-01 -9.55921292e-01 -1.36705542e+00
-2.88195819e-01 -8.26349914e-01 8.59355330e-01 1.98832586e-01
-9.15451527e-01 -1.46624863e+00 3.97372246e-01 1.82927355e-01
5.34333169e-01 3.02906811e-01 1.22331750e+00 -1.28245652e+00
5.00868335e-02 -1.79980084e-01 -9.72891506e-03 4.94386137e-01
-3.51411670e-01 2.70234644e-01 -1.30643797e+00 3.34180295e-01
4.32391018e-02 -9.32997465e-01 6.58304572e-01 -9.44932029e-02
8.17427993e-01 -5.48952520e-01 -4.39932197e-02 1.06018290e-01
8.89972746e-01 6.57494739e-02 4.48122531e-01 -9.99053717e-02
3.54786813e-01 1.26164842e+00 3.10877897e-02 5.78262150e-01
1.00920916e+00 6.17118776e-01 1.72775492e-01 1.76494762e-01
2.75942355e-01 -1.72630116e-01 8.01970482e-01 8.87047052e-01
3.16041648e-01 -4.74097669e-01 -1.09139013e+00 3.49884331e-01
-2.06703401e+00 -1.03769374e+00 1.18587628e-01 1.36551976e+00
1.35120022e+00 1.63632482e-01 -1.24654740e-01 9.14183259e-03
3.13514590e-01 4.18377757e-01 -1.20346054e-01 -1.30007255e+00
-1.24978788e-01 2.12717772e-01 -2.48864725e-01 1.03793728e+00
-1.03106260e+00 1.01799142e+00 6.50116444e+00 3.03209573e-01
-9.85326111e-01 -1.93771079e-01 5.23253620e-01 1.16949178e-01
-3.06867868e-01 -2.88779944e-01 -7.79847622e-01 2.67442912e-01
1.37211287e+00 1.89434469e-01 6.04212880e-01 7.43913412e-01
1.70053080e-01 -1.40950233e-01 -1.56390965e+00 8.14588130e-01
5.34739971e-01 -1.48642206e+00 -3.03144336e-01 -2.38163978e-01
2.11678356e-01 -7.77361691e-02 -2.27937877e-01 9.54113901e-01
8.28040421e-01 -1.20592213e+00 5.13503253e-01 8.68246794e-01
2.37275735e-01 -4.45111662e-01 9.51577663e-01 3.28511655e-01
-5.23098350e-01 -9.21329856e-03 1.41470701e-01 -2.53405422e-01
1.54432163e-01 4.51294869e-01 -8.98050725e-01 8.09014589e-02
6.42512679e-01 7.31078565e-01 -3.52915198e-01 2.40914896e-01
-4.15404499e-01 6.18318677e-01 -5.09822667e-01 -6.83694899e-01
2.74780035e-01 2.41215024e-02 4.40117419e-01 1.79619789e+00
-2.82718956e-01 1.25522718e-01 1.61260843e-01 1.03051615e+00
-1.54032201e-01 2.89277956e-02 -8.28788042e-01 -5.54975033e-01
2.80122012e-01 1.16802919e+00 -1.12049185e-01 -5.58768928e-01
-1.86633676e-01 7.28885353e-01 6.91850722e-01 1.68867663e-01
-5.26562095e-01 -4.80315059e-01 7.64198422e-01 -7.14800119e-01
1.81085877e-02 -2.07119718e-01 -8.84906426e-02 -1.13736200e+00
-1.10933557e-01 -1.00287962e+00 3.10923249e-01 -8.33509505e-01
-1.59252357e+00 7.23672092e-01 -7.73628950e-02 -2.19579682e-01
-7.55177796e-01 -9.74050343e-01 -8.65571141e-01 6.64619446e-01
-1.38072610e+00 -9.68574047e-01 6.07869811e-02 3.80310714e-01
3.68541598e-01 -1.77848786e-01 1.47680235e+00 -9.82992053e-02
-6.76874280e-01 4.70102698e-01 -2.12821797e-01 3.81289691e-01
5.44465125e-01 -1.67036748e+00 4.43673916e-02 3.22280943e-01
-3.29372734e-02 7.55365312e-01 7.60655463e-01 -3.11866969e-01
-1.34636402e+00 -6.42506838e-01 1.54888701e+00 -8.27492595e-01
8.80195379e-01 -8.65272224e-01 -8.79762828e-01 7.23779142e-01
7.37599313e-01 -3.49450856e-01 1.14398932e+00 6.79994524e-01
-5.10283887e-01 1.66518226e-01 -1.09261107e+00 7.91839838e-01
4.24364328e-01 -9.66321230e-01 -1.10067964e+00 1.89995572e-01
8.79776418e-01 -1.72604695e-01 -8.47450137e-01 2.03253269e-01
6.48312807e-01 -9.86397624e-01 5.36899149e-01 -8.42754245e-01
5.66687703e-01 1.11835927e-01 -3.21740299e-01 -1.49357152e+00
2.38540635e-01 -8.51835072e-01 -2.77119637e-01 1.40574038e+00
6.96209788e-01 -6.18024886e-01 3.80013227e-01 1.16742599e+00
-1.30952194e-01 -7.78337240e-01 -5.67196965e-01 -1.00728333e-01
1.38091937e-01 -9.17885303e-01 4.54053819e-01 1.04401088e+00
8.36931348e-01 5.57559013e-01 -2.11652771e-01 -1.45440891e-01
4.05943953e-02 -1.66963160e-01 7.47938514e-01 -1.29410493e+00
-2.99374938e-01 -6.40132964e-01 -3.32447588e-01 -6.58776164e-01
8.60570014e-01 -9.22579229e-01 4.48194176e-01 -1.71110725e+00
-3.90505075e-01 -1.56724334e-01 -6.77247532e-03 7.59260774e-01
2.45817140e-01 -1.45158052e-01 2.22510993e-01 -5.62179446e-01
-7.07422614e-01 8.67493451e-01 3.79085630e-01 -1.89990446e-01
-2.98987955e-01 -4.79943484e-01 -7.14792907e-01 1.05318511e+00
7.69611180e-01 -2.20416993e-01 -1.63786411e-01 -1.02961771e-01
7.07574487e-01 9.71652791e-02 3.69875968e-01 -6.52984381e-01
4.46104258e-01 -3.03001329e-02 6.21894468e-03 -2.62910277e-01
8.42067242e-01 -5.90288997e-01 -4.61329490e-01 -2.26318482e-02
-8.80196512e-01 1.34502864e-02 3.09119731e-01 3.34264249e-01
-3.66654575e-01 -1.23880371e-01 4.05310690e-01 2.33210493e-02
-4.76030916e-01 -3.69374096e-01 -8.54936063e-01 1.65434271e-01
6.39623940e-01 2.42780700e-01 -4.05783534e-01 -6.95718110e-01
-1.00542700e+00 6.28250778e-01 3.49172540e-02 6.13934636e-01
6.22007430e-01 -1.10587645e+00 -5.51446140e-01 5.64257205e-02
2.19789788e-01 -2.24864841e-01 2.97127813e-02 7.96278358e-01
-2.14997306e-01 4.46804941e-01 1.17283583e-01 -1.65989384e-01
-1.16196144e+00 2.05012515e-01 5.41070223e-01 -2.79687732e-01
-2.71637589e-01 9.47910666e-01 -2.81500459e-01 -1.22808921e+00
6.35236144e-01 -5.57551026e-01 -6.21553779e-01 4.38997030e-01
5.98946750e-01 -1.11657910e-01 -2.26303041e-01 -4.01998907e-01
-4.73057538e-01 2.74702627e-03 -2.26916410e-02 -5.37159622e-01
1.51602614e+00 1.61191344e-01 -5.48811376e-01 1.00480175e+00
1.28746176e+00 1.54161453e-02 -9.16921198e-01 -3.54799062e-01
2.15036571e-01 2.96871424e-01 -1.40712867e-02 -1.25970268e+00
-3.25067431e-01 9.64417875e-01 1.86135277e-01 6.94201112e-01
4.88738954e-01 2.31470838e-01 3.68373811e-01 8.54061961e-01
-1.71475887e-01 -1.36412561e+00 7.82703683e-02 1.23757076e+00
1.13449097e+00 -1.19227183e+00 -4.32361543e-01 -1.17656469e-01
-9.79988277e-01 1.36573184e+00 3.00249547e-01 5.09762801e-02
7.44853139e-01 4.53793198e-01 2.77899444e-01 -5.99063396e-01
-1.35058331e+00 -1.54593408e-01 -6.62626401e-02 4.03019726e-01
6.45368636e-01 2.86944211e-01 -4.66908365e-02 1.02181482e+00
-4.63350445e-01 5.22750393e-02 5.00099003e-01 8.38716149e-01
-2.23154649e-01 -1.08968091e+00 -3.12648378e-02 2.16900617e-01
-3.15898240e-01 -2.27518037e-01 -9.83466387e-01 5.31651378e-01
-7.66351670e-02 1.26964521e+00 2.21295342e-01 -4.10468221e-01
2.69970208e-01 8.07568371e-01 8.95880535e-02 -7.32911468e-01
-8.28052640e-01 -5.70721686e-01 8.09985876e-01 -5.49555004e-01
-5.87560356e-01 -6.67624295e-01 -1.39214981e+00 -2.35114664e-01
3.53628471e-02 2.65022397e-01 8.10742378e-01 1.11263466e+00
2.98533916e-01 3.85064304e-01 4.13940817e-01 -7.51948297e-01
-4.92716998e-01 -1.09056878e+00 -1.81922421e-01 1.60263777e-01
4.46913421e-01 -4.29864764e-01 -5.18715382e-01 -1.06708594e-01] | [12.980374336242676, 6.275015830993652] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.