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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
da5b2d69-c550-4d50-aceb-e18cb411b29e | deep-graph-clustering-via-mutual-information | 2205.05168 | null | https://arxiv.org/abs/2205.05168v1 | https://arxiv.org/pdf/2205.05168v1.pdf | Deep Graph Clustering via Mutual Information Maximization and Mixture Model | Attributed graph clustering or community detection which learns to cluster the nodes of a graph is a challenging task in graph analysis. In this paper, we introduce a contrastive learning framework for learning clustering-friendly node embedding. Although graph contrastive learning has shown outstanding performance in self-supervised graph learning, using it for graph clustering is not well explored. We propose Gaussian mixture information maximization (GMIM) which utilizes a mutual information maximization approach for node embedding. Meanwhile, it assumes that the representation space follows a Mixture of Gaussians (MoG) distribution. The clustering part of our objective tries to fit a Gaussian distribution to each community. The node embedding is jointly optimized with the parameters of MoG in a unified framework. Experiments on real-world datasets demonstrate the effectiveness of our method in community detection. | ['Abdolreza Mirzaei', 'Mehran Safayani', 'Maedeh Ahmadi'] | 2022-05-10 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [-1.65726915e-01 2.59129316e-01 -2.91238725e-01 -2.02050701e-01
-3.71921510e-01 -3.00264299e-01 4.80005205e-01 4.27605927e-01
-4.40154448e-02 1.28085807e-01 -1.13786772e-01 -1.90441892e-01
-3.10241580e-01 -7.92986989e-01 -2.46476904e-01 -9.55563128e-01
-4.65934277e-01 4.55859512e-01 -1.86653938e-02 2.60631889e-01
-3.05220056e-02 3.63943160e-01 -7.23259509e-01 -3.16364259e-01
8.85669768e-01 1.68743297e-01 3.44102591e-01 7.79893816e-01
-4.03793864e-02 8.94723058e-01 -2.86274046e-01 -3.67528737e-01
4.66326512e-02 -2.00163186e-01 -6.60690784e-01 5.55327296e-01
4.21252698e-02 1.99027553e-01 -7.09747612e-01 1.25528204e+00
3.14573586e-01 1.70200672e-02 1.17745638e+00 -1.91851938e+00
-5.96299469e-01 8.09241772e-01 -7.94502735e-01 -7.45039210e-02
1.83975175e-02 -2.37938628e-01 1.20782661e+00 -6.25856698e-01
3.86934221e-01 1.34556925e+00 5.39813638e-01 1.90409273e-01
-1.39724696e+00 -4.66048121e-01 3.99896234e-01 1.86701983e-01
-1.98730624e+00 7.60879740e-02 1.24553502e+00 -6.50753677e-01
4.46456343e-01 1.54672787e-02 6.45799935e-01 5.22129416e-01
6.31539524e-02 8.02912652e-01 5.57917595e-01 -2.99877524e-01
2.32022196e-01 3.32097411e-01 2.04596743e-01 8.94872069e-01
5.02481699e-01 -4.69082117e-01 -1.93961442e-01 -4.42470878e-01
6.75108969e-01 1.73737168e-01 -8.50386694e-02 -1.03248656e+00
-8.61151218e-01 1.22798920e+00 7.10758269e-01 3.48019034e-01
-2.35074967e-01 3.77800822e-01 8.21759403e-02 1.85158685e-01
5.97954512e-01 -1.95119902e-02 9.26670358e-02 5.58511913e-01
-7.99958825e-01 -2.02442139e-01 9.61723030e-01 1.16281116e+00
1.12543762e+00 6.54714182e-02 2.05277786e-01 7.64974058e-01
9.66530323e-01 3.58169705e-01 -6.34011701e-02 -7.07883894e-01
1.25285938e-01 1.07296109e+00 -4.75947320e-01 -1.55309224e+00
-2.96778083e-01 -6.51751518e-01 -1.27334774e+00 -1.13490693e-01
1.11040235e-01 -1.38293356e-01 -4.89216030e-01 1.66906273e+00
5.33898473e-01 6.05354786e-01 -7.30123231e-03 5.77642798e-01
7.87344456e-01 4.42476779e-01 7.64665380e-02 -2.22768083e-01
1.01764977e+00 -8.31599951e-01 -7.76505172e-01 -5.34192286e-02
5.83413541e-01 -3.96462321e-01 5.14912307e-01 -3.12089864e-02
-5.91724634e-01 -2.42543861e-01 -9.60398257e-01 3.63910258e-01
-1.53413922e-01 -4.12314162e-02 8.40430915e-01 7.57088006e-01
-1.27235901e+00 3.02068532e-01 -9.60617065e-01 -4.98723179e-01
4.08807874e-01 4.02763754e-01 -6.23289108e-01 -5.63832484e-02
-7.19231963e-01 2.57298470e-01 5.87935627e-01 -1.67165279e-01
-9.04203296e-01 -2.14805022e-01 -1.23969805e+00 2.28016183e-01
5.55403829e-01 -7.46138632e-01 4.62565750e-01 -4.81471777e-01
-1.07631707e+00 8.34492922e-01 -7.40898699e-02 -2.73785859e-01
5.96024767e-02 2.73065686e-01 -3.91274482e-01 3.68500531e-01
1.37862220e-01 5.01029849e-01 1.22468174e+00 -1.49210584e+00
-2.00849205e-01 -3.55083883e-01 -3.03388685e-01 2.02167317e-01
-8.07108700e-01 -1.59597635e-01 -7.34154105e-01 -5.59578121e-01
6.20834708e-01 -9.50991690e-01 -6.39520168e-01 -1.33636340e-01
-7.34155834e-01 -5.43084800e-01 8.16725135e-01 -4.09313142e-01
1.67717075e+00 -2.15521860e+00 2.86459923e-01 7.43178666e-01
1.00030124e+00 -2.39946634e-01 -1.08470738e-01 7.69125044e-01
-6.91803247e-02 2.71147966e-01 -4.10508662e-01 -5.31168401e-01
1.42249838e-01 3.82150412e-02 3.00246239e-01 9.26002383e-01
8.11834186e-02 6.80253327e-01 -1.08111799e+00 -9.45809543e-01
2.28133842e-01 7.14486718e-01 -6.26001000e-01 4.67917830e-01
2.66011506e-01 7.99961537e-02 -3.79828304e-01 4.95306760e-01
7.79301286e-01 -7.50893474e-01 8.26956809e-01 -3.44883978e-01
4.71416920e-01 -3.83280873e-01 -1.57280862e+00 1.34901023e+00
3.01117808e-01 4.39725757e-01 2.49995455e-01 -1.21884084e+00
9.59277749e-01 1.92876413e-01 7.12432384e-01 3.42599183e-01
8.43886845e-03 -2.51937658e-01 8.94548893e-02 -3.01718920e-01
1.06536746e-01 2.40866318e-01 4.36737090e-02 5.46269774e-01
3.58829409e-01 1.16013095e-01 2.14940712e-01 8.67994547e-01
1.34855890e+00 -4.13450181e-01 4.45626169e-01 -3.98123235e-01
5.22988319e-01 -3.85945708e-01 4.66494262e-01 6.11811697e-01
-3.52702707e-01 4.12884921e-01 5.74919581e-01 8.99765342e-02
-7.92176366e-01 -1.25005269e+00 1.92453668e-01 8.80121768e-01
2.66043305e-01 -9.73576725e-01 -7.98738241e-01 -8.37005317e-01
1.40802320e-02 2.17607722e-01 -4.62123305e-01 -2.19715789e-01
-3.36679518e-01 -9.47089672e-01 2.30323642e-01 2.25106880e-01
3.88625786e-02 -6.51592374e-01 2.44654313e-01 1.86140880e-01
-2.54942775e-02 -1.05428445e+00 -7.22260892e-01 7.01855198e-02
-7.66592503e-01 -1.35737836e+00 -4.03850943e-01 -1.26353383e+00
1.01675487e+00 6.94004178e-01 1.02663195e+00 3.43856484e-01
-2.34315217e-01 8.44556034e-01 -2.02779725e-01 1.60703897e-01
-4.30859625e-01 2.92734206e-01 4.30719033e-02 3.08491558e-01
3.27749670e-01 -8.31588268e-01 -4.36025798e-01 1.14256702e-01
-8.24814737e-01 -1.52510062e-01 5.52216470e-01 5.44639766e-01
3.15580398e-01 4.69557136e-01 2.89362967e-01 -9.52545822e-01
9.34137166e-01 -9.10707295e-01 -3.40925187e-01 2.72475660e-01
-9.09479260e-01 7.22563937e-02 3.14707130e-01 -4.37403142e-01
-4.72306073e-01 7.71004334e-02 3.44253361e-01 -6.67054594e-01
-6.72969073e-02 8.17832351e-01 -5.29630661e-01 -4.05921750e-02
1.34307891e-01 2.32679293e-01 3.62204403e-01 -3.56163979e-01
7.00977564e-01 5.42460918e-01 4.10056114e-01 -3.09027076e-01
1.33358765e+00 5.74512720e-01 1.07022807e-01 -1.01650465e+00
-1.99035570e-01 -1.01180542e+00 -8.40146124e-01 -4.12439793e-01
6.72832847e-01 -9.92823899e-01 -9.09486771e-01 2.16899827e-01
-7.69578755e-01 -8.57453719e-02 2.44657055e-01 5.78944266e-01
-3.16564739e-01 9.53588188e-01 -6.34794176e-01 -1.09586835e+00
-1.25577196e-01 -8.54549766e-01 7.40895629e-01 5.61711378e-02
7.31716072e-03 -1.60598743e+00 3.45523596e-01 1.22851454e-01
3.92557681e-02 3.43880802e-01 8.09500217e-01 -7.96164453e-01
-4.85722214e-01 -3.59633207e-01 -3.16686124e-01 2.21286744e-01
4.40299690e-01 2.81401187e-01 -6.08444393e-01 -7.16220677e-01
-1.73438773e-01 9.70924124e-02 8.50029767e-01 4.00121182e-01
8.75923455e-01 -4.53622639e-01 -7.84062028e-01 6.34811401e-01
1.54110193e+00 -2.41358414e-01 3.13492358e-01 -7.28164241e-02
1.29422224e+00 7.14318454e-01 2.81515479e-01 4.11730319e-01
7.67525017e-01 3.24039489e-01 6.52087152e-01 -5.01821600e-02
1.02777168e-01 -3.45987916e-01 3.70944917e-01 1.25949967e+00
3.10081244e-01 -3.89203668e-01 -1.00653958e+00 5.57600796e-01
-2.07647085e+00 -9.08635557e-01 -5.21739841e-01 1.98653638e+00
5.05123794e-01 4.95593511e-02 3.62155050e-01 6.63592517e-02
1.22172797e+00 1.92792162e-01 -3.59304339e-01 2.26020530e-01
-2.32015833e-01 -3.06169838e-01 3.90823364e-01 7.01280057e-01
-1.42756164e+00 8.92634690e-01 5.92369699e+00 8.45385373e-01
-6.03959084e-01 7.66292065e-02 3.33509952e-01 6.84748292e-01
-2.89215088e-01 4.47469294e-01 -4.15458977e-01 3.94894004e-01
6.31436884e-01 -3.38424742e-01 3.40187371e-01 9.85031486e-01
-1.93412192e-02 3.85662794e-01 -9.59128439e-01 1.12003589e+00
1.32774234e-01 -9.98094678e-01 -2.04199329e-02 5.00181794e-01
7.14992940e-01 -6.78421780e-02 -5.62459696e-03 2.74249971e-01
6.40429139e-01 -1.04462874e+00 1.24225572e-01 4.74276572e-01
3.74821961e-01 -8.19189727e-01 4.82027084e-01 4.91897106e-01
-1.65669811e+00 1.13756821e-01 -4.81910527e-01 7.41628334e-02
-3.07621863e-02 7.01629639e-01 -1.23024309e+00 5.76024115e-01
3.69189262e-01 1.04356921e+00 -8.46023083e-01 1.12096238e+00
-1.88978225e-01 8.37303460e-01 -2.49644488e-01 2.60307896e-03
1.35652423e-01 -7.18676984e-01 8.65218759e-01 1.28184581e+00
-1.51891097e-01 -3.59719694e-01 6.08900070e-01 1.01851857e+00
-1.60288721e-01 4.92484480e-01 -7.24754393e-01 -4.02745694e-01
8.06024790e-01 1.73720312e+00 -1.06216252e+00 -1.79822966e-01
-2.49381587e-01 9.55038071e-01 7.69025981e-01 4.55995053e-01
-5.64670146e-01 -2.09736541e-01 4.73012865e-01 6.36739284e-02
2.66255558e-01 -4.22681361e-01 2.59343892e-01 -1.11452711e+00
-2.55610853e-01 -5.88066757e-01 6.65839136e-01 -1.97888330e-01
-1.67989337e+00 3.97847027e-01 1.00759938e-01 -1.21020138e+00
6.37193862e-03 -3.90095621e-01 -1.05094421e+00 5.25871396e-01
-1.24606669e+00 -1.33580971e+00 -3.99242461e-01 7.06511140e-01
-1.24403745e-01 -3.75544161e-01 6.50392175e-01 1.84969068e-01
-8.48580480e-01 5.46905041e-01 2.13838428e-01 5.04914463e-01
5.46084225e-01 -1.72766960e+00 1.10611409e-01 7.55548239e-01
3.04415315e-01 8.50248873e-01 6.34687066e-01 -8.59838188e-01
-1.50652671e+00 -1.38314152e+00 5.46596825e-01 -5.41799590e-02
7.76631355e-01 -5.45553207e-01 -1.02539444e+00 7.22186983e-01
3.45933348e-01 8.10129941e-02 1.04264045e+00 1.99597642e-01
-4.78420019e-01 -3.11420057e-02 -1.03057766e+00 5.22068322e-01
8.62208903e-01 -5.74918926e-01 6.20185398e-02 5.51699698e-01
6.55756116e-01 3.92285228e-01 -1.32735837e+00 2.93805510e-01
1.18290819e-01 -5.92927277e-01 9.99093354e-01 -5.66527069e-01
-1.81233078e-01 -6.51718020e-01 -1.26190916e-01 -1.32356143e+00
-7.86078513e-01 -7.74193883e-01 -5.24947166e-01 1.52378452e+00
3.95241287e-03 -5.66881478e-01 1.03521621e+00 8.01559538e-02
1.66228309e-01 -5.92536747e-01 -5.98915756e-01 -7.68476963e-01
-5.79675734e-02 -2.64128566e-01 3.68172407e-01 1.36965549e+00
1.59525454e-01 7.32368171e-01 -3.35178912e-01 6.27425194e-01
1.49451458e+00 -1.41061962e-01 1.05290723e+00 -1.53654408e+00
-2.69003689e-01 -6.14775658e-01 -7.73397088e-01 -6.81933999e-01
4.69377011e-01 -1.35589325e+00 -2.28822872e-01 -1.68380272e+00
7.00325131e-01 -3.63752097e-01 -3.13902140e-01 -2.17444319e-02
-6.16207480e-01 -1.22726582e-01 1.73830822e-01 3.86225581e-01
-1.10886455e+00 6.57757938e-01 6.05547726e-01 -5.21580875e-01
-1.37006685e-01 5.15635088e-02 -6.25583887e-01 6.86212838e-01
8.70775163e-01 -6.57324851e-01 -5.96048295e-01 1.63269162e-01
5.21861762e-02 -1.07424244e-01 3.07499766e-01 -8.00922453e-01
5.79389751e-01 9.22799408e-02 -3.10001411e-02 -7.56421506e-01
1.99213058e-01 -8.88104200e-01 3.83626938e-01 4.72021282e-01
-8.23308229e-02 -1.25558376e-01 -4.61319596e-01 1.05966067e+00
-1.38168842e-01 -1.79306015e-01 7.13216364e-01 2.46295743e-02
-4.57374394e-01 6.27488613e-01 -2.96018630e-01 -1.61995254e-02
1.04320645e+00 -1.51402444e-01 7.67273009e-02 -6.59457326e-01
-9.05488014e-01 6.09711051e-01 5.14580905e-01 7.30550289e-02
9.09571588e-01 -1.53762329e+00 -1.08032596e+00 -1.06437005e-01
3.61078143e-01 -1.33446410e-01 1.33508712e-01 9.42730069e-01
-2.39836320e-01 -4.39963937e-02 4.15325612e-01 -9.16687369e-01
-1.35777366e+00 8.12730789e-01 1.48853913e-01 -4.80068207e-01
-4.32668477e-01 7.49905288e-01 2.92354017e-01 -7.27082968e-01
3.39963108e-01 3.82800430e-01 -5.16333163e-01 -1.17286490e-02
1.78247854e-01 4.61349577e-01 -4.67806906e-01 -8.50042403e-01
-4.20571327e-01 2.31345847e-01 -1.32435232e-01 3.14459980e-01
1.32196724e+00 -3.97407472e-01 -6.02350533e-01 4.41901326e-01
1.48633611e+00 -2.56135222e-02 -1.06974018e+00 -4.96022671e-01
3.81700307e-01 -2.19487205e-01 2.43768618e-02 2.75227040e-01
-1.10930467e+00 6.06721222e-01 3.95364016e-01 5.91542006e-01
7.04970181e-01 3.19146663e-01 2.26110235e-01 3.07179898e-01
1.68484524e-02 -8.08719754e-01 3.84324282e-01 7.40279779e-02
3.49835843e-01 -1.37280047e+00 9.65144858e-02 -8.66472244e-01
-5.11320651e-01 1.01741517e+00 6.01618588e-01 -4.32744473e-01
1.22250736e+00 8.79611596e-02 -2.10070938e-01 -5.54341853e-01
-5.87937653e-01 -4.36529219e-01 4.96903032e-01 8.93964171e-01
3.41926247e-01 3.45969319e-01 -6.83089718e-02 4.27855432e-01
1.10277660e-01 -8.72561276e-01 5.07255673e-01 7.42737293e-01
-5.01991749e-01 -1.08352005e+00 -2.59048194e-01 3.95770013e-01
-1.12466380e-01 -1.11289687e-01 -5.08524537e-01 6.85779452e-01
-2.53963083e-01 1.22104251e+00 -1.00309119e-01 -5.48260927e-01
-2.50891775e-01 -2.30700240e-01 1.94924295e-01 -8.36452603e-01
-1.69851229e-01 4.87461925e-01 -4.32003856e-01 -2.14017645e-01
-5.34261405e-01 -5.65330207e-01 -1.16595685e+00 -3.04407686e-01
-5.83535850e-01 5.38715720e-01 5.11635363e-01 5.77436507e-01
2.11617693e-01 3.73861134e-01 1.03124082e+00 -6.15172982e-01
-3.27717572e-01 -9.33312595e-01 -1.01079786e+00 4.25003052e-01
8.76909271e-02 -5.20459890e-01 -6.25122845e-01 -7.62095749e-02] | [7.314606666564941, 5.825501918792725] |
c1b75bde-33f1-4f7e-83bc-321de0b8f14b | distilling-effective-supervision-for-robust | 2106.11099 | null | https://arxiv.org/abs/2106.11099v1 | https://arxiv.org/pdf/2106.11099v1.pdf | Distilling effective supervision for robust medical image segmentation with noisy labels | Despite the success of deep learning methods in medical image segmentation tasks, the human-level performance relies on massive training data with high-quality annotations, which are expensive and time-consuming to collect. The fact is that there exist low-quality annotations with label noise, which leads to suboptimal performance of learned models. Two prominent directions for segmentation learning with noisy labels include pixel-wise noise robust training and image-level noise robust training. In this work, we propose a novel framework to address segmenting with noisy labels by distilling effective supervision information from both pixel and image levels. In particular, we explicitly estimate the uncertainty of every pixel as pixel-wise noise estimation, and propose pixel-wise robust learning by using both the original labels and pseudo labels. Furthermore, we present an image-level robust learning method to accommodate more information as the complements to pixel-level learning. We conduct extensive experiments on both simulated and real-world noisy datasets. The results demonstrate the advantageous performance of our method compared to state-of-the-art baselines for medical image segmentation with noisy labels. | ['Ji Wu', 'Jialin Shi'] | 2021-06-21 | null | null | null | null | ['noise-estimation'] | ['medical'] | [ 5.01243114e-01 2.61005074e-01 -1.34721115e-01 -7.44313240e-01
-1.55830538e+00 -1.58603072e-01 -8.98671448e-02 3.66215408e-02
-6.59882724e-01 6.18878543e-01 4.58110683e-02 -7.42488801e-02
1.98190421e-01 -5.27897537e-01 -7.32374370e-01 -9.57332492e-01
2.72531092e-01 2.40780115e-01 2.90262878e-01 2.03815147e-01
-1.80669099e-01 5.37582524e-02 -1.25013626e+00 2.34888300e-01
1.00849867e+00 1.15714037e+00 2.88987428e-01 5.32337427e-01
-1.76405296e-01 6.44220352e-01 -6.50236130e-01 -2.04333663e-01
4.18027699e-01 -3.95267457e-01 -8.52745712e-01 6.54455364e-01
4.27166164e-01 -3.57636094e-01 -1.06463023e-01 1.48269963e+00
6.40921593e-01 -4.70272787e-02 4.28889275e-01 -8.04753721e-01
-3.97189915e-01 6.05503142e-01 -7.67637134e-01 1.35466039e-01
-1.92687929e-01 3.76620770e-01 6.41705394e-01 -5.90620577e-01
4.78642911e-01 1.22553504e+00 7.70268440e-01 6.13384604e-01
-1.17743397e+00 -4.87577766e-01 2.50590116e-01 -1.87862590e-01
-1.16547501e+00 -1.90301627e-01 8.23793888e-01 -3.90637308e-01
2.25701988e-01 3.71551663e-02 3.41762602e-01 9.02650535e-01
6.95559531e-02 1.10911155e+00 1.45534074e+00 -3.88439417e-01
2.73730695e-01 6.09767362e-02 1.26873553e-01 8.67171705e-01
2.03934595e-01 2.41402909e-02 -1.41049251e-01 4.57034865e-03
8.92129719e-01 1.24520687e-02 -3.02994192e-01 -1.37559325e-01
-1.17267346e+00 5.76151013e-01 7.70732522e-01 2.23265097e-01
-3.37020397e-01 3.15088958e-01 4.54472333e-01 -1.76167116e-01
7.22271323e-01 2.46844009e-01 -5.23269534e-01 2.00941026e-01
-1.17643785e+00 -2.65984178e-01 5.42482018e-01 8.41976881e-01
8.23226988e-01 -3.97619419e-02 -4.96321440e-01 9.33806121e-01
4.58447784e-01 5.29525399e-01 3.75024170e-01 -1.26310384e+00
3.89400929e-01 3.51106197e-01 2.24629603e-02 -7.86008477e-01
-5.50069809e-01 -6.70731246e-01 -1.09965503e+00 1.69368401e-01
5.60548782e-01 -2.22844332e-01 -1.58048391e+00 1.60931528e+00
4.33112711e-01 5.43821692e-01 -9.23389271e-02 9.56474066e-01
8.60781789e-01 4.27118570e-01 2.11153731e-01 -3.34486842e-01
1.30093527e+00 -1.26721025e+00 -1.03390515e+00 -2.65748829e-01
6.55400991e-01 -7.22317278e-01 1.16032612e+00 4.38923061e-01
-1.12827587e+00 -7.40153193e-01 -8.90279293e-01 -1.01823121e-01
-5.19516319e-02 1.98073015e-01 3.62552762e-01 7.47687101e-01
-8.86158288e-01 5.87267876e-01 -1.15105307e+00 1.11719951e-01
9.32610691e-01 1.88405976e-01 -3.51652130e-02 -2.48336375e-01
-1.01014221e+00 5.41150212e-01 3.59416574e-01 4.17389035e-01
-9.62460160e-01 -6.28134131e-01 -1.03526366e+00 -3.36711943e-01
6.81817710e-01 -5.23920774e-01 1.26045573e+00 -8.73062789e-01
-1.27221537e+00 9.83948529e-01 -9.12346616e-02 -3.05805206e-01
8.28330636e-01 -2.03436375e-01 -3.01181991e-02 3.15221310e-01
4.11326140e-01 9.41366076e-01 8.04071665e-01 -1.81156325e+00
-6.45326674e-01 -3.52227211e-01 -1.95778579e-01 6.85750172e-02
-2.81966981e-02 -2.86568463e-01 -8.03743005e-01 -7.66365707e-01
5.84500730e-01 -8.55607688e-01 -7.90038586e-01 1.94366276e-01
-6.83361232e-01 -1.87112242e-02 6.65053070e-01 -7.64876783e-01
9.72906828e-01 -2.05146027e+00 -2.38996819e-01 1.55992880e-01
2.59943426e-01 3.58432919e-01 -1.40591398e-01 -4.82046753e-01
2.28295013e-01 3.10834467e-01 -7.82163262e-01 -6.22556508e-01
-2.30180621e-01 7.15123475e-01 1.08617671e-01 4.35247481e-01
3.29491824e-01 1.02361953e+00 -1.14032221e+00 -1.00938523e+00
3.42436075e-01 4.92964447e-01 -2.66528904e-01 2.43379369e-01
-2.53742069e-01 8.72545898e-01 -6.44954026e-01 9.41657543e-01
7.81357706e-01 -3.71159196e-01 -1.58970058e-01 -4.95735645e-01
3.25725436e-01 -5.09708263e-02 -1.34781051e+00 1.87351656e+00
-3.19908619e-01 1.13524497e-01 4.91351277e-01 -9.52086031e-01
5.63704371e-01 3.98642033e-01 5.72011054e-01 -6.01269305e-01
3.18798810e-01 3.98628592e-01 -3.55058491e-01 -7.96131372e-01
1.11690260e-01 -2.58707672e-01 -6.96409196e-02 2.46679679e-01
-8.07755347e-03 -5.09420037e-01 1.90801099e-01 5.99468052e-02
9.87056971e-01 2.62585640e-01 7.27209821e-02 -2.07128018e-01
4.79271293e-01 -1.86404839e-01 1.05299568e+00 9.49841917e-01
-6.95892572e-01 1.10275519e+00 3.41715723e-01 -2.35521078e-01
-8.20468187e-01 -7.68144310e-01 -4.79048073e-01 7.65286922e-01
3.86829555e-01 6.28375560e-02 -1.23439348e+00 -1.03656459e+00
-3.60247403e-01 2.46347904e-01 -5.66129923e-01 9.41940248e-02
-7.23719835e-01 -1.14874911e+00 6.48497462e-01 6.81633294e-01
7.60533869e-01 -1.10298896e+00 -4.30080622e-01 2.94165969e-01
-5.36108375e-01 -1.35253370e+00 -5.39042950e-01 4.96091664e-01
-9.99974966e-01 -9.72203135e-01 -8.08957100e-01 -9.15181756e-01
9.79867399e-01 9.27853510e-02 1.21605694e+00 3.94693643e-01
-5.06901681e-01 1.67037755e-01 -2.42513776e-01 -3.33640277e-01
-4.77698147e-01 -9.76706445e-02 -2.52330750e-01 -1.49266332e-01
-3.39605398e-02 -7.04838187e-02 -8.02950799e-01 4.36238885e-01
-1.27111089e+00 -1.09175637e-01 8.04280698e-01 1.18374097e+00
1.28614378e+00 2.38795310e-01 6.09377444e-01 -1.43500459e+00
2.70113349e-01 -1.57751352e-01 -5.39750695e-01 3.10634851e-01
-5.44848800e-01 6.04534112e-02 3.10102880e-01 -1.90637127e-01
-1.19738495e+00 4.77864534e-01 -5.67597151e-01 -2.52466679e-01
-4.59743887e-01 3.44433367e-01 -2.54052609e-01 4.97440621e-02
6.74437642e-01 -1.01938903e-01 -2.06142411e-01 -4.03128415e-01
4.63672221e-01 6.42817318e-01 7.68139780e-01 -6.82615519e-01
5.75416565e-01 6.68225646e-01 -1.11662783e-01 -4.65164721e-01
-1.58982849e+00 -7.66207278e-01 -7.72756875e-01 -4.81780171e-02
1.11798632e+00 -9.77820158e-01 -1.72380462e-01 6.87115610e-01
-9.87229168e-01 -4.86092299e-01 -4.67530161e-01 3.08191210e-01
-5.20584166e-01 5.80701590e-01 -1.00812888e+00 -6.70001745e-01
-3.26420307e-01 -1.71973300e+00 1.48768723e+00 2.67702788e-01
2.65888870e-01 -9.72254038e-01 -2.86349624e-01 8.27861905e-01
1.15195863e-01 4.27747875e-01 5.86894631e-01 -3.98419917e-01
-6.77058995e-01 -1.43987313e-01 -5.99105895e-01 7.80613244e-01
2.12374717e-01 -3.89786810e-01 -1.16802931e+00 -1.64448097e-01
3.79339695e-01 -6.50487840e-01 1.15735936e+00 8.60983849e-01
1.57167876e+00 1.14573784e-01 -4.11391288e-01 7.87684023e-01
1.42549956e+00 -1.82777315e-01 4.93459463e-01 1.26428187e-01
1.02754796e+00 5.02822578e-01 7.95017362e-01 9.81037244e-02
3.01677227e-01 3.46602350e-01 3.95089269e-01 -7.07757771e-01
-5.70869625e-01 -1.83742251e-02 -2.90684193e-01 8.50133479e-01
3.52524340e-01 -1.52671590e-01 -8.51894259e-01 5.85428655e-01
-1.94719470e+00 -3.99924099e-01 -2.36609891e-01 1.86365318e+00
1.31206012e+00 2.45816290e-01 -2.82107651e-01 1.68061137e-01
7.69391656e-01 -4.38743550e-03 -7.49458373e-01 2.67861873e-01
2.17913724e-02 6.74205199e-02 7.55587876e-01 5.19405544e-01
-1.56767786e+00 8.40595067e-01 6.57538366e+00 1.00226223e+00
-7.95834959e-01 3.61233920e-01 1.32737136e+00 1.93481535e-01
-1.36275113e-01 -3.58171761e-01 -5.98456264e-01 6.12605989e-01
5.91841996e-01 5.82570314e-01 -7.41542503e-02 8.07773829e-01
3.23266625e-01 -3.68273616e-01 -9.15265858e-01 1.05431664e+00
5.14754392e-02 -1.07432640e+00 -3.21743399e-01 -7.02984482e-02
1.31121826e+00 3.37249413e-02 -4.19844873e-02 8.98990184e-02
3.62230957e-01 -9.80500102e-01 4.99175549e-01 3.68396103e-01
5.80179155e-01 -4.46281433e-01 1.19552875e+00 3.90028775e-01
-9.02465284e-01 8.14113095e-02 -4.15036112e-01 4.99646217e-01
3.79041821e-01 1.19384956e+00 -4.26483005e-01 4.31091666e-01
7.79469371e-01 6.27336681e-01 -5.03835857e-01 1.09112954e+00
-5.85294843e-01 8.47741306e-01 -2.93615997e-01 5.48302829e-01
4.00349468e-01 -1.93845570e-01 7.84022138e-02 1.34242320e+00
1.58369057e-02 3.13473225e-01 7.54610658e-01 8.41525972e-01
-4.56784427e-01 -1.37073860e-01 -1.30017549e-01 4.41152483e-01
1.89549387e-01 1.49265003e+00 -1.22805047e+00 -4.09523487e-01
-3.31007898e-01 9.55477774e-01 1.16926894e-01 4.11716700e-01
-6.94983721e-01 1.24237156e-02 1.48074880e-01 -5.17120063e-02
-3.77702191e-02 -7.55432807e-03 -8.10448587e-01 -8.78470898e-01
1.02990709e-01 -7.12419927e-01 3.92208755e-01 -5.41627586e-01
-1.56020713e+00 5.76675057e-01 -2.11457700e-01 -9.97301221e-01
1.21248916e-01 -4.66803879e-01 -2.83703655e-01 6.98043287e-01
-1.95914984e+00 -1.23247993e+00 -4.38912809e-01 3.30582201e-01
7.35949218e-01 3.30987155e-01 5.46607733e-01 4.74899441e-01
-5.15032411e-01 6.36168063e-01 1.41030684e-01 3.88984889e-01
7.83811748e-01 -1.51745296e+00 1.49573535e-01 8.75773370e-01
1.32745922e-01 1.43045008e-01 5.00420451e-01 -7.49294102e-01
-6.86298192e-01 -1.32341743e+00 3.03892583e-01 -3.71513605e-01
2.98351884e-01 -1.85057074e-01 -1.12225497e+00 4.11970466e-01
-5.60538955e-02 6.37778342e-01 4.68287498e-01 -2.05760613e-01
-2.04093438e-02 -5.58181368e-02 -1.51770580e+00 3.57401460e-01
8.43718171e-01 -2.73841769e-01 -3.45156759e-01 6.76090479e-01
9.34736013e-01 -7.28750050e-01 -8.06951463e-01 6.80303633e-01
-7.26125995e-03 -7.94803739e-01 8.97405863e-01 3.40466984e-02
2.58611828e-01 -4.54287469e-01 1.80881992e-02 -1.19753659e+00
-1.70875601e-02 -2.60129809e-01 2.53311723e-01 1.13664460e+00
4.99533147e-01 -2.06093878e-01 1.00918257e+00 7.30410516e-01
-3.78821850e-01 -9.14879978e-01 -9.28938210e-01 -5.15646815e-01
7.42583573e-02 -6.20936692e-01 1.73771575e-01 7.58790672e-01
-4.94336188e-01 4.22915109e-02 -4.20296043e-01 2.88040757e-01
8.72932076e-01 -2.42628530e-01 3.13247323e-01 -1.00681973e+00
-2.35297352e-01 -1.62648991e-01 -3.04137051e-01 -9.97317553e-01
1.85438097e-01 -6.68493867e-01 8.54286671e-01 -1.80534923e+00
3.35433692e-01 -6.57777846e-01 -5.16815424e-01 5.74849606e-01
-7.08942354e-01 6.44504368e-01 -1.88792408e-01 1.92544356e-01
-9.07389462e-01 2.95302093e-01 1.70410669e+00 -4.03432161e-01
-3.64557728e-02 1.15874983e-01 -5.52817583e-01 1.04876852e+00
6.99259460e-01 -7.08595216e-01 -4.31839347e-01 -5.90719402e-01
-1.28039345e-01 -1.55535460e-01 2.27023944e-01 -9.55100238e-01
1.77728876e-01 -3.54672931e-02 4.06867415e-01 -6.58905089e-01
1.66277021e-01 -8.35082293e-01 -4.89653349e-01 4.12709713e-01
-5.18765748e-01 -5.47431231e-01 -3.06877326e-02 8.29957008e-01
-3.42166305e-01 -4.77061629e-01 1.25300896e+00 -5.19710779e-01
-6.45618141e-01 3.82698268e-01 2.39297040e-02 3.33628327e-01
8.84460092e-01 -8.35536122e-02 -8.24468061e-02 -1.63350731e-01
-9.49098349e-01 4.94542331e-01 2.36312568e-01 8.72590020e-02
5.02341628e-01 -1.03140402e+00 -6.58377707e-01 7.61007145e-02
4.44810763e-02 6.32259011e-01 2.85075486e-01 8.00356984e-01
-4.81577396e-01 1.37198465e-02 2.88659602e-01 -1.00544202e+00
-1.11708963e+00 5.60894728e-01 4.82530236e-01 -4.39444005e-01
-7.13480413e-01 1.08493292e+00 1.79665908e-01 -6.19529366e-01
6.61671758e-01 -5.54232657e-01 1.60082668e-01 -2.24897087e-01
4.71823603e-01 2.59304076e-01 1.94407225e-01 -5.69644928e-01
-1.40260234e-01 7.85498798e-01 -2.22465899e-02 -6.10699691e-02
1.05026710e+00 -3.97984207e-01 -7.25641176e-02 3.73444796e-01
1.04983771e+00 -4.59873855e-01 -1.62222850e+00 -6.10521972e-01
1.71763673e-01 -4.74869043e-01 3.76780063e-01 -1.02866924e+00
-1.45918572e+00 9.01957154e-01 9.56417382e-01 -1.16420597e-01
1.15610147e+00 1.05360756e-02 1.14173400e+00 2.76420504e-01
3.09675246e-01 -1.57925391e+00 3.00919056e-01 9.89851654e-02
3.29305738e-01 -1.98015094e+00 6.83860928e-02 -7.60156691e-01
-6.72082901e-01 5.88009179e-01 5.63354254e-01 2.45980751e-02
7.96396792e-01 5.51534891e-01 8.21353316e-01 -1.20589159e-01
-1.58881232e-01 -4.66686428e-01 3.52763206e-01 6.24419510e-01
4.07320291e-01 4.76147570e-02 -2.00735956e-01 5.83175421e-01
3.40671271e-01 3.12996320e-02 3.34743947e-01 9.80907738e-01
-5.08312821e-01 -1.11688769e+00 -5.65546095e-01 5.83141208e-01
-8.65213513e-01 1.93770733e-02 1.54896781e-01 4.55888450e-01
5.47798932e-01 1.20468378e+00 -3.25271815e-01 1.22121230e-01
2.97908545e-01 -9.90930870e-02 2.79883116e-01 -8.63945842e-01
-4.94780928e-01 5.69946945e-01 -2.76798666e-01 -5.74314296e-01
-6.33347809e-01 -4.11633074e-01 -1.66020024e+00 3.73325229e-01
-5.57936490e-01 -9.54992846e-02 6.78258479e-01 1.12371767e+00
-3.93668227e-02 8.97260606e-01 4.39905465e-01 -8.23948741e-01
-5.52289486e-01 -8.66492450e-01 -7.15847731e-01 7.95566499e-01
4.23663080e-01 -3.52954984e-01 -3.26761991e-01 3.62117618e-01] | [14.589777946472168, -2.1051273345947266] |
9551a6ca-7ed4-4d33-8a21-8c9cfe5befe2 | addressing-distribution-shift-in-online | null | null | https://openreview.net/forum?id=9hgEG-k57Zj | https://openreview.net/pdf?id=9hgEG-k57Zj | Addressing Distribution Shift in Online Reinforcement Learning with Offline Datasets | Recent progress in offline reinforcement learning (RL) has made it possible to train strong RL agents from previously-collected, static datasets. However, depending on the quality of the trained agents and the application being considered, it is often desirable to improve such offline RL agents with further online interaction. As it turns out, fine-tuning offline RL agents is a non-trivial challenge, due to distribution shift – the agent encounters out-of-distribution samples during online interaction, which may cause bootstrapping error in Q-learning and instability during fine-tuning. In order to address the issue, we present a simple yet effective framework, which incorporates a balanced replay scheme and an ensemble distillation scheme. First, we propose to keep separate offline and online replay buffers, and carefully balance the number of samples from each buffer during updates. By utilizing samples from a wider distribution, i.e., both online and offline samples, we stabilize the Q-learning. Next, we present an ensemble distillation scheme, where we train an ensemble of independent actor-critic agents, then distill the policies into a single policy. In turn, we improve the policy using the Q-ensemble during fine-tuning, which allows the policy updates to be more robust to error in each individual Q-function. We demonstrate the superiority of our method on MuJoCo datasets from the recently proposed D4RL benchmark suite.
| ['Jinwoo Shin', 'Pieter Abbeel', 'Kimin Lee', 'Younggyo Seo', 'SeungHyun Lee'] | 2021-01-01 | null | null | null | null | ['d4rl'] | ['robots'] | [-3.37104201e-01 -2.15459913e-01 -2.26094186e-01 -8.82865787e-02
-9.78556871e-01 -8.39914620e-01 5.10307789e-01 2.07994506e-01
-7.51019657e-01 1.18280911e+00 6.13870993e-02 -2.99511015e-01
4.84286342e-03 -6.88033581e-01 -8.56445134e-01 -9.87632573e-01
4.02057320e-02 7.14697540e-01 2.76136011e-01 -3.26171398e-01
9.31107625e-02 4.19223517e-01 -1.51010513e+00 -4.26436923e-02
9.80804026e-01 7.20548987e-01 2.48831008e-02 7.96283603e-01
3.23873274e-02 9.22788501e-01 -1.03908241e+00 -1.18160397e-01
4.49502468e-01 -6.10364318e-01 -4.05889690e-01 9.23787244e-03
-6.51786923e-02 -7.40560353e-01 -1.72096848e-01 8.56260359e-01
7.75325954e-01 4.98380095e-01 2.68356115e-01 -1.16016805e+00
-2.08674565e-01 8.76420200e-01 -5.19286692e-01 7.32589588e-02
1.41105920e-01 7.94461787e-01 8.69574070e-01 -4.02512789e-01
3.81703645e-01 1.28410137e+00 3.83823574e-01 6.80509984e-01
-1.38409817e+00 -8.67976725e-01 4.01106566e-01 -1.11503080e-01
-7.49975801e-01 -4.42425787e-01 6.13253653e-01 -8.92770737e-02
8.93424988e-01 -1.80630207e-01 8.18930328e-01 1.25000370e+00
1.62566975e-01 1.05580389e+00 1.19397247e+00 -2.21201345e-01
7.96957314e-01 7.38326460e-02 -3.73387635e-01 3.67922217e-01
2.08645388e-02 6.34867191e-01 -2.91449606e-01 -4.09016401e-01
7.58148968e-01 -1.07252650e-01 3.00528314e-02 -5.45998454e-01
-9.51553941e-01 9.63823259e-01 3.23675305e-01 -2.35497206e-02
-5.39642751e-01 3.93876463e-01 6.41279519e-01 5.24790585e-01
4.28781569e-01 6.34256005e-01 -6.42751336e-01 -6.09039605e-01
-7.40022242e-01 6.60935044e-01 9.17522788e-01 4.08586264e-01
7.47416556e-01 8.04997832e-02 -3.87378871e-01 9.16310966e-01
5.81299104e-02 4.72983927e-01 6.76847517e-01 -1.07408035e+00
4.97236073e-01 3.13584954e-01 4.57604975e-01 -4.87877935e-01
-2.80795038e-01 -4.59800422e-01 -5.39875925e-01 5.88868022e-01
6.59855664e-01 -5.61141849e-01 -6.75115883e-01 1.91486907e+00
5.94200790e-01 6.22469261e-02 9.84332338e-02 8.47582817e-01
-1.94841251e-02 4.56088036e-01 9.22296569e-02 -3.81343722e-01
6.98077321e-01 -1.16652107e+00 -5.43494940e-01 -8.67924932e-03
5.34430563e-01 -3.45548600e-01 1.17208087e+00 5.54854870e-01
-1.22149873e+00 -5.10213912e-01 -1.06353486e+00 4.71606433e-01
-5.86541779e-02 -8.82328600e-02 4.26002592e-01 3.68162602e-01
-8.17037463e-01 9.20779288e-01 -1.02780461e+00 7.14566559e-02
2.80053794e-01 5.32982647e-01 5.85513748e-02 1.84617773e-01
-1.12378192e+00 7.62916565e-01 5.25799751e-01 -9.74628627e-02
-1.22164762e+00 -5.67740083e-01 -5.08450270e-01 -4.73528169e-02
8.04203153e-01 -5.82074642e-01 1.76210117e+00 -1.38886654e+00
-2.32033563e+00 1.47623241e-01 1.59365311e-01 -5.99498868e-01
9.39355493e-01 -1.81812331e-01 -8.57038572e-02 -1.31137297e-01
-3.02744508e-01 3.84623200e-01 1.07660115e+00 -1.23613989e+00
-7.50326335e-01 -3.04601312e-01 3.49442452e-01 4.88079846e-01
-8.91832039e-02 -2.71092951e-01 -1.62313417e-01 -5.45489132e-01
-5.45780540e-01 -1.02151263e+00 -4.03279930e-01 -4.03028309e-01
8.13091174e-03 -3.62441063e-01 4.66040909e-01 -2.51374990e-01
1.21841168e+00 -2.14054441e+00 1.69544339e-01 2.92769641e-01
1.75734520e-01 3.54480237e-01 -3.23849708e-01 4.49014068e-01
1.32526278e-01 -1.66763172e-01 -1.52427807e-01 -5.24779737e-01
1.23692244e-01 5.28629839e-01 -4.07152236e-01 3.77615154e-01
9.20360312e-02 8.23304057e-01 -1.32665229e+00 -8.69542137e-02
1.55757368e-01 2.04565786e-02 -8.79692197e-01 5.07470191e-01
-7.46162236e-01 8.55544686e-01 -5.86183250e-01 3.01220655e-01
4.97312307e-01 -1.64273202e-01 2.93542534e-01 3.17055434e-01
-6.63751587e-02 2.13767201e-01 -1.23217404e+00 1.54084659e+00
-6.17583990e-01 6.83877543e-02 -3.89817427e-03 -9.20638680e-01
8.38711679e-01 1.49400100e-01 6.52942419e-01 -8.79243672e-01
1.41802222e-01 3.02665621e-01 8.48881602e-02 -2.32090458e-01
5.14868081e-01 -1.08683124e-01 3.35208662e-02 9.27373946e-01
7.07229376e-02 -1.15811020e-01 3.51414740e-01 -9.35682431e-02
1.14391720e+00 4.36342359e-01 3.13884228e-01 1.95188016e-01
4.17360336e-01 -1.14959180e-01 5.17901123e-01 1.00834715e+00
-4.01623607e-01 1.10193014e-01 6.60063088e-01 -3.74790549e-01
-1.13519061e+00 -9.51572776e-01 2.92881668e-01 1.55556273e+00
-5.81167266e-02 -2.17532843e-01 -6.91389561e-01 -1.13762355e+00
3.48877758e-01 7.49915421e-01 -5.31006098e-01 -3.90461087e-01
-8.66800070e-01 -8.22627246e-01 4.80124921e-01 4.45409954e-01
3.75902623e-01 -1.33288860e+00 -8.35607350e-01 6.59491897e-01
1.75009727e-01 -6.92105472e-01 -4.36012447e-01 3.63300860e-01
-8.73773217e-01 -8.90575230e-01 -6.92413449e-01 -1.43111587e-01
4.78458881e-01 -6.96867108e-02 1.30093956e+00 7.54528567e-02
2.15098888e-01 2.37523884e-01 -4.11072403e-01 -2.03048244e-01
-7.24683881e-01 2.26695061e-01 1.44542173e-01 -1.83351234e-01
3.28865424e-02 -5.13540089e-01 -6.91314936e-01 2.92967021e-01
-8.44692409e-01 -2.99980819e-01 4.37376350e-01 1.18685985e+00
4.72041816e-01 -3.09045520e-02 9.05867219e-01 -7.42662251e-01
1.06728578e+00 -5.32164872e-01 -9.68192756e-01 2.07093999e-01
-5.74633181e-01 4.73601699e-01 1.17813957e+00 -7.67254651e-01
-9.20834959e-01 -7.71860927e-02 -2.53510684e-01 -5.36920369e-01
1.17454633e-01 2.37038359e-01 1.82694942e-01 1.01425447e-01
7.65312195e-01 1.77238315e-01 3.48063827e-01 -2.12233618e-01
6.01245105e-01 5.15618980e-01 2.26562366e-01 -9.52020466e-01
5.44224858e-01 1.68126777e-01 -4.76283401e-01 -2.50988692e-01
-7.40969598e-01 -1.50116056e-01 -8.47625434e-02 -2.63047934e-01
3.17626953e-01 -6.34413898e-01 -1.09364712e+00 5.45670450e-01
-6.69724107e-01 -1.09816468e+00 -6.71856821e-01 3.63721460e-01
-8.24659228e-01 4.13827561e-02 -5.07799864e-01 -1.01247883e+00
-2.79587924e-01 -1.41836214e+00 7.62323439e-01 4.37218428e-01
6.82347687e-03 -1.00697470e+00 7.37224758e-01 -9.32382420e-02
4.86057311e-01 7.08956867e-02 5.92433214e-01 -7.62983918e-01
-2.81595737e-01 9.05004367e-02 1.06780492e-01 4.05296177e-01
1.75882533e-01 2.59120204e-02 -8.40884566e-01 -8.76373172e-01
-1.56330019e-01 -9.59207296e-01 7.71322012e-01 1.80501685e-01
1.22408915e+00 -3.37896287e-01 -1.66363567e-02 4.18045461e-01
1.17059672e+00 2.58976668e-01 2.65618265e-01 5.17235637e-01
2.06079438e-01 1.96223110e-01 7.45028496e-01 9.24577475e-01
3.31973404e-01 5.20862937e-01 3.67397368e-01 1.77827597e-01
2.40686521e-01 -4.85629976e-01 7.95926869e-01 5.51593423e-01
7.53435194e-02 -2.03463718e-01 -5.16021013e-01 2.03496337e-01
-2.08627534e+00 -8.70838284e-01 9.00013924e-01 2.52196074e+00
1.22600567e+00 2.42624551e-01 6.89154923e-01 -2.08591998e-01
3.13073903e-01 1.02656208e-01 -1.20452368e+00 -3.20703834e-01
2.02822760e-01 3.82949710e-01 5.97058654e-01 5.52306890e-01
-7.65075147e-01 9.83830154e-01 6.15483761e+00 8.52978587e-01
-1.31776321e+00 3.57563160e-02 5.38357437e-01 -4.09401059e-01
-2.53121138e-01 -6.94245175e-02 -6.65922761e-01 6.58172607e-01
9.72467124e-01 -1.26041574e-02 1.12980580e+00 7.74067044e-01
4.03723627e-01 -3.79710674e-01 -1.10279691e+00 7.12543607e-01
-3.04411232e-01 -1.02128756e+00 -3.53947729e-01 1.67677663e-02
8.55377436e-01 2.49582395e-01 -1.82099342e-02 9.46741879e-01
1.06800520e+00 -8.55502963e-01 7.22685277e-01 2.89689153e-01
4.37046349e-01 -1.06987679e+00 6.26294792e-01 7.59311318e-01
-8.04847062e-01 -4.87897217e-01 -3.00931871e-01 1.50837312e-02
1.03000682e-02 4.49710310e-01 -7.15382457e-01 3.99324119e-01
5.34243822e-01 4.77036983e-01 -2.84069985e-01 8.86612058e-01
-2.42069334e-01 5.42060256e-01 -4.25670803e-01 -1.96335062e-01
3.20652843e-01 -2.95078367e-01 3.41241390e-01 7.36809015e-01
8.29824582e-02 -6.01370223e-02 4.75072056e-01 6.47547960e-01
-8.57205614e-02 -8.40026885e-02 -2.46612385e-01 -1.24960057e-01
5.48814714e-01 1.17828476e+00 -4.29114908e-01 -3.81253034e-01
-1.38935819e-01 7.90245533e-01 7.91734457e-01 4.45094138e-01
-9.85985696e-01 -2.10529733e-02 5.68146586e-01 -3.18175882e-01
3.30517948e-01 -2.94366121e-01 2.01298580e-01 -1.24235642e+00
-1.11144699e-01 -1.39209330e+00 4.43194300e-01 -1.67047292e-01
-1.42289269e+00 3.54394615e-01 -1.13284573e-01 -1.15655470e+00
-7.09645927e-01 -1.83171943e-01 -4.42239583e-01 5.56289196e-01
-1.58760941e+00 -4.52573597e-01 -6.45217374e-02 4.90766466e-01
5.10586023e-01 -1.08175449e-01 5.86026490e-01 1.23120315e-01
-7.05899477e-01 8.36237073e-01 5.43808579e-01 -2.27709934e-01
9.73299503e-01 -1.36989343e+00 1.96770459e-01 3.26974183e-01
-2.17863545e-02 3.54341835e-01 6.93801761e-01 -5.83941996e-01
-1.39509988e+00 -9.38882709e-01 -4.30953652e-02 -2.36346513e-01
8.25683296e-01 -2.58313715e-01 -9.36236560e-01 4.93600100e-01
1.06652446e-01 1.57926351e-01 2.98303634e-01 1.03390245e-02
-8.57270062e-02 -2.93365389e-01 -1.15779459e+00 6.97324336e-01
6.49561346e-01 -2.72827953e-01 -3.37734401e-01 1.84338972e-01
5.83517313e-01 -7.16125309e-01 -7.96162069e-01 1.72884136e-01
5.70638478e-01 -1.03589976e+00 6.83557868e-01 -6.22623861e-01
1.31292120e-01 -2.36965403e-01 2.01006413e-01 -1.91003501e+00
3.03996801e-02 -9.35714960e-01 -4.85363543e-01 9.88955855e-01
2.48154759e-01 -9.42843258e-01 7.49310434e-01 3.78450453e-01
2.58683920e-01 -9.27276492e-01 -7.18452096e-01 -7.49684870e-01
3.07210118e-01 -2.90663749e-01 7.68193722e-01 4.58831578e-01
-1.77812595e-02 1.79383442e-01 -4.78474945e-01 -2.25783974e-01
4.71526980e-01 1.14328407e-01 1.18767047e+00 -7.23329782e-01
-9.55793202e-01 -5.47328532e-01 4.39555764e-01 -1.36721647e+00
2.45808944e-01 -4.71347153e-01 3.81919444e-01 -9.63279724e-01
8.45321417e-02 -7.80346692e-01 -5.26272714e-01 5.02254963e-01
-4.32983130e-01 -2.67052740e-01 4.07256693e-01 2.34316006e-01
-8.24840248e-01 1.02129543e+00 1.47529161e+00 1.55529886e-01
-7.43341386e-01 2.64573902e-01 -3.11111182e-01 3.86729568e-01
9.58227634e-01 -5.19795954e-01 -5.14523387e-01 -1.97012946e-01
2.60269284e-01 4.05806780e-01 1.34498671e-01 -7.35793650e-01
6.48334473e-02 -3.88940960e-01 2.70568639e-01 -3.29710543e-01
1.04536347e-01 -6.19810104e-01 -2.73760676e-01 5.57535827e-01
-4.82146353e-01 2.39028960e-01 1.92671850e-01 6.39516592e-01
-8.28386564e-03 -2.04249859e-01 8.93175423e-01 -1.40927732e-01
-2.55188197e-01 3.31428796e-01 -2.86974072e-01 2.82082289e-01
1.07203174e+00 1.42302096e-01 -1.87587589e-01 -3.74928266e-01
-5.59181750e-01 6.84054852e-01 5.80188870e-01 2.92373002e-01
2.46657789e-01 -1.11868215e+00 -5.23398399e-01 3.25813651e-01
-6.94819689e-02 9.89949927e-02 1.19308203e-01 8.32598150e-01
-1.48855865e-01 -1.39216468e-01 -2.16223687e-01 -4.10896599e-01
-6.26985967e-01 5.01154304e-01 5.38156092e-01 -7.92222202e-01
-5.54968596e-01 4.92212892e-01 -1.63874730e-01 -6.51100695e-01
4.39446718e-01 -3.81889790e-01 -1.36943445e-01 5.50899655e-02
4.86807883e-01 3.89044583e-01 -2.12389436e-02 1.04478024e-01
-3.06535549e-02 1.86186627e-01 -1.73947945e-01 -4.30654466e-01
1.15754771e+00 2.23223884e-02 2.42383450e-01 6.26002073e-01
7.51673460e-01 2.70919278e-02 -1.95073712e+00 -2.07381681e-01
-1.06616452e-01 -3.19361866e-01 -2.28828460e-01 -9.13940668e-01
-1.02490389e+00 4.88332003e-01 4.64353859e-01 3.37075174e-01
1.00984144e+00 -3.25200498e-01 8.50831509e-01 3.98729891e-01
4.79664892e-01 -1.45585215e+00 3.35080504e-01 6.94623113e-01
6.59864783e-01 -1.30395460e+00 -7.85315633e-02 6.92965627e-01
-7.41750479e-01 1.08139312e+00 7.40928650e-01 -5.03342390e-01
2.49743283e-01 2.62488127e-01 8.86773765e-02 2.80900389e-01
-1.12963200e+00 2.28748135e-02 -2.45387211e-01 1.89032570e-01
4.39007990e-02 1.51923597e-01 -9.18093398e-02 3.60342264e-01
-1.08819589e-01 1.83115587e-01 2.51454681e-01 1.03593171e+00
-4.08852905e-01 -1.38452458e+00 -3.77157688e-01 3.37156177e-01
-3.17472368e-01 3.51603448e-01 -8.88479576e-02 6.75728083e-01
-1.07971504e-01 8.29751313e-01 6.44934773e-02 -1.55935213e-01
3.10855478e-01 -1.84468895e-01 5.73376060e-01 -3.90020341e-01
-1.03296793e+00 1.92193761e-01 -6.83349743e-02 -8.50357771e-01
-1.73899412e-01 -5.84843636e-01 -1.39271808e+00 -3.79616022e-01
-2.19511971e-01 3.10811311e-01 4.18386638e-01 1.06950986e+00
2.83342540e-01 6.19986534e-01 1.00823355e+00 -9.86222327e-01
-1.51088250e+00 -8.96042585e-01 -3.99481297e-01 2.84291983e-01
6.82739139e-01 -7.70796239e-01 -5.02759218e-01 -5.37462771e-01] | [4.060313701629639, 2.1995160579681396] |
da3175c4-1eab-416b-8fa2-ac89e606abf1 | an-empirical-comparison-of-deep-neural | 2005.01194 | null | https://arxiv.org/abs/2005.01194v1 | https://arxiv.org/pdf/2005.01194v1.pdf | An empirical comparison of deep-neural-network architectures for next activity prediction using context-enriched process event logs | Researchers have proposed a variety of predictive business process monitoring (PBPM) techniques aiming to predict future process behaviour during the process execution. Especially, techniques for the next activity prediction anticipate great potential in improving operational business processes. To gain more accurate predictions, a plethora of these techniques rely on deep neural networks (DNNs) and consider information about the context, in which the process is running. However, an in-depth comparison of such techniques is missing in the PBPM literature, which prevents researchers and practitioners from selecting the best solution for a given event log. To remedy this problem, we empirically evaluate the predictive quality of three promising DNN architectures, combined with five proven encoding techniques and based on five context-enriched real-life event logs. We provide four findings that can support researchers and practitioners in designing novel PBPM techniques for predicting the next activities. | ['B. Eskofier', 'J. Brunk', 'A. Nguyen', 'K. Revoredo', 'J. Becker', 'S. Zilker', 'S. Weinzierl', 'M. Matzner'] | 2020-05-03 | null | null | null | null | ['activity-prediction', 'activity-prediction'] | ['computer-vision', 'time-series'] | [ 5.73226511e-01 1.25955820e-01 -4.52302784e-01 -3.63260061e-01
-1.20402761e-02 -1.50432229e-01 8.08031440e-01 6.85018539e-01
-4.53777499e-02 4.43140209e-01 5.24115205e-01 -5.09129345e-01
-5.69744647e-01 -1.11290109e+00 -3.02941829e-01 -2.03372180e-01
-2.72969007e-01 6.42226219e-01 8.23762640e-03 2.55512297e-01
5.12921751e-01 8.32356513e-01 -1.43134284e+00 7.12954283e-01
2.92923450e-01 1.34709954e+00 1.03624091e-01 5.19703329e-01
-4.70873475e-01 1.68132913e+00 -4.71485466e-01 -5.17259955e-01
1.66617185e-01 -4.13898230e-01 -8.10545802e-01 -1.55923679e-01
-2.45969966e-01 -2.49303788e-01 -5.73686957e-01 6.00803375e-01
4.83240224e-02 2.42568925e-01 2.20697746e-01 -1.12830472e+00
-6.17627740e-01 9.07164752e-01 -1.64127350e-01 7.39524364e-01
4.47104752e-01 4.71348047e-01 1.17294717e+00 -3.52557331e-01
5.73664427e-01 1.04011416e+00 8.39058757e-01 4.63535339e-01
-9.35417473e-01 -4.68250245e-01 2.59464502e-01 5.85873783e-01
-8.10556650e-01 -2.82941043e-01 8.10318410e-01 -1.77777842e-01
1.41761947e+00 1.13408267e-01 9.10285711e-01 1.59209955e+00
5.32812417e-01 1.11914062e+00 7.30642378e-01 -3.97599012e-01
4.36535269e-01 -4.33430314e-01 2.06161678e-01 1.20954439e-01
2.24540100e-01 1.65957019e-01 -8.76692295e-01 -3.09994118e-03
7.03775525e-01 6.36119246e-01 -4.60500047e-02 7.53450245e-02
-1.16685843e+00 3.27093214e-01 -3.74153703e-02 8.10295224e-01
-9.65430796e-01 3.19016874e-01 5.90151906e-01 2.81312615e-01
2.09956527e-01 8.80472362e-01 -5.94125032e-01 -1.01803517e+00
-9.16047215e-01 3.88875842e-01 1.44128156e+00 8.00947607e-01
4.94743794e-01 4.50679176e-02 -7.12133765e-01 1.98225260e-01
2.07869604e-01 -1.74991786e-01 5.58390379e-01 -1.11055076e+00
4.89370525e-01 7.81261802e-01 2.24417001e-01 -9.43948567e-01
-3.21408391e-01 -2.93112695e-01 -6.71227932e-01 -4.41885799e-01
4.05619055e-01 4.41316962e-02 -5.50469279e-01 1.05952275e+00
-5.09417772e-01 5.68553388e-01 -1.19375728e-01 3.38297099e-01
1.43888205e-01 8.86747301e-01 3.62851262e-01 -3.38779062e-01
1.35225725e+00 -1.07393229e+00 -1.06486464e+00 -6.53857768e-01
1.15122251e-01 -1.19286783e-01 7.42496848e-01 7.71713853e-01
-1.22797465e+00 -6.79488480e-01 -9.26884472e-01 3.63627762e-01
-5.72119094e-02 -3.63188386e-01 1.03983545e+00 4.53547388e-01
-7.52176464e-01 1.28688979e+00 -1.45603108e+00 -3.00544202e-01
6.17385089e-01 2.22206488e-01 -3.10602859e-02 -5.62754609e-02
-1.20605397e+00 9.74835277e-01 6.35824084e-01 3.25931072e-01
-1.13994730e+00 -5.82176805e-01 -3.34647208e-01 8.65229249e-01
5.24245501e-01 -5.64470530e-01 1.74777842e+00 -9.54427540e-01
-1.44623756e+00 1.41859084e-01 -3.81954700e-01 -1.09464252e+00
3.69056195e-01 -5.16924679e-01 -7.65745938e-01 6.80090860e-02
-4.73859191e-01 -1.69176124e-02 5.16899824e-01 -7.99012363e-01
-1.08462036e+00 -3.86920720e-01 -2.63868511e-01 4.73182760e-02
-3.41925263e-01 2.51311034e-01 -2.97171801e-01 -2.18046159e-01
-3.17550637e-02 -3.51094335e-01 -5.25170982e-01 -6.70742571e-01
-2.21620247e-01 -3.62551749e-01 5.59073150e-01 -7.82716990e-01
1.55798829e+00 -1.91403925e+00 -3.43006939e-01 9.59648490e-02
3.37287545e-01 2.06018388e-01 1.86324924e-01 9.29948390e-01
5.07897772e-02 3.00401121e-01 2.13583231e-01 -3.22293282e-01
2.65610397e-01 3.49577963e-01 -6.17064536e-01 -1.16221480e-01
3.91024292e-01 1.05267489e+00 -8.36672902e-01 -2.72134282e-02
3.27056259e-01 2.80537546e-01 -1.40597701e-01 5.01027882e-01
-4.01105493e-01 4.20953929e-01 -4.36555535e-01 7.08199024e-01
1.90278471e-01 -4.15323377e-01 4.93743926e-01 3.06909949e-01
-1.70249958e-03 7.83190072e-01 -8.31389129e-01 1.17536640e+00
-3.80025268e-01 7.73187518e-01 -4.39397305e-01 -8.81868720e-01
1.03640354e+00 5.79175055e-01 8.13626170e-01 -8.17605853e-01
-1.16882823e-01 -2.71749109e-01 2.00995542e-02 -4.28399146e-01
6.96280658e-01 -3.29867691e-01 2.94955820e-01 4.65311527e-01
-3.94517742e-02 7.47845531e-01 4.73097682e-01 -5.16148508e-01
1.85613728e+00 1.92524418e-01 5.90978920e-01 3.62336099e-01
3.86529505e-01 -2.88394988e-01 1.01118624e+00 1.05480707e+00
-6.66783929e-01 1.56558454e-01 9.32440162e-01 -9.42835808e-01
-8.89721990e-01 -9.51175392e-01 3.08292478e-01 1.08281231e+00
-1.50284722e-01 -7.26761937e-01 -3.56627822e-01 -5.07407963e-01
-1.03943035e-01 1.20358348e+00 -5.54163694e-01 -1.98542029e-01
-7.85472810e-01 -5.79521418e-01 5.85101426e-01 1.14678156e+00
4.30327833e-01 -1.51408637e+00 -7.44390428e-01 8.12130272e-01
-2.40847412e-02 -1.14580667e+00 3.09598893e-01 5.73981881e-01
-1.31467104e+00 -8.87966692e-01 2.19487756e-01 -3.12689543e-01
-9.09635052e-02 -2.14568779e-01 1.33430958e+00 -4.05815005e-01
3.13765228e-01 1.44947037e-01 -3.01513761e-01 -6.65722072e-01
-7.27596104e-01 8.40479806e-02 -2.71055818e-01 -1.24969482e-01
1.26417446e+00 -7.33733594e-01 -4.56163675e-01 8.42212811e-02
-5.53490698e-01 -1.20193753e-02 8.23832154e-01 5.54753602e-01
4.29261655e-01 5.33353448e-01 5.64290285e-01 -6.95284903e-01
9.71362770e-01 -5.39094090e-01 -2.45340690e-01 2.43064687e-01
-1.04505467e+00 -7.36558586e-02 6.90558553e-01 -2.92472571e-01
-1.46112120e+00 -3.66314560e-01 -2.87359506e-01 -3.14779341e-01
-7.43285954e-01 7.23326385e-01 -1.14505470e-01 8.73731852e-01
4.07338053e-01 4.55889851e-01 -3.19422603e-01 -3.70944262e-01
-4.35768589e-02 3.28540415e-01 3.61478150e-01 -4.93779361e-01
9.54405405e-03 4.14253056e-01 -5.69738261e-02 -2.56807894e-01
-7.64850140e-01 -5.87104917e-01 -2.87274450e-01 -2.27266923e-01
7.68123925e-01 -5.38367808e-01 -1.12476826e+00 1.82791620e-01
-1.15141249e+00 -2.41171420e-01 -6.66198313e-01 2.79165506e-01
-8.41953337e-01 1.79098308e-01 -1.23090911e+00 -1.03478122e+00
-3.34170759e-01 -7.90051401e-01 6.80264056e-01 2.71896511e-01
-8.03005159e-01 -1.03649652e+00 3.92619111e-02 4.10352945e-01
5.15827417e-01 -5.73087335e-02 8.56260240e-01 -1.34466743e+00
-8.74205649e-01 -5.65307677e-01 5.28281145e-02 3.82347673e-01
2.02564433e-01 -8.40720609e-02 -1.04891837e+00 1.11794420e-01
5.06546617e-01 4.88104671e-01 5.10972202e-01 3.24248463e-01
1.68649995e+00 -3.34502786e-01 -6.08325481e-01 4.08146530e-01
1.19805014e+00 8.29301417e-01 9.45751727e-01 6.26705468e-01
5.48073828e-01 6.39029980e-01 4.79607522e-01 6.85813785e-01
2.24445879e-01 4.05284874e-02 4.64595735e-01 7.39814997e-01
3.92645299e-01 -5.91385305e-01 4.53538537e-01 6.55467808e-01
-6.42832935e-01 -5.30765653e-01 -1.29575074e+00 4.33595598e-01
-1.91057074e+00 -1.64033699e+00 4.24780212e-02 1.79683316e+00
4.09171671e-01 5.92948437e-01 -2.97516823e-01 3.52907211e-01
5.66504419e-01 3.48004133e-01 -3.63733143e-01 -5.03269911e-01
1.75479487e-01 3.05137932e-01 4.25976962e-01 -1.98906437e-01
-1.09127367e+00 5.10209560e-01 6.69668579e+00 4.45305437e-01
-7.12690055e-01 2.82120071e-02 7.28480160e-01 -1.78988457e-01
-7.44422823e-02 5.35051376e-02 -9.85114455e-01 4.71807897e-01
1.96324694e+00 -2.05971450e-01 3.85984510e-01 1.04008162e+00
6.09465778e-01 1.37474120e-01 -1.73972392e+00 8.95703554e-01
-4.13234919e-01 -1.73891878e+00 -5.80082834e-02 2.29411080e-01
5.51892877e-01 -3.02003846e-02 -3.43265921e-01 6.31802320e-01
3.70248914e-01 -1.27992260e+00 4.78383631e-01 9.01249707e-01
-2.55545676e-01 -9.77896810e-01 8.32825780e-01 4.30330843e-01
-9.90561783e-01 -9.10271287e-01 -1.54560074e-01 -5.88521361e-01
4.28563565e-01 7.20954537e-01 -1.19579852e+00 5.99447429e-01
7.66807079e-01 9.25844133e-01 -3.19874614e-01 8.51892710e-01
-3.04455668e-01 1.09817874e+00 1.19023629e-01 -6.96320133e-03
2.69923091e-01 -1.12635508e-01 3.13772619e-01 1.28042483e+00
4.30696875e-01 -5.59108675e-01 -3.40408646e-02 1.02762711e+00
-2.54951771e-02 -2.11592898e-01 -5.67941189e-01 -6.15502596e-01
4.60372746e-01 9.16665494e-01 -7.66535580e-01 -2.99722970e-01
-5.03871560e-01 8.09113622e-01 1.54959634e-01 2.82839388e-01
-5.99514365e-01 1.63764387e-01 8.82549822e-01 3.34977835e-01
2.64852941e-01 -2.15201139e-01 -7.44644403e-01 -8.30591261e-01
1.22704916e-01 -9.04056787e-01 4.49810565e-01 -6.33080006e-01
-1.56852007e+00 3.98624569e-01 -2.69875407e-01 -8.50371361e-01
-4.50176567e-01 -6.27817094e-01 -7.26006925e-01 7.52389729e-01
-1.27832711e+00 -9.92014825e-01 -2.77794003e-01 -3.67870927e-02
5.37195265e-01 -1.56220600e-01 7.33765423e-01 1.73386350e-01
-6.60322130e-01 -2.71848381e-01 -7.42714703e-02 1.90099608e-02
4.62273121e-01 -1.41519177e+00 7.95386672e-01 9.01907086e-01
7.15757757e-02 7.28770912e-01 6.35724127e-01 -8.16875577e-01
-1.58186400e+00 -1.03145039e+00 1.28201556e+00 -7.02294350e-01
9.47010219e-01 4.94289398e-02 -1.20174384e+00 1.32034051e+00
3.03082734e-01 -4.18465108e-01 8.69241059e-01 4.37029809e-01
2.42038041e-01 -3.87535304e-01 -7.89388299e-01 6.78757668e-01
1.10868478e+00 -4.94903207e-01 -9.51590300e-01 -1.62080184e-01
4.56067473e-01 -6.71241730e-02 -1.26440179e+00 2.98638046e-01
2.78421193e-01 -1.08529627e+00 8.42188299e-01 -7.36445963e-01
8.30590308e-01 5.23749217e-02 4.13085297e-02 -1.01150715e+00
-4.96279508e-01 -7.73805141e-01 -1.09238946e+00 1.05430114e+00
-3.77708822e-02 -3.54959399e-01 1.11776531e+00 7.44559765e-01
-2.46571228e-01 -8.62594187e-01 -5.87069035e-01 -5.48225522e-01
-3.65002304e-01 -1.14363015e+00 1.07505643e+00 8.95415545e-01
1.42605856e-01 1.70687705e-01 -3.41315210e-01 2.34750621e-02
1.20416209e-01 -5.78634292e-02 3.89951050e-01 -1.55183327e+00
-4.02234733e-01 -7.99465120e-01 -3.71316582e-01 -1.12058592e+00
3.53654660e-02 -6.18733227e-01 -4.04628702e-02 -1.96593070e+00
6.69766888e-02 5.39256632e-02 -8.18232656e-01 4.10073429e-01
3.14310119e-02 -6.74347699e-01 1.07626215e-01 2.37579748e-01
-6.40722692e-01 5.05574048e-01 1.14448094e+00 -4.62433361e-02
-2.66221225e-01 3.02362531e-01 -9.42949176e-01 8.04877818e-01
8.74453306e-01 -2.69041955e-01 -4.20029223e-01 -5.77383079e-02
5.91136932e-01 4.60751474e-01 2.93969303e-01 -1.33404911e+00
4.65516120e-01 -4.38438326e-01 5.74153125e-01 -7.07998931e-01
1.75432548e-01 -8.78230810e-01 3.82786840e-01 5.64987838e-01
-6.37096763e-01 2.85088331e-01 3.76144126e-02 9.05104280e-01
-6.31842136e-01 -3.97042692e-01 2.03744501e-01 -3.31073076e-01
-1.23314798e+00 3.81169945e-01 -8.74552846e-01 -4.06767726e-01
9.80093718e-01 -4.97385293e-01 -9.03557763e-02 -3.60052139e-01
-9.09214795e-01 1.30974380e-02 4.51626852e-02 4.07094479e-01
4.85287786e-01 -1.03508604e+00 -2.55478650e-01 4.22723889e-02
6.58213645e-02 6.63696229e-02 2.66548455e-01 6.42321348e-01
-6.49534941e-01 6.36682928e-01 -4.08410579e-01 -2.03892529e-01
-5.64009070e-01 7.80680060e-01 3.66394937e-01 -1.05326784e+00
-7.46945798e-01 3.00701529e-01 -3.17752361e-01 -2.83059143e-02
2.31556311e-01 -6.75553858e-01 -2.37617180e-01 -8.65095016e-03
8.65962923e-01 7.43731380e-01 2.24478021e-01 2.29875118e-01
-9.02610868e-02 -7.08875418e-01 -4.63563539e-02 4.22791094e-02
1.57906663e+00 2.59797014e-02 -2.53776759e-01 8.17671061e-01
3.01284432e-01 -3.14252019e-01 -1.56181729e+00 -1.39624864e-01
1.08907306e+00 -3.75386864e-01 -5.66673763e-02 -9.18980420e-01
-7.38228858e-01 9.83826041e-01 2.56528825e-01 5.60482621e-01
1.15180218e+00 -3.06178212e-01 7.93121517e-01 6.48051620e-01
2.62428850e-01 -1.28242004e+00 9.80569124e-02 8.88337612e-01
4.06554997e-01 -7.91327477e-01 -2.45686665e-01 -8.42438266e-02
-6.65059984e-01 1.27808917e+00 6.32456005e-01 9.15455595e-02
6.27986491e-01 3.87404591e-01 -4.12199289e-01 -2.80280381e-01
-1.23996675e+00 2.99611360e-01 -1.38960287e-01 5.91760814e-01
6.42947674e-01 2.33462885e-01 1.50332108e-01 1.06519330e+00
2.24709585e-02 5.22149384e-01 5.30767560e-01 1.21399629e+00
-3.74921501e-01 -1.18101954e+00 -2.37029672e-01 9.11973000e-01
-8.72752905e-01 -5.56958057e-02 -7.12416098e-02 5.55064559e-01
9.06589534e-03 9.05852556e-01 5.03051400e-01 -3.34411323e-01
6.06244504e-01 6.69504762e-01 3.52594197e-01 -5.36183894e-01
-9.87743199e-01 -2.77860731e-01 4.03858662e-01 -7.27618515e-01
-2.89379656e-01 -9.13162827e-01 -9.83803093e-01 -7.20611453e-01
3.40644360e-01 -2.15694621e-01 3.14828396e-01 1.04817605e+00
3.96232039e-01 1.13425159e+00 8.49420875e-02 -5.42243898e-01
-6.26350462e-01 -1.21909964e+00 -4.78198528e-01 4.54854935e-01
-2.39515752e-01 -3.32915634e-01 1.82904750e-01 5.60103022e-02] | [8.586065292358398, 5.955399036407471] |
2d16f461-7ce1-40af-8b9a-413e65051ce7 | regformer-an-efficient-projection-aware | 2303.12384 | null | https://arxiv.org/abs/2303.12384v1 | https://arxiv.org/pdf/2303.12384v1.pdf | RegFormer: An Efficient Projection-Aware Transformer Network for Large-Scale Point Cloud Registration | Although point cloud registration has achieved remarkable advances in object-level and indoor scenes, large-scale registration methods are rarely explored. Challenges mainly arise from the huge point number, complex distribution, and outliers of outdoor LiDAR scans. In addition, most existing registration works generally adopt a two-stage paradigm: They first find correspondences by extracting discriminative local features, and then leverage estimators (eg. RANSAC) to filter outliers, which are highly dependent on well-designed descriptors and post-processing choices. To address these problems, we propose an end-to-end transformer network (RegFormer) for large-scale point cloud alignment without any further post-processing. Specifically, a projection-aware hierarchical transformer is proposed to capture long-range dependencies and filter outliers by extracting point features globally. Our transformer has linear complexity, which guarantees high efficiency even for large-scale scenes. Furthermore, to effectively reduce mismatches, a bijective association transformer is designed for regressing the initial transformation. Extensive experiments on KITTI and NuScenes datasets demonstrate that our RegFormer achieves state-of-the-art performance in terms of both accuracy and efficiency. | ['Hesheng Wang', 'Marc Pollefeys', 'Chaokang Jiang', 'Zhe Liu', 'Guangming Wang', 'Jiuming Liu'] | 2023-03-22 | null | null | null | null | ['point-cloud-registration'] | ['computer-vision'] | [-2.83832792e-02 -6.78398669e-01 2.11799070e-02 -5.97725868e-01
-1.04717946e+00 -4.11050290e-01 4.99016970e-01 1.19649865e-01
-4.47541177e-01 2.44839728e-01 5.31809330e-02 1.28495589e-01
-2.32973769e-01 -7.84214914e-01 -7.65652597e-01 -5.78787148e-01
2.18147919e-01 5.83185673e-01 4.21932876e-01 -1.53917775e-01
2.59560615e-01 7.35443592e-01 -1.37916410e+00 -3.57983202e-01
1.23851633e+00 1.01453555e+00 1.77018255e-01 -1.20788179e-02
-1.00737691e-01 7.42353275e-02 -5.28897569e-02 -3.73521626e-01
5.61548054e-01 1.33269131e-01 -1.57426313e-01 7.29293227e-02
8.33197057e-01 -2.32382730e-01 -3.91218305e-01 1.10368335e+00
6.42193198e-01 1.42484009e-01 4.68555719e-01 -1.12947452e+00
-3.68995607e-01 2.76601072e-02 -8.15705657e-01 -1.82910353e-01
1.56417534e-01 2.23763332e-01 9.94993269e-01 -1.44721401e+00
2.57814974e-01 1.13323331e+00 1.03354931e+00 -1.41823858e-01
-1.41258109e+00 -1.02850294e+00 2.23652959e-01 1.35607317e-01
-1.75207305e+00 -4.25950229e-01 9.12434101e-01 -3.87858272e-01
7.21397877e-01 2.84306765e-01 6.30607367e-01 6.43828213e-01
1.36137024e-01 2.55326062e-01 7.86871672e-01 1.18548371e-01
1.24553584e-01 -4.03369039e-01 -1.70632631e-01 3.16530794e-01
3.31600279e-01 4.02274728e-02 -6.46257281e-01 -2.93931633e-01
9.63309169e-01 5.88549316e-01 -1.23279236e-01 -8.47495794e-01
-1.42233837e+00 6.85126603e-01 8.84815633e-01 -1.02781281e-01
-4.74067748e-01 1.13508791e-01 2.18664333e-01 1.63921341e-01
5.06341219e-01 2.02040687e-01 -3.70981485e-01 -9.75452270e-03
-1.09607649e+00 2.24759817e-01 3.04677129e-01 1.31393588e+00
1.20813823e+00 -3.14605057e-01 1.01741761e-01 9.70080614e-01
5.09540856e-01 8.97272885e-01 8.72698352e-02 -6.99083269e-01
8.01447332e-01 7.94735432e-01 -6.02609552e-02 -1.31826115e+00
-3.53888661e-01 -4.29293990e-01 -1.06963885e+00 -4.04807553e-02
1.22194007e-01 3.50384742e-01 -8.38745654e-01 1.38916421e+00
5.00488520e-01 4.63373750e-01 -3.95898789e-01 1.06110752e+00
5.61040699e-01 3.96699309e-01 -1.98254004e-01 -1.00850828e-01
1.21676683e+00 -6.61639571e-01 -3.48672360e-01 -4.29723412e-01
3.38091344e-01 -1.00519300e+00 1.00375795e+00 3.85702075e-03
-8.83258820e-01 -4.78270829e-01 -9.52960253e-01 -4.24084276e-01
3.83711010e-02 -1.18706599e-02 3.59061122e-01 1.54528126e-01
-5.53682327e-01 5.17343819e-01 -1.16550553e+00 -2.90198088e-01
5.69032311e-01 5.80583274e-01 -3.73825908e-01 -3.99844348e-01
-5.69049716e-01 4.96123999e-01 -9.34389085e-02 4.88564670e-01
-3.24207693e-01 -8.33897650e-01 -9.13084745e-01 -6.54335273e-03
4.30600137e-01 -7.47207522e-01 7.81876147e-01 -2.69487709e-01
-1.28079176e+00 5.73922157e-01 -4.47955161e-01 -1.13634951e-01
6.56486988e-01 -5.40966392e-01 -1.41379684e-01 -1.77308023e-01
4.37361866e-01 4.14771855e-01 6.49046600e-01 -1.16226602e+00
-5.96584380e-01 -7.66531408e-01 -4.76413459e-01 1.96420819e-01
-2.36226633e-01 -1.68891490e-01 -8.96824062e-01 -7.12266266e-01
1.04018545e+00 -1.06365716e+00 -5.98869205e-01 2.55199641e-01
-3.22253436e-01 -3.01661696e-02 7.62339354e-01 -4.17752445e-01
9.19163346e-01 -2.41521311e+00 -1.13041356e-01 6.10322535e-01
2.86632717e-01 -1.26423433e-01 -2.63302445e-01 3.98785442e-01
1.02176502e-01 -2.65753388e-01 -3.13975126e-01 -6.29605532e-01
3.99044976e-02 2.51484156e-01 -5.79802513e-01 8.66697729e-01
1.91142321e-01 6.99901104e-01 -7.77708769e-01 -4.41796690e-01
5.14401495e-01 6.34870350e-01 -7.73445666e-01 1.38759896e-01
8.60338286e-02 5.37220716e-01 -5.79038382e-01 7.66502798e-01
1.19369256e+00 -1.55877739e-01 -2.86360949e-01 -6.17232263e-01
-3.16620618e-01 3.57490838e-01 -1.33332598e+00 2.02569318e+00
-4.38964456e-01 2.50282109e-01 -1.09781750e-01 -5.84567010e-01
1.17896831e+00 -1.65809691e-01 8.00811648e-01 -5.78068912e-01
2.57012285e-02 5.74505568e-01 -1.52675390e-01 8.72694701e-02
5.45151293e-01 1.13072783e-01 8.37624911e-03 -8.77373591e-02
-3.27228189e-01 -4.58833814e-01 -1.93987042e-01 -7.37949163e-02
1.10648882e+00 2.50088245e-01 3.04737836e-01 -2.04318270e-01
5.83700597e-01 -8.89135059e-03 1.01097429e+00 4.81767058e-01
6.37291670e-02 1.05997300e+00 -7.87372291e-02 -4.09118354e-01
-7.65576363e-01 -1.25227475e+00 -2.55946636e-01 6.09689832e-01
6.10862732e-01 -5.98280549e-01 -2.12821901e-01 -3.64908397e-01
2.12158173e-01 1.82352230e-01 -5.71522582e-03 2.99479980e-02
-9.81751800e-01 -5.28938711e-01 6.40888885e-02 5.56115150e-01
5.00058174e-01 -5.41427672e-01 -3.83522570e-01 3.90944093e-01
-1.11518070e-01 -1.36235833e+00 -7.69082606e-01 -1.26996711e-02
-1.19314444e+00 -9.86313105e-01 -3.31676275e-01 -5.92220843e-01
9.07916546e-01 8.34554732e-01 9.82499957e-01 8.19862932e-02
-6.45745471e-02 -5.88229410e-02 -1.45027384e-01 -1.55492514e-01
4.99409199e-01 2.15695664e-01 1.69505522e-01 -5.07311746e-02
5.38063228e-01 -1.18149877e+00 -7.87030935e-01 7.36999631e-01
-6.24430418e-01 -2.02052176e-01 8.32620740e-01 8.33064675e-01
1.10637724e+00 -2.46798918e-01 -9.52794254e-02 -4.68377262e-01
3.07833087e-02 -1.42202467e-01 -1.00801826e+00 -2.58906633e-02
-5.50526142e-01 -5.62491305e-02 6.31881535e-01 -2.27840289e-01
-5.75813174e-01 4.53195482e-01 -1.34168386e-01 -7.76940107e-01
2.51252148e-02 3.70958298e-01 -3.74070287e-01 -4.59945560e-01
3.05163622e-01 1.95837751e-01 -2.27100119e-01 -5.82534134e-01
3.72482628e-01 4.69726622e-01 8.75702679e-01 -7.55128205e-01
1.65583801e+00 7.60050833e-01 8.18367526e-02 -7.47363985e-01
-7.74993896e-01 -8.51467073e-01 -9.85295057e-01 1.36098012e-01
5.85595906e-01 -1.28657138e+00 -7.34511256e-01 3.87227893e-01
-1.10642099e+00 1.65391028e-01 -2.88977116e-01 6.64217472e-01
-4.45732504e-01 3.46471965e-01 -2.83874303e-01 -4.63417590e-01
-3.09407622e-01 -1.30701816e+00 1.45326483e+00 1.09985679e-01
8.99901688e-02 -4.92053479e-01 1.78917989e-01 2.31724724e-01
3.36846799e-01 1.98779508e-01 4.01633292e-01 -3.13934535e-01
-1.26702213e+00 -3.01860690e-01 -5.74212790e-01 -1.11552496e-02
2.63563246e-01 -1.88020766e-01 -7.73162663e-01 -4.39683825e-01
3.48828509e-02 2.14513689e-02 6.43976986e-01 4.93109114e-02
1.27530813e+00 2.50166208e-02 -4.21294153e-01 1.18826902e+00
1.40490305e+00 -3.16030741e-01 5.46869040e-01 3.49967092e-01
1.15461707e+00 3.12752813e-01 8.72858346e-01 4.39026207e-01
6.91704452e-01 9.55191612e-01 5.62959611e-01 -8.49053413e-02
1.95204794e-01 -3.62910241e-01 1.32932723e-01 1.20335841e+00
-1.65350422e-01 3.03983659e-01 -1.09790075e+00 3.70940983e-01
-2.01858473e+00 -6.35828972e-01 -2.76370347e-01 2.48701096e+00
5.78154981e-01 9.47388485e-02 -2.29091793e-01 -1.71464249e-01
5.30117869e-01 2.79513121e-01 -6.74299002e-01 4.69550729e-01
-4.54554185e-02 1.78976551e-01 8.33126783e-01 3.31131727e-01
-1.09095430e+00 1.04002500e+00 4.91178322e+00 7.70969450e-01
-1.19093025e+00 1.05675481e-01 -2.41123810e-02 -9.78401825e-02
-1.82410046e-01 3.03495318e-01 -7.65909672e-01 3.82572263e-01
3.77707362e-01 3.44280414e-02 3.15666378e-01 9.78837013e-01
1.61622092e-01 1.81812823e-01 -9.56146359e-01 1.38377714e+00
-8.73479247e-02 -1.15573919e+00 -1.88007936e-01 3.85464311e-01
5.51083326e-01 6.33316517e-01 -2.26831153e-01 6.98606819e-02
2.94446468e-01 -6.33920610e-01 6.70161963e-01 3.24763715e-01
6.54726982e-01 -7.86825061e-01 6.91701949e-01 2.59707868e-01
-1.65169859e+00 1.11972958e-01 -7.42920935e-01 1.12605512e-01
2.97031611e-01 9.34620619e-01 -3.82278740e-01 8.03104818e-01
8.59769523e-01 1.06169343e+00 -6.10765934e-01 1.42446136e+00
-1.05959661e-01 2.74301857e-01 -8.78818512e-01 4.33539718e-01
1.50777260e-02 -6.07054532e-01 5.37167192e-01 8.81293774e-01
7.01479197e-01 1.78069517e-01 6.38759673e-01 6.74081147e-01
-6.52872473e-02 1.45271033e-01 -3.23856741e-01 5.41079283e-01
8.33718956e-01 1.40843010e+00 -6.06757700e-01 1.20915070e-01
-5.53474247e-01 8.34483504e-01 3.21703464e-01 1.02807499e-01
-7.33173549e-01 -2.88288862e-01 1.08918345e+00 3.59237701e-01
3.47821683e-01 -6.87272727e-01 -5.15237331e-01 -1.43390393e+00
4.99060988e-01 -7.42655098e-01 6.29748330e-02 -4.25938725e-01
-1.56667554e+00 4.54807937e-01 -1.80049255e-01 -1.81426585e+00
1.18189789e-01 -2.76261538e-01 -6.33945823e-01 7.38124073e-01
-1.65826786e+00 -1.34581828e+00 -8.43165040e-01 7.25577593e-01
3.01854044e-01 2.30264604e-01 3.75959188e-01 6.72690213e-01
-5.64283729e-01 4.77771968e-01 1.46816343e-01 1.46987855e-01
1.00981474e+00 -9.16550100e-01 6.44303441e-01 1.02930641e+00
1.52524814e-01 8.70179534e-01 4.65227455e-01 -6.73939347e-01
-1.69432878e+00 -1.22913373e+00 7.47065365e-01 -4.25602496e-01
5.97283185e-01 -5.91756582e-01 -1.05478716e+00 6.06925428e-01
-4.29761380e-01 5.45945227e-01 3.82884771e-01 2.51088977e-01
-5.52333951e-01 -5.30307353e-01 -9.10833478e-01 6.26472235e-01
1.50606775e+00 -4.36585784e-01 -3.89659613e-01 3.48144591e-01
5.50497591e-01 -8.75615358e-01 -9.91530299e-01 7.44177282e-01
4.57877398e-01 -9.38328505e-01 1.25663948e+00 2.69481111e-02
1.33255601e-01 -7.46542394e-01 -3.54525626e-01 -1.05563688e+00
-3.38960975e-01 -6.53642356e-01 1.59031957e-01 1.40447497e+00
-5.81257194e-02 -8.60145330e-01 7.39955962e-01 5.44198573e-01
-3.45690101e-01 -5.95636845e-01 -1.11788583e+00 -8.92525494e-01
-3.09413135e-01 -5.53378940e-01 9.41949904e-01 9.21780765e-01
-6.62257373e-01 3.31802994e-01 -3.24741036e-01 5.85914075e-01
9.04850900e-01 4.52508301e-01 1.48223662e+00 -1.49238145e+00
2.21790284e-01 -1.98378295e-01 -8.09491754e-01 -1.44259727e+00
-3.96046303e-02 -6.56751692e-01 4.61700499e-01 -1.34817469e+00
-2.45130584e-02 -9.93951738e-01 -1.77329719e-01 3.45695198e-01
-2.72679210e-01 3.63005757e-01 2.20270127e-01 7.58182168e-01
-4.90138143e-01 9.31514382e-01 8.95834863e-01 1.52049899e-01
-2.65705854e-01 5.02826422e-02 -3.93649429e-01 9.29748774e-01
4.43044960e-01 -7.17110276e-01 -7.92875588e-02 -7.45374382e-01
7.73686245e-02 -2.62109458e-01 4.74270880e-01 -1.31423688e+00
5.56452811e-01 -3.90069753e-01 4.20026451e-01 -1.23769987e+00
5.27608454e-01 -1.23937619e+00 2.94922143e-01 1.03447624e-01
2.69945771e-01 4.98014450e-01 -1.63886264e-01 6.88609362e-01
-4.05634135e-01 3.84778470e-01 7.19152272e-01 3.06524903e-01
-3.15457642e-01 9.77297008e-01 4.86541212e-01 -9.81779322e-02
8.05941880e-01 -2.92238951e-01 -9.20817032e-02 -9.74271744e-02
-9.20448229e-02 5.23076832e-01 9.86892939e-01 4.28378344e-01
6.53398454e-01 -1.57826245e+00 -6.23910308e-01 4.39298540e-01
3.89321357e-01 8.62124860e-01 1.16874129e-01 1.12373376e+00
-4.65428829e-01 5.71877956e-02 7.70932734e-02 -1.05725288e+00
-1.05946660e+00 1.73620895e-01 2.25960463e-02 -1.55741856e-01
-9.49006200e-01 6.75655484e-01 2.02035517e-01 -7.73456633e-01
1.20146554e-02 -3.94123554e-01 3.58808607e-01 -5.17783016e-02
2.09766194e-01 2.33384892e-01 2.57407427e-01 -8.91567171e-01
-6.57938659e-01 1.23692286e+00 -7.50907436e-02 1.67237818e-01
1.51777625e+00 -9.97996256e-02 -3.06869984e-01 1.16758049e-01
1.13578212e+00 4.38326478e-01 -1.26881003e+00 -6.61977649e-01
9.77764875e-02 -1.07633173e+00 3.70305553e-02 9.75229964e-02
-1.28010309e+00 8.48865509e-01 4.62200135e-01 -3.03497970e-01
1.11004353e+00 -1.81440294e-01 1.12979150e+00 5.75770259e-01
5.93945682e-01 -8.30178380e-01 -1.22938976e-01 6.48099422e-01
8.58066320e-01 -1.33394182e+00 4.77604896e-01 -8.63214791e-01
-1.47658601e-01 1.03756571e+00 6.96597397e-01 -4.81032461e-01
6.85491621e-01 1.42052799e-01 8.80337730e-02 -2.13138595e-01
-2.13596046e-01 -1.41255051e-01 4.61743206e-01 3.82829428e-01
1.19465895e-01 -1.32034704e-01 -1.16422683e-01 2.73656785e-01
-4.25735146e-01 -8.64796638e-02 -4.27029133e-02 9.18528378e-01
-2.89526284e-01 -1.18791175e+00 -5.18864334e-01 4.39032704e-01
-1.69655159e-02 -1.80438861e-01 -5.43468446e-03 6.67406559e-01
2.91933725e-03 6.41488731e-01 1.85716197e-01 -4.90913391e-01
7.53705919e-01 -5.03079414e-01 1.14087574e-01 -5.61948180e-01
-5.29256344e-01 4.09966618e-01 -4.18348193e-01 -1.11207342e+00
-4.27381307e-01 -9.65536833e-01 -1.09012151e+00 -3.31780046e-01
-5.77997148e-01 2.13846471e-03 6.87075496e-01 8.75648737e-01
5.77225685e-01 8.62071738e-02 8.38050961e-01 -1.21441114e+00
-6.02182508e-01 -6.95310354e-01 -2.22470835e-01 3.53989035e-01
3.33920181e-01 -7.96005189e-01 -2.80861557e-01 -3.41802239e-01] | [7.6908769607543945, -3.015469551086426] |
c8d45f55-c253-4f1c-be87-c5352f74c51e | a-multimodal-corpus-for-mutual-gaze-and-joint | null | null | https://aclanthology.org/L18-1019 | https://aclanthology.org/L18-1019.pdf | A Multimodal Corpus for Mutual Gaze and Joint Attention in Multiparty Situated Interaction | null | ['Patrik Jonell', 'Alex', 'Vanya Avramova', 'Gabriel Skantze', 'Dimosthenis Kontogiorgos', 'Catharine Oertel', 'Simon erson', 'Jonas Beskow', 'Joakim Gustafson'] | 2018-05-01 | a-multimodal-corpus-for-mutual-gaze-and-joint-1 | https://aclanthology.org/L18-1019 | https://aclanthology.org/L18-1019.pdf | lrec-2018-5 | ['mutual-gaze'] | ['computer-vision'] | [-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] | [-7.164504528045654, 3.819340229034424] |
86b66993-4391-41b3-a2fc-97868f97e4c1 | graphon-aided-joint-estimation-of-multiple | 2202.05686 | null | https://arxiv.org/abs/2202.05686v1 | https://arxiv.org/pdf/2202.05686v1.pdf | Graphon-aided Joint Estimation of Multiple Graphs | We consider the problem of estimating the topology of multiple networks from nodal observations, where these networks are assumed to be drawn from the same (unknown) random graph model. We adopt a graphon as our random graph model, which is a nonparametric model from which graphs of potentially different sizes can be drawn. The versatility of graphons allows us to tackle the joint inference problem even for the cases where the graphs to be recovered contain different number of nodes and lack precise alignment across the graphs. Our solution is based on combining a maximum likelihood penalty with graphon estimation schemes and can be used to augment existing network inference methods. We validate our proposed approach by comparing its performance against competing methods in synthetic and real-world datasets. | ['Santiago Segarra', 'Madeline Navarro'] | 2022-02-11 | null | null | null | null | ['graphon-estimation'] | ['graphs'] | [ 2.40888298e-01 5.02773285e-01 -2.84777135e-01 -9.13780183e-02
-2.87713468e-01 -9.00006831e-01 7.74582028e-01 1.25238225e-01
5.26166819e-02 1.18307126e+00 -2.98221141e-01 -3.46229553e-01
-5.23206294e-01 -1.10051394e+00 -7.85510957e-01 -5.68683624e-01
-5.20893455e-01 1.10823047e+00 4.83941853e-01 3.54872018e-01
8.11048672e-02 5.94076395e-01 -5.97253859e-01 -5.07378042e-01
6.40676379e-01 3.43987763e-01 8.14942420e-02 8.29731762e-01
1.24210201e-01 6.03912115e-01 -6.39350235e-01 -5.31156540e-01
2.52378523e-01 -1.03850588e-01 -6.12795353e-01 2.39505529e-01
4.19460475e-01 -1.08150259e-01 -5.59754133e-01 1.09201598e+00
2.60721654e-01 -2.20081314e-01 9.48123395e-01 -1.67590845e+00
-1.28523916e-01 9.06318307e-01 -8.21792483e-01 1.75786465e-01
4.83648986e-01 -2.46103778e-01 1.19679165e+00 -1.70226231e-01
8.46404970e-01 1.57749593e+00 8.23655903e-01 -2.19526187e-01
-1.90213954e+00 -5.57125747e-01 1.48069739e-01 -1.29471660e-01
-1.71375966e+00 -2.58068204e-01 8.51426542e-01 -2.97663718e-01
3.77958357e-01 1.87303156e-01 4.46342528e-01 1.12990701e+00
1.64736852e-01 3.01521033e-01 9.60956037e-01 -1.78358272e-01
2.63751864e-01 -5.46326227e-02 -6.79683387e-02 7.14042604e-01
8.38608563e-01 -1.03212580e-01 -1.56425655e-01 -8.19694102e-01
1.00270784e+00 -7.56243095e-02 -2.44606018e-01 -5.68043649e-01
-1.05317664e+00 8.69781613e-01 3.24160159e-01 -1.77903622e-01
-3.12388152e-01 4.63364542e-01 -7.63290655e-03 3.41478020e-01
2.08644316e-01 -4.17689011e-02 -3.48977149e-01 3.28188628e-01
-9.46986139e-01 1.66598111e-01 1.39061630e+00 1.23973119e+00
7.33342290e-01 -3.07825347e-03 2.00561240e-01 4.59601134e-01
7.41397858e-01 5.27739048e-01 -4.10511732e-01 -8.43502760e-01
5.92622876e-01 4.98792082e-01 9.04384330e-02 -1.40151751e+00
-1.29240766e-01 -3.46259505e-01 -1.07072604e+00 -2.25508690e-01
6.21335387e-01 -3.69471282e-01 -8.77302945e-01 1.78471255e+00
3.53302807e-01 8.27869713e-01 -3.12974364e-01 2.83487707e-01
5.70024133e-01 3.48478496e-01 -1.94853783e-01 -2.98219562e-01
7.84805179e-01 -5.79966724e-01 -5.09195566e-01 -1.34316329e-02
2.45009407e-01 -4.79452372e-01 1.99906394e-01 3.54306608e-01
-8.01017940e-01 -1.18365087e-01 -8.17489088e-01 4.38218892e-01
-1.48655862e-01 -2.99134076e-01 5.02944469e-01 8.13718915e-01
-1.31969774e+00 6.19184494e-01 -9.17205930e-01 -5.58053613e-01
1.97095588e-01 3.45738173e-01 -5.39719999e-01 -2.13851139e-01
-7.25964963e-01 4.56367016e-01 6.57583296e-01 3.48846316e-01
-9.87897754e-01 -2.08561361e-01 -7.32094705e-01 4.37419750e-02
8.25677693e-01 -7.98423767e-01 4.66736585e-01 -5.40063620e-01
-1.20928657e+00 3.69347304e-01 -1.97826058e-01 -3.79442751e-01
4.89291072e-01 3.63010675e-01 -3.86496484e-01 1.64126083e-01
9.06689093e-02 -5.73437614e-03 7.91434765e-01 -1.35603821e+00
-4.78722751e-02 -2.11689830e-01 7.77173936e-02 -2.15186119e-01
3.26660961e-01 -3.29669982e-01 -5.69717169e-01 -5.33022106e-01
3.36293757e-01 -1.23072386e+00 -3.42357993e-01 -1.29724378e-02
-1.20515907e+00 -7.89815336e-02 8.44970345e-01 -5.52050471e-01
1.16495931e+00 -1.54213071e+00 4.39253181e-01 9.76777613e-01
7.12253392e-01 -2.94283003e-01 -1.99418128e-01 1.05848217e+00
1.33209825e-01 4.88647819e-01 -2.15590492e-01 -2.65960693e-01
-8.29194039e-02 5.45184314e-01 5.38346320e-02 8.27051938e-01
-2.05612212e-01 6.12759590e-01 -9.01894093e-01 -6.95831597e-01
-4.23513120e-03 9.07439515e-02 -1.09285682e-01 2.51905710e-01
-1.77022070e-01 5.02014995e-01 -6.89697742e-01 4.85622227e-01
9.38627124e-01 -6.66834414e-01 1.15156698e+00 -9.16114599e-02
4.74770635e-01 -7.01957047e-02 -1.71219695e+00 1.18482542e+00
-1.15439013e-01 4.12504196e-01 4.24721926e-01 -1.16190350e+00
7.77436852e-01 2.62138218e-01 2.66820669e-01 3.04005384e-01
8.60488266e-02 -9.07187983e-02 3.06790382e-01 -8.73845071e-02
-4.76560332e-02 8.17509517e-02 -1.01127088e-01 6.98545694e-01
2.11713880e-01 -2.73571163e-01 4.14731085e-01 6.43551409e-01
1.59211421e+00 -3.07364166e-01 6.76988602e-01 -3.66857082e-01
2.04894856e-01 -5.24471223e-01 5.34388781e-01 1.09431612e+00
2.60535404e-02 2.35638380e-01 1.04519677e+00 -1.88872106e-02
-1.01471817e+00 -1.53591752e+00 -1.89989865e-01 3.46190184e-01
7.98829868e-02 -3.75584513e-01 -2.53899783e-01 -7.87575245e-01
3.57041299e-01 2.65375108e-01 -6.32008314e-01 3.66137356e-01
-2.32287303e-01 -8.16985309e-01 5.56351900e-01 6.11053407e-02
-1.06343292e-01 -5.30332983e-01 4.71776396e-01 2.19450518e-01
-1.11461744e-01 -1.48218381e+00 -1.57543883e-01 -1.13813676e-01
-9.61330950e-01 -1.58355772e+00 -2.49875188e-01 -6.17442667e-01
9.59710181e-01 7.92548619e-03 1.35267377e+00 3.34965318e-01
-8.15385059e-02 4.73088026e-01 -1.03562668e-01 1.58654936e-02
-6.44614995e-01 2.66063213e-01 5.01617752e-02 7.21942037e-02
-3.23424667e-01 -1.32295835e+00 -3.02339703e-01 3.27094644e-01
-9.70602632e-01 -9.23022553e-02 5.37307084e-01 6.90889359e-01
4.55744892e-01 3.94100517e-01 5.27297556e-01 -1.49334419e+00
7.80752361e-01 -1.04990923e+00 -9.22837138e-01 4.81214643e-01
-5.72655261e-01 2.00755462e-01 6.70273364e-01 -3.95802557e-01
-5.08646965e-01 -1.24655046e-01 4.33355033e-01 -2.91454881e-01
-4.28146310e-02 7.52981782e-01 -2.23983631e-01 -4.49818164e-01
1.17704324e-01 -1.77494362e-01 9.09509212e-02 -4.53195661e-01
5.71462810e-01 2.97095418e-01 2.74473250e-01 -6.55770183e-01
1.16562927e+00 5.00358820e-01 6.66773438e-01 -9.00251091e-01
-1.49044037e-01 -2.51940370e-01 -8.49399626e-01 -1.18999295e-01
2.52638340e-01 -7.46100843e-01 -8.82473350e-01 2.65791297e-01
-1.02004373e+00 -2.35338718e-01 3.86031359e-01 3.51292014e-01
-1.21398926e-01 8.54499102e-01 -7.38979340e-01 -8.98658931e-01
1.73904583e-01 -8.04948926e-01 7.93406308e-01 2.89988816e-02
-2.46952444e-01 -1.73692119e+00 2.95592785e-01 -1.06641866e-01
1.02567501e-01 8.37069154e-01 9.35153425e-01 -7.98042238e-01
-9.30261672e-01 -3.53918463e-01 -4.57800925e-01 -1.65403217e-01
3.18963796e-01 4.87519354e-01 -3.55224460e-01 -6.78421319e-01
-6.68180048e-01 -8.14948753e-02 5.42607307e-01 5.12253642e-01
8.76289189e-01 -4.51778859e-01 -8.48322272e-01 4.23683763e-01
1.74252725e+00 -4.68725324e-01 5.42385042e-01 -2.58143455e-01
8.62423778e-01 3.92417967e-01 4.50125262e-02 4.88291234e-01
5.16689777e-01 6.21936202e-01 6.20863497e-01 9.90122706e-02
1.93096980e-01 -5.90030611e-01 -1.35455251e-01 9.62248683e-01
-5.91794625e-02 -9.90992367e-01 -9.54249322e-01 5.37587643e-01
-1.99380600e+00 -7.53281057e-01 -5.58958411e-01 2.22573376e+00
4.01084870e-01 2.95800388e-01 2.17854723e-01 -2.66922325e-01
1.13804388e+00 2.88463771e-01 -5.00674129e-01 -4.15432975e-02
-7.31583089e-02 1.42490104e-01 7.85025477e-01 6.94704413e-01
-7.59483874e-01 5.14591396e-01 7.69631958e+00 7.67287672e-01
-4.65825737e-01 -3.33000451e-01 4.29322958e-01 2.97993064e-01
-2.69433320e-01 6.40969455e-01 -4.10517424e-01 5.76080620e-01
1.11723912e+00 -3.17214429e-01 5.08712113e-01 3.58116895e-01
2.08761320e-01 -2.85810232e-01 -1.07490444e+00 6.19648099e-01
-2.99271375e-01 -1.16640699e+00 5.14429323e-02 5.29029250e-01
8.65958989e-01 1.53489783e-01 -4.42412794e-01 -4.31566387e-02
1.18294525e+00 -1.11937904e+00 6.21188246e-02 8.10186028e-01
8.16151083e-01 -4.83794123e-01 6.93147540e-01 5.28063834e-01
-1.47369635e+00 2.23303229e-01 -2.59287328e-01 3.59314829e-02
4.37468529e-01 8.64510715e-01 -1.26124179e+00 1.07315934e+00
2.89104134e-01 7.39817858e-01 -4.00905848e-01 1.12774599e+00
-2.26564765e-01 8.47988129e-01 -7.34175563e-01 1.17239945e-01
-2.11602807e-01 -6.77828252e-01 8.45682263e-01 7.89027452e-01
3.01682949e-01 -2.77904183e-01 6.12773001e-01 9.61397767e-01
-2.76072741e-01 -2.50800908e-01 -9.32482719e-01 -1.74931288e-01
9.82711554e-01 1.30954921e+00 -1.09878075e+00 -2.22873852e-01
-3.92020494e-01 6.91676557e-01 7.13308930e-01 6.22069597e-01
-8.63907099e-01 -3.39015394e-01 1.00235783e-01 2.42759541e-01
4.37580764e-01 -4.64160681e-01 2.92447895e-01 -1.01782644e+00
-1.40336171e-01 -5.63946426e-01 3.89562666e-01 -5.46453834e-01
-1.66780686e+00 4.84918743e-01 3.39378566e-01 -9.14024830e-01
-3.56868654e-01 -2.80109107e-01 -8.86729062e-01 7.43326902e-01
-1.05356550e+00 -1.14408517e+00 -1.25221089e-01 4.32147741e-01
-2.57525414e-01 4.67849597e-02 6.53497517e-01 -3.80279287e-03
-5.50970018e-01 2.67564714e-01 4.30009872e-01 1.76790535e-01
3.68474990e-01 -1.42482746e+00 3.99451852e-01 9.07658041e-01
1.21640198e-01 4.97015685e-01 8.40773463e-01 -1.11057067e+00
-1.39884639e+00 -1.14712501e+00 4.12213117e-01 -4.19298202e-01
1.07947922e+00 -6.27498865e-01 -7.00487554e-01 1.17783141e+00
4.33544368e-02 3.29522640e-01 5.18031657e-01 2.87727773e-01
-1.15275115e-01 5.36691584e-02 -1.35795033e+00 3.91449928e-01
1.06783462e+00 -3.50795418e-01 -7.54010305e-02 3.51907879e-01
3.99025887e-01 1.37352729e-02 -1.27279413e+00 2.02946529e-01
4.04132754e-01 -7.11130798e-01 8.81532192e-01 -4.86026943e-01
-2.03507781e-01 -4.80470479e-01 -5.64413518e-02 -1.30267596e+00
-3.06186557e-01 -8.75187933e-01 -3.53461921e-01 1.44513011e+00
4.12249506e-01 -9.10075009e-01 8.47784460e-01 2.76480943e-01
8.18997741e-01 -2.47294083e-01 -1.07990944e+00 -8.53076458e-01
-3.89854997e-01 -4.46682833e-02 6.00268543e-01 1.03877485e+00
-2.99380541e-01 6.79224551e-01 -5.41697979e-01 6.58063233e-01
1.26103222e+00 9.14915279e-02 1.14049685e+00 -1.82222569e+00
-6.68622971e-01 -7.73567334e-02 -6.85399473e-01 -9.57920671e-01
4.96193200e-01 -8.70340765e-01 -1.99360400e-01 -1.48076284e+00
5.44451535e-01 -5.96045852e-01 2.21034721e-01 1.56565100e-01
-1.52826672e-02 8.96369293e-02 -1.89188778e-01 1.79016113e-01
-6.20407343e-01 4.05118048e-01 1.02416503e+00 1.28401980e-01
4.53060381e-02 3.75999123e-01 -3.94959390e-01 7.00933158e-01
5.32901824e-01 -8.19199324e-01 -6.44018054e-01 -6.14223331e-02
2.98116118e-01 7.36505330e-01 5.45713782e-01 -6.51358545e-01
2.27479890e-01 -1.59916580e-01 7.82347545e-02 -4.62298423e-01
1.82120442e-01 -9.58244562e-01 9.05094981e-01 2.52519906e-01
-1.62402242e-02 1.51392922e-01 -1.23886518e-01 1.42903364e+00
2.31112186e-02 -9.89638120e-02 4.18744147e-01 -8.68235379e-02
-6.23334609e-02 6.55960381e-01 -1.47665039e-01 8.97185728e-02
9.07668531e-01 -1.87984124e-01 -3.90671015e-01 -8.17464292e-01
-1.00298584e+00 3.78137261e-01 6.33147061e-01 -1.63498610e-01
3.76182139e-01 -1.11609030e+00 -8.61933351e-01 -2.47676238e-01
-1.72973394e-01 4.66722734e-02 -1.46501765e-01 9.54654336e-01
-4.94538724e-01 -6.33672252e-02 1.02268606e-01 -6.45377874e-01
-1.14367592e+00 4.33233738e-01 1.24543779e-01 -7.16026723e-01
-4.33972329e-01 3.45591277e-01 -2.20137000e-01 -7.68240213e-01
-5.12968823e-02 4.09495607e-02 -6.10910989e-02 -3.55083615e-01
-1.51984096e-01 7.07986951e-01 -4.75490868e-01 -6.17248654e-01
-3.13860267e-01 2.98699170e-01 -5.74670210e-02 -2.64694938e-03
1.44124126e+00 -4.78550971e-01 -5.32330573e-01 6.31014585e-01
1.03490055e+00 4.62709665e-01 -9.64522779e-01 -3.84693801e-01
1.47343174e-01 -4.99449670e-01 -2.11889088e-01 -3.98701519e-01
-9.92312610e-01 2.75465906e-01 -1.83078507e-03 8.61680686e-01
5.85568368e-01 1.80148706e-01 2.78082609e-01 1.11889742e-01
5.62867105e-01 -5.07950962e-01 -4.23961997e-01 4.55212072e-02
5.37899852e-01 -1.13578367e+00 4.67419893e-01 -9.81115937e-01
8.51825848e-02 1.22195745e+00 2.58089572e-01 -5.99063516e-01
9.96981680e-01 3.32128614e-01 -6.10319316e-01 -2.89695740e-01
-8.60875070e-01 3.63788009e-02 2.07472995e-01 9.39011753e-01
1.68639362e-01 2.63555199e-01 -6.52300343e-02 2.04520285e-01
-4.15300764e-02 -3.19923729e-01 9.50777054e-01 5.51318467e-01
-1.21358940e-02 -1.35889149e+00 -1.76751599e-01 8.07791471e-01
-2.61623532e-01 -1.37598785e-02 -5.65591276e-01 1.00467992e+00
-3.73706728e-01 9.57592964e-01 -1.70757815e-01 -6.10311329e-02
-7.08249882e-02 -5.69744051e-01 4.56236213e-01 -6.40091121e-01
1.90175042e-01 1.42271578e-01 4.38660771e-01 -3.14552575e-01
-4.39343721e-01 -4.75453138e-01 -6.37509763e-01 -8.25697064e-01
-6.37132347e-01 2.87553787e-01 2.16651484e-01 8.47375691e-01
3.55499059e-01 3.70189250e-01 7.02855051e-01 -7.19930828e-01
-2.97230005e-01 -1.04028094e+00 -1.31421316e+00 2.39974558e-01
2.37683371e-01 -8.31951976e-01 -7.21337795e-01 -2.68359810e-01] | [6.935458183288574, 5.330173015594482] |
dc50ab44-4f5a-4a4b-a053-24bac1ae1451 | bighand22m-benchmark-hand-pose-dataset-and | 1704.02612 | null | http://arxiv.org/abs/1704.02612v2 | http://arxiv.org/pdf/1704.02612v2.pdf | BigHand2.2M Benchmark: Hand Pose Dataset and State of the Art Analysis | In this paper we introduce a large-scale hand pose dataset, collected using a
novel capture method. Existing datasets are either generated synthetically or
captured using depth sensors: synthetic datasets exhibit a certain level of
appearance difference from real depth images, and real datasets are limited in
quantity and coverage, mainly due to the difficulty to annotate them. We
propose a tracking system with six 6D magnetic sensors and inverse kinematics
to automatically obtain 21-joints hand pose annotations of depth maps captured
with minimal restriction on the range of motion. The capture protocol aims to
fully cover the natural hand pose space. As shown in embedding plots, the new
dataset exhibits a significantly wider and denser range of hand poses compared
to existing benchmarks. Current state-of-the-art methods are evaluated on the
dataset, and we demonstrate significant improvements in cross-benchmark
performance. We also show significant improvements in egocentric hand pose
estimation with a CNN trained on the new dataset. | ['Tae-Kyun Kim', 'Siddhant Jain', 'Bjorn Stenger', 'Shanxin Yuan', 'Qi Ye'] | 2017-04-09 | bighand2-2m-benchmark-hand-pose-dataset-and | http://openaccess.thecvf.com/content_cvpr_2017/html/Yuan_BigHand2.2M_Benchmark_Hand_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Yuan_BigHand2.2M_Benchmark_Hand_CVPR_2017_paper.pdf | cvpr-2017-7 | ['art-analysis'] | ['computer-vision'] | [-6.96076974e-02 -2.36217082e-01 -3.08123320e-01 -7.94893429e-02
-6.55984044e-01 -7.53569901e-01 3.49445075e-01 -6.56424522e-01
-3.93142134e-01 7.55350828e-01 6.92599356e-01 5.38208485e-01
5.54719232e-02 -3.04041743e-01 -6.66861773e-01 -3.82698417e-01
-5.89395931e-04 9.06569898e-01 1.21164866e-01 -1.97532818e-01
1.54933512e-01 8.43366563e-01 -1.32160378e+00 -1.44633263e-01
-2.61539612e-02 6.34879112e-01 2.33929306e-01 7.24157572e-01
5.57979524e-01 5.54537237e-01 -5.77263176e-01 -1.28516510e-01
5.42798579e-01 6.06392790e-03 -1.09060955e+00 2.00193986e-01
8.32989633e-01 -1.09312773e+00 -1.06338799e+00 8.01753283e-01
1.21538532e+00 -1.80948079e-01 4.27069336e-01 -1.15840578e+00
-2.71159530e-01 2.39585355e-01 -6.43021345e-01 -1.93640053e-01
9.49453294e-01 3.01533550e-01 7.24537909e-01 -7.34828353e-01
1.29792464e+00 1.19943631e+00 7.12849140e-01 1.02162683e+00
-1.20962882e+00 -4.98498619e-01 8.61041546e-02 -3.07843667e-02
-1.55753708e+00 -2.53550649e-01 9.79918301e-01 -5.68464339e-01
1.12870133e+00 -3.93324494e-02 9.19264197e-01 1.94221354e+00
-8.52195323e-02 9.57222939e-01 8.16989303e-01 -4.14829791e-01
8.98777097e-02 -4.20449227e-01 -2.69459218e-01 6.02835476e-01
2.89663494e-01 5.93859702e-02 -9.38956499e-01 2.73704678e-02
1.47263241e+00 1.69599548e-01 -5.05718827e-01 -1.04257429e+00
-1.81988811e+00 1.73893929e-01 5.14713585e-01 -1.05518736e-02
-5.59440076e-01 3.90284061e-01 3.76106739e-01 -4.79567684e-02
1.56954423e-01 2.57548422e-01 -6.21105731e-01 -6.04768872e-01
-6.55023038e-01 6.79989040e-01 7.42751181e-01 1.47038972e+00
1.21882886e-01 -2.92286456e-01 -1.71669930e-01 4.30038899e-01
2.89374739e-01 5.91624200e-01 2.72049844e-01 -1.14649761e+00
7.72617161e-01 4.20223653e-01 4.11149353e-01 -7.29244649e-01
-5.44859290e-01 -3.34577888e-01 -5.05970359e-01 2.28117004e-01
9.31881070e-01 -9.51040536e-02 -8.55674922e-01 1.75162661e+00
2.86167920e-01 -1.96105972e-01 -4.75836933e-01 1.24057198e+00
3.21014196e-01 -2.92361617e-01 -1.75334513e-01 2.54951447e-01
1.18506944e+00 -9.14548218e-01 -7.40583777e-01 -2.68106520e-01
2.99567938e-01 -5.59823275e-01 1.35696828e+00 4.82391357e-01
-8.69908273e-01 -1.88278899e-01 -9.35065866e-01 -7.04589188e-02
-7.17676729e-02 2.01600134e-01 8.20294917e-01 6.40472949e-01
-9.16696191e-01 4.78358805e-01 -1.21398640e+00 -4.76239413e-01
5.18596053e-01 3.51278067e-01 -6.98933423e-01 -6.73734248e-02
-7.58907437e-01 7.11685359e-01 1.44237041e-01 1.37604997e-01
-8.27592969e-01 -5.83548427e-01 -8.61589372e-01 -5.85581124e-01
1.31500930e-01 -6.11643076e-01 1.31301415e+00 3.23059782e-02
-1.54072011e+00 1.03217435e+00 -5.78184389e-02 1.20082021e-01
1.12258220e+00 -5.78491390e-01 1.79703161e-01 3.24217111e-01
1.53684169e-01 8.59682739e-01 6.76814437e-01 -1.10627854e+00
-1.70261055e-01 -9.69828308e-01 6.27790242e-02 7.83509836e-02
-3.54582489e-01 -1.69407815e-01 -7.79806614e-01 -9.90020335e-01
2.71903217e-01 -1.27338183e+00 6.22325055e-02 3.79672259e-01
-4.95347738e-01 8.38196874e-02 9.08508599e-01 -6.94616675e-01
7.10029721e-01 -1.60301185e+00 5.66532254e-01 9.41707566e-03
3.75781357e-01 -9.09390301e-02 2.01410353e-02 1.06930427e-01
2.02170238e-01 -3.47670346e-01 3.20803747e-02 -6.85674131e-01
6.17249236e-02 1.09629475e-01 -1.35738775e-01 6.11752450e-01
-1.37824640e-01 9.02274966e-01 -1.13320971e+00 -4.41598952e-01
2.49870300e-01 8.63668919e-01 -5.73367000e-01 2.25472361e-01
-2.17225447e-01 8.64607871e-01 -4.33942765e-01 9.01357651e-01
4.21234906e-01 -1.43464282e-01 1.42895252e-01 -3.91852736e-01
3.23926687e-01 3.55634242e-01 -1.22611427e+00 2.79281569e+00
-2.49139652e-01 5.50162494e-01 -1.01100378e-01 -9.09132510e-02
4.59779680e-01 5.24949372e-01 6.84404314e-01 -1.82885930e-01
3.54441285e-01 9.87941995e-02 -2.73634791e-01 -3.70654672e-01
3.84985626e-01 2.96197504e-01 1.84448883e-01 6.39291525e-01
2.77802702e-02 -2.98398137e-01 -2.06937298e-01 -9.83077474e-03
1.09336507e+00 6.94286466e-01 -1.85928449e-01 5.06582670e-02
-1.26847804e-01 1.20975532e-01 1.97582662e-01 3.41435015e-01
-5.26844263e-01 1.15768826e+00 3.36660564e-01 -4.30014461e-01
-1.27874112e+00 -1.18683505e+00 -1.57260880e-01 6.17255986e-01
-6.64939079e-03 -1.87816888e-01 -9.61321235e-01 -6.24102890e-01
1.61352828e-01 -1.04637027e-01 -8.56032610e-01 2.62079269e-01
-7.70710230e-01 -2.87111104e-01 7.73375809e-01 1.25346720e+00
6.11045957e-01 -1.09068489e+00 -1.02574217e+00 6.89699650e-02
-2.57007331e-01 -1.41429245e+00 -7.93302834e-01 -2.49937296e-01
-8.60883355e-01 -1.17930245e+00 -1.27581739e+00 -9.11095738e-01
5.97513199e-01 1.33307621e-01 1.04977751e+00 -5.65644145e-01
-5.32204390e-01 7.12986350e-01 -2.80358076e-01 -3.18597972e-01
2.27840155e-01 5.33233941e-01 4.40557092e-01 -4.87392157e-01
4.50073808e-01 -6.53411388e-01 -9.98047173e-01 3.20342273e-01
-2.59489805e-01 -9.83867720e-02 3.71392399e-01 7.91609764e-01
5.09425640e-01 -8.35611641e-01 4.23142403e-01 -1.83388963e-01
4.80816633e-01 1.64829046e-01 -2.66142726e-01 -2.00712271e-02
-3.09364229e-01 7.17127398e-02 -1.11280188e-01 -5.95748961e-01
-9.80939925e-01 6.06592596e-01 -9.05318558e-02 -7.34011531e-01
-1.44885629e-01 -2.61472821e-01 -3.82655352e-01 -1.93417221e-01
7.77489305e-01 -2.50824671e-02 2.67052293e-01 -6.74482584e-01
3.83659482e-01 7.92099059e-01 9.51774955e-01 -8.03711474e-01
4.37354237e-01 7.84893692e-01 8.89011193e-03 -3.68541807e-01
-4.18549865e-01 -2.58941412e-01 -1.33779860e+00 -2.38865539e-01
6.93353057e-01 -9.74125922e-01 -1.05026758e+00 9.80139971e-01
-1.38249469e+00 -5.39042175e-01 -2.08484426e-01 8.48772228e-01
-9.19927537e-01 3.29193383e-01 -7.61092901e-01 -5.36449313e-01
-4.40150023e-01 -1.33678615e+00 1.68844616e+00 -2.09039956e-01
-6.52958035e-01 -6.77699029e-01 2.33433962e-01 1.41464159e-01
4.72623250e-03 7.63023555e-01 2.68056154e-01 8.98669139e-02
-4.38401282e-01 -6.50111675e-01 -3.74945626e-02 4.18511368e-02
4.16841894e-01 -3.66158068e-01 -1.13226163e+00 -7.95447767e-01
-3.45059663e-01 -5.18794119e-01 2.53967553e-01 2.74013460e-01
1.01430368e+00 1.78278610e-01 -4.39737469e-01 5.38813353e-01
9.86360729e-01 -1.56568140e-01 6.19017661e-01 4.97486174e-01
1.09954178e+00 6.29372060e-01 4.36083853e-01 5.87488830e-01
3.48521411e-01 1.00514662e+00 4.85321641e-01 3.53180468e-01
-4.90229934e-01 -4.09872383e-01 7.41936788e-02 5.08333027e-01
-6.26112461e-01 1.53912038e-01 -1.00844169e+00 7.08387434e-01
-1.64780951e+00 -6.46848083e-01 1.81789264e-01 2.16525197e+00
1.08722425e+00 -1.77653059e-02 5.85318565e-01 4.94227856e-01
5.41246772e-01 1.47930562e-01 -7.96325505e-01 4.83739913e-01
-6.52141646e-02 3.85397524e-01 5.51396310e-01 3.36485296e-01
-8.90331388e-01 9.35198009e-01 6.98138857e+00 4.23023216e-02
-1.09249985e+00 1.18971325e-01 -2.81164855e-01 -7.51953781e-01
2.76253283e-01 -5.41718841e-01 -6.44430637e-01 9.24049839e-02
5.00425659e-02 4.93656665e-01 4.31260854e-01 9.35006917e-01
-1.92594007e-01 3.27954561e-01 -1.48950064e+00 1.25594735e+00
1.56514525e-01 -9.27596867e-01 -1.54405728e-01 3.20079505e-01
7.14527488e-01 1.87986955e-01 -4.10283171e-02 -2.82339007e-01
-8.47512707e-02 -8.42379749e-01 9.47709799e-01 4.61879015e-01
1.26935947e+00 -5.52039385e-01 3.68897051e-01 1.94641292e-01
-1.28123820e+00 3.79756182e-01 -1.05249614e-01 -3.18300664e-01
1.53736845e-01 -6.82483800e-03 -7.73164630e-01 9.40426216e-02
8.95829558e-01 7.57097900e-01 -4.72029030e-01 6.11911297e-01
-3.34005862e-01 -8.01757500e-02 -3.80841285e-01 -1.19772650e-01
-2.17909470e-01 4.27486360e-01 7.61035085e-01 8.57172251e-01
-4.73238528e-02 -2.68563390e-01 -2.01017871e-01 7.34277487e-01
-7.10009336e-02 -4.80688542e-01 -7.27056623e-01 1.64507851e-01
5.52665174e-01 8.56702924e-01 -4.94265676e-01 -6.20763563e-02
-1.55370638e-01 1.41552413e+00 2.67639697e-01 2.45826200e-01
-4.97223437e-01 -5.58013141e-01 9.79066610e-01 2.06735700e-01
8.51199627e-02 -7.09178686e-01 -4.26045775e-01 -1.37315941e+00
6.29884958e-01 -7.41108894e-01 -1.13571892e-02 -8.00173938e-01
-1.13745117e+00 5.33179939e-01 2.90837623e-02 -1.24576366e+00
-6.46535575e-01 -9.81077492e-01 8.17684904e-02 9.95970130e-01
-9.10535753e-01 -1.24135685e+00 -9.93923306e-01 9.63590086e-01
5.89134812e-01 -1.56241655e-01 1.13199461e+00 2.89585978e-01
-2.66995549e-01 8.60878944e-01 -1.51554465e-01 5.91535211e-01
9.18143272e-01 -1.28438365e+00 8.46173346e-01 3.57253015e-01
5.47063304e-03 7.43596017e-01 4.50133562e-01 -6.34683847e-01
-1.84066772e+00 -8.34943473e-01 3.01507533e-01 -1.35005176e+00
1.80523053e-01 -6.67851150e-01 -3.81797761e-01 1.10234582e+00
-1.32258639e-01 2.57653564e-01 7.39702433e-02 -2.04137806e-02
-6.37853503e-01 2.18762904e-01 -1.42948020e+00 5.84118009e-01
1.86798966e+00 -8.68939221e-01 -6.54916823e-01 4.63681430e-01
4.66522545e-01 -9.98520315e-01 -9.76704359e-01 3.30316991e-01
1.51816976e+00 -5.86618185e-01 1.16116929e+00 -7.67056346e-01
3.53940040e-01 -2.22131714e-01 -3.45163494e-01 -1.12844300e+00
-1.73524931e-01 -5.94184637e-01 -6.11057997e-01 9.80021715e-01
-1.73991799e-01 -3.32687974e-01 1.43661642e+00 5.26656806e-01
4.83939439e-01 -6.85598314e-01 -8.70423794e-01 -9.77222919e-01
-2.39743702e-02 -4.70214605e-01 7.77821839e-01 7.03587472e-01
2.54410684e-01 8.72491375e-02 -2.25599557e-01 3.01817842e-02
1.10063958e+00 -1.72580138e-01 1.07746768e+00 -1.08119261e+00
-1.22542329e-01 -2.37881213e-01 -8.11605573e-01 -1.31833804e+00
1.76549360e-01 -4.52476293e-01 1.17235212e-02 -1.37671292e+00
1.57377258e-01 -1.73676670e-01 6.57068416e-02 2.79314846e-01
1.42611369e-01 6.90359771e-01 3.72364707e-02 4.66474980e-01
-1.98232215e-02 3.46182764e-01 1.55202520e+00 -1.89594373e-01
-7.80552104e-02 -2.63422400e-01 -1.15791075e-01 6.76372588e-01
5.57650924e-01 -1.77576035e-01 -3.34788740e-01 -8.34230304e-01
-7.22081587e-02 -3.16174813e-02 6.92296684e-01 -1.00162506e+00
1.87890232e-01 1.27613366e-01 5.85167944e-01 -7.39609718e-01
5.05215108e-01 -6.58968329e-01 3.89813334e-02 5.49933136e-01
-1.62346438e-01 1.68444917e-01 2.41411272e-02 5.35663724e-01
1.82003811e-01 4.13682550e-01 5.05697310e-01 -8.38946775e-02
-5.85933447e-01 4.96407032e-01 1.21476330e-01 1.39554292e-01
8.94970596e-01 -4.13120449e-01 -9.52877179e-02 -4.32206452e-01
-6.59128606e-01 -6.96524084e-02 6.85591698e-01 7.54873931e-01
4.70743358e-01 -1.60421765e+00 -4.81356174e-01 1.85803369e-01
2.89953649e-01 5.02161443e-01 5.94720524e-03 7.03561187e-01
-8.66759598e-01 6.19119465e-01 -6.32358551e-01 -8.59806299e-01
-1.02163911e+00 2.62104630e-01 3.78183603e-01 1.53247714e-01
-8.99457514e-01 8.66249263e-01 -1.64092809e-01 -6.74903154e-01
8.08486700e-01 -4.89761770e-01 3.73582959e-01 -2.19633818e-01
5.04069924e-01 6.53104484e-01 1.48790210e-01 -7.26878643e-01
-6.11558735e-01 9.19984102e-01 5.61860800e-02 -3.46889347e-01
1.23488367e+00 1.76830113e-01 2.89738089e-01 1.61532223e-01
1.23070526e+00 3.13630700e-02 -1.86356580e+00 -3.19622457e-01
-3.94447923e-01 -8.69405091e-01 -1.80910379e-01 -9.11703944e-01
-1.09609079e+00 9.91457641e-01 9.45217907e-01 -5.81176996e-01
7.85984278e-01 1.75850481e-01 9.13586974e-01 6.05786264e-01
1.09569657e+00 -1.08621430e+00 2.57204145e-01 5.15343547e-01
1.32912326e+00 -1.21985400e+00 -6.23898618e-02 -4.29466039e-01
-3.70419949e-01 9.25543666e-01 7.44058907e-01 -1.20959163e-01
5.16693950e-01 4.98925269e-01 1.64563656e-01 -3.10877562e-01
-1.26479771e-02 7.18958229e-02 7.52662420e-02 9.42273796e-01
3.22625160e-01 1.30574018e-01 1.18462175e-01 3.71309727e-01
-4.33909029e-01 2.22868621e-01 6.63543344e-02 1.37076545e+00
2.42406249e-01 -1.07505143e+00 -4.57256585e-01 1.33934125e-01
-2.93727785e-01 4.68515158e-01 -7.25889266e-01 1.02986968e+00
-1.98389202e-01 5.14550805e-01 4.31322418e-02 -6.86804414e-01
7.65207767e-01 -7.35320523e-02 1.49988627e+00 -5.16373217e-01
-1.99074119e-01 -1.69953361e-01 -2.84104675e-01 -6.94425285e-01
-3.62638116e-01 -7.27056921e-01 -1.11441731e+00 -4.36760485e-01
-2.16160685e-01 -5.50669730e-01 7.75961637e-01 7.62195408e-01
4.66119260e-01 5.45521617e-01 1.29865080e-01 -1.70749366e+00
-7.77442098e-01 -1.31416607e+00 -7.60084033e-01 4.78845060e-01
2.98787713e-01 -1.27552307e+00 3.54029201e-02 -1.05176680e-02] | [6.571592330932617, -0.8496820330619812] |
cbd2cd9f-e6e1-45c9-a7ab-52b916143b90 | polyphonic-pitch-detection-with-convolutional | 2202.02115 | null | https://arxiv.org/abs/2202.02115v1 | https://arxiv.org/pdf/2202.02115v1.pdf | Polyphonic pitch detection with convolutional recurrent neural networks | Recent directions in automatic speech recognition (ASR) research have shown that applying deep learning models from image recognition challenges in computer vision is beneficial. As automatic music transcription (AMT) is superficially similar to ASR, in the sense that methods often rely on transforming spectrograms to symbolic sequences of events (e.g. words or notes), deep learning should benefit AMT as well. In this work, we outline an online polyphonic pitch detection system that streams audio to MIDI by ConvLSTMs. Our system achieves state-of-the-art results on the 2007 MIREX multi-F0 development set, with an F-measure of 83\% on the bassoon, clarinet, flute, horn and oboe ensemble recording without requiring any musical language modelling or assumptions of instrument timbre. | ['Sven Ahlbäck', 'Carl Thomé'] | 2022-02-04 | null | null | null | null | ['music-transcription'] | ['music'] | [ 2.91507691e-01 -2.10718185e-01 1.71112180e-01 7.09993392e-02
-1.06263220e+00 -6.86309695e-01 6.13286316e-01 -3.52506995e-01
-3.54277015e-01 2.02160150e-01 3.76236141e-01 -1.49484694e-01
-1.62617788e-02 -1.54411599e-01 -7.13460505e-01 -6.34870708e-01
1.17260721e-02 1.18542150e-01 -3.94600958e-01 -3.25406224e-01
1.06548414e-01 3.14657390e-01 -1.49855435e+00 6.26962006e-01
1.89523458e-01 1.10411501e+00 2.40748838e-01 1.41727376e+00
-1.98805220e-02 9.53335941e-01 -9.97381806e-01 -5.34938753e-01
1.96064606e-01 -6.60687327e-01 -7.28196800e-01 -5.03096171e-02
6.06025934e-01 -1.66843578e-01 -3.66729736e-01 1.00332212e+00
8.10734689e-01 -4.91648307e-03 6.54009044e-01 -6.74659252e-01
-5.54977536e-01 1.22219610e+00 -3.75113457e-01 1.44464657e-01
4.06266510e-01 4.78047915e-02 1.19506145e+00 -1.17375803e+00
3.04472476e-01 1.06581950e+00 9.42371964e-01 2.31611505e-01
-1.28905499e+00 -8.83063138e-01 -4.67870057e-01 2.40934208e-01
-1.40107477e+00 -8.10171068e-01 8.69701505e-01 -4.95839804e-01
1.15796733e+00 2.90775239e-01 7.64758766e-01 1.16671586e+00
-7.16682822e-02 9.98900473e-01 7.45770633e-01 -7.93633103e-01
-5.09408675e-02 -3.20571721e-01 -2.07373247e-01 2.31443077e-01
-6.33145511e-01 2.14413509e-01 -1.16820836e+00 1.59714401e-01
8.39984953e-01 -5.79141557e-01 -2.35048801e-01 3.33993852e-01
-1.44985771e+00 7.47593999e-01 -5.92815168e-02 4.08740103e-01
-4.80259091e-01 5.05044341e-01 8.49392951e-01 6.45810246e-01
2.52449930e-01 6.18806720e-01 -4.20883715e-01 -5.12357056e-01
-1.41369796e+00 1.71842769e-01 6.54443145e-01 6.56201243e-01
-1.30308615e-02 8.51980746e-01 3.26878428e-02 1.36474800e+00
7.76342452e-02 5.55044532e-01 8.06552172e-01 -1.22561967e+00
2.28249878e-01 -3.82301658e-01 -9.68404934e-02 -6.25452638e-01
-2.42217615e-01 -6.40941381e-01 -8.66185546e-01 2.36057311e-01
2.86943376e-01 -1.80109162e-02 -8.97148728e-01 1.62233067e+00
-3.81631345e-01 3.45781446e-01 1.32704467e-01 7.54826546e-01
8.23522806e-01 8.21655452e-01 -3.15175325e-01 -4.45394874e-01
1.19536483e+00 -8.00346315e-01 -1.00700462e+00 1.85703352e-01
8.80457610e-02 -1.27055681e+00 1.14253175e+00 1.09859252e+00
-1.41075635e+00 -1.05999935e+00 -1.02247655e+00 -1.74648985e-01
8.65092035e-03 5.33772111e-01 4.28206503e-01 7.44880438e-01
-9.91688728e-01 8.20916533e-01 -5.82247972e-01 -1.31059840e-01
5.47216870e-02 2.25000054e-01 9.25468374e-03 7.66195595e-01
-1.25990140e+00 5.07050753e-01 5.12689233e-01 -4.50368039e-02
-1.18762946e+00 -8.59446168e-01 -5.85730314e-01 1.00729235e-01
-2.24083718e-02 -3.66305739e-01 1.93668902e+00 -1.32347643e+00
-2.01819706e+00 9.07823980e-01 3.76538374e-02 -1.16125107e+00
3.25224966e-01 -6.77971780e-01 -7.16876447e-01 2.78378278e-01
-4.79601353e-01 7.45200872e-01 1.33288872e+00 -7.15612710e-01
-7.37683356e-01 1.22416295e-01 -4.33300346e-01 2.52944797e-01
-1.77073464e-01 4.04441774e-01 6.23032032e-03 -1.18114901e+00
4.91605438e-02 -9.29710329e-01 3.38123798e-01 -5.86666584e-01
-4.07568634e-01 -1.48385853e-01 5.28288960e-01 -1.09991312e+00
1.32789779e+00 -2.36642480e+00 1.31392241e-01 -1.16636097e-01
-2.09117174e-01 4.33395326e-01 -1.02401711e-01 3.41645718e-01
-3.67212564e-01 -1.31066859e-01 -6.33841082e-02 -3.12191308e-01
2.66649067e-01 -1.88225493e-01 -1.03101003e+00 2.20912680e-01
9.18813050e-03 7.10290194e-01 -6.17813945e-01 -1.04392432e-01
4.53607082e-01 6.49868011e-01 -3.12161416e-01 -7.36750104e-03
-8.20144042e-02 2.08243012e-01 5.73497236e-01 4.37885523e-01
9.89091992e-02 3.79307210e-01 -1.80970281e-01 -1.44001052e-01
-2.77272731e-01 8.63693893e-01 -1.13487935e+00 2.10141039e+00
-5.75782657e-01 1.03470206e+00 1.87599018e-01 -8.62051308e-01
9.68552887e-01 1.05533528e+00 5.31107008e-01 -3.21606576e-01
1.80542399e-03 5.56189537e-01 4.17795449e-01 -2.49342859e-01
4.13859665e-01 -4.37247276e-01 5.25835156e-02 2.40041047e-01
4.71764922e-01 -4.96591926e-01 3.72868255e-02 -2.51162410e-01
8.17303002e-01 5.77407181e-02 3.01362723e-01 1.01837426e-01
3.93496901e-01 -1.70722157e-01 2.57176399e-01 6.53746188e-01
-8.02554004e-03 1.05738437e+00 2.88368855e-02 -1.68903202e-01
-1.27707601e+00 -1.39951122e+00 -1.01932772e-01 1.31354117e+00
-9.48928714e-01 -4.74858791e-01 -7.81960011e-01 1.32884577e-01
-4.09965575e-01 6.94154203e-01 -1.06258549e-01 2.94678528e-02
-6.62704527e-01 -2.95446724e-01 1.22168541e+00 5.52238822e-01
1.75139681e-01 -1.64225233e+00 -3.64804387e-01 5.45613825e-01
-1.91808805e-01 -1.00411665e+00 -6.71526253e-01 4.17019814e-01
-5.36990762e-01 -5.06909728e-01 -1.27924609e+00 -9.34737325e-01
-4.85690236e-01 -2.33485848e-02 1.13668382e+00 -7.22271442e-01
-4.30251122e-01 3.63040328e-01 -3.38860482e-01 -7.47298539e-01
-7.02570558e-01 1.15304969e-01 4.51181620e-01 5.65433577e-02
2.59132117e-01 -8.82303834e-01 -2.94951707e-01 -1.34714052e-01
-6.58058763e-01 1.39825702e-01 4.68341440e-01 8.26030493e-01
7.64254570e-01 4.93897870e-02 1.04022872e+00 -4.75552350e-01
7.45283067e-01 1.04832262e-01 -4.49259102e-01 -2.91896135e-01
-2.15410948e-01 -5.34610093e-01 7.86439419e-01 -7.10492313e-01
-9.01804209e-01 2.23567754e-01 -3.46247792e-01 -9.60472763e-01
-2.40135372e-01 5.63254893e-01 4.51052487e-02 4.25633639e-01
7.79905617e-01 4.11030620e-01 -3.57601583e-01 -6.09875262e-01
4.81366992e-01 1.09926772e+00 1.23868597e+00 -3.06874305e-01
5.70572019e-01 2.45679006e-01 -3.08460027e-01 -1.32743144e+00
-8.45110476e-01 -5.91984391e-01 -5.36416233e-01 -2.29835331e-01
9.02345896e-01 -1.08955932e+00 -8.56809616e-01 6.71630502e-01
-1.22658205e+00 -2.78118908e-01 -5.62356889e-01 8.56593311e-01
-1.19299936e+00 9.14157182e-02 -1.12165606e+00 -1.01569343e+00
-6.96942508e-01 -8.60587597e-01 1.02194273e+00 -6.07801490e-02
-7.19841123e-01 -4.80534434e-01 3.84645283e-01 2.84852892e-01
3.41013312e-01 -1.38435170e-01 6.37806296e-01 -4.01936918e-01
-2.37792935e-02 -6.91015795e-02 3.46446335e-01 8.22087944e-01
7.44866952e-02 2.98946425e-02 -1.62091208e+00 -1.51501298e-01
1.61732912e-01 -4.42190766e-01 9.33924019e-01 7.25727797e-01
1.14892566e+00 -2.01340646e-01 5.14032602e-01 7.98537910e-01
8.44955921e-01 5.56603372e-01 7.19917119e-01 1.28084078e-01
5.61731219e-01 4.32768911e-01 2.88492024e-01 3.77368987e-01
-2.34552786e-01 7.15315998e-01 -8.84213001e-02 -7.95863196e-02
-5.97268164e-01 -3.09749901e-01 9.47737753e-01 1.51557231e+00
-6.54817596e-02 1.21920750e-01 -8.20887268e-01 7.41551816e-01
-1.43316221e+00 -1.27646565e+00 8.60496797e-03 2.20855951e+00
1.09315300e+00 1.07915111e-01 3.47071707e-01 9.31072533e-01
7.04465866e-01 1.66694909e-01 -3.81129235e-01 -8.00553203e-01
-2.97631681e-01 7.85885990e-01 1.91157907e-01 3.20692152e-01
-1.13292098e+00 9.31676388e-01 6.46932936e+00 1.19543910e+00
-1.31801999e+00 1.28840521e-01 3.30349982e-01 -6.09808803e-01
2.59539962e-01 -5.41562021e-01 -4.68508989e-01 4.20121290e-02
1.46096396e+00 -4.03841138e-02 1.05050600e+00 5.75847805e-01
4.02591676e-01 4.33442324e-01 -1.23757160e+00 1.52928293e+00
5.47841750e-02 -1.26909709e+00 -5.52735366e-02 -1.06501177e-01
6.41360104e-01 2.89789021e-01 5.49512506e-01 4.26600903e-01
1.29748685e-02 -1.35044289e+00 1.22411168e+00 3.33739817e-01
1.05566645e+00 -1.07135880e+00 3.44514370e-01 1.82949945e-01
-1.32312286e+00 9.72039253e-02 -1.74111247e-01 -5.91744073e-02
2.27181762e-02 3.05103421e-01 -1.05717862e+00 3.30187410e-01
7.90547490e-01 6.62742496e-01 -1.31004304e-01 9.75637197e-01
-1.09152980e-01 1.33104146e+00 -2.06984028e-01 5.30197978e-01
5.91118149e-02 -2.41115056e-02 7.44132698e-01 1.64522696e+00
4.33260143e-01 -1.53047010e-01 -1.04646996e-01 6.74566686e-01
-1.43391833e-01 1.55337021e-01 -4.86890137e-01 -5.30167460e-01
3.13465148e-01 1.05509329e+00 -4.68551487e-01 -1.92549214e-01
-1.77247182e-01 7.97089875e-01 -2.13050127e-01 4.00379241e-01
-7.14652896e-01 -5.97391188e-01 7.11868227e-01 -4.41125035e-02
3.72402668e-01 -3.25992197e-01 -3.09536546e-01 -7.78701425e-01
-1.87004715e-01 -1.18175650e+00 5.09993993e-02 -1.01525998e+00
-1.15483332e+00 5.63450575e-01 -5.76271176e-01 -1.42181253e+00
-6.45451665e-01 -6.95799828e-01 -5.18097162e-01 9.29183483e-01
-1.19534826e+00 -9.59934056e-01 4.81828779e-01 5.79153299e-01
1.01612341e+00 -6.17666781e-01 1.00164306e+00 4.13993150e-01
-1.17202848e-02 5.72537422e-01 8.33101422e-02 4.60348815e-01
6.93591118e-01 -1.41870177e+00 6.07560098e-01 6.04943275e-01
1.13875532e+00 4.98491734e-01 8.76526237e-01 -1.69888422e-01
-1.18906033e+00 -9.55436826e-01 7.59071052e-01 -1.73778266e-01
7.72089064e-01 -3.99884880e-01 -7.43239760e-01 4.91398692e-01
4.70061719e-01 -4.08806533e-01 6.74369872e-01 1.74812436e-01
-3.88978362e-01 -1.79971293e-01 -3.73344421e-01 6.79148912e-01
7.37309933e-01 -1.18582761e+00 -7.60728896e-01 1.45185709e-01
8.32386255e-01 -3.83254379e-01 -9.06530440e-01 2.98333108e-01
7.08118975e-01 -8.80366206e-01 1.12924016e+00 -4.64458406e-01
3.42869222e-01 -4.10685748e-01 -4.08532351e-01 -1.35217464e+00
-1.15758814e-01 -1.26836896e+00 -2.12061957e-01 1.32653463e+00
3.53284210e-01 8.26468617e-02 4.51148272e-01 -4.80760247e-01
-4.91780400e-01 4.25599329e-02 -9.84488964e-01 -1.08413637e+00
-1.08682431e-01 -1.16203427e+00 2.02967808e-01 8.87253642e-01
3.71673182e-02 5.58891416e-01 -6.55183315e-01 -5.14815226e-02
3.33101213e-01 -7.12410584e-02 7.09134638e-01 -1.29350579e+00
-6.58784509e-01 -8.64054143e-01 -2.48319179e-01 -6.40009284e-01
2.07456440e-01 -1.01127958e+00 1.16709583e-01 -1.05032384e+00
-2.20369458e-01 2.35287756e-01 -5.93157053e-01 2.41896003e-01
3.97860140e-01 3.48160774e-01 4.75288123e-01 1.30419299e-01
-2.82259881e-01 4.92722988e-01 8.70458841e-01 -3.25428247e-01
-2.11196423e-01 1.70512781e-01 -6.56849146e-02 1.09203470e+00
7.19435334e-01 -3.93366486e-01 -2.78869361e-01 -2.46850565e-01
2.36427277e-01 5.89177310e-01 2.11831093e-01 -1.37172556e+00
2.20625550e-01 3.75239819e-01 3.48886102e-01 -7.68540084e-01
8.12962890e-01 -3.28004926e-01 3.31811279e-01 3.39490235e-01
-9.13604558e-01 -5.56071028e-02 3.80601972e-01 2.60680646e-01
-6.78535998e-01 -2.47001782e-01 7.45485723e-01 -1.41903028e-01
-3.31499726e-01 -6.47267476e-02 -7.22084105e-01 3.54281105e-02
9.51063037e-02 -7.66193122e-02 1.38266413e-02 -6.88798010e-01
-9.24413800e-01 -6.38441861e-01 -7.44533911e-02 4.30139124e-01
4.93463695e-01 -1.42542100e+00 -1.04589522e+00 1.52476966e-01
-1.86940163e-01 -4.03053343e-01 2.16947302e-01 6.67458653e-01
-4.10951257e-01 6.35122240e-01 -6.60119206e-02 -6.09792709e-01
-1.38697898e+00 3.52632374e-01 4.66842741e-01 1.01034246e-01
-7.36168385e-01 1.07753384e+00 3.77819657e-01 -1.31800234e-01
7.32289791e-01 -4.06857342e-01 -2.24161491e-01 2.86157489e-01
6.12080693e-01 3.72839779e-01 3.06991398e-01 -5.03795028e-01
-6.91521242e-02 4.02230442e-01 5.74913099e-02 -8.12025785e-01
1.25684559e+00 1.64827630e-02 1.57406732e-01 1.22764087e+00
9.53141510e-01 2.18424886e-01 -9.00373995e-01 -6.03502325e-04
1.73007652e-01 -1.37512470e-02 3.17212045e-01 -8.72175872e-01
-8.84754539e-01 1.29983461e+00 6.82564855e-01 4.16484028e-01
1.23154497e+00 -2.88955241e-01 1.01290166e+00 5.43833435e-01
5.71368225e-02 -1.34871840e+00 2.07637250e-01 8.95098150e-01
1.19906354e+00 -8.60976696e-01 -5.72457492e-01 2.73202330e-01
-7.29214549e-01 1.45083725e+00 -1.46397993e-01 -3.06710511e-01
5.38586318e-01 4.19728875e-01 3.48901212e-01 4.18724418e-02
-7.26588488e-01 -2.07474634e-01 3.20039153e-01 3.70499104e-01
8.66077065e-01 2.15152293e-01 1.64386064e-01 7.17798173e-01
-1.09578145e+00 -6.56916127e-02 5.74874520e-01 2.37886637e-01
-4.67700034e-01 -5.54787755e-01 -7.65231252e-01 4.32019904e-02
-9.90906596e-01 -4.59606469e-01 -5.03280938e-01 4.41306829e-01
-1.80532616e-02 1.03855467e+00 -2.34212689e-02 -4.33560342e-01
1.60876095e-01 5.14381230e-01 5.08138597e-01 -5.99129081e-01
-9.51122940e-01 1.05710995e+00 4.64650290e-03 -1.53887257e-01
-2.29123250e-01 -6.34186327e-01 -1.21495378e+00 6.33015633e-02
-1.87368631e-01 -9.42237005e-02 8.63332331e-01 7.11748719e-01
-1.11161090e-01 9.69710648e-01 4.03634250e-01 -1.04811990e+00
-4.34997410e-01 -1.34252334e+00 -1.02507985e+00 2.13420570e-01
6.28167033e-01 -3.62333022e-02 -3.03479999e-01 6.25781655e-01] | [15.59126091003418, 5.556508541107178] |
abd25c4b-48e8-4460-be48-7dc7aa711568 | linear-disentangled-representation-learning | 1701.03102 | null | http://arxiv.org/abs/1701.03102v1 | http://arxiv.org/pdf/1701.03102v1.pdf | Linear Disentangled Representation Learning for Facial Actions | Limited annotated data available for the recognition of facial expression and
action units embarrasses the training of deep networks, which can learn
disentangled invariant features. However, a linear model with just several
parameters normally is not demanding in terms of training data. In this paper,
we propose an elegant linear model to untangle confounding factors in
challenging realistic multichannel signals such as 2D face videos. The simple
yet powerful model does not rely on huge training data and is natural for
recognizing facial actions without explicitly disentangling the identity. Base
on well-understood intuitive linear models such as Sparse Representation based
Classification (SRC), previous attempts require a prepossessing of explicit
decoupling which is practically inexact. Instead, we exploit the low-rank
property across frames to subtract the underlying neutral faces which are
modeled jointly with sparse representation on the action components with group
sparsity enforced. On the extended Cohn-Kanade dataset (CK+), our one-shot
automatic method on raw face videos performs as competitive as SRC applied on
manually prepared action components and performs even better than SRC in terms
of true positive rate. We apply the model to the even more challenging task of
facial action unit recognition, verified on the MPI Face Video Database
(MPI-VDB) achieving a decent performance. All the programs and data have been
made publicly available. | ['Trac. D. Tran', 'Xiang Xiang'] | 2017-01-11 | null | null | null | null | ['sparse-representation-based-classification', 'facial-action-unit-detection'] | ['computer-vision', 'computer-vision'] | [ 2.81236202e-01 1.29659876e-01 -1.29291281e-01 -5.12949824e-01
-6.70959353e-01 -3.94483507e-01 6.26342118e-01 -7.42844462e-01
-3.33701551e-01 7.10204422e-01 3.15970391e-01 2.14245617e-01
-1.55125886e-01 -5.65819740e-02 -7.09609807e-01 -1.15780246e+00
-2.35751107e-01 2.14034006e-01 -5.16241133e-01 -2.35818356e-01
-1.24859400e-01 5.99999249e-01 -1.75283468e+00 5.68361521e-01
2.58905590e-01 1.07040024e+00 -3.75569880e-01 5.20464778e-01
2.99872875e-01 8.83728683e-01 -1.82217807e-01 -5.54790556e-01
5.31817138e-01 -5.57198465e-01 -4.46530312e-01 2.24029988e-01
9.38805461e-01 -4.05506462e-01 -5.60066640e-01 9.56723809e-01
5.26964128e-01 -1.33460224e-01 5.86026251e-01 -1.35589635e+00
-4.81897652e-01 3.68037969e-01 -6.05646431e-01 9.63720530e-02
2.48856544e-01 4.41689007e-02 9.83189642e-01 -1.28882682e+00
5.87830007e-01 1.25929701e+00 6.68468237e-01 6.99030817e-01
-1.63312697e+00 -9.51087952e-01 -9.03462246e-02 4.13819492e-01
-1.59674072e+00 -1.19569170e+00 7.15240359e-01 -4.81043220e-01
5.91507673e-01 3.62706602e-01 5.14401138e-01 1.59849441e+00
-3.41426939e-01 6.52328014e-01 1.17745018e+00 -2.47399688e-01
-6.00712895e-02 5.22481576e-02 -4.28448617e-02 7.30720520e-01
1.87634617e-01 1.95636764e-01 -9.55006778e-01 -1.79016754e-01
5.86625338e-01 -1.70147023e-03 -4.96768504e-01 -5.87351263e-01
-1.24039996e+00 8.08623135e-01 1.64804175e-01 4.02785450e-01
-3.75410140e-01 3.56316380e-03 3.48300695e-01 4.70530629e-01
4.39822942e-01 2.69104689e-01 -4.00629073e-01 -5.54111637e-02
-1.06126630e+00 -2.34496705e-02 8.09896827e-01 5.42355120e-01
8.40762854e-01 4.20907587e-01 1.15129121e-01 7.56730616e-01
2.05863386e-01 6.00815833e-01 3.99087578e-01 -1.21453047e+00
2.54401732e-02 2.26620913e-01 -1.64241433e-01 -1.30124533e+00
-3.10586333e-01 -2.37004876e-01 -1.17214298e+00 4.60811377e-01
6.42798483e-01 -5.20393737e-02 -7.12498784e-01 2.06763148e+00
1.94330662e-01 6.22923434e-01 1.27067372e-01 1.04545498e+00
7.56295562e-01 3.38119268e-01 -2.34292492e-01 -4.41887110e-01
1.34674120e+00 -5.38295627e-01 -8.13443363e-01 -1.99020859e-02
4.83113378e-01 -6.11893535e-01 6.37036443e-01 6.99825704e-01
-8.58538091e-01 -3.33847433e-01 -9.73946691e-01 -2.02692822e-01
-4.97312145e-03 4.02575314e-01 8.19383085e-01 5.41952550e-01
-9.55847085e-01 4.68738258e-01 -8.57356906e-01 -2.06886053e-01
6.09358788e-01 6.66660547e-01 -1.24703813e+00 -2.66637236e-01
-9.14271653e-01 6.59089684e-01 -6.72155991e-02 5.69738030e-01
-1.01581800e+00 -6.75146163e-01 -9.52760339e-01 5.41225076e-02
5.08312285e-01 -3.80177200e-01 7.67890275e-01 -1.54720545e+00
-1.66714287e+00 1.03149748e+00 -2.48930931e-01 -2.03781739e-01
3.57889205e-01 -1.24018475e-01 -2.92727172e-01 2.63341755e-01
-2.52253950e-01 5.96317410e-01 1.39296579e+00 -1.10122991e+00
1.89165592e-01 -6.82040811e-01 -3.09176389e-02 -8.61788020e-02
-4.36435401e-01 4.81175035e-02 -1.13549143e-01 -6.54014945e-01
6.94637150e-02 -1.07533145e+00 9.58152935e-02 3.34879696e-01
-1.87254295e-01 2.77088970e-01 6.83397651e-01 -7.24627614e-01
6.93008900e-01 -2.57967520e+00 6.54449940e-01 8.88466462e-02
3.05563003e-01 2.11599499e-01 -6.13047004e-01 1.16825856e-01
-5.98226964e-01 -2.72755265e-01 -4.72785905e-02 -5.86239040e-01
-5.82683012e-02 2.57255465e-01 -2.59502798e-01 9.74832475e-01
3.72421950e-01 7.38496125e-01 -5.09683549e-01 -1.84179202e-01
-6.29105270e-02 8.28875184e-01 -8.56613219e-01 2.21673518e-01
2.22621784e-01 5.91354191e-01 -5.72075807e-02 5.93106806e-01
7.33407259e-01 -1.90686196e-01 4.44007725e-01 -5.87342203e-01
1.39429823e-01 -9.59242657e-02 -1.14476740e+00 1.85941923e+00
-2.73698568e-01 6.69049740e-01 6.57534420e-01 -1.38088858e+00
6.78074241e-01 6.09000206e-01 6.61595464e-01 -3.90329361e-01
1.96872666e-01 1.94429129e-01 1.50796354e-01 -5.20254433e-01
-1.12237252e-01 -5.00631392e-01 1.44993126e-01 1.55341312e-01
7.23296046e-01 1.92162126e-01 -1.08615138e-01 1.52579263e-01
9.92010415e-01 1.46125451e-01 2.62450904e-01 -3.80584866e-01
7.07508862e-01 -4.99780685e-01 6.82690859e-01 2.87413031e-01
-1.18602224e-01 6.33203387e-01 8.81644905e-01 -3.65093708e-01
-7.07879722e-01 -7.83555150e-01 -9.17513669e-02 1.01278591e+00
-3.78643781e-01 -5.05134881e-01 -5.94566464e-01 -4.97474760e-01
-9.64669436e-02 2.01590076e-01 -9.53799069e-01 -9.52093378e-02
-4.07348186e-01 -7.67590284e-01 6.52647674e-01 1.55996934e-01
2.15853736e-01 -6.36272430e-01 1.87032055e-02 -3.57477725e-01
7.30220228e-03 -1.43165088e+00 -2.98097491e-01 2.83896476e-02
-3.96633714e-01 -1.23518860e+00 -6.46852016e-01 -3.72773498e-01
6.54568136e-01 2.13487223e-01 7.34636366e-01 -1.08645417e-01
-4.44357127e-01 4.20018196e-01 -9.45407748e-02 -3.40222977e-02
-9.16590318e-02 -3.68734062e-01 3.96235436e-01 7.78409541e-01
4.07531530e-01 -9.03084338e-01 -3.77758384e-01 3.24584275e-01
-7.16597378e-01 -2.13935878e-02 6.80006862e-01 1.05621028e+00
2.41906583e-01 -4.53021854e-01 4.23667043e-01 -6.99274302e-01
-4.33069319e-02 -4.88098145e-01 -4.79364187e-01 1.36521934e-02
-1.06503442e-01 1.16267003e-01 5.22461295e-01 -6.48129344e-01
-9.46034610e-01 4.32241321e-01 -2.33430192e-02 -9.74189043e-01
-1.85155466e-01 1.99698210e-01 -3.97049963e-01 -3.20874006e-01
5.20948052e-01 2.10021496e-01 4.79979008e-01 -5.24545193e-01
3.46646935e-01 3.39582562e-01 4.95606929e-01 -3.48427117e-01
8.19622159e-01 7.49719441e-01 2.77376741e-01 -1.11693525e+00
-8.48466516e-01 -2.65781343e-01 -8.83411646e-01 -7.27260187e-02
9.59906697e-01 -1.33138084e+00 -9.87766922e-01 4.55877632e-01
-9.52744246e-01 -7.85696283e-02 -7.37516955e-02 7.29048312e-01
-6.27731800e-01 5.08085310e-01 -4.97263581e-01 -6.04358375e-01
8.43960568e-02 -1.02004528e+00 1.14414465e+00 -3.16659153e-01
-2.16931820e-01 -5.92373013e-01 1.02011852e-01 5.61125815e-01
2.82031983e-01 3.02130669e-01 6.17032111e-01 -6.29856229e-01
-5.12709498e-01 -2.54363507e-01 -1.14503659e-01 7.25279748e-01
1.35325447e-01 -1.13303013e-01 -1.37809300e+00 -2.50631094e-01
3.26625735e-01 -6.98086858e-01 8.89013946e-01 1.90212540e-02
1.16703784e+00 -5.21940768e-01 1.56997144e-01 8.63765061e-01
1.01836312e+00 -3.99811298e-01 5.46581805e-01 -4.11110759e-01
8.92714560e-01 9.59765971e-01 4.08656299e-02 4.46984202e-01
2.96485741e-02 9.02392030e-01 3.25673401e-01 -3.33628245e-02
-7.57052079e-02 -7.98308775e-02 7.14689732e-01 7.91614711e-01
-4.76412147e-01 1.95260733e-01 -3.96946311e-01 7.79884309e-02
-1.80159819e+00 -1.30902374e+00 -1.25625851e-02 2.11780000e+00
7.32210159e-01 -4.17141527e-01 4.40625586e-02 1.06428072e-01
2.89584875e-01 2.61261344e-01 -3.54011327e-01 -2.00510379e-02
-4.45252150e-01 2.94569880e-01 1.01266950e-01 6.05775714e-01
-9.91402745e-01 7.61552632e-01 5.92391682e+00 6.68464422e-01
-1.30075300e+00 2.87984312e-01 5.42231083e-01 -6.01926684e-01
7.90864825e-02 -1.76710650e-01 -5.77849567e-01 2.43219391e-01
1.05410910e+00 5.93078583e-02 6.53323352e-01 7.05882072e-01
2.38340750e-01 1.54636458e-01 -1.42829680e+00 1.47383273e+00
4.92433667e-01 -1.16731799e+00 -1.44574642e-02 2.57839650e-01
4.09202546e-01 -6.90554082e-02 1.68785512e-01 2.90751427e-01
-5.86907119e-02 -1.33794534e+00 5.18125415e-01 5.76251090e-01
8.35947394e-01 -4.85946327e-01 5.08139908e-01 3.07297170e-01
-6.75969899e-01 1.91607140e-02 -3.54255319e-01 -1.74214348e-01
-9.58359390e-02 4.43965971e-01 -3.39864969e-01 3.55028540e-01
5.55649221e-01 1.15292633e+00 -4.01847243e-01 2.13532060e-01
-1.46853164e-01 5.24048209e-01 -3.47847253e-01 6.09939635e-01
-6.38854504e-02 -3.78084838e-01 4.07000750e-01 9.76925254e-01
3.28785241e-01 5.18724859e-01 -2.38502890e-01 7.15081513e-01
-1.03622727e-01 9.45234522e-02 -7.91736841e-01 -2.35725820e-01
-2.85557181e-01 1.53144479e+00 -2.36288697e-01 -9.00106579e-02
-7.14099050e-01 1.14543116e+00 4.06207949e-01 4.62869376e-01
-6.84276521e-01 3.47487658e-01 1.20685756e+00 -5.07985987e-02
4.35492337e-01 -3.37232530e-01 2.24995136e-01 -1.81584895e+00
-1.70871541e-02 -1.32326436e+00 1.27157927e-01 -5.09864032e-01
-1.39673841e+00 6.35762393e-01 -1.27830356e-01 -1.10150909e+00
-3.32880437e-01 -9.04139042e-01 -2.69798458e-01 6.79505289e-01
-1.27703762e+00 -1.18392110e+00 -3.34692776e-01 1.04766977e+00
1.46239087e-01 -1.82540610e-01 1.11385357e+00 6.15855813e-01
-8.03758562e-01 6.68559134e-01 -1.75460011e-01 1.16136089e-01
8.79042745e-01 -7.11507380e-01 -4.69026774e-01 5.97481191e-01
5.78739285e-01 6.17711902e-01 6.40072167e-01 -3.98938134e-02
-1.75708711e+00 -5.97144961e-01 7.34395206e-01 -4.75948572e-01
7.81707168e-01 -9.70459461e-01 -9.50835407e-01 7.52468705e-01
6.85411915e-02 5.57950795e-01 1.11215019e+00 1.90178782e-01
-7.54504740e-01 -2.24874198e-01 -8.10057044e-01 3.51209402e-01
1.19484222e+00 -8.50433230e-01 -3.94264013e-01 4.08373863e-01
2.49439418e-01 4.06566970e-02 -7.68917084e-01 3.49245906e-01
8.05702746e-01 -1.30928886e+00 9.58138704e-01 -9.59448516e-01
2.50682145e-01 -9.27180871e-02 -5.67531765e-01 -8.67742121e-01
-1.52305156e-01 -7.84922481e-01 -1.61858782e-01 1.07192302e+00
2.59750307e-01 -5.22385895e-01 7.41540909e-01 6.19173288e-01
3.21076512e-01 -5.76810181e-01 -1.19934177e+00 -6.40475571e-01
-2.77386278e-01 -3.68746787e-01 3.69925983e-02 1.24456322e+00
2.04276666e-02 5.03950715e-01 -9.59161460e-01 2.00569883e-01
6.21402800e-01 2.14419626e-02 9.25032735e-01 -1.14127767e+00
-5.29915094e-01 -1.11031868e-01 -7.37031579e-01 -7.38352537e-01
8.31929803e-01 -9.34431911e-01 -2.42597416e-01 -4.09366369e-01
2.83400655e-01 2.45156474e-02 -2.54970104e-01 7.21009552e-01
3.46511275e-01 7.04863071e-01 2.44710058e-01 2.20830843e-01
-3.43083948e-01 8.43999982e-01 8.32289219e-01 -1.37241468e-01
3.11550051e-01 -2.04049334e-01 -6.66370451e-01 8.77384424e-01
2.50169456e-01 -3.92833292e-01 -3.23476791e-01 -3.49630326e-01
7.15353936e-02 1.57857597e-01 6.23580873e-01 -6.82350039e-01
3.28505971e-02 1.11383736e-01 4.72489685e-01 1.82921782e-01
9.05029714e-01 -1.00717306e+00 3.85728806e-01 1.76863059e-01
-2.80833155e-01 -2.77936220e-01 1.10764056e-01 5.77120602e-01
-4.36993539e-01 1.09184638e-01 7.81821311e-01 8.37531239e-02
-4.70586538e-01 4.98341888e-01 -1.51548564e-01 -2.31291518e-01
9.86822426e-01 -1.05555534e-01 2.02767886e-02 -5.55419385e-01
-1.19498050e+00 -3.53431165e-01 3.77673268e-01 3.66233975e-01
5.44460177e-01 -1.36183131e+00 -8.35421145e-01 7.96275079e-01
-3.73628438e-02 -6.11386895e-01 4.92890239e-01 1.35587454e+00
-4.95674014e-02 3.74796659e-01 -4.10873681e-01 -4.80344325e-01
-1.52758706e+00 6.47032082e-01 3.67284030e-01 1.49453640e-01
-3.78340274e-01 8.26732457e-01 6.02399409e-01 -2.26471648e-01
4.00939360e-02 6.70018122e-02 -1.53836101e-01 4.98071760e-01
6.62453413e-01 1.72490373e-01 -5.70262410e-02 -1.16519153e+00
-5.25942862e-01 6.71744108e-01 5.54717965e-02 -2.33243760e-02
1.64083612e+00 1.22315504e-01 -4.22274113e-01 3.88542473e-01
1.73660016e+00 1.00828402e-01 -1.37070775e+00 -1.24274142e-01
-1.99297756e-01 -7.04519510e-01 1.10538490e-01 -3.08415234e-01
-1.29481995e+00 1.03753710e+00 4.73395258e-01 -3.51880997e-01
1.10875773e+00 -1.28031913e-02 1.21544890e-01 5.20547748e-01
3.27568710e-01 -4.57225353e-01 2.57341146e-01 2.36565217e-01
1.17882431e+00 -1.45405841e+00 2.13313438e-02 -5.13061762e-01
-5.63638568e-01 1.09496093e+00 3.85669142e-01 -2.68675238e-01
8.59505594e-01 3.32961351e-01 2.49191877e-02 -2.74098426e-01
-9.06349957e-01 -3.13898325e-02 4.42474365e-01 3.37106794e-01
3.37952375e-01 -2.39212602e-01 -1.99341439e-02 6.31442726e-01
-1.86933130e-02 1.66857056e-02 3.93343359e-01 3.47721487e-01
1.38711259e-01 -9.41423774e-01 -1.75783917e-01 1.14572741e-01
-6.44463301e-01 -4.53079827e-02 -4.82791543e-01 7.77336121e-01
6.84655979e-02 6.10763848e-01 -6.58745915e-02 -2.97473401e-01
7.38307834e-02 3.09091866e-01 7.20384598e-01 -5.28715491e-01
-1.34640306e-01 1.13841631e-01 1.11731648e-01 -1.12346566e+00
-8.26285481e-01 -9.21887398e-01 -7.45633304e-01 -3.18866670e-01
-1.65116668e-01 -3.96202132e-02 4.88814473e-01 8.61756682e-01
4.64035153e-01 -1.41359732e-01 5.76517940e-01 -1.32436812e+00
-6.07634723e-01 -8.91190588e-01 -8.42681646e-01 5.31247854e-01
6.00670695e-01 -8.49952817e-01 -7.51485288e-01 1.59606710e-01] | [13.192313194274902, 0.5150875449180603] |
cfced395-477d-4ffc-8daa-9dce5ff2323b | cosmopower-jax-high-dimensional-bayesian | 2305.06347 | null | https://arxiv.org/abs/2305.06347v2 | https://arxiv.org/pdf/2305.06347v2.pdf | CosmoPower-JAX: high-dimensional Bayesian inference with differentiable cosmological emulators | We present CosmoPower-JAX, a JAX-based implementation of the CosmoPower framework, which accelerates cosmological inference by building neural emulators of cosmological power spectra. We show how, using the automatic differentiation, batch evaluation and just-in-time compilation features of JAX, and running the inference pipeline on graphics processing units (GPUs), parameter estimation can be accelerated by orders of magnitude with advanced gradient-based sampling techniques. These can be used to efficiently explore high-dimensional parameter spaces, such as those needed for the analysis of next-generation cosmological surveys. We showcase the accuracy and computational efficiency of CosmoPower-JAX on two simulated Stage IV configurations. We first consider a single survey performing a cosmic shear analysis totalling 37 model parameters. We validate the contours derived with CosmoPower-JAX and a Hamiltonian Monte Carlo sampler against those derived with a nested sampler and without emulators, obtaining a speed-up factor of $\mathcal{O}(10^3)$. We then consider a combination of three Stage IV surveys, each performing a joint cosmic shear and galaxy clustering (3x2pt) analysis, for a total of 157 model parameters. Even with such a high-dimensional parameter space, CosmoPower-JAX provides converged posterior contours in 3 days, as opposed to the estimated 6 years required by standard methods. CosmoPower-JAX is fully written in Python, and we make it publicly available to help the cosmological community meet the accuracy requirements set by next-generation surveys. | ['A. Spurio Mancini', 'D. Piras'] | 2023-05-10 | null | null | null | null | ['bayesian-inference'] | ['methodology'] | [-3.86592984e-01 -3.25927168e-01 4.65782493e-01 -2.30597794e-01
-7.40488648e-01 -7.49526203e-01 1.03263736e+00 -1.53486148e-01
-6.10653341e-01 7.04340577e-01 -6.16055615e-02 -7.75099158e-01
-7.85786361e-02 -1.09917533e+00 -2.84865409e-01 -8.64347816e-01
-2.83335924e-01 9.81519163e-01 6.23186707e-01 5.99599481e-02
5.11212409e-01 1.88806474e-01 -1.47988462e+00 -4.55818087e-01
7.46351123e-01 7.60642231e-01 -1.37676317e-02 1.16878700e+00
2.99603581e-01 -9.10028741e-02 -3.46081793e-01 -5.42253137e-01
2.79214650e-01 -4.74377215e-01 -3.09726804e-01 -3.09671462e-01
2.31230870e-01 -2.78779089e-01 -2.06701353e-01 1.07710779e+00
6.54502094e-01 5.14444888e-01 6.14670277e-01 -5.08254111e-01
-1.02975639e-02 3.27699721e-01 -3.44431609e-01 3.41477931e-01
-2.77876467e-01 6.42768085e-01 7.52210379e-01 -6.99310780e-01
3.71905714e-01 1.23997390e+00 8.64141166e-01 5.65745309e-02
-1.65734124e+00 -5.40748239e-01 -5.20488143e-01 -3.95706356e-01
-1.40798104e+00 -2.72509426e-01 3.24518144e-01 -6.69354618e-01
1.27301538e+00 6.14503860e-01 8.83938193e-01 8.90713394e-01
1.29342631e-01 -2.89425999e-01 1.11774838e+00 -3.18018019e-01
9.30870354e-01 -1.04134209e-01 1.28419325e-01 5.63033104e-01
3.08165044e-01 4.39315319e-01 -2.05570027e-01 -7.71551907e-01
1.08259952e+00 -7.61509061e-01 1.80789679e-01 1.64962336e-01
-1.16371465e+00 1.12947774e+00 -9.96690169e-02 -1.14623182e-01
1.59540519e-01 5.58386207e-01 4.09260422e-01 -3.00367534e-01
7.74579525e-01 5.89491665e-01 -3.84428471e-01 -5.06454110e-01
-1.24984884e+00 7.92477012e-01 8.63401413e-01 4.30114061e-01
6.78106070e-01 2.99340308e-01 5.10988355e-01 8.34173083e-01
5.03186285e-01 8.49250317e-01 3.54887784e-01 -1.59277558e+00
-1.84169888e-01 -6.98230937e-02 1.39667436e-01 -6.09690785e-01
-5.47576785e-01 -5.40468454e-01 -8.94641936e-01 5.37816465e-01
5.48218906e-01 -4.58654344e-01 -3.79131854e-01 1.39117241e+00
6.22041702e-01 9.58478451e-02 -3.86678427e-01 7.27615952e-01
2.03528002e-01 7.86288500e-01 -1.55332893e-01 -1.01574562e-01
1.49472260e+00 -6.68523908e-01 1.23791091e-01 -2.28686631e-01
4.84267622e-01 -9.23974156e-01 8.42324734e-01 4.52013344e-01
-1.33571327e+00 -4.81801540e-01 -1.11891663e+00 7.32256398e-02
-1.61787182e-01 -1.19542427e-01 9.33737755e-01 1.07150757e+00
-1.25233424e+00 8.41386437e-01 -1.42523038e+00 -9.52132270e-02
-1.15788460e-01 2.02015460e-01 5.24804711e-01 6.88802958e-01
-8.41060281e-01 5.82352459e-01 3.36563796e-01 -2.50697911e-01
-7.18073368e-01 -9.83565152e-01 -5.08618891e-01 1.12529114e-01
5.10948934e-02 -7.89096475e-01 1.35867774e+00 -1.67758420e-01
-1.59143114e+00 5.15379727e-01 6.10543974e-02 -4.76730168e-01
5.08368552e-01 2.26771772e-01 -2.69878000e-01 1.29776755e-02
-7.35919327e-02 1.97595686e-01 2.11437002e-01 -7.61509061e-01
-6.73459694e-02 -2.26400807e-01 -3.21723372e-01 -2.40611300e-01
1.81792006e-01 1.92012370e-01 -6.60125732e-01 -3.74991357e-01
1.54588193e-01 -9.96990144e-01 -5.55680215e-01 -1.08453631e-01
-2.71427482e-01 1.22424908e-01 8.86338428e-02 -9.09290552e-01
9.13617074e-01 -1.73852658e+00 -1.11824758e-01 5.13236880e-01
3.09660155e-02 1.61753912e-02 4.60544825e-01 2.87096858e-01
1.66716293e-01 1.09225996e-01 -6.01306021e-01 -5.81155300e-01
3.71213406e-01 -7.46295154e-02 3.90860252e-03 4.87945169e-01
-2.74286628e-01 3.92368615e-01 -6.71241283e-01 -2.04498008e-01
3.98760051e-01 5.67143798e-01 -1.03909993e+00 -1.29325137e-01
-4.48621809e-01 5.40810287e-01 -2.33622655e-01 1.94030419e-01
8.31831634e-01 -4.37162042e-01 2.92190313e-01 1.08475544e-01
-6.72065616e-01 3.65417540e-01 -1.27396083e+00 1.52088368e+00
-5.19098818e-01 6.03384376e-01 3.07061374e-01 -4.66517687e-01
8.62890005e-01 -1.06953926e-01 -5.79409786e-02 -5.11326492e-01
2.60858595e-01 4.84045297e-01 7.32649192e-02 7.58577809e-02
6.63443148e-01 -2.47018784e-01 7.01855496e-02 6.91944122e-01
1.40409008e-01 -6.48870409e-01 3.70431572e-01 2.22498447e-01
8.89174819e-01 2.55725533e-01 -6.05166592e-02 -9.18308020e-01
2.31256023e-01 -9.20537487e-02 1.79497495e-01 8.40296507e-01
2.38277286e-01 5.18495202e-01 6.37950838e-01 -3.11393172e-01
-1.95222807e+00 -1.11810482e+00 -7.77313650e-01 7.74974287e-01
-3.43323588e-01 -5.32565832e-01 -8.74783754e-01 1.94299757e-01
-4.27166410e-02 9.89186645e-01 -3.79355878e-01 3.26118201e-01
-4.54358071e-01 -1.88286078e+00 5.03536642e-01 1.97028033e-02
5.80070496e-01 -4.89118606e-01 -5.96054792e-01 4.51498702e-02
3.77872556e-01 -6.61296308e-01 -9.89889950e-02 -3.00528817e-02
-7.20060110e-01 -5.96533000e-01 -5.58848441e-01 -1.79212987e-01
3.79183233e-01 -4.86436009e-01 1.25155115e+00 1.63379654e-01
-4.68987465e-01 6.57047108e-02 1.23444133e-01 1.30944520e-01
-5.73719144e-01 -1.26301616e-01 1.19328894e-01 -4.54653352e-01
7.16194063e-02 -1.05852544e+00 -1.06154275e+00 2.13374659e-01
-4.68296766e-01 1.51865557e-01 -1.75194860e-01 6.50845408e-01
1.59364671e-01 1.93482459e-01 1.48879141e-01 -4.14677769e-01
2.38697127e-01 -5.65814614e-01 -1.79316843e+00 -2.13734090e-01
-5.54209173e-01 2.14293763e-01 4.85929668e-01 -2.27910712e-01
-1.20554733e+00 -4.09657091e-01 -8.32812130e-01 1.79875314e-01
1.92631617e-01 3.67220670e-01 3.51485848e-01 -1.97995320e-01
7.65953898e-01 1.12444103e-01 -9.86937899e-03 -8.27275932e-01
2.08503261e-01 3.39545429e-01 7.09720075e-01 -1.04498410e+00
7.62736022e-01 4.38272029e-01 1.94744423e-01 -8.77726853e-01
-3.33307326e-01 -3.14675421e-02 -3.19691300e-01 -1.47125110e-01
1.12583268e+00 -7.74858296e-01 -1.08168375e+00 5.04541397e-01
-7.40187526e-01 -6.57174110e-01 1.52890580e-02 7.81038284e-01
-3.30557317e-01 4.82643902e-01 -6.29423678e-01 -1.23294008e+00
-2.51632452e-01 -1.27266765e+00 8.81203830e-01 3.57316613e-01
-1.05413899e-01 -1.36579812e+00 5.56424379e-01 1.51938602e-01
5.66318274e-01 1.73425063e-01 9.25850987e-01 -4.44434255e-01
-6.38060391e-01 -5.53881600e-02 -1.90263003e-01 -8.98117945e-02
-6.99448645e-01 3.36147636e-01 -1.03806949e+00 -3.39520067e-01
2.36356616e-01 -1.24749944e-01 8.30733001e-01 6.96913004e-01
1.13321519e+00 4.10648212e-02 -2.51913760e-02 9.40004110e-01
1.59793913e+00 1.21299438e-02 4.53643173e-01 2.15931028e-01
1.56935558e-01 3.66409540e-01 -1.48494557e-01 6.44165158e-01
1.95337221e-01 6.33144259e-01 1.09693132e-01 6.13619052e-02
1.44244477e-01 2.11323321e-01 2.98906527e-02 1.26054585e+00
-2.36053422e-01 2.35330299e-01 -1.13327396e+00 3.44286799e-01
-1.43133366e+00 -8.46124768e-01 -4.31406349e-01 2.68851590e+00
9.98788178e-01 3.70147079e-01 1.85919821e-01 -5.04932761e-01
4.16489422e-01 1.61694005e-01 -5.32040775e-01 -3.92864197e-01
8.88342932e-02 6.61002100e-01 5.73994458e-01 1.13728917e+00
-6.52670443e-01 4.15005952e-01 6.80661440e+00 9.19074118e-01
-7.56042778e-01 3.90138447e-01 9.53833103e-01 -3.24395388e-01
-4.34935272e-01 5.38303077e-01 -8.73822570e-01 7.92783201e-01
1.58911204e+00 -2.84390450e-01 7.05675125e-01 9.23359931e-01
3.10080558e-01 -7.56621242e-01 -6.79037273e-01 7.06699789e-01
-4.15954977e-01 -1.68106270e+00 -6.67340040e-01 4.04683948e-01
7.10570097e-01 6.27129078e-01 -1.82593316e-01 3.79495412e-01
7.65655696e-01 -7.75580585e-01 3.88497382e-01 3.38497460e-01
8.02749753e-01 -8.94764960e-01 4.93832946e-01 3.36589992e-01
-8.47045302e-01 4.91948664e-01 -8.12201202e-01 -2.16346264e-01
5.26542664e-01 1.10289705e+00 -5.44805527e-01 4.27027643e-01
6.76585793e-01 -2.43487149e-01 -6.03329957e-01 1.03532279e+00
4.72791165e-01 1.02329278e+00 -9.17458236e-01 -2.70691335e-01
8.67334455e-02 -8.05942953e-01 5.09042203e-01 1.26628137e+00
6.81064665e-01 4.06416170e-02 -1.58467814e-01 1.37747562e+00
3.80515099e-01 -3.75866383e-01 1.30081326e-01 1.95949286e-01
5.75428367e-01 1.27257526e+00 -1.13416016e+00 -7.01527059e-01
-2.77504921e-01 3.74357671e-01 -6.44309223e-02 2.69381404e-01
-8.08812857e-01 -3.29125106e-01 5.72579205e-01 7.77408481e-02
4.23618764e-01 -8.56537044e-01 -4.73906010e-01 -1.25879741e+00
-4.96752471e-01 -5.87636352e-01 -2.04868764e-02 -8.09429348e-01
-1.17738712e+00 2.40249723e-01 4.29054111e-01 -3.18594962e-01
-4.00404990e-01 -8.43767941e-01 -9.23611522e-01 1.24387658e+00
-7.41854727e-01 -5.65979183e-01 -9.38091427e-02 -1.50330365e-01
3.89210641e-01 -1.51219368e-01 8.24268878e-01 8.12525600e-02
-6.13440037e-01 1.45373955e-01 7.88733959e-01 -1.75485834e-01
1.01979643e-01 -1.41896760e+00 1.25777698e+00 8.63005459e-01
-3.30159485e-01 7.59202421e-01 1.18414414e+00 -1.02400541e+00
-1.27343953e+00 -6.65126681e-01 4.06357527e-01 -2.58527398e-01
1.14867234e+00 -7.40378857e-01 -8.87065053e-01 4.69111621e-01
3.26584697e-01 -8.98891985e-02 6.05771184e-01 3.93282205e-01
-1.57965392e-01 2.88540691e-01 -9.38553452e-01 7.68389940e-01
6.97387397e-01 -3.35723847e-01 -2.00253159e-01 5.93485832e-01
5.01658261e-01 -2.09063321e-01 -1.05627370e+00 8.14292207e-03
6.69667363e-01 -1.26794171e+00 1.22484910e+00 2.02334568e-01
6.81663165e-03 -3.03471386e-01 -2.13906825e-01 -1.07454312e+00
-2.03229204e-01 -9.13698971e-01 3.01907182e-01 9.85169411e-01
2.45693594e-01 -9.56935823e-01 5.93623459e-01 5.88742673e-01
-2.36937357e-03 -1.83528170e-01 -1.08678007e+00 -6.24815285e-01
6.35771871e-01 -6.59864008e-01 5.40187240e-01 6.12831771e-01
-4.32324797e-01 1.22568928e-01 -7.21397623e-02 3.06423753e-01
1.24629283e+00 -1.69736803e-01 6.94345832e-01 -1.10827267e+00
-1.12970686e+00 -6.12162709e-01 1.83865782e-02 -8.35081279e-01
-2.80229121e-01 -7.83011317e-01 -1.46554038e-01 -7.92186797e-01
2.64302790e-01 -4.39910412e-01 5.07465601e-01 -1.92622721e-01
3.97124514e-02 4.67938483e-01 -1.91499546e-01 -4.98975366e-02
-7.73983374e-02 4.27319646e-01 6.49159312e-01 4.60358620e-01
1.51547119e-01 -1.75057217e-01 -6.20315410e-02 9.59711611e-01
7.04339981e-01 -1.94391683e-01 -2.35221326e-01 -2.19848052e-01
7.16146946e-01 1.63228303e-01 7.08150148e-01 -1.19652379e+00
6.94851950e-02 -1.15463248e-04 7.40752041e-01 -7.84265280e-01
5.38452029e-01 2.26574808e-01 5.64107299e-01 5.14949024e-01
1.34907126e-01 9.05725509e-02 4.23830688e-01 3.48348990e-02
4.72975045e-01 -6.84188962e-01 1.05440497e+00 -4.42852408e-01
-6.61694407e-02 -1.41820088e-01 -7.34947264e-01 1.06247164e-01
3.60031515e-01 4.18828905e-01 -4.59762186e-01 -2.47746959e-01
-6.63794816e-01 -7.02703670e-02 8.05276453e-01 -5.24645090e-01
-7.11402297e-02 -1.10999382e+00 -7.51088023e-01 2.90954411e-01
-5.18166065e-01 -4.12544049e-02 4.86832231e-01 7.24900842e-01
-1.17982614e+00 6.42895848e-02 1.70785025e-01 -6.29735708e-01
-7.40480959e-01 2.36296088e-01 5.18283129e-01 -2.00440437e-01
-6.11968935e-01 8.53831232e-01 2.39692643e-01 -7.21227884e-01
-3.48452896e-01 -3.45357805e-02 4.87541765e-01 -2.88598090e-01
7.57671297e-01 6.07667565e-01 2.47481093e-03 -1.81138054e-01
-1.44175828e-01 4.59099203e-01 3.78446162e-01 -8.71943593e-01
1.24201763e+00 -1.62539512e-01 -3.93296242e-01 5.07785857e-01
1.05181098e+00 4.01446283e-01 -1.41161954e+00 1.13715343e-01
-9.39310789e-02 -2.96369225e-01 3.15451026e-01 -7.86393166e-01
-5.39143264e-01 8.32331955e-01 5.30421793e-01 3.65569860e-01
6.03158414e-01 -4.55673598e-02 5.33408344e-01 1.90438911e-01
3.16983134e-01 -1.05253565e+00 -2.80867010e-01 4.59145278e-01
5.08275807e-01 -7.88126528e-01 5.13911843e-01 1.47000765e-02
4.15026918e-02 1.11333871e+00 3.76080006e-01 -4.43843931e-01
7.46870100e-01 3.69614661e-01 -4.07055855e-01 -1.39715001e-01
-8.06682527e-01 2.41087660e-01 1.16273738e-01 1.15781493e-01
2.05905914e-01 3.54625471e-02 -2.99025029e-01 1.96813628e-01
-7.60002255e-01 -6.58104300e-01 5.84408462e-01 5.69060802e-01
-5.92186153e-01 -1.04263115e+00 -6.34671271e-01 3.94744515e-01
-1.97593853e-01 -3.21557820e-01 1.97804630e-01 7.00081527e-01
-2.56787181e-01 6.27679169e-01 6.51379943e-01 4.23621684e-02
-5.00892043e-01 1.70246750e-01 4.26212758e-01 -2.33310744e-01
-5.28682828e-01 6.60525739e-01 4.63091880e-01 -3.63800675e-01
1.54808968e-01 -1.01226878e+00 -1.08003426e+00 -8.44053447e-01
-1.67764246e-01 2.72538185e-01 1.13284326e+00 7.56708562e-01
-9.54981055e-03 4.32122588e-01 4.08775270e-01 -1.48733699e+00
-5.18136919e-01 -9.30460751e-01 -5.11857629e-01 -2.25045994e-01
-2.68639531e-02 -4.49289829e-01 -9.02393758e-01 -3.61839592e-01] | [7.065707683563232, 3.5368220806121826] |
2510e0a7-404b-4727-b277-0275b79d060e | recyclable-semi-supervised-method-based-on | 2306.02894 | null | https://arxiv.org/abs/2306.02894v1 | https://arxiv.org/pdf/2306.02894v1.pdf | Recyclable Semi-supervised Method Based on Multi-model Ensemble for Video Scene Parsing | Pixel-level Scene Understanding is one of the fundamental problems in computer vision, which aims at recognizing object classes, masks and semantics of each pixel in the given image. Since the real-world is actually video-based rather than a static state, learning to perform video semantic segmentation is more reasonable and practical for realistic applications. In this paper, we adopt Mask2Former as architecture and ViT-Adapter as backbone. Then, we propose a recyclable semi-supervised training method based on multi-model ensemble. Our method achieves the mIoU scores of 62.97% and 65.83% on Development test and final test respectively. Finally, we obtain the 2nd place in the Video Scene Parsing in the Wild Challenge at CVPR 2023. | ['Ning Wang', 'Xiaofeng Zhang', 'Si Gao', 'Chengjian Zheng', 'Diankai Zhang', 'Shaoli Liu', 'Biao Wu'] | 2023-06-05 | null | null | null | null | ['scene-parsing', 'video-semantic-segmentation', 'scene-understanding'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 6.45558000e-01 1.61242321e-01 -1.87089831e-01 -6.50770962e-01
-8.11405599e-01 -4.81978416e-01 2.38837838e-01 -3.39347184e-01
-5.20415425e-01 4.77697700e-01 -2.54774272e-01 -3.15022498e-01
4.87366945e-01 -5.19833148e-01 -1.06323993e+00 -6.57098532e-01
4.59314525e-01 3.89935434e-01 9.37406957e-01 7.12578148e-02
6.80699646e-02 -1.33414909e-01 -1.44284725e+00 7.63982713e-01
7.58463621e-01 1.10249102e+00 7.22690940e-01 8.83422256e-01
-2.70839423e-01 9.62555110e-01 -4.06069368e-01 -5.10426700e-01
1.73932642e-01 -4.15543020e-01 -9.67005432e-01 5.38652122e-01
8.27816188e-01 -3.26234639e-01 -1.77001163e-01 1.38607883e+00
1.37224242e-01 9.17876288e-02 3.95690829e-01 -1.11411595e+00
-2.55176365e-01 7.64065266e-01 -8.95799577e-01 3.79332572e-01
-1.01085871e-01 2.68138014e-02 8.17610562e-01 -6.57497764e-01
5.81209540e-01 1.14076018e+00 2.13720515e-01 7.61491656e-01
-6.18348479e-01 -4.96424198e-01 5.69453537e-01 5.28059483e-01
-1.05901265e+00 -5.01792073e-01 5.82198739e-01 -3.56964678e-01
7.63928831e-01 1.66255131e-01 3.02340567e-01 1.08718610e+00
-1.67141650e-02 1.30331254e+00 1.12408412e+00 -2.51906902e-01
8.88668373e-02 -4.58371192e-02 3.04167122e-01 7.17037797e-01
1.58666149e-01 -3.04453731e-01 -3.40795457e-01 4.37542289e-01
7.43488014e-01 -1.19547255e-01 -1.68926701e-01 -1.70064509e-01
-9.13617849e-01 5.48684657e-01 7.02230483e-02 1.67947859e-01
-2.73556143e-01 2.17193559e-01 2.65531391e-01 -1.04666904e-01
4.01878119e-01 -2.80612022e-01 -7.66855061e-01 -1.25861108e-01
-1.12904286e+00 -1.29006624e-01 5.22669077e-01 9.71115828e-01
5.26964307e-01 1.70014516e-01 -9.54779834e-02 8.49798143e-01
6.82260990e-01 4.52269912e-01 4.26377565e-01 -1.20947242e+00
5.56657255e-01 4.03732300e-01 -1.33768320e-01 -5.57131112e-01
-4.74419259e-02 -4.88822579e-01 -5.93245566e-01 -1.12586198e-02
4.01776552e-01 -1.90405920e-01 -1.59780109e+00 1.57845831e+00
4.11633909e-01 1.06473839e+00 1.41739294e-01 8.03870320e-01
1.10895169e+00 8.96143138e-01 3.37004811e-01 -1.99189305e-01
1.34099889e+00 -1.48184848e+00 -5.81876397e-01 -7.65412092e-01
3.54534090e-01 -6.08046830e-01 6.79218829e-01 5.42395711e-01
-1.06813824e+00 -8.13967586e-01 -1.00420034e+00 -4.73352075e-02
-2.26287954e-02 6.31574839e-02 6.60269439e-01 7.99204469e-01
-1.03784728e+00 3.86290431e-01 -9.64623868e-01 -2.01750189e-01
7.04125583e-01 2.34388798e-01 -3.10851961e-01 -3.46431404e-01
-6.98827922e-01 5.30602932e-01 7.16422677e-01 3.11746776e-01
-1.18236864e+00 -3.43015403e-01 -8.23957205e-01 -1.48215458e-01
5.22321045e-01 -3.18440050e-01 1.37269819e+00 -1.22539687e+00
-1.35326314e+00 1.22276819e+00 -3.71863663e-01 -5.59685349e-01
5.50587416e-01 -3.31289589e-01 -3.33409101e-01 1.78849548e-01
1.11233339e-01 9.79330897e-01 7.39497900e-01 -1.34694862e+00
-1.11054611e+00 -3.20866823e-01 7.93522224e-02 2.41107911e-01
1.16232231e-01 6.29794374e-02 -1.20105875e+00 -2.78902769e-01
4.03565168e-01 -7.35777736e-01 -3.60042870e-01 -4.79141086e-01
-4.10130024e-01 -2.85320312e-01 9.40855086e-01 -1.03789759e+00
8.07264745e-01 -2.18736982e+00 -4.81919432e-03 -2.52937108e-01
4.54918034e-02 4.76251274e-01 -1.60158366e-01 -1.80487558e-01
2.57615075e-02 3.55863795e-02 -3.85649174e-01 -4.27841902e-01
-3.57138515e-01 2.97760695e-01 -1.52596220e-01 3.16442370e-01
2.94336528e-01 8.71464193e-01 -7.55561233e-01 -8.62044990e-01
4.15858686e-01 2.08063632e-01 -6.06759071e-01 1.59237564e-01
-5.30847371e-01 5.73002696e-01 -5.77151775e-01 7.90609181e-01
8.97733986e-01 -2.88866818e-01 1.98646501e-01 -1.87556997e-01
1.35110587e-01 -4.39115055e-02 -1.18839633e+00 2.14393473e+00
-7.53548965e-02 6.14816010e-01 3.67416814e-02 -1.32744765e+00
5.22753775e-01 2.07487881e-01 3.63754541e-01 -6.56979620e-01
2.53520399e-01 -2.94707473e-02 1.39076158e-01 -6.21087372e-01
3.57523650e-01 2.99211562e-01 4.46574911e-02 -5.74293844e-02
3.61068577e-01 -4.35537323e-02 3.39562386e-01 5.82712889e-02
1.09074545e+00 5.38404405e-01 -2.53722370e-02 -1.57488152e-01
5.38338780e-01 1.07991956e-01 9.07327712e-01 7.50762224e-01
-5.39158285e-01 1.04345787e+00 2.45339394e-01 -1.76299825e-01
-6.67343318e-01 -1.08724558e+00 -4.05023284e-02 1.03841615e+00
4.18110877e-01 -2.77885527e-04 -1.12788117e+00 -7.89688408e-01
-5.40859461e-01 6.17671609e-01 -3.64069253e-01 1.95048824e-01
-7.50223935e-01 -7.01002955e-01 3.93871546e-01 5.93569815e-01
1.01899278e+00 -1.09320998e+00 -4.73099977e-01 1.92437261e-01
-5.27177215e-01 -1.83522046e+00 -3.19080114e-01 6.82028681e-02
-9.49113071e-01 -1.29767728e+00 -5.05288184e-01 -1.21997094e+00
4.68752533e-01 4.70417857e-01 1.04562652e+00 7.98745975e-02
-1.96564928e-01 1.32784322e-01 -4.82167274e-01 -3.30153763e-01
-1.83351666e-01 -1.44925445e-01 -5.18910527e-01 3.13340813e-01
2.72762060e-01 -2.23121494e-01 -6.88222647e-01 1.72477677e-01
-8.25615406e-01 4.88722384e-01 6.14325106e-01 4.68800873e-01
8.80271077e-01 4.47363704e-02 2.23253623e-01 -1.18472254e+00
-3.26357663e-01 -2.33852386e-01 -5.66578150e-01 4.93181050e-01
-8.35509077e-02 -3.94986749e-01 2.05871761e-01 -2.11923614e-01
-1.28270853e+00 5.79620481e-01 -3.21062863e-01 -3.98868293e-01
-4.56444532e-01 1.43803984e-01 -6.88320339e-01 1.52652204e-01
1.84173077e-01 2.87289917e-01 -5.96234918e-01 -4.56564337e-01
3.81226420e-01 6.85960770e-01 8.06761384e-01 -4.27677542e-01
4.50691313e-01 4.65549290e-01 -4.33287621e-01 -8.57718170e-01
-1.18342900e+00 -6.26991570e-01 -6.06486559e-01 -3.08050692e-01
1.39233172e+00 -1.36757529e+00 -3.22051197e-01 7.64258862e-01
-1.23024142e+00 -5.77326953e-01 1.17636196e-01 2.83990562e-01
-4.76051331e-01 4.50423896e-01 -4.75428909e-01 -6.38656735e-01
-1.64600402e-01 -1.42042911e+00 1.15367234e+00 6.06300712e-01
2.76083738e-01 -6.43248320e-01 -2.66497165e-01 1.02976012e+00
7.95131642e-03 1.95432305e-01 4.49466616e-01 -6.60222113e-01
-8.49993169e-01 -3.61718098e-03 -4.39256996e-01 6.41384482e-01
-1.13432534e-01 7.36923004e-03 -1.33626330e+00 6.08850457e-02
2.39478484e-01 -2.45410338e-01 1.17916906e+00 7.56404757e-01
1.41376102e+00 4.17149842e-01 -3.06681335e-01 6.21578872e-01
1.47152305e+00 4.31783020e-01 8.95737410e-01 7.26596639e-02
9.02522802e-01 5.38849354e-01 7.38549411e-01 4.88337427e-02
5.29753506e-01 4.40416992e-01 5.27198315e-01 -5.90182915e-02
-4.03141856e-01 -2.47433692e-01 3.59659642e-01 6.14017725e-01
6.71136007e-02 -5.46223044e-01 -9.50589597e-01 5.27099729e-01
-1.95428669e+00 -8.53722572e-01 -4.54579949e-01 1.76138568e+00
6.02748811e-01 4.38092560e-01 -2.37503022e-01 -1.37956247e-01
1.08768892e+00 2.07473427e-01 -6.23945534e-01 -3.99950184e-02
-3.96397501e-01 3.17566782e-01 7.28131235e-01 4.01590616e-01
-1.60872519e+00 1.52721572e+00 5.95607662e+00 9.06109035e-01
-8.37650895e-01 5.21831036e-01 1.19831896e+00 2.09278390e-01
1.38169467e-01 3.28161605e-02 -9.10592377e-01 3.35275710e-01
9.60916400e-01 4.10187721e-01 7.71685988e-02 7.49012053e-01
2.17412055e-01 -3.92599970e-01 -8.63615513e-01 1.10749209e+00
1.29011199e-01 -1.27511060e+00 -1.91292353e-02 -2.08195224e-01
9.40603495e-01 3.90274107e-01 -2.98604332e-02 2.48970732e-01
2.19775826e-01 -9.94176269e-01 9.06964421e-01 7.53646716e-02
5.13356984e-01 -4.45482224e-01 6.48588061e-01 4.94183511e-01
-1.17758441e+00 1.58856198e-01 -3.34476054e-01 1.92403346e-01
3.16344887e-01 2.49566540e-01 -3.22466791e-01 3.39547396e-01
9.07419145e-01 7.01964915e-01 -5.95050156e-01 1.21232426e+00
-3.33624333e-01 1.08294904e+00 -2.37129614e-01 2.90249944e-01
2.86924541e-01 -1.73923865e-01 5.25677763e-02 1.20667553e+00
-1.46926427e-02 2.46526673e-01 3.49724740e-01 3.70896697e-01
-1.76377177e-01 -1.58319607e-01 -2.42775261e-01 2.18646694e-02
9.87150446e-02 1.31467044e+00 -1.23840952e+00 -4.92454171e-01
-4.49953943e-01 1.31910193e+00 -2.68928707e-02 4.53648299e-01
-1.19136965e+00 9.25179645e-02 4.81586576e-01 -3.53009194e-01
6.16481781e-01 -2.40100741e-01 -2.97563016e-01 -1.28880811e+00
5.22193611e-02 -8.69600832e-01 2.41014034e-01 -7.02029705e-01
-7.86113799e-01 6.35245502e-01 -8.31749663e-02 -9.22854066e-01
5.71122505e-02 -8.63504171e-01 -7.35731602e-01 4.61904347e-01
-1.60802233e+00 -1.30460024e+00 -3.63242298e-01 5.65040350e-01
1.29779625e+00 -3.11905257e-02 4.21027273e-01 5.13349354e-01
-8.13625991e-01 4.80362535e-01 -6.90469816e-02 4.59867805e-01
3.94687474e-01 -1.17750251e+00 6.18779540e-01 1.37810147e+00
5.29422879e-01 -5.00694253e-02 7.28552282e-01 -6.51697099e-01
-1.35540211e+00 -1.28147721e+00 5.06229401e-01 -4.76722002e-01
3.78873140e-01 -4.52297181e-01 -7.36689210e-01 7.68653154e-01
3.64937246e-01 1.29017115e-01 2.84221411e-01 -3.06450397e-01
-1.32059082e-01 4.25138213e-02 -1.00720739e+00 4.10810411e-01
1.24953198e+00 -2.70480067e-01 -3.79086971e-01 5.00290811e-01
9.90415752e-01 -7.07611799e-01 -4.06071633e-01 5.81606150e-01
2.19465137e-01 -9.59019721e-01 9.19003129e-01 -5.45776308e-01
4.06681687e-01 -4.04318660e-01 -5.74360311e-01 -7.50623941e-01
1.06873736e-01 -3.41308028e-01 1.82623401e-01 1.36162639e+00
5.83361566e-01 -1.50534540e-01 1.35978866e+00 3.89121562e-01
-3.53172719e-01 -3.83152783e-01 -8.82749617e-01 -5.10382831e-01
-7.98828900e-02 -8.61099124e-01 1.17659740e-01 6.38890684e-01
-8.55738938e-01 3.03814739e-01 -2.96324432e-01 5.42961359e-01
8.87597799e-01 2.45597530e-02 5.85927248e-01 -9.30757046e-01
-5.39633870e-01 -1.17883995e-01 -4.52556193e-01 -1.32139599e+00
3.15638989e-01 -7.06563413e-01 3.51573884e-01 -1.68440843e+00
4.24429268e-01 -1.76741779e-01 -4.32380527e-01 3.08585703e-01
-1.87889069e-01 3.66138250e-01 2.26256490e-01 -7.27565363e-02
-1.08995688e+00 1.21111184e-01 1.17859650e+00 -3.50025684e-01
2.44677346e-02 7.75098875e-02 -6.14113510e-01 9.80783403e-01
8.61738980e-01 -3.15237612e-01 -6.05144382e-01 -8.37752759e-01
-4.43964779e-01 9.51291546e-02 2.73007751e-01 -9.51227963e-01
2.22712055e-01 -3.26472044e-01 3.51463735e-01 -8.32585692e-01
4.22025383e-01 -6.50196016e-01 5.97238094e-02 3.56632054e-01
-3.72249708e-02 -2.14391246e-01 1.62917405e-01 6.21180773e-01
-3.83358747e-01 -3.45104039e-01 8.27626586e-01 -2.79000640e-01
-1.54571223e+00 5.01614273e-01 -7.45315328e-02 1.80026740e-01
1.14746416e+00 -3.45595241e-01 -4.23458278e-01 2.65056174e-03
-7.63517380e-01 4.18033510e-01 9.87922177e-02 5.75244427e-01
7.98488081e-01 -7.44064152e-01 -7.91885912e-01 5.23109585e-02
4.41343151e-02 3.34071487e-01 6.51391566e-01 5.63093305e-01
-7.87277043e-01 1.76779866e-01 -3.22881341e-02 -9.24629271e-01
-1.36099577e+00 9.79243815e-02 3.57613534e-01 1.85846109e-02
-3.99402142e-01 1.30015457e+00 5.79274654e-01 -7.93147534e-02
3.31417829e-01 -2.64474213e-01 -3.18803549e-01 -2.60278314e-01
5.55865109e-01 -1.04166217e-01 5.62636144e-02 -8.26111019e-01
-3.94750983e-01 6.30053639e-01 -1.62184387e-01 -1.40907854e-01
1.22059381e+00 -1.76878482e-01 -1.07745789e-02 4.06976908e-01
1.12785244e+00 -4.87437129e-01 -1.44841504e+00 -1.76511064e-01
1.63125604e-01 -2.75141627e-01 1.46067962e-01 -7.94009507e-01
-1.34662879e+00 9.93631959e-01 8.68677020e-01 -2.54143834e-01
1.12878406e+00 1.38396978e-01 8.84568155e-01 3.19483578e-01
3.98506552e-01 -1.15776026e+00 -2.18050495e-01 3.86211634e-01
3.47707123e-01 -1.60008049e+00 -2.79577345e-01 -7.94675291e-01
-8.30411077e-01 7.40246713e-01 9.59750533e-01 -2.07035467e-01
5.64145446e-01 1.52434170e-01 3.07292432e-01 9.73429531e-02
-8.31213772e-01 -4.76330787e-01 2.50059456e-01 6.74777687e-01
3.82455796e-01 1.58722922e-02 4.39967252e-02 5.35721660e-01
1.70152292e-01 -8.75899419e-02 4.28731740e-01 8.15012217e-01
-8.00576210e-01 -1.04812968e+00 -1.01608291e-01 3.98701906e-01
-8.34211111e-01 -1.25277430e-01 -1.62778467e-01 4.45077151e-01
4.45471555e-01 1.35664928e+00 -3.72686386e-02 -4.24615264e-01
1.40159622e-01 1.04810625e-01 6.14581048e-01 -8.61526966e-01
-2.78192937e-01 3.46374273e-01 6.95799291e-02 -7.10990429e-01
-6.98218405e-01 -7.85692930e-01 -1.54815805e+00 2.45270178e-01
-3.30438763e-01 -7.95629025e-02 8.40905488e-01 1.23125088e+00
5.79087324e-02 6.35798573e-01 3.25274885e-01 -5.45858741e-01
-3.42827551e-02 -8.34600687e-01 -4.87756699e-01 3.03306788e-01
9.27808583e-02 -3.92043978e-01 2.32631098e-02 5.21106243e-01] | [9.200881958007812, -0.020463505759835243] |
c6669ec7-b874-418f-8eee-a662e89e1938 | generating-music-with-sentiment-using | 2212.11134 | null | https://arxiv.org/abs/2212.11134v1 | https://arxiv.org/pdf/2212.11134v1.pdf | Generating music with sentiment using Transformer-GANs | The field of Automatic Music Generation has seen significant progress thanks to the advent of Deep Learning. However, most of these results have been produced by unconditional models, which lack the ability to interact with their users, not allowing them to guide the generative process in meaningful and practical ways. Moreover, synthesizing music that remains coherent across longer timescales while still capturing the local aspects that make it sound ``realistic'' or ``human-like'' is still challenging. This is due to the large computational requirements needed to work with long sequences of data, and also to limitations imposed by the training schemes that are often employed. In this paper, we propose a generative model of symbolic music conditioned by data retrieved from human sentiment. The model is a Transformer-GAN trained with labels that correspond to different configurations of the valence and arousal dimensions that quantitatively represent human affective states. We try to tackle both of the problems above by employing an efficient linear version of Attention and using a Discriminator both as a tool to improve the overall quality of the generated music and its ability to follow the conditioning signals. | ['João Florindo', 'Jose Fornari', 'Pedro Neves'] | 2022-12-21 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 2.88427800e-01 2.05167308e-02 2.07847014e-01 -2.68976837e-01
-6.27376080e-01 -6.53202236e-01 7.23807395e-01 8.87733027e-02
-1.87313870e-01 7.37144649e-01 2.10701123e-01 2.72460103e-01
-6.83459416e-02 -7.90841699e-01 -4.76220250e-01 -8.45565319e-01
1.34528816e-01 5.65761566e-01 -9.57595259e-02 -3.09707761e-01
-6.89655095e-02 2.78929621e-01 -1.96776903e+00 4.23166841e-01
7.66057849e-01 1.07702160e+00 2.23545939e-01 6.66480660e-01
-1.64596334e-01 7.92452753e-01 -8.44423294e-01 -3.38937551e-01
8.76563936e-02 -1.13580883e+00 -3.52094352e-01 -1.01949021e-01
4.27359566e-02 1.81412045e-02 1.80414274e-01 9.26065505e-01
5.29339731e-01 2.61819750e-01 7.46304572e-01 -1.03874362e+00
-2.20585719e-01 5.80259800e-01 -1.36605889e-01 -3.12105000e-01
4.11067456e-01 6.90987483e-02 1.14262855e+00 -4.48168397e-01
6.18127525e-01 9.15938258e-01 6.49264395e-01 8.15835595e-01
-1.62167346e+00 -6.39533341e-01 -1.92094147e-01 1.49122059e-01
-1.16657543e+00 -4.93224740e-01 9.47157502e-01 -4.95660037e-01
5.57794273e-01 3.53730857e-01 9.43959951e-01 1.59097469e+00
-1.96608648e-01 5.58490813e-01 1.10573626e+00 -4.34427261e-01
4.13550138e-01 4.33153152e-01 -4.91382688e-01 1.57959163e-01
-1.66845039e-01 7.52745382e-03 -5.93887389e-01 -7.81200305e-02
5.89217424e-01 -3.20103407e-01 -2.95486093e-01 -3.27133954e-01
-9.68037248e-01 8.83816957e-01 3.51753205e-01 7.05394566e-01
-7.03384399e-01 2.24928200e-01 3.08061987e-01 2.59450376e-01
1.73774377e-01 9.79688644e-01 -2.00712070e-01 -3.06226403e-01
-1.27763104e+00 3.07741880e-01 7.46018767e-01 4.03692544e-01
4.52459455e-01 2.96813637e-01 -1.34734571e-01 7.90790379e-01
4.55983961e-03 4.36445832e-01 6.42475963e-01 -8.92477989e-01
-9.57890376e-02 3.61823231e-01 1.55466631e-01 -1.05895758e+00
-3.61633867e-01 -6.24214053e-01 -8.72226357e-01 3.43946457e-01
6.59527659e-01 -1.38773978e-01 -6.20909452e-01 2.19073153e+00
9.59048420e-02 7.15057626e-02 6.88529983e-02 7.69007862e-01
4.22090024e-01 7.84275711e-01 2.54577771e-02 -2.59058416e-01
1.14775240e+00 -4.71589386e-01 -7.50579715e-01 -1.72040328e-01
3.08013350e-01 -8.64855349e-01 1.24619710e+00 7.01120198e-01
-1.16175938e+00 -7.59775043e-01 -9.62761700e-01 1.76993847e-01
-3.30668896e-01 1.81792513e-01 5.05049586e-01 5.91211319e-01
-1.00402248e+00 8.16587746e-01 -6.20061994e-01 -6.87971041e-02
2.04582319e-01 3.17487091e-01 -1.55391678e-01 4.25994098e-01
-1.34993947e+00 8.41835499e-01 3.27135235e-01 1.99303970e-01
-6.48331046e-01 -4.48902518e-01 -5.47530055e-01 2.30073839e-01
-3.13510909e-03 -7.14968443e-01 1.15402639e+00 -1.54840434e+00
-1.79173160e+00 6.07814312e-01 8.30867738e-02 -3.24550271e-01
6.29743457e-01 -1.10110000e-01 -1.54655039e-01 6.96353316e-02
-7.92907923e-02 6.20407939e-01 8.56724977e-01 -1.22467864e+00
-2.08746895e-01 -8.75762030e-02 -1.62028372e-01 1.00251377e-01
-3.39379191e-01 -1.86862409e-01 -2.26234227e-01 -8.18905473e-01
-2.53210664e-01 -1.15986228e+00 -3.58823240e-01 -3.99530232e-01
-1.64772108e-01 7.49939233e-02 4.58310544e-01 -4.43439007e-01
1.10431755e+00 -2.21185064e+00 5.89200795e-01 4.15430069e-01
-2.18494013e-01 3.80270898e-01 -3.36001068e-02 7.15822041e-01
-1.42395571e-01 -1.91565290e-01 -7.94739500e-02 -5.10871410e-01
1.55261248e-01 1.97736144e-01 -6.48931086e-01 5.33571281e-02
1.58306316e-01 6.68669403e-01 -8.81726086e-01 -2.55310744e-01
7.89488852e-02 8.95530164e-01 -6.10588789e-01 4.47044402e-01
-5.00581682e-01 9.60811675e-01 -3.84220451e-01 -6.62093461e-02
5.92130907e-02 -9.90800112e-02 3.20979416e-01 -2.37846881e-01
1.40525931e-02 3.88393462e-01 -1.08005059e+00 1.84264231e+00
-7.38956630e-01 5.16201437e-01 -9.03970227e-02 -9.46281552e-01
9.85924721e-01 6.98454499e-01 6.10778809e-01 -7.30624199e-01
4.35877055e-01 3.79053921e-01 7.59363547e-02 -2.40720943e-01
2.61046380e-01 -6.37875199e-01 -1.46296635e-01 6.86431587e-01
6.42661378e-02 -5.27372360e-01 1.67691082e-01 -9.24302712e-02
8.34546626e-01 2.53164828e-01 1.00370862e-01 6.24175817e-02
5.63166142e-01 -3.23783308e-01 3.06321591e-01 4.77181733e-01
3.04013699e-01 6.46722913e-01 6.44119680e-01 -1.65880889e-01
-1.05609941e+00 -8.58776569e-01 3.77241597e-02 9.43278551e-01
-3.35962176e-01 -4.03230786e-01 -8.76663864e-01 -1.68939397e-01
-4.14265394e-01 8.68397057e-01 -7.49035716e-01 -3.73810709e-01
-4.46924716e-01 -5.43540776e-01 5.12065113e-01 4.61209148e-01
5.11953980e-02 -1.49784207e+00 -9.78850663e-01 4.70883161e-01
-4.01892334e-01 -8.78029585e-01 -7.21416771e-02 4.12689209e-01
-8.45565438e-01 -6.54995203e-01 -7.78646111e-01 -4.22354698e-01
3.58492285e-01 -3.79514873e-01 1.23388076e+00 -1.99188545e-01
-2.23076552e-01 4.09208499e-02 -4.45939362e-01 -4.63718027e-01
-5.90717137e-01 1.57915056e-01 -9.11719427e-02 3.50400031e-01
8.32309276e-02 -1.08160150e+00 -4.94164675e-01 7.87446424e-02
-1.12993729e+00 2.41738304e-01 6.82177603e-01 8.88441086e-01
3.82380575e-01 8.63539279e-02 8.61463845e-01 -8.85246098e-01
7.10843205e-01 -2.73744613e-01 -2.70881623e-01 -1.37003779e-01
-4.88995612e-01 1.20567970e-01 9.09352183e-01 -6.13133132e-01
-7.97703445e-01 1.65882036e-02 -4.00982648e-01 -3.76391977e-01
-3.33191752e-01 3.64570558e-01 -4.71997410e-02 4.80151236e-01
6.61238670e-01 1.77954480e-01 -1.12105355e-01 -2.60347664e-01
4.76127207e-01 5.37249625e-01 5.74699521e-01 -6.43966734e-01
6.62440479e-01 2.64762163e-01 6.43745363e-02 -7.24007010e-01
-9.42762256e-01 3.17330763e-04 -4.89315420e-01 -2.95506895e-01
8.61404479e-01 -6.23258889e-01 -6.30635858e-01 2.86423326e-01
-8.77373695e-01 -5.24036229e-01 -7.07908452e-01 3.20500314e-01
-8.47065210e-01 2.51708785e-03 -6.01711452e-01 -9.87172186e-01
-3.87005538e-01 -1.02066135e+00 9.26284015e-01 1.44704521e-01
-7.63642013e-01 -7.80400693e-01 3.66410822e-01 2.51266360e-02
7.40007102e-01 5.95680416e-01 8.92885447e-01 -2.99593657e-01
-1.71899706e-01 -3.66786808e-01 2.19105273e-01 5.09990036e-01
1.03720658e-01 4.12966162e-02 -1.20285833e+00 -9.16128084e-02
2.10354954e-01 -5.59709728e-01 4.98463213e-01 2.96618342e-01
8.81786525e-01 -2.94612736e-01 8.12148452e-02 3.08092207e-01
1.10353851e+00 3.26741427e-01 5.98541975e-01 5.31051494e-02
3.45497966e-01 8.14400792e-01 2.76941717e-01 3.95683676e-01
-1.15254754e-03 9.74402070e-01 2.68300831e-01 -1.08169332e-01
-4.48787026e-02 -4.07994300e-01 3.75658214e-01 9.21554506e-01
-2.86557585e-01 -1.40732318e-01 -6.11746788e-01 4.14054811e-01
-1.83672667e+00 -1.24623251e+00 3.32165859e-03 2.21911645e+00
1.01160431e+00 2.56174982e-01 2.55520046e-01 5.28447032e-01
2.30152160e-01 -5.14642969e-02 -2.24700794e-01 -6.02606833e-01
-1.13093711e-01 6.67890489e-01 -2.56384790e-01 1.57703936e-01
-8.93335700e-01 6.01141274e-01 5.60656929e+00 5.54304123e-01
-1.47404337e+00 -1.42661676e-01 3.40075046e-01 -1.66655868e-01
-3.78072172e-01 -1.82212234e-01 -1.31213352e-01 5.40260911e-01
1.13403165e+00 2.39831477e-01 6.51638806e-01 6.01477027e-01
2.50766009e-01 1.34647014e-02 -1.18949962e+00 9.58713531e-01
2.27257069e-02 -7.64587104e-01 -1.18473984e-01 -8.66959095e-02
7.67995775e-01 -3.80230099e-01 3.98349077e-01 2.06161112e-01
-3.50580364e-02 -1.30187821e+00 1.08314896e+00 6.99115515e-01
7.86830485e-01 -8.15427184e-01 6.53163910e-01 4.67831343e-01
-8.43343437e-01 -2.95339786e-02 -4.54300232e-02 -9.37538743e-02
1.33024916e-01 5.28744340e-01 -7.98567832e-01 2.74143100e-01
4.17017490e-01 4.06286329e-01 -3.21714073e-01 6.93255544e-01
-3.36785793e-01 6.20500684e-01 -3.15491974e-01 -5.76063953e-02
2.97524571e-01 -2.47931987e-01 2.44107038e-01 1.17428041e+00
5.97547472e-01 -1.20045654e-01 -5.49584776e-02 9.84223604e-01
1.06677450e-01 2.73784280e-01 -6.41786575e-01 -3.54284078e-01
1.64361879e-01 1.26761293e+00 -8.84708762e-01 -1.87917039e-01
-1.25222094e-02 1.05168796e+00 1.06768169e-01 2.70582587e-01
-8.04958940e-01 -3.02535325e-01 3.66344899e-01 1.82881042e-01
3.26524556e-01 -2.59085536e-01 -1.41041681e-01 -1.03004384e+00
-2.04295725e-01 -1.08143568e+00 8.88057202e-02 -8.39301169e-01
-1.20433402e+00 9.75105464e-01 -2.77916908e-01 -1.19207346e+00
-6.87225878e-01 -2.43461430e-01 -5.34115195e-01 9.10803556e-01
-1.02948987e+00 -1.01393390e+00 -1.40027732e-01 5.68209469e-01
1.22615501e-01 9.55137834e-02 1.25437772e+00 4.98226285e-01
-1.40609801e-01 4.18646246e-01 -1.86663792e-01 -1.33403629e-01
7.39231586e-01 -1.21570861e+00 -4.94522564e-02 4.51538444e-01
5.21324277e-01 5.85441887e-01 1.16868460e+00 -4.46996987e-02
-9.74889874e-01 -7.21425116e-01 9.20927167e-01 -3.92254770e-01
4.92486477e-01 -4.95954484e-01 -8.88656259e-01 3.50141376e-01
2.36681059e-01 -4.43879366e-01 8.82368922e-01 2.35096574e-01
-3.05692554e-01 -3.06103289e-01 -8.04434776e-01 5.02141714e-01
4.67018247e-01 -5.95176220e-01 -4.88258958e-01 -1.14367098e-01
1.86448827e-01 -2.14006275e-01 -6.38664842e-01 2.92880058e-01
6.36015594e-01 -1.17542303e+00 8.10296059e-01 -4.32417989e-01
5.05671024e-01 -2.89909095e-01 -3.84205133e-02 -1.37562323e+00
-1.93622500e-01 -4.63249534e-01 8.62385333e-02 1.41832912e+00
3.65248591e-01 -3.44103456e-01 6.60684109e-01 2.16396987e-01
9.35912430e-02 -6.25290275e-01 -8.97302449e-01 -3.96806538e-01
-7.61943385e-02 -5.89331627e-01 5.15746295e-01 8.28072369e-01
-2.02883463e-02 6.86746716e-01 -6.65953815e-01 -2.68472999e-01
3.23185116e-01 3.92887712e-01 8.15229654e-01 -1.28444862e+00
-7.79423356e-01 -8.00124466e-01 -2.81095743e-01 -5.79111576e-01
-5.07458067e-03 -7.83361375e-01 2.88103878e-01 -1.08754051e+00
-3.48191448e-02 -5.93666375e-01 -3.91462952e-01 3.46997768e-01
1.01410449e-01 6.33070588e-01 3.07239562e-01 1.43436775e-01
-2.18680605e-01 6.47847474e-01 1.08086514e+00 1.06622599e-01
-2.99459666e-01 8.92398134e-03 -6.01357818e-01 5.96447825e-01
7.15359747e-01 -4.41338688e-01 -6.08841717e-01 -6.24358170e-02
7.08621860e-01 2.60171205e-01 3.11969221e-01 -1.30301118e+00
-3.40883881e-02 7.31819198e-02 4.44552362e-01 -1.71530098e-01
7.24235058e-01 -7.77152300e-01 7.18146920e-01 3.22003514e-01
-5.52282274e-01 -1.27343640e-01 3.37100811e-02 2.51675755e-01
-4.46059138e-01 -2.87783802e-01 1.00151980e+00 -1.03851847e-01
-2.56503373e-02 -8.28896239e-02 -3.15430999e-01 -9.68565717e-02
7.71055758e-01 1.18403152e-01 3.71931046e-01 -8.71088147e-01
-8.31955433e-01 -2.01428056e-01 5.77324450e-01 3.53826523e-01
1.04251616e-01 -1.28583074e+00 -5.12819350e-01 3.05096775e-01
-6.75527006e-02 -2.68365234e-01 2.34748840e-01 7.64876962e-01
-1.89141542e-01 2.87447035e-01 -3.24374676e-01 -4.68664914e-01
-9.21093225e-01 5.50544620e-01 2.45327473e-01 -4.30421859e-01
-6.04075491e-01 6.77736759e-01 -2.00837627e-02 -6.55238032e-02
2.68386662e-01 -1.65014878e-01 -3.56635690e-01 4.74190980e-01
3.76766741e-01 3.60681079e-02 6.96155056e-02 -6.69530630e-01
-7.91470110e-02 5.16054153e-01 4.66472268e-01 -4.64933693e-01
1.44272614e+00 1.96035028e-01 1.37717025e-02 8.92919958e-01
8.23625386e-01 2.56554842e-01 -1.01325715e+00 1.17413230e-01
-3.95824127e-02 -2.57469654e-01 -1.16491832e-01 -9.17091429e-01
-9.74867582e-01 1.10474527e+00 4.38363612e-01 4.29705828e-01
1.29485571e+00 -1.41364545e-01 8.97667468e-01 2.71147359e-02
4.02810872e-01 -1.01772070e+00 2.00481340e-01 2.15081438e-01
7.70358443e-01 -7.52925873e-01 -4.23453957e-01 1.71294108e-01
-6.28359139e-01 1.14021063e+00 1.07720204e-01 -8.34208354e-02
3.15159142e-01 2.24366143e-01 4.46731836e-01 -6.00195415e-02
-7.16309726e-01 -2.26483241e-01 3.83387327e-01 3.36086601e-01
6.95496023e-01 -2.22097374e-02 -3.55368704e-01 5.95893741e-01
-5.71783662e-01 1.00408636e-01 3.69902253e-01 4.23007429e-01
-1.64202616e-01 -1.38738418e+00 -2.19743848e-01 1.49629325e-01
-6.27898753e-01 2.88606733e-01 -4.06463683e-01 4.84153926e-01
4.98592615e-01 8.33732545e-01 -9.52765048e-02 -3.57963800e-01
3.57461661e-01 3.31976682e-01 5.60828924e-01 -5.95544815e-01
-5.57512343e-01 4.21164215e-01 1.01158826e-03 -4.00514215e-01
-5.45355201e-01 -7.17445314e-01 -1.01877248e+00 3.37500386e-02
-1.72310308e-01 5.25885582e-01 6.69629753e-01 7.64421105e-01
1.94561370e-02 7.64035046e-01 6.78677797e-01 -1.06214416e+00
-5.19687653e-01 -1.02204192e+00 -6.00101769e-01 7.89592445e-01
5.98896742e-02 -5.52030981e-01 -1.43300399e-01 1.17031686e-01] | [15.913276672363281, 5.500870227813721] |
34279f95-c0ba-4bce-9a2f-cd5f8f7f62dd | video-object-segmentation-using-teacher | 1810.07733 | null | http://arxiv.org/abs/1810.07733v4 | http://arxiv.org/pdf/1810.07733v4.pdf | Video Object Segmentation using Teacher-Student Adaptation in a Human Robot Interaction (HRI) Setting | Video object segmentation is an essential task in robot manipulation to
facilitate grasping and learning affordances. Incremental learning is important
for robotics in unstructured environments, since the total number of objects
and their variations can be intractable. Inspired by the children learning
process, human robot interaction (HRI) can be utilized to teach robots about
the world guided by humans similar to how children learn from a parent or a
teacher. A human teacher can show potential objects of interest to the robot,
which is able to self adapt to the teaching signal without providing manual
segmentation labels. We propose a novel teacher-student learning paradigm to
teach robots about their surrounding environment. A two-stream motion and
appearance "teacher" network provides pseudo-labels to adapt an appearance
"student" network. The student network is able to segment the newly learned
objects in other scenes, whether they are static or in motion. We also
introduce a carefully designed dataset that serves the proposed HRI setup,
denoted as (I)nteractive (V)ideo (O)bject (S)egmentation. Our IVOS dataset
contains teaching videos of different objects, and manipulation tasks. Unlike
previous datasets, IVOS provides manipulation tasks sequences with segmentation
annotation along with the waypoints for the robot trajectories. It also
provides segmentation annotation for the different transformations such as
translation, scale, planar rotation, and out-of-plane rotation. Our proposed
adaptation method outperforms the state-of-the-art on DAVIS and FBMS with 6.8%
and 1.2% in F-measure respectively. It improves over the baseline on IVOS
dataset with 46.1% and 25.9% in mIoU. | ['Mennatullah Siam', 'Chen Jiang', 'Martin Jagersand', 'Laura Petrich', 'Steven Lu', 'Mohamed Elhoseiny', 'Mahmoud Gamal'] | 2018-10-17 | null | null | null | null | ['unsupervised-video-object-segmentation'] | ['computer-vision'] | [ 2.05935985e-01 3.02152395e-01 -3.43398869e-01 -3.00982386e-01
-1.67635873e-01 -7.22030640e-01 2.64666229e-01 -1.24602363e-01
-4.95322436e-01 5.08896828e-01 -4.69584405e-01 1.10572994e-01
-1.08797684e-01 -6.01509452e-01 -1.34640455e+00 -8.02909076e-01
-2.25779384e-01 9.14128184e-01 8.30068827e-01 -1.25604495e-01
2.41286203e-01 7.17851162e-01 -1.72330344e+00 1.45831347e-01
9.78355229e-01 6.63834095e-01 9.41436648e-01 7.83390760e-01
-6.98797107e-02 8.95340383e-01 -5.64558148e-01 1.27653241e-01
3.02702993e-01 5.75899966e-02 -1.10962403e+00 3.28927040e-01
5.44738352e-01 -7.12415576e-01 -4.96110141e-01 1.19997871e+00
1.73383400e-01 5.39740503e-01 1.03101623e+00 -1.52665472e+00
-5.64404368e-01 9.79762733e-01 -7.50304520e-01 -1.17515065e-01
3.72403413e-01 3.24337095e-01 4.30576473e-01 -7.79744983e-01
9.53104496e-01 1.47749865e+00 1.13844387e-01 8.44879329e-01
-9.59564209e-01 -7.40412831e-01 5.04102170e-01 5.02141058e-01
-7.84863293e-01 1.09835356e-01 5.90826392e-01 -5.78521848e-01
6.93437099e-01 -1.56857580e-01 7.65724301e-01 1.12954116e+00
-3.39366198e-02 1.43565178e+00 7.24757016e-01 -2.77125150e-01
4.08735089e-02 -1.01794396e-02 1.87086150e-01 6.58627391e-01
-9.02549177e-02 1.26735745e-02 -1.39047489e-01 6.44425869e-01
1.05503690e+00 9.67054591e-02 -3.06266487e-01 -8.52524936e-01
-1.38929451e+00 4.19280648e-01 6.32358491e-01 1.42620709e-02
-2.42485091e-01 4.17640358e-01 1.81282207e-01 2.72475302e-01
-2.54254371e-01 3.21699232e-01 -8.58604670e-01 -8.06698501e-02
-1.83539107e-01 2.94612408e-01 7.96213448e-01 1.75637829e+00
7.13601112e-01 2.24540964e-01 3.01095773e-03 8.88365686e-01
2.91593313e-01 6.18451536e-01 5.62009394e-01 -1.36151540e+00
3.72467577e-01 5.28794408e-01 1.17528066e-01 -5.07619023e-01
-5.36552072e-01 5.62034734e-03 -3.87036622e-01 5.05184352e-01
5.55909634e-01 -2.38513853e-02 -1.55272412e+00 1.50602531e+00
9.28729773e-01 2.45206282e-01 1.88007429e-01 9.65862572e-01
1.22061753e+00 7.52017558e-01 -1.17929326e-02 4.41487655e-02
1.13234389e+00 -1.69856107e+00 -3.91713887e-01 -1.59883112e-01
3.55589688e-01 -5.91778874e-01 1.06952763e+00 7.96541035e-01
-1.05403209e+00 -8.42716753e-01 -6.96380436e-01 -3.46102603e-02
-1.98894218e-01 1.60416156e-01 6.35293603e-01 2.23237816e-02
-7.79304683e-01 6.13765240e-01 -1.09172416e+00 -3.91483307e-01
5.09230435e-01 6.57949209e-01 -5.10190308e-01 -1.37830541e-01
-4.24758971e-01 6.03624165e-01 6.59993291e-01 -1.80682302e-01
-1.75671017e+00 -6.26000285e-01 -8.60774636e-01 -7.81973004e-02
7.21741796e-01 -1.82158783e-01 1.57855499e+00 -8.68092000e-01
-1.77726996e+00 6.02169454e-01 3.39869112e-01 -2.02319786e-01
6.49692237e-01 -5.31517208e-01 3.69682074e-01 6.24799848e-01
1.57991230e-01 1.50762105e+00 1.06847560e+00 -1.59541667e+00
-9.88937080e-01 -3.16811949e-01 2.72261828e-01 4.75175351e-01
-2.39131935e-02 -1.75239027e-01 -6.34301007e-01 -6.42377317e-01
4.06670153e-01 -1.25123453e+00 -2.01672822e-01 6.82050809e-02
-3.48469466e-01 -5.99053681e-01 1.43820453e+00 -4.33731407e-01
3.06631297e-01 -1.87332201e+00 5.01888335e-01 -2.94022739e-01
-1.07720114e-01 1.69214189e-01 -2.97867477e-01 3.97809111e-02
-9.97483879e-02 -3.81058276e-01 -1.13504916e-01 4.76519987e-02
-1.32220820e-01 4.54828620e-01 -8.33477527e-02 3.90248686e-01
3.52224819e-02 8.11276674e-01 -1.24205065e+00 -5.44900537e-01
3.35968971e-01 1.78696766e-01 -6.91679478e-01 6.10958159e-01
-6.15716219e-01 9.88349736e-01 -3.56165767e-01 7.80627072e-01
4.39523965e-01 9.84809324e-02 -1.41387209e-01 -6.05023503e-02
-5.44656552e-02 -1.77496731e-01 -1.20167220e+00 1.81300652e+00
-1.19293511e-01 5.05518138e-01 2.69875973e-01 -1.29554927e+00
8.02239716e-01 3.02936941e-01 5.08705616e-01 -2.45590955e-02
2.65036196e-01 7.65040144e-02 2.42205992e-01 -1.16984630e+00
1.37624919e-01 5.63235521e-01 1.50257096e-01 1.79669142e-01
5.53153396e-01 -6.58815682e-01 4.26419109e-01 -5.22507471e-04
8.21872294e-01 8.26482952e-01 -4.75676544e-02 -1.31568462e-01
3.22746724e-01 3.32644172e-02 5.61280191e-01 6.42228961e-01
-2.95745134e-01 4.51775819e-01 1.67669058e-01 -2.64253795e-01
-8.07501853e-01 -1.07926118e+00 7.83583894e-02 1.29059041e+00
6.11409485e-01 4.90906000e-01 -8.39603007e-01 -8.24954987e-01
-4.16699462e-02 5.85735440e-01 -4.59381163e-01 9.13857967e-02
-9.56389368e-01 -2.72418950e-02 7.85154349e-04 7.83391953e-01
6.19059563e-01 -1.77304173e+00 -1.06630719e+00 1.14999734e-01
-1.05650827e-01 -1.34646165e+00 -3.25370491e-01 5.37444055e-01
-9.10406232e-01 -1.18525755e+00 -8.36557865e-01 -1.41689467e+00
9.48589265e-01 4.40264493e-01 5.84331036e-01 -2.18127593e-01
-3.36190283e-01 8.22926342e-01 -6.79076910e-01 -5.96741796e-01
-3.70003194e-01 5.52662052e-02 2.14620978e-01 -5.02272308e-01
8.79366323e-03 -6.78821027e-01 -5.40916264e-01 4.86686170e-01
-8.40385258e-01 1.03143804e-01 7.24789321e-01 8.00433159e-01
5.67255020e-01 7.83999115e-02 4.25541878e-01 -7.31080830e-01
-1.49133757e-01 -4.10479188e-01 -6.24777615e-01 2.34021410e-01
-1.66714340e-01 1.57885160e-02 5.53250134e-01 -1.17095149e+00
-1.10086048e+00 6.24646842e-01 2.04706356e-01 -7.70311296e-01
-7.89134622e-01 -8.06815177e-02 -1.77616507e-01 -1.87967017e-01
3.77994269e-01 -1.63685098e-01 -2.86873609e-01 -3.96447212e-01
5.39833069e-01 5.18133759e-01 1.00028837e+00 -1.03298652e+00
7.94541717e-01 3.23007017e-01 -1.88093886e-01 -8.53218913e-01
-5.73718190e-01 -4.79030788e-01 -1.24817348e+00 -5.19769251e-01
1.10882986e+00 -6.37455344e-01 -9.87735391e-01 8.48727345e-01
-1.29524624e+00 -1.02273464e+00 -3.53038222e-01 5.66740632e-01
-9.46995974e-01 2.03476027e-01 -6.99527323e-01 -5.46488583e-01
-2.73387767e-02 -1.73595500e+00 9.22560573e-01 6.28517449e-01
1.39516205e-01 -6.00281894e-01 -4.68951494e-01 4.01652217e-01
-2.26605088e-01 2.78836280e-01 8.40685666e-01 -6.58285320e-01
-7.42903173e-01 2.55899996e-01 -9.78544578e-02 1.93153113e-01
1.49902239e-01 -4.02483009e-02 -7.36911535e-01 -3.24040741e-01
-1.49979711e-01 -7.90435910e-01 5.99889874e-01 3.64099741e-01
1.42070234e+00 -3.17306012e-01 -4.17996347e-01 5.75206637e-01
9.02161360e-01 7.85032749e-01 2.09156752e-01 3.50189120e-01
1.09442294e+00 9.63994563e-01 9.56709146e-01 1.00142226e-01
3.39354932e-01 6.29706383e-01 7.62489736e-01 2.19321176e-01
-1.63975120e-01 -2.24991441e-01 5.70397139e-01 6.15193665e-01
-1.15498498e-01 -1.40163466e-01 -9.20163989e-01 7.17683911e-01
-1.85741317e+00 -6.45472288e-01 -1.08515114e-01 1.69083381e+00
6.96158588e-01 -2.26807222e-02 1.19504243e-01 5.33811077e-02
6.20436847e-01 -3.91891539e-01 -1.04447901e+00 -2.28622869e-01
3.47463250e-01 1.12261325e-01 4.87647206e-01 3.28084767e-01
-1.20934582e+00 1.39611483e+00 4.87416935e+00 5.68782687e-01
-1.10856819e+00 -1.26152486e-01 2.22302064e-01 3.38909239e-01
2.62818068e-01 -2.14735180e-01 -7.95929790e-01 1.12317748e-01
1.62116736e-01 4.74887222e-01 5.93481600e-01 1.24454093e+00
-7.82306939e-02 -3.97953451e-01 -1.51956558e+00 8.37958038e-01
1.54500335e-01 -6.22942209e-01 -1.32904220e-02 -4.85789925e-01
1.01635146e+00 -5.24427230e-03 7.62837157e-02 5.48153579e-01
6.05410814e-01 -8.50524724e-01 1.00046694e+00 1.98706463e-01
6.01897717e-01 -4.75237995e-01 4.03077275e-01 6.49718404e-01
-1.23194492e+00 -4.37458009e-01 -2.24943042e-01 2.62102038e-01
7.00225160e-02 -4.44907695e-01 -1.18613577e+00 2.35397682e-01
1.29432428e+00 6.79864705e-01 -3.17490876e-01 9.19332385e-01
-6.52471125e-01 4.84703124e-01 -4.33166862e-01 -8.35288316e-02
3.31107438e-01 -3.90246570e-01 6.80934966e-01 9.45115566e-01
2.56421387e-01 4.65955138e-01 5.71662128e-01 6.20108008e-01
-3.40613467e-03 -2.20522478e-01 -4.89832222e-01 8.82837549e-02
5.22484660e-01 1.13856661e+00 -1.20744848e+00 -3.02209020e-01
-2.45230347e-01 1.22908556e+00 3.22587788e-01 5.59005976e-01
-6.67219102e-01 -4.58470583e-01 1.35985583e-01 -2.34155908e-01
6.41580641e-01 -3.18224221e-01 2.74352252e-01 -8.06889057e-01
-2.64272511e-01 -1.04578757e+00 1.45979896e-01 -1.01082325e+00
-7.05180824e-01 2.71878958e-01 5.66617131e-01 -1.16757679e+00
-1.64047599e-01 -9.01616275e-01 -3.91424835e-01 -4.74647917e-02
-1.23485124e+00 -1.16455257e+00 -7.04983950e-01 4.32397902e-01
1.30440366e+00 -1.69715673e-01 2.46103838e-01 2.44875494e-02
-3.31066996e-01 3.13836277e-01 -1.95265725e-01 2.64677882e-01
5.70839763e-01 -1.47148359e+00 -4.71687280e-02 4.04733777e-01
1.96207300e-01 2.51527280e-01 6.85235441e-01 -5.72977602e-01
-1.42913795e+00 -1.02160180e+00 -2.35324830e-01 -2.92425722e-01
4.94821846e-01 -2.49185994e-01 -9.56562936e-01 9.96711195e-01
9.59664732e-02 5.28740175e-02 -2.20807105e-01 -7.00416028e-01
1.20387234e-01 9.46455598e-02 -1.20537567e+00 6.45945609e-01
1.16551244e+00 1.75401717e-01 -6.43393815e-01 3.81521046e-01
1.18661141e+00 -1.00279665e+00 -7.27211177e-01 7.89995253e-01
6.62500441e-01 -5.76883733e-01 9.44296956e-01 -6.15078688e-01
4.55862194e-01 -3.39992523e-01 -6.06024563e-02 -1.12854671e+00
5.92052415e-02 -5.29099405e-01 -1.45433024e-01 1.09277809e+00
1.33396968e-01 -2.62721241e-01 9.45189536e-01 2.68105865e-01
-4.67316210e-01 -8.94110084e-01 -3.98242503e-01 -8.16852450e-01
1.48027897e-01 -3.56486887e-01 3.50751311e-01 8.84816885e-01
-2.14807019e-01 1.13112457e-01 3.24823000e-02 4.40333366e-01
5.04994452e-01 1.26127526e-01 1.26997721e+00 -1.19921434e+00
-2.70995229e-01 -2.09268510e-01 -3.67276102e-01 -1.61405087e+00
4.09896761e-01 -8.65819395e-01 5.42992175e-01 -1.63636220e+00
1.91073626e-01 -6.68206096e-01 1.65880203e-01 5.97488105e-01
-1.13904171e-01 -9.06026214e-02 2.48086229e-01 1.43666863e-01
-5.68324685e-01 6.05156302e-01 1.81121111e+00 -4.12820756e-01
-5.65344334e-01 1.56760469e-01 2.69496322e-01 1.09920108e+00
6.65574372e-01 -5.14787078e-01 -5.94920933e-01 -6.25576437e-01
-3.23559791e-01 2.60471135e-01 2.47068077e-01 -1.01684642e+00
2.83390254e-01 -4.86641914e-01 3.05152535e-01 -8.84929240e-01
3.66261303e-01 -1.05290759e+00 -2.00232849e-01 7.49080837e-01
-2.64253110e-01 -1.28055573e-01 1.47301659e-01 5.83973289e-01
5.71923330e-02 -8.00217330e-01 8.81971836e-01 -4.12410736e-01
-1.08480930e+00 4.93256658e-01 -3.39339167e-01 -1.22814602e-03
1.48741376e+00 -1.72012880e-01 -2.16758206e-01 -3.84069949e-01
-9.04227197e-01 7.26799130e-01 3.27047944e-01 6.03700399e-01
8.13551247e-01 -9.46445405e-01 -2.51764029e-01 4.31725681e-02
-7.95618519e-02 1.03540230e+00 9.82984602e-02 7.17935562e-01
-8.96875560e-01 3.77607085e-02 -3.97393703e-01 -1.07471597e+00
-1.25374973e+00 6.78961158e-01 1.67503834e-01 3.79865140e-01
-7.24680245e-01 1.19166315e+00 5.13695180e-01 -9.29344058e-01
9.24748361e-01 -7.42247880e-01 -4.54635710e-01 -1.14838377e-01
9.82481092e-02 6.11609995e-01 -5.43839395e-01 -4.84021008e-01
1.52195141e-01 7.41776943e-01 -1.35534689e-01 9.35039204e-03
1.44653904e+00 2.28522494e-02 -1.36363998e-01 6.28118277e-01
8.88719082e-01 -4.56765145e-01 -1.68416548e+00 -7.73484781e-02
5.67339435e-02 -1.40553311e-01 -6.40360415e-01 -6.75050557e-01
-1.01100647e+00 1.08907402e+00 6.48435235e-01 -2.04555184e-01
8.77122045e-01 3.12097430e-01 7.55997896e-01 8.34160745e-01
6.92119777e-01 -1.16521358e+00 7.63374150e-01 6.27864838e-01
1.04731071e+00 -1.41408348e+00 -1.66541100e-01 -7.12058425e-01
-7.14879453e-01 1.26069224e+00 1.37054515e+00 -2.24750027e-01
3.64171237e-01 2.67169774e-01 1.18178293e-01 -1.15697607e-01
-2.84121066e-01 -1.36008905e-03 1.99104741e-01 9.24052417e-01
-4.03891951e-02 4.92406860e-02 2.53052950e-01 3.12530458e-01
-3.67597401e-01 -3.51350665e-01 6.13365710e-01 9.59367275e-01
-7.70320594e-01 -8.03984702e-01 -4.56597984e-01 1.98429301e-01
-7.73135247e-03 5.37041903e-01 -1.77221984e-01 1.10264552e+00
3.00842911e-01 7.72972643e-01 -1.09885475e-02 -1.04428820e-01
3.28583091e-01 -2.19132885e-01 9.24944818e-01 -9.83034670e-01
-3.31917554e-01 1.08587191e-01 -4.66470748e-01 -5.27821839e-01
-6.23170137e-01 -7.35210121e-01 -1.93814683e+00 5.17271817e-01
-4.04805660e-01 -1.72307380e-02 5.57103574e-01 8.83439422e-01
-4.27704781e-01 7.63465285e-01 4.67762411e-01 -1.51020515e+00
-4.21707213e-01 -1.09854221e+00 -2.12450176e-01 3.87648404e-01
3.94258678e-01 -1.00760901e+00 -1.98548779e-01 4.02683854e-01] | [6.0224199295043945, -0.9436632990837097] |
ac5af762-5374-440a-a914-7ff6d3a2025d | facing-changes-continual-entity-alignment-for | 2207.11436 | null | https://arxiv.org/abs/2207.11436v1 | https://arxiv.org/pdf/2207.11436v1.pdf | Facing Changes: Continual Entity Alignment for Growing Knowledge Graphs | Entity alignment is a basic and vital technique in knowledge graph (KG) integration. Over the years, research on entity alignment has resided on the assumption that KGs are static, which neglects the nature of growth of real-world KGs. As KGs grow, previous alignment results face the need to be revisited while new entity alignment waits to be discovered. In this paper, we propose and dive into a realistic yet unexplored setting, referred to as continual entity alignment. To avoid retraining an entire model on the whole KGs whenever new entities and triples come, we present a continual alignment method for this task. It reconstructs an entity's representation based on entity adjacency, enabling it to generate embeddings for new entities quickly and inductively using their existing neighbors. It selects and replays partial pre-aligned entity pairs to train only parts of KGs while extracting trustworthy alignment for knowledge augmentation. As growing KGs inevitably contain non-matchable entities, different from previous works, the proposed method employs bidirectional nearest neighbor matching to find new entity alignment and update old alignment. Furthermore, we also construct new datasets by simulating the growth of multilingual DBpedia. Extensive experiments demonstrate that our continual alignment method is more effective than baselines based on retraining or inductive learning. | ['Wei Hu', 'Kexin Han', 'Yiqiao Jiang', 'Zequn Sun', 'Wenqiang Liu', 'Yuanning Cui', 'Yuxin Wang'] | 2022-07-23 | null | null | null | null | ['entity-alignment', 'entity-alignment'] | ['knowledge-base', 'natural-language-processing'] | [-1.78119972e-01 6.51005626e-01 -5.43891668e-01 -2.97346056e-01
-4.33328450e-01 -5.41423976e-01 4.16783899e-01 6.51235759e-01
-4.53059822e-01 9.26108956e-01 2.92865872e-01 -2.05407351e-01
4.74580787e-02 -1.21662736e+00 -1.08552432e+00 -4.00458544e-01
-1.18767016e-01 7.95410335e-01 2.73119688e-01 -4.58021790e-01
-1.05380662e-01 1.13146462e-01 -1.17853165e+00 -1.30380243e-01
1.18928480e+00 4.49825287e-01 -2.52408292e-02 3.70635167e-02
-3.66356164e-01 5.67489326e-01 -3.97232264e-01 -1.12125373e+00
4.47222471e-01 -3.79776508e-02 -1.14269376e+00 -3.52400512e-01
5.48892319e-01 -6.36788160e-02 -4.53709871e-01 1.14768851e+00
4.79208171e-01 -5.77351302e-02 2.91561812e-01 -1.57027388e+00
-1.14401853e+00 1.32448459e+00 -5.06883979e-01 -1.31211523e-02
4.89853770e-01 -2.83046544e-01 1.30406082e+00 -9.99738812e-01
1.04939616e+00 7.17590988e-01 1.02063644e+00 3.90204132e-01
-1.03311920e+00 -7.34635770e-01 4.92218941e-01 5.48490822e-01
-1.69378924e+00 -3.46674234e-01 7.60379255e-01 -9.21164900e-02
9.20085430e-01 5.37513345e-02 9.58843827e-01 9.69115078e-01
-2.24938318e-01 7.69521236e-01 4.43944395e-01 -6.29647195e-01
-8.13394487e-02 2.33030707e-01 3.09293836e-01 6.97133780e-01
8.37750554e-01 -3.44628572e-01 -4.36270058e-01 -2.04692349e-01
2.52584487e-01 -3.84092122e-01 -3.65645766e-01 -7.58675396e-01
-1.40203297e+00 5.53111255e-01 6.92308784e-01 3.88845563e-01
-4.06896770e-01 -1.79860666e-01 2.20322981e-01 3.61639500e-01
3.64365220e-01 6.64513230e-01 -5.84931731e-01 2.06444025e-01
-7.03520358e-01 4.76793677e-01 8.70606422e-01 1.31561375e+00
1.26459646e+00 -3.03843141e-01 1.83114916e-01 5.80661714e-01
5.47064990e-02 4.10189003e-01 3.68106514e-01 -3.60454559e-01
7.90239573e-01 1.02270460e+00 1.14308439e-01 -1.38067424e+00
-5.20992696e-01 -5.73983967e-01 -8.24970484e-01 -6.28749609e-01
1.97765917e-01 1.95692331e-02 -7.86484957e-01 1.89535034e+00
8.56784284e-01 4.63534355e-01 3.72622401e-01 5.35991073e-01
9.45457697e-01 2.73135811e-01 5.28983250e-02 -1.05618574e-01
9.75963473e-01 -1.06284285e+00 -6.83673143e-01 -2.94320136e-01
9.66782749e-01 -3.80929172e-01 7.17432737e-01 -1.48540884e-01
-8.50975394e-01 -3.92702639e-01 -1.12466002e+00 -1.02413647e-01
-7.75407314e-01 -4.18037146e-01 7.51986086e-01 4.12394971e-01
-1.01467478e+00 6.33839846e-01 -6.51380181e-01 -4.75473285e-01
1.62634850e-01 3.80868196e-01 -8.90901208e-01 -2.01462179e-01
-1.80461347e+00 1.20587993e+00 1.04576254e+00 2.93831646e-01
-6.21852875e-02 -9.10001099e-01 -1.20603120e+00 -2.57283747e-01
6.24198139e-01 -1.19589782e+00 8.74157906e-01 -5.23273110e-01
-9.07113969e-01 7.15815723e-01 -1.13069110e-01 -6.19636714e-01
1.36642918e-01 -2.68670440e-01 -8.21964443e-01 -3.66066068e-01
3.39840531e-01 5.75092256e-01 4.10699189e-01 -1.49552858e+00
-7.10031569e-01 -5.00105262e-01 2.42595598e-01 4.71634626e-01
-5.31221569e-01 -5.12533844e-01 -6.87424004e-01 -4.55189943e-01
4.22338367e-01 -8.32052469e-01 -2.57011682e-01 -6.50073886e-01
-5.30719399e-01 -2.02867851e-01 4.30799276e-01 -7.94034183e-01
1.67922676e+00 -1.58979166e+00 3.40446740e-01 4.45976228e-01
3.93240929e-01 2.52284974e-01 -1.98943332e-01 7.72464812e-01
-7.18454123e-02 1.16973534e-01 -2.69172877e-01 -3.06521803e-01
6.74216449e-02 3.74188125e-01 -1.75815746e-01 2.13857472e-01
4.82519083e-02 1.38656914e+00 -1.22810030e+00 -7.11477935e-01
-2.53685772e-01 1.13512613e-01 -5.46794891e-01 -6.82903454e-02
-5.58831953e-02 2.73627609e-01 -2.21629441e-01 9.08080995e-01
8.26368213e-01 -2.35193923e-01 5.41352689e-01 -6.68080389e-01
2.09697500e-01 2.47436926e-01 -1.33391821e+00 1.96060395e+00
-2.75484353e-01 3.17271203e-02 -4.57776129e-01 -1.07020223e+00
7.77281106e-01 1.07668666e-02 5.29928386e-01 -6.65971756e-01
-2.58735567e-01 3.73762518e-01 -1.26182362e-01 -4.63778973e-01
1.09704244e+00 5.11759892e-02 -2.13751778e-01 1.70885399e-01
1.83105260e-01 2.82573491e-01 2.82837719e-01 5.29041409e-01
1.06544352e+00 2.09379628e-01 4.74935472e-01 1.57811329e-01
4.55991060e-01 1.84554547e-01 8.29464138e-01 4.22286600e-01
1.68402009e-02 8.81095827e-02 1.06384143e-01 -4.31394726e-01
-1.01315427e+00 -1.11388385e+00 6.68974891e-02 6.43445969e-01
6.10322654e-01 -8.18498492e-01 -4.81062084e-01 -1.19074070e+00
3.71071965e-01 6.69162512e-01 -6.44889951e-01 -4.46877301e-01
-9.43341076e-01 -8.15396130e-01 6.00682020e-01 5.60369730e-01
4.05690998e-01 -8.78602028e-01 2.01441109e-01 3.97253007e-01
-5.64892292e-01 -1.17439806e+00 -2.54070699e-01 -1.41730577e-01
-5.71164787e-01 -1.21583128e+00 -2.38515973e-01 -9.32219744e-01
9.52752590e-01 8.56769010e-02 1.42701137e+00 -1.41313970e-02
1.35526493e-01 3.10324639e-01 -4.86047924e-01 -2.55406231e-01
-4.88042027e-01 6.78015590e-01 2.18388706e-01 -1.89083859e-01
5.89316070e-01 -8.49916339e-01 -4.27515209e-01 1.16453618e-01
-7.08945572e-01 2.03170046e-01 7.29364097e-01 7.96231329e-01
8.04567277e-01 5.75321727e-02 9.14118886e-01 -1.42086458e+00
4.33101565e-01 -8.25671494e-01 -3.31030279e-01 7.27836192e-01
-1.04008162e+00 1.73440561e-01 4.77754980e-01 -2.56610125e-01
-7.18703747e-01 -1.31535619e-01 -3.56281884e-02 -3.11851770e-01
3.04500818e-01 9.91063118e-01 -4.32790041e-01 -2.90365536e-02
4.68230158e-01 1.52222261e-01 -4.68733042e-01 -3.55623931e-01
8.32037926e-01 5.04968405e-01 8.31220567e-01 -7.40999401e-01
1.32817304e+00 1.92516685e-01 -1.95695370e-01 -2.34381184e-01
-7.43723929e-01 -5.40341556e-01 -9.98949170e-01 4.13021632e-02
3.71230990e-01 -9.85391080e-01 -2.82444865e-01 4.86636043e-01
-1.04465902e+00 9.72010121e-02 -5.41556776e-01 2.82931864e-01
1.83762405e-02 5.48216701e-01 -3.39412183e-01 -2.73267567e-01
-4.35843349e-01 -5.66365302e-01 9.35477078e-01 1.41572431e-01
-1.67815313e-01 -1.07919073e+00 4.55064505e-01 2.75451660e-01
2.13351429e-01 3.23094398e-01 9.08628166e-01 -9.37876284e-01
-6.69491410e-01 -3.94652247e-01 1.58370864e-02 9.56441388e-02
4.28278208e-01 -9.85045284e-02 -4.76565570e-01 -2.60595381e-01
-5.76751471e-01 -3.55000695e-04 5.43924749e-01 -4.37219828e-01
5.55513501e-01 -6.41920447e-01 -7.18473911e-01 5.82510948e-01
1.38214135e+00 -1.65161908e-01 6.21924996e-01 6.36512280e-01
1.10152221e+00 4.22384441e-01 7.21735001e-01 3.68801877e-02
1.40040791e+00 7.02571213e-01 4.28431004e-01 -3.14557403e-02
8.80241841e-02 -7.28108346e-01 6.45766631e-02 1.41734409e+00
-5.47919944e-02 -1.78071111e-01 -1.04049730e+00 9.10674334e-01
-1.99843371e+00 -8.38731110e-01 -1.45559534e-02 2.26535439e+00
1.13815320e+00 -2.19134558e-02 -8.37867558e-02 -3.43598984e-03
7.74707556e-01 -8.55917763e-03 -5.06038487e-01 1.44374505e-01
-4.67657506e-01 2.11393848e-01 6.16167426e-01 4.70084578e-01
-1.06307840e+00 1.28563201e+00 5.03853607e+00 4.63195473e-01
-8.07623684e-01 7.46898539e-03 4.87541817e-02 3.00278455e-01
-6.45538151e-01 3.67121875e-01 -1.08813858e+00 4.53917295e-01
6.47396445e-01 -5.63291669e-01 2.96329945e-01 6.92358434e-01
-6.37534320e-01 1.25681132e-01 -1.07017529e+00 6.46393418e-01
1.42905504e-01 -1.37617636e+00 2.99061269e-01 -1.41006261e-01
8.73342812e-01 7.82283209e-03 -3.60475808e-01 7.79563129e-01
5.82606018e-01 -6.67571843e-01 4.66894388e-01 5.10783672e-01
5.35044730e-01 -7.43861616e-01 9.21415865e-01 1.58103853e-01
-1.37438977e+00 1.87569186e-01 -3.41228694e-01 3.66143048e-01
2.72423059e-01 6.62678540e-01 -1.17562556e+00 1.51754093e+00
5.77107251e-01 8.24631572e-01 -8.56254458e-01 9.02626514e-01
-3.49671483e-01 1.18297979e-01 -3.60017687e-01 4.29984480e-01
3.35440636e-02 -2.33570293e-01 4.76476401e-01 1.01740015e+00
3.65782171e-01 -1.94663420e-01 1.88958332e-01 4.68750775e-01
-5.33088982e-01 4.08026904e-01 -7.88336039e-01 7.03698993e-02
1.02943742e+00 1.28874981e+00 -3.02020907e-01 -3.85137320e-01
-4.98665690e-01 1.02461100e+00 9.11365569e-01 2.31190249e-01
-8.67528558e-01 -3.58257592e-01 4.89802182e-01 1.70029327e-01
2.88229436e-01 -5.77409454e-02 1.25029221e-01 -1.48161888e+00
5.21840334e-01 -7.54839540e-01 6.89726114e-01 -4.93796051e-01
-1.35891700e+00 6.38024211e-01 7.47882947e-02 -1.02955663e+00
-1.90954462e-01 7.86525235e-02 -3.39210480e-01 6.24283731e-01
-1.72249794e+00 -1.57224905e+00 -3.49549800e-01 4.69415635e-01
-8.86281505e-02 1.80842429e-01 8.79026353e-01 7.16219842e-01
-6.10587895e-01 9.59269941e-01 -1.06544338e-01 3.97553384e-01
9.61987197e-01 -1.32778943e+00 8.41831625e-01 9.74538028e-01
5.45316100e-01 9.63805795e-01 5.72058558e-01 -1.05490780e+00
-1.38331652e+00 -1.31897795e+00 1.37000608e+00 -6.09816968e-01
8.78536642e-01 -3.40946078e-01 -1.22764730e+00 1.15823376e+00
6.87551424e-02 2.04610661e-01 6.25727773e-01 5.09461939e-01
-6.53507292e-01 -2.90682107e-01 -9.45941210e-01 7.44383872e-01
1.59065247e+00 -4.15274829e-01 -7.86468267e-01 1.86384469e-01
8.48460257e-01 -6.58247590e-01 -1.19132519e+00 1.01517928e+00
4.21315014e-01 -4.55294788e-01 8.77610922e-01 -9.75542605e-01
2.39981152e-02 -5.88589728e-01 -8.93381834e-02 -1.51523626e+00
-2.21422896e-01 -4.56305712e-01 -6.25773728e-01 1.60910320e+00
7.54477441e-01 -7.91504264e-01 9.27774310e-01 6.02401435e-01
-2.47319922e-01 -8.41706097e-01 -9.05873179e-01 -8.73413265e-01
-7.80873466e-03 -5.77545166e-02 1.08074224e+00 1.74167168e+00
3.20426285e-01 4.45258766e-01 -1.78178042e-01 6.90384507e-01
5.22897184e-01 4.32601064e-01 1.06997085e+00 -1.21188176e+00
-7.43576046e-03 -1.21309660e-01 -6.70899630e-01 -8.29263151e-01
3.55941534e-01 -1.31089020e+00 -2.64245540e-01 -1.81167150e+00
1.55047640e-01 -8.62029314e-01 -4.72955316e-01 6.68736815e-01
-6.24599159e-01 -1.37636755e-02 -1.44290298e-01 1.53096393e-01
-6.81479514e-01 6.15278184e-01 9.05382395e-01 -2.84397483e-01
-1.66662887e-01 -3.34049672e-01 -9.21324968e-01 5.14838099e-01
5.95334351e-01 -4.83236074e-01 -3.97096366e-01 -4.47723269e-01
7.79530585e-01 -2.95985013e-01 2.02680424e-01 -8.21889460e-01
6.57332659e-01 6.23594113e-02 -5.17298318e-02 -5.55595338e-01
5.81666678e-02 -9.34212029e-01 6.69835091e-01 -5.18922843e-02
4.79278341e-02 2.92966068e-01 7.92111382e-02 6.38530433e-01
-3.97559673e-01 -1.60542324e-01 1.65448397e-01 -6.03058822e-02
-9.81554031e-01 5.49573839e-01 4.90577191e-01 2.67540842e-01
1.11696327e+00 -3.44600588e-01 -5.59459925e-01 1.49102099e-02
-8.95604432e-01 6.48959637e-01 7.23842621e-01 5.38759351e-01
3.44855815e-01 -1.57572949e+00 -6.77203178e-01 -4.59733382e-02
4.80934441e-01 4.67211455e-01 2.85790771e-01 8.69638801e-01
-2.33894646e-01 9.09987465e-02 3.50448713e-02 -1.82266027e-01
-1.08883357e+00 8.17328870e-01 1.42454594e-01 -6.93723142e-01
-6.00598097e-01 7.45830297e-01 -2.14784637e-01 -8.82517040e-01
8.53785202e-02 -1.08123675e-01 -2.69694448e-01 1.90000743e-01
4.94689457e-02 1.36444956e-01 3.33603740e-01 -7.38649428e-01
-3.87332857e-01 3.94132197e-01 -4.18821961e-01 3.72301489e-01
1.42951107e+00 -3.09984952e-01 -3.37113410e-01 4.23610061e-01
8.94253433e-01 4.25773114e-01 -3.67396325e-01 -6.10317647e-01
4.14788067e-01 -4.18795288e-01 -5.35668135e-01 -6.80444479e-01
-1.05725372e+00 1.49656981e-01 -1.99742094e-02 6.97477460e-02
9.12855804e-01 9.36561376e-02 1.05990708e+00 6.56158507e-01
7.15216398e-01 -9.51366544e-01 -4.83068854e-01 3.90212208e-01
5.59620917e-01 -1.20606351e+00 9.48428214e-02 -6.62417471e-01
-4.79804575e-01 6.49453402e-01 1.03960645e+00 3.54974538e-01
5.81122100e-01 -4.87298658e-03 -2.04245880e-01 -2.89230973e-01
-4.48041916e-01 -2.90699035e-01 1.27948254e-01 7.84521997e-01
-3.93759944e-02 -2.56982241e-02 -2.60098875e-01 7.60876894e-01
-4.48124677e-01 -1.40676066e-01 3.48027229e-01 8.83238077e-01
-8.99302773e-03 -1.58486092e+00 2.11083412e-01 3.77257526e-01
-1.14971288e-01 -4.59165245e-01 -4.57881451e-01 1.18575013e+00
4.53103006e-01 3.03542793e-01 -1.62821129e-01 -6.23202980e-01
5.94575405e-01 3.28990489e-01 5.74359894e-01 -5.55576146e-01
-4.70833272e-01 -8.62681985e-01 4.36638981e-01 -2.58600414e-01
-5.03945529e-01 -5.58738410e-01 -1.28806460e+00 -4.44024801e-01
-8.23526263e-01 5.56353569e-01 4.61381465e-01 8.04539919e-01
6.33945346e-01 2.99963415e-01 4.10768360e-01 -1.38276741e-01
-1.95407495e-01 -7.95980334e-01 -3.71740639e-01 4.98982877e-01
-5.26040122e-02 -6.87117279e-01 -2.97105704e-02 -4.59790938e-02] | [8.792752265930176, 8.021483421325684] |
b663c027-fb3f-40e4-960d-1c143f9088bd | ecg-classification-with-a-convolutional | 2009.13320 | null | https://arxiv.org/abs/2009.13320v2 | https://arxiv.org/pdf/2009.13320v2.pdf | ECG Classification with a Convolutional Recurrent Neural Network | We developed a convolutional recurrent neural network to classify 12-lead ECG signals for the challenge of PhysioNet/ Computing in Cardiology 2020 as team Pink Irish Hat. The model combines convolutional and recurrent layers, takes sliding windows of ECG signals as input and yields the probability of each class as output. The convolutional part extracts features from each sliding window. The bi-directional gated recurrent unit (GRU) layer and an attention layer aggregate these features from all windows into a single feature vector. Finally, a dense layer outputs class probabilities. The final decision is made using test time augmentation (TTA) and an optimized decision threshold. Several hyperparameters of our architecture were optimized, the most important of which turned out to be the choice of optimizer and the number of filters per convolutional layer. Our network achieved a challenge score of 0.511 on the hidden validation set and 0.167 on the full hidden test set, ranking us 23rd out of 41 in the official ranking. | ['Ricard Delgado-Gonzalo', 'Jérôme Van Zaen', 'Mathieu Lemay', 'Halla Sigurthorsdottir'] | 2020-09-28 | null | null | null | null | ['ecg-classification'] | ['medical'] | [ 2.77917862e-01 1.33635864e-01 -2.12024711e-02 -3.37442696e-01
-9.87154543e-01 -2.92605042e-01 -8.11320767e-02 1.23151504e-01
-5.72915912e-01 6.13714397e-01 2.51907885e-01 -4.79594380e-01
-4.00462486e-02 -3.93417269e-01 -6.09499514e-01 -6.41725183e-01
-6.08660758e-01 1.77719980e-01 -6.33078963e-02 4.52878885e-02
-1.38292536e-01 3.75983387e-01 -9.49703515e-01 7.93127179e-01
5.23026764e-01 1.35587776e+00 -2.53260016e-01 1.31742215e+00
5.81608236e-01 1.01285303e+00 -8.12542379e-01 -1.43893868e-01
1.71622019e-02 -7.26922750e-01 -7.70252347e-01 -4.22072530e-01
-1.24372736e-01 -2.35488489e-02 -3.92712027e-01 3.35578293e-01
1.14531064e+00 -2.41895411e-02 3.27453315e-01 -7.51621723e-01
-9.01115909e-02 9.14604962e-01 1.18782641e-02 8.67655218e-01
8.08222443e-02 1.68004215e-01 9.52823997e-01 -8.96244347e-01
6.01761937e-01 4.03267533e-01 1.06578207e+00 2.71483332e-01
-1.18759561e+00 -5.09691536e-01 -2.56284148e-01 2.03944489e-01
-1.42811167e+00 -2.10184887e-01 4.93326038e-01 -3.93366724e-01
1.24688220e+00 4.31001157e-01 1.02808475e+00 9.92972970e-01
6.10874236e-01 5.14148653e-01 7.88396180e-01 -2.58057386e-01
1.13767453e-01 -5.08979410e-02 1.61871329e-01 4.42221552e-01
-4.00098637e-02 6.48633530e-03 -2.65101075e-01 -2.72961438e-01
7.65040755e-01 8.23677704e-02 -2.03061506e-01 3.30356628e-01
-1.31813157e+00 6.66474164e-01 6.66838109e-01 2.15739280e-01
-7.96466947e-01 2.80299395e-01 4.27878261e-01 4.17681187e-01
1.55652940e-01 7.96975136e-01 -6.50741756e-01 -2.31553301e-01
-9.68676925e-01 1.68261304e-01 6.17512763e-01 1.43568739e-01
2.52580464e-01 -7.64940605e-02 -6.64663553e-01 7.41172254e-01
8.11389610e-02 1.93426684e-01 5.40004909e-01 -7.56972194e-01
3.20622444e-01 5.33011198e-01 -1.05142519e-01 -7.20750749e-01
-1.05125546e+00 -1.00207353e+00 -8.51061046e-01 1.06580049e-01
1.93168521e-01 -8.87843966e-01 -1.08046365e+00 1.33381176e+00
-1.23186700e-01 1.69898242e-01 3.63891781e-03 9.09085453e-01
9.79114711e-01 4.99467522e-01 1.05630755e-01 -1.26543894e-01
1.40590000e+00 -5.35873473e-01 -4.99011636e-01 -1.32640794e-01
6.09196961e-01 -6.84265912e-01 5.92454910e-01 4.23800230e-01
-1.15897143e+00 -6.65766716e-01 -1.23316920e+00 1.83199257e-01
-1.23416379e-01 6.10409498e-01 3.85833740e-01 4.59757030e-01
-1.07640350e+00 8.94069493e-01 -9.21162665e-01 9.57244262e-02
4.37511206e-01 6.17294312e-01 -1.32609889e-01 4.02458102e-01
-1.44200671e+00 9.06303048e-01 1.97901756e-01 5.02740860e-01
-8.87294590e-01 -6.78654134e-01 -7.80152738e-01 1.85435832e-01
-1.63333416e-01 -5.73999763e-01 1.11663365e+00 -6.64513707e-01
-1.45391250e+00 8.55884850e-01 8.92813057e-02 -9.54256296e-01
5.08234203e-01 -2.12960601e-01 -6.65610135e-01 8.42897445e-02
5.31193465e-02 2.70012707e-01 7.14124024e-01 -4.29594219e-01
-5.70474207e-01 -1.60473138e-01 -3.65876168e-01 -5.72541216e-03
1.07749514e-01 -1.69426389e-03 -4.52478439e-01 -6.90797269e-01
2.28575528e-01 -9.79798317e-01 -6.59865320e-01 -6.81455493e-01
-5.23437738e-01 3.91255692e-02 3.06088030e-01 -8.34107399e-01
1.51683712e+00 -2.23279452e+00 1.87582999e-01 5.60721338e-01
3.82589936e-01 4.95482795e-02 1.07497446e-01 6.67199343e-02
-6.82672977e-01 3.72790433e-02 7.56850764e-02 -1.33068293e-01
-5.19720733e-01 -1.83543235e-01 -2.38935843e-01 4.08175319e-01
6.17769003e-01 1.21398222e+00 -6.15094662e-01 -1.01206489e-01
1.37546062e-01 8.89721632e-01 -5.14129162e-01 2.66946346e-01
1.55167028e-01 3.83230358e-01 -3.58915150e-01 2.59954274e-01
1.41831651e-01 -5.60188055e-01 3.38424146e-01 -2.54694402e-01
-4.97382544e-02 6.71156764e-01 -1.15266860e+00 1.43962801e+00
-2.77526289e-01 6.23047888e-01 -6.84038401e-01 -7.33697772e-01
9.85815644e-01 6.72206461e-01 6.78515673e-01 -4.07485127e-01
3.20296794e-01 2.62635231e-01 4.90447462e-01 -5.42824745e-01
5.89279942e-02 -9.41462442e-02 -2.80501634e-01 3.44277978e-01
2.89524477e-02 2.70198643e-01 -5.57281263e-02 -1.56096667e-02
1.47651637e+00 5.41448630e-02 5.85255120e-03 -1.25222161e-01
2.77864546e-01 -3.17251444e-01 6.95308387e-01 7.99371719e-01
8.22091699e-02 1.34593797e+00 1.15974224e+00 -1.09822702e+00
-8.47655416e-01 -8.20308983e-01 -2.67615706e-01 7.83536196e-01
-5.36387384e-01 -6.63384855e-01 -5.42173743e-01 -5.60899436e-01
-2.12535903e-01 2.79797345e-01 -1.08353829e+00 -3.71783853e-01
-7.28290617e-01 -1.07632613e+00 8.11472595e-01 8.09820950e-01
7.42746592e-02 -1.49250698e+00 -1.27601016e+00 6.96188807e-01
-2.17963129e-01 -9.60657001e-01 -1.47867277e-01 7.49479711e-01
-8.22353244e-01 -1.08626485e+00 -7.17203557e-01 -5.74572861e-01
2.42221519e-01 -7.89958239e-01 1.21726370e+00 -4.63195890e-03
-5.64451635e-01 -3.87359709e-01 -2.53788799e-01 -5.34291446e-01
-1.35929585e-01 4.70051438e-01 -2.73223668e-01 2.16578975e-01
7.69839734e-02 -3.97649676e-01 -9.89324152e-01 1.05119012e-02
-2.74166018e-01 -1.61711067e-01 6.17885888e-01 9.04987276e-01
6.16128147e-01 -5.09622633e-01 5.06703913e-01 -1.02654123e+00
6.32299066e-01 -3.38063031e-01 -2.52546072e-01 1.20619647e-02
-4.30360168e-01 -8.00060034e-02 5.30186296e-01 -1.71403512e-01
-1.43959686e-01 2.45236680e-01 -4.53198224e-01 -3.61870766e-01
2.90302128e-01 6.07410789e-01 3.65008801e-01 4.40908879e-01
8.96279514e-01 -3.06629613e-02 -1.03017844e-01 -2.98364967e-01
-1.77192539e-01 3.62178296e-01 6.15132511e-01 -6.89657107e-02
1.44808635e-01 8.60615224e-02 -2.01102078e-01 -4.56797898e-01
-8.03140521e-01 -2.77892649e-01 -5.98203897e-01 -2.05477491e-01
1.10907125e+00 -8.42146039e-01 -7.76017189e-01 2.32250035e-01
-1.11351883e+00 -3.90173048e-01 -6.73435092e-01 6.71764553e-01
-2.82029003e-01 -3.53011400e-01 -7.51830757e-01 -7.09737480e-01
-7.77288139e-01 -1.24792409e+00 8.79194319e-01 1.57057777e-01
-6.27718270e-01 -4.95065331e-01 1.25561431e-01 -1.04213260e-01
6.30065382e-01 7.19648421e-01 6.66207671e-01 -7.57044375e-01
-8.44187364e-02 -6.38503551e-01 1.82206586e-01 4.20187593e-01
-1.75506741e-01 -2.11216472e-02 -1.20969462e+00 -1.37228563e-01
-9.63828340e-02 -1.31855339e-01 1.04750216e+00 8.76177549e-01
1.39076614e+00 6.90301657e-02 -1.81136057e-01 8.54632735e-01
1.03159857e+00 2.97002375e-01 9.41221118e-01 1.40099317e-01
5.15541077e-01 -2.40138069e-01 3.32566425e-02 5.16118586e-01
1.23079978e-01 4.03694451e-01 2.22525358e-01 -6.23618126e-01
1.28864627e-02 2.84823984e-01 1.71907142e-01 6.38196945e-01
-5.59436083e-01 2.98509091e-01 -1.24130094e+00 3.36206138e-01
-1.55533564e+00 -9.38782990e-01 -6.30069226e-02 2.31105995e+00
6.57318890e-01 5.75093329e-01 2.03936324e-01 5.06157219e-01
4.43133384e-01 -5.75800128e-02 -5.09051323e-01 -6.26988590e-01
-2.24134505e-01 7.62043118e-01 2.73031563e-01 2.79455513e-01
-1.25864708e+00 5.15327930e-01 6.65257788e+00 5.99827655e-02
-1.39839077e+00 -7.94013813e-02 1.15091276e+00 -3.09014827e-01
2.13410392e-01 -3.04736763e-01 -6.57892346e-01 3.59720737e-01
1.52256727e+00 2.39781931e-01 7.09449574e-02 5.61452627e-01
1.06871039e-01 2.31793612e-01 -8.43315363e-01 8.77297997e-01
-9.87647846e-02 -1.50536406e+00 -3.39835882e-01 -3.71212848e-02
5.44472337e-01 6.71043217e-01 1.01369329e-01 4.26265061e-01
-4.87219309e-03 -1.54604244e+00 5.03115475e-01 8.30664814e-01
1.01713610e+00 -8.65330994e-01 1.19940937e+00 -1.70333937e-01
-9.05993223e-01 -3.73399287e-01 -1.27304476e-02 -1.12017214e-01
-1.77099809e-01 5.40966868e-01 -1.06085062e+00 2.66598552e-01
9.16268170e-01 8.46728563e-01 -6.49858117e-01 1.12674963e+00
-2.17321545e-01 7.82727778e-01 -3.20535243e-01 2.23742843e-01
4.18890454e-02 4.14677233e-01 3.93673122e-01 1.48956633e+00
3.38837095e-02 1.96186945e-01 8.04704055e-02 4.23658967e-01
-1.34185374e-01 -1.33794248e-01 -1.96270689e-01 1.58859834e-01
1.95596963e-01 1.43570745e+00 -6.31934285e-01 -4.72646534e-01
-6.64084358e-03 7.26110280e-01 1.72150165e-01 4.25333828e-01
-1.06275570e+00 -8.61162901e-01 4.88170743e-01 2.33524203e-01
3.44529539e-01 3.56680274e-01 -6.69818223e-01 -9.48185086e-01
4.89324750e-03 -8.75283420e-01 7.53774941e-01 -5.19290149e-01
-7.19446182e-01 1.22959292e+00 -7.22152650e-01 -1.19216907e+00
-4.83995646e-01 -4.25003022e-01 -7.28202283e-01 1.26114976e+00
-1.16469038e+00 -6.05376184e-01 -1.45573407e-01 3.59455079e-01
1.35185435e-01 -1.63138628e-01 1.23508525e+00 4.25869405e-01
-7.48729646e-01 7.36712396e-01 -4.32238191e-01 5.94881117e-01
3.23713452e-01 -1.20801485e+00 7.27398217e-01 7.01684952e-01
-6.80100396e-02 6.88965440e-01 5.03276289e-01 -3.17459494e-01
-1.02891469e+00 -1.07649624e+00 1.21565890e+00 -5.37939906e-01
3.61497015e-01 -1.52355969e-01 -8.09031248e-01 6.59497678e-01
-9.87585783e-02 4.48800445e-01 7.50514507e-01 1.73533782e-01
-5.31520359e-02 -1.60175726e-01 -5.13866842e-01 1.53351665e-01
3.87417644e-01 -6.47239864e-01 -4.75992888e-01 8.87994394e-02
4.30133551e-01 -9.62191343e-01 -1.29634941e+00 5.98673284e-01
1.00959218e+00 -7.71673739e-01 9.13677275e-01 -8.58794212e-01
6.14636064e-01 -2.13347197e-01 2.50995487e-01 -1.09867632e+00
-4.83848035e-01 -8.79616141e-01 1.42584415e-02 3.14685136e-01
1.11473346e+00 -3.54145944e-01 7.40120471e-01 4.43866044e-01
-3.51757407e-01 -1.42518127e+00 -7.79471040e-01 -4.57555465e-02
-1.84000477e-01 -6.49043858e-01 3.31065506e-01 6.07931137e-01
7.98243843e-03 5.54970801e-01 -4.59393501e-01 6.22576699e-02
1.40384302e-01 1.09216208e-02 1.88357040e-01 -1.17597795e+00
-4.61283147e-01 -4.59856749e-01 -5.55770278e-01 -1.33126736e-01
-6.57065630e-01 -9.62265313e-01 -2.22111136e-01 -1.47034955e+00
-5.43519035e-02 -4.27878916e-01 -1.07096612e+00 8.40076029e-01
-4.72392499e-01 7.04142869e-01 1.15339279e-01 -4.84035872e-02
-4.42061961e-01 -2.18771681e-01 7.92010605e-01 5.72883040e-02
-6.69415176e-01 4.00655538e-01 -7.26609051e-01 4.87141937e-01
9.36013222e-01 -5.26376307e-01 4.71242554e-02 -2.52921969e-01
4.30337012e-01 2.49074534e-01 3.43194067e-01 -1.25542939e+00
1.00338213e-01 5.34107506e-01 1.15371966e+00 -4.57996905e-01
2.82818466e-01 -3.97375792e-01 3.80689770e-01 8.54349375e-01
-6.17213607e-01 4.35898602e-01 2.31540814e-01 -3.01768794e-03
-8.94813091e-02 4.15114969e-01 7.81221628e-01 1.00642197e-01
-1.30189672e-01 2.71492749e-01 -2.80187786e-01 1.88876510e-01
6.06960237e-01 -4.81317937e-02 1.63171530e-01 -9.37868655e-03
-1.53082919e+00 2.81171739e-01 -3.63089710e-01 4.40139204e-01
7.27060497e-01 -1.16872740e+00 -1.13259375e+00 6.10777497e-01
-6.99890777e-02 -1.40728787e-01 3.20348799e-01 1.14659369e+00
-5.20018280e-01 2.20495507e-01 -3.04812461e-01 -7.94453382e-01
-9.91199076e-01 -1.94022551e-01 8.30626488e-01 -4.92962509e-01
-1.09071159e+00 9.42130208e-01 -3.94861579e-01 7.68679380e-02
2.97332287e-01 -5.62034428e-01 -4.73457813e-01 2.17776418e-01
8.49048972e-01 2.66168624e-01 5.77121258e-01 -2.91459143e-01
-6.09507501e-01 4.16243672e-01 3.80848385e-02 -1.33547876e-02
1.58752131e+00 5.77484488e-01 3.67309600e-02 6.73711896e-01
1.26398540e+00 -4.08654541e-01 -9.12790716e-01 7.69231021e-02
1.75947063e-02 2.98784763e-01 7.66777396e-02 -1.02846849e+00
-1.17660511e+00 9.38037217e-01 7.59310007e-01 3.36704075e-01
1.08907938e+00 -2.19605073e-01 6.41683340e-01 1.45190582e-01
-2.84320474e-01 -9.35415566e-01 -2.47658312e-01 5.88437080e-01
8.75739515e-01 -7.44501293e-01 -5.36315329e-02 2.62500256e-01
-9.20220912e-01 1.27771270e+00 1.86921671e-01 -5.77008307e-01
7.82184303e-01 5.32254100e-01 3.95471215e-01 -4.11902308e-01
-9.28952396e-01 -2.18728017e-02 5.74072182e-01 1.12685822e-01
9.86859620e-01 3.16457301e-01 -1.75513640e-01 7.91404963e-01
-4.06886220e-01 2.97439784e-01 2.99129277e-01 6.41844451e-01
-5.63934855e-02 -7.39688575e-01 -5.20307161e-02 9.22196686e-01
-1.02886009e+00 -1.06054507e-01 6.74127713e-02 4.23006207e-01
1.21216781e-01 5.95085204e-01 1.68345883e-01 -7.54853368e-01
7.27879524e-01 3.59900117e-01 1.27779186e-01 -6.82439923e-01
-1.41482818e+00 3.21500182e-01 9.75711867e-02 -7.07166255e-01
2.40724891e-01 -6.89682782e-01 -1.40142941e+00 3.62798095e-01
-1.80289626e-01 3.14901084e-01 6.19956911e-01 6.57109201e-01
5.07760167e-01 1.30927098e+00 6.92991912e-01 -6.60564363e-01
-4.85102773e-01 -1.04067719e+00 -2.94425488e-01 1.67658329e-01
6.61124170e-01 -4.90021072e-02 -1.60244972e-01 4.95548062e-02] | [14.351513862609863, 3.3171379566192627] |
2b26ec3b-80cc-4fd8-b7b5-f1728ce2eb4b | cipcad-bench-continuous-industrial-process | 2208.01529 | null | https://arxiv.org/abs/2208.01529v1 | https://arxiv.org/pdf/2208.01529v1.pdf | CIPCaD-Bench: Continuous Industrial Process datasets for benchmarking Causal Discovery methods | Causal relationships are commonly examined in manufacturing processes to support faults investigations, perform interventions, and make strategic decisions. Industry 4.0 has made available an increasing amount of data that enable data-driven Causal Discovery (CD). Considering the growing number of recently proposed CD methods, it is necessary to introduce strict benchmarking procedures on publicly available datasets since they represent the foundation for a fair comparison and validation of different methods. This work introduces two novel public datasets for CD in continuous manufacturing processes. The first dataset employs the well-known Tennessee Eastman simulator for fault detection and process control. The second dataset is extracted from an ultra-processed food manufacturing plant, and it includes a description of the plant, as well as multiple ground truths. These datasets are used to propose a benchmarking procedure based on different metrics and evaluated on a wide selection of CD algorithms. This work allows testing CD methods in realistic conditions enabling the selection of the most suitable method for specific target applications. The datasets are available at the following link: https://github.com/giovanniMen | ['Paolo Fiorini', "Diego Dall'Alba", 'Giovanni Menegozzo'] | 2022-08-02 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [ 3.39781582e-01 -4.93232943e-02 -1.61853850e-01 -1.14509895e-01
-3.67218375e-01 -4.11528289e-01 8.37788105e-01 7.49235094e-01
3.11256230e-01 8.52394760e-01 -4.63084653e-02 -3.28992397e-01
-7.83172488e-01 -1.05308557e+00 -6.05509102e-01 -6.06333375e-01
-4.09541041e-01 5.96707702e-01 7.84096494e-02 1.12970658e-02
4.87907439e-01 8.17139149e-01 -1.72778881e+00 4.56497759e-01
7.85673201e-01 9.26174998e-01 4.92126405e-01 3.98553282e-01
3.83425653e-01 5.94377220e-01 -6.51534975e-01 1.31800413e-01
1.35803789e-01 -6.19303226e-01 -7.17243373e-01 4.21855971e-02
-2.35753924e-01 3.33877712e-01 -1.11904301e-01 7.54356980e-01
3.42419177e-01 -4.16728519e-02 3.13476145e-01 -1.60257280e+00
-3.85991573e-01 7.84063876e-01 -5.59288496e-03 2.26431429e-01
5.16139627e-01 4.62738991e-01 8.16454232e-01 -6.06964350e-01
8.34045827e-01 1.15923190e+00 3.09304655e-01 -1.53193668e-01
-1.13234210e+00 -4.73178446e-01 -7.22360015e-02 3.09824944e-01
-1.18271482e+00 -4.89702635e-03 4.94298756e-01 -5.67451835e-01
9.14349377e-01 4.23936903e-01 5.62026203e-01 1.24073112e+00
7.95387208e-01 2.60153949e-01 1.40125751e+00 -4.06588435e-01
5.43540299e-01 -2.05579281e-01 -5.21235773e-03 4.27097350e-01
7.35891998e-01 5.30406952e-01 -4.09809947e-01 -3.01077366e-01
6.77484214e-01 -7.14292154e-02 -3.91638398e-01 -3.32030982e-01
-1.30614698e+00 6.41200900e-01 3.80407095e-01 4.28592354e-01
-7.57441163e-01 1.63132995e-01 4.34877753e-01 3.43650043e-01
6.28917441e-02 6.73671484e-01 -6.29691541e-01 -6.10509329e-02
-4.50055778e-01 3.89189452e-01 1.02195454e+00 8.87002826e-01
4.58651394e-01 -1.46504417e-01 -1.09776706e-01 3.32396954e-01
5.01380444e-01 2.10714787e-01 -7.70582492e-03 -6.04838967e-01
2.59934604e-01 7.50987947e-01 1.46673024e-01 -8.97356391e-01
-3.98176402e-01 -4.14450496e-01 -7.16669202e-01 2.82965183e-01
1.57062277e-01 -1.01030737e-01 -6.54471278e-01 9.14005101e-01
1.97720543e-01 3.61362606e-01 8.10356960e-02 7.90902555e-01
3.35609198e-01 4.85116988e-01 -1.02638088e-01 -3.29233825e-01
1.23712301e+00 -3.19907278e-01 -8.84660542e-01 9.83310565e-02
4.35494035e-01 -9.58561063e-01 7.15123177e-01 9.88786697e-01
-7.21288204e-01 -4.96139795e-01 -1.26175392e+00 6.85865700e-01
-4.55136478e-01 9.10679623e-02 6.21165037e-01 4.01232958e-01
-3.70556384e-01 9.89665926e-01 -8.88903081e-01 -5.14488518e-01
3.01392227e-01 1.89398661e-01 -2.50905037e-01 -5.15190065e-01
-1.24452794e+00 9.97328103e-01 5.38458288e-01 4.68656093e-01
-1.39122188e+00 -8.11226964e-01 -5.99153697e-01 -2.23380417e-01
5.94069421e-01 -6.38377249e-01 1.08343732e+00 -1.45066887e-01
-1.23917830e+00 1.37456357e-01 3.39552790e-01 -5.17813742e-01
7.06462026e-01 -4.43993419e-01 -7.80276477e-01 -1.13261707e-01
-3.35653201e-02 -1.77499548e-01 3.23329091e-01 -1.54294991e+00
-6.01570487e-01 -2.30474234e-01 -8.11480880e-02 -5.47007680e-01
3.72232080e-01 6.46572188e-02 -6.42189831e-02 -5.27834475e-01
7.81950075e-03 -9.44457114e-01 -2.64467537e-01 -7.49482691e-01
-8.69253159e-01 7.91017152e-03 8.99580598e-01 -5.56720436e-01
1.10795081e+00 -1.84417701e+00 8.96680206e-02 3.30076426e-01
-2.53325760e-01 -1.82855487e-01 1.09696306e-01 1.20891964e+00
-4.42553550e-01 2.59846635e-02 -4.66910332e-01 1.40676767e-01
-1.31954942e-02 3.55901718e-01 -1.15671940e-01 5.36014616e-01
7.83363521e-01 3.82291377e-01 -1.01656818e+00 4.52200100e-02
7.77779579e-01 2.49291897e-01 -6.19756430e-02 1.17565610e-01
-3.63723367e-01 7.11861730e-01 -5.13085365e-01 7.60868847e-01
4.45160687e-01 3.91729362e-02 4.08731967e-01 -2.73687601e-01
-3.19682747e-01 2.78652072e-01 -1.50705671e+00 1.41167569e+00
-4.68014568e-01 1.41317025e-01 -2.34343305e-01 -7.79592395e-01
1.19317138e+00 5.31210959e-01 6.46041751e-01 -5.20975232e-01
3.39761138e-01 2.23850742e-01 2.26356179e-01 -5.22824764e-01
9.36996788e-02 -1.49785671e-02 7.07329437e-03 1.32275419e-02
-4.71311286e-02 -4.48629498e-01 6.35527730e-01 -1.83863625e-01
1.62457049e+00 1.48975402e-01 2.71223426e-01 -5.95953763e-01
5.81523001e-01 2.72213310e-01 6.77161157e-01 2.49691337e-01
1.79931998e-01 5.07235706e-01 7.44395256e-01 -4.05991346e-01
-6.97840214e-01 -8.95636499e-01 -3.26852053e-01 -6.90089364e-04
1.20878249e-01 -3.88691306e-01 -3.30281973e-01 -5.52051485e-01
4.78692055e-01 1.24872530e+00 -5.47260821e-01 -2.09225997e-01
-1.14172325e-01 -9.39162552e-01 2.04437047e-01 1.95499510e-01
1.79936767e-01 -9.98781979e-01 -6.30965114e-01 3.40634942e-01
2.75737762e-01 -1.00098693e+00 6.17554188e-01 5.59880495e-01
-8.97861481e-01 -1.64858305e+00 1.55852064e-01 -6.00026436e-02
4.11920279e-01 -1.32108942e-01 1.21495342e+00 -2.39820685e-02
-7.08882570e-01 3.21504027e-02 -5.01893401e-01 -6.78116500e-01
-9.00582612e-01 -2.44702041e-01 1.01378597e-01 -3.62716645e-01
3.03399656e-02 -2.90745139e-01 -2.90453225e-01 6.10535383e-01
-8.35037172e-01 -3.40460032e-01 8.43299627e-01 5.64665854e-01
6.57573104e-01 7.26235092e-01 6.69602931e-01 -9.72471654e-01
6.81598723e-01 -8.74509633e-01 -9.77703094e-01 1.84888884e-01
-1.03086090e+00 -1.41285539e-01 4.75036025e-01 7.85012990e-02
-9.61939394e-01 1.30770933e-02 1.54014692e-01 -4.10464674e-01
-5.99637091e-01 9.55414116e-01 -3.64926636e-01 4.95121747e-01
6.13447070e-01 -4.66096550e-01 -1.37201220e-01 -5.31755507e-01
1.26159012e-01 3.05145532e-01 3.32348377e-01 -5.85229218e-01
6.67799354e-01 9.11410600e-02 3.34068537e-01 -6.06029868e-01
-2.76216298e-01 -3.87581825e-01 -6.20600879e-01 -3.93872142e-01
4.82430726e-01 -6.84098065e-01 -7.37367392e-01 1.38822734e-01
-9.40072834e-01 -3.96353513e-01 -4.00416374e-01 6.29269004e-01
-5.58370352e-01 -1.30476519e-01 -4.95884508e-01 -6.75225437e-01
9.67988595e-02 -1.25781465e+00 8.56016159e-01 -3.42955410e-01
-4.64767605e-01 -8.69620621e-01 5.03521133e-03 1.49946138e-01
1.69926435e-01 9.78532970e-01 1.05131030e+00 -5.21847010e-01
-6.70880914e-01 -5.08738041e-01 8.81793201e-02 4.80217725e-01
6.15857780e-01 6.78096414e-01 -7.96039760e-01 -2.07488954e-01
-1.54204117e-02 1.98094621e-01 4.74964976e-01 3.25304508e-01
9.36456501e-01 1.54126376e-01 -5.18681347e-01 -1.43893093e-01
1.57793283e+00 3.31390232e-01 7.07727194e-01 4.09764081e-01
4.56005335e-01 7.29119778e-01 1.30472267e+00 7.30581045e-01
-1.08440883e-01 4.68874663e-01 1.07216752e+00 1.45583928e-01
5.81080541e-02 1.37306318e-01 2.00804964e-01 7.80334711e-01
-2.16690451e-01 -5.02504051e-01 -1.26926422e+00 4.03477401e-01
-1.80192685e+00 -7.95957446e-01 -9.41223025e-01 2.21717262e+00
3.47798496e-01 3.37406695e-01 -2.39382267e-01 6.01907253e-01
6.29026115e-01 -3.46415013e-01 -2.67029166e-01 -2.97337532e-01
7.46832192e-02 2.22012565e-01 5.37131786e-01 1.18976630e-01
-1.02472067e+00 1.35618165e-01 5.78715754e+00 2.05698386e-01
-8.31854522e-01 3.84855270e-03 3.81894946e-01 8.74265581e-02
-4.19683978e-02 4.90481369e-02 -6.03136063e-01 3.41994613e-01
1.55707943e+00 -3.65289569e-01 5.13870083e-02 6.17063880e-01
7.52841532e-01 -4.05340672e-01 -1.30406606e+00 3.47930133e-01
-4.69229341e-01 -1.52432406e+00 -2.51873523e-01 2.10565239e-01
9.24279213e-01 1.14423707e-01 -4.58212376e-01 -2.05553487e-01
5.75310051e-01 -9.23287272e-01 6.60053611e-01 7.19023168e-01
1.91187561e-01 -9.15880203e-01 1.10624838e+00 -9.31363180e-02
-8.74359190e-01 -4.44441348e-01 -1.34805618e-02 -1.30533338e-01
3.52446169e-01 1.36390960e+00 -1.03880572e+00 1.68596268e+00
7.11368024e-01 7.76784956e-01 -2.58898854e-01 1.14903498e+00
-3.73207837e-01 7.18775451e-01 3.28167006e-02 3.41873020e-01
-1.25289515e-01 -1.78728953e-01 3.70467365e-01 1.01549375e+00
7.06701458e-01 -4.57129985e-01 1.85842112e-01 9.73352432e-01
2.14242131e-01 -9.09905359e-02 -7.01240420e-01 -3.03270906e-01
5.81184387e-01 1.19305587e+00 -8.66763771e-01 1.31151214e-01
-2.18570292e-01 5.59496403e-01 -5.01771748e-01 -1.11917220e-01
-9.01910365e-01 3.69794331e-02 8.77450705e-01 3.40196878e-01
-4.99665290e-02 -3.94888788e-01 -2.69586265e-01 -3.03677887e-01
-3.53373736e-01 -1.00958824e+00 4.31253314e-01 -7.53041208e-01
-1.45402002e+00 4.85062420e-01 3.94026130e-01 -1.34747684e+00
-1.91249475e-01 -7.02970028e-01 -3.86820018e-01 9.64963734e-01
-1.39619541e+00 -7.10921168e-01 -7.13714361e-01 8.64517614e-02
5.41246414e-01 -1.99544523e-02 9.13836062e-01 3.72171491e-01
-1.10519719e+00 -5.20241797e-01 -1.83820408e-02 -4.72079158e-01
6.75556362e-01 -1.25588572e+00 3.84657770e-01 9.29813564e-01
-1.89927086e-01 5.14554083e-01 1.17955744e+00 -1.07210493e+00
-1.75492358e+00 -1.44714701e+00 3.33082557e-01 -5.55401802e-01
1.01086509e+00 -8.49474743e-02 -8.58074248e-01 6.34010851e-01
3.13254327e-01 -3.59485717e-03 4.57757920e-01 -6.22070767e-02
4.47271258e-01 -9.85758975e-02 -1.07810533e+00 9.45900828e-02
7.34100103e-01 -4.77889292e-02 -4.07557487e-01 5.98088682e-01
4.55667943e-01 -2.85187751e-01 -1.50615263e+00 6.53473198e-01
-3.34535576e-02 -9.05428648e-01 5.43959320e-01 -1.32186428e-01
7.53871202e-01 -6.87834442e-01 -1.02055736e-01 -1.50903487e+00
-1.03756703e-01 -4.85607237e-01 -8.92229155e-02 1.37550962e+00
2.58801699e-01 -7.98439384e-01 1.85336664e-01 1.43692032e-01
-5.02877533e-01 -7.49425292e-01 -6.59890473e-01 -1.04970551e+00
-4.10697550e-01 -6.27037346e-01 8.67331207e-01 1.09528768e+00
-2.57697910e-01 1.81907807e-02 2.45150834e-01 7.38330305e-01
4.31712687e-01 6.55609220e-02 6.72243178e-01 -1.53300512e+00
-1.13512799e-02 -2.05604881e-01 -6.51533723e-01 2.68542230e-01
-2.39746615e-01 -8.18639040e-01 1.72866270e-01 -1.75212348e+00
-3.67577821e-01 -6.85279310e-01 -5.66493273e-01 4.46689665e-01
1.79721236e-01 -1.35545820e-01 7.17875501e-03 2.47981474e-02
2.31068611e-01 4.80329722e-01 1.11794388e+00 -9.65943839e-03
3.92691083e-02 7.25553855e-02 -2.87627995e-01 3.72565359e-01
1.02398348e+00 -5.48841834e-01 -4.71523851e-01 1.66054130e-01
2.38424674e-01 -5.69546036e-03 6.48223579e-01 -1.26923263e+00
-7.45991617e-02 -3.77869040e-01 2.06718326e-01 -5.83497822e-01
-1.94220901e-01 -1.06923008e+00 1.09489167e+00 9.74098444e-01
-3.49253900e-02 4.69877809e-01 3.74490768e-01 6.06092155e-01
-4.53023046e-01 -2.78322995e-01 3.60929757e-01 2.26123407e-02
-8.41493189e-01 -2.91899405e-02 -1.11740619e-01 -5.69470048e-01
1.49055040e+00 1.06051847e-01 -3.90407026e-01 2.96662688e-01
-6.65302217e-01 1.97281763e-01 3.77694011e-01 7.43880987e-01
3.82205099e-01 -1.13022506e+00 -8.47924292e-01 4.62539680e-02
3.15757364e-01 -1.83030069e-02 1.06813140e-01 9.31570590e-01
-6.78312123e-01 4.92316961e-01 -4.65959996e-01 -5.55264294e-01
-8.90409827e-01 7.61298001e-01 1.71597481e-01 -1.70454785e-01
-5.35869777e-01 4.34865952e-01 -2.58341521e-01 -2.56401777e-01
-5.34628808e-01 -8.79708469e-01 1.37110323e-01 7.01167714e-03
2.05260560e-01 7.28184164e-01 7.50843763e-01 -1.32542729e-01
-4.83956307e-01 -2.79798806e-01 4.43177670e-01 3.43412906e-01
1.58928490e+00 1.46725044e-01 -3.18861365e-01 9.85918581e-01
4.99135166e-01 -1.40122518e-01 -1.16524470e+00 4.89731818e-01
5.46335220e-01 -5.76802909e-01 1.55282602e-01 -1.18029010e+00
-1.07821274e+00 3.75091940e-01 6.52663887e-01 6.58747911e-01
1.05580115e+00 -3.11912328e-01 1.66699793e-02 -4.60809544e-02
7.24886775e-01 -7.13419259e-01 -1.38703212e-01 3.17006886e-01
1.35354924e+00 -7.89761603e-01 2.83556789e-01 -1.03850949e+00
-4.73377347e-01 1.09485877e+00 2.14999303e-01 -1.73854068e-01
5.38798153e-01 6.75354719e-01 -1.97145805e-01 -5.90272069e-01
-9.81425285e-01 -3.18992026e-02 -1.47600219e-01 4.43469763e-01
5.68322003e-01 2.20289975e-01 -3.20799440e-01 3.31321210e-01
3.71807702e-02 3.84623408e-01 5.92683017e-01 1.21357012e+00
7.91214108e-02 -1.46651363e+00 -4.59254652e-01 5.94896317e-01
-2.38796860e-01 3.24550778e-01 -2.01268032e-01 1.22661734e+00
2.31776997e-01 1.26054287e+00 -1.77766338e-01 -3.15994471e-01
1.03298950e+00 -6.48580119e-02 4.97665226e-01 -7.30180562e-01
-7.28256702e-01 -4.16203171e-01 4.72278655e-01 -7.74475634e-01
-2.69124776e-01 -1.13960457e+00 -1.33342469e+00 -2.89417624e-01
-3.97481352e-01 1.37628219e-03 8.78335953e-01 7.82829463e-01
4.97091383e-01 1.48561680e+00 6.06162190e-01 -7.22384274e-01
-3.15390170e-01 -1.06069720e+00 -6.80046260e-01 4.32023704e-01
-2.33388856e-01 -1.16292775e+00 -1.96293831e-01 2.00411990e-01] | [6.946234226226807, 2.4979119300842285] |
6bec4b6b-9635-4a80-8eb6-b89a7cb64a35 | gazing-at-social-interactions-between | null | null | https://www.frontiersin.org/articles/10.3389/fnbot.2021.639999/full | https://www.frontiersin.org/articles/10.3389/fnbot.2021.639999/full | Gazing at Social Interactions Between Foraging and Decision Theory | Finding the underlying principles of social attention in humans seems to be essential for the design of the interaction between natural and artificial agents. Here, we focus on the computational modeling of gaze dynamics as exhibited by humans when perceiving socially relevant multimodal information. The audio-visual landscape of social interactions is distilled into a number of multimodal patches that convey different social value, and we work under the general frame of foraging as a tradeoff between local patch exploitation and landscape exploration. We show that the spatio-temporal dynamics of gaze shifts can be parsimoniously described by Langevin-type stochastic differential equations triggering a decision equation over time. In particular, value-based patch choice and handling is reduced to a simple multi-alternative perceptual decision making that relies on a race-to-threshold between independent continuous-time perceptual evidence integrators, each integrator being associated with a patch. | ['Giuseppe Boccignone', "Alessandro D'Amelio"] | 2022-03-30 | null | null | null | frontiers-in-neurorobotics-2022-3 | ['eye-tracking'] | ['computer-vision'] | [ 4.34501767e-01 1.21606579e-02 2.69045621e-01 -3.14687341e-02
1.59357116e-01 -6.66104138e-01 5.35128832e-01 1.89100862e-01
-6.76843047e-01 5.88913023e-01 3.71290594e-02 -1.48336828e-01
-4.44985867e-01 -1.24847345e-01 -2.74885803e-01 -9.80943024e-01
-4.94891584e-01 -9.00681168e-02 2.75533050e-01 -4.51035112e-01
5.00180781e-01 2.77115613e-01 -2.15362120e+00 8.43686759e-02
5.40858269e-01 1.04411685e+00 2.45088339e-01 1.19997919e+00
4.70572621e-01 5.30528486e-01 -6.05956316e-01 -3.79668660e-02
9.65404809e-02 -9.02423084e-01 -3.79874736e-01 -1.07449666e-01
-7.69731924e-02 6.00146465e-02 3.43764901e-01 8.72007489e-01
5.42338789e-01 4.14590955e-01 1.06317532e+00 -1.50619698e+00
-6.97261095e-01 2.82029301e-01 -5.58807969e-01 5.70854247e-01
4.52298939e-01 5.88578880e-01 9.34541464e-01 -4.18917328e-01
5.10645032e-01 1.34128666e+00 3.06858689e-01 5.27656615e-01
-1.71233702e+00 -5.62268756e-02 2.66731709e-01 2.46624306e-01
-1.43815732e+00 -5.38719893e-01 6.44651532e-01 -6.98368609e-01
1.03309131e+00 4.78571594e-01 1.37918460e+00 8.77931893e-01
8.15511823e-01 4.19915140e-01 1.28609037e+00 -3.68621409e-01
5.47256887e-01 -1.60347760e-01 -2.22384766e-01 5.37638128e-01
1.26821801e-01 4.05000806e-01 -1.16416025e+00 -5.24179518e-01
7.87482440e-01 -3.47952068e-01 -2.11507708e-01 -4.98761296e-01
-1.26834905e+00 5.83079934e-01 5.20308673e-01 1.52074814e-01
-6.00286722e-01 3.02552074e-01 -2.36528412e-01 6.07181907e-01
1.78164244e-01 6.42206848e-01 -1.10408425e-01 -1.39088094e-01
-5.68135440e-01 3.13002974e-01 5.89039028e-01 3.65379900e-01
6.76931620e-01 -3.44108939e-01 -1.14989191e-01 3.61530095e-01
7.84104466e-01 8.52368712e-01 2.59177387e-02 -1.43509388e+00
-3.21352541e-01 1.49271980e-01 4.36447144e-01 -9.59409058e-01
-5.29077291e-01 -3.15573364e-01 -4.05701518e-01 6.79561555e-01
5.38254797e-01 -5.64616263e-01 -1.91146046e-01 2.11986279e+00
3.65173548e-01 -3.47779900e-01 -2.79049203e-02 9.56594765e-01
-2.19909959e-02 5.15901625e-01 1.99785456e-01 -6.15332544e-01
1.37913370e+00 -2.03037396e-01 -7.34018803e-01 -4.27798182e-02
1.66263029e-01 -4.03608024e-01 9.74358857e-01 3.67462456e-01
-1.39645290e+00 -6.56488240e-02 -9.75919306e-01 8.00920650e-02
-2.88724422e-01 -5.25928557e-01 3.49223346e-01 3.53330225e-01
-1.45773244e+00 4.21307892e-01 -5.07675469e-01 -6.85610116e-01
6.80413023e-02 2.50927418e-01 7.71134719e-02 8.67940068e-01
-9.73099470e-01 8.23905408e-01 -2.33552262e-01 3.97011608e-01
-7.29471326e-01 -1.96045145e-01 -4.62853760e-01 1.24166487e-02
2.04168364e-01 -9.89170074e-01 1.53648698e+00 -1.44890130e+00
-1.73850870e+00 9.33657110e-01 -2.03439131e-01 -4.96359095e-02
2.56592542e-01 2.85620719e-01 3.21000405e-02 4.00857568e-01
-5.46278395e-02 9.71607447e-01 1.24937356e+00 -1.47749114e+00
-5.87829351e-01 -4.52549070e-01 2.41154119e-01 8.41743827e-01
1.94157705e-01 2.31740326e-01 4.73234564e-01 -5.08941054e-01
-1.03412606e-01 -1.09142911e+00 -1.34973019e-01 5.73724568e-01
-6.84933215e-02 -2.09276646e-01 1.75265774e-01 1.05090871e-01
1.33249390e+00 -2.12296271e+00 7.74882555e-01 9.75626558e-02
3.21046293e-01 -4.40624356e-01 3.90646532e-02 6.48318887e-01
3.71169567e-01 -7.93872401e-02 -2.84382254e-01 -4.73288953e-01
2.23249212e-01 -1.68100640e-01 -2.26317167e-01 7.02288091e-01
3.23585987e-01 7.02840865e-01 -1.01782012e+00 -5.89885294e-01
-2.48417571e-01 5.49064577e-01 -6.23974860e-01 2.77377944e-02
-1.35399237e-01 6.51640177e-01 -3.88501287e-01 5.59509277e-01
2.23444536e-01 -1.45259082e-01 8.76476765e-02 3.37124705e-01
-6.69938266e-01 3.96534838e-02 -9.39201534e-01 1.51801038e+00
4.21037413e-02 1.00464666e+00 5.08349240e-01 -1.34549931e-01
2.09933564e-01 -2.78956965e-02 8.95003304e-02 -5.45296788e-01
4.86180335e-01 -3.97736616e-02 5.42196512e-01 -5.79075813e-01
6.27756059e-01 -1.98782861e-01 -8.86889622e-02 9.14844871e-01
-1.07602850e-01 -8.40179622e-02 -7.78872073e-02 4.12847325e-02
6.81936562e-01 1.02124922e-01 3.07022572e-01 -7.72408128e-01
3.48816365e-02 -6.14358485e-02 1.36030093e-01 7.30260551e-01
-5.88156044e-01 1.39588237e-01 7.22547352e-01 1.76097341e-02
-6.28549814e-01 -1.07979751e+00 -2.10426345e-01 1.43656802e+00
6.95919991e-01 2.80746408e-02 -1.00324035e+00 3.46469253e-01
-8.82511586e-02 6.84577942e-01 -1.00906134e+00 -4.73700136e-01
-2.02023566e-01 -6.56813264e-01 2.51000106e-01 -1.53642163e-01
-9.09279361e-02 -1.49111462e+00 -1.99776208e+00 -4.05701250e-02
1.03273034e-01 -2.14601576e-01 -5.37268460e-01 4.23764259e-01
-4.52826291e-01 -9.12510991e-01 -7.17506766e-01 -3.53441268e-01
4.33084875e-01 4.17586893e-01 8.83671403e-01 -6.13824204e-02
-3.20815802e-01 8.88498187e-01 -1.97474286e-02 -5.22913694e-01
-1.87735558e-01 -2.29033336e-01 2.48417571e-01 3.88531446e-01
1.57103047e-01 -5.63797712e-01 -9.76141751e-01 3.35824817e-01
-8.46992552e-01 8.83864518e-03 5.30378334e-02 3.47129464e-01
3.43967557e-01 -4.08340722e-01 1.97314665e-01 2.66870797e-01
1.09097934e+00 -6.49256110e-01 -5.83234847e-01 2.84969509e-01
-1.54921068e-02 7.33998790e-02 -2.36441255e-01 -8.88325512e-01
-1.05637026e+00 -2.25489885e-01 6.70976818e-01 9.52391848e-02
-9.32064503e-02 4.05792803e-01 2.53146410e-01 -2.62093514e-01
8.53313208e-01 -1.08934104e-01 3.27045619e-01 -6.85983822e-02
5.63477159e-01 5.98695338e-01 5.28456047e-02 -3.89344603e-01
2.30676547e-01 5.99500000e-01 2.75798321e-01 -1.20917201e+00
-1.72538236e-01 1.63763717e-01 -6.27100408e-01 -9.05634165e-01
1.18715930e+00 -4.03610319e-01 -1.62620556e+00 6.60673440e-01
-1.04647148e+00 -8.14260066e-01 -5.66301346e-01 3.25077444e-01
-9.36601937e-01 -1.33848175e-01 -2.32424721e-01 -1.58336270e+00
2.85063773e-01 -1.09186196e+00 1.14503229e+00 4.58119869e-01
-4.37789440e-01 -7.51573801e-01 4.52481747e-01 -3.72459143e-01
6.71428382e-01 1.36342555e-01 7.68385589e-01 1.22219875e-01
-7.54783869e-01 3.65733653e-01 1.44426271e-01 -5.21399856e-01
-1.35138750e-01 4.66229022e-01 -1.05762565e+00 -6.47421330e-02
2.52939969e-01 -2.28454456e-01 6.70199573e-01 7.08459139e-01
1.47453636e-01 -1.69237033e-01 -2.52572984e-01 3.37626517e-01
1.08654511e+00 4.42821324e-01 7.41488859e-02 1.31024167e-01
1.00734569e-01 1.35266650e+00 2.71954894e-01 4.55169290e-01
6.10783339e-01 5.64149082e-01 7.14281440e-01 2.50377029e-01
4.34366345e-01 -4.99222009e-03 4.42251414e-01 1.42096147e-01
-1.27696365e-01 -4.60125059e-01 -6.88656092e-01 4.03945029e-01
-1.86110175e+00 -1.16873384e+00 2.09128618e-01 2.42001438e+00
6.58170581e-01 2.10014768e-02 5.75679898e-01 -1.23147130e-01
7.30560541e-01 4.64630723e-02 -8.12249720e-01 -5.64993203e-01
-3.25392842e-01 -4.74738896e-01 2.54851192e-01 8.34121525e-01
-6.69929326e-01 6.53354406e-01 7.96179485e+00 2.08840594e-01
-1.08237004e+00 5.93847185e-02 2.45290518e-01 -5.88112831e-01
-3.57494235e-01 7.29039358e-03 -4.86295342e-01 4.18987453e-01
8.92597437e-01 -1.17080256e-01 8.26315939e-01 -5.30995652e-02
3.96115690e-01 -1.00358188e+00 -9.60526943e-01 8.08551013e-01
-6.05047792e-02 -7.28940547e-01 -4.66737032e-01 3.46319407e-01
2.14560792e-01 -1.35828346e-01 4.61223692e-01 -4.24803138e-01
2.47461900e-01 -8.81662428e-01 1.28064513e+00 8.89021039e-01
5.81282258e-01 -9.18258727e-02 -7.55026266e-02 4.73539650e-01
-9.46886837e-01 -4.72649366e-01 3.52395959e-02 -5.88143706e-01
3.72191846e-01 -1.98552027e-01 -8.42784792e-02 -3.45019311e-01
7.76653886e-01 3.46062958e-01 -3.93222541e-01 1.05165029e+00
1.38983652e-01 -8.78548324e-02 -5.67333221e-01 -4.67212319e-01
-6.23814836e-02 -4.24066842e-01 1.00350153e+00 6.22339249e-01
1.92981675e-01 3.02541614e-01 -4.74190652e-01 1.28916728e+00
6.72636688e-01 -1.67431012e-01 -8.23919594e-01 5.22671528e-02
3.75318617e-01 8.64467680e-01 -9.50787902e-01 4.27065045e-02
5.02596349e-02 8.87387395e-01 2.04489768e-01 6.59037411e-01
-6.19279623e-01 -1.32159472e-01 1.09870744e+00 9.51830372e-02
2.13682309e-01 -4.15691108e-01 2.17547435e-02 -1.04131019e+00
-3.06395292e-01 -4.59236026e-01 4.03400697e-02 -1.07682371e+00
-1.11507428e+00 6.83212817e-01 3.59282464e-01 -9.47780848e-01
-1.31534189e-01 -2.40342632e-01 -3.77959281e-01 9.24861133e-01
-1.16207671e+00 -7.52281368e-01 5.07296957e-02 5.66658795e-01
2.91088801e-02 1.99441671e-01 6.11807764e-01 -4.72669214e-01
-5.45709729e-01 3.30456227e-01 -1.41935140e-01 -9.10368085e-01
3.94063920e-01 -1.22606289e+00 1.05906866e-01 5.31451046e-01
-1.85331717e-01 6.82064414e-01 1.10234952e+00 -5.09258807e-01
-1.27630317e+00 -1.76562175e-01 8.33320796e-01 -4.95418727e-01
8.26633811e-01 -4.78617698e-01 -5.94238818e-01 2.14408159e-01
8.09794128e-01 -4.49903280e-01 8.00092995e-01 -2.84228534e-01
-7.45608937e-03 2.45599061e-01 -1.13563526e+00 9.16250706e-01
1.19876456e+00 -8.07248831e-01 -5.58653474e-01 -1.19362324e-01
5.01505554e-01 -9.41923410e-02 -2.68703669e-01 -1.04045779e-01
1.01959085e+00 -1.27099824e+00 6.26428664e-01 -5.14242768e-01
-1.89742237e-01 -5.49432337e-01 -1.29936472e-01 -1.24212074e+00
-3.01303148e-01 -1.26534462e+00 2.51859695e-01 4.67252433e-01
1.78023785e-01 -9.75729942e-01 7.40523860e-02 4.59947377e-01
4.09209460e-01 -4.29761142e-01 -1.21140909e+00 -2.57261932e-01
-1.50587037e-01 -1.59661382e-01 2.63697833e-01 4.23094004e-01
3.81322682e-01 3.36810112e-01 -4.88067567e-02 5.11875749e-02
6.93691373e-01 -4.59369458e-02 2.88567454e-01 -1.45798743e+00
-4.15525258e-01 -9.59303200e-01 -2.46000335e-01 -1.00547242e+00
-1.83696106e-01 -2.62866855e-01 3.43275607e-01 -9.54347849e-01
-1.38873398e-01 -1.59114659e-01 -3.17262769e-01 3.10450271e-02
1.29662350e-01 2.57999063e-01 3.82900953e-01 4.08638209e-01
-5.80359399e-01 6.36424243e-01 1.15231562e+00 2.60742366e-01
-6.06433153e-01 -6.06069565e-02 -7.30184436e-01 5.72907746e-01
4.39001173e-01 -3.97666305e-01 -5.54221272e-01 -3.30862612e-01
1.14317715e+00 3.44569862e-01 6.58525586e-01 -6.62774324e-01
4.48879421e-01 -4.56833601e-01 -2.71707922e-01 -2.93150336e-01
7.61744499e-01 -7.79475629e-01 1.67129353e-01 4.96506274e-01
-7.17615783e-01 3.85483772e-01 5.77838086e-02 8.36582124e-01
2.60162354e-01 4.48327400e-02 7.19224751e-01 1.24599049e-02
-8.26186463e-02 -2.89573014e-01 -1.16803813e+00 -4.10522185e-02
1.21945202e+00 -5.08361638e-01 -3.45020831e-01 -5.11099815e-01
-1.02404010e+00 3.13321024e-01 9.02710497e-01 1.64864779e-01
2.24457011e-01 -9.13009822e-01 -2.68909425e-01 3.01225543e-01
5.52960113e-03 -4.63961035e-01 3.35778147e-01 9.77279305e-01
-3.94276649e-01 3.78150046e-02 -5.12973070e-01 -8.26122701e-01
-1.10454440e+00 4.71330434e-01 8.12743247e-01 5.30146837e-01
9.28963441e-03 1.06252873e+00 4.54735070e-01 4.10991192e-01
2.16637745e-01 -5.47888458e-01 -4.28142637e-01 5.29756188e-01
5.13498366e-01 4.63727057e-01 -6.67176008e-01 -1.02594781e+00
-4.32509661e-01 9.32473302e-01 7.49725282e-01 -7.71293700e-01
7.64261127e-01 -8.34679186e-01 -4.05513406e-01 1.21367085e+00
5.49552917e-01 -1.53632924e-01 -1.59413338e+00 1.18753076e-01
-3.63273978e-01 -2.79734403e-01 -8.44834000e-02 -7.50551879e-01
-3.20543110e-01 1.03962898e+00 7.12979853e-01 1.02967143e+00
1.13367617e+00 1.82048827e-01 5.86710572e-02 2.97792524e-01
3.75082910e-01 -8.68163168e-01 8.90384912e-02 2.05129817e-01
1.03260756e+00 -1.03303230e+00 -3.08384120e-01 3.79682928e-02
-7.16077030e-01 7.68637538e-01 1.36131272e-01 -1.98824421e-01
1.07853425e+00 1.85235083e-01 -1.76057953e-03 -2.47679800e-01
-1.28093922e+00 -5.71579576e-01 3.29010814e-01 7.88567662e-01
-2.73972936e-02 2.31946230e-01 -2.41752669e-01 1.35125443e-01
-1.13437355e-01 -2.19255343e-01 3.97270828e-01 9.61789489e-01
-7.25020766e-01 -6.39575899e-01 -4.38894391e-01 -1.25669250e-02
-1.94780633e-01 5.10231499e-03 -6.05208099e-01 4.10384834e-01
1.61017686e-01 1.12045979e+00 4.70922947e-01 -1.01524562e-01
1.69266164e-01 -1.97337344e-02 7.61346877e-01 -4.81191754e-01
-7.30085015e-01 4.23489004e-01 -3.74571860e-01 -5.47306240e-01
-8.66750062e-01 -1.26108539e+00 -8.77413452e-01 -5.94746806e-02
-1.09142445e-01 -1.32717686e-02 6.34347439e-01 7.03335881e-01
5.56970179e-01 3.40433925e-01 3.84676546e-01 -1.40359974e+00
-3.82109880e-01 -7.85643995e-01 -7.21257925e-01 -1.76906034e-01
1.05308235e+00 -8.77991319e-01 -7.58079886e-01 9.46938470e-02] | [10.060013771057129, 1.6292059421539307] |
2290e887-098e-4fe4-b7de-901576e27e16 | generating-3d-faces-using-convolutional-mesh | 1807.10267 | null | http://arxiv.org/abs/1807.10267v3 | http://arxiv.org/pdf/1807.10267v3.pdf | Generating 3D faces using Convolutional Mesh Autoencoders | Learned 3D representations of human faces are useful for computer vision
problems such as 3D face tracking and reconstruction from images, as well as
graphics applications such as character generation and animation. Traditional
models learn a latent representation of a face using linear subspaces or
higher-order tensor generalizations. Due to this linearity, they can not
capture extreme deformations and non-linear expressions. To address this, we
introduce a versatile model that learns a non-linear representation of a face
using spectral convolutions on a mesh surface. We introduce mesh sampling
operations that enable a hierarchical mesh representation that captures
non-linear variations in shape and expression at multiple scales within the
model. In a variational setting, our model samples diverse realistic 3D faces
from a multivariate Gaussian distribution. Our training data consists of 20,466
meshes of extreme expressions captured over 12 different subjects. Despite
limited training data, our trained model outperforms state-of-the-art face
models with 50% lower reconstruction error, while using 75% fewer parameters.
We also show that, replacing the expression space of an existing
state-of-the-art face model with our autoencoder, achieves a lower
reconstruction error. Our data, model and code are available at
http://github.com/anuragranj/coma | ['Michael J. Black', 'Anurag Ranjan', 'Timo Bolkart', 'Soubhik Sanyal'] | 2018-07-26 | generating-3d-faces-using-convolutional-mesh-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Anurag_Ranjan_Generating_3D_Faces_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Anurag_Ranjan_Generating_3D_Faces_ECCV_2018_paper.pdf | eccv-2018-9 | ['3d-face-modeling'] | ['computer-vision'] | [-1.54862508e-01 1.34896800e-01 -3.37780118e-02 -5.92409790e-01
-4.02585596e-01 -3.43930811e-01 3.72568548e-01 -5.80595970e-01
2.86274403e-01 2.69593924e-01 2.66333848e-01 3.36061805e-01
3.40378702e-01 -7.28347898e-01 -9.94384050e-01 -5.94634414e-01
-1.69675097e-01 7.54950523e-01 -5.57761252e-01 -8.92805234e-02
-2.40491912e-01 1.00730002e+00 -1.60921407e+00 4.45735365e-01
2.73735583e-01 1.15900087e+00 -4.12387997e-01 4.52821732e-01
-1.21339478e-01 2.07530677e-01 -2.09328786e-01 -5.83745301e-01
2.85417080e-01 -1.05835475e-01 -4.93648231e-01 4.88193899e-01
1.17247641e+00 -5.74950218e-01 -4.08995360e-01 1.06839621e+00
4.24797356e-01 1.21193854e-02 9.43454385e-01 -1.05871320e+00
-9.73831892e-01 -1.45483449e-01 -7.38772810e-01 -4.00175452e-01
3.38693708e-01 -1.56621084e-01 6.82847977e-01 -1.33680975e+00
8.37188721e-01 1.80872488e+00 9.78012621e-01 1.12289202e+00
-1.78557801e+00 -7.33407557e-01 -1.67098776e-01 -2.95388609e-01
-1.51258194e+00 -6.72669649e-01 1.03586149e+00 -5.85775554e-01
7.15246022e-01 1.72771141e-01 8.57229114e-01 1.43514419e+00
2.03406513e-01 6.59956396e-01 1.02557456e+00 -1.26914844e-01
1.13923393e-01 7.49985687e-03 -3.36219192e-01 1.11608577e+00
-1.08109400e-01 -6.47486001e-02 -5.87548494e-01 -6.84266031e-01
1.40916085e+00 -2.97806188e-02 -2.29842782e-01 -4.36087251e-01
-6.37929142e-01 8.91049743e-01 2.29826465e-01 1.57831833e-01
-5.28529346e-01 3.68485928e-01 1.60091758e-01 1.15285076e-01
9.58665848e-01 -3.17182019e-02 -4.83599156e-01 1.99135598e-02
-9.44641590e-01 3.95471811e-01 9.15686190e-01 8.31294775e-01
9.65117216e-01 4.85282034e-01 2.24153221e-01 8.12002480e-01
5.58997393e-01 7.19202638e-01 1.25653028e-01 -1.65369868e+00
-3.11054260e-01 4.33992386e-01 -1.64947674e-01 -1.06814265e+00
-1.49908796e-01 -1.94414988e-01 -9.15516675e-01 5.10237932e-01
1.94657266e-01 -5.18151596e-02 -9.14028287e-01 1.97803414e+00
5.59966266e-01 5.68130136e-01 -4.12045807e-01 7.99712837e-01
7.77643681e-01 5.20119190e-01 -2.20760569e-01 -2.31768504e-01
1.31080258e+00 -2.23207131e-01 -5.95089078e-01 1.33054540e-01
9.99031961e-03 -7.83314407e-01 7.60587811e-01 3.62040967e-01
-1.42956591e+00 -4.25591290e-01 -5.83414078e-01 -2.20695928e-01
1.37486264e-01 2.43336689e-02 7.31885731e-01 4.56990868e-01
-1.37116122e+00 7.79114246e-01 -1.02025521e+00 -3.38178962e-01
6.87779725e-01 4.31765646e-01 -6.24367833e-01 1.14885971e-01
-6.55135989e-01 5.73580623e-01 -6.24022007e-01 -2.34575979e-02
-9.01404798e-01 -1.05575264e+00 -1.00518370e+00 -7.91815072e-02
-1.98406488e-01 -9.62504923e-01 1.22102356e+00 -1.01316512e+00
-1.79049563e+00 1.21699309e+00 -5.96495152e-01 6.23926483e-02
2.89728075e-01 -2.00583577e-01 -7.30387941e-02 2.80508220e-01
-2.85785198e-01 7.20028579e-01 1.38613009e+00 -1.41606891e+00
5.55035233e-01 -7.30576158e-01 -1.95416898e-01 -9.31501612e-02
-2.81155854e-01 5.35950102e-02 -4.26675230e-01 -6.63757682e-01
9.45441797e-02 -1.13294542e+00 2.75696889e-02 8.23943019e-01
-6.94596022e-02 -1.45836368e-01 1.09163010e+00 -6.22096717e-01
5.06429613e-01 -2.05884862e+00 5.56326568e-01 1.94518968e-01
4.17071939e-01 -1.10655673e-01 -1.18500717e-01 5.44780381e-02
-1.76839426e-01 1.24311239e-01 -2.82757163e-01 -1.01431608e+00
1.78808207e-03 2.88142294e-01 -2.09591180e-01 8.12076867e-01
3.70864868e-01 8.33829939e-01 -4.89811122e-01 -3.72440517e-01
2.12381303e-01 1.33389413e+00 -9.36169624e-01 1.82769701e-01
-6.02334917e-01 7.31969774e-01 -4.16439533e-01 8.47292185e-01
8.77862990e-01 -2.83106476e-01 8.60995203e-02 -3.93627554e-01
3.00043344e-01 -2.86349028e-01 -9.99684870e-01 1.86917067e+00
-5.25127769e-01 5.96566021e-01 6.81349993e-01 -8.37626040e-01
8.27701986e-01 5.78416109e-01 8.88809443e-01 -9.64766368e-02
3.31609279e-01 1.09079733e-01 -5.85819244e-01 -3.37554634e-01
-1.17280073e-02 -4.39722985e-01 3.57765555e-01 3.54705751e-01
2.23104998e-01 -3.49342585e-01 -4.20916378e-01 -5.13029397e-02
6.93398118e-01 2.72666246e-01 -2.71930933e-01 -2.87768245e-01
2.83356309e-01 -6.26621485e-01 5.42489111e-01 -5.38843907e-02
4.63366285e-02 6.61159992e-01 3.59480917e-01 -6.64910972e-01
-1.13428819e+00 -1.16568637e+00 -5.70052743e-01 7.73508430e-01
-4.53519523e-01 -4.14909989e-01 -8.47857833e-01 -1.07941531e-01
4.17432457e-01 3.32979620e-01 -1.02720428e+00 1.09970652e-01
-6.86824083e-01 -3.63560587e-01 6.12118840e-01 3.78088742e-01
1.72743335e-01 -7.96051323e-01 -1.35524869e-01 -1.54665843e-01
1.49723571e-02 -1.19889462e+00 -5.33758461e-01 -6.05292380e-01
-1.05236816e+00 -8.34231496e-01 -7.80112803e-01 -7.28494883e-01
9.95628059e-01 -2.55050957e-01 1.30308831e+00 2.03880414e-01
-4.33572799e-01 9.16499078e-01 2.94904530e-01 -2.95684457e-01
-5.73490441e-01 -5.35688400e-01 6.11617804e-01 3.69501114e-01
2.71400064e-01 -1.08302677e+00 -4.89288747e-01 1.84115916e-01
-6.96168303e-01 -1.18693709e-01 1.11117043e-01 7.46668458e-01
9.05743122e-01 -5.01255810e-01 -4.41361852e-02 -6.09550476e-01
4.07052636e-01 -4.50749308e-01 -5.81341028e-01 -5.53732589e-02
-1.48925900e-01 2.09865645e-01 4.10735518e-01 -7.50662029e-01
-1.01969075e+00 1.57555073e-01 -2.56987482e-01 -1.35334611e+00
-1.22816734e-01 1.86093450e-02 8.60009566e-02 -4.97059673e-01
7.05016851e-01 4.38299477e-02 5.83489358e-01 -6.79614246e-01
3.08911026e-01 3.49163562e-01 3.13238949e-01 -1.02828956e+00
8.18757236e-01 8.62297475e-01 4.31628764e-01 -1.32114899e+00
-4.33860123e-01 1.74517781e-02 -6.08638108e-01 -3.74301195e-01
7.73424447e-01 -9.44467008e-01 -1.06959856e+00 5.43424487e-01
-1.43273747e+00 -4.48198825e-01 -2.40841150e-01 2.38199264e-01
-8.28257978e-01 2.31760100e-01 -9.54592586e-01 -8.57168734e-01
-4.48902935e-01 -1.07557440e+00 1.63516867e+00 -6.69802576e-02
-3.35983515e-01 -9.76546049e-01 1.06782451e-01 2.27024913e-01
4.04476255e-01 6.71304286e-01 8.04494679e-01 1.97702810e-01
-4.71438676e-01 4.13684770e-02 2.08568305e-01 4.63248193e-01
1.02825463e-01 2.91201591e-01 -1.05573535e+00 -3.75016034e-01
1.73042282e-01 -5.45993030e-01 5.75108707e-01 5.28352141e-01
1.43459451e+00 -4.95609254e-01 -6.25103489e-02 1.10593998e+00
1.16451085e+00 -4.24551636e-01 3.66447479e-01 -6.18654370e-01
8.20104361e-01 6.50565684e-01 -3.55051190e-01 6.10309720e-01
2.57082075e-01 8.11865032e-01 4.55451131e-01 2.35831141e-01
-1.86073884e-01 -2.35672399e-01 4.77078944e-01 8.68136108e-01
-4.40116584e-01 4.15907532e-01 -7.44924903e-01 2.51148254e-01
-1.38440549e+00 -1.02166319e+00 2.62567133e-01 1.99528658e+00
7.90261447e-01 -3.49884868e-01 5.83708212e-02 -3.19755554e-01
4.25078392e-01 6.87247440e-02 -8.25718343e-01 -5.22200823e-01
5.79544902e-03 5.33221245e-01 -2.66186502e-02 6.39324546e-01
-6.79796040e-01 8.04488182e-01 6.62214708e+00 4.83612955e-01
-1.49167502e+00 1.02491893e-01 5.63588798e-01 -2.61417925e-01
-6.14613950e-01 -4.50888842e-01 -5.57757735e-01 1.34276614e-01
8.76648188e-01 -3.92339975e-02 7.64023781e-01 8.55021656e-01
9.55431908e-02 6.50487483e-01 -1.19977069e+00 1.23454070e+00
2.78620392e-01 -1.53651726e+00 2.50177234e-01 3.97603363e-01
8.40622663e-01 1.99811067e-02 3.43669653e-01 1.24521479e-02
1.13188148e-01 -1.33406973e+00 6.64870083e-01 7.80103445e-01
1.25557685e+00 -5.08331716e-01 -5.76449707e-02 2.52456307e-01
-1.05653524e+00 3.54224682e-01 -3.85352880e-01 4.59867381e-02
5.46977744e-02 5.03257453e-01 -3.50366741e-01 8.85680467e-02
7.38310754e-01 7.11154222e-01 -6.41726255e-02 2.25030452e-01
5.05570695e-02 5.77116132e-01 -5.97535729e-01 2.65166104e-01
-2.83921301e-01 -3.56620371e-01 7.68003047e-01 8.49664330e-01
4.08137560e-01 3.70139897e-01 8.22595283e-02 1.33861578e+00
-5.61873794e-01 -8.37707252e-04 -7.43974268e-01 -8.21526125e-02
3.39275062e-01 1.18938756e+00 -1.23916462e-01 -6.98550269e-02
-3.92747641e-01 1.03041065e+00 4.83332723e-01 5.07459164e-01
-6.19255960e-01 3.91584575e-01 1.35670388e+00 5.57688177e-01
2.34365329e-01 -6.34602964e-01 -1.30163863e-01 -1.32103395e+00
1.34239309e-02 -8.67787659e-01 -2.18581855e-01 -8.49338174e-01
-1.41820550e+00 6.80325389e-01 6.07591867e-02 -9.16845620e-01
-4.15599585e-01 -9.50412035e-01 -4.49646950e-01 8.23654473e-01
-9.08876121e-01 -1.42253196e+00 -3.36839080e-01 7.33538985e-01
3.68643641e-01 -1.37496099e-01 1.27744353e+00 2.07523584e-01
-3.03489208e-01 6.48482561e-01 -1.50395676e-01 2.62707740e-01
4.53406483e-01 -9.74474013e-01 4.83647287e-01 1.95110112e-01
3.02405894e-01 9.21490669e-01 6.94148242e-01 -5.15287519e-01
-1.97719753e+00 -9.72824752e-01 3.92880946e-01 -5.80402493e-01
3.39046389e-01 -4.39086199e-01 -1.01571655e+00 1.02870643e+00
-1.33157939e-01 5.97076178e-01 7.21706748e-01 1.38039261e-01
-6.24608457e-01 1.10919200e-01 -1.41129935e+00 5.59791028e-01
1.19938171e+00 -8.10752630e-01 -1.83068320e-01 2.52984643e-01
4.24116790e-01 -6.97760880e-01 -1.24978542e+00 4.62259680e-01
9.30089056e-01 -8.81215036e-01 1.10150695e+00 -6.41837835e-01
4.76238817e-01 5.82275540e-02 -4.04233009e-01 -1.25060844e+00
-3.88443053e-01 -8.44640672e-01 -7.51601696e-01 8.06803524e-01
-1.16836704e-01 -4.13458675e-01 1.04411876e+00 7.72375286e-01
1.10708065e-01 -1.25961018e+00 -1.11508143e+00 -5.15408158e-01
4.45248991e-01 -5.37142873e-01 6.42655253e-01 9.59916592e-01
-4.34655249e-01 7.83690512e-02 -3.23687077e-01 1.44470885e-01
1.10283601e+00 1.23454392e-01 7.92043149e-01 -1.40798175e+00
-2.86222190e-01 -4.47965235e-01 -5.25525928e-01 -1.16276491e+00
8.70917499e-01 -1.01710236e+00 -4.10105318e-01 -9.85457182e-01
8.69120657e-02 -1.70994118e-01 3.55875671e-01 3.92985553e-01
4.22465950e-01 4.25339282e-01 4.58765924e-02 1.04025774e-01
1.17974721e-01 9.17894423e-01 1.39287043e+00 -1.24553651e-01
1.32644162e-01 -2.58929580e-01 -4.79340941e-01 1.13372636e+00
5.18324137e-01 -1.65009797e-01 -2.60959387e-01 -7.25715339e-01
1.18711181e-02 2.23172307e-01 5.66814244e-01 -6.66284382e-01
-6.32671341e-02 -1.48872718e-01 8.53684664e-01 -2.98310332e-02
9.95080173e-01 -7.01295674e-01 5.02051473e-01 6.65477961e-02
-2.29077898e-02 -3.11986334e-03 4.09934044e-01 3.82317841e-01
1.54844120e-01 1.43083617e-01 9.21566129e-01 -2.44167775e-01
-1.23870321e-01 9.52365756e-01 -6.29668385e-02 -6.18999414e-02
6.91627622e-01 -1.52797878e-01 1.03566580e-01 -5.84205806e-01
-8.03390265e-01 -2.65097678e-01 9.16298330e-01 2.77441263e-01
7.89853632e-01 -1.74927390e+00 -9.40404117e-01 6.97409689e-01
-2.18743101e-01 1.38932690e-01 2.64399558e-01 5.01332343e-01
-4.72214818e-01 -1.52999148e-01 -2.84645379e-01 -7.86736310e-01
-1.37616193e+00 1.02851905e-01 6.37410045e-01 2.85437912e-01
-6.72876120e-01 1.09960139e+00 2.49085441e-01 -5.93154311e-01
8.05047303e-02 -1.22708052e-01 2.39796251e-01 -1.88485757e-01
5.00702381e-01 2.25749835e-01 -2.04880938e-01 -1.27023578e+00
-3.55625480e-01 1.14346623e+00 2.96385825e-01 -5.14421314e-02
1.41871321e+00 3.49583566e-01 -3.83358270e-01 4.76076722e-01
1.57546401e+00 1.87994346e-01 -1.42094076e+00 -1.60505190e-01
-6.80073857e-01 -4.76651877e-01 9.16132331e-02 -1.57145113e-01
-1.30804515e+00 9.49857593e-01 3.75421494e-01 -1.74800605e-01
9.20361876e-01 2.24871308e-01 8.23621094e-01 1.28474295e-01
5.14181197e-01 -5.20096660e-01 2.47388735e-01 4.21890259e-01
1.40888214e+00 -1.00192750e+00 7.70964772e-02 -4.61149395e-01
-2.91880190e-01 1.18305731e+00 4.72412556e-01 -3.89186651e-01
1.13735032e+00 4.12242502e-01 -4.55008773e-03 -4.64073598e-01
-7.82807350e-01 3.58602047e-01 4.57126558e-01 4.56379831e-01
5.93262255e-01 1.76117346e-01 3.73320401e-01 3.28404397e-01
-3.74397099e-01 7.49800950e-02 1.96891874e-01 5.87629676e-01
3.10322288e-02 -1.06667256e+00 -4.78267819e-01 3.74037027e-01
-5.35809875e-01 2.78085411e-01 -2.22652644e-01 2.39721701e-01
7.10844994e-02 3.67114037e-01 3.30732405e-01 -3.79780829e-01
1.92317277e-01 3.06356370e-01 1.09885430e+00 -4.46203351e-01
2.44318824e-02 7.90217519e-02 -1.88866898e-01 -7.64377654e-01
-4.25191820e-01 -9.51485872e-01 -1.21927345e+00 -7.94876456e-01
1.23310424e-01 -2.09758371e-01 7.21819043e-01 4.97305453e-01
6.60516500e-01 4.66345176e-02 6.47871912e-01 -1.57044172e+00
-6.54099822e-01 -8.52189898e-01 -6.14639938e-01 7.69303143e-01
4.97521341e-01 -9.54461336e-01 -4.25537765e-01 2.17506200e-01] | [13.049311637878418, -0.03677217662334442] |
179cfaf7-9e27-4dfe-b10a-888ed8b9001b | a-two-stage-bayesian-optimisation-for | 2206.15115 | null | https://arxiv.org/abs/2206.15115v1 | https://arxiv.org/pdf/2206.15115v1.pdf | A Two-Stage Bayesian Optimisation for Automatic Tuning of an Unscented Kalman Filter for Vehicle Sideslip Angle Estimation | This paper presents a novel methodology to auto-tune an Unscented Kalman Filter (UKF). It involves using a Two-Stage Bayesian Optimisation (TSBO), based on a t-Student Process to optimise the process noise parameters of a UKF for vehicle sideslip angle estimation. Our method minimises performance metrics, given by the average sum of the states' and measurement' estimation error for various vehicle manoeuvres covering a wide range of vehicle behaviour. The predefined cost function is minimised through a TSBO which aims to find a location in the feasible region that maximises the probability of improving the current best solution. Results on an experimental dataset show the capability to tune the UKF in 79.9% less time than using a genetic algorithm (GA) and the overall capacity to improve the estimation performance in an experimental test dataset of 9.9% to the current state-of-the-art GA. | ['R. Happee', 'M. Alirezaei', 'B. Shyrokau', 'A. Bertipaglia'] | 2022-06-30 | null | null | null | null | ['bayesian-optimisation'] | ['methodology'] | [ 2.36318022e-01 3.21403667e-02 3.06151628e-01 -2.22756028e-01
-9.01379883e-01 -3.75115097e-01 9.28798974e-01 -9.07531455e-02
-5.85830986e-01 1.01401627e+00 -5.03869019e-02 -5.65621138e-01
-5.97686052e-01 -6.15569115e-01 -6.70730114e-01 -1.10915089e+00
1.53951170e-02 5.61517954e-01 3.31739515e-01 -8.58361498e-02
5.39831161e-01 5.72964311e-01 -1.66662216e+00 -5.41207910e-01
7.62026191e-01 7.98821390e-01 3.73322219e-01 1.03497362e+00
3.46899480e-01 1.66877173e-02 -6.86910987e-01 -9.78250876e-02
2.13250592e-01 -3.19131762e-01 -1.78955361e-01 7.70666301e-02
-2.52397195e-03 2.69414574e-01 1.57289460e-01 1.11917329e+00
6.95931375e-01 5.34163773e-01 8.37039590e-01 -1.12561011e+00
6.11342430e-01 -1.23335265e-01 -8.02907199e-02 2.95776337e-01
1.34941369e-01 3.55717301e-01 1.68030202e-01 -4.12562132e-01
3.70065719e-01 1.24348629e+00 5.60132146e-01 1.95055187e-01
-1.33862484e+00 -7.16986001e-01 -5.52204430e-01 1.99874952e-01
-1.44122350e+00 -4.56937939e-01 2.21607134e-01 -5.01645803e-01
9.23702657e-01 2.39506632e-01 8.03874791e-01 3.53254765e-01
9.15535033e-01 1.40684351e-01 1.29954982e+00 -2.59936512e-01
6.64069593e-01 1.44639879e-01 -4.92306203e-01 3.24589550e-01
3.44463229e-01 6.85068548e-01 -1.16717279e-01 -7.97543898e-02
4.22579080e-01 -6.12166822e-01 6.25292361e-02 -5.53312123e-01
-8.91758025e-01 9.56132352e-01 -1.91102520e-01 -7.64652565e-02
-5.12281060e-01 2.38580972e-01 3.42184126e-01 2.84035891e-01
3.27978164e-01 5.35433233e-01 -6.99256003e-01 -5.96682906e-01
-9.44851756e-01 4.89660084e-01 9.35121059e-01 3.71892899e-01
7.09733009e-01 3.77439648e-01 1.19589493e-01 3.32151711e-01
8.18914175e-01 1.09449732e+00 9.89216566e-02 -7.87042797e-01
1.16630577e-01 4.14937884e-01 3.72040480e-01 -7.38169611e-01
-3.80887806e-01 -4.80999917e-01 -3.05981636e-01 5.32212734e-01
3.04978162e-01 -8.60295892e-01 -9.98219430e-01 1.21677065e+00
5.26898086e-01 6.72910929e-01 5.36769331e-01 5.36471128e-01
2.80666262e-01 9.30183887e-01 -1.04872525e-01 -6.68059289e-01
1.11924553e+00 -3.74287665e-01 -8.37246537e-01 -2.13523671e-01
5.65839171e-01 -1.13161361e+00 1.59083307e-01 5.41406393e-01
-7.78893709e-01 -5.92351615e-01 -1.27083099e+00 1.04325783e+00
-2.23505497e-01 6.37939647e-02 7.97738954e-02 1.22348523e+00
-8.68207455e-01 5.12711763e-01 -7.70028532e-01 -1.08570166e-01
-1.68364376e-01 5.81125021e-01 -1.10423751e-01 2.60854572e-01
-1.17340124e+00 1.42561555e+00 7.72877276e-01 3.10948372e-01
-1.05027187e+00 -7.07755983e-01 -8.69587600e-01 -3.06879938e-01
6.32457137e-01 -4.81899858e-01 1.05775142e+00 -5.90381622e-01
-2.12181067e+00 2.04426736e-01 -1.98125377e-01 -8.20462525e-01
5.04645228e-01 -2.26979509e-01 -5.77108204e-01 -2.11018458e-01
-3.37887943e-01 3.61160427e-01 1.12450838e+00 -1.13916206e+00
-9.01110172e-01 -6.81008771e-02 -5.11753738e-01 1.66614369e-01
5.64751148e-01 -1.74146831e-01 -2.68394887e-01 -1.76112086e-01
-2.13693976e-01 -1.43371165e+00 -4.98812258e-01 -8.57807338e-01
1.61262736e-01 -3.00406575e-01 1.02885139e+00 -6.15507007e-01
1.29319131e+00 -1.88646579e+00 1.39655858e-01 7.03255117e-01
-5.54277539e-01 7.80862927e-01 2.78369009e-01 5.55485666e-01
1.85577676e-01 -4.58082110e-01 -1.12864263e-01 5.63231409e-02
-7.41355270e-02 5.18136621e-01 2.91634426e-02 8.91613781e-01
1.35563031e-01 5.35657465e-01 -9.16113138e-01 -1.26632214e-01
7.80106962e-01 6.10776722e-01 -2.39902422e-01 1.59629002e-01
-9.25657600e-02 3.43545616e-01 -3.24014515e-01 -1.12710789e-01
6.94386840e-01 6.32913172e-01 2.30960310e-01 3.26700322e-02
-5.13260961e-01 -1.06323928e-01 -1.79507542e+00 9.10691619e-01
-4.41950142e-01 8.30473006e-01 -9.91307572e-03 -8.27042937e-01
1.47334850e+00 3.37146521e-01 1.89000979e-01 -3.77566695e-01
5.88508785e-01 2.79430449e-01 1.21817783e-01 -4.14905518e-01
3.24007988e-01 -1.88059792e-01 -9.56815183e-02 -8.97457823e-02
-5.94129153e-02 -5.83246648e-01 1.19441658e-01 -1.53616413e-01
6.69023812e-01 1.13182068e-01 3.07538956e-01 -6.87114060e-01
1.03282785e+00 -3.47080767e-01 4.00052369e-01 5.90012372e-01
-6.18261546e-02 -2.14755133e-01 4.00187522e-01 -1.99162617e-01
-1.01310086e+00 -8.09645474e-01 -2.09940657e-01 2.68288583e-01
6.46119490e-02 -8.83496404e-02 -5.89635611e-01 -2.68607169e-01
-2.29141843e-02 1.16484201e+00 -5.61047852e-01 -2.65537322e-01
-4.28905487e-01 -8.22720587e-01 1.08161665e-01 1.56830028e-02
2.56241620e-01 -6.09350681e-01 -1.01863241e+00 4.08170760e-01
3.76645207e-01 -8.42639506e-01 4.85619493e-02 1.81703389e-01
-7.89072275e-01 -8.08287621e-01 -2.55724162e-01 -3.42201114e-01
5.78280985e-01 -3.39631736e-01 6.52007759e-01 -2.83541948e-01
8.37939829e-02 1.94205433e-01 -1.98923618e-01 -9.26877856e-01
-8.77297103e-01 -1.34920016e-01 6.11685701e-02 2.61480305e-02
3.63505393e-01 1.06938854e-01 -4.08051223e-01 7.03874290e-01
-4.35176164e-01 -4.14844275e-01 6.11979008e-01 7.09662676e-01
3.93396556e-01 7.79488802e-01 5.06172419e-01 -3.51567358e-01
5.70611596e-01 -2.42768854e-01 -1.49744439e+00 -2.13236153e-01
-8.15752745e-01 3.12814683e-01 1.76366299e-01 -2.79284358e-01
-1.02565646e+00 2.42798701e-01 -6.53188527e-02 -1.14221290e-01
-2.51540869e-01 3.48282337e-01 -1.01351343e-01 -4.67660815e-01
3.78942698e-01 1.47030324e-01 4.95407432e-01 -1.36227846e-01
2.52656281e-01 6.31905913e-01 4.54308033e-01 -8.34158137e-02
9.96798158e-01 1.41147658e-01 6.94975257e-01 -9.67729449e-01
-2.25371063e-01 -8.88549984e-01 -2.89775580e-01 -7.36193120e-01
7.51434624e-01 -7.37380922e-01 -1.07532346e+00 4.88159567e-01
-5.04342914e-01 -1.40270442e-01 -1.86013104e-03 8.52186799e-01
-7.22275734e-01 1.98359400e-01 3.16833884e-01 -1.31756377e+00
-2.02392414e-01 -1.37545037e+00 9.47067678e-01 3.83460522e-01
-1.73805475e-01 -1.12241471e+00 2.22203851e-01 1.58274293e-01
3.85539204e-01 4.58867997e-01 2.57555515e-01 -3.41201544e-01
-4.60900426e-01 -5.01933157e-01 2.40576997e-01 5.21518648e-01
-3.21874380e-01 1.16255969e-01 -4.54307795e-01 -6.94332540e-01
-5.17503638e-03 2.87686706e-01 6.45977914e-01 6.19829297e-01
1.45793572e-01 -1.54965430e-01 -4.63691801e-01 3.08640987e-01
1.84449255e+00 7.36039639e-01 9.06641722e-01 6.32581294e-01
5.06148189e-02 3.50624800e-01 1.19134629e+00 4.12595600e-01
-9.77061018e-02 7.72848010e-01 4.99958307e-01 2.67254800e-01
3.77053499e-01 9.21040550e-02 3.53787959e-01 3.66962641e-01
1.46855842e-02 -1.45620078e-01 -8.76738131e-01 6.78155601e-01
-1.80510592e+00 -7.74568439e-01 -3.42292070e-01 2.49583721e+00
4.10367370e-01 6.34658575e-01 7.01768547e-02 4.06011313e-01
5.28836071e-01 -1.73551425e-01 3.12628248e-03 -9.59733903e-01
5.02534747e-01 1.31324574e-01 1.35146832e+00 8.44385862e-01
-1.08635628e+00 5.26951134e-01 5.65880251e+00 1.09576511e+00
-8.21920931e-01 -4.11590189e-01 -1.38814196e-01 1.89929023e-01
1.57958359e-01 2.19243929e-01 -1.14501846e+00 5.32031834e-01
1.72618473e+00 -2.21502170e-01 3.26808065e-01 4.25637752e-01
7.96910167e-01 -7.80570269e-01 -2.46968895e-01 5.46787798e-01
6.44580349e-02 -1.08297205e+00 -4.32503223e-01 3.60766709e-01
1.02973783e+00 -1.34667724e-01 -1.40055701e-01 2.55507231e-01
2.48475552e-01 -8.01199496e-01 5.07915676e-01 9.33064342e-01
2.50840098e-01 -1.42890477e+00 1.35188437e+00 4.91877258e-01
-9.62037086e-01 -2.35237196e-01 -8.46421719e-02 4.52292375e-02
6.30630672e-01 4.27958429e-01 -1.60352159e+00 6.59290910e-01
3.62915128e-01 9.58743244e-02 -1.19686186e-01 1.57811701e+00
-1.72593802e-01 6.76649868e-01 -6.77831709e-01 -5.31813681e-01
7.39264786e-01 -2.78063089e-01 1.12528110e+00 1.20843267e+00
3.40873063e-01 -2.68497616e-01 -8.66464898e-02 -3.95550802e-02
8.51467073e-01 1.29440511e-02 -2.25444227e-01 1.67371854e-01
4.78994071e-01 8.54773343e-01 -6.27087891e-01 -3.27778190e-01
1.58958539e-01 1.04895001e-02 -6.58181846e-01 1.65224671e-01
-7.60936081e-01 -5.83865702e-01 5.55910826e-01 -3.61921871e-03
9.01265919e-01 -2.45828584e-01 6.97511109e-03 2.09921245e-02
-4.07309443e-01 -6.20812654e-01 7.15811178e-02 -5.62997818e-01
-4.54616189e-01 2.75502712e-01 4.72250730e-01 -1.18917477e+00
-7.87416577e-01 -5.01754582e-01 -3.85042250e-01 1.05679798e+00
-1.17104340e+00 -6.38706625e-01 1.65916920e-01 -9.65053309e-03
5.36932111e-01 -4.19853419e-01 6.31009161e-01 -9.71790031e-02
-1.90617248e-01 1.01809770e-01 7.70069003e-01 -4.56758827e-01
3.02983701e-01 -1.18882298e+00 3.67867887e-01 9.70723867e-01
-6.19935989e-01 3.62910062e-01 1.66416144e+00 -7.76091397e-01
-1.44529414e+00 -1.00920177e+00 9.07680809e-01 -2.72935539e-01
6.97903216e-01 1.31074646e-02 -4.54250723e-01 1.45906314e-01
9.34548527e-02 -3.88158530e-01 7.11069703e-02 -2.69344956e-01
5.44109404e-01 -1.82136297e-01 -1.15012538e+00 4.60009634e-01
6.99334368e-02 -4.89734709e-02 -7.13339388e-01 -1.29134759e-01
2.36270297e-02 -6.27155483e-01 -1.06731260e+00 8.21081936e-01
7.36607134e-01 -6.22782290e-01 7.55958617e-01 -1.26483500e-01
-5.47976553e-01 -8.81087482e-01 1.89819455e-01 -1.68998849e+00
8.91963020e-02 -1.15456450e+00 2.09532641e-02 8.72571588e-01
3.36876243e-01 -7.83722878e-01 6.91166461e-01 9.35662631e-03
1.73984468e-01 -8.86701941e-01 -1.40051591e+00 -7.35018373e-01
-4.47597623e-01 -6.45900667e-01 3.45743597e-01 -2.53653899e-03
-5.85349917e-01 2.68839151e-01 -3.11181515e-01 4.94437605e-01
8.38853538e-01 -4.60970640e-01 1.13863659e+00 -1.23337340e+00
2.08028089e-02 -1.93887189e-01 -9.81963158e-01 -5.33627331e-01
1.05500922e-01 -2.24718198e-01 4.74911094e-01 -1.33443677e+00
-4.47495580e-01 -1.16886370e-01 1.07495315e-01 -3.81482929e-01
-8.83861631e-02 -5.44405840e-02 3.26174125e-02 -4.92920727e-01
-2.53485948e-01 2.51751721e-01 9.59726453e-01 1.85496643e-01
-3.80277365e-01 7.17712879e-01 -7.95717724e-03 4.96823877e-01
8.12558651e-01 -7.32746303e-01 -6.01614296e-01 3.72867733e-01
2.72145182e-01 2.67516226e-01 2.49763951e-01 -1.24469507e+00
1.41177282e-01 -7.95860067e-02 2.45998397e-01 -1.14573514e+00
2.95235932e-01 -1.06603479e+00 8.52268755e-01 8.30563009e-01
1.73104301e-01 1.08260550e-01 6.11128569e-01 8.96947920e-01
8.23425427e-02 -5.48239112e-01 9.42062140e-01 3.03422183e-01
-9.93914902e-01 -3.96279991e-01 -8.72165561e-01 -4.53665167e-01
1.39531672e+00 -5.83337903e-01 1.26791447e-01 -4.67846125e-01
-4.85932589e-01 4.50548440e-01 1.97819740e-01 4.07980263e-01
3.34276915e-01 -7.53733754e-01 -6.38090491e-01 5.00673532e-01
-2.81233907e-01 -4.15653259e-01 2.38064200e-01 7.26806104e-01
-5.78642666e-01 8.08388531e-01 -2.89572100e-03 -7.59886324e-01
-1.46399605e+00 2.29753792e-01 5.47257185e-01 -2.45627895e-01
-1.76668510e-01 5.08419693e-01 -7.60097563e-01 -3.20318639e-01
-1.25532195e-01 -5.49492799e-02 -4.33021098e-01 7.12525547e-02
2.69974887e-01 8.57278943e-01 3.63691479e-01 -9.49847102e-01
-2.86749035e-01 7.43152499e-01 5.37725091e-01 -3.33299547e-01
9.98116314e-01 -2.73854762e-01 2.72777289e-01 2.47665361e-01
1.25659835e+00 -1.82162933e-02 -1.63425934e+00 3.04311454e-01
-1.14125956e-03 -3.82725656e-01 6.92744851e-01 -8.01785409e-01
-5.19056082e-01 1.15548186e-01 9.93518710e-01 7.58045092e-02
7.83100188e-01 -4.74743605e-01 4.66771752e-01 1.67965323e-01
2.27047831e-01 -1.19823503e+00 -7.13888526e-01 7.04759777e-01
4.57461029e-01 -8.23015690e-01 3.17885101e-01 -1.63661659e-01
-5.87398291e-01 1.20863116e+00 2.98953149e-02 -2.30202764e-01
8.85257900e-01 3.83705020e-01 1.48009464e-01 4.16575149e-02
-7.74291813e-01 -3.78898501e-01 4.36018705e-01 4.04901177e-01
-2.82917947e-01 2.43908197e-01 -6.83670104e-01 -4.44458902e-01
-2.84424186e-01 1.43384427e-01 2.98660368e-01 8.08526993e-01
-7.84763038e-01 -8.28668118e-01 -6.82661116e-01 4.94306237e-01
-3.08960289e-01 3.93491715e-01 4.19526994e-01 8.37112248e-01
2.15597615e-01 1.11744130e+00 1.49345502e-01 -2.64061868e-01
5.68016231e-01 7.02611506e-02 2.20487863e-01 -1.78183019e-02
-4.45100635e-01 2.33624712e-01 4.01293814e-01 -3.75650734e-01
-5.84589243e-01 -1.08821952e+00 -8.35769475e-01 -2.88872421e-01
-7.49565125e-01 5.99099338e-01 1.20112693e+00 1.24166608e+00
-6.25876412e-02 5.48944890e-01 7.35684276e-01 -8.88140559e-01
-6.80666625e-01 -7.31626987e-01 -3.01547796e-01 -3.41182113e-01
3.62155557e-01 -9.94243264e-01 -6.00407302e-01 -2.00088143e-01] | [5.202939033508301, 2.1534390449523926] |
cf7b2e55-a328-4f25-9f2e-7c5ef3680695 | improving-heterogeneous-graph-learning-with | 2307.04514 | null | https://arxiv.org/abs/2307.04514v1 | https://arxiv.org/pdf/2307.04514v1.pdf | Improving Heterogeneous Graph Learning with Weighted Mixed-Curvature Product Manifold | In graph representation learning, it is important that the complex geometric structure of the input graph, e.g. hidden relations among nodes, is well captured in embedding space. However, standard Euclidean embedding spaces have a limited capacity in representing graphs of varying structures. A promising candidate for the faithful embedding of data with varying structure is product manifolds of component spaces of different geometries (spherical, hyperbolic, or euclidean). In this paper, we take a closer look at the structure of product manifold embedding spaces and argue that each component space in a product contributes differently to expressing structures in the input graph, hence should be weighted accordingly. This is different from previous works which consider the roles of different components equally. We then propose WEIGHTED-PM, a data-driven method for learning embedding of heterogeneous graphs in weighted product manifolds. Our method utilizes the topological information of the input graph to automatically determine the weight of each component in product spaces. Extensive experiments on synthetic and real-world graph datasets demonstrate that WEIGHTED-PM is capable of learning better graph representations with lower geometric distortion from input data, and performs better on multiple downstream tasks, such as word similarity learning, top-$k$ recommendation, and knowledge graph embedding. | ['The-Anh Ta', 'Dung D. Le', 'Tuc Nguyen-Van'] | 2023-07-10 | null | null | null | null | ['graph-embedding', 'graph-learning', 'knowledge-graph-embedding', 'representation-learning', 'graph-representation-learning', 'word-similarity'] | ['graphs', 'graphs', 'graphs', 'methodology', 'methodology', 'natural-language-processing'] | [-1.89721137e-01 4.44537729e-01 -3.20922732e-01 -4.01739962e-02
1.23634739e-02 -7.93581128e-01 4.95203823e-01 5.02563655e-01
1.62736341e-01 -3.10092280e-03 3.81910980e-01 -4.93837535e-01
-5.41604996e-01 -1.25022662e+00 -3.59796941e-01 -6.39287770e-01
-3.65783602e-01 3.85187715e-01 1.46096453e-01 -5.17498970e-01
2.21426308e-01 6.63031697e-01 -1.03626275e+00 -1.55738339e-01
6.57885611e-01 6.35429859e-01 -4.26538251e-02 5.70339918e-01
-2.46676415e-01 5.21602869e-01 -3.15046489e-01 -7.12686181e-01
3.05542409e-01 -3.40880156e-01 -9.39439237e-01 1.77393004e-01
4.05115187e-01 2.10010871e-01 -9.65645075e-01 1.21890545e+00
1.85425162e-01 1.89210683e-01 1.01754689e+00 -1.38475001e+00
-1.19898129e+00 6.84425712e-01 -4.12825704e-01 2.09953025e-01
2.09506378e-01 -3.53174061e-01 1.72835672e+00 -8.82631958e-01
6.10376775e-01 1.38304484e+00 4.21640009e-01 1.37047306e-01
-1.51190364e+00 -1.73370987e-01 3.07191700e-01 1.78942189e-01
-1.42109227e+00 5.17436564e-02 1.22977507e+00 -6.18747652e-01
9.10232306e-01 3.79554540e-01 7.39193738e-01 5.86555541e-01
2.30931342e-01 3.24988127e-01 4.51980978e-01 -4.07932132e-01
8.70144442e-02 1.28869981e-01 3.57386798e-01 1.21036255e+00
6.10408306e-01 -2.08919555e-01 -2.78285205e-01 -2.60364503e-01
7.17044592e-01 2.51984835e-01 -4.10981745e-01 -1.19359112e+00
-1.34766364e+00 1.17143214e+00 9.49976206e-01 5.98288476e-01
5.83771020e-02 2.66886771e-01 3.05959076e-01 5.39990485e-01
5.40534258e-01 7.31194615e-01 -1.52617171e-01 4.36237186e-01
-1.41063243e-01 -5.51627465e-02 8.03222418e-01 1.15709460e+00
7.91900873e-01 -1.03004850e-01 1.59293890e-01 6.66549623e-01
5.45650780e-01 2.72797257e-01 4.33954358e-01 -5.41169107e-01
6.22066259e-01 1.04644787e+00 -2.39724576e-01 -1.79444063e+00
-3.84485424e-01 -2.05881670e-01 -8.54951143e-01 -5.15349284e-02
2.70388901e-01 4.31895286e-01 -6.93360150e-01 1.70897436e+00
2.06482783e-01 3.45993624e-03 -4.43074331e-02 6.34135783e-01
7.76597440e-01 6.36480212e-01 -1.94172502e-01 3.01864684e-01
1.38952255e+00 -8.56817782e-01 -4.53995347e-01 -4.77325581e-02
1.11430347e+00 -4.29908067e-01 9.34093356e-01 -1.97915524e-01
-8.58709395e-01 -3.71570557e-01 -1.37654424e+00 -1.04955815e-01
-9.39818263e-01 -2.36089528e-01 7.24398673e-01 7.00274229e-01
-1.16306329e+00 7.08837748e-01 -4.65532511e-01 -5.16987741e-01
2.33328044e-01 2.22261190e-01 -6.07836246e-01 -3.79366070e-01
-1.16312885e+00 7.98616171e-01 3.98487210e-01 -1.95929930e-01
-2.30334446e-01 -7.42492259e-01 -1.42953789e+00 1.47618502e-01
1.39402643e-01 -6.56280160e-01 3.87243301e-01 -4.67526197e-01
-9.44266915e-01 7.45347559e-01 2.61805058e-01 -9.80046019e-02
-9.98915806e-02 4.99606401e-01 -6.18593991e-01 2.23372906e-01
-2.26050243e-01 2.90426075e-01 9.64784741e-01 -9.72232163e-01
-1.52522072e-01 -5.43696463e-01 5.79723001e-01 -3.85540947e-02
-5.19438088e-01 -3.44421208e-01 -1.30462661e-01 -7.26955771e-01
4.48042303e-01 -1.07173634e+00 -1.35232463e-01 1.02776721e-01
-3.23119938e-01 -4.68366116e-01 6.89659536e-01 -3.54227096e-01
1.43802106e+00 -2.51316381e+00 5.44592321e-01 6.20857477e-01
7.93861985e-01 1.58784632e-02 -4.25916046e-01 8.29814553e-01
-4.43775088e-01 3.68933022e-01 -5.67535125e-02 6.98098913e-02
2.99573034e-01 2.49396503e-01 -2.11988807e-01 6.22322261e-01
3.32976162e-01 1.16143394e+00 -1.26370108e+00 -4.21452850e-01
1.37217760e-01 5.92711747e-01 -5.34446359e-01 5.06181158e-02
1.64709330e-01 -2.67065823e-01 -4.41417098e-01 3.09244484e-01
4.86706793e-01 -5.55504024e-01 5.13321936e-01 -5.69554925e-01
4.47740883e-01 1.94030464e-01 -1.33932793e+00 1.65134346e+00
-3.25660259e-01 6.25011325e-01 -3.26549977e-01 -1.29570246e+00
1.03577268e+00 1.00807108e-01 4.14023310e-01 -4.71065223e-01
-1.79926939e-02 1.57690004e-01 2.48511478e-01 -1.44675538e-01
6.91508651e-01 6.95602968e-02 -1.50672466e-01 5.44817984e-01
2.60481030e-01 -1.53662905e-01 2.18525127e-01 6.42117560e-01
1.20951760e+00 -5.74323356e-01 1.32143691e-01 -5.42327881e-01
4.18584347e-01 -3.79988611e-01 1.78762153e-01 1.28893971e-01
-8.47800076e-02 4.89413857e-01 8.18010449e-01 -4.39304560e-01
-1.07058060e+00 -1.31455994e+00 -9.81101394e-02 1.03560567e+00
3.80148739e-01 -8.76797378e-01 -4.21243787e-01 -7.95290112e-01
2.91920066e-01 3.85698438e-01 -8.02039921e-01 -7.64060259e-01
-4.07422453e-01 -4.58644301e-01 1.02363445e-01 4.79877561e-01
-1.79915309e-01 -5.84467888e-01 1.36738092e-01 4.13863324e-02
1.01643398e-01 -8.51685226e-01 -1.10535944e+00 -9.50251445e-02
-9.47997332e-01 -1.46622133e+00 -3.90631139e-01 -1.20310247e+00
1.25916588e+00 8.35779846e-01 1.18927395e+00 2.84416646e-01
-4.05348063e-01 8.17695022e-01 -5.27068436e-01 4.08583656e-02
-2.04082772e-01 1.08379629e-02 -1.23408064e-02 2.50136733e-01
2.45529801e-01 -5.64954758e-01 -5.88846147e-01 3.41633141e-01
-1.14267874e+00 -2.13498682e-01 3.12724471e-01 8.63604665e-01
3.48756015e-01 3.11524898e-01 2.06171274e-01 -9.31426883e-01
8.79396677e-01 -6.57300949e-01 -1.74595863e-01 3.93895060e-01
-7.41345823e-01 5.78989804e-01 6.11009836e-01 -4.17521924e-01
-1.22320816e-01 -4.88250792e-01 4.89483833e-01 -5.06301284e-01
5.84981382e-01 7.16563702e-01 -3.14146608e-01 -2.27884784e-01
6.48688972e-01 1.65919550e-02 2.57980585e-01 -3.26858968e-01
8.23998570e-01 1.37500703e-01 -9.17586386e-02 -4.04088050e-01
1.00652766e+00 3.68434578e-01 4.08111900e-01 -9.20346737e-01
-3.31394345e-01 -5.07648408e-01 -8.58654916e-01 2.11358160e-01
6.01781666e-01 -4.95264351e-01 -3.29363912e-01 -8.43606889e-02
-8.69652987e-01 1.21961839e-01 -4.53712076e-01 3.59224051e-01
-4.20466304e-01 5.38267255e-01 -5.29439807e-01 -2.01664984e-01
-1.28040403e-01 -8.65584731e-01 7.79223084e-01 -2.99887985e-01
-1.99054867e-01 -1.63054514e+00 3.71803313e-01 -6.15199329e-03
1.99352443e-01 2.58890331e-01 1.67956901e+00 -6.20530665e-01
-4.55122024e-01 -4.34123129e-01 -2.38426283e-01 2.29143217e-01
3.96126479e-01 5.97257130e-02 -3.08490902e-01 -5.81092119e-01
-4.47529405e-01 1.47631943e-01 7.14150548e-01 -6.22918569e-02
8.88087213e-01 -4.54551756e-01 -3.72225821e-01 5.96084237e-01
1.48073637e+00 -2.98227906e-01 5.30394375e-01 -1.52661219e-01
1.09738231e+00 8.86037767e-01 2.65241951e-01 5.77626331e-03
5.53279042e-01 5.50762117e-01 5.70576549e-01 6.35282695e-02
-2.56770314e-03 -5.73284566e-01 3.88934731e-01 1.25901449e+00
2.96723302e-02 -1.42532378e-01 -7.96060085e-01 5.84467828e-01
-1.73091507e+00 -8.28742623e-01 -5.97388782e-02 2.03653073e+00
2.44374737e-01 2.00932864e-02 1.55071035e-01 2.98318774e-01
8.26531768e-01 6.36197627e-01 -2.99693495e-01 -5.85685015e-01
-6.37996420e-02 2.57272750e-01 5.47378421e-01 5.35289884e-01
-7.52478361e-01 6.39144421e-01 5.79377079e+00 5.18647432e-01
-8.06066334e-01 -4.49767187e-02 2.61921376e-01 1.78707108e-01
-8.45752597e-01 -2.69329101e-02 -5.79819009e-02 3.87275755e-01
8.37134242e-01 -5.73592365e-01 6.19412363e-01 6.90759063e-01
-4.69485283e-01 7.12615907e-01 -1.35198736e+00 1.12991083e+00
2.02292234e-01 -1.42438543e+00 4.77075756e-01 2.92549193e-01
6.20522380e-01 -9.80401188e-02 2.27870718e-01 2.63573468e-01
3.35739762e-01 -1.16873729e+00 4.12673384e-01 4.45141762e-01
6.62375331e-01 -8.07959676e-01 4.82456416e-01 -2.90413238e-02
-1.68629849e+00 1.20483071e-01 -7.45501220e-01 1.77490994e-01
-2.38616168e-01 3.42703581e-01 -6.43829226e-01 7.93974161e-01
1.71020851e-01 9.45216537e-01 -7.05640316e-01 5.81390440e-01
-1.12700589e-01 2.10607782e-01 1.10623650e-01 -1.03041135e-01
2.23871753e-01 -7.36993313e-01 4.50631887e-01 9.96438444e-01
3.17395031e-01 1.86772123e-01 1.00513950e-01 8.98362994e-01
-2.06662863e-01 3.07547390e-01 -1.23392451e+00 -6.50861740e-01
4.23367739e-01 1.35906553e+00 -5.92468500e-01 2.42933761e-02
-6.77739501e-01 7.95548439e-01 7.03420341e-01 4.22567397e-01
-5.75177431e-01 -6.82967842e-01 9.94973421e-01 3.28051835e-01
3.93796861e-01 -7.26468742e-01 2.72529840e-01 -1.15037882e+00
6.02728240e-02 -4.38780546e-01 6.30001009e-01 -5.39397359e-01
-1.50085509e+00 5.57663202e-01 -2.01236099e-01 -1.29248297e+00
4.79078926e-02 -1.09418988e+00 -7.10995615e-01 8.99649918e-01
-1.30127037e+00 -1.24174559e+00 -9.84152332e-02 5.96155882e-01
1.07478211e-03 -1.61218897e-01 9.19882655e-01 8.14128593e-02
-3.91031414e-01 5.36168158e-01 2.37121016e-01 4.31661576e-01
2.74460286e-01 -1.57984757e+00 5.99978209e-01 3.98711592e-01
7.91372478e-01 7.32667983e-01 3.39086205e-01 -2.74136126e-01
-2.08770752e+00 -1.25048769e+00 8.61722231e-01 -6.10426903e-01
1.22220349e+00 -5.06698132e-01 -9.31487381e-01 7.33948231e-01
-1.06853386e-02 4.25772369e-01 9.99661386e-01 2.53660858e-01
-9.39196229e-01 -1.00017101e-01 -1.04904532e+00 6.38286233e-01
1.36961210e+00 -9.14120078e-01 -5.71281552e-01 3.77312392e-01
1.06559455e+00 1.13161534e-01 -1.26596093e+00 1.86764017e-01
3.17049235e-01 -5.73700428e-01 1.22830856e+00 -1.25020421e+00
1.82195902e-01 -2.92910010e-01 -3.80964428e-01 -1.79978228e+00
-9.14088011e-01 -4.96112734e-01 -4.36691523e-01 8.33915114e-01
4.67003047e-01 -8.76186848e-01 7.06890941e-01 5.02679110e-01
4.91584651e-02 -8.89867246e-01 -6.51938319e-01 -8.91981661e-01
2.23489612e-01 -7.24005327e-02 8.39016318e-01 1.29444838e+00
4.91755426e-01 5.26542485e-01 1.08283527e-01 2.68439561e-01
5.94329953e-01 2.28361219e-01 5.79098344e-01 -1.41011584e+00
-1.38432141e-02 -6.17046535e-01 -1.26918173e+00 -8.11889708e-01
2.77884543e-01 -1.61631501e+00 -6.56719089e-01 -1.60205233e+00
-7.88655132e-02 -3.72671455e-01 -5.47361612e-01 -1.13801360e-01
-7.65777454e-02 7.27482438e-02 1.13552220e-01 8.96301270e-02
-4.71902639e-01 6.99201763e-01 1.29215539e+00 -5.51804602e-01
1.38967216e-01 -3.81032586e-01 -8.76619399e-01 4.09375280e-01
6.77509665e-01 -1.37349084e-01 -7.44241238e-01 -4.51931775e-01
5.68752587e-01 -3.54278535e-01 1.30104080e-01 -4.82067078e-01
2.78574359e-02 -1.36983633e-01 -1.28734767e-01 -7.03615397e-02
1.82849035e-01 -1.05944192e+00 1.68509290e-01 4.24121916e-01
-2.15319499e-01 3.21101427e-01 -2.57734716e-01 1.00161624e+00
-1.76110283e-01 -2.48753071e-01 5.80397546e-01 2.95468001e-03
-5.75209618e-01 7.63833821e-01 2.80723751e-01 3.08340192e-01
9.86726046e-01 -3.23855370e-01 -5.12906134e-01 -1.50143132e-01
-6.73906326e-01 4.57759500e-02 5.93652666e-01 7.59491861e-01
7.41778314e-01 -1.89827836e+00 -4.45489138e-01 2.28869453e-01
6.83630347e-01 -3.19119662e-01 1.56652573e-02 5.78309357e-01
-5.46852827e-01 4.53683764e-01 -3.07416208e-02 -2.96109825e-01
-1.01591814e+00 1.04762125e+00 3.17304254e-01 -2.75611490e-01
-6.35908902e-01 6.28822029e-01 3.80245894e-01 -5.04000962e-01
-3.41551378e-02 -5.03608227e-01 -2.19834507e-01 2.10882619e-01
3.01450104e-01 4.65875536e-01 5.34269679e-03 -9.28138435e-01
-1.68803439e-01 7.95596898e-01 -1.09533638e-01 2.25649402e-01
1.19412458e+00 5.79201430e-02 -3.67949873e-01 6.19082510e-01
1.71313107e+00 3.89736742e-02 -6.94078267e-01 -5.88056445e-01
5.41958921e-02 -7.91879952e-01 -6.32632151e-02 3.17109860e-02
-1.18482792e+00 7.98877239e-01 1.69686526e-01 8.83489430e-01
5.16280770e-01 4.45766300e-01 5.32120407e-01 4.28227007e-01
5.59179068e-01 -9.26096737e-01 5.38642049e-01 3.54747534e-01
1.01677716e+00 -9.81132686e-01 1.43246185e-02 -6.52408898e-01
-3.89061868e-01 1.11995459e+00 2.91492254e-01 -3.96937042e-01
1.17336988e+00 -6.22681379e-01 -4.26093310e-01 -7.68039525e-01
-5.00412166e-01 -1.44021407e-01 8.67085338e-01 6.81582034e-01
3.57202828e-01 4.48249996e-01 -1.55691519e-01 2.78050810e-01
-3.14345866e-01 -1.10585117e+00 4.10716504e-01 8.60861301e-01
-4.17694151e-01 -1.17638075e+00 2.00481545e-02 6.11826360e-01
7.68269971e-02 -3.82438339e-02 -6.65789783e-01 7.16739237e-01
-2.83418775e-01 8.57851684e-01 1.63147554e-01 -6.12427235e-01
4.34500128e-01 -6.81427494e-02 6.84415936e-01 -8.97042811e-01
-3.81400548e-02 -7.13216722e-01 -1.04874521e-01 -5.25615990e-01
-1.00942530e-01 -3.52576822e-01 -1.11253846e+00 -5.26423872e-01
-2.43171915e-01 4.19180423e-01 4.98475015e-01 4.30371642e-01
5.92968643e-01 3.23952258e-01 9.51269925e-01 -4.76081133e-01
-5.31092048e-01 -7.41726339e-01 -1.23348546e+00 8.85084271e-01
2.62966812e-01 -8.21276605e-01 -7.09823847e-01 -2.61290550e-01] | [7.191052436828613, 6.056935787200928] |
036ae0a1-d4bc-4cbb-80ab-7af7a35b5837 | constructing-the-f-graph-with-a-symmetric | 1912.07871 | null | https://arxiv.org/abs/1912.07871v1 | https://arxiv.org/pdf/1912.07871v1.pdf | Constructing the F-Graph with a Symmetric Constraint for Subspace Clustering | Based on further studying the low-rank subspace clustering (LRSC) and L2-graph subspace clustering algorithms, we propose a F-graph subspace clustering algorithm with a symmetric constraint (FSSC), which constructs a new objective function with a symmetric constraint basing on F-norm, whose the most significant advantage is to obtain a closed-form solution of the coefficient matrix. Then, take the absolute value of each element of the coefficient matrix, and retain the k largest coefficients per column, set the other elements to 0, to get a new coefficient matrix. Finally, FSSC performs spectral clustering over the new coefficient matrix. The experimental results on face clustering and motion segmentation show FSSC algorithm can not only obviously reduce the running time, but also achieve higher accuracy compared with the state-of-the-art representation-based subspace clustering algorithms, which verifies that the FSSC algorithm is efficacious and feasible. | ['Wen-Bo Hu', 'Xiao-Jun Wu', 'Kai Xu'] | 2019-12-17 | null | null | null | null | ['motion-segmentation', 'face-clustering'] | ['computer-vision', 'computer-vision'] | [ 2.85839625e-02 -4.29642856e-01 -5.46694174e-02 5.81084527e-02
-2.42111474e-01 -3.67576927e-01 -1.91393849e-02 -4.94892627e-01
-9.13322717e-02 1.81504130e-01 4.03193921e-01 -1.03101708e-01
-5.06554425e-01 -3.16495478e-01 -1.64946571e-01 -1.30651081e+00
-1.40248630e-02 1.86268330e-01 1.10201836e-01 9.77482498e-02
8.42841938e-02 4.02896106e-01 -1.21758270e+00 -1.76052257e-01
1.22304630e+00 7.91449666e-01 1.27899602e-01 -4.48986776e-02
-9.40743536e-02 3.61813545e-01 -2.06985801e-01 6.55646920e-02
2.46104121e-01 -5.86422324e-01 -5.19480050e-01 4.00044024e-01
4.34630178e-02 4.06350680e-02 -6.30210876e-01 1.38929296e+00
4.65101749e-01 2.82081187e-01 6.40704870e-01 -1.23614848e+00
-4.47244823e-01 3.16691786e-01 -9.82502997e-01 -1.99201390e-01
3.94623637e-01 -2.37672955e-01 4.75973636e-01 -8.97015452e-01
6.91032588e-01 1.36137378e+00 3.38466465e-01 3.19477260e-01
-1.42667782e+00 -9.32674229e-01 3.40726882e-01 3.54139894e-01
-1.92555952e+00 -4.15185601e-01 1.02547443e+00 -3.56385529e-01
3.06380749e-01 6.16876066e-01 7.61929214e-01 4.50504541e-01
-3.95833462e-01 7.43798792e-01 9.17828619e-01 -2.18451113e-01
2.51023918e-01 -3.00047159e-01 1.72181889e-01 7.95856714e-01
3.83775622e-01 -3.64408731e-01 -2.25005776e-01 -3.55258077e-01
7.07817495e-01 2.07084373e-01 -7.46289611e-01 -7.49541402e-01
-1.50223672e+00 6.45463049e-01 2.67941773e-01 4.73850489e-01
-2.34640539e-01 -1.84188798e-01 1.48224443e-01 -3.65052521e-02
1.26415327e-01 -2.62928933e-01 3.46477404e-02 2.07509667e-01
-1.05936992e+00 -2.61666235e-02 5.27534425e-01 1.01971388e+00
8.81339192e-01 1.62854388e-01 6.58871420e-03 8.58235538e-01
5.49727499e-01 7.80081570e-01 4.78044927e-01 -1.13040364e+00
3.84280413e-01 9.81203854e-01 -2.68022954e-01 -1.52105045e+00
-2.89918244e-01 -3.46480966e-01 -1.26679134e+00 -4.49157208e-01
1.48462549e-01 -7.53493309e-02 -6.34058535e-01 1.69804084e+00
4.61834937e-01 6.43064022e-01 7.30674015e-03 1.25522196e+00
7.84479082e-01 7.53229082e-01 -1.86241940e-01 -1.05001354e+00
7.99732208e-01 -5.38684905e-01 -8.13960850e-01 2.16274723e-01
4.58386719e-01 -6.51221335e-01 7.90471256e-01 3.62263888e-01
-5.66156566e-01 -3.72769386e-01 -8.69800329e-01 3.66156697e-01
1.69205740e-01 4.57578689e-01 6.12389982e-01 6.10886395e-01
-9.39673781e-01 2.61620522e-01 -7.74489880e-01 -2.43413225e-01
8.24454948e-02 4.93579149e-01 -5.38956285e-01 -1.99740618e-01
-8.60366404e-01 -1.38427410e-02 4.27968681e-01 2.61916548e-01
-3.79161716e-01 -4.99469101e-01 -7.71872163e-01 -5.33826575e-02
6.01437628e-01 -2.38685653e-01 3.81680042e-01 -7.63729274e-01
-1.19837558e+00 5.40008485e-01 -6.67758107e-01 3.63757551e-01
1.95651263e-01 2.66263187e-01 -7.35675037e-01 2.63260365e-01
2.94766814e-01 3.24775279e-01 8.02553594e-01 -1.48269224e+00
-6.23556860e-02 -5.81110299e-01 -7.04270065e-01 2.72986323e-01
-7.51624882e-01 -8.01783130e-02 -1.12126124e+00 -5.57146013e-01
9.12501156e-01 -1.07006335e+00 -3.48500639e-01 -3.65302116e-01
-4.85578775e-01 -1.86587751e-01 1.27795458e+00 -8.21502864e-01
1.99960709e+00 -2.51479316e+00 8.11131358e-01 8.22810233e-01
1.08487710e-01 1.57200396e-01 -1.16689652e-01 3.40568095e-01
-3.71043384e-01 -5.26311137e-02 -5.92102945e-01 2.53866851e-01
-3.57714683e-01 2.58687278e-03 3.01172137e-02 5.65309763e-01
-3.99655432e-01 5.15809476e-01 -8.96025181e-01 -7.94295371e-01
2.25614458e-01 3.39496374e-01 -5.66007018e-01 -8.61326009e-02
2.10420400e-01 4.93351966e-01 -7.44776547e-01 3.66430074e-01
1.04574132e+00 -2.15469748e-01 5.62257588e-01 -5.08026779e-01
-7.35814497e-02 -5.20669580e-01 -2.12291360e+00 1.76666594e+00
2.53338158e-01 8.36591125e-02 5.24570882e-01 -9.42794144e-01
8.04502308e-01 1.49925798e-01 1.03685582e+00 -1.59837499e-01
4.74528968e-02 1.49063125e-01 -7.73705496e-03 -3.50971639e-01
4.86056656e-02 1.24742940e-01 1.79426566e-01 2.25800872e-01
-2.80495226e-01 3.28131914e-01 2.64363378e-01 7.99491704e-01
6.12131715e-01 -6.48586005e-02 -2.02702403e-01 -7.49166250e-01
1.37971151e+00 -2.85553306e-01 9.53742802e-01 -8.22660178e-02
-9.70488638e-02 4.81092662e-01 5.67244470e-01 -1.29271587e-02
-4.79057074e-01 -8.76637101e-01 -5.31446636e-02 5.21570623e-01
5.75203598e-01 -6.84706748e-01 -1.10468757e+00 -5.79560518e-01
4.92139906e-02 2.98011303e-01 -2.65915900e-01 -2.76144147e-01
-5.62060952e-01 -7.93808818e-01 -6.59748241e-02 3.17280114e-01
7.87562430e-01 -5.77248514e-01 1.74688175e-01 2.24878266e-02
-3.95193636e-01 -7.89690077e-01 -9.51055288e-01 -4.78988767e-01
-9.10331786e-01 -1.17134631e+00 -6.38959825e-01 -1.11669743e+00
1.16844261e+00 9.22055602e-01 1.79607674e-01 3.47929239e-01
-1.40707612e-01 3.86765033e-01 -3.01586986e-01 4.53129262e-01
1.78203359e-01 -2.62880474e-01 3.36205542e-01 5.63774049e-01
2.56420106e-01 -5.89972794e-01 -6.17400467e-01 5.18138826e-01
-9.13456142e-01 -1.26871005e-01 3.49252909e-01 6.89068973e-01
8.98693264e-01 5.34556329e-01 4.25903857e-01 -6.82549119e-01
5.46572208e-01 -1.83799461e-01 -6.40688002e-01 2.25486919e-01
-7.40950704e-01 -2.66603470e-01 7.41739690e-01 -3.22071314e-01
-9.33720767e-01 5.12320340e-01 3.49383384e-01 -9.54322815e-01
3.29754710e-01 4.74723727e-01 -8.33194554e-01 -9.65619013e-02
1.96475923e-01 7.17969120e-01 2.33182073e-01 -4.96913075e-01
5.29993296e-01 6.17035389e-01 6.49198294e-01 -5.04226089e-01
1.17804134e+00 5.87787032e-01 3.48092854e-01 -9.64471042e-01
-3.59654874e-01 -9.45519984e-01 -8.80900085e-01 -3.63255292e-01
9.36130345e-01 -1.03067744e+00 -9.68950272e-01 4.70952213e-01
-7.89141119e-01 2.86633402e-01 1.89093053e-01 6.27672136e-01
-2.86494792e-01 1.04813421e+00 -5.14393508e-01 -7.91241050e-01
-1.39481992e-01 -9.24872041e-01 7.29933918e-01 1.89531580e-01
2.35688671e-01 -6.56269431e-01 -1.26944646e-01 4.68040705e-01
-5.74414842e-02 1.97659209e-01 9.16080296e-01 -5.78417890e-02
-5.31120241e-01 -4.15070690e-02 -1.77856967e-01 1.76753730e-01
7.80623108e-02 5.16680479e-02 -4.33975875e-01 -7.45822430e-01
3.72776166e-02 2.64575869e-01 8.34178567e-01 4.25166517e-01
1.30876410e+00 -2.96701431e-01 -6.07670844e-01 8.26147854e-01
1.42356503e+00 3.09333861e-01 4.90750194e-01 8.62289965e-02
1.19713223e+00 5.82572699e-01 6.55364752e-01 2.52483934e-01
7.67018199e-02 5.41999221e-01 8.25795978e-02 -3.60036381e-02
4.55051176e-02 -3.14459532e-01 3.50930244e-01 1.49148440e+00
-2.02272132e-01 3.50452334e-01 -6.81781173e-01 4.41011637e-01
-2.23394537e+00 -1.03510082e+00 -6.65463328e-01 2.43851304e+00
4.52419788e-01 -3.81044209e-01 2.36761287e-01 2.63549030e-01
1.10336816e+00 2.70005941e-01 -6.26810551e-01 3.13833594e-01
-1.96256846e-01 -3.37187350e-01 4.17017758e-01 3.55599701e-01
-1.02157164e+00 1.07473707e+00 5.94851351e+00 1.18358290e+00
-8.98490906e-01 -1.11805938e-01 2.40917265e-01 -2.44698878e-02
-4.93010640e-01 2.06438735e-01 -5.30893922e-01 7.37537742e-01
4.10362482e-01 -2.20626682e-01 1.00037491e+00 8.13579738e-01
3.30384821e-01 2.35906348e-01 -6.11380756e-01 1.45131600e+00
3.23729575e-01 -8.59642863e-01 2.83108413e-01 1.81835234e-01
9.64731276e-01 -4.92536753e-01 2.74724886e-02 -4.48890142e-02
-8.00141990e-02 -8.15979421e-01 3.15421551e-01 1.96500391e-01
8.97868872e-01 -1.08724940e+00 4.04807806e-01 4.02417898e-01
-1.59804845e+00 -4.98832203e-02 -6.72561169e-01 3.61266583e-01
-3.07488497e-02 6.74302399e-01 -7.80910477e-02 8.01784158e-01
6.18227601e-01 1.09379268e+00 -5.12808979e-01 9.45506036e-01
-1.92285314e-01 6.15679383e-01 -2.80956715e-01 1.87505946e-01
1.03997946e-01 -1.22650373e+00 8.33085239e-01 9.80494797e-01
2.87703723e-01 6.15941107e-01 5.63427210e-01 6.29416585e-01
1.10579245e-01 6.56435907e-01 -3.60670388e-01 -7.89078027e-02
7.35767722e-01 1.54197621e+00 -7.86944985e-01 -2.61450261e-01
-4.50828880e-01 1.07404065e+00 2.89140344e-02 6.95684850e-01
-4.91496295e-01 -4.43056136e-01 5.07554114e-01 4.81327362e-02
3.26268852e-01 -4.06829387e-01 -1.97143536e-02 -1.31862020e+00
1.10998921e-01 -1.03124750e+00 4.14036006e-01 -4.83255178e-01
-8.71256351e-01 1.74062192e-01 -1.90764412e-01 -1.46823347e+00
-3.20047140e-02 -3.87387812e-01 -6.36058986e-01 6.19997621e-01
-9.04787660e-01 -1.03032386e+00 -3.76889646e-01 1.20296562e+00
-3.84931490e-02 -3.52473706e-01 4.87709373e-01 4.67769712e-01
-1.10400009e+00 4.57049578e-01 7.05122888e-01 2.28556484e-01
5.59522569e-01 -7.67145336e-01 -4.58443105e-01 1.01697707e+00
-1.80722345e-02 9.53534067e-01 2.98682928e-01 -8.13136041e-01
-1.88351750e+00 -1.05920541e+00 3.08185279e-01 -4.48554792e-02
5.59227765e-01 -1.80002168e-01 -9.15144801e-01 4.47746992e-01
-2.29835391e-01 -2.35068634e-01 7.09519207e-01 1.89603548e-02
-4.14355397e-01 -3.13049614e-01 -9.51215982e-01 6.69455051e-01
1.31357133e+00 -4.56750542e-01 -2.68501103e-01 4.44556683e-01
6.85312808e-01 -7.44648725e-02 -9.05089676e-01 5.96342325e-01
3.98061544e-01 -7.78372943e-01 1.05286145e+00 -3.82074803e-01
-8.59618187e-02 -1.09101880e+00 -1.11301713e-01 -1.10175443e+00
-9.40514207e-01 -6.40438378e-01 3.39441337e-02 1.41991591e+00
1.25663564e-01 -6.36957407e-01 9.73596334e-01 3.61602217e-01
3.41286026e-02 -5.41960776e-01 -9.33263779e-01 -1.00914180e+00
-2.86178023e-01 2.00057663e-02 6.13548160e-01 1.35876894e+00
1.12657450e-01 3.55231196e-01 -4.12992895e-01 2.01392263e-01
8.60084176e-01 5.30736506e-01 6.25117004e-01 -1.36766934e+00
-3.28688323e-02 -2.56240368e-01 -3.88888687e-01 -8.27212870e-01
4.84186828e-01 -1.06647527e+00 -3.35237086e-01 -1.50755143e+00
5.78630090e-01 -2.45296970e-01 -3.43871325e-01 1.68491721e-01
-3.76127064e-01 7.32346922e-02 2.35191762e-01 6.70038044e-01
-8.10809493e-01 6.51630402e-01 1.33858252e+00 -5.96899465e-02
-4.87020731e-01 -2.91947067e-01 -7.46786714e-01 6.85251236e-01
2.68104792e-01 -7.28946254e-02 -6.00715816e-01 -3.61066125e-02
-3.18491071e-01 5.68628460e-02 -1.36422962e-01 -8.84326279e-01
3.15480620e-01 -2.64453799e-01 6.46605313e-01 -6.10561430e-01
1.56146795e-01 -1.04064929e+00 4.93407816e-01 6.91629946e-01
1.14788078e-01 -3.35349292e-01 -1.18083246e-01 8.82235825e-01
-2.36562237e-01 7.15036765e-02 7.19880819e-01 2.46620268e-01
-6.77910089e-01 5.46009898e-01 -1.40083432e-01 -9.41990837e-02
1.15719986e+00 -4.06995654e-01 -1.14298560e-01 -3.54127228e-01
-7.70304322e-01 5.17050862e-01 7.22744763e-01 2.38697365e-01
6.24757230e-01 -1.80798328e+00 -6.08534336e-01 4.65005338e-01
-5.55757657e-02 -6.92425221e-02 4.31211054e-01 9.05735254e-01
-3.54594022e-01 4.19528127e-01 1.13534689e-01 -7.64422596e-01
-1.45633578e+00 8.00914824e-01 -1.00483559e-01 1.72520131e-01
-4.80646551e-01 5.90354860e-01 3.35523754e-01 -4.12795186e-01
-1.43617392e-03 4.22115505e-01 -4.50936198e-01 1.08793594e-01
2.60751158e-01 7.44342804e-01 -3.01027924e-01 -1.22855294e+00
-6.24331415e-01 1.25505316e+00 2.12349266e-01 -7.64257535e-02
1.12105334e+00 -1.20865501e-01 -6.71986520e-01 -6.65894076e-02
1.37788272e+00 3.05815876e-01 -8.54061782e-01 -1.18179396e-01
-1.94381624e-01 -7.14446843e-01 1.87307075e-01 -2.97582060e-01
-1.47666764e+00 5.16221106e-01 5.94809234e-01 -1.04883626e-01
1.42869282e+00 -1.73638731e-01 6.98956132e-01 3.51168752e-01
4.84991401e-01 -1.08521116e+00 1.35883586e-02 2.24379539e-01
6.72379434e-01 -7.00743616e-01 2.63954699e-01 -1.04863560e+00
-7.26713240e-01 8.21881473e-01 6.44157052e-01 -1.65759027e-01
7.25358069e-01 -2.70095557e-01 -1.35232449e-01 -7.65947998e-02
-1.20859571e-01 -1.42776206e-01 5.29375494e-01 4.43252474e-01
3.15137327e-01 6.87588230e-02 -7.52649069e-01 4.42898750e-01
-3.99068370e-02 -3.60929817e-01 1.65913209e-01 3.24464202e-01
-4.64486420e-01 -1.14117014e+00 -6.28370285e-01 4.73201543e-01
9.56726521e-02 1.25747263e-01 -5.25199354e-01 3.08343947e-01
6.03674212e-03 1.14719975e+00 -2.30843261e-01 -6.85799778e-01
2.74164051e-01 1.11774333e-01 1.47959858e-01 -4.24528867e-01
-1.82185154e-02 7.47859478e-01 -3.64297539e-01 -7.74955750e-01
-3.61347705e-01 -9.49745119e-01 -1.64849079e+00 -2.46809125e-01
-4.24379259e-01 5.22122800e-01 2.94463396e-01 5.40260315e-01
4.87409711e-01 1.19598098e-01 9.78894770e-01 -4.51639473e-01
-1.99866012e-01 -6.24035656e-01 -1.02202177e+00 8.02720606e-01
-1.90336123e-01 -5.21801949e-01 -6.06038332e-01 2.86174826e-02] | [7.9735260009765625, 4.488919734954834] |
62b18ff7-80a9-44ac-98cf-9d51e3677fd5 | subsurface-structure-analysis-using | 1812.08756 | null | http://arxiv.org/abs/1812.08756v1 | http://arxiv.org/pdf/1812.08756v1.pdf | Subsurface structure analysis using computational interpretation and learning: A visual signal processing perspective | Understanding Earth's subsurface structures has been and continues to be an
essential component of various applications such as environmental monitoring,
carbon sequestration, and oil and gas exploration. By viewing the seismic
volumes that are generated through the processing of recorded seismic traces,
researchers were able to learn from applying advanced image processing and
computer vision algorithms to effectively analyze and understand Earth's
subsurface structures. In this paper, first, we summarize the recent advances
in this direction that relied heavily on the fields of image processing and
computer vision. Second, we discuss the challenges in seismic interpretation
and provide insights and some directions to address such challenges using
emerging machine learning algorithms. | ['M. Alfarraj', 'Z. Wang', 'M. Deriche', 'M. Shafiq', 'Z. Long', 'Y. Alaudah', 'H. Di', 'G. AlRegib'] | 2018-12-20 | null | null | null | null | ['seismic-interpretation'] | ['miscellaneous'] | [ 6.57181561e-01 1.20065406e-01 2.68133640e-01 -2.65200496e-01
-5.22196949e-01 -3.53726000e-01 6.93965077e-01 2.58411795e-01
-3.26571167e-01 2.29149014e-01 3.35706919e-01 -5.37060738e-01
-1.25578851e-01 -9.85003591e-01 -3.13446224e-01 -8.13409567e-01
-6.83783114e-01 2.63336599e-01 4.08224881e-01 -3.40101719e-01
9.32446003e-01 4.98208791e-01 -1.19295156e+00 3.58273059e-01
5.32657981e-01 1.11909950e+00 3.69701356e-01 7.75866389e-01
-3.26666497e-02 4.64046091e-01 1.19979738e-03 1.70221671e-01
2.24194184e-04 1.51304081e-01 -7.40043938e-01 1.18063595e-02
-4.14983220e-02 -3.64979476e-01 -2.79923618e-01 9.36517596e-01
-3.55441943e-02 3.38655189e-02 6.83500528e-01 -5.47630489e-01
-4.31014419e-01 2.43947372e-01 -1.05711377e+00 4.66022432e-01
3.38312685e-01 -1.35031685e-01 8.82563412e-01 -1.28079271e+00
3.77534896e-01 1.06104410e+00 7.94339061e-01 -5.61185889e-02
-8.16990435e-01 -7.25165009e-02 -6.99124858e-02 4.23027664e-01
-8.63048911e-01 -5.43170691e-01 1.06527913e+00 -7.41510570e-01
9.57791448e-01 2.59209037e-01 6.35233164e-01 3.64639729e-01
1.66559204e-01 7.68073976e-01 8.33280504e-01 -8.05014908e-01
2.74867594e-01 -4.61114824e-01 1.90028027e-01 5.62970638e-01
3.52573276e-01 -6.68033585e-02 -6.58682346e-01 -8.37528706e-02
8.10281813e-01 -1.22771814e-01 -2.53086418e-01 1.38457358e-01
-1.39270866e+00 8.70985448e-01 2.01224506e-01 4.51170594e-01
-5.30817926e-01 3.04928988e-01 3.09072793e-01 -8.69583338e-02
3.92770499e-01 3.86323631e-01 -5.00245802e-02 -1.90384630e-02
-1.04937553e+00 7.61168897e-02 6.17741346e-01 2.79563785e-01
8.25840890e-01 5.44505537e-01 8.70028496e-01 5.16275942e-01
8.12141359e-01 8.47131312e-01 2.07329929e-01 -1.07066333e+00
5.67816377e-01 3.14773381e-01 -6.54168651e-02 -1.36862564e+00
-5.63617229e-01 6.99167624e-02 -6.70237839e-01 3.50135386e-01
2.28859186e-01 7.02710748e-02 -9.60932136e-01 7.93618262e-01
1.23546019e-01 1.09474555e-01 1.50296047e-01 4.00686860e-01
6.41435742e-01 8.35631371e-01 -3.32123995e-01 1.46330753e-02
1.35068905e+00 -5.18295169e-01 -5.79150677e-01 -6.83258057e-01
2.86999196e-01 -4.56902266e-01 7.24325716e-01 3.90138149e-01
-8.20834100e-01 -1.44751325e-01 -1.22850120e+00 1.23243891e-01
-1.26157597e-01 -3.50519866e-01 7.80756056e-01 4.54323083e-01
-8.06940556e-01 5.12831926e-01 -1.57302105e+00 -2.37837598e-01
4.16516930e-01 -4.95646521e-03 -2.80990601e-01 1.37165114e-01
-9.39436436e-01 7.31636226e-01 2.69917637e-01 4.75307435e-01
-6.42464519e-01 -5.25894225e-01 -8.65822732e-01 -2.09376286e-03
2.10101858e-01 -2.02660605e-01 1.13422453e+00 -5.29417694e-02
-1.05329955e+00 7.70991981e-01 -2.83095539e-01 -3.67813498e-01
1.75671428e-01 -3.70279640e-01 -2.37557024e-01 7.78975248e-01
-3.23599577e-02 -5.89691661e-02 4.54455405e-01 -1.29979610e+00
-7.35365927e-01 -8.19254816e-01 -6.26433015e-01 -1.58993691e-01
-1.34242982e-01 -6.60596862e-02 -8.35906193e-02 -4.97763246e-01
1.12285578e+00 -7.54882693e-01 -3.90987068e-01 -7.71832913e-02
-2.67481297e-01 3.08003813e-01 1.16997027e+00 -1.17249680e+00
7.08213389e-01 -1.82395267e+00 -1.13748834e-01 6.01579547e-01
2.78999120e-01 1.09441821e-02 5.52815437e-01 6.80714428e-01
1.39464572e-01 1.04750283e-01 -7.47829556e-01 -9.71245691e-02
-4.05812979e-01 1.89486369e-02 -6.22664571e-01 5.35229445e-01
-5.17665818e-02 6.33111179e-01 -7.85644174e-01 -1.27248779e-01
4.20043677e-01 2.07884088e-01 -2.37875357e-01 -1.57845050e-01
-7.95476511e-02 7.71375477e-01 -8.95675063e-01 1.07031703e+00
5.54156303e-01 -1.50304720e-01 1.91428423e-01 -1.48547977e-01
-4.88985300e-01 -2.70812977e-02 -9.07633066e-01 1.26670992e+00
-5.16281314e-02 1.04207063e+00 1.86618388e-01 -1.58949816e+00
9.38502371e-01 2.84541160e-01 5.05091608e-01 -9.37263191e-01
-2.14786053e-01 5.33225238e-01 -2.31201977e-01 -1.13333809e+00
5.61429977e-01 -4.49647337e-01 1.48180559e-01 5.58550119e-01
-5.14248729e-01 -3.96174520e-01 -6.81777969e-02 1.23211592e-01
9.44110632e-01 -1.20462077e-02 2.90051371e-01 -3.05313170e-01
4.09107506e-01 3.14623088e-01 1.99982688e-01 6.30123019e-01
1.55028760e-01 7.01346338e-01 8.75078887e-02 -7.46238232e-01
-1.26495051e+00 -1.16858387e+00 -2.20694214e-01 6.92640781e-01
-1.06746733e-01 1.99799404e-01 -3.59758228e-01 1.94160163e-01
-1.13934509e-01 3.58914971e-01 -5.49540937e-01 2.18258098e-01
-7.15953529e-01 -1.29165339e+00 5.83238363e-01 8.33739817e-01
5.48912287e-01 -9.89025950e-01 -1.08224010e+00 3.45254183e-01
-4.04973239e-01 -1.07308555e+00 4.16677654e-01 -2.72125881e-02
-1.21370029e+00 -9.70075011e-01 -6.79402530e-01 -6.11174881e-01
5.35749912e-01 5.06012201e-01 9.05791163e-01 2.16898039e-01
-4.82163787e-01 6.87966764e-01 -3.83355498e-01 -6.29498899e-01
-1.81745335e-01 -3.69385540e-01 -4.06197570e-02 -3.77673801e-04
1.66137949e-01 -9.36559796e-01 -6.09639823e-01 5.94764948e-02
-9.12599981e-01 4.02325690e-02 3.97888064e-01 6.42077565e-01
5.13681591e-01 5.56641743e-02 4.91433442e-01 -7.22314179e-01
2.96706140e-01 -7.19615161e-01 -5.46149671e-01 3.22718620e-01
-3.42933655e-01 -2.57820301e-02 1.30513623e-01 2.36615509e-01
-1.34789681e+00 -1.85065359e-01 -3.46498579e-01 3.82062346e-01
-1.52544901e-01 1.27183783e+00 1.32647380e-01 -2.69141197e-01
5.23460329e-01 3.86387885e-01 -1.49397746e-01 -6.18579626e-01
7.14570582e-02 7.69662678e-01 1.01303816e+00 -6.49071217e-01
7.53444314e-01 1.30983198e+00 1.32834762e-01 -1.74298036e+00
-6.46531761e-01 -5.26262343e-01 -8.26655805e-01 -4.86052096e-01
8.79884362e-01 -5.38272262e-01 -4.48651612e-01 7.27573812e-01
-1.02146983e+00 -6.35321587e-02 1.63058817e-01 6.51679099e-01
-4.77105349e-01 7.69400656e-01 -5.11765301e-01 -1.06239998e+00
-2.70480663e-01 -9.28304851e-01 1.04077053e+00 3.13990980e-01
-8.95971209e-02 -1.47443569e+00 4.29434061e-01 4.64021564e-01
4.24276263e-01 6.06249332e-01 8.94206405e-01 -1.91600382e-01
-8.20715368e-01 -1.45185083e-01 -1.21721976e-01 -1.05403669e-01
1.75774425e-01 2.38873750e-01 -1.11302495e+00 -1.31692700e-02
4.23990458e-01 -5.35536818e-02 1.22653270e+00 6.55182600e-01
8.04575920e-01 -3.04633938e-02 -4.51833695e-01 8.66690516e-01
1.64489913e+00 6.18393302e-01 1.00340378e+00 8.58082771e-01
5.42434335e-01 1.07361698e+00 2.24590346e-01 6.28296971e-01
3.40829074e-01 8.08866620e-02 4.18571353e-01 -6.75991774e-02
3.47002596e-01 -2.39432622e-02 -1.54117197e-01 9.57620144e-01
-8.20725441e-01 1.81898043e-01 -1.67464900e+00 7.42141366e-01
-1.71026897e+00 -1.27238500e+00 -4.25570756e-01 1.78210199e+00
1.05519965e-01 1.17555283e-01 -5.48518598e-01 4.75965828e-01
4.97252643e-01 2.81596720e-01 -4.83282000e-01 1.79499630e-02
-3.17553073e-01 1.08678684e-01 4.87336695e-01 4.21145558e-01
-1.00894940e+00 5.52830756e-01 7.41280413e+00 1.32969841e-01
-1.19960260e+00 -3.15643072e-01 5.63815117e-01 7.29556739e-01
-6.73933506e-01 1.51120052e-01 -3.55936676e-01 3.80937383e-02
6.35153115e-01 -3.46436985e-02 1.19058281e-01 5.05581141e-01
5.36428392e-01 -7.81996548e-01 -7.32106626e-01 7.13107288e-01
2.67379964e-03 -1.70650840e+00 -2.10874565e-02 9.23112631e-02
6.02182329e-01 2.01900810e-01 1.61869731e-02 -5.10078371e-01
-3.85861546e-02 -8.91068935e-01 8.53284478e-01 6.65787578e-01
4.00245905e-01 -3.28114122e-01 4.98668879e-01 1.25350267e-01
-1.55877399e+00 -1.66541934e-01 -1.51060089e-01 -4.46851999e-01
6.18301630e-01 8.93936276e-01 -7.28687942e-01 5.17077565e-01
9.91651535e-01 6.86051130e-01 -2.13847399e-01 9.39714789e-01
-1.58822328e-01 5.63656330e-01 -5.62900007e-01 4.61174667e-01
3.06477547e-01 -4.68857914e-01 4.52312857e-01 1.03955531e+00
6.15556657e-01 5.87617874e-01 5.22086993e-02 6.46855354e-01
3.60137373e-01 -3.61586750e-01 -7.59644210e-01 -7.45032281e-02
4.51239884e-01 9.12426412e-01 -1.13946700e+00 -2.45207459e-01
-7.99714088e-01 2.52345502e-01 -2.23586902e-01 4.06542271e-01
-3.15132141e-01 -4.04111028e-01 2.36002266e-01 2.84950823e-01
8.54644775e-02 -1.01674628e+00 -9.43683803e-01 -9.72896457e-01
1.49742290e-01 -2.27392390e-01 2.60774970e-01 -8.03432226e-01
-7.49889970e-01 1.83658361e-01 1.91624805e-01 -9.50748026e-01
-4.74443510e-02 -5.92291415e-01 -1.07078302e+00 4.59344715e-01
-1.67608571e+00 -1.03456819e+00 -4.64456886e-01 1.98149234e-01
4.55523163e-01 -1.37991995e-01 4.75348890e-01 7.79910292e-03
-2.84528702e-01 -2.45146185e-01 6.10670447e-01 3.16195935e-01
7.29571749e-03 -9.44815218e-01 7.87209809e-01 9.68993306e-01
1.17158644e-01 3.28286171e-01 9.63322937e-01 -7.82966793e-01
-1.61190856e+00 -4.68058109e-01 3.86117965e-01 -5.49171679e-02
9.79648888e-01 2.48868451e-01 -1.33643997e+00 9.22057033e-01
3.05681117e-02 -1.74500067e-02 5.66551447e-01 -2.70348430e-01
3.15321796e-02 1.42827541e-01 -9.98730600e-01 4.44634914e-01
5.41321278e-01 -4.91241813e-01 -1.04431760e+00 2.73052845e-02
1.36605725e-01 -1.73140675e-01 -7.30692148e-01 5.87626219e-01
7.48915434e-01 -1.04591048e+00 1.21832025e+00 -1.81211084e-01
5.38044214e-01 -2.01383784e-01 -3.48095119e-01 -8.18542004e-01
-2.67289951e-02 -1.41799539e-01 5.18995263e-02 6.73126698e-01
3.81868899e-01 -6.60114944e-01 1.05300772e+00 7.48580813e-01
-5.33176303e-01 -3.56913060e-01 -9.40924466e-01 -5.13468921e-01
-1.54238893e-03 -6.79013729e-01 1.36228666e-01 7.62931406e-01
5.17121665e-02 -2.52844930e-01 -1.56119242e-01 4.54238892e-01
1.15785778e+00 1.50377080e-01 3.00291032e-01 -1.40043449e+00
6.59964532e-02 -1.46890715e-01 -3.86637717e-01 -9.26438034e-01
-3.85387063e-01 -5.44656157e-01 1.88081995e-01 -1.70720887e+00
1.64776206e-01 -4.56904203e-01 -9.33798123e-03 4.10094202e-01
1.24330953e-01 3.63829494e-01 -5.77921793e-02 5.86129189e-01
1.27717033e-01 4.40756142e-01 8.82911325e-01 -2.91914076e-01
-7.03389645e-02 -1.75835088e-01 -5.39176762e-01 1.26720250e+00
7.87809014e-01 -2.14318037e-01 2.61875857e-02 -9.33580756e-01
6.12127244e-01 3.11269671e-01 2.93176770e-01 -9.25877690e-01
4.43789989e-01 -4.94321972e-01 4.01564270e-01 -9.61213708e-01
2.34785125e-01 -6.08048677e-01 6.40847608e-02 6.85593843e-01
2.23745018e-01 9.72179770e-02 3.86083648e-02 6.15846753e-01
-6.07791722e-01 -3.47290337e-01 8.03592324e-01 -5.92318952e-01
-1.38158262e+00 -1.21506520e-01 -7.16873407e-01 -2.24115938e-01
1.01950133e+00 -5.25159955e-01 -3.13779056e-01 -1.96928784e-01
-7.75138676e-01 2.01088935e-01 2.33324736e-01 -6.44969642e-02
1.25212562e+00 -8.40114355e-01 -8.30657601e-01 2.27893949e-01
-1.04714580e-01 2.82792002e-01 5.52694380e-01 8.64310443e-01
-1.29257286e+00 3.02964389e-01 -2.17116326e-01 -8.91658545e-01
-8.80320966e-01 5.27425334e-02 1.80663913e-01 -5.59715889e-02
-1.06199825e+00 6.96072698e-01 2.87633866e-01 1.01981498e-01
-2.20131859e-01 -2.02951178e-01 -4.34486866e-01 -8.54169130e-02
9.90419924e-01 7.85782993e-01 1.80734843e-01 -4.63470399e-01
-4.64227974e-01 9.79641974e-01 2.25684822e-01 -4.91803050e-01
1.88374543e+00 -3.46419364e-01 -1.22031160e-01 5.36275446e-01
7.65084505e-01 -1.86575726e-01 -1.11983216e+00 -1.78539827e-01
7.48488903e-01 -4.25013453e-01 1.69565693e-01 -9.89062041e-02
-1.23685920e+00 1.28174818e+00 1.68295428e-01 1.66230753e-01
1.01523066e+00 2.16223463e-01 8.10789108e-01 7.23137438e-01
5.07935405e-01 -1.25931263e+00 3.44939500e-01 5.71206033e-01
9.68098342e-01 -1.31529033e+00 2.84746259e-01 -3.52953523e-01
-2.73684591e-01 1.66401780e+00 -1.17123369e-02 -8.15217849e-03
9.53859448e-01 5.53685009e-01 1.04151316e-01 -6.00951076e-01
-2.53123611e-01 2.70548999e-01 -1.10015541e-01 6.79703116e-01
2.67854452e-01 -1.16422102e-01 1.41523674e-01 1.43408865e-01
1.98367480e-02 -1.75459653e-01 7.87728608e-01 1.62139904e+00
-1.32964766e+00 -7.04899728e-01 -1.11661780e+00 4.43367153e-01
-4.21135575e-01 -1.89401582e-01 1.09761663e-01 4.06562954e-01
-5.50844550e-01 7.30875552e-01 1.50226623e-01 -1.80085272e-01
7.38436654e-02 -2.88553745e-01 3.52552593e-01 -4.78082150e-01
3.45842123e-01 8.79638419e-02 4.59355265e-02 -1.60687268e-01
-7.41134942e-01 -8.39613378e-01 -1.70675278e+00 -7.00806454e-03
-1.09654024e-01 -1.91864260e-02 1.00594854e+00 1.34668958e+00
-2.70687312e-01 3.91479254e-01 4.78717566e-01 -1.49972892e+00
-1.36473179e-01 -7.73613274e-01 -9.02289569e-01 3.18107307e-02
4.87946570e-01 -6.52861893e-01 -4.00662392e-01 6.03861451e-01] | [6.939342975616455, 2.4223568439483643] |
866751f8-ce23-483b-aa12-fe313ea784f9 | boosting-factor-specific-functional | 1609.06070 | null | http://arxiv.org/abs/1609.06070v2 | http://arxiv.org/pdf/1609.06070v2.pdf | Boosting Factor-Specific Functional Historical Models for the Detection of Synchronisation in Bioelectrical Signals | The link between different psychophysiological measures during emotion
episodes is not well understood. To analyse the functional relationship between
electroencephalography (EEG) and facial electromyography (EMG), we apply
historical function-on-function regression models to EEG and EMG data that were
simultaneously recorded from 24 participants while they were playing a
computerised gambling task. Given the complexity of the data structure for this
application, we extend simple functional historical models to models including
random historical effects, factor-specific historical effects, and
factor-specific random historical effects. Estimation is conducted by a
component-wise gradient boosting algorithm, which scales well to large data
sets and complex models. | ['Kornelia Gentsch', 'David Rügamer', 'Sarah Brockhaus', 'Klaus Scherer', 'Sonja Greven'] | 2016-09-20 | null | null | null | null | ['electromyography-emg'] | ['medical'] | [ 2.76694566e-01 -5.07391870e-01 -1.16262451e-01 -5.19698262e-01
-5.42863309e-01 -1.70036286e-01 4.71500695e-01 -1.97788298e-01
-9.03129339e-01 1.12185776e+00 1.30854055e-01 -1.36493742e-01
-6.08004510e-01 -6.90852180e-02 -7.99774110e-01 -5.20893335e-01
-9.24605131e-01 -4.05510282e-03 -3.89132023e-01 -1.34765685e-01
3.90962690e-01 3.14277142e-01 -1.15124476e+00 9.18342471e-02
6.39473081e-01 1.03860724e+00 2.61559248e-01 6.65912747e-01
7.07386374e-01 4.35366839e-01 -6.43878281e-01 -2.82549500e-01
5.51250204e-02 -7.11093307e-01 -4.01367486e-01 -7.20431209e-02
-1.06059052e-01 -7.71932676e-02 -9.50723588e-02 5.63538611e-01
5.70582628e-01 4.15106684e-01 6.71005547e-01 -1.30963397e+00
-2.38447353e-01 1.91811219e-01 -6.32419467e-01 6.27075315e-01
6.29860759e-01 2.11153805e-01 7.05170512e-01 -5.23032546e-01
4.83890742e-01 9.66713309e-01 7.68905818e-01 6.06807232e-01
-1.76628315e+00 -9.83930886e-01 3.28686118e-01 7.40191698e-01
-1.11192071e+00 -4.01190847e-01 7.35007823e-01 -4.16391075e-01
1.19640946e+00 1.63127229e-01 1.29940760e+00 1.64343083e+00
9.98312533e-01 3.19602877e-01 1.67449784e+00 -3.36827822e-02
3.43943506e-01 -3.40645343e-01 2.27330089e-01 -9.31105465e-02
-2.59770840e-01 5.60436547e-01 -1.00523746e+00 -4.38345313e-01
7.34030902e-01 -3.40092123e-01 -3.23090643e-01 1.59504980e-01
-6.96462333e-01 6.64859354e-01 8.32339376e-02 3.74662131e-01
-1.01457334e+00 4.49093074e-01 4.56927627e-01 6.56390786e-01
4.60892916e-01 4.00370449e-01 -7.33244479e-01 -6.59985363e-01
-1.02773702e+00 4.54392701e-01 4.77225780e-01 7.80711621e-02
2.03776389e-01 2.01504782e-01 2.14497164e-01 9.54467297e-01
-4.22719121e-02 1.41358107e-01 6.17556453e-01 -9.11387384e-01
3.19746286e-02 2.62484625e-02 -1.96260749e-03 -8.62918377e-01
-8.90671432e-01 -3.46503407e-01 -3.43268305e-01 3.89143050e-01
2.84079313e-01 -3.99992913e-01 -5.06321907e-01 2.00824571e+00
-4.84234691e-02 3.15518737e-01 -5.65677345e-01 9.92025435e-01
3.16866003e-02 3.36620510e-01 5.65909386e-01 -7.76666939e-01
1.31454813e+00 -1.48600146e-01 -7.08523810e-01 -4.22823191e-01
1.74277887e-01 2.66036000e-02 7.24842012e-01 9.61349785e-01
-1.07679176e+00 -2.78146148e-01 -8.75925601e-01 3.63767594e-01
1.75737560e-01 -4.57159728e-01 8.56729627e-01 6.72397077e-01
-8.12469244e-01 9.61180210e-01 -1.16324389e+00 3.87361757e-02
2.14079529e-01 6.56170964e-01 -5.77062011e-01 2.63056427e-01
-1.34777439e+00 1.40996075e+00 1.01111224e-02 2.83590317e-01
-6.95450366e-01 -6.47891879e-01 -4.36065257e-01 -2.09756553e-01
-1.24083906e-02 -4.50623512e-01 9.03214872e-01 -1.02601075e+00
-1.50022662e+00 5.26198983e-01 -6.78116828e-02 -2.38229007e-01
3.44141833e-02 1.94033250e-01 -5.56673467e-01 -8.23439062e-02
-3.09785038e-01 3.15532625e-01 9.93762612e-01 -8.01685929e-01
-5.71027212e-02 -9.82576072e-01 -3.44155163e-01 3.87796044e-01
1.33303791e-01 3.79469931e-01 4.05027479e-01 -5.91390729e-01
-7.12576434e-02 -8.30760539e-01 -1.79018259e-01 -3.34304780e-01
4.87813830e-01 -1.03848286e-01 -2.08921134e-01 -1.17029774e+00
1.21379578e+00 -2.14526463e+00 5.51390290e-01 3.25449765e-01
-3.23495492e-02 -5.15440941e-01 -1.89724833e-01 3.15335184e-01
-5.68861663e-01 -3.76346529e-01 -1.37379646e-01 1.47210121e-01
8.28438327e-02 -9.46928561e-03 2.93644935e-01 7.96457052e-01
1.22089183e-03 8.92516315e-01 -4.19804424e-01 4.19845954e-02
-1.97056532e-02 3.04963708e-01 -4.81811494e-01 -1.33361980e-01
2.88532197e-01 3.39013487e-01 4.59400527e-02 4.31195736e-01
3.34423304e-01 3.36238861e-01 3.11744034e-01 6.38598874e-02
-4.58964072e-02 3.62626702e-01 -8.27630758e-01 1.65523720e+00
-5.26725292e-01 7.37771809e-01 5.23476720e-01 -1.26001513e+00
7.60341167e-01 6.01791680e-01 7.19516039e-01 -1.01777089e+00
3.96930277e-01 1.84805781e-01 6.83427274e-01 -3.40644866e-01
1.97921321e-02 -9.71187949e-01 -1.32301301e-01 4.46332663e-01
3.67800325e-01 -6.42694756e-02 1.02706455e-01 -3.00015211e-01
1.33442533e+00 4.59732890e-01 2.30338991e-01 -3.84208441e-01
-5.67193739e-02 -3.07500362e-01 5.33985853e-01 3.80576611e-01
-2.01733410e-01 5.85985146e-02 7.81510770e-01 1.61489211e-02
-4.15216982e-01 -8.36694062e-01 -7.00738132e-01 1.10879195e+00
-2.80810386e-01 -2.32075945e-01 -6.76047087e-01 1.00251369e-01
1.94813669e-01 9.49295580e-01 -1.02693796e+00 -6.08283043e-01
-1.71358347e-01 -1.22459078e+00 1.94784775e-01 5.51844716e-01
-2.99666137e-01 -1.27467179e+00 -7.91617572e-01 8.21765780e-01
2.75778230e-02 -6.09018028e-01 -2.98460461e-02 7.87526429e-01
-1.16213894e+00 -9.89104509e-01 -6.03203058e-01 -1.32594824e-01
3.86398509e-02 -5.20478964e-01 8.44061375e-01 -3.25123340e-01
-5.14612556e-01 5.54846346e-01 -4.73245718e-02 -5.65516174e-01
2.04327375e-01 -4.70440149e-01 4.77055281e-01 -1.15031667e-01
3.55777860e-01 -1.05286455e+00 -4.61501837e-01 1.55833483e-01
-5.23287892e-01 -4.27338123e-01 5.09230435e-01 1.13912928e+00
6.65650368e-02 -1.53203651e-01 8.61765385e-01 -2.59091288e-01
1.18632090e+00 -7.16947138e-01 -3.06560636e-01 -2.00892538e-01
-6.55852437e-01 -3.46891224e-01 2.47908220e-01 -8.53391111e-01
-9.29677129e-01 -1.86596125e-01 -1.73463732e-01 -2.22957179e-01
1.66735835e-02 7.31005013e-01 1.95623353e-01 -1.19515114e-01
7.79271185e-01 2.57499274e-02 3.30344811e-02 -2.83250183e-01
-1.73712447e-01 6.80247724e-01 4.81482774e-01 -7.49508023e-01
-1.13464326e-01 -1.20025948e-01 -1.15432106e-01 -7.84755826e-01
-6.47045448e-02 1.29647568e-01 -3.14721823e-01 -6.76980197e-01
6.01318002e-01 -8.51457119e-01 -1.14656818e+00 4.86470312e-01
-8.50893855e-01 -6.21194422e-01 2.01627463e-01 1.17345977e+00
-1.10054111e+00 -1.37485936e-01 -6.67746544e-01 -1.10480130e+00
-3.29873562e-01 -9.51213241e-01 6.56287611e-01 -1.85604677e-01
-8.38403761e-01 -7.83730507e-01 2.48587400e-01 1.84801623e-01
6.43734336e-02 2.79227644e-01 8.25827837e-01 -1.53480932e-01
2.49695808e-01 -3.68996292e-01 5.09485424e-01 3.77883047e-01
-6.49102926e-02 -3.08633566e-01 -6.69738531e-01 -4.33788188e-02
8.19215059e-01 -4.93566275e-01 2.98054785e-01 8.50830972e-01
8.99725139e-01 3.75037268e-02 -1.07028797e-01 3.55873764e-01
1.16055167e+00 6.88499928e-01 8.11701536e-01 3.04892153e-01
9.90398899e-02 8.07948470e-01 2.44272575e-01 3.95390689e-01
-2.00469960e-02 7.69503236e-01 1.75852805e-01 3.67953479e-01
6.02572143e-01 9.01538134e-02 4.89858389e-01 5.72557628e-01
-3.86463791e-01 1.12589046e-01 -4.79888678e-01 4.97698635e-01
-1.50128269e+00 -9.67088282e-01 -2.47212753e-01 1.93956292e+00
7.62409806e-01 2.73946196e-01 4.53076333e-01 2.29311824e-01
5.42621970e-01 -1.77083567e-01 -7.85212755e-01 -8.78933728e-01
2.04319760e-01 6.57236278e-01 2.63014048e-01 1.24888249e-01
-1.10638820e-01 2.24898428e-01 7.87808323e+00 6.90786123e-01
-1.11775756e+00 2.10324913e-01 4.40414101e-01 -7.15872705e-01
6.09337986e-02 -5.51123521e-04 -6.74834326e-02 7.10658431e-01
1.66926515e+00 -3.75314623e-01 1.12650752e+00 3.80296946e-01
5.56716740e-01 -7.31855929e-01 -9.97759521e-01 1.09048688e+00
-2.07672283e-01 -5.70474684e-01 -9.67490673e-01 1.55011296e-01
2.15604380e-01 1.53083712e-01 -1.35896772e-01 3.30506116e-01
-1.70726433e-01 -1.03291726e+00 8.69862378e-01 7.93400109e-01
8.20386112e-01 -7.36217141e-01 2.43826285e-01 4.53517616e-01
-4.80057567e-01 -3.94716144e-01 2.67575048e-02 -7.04933882e-01
3.61528158e-01 4.16089714e-01 -9.08839032e-02 1.14942856e-01
6.80749238e-01 5.43670237e-01 1.71172004e-02 6.73109949e-01
-1.18235745e-01 6.47326827e-01 -4.90786701e-01 -1.75767064e-01
-1.72712877e-02 -3.35933030e-01 2.53079325e-01 6.37036085e-01
2.55901366e-01 4.60427165e-01 -7.84083128e-01 9.98553634e-01
5.27701139e-01 -5.12510054e-02 -2.43698299e-01 -1.89278051e-01
3.01153455e-02 1.19450283e+00 -6.41268849e-01 2.28098139e-01
-4.51785475e-01 9.13573623e-01 1.47903219e-01 3.49940509e-01
-7.26981163e-01 -2.89965183e-01 6.96745872e-01 5.94503321e-02
-8.71656761e-02 -2.30016515e-01 -1.95479915e-01 -8.69650543e-01
2.78998464e-01 -9.97993588e-01 1.45986468e-01 -1.09022152e+00
-1.29408634e+00 3.59907568e-01 4.62333769e-01 -5.37138164e-01
-5.95788062e-01 -5.72530508e-01 -5.80658376e-01 1.09758878e+00
-5.74217975e-01 -2.79951036e-01 3.07038546e-01 7.47583449e-01
2.43646219e-01 3.75540942e-01 8.35045159e-01 1.74640015e-01
-4.94548440e-01 1.42936170e-01 -8.04621205e-02 -5.25069356e-01
3.91859978e-01 -7.68728614e-01 2.20988974e-01 3.60196233e-01
-1.61501125e-01 7.24012196e-01 7.30812311e-01 -6.30911946e-01
-1.37598324e+00 -3.24478857e-02 5.08946717e-01 -1.60790190e-01
8.69990468e-01 -6.51986599e-01 -7.79592454e-01 4.93802458e-01
-7.24283829e-02 -2.38361731e-01 7.58027852e-01 3.71248186e-01
2.69227564e-01 -5.87990731e-02 -1.10353470e+00 4.27495033e-01
1.14295936e+00 -5.27505577e-01 -7.35409856e-01 1.19829318e-02
-2.50105590e-01 -1.83805391e-01 -1.28488088e+00 2.80556947e-01
1.22251701e+00 -6.99405074e-01 5.45995891e-01 -6.69843316e-01
7.97839686e-02 4.58781093e-01 1.96131468e-01 -1.79164314e+00
-4.68246669e-01 -6.10088646e-01 1.54005423e-01 7.00095773e-01
1.57782614e-01 -6.18802130e-01 4.05849874e-01 9.83070135e-01
-1.38551719e-03 -8.11677277e-01 -1.34107852e+00 -9.10598516e-01
1.32068425e-01 -9.40752566e-01 2.07479715e-01 6.41338229e-01
8.67695272e-01 3.33759516e-01 -5.56087554e-01 -5.57931006e-01
2.48636246e-01 -3.48934352e-01 5.22683188e-02 -1.11724246e+00
-5.87271154e-01 -4.78459805e-01 -6.92057192e-01 -3.29968601e-01
4.72066611e-01 -7.19321609e-01 1.75089575e-02 -1.11771214e+00
2.32996970e-01 5.28887920e-02 -5.08246899e-01 4.32408988e-01
-3.07197720e-01 -1.28697213e-02 -5.94256446e-02 -2.60999858e-01
1.02933176e-01 6.23452067e-01 6.71690702e-01 3.48190725e-01
-4.31032449e-01 2.77745873e-02 -7.34323263e-01 3.63175213e-01
6.92195058e-01 -8.59716117e-01 -3.83928597e-01 5.67826033e-02
2.41725415e-01 8.38337660e-01 2.53242552e-01 -7.82717168e-01
-2.00765848e-01 -2.34014481e-01 6.97787344e-01 5.57072684e-02
7.40585327e-01 -6.67045414e-01 6.57499135e-01 3.97625118e-01
-2.84493715e-01 2.32190624e-01 5.22012234e-01 5.04182994e-01
8.41218457e-02 -2.36960977e-01 4.51963842e-01 -6.52015284e-02
-2.73589164e-01 8.54241475e-02 -9.28093910e-01 -1.18057631e-01
7.88684428e-01 -3.96172613e-01 2.05720559e-01 -5.21019697e-01
-1.22251916e+00 3.52953747e-03 1.57922864e-01 3.31078619e-01
4.29360449e-01 -1.24585819e+00 -6.45135939e-01 1.61092088e-01
-2.67795146e-01 -1.06646657e+00 6.11945152e-01 1.41036761e+00
6.44092262e-02 1.77362084e-01 -6.94604039e-01 -2.83392519e-01
-1.17886317e+00 5.75895131e-01 4.46657240e-01 1.43965960e-01
-2.96577543e-01 8.60622883e-01 -1.18911974e-01 1.35234684e-01
5.16627589e-03 3.28252316e-02 -9.21066925e-02 3.97857249e-01
4.43333447e-01 6.01873457e-01 2.72342414e-01 -4.60718215e-01
-5.41170657e-01 1.62420049e-01 4.28501427e-01 -7.26836979e-01
1.70243347e+00 5.22762649e-02 -1.32862702e-01 9.00392115e-01
1.08152890e+00 -6.24652565e-01 -1.43582916e+00 5.92682004e-01
-2.10088953e-01 -2.38202617e-01 3.67130458e-01 -9.32530403e-01
-8.12415183e-01 7.51683533e-01 6.71673596e-01 -2.59855062e-01
1.45444942e+00 -3.67245138e-01 1.80542842e-01 -1.45066023e-01
6.01062298e-01 -1.30143690e+00 -4.09834474e-01 -2.35091686e-01
1.19353926e+00 -5.25114119e-01 3.26480865e-01 1.68888390e-01
-6.24887288e-01 7.01144695e-01 2.05127046e-01 -5.95136225e-01
6.22905552e-01 3.99691790e-01 -4.03140038e-01 -1.67118847e-01
-1.28706825e+00 1.28844872e-01 3.01370889e-01 6.08892620e-01
3.08084190e-01 2.10388720e-01 -1.21407557e+00 1.09556091e+00
-3.23976398e-01 4.84949619e-01 3.62688005e-01 9.52642262e-01
1.78283751e-01 -1.05741596e+00 -4.34403718e-01 9.15563524e-01
-7.26571918e-01 -2.37610191e-01 -2.66543776e-01 1.04474640e+00
-8.05162936e-02 9.30809498e-01 2.07187414e-01 -4.62989479e-01
4.90497410e-01 5.34819424e-01 1.04857266e+00 -3.62944692e-01
-9.58397508e-01 3.09792697e-01 2.33656675e-01 -6.34308696e-01
-4.29219842e-01 -1.33867109e+00 -1.03128183e+00 -2.58119069e-02
-5.20341992e-01 -1.12638831e-01 1.32879889e+00 9.04057920e-01
1.11842752e-01 5.00248849e-01 5.35192728e-01 -1.03908181e+00
-3.91106665e-01 -1.49907136e+00 -1.25611281e+00 2.21611515e-01
3.73787694e-02 -9.10962641e-01 -7.31324375e-01 -3.18143010e-01] | [13.0013427734375, 3.4048194885253906] |
a8de1332-336c-42a8-9c99-f072da3c23c2 | congrat-self-supervised-contrastive | 2305.14321 | null | https://arxiv.org/abs/2305.14321v1 | https://arxiv.org/pdf/2305.14321v1.pdf | ConGraT: Self-Supervised Contrastive Pretraining for Joint Graph and Text Embeddings | We propose ConGraT(Contrastive Graph-Text pretraining), a general, self-supervised method for jointly learning separate representations of texts and nodes in a parent (or ``supervening'') graph, where each text is associated with one of the nodes. Datasets fitting this paradigm are common, from social media (users and posts), to citation networks over articles, to link graphs over web pages. We expand on prior work by providing a general, self-supervised, joint pretraining method, one which does not depend on particular dataset structure or a specific task. Our method uses two separate encoders for graph nodes and texts, which are trained to align their representations within a common latent space. Training uses a batch-wise contrastive learning objective inspired by prior work on joint text and image encoding. As graphs are more structured objects than images, we also extend the training objective to incorporate information about node similarity and plausible next guesses in matching nodes and texts. Experiments on various datasets reveal that ConGraT outperforms strong baselines on various downstream tasks, including node and text category classification and link prediction. Code and certain datasets are available at https://github.com/wwbrannon/congrat. | ['Deb Roy', 'Jad Kabbara', 'Brandon Roy', 'Wonjune Kang', 'Hang Jiang', 'Suyash Fulay', 'William Brannon'] | 2023-05-23 | null | null | null | null | ['link-prediction'] | ['graphs'] | [ 3.97244930e-01 4.90943789e-01 -5.47168791e-01 -5.95241189e-01
-4.28990066e-01 -6.54236317e-01 1.08723330e+00 6.80040479e-01
-1.43242538e-01 1.80352792e-01 5.24070323e-01 -3.74713302e-01
-2.70574782e-02 -8.57348800e-01 -1.01320148e+00 -3.28898907e-01
-6.29289076e-02 7.18680680e-01 3.20202410e-02 1.75035298e-01
1.55082151e-01 -4.31301929e-02 -1.09151554e+00 4.40541327e-01
5.48341036e-01 5.17999768e-01 3.61596882e-01 8.50792885e-01
-2.79363483e-01 9.37282443e-01 -1.34266570e-01 -9.78848279e-01
4.67055440e-02 -3.44528407e-01 -1.08815718e+00 3.03679258e-01
9.25718606e-01 -2.42002979e-01 -9.62055683e-01 9.82706547e-01
1.29901439e-01 6.37435466e-02 9.28969383e-01 -1.43410051e+00
-1.11846530e+00 1.26013124e+00 -7.62348413e-01 2.55527496e-01
2.38798887e-01 -4.95757088e-02 1.71462619e+00 -7.96127915e-01
8.73903155e-01 1.29775393e+00 8.31251860e-01 2.43198857e-01
-1.50528526e+00 -5.04556179e-01 4.13959295e-01 4.51519042e-02
-1.30403495e+00 -6.02548122e-01 8.06772649e-01 -7.04374731e-01
1.03026450e+00 -2.16312781e-02 3.35537225e-01 1.27434230e+00
2.46350124e-01 9.32867885e-01 5.40486336e-01 -4.27619010e-01
-2.56627470e-01 7.20381960e-02 1.39199346e-01 1.19887984e+00
3.80607814e-01 -3.28652829e-01 -5.23396671e-01 -1.82150975e-01
6.67592525e-01 1.13345519e-01 -2.29799658e-01 -5.56204200e-01
-1.23591673e+00 9.41694975e-01 6.69932485e-01 3.67872387e-01
-2.28853703e-01 3.85659248e-01 3.42375159e-01 3.35759997e-01
7.18282044e-01 1.92923695e-01 -2.25790441e-01 3.87085438e-01
-1.03479636e+00 -5.24074994e-02 9.52453554e-01 1.09880424e+00
1.01690471e+00 -3.42768133e-01 -1.60469562e-01 6.68586433e-01
6.71379805e-01 1.76978886e-01 3.09838027e-01 -8.35835397e-01
6.45439088e-01 5.70941746e-01 -5.18807590e-01 -1.27172220e+00
-3.00634056e-01 -7.06914186e-01 -7.98548818e-01 -3.09343874e-01
2.44354427e-01 6.37615249e-02 -9.98748958e-01 1.77851760e+00
-6.02126196e-02 2.24436134e-01 -1.49968013e-01 3.91724855e-01
1.27074826e+00 6.64409518e-01 2.14490786e-01 2.73121875e-02
1.17511106e+00 -1.30692053e+00 -5.93869865e-01 -4.34533328e-01
9.14789081e-01 -8.17270875e-01 8.17762792e-01 -1.00178393e-02
-1.12314451e+00 -3.58069599e-01 -7.31102288e-01 -4.62341785e-01
-5.77591121e-01 9.32690725e-02 7.73902059e-01 3.97115340e-03
-1.46536827e+00 7.51961231e-01 -6.88855708e-01 -8.91496480e-01
7.09630072e-01 2.03366756e-01 -4.44484800e-01 -4.43545841e-02
-1.05270040e+00 6.45584583e-01 2.90776730e-01 -2.09135830e-01
-7.64030814e-01 -4.19923961e-01 -1.02903628e+00 2.94747472e-01
3.85459721e-01 -9.27307248e-01 1.03764856e+00 -1.05525911e+00
-9.33744490e-01 1.34281003e+00 -2.61886537e-01 -4.97847468e-01
1.53373599e-01 1.13861218e-01 -2.19606355e-01 2.59791881e-01
3.19885999e-01 9.38799202e-01 9.35189605e-01 -1.35186398e+00
-1.11974791e-01 -1.75768554e-01 4.99975197e-02 4.03942227e-01
-5.89233398e-01 -8.92581046e-02 -8.79724503e-01 -7.78973877e-01
3.55179906e-02 -7.81953275e-01 -1.10441625e-01 3.24672669e-01
-7.93061733e-01 -6.06652319e-01 5.93692124e-01 -6.57732248e-01
1.14607847e+00 -1.84709203e+00 3.72185051e-01 1.89025998e-01
7.38339841e-01 -2.12853506e-01 -4.83250737e-01 9.20017481e-01
-2.99885958e-01 3.44256580e-01 -1.51351631e-01 -7.40109205e-01
8.41954574e-02 2.85044163e-01 -8.24099109e-02 5.49192011e-01
8.18488970e-02 1.21789670e+00 -9.59101081e-01 -8.98011386e-01
-5.57059869e-02 2.74539500e-01 -4.22698289e-01 9.45323408e-02
-4.06303287e-01 -1.88606195e-02 -3.22274417e-01 3.87196630e-01
3.76781970e-01 -1.16633344e+00 3.90821338e-01 -4.03351575e-01
3.31931204e-01 5.17735064e-01 -7.26614833e-01 1.84008276e+00
-1.68599650e-01 9.06207860e-01 2.40445673e-01 -1.35263205e+00
7.39665270e-01 1.32968813e-01 4.86924917e-01 -2.86658823e-01
1.71777427e-01 -5.23383133e-02 -1.21333100e-01 -3.79572272e-01
3.57191026e-01 2.39257336e-01 2.21462622e-01 8.08065176e-01
5.94824553e-01 -6.39597401e-02 4.71630216e-01 1.08722782e+00
1.30784702e+00 1.52469397e-01 1.12996802e-01 -2.07357213e-01
1.63152620e-01 -2.59296864e-01 4.42399867e-02 7.80680299e-01
7.67673254e-02 4.45550829e-01 6.69953227e-01 -6.06449954e-02
-1.00458360e+00 -1.00193131e+00 -1.02098636e-01 1.33676922e+00
-6.57303631e-02 -8.30869496e-01 -4.14040208e-01 -9.79851663e-01
2.65451789e-01 5.45661151e-01 -7.42864609e-01 -1.95941012e-02
-4.09520358e-01 -3.14251214e-01 2.85575837e-01 5.36555588e-01
9.03215781e-02 -7.89936125e-01 3.57653528e-01 5.29586431e-03
-1.39642358e-01 -1.25300884e+00 -7.16751397e-01 3.16366971e-01
-9.32370007e-01 -1.04292285e+00 -4.92918879e-01 -1.20938313e+00
9.93604481e-01 4.11745012e-01 1.57739401e+00 6.34304464e-01
-1.34872377e-01 8.93023670e-01 -2.81080425e-01 1.13529801e-01
-3.72717738e-01 3.30251276e-01 -3.71424139e-01 -6.02938682e-02
1.10480726e-01 -5.96036375e-01 -3.48278821e-01 -8.93341228e-02
-8.15229654e-01 4.20097142e-01 6.13227606e-01 7.86359847e-01
3.52022767e-01 4.25907373e-02 1.90675750e-01 -1.44770753e+00
5.40337265e-01 -1.01268923e+00 -2.74871349e-01 3.61035764e-01
-7.89063632e-01 1.69554666e-01 4.44616944e-01 -1.98150262e-01
-6.45681858e-01 3.97710092e-02 1.08065896e-01 -4.20954168e-01
-2.27951910e-03 9.75568295e-01 2.76634451e-02 2.06581622e-01
5.38497448e-01 1.15907967e-01 2.69760024e-02 -3.45405668e-01
6.07513070e-01 4.37232822e-01 5.59824526e-01 -4.63672489e-01
1.15146041e+00 3.17218244e-01 1.04642980e-01 -6.25504434e-01
-9.69665408e-01 -5.46473861e-01 -9.23892379e-01 -1.65775821e-01
6.89360559e-01 -1.08993387e+00 -3.07099879e-01 1.54882416e-01
-1.07411313e+00 -7.83786237e-01 -1.50083527e-01 2.94034898e-01
-3.07329684e-01 6.01683080e-01 -8.21310937e-01 -2.60506868e-01
-2.39369452e-01 -6.84405625e-01 1.09671259e+00 -5.15033193e-02
-1.69098094e-01 -1.77796209e+00 -1.98913440e-02 3.54491115e-01
1.90234840e-01 2.67264321e-02 9.43508685e-01 -1.01164985e+00
-5.46333551e-01 -1.26702532e-01 -4.97020155e-01 -9.07013863e-02
2.29705572e-01 3.49144250e-01 -5.50734162e-01 -5.70587397e-01
-5.47753036e-01 -6.47618175e-01 1.33883667e+00 2.49990642e-01
1.13800478e+00 -6.63411260e-01 -9.17879701e-01 6.10172510e-01
1.35671616e+00 -4.75310922e-01 2.29425624e-01 7.01657403e-03
1.09279180e+00 7.06412792e-01 -1.09612010e-02 2.52225161e-01
8.87752712e-01 3.98217589e-01 6.09299302e-01 -3.03448170e-01
-4.29698855e-01 -6.19141936e-01 3.34659010e-01 1.04600239e+00
4.03274864e-01 -9.30310190e-01 -8.90466332e-01 7.50492454e-01
-1.88809228e+00 -1.02698457e+00 -4.13910210e-01 1.81782639e+00
9.23849344e-01 1.77367032e-01 -7.02502625e-03 -2.08938524e-01
1.05557203e+00 5.49627662e-01 -5.87531090e-01 4.99688499e-02
-1.68770567e-01 1.96396366e-01 4.98552591e-01 7.89673984e-01
-1.20119846e+00 1.03755569e+00 5.81011534e+00 5.21329343e-01
-9.10111904e-01 1.27270147e-01 7.47497082e-01 1.78184718e-01
-8.17089260e-01 3.11624765e-01 -5.84669352e-01 3.06294054e-01
9.60288584e-01 -3.17160964e-01 4.95011479e-01 4.54976320e-01
-2.03634605e-01 2.41769925e-01 -1.41088653e+00 7.15960145e-01
3.70428473e-01 -1.65643477e+00 2.37056211e-01 1.11903064e-01
7.82899678e-01 4.03952330e-01 9.97115523e-02 2.16060936e-01
8.46007705e-01 -1.07436764e+00 4.75522578e-01 5.51512063e-01
7.73699462e-01 -1.47645056e-01 3.66153002e-01 1.48524165e-01
-1.30791998e+00 2.78301299e-01 -2.64617234e-01 1.66527793e-01
-1.73676938e-01 5.49298227e-01 -8.96161258e-01 4.79436100e-01
5.46668053e-01 1.51039970e+00 -1.05241799e+00 6.93278968e-01
-3.87642443e-01 8.38680327e-01 -2.25818023e-01 1.60610396e-02
2.35439688e-01 -1.03225708e-02 4.86227900e-01 1.38178098e+00
2.92308331e-02 -5.17988801e-02 4.13028985e-01 1.00977969e+00
-7.63994575e-01 2.87038177e-01 -7.32242942e-01 -4.78464842e-01
6.21057689e-01 1.59777200e+00 -8.77186060e-01 -4.99914527e-01
-7.49188125e-01 1.10415292e+00 6.76040411e-01 5.57342291e-01
-5.36921084e-01 -3.11379641e-01 -3.68092358e-02 2.75253892e-01
4.27825630e-01 -2.62283146e-01 -1.55371398e-01 -1.28967381e+00
-2.76486158e-01 -4.20689702e-01 7.31033802e-01 -8.93386602e-01
-1.68777299e+00 4.99579698e-01 -1.84716042e-02 -7.91751981e-01
-1.85386628e-01 -6.71992958e-01 -8.46895158e-01 7.44965315e-01
-1.52354610e+00 -1.57505727e+00 -3.11780542e-01 7.29604661e-01
5.76654792e-01 -1.82992846e-01 5.65471232e-01 2.32575573e-02
-6.10320985e-01 6.44963622e-01 1.75844729e-01 4.84501690e-01
8.70207012e-01 -1.47852564e+00 6.74826026e-01 7.85145462e-01
6.51449084e-01 6.27443969e-01 5.60110033e-01 -9.12160814e-01
-1.43806446e+00 -1.21800196e+00 1.16385674e+00 -5.24350703e-01
1.02765179e+00 -5.41745901e-01 -1.09030855e+00 1.24676609e+00
7.74150968e-01 2.48532712e-01 6.73212171e-01 3.87931705e-01
-4.39237267e-01 1.62755921e-01 -7.40269542e-01 5.60324728e-01
1.21051514e+00 -8.17525864e-01 -4.15724963e-01 8.90445828e-01
6.80271864e-01 -1.45535886e-01 -9.60732520e-01 1.08994491e-01
1.94440514e-01 -6.20845675e-01 6.89489961e-01 -5.98307192e-01
7.78402269e-01 1.32225826e-01 2.89185327e-02 -1.27785468e+00
-7.19819903e-01 -6.74993277e-01 -4.56444055e-01 1.50010765e+00
4.90451604e-01 -4.98767644e-01 1.01976800e+00 1.48129210e-01
-2.56590635e-01 -7.48151004e-01 -4.04093087e-01 -4.80998099e-01
1.64027005e-01 3.40038389e-02 1.87771484e-01 1.42392468e+00
7.66664296e-02 9.76780415e-01 7.23583205e-03 2.05667511e-01
8.19474638e-01 1.19394518e-01 7.11246848e-01 -1.62340367e+00
-3.98138553e-01 -6.81015790e-01 -1.39627069e-01 -1.38789916e+00
5.99530995e-01 -1.75044167e+00 -1.27137713e-02 -2.09756541e+00
5.16977012e-01 -3.85562330e-01 -1.09783120e-01 7.71676242e-01
-2.90790468e-01 1.43047655e-02 -1.23667747e-01 5.89703858e-01
-8.44213605e-01 4.82548177e-01 1.06588042e+00 -3.70628119e-01
2.75906384e-01 -2.25776613e-01 -7.63389289e-01 5.66140592e-01
5.93244970e-01 -4.69003797e-01 -6.12364411e-01 -7.47782707e-01
4.56942141e-01 -3.64824235e-02 3.46872449e-01 -5.36771357e-01
6.41059160e-01 9.74422023e-02 3.56517524e-01 -5.88716030e-01
9.61773098e-02 -5.88903546e-01 -9.37925875e-02 3.21668446e-01
-8.28222096e-01 3.86581719e-02 -1.23176746e-01 8.39692891e-01
7.14204684e-02 -3.77469063e-01 5.43762445e-01 -1.73561081e-01
-4.97790843e-01 6.46872282e-01 -2.12987050e-01 1.55045986e-01
5.49049437e-01 -3.04727368e-02 -5.71442723e-01 -7.82851934e-01
-8.08537364e-01 5.89922488e-01 5.52230358e-01 3.40044737e-01
5.44316828e-01 -1.22477984e+00 -8.83291602e-01 -8.07777420e-02
1.75741270e-01 -4.34953906e-02 -1.80375576e-01 7.33447433e-01
-1.91337287e-01 3.04611504e-01 1.99162513e-01 -6.43381238e-01
-1.10558558e+00 6.68441892e-01 1.62622496e-01 -4.28788424e-01
-6.03923440e-01 1.18255472e+00 2.76696831e-01 -5.77180505e-01
3.10165107e-01 -5.79860136e-02 -2.04202175e-01 4.83046956e-02
-8.22015926e-02 -1.26489559e-02 -1.19290337e-01 -7.20452487e-01
-1.74824983e-01 3.74332190e-01 -3.98623198e-01 7.07670525e-02
1.50036263e+00 -4.06336635e-01 -3.80911827e-01 4.95742649e-01
1.35115945e+00 -6.17466634e-03 -1.20866132e+00 -8.02188516e-01
1.36577487e-01 -1.79753110e-01 3.49474013e-01 -5.40339828e-01
-1.24777150e+00 7.47528315e-01 -4.89743575e-02 4.94236022e-01
7.16951430e-01 5.12083292e-01 5.81930757e-01 3.44303399e-01
-2.18721673e-01 -6.06245041e-01 5.39473176e-01 4.24304754e-01
8.09259057e-01 -1.40695977e+00 3.55754048e-01 -6.04427993e-01
-3.85315716e-01 1.22411180e+00 6.42178237e-01 -1.36219546e-01
8.69723976e-01 1.37478635e-01 -3.45458001e-01 -6.29451156e-01
-1.27798045e+00 -3.17193270e-01 5.08386731e-01 4.34743911e-01
9.34746742e-01 -2.70850450e-01 6.37823762e-03 2.10969210e-01
-6.61938563e-02 -4.81689751e-01 4.13589418e-01 8.05359960e-01
-3.39187801e-01 -1.19436836e+00 1.91163689e-01 9.35551226e-01
-3.25860232e-01 -5.44130981e-01 -6.22803926e-01 5.71721613e-01
-1.20467134e-01 8.03548992e-01 4.97031361e-01 -3.58296335e-01
-3.03879380e-01 1.57577526e-02 4.68527526e-01 -1.10885990e+00
-4.09466833e-01 5.45956306e-02 2.83429444e-01 -9.19542238e-02
-5.33234715e-01 -8.37030232e-01 -1.15771770e+00 -4.39776868e-01
-4.72955465e-01 2.38900170e-01 4.37635928e-01 8.09776604e-01
3.29766542e-01 4.58601743e-01 6.77886367e-01 -7.85554290e-01
-1.39759496e-01 -1.00240624e+00 -4.01020080e-01 5.67523718e-01
2.09533840e-01 -3.87342840e-01 -3.83364797e-01 3.55459481e-01] | [7.18619441986084, 6.244821548461914] |
080f7b46-94fb-4003-becc-7e5ad80e9a5a | nested-named-entity-recognition-as-holistic | 2204.08006 | null | https://arxiv.org/abs/2204.08006v1 | https://arxiv.org/pdf/2204.08006v1.pdf | Nested Named Entity Recognition as Holistic Structure Parsing | As a fundamental natural language processing task and one of core knowledge extraction techniques, named entity recognition (NER) is widely used to extract information from texts for downstream tasks. Nested NER is a branch of NER in which the named entities (NEs) are nested with each other. However, most of the previous studies on nested NER usually apply linear structure to model the nested NEs which are actually accommodated in a hierarchical structure. Thus in order to address this mismatch, this work models the full nested NEs in a sentence as a holistic structure, then we propose a holistic structure parsing algorithm to disclose the entire NEs once for all. Besides, there is no research on applying corpus-level information to NER currently. To make up for the loss of this information, we introduce Point-wise Mutual Information (PMI) and other frequency features from corpus-aware statistics for even better performance by holistic modeling from sentence-level to corpus-level. Experiments show that our model yields promising results on widely-used benchmarks which approach or even achieve state-of-the-art. Further empirical studies show that our proposed corpus-aware features can substantially improve NER domain adaptation, which demonstrates the surprising advantage of our proposed corpus-level holistic structure modeling. | ['Hai Zhao', 'Zuchao Li', 'Yifei Yang'] | 2022-04-17 | null | null | null | null | ['nested-named-entity-recognition'] | ['natural-language-processing'] | [-3.74746043e-03 9.40935910e-02 -3.48754376e-01 -4.05753851e-01
-8.58252168e-01 -7.14089751e-01 4.54086363e-01 4.87092763e-01
-6.29156351e-01 7.61955082e-01 7.66461253e-01 -4.51611787e-01
9.10529401e-03 -9.25772667e-01 -5.98239362e-01 -2.49774575e-01
-1.61735788e-01 1.08978838e-01 3.49163204e-01 -5.27225792e-01
3.14261675e-01 4.36964005e-01 -1.12854803e+00 4.43929166e-01
1.22959709e+00 4.07388777e-01 3.11748087e-01 4.41696852e-01
-9.40799356e-01 8.40688050e-01 -6.13559306e-01 -6.80483699e-01
1.89558286e-02 -3.49136233e-01 -1.26645029e+00 -3.23597938e-01
5.00956587e-02 1.04240432e-01 -1.59165293e-01 9.97127235e-01
3.44377905e-01 2.23397523e-01 5.76379240e-01 -5.51215589e-01
-5.73773503e-01 1.30781269e+00 -5.17198741e-01 4.48600829e-01
4.44096535e-01 -3.06252509e-01 1.27621591e+00 -7.57810652e-01
7.18810141e-01 9.89134252e-01 7.66812444e-01 5.03174543e-01
-7.72446215e-01 -4.67999995e-01 2.00183764e-01 5.62616251e-02
-1.15360510e+00 -4.00550306e-01 5.49321115e-01 -4.58783805e-02
1.21105695e+00 2.30318606e-01 -4.99360375e-02 8.43727112e-01
2.42597178e-01 7.75143027e-01 1.01028419e+00 -6.74467862e-01
-2.93660052e-02 9.19958130e-02 6.95339024e-01 4.88062173e-01
3.90263438e-01 -2.38132820e-01 -3.02839071e-01 2.21794881e-02
3.29379499e-01 -2.47282580e-01 -5.26519790e-02 2.42208764e-01
-1.07528186e+00 7.31159985e-01 2.73124605e-01 8.80188465e-01
-3.24453205e-01 -5.02846479e-01 5.43957591e-01 1.73021406e-01
3.49764079e-01 5.07127881e-01 -9.89775002e-01 -3.02611202e-01
-1.03312862e+00 -2.42457584e-01 1.19767201e+00 1.13170385e+00
7.61387944e-01 -2.33172953e-01 -2.64135212e-01 8.46571982e-01
3.16929147e-02 3.96131836e-02 7.52187848e-01 -4.47911143e-01
8.87801290e-01 1.01616371e+00 -1.41355783e-01 -9.08901155e-01
-6.82133377e-01 -5.72963536e-01 -9.27611470e-01 -7.04640567e-01
2.25418821e-01 -3.71987432e-01 -6.04061604e-01 1.70611393e+00
4.24457282e-01 1.13224462e-01 2.88512319e-01 5.57282865e-01
1.17650640e+00 7.15606570e-01 4.74868894e-01 -3.10346901e-01
1.77999973e+00 -9.60968614e-01 -8.72461438e-01 -2.19536692e-01
1.20102370e+00 -6.99361563e-01 7.50732303e-01 1.35116167e-02
-6.82917774e-01 -4.85522479e-01 -7.49916732e-01 -2.78956026e-01
-7.56048560e-01 2.35683009e-01 6.42464995e-01 9.42497849e-01
-6.43279374e-01 6.61418021e-01 -7.58239985e-01 -4.76926476e-01
1.48838433e-02 1.01358958e-01 -6.39484525e-01 7.53107443e-02
-1.58359659e+00 7.85031140e-01 7.68357694e-01 9.20114741e-02
-6.63052797e-02 -7.74653912e-01 -1.14833724e+00 2.96388507e-01
5.58267713e-01 -5.89299083e-01 1.03987753e+00 -3.69431913e-01
-1.16929436e+00 6.17835581e-01 -6.43924832e-01 -4.87612873e-01
-1.42421247e-02 -2.41784588e-01 -7.21540749e-01 -5.27451858e-02
3.23172629e-01 2.49039114e-01 1.77708149e-01 -1.16729808e+00
-7.40631819e-01 -3.75492394e-01 1.19265251e-01 1.26061887e-01
-7.21879184e-01 3.98664325e-01 -2.60383695e-01 -5.74507236e-01
1.96910650e-01 -4.32318330e-01 -3.72958452e-01 -1.01630843e+00
-7.14125752e-01 -6.52162433e-01 3.43624651e-01 -8.17725122e-01
2.00534201e+00 -1.91558611e+00 -2.70818412e-01 -1.42646849e-01
1.34909585e-01 2.35854000e-01 -1.32813364e-01 9.23246861e-01
-1.42311335e-01 7.43723035e-01 -2.82812566e-01 -3.73756170e-01
8.51079151e-02 1.72299683e-01 -3.85725945e-01 -1.80200394e-02
3.96463782e-01 9.44453359e-01 -7.69905031e-01 -7.76163757e-01
-1.01989470e-01 2.98630178e-01 -3.15789133e-01 4.43682596e-02
2.37373114e-01 3.33900392e-01 -6.27718806e-01 3.47947508e-01
7.43026793e-01 -8.94878898e-03 3.02307367e-01 -3.18904638e-01
-3.71484727e-01 7.40300417e-01 -1.18482637e+00 1.59697449e+00
-5.24280608e-01 1.78514346e-01 -1.44089639e-01 -1.07089269e+00
8.18433046e-01 4.02100533e-01 2.79186904e-01 -4.87582088e-01
2.77239047e-02 1.24846034e-01 -6.56101257e-02 -4.53157991e-01
7.34137893e-01 -2.52393603e-01 -5.49958289e-01 4.37343009e-02
1.47025064e-01 4.62446004e-01 5.67264855e-01 3.34539443e-01
1.52009106e+00 -9.20396857e-03 8.18209171e-01 -3.01742345e-01
7.17686474e-01 2.69074738e-01 9.24234331e-01 8.63509953e-01
-1.85404375e-01 3.30872536e-01 4.23456341e-01 -1.66323900e-01
-9.03044999e-01 -7.33377218e-01 -3.00179183e-01 1.17821991e+00
-1.05801761e-01 -5.98013639e-01 -1.00176156e+00 -9.71145630e-01
-4.30751294e-01 9.26866233e-01 -4.53204572e-01 2.55417854e-01
-9.24113870e-01 -6.94281995e-01 8.84951532e-01 6.47363305e-01
5.55851519e-01 -1.18352067e+00 -9.76717249e-02 4.61413920e-01
-5.73575020e-01 -1.45920193e+00 -5.93999803e-01 3.92572910e-01
-1.05581379e+00 -8.25344145e-01 -3.62942398e-01 -9.25057769e-01
2.44066715e-01 8.55416134e-02 1.16297960e+00 -1.22493759e-01
4.32688408e-02 9.30124149e-02 -8.81906033e-01 -1.96401119e-01
-4.86421138e-01 6.33902550e-01 -3.82915363e-02 -2.44894758e-01
6.66259825e-01 -5.70589125e-01 -1.95288524e-01 1.18362583e-01
-9.83705699e-01 -3.67102712e-01 8.64940226e-01 8.11223447e-01
4.33742017e-01 4.40586686e-01 7.52922654e-01 -1.30645871e+00
6.61844969e-01 -6.41565442e-01 -2.85332482e-02 3.44324052e-01
-4.31104153e-01 3.39135259e-01 1.08918405e+00 -1.89163357e-01
-1.44448841e+00 -1.12264626e-01 -4.13504303e-01 3.18397164e-01
-7.05282331e-01 9.70383942e-01 -5.81621170e-01 5.50428689e-01
4.55140591e-01 4.71324027e-01 -6.14165902e-01 -8.46042931e-01
4.10456777e-01 9.28398192e-01 3.84293288e-01 -6.30703568e-01
7.40151584e-01 1.97299570e-01 1.11707794e-02 -9.56960440e-01
-1.37231421e+00 -1.05283332e+00 -1.07812095e+00 4.40593570e-01
8.36465538e-01 -8.65730762e-01 -4.71757233e-01 -1.33002967e-01
-1.43075049e+00 2.56348997e-01 -6.85720816e-02 3.87077332e-01
-4.43162620e-02 8.83949101e-01 -8.63272130e-01 -8.95675421e-01
-5.15412688e-01 -5.68900168e-01 9.85217035e-01 3.76425654e-01
-1.15411945e-01 -1.11054564e+00 2.02893674e-01 4.58286166e-01
1.12891689e-01 -8.10305998e-02 1.03672373e+00 -1.34435320e+00
-1.21876515e-01 -1.84373021e-01 -1.45055696e-01 2.17997670e-01
1.07854947e-01 -3.91770065e-01 -7.11536169e-01 9.20058712e-02
1.06690854e-01 2.26899818e-01 9.19126332e-01 3.63998301e-02
7.34989524e-01 -4.27771538e-01 -4.51340497e-01 2.49532059e-01
1.52986133e+00 7.25995526e-02 6.68592691e-01 5.95308185e-01
6.82705879e-01 9.82549071e-01 4.71084744e-01 3.39187056e-01
6.69883788e-01 2.41551876e-01 3.15290429e-02 1.88671295e-02
5.00070304e-02 -4.69055682e-01 5.08467674e-01 1.33982062e+00
1.35379478e-01 -3.86472195e-01 -9.20846999e-01 6.84162796e-01
-1.45537591e+00 -1.15903878e+00 -5.67963123e-01 1.84258604e+00
9.27213848e-01 1.95113420e-01 -3.07945777e-02 4.26790603e-02
9.34801579e-01 2.28116568e-03 -9.68690738e-02 -5.78306913e-01
-3.34158570e-01 3.36049169e-01 5.81625462e-01 1.24820478e-01
-1.35139775e+00 1.13249016e+00 5.79461193e+00 1.20119119e+00
-6.88220799e-01 9.40837786e-02 2.45558634e-01 6.15491986e-01
-2.42447689e-01 3.10182393e-01 -1.47451234e+00 2.51501054e-01
1.35743666e+00 -2.31175452e-01 7.88198784e-02 8.19050074e-01
2.35264063e-01 1.42243952e-01 -9.28371906e-01 5.24436951e-01
-1.20665222e-01 -1.04197991e+00 1.88695267e-01 1.82884589e-01
5.70095003e-01 3.29066440e-02 -4.89882082e-01 6.67848349e-01
2.24856868e-01 -8.11677277e-01 3.96449447e-01 3.34044904e-01
4.08260256e-01 -8.40918183e-01 1.11822236e+00 8.78579974e-01
-1.63835216e+00 -1.27356708e-01 -5.30232847e-01 -4.07339409e-02
3.82729828e-01 8.79887223e-01 -6.63607955e-01 1.35020578e+00
5.90312183e-01 4.51771826e-01 -5.14991343e-01 8.51029158e-01
-1.40391901e-01 1.04954958e+00 -3.22707087e-01 -2.69278824e-01
3.08636039e-01 -7.25265071e-02 6.31777287e-01 1.84348333e+00
2.23381579e-01 2.81641513e-01 1.07178323e-01 6.68884933e-01
-2.46396050e-01 7.49673009e-01 -5.31294703e-01 -2.97640890e-01
6.29820049e-01 1.49390197e+00 -7.32724845e-01 -2.91972637e-01
-5.84802866e-01 9.50507462e-01 5.63778937e-01 -1.35036567e-02
-5.49820542e-01 -8.52926373e-01 4.21985209e-01 -7.67301247e-02
5.50153911e-01 -4.00794953e-01 -4.78279740e-01 -1.42139649e+00
1.60525829e-01 -5.30369818e-01 6.63859785e-01 -1.32389247e-01
-1.82392049e+00 8.25239778e-01 -2.02536151e-01 -1.23061144e+00
1.39043238e-02 -6.12909734e-01 -6.37853742e-01 7.20033824e-01
-1.64835835e+00 -1.00726521e+00 2.62845933e-01 2.90148735e-01
6.14380121e-01 1.20044328e-01 9.31756914e-01 2.87805885e-01
-7.82390714e-01 7.91731715e-01 1.67141140e-01 8.60025823e-01
6.58335626e-01 -1.39499617e+00 5.00979543e-01 1.00035059e+00
4.97203231e-01 1.21815813e+00 3.71338069e-01 -7.89203346e-01
-1.22228980e+00 -9.77738917e-01 1.70209241e+00 -7.63541102e-01
8.08265507e-01 -2.42426142e-01 -9.90499496e-01 5.30416369e-01
3.67973179e-01 -2.89229125e-01 1.06264758e+00 5.03527105e-01
-4.38417315e-01 8.56309161e-02 -9.91951585e-01 5.76674879e-01
1.21996880e+00 -3.05647075e-01 -1.26633167e+00 -8.15823898e-02
1.09966457e+00 -1.78461134e-01 -1.19276464e+00 3.64785552e-01
1.38441592e-01 -8.94088507e-01 5.59466720e-01 -7.90148318e-01
5.20373344e-01 -1.80160403e-01 -1.26717120e-01 -1.10476792e+00
-2.06117213e-01 -5.12381971e-01 -3.44199762e-02 1.88281369e+00
7.38961875e-01 -5.14867127e-01 6.32030308e-01 3.79818171e-01
-3.99607450e-01 -5.72069168e-01 -1.02261615e+00 -9.93971884e-01
5.13205647e-01 -5.41485190e-01 7.05714524e-01 1.26158559e+00
4.58309174e-01 7.72755682e-01 -8.58622715e-02 2.58891284e-01
4.55438703e-01 -4.51668575e-02 3.04999679e-01 -1.18970394e+00
-7.31015652e-02 -3.24070185e-01 -1.89141244e-01 -1.35469961e+00
3.25355291e-01 -1.07125115e+00 1.39409482e-01 -1.67887771e+00
5.37170589e-01 -2.32220337e-01 -3.53602767e-01 4.73012626e-01
-3.92342657e-01 -3.40329498e-01 2.97411606e-02 1.80493847e-01
-6.73247337e-01 4.69729602e-01 1.01369703e+00 1.32350579e-01
-2.92015523e-01 -2.12689817e-01 -1.00648713e+00 5.70712090e-01
6.63108528e-01 -7.28909314e-01 5.60900308e-02 -4.01180685e-01
1.58102632e-01 1.10601215e-02 -2.28922367e-01 -7.86305487e-01
5.73683977e-01 -5.76785468e-02 1.65468380e-01 -7.08525896e-01
-2.37647548e-01 -7.85113275e-01 -2.81466186e-01 9.57962647e-02
-2.65909553e-01 -9.39514041e-02 1.04728699e-01 7.31153488e-01
-5.71241200e-01 -6.30520821e-01 2.76229203e-01 -2.21676394e-01
-7.60464907e-01 1.25205219e-01 -3.16701174e-01 3.30806822e-01
5.81797063e-01 -2.10532546e-01 -2.26033553e-01 9.82632861e-02
-5.72067797e-01 1.48840114e-01 -6.49508759e-02 4.33721632e-01
2.46789768e-01 -9.89905775e-01 -7.40684807e-01 -1.35988131e-01
7.44136721e-02 -7.94109926e-02 4.16674793e-01 7.58977592e-01
-1.55175447e-01 7.53250420e-01 3.67523789e-01 -2.24963054e-01
-1.02723801e+00 5.74212492e-01 -4.66447212e-02 -9.75394130e-01
-4.98626739e-01 5.26045740e-01 1.11240059e-01 -9.30724382e-01
-2.23383769e-01 -3.32458258e-01 -7.62661576e-01 9.38700363e-02
5.82568765e-01 2.90507048e-01 2.41355613e-01 -8.14889908e-01
-3.83245707e-01 4.42857116e-01 -2.16590479e-01 1.03187017e-01
1.26526189e+00 -4.83534366e-01 -3.77724558e-01 3.63550514e-01
1.06923795e+00 5.80769360e-01 -4.43341374e-01 -3.60347539e-01
7.78109610e-01 1.24868877e-01 -1.42083913e-01 -6.63128853e-01
-4.83686298e-01 8.28051269e-01 -3.98374423e-02 3.15440416e-01
1.08179188e+00 6.83668926e-02 1.12188518e+00 6.68085515e-01
4.98782188e-01 -1.03905118e+00 -4.87777352e-01 1.02575397e+00
2.98469514e-01 -1.11004317e+00 -3.07978243e-01 -8.23999584e-01
-6.43601418e-01 1.29397917e+00 4.50752378e-01 1.10646605e-01
7.44422257e-01 4.17955160e-01 -8.37004781e-02 1.28229856e-01
-5.95584631e-01 -4.54067796e-01 3.91394943e-01 3.88371736e-01
7.55814910e-01 -2.57512480e-02 -9.23746109e-01 1.47060311e+00
-3.34650457e-01 -3.36858273e-01 3.88945699e-01 9.04330969e-01
-7.45108008e-01 -1.39338255e+00 -1.45395175e-01 3.99601847e-01
-9.93071496e-01 -6.13626361e-01 -3.12620521e-01 7.67287731e-01
6.34820610e-02 1.29749310e+00 -2.63125420e-01 -3.46516430e-01
6.35435522e-01 3.67574573e-01 5.14203757e-02 -9.63791013e-01
-9.69289064e-01 -1.68244869e-01 2.68484503e-01 -2.54546911e-01
-3.81214350e-01 -5.12169957e-01 -1.70874953e+00 -2.13009223e-01
-5.71289718e-01 6.29368782e-01 5.88320971e-01 1.32761586e+00
4.48544741e-01 4.03085649e-01 6.73295200e-01 -1.36792794e-01
-6.78011417e-01 -1.24276829e+00 -7.08842933e-01 3.42711180e-01
-1.34485990e-01 -2.77513206e-01 -4.71323997e-01 9.96765345e-02] | [9.665306091308594, 9.487525939941406] |
d16cd15e-b56b-4e0a-8073-b763660d501d | the-objectfolder-benchmark-multisensory-1 | 2306.00956 | null | https://arxiv.org/abs/2306.00956v1 | https://arxiv.org/pdf/2306.00956v1.pdf | The ObjectFolder Benchmark: Multisensory Learning with Neural and Real Objects | We introduce the ObjectFolder Benchmark, a benchmark suite of 10 tasks for multisensory object-centric learning, centered around object recognition, reconstruction, and manipulation with sight, sound, and touch. We also introduce the ObjectFolder Real dataset, including the multisensory measurements for 100 real-world household objects, building upon a newly designed pipeline for collecting the 3D meshes, videos, impact sounds, and tactile readings of real-world objects. We conduct systematic benchmarking on both the 1,000 multisensory neural objects from ObjectFolder, and the real multisensory data from ObjectFolder Real. Our results demonstrate the importance of multisensory perception and reveal the respective roles of vision, audio, and touch for different object-centric learning tasks. By publicly releasing our dataset and benchmark suite, we hope to catalyze and enable new research in multisensory object-centric learning in computer vision, robotics, and beyond. Project page: https://objectfolder.stanford.edu | ['Jiajun Wu', 'Li Fei-Fei', 'Yunzhu Li', 'Jeannette Bohg', 'Tanmay Agarwal', 'Hao Li', 'Yiming Dou', 'Ruohan Gao'] | 2023-06-01 | the-objectfolder-benchmark-multisensory | http://openaccess.thecvf.com//content/CVPR2023/html/Gao_The_ObjectFolder_Benchmark_Multisensory_Learning_With_Neural_and_Real_Objects_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Gao_The_ObjectFolder_Benchmark_Multisensory_Learning_With_Neural_and_Real_Objects_CVPR_2023_paper.pdf | cvpr-2023-1 | ['object-recognition'] | ['computer-vision'] | [ 1.44057140e-01 -4.50953215e-01 1.43566042e-01 -2.80449629e-01
-9.19523537e-01 -6.66853249e-01 4.54094142e-01 1.35304302e-01
-1.30969167e-01 1.20862871e-01 2.48188272e-01 4.57029015e-01
-1.51910186e-01 -5.37137032e-01 -1.51969171e+00 -4.74979341e-01
-3.69150117e-02 3.96658480e-01 8.90935510e-02 9.36154127e-02
1.98721513e-01 4.83155996e-01 -2.15850043e+00 6.44025266e-01
-1.32999375e-01 1.71741688e+00 4.83126938e-01 1.00634789e+00
1.41982764e-01 5.28285444e-01 -2.08506748e-01 -3.53530720e-02
2.66898841e-01 4.46230292e-01 -6.29903793e-01 -2.65924204e-02
1.15090287e+00 -4.28662151e-01 -2.69727200e-01 9.69317853e-01
6.77341163e-01 -2.57116780e-02 5.43307722e-01 -1.43909323e+00
-1.07888687e+00 6.45646155e-01 -2.68920541e-01 -2.94825673e-01
5.88368058e-01 6.06494963e-01 1.27422535e+00 -1.17299771e+00
5.02845526e-01 1.73738027e+00 7.96884418e-01 5.62613547e-01
-1.08639073e+00 -6.28969312e-01 -1.48091279e-02 1.98960081e-01
-9.69955564e-01 -6.11726820e-01 5.88998497e-01 -6.73002839e-01
9.12321806e-01 2.85099864e-01 6.84533179e-01 1.65248466e+00
-5.98186515e-02 1.00570393e+00 9.55444694e-01 8.33642036e-02
3.53317410e-01 -1.42684400e-01 4.78924140e-02 6.99999750e-01
6.51800111e-02 2.49677807e-01 -8.21284235e-01 -5.02343737e-02
7.99803793e-01 3.45943332e-01 -6.58977479e-02 -4.60389018e-01
-1.82542992e+00 2.00637192e-01 6.55355334e-01 -2.74189591e-01
-3.97914052e-01 7.46560156e-01 2.46768445e-01 2.14055613e-01
-5.03854491e-02 5.26190341e-01 -7.32488990e-01 -2.09956199e-01
-5.70591241e-02 3.53800118e-01 8.99790525e-01 1.35592782e+00
6.27632856e-01 -6.49370849e-02 1.56757459e-01 8.58345926e-01
6.01773143e-01 1.19867432e+00 1.63562879e-01 -1.38402581e+00
3.78110737e-01 5.59851289e-01 2.62284458e-01 -7.06364095e-01
-5.00541627e-01 1.25580594e-01 -4.40480441e-01 2.02251434e-01
1.91926256e-01 1.84150010e-01 -9.96315241e-01 1.26682639e+00
4.54249084e-01 4.63128358e-01 -5.36871813e-02 1.18004429e+00
1.68809533e+00 3.63154292e-01 -3.11208684e-02 5.00368357e-01
1.16095567e+00 -9.01350439e-01 -2.51984060e-01 6.72731772e-02
1.11533828e-01 -7.22674012e-01 1.48906147e+00 8.47963929e-01
-8.46387506e-01 -7.97447622e-01 -1.00460351e+00 -2.48845354e-01
-6.19345248e-01 -4.26515704e-03 8.56277168e-01 -1.67882577e-01
-6.04956269e-01 4.98643577e-01 -9.11318243e-01 -4.72900361e-01
5.71596444e-01 2.47066081e-01 -7.04731703e-01 -2.44974986e-01
-3.60545963e-01 6.65403247e-01 1.89632818e-01 7.65939504e-02
-1.44617510e+00 -8.53499353e-01 -8.69955003e-01 -2.89275557e-01
2.23990828e-01 -4.39330041e-01 1.27899849e+00 -1.80459112e-01
-1.37312877e+00 9.34519887e-01 4.56931174e-01 -1.12219334e-01
-6.34261640e-03 -5.81382453e-01 -2.39653111e-01 5.07237986e-02
-1.83430500e-02 9.10394549e-01 8.74123335e-01 -1.65279269e+00
-3.90844166e-01 -5.32365441e-01 -1.08736120e-01 -1.78472415e-01
-1.42925486e-01 -1.69283286e-01 -3.57407242e-01 -3.22827041e-01
2.02004895e-01 -8.78183961e-01 1.64431989e-01 5.99114299e-01
-8.17984045e-01 -3.37717354e-01 8.27987909e-01 -2.48891056e-01
-3.86554636e-02 -2.42759681e+00 3.79944026e-01 -2.86410272e-01
2.69633591e-01 -2.29165614e-01 -7.44970322e-01 2.76350439e-01
2.89965332e-01 -2.12343216e-01 2.95421779e-01 -6.73920214e-01
3.62761289e-01 3.35262895e-01 -5.49476266e-01 4.50949132e-01
9.34222043e-02 1.14647686e+00 -8.89948964e-01 -3.72715414e-01
5.66235125e-01 3.79933953e-01 -5.19341171e-01 3.62195164e-01
-7.56899834e-01 4.11672533e-01 -3.55811417e-01 1.63114607e+00
5.00777364e-01 -2.63492137e-01 -3.97783011e-01 -6.01838768e-01
1.25223145e-01 1.56705648e-01 -1.15320158e+00 2.17053533e+00
-4.59410548e-01 3.51967156e-01 5.87819397e-01 -5.22551596e-01
7.15585828e-01 2.06733242e-01 6.36210978e-01 -6.80750549e-01
2.15490207e-01 1.10592872e-01 -4.32093680e-01 -9.55014586e-01
3.55123818e-01 2.35914931e-01 -2.80608386e-01 2.31290519e-01
4.09540921e-01 -7.00868607e-01 -3.48329544e-01 -9.65642929e-02
1.15635753e+00 2.96377361e-01 -8.30139816e-02 2.82083869e-01
-3.36706161e-01 -7.89731294e-02 1.30823314e-01 7.78949440e-01
-7.56527111e-02 7.27818727e-01 -1.46399677e-01 -4.44826663e-01
-7.05445945e-01 -1.72268534e+00 -2.38713160e-01 1.56460643e+00
3.22154641e-01 -2.54753739e-01 -4.14102077e-02 -2.63447464e-01
8.93203735e-01 2.47842059e-01 -6.71594739e-01 7.61800557e-02
-2.17096776e-01 -2.60269701e-01 5.36871672e-01 8.48535538e-01
-8.44888762e-03 -1.55806375e+00 -6.91878915e-01 7.60232359e-02
7.62653584e-03 -1.37023997e+00 -2.87407767e-02 3.02630216e-01
-6.17524445e-01 -1.34159172e+00 -2.37876013e-01 -4.79146898e-01
7.41860792e-02 3.16447854e-01 1.14707446e+00 -1.97986305e-01
-7.02967882e-01 1.20551133e+00 -3.86590153e-01 -9.18751776e-01
-1.38390914e-01 -2.83862382e-01 4.41134900e-01 -1.37292966e-01
6.59461170e-02 -7.06910789e-01 -2.46290818e-01 2.70584762e-01
-6.82869852e-01 -1.31473511e-01 5.25394619e-01 3.96674454e-01
9.72219050e-01 -6.30628705e-01 4.16694582e-01 1.20619750e-02
-6.08478710e-02 -6.96782470e-01 -6.69914126e-01 1.40182570e-01
1.23649284e-01 -4.19449449e-01 1.27221435e-01 -7.56865144e-01
-3.87903214e-01 3.61560374e-01 -1.73231304e-01 -8.44586134e-01
-6.19491458e-01 -9.84426681e-03 -4.05719578e-01 -3.48097205e-01
6.30199194e-01 8.30952674e-02 -1.45423636e-01 -9.16316986e-01
6.85684264e-01 7.08383679e-01 9.97009993e-01 -7.96088755e-01
5.00992298e-01 6.98916674e-01 3.55471633e-02 -8.44879150e-01
-9.18326437e-01 -5.62504828e-01 -5.46553016e-01 -3.44017655e-01
9.55733359e-01 -1.14849293e+00 -1.56663013e+00 7.34307587e-01
-1.18086624e+00 -5.88258386e-01 -4.92379993e-01 5.18653572e-01
-9.69164610e-01 -2.18592018e-01 -5.36036491e-01 -6.58162415e-01
-2.66754806e-01 -1.00901246e+00 1.99288261e+00 3.43349800e-02
-6.26177415e-02 -4.83374536e-01 2.36836285e-03 5.54733574e-01
1.09858111e-01 5.82853734e-01 5.55072665e-01 -4.59490836e-01
-1.07088161e+00 -3.08151972e-02 -3.35556179e-01 4.02155936e-01
1.10923164e-01 -6.87116235e-02 -1.19554603e+00 -1.04049481e-01
-3.63703907e-01 -1.29746449e+00 1.02897859e+00 9.34162363e-03
1.34392607e+00 -1.93537205e-01 -2.44687811e-01 8.71753693e-01
1.45689666e+00 -3.31045806e-01 -7.16053769e-02 4.06819731e-02
1.31233156e+00 3.73909891e-01 6.21664762e-01 5.86488724e-01
7.08583236e-01 4.65962261e-01 1.40847838e+00 2.38267362e-01
-2.37893984e-01 -2.41733015e-01 1.31253868e-01 8.07030737e-01
1.68725386e-01 -3.10083218e-02 -8.61216784e-01 6.87415183e-01
-1.64075720e+00 -7.36970007e-01 -8.62404332e-03 1.69487762e+00
7.24998176e-01 -1.72136560e-01 3.92081328e-02 1.22713819e-01
1.14301600e-01 1.27684884e-02 -1.15927076e+00 1.20398492e-01
-2.25011915e-01 2.68946849e-02 1.98841959e-01 5.97549714e-02
-1.29158318e+00 7.70511508e-01 6.67149782e+00 1.53745502e-01
-1.38136280e+00 1.31288216e-01 -3.00863773e-01 -7.99157679e-01
-2.40978718e-01 -7.18775809e-01 -8.08028460e-01 3.50392014e-01
6.98839128e-01 6.15176201e-01 7.78553545e-01 1.03883386e+00
-3.22595149e-01 -6.03335304e-03 -1.68519354e+00 1.36183047e+00
5.86008839e-02 -1.40896022e+00 -1.57646984e-01 -1.40399665e-01
4.59053546e-01 9.25566316e-01 1.38816554e-02 2.14397207e-01
5.48994482e-01 -1.10281873e+00 1.13662219e+00 8.46419215e-01
8.43878031e-01 4.76490855e-02 1.62351862e-01 2.22068235e-01
-1.07944155e+00 -3.36170822e-01 -1.77817345e-01 -2.51878966e-02
-1.74709424e-01 4.15887654e-01 -5.78520775e-01 5.97482920e-02
1.20758653e+00 1.17228496e+00 -4.00142252e-01 1.00675941e+00
1.57817110e-01 4.35909867e-01 -6.31962359e-01 -5.10771312e-02
-1.98516145e-01 5.37245035e-01 7.56140471e-01 1.09046376e+00
5.68811819e-02 -1.71829611e-01 3.43578041e-01 9.14776564e-01
-3.21428955e-01 -2.49002084e-01 -6.77665114e-01 -1.92585304e-01
6.45906031e-01 1.28449583e+00 -1.74852744e-01 -1.20043531e-01
-4.38646108e-01 6.84696138e-01 2.73336500e-01 2.44152248e-01
-5.76228380e-01 -6.52389228e-02 1.07556403e+00 -1.59397528e-01
4.33687508e-01 -3.39150578e-01 -6.12673938e-01 -1.04998195e+00
2.53802717e-01 -7.74462342e-01 2.27932520e-02 -1.22827959e+00
-1.71661592e+00 -5.99459000e-02 -2.57792175e-01 -9.76336241e-01
1.77015364e-01 -1.03181481e+00 -3.30928653e-01 3.44877928e-01
-1.22018898e+00 -1.57876229e+00 -8.01747918e-01 7.47126460e-01
4.01821673e-01 -3.50253023e-02 1.05869496e+00 2.22033054e-01
-8.64225626e-02 2.67720908e-01 -1.12603769e-01 7.46893212e-02
7.32260764e-01 -1.29953146e+00 3.81909639e-01 -1.20984368e-01
2.87831903e-01 2.74656624e-01 5.06647944e-01 -4.71905500e-01
-2.50479245e+00 -1.28548837e+00 -1.61095634e-01 -1.10477710e+00
9.56673384e-01 -8.08545053e-01 -6.42626047e-01 8.53495359e-01
-3.26847821e-01 4.16246265e-01 4.64866132e-01 2.10243374e-01
-9.20731783e-01 -2.89402813e-01 -1.16862559e+00 1.37294516e-01
1.38330483e+00 -8.00475955e-01 -7.04760194e-01 6.10135794e-01
1.23127747e+00 -6.20782256e-01 -1.38390875e+00 6.53565764e-01
1.07024717e+00 -7.28690982e-01 1.43796325e+00 -5.37168384e-01
5.21096468e-01 -1.10450216e-01 -1.00241292e+00 -1.10295725e+00
-7.22361058e-02 -1.36056930e-01 -4.99259353e-01 1.17028117e+00
-1.12950988e-01 -4.04534966e-01 4.60007370e-01 -9.67687294e-02
-3.82066786e-01 -7.24605799e-01 -1.03693128e+00 -8.08776975e-01
-2.53114045e-01 -8.68041873e-01 8.12896371e-01 6.87256813e-01
-4.32566494e-01 6.08933792e-02 -2.36120373e-01 3.61959428e-01
1.02973926e+00 4.88108337e-01 1.04971600e+00 -1.59957743e+00
-5.74997962e-01 -1.59425110e-01 -5.26188731e-01 -8.10236096e-01
3.32799032e-02 -7.77749717e-01 5.59701502e-01 -1.55068159e+00
5.90782724e-02 -4.20558780e-01 -5.01875639e-01 8.71812582e-01
3.29848081e-01 5.01887977e-01 5.26367784e-01 3.39235932e-01
-1.08903778e+00 6.45918787e-01 1.46304381e+00 -4.79155421e-01
1.34253338e-01 -1.49913207e-01 -5.85060537e-01 8.50356042e-01
4.62482750e-01 -2.32079431e-01 1.76374286e-01 -9.64105546e-01
9.53539163e-02 -4.12476622e-02 1.19715869e+00 -1.27962887e+00
2.01485083e-01 -2.63009608e-01 5.06371856e-01 -8.81313026e-01
1.09430814e+00 -9.87058043e-01 -9.04150531e-02 1.80799980e-02
-3.32456261e-01 -2.65395015e-01 2.68527806e-01 7.44188726e-01
1.69195950e-01 4.78013813e-01 5.40248036e-01 -2.41397798e-01
-9.37487364e-01 4.30288106e-01 1.95995420e-01 -8.07885379e-02
7.37383544e-01 -5.28392987e-03 -5.13719916e-01 1.17739402e-01
-9.30402160e-01 3.11707348e-01 4.04900998e-01 1.06695008e+00
8.43647957e-01 -1.52210593e+00 -5.15274346e-01 2.71687001e-01
5.71467817e-01 3.46868575e-01 8.13179761e-02 5.25720537e-01
8.88891593e-02 2.34240457e-01 -4.88302737e-01 -1.01570559e+00
-1.19413733e+00 5.76444209e-01 6.70224503e-02 8.72483611e-01
-6.40856087e-01 1.18842363e+00 -5.03393076e-02 -1.04247499e+00
9.24811363e-01 -9.17632043e-01 3.23651463e-01 -2.76785821e-01
4.01026279e-01 5.69372058e-01 -1.75311476e-01 -3.88812810e-01
-4.66292411e-01 8.63846540e-01 3.94940883e-01 6.97973669e-02
1.55201697e+00 1.19792685e-01 -1.80824623e-01 1.12791133e+00
1.15758383e+00 -2.91863084e-01 -1.50055099e+00 -4.30420041e-01
-1.69002905e-01 -9.86285135e-02 -1.86787501e-01 -9.94536340e-01
-7.28346527e-01 9.94119644e-01 6.62539899e-01 4.24406715e-02
9.65942383e-01 6.73344970e-01 9.68396246e-01 9.79344726e-01
8.25584829e-01 -9.24671531e-01 5.66583872e-01 6.29210353e-01
1.43326628e+00 -1.55890346e+00 -2.88308442e-01 -2.07747042e-01
-1.02576226e-01 7.06885099e-01 7.68375397e-01 -1.70149803e-01
8.30047011e-01 8.02680910e-01 4.83568534e-02 -3.88804644e-01
-9.55784202e-01 -2.00528607e-01 3.20208251e-01 7.66926408e-01
-1.22509912e-01 4.45430100e-01 1.08931661e+00 8.39454949e-01
-3.23055923e-01 5.25455847e-02 -5.49298674e-02 9.06060159e-01
-5.94186902e-01 -4.11281586e-01 -6.21093094e-01 3.77247751e-01
1.08213343e-01 3.44102755e-02 -5.57743609e-01 4.53754008e-01
4.07954276e-01 8.04029047e-01 1.07648686e-01 -7.26857305e-01
5.73550045e-01 -2.32338458e-01 8.51363420e-01 -6.33352339e-01
-3.73358518e-01 -3.00521553e-01 5.81733976e-03 -1.23141873e+00
-3.23388249e-01 -7.19979167e-01 -1.27249670e+00 -3.37719619e-02
9.92777422e-02 -6.72700167e-01 1.42371094e+00 9.09577012e-01
5.07125974e-01 4.85735357e-01 2.23772034e-01 -1.53898883e+00
-7.02209175e-01 -1.03542840e+00 -4.79266107e-01 3.41134816e-01
7.82518685e-01 -8.75396013e-01 -2.71669835e-01 7.49056414e-02] | [7.0136942863464355, -2.1501410007476807] |
1d991f75-1b03-4aac-8694-8dc7a1e171c2 | variational-template-machine-for-data-to-text-1 | 2002.01127 | null | https://arxiv.org/abs/2002.01127v2 | https://arxiv.org/pdf/2002.01127v2.pdf | Variational Template Machine for Data-to-Text Generation | How to generate descriptions from structured data organized in tables? Existing approaches using neural encoder-decoder models often suffer from lacking diversity. We claim that an open set of templates is crucial for enriching the phrase constructions and realizing varied generations. Learning such templates is prohibitive since it often requires a large paired <table, description> corpus, which is seldom available. This paper explores the problem of automatically learning reusable "templates" from paired and non-paired data. We propose the variational template machine (VTM), a novel method to generate text descriptions from data tables. Our contributions include: a) we carefully devise a specific model architecture and losses to explicitly disentangle text template and semantic content information, in the latent spaces, and b)we utilize both small parallel data and large raw text without aligned tables to enrich the template learning. Experiments on datasets from a variety of different domains show that VTM is able to generate more diversely while keeping a good fluency and quality. | ['Lei LI', 'Rong Ye', 'Hao Zhou', 'Wenxian Shi', 'Zhongyu Wei'] | 2020-02-04 | null | https://openreview.net/forum?id=HkejNgBtPB | https://openreview.net/pdf?id=HkejNgBtPB | iclr-2020-1 | ['table-to-text-generation'] | ['natural-language-processing'] | [ 2.86467642e-01 5.37889719e-01 -3.34125131e-01 -5.63505948e-01
-1.29220092e+00 -7.73714960e-01 8.29119027e-01 -1.01810686e-01
1.53769013e-02 1.15458512e+00 6.76011920e-01 -1.63080599e-02
6.92156479e-02 -9.52204287e-01 -1.06824362e+00 -3.86766613e-01
3.82938802e-01 1.02020991e+00 -1.53665826e-01 -3.40933681e-01
1.13985024e-01 -1.66180208e-01 -1.69199049e+00 7.83693731e-01
1.05520117e+00 7.26838648e-01 4.84324455e-01 1.41632542e-01
-7.62164891e-01 7.68615365e-01 -6.09087944e-01 -1.14275289e+00
2.43875414e-01 -6.67805374e-01 -8.32079351e-01 2.49930218e-01
2.69659519e-01 -2.34343901e-01 -2.27092490e-01 8.77514005e-01
3.48357379e-01 2.64149532e-02 8.50676477e-01 -1.20355761e+00
-1.18396699e+00 1.58566546e+00 -6.68573603e-02 -2.92275816e-01
2.70536095e-01 4.22591679e-02 1.34801614e+00 -9.53681648e-01
9.37903345e-01 1.30668437e+00 4.91196960e-01 1.06895483e+00
-1.57837284e+00 -6.05634809e-01 9.51769501e-02 -1.68013424e-02
-1.34654117e+00 -6.01821542e-01 8.46937060e-01 -4.04382020e-01
9.30201590e-01 2.59509891e-01 4.88654166e-01 1.94648516e+00
1.40018985e-01 9.10890579e-01 8.13899517e-01 -3.63888443e-01
1.53478801e-01 5.44249535e-01 -1.25051916e-01 5.88368177e-01
4.28216249e-01 -1.16346590e-01 -9.57128048e-01 1.19730942e-01
6.82534397e-01 -4.80471820e-01 -5.65772206e-02 -6.55854225e-01
-1.57804561e+00 1.05394316e+00 -7.27178752e-02 1.76031828e-01
-1.70171484e-01 1.23947822e-01 5.05698085e-01 2.16423973e-01
3.07277173e-01 7.81174839e-01 -5.63774407e-01 -2.83009142e-01
-7.18284070e-01 6.01548254e-01 7.92547584e-01 1.85075700e+00
6.25080526e-01 1.61682323e-01 -5.40771306e-01 8.95775437e-01
2.46768117e-01 5.18578470e-01 5.23496091e-01 -6.99744821e-01
1.11806524e+00 3.91339123e-01 2.51896590e-01 -6.16437137e-01
-3.87724787e-02 -2.68816859e-01 -9.68998134e-01 -4.25752401e-01
2.95561224e-01 -1.53226554e-01 -7.74349511e-01 2.16546822e+00
-7.94781670e-02 -4.89962131e-01 3.20928484e-01 4.28808302e-01
9.81358767e-01 8.45730424e-01 -7.58008957e-02 -1.47031382e-01
1.30174828e+00 -9.73226666e-01 -1.10814118e+00 -2.59607017e-01
6.31459534e-01 -5.02816617e-01 1.58004117e+00 2.53389388e-01
-1.49755526e+00 -5.97148478e-01 -1.05657756e+00 -6.10791862e-01
-5.83471239e-01 3.15286160e-01 6.92098975e-01 5.90678334e-01
-9.09726918e-01 6.28127694e-01 -6.19274437e-01 -6.90504313e-02
3.78800064e-01 3.15608263e-01 -3.92870903e-01 1.58398643e-01
-1.33301795e+00 8.76291275e-01 6.18536770e-01 -2.05993339e-01
-7.74633646e-01 -6.75879657e-01 -1.28844130e+00 2.12696254e-01
3.72617304e-01 -1.02551627e+00 1.20683992e+00 -6.22580111e-01
-1.54799211e+00 7.51571953e-01 -1.21042423e-01 -3.40948254e-01
5.29938221e-01 1.46912206e-02 -1.32663995e-01 -5.05633950e-01
2.52270609e-01 9.68103290e-01 6.35804713e-01 -1.16246784e+00
-2.22922862e-01 -1.78419665e-01 -2.18345955e-01 1.44880772e-01
-3.87737811e-01 -3.56457651e-01 -2.98188180e-01 -9.80358064e-01
-1.34276509e-01 -5.73697269e-01 -8.57623518e-02 -4.18395609e-01
-7.64836013e-01 -3.16824079e-01 9.23905224e-02 -5.19106328e-01
1.22581780e+00 -1.75158501e+00 5.40183723e-01 -1.73366427e-01
9.79941785e-02 -1.40600860e-01 -2.54426390e-01 6.63169920e-01
2.63525903e-01 4.06430781e-01 -3.83629352e-01 -6.08781397e-01
6.50560498e-01 4.00256276e-01 -5.66996574e-01 -3.92789662e-01
4.08870876e-01 1.07318580e+00 -7.96429813e-01 -6.20488584e-01
-1.80828944e-01 3.50861251e-01 -8.84916306e-01 2.88187414e-01
-7.92429268e-01 2.46056676e-01 -3.20419401e-01 4.08081353e-01
3.22166950e-01 -4.82067645e-01 2.62811512e-01 -1.33906186e-01
6.96456581e-02 7.49124169e-01 -1.22116971e+00 2.02304792e+00
-5.01279950e-01 3.85784268e-01 -4.29135054e-01 -5.95306218e-01
1.13037455e+00 4.44371015e-01 -1.76615566e-02 -5.47316551e-01
1.63391918e-01 3.75305325e-01 -3.60811651e-01 -5.77366292e-01
8.60932350e-01 -3.13568085e-01 -5.27932942e-01 3.18474740e-01
3.92215312e-01 -2.32065260e-01 5.88009894e-01 9.35140699e-02
8.45113873e-01 4.25766736e-01 1.77260935e-01 -2.68211871e-01
2.97288924e-01 -5.78184314e-02 6.72186196e-01 7.63119936e-01
2.91667908e-01 6.09872401e-01 7.16320336e-01 -3.68222564e-01
-1.51629734e+00 -1.35793364e+00 -1.60319120e-01 7.79854655e-01
-3.67815867e-02 -7.68238246e-01 -6.39261425e-01 -7.25737572e-01
-1.47019491e-01 1.12798536e+00 -6.22709692e-01 -4.67285886e-02
-4.51121926e-01 -6.78774893e-01 4.24354404e-01 7.58461714e-01
2.40341172e-01 -1.07094431e+00 -3.85227986e-02 4.20696557e-01
-6.60371661e-01 -1.03801930e+00 -6.07454360e-01 3.14722568e-01
-7.44764984e-01 -4.53821748e-01 -4.76860285e-01 -8.69132459e-01
6.14038408e-01 -3.59094292e-01 1.60920119e+00 -5.45946658e-01
4.48129773e-02 -1.60937682e-01 -1.57507360e-01 -4.34823930e-01
-8.13526034e-01 4.15432781e-01 -5.75603321e-02 -1.94322780e-01
4.49296057e-01 -6.44611001e-01 1.43463001e-01 -9.81983393e-02
-8.60884905e-01 8.04241061e-01 7.35634148e-01 1.12302887e+00
6.92032576e-01 -1.41157389e-01 7.78101742e-01 -1.23611510e+00
7.01547921e-01 -4.75926250e-01 -7.13656127e-01 4.82635975e-01
-7.06403255e-01 7.48580635e-01 7.15330958e-01 -3.47550362e-01
-1.24286520e+00 -8.83272290e-03 -7.86973909e-02 -1.11977994e-01
6.28139526e-02 2.96163857e-01 -8.36315572e-01 8.96151304e-01
6.28773332e-01 2.62992591e-01 -1.04220547e-01 -5.76913238e-01
7.02685714e-01 4.80434388e-01 3.07202011e-01 -9.19186294e-01
7.81243861e-01 1.11005358e-01 -1.67498335e-01 -3.51258874e-01
-8.22986901e-01 1.10072717e-01 -6.74967587e-01 1.86910301e-01
1.03864682e+00 -8.94959509e-01 -3.59949738e-01 -9.10856873e-02
-1.35059345e+00 -3.36328804e-01 -4.54011559e-01 3.01513404e-01
-8.95332515e-01 -6.28535077e-03 -6.71023309e-01 -4.50824767e-01
-3.35755885e-01 -1.22351241e+00 1.25081801e+00 -1.47264019e-01
-4.96856719e-01 -9.73976612e-01 8.44993815e-02 3.03570092e-01
3.94870639e-01 1.58160821e-01 1.34292328e+00 -6.35365486e-01
-9.74976361e-01 2.07122281e-01 1.96203915e-03 9.55958813e-02
2.65540957e-01 -7.02473596e-02 -8.16940606e-01 1.85814753e-01
-2.49287695e-01 -5.81244171e-01 7.07839429e-01 -3.90089154e-02
1.21541715e+00 -7.12352276e-01 -2.19919786e-01 6.66265786e-01
1.39323962e+00 1.54053405e-01 6.51170373e-01 1.84470356e-01
7.93060124e-01 8.60697269e-01 2.41983369e-01 5.48931956e-01
6.42576218e-01 6.00080192e-01 3.58253787e-03 4.13614303e-01
-1.20064929e-01 -8.29298377e-01 5.43765366e-01 1.29245269e+00
3.08613628e-01 -5.57804406e-01 -6.34948671e-01 6.20621979e-01
-1.70790374e+00 -1.09402776e+00 1.27617314e-01 1.89511478e+00
1.38109016e+00 1.68865621e-01 -9.16006714e-02 -8.79858993e-03
5.96010983e-01 2.30680648e-02 -2.58125275e-01 -3.38332444e-01
-3.53030235e-01 2.07038954e-01 1.48540676e-01 4.40844685e-01
-8.71852458e-01 9.85974252e-01 6.27774954e+00 8.84009182e-01
-5.77666104e-01 -3.38388868e-02 5.39121330e-01 -2.55442858e-01
-1.10402870e+00 -1.17079549e-01 -1.20859718e+00 5.69550872e-01
1.13759565e+00 -3.90623778e-01 4.28794026e-01 8.64617169e-01
-1.21325426e-01 6.47510886e-01 -1.68953979e+00 1.10497689e+00
1.49715230e-01 -1.79156339e+00 8.06453526e-01 -2.81171035e-02
7.50846922e-01 -4.71063286e-01 1.50371149e-01 6.43947780e-01
4.78939176e-01 -1.06331301e+00 1.09732997e+00 6.34289861e-01
8.87163579e-01 -6.35445297e-01 4.87580955e-01 2.46361822e-01
-9.04596806e-01 1.71826988e-01 -5.30674279e-01 1.62663966e-01
1.69852272e-01 3.07460546e-01 -6.91477478e-01 4.80039835e-01
3.35594743e-01 6.87711179e-01 -5.32806575e-01 4.01577085e-01
-2.07682982e-01 1.73426703e-01 1.66226760e-01 -3.40762734e-01
1.40976846e-01 -2.97264665e-01 3.93199265e-01 1.06260645e+00
4.91263330e-01 -1.58645809e-01 -7.01005608e-02 1.52553260e+00
-4.24904346e-01 1.42431170e-01 -9.04133499e-01 -4.36510146e-01
4.94359732e-01 9.42335010e-01 -4.53511178e-01 -4.36584383e-01
-4.53804493e-01 8.98476958e-01 4.38694894e-01 1.64226070e-01
-8.91596019e-01 -3.33213955e-01 4.86749142e-01 1.80891529e-02
3.39201480e-01 -1.34619251e-01 -6.51523709e-01 -1.52623034e+00
3.24592233e-01 -1.05548286e+00 2.37631649e-01 -8.32357824e-01
-1.56798232e+00 8.91433120e-01 1.62830696e-01 -1.26193988e+00
-6.23336613e-01 -5.16650796e-01 6.66142404e-02 8.68743837e-01
-1.08049023e+00 -1.15499341e+00 2.72334344e-03 3.55239511e-01
8.75100851e-01 -3.34460407e-01 9.37023997e-01 3.45003545e-01
-5.95570564e-01 8.98482025e-01 1.21414512e-01 8.55376124e-02
6.56793058e-01 -1.55342829e+00 7.13711083e-01 6.44811153e-01
3.87206197e-01 7.20095575e-01 7.05361724e-01 -7.11995065e-01
-1.36699355e+00 -1.20634675e+00 1.46818912e+00 -9.73146677e-01
4.28790092e-01 -1.09926367e+00 -8.09045374e-01 9.01889622e-01
4.05027300e-01 -5.66204548e-01 8.67513359e-01 1.36558507e-02
-4.92484123e-01 -4.65262979e-02 -9.70139146e-01 9.69058633e-01
1.24876976e+00 -4.73035872e-01 -7.54740834e-01 3.62039119e-01
1.22232783e+00 -4.29316908e-01 -8.41974676e-01 2.37308741e-01
3.97763640e-01 -8.32328856e-01 7.11603880e-01 -6.67189300e-01
9.64796066e-01 1.40198581e-02 -4.51293170e-01 -1.25658095e+00
-2.56913275e-01 -8.34878027e-01 -2.32718542e-01 1.75209689e+00
9.46250379e-01 -1.83741495e-01 8.39862227e-01 8.13300312e-01
-3.04620951e-01 -7.99503982e-01 -5.96977472e-01 -9.68672097e-01
3.63378555e-01 -3.78128529e-01 1.21373868e+00 8.05724084e-01
8.13807473e-02 8.04102957e-01 -5.49116969e-01 -2.55763113e-01
4.86342311e-01 8.73048156e-02 6.48591936e-01 -1.04009128e+00
-3.89601201e-01 -4.58787113e-01 5.52093908e-02 -1.02330267e+00
2.87731588e-01 -1.25058043e+00 8.47485662e-02 -1.73330629e+00
3.60611588e-01 -4.04197991e-01 1.85243085e-01 3.12638819e-01
-1.58087779e-02 -1.60339892e-01 8.29834789e-02 2.98367050e-02
-3.96387190e-01 1.05047846e+00 1.35238647e+00 -3.10091496e-01
-7.37095326e-02 -1.85823843e-01 -9.42054689e-01 3.90388608e-01
6.64007485e-01 -5.08262634e-01 -8.16093504e-01 -8.06540728e-01
5.12250900e-01 2.90624112e-01 3.46558262e-03 -7.68710136e-01
-1.44182444e-01 -2.41545305e-01 3.82528335e-01 -7.31836319e-01
4.75015432e-01 -4.05144215e-01 2.19815746e-01 -5.61053716e-02
-9.42726672e-01 4.06444281e-01 1.09772228e-01 4.99723941e-01
-1.56358466e-01 -2.25733444e-01 4.37209845e-01 -3.69208604e-01
-4.24513727e-01 2.92847544e-01 -2.60497212e-01 4.34011042e-01
6.35724783e-01 4.52386625e-02 -2.43785128e-01 -2.41998091e-01
-5.61879158e-01 7.03164861e-02 4.32802141e-01 7.12310970e-01
4.66307521e-01 -1.73564637e+00 -7.57707715e-01 3.31375241e-01
3.85144234e-01 -4.92975488e-02 7.03887269e-02 8.85518417e-02
-2.71344334e-01 6.68883801e-01 -2.77697504e-01 -3.85315627e-01
-7.13690341e-01 8.52051556e-01 -7.80944154e-03 -4.69060749e-01
-6.11243367e-01 8.48048031e-01 2.26479307e-01 -5.89350045e-01
3.93228710e-01 -5.38366735e-01 -3.62876505e-02 1.10862002e-01
3.95410240e-01 2.69911941e-02 9.08164657e-04 -3.13371688e-01
1.03713989e-01 1.19751140e-01 -2.28638113e-01 -5.21504700e-01
1.25046718e+00 -2.52924025e-01 -6.08299710e-02 7.04290926e-01
1.10457659e+00 1.94758568e-02 -1.17324448e+00 -2.56845623e-01
2.55179405e-01 -3.33959162e-01 -4.97029305e-01 -6.98203981e-01
-6.66059077e-01 9.09979641e-01 -2.96523012e-02 2.22739857e-02
7.54311442e-01 2.90672462e-02 1.04351032e+00 5.74119747e-01
5.72870374e-01 -1.00604677e+00 3.53945017e-01 5.53106904e-01
1.06855822e+00 -1.15106761e+00 -5.37618399e-01 -3.69939774e-01
-9.86006558e-01 9.67510700e-01 7.40502238e-01 2.80539364e-01
2.06280679e-01 2.93910354e-01 -1.42288864e-01 -6.76879883e-02
-1.26283836e+00 -1.03567623e-01 2.99868882e-01 6.23400629e-01
6.60375059e-01 -1.01165730e-03 -2.21778944e-01 1.20043790e+00
-7.87319303e-01 -3.10367018e-01 7.89558649e-01 6.52225018e-01
-7.60295689e-02 -1.62555146e+00 -1.03578173e-01 4.58734542e-01
-3.03704113e-01 -3.29358160e-01 -3.91387641e-01 8.20904791e-01
3.43107551e-01 7.25589037e-01 1.12634078e-01 -3.37499112e-01
1.68527886e-01 4.67377245e-01 7.28801847e-01 -8.44664812e-01
-3.94894183e-01 -1.60521299e-01 2.94826597e-01 -3.35569292e-01
-3.44146751e-02 -6.51319087e-01 -9.40038562e-01 -2.14632943e-01
-5.04324064e-02 3.48083526e-01 4.92743880e-01 6.23346746e-01
4.30767149e-01 6.25129998e-01 2.92051822e-01 -4.76568192e-01
-8.51414084e-01 -8.62618089e-01 -5.82243621e-01 5.09910941e-01
1.34167224e-01 -7.92557716e-01 -1.52463794e-01 3.61529261e-01] | [11.570198059082031, 8.847511291503906] |
f9aab7f9-62eb-40ee-a6e9-c9a0f49dbbf0 | understanding-person-identification-via-gait | 2203.04179 | null | https://arxiv.org/abs/2203.04179v4 | https://arxiv.org/pdf/2203.04179v4.pdf | Understanding Person Identification through Gait | Gait recognition is the process of identifying humans from their bipedal locomotion such as walking or running. As such, gait data is privacy sensitive information and should be anonymized where possible. With the rise of higher quality gait recording techniques, such as depth cameras or motion capture suits, an increasing amount of detailed gait data is captured and processed. The introduction and rise of the Metaverse is an example of a potentially popular application scenario in which the gait of users is transferred onto digital avatars. As a first step towards developing effective anonymization techniques for high-quality gait data, we study different aspects of movement data to quantify their contribution to gait recognition. We first extract categories of features from the literature on human gait perception and then design experiments for each category to assess how much the information they contain contributes to recognition success. We evaluated the utility of gait perturbation by means of naturalness ratings in a user study. Our results show that gait anonymization will be challenging, as the data is highly redundant and inter-dependent. | ['Admantini Hatzipanayioti', 'Thorsten Strufe', 'Shu-Chen Li', 'Evelyn Muschter', 'Simon Hanisch'] | 2022-03-08 | null | null | null | null | ['person-identification'] | ['computer-vision'] | [ 1.73294917e-01 -2.15337258e-02 -7.15605840e-02 -4.28596348e-01
-6.12444818e-01 -6.28195226e-01 3.45241040e-01 1.40230298e-01
-6.46084547e-01 6.84787929e-01 7.81431913e-01 2.51906991e-01
1.74723089e-01 -7.41510808e-01 -4.21616942e-01 -3.58479708e-01
-2.81397581e-01 2.88955867e-01 -4.58080731e-02 -2.60145813e-01
9.28878635e-02 4.50092345e-01 -1.71069157e+00 4.77720350e-02
3.98134172e-01 7.20727026e-01 -5.96621037e-01 5.54922640e-01
4.84716177e-01 1.85989931e-01 -5.55619419e-01 -7.21692383e-01
6.10153556e-01 -4.61736143e-01 -7.78663516e-01 4.86080982e-02
3.07050109e-01 -4.77330834e-01 -5.30163586e-01 7.26931274e-01
8.14636946e-01 2.36509651e-01 4.44245577e-01 -1.75686276e+00
-2.86242813e-01 9.13895015e-03 -2.44677961e-01 1.02092594e-01
1.00353467e+00 4.15384620e-01 8.41823697e-01 -3.24487180e-01
9.23048675e-01 8.39397490e-01 8.53872240e-01 5.27164340e-01
-1.39348185e+00 -5.40796936e-01 -5.45081913e-01 1.32597134e-01
-1.32811546e+00 -7.85483241e-01 5.72857916e-01 -5.49077630e-01
6.71607852e-01 4.21559364e-01 9.50237155e-01 1.53425109e+00
-1.63567699e-02 6.17376745e-01 7.18631327e-01 -7.34258890e-02
4.21224952e-01 -1.79370388e-01 7.90524855e-02 4.16377962e-01
7.78112590e-01 -4.98593226e-02 -7.93442309e-01 -4.21881199e-01
5.60443044e-01 -2.84688413e-01 -2.44733706e-01 -8.38788390e-01
-1.17769337e+00 5.53185523e-01 -4.58061062e-02 -4.38099913e-02
-2.91562974e-01 1.41060725e-01 8.63709331e-01 3.62463981e-01
-2.83397175e-02 4.64972943e-01 1.69372767e-01 -9.23847198e-01
-8.50548804e-01 5.79253614e-01 8.61570001e-01 9.17460263e-01
4.70184743e-01 -1.57012165e-01 1.81253254e-01 4.81305331e-01
-3.04039251e-02 2.76139081e-01 7.53862619e-01 -1.17103744e+00
6.75791502e-01 5.75885415e-01 3.33865702e-01 -1.52744722e+00
-2.78237373e-01 3.45463842e-01 -6.28568470e-01 1.41143024e-01
6.30231082e-01 9.00419895e-03 -5.80946863e-01 1.80056751e+00
1.98731467e-01 -2.42550030e-01 -1.47610933e-01 9.11046982e-01
2.46615887e-01 6.24100454e-02 2.44346544e-01 8.09338614e-02
1.34796882e+00 -4.97609042e-02 -5.21213472e-01 -1.38273671e-01
9.46739376e-01 -5.31149924e-01 1.25165057e+00 1.83399528e-01
-7.72861719e-01 -1.24792539e-01 -1.10152078e+00 -1.05187275e-01
-4.65996861e-01 -2.10056067e-01 3.98868948e-01 1.34778214e+00
-7.74164319e-01 9.51602221e-01 -1.03972316e+00 -1.00265741e+00
6.35846257e-01 5.13448715e-01 -1.08523822e+00 -2.00956479e-01
-1.19758832e+00 7.93791115e-01 3.05792987e-01 -3.22925836e-01
-1.38118401e-01 -4.60117906e-01 -9.84534562e-01 -2.39764333e-01
-3.79238166e-02 -8.38985562e-01 7.81040132e-01 -5.12029469e-01
-1.05898440e+00 1.14632308e+00 8.54882076e-02 -4.99407560e-01
8.86101246e-01 -3.28719497e-01 -4.43249613e-01 9.57985595e-02
3.05288255e-01 3.53844672e-01 8.57965291e-01 -8.29203904e-01
-5.13736129e-01 -6.27628505e-01 -2.89142370e-01 3.32234651e-01
-4.40576255e-01 -5.25777228e-02 -7.49914348e-02 -5.82092881e-01
-1.53429046e-01 -1.05543363e+00 3.77964154e-02 1.20710336e-01
-2.68569499e-01 5.39261580e-01 1.04168582e+00 -1.06746793e+00
1.24825001e+00 -2.17117739e+00 1.01537183e-01 4.83254164e-01
4.04300869e-01 1.35591492e-01 2.29225412e-01 5.84241927e-01
3.94746587e-02 3.88999760e-01 -3.10612142e-01 -3.97731334e-01
2.24693120e-01 3.69191855e-01 -9.82668102e-02 6.97584867e-01
-1.34639800e-01 7.97509849e-01 -8.41357410e-01 -4.34914201e-01
2.90889680e-01 2.55967766e-01 -6.48615003e-01 -4.58979756e-02
4.98641938e-01 4.74411160e-01 -2.09651649e-01 7.27734566e-01
3.72146904e-01 2.28437543e-01 3.08478214e-02 -2.67641634e-01
1.76135078e-01 4.47909497e-02 -1.06259918e+00 1.58418608e+00
8.98292568e-03 4.82727051e-01 -8.16842690e-02 -7.70583749e-01
6.93990529e-01 3.44838023e-01 1.03185678e+00 -6.25063121e-01
2.06101939e-01 1.07214496e-01 -1.98645428e-01 -6.49568617e-01
8.86137128e-01 6.27270564e-02 -5.34111679e-01 6.41911089e-01
-1.90767810e-01 2.13832006e-01 2.05373466e-01 -3.65642831e-02
1.43250203e+00 2.06298232e-01 6.02582157e-01 6.47326037e-02
-5.59226098e-03 1.12313293e-01 5.45175731e-01 4.08829153e-01
-9.39794719e-01 1.02070630e+00 2.62773842e-01 -4.12445426e-01
-1.35776639e+00 -1.08478236e+00 2.30796605e-01 6.41967177e-01
-4.26576249e-02 -7.90318072e-01 -8.26667666e-01 -1.72105387e-01
1.75980866e-01 4.66267705e-01 -6.93484783e-01 -6.22813642e-01
-5.59044659e-01 -7.94496298e-01 1.16815674e+00 6.36933863e-01
5.78472435e-01 -8.69647861e-01 -1.31023848e+00 1.86461836e-01
-6.94633603e-01 -1.12733865e+00 -6.55627131e-01 -2.45728254e-01
-6.27820075e-01 -9.52758670e-01 -7.12667465e-01 -1.79876029e-01
2.81076193e-01 -7.72833899e-02 8.40493858e-01 -2.12754592e-01
-5.25524616e-01 7.88716197e-01 -5.16175270e-01 1.46471290e-02
-1.89218640e-01 9.55657437e-02 5.58780193e-01 9.77535993e-02
7.45542943e-01 -9.59778249e-01 -6.02926910e-01 3.85390043e-01
-5.83586991e-01 -3.27028811e-01 1.26871333e-01 6.21541142e-01
2.68188357e-01 -5.07889055e-02 7.46084303e-02 -5.82650661e-01
1.02034736e+00 -5.96305311e-01 1.66655064e-01 -1.23794831e-01
-3.92874300e-01 -1.60658777e-01 4.65082109e-01 -2.90812105e-01
-5.14521122e-01 3.42061728e-01 1.64064899e-01 -2.27333397e-01
-4.33918566e-01 2.95581669e-01 -4.52408403e-01 -9.00061354e-02
9.05229926e-01 1.23754270e-01 2.63737738e-01 -2.94027239e-01
2.83556044e-01 6.81837142e-01 6.82404935e-01 -6.43602252e-01
7.89875150e-01 5.16143918e-01 5.37178442e-02 -1.30780208e+00
4.28821325e-01 -4.35283720e-01 -7.51687109e-01 -3.91175389e-01
8.82457197e-01 -5.75589776e-01 -8.12776566e-01 4.69410688e-01
-6.40260875e-01 1.56472102e-01 -6.50976062e-01 4.47008222e-01
-1.05964088e+00 7.40935385e-01 -4.14783984e-01 -7.21529305e-01
-6.90313727e-02 -9.42536116e-01 1.03200948e+00 -1.23323210e-01
-1.17534423e+00 -5.10037720e-01 2.48553917e-01 6.19836032e-01
1.42318383e-01 7.91635215e-01 7.03864932e-01 -5.54572344e-01
-7.27756768e-02 -8.27523649e-01 1.38168752e-01 -5.68724610e-03
4.64816839e-01 -1.44828573e-01 -7.92916000e-01 -3.03546578e-01
-1.33013964e-01 -3.19542527e-01 1.52043924e-01 6.18242957e-02
5.86624980e-01 -5.10553181e-01 -1.68250754e-01 4.62990314e-01
1.09522879e+00 -3.54790427e-02 1.14561701e+00 4.77457285e-01
9.45661545e-01 8.08334231e-01 5.18281162e-01 5.67185581e-01
3.05192083e-01 8.47306311e-01 1.47405848e-01 3.87324721e-01
2.17857584e-01 -5.53805113e-01 3.68040144e-01 6.78901792e-01
-2.91879088e-01 -1.33056045e-01 -1.03350937e+00 5.47803164e-01
-1.87014520e+00 -1.26023412e+00 2.12108672e-01 2.57751465e+00
3.22248012e-01 -2.11110666e-01 9.15234923e-01 5.93289554e-01
6.19846344e-01 2.51080953e-02 -4.53279734e-01 -4.41866517e-01
-1.92961730e-02 2.44578309e-02 7.83030570e-01 -3.91362607e-02
-1.10936141e+00 5.32379568e-01 6.55556488e+00 1.16183959e-01
-7.76307464e-01 -4.92013603e-01 3.28897178e-01 -1.29342958e-01
1.72960073e-01 -2.12642804e-01 -5.00199907e-02 6.69446886e-01
1.02440655e+00 -5.81656575e-01 2.14381099e-01 7.81512439e-01
4.39831287e-01 -6.72631860e-02 -1.28370500e+00 1.18765306e+00
-1.72046497e-01 -8.54206622e-01 1.98396072e-01 5.82681835e-01
-5.90818841e-03 -4.23428744e-01 4.21821401e-02 -1.62775904e-01
3.45040828e-01 -1.05389881e+00 5.38492024e-01 3.30639571e-01
8.73758316e-01 -7.25227416e-01 5.60636818e-01 4.86618169e-02
-1.30843556e+00 1.02130547e-01 -1.51097894e-01 -1.32153600e-01
3.76559854e-01 3.95937003e-02 -6.37178600e-01 6.14664912e-01
8.84640455e-01 8.27050090e-01 -5.60962975e-01 1.07026350e+00
4.41491306e-01 5.69176495e-01 -7.26538599e-01 1.88484401e-01
-3.29095215e-01 -3.08172315e-01 7.20325708e-01 8.25592935e-01
5.05783141e-01 1.30648881e-01 -8.76215845e-02 4.76265281e-01
-1.50155583e-02 2.45830685e-01 -1.16009712e+00 -4.86294657e-01
4.69423711e-01 6.45319223e-01 -7.00907111e-01 9.28753391e-02
-2.97682136e-01 1.35672235e+00 2.51168236e-02 6.97801560e-02
-4.60334808e-01 -3.13341975e-01 1.29513061e+00 4.67189223e-01
3.61560471e-02 -4.54884112e-01 -4.47676212e-01 -1.24758518e+00
4.06969696e-01 -9.11262274e-01 5.24002612e-01 -6.22190297e-01
-1.26360357e+00 4.48851939e-03 -9.75114701e-04 -1.68529594e+00
-5.84773183e-01 -1.35782138e-01 -2.87361294e-01 6.07534051e-01
-2.30282679e-01 -9.63567913e-01 -3.91336054e-01 6.61692739e-01
-1.03556655e-01 -7.82353878e-02 1.16480601e+00 5.95749915e-01
-3.45922768e-01 8.76862526e-01 1.19047491e-02 3.81847948e-01
7.30375707e-01 -8.13517988e-01 8.60263407e-01 1.00065947e+00
-1.92648955e-02 8.02717149e-01 8.98012102e-01 -9.34263051e-01
-1.38404763e+00 -8.83501112e-01 6.87334538e-01 -8.55304658e-01
4.80643600e-01 -3.82031292e-01 -7.45308697e-01 7.80198514e-01
-4.62177843e-01 -2.14016423e-01 1.04600930e+00 -1.72915146e-01
-2.73508251e-01 1.43942922e-01 -1.68766642e+00 9.71975446e-01
1.41583014e+00 -6.60814345e-01 -6.61817610e-01 -2.84199715e-01
2.62133002e-01 -2.04641625e-01 -1.17982304e+00 9.97904912e-02
9.56186354e-01 -1.02188921e+00 1.03151667e+00 -7.48171628e-01
2.86844730e-01 -2.70611167e-01 -4.91173565e-01 -1.14993334e+00
-2.24009931e-01 -7.63012528e-01 3.36957760e-02 1.22443247e+00
-2.13123616e-02 -4.58036780e-01 1.42414570e+00 1.41680229e+00
3.97311449e-01 -5.67257740e-02 -1.01332688e+00 -7.96558022e-01
-1.39858484e-01 -3.17516237e-01 6.65457070e-01 1.21529472e+00
3.98580581e-01 -9.98849347e-02 -8.05150568e-01 -1.44667774e-01
9.68912601e-01 -4.71361548e-01 1.12884343e+00 -1.20581222e+00
4.44643572e-02 -1.28447205e-01 -1.60169899e+00 -2.70465195e-01
-1.87562659e-01 -6.74006224e-01 -2.69803941e-01 -1.15027416e+00
-2.74758279e-01 4.40024026e-02 2.51430869e-01 4.30855572e-01
1.51204929e-01 3.81075293e-01 1.85126647e-01 3.87150139e-01
-1.76321715e-01 5.43987989e-01 5.17275333e-01 1.78263128e-01
-1.94015577e-01 4.26063351e-02 -4.84572470e-01 4.59621042e-01
8.74222040e-01 -3.52834791e-01 -4.59306419e-01 4.96616811e-02
-1.32942393e-01 -1.47964224e-01 4.89502609e-01 -1.36916006e+00
5.47306240e-02 6.98582008e-02 3.53265107e-01 -1.48789749e-01
6.11233294e-01 -1.08011734e+00 5.62215328e-01 4.79347259e-01
-1.53528109e-01 2.40858331e-01 7.05740526e-02 6.69418275e-01
-9.72702131e-02 3.51529598e-01 4.13915038e-01 -2.47513130e-01
-8.90282035e-01 2.41572633e-01 -7.39386737e-01 3.64192516e-01
1.09403324e+00 -1.04147887e+00 4.57460508e-02 -6.93204701e-01
-8.62273633e-01 -9.20507964e-03 1.14924598e+00 5.01230121e-01
4.75975245e-01 -1.52480626e+00 -3.66591662e-01 2.32322752e-01
4.88850474e-01 -6.01509452e-01 1.19169746e-02 4.63491917e-01
-5.99007070e-01 2.92703062e-02 -1.03357923e+00 -3.28007460e-01
-1.37386858e+00 2.58432031e-01 7.86341503e-02 1.87023953e-01
-9.14309323e-01 1.83931306e-01 -4.70287919e-01 -2.97807842e-01
3.37354429e-02 -6.93373829e-02 1.09209217e-01 2.11596638e-01
3.26574862e-01 8.24537992e-01 1.43204644e-01 -1.18640077e+00
-5.64962327e-01 3.20848018e-01 2.86598861e-01 -5.08164585e-01
1.14116251e+00 -4.10326749e-01 2.59403974e-01 3.46597791e-01
1.40405524e+00 -2.67767161e-01 -1.05115676e+00 2.91262448e-01
1.63870513e-01 -8.81884277e-01 -7.21638083e-01 -1.34244055e-01
-7.19961762e-01 5.19922554e-01 6.98193967e-01 4.41258727e-03
9.09852087e-01 -6.31739080e-01 1.10609376e+00 2.41036609e-01
8.70937943e-01 -1.24364996e+00 -2.96063155e-01 2.66776700e-02
4.47738856e-01 -1.04615450e+00 8.63553584e-02 -3.17916512e-01
-7.61249363e-01 6.68167651e-01 2.99193650e-01 -5.43608442e-02
4.27065849e-01 1.34978503e-01 -1.17472202e-01 5.26844077e-02
-1.06880330e-01 1.67043164e-01 -1.22159578e-01 1.26453984e+00
1.74125224e-01 2.30711490e-01 -4.38151568e-01 5.45319915e-01
-9.51676786e-01 2.71427006e-01 7.53067434e-01 1.42627776e+00
-4.72629666e-02 -1.11975026e+00 -4.34429824e-01 7.73583651e-01
-1.76708236e-01 3.65961015e-01 -7.07915246e-01 8.57644916e-01
-3.25594544e-01 6.95236087e-01 -2.08657786e-01 -9.67960179e-01
6.95008278e-01 4.03763413e-01 3.53131056e-01 -3.00633997e-01
-4.53513622e-01 -8.59466314e-01 5.02689838e-01 -1.02803123e+00
-2.35509515e-01 -1.10415423e+00 -9.57487047e-01 -9.45493817e-01
4.74181026e-01 -6.30015358e-02 4.23441201e-01 6.41178131e-01
5.21194160e-01 -1.99898690e-01 1.41066760e-01 -9.15459514e-01
-3.01403642e-01 -2.36313388e-01 -7.20443964e-01 1.19909573e+00
1.24135651e-01 -6.95704520e-01 -2.64016718e-01 3.90873522e-01] | [14.263510704040527, 1.395126223564148] |
2fdfc7e0-2da9-4d61-8be4-205909e3ca5c | detecting-clusters-of-anomalies-on-low | 1511.01047 | null | http://arxiv.org/abs/1511.01047v1 | http://arxiv.org/pdf/1511.01047v1.pdf | Detecting Clusters of Anomalies on Low-Dimensional Feature Subsets with Application to Network Traffic Flow Data | In a variety of applications, one desires to detect groups of anomalous data
samples, with a group potentially manifesting its atypicality (relative to a
reference model) on a low-dimensional subset of the full measured set of
features. Samples may only be weakly atypical individually, whereas they may be
strongly atypical when considered jointly. What makes this group anomaly
detection problem quite challenging is that it is a priori unknown which subset
of features jointly manifests a particular group of anomalies. Moreover, it is
unknown how many anomalous groups are present in a given data batch. In this
work, we develop a group anomaly detection (GAD) scheme to identify the subset
of samples and subset of features that jointly specify an anomalous cluster. We
apply our approach to network intrusion detection to detect BotNet and
peer-to-peer flow clusters. Unlike previous studies, our approach captures and
exploits statistical dependencies that may exist between the measured features.
Experiments on real world network traffic data demonstrate the advantage of our
proposed system, and highlight the importance of exploiting feature dependency
structure, compared to the feature (or test) independence assumption made in
previous studies. | ['George Kesidis', 'Zhicong Qiu', 'David J. Miller'] | 2015-06-10 | null | null | null | null | ['group-anomaly-detection'] | ['methodology'] | [ 2.55792260e-01 -2.17773601e-01 1.71499670e-01 -4.89778459e-01
6.17781319e-02 -5.33536792e-01 6.77039087e-01 3.84200454e-01
7.72311240e-02 5.16748905e-01 -4.03029859e-01 -3.73154551e-01
-6.09821141e-01 -6.72567546e-01 -2.16209307e-01 -9.57561970e-01
-8.30381274e-01 7.24656463e-01 5.10782957e-01 6.88198134e-02
4.63241309e-01 9.15760994e-01 -1.73936379e+00 2.36648813e-01
6.82814360e-01 1.04097414e+00 -3.71598452e-01 8.52569282e-01
-1.70764953e-01 7.81267464e-01 -1.03720343e+00 9.83070210e-02
4.03203994e-01 -3.88532937e-01 -7.19325483e-01 6.10135734e-01
5.22657633e-01 -4.00218666e-01 1.06150713e-02 1.08243084e+00
-1.14603110e-01 -8.24682713e-02 9.01086926e-01 -2.17028260e+00
-6.32847920e-02 2.72833586e-01 -6.65210724e-01 9.28532124e-01
2.16209203e-01 2.92121440e-01 1.14116907e+00 -5.22623181e-01
3.12858790e-01 9.67205524e-01 2.75114834e-01 2.78303564e-01
-1.48883998e+00 -6.67726278e-01 3.94037902e-01 1.55981019e-01
-1.20006025e+00 -3.76118302e-01 9.06567991e-01 -6.66311681e-01
7.42650986e-01 3.04837167e-01 3.71134996e-01 1.02908552e+00
6.04346059e-02 4.26157206e-01 7.96374559e-01 1.18994266e-02
4.22841907e-01 1.46768779e-01 5.21696270e-01 4.16914970e-01
6.59871399e-01 -4.54512499e-02 -1.50763229e-01 -9.78623688e-01
3.92828137e-01 3.29900265e-01 1.44405559e-01 -4.07732368e-01
-8.06148291e-01 8.17383289e-01 -2.81181723e-01 5.22449613e-01
-6.56274140e-01 -2.83511341e-01 4.23657656e-01 8.57556760e-01
4.14440960e-01 2.53744066e-01 -5.16279459e-01 -1.13699891e-01
-6.70495987e-01 1.52462855e-01 1.07024467e+00 7.33540714e-01
9.16928530e-01 1.91924155e-01 1.40576243e-01 3.95252496e-01
2.10516989e-01 2.94547558e-01 3.80718023e-01 -5.14341533e-01
2.41695970e-01 8.88430774e-01 -1.00304475e-02 -1.18303275e+00
-3.94550055e-01 -2.30390519e-01 -8.46836925e-01 2.24385083e-01
7.60803103e-01 -8.65775645e-02 -3.09790045e-01 1.70281982e+00
4.80862439e-01 7.27879524e-01 -7.74448514e-02 3.73501360e-01
1.53825566e-01 2.29619220e-01 -7.21361786e-02 -5.13502657e-01
8.61087561e-01 -1.44744143e-01 -4.54134315e-01 1.33227512e-01
7.18294501e-01 -5.14713228e-01 7.11144209e-01 4.31769460e-01
-5.60236156e-01 -3.33031684e-01 -6.91964269e-01 1.07806575e+00
-3.61660749e-01 -5.85673392e-01 4.97418761e-01 9.06089842e-01
-8.76268983e-01 6.72389507e-01 -6.71562850e-01 -8.43825698e-01
2.56279886e-01 5.53085864e-01 -3.55104119e-01 2.00202674e-01
-7.33197808e-01 1.94498420e-01 4.50486749e-01 -1.32673338e-01
-6.20815694e-01 -4.89845186e-01 -3.07666719e-01 -2.48769280e-02
5.35349548e-01 -5.26245013e-02 7.65254319e-01 -9.67051268e-01
-8.69505525e-01 4.54343170e-01 -2.15602756e-01 -4.56201881e-01
4.20400232e-01 1.57126084e-01 -1.19691777e+00 2.22776279e-01
6.04472607e-02 -2.72455156e-01 1.06979740e+00 -1.24138033e+00
-9.26443219e-01 -5.50315022e-01 -1.82173252e-01 -4.12718564e-01
-5.02847672e-01 2.51850903e-01 2.54975319e-01 -2.34573409e-01
3.14037979e-01 -7.67204344e-01 -2.76345670e-01 -3.62541586e-01
-7.27833569e-01 -3.62238824e-01 1.67976296e+00 -1.19352169e-01
1.21922171e+00 -2.21027279e+00 -4.56873804e-01 1.06077576e+00
5.04012883e-01 3.01388979e-01 -7.60722309e-02 6.98127925e-01
-2.22628877e-01 3.97012472e-01 -4.10745263e-01 -1.35560796e-01
-1.77661508e-01 6.10084653e-01 -5.82832932e-01 8.78423393e-01
6.28644168e-01 3.40926461e-02 -7.42451012e-01 -3.82550240e-01
5.42568676e-02 -1.68346867e-01 -7.04394162e-01 4.13127631e-01
-3.47409770e-03 8.63001227e-01 -8.50319505e-01 7.79901564e-01
7.83449292e-01 -2.09864944e-01 5.31663410e-02 3.09145927e-01
-5.18582277e-02 -5.72151765e-02 -1.57477677e+00 4.77896214e-01
2.41895840e-01 2.41036013e-01 -8.45305771e-02 -1.46317995e+00
1.17068911e+00 4.38226193e-01 9.86211419e-01 -8.71957839e-02
-7.06331059e-02 2.71024764e-01 5.52817166e-01 -6.33516252e-01
3.17579180e-01 2.74044573e-01 1.13020942e-01 9.49634612e-01
3.50758284e-02 5.25780797e-01 3.96090955e-01 2.03122705e-01
1.79154301e+00 -7.84825325e-01 5.80391943e-01 -3.75649899e-01
8.50875318e-01 -3.57954115e-01 9.11837161e-01 1.12047637e+00
-5.67762971e-01 2.44618356e-01 1.32683778e+00 -5.00985026e-01
-9.50190544e-01 -1.15018678e+00 -2.60660738e-01 7.84497499e-01
-1.31754339e-01 -2.30270252e-01 -2.07174122e-01 -1.12214029e+00
1.50345996e-01 3.71455014e-01 -5.70051849e-01 -8.84866342e-02
-6.20728374e-01 -9.49247539e-01 6.06453180e-01 -1.05286874e-01
1.91275194e-01 -1.15386987e+00 -4.48937058e-01 2.38185719e-01
2.96723425e-01 -1.18301606e+00 1.93333309e-02 1.66824728e-01
-9.14777815e-01 -1.66529500e+00 5.02418637e-01 -2.84335226e-01
8.86173785e-01 1.05645098e-01 1.07262790e+00 4.32921618e-01
-3.92886937e-01 5.39652944e-01 -5.66510916e-01 -2.79080182e-01
-6.87992394e-01 -1.35786131e-01 5.58379829e-01 7.54727483e-01
9.16822553e-01 -1.02345157e+00 -3.75490874e-01 5.31478822e-01
-1.07707226e+00 -9.97476816e-01 4.71880794e-01 4.67340201e-01
1.61148965e-01 6.63760364e-01 8.39605331e-01 -1.25796616e+00
7.15901494e-01 -1.26340771e+00 -3.99839312e-01 -4.39521344e-03
-6.47285342e-01 -1.89281568e-01 1.04656827e+00 -5.83023548e-01
-5.48651457e-01 -1.43555641e-01 3.82501304e-01 -4.05742586e-01
-8.98217559e-01 -2.81205922e-02 -2.95579135e-01 1.24169834e-01
5.26096582e-01 2.77261585e-01 2.57694066e-01 -3.97633612e-01
-2.48431921e-01 6.65641367e-01 1.77577198e-01 -7.00570703e-01
1.11006129e+00 5.87906063e-01 3.42802763e-01 -1.23709071e+00
-1.49006650e-01 -8.66845191e-01 -1.09201896e+00 -1.20394118e-01
2.23501921e-01 -3.05289954e-01 -7.10709155e-01 2.71751493e-01
-7.37587154e-01 2.24652529e-01 -3.47789943e-01 3.45593125e-01
-2.52719581e-01 7.64246345e-01 -2.12985083e-01 -1.01100183e+00
1.24053612e-01 -7.99448073e-01 5.58680058e-01 -7.27928281e-02
-4.32608426e-01 -1.07503629e+00 2.25828335e-01 -2.63089180e-01
3.30112636e-01 5.37688017e-01 8.55802655e-01 -1.85824597e+00
-3.60515445e-01 -4.91334885e-01 -6.44911602e-02 3.76894951e-01
6.05687499e-01 6.21341586e-01 -7.93595314e-01 -5.56380391e-01
1.49110734e-01 1.98471859e-01 3.81151408e-01 6.39737472e-02
1.18355381e+00 -5.19332051e-01 -1.78610846e-01 2.46257469e-01
1.11412728e+00 4.20792133e-01 1.40491113e-01 -4.71286960e-02
5.38640022e-01 8.52124870e-01 4.94448841e-01 9.08314943e-01
-1.84231222e-01 4.62122113e-01 5.50771952e-01 1.29412726e-01
6.11574173e-01 2.22156316e-01 4.33652192e-01 4.58370000e-01
6.61266297e-02 -2.00554684e-01 -8.92940223e-01 5.11131406e-01
-1.73746741e+00 -1.33030617e+00 -5.13304174e-01 2.31918097e+00
9.68195125e-02 9.86139700e-02 7.94485390e-01 5.56683540e-01
9.49253082e-01 8.91951844e-02 -7.14960635e-01 -5.55478334e-01
-1.40965164e-01 7.78984569e-04 2.04858795e-01 9.40698460e-02
-1.21884787e+00 4.35923636e-01 6.02990484e+00 3.08057815e-01
-9.36292887e-01 -3.15328956e-01 4.84085143e-01 1.83802947e-01
7.62560740e-02 1.72660172e-01 -6.37733757e-01 5.49544334e-01
1.09542048e+00 -2.74799168e-01 1.59114853e-01 8.24134171e-01
1.87658235e-01 -3.05157658e-02 -1.39872324e+00 4.82094198e-01
-7.22894296e-02 -4.57165331e-01 -5.27674425e-03 5.92526436e-01
5.54531336e-01 -1.00261293e-01 2.57779181e-01 1.21460356e-01
4.57603842e-01 -6.93073094e-01 -1.26624390e-01 3.51365477e-01
6.02427162e-02 -8.70017588e-01 6.63232446e-01 4.05825019e-01
-1.08340204e+00 -4.11750704e-01 -3.20953369e-01 -1.86764672e-01
-1.62079811e-01 8.73516381e-01 -1.23325431e+00 4.15995091e-01
6.44198298e-01 7.24378765e-01 -7.95034647e-01 8.78600657e-01
3.36361378e-01 1.00193477e+00 -5.50665617e-01 2.49022990e-01
2.35509038e-01 -2.66943276e-01 1.04496014e+00 1.01551986e+00
3.58813107e-01 -1.18869834e-01 5.43021739e-01 8.09074938e-01
3.99870872e-01 2.39852831e-01 -1.00589705e+00 -1.93034783e-01
6.33356571e-01 1.29614246e+00 -9.01102543e-01 -3.24384362e-01
-5.78739345e-01 6.67744458e-01 4.22431566e-02 3.25418323e-01
-3.89644980e-01 -2.76928157e-01 1.20809293e+00 2.53359497e-01
3.62816095e-01 -1.56629145e-01 6.03143759e-02 -1.25180125e+00
8.41713324e-02 -7.51299918e-01 6.84130549e-01 3.15241516e-01
-1.94841862e+00 6.41570151e-01 -6.39860183e-02 -1.69963038e+00
-8.18063080e-01 -3.28098506e-01 -1.23908579e+00 5.64670920e-01
-9.90184128e-01 -6.89775407e-01 -1.17226847e-01 9.95304883e-01
1.75270393e-01 -6.69021547e-01 8.29029858e-01 7.05085918e-02
-7.62077749e-01 4.06251848e-01 1.65153846e-01 2.14151427e-01
4.52111423e-01 -1.17735720e+00 2.64896661e-01 1.08149755e+00
1.45468965e-01 5.06740034e-01 9.24683154e-01 -6.91665053e-01
-7.78617680e-01 -1.11533344e+00 5.56746006e-01 -5.30026436e-01
1.09374511e+00 -2.88821191e-01 -1.26780009e+00 7.65918434e-01
-3.71020675e-01 5.96624076e-01 9.87140357e-01 1.93808749e-01
-3.68078381e-01 6.93996903e-04 -1.61684632e+00 3.23146522e-01
9.82481420e-01 -2.89802521e-01 -3.56287658e-01 1.73834264e-01
2.05209240e-01 5.51408827e-01 -8.19668770e-01 4.65987653e-01
1.81896165e-01 -1.52728283e+00 5.78226924e-01 -1.05992353e+00
-2.12149516e-01 -5.08482099e-01 -2.51336217e-01 -1.12302756e+00
-3.18018526e-01 -5.25698304e-01 -4.03472066e-01 1.51371181e+00
2.23804656e-02 -9.79798853e-01 7.40625501e-01 4.39620286e-01
2.89913535e-01 -2.56940752e-01 -1.12200868e+00 -1.08277810e+00
-1.15214035e-01 -2.95366913e-01 8.10799778e-01 1.20508480e+00
3.78326401e-02 -2.00150594e-01 -3.07269514e-01 7.18376815e-01
9.98372436e-01 7.46266916e-02 1.08581924e+00 -2.06036639e+00
-2.62240291e-01 -3.82238150e-01 -1.02100790e+00 -2.54816949e-01
2.79510468e-01 -5.20853162e-01 -3.97657037e-01 -5.35069108e-01
-8.09048712e-02 -6.77963614e-01 -5.47360241e-01 1.95601538e-01
1.06671445e-01 7.04330346e-03 -2.38488838e-01 5.09529650e-01
-6.29147530e-01 7.63968751e-02 4.66318548e-01 2.94496059e-01
-4.13751066e-01 4.99082923e-01 -4.22102362e-01 7.41177797e-01
1.03594482e+00 -5.96303642e-01 -3.61757278e-01 3.98977309e-01
-2.18303531e-01 -1.62947088e-01 3.84064913e-01 -1.11797309e+00
1.37537405e-01 -3.54345560e-01 1.61569774e-01 -5.79325318e-01
-4.85925041e-02 -1.19182181e+00 2.46407278e-02 5.27384222e-01
-5.58522716e-02 2.95407832e-01 -2.40472659e-01 8.50443602e-01
-2.25032508e-01 3.05793919e-02 6.84786975e-01 2.13745117e-01
-6.06706142e-01 6.81807280e-01 -7.45353281e-01 7.35851005e-02
1.41471338e+00 -2.87291229e-01 -1.67748898e-01 -3.77809048e-01
-6.98230743e-01 7.32274279e-02 3.78109932e-01 4.32312489e-01
4.62113440e-01 -1.09505796e+00 -7.30374813e-01 9.04580235e-01
4.23392653e-01 -2.84088194e-01 2.10765764e-01 1.01040590e+00
-9.11246613e-02 1.11857653e-01 -4.90605593e-01 -9.53650296e-01
-1.44060779e+00 5.32019973e-01 -3.28966789e-02 -2.66831756e-01
-7.01367795e-01 1.29970759e-01 -1.75273083e-02 -3.13426524e-01
-5.90681136e-02 2.10232735e-02 -4.41525787e-01 4.61337343e-02
6.34409785e-01 6.05746925e-01 -1.03030644e-01 -9.15323913e-01
-3.58395040e-01 2.30020121e-01 -4.84219551e-01 3.12714607e-01
1.29656231e+00 -1.13688566e-01 -4.37512100e-01 7.20712841e-01
1.14328945e+00 4.25776765e-02 -1.05100334e+00 -5.83627522e-01
5.56536257e-01 -7.00615704e-01 -6.40759468e-01 -1.40181649e-02
-1.10754240e+00 4.70682055e-01 4.24849361e-01 1.08130825e+00
1.07844913e+00 1.12221740e-01 2.75406629e-01 5.51165164e-01
2.14388654e-01 -7.65605509e-01 1.24787167e-01 3.68605822e-01
3.24656129e-01 -1.25633407e+00 -3.20045173e-01 -3.37924093e-01
-5.02354264e-01 1.30174100e+00 7.80977786e-01 -5.92862964e-01
9.62041557e-01 1.91441387e-01 -2.05417633e-01 -4.23801571e-01
-9.79078531e-01 -1.71505988e-01 -1.25090897e-01 8.16054523e-01
-2.28094738e-02 8.23972672e-02 -1.24706089e-01 1.70376837e-01
-1.05028994e-01 -5.64402103e-01 8.12713087e-01 9.56744373e-01
-6.23971224e-01 -1.13487399e+00 -5.23717403e-01 1.01025140e+00
-6.87117755e-01 3.72408122e-01 -5.11659980e-01 8.68003666e-01
2.92840917e-02 1.26551592e+00 6.95815027e-01 -5.30988395e-01
2.66562402e-01 2.98299074e-01 -2.24726126e-01 -5.87748706e-01
-4.22951519e-01 -2.05441996e-01 -6.09919205e-02 -7.59411156e-01
-2.06300870e-01 -1.04497635e+00 -9.37821507e-01 -5.80489635e-01
-2.40272898e-02 3.42757106e-01 3.21536422e-01 1.11924136e+00
2.73671269e-01 2.58094847e-01 1.29325807e+00 -2.42828205e-01
-5.49238682e-01 -9.38190162e-01 -1.10108173e+00 7.94219613e-01
6.81764901e-01 -6.79650545e-01 -1.09941137e+00 -2.27211326e-01] | [7.50612735748291, 2.667022705078125] |
62cd1a66-5dd6-449a-ad29-e941bfbadc70 | fast-multi-view-clustering-via-ensembles | 2203.11572 | null | https://arxiv.org/abs/2203.11572v4 | https://arxiv.org/pdf/2203.11572v4.pdf | Fast Multi-view Clustering via Ensembles: Towards Scalability, Superiority, and Simplicity | Despite significant progress, there remain three limitations to the previous multi-view clustering algorithms. First, they often suffer from high computational complexity, restricting their feasibility for large-scale datasets. Second, they typically fuse multi-view information via one-stage fusion, neglecting the possibilities in multi-stage fusions. Third, dataset-specific hyperparameter-tuning is frequently required, further undermining their practicability. In light of this, we propose a fast multi-view clustering via ensembles (FastMICE) approach. Particularly, the concept of random view groups is presented to capture the versatile view-wise relationships, through which the hybrid early-late fusion strategy is designed to enable efficient multi-stage fusions. With multiple views extended to many view groups, three levels of diversity (w.r.t. features, anchors, and neighbors, respectively) are jointly leveraged for constructing the view-sharing bipartite graphs in the early-stage fusion. Then, a set of diversified base clusterings for different view groups are obtained via fast graph partitioning, which are further formulated into a unified bipartite graph for final clustering in the late-stage fusion. Notably, FastMICE has almost linear time and space complexity, and is free of dataset-specific tuning. Experiments on 22 multi-view datasets demonstrate its advantages in scalability (for extremely large datasets), superiority (in clustering performance), and simplicity (to be applied) over the state-of-the-art. Code available: https://github.com/huangdonghere/FastMICE. | ['Jian-Huang Lai', 'Chang-Dong Wang', 'Dong Huang'] | 2022-03-22 | null | null | null | null | ['graph-partitioning'] | ['graphs'] | [-2.45611131e-01 -3.22699487e-01 -2.39578202e-01 -2.82023609e-01
-8.52606654e-01 -8.43518674e-01 4.84686494e-01 4.95108739e-02
1.62302136e-01 3.64404470e-01 3.41110885e-01 1.41773354e-02
-4.86544043e-01 -6.89517915e-01 -2.93050021e-01 -1.05388117e+00
1.02266818e-01 4.73312438e-01 2.14267522e-01 1.04085337e-02
-9.80665628e-03 2.97913969e-01 -1.47151959e+00 2.77582556e-01
9.18834865e-01 7.32244074e-01 1.81565762e-01 3.69577527e-01
1.31122693e-01 2.56502539e-01 -2.48894617e-01 -4.33008224e-01
2.38331348e-01 -3.19026977e-01 -5.42695045e-01 6.49028182e-01
8.80380943e-02 2.07722634e-02 -1.54627621e-01 9.73206699e-01
5.88151336e-01 2.18402855e-02 3.46383363e-01 -1.46203363e+00
-4.40798312e-01 5.58840752e-01 -1.03558815e+00 -2.17897892e-01
3.69792730e-01 -1.44140404e-02 1.22694433e+00 -1.21707845e+00
6.75449908e-01 9.64126170e-01 5.02977014e-01 1.01760484e-01
-1.51634026e+00 -5.66899836e-01 3.72394085e-01 8.69455338e-02
-1.75289094e+00 -4.88915563e-01 8.65719795e-01 -4.57812160e-01
3.53295952e-01 4.08171654e-01 8.33289206e-01 7.48882890e-01
-5.75082079e-02 6.33637726e-01 1.09988105e+00 -1.96722411e-02
2.61003673e-01 -5.95629439e-02 7.59990215e-02 6.18420839e-01
5.03817558e-01 -3.03918242e-01 -5.61766505e-01 -4.67375576e-01
3.98443729e-01 4.77813184e-01 -5.05574048e-01 -9.55022633e-01
-1.43419063e+00 8.26410115e-01 4.85674322e-01 3.19770515e-01
-4.69552398e-01 -3.71819943e-01 3.67204934e-01 1.69125989e-01
3.32486689e-01 -2.12378558e-02 -1.52426258e-01 2.63302237e-01
-1.12173879e+00 -7.80987442e-02 6.07288122e-01 1.09415913e+00
9.96898532e-01 -3.38895321e-01 1.86965391e-01 7.58614302e-01
3.72633696e-01 3.72610033e-01 2.86440626e-02 -9.27306533e-01
5.44962585e-01 1.00491309e+00 -2.93166280e-01 -1.23524523e+00
-5.48290789e-01 -4.72731084e-01 -1.38215363e+00 -2.44022816e-01
5.42700291e-02 4.42403480e-02 -7.77851701e-01 1.54959404e+00
7.12191284e-01 1.07850013e-02 8.49449635e-02 8.76940310e-01
9.12818313e-01 3.84572715e-01 -3.71100515e-01 -6.04814172e-01
1.45188105e+00 -9.72753704e-01 -5.31164169e-01 2.03222379e-01
4.46270615e-01 -6.36940718e-01 5.96724749e-01 4.64861423e-01
-9.40538466e-01 -4.42988068e-01 -9.72002327e-01 2.85853654e-01
-3.08331758e-01 5.08328266e-02 6.54187322e-01 5.44421434e-01
-1.21332192e+00 -1.24427499e-02 -7.72299409e-01 -5.44726193e-01
3.64359528e-01 4.14300740e-01 -6.12566650e-01 -4.73574996e-01
-8.19171190e-01 2.08696142e-01 5.63278139e-01 5.21028414e-02
-5.22960126e-01 -6.49174631e-01 -6.43743455e-01 4.68395166e-02
8.57737243e-01 -1.09004629e+00 5.13211429e-01 -5.03823340e-01
-1.09294486e+00 6.40682817e-01 -2.21404389e-01 9.11410302e-02
2.96851575e-01 1.13260038e-01 -6.32452190e-01 3.96159351e-01
2.80325085e-01 6.24248385e-01 5.93391299e-01 -1.82119215e+00
-6.61515176e-01 -5.38438261e-01 8.86498094e-02 5.47266901e-01
-4.00385201e-01 -2.90610433e-01 -1.21453249e+00 -5.46217680e-01
5.75494647e-01 -1.19998348e+00 -4.12879020e-01 -2.89478779e-01
-5.16661346e-01 -7.07350746e-02 7.90888011e-01 -3.96976739e-01
1.64832532e+00 -2.28846478e+00 6.34311974e-01 5.88197768e-01
7.95378089e-01 2.90181935e-02 8.45501125e-02 8.16057801e-01
-2.47860122e-02 4.38532494e-02 -3.46963793e-01 -3.78694087e-01
-1.84552193e-01 2.30189145e-01 1.15993842e-01 6.20142162e-01
-3.60722452e-01 7.44896233e-01 -9.11053360e-01 -6.01993740e-01
3.81296217e-01 3.59641612e-01 -5.52958906e-01 -3.78766619e-02
3.13605845e-01 6.27590775e-01 -5.01132965e-01 8.42719793e-01
8.66364479e-01 -7.67205179e-01 6.38807416e-01 -5.59256673e-01
1.23852119e-01 -3.79467368e-01 -1.38166559e+00 1.84275591e+00
-1.14621982e-01 -1.52709549e-02 3.63814205e-01 -8.10900509e-01
6.04190052e-01 3.61439854e-01 9.58549798e-01 -4.43768799e-02
-1.94289268e-03 1.91190690e-01 -4.00792062e-02 -3.99664417e-02
3.93605113e-01 5.04129641e-02 -2.25311086e-01 5.20763099e-01
1.39074713e-01 6.85275868e-02 4.70515639e-01 6.71133995e-01
7.28794158e-01 -9.23006386e-02 5.57002366e-01 -2.87196487e-01
6.01442337e-01 -1.12478204e-01 5.88410854e-01 2.60367125e-01
-1.51827093e-02 8.33066583e-01 3.46100986e-01 -1.86408564e-01
-6.82379246e-01 -1.08197439e+00 1.28484845e-01 8.11125159e-01
3.95314157e-01 -9.26253557e-01 -5.98444879e-01 -7.56930768e-01
-3.35574113e-02 2.30260879e-01 -5.46243429e-01 3.35767977e-02
-2.15314016e-01 -8.68878901e-01 1.75562147e-02 3.92522633e-01
2.95190781e-01 -3.51229906e-01 -2.58085489e-01 1.46781936e-01
-4.33224976e-01 -1.11146271e+00 -5.53701818e-01 -9.79556590e-02
-8.50424528e-01 -1.33175409e+00 -7.06921577e-01 -3.29100519e-01
9.56371486e-01 1.07614374e+00 1.06812263e+00 1.40985593e-01
1.09804362e-01 7.96948314e-01 -5.57173610e-01 1.00419305e-01
-1.18374966e-01 1.14140056e-01 2.38352597e-01 3.54240984e-01
1.60298020e-01 -8.72207761e-01 -8.63865912e-01 6.19926095e-01
-9.87189651e-01 2.03749776e-01 5.59985757e-01 8.85732114e-01
8.57627630e-01 3.46414059e-01 4.58104819e-01 -9.48440790e-01
2.64500529e-01 -6.68322504e-01 -4.14118767e-01 5.66593945e-01
-6.94134176e-01 -4.41642255e-01 6.29014075e-01 -8.29360075e-03
-9.12316144e-01 4.53084819e-02 3.52495253e-01 -6.99970424e-01
-1.53526263e-02 6.60591304e-01 -4.78776455e-01 2.78680474e-02
2.38246053e-01 1.73667699e-01 1.01156250e-01 -3.11181009e-01
6.70206070e-01 4.11014318e-01 3.11373800e-01 -3.25108141e-01
8.75026405e-01 7.90064633e-01 -6.40245248e-03 -5.05010128e-01
-7.42662132e-01 -7.58460402e-01 -8.34957898e-01 -4.04746830e-01
8.25492144e-01 -1.36488831e+00 -5.74943185e-01 3.95929217e-01
-5.24795115e-01 3.61626565e-01 -2.53960863e-02 4.33040202e-01
-2.54145801e-01 6.51598573e-01 -3.56052279e-01 -4.30893660e-01
-2.46520877e-01 -1.15863109e+00 9.79261816e-01 1.20433442e-01
5.90363443e-02 -1.05232799e+00 -1.58803388e-01 7.47106373e-01
-5.69209158e-02 3.80621880e-01 5.19255817e-01 -5.79014122e-01
-6.59017026e-01 -2.97332474e-04 -8.82015154e-02 4.80027124e-03
4.24341589e-01 1.58027738e-01 -7.38402188e-01 -8.74455810e-01
-1.23401538e-01 -5.67182750e-02 7.72876382e-01 4.55123574e-01
9.65989470e-01 -2.27326117e-02 -6.25973701e-01 6.70577645e-01
1.50957334e+00 -5.17643131e-02 1.71874702e-01 -1.09548464e-01
1.05236185e+00 5.59109688e-01 7.52059340e-01 6.56646311e-01
9.07045245e-01 6.26867354e-01 2.32945353e-01 -2.47687727e-01
4.95664291e-02 1.36168487e-02 2.09847808e-01 1.50229764e+00
-2.71661758e-01 -4.66463596e-01 -8.52171719e-01 5.27426183e-01
-2.06140089e+00 -8.62153172e-01 -2.72744030e-01 2.26057220e+00
1.95730969e-01 -1.51077494e-01 4.51506913e-01 8.44521299e-02
1.02013683e+00 4.30993021e-01 -5.47848403e-01 2.56911576e-01
-1.85932606e-01 -5.04492462e-01 2.93017179e-01 8.38200673e-02
-1.14463842e+00 5.76788425e-01 5.32151175e+00 8.49784672e-01
-7.61901319e-01 1.73894241e-01 5.59868872e-01 -2.51626790e-01
-5.77261448e-01 2.00679883e-01 -5.90939105e-01 3.38997155e-01
6.12855792e-01 -2.01993242e-01 5.05275071e-01 4.49427903e-01
1.49199329e-02 -2.27794856e-01 -7.94475436e-01 1.15116680e+00
2.92193666e-02 -1.24873900e+00 1.30253509e-01 4.25578177e-01
9.20532703e-01 5.04550859e-02 -1.07333675e-01 -1.05032660e-01
4.76555288e-01 -4.01229531e-01 3.68102282e-01 1.75135061e-01
8.63934398e-01 -1.02918053e+00 5.75339079e-01 3.65575731e-01
-1.69122827e+00 -1.76951110e-01 -2.00352013e-01 4.29110050e-01
4.54691052e-01 9.46572363e-01 -2.44613245e-01 1.69968271e+00
8.40062976e-01 9.24431622e-01 -6.40552700e-01 5.53399384e-01
1.21803142e-01 3.53743523e-01 -4.32339668e-01 5.77894628e-01
2.08717659e-01 -5.71112692e-01 5.88417768e-01 7.98975885e-01
5.13280749e-01 2.78424740e-01 6.03871763e-01 3.00061196e-01
4.54763621e-02 4.23894338e-02 -6.68241084e-01 1.75389349e-01
6.62388384e-01 1.72603464e+00 -1.13548052e+00 -4.07372594e-01
-6.90660536e-01 7.62480915e-01 2.91521102e-01 3.69179368e-01
-8.36063921e-01 5.74249662e-02 3.97557765e-01 -6.38312660e-04
5.93553960e-01 -1.25386909e-01 -2.39723369e-01 -1.49952543e+00
1.04236484e-01 -1.07585907e+00 9.35059726e-01 -4.30428386e-01
-1.45696163e+00 5.59708655e-01 1.32875532e-01 -1.69930124e+00
2.10195873e-02 -6.33237362e-02 -4.19455141e-01 4.60554749e-01
-9.81384277e-01 -1.48524702e+00 -4.03193861e-01 8.97018552e-01
2.82456577e-01 -6.51237667e-02 6.81388259e-01 3.97825480e-01
-6.38076723e-01 4.99161661e-01 1.29499272e-01 -2.08443716e-01
7.69744515e-01 -1.09570360e+00 -6.91034272e-03 1.01720107e+00
2.18613505e-01 7.00972617e-01 2.35533938e-01 -5.24905086e-01
-1.59537852e+00 -1.03967357e+00 4.00559008e-01 -3.34893852e-01
6.16957963e-01 -4.97836590e-01 -8.36671889e-01 4.94178474e-01
2.65301913e-01 1.42228171e-01 1.29002571e+00 3.01734149e-01
-4.19260532e-01 -2.32632235e-01 -1.00547266e+00 6.29140854e-01
1.12224543e+00 -4.49415058e-01 -1.48880407e-01 1.28972396e-01
5.26329398e-01 -1.77556813e-01 -1.45539248e+00 5.89832604e-01
4.50446546e-01 -1.44239545e+00 9.45852101e-01 -2.08139755e-02
-1.11898445e-02 -7.33594179e-01 -3.90248805e-01 -1.20785546e+00
-6.80667102e-01 -7.46585310e-01 -3.08854848e-01 1.66743791e+00
2.09703445e-01 -8.57307911e-01 4.74506944e-01 3.32488954e-01
-4.11898419e-02 -9.18151855e-01 -7.78624654e-01 -5.47561228e-01
-4.51762974e-01 -3.09075154e-02 6.90441787e-01 1.22093809e+00
-1.02583796e-01 5.05640984e-01 -4.22144443e-01 4.26929802e-01
7.62444973e-01 7.72545636e-01 9.72982526e-01 -1.34334528e+00
-2.88334489e-01 -3.52208883e-01 -3.00246567e-01 -8.12021613e-01
-2.55199671e-01 -9.26716387e-01 -3.91881913e-01 -1.61551785e+00
6.92180216e-01 -3.81443948e-01 -4.95627522e-01 3.43314111e-01
-5.23482621e-01 3.06652755e-01 4.58599448e-01 6.06041074e-01
-9.65555847e-01 5.07621050e-01 1.31428444e+00 1.72420397e-01
-1.32257849e-01 6.32627010e-02 -9.68192339e-01 6.26954913e-01
6.02278769e-01 -3.12077641e-01 -7.32608378e-01 -4.43567522e-02
9.46907774e-02 3.22641522e-01 1.17089435e-01 -9.47059810e-01
3.97319913e-01 -7.62862712e-03 2.77789235e-01 -9.94620144e-01
3.07779700e-01 -1.03027022e+00 7.66878307e-01 2.35356078e-01
2.99886495e-01 2.85487860e-01 -1.67716771e-01 9.88921046e-01
-5.39489090e-01 3.74548912e-01 5.76669753e-01 -7.42876977e-02
-6.10690057e-01 6.45117939e-01 -5.86705729e-02 -8.09379146e-02
1.36559510e+00 -5.68304420e-01 -1.10448688e-01 -4.28610921e-01
-8.34893823e-01 6.34366632e-01 8.99670780e-01 3.55312318e-01
3.34059834e-01 -1.47525513e+00 -6.19427919e-01 1.07798927e-01
3.83891582e-01 3.18747163e-01 7.82263994e-01 1.33864379e+00
-8.11675638e-02 1.16072513e-01 5.52591346e-02 -9.25456703e-01
-1.40114665e+00 9.50248837e-01 -2.54593134e-01 -5.24692357e-01
-5.76730788e-01 4.97177988e-01 5.65150201e-01 -4.49613184e-01
-2.15169847e-01 2.54409611e-01 -6.00487180e-02 5.02305567e-01
1.08460486e-01 5.08001685e-01 -3.41234766e-02 -8.71620357e-01
-5.75603366e-01 8.65174472e-01 -2.89082788e-02 1.10626265e-01
1.41802490e+00 -6.78848863e-01 -4.18349832e-01 4.39690083e-01
1.05218732e+00 2.08982587e-01 -1.13165128e+00 -2.55443662e-01
-1.29082397e-01 -4.15710270e-01 -1.63248003e-01 -4.46550280e-01
-1.52130938e+00 5.58834553e-01 1.22705065e-01 3.51219684e-01
1.62927699e+00 1.66156426e-01 6.85187578e-01 5.02077900e-02
5.92682660e-01 -8.96677375e-01 -1.66418239e-01 4.56411466e-02
6.73522174e-01 -1.09952569e+00 4.84318197e-01 -7.90732861e-01
-8.39530766e-01 7.97035456e-01 4.88394469e-01 2.25380138e-01
8.04299831e-01 9.93958265e-02 4.67791408e-02 -6.28148556e-01
-8.54501903e-01 -1.50877938e-01 3.45784098e-01 4.07064766e-01
1.05586201e-01 2.66255915e-01 -1.51188508e-01 3.64046454e-01
1.80629894e-01 -3.96663487e-01 2.90717542e-01 7.74918377e-01
-9.50066671e-02 -1.06037521e+00 -4.34851348e-01 4.97185439e-01
-2.12552860e-01 1.04093440e-01 -3.94607067e-01 7.25770175e-01
1.73148200e-01 1.19954264e+00 -1.95867136e-01 -5.97366571e-01
1.75873220e-01 -1.45698011e-01 1.96514994e-01 -4.76753592e-01
-6.99293971e-01 5.89684784e-01 1.29319513e-02 -7.46035278e-01
-9.41043198e-01 -7.16818988e-01 -8.95287335e-01 -4.04193640e-01
-5.45883119e-01 2.92572051e-01 2.11956933e-01 5.36202133e-01
8.95243406e-01 3.45918000e-01 1.03072870e+00 -9.07763898e-01
-1.16125755e-01 -5.40789843e-01 -6.59666002e-01 4.61793214e-01
-1.08672194e-02 -8.15883934e-01 -3.31324875e-01 -4.55181636e-02] | [8.209094047546387, 4.630855560302734] |
f0cec7e9-322b-4f72-8f0d-b25d2b623fe1 | uncertainty-quantification-using-bayesian | null | null | https://openreview.net/forum?id=Sk_P2Q9sG | https://openreview.net/pdf?id=Sk_P2Q9sG | Uncertainty quantification using Bayesian neural networks in classification: Application to ischemic stroke lesion segmentation | Most recent research of neural networks in the field of computer vision has focused on improving accuracy of point predictions by developing various network architectures or learning algorithms. Uncertainty quantification accompanied by point estimation can lead to a more informed decision, and the quality of prediction can be improved. In medical imaging applications, assessment of uncertainty could potentially reduce untoward outcomes due to suboptimal decisions. In this paper, we invoke a Bayesian neural network and propose a natural way to quantify uncertainty in classification problems by decomposing predictive uncertainty into two parts, aleatoric and epistemic uncertainty. The proposed method takes into account discrete nature of the outcome, yielding correct interpretation of each uncertainty. We demonstrate that the proposed uncertainty quantification method provides additional insight to the point prediction using images from the Ischemic Stroke Lesion Segmentation Challenge. | ['Joong-Ho Won', 'Beom Joon Kim', 'Yongchan Kwon', 'Myunghee Cho Paik'] | 2018-04-10 | null | null | null | midl-2018-conference-2018-4 | ['ischemic-stroke-lesion-segmentation'] | ['medical'] | [ 2.74737954e-01 6.50153458e-01 -2.80762643e-01 -6.70276105e-01
-9.76201117e-01 -2.79549271e-01 5.93004405e-01 3.69403273e-01
-2.70983636e-01 1.04351687e+00 3.09614122e-01 -4.78073716e-01
-6.12185836e-01 -8.98321748e-01 -6.50185287e-01 -7.48060346e-01
8.62622336e-02 5.29862463e-01 7.53083155e-02 4.98233616e-01
4.04879212e-01 4.63867784e-01 -9.82612252e-01 2.72050172e-01
9.91389930e-01 1.51749408e+00 -2.85825908e-01 3.72311562e-01
-6.07587397e-02 7.16858447e-01 -5.53834498e-01 -4.49325919e-01
1.47285849e-01 -4.80611697e-02 -7.49391735e-01 -1.92811221e-01
-6.89700320e-02 -5.01784742e-01 6.93563968e-02 1.29923296e+00
1.34751961e-01 1.10373022e-02 1.32588434e+00 -1.11624837e+00
-5.48115075e-01 1.16165555e+00 -3.14517468e-01 2.08133429e-01
-1.56788632e-01 9.14554745e-02 8.38248789e-01 -5.32223105e-01
1.17548853e-01 1.21670198e+00 7.45352864e-01 4.19452608e-01
-1.07430768e+00 -2.51276582e-01 1.02664120e-01 3.89845014e-01
-1.07871640e+00 -1.72357634e-01 7.99705923e-01 -7.16526151e-01
3.63350838e-01 5.49577475e-02 5.28450370e-01 1.12428308e+00
4.79745299e-01 7.79102445e-01 1.23768294e+00 -4.81929272e-01
7.03139782e-01 1.33291688e-02 3.79686743e-01 3.70412409e-01
6.49263501e-01 7.17152297e-01 -1.35926574e-01 -2.44918868e-01
5.56338966e-01 1.64271146e-02 -3.04563373e-01 -1.97115317e-01
-8.97469401e-01 9.57451582e-01 6.68215990e-01 -3.86711545e-02
-4.54062343e-01 4.56276298e-01 2.35228926e-01 -3.42063755e-01
6.22130930e-01 3.55197638e-01 -2.98811644e-01 5.82211791e-03
-9.50423300e-01 5.62831387e-02 6.65101409e-01 4.59399104e-01
1.62782997e-01 -8.93591344e-02 -5.04193246e-01 5.76763034e-01
6.84621751e-01 2.16968149e-01 2.51981050e-01 -1.37946641e+00
1.13493547e-01 1.31110951e-01 3.51130664e-01 -7.84566164e-01
-5.59915543e-01 -5.29050589e-01 -8.92208636e-01 6.42388999e-01
5.98748446e-01 -2.93644100e-01 -1.11852241e+00 1.55889809e+00
4.55149114e-02 3.41285348e-01 3.88431782e-03 7.04508424e-01
5.33008218e-01 3.23243678e-01 2.18701839e-01 -1.73782691e-01
1.17622876e+00 -5.04232168e-01 -8.01654398e-01 -4.17262912e-02
2.29626790e-01 -1.58206686e-01 2.95030594e-01 5.47373474e-01
-8.57946277e-01 -1.44654498e-01 -1.11959565e+00 3.38193625e-01
-6.04704507e-02 -1.95510045e-01 5.34031451e-01 7.67045736e-01
-7.26793051e-01 1.12804055e+00 -1.01830256e+00 1.26380384e-01
8.66166413e-01 2.31303260e-01 1.74076974e-01 9.47583914e-02
-1.36385131e+00 1.33061743e+00 6.42926097e-01 3.27342033e-01
-8.29394221e-01 -7.74680734e-01 -6.66528583e-01 1.79944098e-01
3.53120804e-01 -8.07784259e-01 1.39292479e+00 -6.34859622e-01
-1.49514902e+00 3.31798673e-01 2.32901886e-01 -1.00254476e+00
8.95902097e-01 -3.15541536e-01 -7.24160904e-03 1.68038815e-01
-1.11887775e-01 7.52273202e-01 1.03301072e+00 -1.19644129e+00
-6.04438126e-01 -4.83969718e-01 -5.80429062e-02 -5.36008775e-02
1.83364496e-01 -3.71043533e-01 3.01132947e-01 -4.61577743e-01
2.34133244e-01 -7.23676682e-01 -5.00440300e-01 1.34301439e-01
-3.89316916e-01 -2.45745033e-01 2.39884660e-01 -4.87911999e-01
8.42445493e-01 -1.94635642e+00 -1.18069880e-01 2.43741557e-01
4.38539267e-01 -1.28376797e-01 4.43649203e-01 -1.78468049e-01
1.53961917e-02 3.92116517e-01 -6.51257157e-01 1.40821468e-03
3.32912989e-02 3.91667128e-01 -4.02499616e-01 3.83049458e-01
3.88358682e-01 7.78610885e-01 -8.61551225e-01 -5.74783266e-01
4.05100971e-01 4.98380750e-01 -3.65726799e-01 -5.58755584e-02
-4.02151167e-01 5.12048602e-01 -6.48434043e-01 5.94043076e-01
5.86482704e-01 -3.47198099e-01 -1.44312471e-01 -2.85785824e-01
2.37999350e-01 -2.64892913e-02 -9.52143729e-01 1.30683160e+00
-1.35889798e-01 4.97371465e-01 -4.16449875e-01 -1.12384117e+00
6.72956645e-01 3.07302207e-01 4.80708271e-01 8.20581466e-02
5.52659452e-01 1.18580870e-01 4.35429141e-02 -2.96241820e-01
5.23089617e-02 -4.06586468e-01 3.47952396e-02 2.03942478e-01
-1.26193643e-01 -2.65935302e-01 -4.07490402e-01 -1.42824396e-01
8.12044501e-01 2.04433709e-01 5.74385345e-01 -2.84320116e-01
2.06732795e-01 -1.77261010e-02 5.62067807e-01 1.13871861e+00
-7.65079319e-01 6.80495203e-01 7.36813784e-01 -4.54493284e-01
-7.65831828e-01 -1.39879179e+00 -8.37677777e-01 3.20464253e-01
-5.99067286e-03 3.99552613e-01 -7.23698974e-01 -7.08442390e-01
1.52548775e-01 1.06882620e+00 -8.61440539e-01 -4.40378785e-01
-1.60510734e-01 -1.03907073e+00 5.16197920e-01 8.86808515e-01
4.87307489e-01 -7.84870923e-01 -8.47571194e-01 1.88257828e-01
-2.32626140e-01 -8.07427704e-01 5.52418046e-02 3.48745793e-01
-1.25286210e+00 -1.05990565e+00 -7.43265390e-01 1.01540901e-01
2.37880170e-01 -4.06628013e-01 9.71880317e-01 -2.36882076e-01
-5.08594438e-02 3.65701109e-01 -2.35941976e-01 -8.28744829e-01
-3.79483432e-01 -3.57823074e-01 3.63804065e-02 -1.53108299e-01
3.45912158e-01 -4.69138950e-01 -8.50383043e-01 -4.68346430e-03
-7.30968297e-01 -2.67328441e-01 5.69640219e-01 1.02725136e+00
5.07478297e-01 8.33145529e-02 8.77548039e-01 -8.32808673e-01
7.90555179e-01 -6.16291940e-01 -6.51400208e-01 3.02226931e-01
-1.21628189e+00 4.47491944e-01 6.31882846e-02 -1.79821730e-01
-1.47582757e+00 -6.65401220e-02 -6.14771247e-03 -3.69559884e-01
-3.18477720e-01 8.36540282e-01 1.63401872e-01 4.55459952e-02
8.82028282e-01 -2.47564286e-01 -2.02108119e-02 -1.03843935e-01
6.40254974e-01 6.52050376e-01 6.55119061e-01 -7.83118248e-01
1.69871040e-02 4.68627423e-01 4.96767193e-01 -3.04675102e-01
-1.13139796e+00 2.79099923e-02 -6.25323117e-01 -3.58173490e-01
9.35084999e-01 -5.02134085e-01 -7.50721812e-01 2.89306790e-01
-1.16510773e+00 6.42899796e-02 -4.43918943e-01 7.17341363e-01
-8.76237750e-01 4.11196351e-01 -4.42868054e-01 -1.03926206e+00
-3.09917390e-01 -1.31197178e+00 7.42099822e-01 2.82110125e-01
-2.37404540e-01 -1.12156415e+00 -1.66794851e-01 2.81707138e-01
2.94946134e-01 4.21138704e-01 1.15052271e+00 -8.35102499e-01
-6.49901748e-01 -4.29584622e-01 -4.62408930e-01 4.58032429e-01
-1.23677149e-01 -3.05459797e-02 -1.06357944e+00 3.68986011e-01
3.11114192e-01 -1.24892928e-01 1.19790673e+00 1.26082718e+00
1.28685284e+00 -1.46602049e-01 -3.90890747e-01 2.97736615e-01
1.33867764e+00 3.64147455e-01 5.78961730e-01 2.75609136e-01
2.08426058e-01 6.36220694e-01 3.99873883e-01 5.20136893e-01
1.60123914e-01 3.44658732e-01 8.58641565e-01 5.96278965e-01
1.13520557e-02 1.25664741e-01 -1.79487184e-01 -7.31314858e-03
-1.63673356e-01 -1.24228671e-01 -1.08979726e+00 4.28779989e-01
-2.06912160e+00 -8.78467262e-01 1.64123718e-02 2.05889201e+00
7.75206685e-01 4.66100574e-01 -2.16942325e-01 1.50410011e-01
7.01757491e-01 -1.15085758e-01 -6.52438283e-01 -1.52426720e-01
2.08748728e-01 -1.90402761e-01 5.42184770e-01 5.50271332e-01
-1.08650970e+00 3.99977118e-01 7.54334497e+00 6.83379412e-01
-7.97571957e-01 9.77347642e-02 1.06982028e+00 1.88954803e-03
-2.85722077e-01 -1.42198533e-01 -4.54539746e-01 4.37159032e-01
9.91722465e-01 -7.73694515e-02 6.48000687e-02 9.94137645e-01
1.70387402e-01 -4.65885103e-01 -1.40724921e+00 7.07284570e-01
-3.33975494e-01 -1.68851340e+00 3.23578976e-02 -1.25483066e-01
6.55919015e-01 -3.02776266e-02 1.91587195e-01 8.12503323e-03
5.45468748e-01 -1.28806746e+00 7.60685503e-01 1.13473654e+00
3.80495518e-01 -7.91291177e-01 1.04200566e+00 3.92555237e-01
-3.66944820e-01 -2.53084242e-01 -2.81393319e-01 -3.89764309e-02
2.90344775e-01 1.13269830e+00 -9.38126028e-01 3.45737755e-01
6.72732234e-01 4.84805286e-01 -1.62161887e-01 1.37547672e+00
-3.26930940e-01 6.85347497e-01 -4.00203854e-01 -3.05843540e-02
1.75012648e-01 -1.34191260e-01 7.78248131e-01 8.46958160e-01
2.36099496e-01 3.01374495e-01 -2.13683784e-01 1.30564487e+00
1.08560771e-01 -4.35012609e-01 -4.57788974e-01 1.79493546e-01
4.16851103e-01 8.52496803e-01 -8.29360843e-01 -2.73245007e-01
-5.53260557e-02 3.98085177e-01 3.19681466e-01 2.69371092e-01
-7.03642309e-01 3.16341151e-03 2.10894451e-01 -3.79044086e-01
-7.57431611e-02 1.56254038e-01 -1.21173406e+00 -9.82187390e-01
-1.66536003e-01 -2.17783630e-01 5.15507460e-01 -8.89245033e-01
-1.44647288e+00 5.81486762e-01 4.66516912e-01 -9.19353247e-01
-4.56211418e-01 -1.04121590e+00 -6.01615667e-01 1.00277019e+00
-1.34396362e+00 -8.38575482e-01 2.94630360e-02 6.78018630e-02
3.51225823e-01 -2.51695011e-02 8.06152701e-01 -2.46636614e-01
-3.82077128e-01 1.06886663e-01 2.07809374e-01 -9.51102078e-02
4.62600172e-01 -1.42882490e+00 -3.37134868e-01 7.24940300e-01
-1.52886614e-01 2.44847447e-01 9.97399509e-01 -7.13026881e-01
-4.03584331e-01 -9.30389822e-01 3.56412739e-01 -8.18429828e-01
6.78537965e-01 5.42209864e-01 -8.27166557e-01 5.42364419e-01
-1.05829500e-01 4.96404581e-02 5.33473790e-01 2.51263738e-01
-2.11541995e-01 6.28327876e-02 -1.49391103e+00 4.36785847e-01
4.69221503e-01 -2.91845411e-01 -1.04496634e+00 2.32016340e-01
7.40906239e-01 -2.89892852e-01 -9.14628327e-01 7.31908321e-01
6.62650347e-01 -9.66100693e-01 9.29060638e-01 -7.10493803e-01
7.54832983e-01 4.03085463e-02 -3.50342005e-01 -1.52598858e+00
-2.67405391e-01 1.06971301e-01 -4.53788549e-01 7.05374122e-01
5.00345051e-01 -5.00389159e-01 7.36743093e-01 1.22767174e+00
-2.04532564e-01 -8.10823023e-01 -1.19818020e+00 -5.69704115e-01
5.23870409e-01 -1.00260067e+00 6.36555851e-01 4.50871080e-01
5.98804951e-02 -1.62010819e-01 -3.23828220e-01 4.73776996e-01
1.11646295e+00 -1.73362330e-01 -2.91853786e-01 -1.62070513e+00
-1.49941072e-01 -7.63296664e-01 -5.10678232e-01 -4.17299837e-01
1.71390295e-01 -6.99082494e-01 2.72598863e-01 -1.66086090e+00
1.93385676e-01 -3.32519025e-01 -5.59254587e-01 2.92712808e-01
-2.20788226e-01 -2.83483621e-02 -1.15648145e-02 2.52524406e-01
-1.33017957e-01 4.08402175e-01 9.89479303e-01 -2.93785691e-01
1.67261273e-01 5.33800006e-01 -7.38015175e-01 1.14458299e+00
8.09417307e-01 -5.62085152e-01 -3.90130818e-01 -1.88293144e-01
2.72766918e-01 5.23019135e-01 7.36350298e-01 -8.80460680e-01
1.44050002e-01 -2.93355674e-01 5.92841685e-01 -5.54917336e-01
1.83140412e-01 -1.11416686e+00 2.32431442e-02 5.16289353e-01
-6.35311365e-01 -6.78679407e-01 -7.77115347e-04 1.09642553e+00
-9.64068621e-02 -8.33333075e-01 9.72710609e-01 -1.27719060e-01
-4.90307033e-01 1.13475747e-01 -3.31336975e-01 -2.57566005e-01
9.70518470e-01 -4.41626133e-03 -2.72859037e-01 -4.23294067e-01
-1.34511733e+00 6.82744533e-02 -2.94903725e-01 2.10030992e-02
7.05078006e-01 -1.32366860e+00 -6.53218448e-01 -3.01173151e-01
6.59265444e-02 3.70360874e-02 3.21657479e-01 7.60574877e-01
-4.23240811e-01 5.73051989e-01 -1.29756182e-01 -9.31107283e-01
-5.16234577e-01 3.75565231e-01 6.71014130e-01 -3.07932734e-01
-4.60413814e-01 7.79287994e-01 -1.36237800e-01 -6.30380511e-02
3.96022588e-01 -6.21968150e-01 -4.72812682e-01 4.33236882e-02
3.98730934e-01 6.92639530e-01 -9.00065154e-02 -1.18717380e-01
-1.83599651e-01 2.52104282e-01 -4.36441675e-02 -4.04901475e-01
1.10342836e+00 -2.92613953e-01 3.95655707e-02 5.46144783e-01
4.85617965e-01 -6.10076606e-01 -1.51297617e+00 -1.44755140e-01
3.51569235e-01 -2.32798859e-01 5.53556979e-01 -1.36862576e+00
-8.53157282e-01 1.08287215e+00 8.99423301e-01 1.99600428e-01
9.18341041e-01 5.23787700e-02 2.09524393e-01 4.36198235e-01
4.51343179e-01 -9.07744348e-01 -2.00318143e-01 2.54033476e-01
8.99837971e-01 -1.78323340e+00 2.43922509e-02 -1.68274879e-01
-6.40271068e-01 1.39962304e+00 3.11514288e-01 -4.34519500e-02
1.13161504e+00 3.07206307e-02 7.43979774e-03 -9.43233818e-02
-5.29451847e-01 1.90842479e-01 6.66739166e-01 7.49650419e-01
1.34729773e-01 4.58135009e-01 -3.57661486e-01 1.19032812e+00
1.07549794e-01 4.12205577e-01 5.46735108e-01 5.76167762e-01
-7.12025642e-01 -6.69235349e-01 -5.75794756e-01 8.36839259e-01
-6.06378257e-01 -1.48235619e-01 1.44131660e-01 3.11159551e-01
1.21006913e-01 9.40150201e-01 -1.34054080e-01 -5.79681881e-02
-2.21328065e-02 3.50192666e-01 4.74745959e-01 -4.07961935e-01
5.85326850e-02 -9.80146751e-02 9.49832201e-02 -3.70680004e-01
-4.24596339e-01 -6.50617421e-01 -1.21627069e+00 1.82747662e-01
-4.59860027e-01 5.56144044e-02 7.06180036e-01 1.36230195e+00
1.48104861e-01 7.33778596e-01 1.78539306e-01 -7.34030545e-01
-1.15520442e+00 -1.03953362e+00 -6.69933319e-01 -9.08343792e-02
4.45865840e-01 -1.00029838e+00 -5.17716229e-01 -8.31808671e-02] | [14.239113807678223, -2.054959774017334] |
a48cf463-f67a-4651-89e5-4ac1e7998f1d | line-drawing-guided-progressive-inpainting-of | 2211.06649 | null | https://arxiv.org/abs/2211.06649v1 | https://arxiv.org/pdf/2211.06649v1.pdf | Line Drawing Guided Progressive Inpainting of Mural Damages | Mural image inpainting refers to repairing the damage or missing areas in a mural image to restore the visual appearance. Most existing image-inpainting methods tend to take a target image as the only input and directly repair the damage to generate a visually plausible result. These methods obtain high performance in restoration or completion of some specific objects, e.g., human face, fabric texture, and printed texts, etc., however, are not suitable for repairing murals with varied subjects, especially for murals with large damaged areas. Moreover, due to the discrete colors in paints, mural inpainting may suffer from apparent color bias as compared to natural image inpainting. To this end, in this paper, we propose a line drawing guided progressive mural inpainting method. It divides the inpainting process into two steps: structure reconstruction and color correction, executed by a structure reconstruction network (SRN) and a color correction network (CCN), respectively. In the structure reconstruction, line drawings are used by SRN as a guarantee for large-scale content authenticity and structural stability. In the color correction, CCN operates a local color adjustment for missing pixels which reduces the negative effects of color bias and edge jumping. The proposed approach is evaluated against the current state-of-the-art image inpainting methods. Qualitative and quantitative results demonstrate the superiority of the proposed method in mural image inpainting. The codes and data are available at {https://github.com/qinnzou/mural-image-inpainting}. | ['Xiaoguang Wang', 'Xianfeng Huang', 'Chengfang Song', 'Long Chen', 'Hongkai Yu', 'Fan Zhang', 'Qin Zou', 'Luxi Li'] | 2022-11-12 | null | null | null | null | ['image-inpainting'] | ['computer-vision'] | [ 6.13981247e-01 -7.30902329e-02 7.91635737e-02 1.49768218e-01
-2.60799021e-01 -3.11043352e-01 2.83746451e-01 -1.02216890e-03
-8.45008194e-02 8.08053076e-01 2.40047164e-02 -2.42979098e-02
2.29953289e-01 -8.69515121e-01 -7.53225923e-01 -6.17662668e-01
4.75575238e-01 -8.40992108e-02 1.76507369e-01 -1.49915874e-01
5.29023230e-01 6.92780256e-01 -1.38565040e+00 3.18372309e-01
1.15762997e+00 7.49408782e-01 3.75329703e-01 5.68136275e-01
-2.44311556e-01 8.54469240e-01 -7.00119734e-01 -3.98436815e-01
2.39589676e-01 -8.20641279e-01 -4.81719732e-01 3.89975280e-01
3.87985885e-01 -6.43841684e-01 -4.36001509e-01 1.31677651e+00
4.09757316e-01 2.76307110e-02 6.00042462e-01 -1.22633159e+00
-9.25673544e-01 1.46916956e-01 -1.23920417e+00 -5.88076971e-02
3.92674476e-01 2.03627199e-01 4.37425166e-01 -9.95737851e-01
7.16706157e-01 1.27009356e+00 5.27141631e-01 5.77069819e-01
-1.33806503e+00 -5.74384868e-01 -2.59813160e-01 2.32464030e-01
-1.25274193e+00 -2.98967391e-01 1.31372619e+00 -2.49356136e-01
1.41924605e-01 5.71461022e-01 6.49280190e-01 6.85498536e-01
3.09183538e-01 7.60311604e-01 1.19929254e+00 -5.72435081e-01
7.36667514e-02 -1.26952335e-01 -4.44016665e-01 6.05055153e-01
2.54435688e-01 4.08121832e-02 -3.50904316e-01 1.83096007e-02
1.04671204e+00 2.42304459e-01 -5.82192302e-01 -4.24436741e-02
-9.15662110e-01 4.72256392e-01 3.84567708e-01 1.75271586e-01
-4.83058870e-01 4.73677590e-02 4.36681062e-01 4.28660780e-01
6.29365861e-01 1.46857172e-01 1.55864865e-01 3.56389545e-02
-1.23030901e+00 1.69569835e-01 3.00195962e-01 5.73149979e-01
6.90778673e-01 2.34749004e-01 -8.50141272e-02 1.19050181e+00
9.46027711e-02 4.19455707e-01 2.04999968e-01 -9.49859142e-01
4.73089993e-01 5.68367779e-01 1.72199577e-01 -1.44429874e+00
2.97074560e-02 -1.23358537e-02 -1.12722898e+00 7.84507334e-01
1.92230865e-01 -8.93235356e-02 -7.86103129e-01 1.28686345e+00
3.59889001e-01 8.25108290e-02 -1.11825928e-01 9.46181953e-01
7.42487073e-01 9.43563700e-01 -1.50862053e-01 -4.45890397e-01
1.29157948e+00 -1.11085594e+00 -1.03314412e+00 -2.37440035e-01
5.03205806e-02 -1.34874976e+00 1.26389849e+00 5.61162055e-01
-1.43589377e+00 -6.09862387e-01 -1.03754711e+00 -2.05221564e-01
1.40182242e-01 2.84893870e-01 1.01233989e-01 3.96557391e-01
-7.44276762e-01 4.99869525e-01 -4.02579308e-01 -1.65093869e-01
4.29638684e-01 -1.79082572e-01 -3.64154607e-01 -5.27385235e-01
-9.80199277e-01 7.41661370e-01 4.09992963e-01 4.38292891e-01
-7.76145458e-01 -5.98867059e-01 -6.56109035e-01 -7.16757104e-02
3.66394192e-01 -3.58919948e-01 8.50512743e-01 -1.26558936e+00
-1.35480785e+00 8.64989698e-01 1.23317480e-01 -1.29798770e-01
9.45356846e-01 -2.98187584e-01 -3.98822337e-01 4.32945371e-01
-5.46673015e-02 5.87190390e-01 1.22224033e+00 -1.80375183e+00
-3.11426073e-01 8.45740214e-02 -2.04906687e-01 3.91713589e-01
-7.34207258e-02 2.07107309e-02 -6.01502180e-01 -1.21749389e+00
1.19853295e-01 -4.50160265e-01 1.19844154e-01 5.20487368e-01
-5.08843482e-01 3.97696674e-01 1.37829721e+00 -1.28013217e+00
1.32585430e+00 -2.29002094e+00 3.17441896e-02 1.20644882e-01
1.03551902e-01 3.52117270e-01 -3.11001778e-01 6.35203779e-01
-1.82943091e-01 -5.94094545e-02 -5.78106582e-01 -3.28273535e-01
-4.71278250e-01 1.97426230e-01 -2.34874338e-01 5.07632256e-01
1.25815764e-01 6.53964698e-01 -7.05266893e-01 -6.30141973e-01
3.13237578e-01 6.27774358e-01 -2.29749009e-01 3.57611805e-01
-6.75075948e-02 3.93503189e-01 -9.60518271e-02 8.28888237e-01
1.09797454e+00 4.34320211e-01 -1.55810565e-01 -4.43030447e-01
-1.43672913e-01 -5.15145183e-01 -1.23886526e+00 1.51085591e+00
-2.96725810e-01 5.89015424e-01 4.74900454e-01 -6.99155986e-01
1.17371118e+00 2.13265181e-01 3.73697728e-01 -7.80866385e-01
1.51172712e-01 3.27916920e-01 -2.06726804e-01 -6.99697554e-01
6.18534982e-01 -7.64622390e-02 3.30198765e-01 4.75607872e-01
-6.15144849e-01 -4.20079708e-01 1.45189866e-01 2.00242281e-01
7.45898485e-01 1.34927273e-01 -1.22678531e-02 1.22860996e-02
6.88652039e-01 -1.67501301e-01 5.95956445e-01 2.49356970e-01
2.97790430e-02 1.08531749e+00 5.71372271e-01 -2.87263244e-01
-1.45292711e+00 -8.71119380e-01 8.33630189e-02 3.96241397e-01
4.91669029e-01 -2.88419984e-02 -1.05343413e+00 -3.03510994e-01
-2.15227991e-01 6.14994884e-01 -5.62972605e-01 -3.06320548e-01
-5.91153264e-01 -4.26546574e-01 5.16164184e-01 1.26104355e-01
9.30690587e-01 -1.52664208e+00 -5.48995197e-01 1.45113260e-01
-2.82121837e-01 -7.12981820e-01 -8.76090348e-01 -5.05055606e-01
-8.31195593e-01 -1.19982600e+00 -1.15133774e+00 -1.06639874e+00
1.08430243e+00 4.88003552e-01 6.92145884e-01 6.66709900e-01
-4.96043652e-01 7.69156143e-02 -4.32614386e-01 5.48372641e-02
-7.01117039e-01 -4.52332824e-01 -3.36870462e-01 2.63193727e-01
-4.24658149e-01 -6.99983597e-01 -8.37037981e-01 3.16522479e-01
-1.54771805e+00 4.80791897e-01 6.43892944e-01 9.67687309e-01
4.98955846e-01 3.69731933e-01 2.75261223e-01 -9.08690095e-01
9.63367760e-01 -1.01470806e-01 -4.31141794e-01 2.55503893e-01
-4.48672026e-01 -3.11813086e-01 8.68573427e-01 -5.11026144e-01
-1.29486275e+00 -2.62098551e-01 -8.65530968e-02 -5.70547879e-01
-7.79871196e-02 2.83123076e-01 -1.57635912e-01 -1.71466559e-01
3.78897160e-01 5.51468730e-01 3.66851211e-01 -5.49250185e-01
2.06555188e-01 4.59315121e-01 7.82127261e-01 -4.76097405e-01
1.06402326e+00 4.13294673e-01 -2.35644355e-02 -6.98567390e-01
-2.25956663e-01 1.23876579e-01 -2.94669986e-01 -5.04145622e-01
6.58635080e-01 -6.60346687e-01 -4.22174335e-01 9.18694258e-01
-1.22094262e+00 -3.63934577e-01 -3.33227247e-01 9.03598666e-02
-4.43378687e-01 8.77345860e-01 -1.02484369e+00 -7.33196735e-01
-5.43975353e-01 -8.82406473e-01 7.87139058e-01 5.16031206e-01
1.06175728e-01 -6.46297395e-01 -1.30192026e-01 3.77570212e-01
2.66960710e-01 6.67134166e-01 1.02765775e+00 3.91890287e-01
-4.14377809e-01 -9.14389491e-02 -4.84188855e-01 6.10855997e-01
3.59813720e-01 2.73851275e-01 -5.45634866e-01 -3.65382999e-01
-8.16670880e-02 -2.33207360e-01 6.32882476e-01 2.33443990e-01
1.00483239e+00 -4.24962729e-01 7.92699009e-02 4.77538705e-01
1.60684025e+00 5.14575422e-01 1.30259144e+00 3.15312147e-01
6.87953770e-01 6.42115116e-01 9.14353669e-01 3.98010761e-01
-2.92277783e-01 4.13111120e-01 4.59271759e-01 -7.15484679e-01
-5.92695713e-01 -4.90396172e-01 3.58911127e-01 7.01286554e-01
-1.47508115e-01 -2.78971702e-01 -4.85500753e-01 4.26787049e-01
-1.70417774e+00 -1.03786874e+00 -4.02135938e-01 2.10615611e+00
9.68049109e-01 -4.63797860e-02 -1.64397851e-01 3.59593511e-01
1.11487913e+00 1.14620864e-01 -4.61090773e-01 -6.03481650e-01
-3.22137058e-01 9.89419967e-03 3.81830096e-01 6.17278337e-01
-5.93434155e-01 7.29040027e-01 5.13419008e+00 1.25512850e+00
-1.08023691e+00 4.65244241e-02 8.74913514e-01 1.83979735e-01
-3.26920748e-01 1.05857803e-02 -2.50723250e-02 7.29484618e-01
-5.72858565e-02 2.54482001e-01 6.37501240e-01 3.82089823e-01
7.08897710e-01 -5.26740849e-01 -2.76524305e-01 1.24320817e+00
2.24323839e-01 -1.26206934e+00 1.76462412e-01 -2.91820645e-01
8.07975173e-01 -8.10342729e-01 8.91072974e-02 -4.08482671e-01
-3.40272069e-01 -8.98598492e-01 8.75833869e-01 6.28819942e-01
1.06828475e+00 -1.05503750e+00 5.35271525e-01 1.47173151e-01
-1.00761902e+00 1.05433576e-01 -3.26013118e-01 1.57619879e-01
3.54931146e-01 7.21394062e-01 -3.14365596e-01 6.14574373e-01
5.30564487e-01 5.26566625e-01 -6.02995932e-01 1.14617717e+00
-5.39070547e-01 4.36287791e-01 3.69915739e-02 5.69093764e-01
3.56021300e-02 -5.90506792e-01 5.78301370e-01 9.07812774e-01
3.83573085e-01 9.16725025e-02 -1.21740073e-01 9.59992826e-01
-8.04960504e-02 2.65539855e-01 -4.45632160e-01 5.13920151e-02
4.30281669e-01 1.22367978e+00 -9.12557662e-01 -3.34369898e-01
-2.47788206e-01 1.52051580e+00 -1.45576358e-01 3.98815155e-01
-9.53373671e-01 -6.88434899e-01 1.32292300e-01 3.27832758e-01
6.64517470e-03 -1.03940614e-01 -3.89038175e-01 -9.57490742e-01
2.00640231e-01 -1.10479271e+00 5.42102382e-02 -1.22366750e+00
-1.09435761e+00 4.76881593e-01 -1.94117606e-01 -1.46820772e+00
3.77283305e-01 -9.61620659e-02 -8.80252182e-01 8.14475596e-01
-1.38102829e+00 -1.16668153e+00 -6.14214182e-01 5.72931111e-01
8.46713066e-01 5.26504107e-02 3.30386400e-01 5.89555562e-01
-5.59516490e-01 4.27587509e-01 2.49587536e-01 2.17364859e-02
8.30707848e-01 -7.85836458e-01 1.53587665e-02 1.25712621e+00
-4.09531415e-01 1.48253962e-01 9.45411861e-01 -9.84844208e-01
-1.13016796e+00 -9.54459667e-01 3.63229841e-01 4.55335736e-01
1.70492575e-01 -1.01966925e-01 -1.06305909e+00 2.95359910e-01
5.62561750e-01 -2.82657519e-02 -1.22294985e-02 -7.99967229e-01
-7.18820244e-02 -1.38841555e-01 -1.57369947e+00 7.53520489e-01
7.39689648e-01 -2.61658043e-01 -2.82501072e-01 8.80524814e-02
4.34593141e-01 -3.52483600e-01 -6.31616831e-01 1.12317964e-01
5.58998883e-01 -1.07830083e+00 9.27816033e-01 1.52982593e-01
8.65036428e-01 -8.70877087e-01 2.12233365e-01 -1.17165828e+00
-1.33679837e-01 -7.43190885e-01 1.02762632e-01 1.47169995e+00
-8.76018871e-03 -4.21720207e-01 7.37137377e-01 2.40596041e-01
-9.51185822e-02 -4.95844960e-01 -5.75864553e-01 -4.25985932e-01
-3.42338592e-01 1.51145738e-02 3.26880813e-01 1.05181479e+00
-4.23717499e-01 1.05906343e-02 -8.42807114e-01 -2.83803530e-02
7.48113275e-01 2.64510047e-02 6.75943553e-01 -7.34241247e-01
-2.86487967e-01 -3.01569909e-01 -1.76828623e-01 -6.52481377e-01
-3.33855838e-01 -3.76446903e-01 4.93547246e-02 -1.58908033e+00
1.08320527e-01 -3.69710177e-01 1.59573585e-01 4.99598026e-01
-2.42974684e-01 7.69285798e-01 4.94927824e-01 4.29450542e-01
5.92571571e-02 5.85104108e-01 1.80202210e+00 -3.18594426e-01
-2.44143069e-01 -3.07506859e-01 -4.96063679e-01 5.17722607e-01
9.49229896e-01 -4.49607372e-01 -4.76435393e-01 -3.06106627e-01
4.27919999e-03 3.07945877e-01 5.16602099e-01 -9.27131116e-01
-6.62345439e-02 -2.96254694e-01 5.42479157e-01 -5.81246614e-01
2.10304007e-01 -8.97594988e-01 5.66779673e-01 8.18776131e-01
-1.12135403e-01 2.88098812e-01 1.97084129e-01 4.01922226e-01
-4.46744710e-01 -4.74287271e-01 1.21524751e+00 -9.46584344e-02
-6.07871532e-01 3.33434495e-04 -2.91273504e-01 -1.52858555e-01
1.11426318e+00 -5.82378626e-01 -3.21791261e-01 -5.37494242e-01
-4.66252089e-01 -7.47554377e-02 8.18913460e-01 2.40828797e-01
1.08867216e+00 -1.40540314e+00 -8.28314900e-01 3.06719959e-01
-2.98268348e-01 5.86084388e-02 6.99557126e-01 9.23268318e-01
-1.35022318e+00 -4.06686842e-01 -6.50437474e-01 -9.42951068e-02
-1.57283175e+00 5.89651406e-01 7.49025270e-02 1.64372660e-02
-8.33282590e-01 4.11761314e-01 2.65614271e-01 2.02522755e-01
8.79024416e-02 -3.94947492e-02 -4.06109728e-02 -6.54186308e-02
6.08899236e-01 5.20025551e-01 -1.60150617e-01 -5.48410654e-01
7.12797940e-02 6.91526949e-01 -1.12997577e-01 -1.13119213e-02
1.18029356e+00 -3.65507394e-01 -5.14845192e-01 3.74652818e-02
8.94873202e-01 3.79271120e-01 -1.49613225e+00 1.38112515e-01
-4.91341501e-01 -8.33064437e-01 -1.39509793e-02 -7.70917833e-01
-1.51620328e+00 6.95835531e-01 6.06556296e-01 1.23818450e-01
1.57017291e+00 -5.30174315e-01 1.15124440e+00 -3.30113888e-01
-2.71738321e-02 -1.05578744e+00 3.04260403e-01 -9.07911807e-02
1.21413028e+00 -6.61473572e-01 2.23314479e-01 -5.83899677e-01
-6.66106522e-01 1.31598473e+00 6.14763439e-01 -4.10426259e-01
7.33364373e-02 2.99676210e-01 8.92474726e-02 9.81576890e-02
-2.20413491e-01 2.49728724e-01 8.87466967e-02 5.84313869e-01
1.68003067e-01 -2.47634158e-01 -8.67723286e-01 5.41660748e-02
2.43656635e-01 -5.45634814e-02 7.86737561e-01 9.70792532e-01
-4.03846949e-01 -1.07922220e+00 -1.00965559e+00 1.17455542e-01
-3.02683800e-01 -1.56473920e-01 -3.83486956e-01 6.91092670e-01
3.57573867e-01 8.90370369e-01 -1.82360813e-01 -4.34368514e-02
3.59697998e-01 -3.77765954e-01 5.20273447e-01 -1.06016926e-01
-4.48669255e-01 2.04405814e-01 -4.12089936e-02 -3.70846689e-01
-2.49462709e-01 -3.74057174e-01 -1.13200176e+00 -5.08329749e-01
-1.19048357e-01 -5.10819480e-02 4.31745857e-01 4.25286978e-01
1.70921162e-01 5.58730721e-01 6.53977334e-01 -8.65685403e-01
-3.16656940e-02 -7.52044797e-01 -9.02305663e-01 7.55875707e-01
1.91849142e-01 -4.29953456e-01 -2.28538811e-01 3.90566736e-01] | [11.23507308959961, -1.4116097688674927] |
11b9746b-278a-46a8-a9f1-2c1490c5b1d7 | a-model-driven-deep-neural-network-for-single | 2005.01333 | null | https://arxiv.org/abs/2005.01333v1 | https://arxiv.org/pdf/2005.01333v1.pdf | A Model-driven Deep Neural Network for Single Image Rain Removal | Deep learning (DL) methods have achieved state-of-the-art performance in the task of single image rain removal. Most of current DL architectures, however, are still lack of sufficient interpretability and not fully integrated with physical structures inside general rain streaks. To this issue, in this paper, we propose a model-driven deep neural network for the task, with fully interpretable network structures. Specifically, based on the convolutional dictionary learning mechanism for representing rain, we propose a novel single image deraining model and utilize the proximal gradient descent technique to design an iterative algorithm only containing simple operators for solving the model. Such a simple implementation scheme facilitates us to unfold it into a new deep network architecture, called rain convolutional dictionary network (RCDNet), with almost every network module one-to-one corresponding to each operation involved in the algorithm. By end-to-end training the proposed RCDNet, all the rain kernels and proximal operators can be automatically extracted, faithfully characterizing the features of both rain and clean background layers, and thus naturally lead to its better deraining performance, especially in real scenarios. Comprehensive experiments substantiate the superiority of the proposed network, especially its well generality to diverse testing scenarios and good interpretability for all its modules, as compared with state-of-the-arts both visually and quantitatively. The source codes are available at \url{https://github.com/hongwang01/RCDNet}. | ['Qian Zhao', 'Qi Xie', 'Hong Wang', 'Deyu Meng'] | 2020-05-04 | a-model-driven-deep-neural-network-for-single-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Wang_A_Model-Driven_Deep_Neural_Network_for_Single_Image_Rain_Removal_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Wang_A_Model-Driven_Deep_Neural_Network_for_Single_Image_Rain_Removal_CVPR_2020_paper.pdf | cvpr-2020-6 | ['single-image-deraining'] | ['computer-vision'] | [-3.20066988e-01 -2.17931375e-01 2.98728883e-01 -5.49141824e-01
-1.62437513e-01 -1.82424277e-01 1.62714839e-01 -2.35742554e-01
-1.09801963e-01 6.34352386e-01 2.80430298e-02 -4.59247559e-01
-1.44305483e-01 -6.90730691e-01 -6.46533251e-01 -9.67432857e-01
-9.74675789e-02 7.45328292e-02 -1.08241521e-01 -4.73023176e-01
-1.52091473e-01 5.85067034e-01 -1.32737327e+00 -5.11990078e-02
1.31748903e+00 8.55005026e-01 3.84752661e-01 4.67147827e-01
-1.37366131e-01 1.12746751e+00 -2.22599000e-01 -2.60638028e-01
2.09713206e-01 -4.81096804e-01 -2.62816429e-01 6.98604360e-02
4.51377153e-01 -6.66239917e-01 -6.07491910e-01 1.05172849e+00
6.09666944e-01 -4.58947569e-02 3.02383304e-01 -6.38779342e-01
-7.40811765e-01 3.37531805e-01 -6.63312674e-01 3.29726785e-01
-5.83995506e-02 2.09635869e-01 9.75027621e-01 -1.33888721e+00
1.69418737e-01 1.11232007e+00 7.46069610e-01 2.51678914e-01
-1.00170386e+00 -6.36207581e-01 3.80320460e-01 1.61969408e-01
-1.37372077e+00 -3.73012692e-01 7.61196136e-01 -2.97387540e-01
4.05717432e-01 3.63754630e-01 8.89864743e-01 7.75600493e-01
4.66499887e-02 8.81469429e-01 1.26971185e+00 -1.51301011e-01
7.91300610e-02 2.37048065e-04 3.56017649e-01 1.02283192e+00
6.34063065e-01 1.05470225e-01 -4.51713577e-02 3.21108162e-01
9.03366506e-01 2.95387775e-01 -7.03516126e-01 -1.79197669e-01
-7.28544831e-01 8.40320349e-01 8.82421076e-01 2.07543105e-01
-5.16161740e-01 -8.62053707e-02 3.75903845e-02 2.53190637e-01
6.64845765e-01 4.35049012e-02 -2.38265172e-01 4.33151960e-01
-9.93823767e-01 3.79026145e-01 7.60397196e-01 6.12486839e-01
1.07307291e+00 5.09378254e-01 -2.52430942e-02 6.40915751e-01
5.00112832e-01 8.83738458e-01 6.53010905e-02 -5.45154810e-01
3.75456482e-01 5.68483829e-01 1.08811297e-01 -1.04707134e+00
-5.87202370e-01 -9.35050189e-01 -1.44958866e+00 4.42195207e-01
1.52306585e-02 -2.70583481e-01 -1.08229446e+00 1.40962744e+00
5.09559691e-01 4.01516140e-01 1.89666107e-01 1.28488421e+00
1.05464637e+00 7.53536284e-01 -9.65793952e-02 -2.82609463e-01
1.20409560e+00 -9.60677266e-01 -7.83172250e-01 -3.96350890e-01
3.23112041e-01 -7.02329159e-01 1.06892836e+00 4.68719065e-01
-7.47711182e-01 -7.78772235e-01 -1.15627706e+00 -4.77519631e-03
-8.64255875e-02 3.64040464e-01 7.60512888e-01 2.32739344e-01
-7.45685041e-01 5.62477469e-01 -7.20253170e-01 -1.97643667e-01
5.09326994e-01 8.36079717e-02 -1.58976659e-01 -3.01819682e-01
-1.15698481e+00 7.74979770e-01 3.40846807e-01 1.11501253e+00
-1.13922417e+00 -6.10455096e-01 -7.34329402e-01 -3.46331522e-02
2.76499540e-02 -7.81377256e-01 9.74744737e-01 -1.06988561e+00
-1.25485694e+00 5.77365756e-01 -1.99006155e-01 -3.95331144e-01
2.56199360e-01 -7.51879692e-01 -5.82030952e-01 1.25663355e-01
-2.22516611e-01 8.50327015e-02 1.14994967e+00 -1.75658441e+00
-5.44957280e-01 -2.44653434e-01 3.87594074e-01 2.43221506e-01
-2.48711959e-01 -2.87986457e-01 -3.56202155e-01 -8.16869199e-01
1.83862284e-01 -7.14151680e-01 -3.42775196e-01 8.55404437e-02
-3.14978838e-01 2.68661201e-01 8.39183629e-01 -9.95904922e-01
1.50338697e+00 -2.17046857e+00 2.62725890e-01 -3.43930498e-02
5.83885908e-01 7.73980200e-01 -4.53067385e-02 2.32635289e-01
-1.52740434e-01 -2.10467845e-01 -6.07634187e-01 -4.39215839e-01
-1.48341745e-01 4.90294993e-01 -4.05426890e-01 7.15561807e-01
3.64509225e-01 7.45591760e-01 -6.63948536e-01 -3.36800367e-01
3.87166202e-01 5.34827530e-01 -2.04281822e-01 7.30586052e-01
-2.15200827e-01 5.87437749e-01 -5.81291437e-01 7.39665210e-01
1.33430088e+00 1.23984283e-02 1.75067913e-02 -4.50981230e-01
-1.35803685e-01 -2.60313869e-01 -1.34564781e+00 1.28262007e+00
-7.08597481e-01 4.83500481e-01 4.62170601e-01 -9.81353343e-01
9.68808651e-01 1.98921993e-01 -1.47370063e-02 -7.62687147e-01
4.02472876e-02 3.57197851e-01 -5.38533181e-02 -9.12257850e-01
6.05954155e-02 -2.25990340e-01 5.50806165e-01 1.07792668e-01
-6.29080310e-02 1.38508096e-01 -5.86780757e-02 1.00393616e-01
5.92301488e-01 1.73873648e-01 1.57559842e-01 -3.03021759e-01
6.74125135e-01 -2.06255898e-01 6.16973341e-01 6.62220895e-01
4.96913083e-02 6.77762926e-01 2.62257308e-02 -8.19286585e-01
-8.55306506e-01 -9.18090999e-01 -3.03993076e-01 7.51328409e-01
4.54004467e-01 -1.15384227e-02 -6.97540760e-01 -4.39290047e-01
-3.22724357e-02 5.78723192e-01 -6.96309566e-01 7.44824409e-02
-7.75735617e-01 -1.11698151e+00 4.32513297e-01 2.92947114e-01
9.95696604e-01 -1.12849104e+00 -3.65735829e-01 1.03271231e-01
-5.53627796e-02 -1.07507324e+00 5.28279878e-03 1.80610314e-01
-9.49992359e-01 -9.76461112e-01 -5.57902217e-01 -7.39092350e-01
6.78653896e-01 6.02802932e-01 1.24804914e+00 4.76000726e-01
-2.66363055e-01 -6.17878363e-02 -5.62626600e-01 -4.06634867e-01
-2.14914102e-02 -2.34368905e-01 -1.21304274e-01 3.12866360e-01
1.45915791e-01 -8.29669237e-01 -9.55754042e-01 -7.76548609e-02
-1.17982328e+00 1.33369192e-01 9.66029763e-01 8.36043537e-01
4.09924150e-01 -7.66350031e-02 2.54832387e-01 -1.00906134e+00
3.73955071e-01 -5.78648448e-01 -5.10057986e-01 1.14277638e-01
-4.64070678e-01 -5.32384254e-02 8.33980203e-01 1.12523779e-01
-1.34795654e+00 -6.39300346e-02 -3.81008595e-01 -4.97713983e-01
-2.13322401e-01 7.09697962e-01 -3.36636037e-01 -2.67927438e-01
5.50086498e-01 4.62513208e-01 -1.66698694e-01 -7.40529954e-01
4.76371348e-01 4.54084158e-01 3.78824234e-01 -3.20326626e-01
1.28520310e+00 6.96486533e-01 -9.56849158e-02 -8.53483737e-01
-1.32642937e+00 -3.51123363e-01 -6.31877661e-01 -1.68187067e-01
8.11607957e-01 -1.40348971e+00 -3.83618802e-01 7.75204301e-01
-1.14727128e+00 -3.32934231e-01 1.55015532e-02 4.19500202e-01
3.15481499e-02 4.75260854e-01 -7.12603867e-01 -9.19410944e-01
-6.77242279e-01 -9.29880261e-01 8.53123724e-01 5.13862967e-01
4.88047361e-01 -1.14262331e+00 1.05643064e-01 1.72770143e-01
5.71159840e-01 3.34774047e-01 7.84668684e-01 -1.03443481e-01
-7.19017327e-01 3.43642943e-02 -4.03271616e-01 7.55868614e-01
1.01288013e-01 2.16505062e-02 -1.20206273e+00 -5.41261613e-01
3.94268513e-01 -2.37365052e-01 1.20551133e+00 3.17940474e-01
9.99177873e-01 -3.97543669e-01 8.23368356e-02 1.12523699e+00
1.84912026e+00 -1.52648896e-01 8.53256106e-01 4.49579418e-01
1.01515305e+00 3.24510843e-01 6.50417447e-01 5.63149154e-01
2.55425513e-01 3.72811317e-01 9.44965959e-01 -8.20450842e-01
-1.41116351e-01 5.63322753e-02 3.45242172e-01 1.03018665e+00
-3.70057136e-01 -3.29591751e-01 -5.44008493e-01 5.26029468e-01
-1.81553745e+00 -8.07640374e-01 -3.42273831e-01 1.90137410e+00
5.73574722e-01 6.96127042e-02 -3.30886573e-01 -1.14698611e-01
4.03351754e-01 6.38133943e-01 -4.79600012e-01 -2.24474043e-01
-2.92057931e-01 3.86332244e-01 3.29551071e-01 6.27007782e-01
-1.22481465e+00 8.78940046e-01 4.84498644e+00 7.63529539e-01
-1.19171321e+00 8.20942894e-02 4.81655449e-01 2.62721449e-01
-3.05502266e-01 9.50137749e-02 -7.73963034e-01 3.24802220e-01
4.93285835e-01 3.03918928e-01 5.04686594e-01 5.96623242e-01
7.66766548e-01 1.25570819e-01 -5.58185756e-01 7.68006623e-01
-1.01674765e-01 -1.12121785e+00 2.84943938e-01 -3.36751342e-01
6.42410815e-01 7.10645467e-02 -2.47904975e-02 2.96595901e-01
7.67429546e-02 -1.03123450e+00 6.16032362e-01 7.90412247e-01
5.27474523e-01 -5.62714934e-01 9.88715887e-01 2.85990924e-01
-1.28932714e+00 -7.95888975e-02 -5.65224290e-01 -1.47549182e-01
-2.17127576e-02 1.12691212e+00 -2.79388338e-01 9.87528861e-01
8.63387942e-01 8.97933066e-01 -5.19778669e-01 1.04059291e+00
-5.88809133e-01 8.78865182e-01 -1.78945884e-01 3.74300897e-01
3.30466539e-01 -5.76020241e-01 5.39409101e-01 1.58767474e+00
2.55340666e-01 3.07129979e-01 8.76416713e-02 7.02115536e-01
7.16452971e-02 -4.22347598e-02 -5.76201558e-01 3.73059332e-01
2.88740337e-01 1.46474743e+00 -3.17488164e-01 -3.11259180e-01
-5.29446900e-01 9.91881490e-01 3.38311791e-01 7.03206003e-01
-1.07082248e+00 -6.07046723e-01 9.49755192e-01 -3.51342633e-02
6.14035249e-01 -3.32573235e-01 -3.00049096e-01 -1.30287111e+00
2.52367765e-01 -1.09811318e+00 -1.95143986e-02 -7.02643633e-01
-1.29154968e+00 9.27992225e-01 -2.17755497e-01 -1.36502111e+00
3.80290151e-01 -7.02342629e-01 -1.02478278e+00 1.01365042e+00
-2.08384919e+00 -1.35973990e+00 -1.11108375e+00 6.78955317e-01
4.04766351e-01 7.91329518e-02 7.03103960e-01 5.81041396e-01
-1.00556254e+00 3.73770803e-01 4.43063915e-01 1.05208807e-01
5.90274334e-01 -1.24947238e+00 2.23004639e-01 1.39150274e+00
5.57414629e-02 6.08686745e-01 7.64502525e-01 -2.83884883e-01
-1.47017705e+00 -1.40781152e+00 4.47591543e-01 4.19017933e-02
6.17567599e-01 -2.88854748e-01 -1.26161742e+00 5.69680154e-01
2.50121772e-01 2.55711347e-01 2.24137887e-01 5.71227819e-02
-3.21538478e-01 -6.31232679e-01 -8.04873466e-01 4.60690916e-01
9.37005162e-01 -3.53295147e-01 -4.62185025e-01 4.96950895e-01
7.85119355e-01 -6.15454674e-01 -6.16761386e-01 5.54709792e-01
3.10319096e-01 -1.34701312e+00 8.95123363e-01 -3.61233234e-01
5.93264580e-01 -6.66788459e-01 -1.40655026e-01 -1.20743954e+00
-6.07709765e-01 -3.09611201e-01 -4.31575090e-01 1.16734493e+00
1.81490436e-01 -6.45752251e-01 4.35343444e-01 -1.23805821e-01
-4.65812832e-01 -9.94084716e-01 -6.47222638e-01 -3.52323681e-01
-1.85170516e-01 -2.53023237e-01 4.55582589e-01 8.87224436e-01
-7.49913871e-01 3.18876803e-01 -7.02289343e-01 8.83800924e-01
8.06099772e-01 5.66319346e-01 7.64900982e-01 -1.02641201e+00
-4.12341952e-01 -1.37828603e-01 -1.54678941e-01 -1.31401098e+00
-1.41028956e-01 -5.81378222e-01 2.55131245e-01 -1.74861991e+00
4.08615842e-02 -4.53783035e-01 -3.74929070e-01 4.72255945e-01
-5.72781384e-01 2.33703762e-01 1.57646954e-01 4.49392319e-01
-4.76221979e-01 9.33616877e-01 1.43989980e+00 -9.96679142e-02
-1.04909971e-01 6.22822121e-02 -7.51342773e-01 8.58233809e-01
8.29744041e-01 -2.91315675e-01 -2.85068870e-01 -1.04760826e+00
-6.45513758e-02 -1.92877516e-01 6.24314249e-01 -1.13124621e+00
-1.05882160e-01 -1.57224000e-01 3.51254314e-01 -2.31782600e-01
2.15623304e-01 -8.49526942e-01 1.19222119e-01 4.51741874e-01
2.12424979e-01 -1.37440324e-01 2.91656047e-01 5.93370616e-01
-5.46436191e-01 -1.40073016e-01 9.90122557e-01 -1.07996114e-01
-9.16319549e-01 5.31114817e-01 -1.63028121e-01 -1.47381008e-01
6.67658627e-01 3.00535485e-02 -3.23212683e-01 -2.61695564e-01
-7.87584782e-01 2.35878780e-01 1.58104539e-01 4.19762777e-03
6.93539798e-01 -1.01978731e+00 -1.03277183e+00 3.06355327e-01
-1.35902002e-01 3.43877524e-01 6.26940250e-01 9.15718615e-01
-1.05131233e+00 3.28516327e-02 -6.93451539e-02 -3.82503062e-01
-9.32751834e-01 3.22007179e-01 6.06428862e-01 -2.41097957e-01
-9.35443163e-01 8.08003247e-01 6.84229732e-01 -2.45001718e-01
2.08645776e-01 -4.09827560e-01 -3.22625428e-01 -1.46568626e-01
4.96905535e-01 1.02078073e-01 1.47532403e-01 -3.42408508e-01
-1.41942531e-01 6.44735813e-01 6.89850328e-03 5.54128170e-01
1.60680497e+00 -1.74590752e-01 -3.91879559e-01 2.41042480e-01
9.01897788e-01 -1.05560338e-02 -1.40225673e+00 -3.14717650e-01
-4.33886170e-01 -4.13101047e-01 1.99478611e-01 -5.86028457e-01
-1.54602420e+00 1.10505092e+00 1.01710832e+00 5.30242287e-02
1.48069406e+00 -4.81284440e-01 9.22801614e-01 5.36908150e-01
4.92109880e-02 -4.13714528e-01 -1.24576673e-01 6.15670204e-01
8.90723705e-01 -1.37258315e+00 2.42872760e-01 -2.65743732e-01
-5.89517713e-01 1.20334458e+00 6.16724849e-01 -3.69003594e-01
8.66794646e-01 1.01546988e-01 4.05087978e-01 -4.13067967e-01
-3.09846342e-01 -3.16926241e-01 -1.05613783e-01 3.60379934e-01
3.39725018e-01 8.31013471e-02 -2.34920532e-01 5.77997088e-01
1.90999344e-01 3.28899473e-02 2.83337653e-01 7.48022795e-01
-7.91413903e-01 -7.07310200e-01 -4.01930779e-01 2.75659502e-01
-9.27389339e-02 -4.60313827e-01 1.57665044e-01 8.01747859e-01
4.62963641e-01 8.06205988e-01 -2.54515499e-01 -3.27243268e-01
5.21399200e-01 -5.22341132e-01 2.42157266e-01 -3.31824809e-01
-3.70130092e-01 -6.56905174e-02 -8.75224397e-02 -3.52872252e-01
-6.66748703e-01 -3.30297291e-01 -1.06333363e+00 -4.15753245e-01
-1.98709428e-01 2.29943395e-01 2.47430891e-01 1.00578463e+00
2.52056748e-01 5.42846680e-01 8.48568380e-01 -1.06858981e+00
-4.13680375e-01 -1.04581082e+00 -9.25464392e-01 2.37519443e-01
6.57543242e-01 -4.92688984e-01 -4.51249301e-01 3.52780409e-02] | [10.920653343200684, -3.2582292556762695] |
6f1f0e4b-cf3c-4b8c-b6e8-e1783b3659a0 | oriented-objects-as-pairs-of-middle-lines | 1912.10694 | null | https://arxiv.org/abs/1912.10694v3 | https://arxiv.org/pdf/1912.10694v3.pdf | Oriented Objects as pairs of Middle Lines | The detection of oriented objects is frequently appeared in the field of natural scene text detection as well as object detection in aerial images. Traditional detectors for oriented objects are common to rotate anchors on the basis of the RCNN frameworks, which will multiple the number of anchors with a variety of angles, coupled with rotating NMS algorithm, the computational complexities of these models are greatly increased. In this paper, we propose a novel model named Oriented Objects Detection Network O^2-DNet to detect oriented objects by predicting a pair of middle lines inside each target. O^2-DNet is an one-stage, anchor-free and NMS-free model. The target line segments of our model are defined as two corresponding middle lines of original rotating bounding box annotations which can be transformed directly instead of additional manual tagging. Experiments show that our O^2-DNet achieves excellent performance on ICDAR 2015 and DOTA datasets. It is noteworthy that the objects in COCO can be regard as a special form of oriented objects with an angle of 90 degrees. O^2-DNet can still achieve competitive results in these general natural object detection datasets. | ['Hao-Ran Wei', 'Yue Zhang', 'Xian Sun', 'Hongqi Wang', 'Zhonghan Chang', 'Hao Li'] | 2019-12-23 | null | null | null | null | ['object-detection-in-aerial-images'] | ['computer-vision'] | [ 2.47196525e-01 1.38967127e-01 -2.36343041e-01 -2.02546045e-01
-1.54554799e-01 -6.48977935e-01 4.09021109e-01 8.48545320e-03
-3.30127418e-01 -1.75549220e-02 -1.52459487e-01 -8.92437026e-02
6.55741543e-02 -6.77599847e-01 -4.53264743e-01 -5.48527539e-01
-1.08046733e-01 3.00571501e-01 9.45007205e-01 -2.70784795e-01
1.50701046e-01 6.32900655e-01 -1.09712589e+00 2.51157850e-01
4.97122020e-01 1.09152412e+00 1.19190618e-01 8.30744147e-01
-1.25084341e-01 5.09309947e-01 -7.56010652e-01 -4.08144176e-01
4.72068518e-01 1.75231084e-01 -6.76482737e-01 3.31599683e-01
5.95043898e-01 -5.69068789e-01 -3.77565175e-01 1.17378247e+00
5.27032673e-01 -2.84103721e-01 8.26417804e-01 -1.30665946e+00
-7.58825600e-01 7.81291664e-01 -1.08489823e+00 3.39869946e-01
3.90365012e-02 -2.32259303e-01 8.58422399e-01 -1.17505372e+00
4.07408893e-01 1.33191729e+00 7.22650945e-01 4.12611842e-01
-8.10813367e-01 -4.98993158e-01 6.21139586e-01 1.13487311e-01
-1.67591298e+00 -5.16011380e-02 6.47457659e-01 -5.34206212e-01
4.81581956e-01 4.46633130e-01 6.27000511e-01 8.45609367e-01
-7.89725482e-02 1.21035957e+00 4.66800630e-01 -5.66090405e-01
-8.90717357e-02 -2.79533304e-02 5.04819572e-01 8.49317193e-01
4.10672098e-01 -2.78773397e-01 -2.89149940e-01 4.84754033e-02
5.41953623e-01 5.11620402e-01 -1.53833523e-01 -4.51715857e-01
-1.43824291e+00 6.19585812e-01 9.35397685e-01 1.03636362e-01
6.35511279e-02 3.71330157e-02 3.34123552e-01 -3.60947430e-01
4.75645691e-01 2.69051731e-01 -5.32043636e-01 7.11873353e-01
-7.61613667e-01 1.45875633e-01 4.15668696e-01 1.57225275e+00
1.03734590e-01 -3.05755157e-03 -5.06334662e-01 8.00122321e-01
3.56540293e-01 7.34513164e-01 3.86830479e-01 -2.41430134e-01
6.12368107e-01 1.06750166e+00 2.31058180e-01 -1.45162451e+00
-7.78374970e-01 -5.99403858e-01 -9.22504365e-01 -1.09115355e-01
4.11463559e-01 -2.06250399e-01 -1.07363880e+00 1.19898200e+00
6.53223515e-01 -2.35869437e-01 -2.66753230e-02 1.01521730e+00
8.86260986e-01 6.97555602e-01 -5.39522022e-02 1.38128221e-01
1.75837243e+00 -1.11272633e+00 -5.69945693e-01 -3.74097019e-01
8.21574211e-01 -8.98718417e-01 7.85632074e-01 3.79028976e-01
-6.71020389e-01 -6.47235274e-01 -1.11772919e+00 -3.49227265e-02
-6.61197603e-01 9.09573257e-01 5.68191051e-01 3.45477939e-01
-6.09216988e-01 2.51531094e-01 -6.35456622e-01 -4.11521554e-01
6.75830245e-01 4.01151717e-01 -8.89729559e-02 1.05288804e-01
-7.81533241e-01 5.39534748e-01 7.42785513e-01 6.16112292e-01
-6.33040130e-01 5.47791598e-03 -5.60263574e-01 -1.01105064e-01
7.32332170e-01 -2.88462400e-01 1.12974656e+00 -8.40051472e-01
-7.89428473e-01 7.85361409e-01 -3.80651914e-02 -4.59022880e-01
7.27447748e-01 -5.76570153e-01 -5.85714161e-01 5.34658805e-02
4.43432629e-01 8.86982858e-01 9.91289914e-01 -1.05458701e+00
-1.18783665e+00 -5.59368253e-01 7.43277073e-02 3.45733851e-01
-6.54134810e-01 1.97030365e-01 -8.05387676e-01 -8.05097401e-01
6.64068520e-01 -1.14023316e+00 -3.03197771e-01 3.85779351e-01
-1.01348424e+00 -6.91760302e-01 1.37437809e+00 -4.08321142e-01
1.40037644e+00 -1.89491689e+00 -1.51478544e-01 -9.81063619e-02
4.77755278e-01 3.44155431e-01 3.93078923e-02 -1.22275092e-01
-6.78129047e-02 2.06610844e-01 7.17514977e-02 5.12123480e-02
-1.51131526e-01 -3.79527599e-01 -4.91925210e-01 5.51812053e-01
3.65270436e-01 5.98077416e-01 -7.95030296e-01 -8.00412655e-01
-3.14242020e-02 1.24279268e-01 -2.13317767e-01 1.00513436e-01
-6.83188140e-02 -2.41482735e-01 -7.00944960e-01 1.07430398e+00
8.39792013e-01 -2.88211077e-01 -5.04010133e-02 -3.90121430e-01
-2.90732116e-01 -1.97414711e-01 -1.27523136e+00 1.23646533e+00
2.88382828e-01 7.87521839e-01 -3.11532527e-01 -7.30589509e-01
1.13147223e+00 9.37545002e-02 2.25463018e-01 -1.38293028e-01
1.26483347e-02 -6.82131201e-02 2.21002176e-01 -6.23378396e-01
6.80502892e-01 5.06309748e-01 -1.95025519e-01 -2.25731775e-01
-3.64250660e-01 -2.32867990e-02 2.59770632e-01 1.68249071e-01
8.39840233e-01 2.29002535e-01 5.12906671e-01 -2.59942383e-01
3.90526235e-01 9.96788889e-02 5.40118158e-01 9.32572901e-01
-2.67534196e-01 7.05843031e-01 4.44605738e-01 -8.21598411e-01
-1.08500588e+00 -7.13459611e-01 -3.15058827e-01 1.34501791e+00
4.28789109e-01 -5.11772633e-01 -7.90150404e-01 -1.14919877e+00
-2.69238085e-01 2.87004203e-01 -5.13663411e-01 6.65811971e-02
-7.35200047e-01 -9.83592629e-01 5.11185884e-01 8.95628452e-01
6.74831688e-01 -9.64408815e-01 -5.90041399e-01 -2.05794591e-02
-6.41430095e-02 -1.18366671e+00 -7.54477382e-01 5.67232788e-01
-7.75999665e-01 -1.15024424e+00 -8.23301911e-01 -1.18947279e+00
1.07994044e+00 8.20451796e-01 8.71493101e-01 7.09668174e-02
-3.41936260e-01 -2.81691581e-01 -4.68631685e-01 -9.56353009e-01
1.18225098e-01 3.28096926e-01 1.86992660e-01 -9.51575935e-02
5.87658823e-01 1.07878737e-01 -7.45754778e-01 7.88555622e-01
-8.63282144e-01 2.04344615e-01 6.04072094e-01 4.64850307e-01
5.39584219e-01 -8.15576240e-02 3.48545700e-01 -8.79571080e-01
2.12005019e-01 -2.06125870e-01 -9.17638659e-01 2.54860580e-01
-2.78767556e-01 -1.86266363e-01 7.92958319e-01 -7.23528087e-01
-3.63392740e-01 6.10156119e-01 1.97682992e-01 -4.23771471e-01
-3.26638937e-01 6.20043501e-02 -2.39805639e-01 1.62263170e-01
7.87518859e-01 2.54664514e-02 -9.56281781e-01 -3.26656044e-01
2.37815723e-01 1.14315021e+00 4.71118510e-01 -8.17730278e-02
1.02832258e+00 5.85775495e-01 4.08777483e-02 -9.28372800e-01
-1.29334390e+00 -7.78283656e-01 -9.60675299e-01 -1.39197752e-01
8.41748416e-01 -8.78795564e-01 -6.34174764e-01 5.31784296e-01
-1.53873825e+00 1.52030870e-01 -2.88657565e-02 2.59863347e-01
1.00289322e-01 5.01382351e-01 -2.53526986e-01 -8.22235584e-01
-5.29850483e-01 -1.08179820e+00 1.46858275e+00 4.68065739e-01
4.86588962e-02 -4.69373047e-01 -4.08488631e-01 2.98051592e-02
-1.39063507e-01 2.79095501e-01 6.23429537e-01 -1.05518734e+00
-5.26083231e-01 -7.93438554e-01 -4.09013599e-01 1.81735620e-01
-4.11485992e-02 2.12542057e-01 -8.00299466e-01 -1.68541282e-01
-4.04299825e-01 -6.76728785e-02 7.03916311e-01 3.02405655e-01
1.54104722e+00 -4.37110275e-01 -8.00073743e-01 5.64929247e-01
9.81809974e-01 3.86822909e-01 2.61236846e-01 4.52398121e-01
1.05419862e+00 3.81182909e-01 8.59210074e-01 3.00323367e-01
1.10666968e-01 8.78942370e-01 6.09047651e-01 -3.94185990e-01
-2.12231576e-02 -1.61405802e-01 2.68117726e-01 6.45616233e-01
-2.59863250e-02 -6.22475564e-01 -1.04226649e+00 4.12860334e-01
-1.87254143e+00 -7.31341600e-01 -6.28513336e-01 2.02544093e+00
4.04872835e-01 5.44956744e-01 2.63912499e-01 2.61015654e-01
1.30480516e+00 1.10694207e-01 -6.30594671e-01 2.13246003e-01
-3.02643508e-01 -5.29852271e-01 7.96104968e-01 -1.62290841e-01
-1.80892420e+00 1.05617619e+00 6.19812489e+00 8.52517962e-01
-9.07374322e-01 -2.58498847e-01 5.28743744e-01 2.28067487e-01
6.95297062e-01 -4.93467301e-02 -1.58881843e+00 4.33878571e-01
1.42702535e-01 3.43164533e-01 -1.04624048e-01 1.55965412e+00
1.38942599e-01 3.14682275e-01 -1.05665100e+00 8.94600034e-01
2.46815324e-01 -9.62543011e-01 1.42969742e-01 -1.99126095e-01
6.28992319e-01 -1.55375585e-01 1.35472917e-04 1.14124477e-01
1.90695107e-01 -7.01563239e-01 9.51466203e-01 1.84425637e-01
6.96047127e-01 -4.81455594e-01 7.53416598e-01 5.09406328e-01
-1.47649407e+00 -2.23662704e-01 -7.86548257e-01 3.01020712e-01
-2.28686184e-01 3.50062221e-01 -1.12061977e+00 3.46840054e-01
9.29209411e-01 7.40919650e-01 -8.93437147e-01 1.30804420e+00
-3.24007869e-01 5.81623137e-01 -4.97795194e-01 -4.02040660e-01
2.78195679e-01 -2.04997417e-02 6.60192668e-01 1.29382145e+00
2.34012291e-01 -7.56521448e-02 4.53771502e-01 4.75900620e-01
-2.56175637e-01 2.14049801e-01 -6.89622343e-01 2.43892074e-01
4.53096181e-01 1.77896547e+00 -1.34285200e+00 -5.28859675e-01
-1.70881554e-01 8.55302274e-01 2.45110378e-01 -6.94652572e-02
-1.12680745e+00 -8.69922101e-01 -1.58946514e-01 2.34504238e-01
5.36172509e-01 -1.10199139e-01 8.54813606e-02 -1.15998280e+00
1.23397976e-01 -7.75409818e-01 4.14728701e-01 -9.23854709e-01
-1.12012684e+00 7.30251670e-01 -4.70251441e-02 -1.69965851e+00
4.90520030e-01 -1.03830087e+00 -5.27513027e-01 3.05389315e-01
-1.15940177e+00 -1.34122086e+00 -6.80608928e-01 3.22907269e-01
9.69077885e-01 -6.19402230e-02 5.33186316e-01 2.70877630e-02
-1.05363953e+00 4.31885660e-01 1.70024589e-01 7.84632087e-01
7.35053301e-01 -1.28775191e+00 4.45148528e-01 1.11640227e+00
2.74901420e-01 5.03478229e-01 5.82557380e-01 -5.91194868e-01
-9.06492114e-01 -1.67070448e+00 4.32687551e-01 -6.01340115e-01
4.16364610e-01 -5.89722633e-01 -6.39692426e-01 6.97739959e-01
-9.18121636e-02 5.28551042e-01 2.89050102e-01 -1.52546704e-01
-3.01254213e-01 -1.26609981e-01 -7.06732690e-01 7.97201395e-01
1.32533026e+00 5.36130033e-02 -4.95836616e-01 1.06322718e+00
8.06116998e-01 -5.11875808e-01 -3.57384950e-01 5.26205003e-01
3.59300345e-01 -6.81817293e-01 9.75764871e-01 -7.34763503e-01
3.24186653e-01 -9.37308073e-01 -1.26910493e-01 -6.41433299e-01
-2.80868530e-01 -4.18218970e-01 -2.52753019e-01 1.08498228e+00
2.68898040e-01 -2.31022358e-01 8.07272911e-01 -1.71530366e-01
-3.04215997e-01 -6.94621682e-01 -5.76898396e-01 -6.85273409e-01
-3.70875239e-01 -1.18342631e-01 4.10459787e-01 6.21256769e-01
-3.00192833e-01 7.37814724e-01 -3.14534366e-01 7.72780180e-01
4.55949634e-01 9.11426619e-02 9.42226946e-01 -1.38903654e+00
-8.71469080e-03 -3.29685122e-01 -6.76644504e-01 -1.73325861e+00
-3.42550159e-01 -8.00924659e-01 1.97820678e-01 -1.44449508e+00
2.27690443e-01 -3.97206187e-01 3.37845944e-02 6.20844781e-01
-2.71915019e-01 3.71511370e-01 1.78357184e-01 5.29001176e-01
-1.05215085e+00 2.78502554e-01 9.93224800e-01 -3.83809268e-01
-3.06071371e-01 2.59682238e-01 -4.21817958e-01 1.58143604e+00
6.07146561e-01 -7.00957954e-01 -3.51043679e-02 -5.32053173e-01
1.48011103e-01 -4.00312960e-01 2.77839363e-01 -1.07197225e+00
1.97773099e-01 -7.77935013e-02 1.02493417e+00 -1.30499566e+00
1.77498702e-02 -1.00538135e+00 -5.75871885e-01 5.49975336e-01
-4.38955367e-01 2.40222037e-01 1.05804969e-02 7.24743783e-01
1.92506745e-01 -4.03807163e-01 7.21651018e-01 3.58385108e-02
-7.79670954e-01 3.04931223e-01 -2.05152258e-01 -2.15780884e-01
1.36263633e+00 -4.27191347e-01 -5.98462105e-01 1.47244975e-01
-7.46285737e-01 2.90806890e-01 8.73060599e-02 5.21345079e-01
5.78892052e-01 -1.15888214e+00 -2.70837754e-01 -1.01515464e-01
3.25728267e-01 8.67276311e-01 -5.14618717e-02 7.45470524e-01
-7.97530532e-01 5.76056480e-01 2.48290375e-01 -9.60968792e-01
-1.49950612e+00 7.69229949e-01 4.62586552e-01 -1.42952070e-01
-6.87399805e-01 8.75059128e-01 5.53173542e-01 -3.41523558e-01
3.93857449e-01 -4.77809846e-01 -5.49127936e-01 1.20139979e-01
3.80645961e-01 3.67052346e-01 -1.23789385e-01 -5.87553859e-01
-5.39655507e-01 7.53820956e-01 -4.80186611e-01 4.89373267e-01
1.04741633e+00 6.04809411e-02 1.43484816e-01 2.63848990e-01
7.53659308e-01 9.64870602e-02 -1.20632875e+00 -1.56040251e-01
1.21958360e-01 -1.62299126e-01 -1.20513439e-01 -5.06969750e-01
-9.60105300e-01 7.72944391e-01 8.52065325e-01 5.31027615e-01
8.35354865e-01 3.75353098e-02 3.37427944e-01 8.43482256e-01
-5.37762884e-03 -1.05408502e+00 4.06613201e-01 2.57425308e-01
8.82998109e-01 -1.20663905e+00 2.19892442e-01 -6.18391514e-01
-4.91001099e-01 1.43470919e+00 1.13222146e+00 -1.20700330e-01
3.66632313e-01 2.39632905e-01 5.74754551e-03 -1.24448948e-01
-1.69800177e-01 -2.85166264e-01 4.05187190e-01 4.75497723e-01
3.31419140e-01 -1.15014042e-03 -8.04298371e-02 6.30873322e-01
-1.10256538e-01 -3.42283964e-01 4.54224944e-01 1.02065921e+00
-7.02987790e-01 -4.11167949e-01 -7.55416155e-01 4.79591668e-01
-4.98712629e-01 -9.04170200e-02 -6.01635039e-01 1.01644492e+00
4.21738684e-01 7.53695250e-01 1.39592141e-01 -4.19193149e-01
5.35380781e-01 -1.78140342e-01 -1.07256502e-01 -6.40810311e-01
-2.06607714e-01 3.07619780e-01 -3.14450562e-01 -1.29349887e-01
-3.79614502e-01 -3.30875218e-01 -1.25142288e+00 2.10322827e-01
-1.27675200e+00 -1.25085235e-01 5.11116922e-01 7.54099965e-01
2.07066938e-01 6.43571854e-01 6.29827142e-01 -7.54504263e-01
-6.42697990e-01 -1.17340100e+00 -4.66729015e-01 1.96610138e-01
2.03759074e-01 -4.57340837e-01 -1.87760264e-01 2.68692851e-01] | [8.695822715759277, -0.632073163986206] |
7f0d31de-6ef3-4487-a7e5-d55690b634f0 | perch-perception-via-search-for-multi-object | 1510.05613 | null | http://arxiv.org/abs/1510.05613v2 | http://arxiv.org/pdf/1510.05613v2.pdf | PERCH: Perception via Search for Multi-Object Recognition and Localization | In many robotic domains such as flexible automated manufacturing or personal
assistance, a fundamental perception task is that of identifying and localizing
objects whose 3D models are known. Canonical approaches to this problem include
discriminative methods that find correspondences between feature descriptors
computed over the model and observed data. While these methods have been
employed successfully, they can be unreliable when the feature descriptors fail
to capture variations in observed data; a classic cause being occlusion. As a
step towards deliberative reasoning, we present PERCH: PErception via SeaRCH,
an algorithm that seeks to find the best explanation of the observed sensor
data by hypothesizing possible scenes in a generative fashion. Our
contributions are: i) formulating the multi-object recognition and localization
task as an optimization problem over the space of hypothesized scenes, ii)
exploiting structure in the optimization to cast it as a combinatorial search
problem on what we call the Monotone Scene Generation Tree, and iii) leveraging
parallelization and recent advances in multi-heuristic search in making
combinatorial search tractable. We prove that our system can guaranteedly
produce the best explanation of the scene under the chosen cost function, and
validate our claims on real world RGB-D test data. Our experimental results
show that we can identify and localize objects under heavy occlusion--cases
where state-of-the-art methods struggle. | ['Venkatraman Narayanan', 'Maxim Likhachev'] | 2015-10-19 | null | null | null | null | ['scene-generation'] | ['computer-vision'] | [ 6.45633519e-01 2.53724605e-01 -5.25972433e-02 -3.59634608e-01
-9.31791604e-01 -7.95917332e-01 5.16867220e-01 -7.01768138e-03
1.84593990e-01 3.62745464e-01 -1.54774740e-01 -1.73794135e-01
-6.21729076e-01 -5.50766528e-01 -9.88505900e-01 -6.25402689e-01
-8.77417028e-02 9.75048125e-01 2.94461250e-01 -5.97977042e-02
5.00078857e-01 7.82861471e-01 -1.84593618e+00 5.43569140e-02
5.26532650e-01 1.15001166e+00 6.84706330e-01 6.79414034e-01
1.26891747e-01 4.27398115e-01 -1.88829571e-01 2.96674557e-02
5.90946078e-01 -3.48846734e-01 -9.71333206e-01 9.16944206e-01
4.28914338e-01 3.23302224e-02 -1.24698162e-01 9.64820147e-01
6.38294071e-02 3.25294882e-02 5.28733969e-01 -1.42097247e+00
-5.52113831e-01 2.05531381e-02 -2.18888223e-01 -3.97289068e-01
8.17012668e-01 1.20042125e-02 1.13199425e+00 -1.02313375e+00
9.00661230e-01 1.51508474e+00 4.63306338e-01 3.87205303e-01
-1.48008728e+00 5.70517406e-02 9.81211737e-02 2.38885328e-01
-1.41422498e+00 -3.96174312e-01 8.83700967e-01 -3.60704064e-01
9.26912129e-01 3.65913898e-01 6.27187014e-01 7.90781438e-01
3.53848189e-01 8.03791463e-01 1.05041909e+00 -4.59121644e-01
3.60633254e-01 2.49000341e-02 -1.84192881e-01 1.00029314e+00
2.67376065e-01 1.66036949e-01 -5.71113050e-01 -2.82452554e-01
9.92190301e-01 -6.50487700e-03 -1.25950739e-01 -1.18779469e+00
-1.36957896e+00 6.89741254e-01 4.64124441e-01 -2.09867749e-02
-6.23838603e-01 2.31336221e-01 -2.70548224e-01 2.44574115e-01
-8.24851915e-02 8.73114586e-01 -6.54372752e-01 7.66225010e-02
-2.70064950e-01 5.22242665e-01 9.77085471e-01 1.42108369e+00
1.00899744e+00 -4.09617007e-01 4.93334472e-01 5.09484112e-01
4.43827897e-01 5.45071185e-01 1.20944846e-02 -1.30828512e+00
3.15596759e-01 6.02389514e-01 2.95932084e-01 -9.43801582e-01
-4.07331318e-01 -6.62173703e-02 -1.97399884e-01 3.62086326e-01
2.40498960e-01 2.39193305e-01 -9.47233617e-01 1.39061916e+00
4.66258585e-01 -1.55039743e-01 1.87940463e-01 1.01740718e+00
1.61964536e-01 3.31246197e-01 -4.96662557e-01 -2.32881024e-01
1.14796317e+00 -8.21930408e-01 -3.37298602e-01 -6.12137496e-01
2.09121004e-01 -8.28945339e-01 7.21342266e-01 5.51064014e-01
-1.03836262e+00 -5.38117230e-01 -1.10214496e+00 -8.76589864e-02
-1.72177941e-01 -2.07831249e-01 8.31319094e-01 2.31643379e-01
-8.94779682e-01 6.49873435e-01 -9.15943205e-01 -6.36228740e-01
1.58637702e-01 5.42295754e-01 -5.45278251e-01 -2.75874019e-01
-1.80783138e-01 1.17182863e+00 5.55085301e-01 1.98246807e-01
-1.03538668e+00 -1.76329210e-01 -9.05081928e-01 -3.23352307e-01
1.02775311e+00 -7.61538863e-01 1.32939970e+00 -6.33648157e-01
-1.29998565e+00 7.88321614e-01 -4.01401818e-01 -1.91518202e-01
2.86780357e-01 -3.10684025e-01 -3.28440778e-02 3.17052528e-02
1.21687129e-01 5.85675657e-01 6.76899672e-01 -1.82846963e+00
-6.53121650e-01 -5.62291682e-01 9.60301757e-02 2.54755169e-01
5.83062112e-01 -4.91869330e-01 -5.52913189e-01 -7.80421719e-02
1.17371786e+00 -1.33941138e+00 -6.65980816e-01 1.84401214e-01
-7.41254568e-01 -2.51134843e-01 8.69041562e-01 -2.03042984e-01
2.33728439e-01 -1.87816119e+00 4.78548110e-01 5.63525617e-01
-1.10868201e-01 -4.34749633e-01 -7.51069263e-02 5.21164834e-01
2.97616959e-01 -1.84601590e-01 -7.60834590e-02 -3.69696617e-01
2.91033208e-01 8.62719774e-01 -6.66224420e-01 6.38888597e-01
3.04632097e-01 7.74078965e-01 -9.28392768e-01 -5.10489345e-01
4.95131046e-01 4.36697491e-02 -5.18336773e-01 3.11288208e-01
-6.64068758e-01 4.06922638e-01 -7.86449730e-01 9.09087360e-01
3.97724956e-01 -3.26154232e-01 3.05543333e-01 -1.38202190e-01
-7.45503977e-02 1.57255352e-01 -1.64680994e+00 2.13883424e+00
-2.61114746e-01 3.01012635e-01 5.55509850e-02 -9.26587939e-01
1.12200677e+00 -5.63156307e-02 6.34888470e-01 -3.38047296e-01
-5.63747175e-02 4.74809766e-01 -2.44761065e-01 -7.39261329e-01
6.05963647e-01 -1.10068485e-01 -2.88935244e-01 8.52322727e-02
-6.91755041e-02 -9.15095806e-01 -2.82130837e-01 -1.70234099e-01
1.24194717e+00 6.04053795e-01 3.65725964e-01 -1.45217180e-01
6.62889555e-02 5.13068378e-01 5.52673340e-01 1.13867843e+00
1.61817580e-01 6.14068091e-01 1.78816929e-01 -5.39108217e-01
-1.07817984e+00 -1.23577511e+00 -8.38830173e-02 6.18833899e-01
8.90758812e-01 -2.22005293e-01 -2.42873862e-01 -5.23350358e-01
2.53381461e-01 5.91497779e-01 -4.61266279e-01 1.86906621e-01
-4.79781568e-01 -3.54925454e-01 -1.85213625e-01 4.37179714e-01
2.57463664e-01 -8.98036718e-01 -1.20123041e+00 3.79854023e-01
8.95185471e-02 -1.24528658e+00 1.22464955e-01 5.40309668e-01
-8.29492927e-01 -1.34833491e+00 3.29310112e-02 -8.14385951e-01
9.19097304e-01 5.33962905e-01 1.01031613e+00 -6.41555339e-02
-6.30027473e-01 8.76344502e-01 -2.36719444e-01 -3.19535971e-01
-3.48865151e-01 -4.93027717e-01 1.46313071e-01 -3.04777950e-01
-5.99948736e-03 -5.55822909e-01 -3.01353067e-01 6.17134869e-01
-6.81231081e-01 1.89815104e-01 7.88745284e-01 7.21198499e-01
1.14821541e+00 1.63495526e-01 -1.15855271e-02 -4.99952644e-01
2.13643417e-01 -2.95122266e-01 -8.04434180e-01 5.06985545e-01
-5.10570526e-01 4.05482084e-01 2.26700813e-01 -3.09911609e-01
-7.15193629e-01 7.96925128e-01 3.13139945e-01 -5.71814835e-01
-3.57644051e-01 3.36178273e-01 -2.61935443e-01 -2.13646144e-01
5.16603291e-01 2.35322386e-01 1.32337362e-02 -4.91826326e-01
4.59314674e-01 2.46122673e-01 8.22238564e-01 -8.48291695e-01
7.41779685e-01 5.82549393e-01 6.68118060e-01 -7.53202975e-01
-1.01170266e+00 -5.86511612e-01 -7.94750750e-01 -1.54654220e-01
8.10786843e-01 -4.94655490e-01 -7.97525883e-01 -1.74030513e-01
-1.33990037e+00 -1.09179050e-01 -3.75517160e-01 4.03038561e-01
-1.28436351e+00 2.00782493e-01 1.49547398e-01 -1.06614494e+00
4.23623055e-01 -1.36476982e+00 1.57686603e+00 4.94906269e-02
-2.16139674e-01 -6.15275383e-01 2.79586855e-02 2.16882735e-01
4.91858423e-02 5.33929765e-01 8.62484336e-01 -5.53634405e-01
-1.19883466e+00 -2.29455292e-01 -4.72120605e-02 -1.96956843e-01
2.89092541e-01 -2.38497466e-01 -7.07009912e-01 -1.75270468e-01
4.86329757e-02 -4.52976763e-01 4.51481402e-01 2.21587211e-01
1.27532697e+00 -1.31808057e-01 -7.05057859e-01 4.96399850e-01
1.68044209e+00 1.94169626e-01 3.35919201e-01 3.32070649e-01
4.87021863e-01 7.66362846e-01 8.93240571e-01 3.82059008e-01
3.83087218e-01 8.81131291e-01 1.00899792e+00 2.64421314e-01
-2.37925146e-02 -3.67932290e-01 -1.29585907e-01 3.37075830e-01
4.77522723e-02 -1.86048031e-01 -9.32459414e-01 6.45435512e-01
-2.31953835e+00 -8.06683898e-01 4.13242960e-03 2.13820386e+00
1.87981144e-01 1.55050993e-01 -1.63470209e-01 -3.73253860e-02
4.20532167e-01 -2.57455915e-01 -8.87588799e-01 -1.95543036e-01
-1.91540584e-01 3.38012353e-02 5.31547844e-01 6.32394731e-01
-1.09600866e+00 7.02512801e-01 6.45984840e+00 2.56994218e-01
-6.11359119e-01 -2.41669357e-01 2.66990885e-02 1.60086066e-01
-3.97557527e-01 3.49614292e-01 -7.09582210e-01 -1.67722583e-01
3.78784120e-01 3.46617810e-02 6.66032553e-01 1.17036033e+00
-1.33494586e-01 -3.17481399e-01 -1.69155288e+00 1.05466163e+00
2.39460096e-01 -1.24788296e+00 -2.33912826e-01 4.17167455e-01
6.61123693e-01 -2.10166518e-02 -1.15400828e-01 -1.17955089e-01
5.17611384e-01 -9.09771800e-01 1.10228491e+00 5.25870085e-01
1.11062147e-01 -3.09488833e-01 3.45996559e-01 4.67652649e-01
-1.06378222e+00 -3.09766293e-01 -4.11238194e-01 3.62925269e-02
3.08859587e-01 3.48148048e-01 -1.29830039e+00 6.32649243e-01
4.04990762e-01 4.03664678e-01 -2.21158043e-01 1.16777623e+00
-3.11350971e-01 -1.77403957e-01 -5.88252604e-01 -1.53255135e-01
1.17693447e-01 -1.66558642e-02 7.91929901e-01 7.62046039e-01
4.62394565e-01 1.57372907e-01 6.51226759e-01 1.04804158e+00
3.59065801e-01 -4.63567823e-01 -8.93371344e-01 2.82681614e-01
5.25223851e-01 9.25090611e-01 -8.76321197e-01 2.62091793e-02
-1.52085021e-01 9.65592563e-01 1.93071455e-01 1.52920678e-01
-6.02705419e-01 -6.43669069e-02 5.98714232e-01 -1.78253338e-01
5.82173407e-01 -7.17652082e-01 -4.95757371e-01 -8.44463527e-01
2.84330755e-01 -5.63405097e-01 9.65399966e-02 -9.93306577e-01
-1.06891918e+00 3.97649348e-01 2.31684119e-01 -1.27653730e+00
-6.39451563e-01 -9.02304113e-01 -4.07582447e-02 5.51517606e-01
-1.22471642e+00 -1.23637521e+00 -3.62278789e-01 4.63251859e-01
8.61257315e-01 2.41502106e-01 9.21740294e-01 -4.72820073e-01
3.95571953e-03 -2.61693060e-01 -1.20169379e-01 -6.31707549e-01
1.24699205e-01 -1.37559271e+00 2.48949587e-01 7.82999694e-01
4.65390354e-01 5.04681706e-01 1.12594652e+00 -6.24293506e-01
-2.26247120e+00 -6.90283120e-01 6.40764832e-01 -6.50784075e-01
6.12113833e-01 -4.46091264e-01 -5.29858589e-01 7.91393757e-01
-3.05918545e-01 1.02239095e-01 9.55319703e-02 8.50149393e-02
-1.93335906e-01 1.98575735e-01 -1.20932043e+00 5.41581392e-01
1.51642680e+00 -4.29072171e-01 -7.13264525e-01 7.44841397e-01
6.79480612e-01 -7.69496799e-01 -8.36799443e-01 4.13260669e-01
6.36678934e-01 -7.90646911e-01 1.06111562e+00 -5.47724426e-01
3.20918560e-01 -4.77549344e-01 -7.79925346e-01 -1.12311637e+00
-1.96101665e-01 -8.58434677e-01 -9.32652503e-02 6.53047860e-01
4.08396572e-01 -5.30152738e-01 8.15653861e-01 8.70416224e-01
-4.78945434e-01 -9.14868355e-01 -1.11896753e+00 -9.72169518e-01
-7.31530666e-01 -6.57638133e-01 6.50933981e-01 4.87058401e-01
-2.35327914e-01 6.82776198e-02 -1.74164489e-01 6.40804231e-01
9.04925108e-01 8.27311575e-01 1.06910467e+00 -1.21253181e+00
-4.60134476e-01 -6.18937835e-02 -6.45394087e-01 -1.36104596e+00
3.34236845e-02 -4.94490594e-01 8.40283453e-01 -1.56794298e+00
2.64466833e-02 -7.64954329e-01 1.74988523e-01 6.07109249e-01
3.72645676e-01 -2.62928188e-01 3.31860334e-01 3.28502178e-01
-7.89657414e-01 4.23304737e-01 1.27043653e+00 -8.14314559e-02
-1.82431012e-01 -1.24083441e-02 -5.18716931e-01 8.64316523e-01
3.70661348e-01 -3.67218137e-01 -3.48021626e-01 -7.01573074e-01
2.24863902e-01 3.11037064e-01 9.18718934e-01 -1.02382898e+00
3.20514381e-01 -6.27914548e-01 2.57112682e-01 -7.14567840e-01
8.25810373e-01 -1.21874201e+00 3.74195665e-01 3.70994478e-01
-2.83773959e-01 2.11859867e-01 -1.54856220e-02 7.02261806e-01
9.45288539e-02 -1.92753777e-01 3.74056667e-01 -2.76690125e-01
-1.18038237e+00 1.52077273e-01 9.67311934e-02 -4.66814727e-01
1.14623141e+00 -5.06834984e-01 -3.50674927e-01 -1.32178530e-01
-8.28777850e-01 2.68233508e-01 8.29304874e-01 6.04734957e-01
9.57493484e-01 -1.24730194e+00 -3.55942458e-01 3.22657973e-01
1.77750707e-01 3.84507686e-01 -3.00148457e-01 5.23618400e-01
-3.14177334e-01 5.65860629e-01 7.33297244e-02 -1.04439354e+00
-1.07972121e+00 5.41124523e-01 1.92175314e-01 1.59710750e-01
-6.43615723e-01 9.12051797e-01 2.41237208e-02 -4.37077165e-01
1.39467105e-01 -4.84939456e-01 5.25747240e-01 -7.06797540e-01
-3.46633792e-03 3.20024431e-01 -1.32101160e-02 -5.83668053e-01
-5.89796901e-01 7.69168854e-01 3.64423007e-01 -3.20623428e-01
1.50282621e+00 -2.66832799e-01 -8.43909234e-02 3.85680467e-01
1.02433133e+00 -1.94203869e-01 -1.35895848e+00 -1.83620587e-01
1.90666661e-01 -7.05040514e-01 -1.94876701e-01 -5.74265599e-01
-3.60760421e-01 3.30342263e-01 4.55471873e-01 6.00059152e-01
9.39717710e-01 7.55022764e-01 4.13714319e-01 8.78700972e-01
1.09253132e+00 -9.08515394e-01 2.97229886e-01 4.03568983e-01
1.28022146e+00 -1.24601781e+00 1.86091200e-01 -6.93746030e-01
-4.59414721e-01 1.31653023e+00 4.83001798e-01 -2.36613363e-01
4.12637055e-01 6.73797354e-02 -2.60235816e-01 -5.76873302e-01
-7.15348721e-01 -3.42091709e-01 4.12895590e-01 5.13637841e-01
-3.10374796e-01 9.06360298e-02 3.55990738e-01 1.24708794e-01
-3.47648770e-01 -2.17239812e-01 2.26702794e-01 1.34168899e+00
-7.22950339e-01 -1.00383866e+00 -5.94478667e-01 2.50351191e-01
2.17601031e-01 5.70763886e-01 -4.17408794e-01 8.36618304e-01
3.13843936e-01 1.11619949e+00 -3.65686528e-02 -4.03892457e-01
6.26953781e-01 -2.19138622e-01 1.03130591e+00 -7.80307651e-01
1.27767667e-01 8.46918076e-02 1.78281575e-01 -1.01966667e+00
-5.60563564e-01 -1.07550883e+00 -1.22567904e+00 3.94226998e-01
-4.70788240e-01 -8.17416534e-02 1.03279960e+00 1.27352846e+00
4.02080506e-01 2.37448663e-01 6.65039897e-01 -1.35603368e+00
-7.47182310e-01 -3.74991000e-01 -3.80511433e-01 3.48502785e-01
5.31639516e-01 -9.85918641e-01 -1.82533026e-01 1.01210788e-01] | [5.90219259262085, -0.9133263826370239] |
895a07fe-7e59-4b39-9d24-294dfa99c098 | farsbase-kbp-a-knowledge-base-population | 2005.01879 | null | https://arxiv.org/abs/2005.01879v1 | https://arxiv.org/pdf/2005.01879v1.pdf | FarsBase-KBP: A Knowledge Base Population System for the Persian Knowledge Graph | While most of the knowledge bases already support the English language, there is only one knowledge base for the Persian language, known as FarsBase, which is automatically created via semi-structured web information. Unlike English knowledge bases such as Wikidata, which have tremendous community support, the population of a knowledge base like FarsBase must rely on automatically extracted knowledge. Knowledge base population can let FarsBase keep growing in size, as the system continues working. In this paper, we present a knowledge base population system for the Persian language, which extracts knowledge from unlabeled raw text, crawled from the Web. The proposed system consists of a set of state-of-the-art modules such as an entity linking module as well as information and relation extraction modules designed for FarsBase. Moreover, a canonicalization system is introduced to link extracted relations to FarsBase properties. Then, the system uses knowledge fusion techniques with minimal intervention of human experts to integrate and filter the proper knowledge instances, extracted by each module. To evaluate the performance of the presented knowledge base population system, we present the first gold dataset for benchmarking knowledge base population in the Persian language, which consisting of 22015 FarsBase triples and verified by human experts. The evaluation results demonstrate the efficiency of the proposed system. | ['Behrouz Minaei-Bidgoli', 'Majid Asgari-Bidhendi', 'Behrooz Janfada'] | 2020-05-04 | null | null | null | null | ['knowledge-base-population'] | ['natural-language-processing'] | [-4.40069169e-01 7.49251187e-01 -3.66891086e-01 6.45658076e-02
-3.37533116e-01 -5.43026984e-01 6.09559953e-01 5.16703546e-01
-5.64749241e-01 1.47754955e+00 -2.62929887e-01 -1.50943488e-01
-5.74638069e-01 -1.35198545e+00 -5.62204540e-01 -1.49567574e-01
2.67448038e-01 1.02444613e+00 8.78185093e-01 -6.75857961e-01
2.02605911e-02 2.53481418e-01 -1.62561297e+00 2.73688942e-01
1.31855237e+00 9.85135496e-01 3.77905322e-03 9.52492505e-02
-5.00163436e-01 8.12151670e-01 -5.30793607e-01 -8.21336746e-01
3.58521715e-02 -7.46530443e-02 -1.30773258e+00 -4.48657006e-01
8.25295821e-02 1.94604561e-01 -1.89918280e-01 1.06429017e+00
2.81121612e-01 -1.21389762e-01 5.45789182e-01 -1.29391837e+00
-4.64952618e-01 1.09251094e+00 -5.58341704e-02 -1.72663763e-01
5.55606723e-01 -4.03978884e-01 8.57777059e-01 -7.15915442e-01
1.16699111e+00 9.63073552e-01 7.42329895e-01 1.13001585e-01
-4.81554836e-01 -6.65532172e-01 -2.28317767e-01 8.40169370e-01
-1.75222957e+00 -2.68058717e-01 3.36923689e-01 -4.55989301e-01
1.17664313e+00 1.73619077e-01 7.53353298e-01 6.64432228e-01
-1.89420983e-01 4.33498323e-01 9.85769272e-01 -9.06505227e-01
2.04211548e-02 7.45573580e-01 8.34998190e-01 8.14074636e-01
1.21841490e+00 -4.10849154e-01 -7.05565870e-01 -1.59257129e-01
2.38325328e-01 -6.83969438e-01 -2.28617013e-01 -4.96982366e-01
-9.39938128e-01 3.46918315e-01 6.66605607e-02 4.85264033e-01
-4.89759564e-01 -6.19886398e-01 3.43460500e-01 8.71878490e-02
-2.23110132e-02 3.30433995e-01 -6.50005639e-01 1.40166506e-01
-9.26349819e-01 3.30065489e-01 1.50105500e+00 1.47840583e+00
8.79064202e-01 -5.53447545e-01 1.90201059e-01 7.67149210e-01
5.50915301e-01 7.43009329e-01 5.20853460e-01 -4.53696549e-01
7.31990516e-01 1.56661570e+00 2.96079427e-01 -1.03201890e+00
-4.98777956e-01 -1.40032604e-01 -5.95379055e-01 -2.38749191e-01
4.15691406e-01 -2.40011126e-01 -5.78463137e-01 1.32965207e+00
5.59906185e-01 -3.20557743e-01 6.06431544e-01 4.71015453e-01
1.37992179e+00 3.50469857e-01 3.42865735e-02 -4.37833309e-01
1.52038360e+00 -7.03656912e-01 -9.02212381e-01 7.92745799e-02
4.90401000e-01 -6.83369100e-01 3.16931039e-01 3.71773303e-01
-1.03521061e+00 -3.04356396e-01 -1.26926804e+00 2.61025596e-03
-1.16214073e+00 2.10320622e-01 2.46925995e-01 6.91290915e-01
-6.06915653e-01 3.45391035e-01 -6.31565213e-01 -8.65941346e-01
2.18974560e-01 3.67318392e-01 -7.84082770e-01 8.01518038e-02
-1.83190966e+00 1.44858468e+00 1.65282238e+00 6.33287281e-02
-1.54375762e-01 -4.25130099e-01 -7.55059600e-01 1.70121491e-01
9.61166799e-01 -9.62268412e-01 7.05271542e-01 -4.69194680e-01
-1.07664740e+00 8.43318224e-01 2.25855783e-01 -7.06302881e-01
3.60153437e-01 -2.14092344e-01 -9.49515343e-01 1.58115312e-01
2.00825736e-01 2.88088694e-02 2.45785907e-01 -1.30279636e+00
-9.66286898e-01 -3.76786143e-01 2.00355276e-01 -3.08794156e-02
-6.19820237e-01 3.24723758e-02 -7.40430593e-01 -1.42455652e-01
-2.45722950e-01 -7.23279953e-01 2.86224216e-01 -6.29527509e-01
-5.14392078e-01 -3.12947601e-01 6.81265831e-01 -9.26555216e-01
1.76907599e+00 -1.60209215e+00 4.11020941e-04 7.42309511e-01
6.99494630e-02 8.93926442e-01 3.10720414e-01 8.65994751e-01
1.73677891e-01 1.38604594e-02 -1.25931859e-01 6.92674279e-01
1.24347828e-01 3.58945370e-01 6.12984374e-02 -2.57656902e-01
-8.25750362e-03 6.94676876e-01 -1.03443313e+00 -1.12566459e+00
5.44988886e-02 1.16175354e-01 -1.72012016e-01 -1.09419800e-01
-2.28283033e-01 -4.51403201e-01 -4.93191093e-01 6.63047731e-01
7.51598418e-01 -2.11376861e-01 7.20483780e-01 -6.74769282e-01
-5.75187504e-02 -1.90022979e-02 -1.60826409e+00 1.32714891e+00
-2.77495116e-01 -3.01480442e-02 -1.41789854e-01 -8.06980133e-01
1.01492739e+00 3.23619843e-01 4.14796144e-01 -2.95503259e-01
1.78293884e-01 6.48683786e-01 -6.62649050e-02 -7.88326144e-01
6.71477377e-01 2.92248160e-01 5.24741933e-02 7.79365003e-02
5.82724512e-01 7.67695904e-02 1.27308214e+00 6.17339909e-01
1.14082503e+00 4.46595967e-01 9.93055344e-01 -3.26026946e-01
1.14672196e+00 6.09815121e-01 5.79206944e-01 3.65805537e-01
3.80108088e-01 -2.11133629e-01 2.24551409e-01 -1.64431348e-01
-7.93697476e-01 -9.92360055e-01 -1.88049540e-01 4.74855423e-01
2.66672879e-01 -1.04498982e+00 -7.47942388e-01 -1.04761934e+00
1.71596810e-01 6.20530903e-01 -2.36229494e-01 3.98698300e-02
-3.19732845e-01 -7.90269375e-01 7.60062635e-01 1.29387975e-01
9.84756172e-01 -1.16072953e+00 -3.93117070e-01 3.75640929e-01
-5.03556907e-01 -1.35505104e+00 2.84167379e-01 -1.08808510e-01
-3.43744457e-01 -1.92437315e+00 -2.05982074e-01 -7.03894079e-01
4.10978884e-01 -2.94562608e-01 1.25937343e+00 9.60937664e-02
2.28456128e-03 4.66202587e-01 -7.38575518e-01 -6.68065250e-01
-7.59933233e-01 6.72354341e-01 -8.01871270e-02 -3.23577702e-01
7.27699101e-01 -5.57378709e-01 -1.18700281e-01 3.42721105e-01
-8.54196310e-01 -2.04347316e-02 8.49349558e-01 7.39961088e-01
3.26191247e-01 3.38747621e-01 8.11743677e-01 -1.40439582e+00
5.48812568e-01 -6.24158502e-01 -5.95036387e-01 9.40374672e-01
-1.03941178e+00 8.33972618e-02 4.99748081e-01 1.62209749e-01
-1.36870790e+00 -4.45110165e-02 2.74025630e-02 3.63464534e-01
-6.80151209e-02 1.10528886e+00 -5.23010135e-01 4.75486703e-02
9.84055638e-01 5.64215854e-02 -2.72905350e-01 -5.24554133e-01
5.46525717e-01 1.00668967e+00 6.20926082e-01 -8.16047370e-01
8.06501031e-01 4.53417227e-02 -1.48925800e-02 -9.14019346e-01
-8.26960444e-01 -5.32681048e-01 -8.65403295e-01 -1.44292265e-01
4.33803201e-01 -9.43603456e-01 -7.53895044e-01 3.07796061e-01
-1.11111236e+00 2.42043838e-01 -4.65672255e-01 3.24389070e-01
-4.33168650e-01 4.55076039e-01 -3.26699018e-01 -6.40299261e-01
-6.26759291e-01 -2.41522610e-01 5.12080073e-01 2.44952530e-01
-2.11253539e-01 -7.72724450e-01 3.37547094e-01 6.80512607e-01
-1.28163874e-01 7.36530125e-02 9.60436583e-01 -1.02144957e+00
-3.71974796e-01 -3.00712883e-01 -3.26032043e-01 4.37495351e-01
1.88783705e-01 -3.40203829e-02 -5.20391524e-01 1.47429720e-01
-7.10189760e-01 -2.23122478e-01 5.60574710e-01 -3.39449108e-01
1.85783848e-01 -1.21051908e-01 -5.77618897e-01 -6.11224174e-02
1.57450843e+00 2.79769748e-01 8.01801145e-01 7.76014447e-01
5.88785291e-01 4.89844203e-01 8.81502271e-01 1.54676512e-01
7.57086694e-01 5.40092230e-01 -1.23906910e-01 3.73971641e-01
-2.65601903e-01 -1.51246056e-01 4.54089977e-02 1.16080236e+00
-7.10916042e-01 -1.27407804e-01 -1.26688647e+00 5.77366590e-01
-2.19143796e+00 -7.88116634e-01 -1.72519550e-01 1.94705069e+00
1.24696517e+00 2.18963593e-01 1.39171034e-01 4.66691941e-01
6.84000969e-01 -5.39756775e-01 -1.63408086e-01 1.09697141e-01
-4.24192607e-01 3.32659543e-01 5.98875582e-01 4.20161366e-01
-9.52246010e-01 1.09905899e+00 5.24200439e+00 1.00830007e+00
-4.62115586e-01 8.49546045e-02 -4.17118967e-01 4.62087899e-01
7.92628527e-02 1.69407576e-01 -1.28833258e+00 2.93094188e-01
9.64341998e-01 -6.96606755e-01 3.47572118e-02 7.69345939e-01
-1.83359295e-01 -2.52355874e-01 -6.66464388e-01 7.22637236e-01
2.54394948e-01 -1.47446299e+00 1.48905247e-01 -9.31948796e-02
6.74331605e-01 -1.51913464e-01 -9.72269833e-01 3.95430416e-01
6.21978581e-01 -4.57600415e-01 7.15690374e-01 9.36705768e-01
5.22929490e-01 -7.37753510e-01 1.27588177e+00 4.09328729e-01
-1.29342031e+00 -1.74379740e-02 -2.17014611e-01 3.35813582e-01
8.26899558e-02 6.87911808e-01 -9.11316991e-01 1.56799948e+00
7.33529031e-01 5.96462727e-01 -9.12496567e-01 1.04966414e+00
-3.50196570e-01 4.03187096e-01 -4.46702331e-01 -1.94987506e-01
-1.91470683e-01 -2.15692028e-01 2.39405528e-01 1.31205165e+00
4.74710584e-01 1.29261062e-01 -3.77038009e-02 4.47091311e-01
-6.55905604e-02 6.85665607e-01 -5.69276631e-01 -1.46046191e-01
6.84479833e-01 1.34341121e+00 -5.39108515e-01 -7.99483657e-01
-4.98478532e-01 3.59598845e-01 3.82515460e-01 1.53377533e-01
-5.21207094e-01 -8.48956764e-01 -8.76068845e-02 3.26893806e-01
2.90980458e-01 2.39719003e-02 9.87632945e-03 -1.22069728e+00
2.85024285e-01 -8.45237970e-01 8.05499315e-01 -8.32322598e-01
-1.22054851e+00 6.46234334e-01 4.29205745e-01 -1.07226384e+00
-4.30849105e-01 -7.13426948e-01 2.98012316e-01 6.14294767e-01
-1.34638667e+00 -1.54064775e+00 -4.80487823e-01 7.22229958e-01
-1.02872103e-01 -4.97203529e-01 8.56852591e-01 6.98142290e-01
-5.33425748e-01 3.52323025e-01 6.75799744e-03 2.63680041e-01
9.24950719e-01 -1.15596604e+00 -1.34049684e-01 6.56031728e-01
1.16416931e-01 8.33202243e-01 4.43821967e-01 -1.12478423e+00
-1.17199874e+00 -1.02261496e+00 1.19440806e+00 -4.41860408e-01
8.97478700e-01 9.32846591e-02 -9.08828318e-01 8.47486019e-01
2.47036517e-01 -1.18272334e-01 5.50845802e-01 3.46502513e-01
-4.56193805e-01 -3.31694871e-01 -1.15170515e+00 2.92910218e-01
1.24201369e+00 -9.89138931e-02 -1.12650454e+00 2.91841149e-01
1.39248177e-01 -3.89002949e-01 -1.37920189e+00 5.27554512e-01
6.31480515e-01 -7.42456794e-01 9.65364099e-01 -5.48112154e-01
6.99197426e-02 -7.50853479e-01 -3.87412347e-02 -1.29661822e+00
-5.37756048e-02 -2.01776698e-01 -5.62507629e-01 1.71821117e+00
8.15022409e-01 -6.69807374e-01 8.17241907e-01 4.05568182e-01
-4.91829414e-04 -2.31838092e-01 -8.30095410e-01 -1.04981244e+00
-2.71518648e-01 -1.82999615e-02 6.65870428e-01 8.94661784e-01
6.07363939e-01 8.66378844e-01 1.97301470e-02 1.14548936e-01
4.21368837e-01 7.97121301e-02 8.06192935e-01 -1.70169914e+00
9.20456126e-02 -5.12232911e-03 -6.79022193e-01 -1.80379748e-01
1.06783822e-01 -1.13650346e+00 -1.81516990e-01 -1.99711168e+00
2.32791364e-01 -4.53440547e-01 -1.37250677e-01 7.92053580e-01
-7.49415830e-02 9.91115496e-02 1.77764073e-01 2.57892627e-03
-8.32727253e-01 2.78784633e-01 8.54817450e-01 -1.14564344e-01
-1.70017004e-01 -4.99164551e-01 -7.15941846e-01 1.06724703e+00
4.75655794e-01 -3.07082951e-01 -4.69027698e-01 2.59852290e-01
7.94380188e-01 -2.73959517e-01 1.69300400e-02 -1.16894233e+00
5.72884917e-01 -1.48969978e-01 1.24397904e-01 -6.74287021e-01
1.08021393e-01 -1.05278468e+00 5.35076559e-01 3.03693503e-01
4.56978172e-01 -2.30388775e-01 2.01489374e-01 3.11841577e-01
-4.90787745e-01 -5.56326807e-01 5.38785756e-01 -1.50009334e-01
-9.71101999e-01 -5.78857735e-02 -1.13894604e-01 3.06358874e-01
1.12919176e+00 -7.23967776e-02 -6.10281348e-01 3.38775963e-01
-8.11816573e-01 2.69319952e-01 1.05002433e-01 1.39531270e-01
2.27179810e-01 -1.38614368e+00 -7.13698924e-01 -1.32881403e-01
4.27211612e-01 -6.80381283e-02 1.27884954e-01 6.31188989e-01
-7.14622617e-01 5.45089602e-01 -5.19487858e-01 4.49887030e-02
-1.24125803e+00 7.03910351e-01 -1.11606427e-01 -5.54643989e-01
-3.66384089e-01 1.59789026e-01 -7.43263364e-01 -4.89555389e-01
-2.08607987e-01 -2.38684028e-01 -8.53478789e-01 5.45911312e-01
3.93139720e-01 6.18394852e-01 4.66117471e-01 -6.74868286e-01
-6.20898187e-01 5.19706130e-01 1.25474915e-01 -4.42589745e-02
1.24027503e+00 -1.81414738e-01 -5.03166556e-01 2.09532931e-01
2.40874961e-01 3.57543260e-01 -8.58204514e-02 -3.06021661e-01
4.97334301e-01 8.89517814e-02 -4.50476676e-01 -1.28902721e+00
-7.03512371e-01 6.20850958e-02 3.05291384e-01 3.14343810e-01
9.08364773e-01 1.43781528e-02 7.03189731e-01 9.65056539e-01
6.87976122e-01 -1.50489795e+00 -7.84970462e-01 6.99484169e-01
7.87931085e-01 -1.07071197e+00 3.18814039e-01 -1.02313185e+00
-4.13459182e-01 1.12815785e+00 6.18700385e-01 1.41998902e-01
1.07819510e+00 2.84993410e-01 -9.88214016e-02 -2.74849385e-01
-6.51456952e-01 -5.82937360e-01 3.60926688e-01 6.51271105e-01
2.11264804e-01 -8.97848457e-02 -1.08000624e+00 1.35542905e+00
-5.10556281e-01 7.08562195e-01 4.60208267e-01 9.58280742e-01
-6.19015515e-01 -1.46745503e+00 -4.47005242e-01 5.58245599e-01
-1.05616055e-01 -1.62164748e-01 -6.58443213e-01 1.33587325e+00
8.09486210e-01 9.32358861e-01 -4.95613366e-01 -3.80189329e-01
8.79763544e-01 4.11761403e-01 7.14050889e-01 -5.98373890e-01
-5.58820963e-01 -5.59783280e-01 9.83624339e-01 -1.58041805e-01
-8.63482475e-01 -3.88971686e-01 -1.30871642e+00 -2.93571800e-01
-5.48246682e-01 7.34828115e-01 3.29102367e-01 1.09466434e+00
1.81903973e-01 3.43608856e-01 -1.61654890e-01 1.28006879e-02
-7.44018191e-03 -9.86756861e-01 -6.76872671e-01 3.42097759e-01
-6.62034094e-01 -9.14769769e-01 2.44740248e-01 4.49134439e-01] | [9.280203819274902, 8.266240119934082] |
b8af47aa-75ab-442e-b499-483fc2012b77 | deep-neural-heart-rate-variability-analysis | 1612.09205 | null | http://arxiv.org/abs/1612.09205v1 | http://arxiv.org/pdf/1612.09205v1.pdf | Deep neural heart rate variability analysis | Despite of the pain and limited accuracy of blood tests for early recognition
of cardiovascular disease, they dominate risk screening and triage. On the
other hand, heart rate variability is non-invasive and cheap, but not
considered accurate enough for clinical practice. Here, we tackle heart beat
interval based classification with deep learning. We introduce an end to end
differentiable hybrid architecture, consisting of a layer of biological neuron
models of cardiac dynamics (modified FitzHugh Nagumo neurons) and several
layers of a standard feed-forward neural network. The proposed model is
evaluated on ECGs from 474 stable at-risk (coronary artery disease) patients,
and 1172 chest pain patients of an emergency department. We show that it can
significantly outperform models based on traditional heart rate variability
predictors, as well as approaching or in some cases outperforming clinical
blood tests, based only on 60 seconds of inter-beat intervals. | ['Tamas Madl'] | 2016-12-29 | null | null | null | null | ['heart-rate-variability', 'electrocardiography-ecg'] | ['medical', 'methodology'] | [ 3.92196402e-02 -2.31687278e-02 -7.78421536e-02 -4.70100731e-01
-3.49924952e-01 -3.05062383e-01 8.18051621e-02 3.94788831e-01
-6.45663321e-01 1.02281618e+00 -2.25707084e-01 -7.61106730e-01
-4.92535502e-01 -5.50648570e-01 -2.42364913e-01 -6.05086505e-01
-7.63612926e-01 7.00079083e-01 -1.67950526e-01 1.98451385e-01
-1.23101987e-01 6.64759636e-01 -8.81424427e-01 7.48255625e-02
6.57690883e-01 1.57727528e+00 -7.56765068e-01 1.19553292e+00
3.97964150e-01 9.89248872e-01 -7.50049591e-01 -3.98057066e-02
1.11526951e-01 -7.64346480e-01 -3.72565120e-01 -5.31303763e-01
1.24416098e-01 -4.43812937e-01 -1.67951822e-01 3.66377532e-01
1.14766669e+00 -1.27278000e-01 3.66136134e-01 -6.88122392e-01
-4.09905732e-01 6.07356012e-01 -1.08230509e-01 8.75548959e-01
-2.34923974e-01 2.46914610e-01 4.94327217e-01 -5.13292432e-01
6.78042099e-02 6.33383155e-01 1.31741345e+00 4.42600816e-01
-1.29774368e+00 -3.09844434e-01 -2.33120680e-01 1.21617526e-01
-1.24510443e+00 -4.86872792e-01 6.11500561e-01 -4.23444092e-01
9.22086835e-01 2.73266733e-01 1.07205570e+00 1.03222406e+00
4.96106565e-01 1.76255435e-01 9.44061756e-01 -1.48033127e-01
1.84396014e-01 7.87988231e-02 4.68415141e-01 4.97691542e-01
3.70236307e-01 4.58052099e-01 -1.73797414e-01 -5.03089428e-01
1.07931972e+00 2.89275706e-01 -3.41629922e-01 -6.01266772e-02
-1.28920209e+00 5.61434567e-01 3.34327787e-01 4.35315043e-01
-7.90126383e-01 2.87826210e-01 6.01044476e-01 5.67093551e-01
3.35179061e-01 3.86162847e-01 -7.05878496e-01 -3.78685504e-01
-9.78864312e-01 1.57184228e-01 9.98190105e-01 6.50738105e-02
2.69693136e-02 3.66702557e-01 -3.83157551e-01 5.49618900e-01
8.37245733e-02 4.04315919e-01 5.34062564e-01 -9.96919394e-01
2.34440025e-02 3.27460319e-01 2.92305171e-01 -9.29009199e-01
-1.10751474e+00 -9.52730775e-01 -1.44106865e+00 2.63454884e-01
6.39701426e-01 -4.42314655e-01 -6.58031881e-01 1.34711540e+00
6.16328306e-02 3.86737972e-01 2.41554365e-03 1.03840828e+00
7.81197369e-01 1.47474617e-01 2.10522190e-01 -5.88498473e-01
1.32964110e+00 -4.31854635e-01 -4.55504626e-01 9.00829136e-02
4.44756061e-01 1.87260851e-01 4.73663926e-01 6.64537191e-01
-1.21008134e+00 -7.27905452e-01 -8.28728974e-01 6.19898513e-02
-1.62665620e-01 1.84918150e-01 3.72484148e-01 9.16307807e-01
-9.42787290e-01 1.29323041e+00 -1.07472336e+00 -4.11650427e-02
5.22237957e-01 3.25625896e-01 1.36348203e-01 4.51066494e-01
-1.40886080e+00 9.27211761e-01 1.44367993e-01 6.54756725e-01
-6.78224146e-01 -1.27275681e+00 -5.11024773e-01 1.97252348e-01
7.10662678e-02 -1.09046006e+00 8.83657813e-01 -6.24870479e-01
-1.61589599e+00 7.22047567e-01 7.59573579e-02 -1.07702804e+00
9.23425853e-01 -3.87098074e-01 -5.37284136e-01 4.77560550e-01
-4.24568981e-01 1.55080631e-01 5.26948571e-01 -5.07377267e-01
-2.42105514e-01 -3.48527700e-01 -2.89079368e-01 -2.96618134e-01
-8.83463491e-03 -8.86531398e-02 3.24120611e-01 -4.97734547e-01
-1.10520802e-01 -7.04189599e-01 -5.74958324e-01 2.34036818e-01
-2.64665961e-01 2.24361822e-01 2.59458959e-01 -9.22867239e-01
1.16615391e+00 -1.97094190e+00 -1.73450634e-01 3.23284268e-01
7.62752831e-01 5.93236864e-01 2.62003034e-01 -1.23719707e-01
-3.51147801e-01 3.73424292e-01 -2.93953210e-01 1.18894137e-01
-2.70445347e-01 7.92242959e-02 -1.34316266e-01 6.16428792e-01
3.03992927e-01 1.28427243e+00 -6.85027659e-01 -2.97273338e-01
5.09416640e-01 8.80858362e-01 -2.24440801e-03 -2.36242879e-02
2.84597635e-01 8.53235066e-01 -1.45432100e-01 5.45677483e-01
1.60498083e-01 -4.06406552e-01 7.59987533e-02 -9.11699682e-02
-3.24195530e-03 2.30487585e-01 -7.89049029e-01 1.20357418e+00
-1.81598604e-01 5.16334891e-01 -3.65308315e-01 -1.27525723e+00
1.20781100e+00 7.47765243e-01 7.58495331e-01 -8.29840899e-01
4.25871998e-01 3.09610963e-01 6.62247658e-01 -4.36086416e-01
-6.39881551e-01 -4.52829361e-01 2.60511100e-01 3.23151767e-01
-2.48352006e-01 4.13753957e-01 1.11870736e-01 -4.50893372e-01
1.23923028e+00 -3.51027250e-02 4.64969367e-01 -2.65718937e-01
6.47251248e-01 -2.85252720e-01 6.56989396e-01 1.05188811e+00
-6.38402104e-01 7.03805268e-01 8.73039484e-01 -1.34672177e+00
-7.94378817e-01 -1.15718675e+00 -4.00151879e-01 3.49861443e-01
-4.48060006e-01 6.41912445e-02 -3.60857964e-01 -2.12414771e-01
1.96932763e-01 3.35165828e-01 -8.18892479e-01 -3.65027666e-01
-9.71867681e-01 -1.18479323e+00 9.48291838e-01 8.93657386e-01
3.20554107e-01 -1.09302521e+00 -1.39605463e+00 7.82544911e-01
1.36213884e-01 -7.83947825e-01 2.21615404e-01 4.65993583e-01
-1.30785775e+00 -1.10907924e+00 -1.26053011e+00 -1.80068955e-01
-1.47712365e-01 -5.51302195e-01 1.51485598e+00 2.34692782e-01
-7.56049633e-01 4.93716076e-02 4.72885370e-02 -7.64966369e-01
-5.43542325e-01 8.27332065e-02 1.18929595e-01 1.50348917e-01
2.92201340e-01 -9.96553302e-01 -1.00089931e+00 4.96171154e-02
-7.72718862e-02 -3.66671741e-01 6.30876303e-01 8.16909730e-01
6.46547794e-01 -5.50195694e-01 9.19172585e-01 -7.36027777e-01
6.12858415e-01 -3.56680691e-01 -4.70772415e-01 4.24673371e-02
-1.09141791e+00 -4.16001201e-01 7.60558605e-01 -4.48895544e-01
-3.24311674e-01 -2.25797206e-01 1.40847014e-02 -6.12690985e-01
-2.15089366e-01 4.37830001e-01 5.48066556e-01 -1.84615944e-02
9.61444736e-01 1.88929558e-01 3.32380950e-01 -4.21771735e-01
-2.31964603e-01 1.41032770e-01 8.69418085e-01 -3.73230994e-01
2.75729597e-01 2.43065745e-01 4.63642001e-01 -7.99692571e-01
-5.69514215e-01 -1.77742958e-01 -9.24023867e-01 -4.31308001e-02
9.71229792e-01 -6.83665454e-01 -1.00627172e+00 4.34692621e-01
-9.67360973e-01 -3.74094456e-01 -6.08774960e-01 6.42745733e-01
-5.34002602e-01 2.09179670e-01 -1.03954387e+00 -9.51178432e-01
-8.38728368e-01 -3.84526700e-01 5.37899852e-01 1.29284933e-01
-2.26990044e-01 -1.09528601e+00 2.44624883e-01 -2.95899123e-01
8.74129474e-01 1.12346065e+00 1.01594043e+00 -7.69823849e-01
1.10898186e-02 -6.29880607e-01 -1.07450180e-01 4.32495534e-01
-8.53093527e-03 -7.98585787e-02 -9.20918345e-01 -1.57217644e-02
2.20001638e-01 2.05841642e-02 7.71984041e-01 1.01808083e+00
9.93568301e-01 8.22138414e-02 -1.14053570e-01 7.40665138e-01
1.15907967e+00 4.93153870e-01 7.24511743e-01 2.36912429e-01
5.08379579e-01 2.11185649e-01 -2.14091733e-01 5.25046587e-01
2.63718843e-01 1.32540345e-01 1.77061304e-01 -5.25189996e-01
1.95854351e-01 6.14555240e-01 6.97969869e-02 5.81281543e-01
-8.67195368e-01 1.53720200e-01 -1.11657631e+00 2.32413426e-01
-1.54561329e+00 -1.09823334e+00 -6.11466527e-01 2.25510645e+00
6.83543444e-01 3.74866366e-01 5.57846427e-01 4.01505649e-01
5.04212916e-01 -2.17991084e-01 -9.28222895e-01 -5.91694593e-01
-1.84619665e-01 5.15200555e-01 2.19638333e-01 1.46415457e-01
-1.07634473e+00 1.68149639e-03 7.07839823e+00 -3.67906779e-01
-1.36003006e+00 5.90932816e-02 1.21788037e+00 -1.99835807e-01
5.37237287e-01 -4.34421510e-01 -2.73211867e-01 4.32110012e-01
1.76178062e+00 -2.52143219e-02 1.59953952e-01 6.75047874e-01
4.39451933e-01 1.41408503e-01 -1.22011554e+00 1.09162557e+00
-3.23064059e-01 -1.32900941e+00 -5.55560768e-01 -1.86549440e-01
1.08139753e-01 5.03163300e-02 -1.63201392e-01 3.96156818e-01
-5.91453254e-01 -1.26627910e+00 2.99215823e-01 1.09117317e+00
8.29795361e-01 -5.21563888e-01 1.00645304e+00 3.06579977e-01
-7.94097245e-01 -2.96704888e-01 -8.83982405e-02 -2.33382925e-01
1.52204961e-01 8.49297822e-01 -5.10595679e-01 3.27423781e-01
7.97152042e-01 7.10954368e-01 -4.74463105e-01 1.12326384e+00
1.90319806e-01 8.77321839e-01 -3.21901888e-01 1.46996140e-01
-8.89515132e-03 1.03899971e-01 3.02772492e-01 1.12231803e+00
3.06547970e-01 2.28120312e-01 1.09510295e-01 1.01405895e+00
4.23062235e-01 -6.18132800e-02 -4.34490383e-01 1.75896138e-01
3.37359421e-02 1.25733161e+00 -6.99539483e-01 -7.86460996e-01
-2.42878944e-01 4.81873691e-01 -1.66758195e-01 2.14838907e-01
-9.65307355e-01 -5.60515940e-01 2.50200003e-01 3.37732464e-01
-7.98792392e-02 8.42876211e-02 -7.26960421e-01 -9.53775167e-01
1.08458832e-01 -7.86008239e-01 3.63590807e-01 -3.76374036e-01
-1.09721112e+00 7.72286952e-01 -2.23496154e-01 -1.13274038e+00
-5.30805230e-01 -7.02867627e-01 -9.31433082e-01 1.37235308e+00
-1.54186201e+00 -4.46115226e-01 -3.85666072e-01 3.06415498e-01
-9.79080424e-02 -6.29054103e-03 1.20098388e+00 4.67487872e-01
-6.47133350e-01 4.52624440e-01 -2.74337649e-01 3.31967682e-01
4.44415718e-01 -1.49055326e+00 4.98985589e-01 3.91919076e-01
-4.08892959e-01 6.49466693e-01 4.04535502e-01 -3.79439980e-01
-8.51928890e-01 -9.57635880e-01 1.03285468e+00 -4.07283753e-01
3.38193566e-01 1.16219081e-01 -1.19200552e+00 2.82327503e-01
-1.70607090e-01 7.10608959e-01 7.40606427e-01 -4.93088877e-03
7.18025938e-02 -3.31011444e-01 -1.07373357e+00 1.36892200e-01
5.90597987e-01 -2.28234157e-01 -5.85211635e-01 1.19069979e-01
5.44013754e-02 -3.55745167e-01 -1.43349278e+00 4.77700263e-01
1.00036740e+00 -1.17594528e+00 1.27210796e+00 -9.19189453e-01
2.55998403e-01 8.29489753e-02 4.97805893e-01 -1.02136087e+00
-5.59321940e-01 -1.01777828e+00 -5.94736278e-01 7.17560232e-01
3.00428391e-01 -8.33138227e-01 8.51192594e-01 7.85877466e-01
-1.76933840e-01 -9.48487818e-01 -9.80501294e-01 -8.06506813e-01
3.11640948e-01 -1.43194646e-01 1.35763898e-01 7.53983140e-01
-1.47140861e-01 3.45910966e-01 -4.65861112e-01 -1.75873309e-01
6.06390119e-01 3.86710092e-02 2.71686673e-01 -1.82007360e+00
-5.51304519e-01 -7.41779745e-01 -5.48148096e-01 -2.28144109e-01
-4.29643035e-01 -7.10121930e-01 -1.94231957e-01 -1.30549252e+00
-7.68709108e-02 -4.23964947e-01 -1.02594697e+00 4.51239854e-01
-3.53443235e-01 4.04269010e-01 -1.52240127e-01 -3.77306307e-04
-6.42482787e-02 2.18549068e-03 9.10268903e-01 7.39337429e-02
-6.30023539e-01 2.88525015e-01 -4.76813555e-01 7.48850942e-01
9.63877618e-01 -5.45026600e-01 -2.68794507e-01 4.36513163e-02
6.39377087e-02 7.71341205e-01 8.64205897e-01 -1.41279507e+00
-2.01176047e-01 2.88375467e-01 1.18619621e+00 -3.53928089e-01
2.01514781e-01 -3.97886693e-01 1.60900235e-01 1.16344821e+00
-4.53845322e-01 3.20405573e-01 1.71988010e-01 4.30899054e-01
-1.13736697e-01 1.24077395e-01 9.14315939e-01 -3.92038763e-01
-1.91192906e-02 3.74589205e-01 -5.22321224e-01 2.27578118e-01
8.01776171e-01 -3.05762053e-01 -2.42575794e-01 -9.58718359e-02
-1.19362712e+00 -4.50196080e-02 -3.38483155e-01 6.24401588e-03
4.47567165e-01 -1.07855237e+00 -1.03765881e+00 2.47943223e-01
-2.14736179e-01 -2.62168676e-01 3.40942562e-01 1.69796968e+00
-8.00784469e-01 4.65297014e-01 -4.23303127e-01 -7.97314823e-01
-8.30929995e-01 5.39453864e-01 9.47385430e-01 -2.88049817e-01
-1.18079591e+00 5.12520909e-01 -2.05552459e-01 1.83515489e-01
3.14892411e-01 -7.96549618e-01 -2.72750169e-01 -3.13338125e-03
8.47188532e-01 6.53248250e-01 3.24819505e-01 -7.50291497e-02
-4.90311921e-01 5.09487867e-01 2.63729453e-01 2.22676083e-01
1.27629483e+00 1.69264317e-01 2.14065965e-02 8.98135960e-01
9.00600612e-01 -6.58626974e-01 -8.73578668e-01 -3.11425515e-03
9.41181108e-02 2.10684925e-01 -6.86834976e-02 -1.17913127e+00
-1.13117278e+00 1.44154978e+00 1.20278263e+00 5.36544740e-01
1.08962190e+00 -6.40460372e-01 8.46358657e-01 5.95270813e-01
-7.53840730e-02 -6.83052003e-01 -2.52057672e-01 1.24378435e-01
6.37359560e-01 -1.05436659e+00 3.29373479e-02 1.22052655e-01
-4.03831512e-01 1.60828137e+00 1.86022490e-01 -5.77875257e-01
1.02566659e+00 1.73302069e-01 4.55346465e-01 -1.08592566e-02
-8.84404302e-01 1.73703313e-01 1.02655023e-01 7.14681149e-01
6.31266415e-01 -7.43329078e-02 -3.83104116e-01 8.37654889e-01
1.63801998e-01 6.12199545e-01 3.67358625e-01 7.12089062e-01
-2.31510282e-01 -5.71940839e-01 -1.66453227e-01 8.43905270e-01
-9.68348384e-01 -1.79235399e-01 -1.05886804e-02 7.74701953e-01
-6.04401529e-02 6.98922575e-01 1.28530279e-01 -8.64678994e-02
3.84829938e-01 5.55475652e-01 2.84189373e-01 -2.55694449e-01
-1.05374372e+00 1.41423136e-01 9.17936116e-02 -4.02866900e-01
-1.92730874e-01 -5.88937640e-01 -1.07177806e+00 -1.45992246e-02
1.60361141e-01 -6.93970323e-02 3.60527158e-01 8.14081967e-01
5.08087635e-01 8.73312116e-01 4.33141202e-01 -8.03336978e-01
-8.60410810e-01 -1.20403922e+00 -7.99848199e-01 6.67548180e-02
8.28775823e-01 -4.19086218e-01 -3.25179011e-01 2.93916434e-01] | [14.302852630615234, 3.255481719970703] |
9ac5be40-80a6-4b70-8ea4-1ff394dc1555 | conversational-search-with-mixed-initiative-1 | null | null | https://aclanthology.org/2022.dialdoc-1.7 | https://aclanthology.org/2022.dialdoc-1.7.pdf | Conversational Search with Mixed-Initiative - Asking Good Clarification Questions backed-up by Passage Retrieval | We deal with the scenario of conversational search, where user queries are under-specified or ambiguous. This calls for a mixed-initiative setup. User-asks (queries) and system-answers, as well as system-asks (clarification questions) and user response, in order to clarify her information needs. We focus on the task of selecting the next clarification question, given conversation context. Our method leverages passage retrieval from background content to fine-tune two deep-learning models for ranking candidate clarification questions. We evaluated our method on two different use-cases. The first is an open domain conversational search in a large web collection. The second is a task-oriented customer-support setup.We show that our method performs well on both use-cases. | ['David Konopnicki', 'Asaf Yehudai', 'Doron Cohen', 'Yosi Mass'] | null | null | null | null | dialdoc-acl-2022-5 | ['passage-retrieval', 'conversational-search'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.11109692e-01 2.95861781e-01 -9.42954198e-02 -4.24357742e-01
-1.51364827e+00 -8.08478475e-01 9.00130093e-01 4.05747920e-01
-5.09439707e-01 7.66722858e-01 8.12608719e-01 -6.31280422e-01
-2.44124740e-01 -3.88635069e-01 -6.41825795e-02 5.57710193e-02
4.17551011e-01 1.14348865e+00 2.28414074e-01 -6.18144870e-01
5.32841146e-01 -1.97476824e-03 -1.24914122e+00 1.03833950e+00
8.89558315e-01 7.36978292e-01 6.19684219e-01 1.13446021e+00
-5.80628693e-01 8.05617690e-01 -4.90332991e-01 -4.65555787e-01
-1.46476045e-01 -3.60271484e-01 -1.81958640e+00 1.53461888e-01
4.08341318e-01 -5.16018569e-01 -1.99885666e-02 4.27743167e-01
6.81691885e-01 4.06778455e-01 3.23709041e-01 -1.00418222e+00
-3.02621484e-01 8.52229059e-01 1.40124306e-01 4.78872567e-01
8.09326887e-01 1.66339621e-01 1.64158750e+00 -6.58337653e-01
5.20776093e-01 1.31894362e+00 4.37637269e-01 5.76465547e-01
-1.17833960e+00 -9.65042338e-02 1.50066108e-01 2.16142610e-01
-5.71346700e-01 -6.42306685e-01 5.73772013e-01 -3.49293441e-01
1.22571421e+00 6.05877042e-01 6.08823746e-02 1.24246526e+00
-4.61538970e-01 1.22684550e+00 7.33690977e-01 -4.86936629e-01
1.40267342e-01 3.48879546e-01 8.42710853e-01 -3.69055979e-02
-5.33942580e-01 -4.30777162e-01 -4.39780802e-01 -5.61273515e-01
7.54187107e-02 -8.25178623e-02 -4.52765942e-01 4.92357500e-02
-9.80772972e-01 9.24489260e-01 -1.01713926e-01 5.90871692e-01
-4.45117503e-01 -2.00309679e-01 4.83453870e-01 6.34351969e-01
1.83591917e-01 1.07952213e+00 -9.21003819e-01 -6.84068918e-01
-5.73253393e-01 5.67969561e-01 1.76685655e+00 8.12428176e-01
5.66211462e-01 -8.92751157e-01 -7.15748250e-01 1.42204869e+00
4.08405066e-02 -2.37410124e-02 3.48539144e-01 -1.22521961e+00
8.26996922e-01 3.84125024e-01 6.15854681e-01 -5.57621658e-01
-4.98901278e-01 -4.89428520e-01 -6.80823252e-02 -7.39580929e-01
5.55584967e-01 -3.31001669e-01 -2.05795527e-01 1.56393993e+00
8.22598636e-02 -2.09508628e-01 1.43321604e-01 7.64457345e-01
1.17699277e+00 5.10167360e-01 1.20377755e-02 -3.14481795e-01
1.68900502e+00 -1.33459723e+00 -7.53471971e-01 -3.08854610e-01
6.98075533e-01 -1.10166931e+00 1.48249280e+00 1.41554728e-01
-1.34149289e+00 -3.57366830e-01 -5.35193503e-01 -3.56408238e-01
-1.33139938e-01 1.41296580e-01 1.79050937e-01 1.34269878e-01
-1.03269982e+00 3.72212321e-01 -3.28180403e-01 -7.41233826e-01
-3.57606947e-01 7.24767223e-02 1.90392733e-01 1.18812779e-02
-1.45471418e+00 8.50044549e-01 -2.05723196e-01 -8.58575106e-02
-4.26793873e-01 -7.80467868e-01 -4.27769423e-01 6.44051969e-01
5.31353891e-01 -7.71981716e-01 2.39724922e+00 -6.60320401e-01
-1.60017836e+00 9.83928740e-01 -2.22725764e-01 -3.15291286e-01
5.41407585e-01 -5.59176087e-01 -3.56114775e-01 1.23717800e-01
2.67131478e-01 2.98424065e-01 2.67690927e-01 -1.14754415e+00
-9.18697476e-01 -1.72988534e-01 5.64400375e-01 4.97961223e-01
-5.06460257e-02 1.62575245e-01 -5.23638189e-01 -6.55303970e-02
-1.69438168e-01 -8.08583677e-01 -4.52621132e-02 -7.72219479e-01
-6.38000190e-01 -6.52404130e-01 6.75964475e-01 -7.23199189e-01
1.43396521e+00 -1.73393846e+00 -2.91002363e-01 -1.57322884e-01
2.05954030e-01 2.10537881e-01 -3.80121201e-01 1.17160559e+00
9.83880088e-02 2.29923099e-01 6.84695691e-02 -3.61662805e-01
1.87841639e-01 -1.49989456e-01 -5.39374292e-01 -4.43555802e-01
3.33605632e-02 9.24039066e-01 -9.65824485e-01 -3.82755458e-01
-4.05475438e-01 -1.04794621e-01 -7.20784605e-01 5.86936295e-01
-7.29522049e-01 4.29927289e-01 -8.02100420e-01 2.57224947e-01
3.01370561e-01 -5.68694592e-01 2.63553053e-01 -9.61215571e-02
2.78965309e-02 1.14728534e+00 -7.28558004e-01 1.60026383e+00
-1.04274666e+00 6.27435386e-01 6.13820910e-01 -7.25240707e-01
5.31571090e-01 5.08358538e-01 4.14921969e-01 -7.46976852e-01
-3.44779268e-02 1.84735954e-01 -1.47783309e-01 -9.84054565e-01
6.92328393e-01 3.58976185e-01 -1.35795832e-01 1.25207460e+00
-5.36974519e-02 -2.34582484e-01 7.92049468e-02 5.79828560e-01
1.46658814e+00 -2.29191750e-01 2.02182487e-01 -2.39110097e-01
6.03344142e-01 -5.30967340e-02 -1.81667190e-02 1.29998505e+00
-9.88944694e-02 4.40462232e-01 5.93522489e-01 -1.56998098e-01
-5.38646638e-01 -6.82125807e-01 2.80748427e-01 1.68496549e+00
-3.21253873e-02 -5.42327523e-01 -6.95779741e-01 -9.57845747e-01
-1.20531008e-01 1.10281146e+00 -1.34327099e-01 1.95088491e-01
-6.83626592e-01 -9.45894700e-03 1.08545400e-01 2.73069162e-02
4.00489330e-01 -1.38735843e+00 -5.59234798e-01 4.57186073e-01
-9.09197271e-01 -1.14886689e+00 -9.21272635e-01 1.05230622e-01
-4.94755626e-01 -1.27341688e+00 -5.64591706e-01 -9.41880941e-01
-1.18260853e-01 4.30715829e-01 1.83671224e+00 3.90033692e-01
-2.38642655e-02 1.21618891e+00 -4.22685266e-01 7.53901154e-02
-5.99628091e-01 6.29263103e-01 -5.04943490e-01 -7.22066090e-02
5.53330362e-01 -5.76079428e-01 -8.05867374e-01 6.13572717e-01
-7.04477608e-01 -1.37080356e-01 3.22498858e-01 9.62661147e-01
-5.99954538e-02 -7.05934167e-01 1.03659594e+00 -1.01143694e+00
1.65845692e+00 -7.01913595e-01 -2.08507687e-01 5.98766863e-01
-5.71232438e-01 1.60674617e-01 1.73004732e-01 -3.78651202e-01
-1.24320042e+00 -2.31577083e-01 -5.02938986e-01 3.39788795e-01
-2.72461325e-01 7.07603037e-01 1.97712891e-02 6.90038621e-01
7.38047957e-01 -8.59970376e-02 -2.64887422e-01 -8.55716586e-01
4.53077286e-01 1.15241230e+00 6.35980427e-01 -7.95321524e-01
1.94221616e-01 -1.53128326e-01 -9.81504440e-01 -7.37322092e-01
-9.75671530e-01 -1.07939398e+00 -3.43334287e-01 -3.29514444e-02
7.90658951e-01 -5.73895693e-01 -1.08812916e+00 -1.68569028e-01
-1.51082063e+00 -5.16045392e-01 -1.06173120e-01 9.52641070e-02
-7.13237107e-01 3.21744651e-01 -8.01010609e-01 -9.40022886e-01
-6.64909661e-01 -1.01501393e+00 1.14435565e+00 2.20026910e-01
-7.08454013e-01 -8.92844379e-01 3.62235099e-01 9.80650425e-01
7.21384704e-01 -5.28088272e-01 1.08163178e+00 -1.71824861e+00
-5.40528119e-01 -1.59120247e-01 -1.82716876e-01 -1.45322129e-01
1.19948372e-01 -4.67988402e-01 -1.12190211e+00 -1.52538285e-01
7.37316906e-02 -8.83009791e-01 7.17567742e-01 1.64011754e-02
1.15131807e+00 -7.09351718e-01 -4.17952240e-01 -2.82662392e-01
5.83769798e-01 9.12674144e-02 3.78053218e-01 2.57963061e-01
-2.17360593e-02 9.33705747e-01 5.21164596e-01 4.09516960e-01
7.25621581e-01 9.14354742e-01 8.80348459e-02 2.87901938e-01
1.39267281e-01 -3.36778998e-01 -7.87647814e-02 6.08180225e-01
6.02663815e-01 -6.05054915e-01 -9.23117161e-01 6.84927344e-01
-2.11631227e+00 -1.09249461e+00 1.02479950e-01 1.79821563e+00
1.20685387e+00 1.13673441e-01 1.19158320e-01 -3.65190566e-01
5.84371567e-01 7.95887709e-02 -6.54655337e-01 -4.54331130e-01
3.54345828e-01 1.46617055e-01 -4.52347666e-01 9.81315494e-01
-1.00316978e+00 6.18442655e-01 5.99336958e+00 4.67334121e-01
-7.22198963e-01 2.70555258e-01 6.67064309e-01 -5.32697178e-02
-6.59388781e-01 9.71602648e-02 -8.56737077e-01 3.82425874e-01
9.74843979e-01 -9.41106603e-02 4.68559086e-01 7.16988862e-01
2.65162766e-01 -9.98966023e-02 -1.71020627e+00 8.83136392e-01
-2.85130758e-02 -1.57016015e+00 -2.49926001e-01 -3.28067005e-01
2.62419045e-01 3.87414433e-02 -3.14951599e-01 1.04919899e+00
5.77376187e-01 -6.68007016e-01 6.09883480e-03 3.67939293e-01
2.84211516e-01 -2.02837691e-01 7.33893394e-01 8.06918383e-01
-6.21854544e-01 -1.01336710e-01 3.09451997e-01 2.05330048e-02
4.73901957e-01 2.17045650e-01 -1.48375523e+00 9.59242694e-03
5.57918966e-01 1.24586150e-01 -2.07068220e-01 9.15382028e-01
2.32043001e-03 6.09667182e-01 -2.58706540e-01 -3.20497334e-01
2.84533530e-01 3.57200950e-02 6.35992825e-01 1.50670135e+00
-6.62024394e-02 4.05818313e-01 4.59667683e-01 8.43944609e-01
-3.89750481e-01 1.74254224e-01 -1.58068404e-01 -1.12238146e-01
7.05968738e-01 1.39408648e+00 -3.35696101e-01 -3.00250918e-01
-3.12556833e-01 9.35712099e-01 3.26198101e-01 5.25667906e-01
-2.57884860e-01 -4.29527193e-01 4.76105601e-01 3.02129626e-01
2.24385202e-01 3.89655903e-02 1.85108125e-01 -1.02991259e+00
1.23637535e-01 -1.27403772e+00 7.31961489e-01 -8.17131519e-01
-1.43821311e+00 5.84865987e-01 -1.14342764e-01 -7.02760339e-01
-8.46956074e-01 -3.35897766e-02 -1.05778849e+00 9.64419901e-01
-1.53560507e+00 -7.49185860e-01 -3.64393145e-01 2.90993571e-01
1.26716900e+00 6.31097332e-02 8.70847404e-01 4.40527320e-01
-3.02708179e-01 4.26525176e-01 -4.91688661e-02 -2.61955932e-02
9.36636209e-01 -1.17214680e+00 3.59356880e-01 1.25794590e-01
7.89006874e-02 7.96698868e-01 8.40489566e-01 -3.81125718e-01
-1.23374045e+00 -3.97807509e-01 1.48401034e+00 -6.57773316e-01
7.61934221e-01 -4.49151427e-01 -9.92156208e-01 5.72275400e-01
8.11930299e-01 -6.55239284e-01 8.73270631e-01 6.25901103e-01
-8.86410475e-02 1.80451423e-01 -9.09235775e-01 6.95438027e-01
7.45672405e-01 -9.08388376e-01 -9.83832419e-01 9.39584374e-01
1.27977800e+00 -4.88477737e-01 -6.06489301e-01 2.76256762e-02
3.92162263e-01 -7.63596654e-01 8.95444095e-01 -1.18489754e+00
4.56439793e-01 3.45898509e-01 -1.67753905e-01 -1.26680791e+00
-8.56488869e-02 -1.32181716e+00 4.05292585e-02 1.30094254e+00
9.60112095e-01 -4.06917244e-01 4.90793735e-01 1.16337240e+00
-1.34836391e-01 -7.53209352e-01 -5.89407980e-01 -7.72008672e-02
-2.74866611e-01 -2.18359262e-01 5.87382972e-01 8.26828778e-01
5.94313025e-01 1.33206201e+00 5.68729974e-02 -3.24693084e-01
-2.30789259e-01 5.85743666e-01 9.05651510e-01 -1.24217987e+00
-4.54567909e-01 -6.99856460e-01 7.65752792e-01 -1.82219267e+00
1.58999890e-01 -5.86217940e-01 1.05507836e-01 -1.74180937e+00
1.80470556e-01 -4.23131973e-01 1.47255464e-02 2.35334679e-01
-1.77323788e-01 -6.11131310e-01 3.68140936e-02 3.02635640e-01
-9.46501434e-01 4.38651860e-01 8.95960391e-01 -3.99726808e-01
-6.82094634e-01 8.06358635e-01 -1.02189887e+00 3.88031840e-01
7.02268422e-01 -9.53305438e-02 -5.16534865e-01 -3.83954346e-01
4.36480552e-01 7.62395620e-01 2.53190454e-02 -3.23393017e-01
4.77882206e-01 -2.24457905e-01 -3.98707122e-01 -8.68477345e-01
4.08688992e-01 -5.92668295e-01 -6.61402762e-01 -3.36481482e-02
-1.55987144e+00 2.41008669e-01 -3.98248099e-02 4.51688021e-01
-3.05831403e-01 -5.87024331e-01 2.40748659e-01 -3.31880540e-01
-3.63559067e-01 1.47771791e-01 -5.72543204e-01 5.13646364e-01
2.65502393e-01 1.99984267e-01 -6.09018803e-01 -1.32991195e+00
-1.01342642e+00 8.49794686e-01 -1.76679134e-01 7.25848734e-01
3.48828584e-01 -9.39071000e-01 -8.78102720e-01 -2.70001680e-01
2.07334161e-01 -2.34697759e-01 -2.14329958e-02 6.84068561e-01
6.13840148e-02 8.64110053e-01 3.45591933e-01 -5.10802209e-01
-1.09934092e+00 2.30756789e-01 2.99434483e-01 -7.49554574e-01
-2.94689506e-01 7.86310911e-01 9.60966498e-02 -8.71992886e-01
6.98096871e-01 -3.96077901e-01 -6.59251750e-01 1.83865964e-01
6.81050181e-01 2.27100655e-01 2.97947407e-01 1.79521106e-02
-1.02371991e-01 -1.86825082e-01 -3.17106068e-01 -4.21711981e-01
1.07874310e+00 -6.74476266e-01 -6.28215671e-02 3.32492977e-01
1.31452906e+00 1.71307489e-01 -6.18092418e-01 -7.06212044e-01
6.64937198e-01 -1.50214985e-01 -3.35050941e-01 -1.38259459e+00
-3.29538763e-01 7.83642054e-01 1.56010687e-01 8.87024045e-01
7.28840292e-01 2.85603255e-01 1.09999788e+00 1.21519220e+00
-7.70067126e-02 -1.30020952e+00 4.85494792e-01 1.07441831e+00
1.30445623e+00 -1.47038770e+00 -3.10608745e-01 -4.76479940e-02
-7.74434030e-01 1.15842295e+00 8.61531913e-01 5.27398050e-01
4.88136470e-01 -1.97483450e-01 2.23330170e-01 -3.93600255e-01
-1.37261558e+00 -2.10785463e-01 2.22774938e-01 8.24095458e-02
7.78490961e-01 -2.28006363e-01 -2.02599823e-01 9.13886368e-01
-3.64119321e-01 -1.40842319e-01 3.96026611e-01 8.63605380e-01
-5.14319479e-01 -9.04302835e-01 -5.31074442e-02 5.05730450e-01
-3.59081805e-01 -2.46624961e-01 -8.37140441e-01 3.90527397e-01
-7.48389840e-01 1.58041573e+00 8.61306041e-02 -1.74415708e-01
4.70191777e-01 4.91153181e-01 -1.00589618e-01 -8.19482565e-01
-1.09945154e+00 -2.72720158e-02 9.84154642e-01 -6.93851113e-01
-1.73739910e-01 -5.28475583e-01 -9.45201218e-01 2.12476894e-01
-4.58796442e-01 8.79511178e-01 5.72077572e-01 1.11844873e+00
7.09465265e-01 2.72468001e-01 6.74262106e-01 -3.44149321e-01
-1.01616526e+00 -1.31190717e+00 1.13933228e-01 4.60667193e-01
6.61530972e-01 -1.84498936e-01 -3.50065768e-01 -1.64708138e-01] | [12.1542387008667, 7.836090564727783] |
a6d146d3-18dd-4919-96f8-8b713a12c985 | kernelized-covariance-for-action-recognition | 1604.06582 | null | http://arxiv.org/abs/1604.06582v2 | http://arxiv.org/pdf/1604.06582v2.pdf | Kernelized Covariance for Action Recognition | In this paper we aim at increasing the descriptive power of the covariance
matrix, limited in capturing linear mutual dependencies between variables only.
We present a rigorous and principled mathematical pipeline to recover the
kernel trick for computing the covariance matrix, enhancing it to model more
complex, non-linear relationships conveyed by the raw data. To this end, we
propose Kernelized-COV, which generalizes the original covariance
representation without compromising the efficiency of the computation. In the
experiments, we validate the proposed framework against many previous
approaches in the literature, scoring on par or superior with respect to the
state of the art on benchmark datasets for 3D action recognition. | ['Andrea Zunino', 'Jacopo Cavazza', 'Vittorio Murino', 'Marco San Biagio'] | 2016-04-22 | null | null | null | null | ['3d-human-action-recognition'] | ['computer-vision'] | [ 1.74560994e-01 -5.99390268e-02 1.27900556e-01 -3.68198931e-01
-1.85804918e-01 -6.04253173e-01 7.47890413e-01 3.76297883e-03
-5.02443612e-01 4.76711243e-01 2.42350593e-01 -2.49983475e-01
-5.63654840e-01 -3.36556256e-01 -3.88756931e-01 -6.58239424e-01
-2.27306813e-01 2.89826572e-01 1.70266926e-01 -9.37032998e-02
2.22711593e-01 5.56442142e-01 -1.43827665e+00 1.19691968e-01
5.89983761e-01 1.09340560e+00 -2.86840767e-01 4.52735484e-01
2.71051496e-01 9.96501029e-01 -2.84071296e-01 -3.62222344e-01
2.23277718e-01 -2.13803828e-01 -5.24369299e-01 -6.35526180e-02
5.07872999e-01 -2.16178834e-01 -4.48889732e-01 7.35715985e-01
6.79059234e-03 7.52391964e-02 1.05705190e+00 -1.21856284e+00
-7.84407735e-01 3.74055952e-01 -3.96742165e-01 8.61428306e-02
4.46257383e-01 -4.89923917e-02 1.13825321e+00 -7.00426102e-01
3.96899194e-01 1.19853401e+00 7.07564473e-01 5.19837618e-01
-1.43266642e+00 -3.10441643e-01 1.69250935e-01 3.48663509e-01
-1.44166541e+00 -2.23774165e-01 7.44412363e-01 -7.00648963e-01
1.10898888e+00 3.42715442e-01 7.95439839e-01 1.28100204e+00
1.27282053e-01 8.92458379e-01 1.44819880e+00 -2.96766967e-01
2.70343721e-01 1.07825182e-01 3.69545341e-01 5.66363037e-01
1.60226807e-01 2.57735729e-01 -7.41772711e-01 -1.97280824e-01
6.67557538e-01 -1.68791458e-01 -2.08227754e-01 -1.09724832e+00
-1.07001507e+00 5.52154064e-01 1.35449186e-01 1.83101863e-01
-3.97151053e-01 1.14690743e-01 3.08539122e-01 1.30173057e-01
4.45320725e-01 1.60713613e-01 -5.07209539e-01 -4.75622356e-01
-8.56041670e-01 2.67596781e-01 8.43339562e-01 8.23163509e-01
4.66330886e-01 -2.65093863e-01 -2.85975456e-01 4.66744691e-01
4.20044482e-01 5.75886846e-01 8.58358294e-02 -6.50304735e-01
3.12342763e-01 9.32275414e-01 -4.58125062e-02 -1.00708365e+00
-6.67619944e-01 -4.99207020e-01 -9.25939202e-01 1.91822574e-01
5.29542983e-01 2.20884830e-01 -6.60571277e-01 1.72462785e+00
1.22857846e-01 3.50422025e-01 1.97008938e-01 8.17467451e-01
4.02118683e-01 1.02449164e-01 -7.03664795e-02 1.03277701e-03
1.07828999e+00 -6.45077586e-01 -7.50293911e-01 -9.54768583e-02
4.46385086e-01 -5.97133815e-01 9.40233588e-01 4.29518253e-01
-6.78856194e-01 -4.47106093e-01 -1.00074232e+00 -7.15901032e-02
-3.96359652e-01 4.43017453e-01 9.79238987e-01 7.58160412e-01
-8.70087147e-01 6.31463230e-01 -1.04971492e+00 -2.87940711e-01
4.65950221e-01 3.09765577e-01 -7.23140478e-01 9.25195217e-02
-1.01880097e+00 9.80925441e-01 2.84423769e-01 2.49363422e-01
-3.46179485e-01 -9.69110191e-01 -8.91825259e-01 -1.33050889e-01
3.99116307e-01 -7.04926491e-01 8.01393151e-01 -2.94986278e-01
-1.61450589e+00 6.44693673e-01 6.75052330e-02 -5.61968684e-01
7.91637957e-01 -7.44877815e-01 -1.46457016e-01 7.95195624e-02
-4.20759648e-01 3.27855855e-01 7.52353132e-01 -7.49529719e-01
-1.15231015e-01 -5.47281206e-01 1.09594166e-01 -1.25082180e-01
-4.01772887e-01 -1.48429587e-01 -4.12833333e-01 -6.27040446e-01
9.94238481e-02 -1.18027580e+00 -1.27584547e-01 -7.91237922e-04
-2.20341504e-01 6.33998215e-02 8.27711821e-01 -4.10993725e-01
1.16037226e+00 -2.28836966e+00 5.72477281e-01 4.20534343e-01
1.21487737e-01 3.36047769e-01 1.36276737e-01 5.68042934e-01
-2.51621962e-01 -3.76457572e-01 -3.53753477e-01 -4.49124485e-01
2.07584903e-01 4.67619389e-01 -2.95917690e-01 7.62859285e-01
5.77495754e-01 1.03542018e+00 -6.47191584e-01 -1.93792641e-01
4.73968536e-01 7.91993618e-01 -5.45809984e-01 1.36524186e-01
1.44438431e-01 5.36863208e-01 -4.85465139e-01 4.18116480e-01
6.99287415e-01 -1.82432264e-01 -5.04972376e-02 -4.51972127e-01
9.72722396e-02 1.07829042e-01 -1.46405828e+00 1.90897453e+00
-7.78630599e-02 5.90698481e-01 -3.38463128e-01 -1.02143610e+00
1.01619220e+00 -3.98925915e-02 6.67971969e-01 -2.84492075e-01
2.34394461e-01 -9.03234556e-02 -1.19827362e-02 -4.56080377e-01
4.27112818e-01 -7.45270178e-02 -6.60845498e-03 3.31658036e-01
2.56834149e-01 -5.47086336e-02 7.47182667e-02 2.17712447e-01
1.14146078e+00 5.03611982e-01 3.57300133e-01 -3.62674177e-01
8.46255004e-01 -3.07854563e-01 1.21242590e-01 5.75575411e-01
-2.09656700e-01 4.31355476e-01 7.39432812e-01 -2.97888312e-02
-7.09082842e-01 -1.36142671e+00 -2.18809366e-01 4.83190924e-01
-2.18521938e-01 -6.36541486e-01 -6.95805252e-01 -6.62167847e-01
1.61332324e-01 5.55447400e-01 -9.11632597e-01 -1.82819307e-01
-2.53190100e-01 -7.26670206e-01 8.30863953e-01 7.29700744e-01
3.09613854e-01 -4.25367683e-01 -6.76113307e-01 -2.14341447e-01
8.03260878e-02 -1.64757168e+00 -2.41631463e-01 1.94546327e-01
-9.55931187e-01 -1.28390396e+00 -4.88518536e-01 4.06861678e-02
5.21315277e-01 -1.29821301e-01 7.46551514e-01 -5.10765672e-01
-4.96780455e-01 7.68916309e-01 -4.80432689e-01 -4.05529916e-01
-1.45106792e-01 -2.38628626e-01 1.08297821e-02 2.41013005e-01
5.52896023e-01 -6.63859665e-01 -4.11183119e-01 1.21629670e-01
-1.16402376e+00 -1.45803452e-01 8.91527176e-01 7.16558874e-01
2.99293101e-01 -1.26822874e-01 -7.24848593e-03 -7.86075413e-01
6.68897808e-01 -2.26616636e-01 -4.99127328e-01 9.00463108e-03
-7.34540939e-01 6.13700390e-01 4.49705124e-01 -6.94011629e-01
-8.38795960e-01 2.53522605e-01 1.91398963e-01 -5.71762919e-01
-1.57752827e-01 3.71344417e-01 2.75922995e-02 -5.21185808e-02
4.74493891e-01 3.38699698e-01 2.31203929e-01 -6.99378252e-01
6.06163323e-01 3.05148751e-01 5.09938657e-01 -4.51812178e-01
8.77404988e-01 7.76381016e-01 6.84041202e-01 -5.65991640e-01
-7.76462138e-01 -5.85462272e-01 -1.27433968e+00 -1.13561526e-01
7.47274101e-01 -6.00098252e-01 -8.90872657e-01 7.10345805e-01
-7.39847064e-01 -7.38043431e-03 -3.84258866e-01 9.07625616e-01
-8.95652294e-01 4.81262982e-01 -5.01982749e-01 -8.62796605e-01
7.24125579e-02 -8.67596626e-01 1.18032205e+00 -5.30840382e-02
-1.99690565e-01 -1.12346566e+00 4.34911519e-01 2.48160437e-01
4.64819908e-01 2.77611554e-01 7.87653327e-01 -6.94208562e-01
-2.98417300e-01 -4.09797192e-01 -2.61632413e-01 7.73117065e-01
3.26460153e-02 -1.49430614e-02 -9.82914031e-01 1.06155701e-01
-2.49538869e-02 -3.31163891e-02 9.82620060e-01 9.63615030e-02
6.79480672e-01 1.86752245e-01 5.33424900e-04 4.76134270e-01
1.18245292e+00 -4.50625747e-01 6.56281531e-01 2.89098889e-01
6.15751088e-01 6.43797338e-01 6.01826489e-01 7.40335464e-01
3.97794873e-01 7.58299291e-01 3.01284820e-01 -3.88639942e-02
-1.93221159e-02 -2.69747376e-01 5.09360433e-01 7.06964314e-01
-2.78674453e-01 2.30286449e-01 -8.20694268e-01 2.25025713e-01
-2.16359711e+00 -8.21152866e-01 -4.65222895e-01 1.96982372e+00
5.36938965e-01 5.75152189e-02 3.28604579e-02 1.70848116e-01
4.51171324e-02 2.48303756e-01 -3.23525757e-01 -3.10228616e-01
-2.16631874e-01 3.97137791e-01 6.13545358e-01 4.52598333e-01
-1.26457572e+00 9.47248578e-01 7.30348110e+00 4.82568651e-01
-8.35563660e-01 -3.00535113e-01 -1.01659060e-01 -1.44776776e-01
1.19069293e-01 1.72374621e-01 -6.22853458e-01 -6.51663542e-02
8.95603120e-01 7.12124631e-02 2.42930368e-01 7.74908066e-01
1.61167502e-01 -1.95355639e-01 -1.36459446e+00 9.43322062e-01
4.29062158e-01 -8.71832490e-01 -2.37044226e-02 9.76885259e-02
3.49037588e-01 -1.78977460e-01 2.12013274e-01 1.30788684e-01
3.57906483e-02 -9.58640456e-01 5.80412984e-01 1.12791932e+00
1.43624432e-02 -5.00436544e-01 7.57858872e-01 4.15125072e-01
-9.33861852e-01 1.51992729e-02 -2.23218516e-01 -4.06870872e-01
1.50182201e-02 6.27012193e-01 -6.01079762e-01 8.79082739e-01
6.88138187e-01 1.03588486e+00 -8.87179196e-01 7.52950788e-01
-2.85363555e-01 3.52026612e-01 -3.61351013e-01 1.82961971e-02
2.47421205e-01 -4.99822289e-01 6.08447433e-01 1.41249800e+00
1.72340184e-01 -1.46389576e-02 -9.09280181e-02 1.04736328e+00
5.20451963e-01 3.09762895e-01 -6.53348625e-01 -4.18179333e-02
-2.60287285e-01 1.13510489e+00 -4.14289117e-01 -1.30167916e-01
-5.99633396e-01 1.17281914e+00 4.15713072e-01 3.05740863e-01
-8.40792418e-01 9.27196145e-02 8.89771760e-01 -3.03455949e-01
4.75613981e-01 -7.37844586e-01 -1.09768212e-02 -1.35270214e+00
4.61755693e-01 -6.91576064e-01 4.31136191e-01 -7.19709814e-01
-1.50810635e+00 2.77287185e-01 4.51492280e-01 -1.22762454e+00
-2.73881078e-01 -1.17200482e+00 -7.38762766e-02 8.70008826e-01
-1.45118415e+00 -1.28367567e+00 -2.19259381e-01 7.78623998e-01
-1.20852217e-02 6.46610791e-03 9.52946126e-01 2.10440770e-01
-4.17019844e-01 4.73867655e-01 1.45867839e-02 3.32486592e-02
8.12775612e-01 -1.26913416e+00 2.29670212e-01 7.50124991e-01
3.80467981e-01 8.48796904e-01 7.50783563e-01 -4.11128104e-01
-1.81312954e+00 -6.19884908e-01 5.56532800e-01 -8.31865966e-01
1.20575893e+00 -5.61365545e-01 -7.06225157e-01 5.63744187e-01
-2.71080256e-01 1.57077357e-01 9.63128686e-01 2.81882107e-01
-7.54812896e-01 4.27647084e-02 -9.37114835e-01 3.72637421e-01
9.42727208e-01 -7.12341249e-01 -7.58233070e-01 -1.78243309e-01
1.72464833e-01 -2.82756001e-01 -1.02611744e+00 5.69383383e-01
8.49243164e-01 -1.14866579e+00 1.02743959e+00 -8.07158470e-01
2.34853551e-01 -4.31481868e-01 -4.26828176e-01 -1.15266323e+00
-2.52311915e-01 -4.82054710e-01 -4.97815013e-01 1.15312004e+00
2.35985711e-01 -4.44862127e-01 7.83802569e-01 8.39011431e-01
2.01741889e-01 -7.84751177e-01 -1.04433882e+00 -1.02366436e+00
4.06252146e-02 -8.94985676e-01 1.90272734e-01 6.79248989e-01
1.61336169e-01 1.15112022e-01 -5.62400937e-01 2.16003776e-01
7.26555824e-01 -8.62896442e-02 9.31456864e-01 -1.16528392e+00
-6.12805486e-01 -4.06960428e-01 -1.16870534e+00 -1.10802996e+00
1.55265331e-01 -8.19148600e-01 -3.97340119e-01 -1.12594795e+00
3.06381077e-01 -1.86750814e-02 -5.51986277e-01 4.00939286e-01
-9.49899703e-02 3.03469658e-01 3.22024345e-01 3.55445147e-02
-5.40187001e-01 6.85096979e-01 1.03354585e+00 -6.00171685e-02
-1.55890658e-01 -3.36154178e-02 -5.29841006e-01 8.17635834e-01
5.28234303e-01 -2.64099538e-01 -5.35733640e-01 -2.75761988e-02
1.29559204e-01 -4.28342164e-01 6.64600551e-01 -1.10048258e+00
2.04235360e-01 6.84861932e-03 3.68482441e-01 -5.07420838e-01
4.61611748e-01 -1.10685468e+00 1.53084680e-01 1.97731256e-01
-4.45195675e-01 -1.73940256e-01 3.82754803e-01 6.65560544e-01
-1.18261486e-01 9.90778431e-02 4.47908908e-01 3.62022519e-01
-5.55583775e-01 8.27808529e-02 -1.98763892e-01 -4.61214595e-02
1.07243812e+00 -1.30008519e-01 -2.33520478e-01 -1.72336429e-01
-8.55294466e-01 -1.53616682e-01 3.38848144e-01 6.76628172e-01
6.24414742e-01 -1.22288430e+00 -5.01021862e-01 3.54030967e-01
4.25254524e-01 -7.13850021e-01 3.11464459e-01 1.32190406e+00
-1.91049635e-01 6.93103850e-01 -2.76230514e-01 -7.81630814e-01
-1.48745990e+00 6.99840069e-01 1.54667959e-01 -4.01306212e-01
-7.55721152e-01 4.42425370e-01 -6.56025391e-03 -3.93638819e-01
2.76166379e-01 -7.31206000e-01 -2.83167213e-01 5.36899548e-03
4.96846139e-01 4.94873762e-01 -1.16844058e-01 -7.48457313e-01
-6.97835982e-01 7.94263184e-01 1.63581688e-02 -5.38983718e-02
1.43556499e+00 -1.15189016e-01 -1.10765077e-01 8.78806949e-01
1.24917161e+00 1.22439675e-02 -1.28684878e+00 -2.21339002e-01
1.34985387e-01 -6.52725697e-01 -6.21308805e-03 -6.99377656e-01
-7.32540786e-01 6.95375025e-01 8.43514860e-01 -1.80580504e-02
1.09548569e+00 2.35925555e-01 1.47657514e-01 6.72002792e-01
2.12374449e-01 -8.41136396e-01 -5.90448268e-02 5.05979240e-01
6.57804728e-01 -1.02001786e+00 3.91725361e-01 -5.02870739e-01
-7.92512178e-01 1.24712801e+00 1.33851841e-01 -4.55468416e-01
1.05565631e+00 5.13325870e-01 1.05558366e-01 -2.43157372e-01
-7.29534090e-01 -4.29333955e-01 6.85266495e-01 8.02978575e-01
3.60741436e-01 -7.61601925e-02 -3.39568645e-01 4.90784019e-01
-2.04631865e-01 7.57334754e-02 2.16994911e-01 8.00638735e-01
-1.56407766e-02 -1.39661205e+00 -1.80299059e-01 1.01802759e-01
-3.03152531e-01 -5.56274923e-03 -7.94915915e-01 9.61910427e-01
-1.62712514e-01 7.90208697e-01 -1.78258836e-01 -3.62894386e-01
6.75042152e-01 1.89180419e-01 7.43769348e-01 -1.07821278e-01
-2.33010724e-01 4.89340462e-02 3.25538255e-02 -1.02442276e+00
-8.11687946e-01 -8.75237405e-01 -1.00493050e+00 8.21677372e-02
-2.45918617e-01 -2.83603281e-01 7.39298999e-01 1.06797409e+00
3.83911610e-01 5.42771578e-01 1.96439713e-01 -5.17670691e-01
-7.82588422e-01 -8.98074985e-01 -6.85053110e-01 7.35775292e-01
1.72685802e-01 -1.08343351e+00 -3.16112876e-01 1.17435999e-01] | [7.82554817199707, 3.891082286834717] |
027203c5-f6f6-4576-b3e6-496b2ef2152a | banglawriting-a-multi-purpose-offline-bangla | 2011.07499 | null | https://arxiv.org/abs/2011.07499v3 | https://arxiv.org/pdf/2011.07499v3.pdf | BanglaWriting: A multi-purpose offline Bangla handwriting dataset | This article presents a Bangla handwriting dataset named BanglaWriting that contains single-page handwritings of 260 individuals of different personalities and ages. Each page includes bounding-boxes that bounds each word, along with the unicode representation of the writing. This dataset contains 21,234 words and 32,787 characters in total. Moreover, this dataset includes 5,470 unique words of Bangla vocabulary. Apart from the usual words, the dataset comprises 261 comprehensible overwriting and 450 handwritten strikes and mistakes. All of the bounding-boxes and word labels are manually-generated. The dataset can be used for complex optical character/word recognition, writer identification, handwritten word segmentation, and word generation. Furthermore, this dataset is suitable for extracting age-based and gender-based variation of handwriting. | ['Muhammad Mohsin Kabir', 'Mazedul Islam Emon', 'M. Ameer Ali', 'Abu Quwsar Ohi', 'M. F. Mridha'] | 2020-11-15 | null | null | null | null | ['handwritten-word-segmentation'] | ['computer-vision'] | [-7.32573345e-02 -3.55758011e-01 -2.55084515e-01 -3.13147426e-01
-6.49301847e-03 -7.84230471e-01 4.94684994e-01 5.15252119e-03
-4.77267176e-01 9.13790882e-01 1.07063293e-01 -3.94109666e-01
3.14807817e-02 -9.02661681e-01 -1.77129701e-01 -7.73347914e-01
5.02196014e-01 7.59644568e-01 -9.08292308e-02 -2.95009375e-01
8.12714338e-01 7.50991523e-01 -1.13487375e+00 -6.64046407e-02
7.33016551e-01 4.38080996e-01 3.08795124e-01 9.95523632e-01
-5.84951520e-01 3.79400760e-01 -1.24389422e+00 -1.04230213e+00
-1.14561521e-01 -2.70264894e-01 -6.10396504e-01 6.22666240e-01
6.05582833e-01 -7.34266400e-01 -5.89723885e-01 8.62950742e-01
6.24863744e-01 -3.04678440e-01 8.61609578e-01 -6.27755821e-01
-1.17253542e+00 9.00342166e-01 -8.60586882e-01 -3.40948664e-02
6.87787756e-02 -7.64636174e-02 5.77907801e-01 -8.61316621e-01
6.39207184e-01 1.26806998e+00 4.96011078e-01 7.91759193e-01
-4.84100074e-01 -6.60447776e-01 -3.00327092e-01 -5.41144051e-02
-1.50863576e+00 -1.65218785e-01 5.03403783e-01 -5.33772171e-01
8.09824526e-01 3.40666026e-01 4.94976908e-01 1.15718174e+00
2.00608060e-01 6.65811002e-01 1.15815675e+00 -6.95895672e-01
-7.10397586e-02 1.39162406e-01 8.78232360e-01 6.34768665e-01
7.70732284e-01 -6.53715432e-01 -5.77253461e-01 6.53669462e-02
8.42771292e-01 -2.32356638e-01 2.07725495e-01 6.39997125e-01
-1.22666252e+00 4.95807052e-01 -8.16931486e-01 2.11012289e-01
1.64726749e-01 -1.87933728e-01 4.49126571e-01 -2.17750192e-01
-2.44781747e-02 1.76170290e-01 -2.34683424e-01 -4.57505614e-01
-9.45070148e-01 4.53138143e-01 6.27669990e-01 1.57925093e+00
4.65614736e-01 3.42838109e-01 -1.51206240e-01 1.32426763e+00
1.50015578e-01 1.16284835e+00 6.92092359e-01 -5.06088078e-01
5.94337821e-01 4.79512215e-01 8.23002756e-02 -1.01635122e+00
-5.12138069e-01 2.77399033e-01 -7.30193734e-01 -4.05107170e-01
7.75761604e-01 -9.72021446e-02 -1.21061635e+00 7.81570554e-01
-1.99216723e-01 -7.67972350e-01 -1.48854092e-01 6.93319142e-01
1.02878881e+00 8.04621279e-01 -1.73890904e-01 -2.03874245e-01
1.80180597e+00 -9.38481748e-01 -1.10814786e+00 -5.92024848e-02
-3.78714465e-02 -1.14599574e+00 1.02223277e+00 6.97628140e-01
-1.02614176e+00 -4.85000253e-01 -1.11677682e+00 -3.54150772e-01
-5.08848250e-01 8.57628882e-01 4.99280691e-01 1.37181544e+00
-4.18963701e-01 4.26063463e-02 -3.71785790e-01 -3.44563514e-01
2.96790868e-01 2.48263553e-01 -2.00464562e-01 1.20099507e-01
-7.93106198e-01 6.46837354e-01 4.97823417e-01 8.17081258e-02
-3.89034778e-01 -2.12515712e-01 -6.07237935e-01 -5.98545909e-01
-3.74496728e-02 1.90373033e-01 7.52553046e-01 -3.72458279e-01
-1.46741724e+00 1.18459046e+00 -1.96926907e-01 1.88630268e-01
6.89450264e-01 7.44557753e-02 -9.31087971e-01 5.47733530e-02
-1.30497456e-01 2.89899170e-01 1.09276605e+00 -9.93017495e-01
-5.70232868e-01 -7.27154553e-01 -6.56814218e-01 -2.03679770e-01
-7.52422392e-01 3.86640608e-01 -9.41639066e-01 -1.25872040e+00
-1.18708394e-01 -6.78204358e-01 4.20013756e-01 -3.98839384e-01
-7.76097178e-01 -2.10934788e-01 7.70592272e-01 -1.54820108e+00
1.58374512e+00 -2.12316275e+00 -2.93058008e-01 5.02494574e-01
5.28663062e-02 4.20058697e-01 -1.37910336e-01 4.76194143e-01
5.94267547e-01 2.09332854e-01 -4.22903784e-02 -6.03675433e-02
1.14921629e-01 4.19675231e-01 -4.15254831e-01 4.46786493e-01
1.82567120e-01 6.88066483e-01 -5.76150894e-01 -8.26775432e-01
-3.15672494e-02 3.62417161e-01 2.29513124e-01 -1.75625414e-01
-3.62866446e-02 -1.94233116e-02 -3.57344687e-01 1.17075503e+00
1.09689820e+00 4.85374540e-01 2.93322295e-01 2.87304640e-01
-1.62965611e-01 -1.29520014e-01 -1.18927896e+00 8.90344977e-01
-3.48499805e-01 8.60255778e-01 -4.11427766e-01 -1.22760661e-01
1.63882697e+00 -3.13537508e-01 -1.54682636e-01 -4.71327573e-01
4.13562059e-01 5.71546257e-01 -1.11742400e-01 -6.46327317e-01
1.43222153e+00 4.68107700e-01 -2.38599807e-01 7.25321591e-01
8.84295776e-02 -8.45735148e-02 8.55755687e-01 1.45911900e-02
4.58488405e-01 -1.96757287e-01 1.39401285e-02 -1.56100616e-01
6.43408537e-01 -1.06852129e-02 3.17469329e-01 6.41316295e-01
2.48982180e-02 7.83366024e-01 5.14480054e-01 -5.87420762e-01
-1.53443635e+00 -1.05240417e+00 -6.08033955e-01 1.00793386e+00
-1.91982295e-02 -2.98585951e-01 -8.92594099e-01 -2.73937643e-01
1.59585074e-01 5.07813513e-01 -3.67832601e-01 4.01514024e-01
-6.63027346e-01 -7.81583726e-01 1.27603507e+00 8.24294209e-01
7.68733203e-01 -1.14291918e+00 -2.27306992e-01 1.13929950e-01
-5.90417795e-02 -1.10649216e+00 -7.18534231e-01 -8.80863965e-02
-6.55816615e-01 -9.03346896e-01 -1.36339045e+00 -1.12082469e+00
1.00381362e+00 -1.30860314e-01 8.61950159e-01 8.11312869e-02
-7.17964053e-01 2.13584881e-02 -4.91096079e-01 -4.17444974e-01
-5.49827874e-01 2.40914777e-01 2.28538558e-01 -2.24322066e-01
8.50793660e-01 2.27884948e-01 7.79048726e-02 3.77152652e-01
-6.34350419e-01 -2.29358807e-01 4.48299915e-01 8.77807617e-01
2.19383493e-01 1.47439852e-01 1.83687210e-01 -7.70897388e-01
8.26642096e-01 1.13124400e-01 -5.23770928e-01 7.27547407e-01
-4.44859087e-01 -1.78579330e-01 6.02333903e-01 -5.49947560e-01
-1.13724434e+00 -2.16821060e-01 -9.63253677e-02 4.37007040e-01
-2.62525618e-01 -1.87770817e-02 -2.21277684e-01 1.66586831e-01
3.98862958e-01 8.27509880e-01 -1.85626373e-02 -6.20821178e-01
2.41961494e-01 1.75512028e+00 1.12427187e+00 -7.49330521e-01
4.79536176e-01 6.11518361e-02 -3.88379544e-01 -1.52240098e+00
4.10458678e-03 -2.80609369e-01 -1.09753287e+00 -3.69592607e-01
6.68535769e-01 -5.21606445e-01 -7.63764143e-01 1.65954316e+00
-1.20200372e+00 -3.65411974e-02 7.95535743e-03 6.31940886e-02
6.26752805e-03 7.38699317e-01 -9.80968714e-01 -1.16849518e+00
-4.00359482e-01 -8.91833603e-01 9.57542121e-01 6.79159701e-01
-5.21400154e-01 -7.28473186e-01 -2.50808656e-01 5.30737102e-01
-2.02229440e-01 -2.39536777e-01 1.18528605e+00 -6.87687874e-01
1.63153991e-01 -4.33109879e-01 -3.82775664e-01 4.42643434e-01
3.38058412e-01 7.96462297e-01 -6.56304419e-01 -1.03719935e-01
-7.38552928e-01 -3.47749650e-01 8.08990300e-01 -4.47524041e-02
1.10106063e+00 -2.67433196e-01 1.28817901e-01 3.22772801e-01
1.15821218e+00 5.66067815e-01 8.72755885e-01 2.88411021e-01
9.06062245e-01 6.43823624e-01 4.14926022e-01 1.09823298e+00
4.98155475e-01 1.07251860e-01 -2.37194672e-01 3.85519356e-01
-1.20297350e-01 -1.49340749e-01 4.31198835e-01 1.13520765e+00
-5.17902076e-01 -3.21667194e-01 -1.40600789e+00 6.95348680e-01
-9.97167766e-01 -7.54821241e-01 -7.63328731e-01 1.87690091e+00
1.25772011e+00 -9.40303802e-02 4.54169601e-01 5.45945585e-01
1.10209405e+00 2.13084102e-01 -4.32715923e-01 -9.94407952e-01
-6.54960573e-01 2.48896047e-01 9.74209785e-01 3.01832974e-01
-8.67216766e-01 1.09021175e+00 7.30164862e+00 1.12219512e+00
-9.44093049e-01 -5.09045243e-01 6.27333403e-01 3.25281620e-01
-1.81271493e-01 -6.90357804e-01 -1.65001619e+00 8.39076340e-01
5.50675690e-01 -2.44058575e-03 4.63308275e-01 5.15334964e-01
-1.28601849e-01 -4.20667261e-01 -4.14494246e-01 1.18801880e+00
5.86242676e-01 -1.07908499e+00 5.24241149e-01 5.31606153e-02
8.81352246e-01 -8.54728878e-01 1.96693510e-01 -1.72139153e-01
1.34549648e-01 -1.14617002e+00 1.03209674e+00 5.08622587e-01
1.27152455e+00 -9.59056020e-01 8.19835544e-01 4.93337624e-02
-5.88548481e-01 3.35275084e-02 -6.89445972e-01 2.06627194e-02
-1.63825005e-01 5.07111967e-01 -5.60328960e-01 2.09584579e-01
5.42576849e-01 3.88453662e-01 -1.02545762e+00 5.40377736e-01
-1.89964965e-01 6.70783818e-01 -1.11157402e-01 -7.30677247e-01
7.48028830e-02 -4.48031843e-01 3.17958146e-02 1.66251040e+00
2.73578495e-01 -1.49381250e-01 -3.87496948e-01 4.31381464e-01
-4.31941748e-02 3.47337663e-01 -4.84991372e-02 -8.24246049e-01
7.93325782e-01 1.10539484e+00 -1.23081720e+00 -3.73800516e-01
-1.23662993e-01 1.10628510e+00 -2.10993826e-01 2.33343154e-01
-6.22091949e-01 -1.15031743e+00 5.31500995e-01 -5.34491204e-02
2.69012034e-01 -6.64750159e-01 -9.65566099e-01 -7.45172024e-01
1.51770890e-01 -9.96656537e-01 1.15257993e-01 -4.77750748e-01
-1.30692816e+00 3.31316501e-01 -2.34336823e-01 -6.00012362e-01
9.10465941e-02 -1.36578119e+00 -3.13496530e-01 1.11372733e+00
-7.87385046e-01 -1.28066170e+00 -3.94258529e-01 4.19642985e-01
7.27499187e-01 -9.46045220e-01 6.82200670e-01 2.05056161e-01
-1.01212037e+00 9.95460212e-01 5.02426863e-01 8.65120292e-01
8.69832039e-01 -1.20176029e+00 7.42705047e-01 5.68184912e-01
-1.36037484e-01 7.95595229e-01 4.79428649e-01 -1.01196146e+00
-1.43933201e+00 -9.44029748e-01 1.50870991e+00 -4.52444047e-01
7.10824609e-01 -5.76191604e-01 -5.12425184e-01 4.03650612e-01
1.65604621e-01 -5.65941334e-01 7.72432745e-01 -2.46948078e-01
-4.58929002e-01 3.51858251e-02 -1.08669055e+00 7.07815707e-01
6.07229054e-01 -5.12084126e-01 -5.02285123e-01 3.34870785e-01
-1.09774277e-01 -6.29423499e-01 -8.00833106e-01 -4.72009152e-01
1.00068486e+00 -4.96868044e-01 5.75506210e-01 -5.19961178e-01
8.15594494e-01 -3.38766091e-02 8.70522112e-02 -7.28400409e-01
-1.80162266e-01 -5.03301144e-01 -5.24543077e-02 1.85229516e+00
3.32743287e-01 -3.73446375e-01 7.23804832e-01 5.33261001e-01
1.63061246e-01 -4.84256387e-01 -4.16920930e-01 -1.13857961e+00
4.40441489e-01 -2.63755500e-01 9.30241048e-01 8.15557837e-01
-1.25443742e-01 -1.82342440e-01 -7.68688977e-01 -2.83388913e-01
6.60048604e-01 -2.25890890e-01 6.50060117e-01 -8.82731378e-01
2.41129361e-02 -6.40138447e-01 -4.82585907e-01 -9.45262969e-01
1.02889344e-01 -5.58300853e-01 -1.74369246e-01 -1.28444910e+00
2.59960264e-01 -3.08218032e-01 5.99724174e-01 5.54128647e-01
1.55766932e-02 6.60422921e-01 6.39609173e-02 1.89021170e-01
8.12901482e-02 -8.23407099e-02 1.22409546e+00 -5.31139076e-01
-2.73814090e-02 -2.20622435e-01 -1.66764498e-01 5.37624657e-01
9.56826985e-01 3.38473246e-02 6.35812208e-02 -5.65350950e-01
3.07269573e-01 -4.68794703e-01 -3.11001502e-02 -3.57547343e-01
8.01175162e-02 -4.13726121e-01 7.76817024e-01 -9.74172413e-01
-1.56378642e-01 -2.48600632e-01 -2.11230427e-01 3.55537832e-01
-1.38330758e-01 2.15333581e-01 2.34803576e-02 -3.41092842e-03
9.76967737e-02 -9.42578673e-01 5.66320419e-01 1.20156147e-02
-8.61591816e-01 -9.66426730e-02 -9.62591887e-01 1.53260127e-01
1.11920953e+00 -8.70514631e-01 -5.36781192e-01 2.06848487e-01
-2.38069057e-01 -1.00500487e-01 4.58030462e-01 5.57220101e-01
6.72282398e-01 -1.14857137e+00 -9.89335716e-01 3.71306062e-01
2.20765084e-01 -4.21118528e-01 1.64375558e-01 1.42359629e-01
-1.53029799e+00 3.35094601e-01 -7.46680319e-01 -1.06210016e-01
-1.49422050e+00 1.77037433e-01 -2.30785325e-01 3.65417331e-01
-4.30502146e-01 7.70965040e-01 -8.96340072e-01 -2.54206896e-01
4.32979256e-01 -6.13841042e-03 -3.90457153e-01 5.57235241e-01
8.98904324e-01 1.10969830e+00 3.35244909e-02 -9.42420959e-01
-2.68810362e-01 7.92808056e-01 -3.98605615e-01 -2.37527013e-01
8.06570232e-01 -1.16673574e-01 -5.47141910e-01 6.01961911e-01
7.02606082e-01 5.81513405e-01 -8.48762035e-01 1.82185724e-01
9.86086577e-02 -7.17714489e-01 -5.49767256e-01 -8.74645472e-01
-9.30846691e-01 8.23074698e-01 1.91613838e-01 -2.26712838e-01
8.24855328e-01 -3.37040722e-01 9.11410451e-01 5.19234657e-01
4.25167412e-01 -1.92929959e+00 -2.53367156e-01 1.00673473e+00
7.81295955e-01 -8.56957316e-01 2.29136631e-01 -1.95909142e-01
-8.78485441e-01 1.89266932e+00 5.09358048e-01 2.05686957e-01
2.63930112e-01 6.41348898e-01 3.74394268e-01 3.55394423e-01
1.04634136e-01 1.62722155e-01 1.60029493e-02 1.04585540e+00
4.68782485e-01 2.39855915e-01 -9.37010646e-01 1.08405173e+00
-7.76692986e-01 -4.69573796e-01 9.02094781e-01 8.45765889e-01
-2.77154982e-01 -1.36024868e+00 -8.68519604e-01 5.15084505e-01
-5.43792725e-01 -7.37115294e-02 -8.49550784e-01 6.11522734e-01
1.40983626e-01 8.48636687e-01 4.30147201e-01 -5.32604694e-01
-1.58171097e-04 1.93896368e-01 7.18901694e-01 -1.96718335e-01
-6.81817353e-01 -2.02120975e-01 6.95501789e-02 5.30002296e-01
1.34132355e-01 -7.22674668e-01 -1.39427686e+00 -1.00891006e+00
-1.75676480e-01 -2.48426467e-01 8.17607403e-01 6.41733527e-01
-3.76580089e-01 1.26679897e-01 4.85441715e-01 -3.07262599e-01
-4.72969115e-01 -1.00332212e+00 -1.37660086e+00 7.32419118e-02
-8.23296905e-02 -8.50663856e-02 1.56949490e-01 4.90004599e-01] | [11.859137535095215, 2.5770435333251953] |
880012af-ec92-46f5-b9de-9d740acf9c3c | smix-enhancing-centralized-value-functions | 1911.04094 | null | https://arxiv.org/abs/1911.04094v5 | https://arxiv.org/pdf/1911.04094v5.pdf | SMIX($λ$): Enhancing Centralized Value Functions for Cooperative Multi-Agent Reinforcement Learning | Learning a stable and generalizable centralized value function (CVF) is a crucial but challenging task in multi-agent reinforcement learning (MARL), as it has to deal with the issue that the joint action space increases exponentially with the number of agents in such scenarios. This paper proposes an approach, named SMIX(${\lambda}$), to address the issue using an efficient off-policy centralized training method within a flexible learner search space. As importance sampling for such off-policy training is both computationally costly and numerically unstable, we proposed to use the ${\lambda}$-return as a proxy to compute the TD error. With this new loss function objective, we adopt a modified QMIX network structure as the base to train our model. By further connecting it with the ${Q(\lambda)}$ approach from an unified expectation correction viewpoint, we show that the proposed SMIX(${\lambda}$) is equivalent to ${Q(\lambda)}$ and hence shares its convergence properties, while without being suffered from the aforementioned curse of dimensionality problem inherent in MARL. Experiments on the StarCraft Multi-Agent Challenge (SMAC) benchmark demonstrate that our approach not only outperforms several state-of-the-art MARL methods by a large margin, but also can be used as a general tool to improve the overall performance of other CTDE-type algorithms by enhancing their CVFs. | ['Chao Wen', 'Yuhui Wang', 'Xiaoyang Tan', 'Xinghu Yao'] | 2019-11-11 | null | null | null | null | ['smac-1', 'smac'] | ['playing-games', 'playing-games'] | [-1.48376673e-01 -6.06911704e-02 -7.09831864e-02 3.38167548e-02
-1.14070392e+00 -3.76926750e-01 4.98074085e-01 3.49952072e-01
-1.02136123e+00 1.26451862e+00 -3.51473540e-01 -3.23871523e-01
-6.68395221e-01 -7.09572494e-01 -9.46742058e-01 -1.06733751e+00
-1.32547498e-01 5.50616026e-01 1.98314309e-01 -4.75605100e-01
1.66502520e-01 1.54416159e-01 -1.40902853e+00 -4.08842981e-01
1.20352066e+00 1.23205924e+00 2.19770625e-01 1.83894962e-01
2.01033905e-01 8.49297225e-01 -6.73506260e-01 -5.58226526e-01
3.77906233e-01 -4.75460023e-01 -7.39037156e-01 -7.09291771e-02
2.09525511e-01 -1.79609418e-01 8.32955241e-02 1.23110831e+00
6.67881787e-01 7.40805686e-01 4.45019633e-01 -1.35997570e+00
-2.37653509e-01 5.83777606e-01 -7.18572497e-01 -3.55699621e-02
-9.52065066e-02 3.54147971e-01 1.12989688e+00 -5.97758532e-01
3.55530262e-01 1.10172808e+00 6.22170091e-01 5.58926105e-01
-1.14340234e+00 -4.14710641e-01 4.75483239e-01 1.67430818e-01
-1.13065922e+00 -4.70593236e-02 7.06929684e-01 -2.71512598e-01
7.52105296e-01 3.76245491e-02 4.91804153e-01 1.00889623e+00
1.26383960e-01 9.63613510e-01 1.17935944e+00 -3.96422982e-01
6.38316751e-01 6.17457926e-02 -2.81937152e-01 9.25229728e-01
1.78817406e-01 3.42317939e-01 -3.04365605e-01 -2.40041465e-01
6.47822797e-01 -2.91008055e-01 -1.32470474e-01 -8.08766484e-01
-9.63482499e-01 1.26484251e+00 3.63858670e-01 8.56192484e-02
-5.20803452e-01 5.43475449e-01 5.14602005e-01 4.58567202e-01
4.19910192e-01 5.89897573e-01 -5.55331290e-01 -3.10381621e-01
-5.54156065e-01 6.63754940e-01 5.74651837e-01 5.29166102e-01
7.46567726e-01 3.25594962e-01 -1.45249546e-01 8.61732066e-01
2.41077662e-01 3.70123208e-01 3.10193688e-01 -1.26601362e+00
6.87900126e-01 3.81422549e-01 5.23169637e-01 -5.04803360e-01
-4.28273261e-01 -6.76985264e-01 -7.84695864e-01 7.74759710e-01
6.54403985e-01 -5.54193377e-01 -2.79244602e-01 2.13342452e+00
6.53021395e-01 -1.02718724e-02 1.43726364e-01 8.28623593e-01
-2.22169049e-02 3.95924151e-01 1.35998681e-01 -4.77620244e-01
1.05051422e+00 -1.01682270e+00 -4.11539197e-01 7.09557068e-03
6.31005347e-01 -4.13408190e-01 1.22193623e+00 5.74384749e-01
-1.19976509e+00 -2.71813810e-01 -1.02740669e+00 4.64932203e-01
-1.65143684e-01 3.70211713e-02 7.63843775e-01 4.87071902e-01
-9.68054533e-01 9.92188513e-01 -8.51440549e-01 7.67972544e-02
3.90746564e-01 3.64096522e-01 1.49919870e-04 3.81518453e-01
-1.21327841e+00 1.18157852e+00 5.04874289e-01 -3.58336866e-02
-1.02844369e+00 -8.04569244e-01 -7.18355298e-01 3.89171541e-02
9.33223069e-01 -6.37869537e-01 1.43137491e+00 -9.42755222e-01
-2.10877204e+00 7.43776336e-02 3.50015074e-01 -6.65371418e-01
8.43152165e-01 -3.19097847e-01 3.65273617e-02 1.07565954e-01
4.67308648e-02 3.23364556e-01 1.02828300e+00 -1.17840600e+00
-8.27654541e-01 -2.25712880e-01 3.36258620e-01 3.50025982e-01
-2.64049828e-01 -2.29373485e-01 1.53499141e-01 -7.43590355e-01
-5.31418622e-01 -7.67951787e-01 -4.59744394e-01 -1.39499784e-01
1.82359651e-01 -6.02600276e-01 4.96463746e-01 -3.20076823e-01
1.15196991e+00 -1.75685561e+00 5.26765704e-01 3.06109756e-01
4.18468341e-02 5.40839911e-01 -1.39711738e-01 5.04837275e-01
2.59519398e-01 -2.66970158e-01 -5.09155869e-01 -3.98346603e-01
3.58973801e-01 1.28892154e-01 -1.21381588e-01 5.99850178e-01
6.26444351e-03 6.42699778e-01 -1.21172047e+00 -3.87813747e-01
2.39103407e-01 3.44250888e-01 -8.01358879e-01 2.00998336e-02
-7.00850844e-01 4.05418962e-01 -6.13774419e-01 2.94255167e-01
3.77014071e-01 -1.95181802e-01 2.55695492e-01 1.11640371e-01
-2.51549244e-01 -1.33942470e-01 -1.45897996e+00 1.82698572e+00
-5.04737496e-01 6.40663430e-02 2.23471582e-01 -1.38951182e+00
8.64833832e-01 2.41872743e-01 8.22417676e-01 -7.04323113e-01
1.45681769e-01 3.25549930e-01 -9.83929560e-02 -1.99605957e-01
4.27085191e-01 -2.95425475e-01 -1.71046585e-01 4.95232046e-01
1.57358289e-01 7.76060224e-02 4.44964647e-01 -1.43910050e-01
1.03600979e+00 5.08787930e-01 2.90041178e-01 -3.86912644e-01
8.51052105e-01 -1.65806308e-01 6.55348182e-01 8.63894641e-01
-2.65710443e-01 -1.26389012e-01 5.22201538e-01 -3.44930947e-01
-9.69973207e-01 -1.08230841e+00 9.40577090e-02 1.22708058e+00
8.88634101e-02 -2.28892624e-01 -7.35594988e-01 -9.80915725e-01
1.88517094e-01 7.31641591e-01 -4.37195599e-01 -2.27458045e-01
-6.79582298e-01 -8.89320850e-01 4.56128955e-01 3.23462337e-01
5.95063627e-01 -1.18651533e+00 -9.42674279e-01 6.10476732e-01
-6.80401698e-02 -6.43578291e-01 -4.62757856e-01 3.57464612e-01
-5.11159599e-01 -9.24976885e-01 -8.12837481e-01 -3.83339286e-01
3.15839857e-01 -2.17323795e-01 1.07080412e+00 -1.45696864e-01
5.24824578e-03 5.93695164e-01 -5.19430637e-01 -3.96361351e-01
-3.98425668e-01 5.90112917e-02 2.50646442e-01 1.41958609e-01
-4.39928994e-02 -6.35847211e-01 -7.07568228e-01 1.29805163e-01
-9.70733047e-01 -3.52342516e-01 5.43535888e-01 1.17616403e+00
5.43382764e-01 9.65674669e-02 1.15574944e+00 -5.51003456e-01
9.29996908e-01 -3.20742548e-01 -1.27601862e+00 2.85472870e-01
-9.11304832e-01 4.94762659e-01 1.15257704e+00 -4.84258652e-01
-9.96189475e-01 -1.69530258e-01 -1.51274666e-01 -5.07467270e-01
2.94553369e-01 5.42878449e-01 6.12202547e-02 -1.59930244e-01
5.06245852e-01 3.09719324e-01 3.73345256e-01 -4.73999381e-01
3.69022399e-01 1.78720906e-01 2.95084208e-01 -9.95155811e-01
7.15032995e-01 3.04767489e-01 3.68523091e-01 -3.87325883e-01
-8.49659383e-01 -1.57490715e-01 -8.70778859e-02 -1.36326402e-01
5.05750179e-01 -6.64278805e-01 -1.48062682e+00 2.75173575e-01
-8.63168836e-01 -6.11833930e-01 -5.25264204e-01 6.73520744e-01
-1.03157210e+00 5.13269603e-01 -3.51180106e-01 -9.35719192e-01
-4.44594085e-01 -1.17886424e+00 6.31605923e-01 2.24960998e-01
2.25518674e-01 -1.18327677e+00 3.56500030e-01 2.19997272e-01
4.12714154e-01 3.90317827e-01 8.69056761e-01 -4.32298213e-01
-3.74254435e-01 1.20534472e-01 1.60867184e-01 6.30482733e-01
-9.37180594e-02 -4.17958885e-01 -4.96071756e-01 -8.47767413e-01
8.74204114e-02 -7.35597014e-01 6.91137433e-01 2.79053360e-01
1.08232546e+00 -4.36927110e-01 2.17580706e-01 3.67450982e-01
1.61443710e+00 2.16526598e-01 1.82949200e-01 7.05105662e-01
2.22494140e-01 2.56162584e-01 9.37491894e-01 9.99819398e-01
4.83983368e-01 7.68454134e-01 7.55374551e-01 2.42292181e-01
4.23955292e-01 -1.92799374e-01 7.29549289e-01 4.90081280e-01
-2.00103372e-01 -7.11307377e-02 -5.48376262e-01 2.73994118e-01
-2.30284905e+00 -1.03253245e+00 4.90244955e-01 2.44952035e+00
1.11693656e+00 -4.16245870e-02 4.15077448e-01 -1.04143940e-01
3.78523201e-01 1.87415466e-01 -8.70308518e-01 -4.17030662e-01
-1.45898676e-02 5.53946435e-01 4.63388145e-01 5.72608888e-01
-1.11866510e+00 8.28304052e-01 4.66325760e+00 1.17642975e+00
-1.06518996e+00 2.74515361e-01 1.89804479e-01 -1.98031530e-01
7.65953287e-02 -1.18577138e-01 -7.78228223e-01 6.59169197e-01
7.35224962e-01 -1.36659056e-01 8.04875553e-01 9.63020265e-01
3.05092126e-01 -1.95510834e-01 -9.35875297e-01 8.10817063e-01
-2.67285973e-01 -1.23091924e+00 -2.28144407e-01 1.00425407e-02
8.12599897e-01 -4.74547446e-02 1.94505066e-01 6.90933526e-01
6.37616992e-01 -8.48977327e-01 7.64442623e-01 3.47386986e-01
4.99137193e-01 -1.06842351e+00 5.71488559e-01 5.74782968e-01
-1.21546280e+00 -3.30607653e-01 -3.48769784e-01 -1.31305621e-03
1.04631307e-02 2.33719081e-01 -4.31871831e-01 8.27012718e-01
4.43176001e-01 3.39554459e-01 -1.66043594e-01 1.04384768e+00
-2.63147503e-01 2.17504874e-01 -2.70548612e-01 -2.67308265e-01
7.45975196e-01 -4.68950689e-01 6.30025744e-01 6.43042862e-01
3.76858503e-01 -3.10507774e-01 4.03978795e-01 7.95524001e-01
-1.23672768e-01 1.60010651e-01 -1.86681077e-01 1.94243938e-01
3.49913359e-01 1.18314040e+00 -3.23197722e-01 -1.65811583e-01
-2.90671617e-01 6.91463709e-01 7.49837995e-01 3.47618520e-01
-1.05899405e+00 -3.41528744e-01 7.32332528e-01 -3.10990602e-01
6.25129342e-01 -1.96923852e-01 2.33019978e-01 -1.07911885e+00
3.46412152e-01 -1.02794313e+00 4.99809206e-01 -7.54861236e-02
-1.32489049e+00 3.91089410e-01 5.28275920e-03 -1.49105847e+00
-5.27503371e-01 -5.16718626e-01 -2.76459128e-01 5.42522967e-01
-1.79259574e+00 -8.00396860e-01 1.96899906e-01 8.60592484e-01
3.75015616e-01 -3.39260697e-01 7.54224718e-01 3.01249295e-01
-5.39529622e-01 8.02309394e-01 7.08383203e-01 -2.65614778e-01
4.26739454e-01 -1.45704305e+00 -4.13285702e-01 4.93272007e-01
-2.02047542e-01 1.80695459e-01 7.04006433e-01 -2.07936838e-01
-1.46184695e+00 -1.18700051e+00 2.55969822e-01 -1.99683771e-01
8.43694448e-01 -3.23749147e-02 -6.06164038e-01 5.14626205e-01
2.32456356e-01 1.05127739e-02 3.30767006e-01 -1.63639598e-02
9.43022966e-03 -4.46479321e-01 -1.14744508e+00 5.87761581e-01
6.54042840e-01 -3.51824611e-02 -4.28820401e-01 2.96222985e-01
6.66901767e-01 -2.27301851e-01 -1.10133982e+00 3.41772676e-01
2.61173755e-01 -8.75354171e-01 9.26153660e-01 -6.11531794e-01
2.51236856e-01 -1.80584937e-01 4.46863566e-03 -1.76957190e+00
-1.06700182e-01 -1.00655258e+00 -3.90440226e-01 9.70655501e-01
2.67610759e-01 -8.56361985e-01 6.53074086e-01 2.44082734e-01
-1.60941586e-01 -1.11065388e+00 -1.44524419e+00 -1.12347448e+00
4.90942240e-01 -1.64943248e-01 3.38881195e-01 6.58454955e-01
2.92817075e-02 2.05930412e-01 -6.53305888e-01 4.40855138e-02
1.01592505e+00 7.74379075e-02 6.43923163e-01 -1.00753522e+00
-6.79315031e-01 -6.47676766e-01 1.22400060e-01 -9.65342224e-01
3.32361400e-01 -6.61942065e-01 -7.18695596e-02 -1.15722883e+00
-1.41112223e-01 -7.41081178e-01 -5.80997169e-01 3.50346327e-01
-2.12432817e-01 -3.12950730e-01 4.85946238e-01 -6.60957247e-02
-1.01690388e+00 1.21349216e+00 1.25996828e+00 -5.86806505e-04
-2.36828029e-01 1.69674367e-01 -4.89568383e-01 8.17298234e-01
8.73469353e-01 -4.75671530e-01 -5.04508793e-01 -1.95170209e-01
4.27765876e-01 3.32951188e-01 4.87454981e-01 -9.57657397e-01
2.35556468e-01 -4.79111642e-01 -1.43568754e-01 -1.94116578e-01
4.72329885e-01 -8.21968198e-01 -2.44242191e-01 6.16348445e-01
-3.39924097e-01 1.66406572e-01 1.99529584e-02 5.20269573e-01
-1.33542061e-01 -5.44835567e-01 1.05006540e+00 -2.71056473e-01
-5.75546145e-01 3.55104089e-01 -2.43688747e-01 3.34123701e-01
1.21110594e+00 2.67211825e-01 -2.71100193e-01 -2.22365350e-01
-4.96586591e-01 6.50410354e-01 1.24551497e-01 1.61038399e-01
4.71863300e-01 -1.28117061e+00 -5.70160866e-01 -7.70629868e-02
-1.69490248e-01 5.18764071e-02 1.70678377e-01 9.69949245e-01
-2.05558077e-01 2.74548948e-01 -1.38090551e-01 -2.46501684e-01
-7.50941157e-01 4.85134304e-01 5.14694929e-01 -9.59733009e-01
-4.62213129e-01 7.14011312e-01 -2.55627036e-01 -4.89025831e-01
5.88191271e-01 -1.71951242e-02 2.65701930e-03 4.22813706e-02
2.98897505e-01 5.66672921e-01 -1.63701057e-01 -6.54917732e-02
-1.44116580e-01 3.96230221e-01 -6.89847097e-02 -3.25686544e-01
1.54617143e+00 -6.13911673e-02 1.96861997e-01 2.24183068e-01
1.00351703e+00 -4.19423044e-01 -1.74061012e+00 -4.50691909e-01
8.95778984e-02 -2.35069737e-01 -6.36753142e-02 -8.55086625e-01
-9.99835491e-01 5.81782043e-01 5.94712973e-01 6.54222295e-02
9.27607119e-01 -2.44106114e-01 6.27001166e-01 5.91891944e-01
7.15937495e-01 -1.62066674e+00 3.33479375e-01 6.22987032e-01
8.05598736e-01 -1.29323220e+00 1.12325447e-02 3.18994999e-01
-7.77810931e-01 1.04218972e+00 6.82889879e-01 -3.31168413e-01
3.18370402e-01 -6.02976084e-02 -1.57523856e-01 -9.15606041e-03
-7.85752892e-01 -2.49223441e-01 -2.39007518e-01 3.85862499e-01
-1.23045165e-02 -3.19987908e-02 -6.70293868e-01 5.24105847e-01
1.72441751e-02 2.16375291e-02 2.95612514e-01 1.06865752e+00
-4.05505598e-01 -1.45414662e+00 -1.51362851e-01 2.01571479e-01
-4.66398180e-01 1.37256116e-01 3.23753148e-01 9.47190344e-01
1.07071705e-01 7.92652488e-01 -3.61042410e-01 -2.36481391e-02
2.74378449e-01 -1.06892742e-01 5.88673651e-01 -4.81662303e-02
-8.32907438e-01 -1.64609589e-02 -2.04766676e-01 -6.06867135e-01
-6.01430595e-01 -5.48279226e-01 -1.30274892e+00 -2.36284003e-01
-5.29391021e-02 4.80284959e-01 4.72128987e-01 9.99669909e-01
2.01557830e-01 5.59500635e-01 8.36627007e-01 -6.28195882e-01
-1.66283453e+00 -8.08299184e-01 -6.82484329e-01 3.68802458e-01
3.37467670e-01 -1.08558691e+00 -2.70612329e-01 -5.30362785e-01] | [4.028697967529297, 2.2886595726013184] |
a2e76229-3ed0-48ec-89da-83d56567169a | audio-classification-of-bit-representation | 1904.04364 | null | https://arxiv.org/abs/1904.04364v2 | https://arxiv.org/pdf/1904.04364v2.pdf | Audio Classification of Bit-Representation Waveform | This study investigated the waveform representation for audio signal classification. Recently, many studies on audio waveform classification such as acoustic event detection and music genre classification have been published. Most studies on audio waveform classification have proposed the use of a deep learning (neural network) framework. Generally, a frequency analysis method such as Fourier transform is applied to extract the frequency or spectral information from the input audio waveform before inputting the raw audio waveform into the neural network. In contrast to these previous studies, in this paper, we propose a novel waveform representation method, in which audio waveforms are represented as a bit sequence, for audio classification. In our experiment, we compare the proposed bit representation waveform, which is directly given to a neural network, to other representations of audio waveforms such as a raw audio waveform and a power spectrum with two classification tasks: one is an acoustic event classification task and the other is a sound/music classification task. The experimental results showed that the bit representation waveform achieved the best classification performance for both the tasks. | ['Hiromitsu Nishizaki', 'Naoki Sawada', 'Masaki Okawa', 'Takuya Saito'] | 2019-04-08 | null | null | null | null | ['genre-classification', 'music-classification'] | ['computer-vision', 'music'] | [ 6.96820498e-01 -4.06432092e-01 4.43286926e-01 -3.15705240e-01
-7.54225552e-01 -4.28011745e-01 1.65298268e-01 2.98942536e-01
-2.43407756e-01 5.92510462e-01 6.24206327e-02 -2.62789071e-01
-2.01330855e-01 -7.44437814e-01 -2.42807657e-01 -6.69892192e-01
-3.19229424e-01 -3.07364076e-01 1.57985032e-01 9.93252248e-02
5.16888142e-01 3.62755984e-01 -1.88031423e+00 5.81316590e-01
3.42814863e-01 1.63383579e+00 4.43504367e-04 1.01495576e+00
-3.17743510e-01 6.02091372e-01 -1.17141283e+00 -7.37713724e-02
4.09706086e-02 -5.15140355e-01 -4.86847222e-01 -4.93238330e-01
1.71947643e-01 -1.45893365e-01 -2.89308697e-01 1.09202540e+00
8.51675272e-01 5.17692626e-01 8.02256823e-01 -1.29483533e+00
-3.92525971e-01 1.01192820e+00 -4.82466668e-01 5.43236375e-01
5.01639545e-01 -3.37831318e-01 8.41265380e-01 -7.66592383e-01
-3.88392769e-02 1.06544816e+00 8.56158793e-01 -3.27898562e-02
-7.69758403e-01 -1.11869550e+00 -3.04594100e-01 7.38001466e-01
-1.45899570e+00 -2.64412314e-01 1.55003250e+00 -5.21705151e-01
7.31880367e-01 2.69437015e-01 6.78756952e-01 6.31447911e-01
3.24152917e-01 7.55104423e-01 7.80524969e-01 -6.38586640e-01
3.25052828e-01 -6.99772015e-02 2.23849952e-01 3.65861282e-02
-2.37868667e-01 9.66727361e-02 -5.53786576e-01 -4.20393169e-01
4.08490181e-01 -3.25704783e-01 -3.81447822e-01 4.36409473e-01
-7.25413680e-01 8.13282609e-01 2.41695061e-01 5.54784894e-01
-6.88853741e-01 2.89468318e-01 8.51624727e-01 5.20011961e-01
3.70170414e-01 1.67540982e-01 -8.21085125e-02 -4.17938322e-01
-1.05171168e+00 3.27285796e-01 8.30432892e-01 1.20238133e-01
1.68217957e-01 7.46110439e-01 -9.12517831e-02 9.98761833e-01
2.92093426e-01 2.45472163e-01 1.05577421e+00 -3.50999951e-01
1.91106528e-01 1.84746847e-01 -1.22713163e-01 -1.39236617e+00
-3.57830971e-01 -2.28235140e-01 -6.74093068e-01 2.74019510e-01
2.30391666e-01 -3.60800147e-01 -6.54052258e-01 1.47779977e+00
1.10452324e-01 6.24985218e-01 1.34353340e-01 9.88072872e-01
1.16073167e+00 1.18387818e+00 -1.42316639e-01 -1.26536697e-01
1.42878282e+00 -5.54637849e-01 -1.02040946e+00 4.30527180e-01
-4.95982021e-02 -1.16039789e+00 8.18124235e-01 9.47964251e-01
-1.03312159e+00 -9.77403164e-01 -1.39391208e+00 1.64852515e-01
-6.34039164e-01 3.46172780e-01 4.19182569e-01 9.04036522e-01
-6.95322752e-01 5.51627040e-01 -3.64789307e-01 4.46019135e-03
1.30501285e-01 2.41049901e-01 -1.19405210e-01 4.47018713e-01
-1.49998510e+00 4.28601116e-01 7.82086253e-01 3.31572257e-02
-7.33842492e-01 -6.41047001e-01 -8.75279665e-01 4.43870693e-01
-2.06933349e-01 1.10973500e-01 1.59772718e+00 -1.09970689e+00
-1.76613438e+00 2.45897681e-01 2.12308913e-01 -6.59204185e-01
-1.58549190e-01 -1.87170699e-01 -1.00500131e+00 1.76524565e-01
-5.35285294e-01 3.52384865e-01 9.85072792e-01 -7.18304694e-01
-7.34174252e-01 -6.30612224e-02 -9.31941420e-02 -2.07175910e-01
-5.71406484e-01 3.09903592e-01 4.05544728e-01 -1.04487598e+00
-4.98598069e-02 -5.42960942e-01 4.59816307e-01 -2.86368847e-01
-1.91441402e-01 -4.91544515e-01 1.05539012e+00 -7.76203930e-01
1.48063922e+00 -2.43650365e+00 -1.60809293e-01 3.20294082e-01
-3.67554277e-01 4.41508323e-01 -1.39398113e-01 5.62376261e-01
-6.20459974e-01 -1.65683761e-01 -2.24893168e-01 2.53829435e-02
-3.90534699e-02 -2.76297837e-01 -7.12099433e-01 2.66636431e-01
2.81746477e-01 3.23394388e-01 -7.02356815e-01 -1.12058237e-01
3.75562087e-02 6.85662329e-01 -1.40924886e-01 3.89374793e-01
5.54617867e-02 1.52632073e-01 -1.02618739e-01 3.77255082e-01
8.00876081e-01 5.18042862e-01 -1.29718125e-01 -3.92262191e-01
-2.30742916e-01 6.04362428e-01 -1.53184378e+00 1.41727912e+00
-5.37067890e-01 1.02795815e+00 -1.04535138e-02 -1.11701512e+00
1.40125537e+00 9.96847391e-01 5.26143670e-01 -3.37722778e-01
5.70195019e-01 2.24824384e-01 3.85359675e-01 -6.55071259e-01
5.14261246e-01 -2.34577730e-01 3.61129865e-02 6.59534454e-01
6.04652427e-02 -1.18109159e-01 9.32775252e-03 -6.12953663e-01
8.50557685e-01 2.10110322e-02 2.94666350e-01 2.03912318e-01
8.15547347e-01 -3.50286543e-01 2.62191445e-01 3.43674242e-01
2.27376278e-02 4.93824482e-01 1.02434747e-01 -4.74874586e-01
-7.30344057e-01 -9.40495610e-01 -2.47164458e-01 1.21137583e+00
-2.78039277e-01 -4.11579370e-01 -6.85877442e-01 -1.11371517e-01
-4.84283082e-02 6.04059279e-01 -2.89937556e-01 -4.38015044e-01
-6.10109568e-01 -4.96127695e-01 1.20777905e+00 6.12764776e-01
3.94756526e-01 -1.39187324e+00 -6.97051644e-01 6.34293854e-01
1.58541426e-02 -6.86958253e-01 -4.66253281e-01 3.49540383e-01
-5.39653718e-01 -8.20464313e-01 -6.92005157e-01 -9.97329533e-01
-1.66351542e-01 -7.09061772e-02 4.82814610e-01 -3.49062085e-01
-5.47919393e-01 1.22482151e-01 -7.24115074e-01 -1.15020597e+00
-2.66403586e-01 -2.26661399e-01 1.85611770e-01 5.10558486e-01
4.18745369e-01 -1.01210344e+00 -5.04005075e-01 -1.33680329e-01
-1.13467884e+00 -4.34761226e-01 5.48665822e-01 6.36975944e-01
2.58668035e-01 6.33738875e-01 1.12695432e+00 -3.22454631e-01
1.16334462e+00 -5.02374351e-01 -2.35395730e-01 -2.27027833e-01
3.00949477e-02 -3.78064126e-01 8.18445146e-01 -9.19709682e-01
-9.05302465e-01 -1.40162945e-01 -4.68404055e-01 -4.99352217e-01
-2.88872391e-01 9.87828612e-01 8.33449513e-02 2.95267627e-02
6.06061876e-01 5.06694913e-01 -5.24831414e-02 -8.17300677e-01
-2.06074640e-02 1.40880501e+00 5.40250659e-01 -4.34719712e-01
4.61112350e-01 3.03910486e-02 -3.20593238e-01 -9.90634263e-01
-1.65362135e-01 -4.75239784e-01 -4.05442715e-01 -4.08494532e-01
7.55504072e-01 -5.16790330e-01 -6.71097636e-01 5.76766968e-01
-1.43932915e+00 3.57140303e-01 -6.55695125e-02 1.00852489e+00
-4.71961528e-01 1.87432155e-01 -5.57069361e-01 -1.36837661e+00
-7.38330126e-01 -9.07758415e-01 9.21393335e-01 2.82554686e-01
-3.62027764e-01 -6.89884424e-01 3.47273290e-01 -2.72950500e-01
4.90178376e-01 3.85607332e-01 1.19952929e+00 -8.58655512e-01
2.63109356e-01 -4.39248264e-01 -1.32703081e-01 5.10701239e-01
3.16044122e-01 1.75957661e-02 -1.42420423e+00 -1.16405897e-01
3.35671484e-01 -3.13850969e-01 7.16470063e-01 2.85965383e-01
1.67517567e+00 -1.36460170e-01 1.20782450e-01 3.52487355e-01
1.20243359e+00 1.05610287e+00 6.38523638e-01 -9.59543139e-02
1.34878039e-01 2.93020755e-01 3.94856721e-01 5.98973095e-01
-2.51577318e-01 7.30994165e-01 1.14639528e-01 1.55972451e-01
-1.94677949e-01 -7.62397200e-02 5.16312718e-01 1.21617424e+00
-1.40922880e-02 -4.04278010e-01 -7.13200986e-01 3.55971307e-01
-1.61656392e+00 -1.05110729e+00 1.15826122e-01 2.08775306e+00
9.21115339e-01 1.71965301e-01 2.63986528e-01 1.19352996e+00
8.28475177e-01 1.32848516e-01 -3.02182257e-01 -9.13398921e-01
4.33279395e-01 8.16187501e-01 -1.89963281e-01 -6.37992546e-02
-1.12626159e+00 2.03825846e-01 6.35025883e+00 9.66874897e-01
-1.79756773e+00 5.85953705e-02 1.17492629e-02 -7.43708685e-02
1.24061190e-01 -6.62841022e-01 -2.44528711e-01 6.50521398e-01
1.37163627e+00 -2.34658584e-01 3.04828703e-01 6.59974575e-01
2.34319717e-01 3.66189659e-01 -1.03802371e+00 1.29604757e+00
9.32140350e-02 -8.93184960e-01 2.76241750e-01 -3.81248474e-01
3.81810777e-02 -5.30261874e-01 1.66911528e-01 4.13481474e-01
-3.10335577e-01 -1.12292767e+00 7.40773439e-01 3.96334708e-01
6.44115865e-01 -1.04681075e+00 9.51591134e-01 2.02591484e-03
-1.65074992e+00 -2.14763269e-01 -3.50027978e-01 -4.16754782e-01
1.88840240e-01 4.71439183e-01 -9.39915717e-01 7.81361222e-01
7.52354383e-01 5.83948314e-01 -7.92516321e-02 1.61186957e+00
1.90572798e-01 1.23658729e+00 -2.47570887e-01 -1.89899012e-01
2.06061453e-01 -3.25638764e-02 6.04834914e-01 1.46827424e+00
6.48322940e-01 -3.48858945e-02 1.20170914e-01 6.14212215e-01
5.21464981e-02 4.33936059e-01 -6.00854933e-01 -3.47976565e-01
4.82336462e-01 1.24208677e+00 -7.15469301e-01 -1.50230333e-01
-2.23536149e-01 2.39436015e-01 -2.03828037e-01 1.50935292e-01
-9.45994854e-01 -1.31589127e+00 4.75814432e-01 -1.95247427e-01
4.40491557e-01 -5.55207729e-02 -8.44725221e-02 -3.27133805e-01
-8.85023698e-02 -5.87670267e-01 4.33333963e-01 -1.04649746e+00
-1.22130847e+00 7.25494146e-01 -6.53413832e-02 -1.54971147e+00
-3.53432119e-01 -7.04692006e-01 -1.24848247e+00 9.11363959e-01
-1.44184315e+00 -6.63325548e-01 -1.44263446e-01 4.22077209e-01
5.93755782e-01 -4.53615755e-01 9.52577412e-01 8.43394220e-01
-2.51037031e-01 6.49799228e-01 -2.25726492e-03 3.15282375e-01
6.16833270e-01 -1.12101889e+00 -4.02130336e-02 2.63902217e-01
4.32117105e-01 4.12952095e-01 5.39700985e-01 -2.40410417e-01
-1.22652125e+00 -9.57589805e-01 4.41636413e-01 3.90017509e-01
5.81121564e-01 -2.20647186e-01 -1.10200083e+00 2.04349801e-01
5.36712110e-01 -3.60794157e-01 1.30921984e+00 -2.73534834e-01
-2.59999484e-01 -3.41649234e-01 -1.08983731e+00 9.20144469e-02
4.33755219e-01 -5.96350729e-01 -1.08444321e+00 -8.99460688e-02
6.70888603e-01 -1.58316478e-01 -9.74282980e-01 3.99822503e-01
8.09992135e-01 -3.88342798e-01 8.56110692e-01 -5.77183247e-01
3.94816130e-01 -5.65221429e-01 -3.02832544e-01 -1.49750984e+00
-2.36321628e-01 -3.82807344e-01 -1.59296930e-01 1.32504797e+00
1.59464300e-01 -3.17236304e-01 3.03759754e-01 -4.47751492e-01
-2.00510651e-01 -6.33428395e-01 -9.32594359e-01 -7.31940448e-01
-2.29332626e-01 -7.77612388e-01 7.93111086e-01 7.41409838e-01
9.57117677e-02 2.83805966e-01 -3.21645647e-01 1.62397370e-01
1.13168217e-01 2.98733443e-01 3.34848225e-01 -1.46138704e+00
-2.68079311e-01 -6.72360957e-01 -8.60558867e-01 -5.36397159e-01
1.09679125e-01 -1.07094979e+00 1.88769639e-01 -1.14817631e+00
-3.82744074e-01 -4.11294028e-02 -8.28357041e-01 2.89048016e-01
1.66189104e-01 6.04704440e-01 9.24023315e-02 -2.10503116e-01
1.20320201e-01 5.06326079e-01 8.03587914e-01 -3.54957253e-01
-2.81147450e-01 3.09824467e-01 -3.89884204e-01 8.28701198e-01
7.92030573e-01 -5.22863925e-01 -5.70080757e-01 -1.51254654e-01
1.56911939e-01 2.34046489e-01 1.94752961e-01 -1.40200162e+00
3.13024670e-01 2.67592639e-01 5.43216646e-01 -9.42250967e-01
7.02310562e-01 -7.62552440e-01 2.23942727e-01 3.86073261e-01
-5.01493037e-01 -7.27496948e-03 5.07658422e-01 4.12787378e-01
-7.39612281e-01 -6.44852161e-01 5.13870001e-01 3.48256260e-01
-5.43509007e-01 -1.08386897e-01 -7.05450475e-01 -3.75112742e-01
7.72491634e-01 -3.07760805e-01 -2.96743447e-03 -4.61482823e-01
-8.12031984e-01 -4.38654721e-01 -6.51910961e-01 4.36222255e-01
9.85390067e-01 -1.66464412e+00 -7.84736276e-01 2.81008154e-01
-1.72940761e-01 -4.08957809e-01 1.37123734e-01 5.01954019e-01
-3.89040142e-01 3.46168935e-01 -4.99872208e-01 -4.33641821e-01
-1.36285365e+00 3.14629108e-01 2.89622128e-01 3.05116922e-01
-5.23484528e-01 6.88535392e-01 -2.22649530e-01 -9.82998610e-02
5.80459893e-01 -6.54864073e-01 -8.50917518e-01 4.48463589e-01
1.02312899e+00 4.51253444e-01 2.98904002e-01 -4.10942525e-01
-2.61134118e-01 7.00840414e-01 1.81603029e-01 -2.78944463e-01
1.34883904e+00 4.43153709e-01 -1.26551911e-01 8.99712920e-01
1.38281441e+00 2.74575758e-03 -4.51693892e-01 -9.65429619e-02
3.40972245e-02 -3.20431739e-01 3.21987092e-01 -7.83923805e-01
-1.06652582e+00 1.11278188e+00 8.98986042e-01 8.71226966e-01
1.45212102e+00 -5.69770992e-01 1.11588120e+00 3.41098189e-01
2.90267110e-01 -8.86447310e-01 1.50505053e-02 4.68048006e-01
1.03276312e+00 -6.66060150e-01 -3.23599458e-01 -9.64507014e-02
-1.87650859e-01 1.62656462e+00 3.64259869e-01 -3.77964407e-01
9.96777475e-01 4.45135921e-01 1.16350420e-01 3.20396945e-02
-5.22118628e-01 -7.19080046e-02 4.78734672e-01 4.89096642e-01
6.42823458e-01 -1.86950445e-01 -4.51539367e-01 1.05542219e+00
-7.84045041e-01 2.12712467e-01 5.25607347e-01 8.55542064e-01
-6.82645380e-01 -9.65170860e-01 -7.77150393e-01 5.01994252e-01
-6.66724980e-01 4.07743677e-02 -3.55831951e-01 2.30296239e-01
3.05399954e-01 1.01886630e+00 4.82133001e-01 -7.09429801e-01
3.21828067e-01 6.59111381e-01 3.70347947e-01 -6.23056591e-01
-7.39469945e-01 -9.07803327e-02 -1.07369915e-01 4.13385183e-02
-3.91489744e-01 -2.05927417e-01 -1.37015355e+00 1.95808485e-01
-3.42133582e-01 5.30270875e-01 1.09917760e+00 6.56036377e-01
8.36470127e-02 1.17330837e+00 6.52905703e-01 -1.07872736e+00
-4.99910653e-01 -1.39197588e+00 -7.14782417e-01 1.97549939e-01
3.91061991e-01 -7.25547433e-01 -2.95197904e-01 1.84498787e-01] | [15.218988418579102, 5.317022800445557] |
c9ef37c4-2c0c-44d6-9178-a4a8a0dd6bd1 | physics-informed-neural-networks-pinns-for-4 | 2110.13361 | null | https://arxiv.org/abs/2110.13361v2 | https://arxiv.org/pdf/2110.13361v2.pdf | A Metalearning Approach for Physics-Informed Neural Networks (PINNs): Application to Parameterized PDEs | Physics-informed neural networks (PINNs) as a means of discretizing partial differential equations (PDEs) are garnering much attention in the Computational Science and Engineering (CS&E) world. At least two challenges exist for PINNs at present: an understanding of accuracy and convergence characteristics with respect to tunable parameters and identification of optimization strategies that make PINNs as efficient as other computational science tools. The cost of PINNs training remains a major challenge of Physics-informed Machine Learning (PiML) - and, in fact, machine learning (ML) in general. This paper is meant to move towards addressing the latter through the study of PINNs on new tasks, for which parameterized PDEs provides a good testbed application as tasks can be easily defined in this context. Following the ML world, we introduce metalearning of PINNs with application to parameterized PDEs. By introducing metalearning and transfer learning concepts, we can greatly accelerate the PINNs optimization process. We present a survey of model-agnostic metalearning, and then discuss our model-aware metalearning applied to PINNs as well as implementation considerations and algorithmic complexity. We then test our approach on various canonical forward parameterized PDEs that have been presented in the emerging PINNs literature. | ['Robert M. Kirby', 'Akil Narayan', 'Shandian Zhe', 'Michael Penwarden'] | 2021-10-26 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [ 1.57427788e-01 -1.08247317e-01 -1.80837676e-01 -9.59159806e-02
-8.32295656e-01 -5.70501924e-01 6.06440902e-01 -2.46510044e-01
-5.84113717e-01 8.73907924e-01 -2.81838596e-01 -3.21926504e-01
-8.80687118e-01 -6.18835628e-01 -1.05933869e+00 -9.68875289e-01
-3.63032222e-01 9.65119958e-01 7.41602927e-02 -3.07833791e-01
4.72932994e-01 1.05206490e+00 -1.15521038e+00 -1.43099546e-01
7.37630963e-01 9.17743266e-01 1.53909296e-01 9.98467028e-01
-2.20974952e-01 6.76378489e-01 -3.01331788e-01 -1.59272194e-01
3.31879944e-01 4.87165600e-02 -9.73825514e-01 -5.91182709e-01
5.36874771e-01 1.62208192e-02 -1.96162105e-01 6.80005789e-01
7.71264195e-01 7.03162789e-01 8.73700917e-01 -1.59955192e+00
-7.29224801e-01 5.51528573e-01 -2.50639439e-01 3.54894370e-01
-4.57600892e-01 2.90507078e-01 8.08534622e-01 -1.06282806e+00
4.83796328e-01 1.22251272e+00 1.03064191e+00 7.18780160e-01
-1.27455997e+00 -5.19219577e-01 -1.21113155e-02 1.52648374e-01
-1.20230412e+00 3.47980000e-02 8.52780104e-01 -7.78718114e-01
9.47010994e-01 1.62016645e-01 5.30902028e-01 1.11940229e+00
4.92094696e-01 6.98058605e-01 8.77319098e-01 -3.73474270e-01
7.77178526e-01 2.79383332e-01 3.44268858e-01 4.79542375e-01
1.32866859e-01 3.71430576e-01 -7.02445686e-01 -2.72471696e-01
1.08055830e+00 -5.54088235e-01 -2.04070285e-01 -7.10794210e-01
-9.27335262e-01 1.02707708e+00 5.89562774e-01 1.04505658e-01
-3.06876838e-01 5.47715187e-01 3.79653871e-01 -8.73210654e-02
6.47825181e-01 8.94955039e-01 -9.60387290e-01 -1.21639125e-01
-1.02812147e+00 7.44343042e-01 1.20993888e+00 8.55091631e-01
9.33520794e-01 4.82300639e-01 1.13214269e-01 6.14335656e-01
1.45335704e-01 3.41833591e-01 1.20484881e-01 -1.30104971e+00
2.02285796e-01 1.39865458e-01 1.55736044e-01 -7.33595014e-01
-5.68637252e-01 -3.11011851e-01 -9.31625247e-01 4.37587738e-01
4.29790050e-01 -5.23563921e-01 -6.98490262e-01 1.66522539e+00
2.64346749e-01 4.82931674e-01 -2.15633273e-01 7.81749606e-01
6.98416471e-01 1.12234521e+00 3.51785988e-01 1.01930469e-01
1.12935174e+00 -7.62861967e-01 -9.91482958e-02 -1.29930332e-01
6.93921387e-01 -3.62412572e-01 9.92176890e-01 4.51617032e-01
-1.11241901e+00 -4.74296212e-01 -8.95506918e-01 -1.41134143e-01
-6.03154480e-01 -4.77761447e-01 7.68649280e-01 4.48059380e-01
-1.21810400e+00 1.34032905e+00 -1.21701598e+00 -2.02977061e-01
3.98308992e-01 8.04475665e-01 -4.46537361e-02 3.66467774e-01
-1.04212832e+00 1.11962295e+00 4.94437575e-01 6.67030588e-02
-6.24718785e-01 -1.75538456e+00 -6.41390443e-01 1.50098307e-02
1.56277984e-01 -8.59152198e-01 1.49661911e+00 -6.52124107e-01
-1.82843339e+00 6.82925880e-01 1.74160346e-01 -6.85208619e-01
4.46189225e-01 -3.53017062e-01 -5.88368159e-03 -1.52674258e-01
-3.16639304e-01 8.05308342e-01 6.59447014e-01 -1.21374261e+00
-3.11520368e-01 1.67303309e-01 -1.47197247e-02 2.03761891e-01
-2.30492622e-01 -8.04744810e-02 -2.11734325e-01 -4.96380746e-01
-3.47643554e-01 -9.32485580e-01 -5.79313099e-01 4.05686736e-01
-3.29741657e-01 -3.30370605e-01 8.74900222e-01 -2.63401657e-01
5.79628468e-01 -1.56423616e+00 4.72767174e-01 1.92581657e-02
3.24910104e-01 5.07114172e-01 4.15792689e-03 4.52083677e-01
-3.84559393e-01 2.81736888e-02 -6.19530618e-01 -3.59518290e-01
4.74906653e-01 5.26519179e-01 -3.87227356e-01 5.44705510e-01
3.16503823e-01 1.24837077e+00 -6.08961344e-01 -5.62480688e-01
5.71363449e-01 6.54439926e-01 -7.73036599e-01 1.89917773e-01
-5.38292587e-01 7.84220338e-01 -6.87224329e-01 3.57386202e-01
7.48832941e-01 -2.06075013e-01 -5.37076294e-01 -2.01622933e-01
-2.58466214e-01 -9.06612799e-02 -1.26304293e+00 1.52977061e+00
-7.43458867e-01 7.69096732e-01 1.66467071e-01 -1.51355791e+00
6.07271075e-01 1.92955226e-01 7.84858823e-01 -2.51792371e-01
2.00104430e-01 3.17322344e-01 2.33290028e-02 -4.27391648e-01
3.95779967e-01 -4.21191484e-01 1.17417499e-01 2.11592495e-01
3.74555767e-01 -4.43812311e-01 -1.44219082e-02 -1.57407485e-03
8.74046803e-01 4.12288725e-01 2.45942950e-01 -8.41761351e-01
6.82585061e-01 4.54448581e-01 4.84184504e-01 8.54114354e-01
-2.30742693e-01 3.51454139e-01 1.39028832e-01 -7.22541869e-01
-1.28104448e+00 -1.26137483e+00 -4.86680537e-01 1.22903144e+00
-1.58371463e-01 1.87144931e-02 -8.31702232e-01 -1.57359123e-01
1.79903239e-01 8.16999316e-01 -6.28224015e-01 -1.13225482e-01
-1.07142937e+00 -9.50939536e-01 3.11506480e-01 6.81788683e-01
3.10196161e-01 -1.32990241e+00 -4.57047313e-01 2.32848704e-01
4.53140199e-01 -1.20742965e+00 -5.07904589e-02 3.06053400e-01
-1.03221226e+00 -8.64017725e-01 -1.10735250e+00 -7.09408402e-01
8.62411410e-02 -5.20874500e-01 1.42435932e+00 -3.48332167e-01
-4.87643659e-01 8.40394497e-01 1.55787662e-01 -7.29996800e-01
-4.19918507e-01 4.82533455e-01 2.37963378e-01 -2.46547133e-01
-4.73277792e-02 -9.80584979e-01 -4.90121663e-01 5.31150512e-02
-7.07530022e-01 -1.16078362e-01 5.66444159e-01 4.95843053e-01
5.57265162e-01 -1.20309368e-01 3.10981095e-01 -8.58757854e-01
5.04269719e-01 -5.44031501e-01 -9.67468262e-01 -4.98876236e-02
-3.55828136e-01 2.29590476e-01 9.70317125e-01 -7.00398266e-01
-1.06967831e+00 6.88763708e-03 -4.49611753e-01 -5.59072673e-01
-1.60719752e-01 3.22884440e-01 4.06708270e-02 -8.12705636e-01
8.46834362e-01 -1.52940929e-01 -4.37413335e-01 -6.37254894e-01
4.41685170e-01 1.33538663e-01 6.68508291e-01 -1.32790184e+00
1.03298759e+00 2.58028448e-01 5.60507655e-01 -1.21751106e+00
-7.66649246e-01 -2.40116164e-01 -7.06154287e-01 -1.04958177e-01
8.96367431e-01 -3.78139496e-01 -9.93301630e-01 4.96657073e-01
-1.29193163e+00 -7.52402484e-01 -7.23781466e-01 5.10205507e-01
-1.03444839e+00 1.38393044e-02 -6.02002382e-01 -8.57061505e-01
-5.14696360e-01 -1.31061959e+00 8.39648724e-01 3.18680942e-01
-2.54622996e-01 -1.69655704e+00 3.31073403e-01 -1.10057041e-01
7.01789498e-01 3.46931815e-01 1.18035376e+00 -6.70161188e-01
-6.46102786e-01 1.21862270e-01 -1.35912552e-01 2.12493747e-01
-3.71914506e-01 1.85876250e-01 -1.14929569e+00 -1.60195082e-01
1.21927127e-01 -1.97443917e-01 6.63932920e-01 9.32304800e-01
1.36616635e+00 -2.16957599e-01 -4.25093383e-01 9.30714965e-01
1.74199557e+00 9.72434320e-03 1.40915751e-01 1.99529007e-01
7.97995269e-01 4.42098767e-01 -2.29807701e-02 2.67314076e-01
-2.70920172e-02 5.32511771e-01 3.87751192e-01 -1.35894895e-01
-6.06345152e-03 6.10416904e-02 -1.07901301e-02 7.84358561e-01
-8.23709890e-02 -3.06339562e-01 -1.14833987e+00 3.99870336e-01
-1.74027801e+00 -4.67089951e-01 -2.75155783e-01 1.78343153e+00
6.30997837e-01 3.00997570e-02 -1.55457944e-01 -1.38800023e-02
7.06347466e-01 -1.08279750e-01 -9.91980910e-01 -7.00815201e-01
1.36189163e-01 8.15437078e-01 6.68251336e-01 5.67797482e-01
-1.13832307e+00 8.96544456e-01 6.14794445e+00 1.03236115e+00
-1.19638932e+00 1.31820263e-02 4.60854381e-01 1.46502316e-01
1.75802931e-01 -1.09214246e-01 -9.14634824e-01 2.87334561e-01
1.12444448e+00 -2.41437837e-01 4.68476802e-01 1.24135709e+00
2.58037806e-01 1.47012770e-01 -1.52885747e+00 9.49807286e-01
-4.83023196e-01 -1.75450563e+00 2.45895669e-01 4.20542620e-02
9.10127938e-01 4.34425443e-01 2.55370259e-01 3.13389897e-01
4.87650335e-01 -1.20000279e+00 5.99469602e-01 7.13482857e-01
4.19577897e-01 -7.16040492e-01 4.19821560e-01 4.39211249e-01
-9.90281522e-01 -1.11705186e-02 -3.09864134e-01 -6.07187599e-02
1.63860410e-01 5.84188819e-01 -4.54454184e-01 4.23703283e-01
7.82248259e-01 5.78646839e-01 1.15541980e-01 1.30053282e+00
2.67597735e-01 7.73861051e-01 -6.73171282e-01 -2.68724799e-01
3.91227335e-01 -2.45948434e-01 7.65107512e-01 1.28104949e+00
2.30711713e-01 9.75552052e-02 4.80473153e-02 1.37829840e+00
5.57290651e-02 -5.86948246e-02 -3.28566164e-01 9.78734940e-02
2.56316245e-01 1.30233514e+00 -6.51069701e-01 -1.05864748e-01
2.45150067e-02 4.87749398e-01 2.87708431e-01 5.26058912e-01
-8.56045604e-01 -2.56175250e-01 9.66446877e-01 -4.96758968e-02
1.29527092e-01 -5.67680717e-01 -4.00600493e-01 -8.76576662e-01
-4.05130029e-01 -3.17162693e-01 1.13800786e-01 -8.82689357e-01
-1.61840963e+00 4.28247154e-01 4.41046774e-01 -8.79520297e-01
-1.46881819e-01 -1.33653355e+00 -7.98650742e-01 9.42346752e-01
-1.48209512e+00 -9.44403529e-01 -1.61418647e-01 3.94662380e-01
4.85579610e-01 1.87086239e-01 6.95864499e-01 2.59432375e-01
-5.78249097e-01 3.46192390e-01 3.13553631e-01 8.16735625e-02
3.57982606e-01 -1.37965024e+00 7.18028545e-01 5.26266932e-01
2.49932706e-02 4.92634654e-01 9.54820275e-01 -2.93927431e-01
-1.79387736e+00 -1.06796765e+00 4.06498432e-01 -5.60479879e-01
9.26774800e-01 -3.45444739e-01 -1.02635288e+00 8.03483129e-01
-3.03957891e-02 1.50459468e-01 1.62781045e-01 -1.59198433e-01
2.16342017e-01 4.79108877e-02 -1.11191058e+00 7.31406510e-01
9.01254952e-01 -2.25657418e-01 -2.82440931e-01 5.60004115e-01
6.09906316e-01 -7.23057806e-01 -9.92298126e-01 5.02947390e-01
3.45551670e-01 -7.17709899e-01 1.39551783e+00 -6.76575422e-01
2.86592245e-01 1.29531920e-01 -7.10294843e-02 -1.35147905e+00
-1.59858167e-01 -8.03797245e-01 -2.90286005e-01 1.07636511e+00
2.81422973e-01 -8.79099667e-01 1.17713118e+00 9.87569928e-01
-3.19589436e-01 -1.02579498e+00 -1.06798780e+00 -9.98035073e-01
9.85848188e-01 -9.70796525e-01 2.81374216e-01 7.07051158e-01
-5.10696292e-01 -9.69152451e-02 4.97258939e-02 2.70095855e-01
7.92870462e-01 -2.07711935e-01 6.24509573e-01 -1.43562603e+00
-3.65493149e-01 -6.75354600e-01 -2.99944580e-01 -9.86423850e-01
2.61201590e-01 -9.48567569e-01 7.65770152e-02 -1.33876824e+00
-1.45401001e-01 -6.49627090e-01 -1.02775857e-01 2.35287368e-01
2.40083262e-01 1.26740366e-01 6.94519132e-02 9.10062343e-02
-1.21936001e-01 5.49024224e-01 1.18047023e+00 4.58139703e-02
-2.61796862e-01 1.14571914e-01 -2.83455700e-01 9.91940618e-01
1.02174115e+00 -4.84993875e-01 -3.79484832e-01 -3.88112485e-01
1.11650042e-01 -1.59578249e-01 7.86438525e-01 -1.30253375e+00
4.38125551e-01 -3.03592652e-01 4.12822425e-01 -5.12042999e-01
8.49636972e-01 -6.13556266e-01 -2.25709066e-01 3.28803360e-01
-2.46262714e-01 2.72834487e-03 6.62226677e-01 1.21499978e-01
3.04551005e-01 -3.85260850e-01 1.11550665e+00 -2.94694543e-01
-8.15863132e-01 7.39471912e-01 -5.40490031e-01 4.49678749e-01
7.89269686e-01 -8.50173831e-02 1.47295371e-01 -1.16852492e-01
-6.06377184e-01 9.40335393e-02 3.99595425e-02 7.07942341e-03
2.35954925e-01 -8.80837381e-01 -5.83619356e-01 3.76903564e-02
-3.94885302e-01 2.86212325e-01 3.98623466e-01 6.08230889e-01
-8.58544588e-01 5.81174135e-01 -2.13384628e-01 -8.06050420e-01
-6.03335023e-01 4.45162296e-01 8.80876660e-01 -2.36244753e-01
-5.59104800e-01 1.23002601e+00 3.64299029e-01 -9.41391230e-01
2.29536444e-01 -4.67537463e-01 2.49391049e-01 -4.18422043e-01
6.89448975e-03 8.12580109e-01 1.58425510e-01 -3.41783941e-01
-2.74873137e-01 9.62663352e-01 2.55372405e-01 -1.52566463e-01
1.56639862e+00 4.01038319e-01 1.47732347e-01 3.69855344e-01
1.28916013e+00 -5.98105371e-01 -1.46458173e+00 -2.17260584e-01
7.90581927e-02 4.22149181e-01 2.01007202e-01 -6.92674756e-01
-1.06925702e+00 1.19896674e+00 4.57141697e-01 -1.66851938e-01
7.52787113e-01 -1.12643957e-01 7.88537979e-01 6.41702414e-01
1.37919784e-01 -9.49397206e-01 -2.20072180e-01 6.33074760e-01
8.96026373e-01 -8.55728626e-01 2.84336600e-02 -3.18157554e-01
-1.18952520e-01 1.37119675e+00 5.91376185e-01 -7.42572248e-01
1.22135913e+00 6.24106526e-01 -3.09122980e-01 -1.94779299e-02
-4.56432790e-01 2.75211185e-01 5.07870555e-01 6.91816807e-01
5.66011555e-02 -4.26053584e-01 1.87150747e-01 6.08603299e-01
-4.03758615e-01 1.76318377e-01 2.25799412e-01 9.36843395e-01
-3.29494238e-01 -1.12432098e+00 -3.95694494e-01 3.45422477e-01
-1.18137300e-01 -1.95703238e-01 -1.87868178e-01 1.31585813e+00
2.29281992e-01 3.59380394e-01 1.40968719e-02 6.61465228e-02
2.53094524e-01 1.28887281e-01 5.90754688e-01 -2.63161033e-01
-5.26060939e-01 -4.54287231e-01 -2.23188162e-01 -4.32156503e-01
-1.54071391e-01 -4.50491965e-01 -1.31385767e+00 -5.88206291e-01
-2.98900642e-02 3.28058600e-01 9.41196263e-01 1.30588663e+00
3.23033988e-01 4.57347065e-01 1.07283935e-01 -1.65942264e+00
-6.35029197e-01 -5.17953455e-01 -6.84606075e-01 -3.01047266e-01
3.84772241e-01 -8.65506113e-01 -3.66065562e-01 -1.26527607e-01] | [6.485883712768555, 3.5526180267333984] |
f5ebaecc-1094-4a0b-9803-90d4a1b1c91d | privacy-leakage-of-sift-features-via-deep | 2009.01030 | null | https://arxiv.org/abs/2009.01030v1 | https://arxiv.org/pdf/2009.01030v1.pdf | Privacy Leakage of SIFT Features via Deep Generative Model based Image Reconstruction | Many practical applications, e.g., content based image retrieval and object recognition, heavily rely on the local features extracted from the query image. As these local features are usually exposed to untrustworthy parties, the privacy leakage problem of image local features has received increasing attention in recent years. In this work, we thoroughly evaluate the privacy leakage of Scale Invariant Feature Transform (SIFT), which is one of the most widely-used image local features. We first consider the case that the adversary can fully access the SIFT features, i.e., both the SIFT descriptors and the coordinates are available. We propose a novel end-to-end, coarse-to-fine deep generative model for reconstructing the latent image from its SIFT features. The designed deep generative model consists of two networks, where the first one attempts to learn the structural information of the latent image by transforming from SIFT features to Local Binary Pattern (LBP) features, while the second one aims to reconstruct the pixel values guided by the learned LBP. Compared with the state-of-the-art algorithms, the proposed deep generative model produces much improved reconstructed results over three public datasets. Furthermore, we address more challenging cases that only partial SIFT features (either SIFT descriptors or coordinates) are accessible to the adversary. It is shown that, if the adversary can only have access to the SIFT descriptors while not their coordinates, then the modest success of reconstructing the latent image can be achieved for highly-structured images (e.g., faces) and would fail in general settings. In addition, the latent image can be reconstructed with reasonably good quality solely from the SIFT coordinates. Our results would suggest that the privacy leakage problem can be largely avoided if the SIFT coordinates can be well protected. | ['Jiantao Zhou', 'Haiwei Wu'] | 2020-09-02 | null | null | null | null | ['content-based-image-retrieval'] | ['computer-vision'] | [ 2.58937627e-01 -1.71060339e-01 -1.74529374e-01 -3.40639263e-01
-1.09664810e+00 -8.90893102e-01 4.46734339e-01 -1.79220885e-01
-3.97445589e-01 5.70585966e-01 -1.16210699e-01 -1.79293707e-01
4.04100679e-02 -1.16230392e+00 -1.16851163e+00 -1.27847481e+00
9.28391740e-02 -4.27621081e-02 -1.51085973e-01 1.29018828e-01
1.65517420e-01 8.14147592e-01 -1.51902103e+00 2.36702874e-01
4.66088146e-01 1.36549938e+00 -2.69803494e-01 4.14477319e-01
1.88867882e-01 3.04562926e-01 -3.73092830e-01 -6.87024713e-01
5.93310177e-01 -2.15472743e-01 -6.33316755e-01 -4.07534167e-02
4.66454357e-01 -8.27824593e-01 -6.23650014e-01 1.46826684e+00
2.83966869e-01 -2.38012671e-01 2.83471346e-01 -1.46528268e+00
-8.19948196e-01 2.88575497e-02 -3.56522858e-01 -4.23645407e-01
2.55804300e-01 2.36145973e-01 9.50363755e-01 -5.81184268e-01
7.52467811e-01 1.04801273e+00 4.32106763e-01 1.89325199e-01
-1.19861317e+00 -8.83557081e-01 -2.20707610e-01 4.22358364e-01
-1.83226192e+00 -3.21796834e-01 8.83873701e-01 -1.31452531e-01
4.04372782e-01 4.47765976e-01 2.24126682e-01 1.13039958e+00
4.02459532e-01 5.22075772e-01 1.20241106e+00 -7.78488219e-02
1.62049443e-01 3.34322214e-01 -2.35036984e-01 6.76985204e-01
2.85160273e-01 4.43186581e-01 -5.04428864e-01 -6.33816719e-01
6.14701271e-01 4.72316891e-01 -4.05016363e-01 -6.87864959e-01
-1.04681873e+00 1.08704305e+00 6.21716976e-01 1.90631852e-01
-3.71193022e-01 2.65064746e-01 2.10339427e-01 5.16557753e-01
1.39504358e-01 -1.03354737e-01 -2.59797275e-01 1.10572383e-01
-7.55365551e-01 1.90777555e-01 7.58909643e-01 8.78966928e-01
1.18642020e+00 -3.97426963e-01 1.48722790e-02 7.49203861e-02
2.39959270e-01 8.26835513e-01 1.95844576e-01 -6.77019954e-01
4.13914740e-01 1.55476063e-01 2.79249847e-01 -1.61256611e+00
2.81555444e-01 -8.87801126e-02 -9.82194781e-01 -9.32258815e-02
4.47668910e-01 7.86114708e-02 -6.19602323e-01 1.96975386e+00
4.50065911e-01 -1.41803816e-01 2.04334080e-01 8.33182096e-01
4.16173846e-01 7.50582755e-01 -2.88794935e-01 5.76131884e-03
1.41870415e+00 -3.32344711e-01 -6.16867721e-01 1.21832110e-01
3.22017163e-01 -7.51408696e-01 5.74225664e-01 1.73187390e-01
-7.29824245e-01 -5.02895057e-01 -1.11274695e+00 -3.19617838e-01
-6.23465657e-01 1.02542147e-01 5.51793277e-01 8.30150783e-01
-9.42592323e-01 5.96231163e-01 -8.27052057e-01 -3.78671810e-02
6.30423248e-01 6.96635783e-01 -8.71303916e-01 -4.20537919e-01
-1.34840477e+00 2.02389911e-01 1.46863833e-01 1.64597228e-01
-9.00447011e-01 -1.42673641e-01 -8.28586459e-01 3.01294297e-01
3.31247747e-01 -4.53538358e-01 6.81209683e-01 -8.29097927e-01
-1.32326877e+00 7.43551195e-01 -1.78576380e-01 -3.35795939e-01
5.37674665e-01 1.66251161e-03 -1.64555460e-01 5.05341709e-01
5.01798317e-02 3.90334398e-01 1.33661711e+00 -1.17820191e+00
-4.14676815e-01 -7.06540763e-01 7.22112432e-02 -3.11139166e-01
-3.95808280e-01 -2.18171716e-01 -2.92160600e-01 -4.95502234e-01
3.39866616e-02 -1.22799337e+00 3.09088938e-02 5.61968684e-01
-5.63129723e-01 2.33290985e-01 1.15490532e+00 -6.59420311e-01
6.33868396e-01 -2.53307390e+00 -2.49919519e-01 5.89470327e-01
2.12964527e-02 1.60497874e-01 -1.06585875e-01 5.28983176e-01
7.41430297e-02 2.34409511e-01 -5.22068515e-02 -2.26998389e-01
1.96273416e-01 2.64773428e-01 -7.40199625e-01 1.09131742e+00
3.50466594e-02 1.05563354e+00 -5.64666808e-01 -2.02048913e-01
-5.45839742e-02 8.12513769e-01 -6.23782873e-01 1.24473087e-01
9.61719230e-02 3.64032894e-01 -8.42365682e-01 5.05865693e-01
1.24187493e+00 -2.80700117e-01 7.79025480e-02 -2.61995673e-01
1.79085463e-01 7.31573403e-02 -9.74368513e-01 1.61586511e+00
-3.15247357e-01 3.83827418e-01 2.59788424e-01 -7.20232427e-01
8.07317436e-01 4.66647685e-01 3.57041359e-01 -6.36335909e-01
1.00343652e-01 8.94277468e-02 -3.17831039e-01 -1.58995718e-01
3.13767910e-01 -6.15332611e-02 -3.92955631e-01 5.37901521e-01
-9.99148861e-02 2.48526752e-01 -7.23084033e-01 8.57289732e-02
9.54351246e-01 -4.97517250e-02 1.42256230e-01 2.20677387e-02
5.21027327e-01 -3.60381812e-01 4.18050170e-01 7.01537371e-01
4.63036820e-03 4.68775094e-01 5.18423915e-01 -5.16714811e-01
-8.79056573e-01 -1.03068590e+00 -1.14859685e-01 6.70605004e-01
3.66152287e-01 -4.85649288e-01 -7.36131668e-01 -9.10239279e-01
1.26296520e-01 2.01271400e-01 -6.52351856e-01 -4.11950499e-01
-2.70687819e-01 -2.72843391e-01 6.71153605e-01 8.84340033e-02
9.55733657e-01 -7.95429468e-01 -4.48480457e-01 -1.33224785e-01
-1.78441346e-01 -1.09902155e+00 -6.09457314e-01 -7.54670128e-02
-5.08450866e-01 -9.00646031e-01 -3.35959911e-01 -5.15881300e-01
8.97127330e-01 4.77688640e-01 3.12685490e-01 2.59405315e-01
-2.00400144e-01 4.03203070e-01 -1.89960390e-01 8.49467441e-02
-2.59004444e-01 -5.66094927e-02 -2.04044536e-01 6.71153426e-01
3.81526977e-01 -5.90209246e-01 -6.96356833e-01 3.09391111e-01
-1.28565085e+00 -2.66361028e-01 6.77920043e-01 9.56861019e-01
9.31252420e-01 3.73970002e-01 1.76873326e-01 -9.40098464e-01
2.88009783e-03 -2.46809304e-01 -6.88069642e-01 2.47673124e-01
-2.40067825e-01 1.86505035e-01 8.61568391e-01 -4.60902184e-01
-7.06505239e-01 3.41566861e-01 -2.08402082e-01 -5.60534477e-01
-2.70334333e-01 1.79716781e-01 -7.08068490e-01 -7.18828619e-01
1.59694269e-01 6.55250967e-01 1.26720471e-02 -4.57784712e-01
3.37735981e-01 7.17998683e-01 4.20676112e-01 -6.33447587e-01
1.22543383e+00 8.68722796e-01 3.58695209e-01 -7.37524629e-01
-3.63840014e-01 -1.92286298e-01 -4.40187484e-01 3.69001687e-01
7.23544717e-01 -8.51327479e-01 -1.07587993e+00 6.57099068e-01
-1.26805377e+00 3.29341561e-01 -2.14466244e-01 1.69376627e-01
-6.28486812e-01 6.54907227e-01 -5.75192034e-01 -6.82937741e-01
-3.01391512e-01 -1.46506035e+00 1.45867777e+00 -2.59812409e-03
2.90867507e-01 -6.19401634e-01 -4.16659296e-01 2.90755719e-01
5.20297825e-01 5.96665919e-01 9.01760638e-01 -5.61029315e-01
-1.22582901e+00 -6.94332480e-01 -2.46295437e-01 4.24775451e-01
3.17625821e-01 -4.12716955e-01 -1.15827596e+00 -7.26566732e-01
4.48724091e-01 -3.91356289e-01 6.88005507e-01 -3.65248322e-02
1.41948807e+00 -9.07804668e-01 -2.49720380e-01 1.01502478e+00
1.49198055e+00 7.83344060e-02 8.71012211e-01 8.52773152e-03
7.31959224e-01 3.80041778e-01 4.92769331e-01 4.61550087e-01
1.45774141e-01 6.55780017e-01 5.33538878e-01 7.35986903e-02
3.51525486e-01 -6.69811547e-01 3.39191705e-01 3.75435621e-01
2.92259574e-01 -9.87085477e-02 -2.47776464e-01 3.38033915e-01
-1.65390205e+00 -9.20167327e-01 4.81861651e-01 2.45692325e+00
7.41417766e-01 -2.92915642e-01 -4.30520862e-01 -2.14181483e-01
6.18534744e-01 3.62250626e-01 -6.60030782e-01 -2.45175526e-01
-1.02418438e-01 2.42379487e-01 7.77085662e-01 2.69763738e-01
-1.23621702e+00 8.54742646e-01 5.10694456e+00 9.98143613e-01
-1.27746642e+00 2.10154951e-01 7.42846131e-01 1.28571004e-01
-3.33846807e-01 2.89594352e-01 -7.42710948e-01 6.23073578e-01
8.15093160e-01 -3.00075375e-02 5.96320510e-01 9.78849232e-01
-3.54837567e-01 2.03736305e-01 -9.87179399e-01 1.06296074e+00
-4.77099344e-02 -1.23638916e+00 2.50316679e-01 8.16404700e-01
4.18228149e-01 -3.76062632e-01 5.68203926e-01 -6.61955550e-02
3.75666693e-02 -9.14652348e-01 5.45669615e-01 5.02830088e-01
1.09420776e+00 -1.13398528e+00 6.78240895e-01 4.87370968e-01
-1.04044163e+00 -1.80740946e-03 -7.08039165e-01 4.17902470e-01
-1.75215080e-01 3.76338542e-01 -4.47563380e-01 7.13847816e-01
7.67975092e-01 3.81674558e-01 -4.65612173e-01 5.71117520e-01
-3.66560936e-01 3.94686937e-01 -5.23863494e-01 1.31993592e-01
2.67912447e-01 -1.28700018e-01 3.74997854e-01 7.87463427e-01
5.62125802e-01 7.31121451e-02 1.97886020e-01 9.94080365e-01
-4.89636004e-01 1.06576540e-01 -9.62446630e-01 -1.34018973e-01
4.12590444e-01 1.12817705e+00 -3.86027485e-01 4.40476909e-02
-3.04002851e-01 1.31209314e+00 1.15925439e-01 3.61904949e-01
-7.34259307e-01 -4.88691092e-01 8.78371060e-01 -4.61757183e-02
7.81561553e-01 -1.96563676e-01 2.03715280e-01 -1.22357619e+00
1.74122930e-01 -1.01076996e+00 1.73444882e-01 -4.33810383e-01
-1.31245017e+00 5.79753757e-01 -2.72917092e-01 -1.23941123e+00
-3.07638884e-01 -4.57423657e-01 -1.84060544e-01 9.04413402e-01
-1.58094847e+00 -1.50717115e+00 -1.89343784e-02 1.10435009e+00
-3.63191456e-01 6.16284162e-02 1.02463305e+00 1.41959131e-01
-4.29133512e-02 1.02176630e+00 4.07131165e-01 4.57912803e-01
6.81092083e-01 -6.57188773e-01 3.17755997e-01 8.75492573e-01
4.58613336e-01 9.99942243e-01 3.20299774e-01 -5.35057783e-01
-1.98732555e+00 -1.09421909e+00 7.46996582e-01 -4.59569991e-02
4.52706754e-01 -7.25702167e-01 -8.15130293e-01 7.19540119e-01
-1.70663655e-01 5.81603348e-01 6.49990022e-01 -6.36951149e-01
-7.03754485e-01 -3.37629795e-01 -1.60071182e+00 2.75324732e-01
6.53846681e-01 -1.09675956e+00 4.16404046e-02 3.30114722e-01
7.73181736e-01 -3.25126231e-01 -8.25337470e-01 -9.48212203e-03
7.01639533e-01 -8.63618553e-01 1.02718198e+00 -4.29449052e-01
7.83175379e-02 -2.96205997e-01 -6.63221002e-01 -7.66171634e-01
-1.03753299e-01 -8.28637064e-01 4.20121029e-02 1.25686347e+00
-9.21005569e-03 -9.63398814e-01 6.95077658e-01 7.01543570e-01
6.46495163e-01 -4.03409660e-01 -1.52035224e+00 -8.27455938e-01
-3.57255116e-02 -1.75383702e-01 1.12118483e+00 6.80909812e-01
-5.69283247e-01 -2.36490116e-01 -7.99861133e-01 6.59709156e-01
7.61752307e-01 7.68956244e-01 9.36520219e-01 -9.18576002e-01
-3.69088024e-01 2.06699044e-01 -8.28989148e-01 -1.15142179e+00
3.56429785e-01 -8.01055312e-01 -5.37962392e-02 -7.64402986e-01
1.88475415e-01 -2.90134639e-01 -2.79614925e-01 7.22908676e-01
1.82212695e-01 4.00233686e-01 2.58288711e-01 1.29258379e-01
-3.34627271e-01 6.75015867e-01 1.25685084e+00 -2.32248336e-01
2.55971611e-01 1.63951039e-01 -7.07628906e-01 2.99722791e-01
6.18675470e-01 -8.31730366e-01 -2.03131422e-01 -1.17028877e-01
6.00276738e-02 1.57977715e-01 8.58304799e-01 -6.85251296e-01
4.51348096e-01 -1.52769580e-01 3.34047705e-01 -3.52013171e-01
5.57608306e-01 -1.34590352e+00 3.98002356e-01 4.57675517e-01
-1.95230469e-01 -2.29921460e-01 -9.46622118e-02 8.66172493e-01
-3.57238561e-01 1.72434710e-02 6.66121244e-01 -4.80526593e-04
-2.59479553e-01 7.35607088e-01 1.49841547e-01 -5.88972628e-01
9.72366095e-01 -1.77592218e-01 -4.72693443e-01 -6.12933218e-01
-4.56870645e-01 -3.29355448e-01 7.98354387e-01 3.17656040e-01
7.65132427e-01 -1.51882327e+00 -3.60713512e-01 7.78778791e-01
6.41833320e-02 -2.16160014e-01 4.10696328e-01 6.06607616e-01
-3.23890686e-01 5.34808218e-01 -1.51817322e-01 -4.56212789e-01
-1.31911623e+00 1.11227238e+00 1.33742644e-02 -1.83809355e-01
-6.01700246e-01 5.25028765e-01 5.30503750e-01 -5.81995659e-02
5.19194528e-02 -1.73245251e-01 5.56299806e-01 -2.54863620e-01
6.45768583e-01 -4.24580090e-02 1.28653916e-02 -1.12762630e+00
-3.15059751e-01 7.65548646e-01 -1.72749326e-01 1.76488310e-02
1.27036560e+00 -2.61951774e-01 -3.00950378e-01 -1.99990705e-01
1.99431920e+00 3.64016056e-01 -1.23696911e+00 -3.90017480e-01
-4.95203793e-01 -1.00554168e+00 1.47500411e-02 -2.79607773e-01
-1.44906616e+00 9.99624848e-01 7.26826787e-01 1.22476414e-01
1.22804773e+00 -1.87168449e-01 1.24375999e+00 5.31822443e-01
9.69492257e-01 -5.14784396e-01 -2.87959009e-01 5.38411215e-02
5.79364419e-01 -1.07384658e+00 -1.61717795e-02 -3.68475050e-01
-3.51522624e-01 9.63922620e-01 -5.49681718e-03 -1.73592955e-01
7.63686299e-01 1.16974646e-02 -2.26702571e-01 -4.92008105e-02
-4.74322557e-01 2.57484227e-01 3.18482906e-01 5.71719050e-01
-3.83052796e-01 1.05318964e-01 1.73039839e-01 6.84873879e-01
-2.95826077e-01 -2.23016053e-01 1.19059317e-01 9.69241917e-01
-3.20094563e-02 -1.27957582e+00 -5.37724018e-01 1.85097650e-01
-7.94678330e-01 6.81715161e-02 -3.71108145e-01 5.54335773e-01
3.04973751e-01 7.40068495e-01 -2.79464543e-01 -4.12169904e-01
-7.21533820e-02 -1.53458759e-01 2.52982914e-01 -1.32607907e-01
-4.42303449e-01 -5.43802492e-02 -5.22082329e-01 -9.23856258e-01
-2.22521201e-01 -5.69911122e-01 -8.57410669e-01 -6.49480343e-01
-2.87122726e-01 1.78317383e-01 6.90868318e-01 7.14427292e-01
6.18802369e-01 -2.64206558e-01 1.07431722e+00 -6.80409670e-01
-6.85433507e-01 -2.10613191e-01 -8.75712514e-01 4.89493161e-01
7.44612515e-01 -3.73413175e-01 -5.36677837e-01 2.19027139e-03] | [12.621984481811523, 0.7164384722709656] |
eb675a80-6be7-4966-bade-a94d12dafc55 | masked-and-permuted-implicit-context-learning | 2305.16172 | null | https://arxiv.org/abs/2305.16172v1 | https://arxiv.org/pdf/2305.16172v1.pdf | Masked and Permuted Implicit Context Learning for Scene Text Recognition | Scene Text Recognition (STR) is a challenging task due to variations in text style, shape, and background. Incorporating linguistic information is an effective way to enhance the robustness of STR models. Existing methods rely on permuted language modeling (PLM) or masked language modeling (MLM) to learn contextual information implicitly, either through an ensemble of permuted autoregressive (AR) LMs training or iterative non-autoregressive (NAR) decoding procedure. However, these methods exhibit limitations: PLM's AR decoding results in the lack of information about future characters, while MLM provides global information of the entire text but neglects dependencies among each predicted character. In this paper, we propose a Masked and Permuted Implicit Context Learning Network for STR, which unifies PLM and MLM within a single decoding architecture, inheriting the advantages of both approaches. We utilize the training procedure of PLM, and to integrate MLM, we incorporate word length information into the decoding process by introducing specific numbers of mask tokens. Experimental results demonstrate that our proposed model achieves state-of-the-art performance on standard benchmarks using both AR and NAR decoding procedures. | ['Weiping Wang', 'Dongbao Yang', 'Zhilong Ji', 'Ye Yuan', 'Yu Zhou', 'Jin Wei', 'Zhi Qiao', 'Xiaomeng Yang'] | 2023-05-25 | null | null | null | null | ['scene-text-recognition'] | ['computer-vision'] | [ 7.08666205e-01 -2.94422776e-01 -2.03209728e-01 -4.92601395e-01
-6.59826159e-01 -3.36022645e-01 8.99050057e-01 -6.65835738e-02
-4.91614938e-01 3.41826797e-01 3.89282644e-01 -5.57758629e-01
5.03369808e-01 -5.01237869e-01 -6.59391999e-01 -6.91599488e-01
4.64871049e-01 2.04899475e-01 3.19520235e-01 -7.40479752e-02
5.52000344e-01 3.53359103e-01 -1.08915353e+00 5.75810313e-01
8.59467745e-01 6.93559229e-01 5.49757361e-01 8.29664767e-01
-7.44409502e-01 1.24503124e+00 -4.10442144e-01 -2.71875054e-01
-5.86163215e-02 -3.57504576e-01 -3.12754631e-01 2.22920895e-01
1.05218142e-01 -2.91871637e-01 -6.22779965e-01 5.82722723e-01
2.68620223e-01 1.80927455e-01 7.70135999e-01 -6.31745517e-01
-7.18969166e-01 8.74251664e-01 -7.16062546e-01 7.19704777e-02
1.42899066e-01 -4.52154055e-02 8.11788440e-01 -1.21026266e+00
8.21443498e-02 1.48224568e+00 5.11358917e-01 5.62989891e-01
-1.04781079e+00 -4.75154042e-01 8.43659163e-01 1.49073824e-01
-1.55541050e+00 -6.47575676e-01 9.17894840e-01 -4.41031605e-01
1.19133973e+00 2.28209496e-01 4.49665450e-02 1.34942794e+00
3.28141689e-01 1.23899662e+00 9.51209784e-01 -8.92349184e-01
1.46444827e-01 1.20931342e-01 3.46346289e-01 5.54853201e-01
5.96124725e-03 -2.70128161e-01 -5.16629457e-01 -2.06873659e-02
6.68755174e-01 7.39018768e-02 6.96013942e-02 -9.24659893e-03
-1.02073658e+00 6.90720856e-01 -1.58863366e-01 3.04053456e-01
-1.21401384e-01 1.10712647e-01 3.63217592e-01 -1.75330264e-03
5.19390464e-01 -1.73951387e-01 -3.02561760e-01 -8.23581219e-02
-1.08776450e+00 -3.61236244e-01 4.96656567e-01 8.64127755e-01
5.98732293e-01 6.07784331e-01 -3.24794978e-01 1.12208056e+00
7.03714252e-01 5.36470532e-01 8.24419916e-01 -2.12504238e-01
8.05845201e-01 5.18087685e-01 -1.29620522e-01 -7.86131680e-01
-1.78155094e-01 -3.38785380e-01 -8.70089471e-01 -1.69798300e-01
-1.30729526e-01 -1.77134737e-01 -1.28345490e+00 1.49147892e+00
-1.54800311e-01 5.36058307e-01 1.14908963e-01 4.82416004e-01
6.89658761e-01 9.90217030e-01 2.91155636e-01 -2.07742915e-01
8.22221637e-01 -1.30545688e+00 -7.84552336e-01 -6.03423774e-01
7.39753306e-01 -8.21400285e-01 1.06587052e+00 4.28469092e-01
-7.97995627e-01 -5.71574509e-01 -9.51687813e-01 -2.67551988e-01
-3.25346589e-01 5.75422347e-01 2.42422998e-01 7.41704464e-01
-8.85429800e-01 1.58557042e-01 -9.97755766e-01 -1.29033506e-01
9.07047763e-02 3.53851676e-01 -3.99254933e-02 -5.56474924e-02
-8.50794375e-01 6.21845186e-01 3.69505554e-01 4.96102333e-01
-6.44107699e-01 -6.99973553e-02 -1.05393040e+00 -3.58702131e-02
3.67245525e-01 -3.29146236e-01 1.03706336e+00 -1.05913770e+00
-1.84020460e+00 4.79467690e-01 -7.89846659e-01 -5.09067237e-01
4.89852011e-01 -5.92875719e-01 -4.06843841e-01 -2.32795000e-01
-3.62375975e-01 3.27732861e-01 1.16283417e+00 -1.29265344e+00
-3.71192575e-01 -3.01572353e-01 -5.61585605e-01 3.82746756e-01
-3.07064891e-01 2.78114289e-01 -8.57162654e-01 -1.02004373e+00
2.66149670e-01 -6.43019915e-01 -3.79707783e-01 -4.23635781e-01
-3.63500923e-01 -3.23880352e-02 1.01617515e+00 -9.64641154e-01
1.68154347e+00 -2.18715262e+00 1.75028905e-01 5.87529205e-02
-2.73091614e-01 4.25173521e-01 -3.72797102e-01 4.41043705e-01
2.43323788e-01 3.88340503e-02 -3.41554254e-01 -9.10616875e-01
-9.80193838e-02 5.58409214e-01 -7.50891268e-01 1.52957946e-01
3.14802885e-01 9.65928912e-01 -4.21579838e-01 -4.77308273e-01
4.43864971e-01 6.62973762e-01 -2.55027890e-01 9.50210765e-02
-4.71016347e-01 6.11304224e-01 -4.31889445e-01 3.71021152e-01
5.49539030e-01 -1.93717152e-01 2.75397360e-01 2.99852103e-01
-2.27867097e-01 4.22996551e-01 -1.33842647e+00 1.42733955e+00
-5.22208691e-01 5.45770884e-01 -2.49330744e-01 -1.00303209e+00
1.18927288e+00 1.69844702e-01 -4.41714264e-02 -5.58443129e-01
1.17445208e-01 -2.49015354e-02 -3.02671462e-01 -3.41292083e-01
7.05471039e-01 1.86125830e-01 6.47480860e-02 2.91463584e-01
-1.85531944e-01 3.53167236e-01 -4.28791866e-02 -4.25510965e-02
7.68101096e-01 2.62837827e-01 2.09016994e-01 3.57053488e-01
9.18859303e-01 -5.87228119e-01 6.25994086e-01 1.07525468e+00
2.25590155e-01 6.11478984e-01 2.16513664e-01 -2.57355481e-01
-9.33400691e-01 -7.25525141e-01 5.12041431e-03 1.34127712e+00
-3.80002260e-02 -4.77994800e-01 -4.67740506e-01 -5.15091121e-01
-3.74082565e-01 1.16766882e+00 -4.45724010e-01 3.89644206e-02
-9.58152950e-01 -8.90956521e-01 5.43797851e-01 7.19270229e-01
5.58452964e-01 -1.06759167e+00 -1.16440475e-01 2.95449913e-01
-2.11259186e-01 -1.34773540e+00 -5.16359389e-01 9.67352390e-02
-9.80114341e-01 -4.46669340e-01 -5.66353440e-01 -8.33523452e-01
6.19842589e-01 3.81982684e-01 6.32761478e-01 -2.16625910e-02
7.49501050e-04 4.46187824e-01 -5.31885207e-01 -2.00475931e-01
-4.55981404e-01 3.01611964e-02 -2.12555781e-01 4.88717079e-01
3.92924517e-01 -3.49120319e-01 -1.05497137e-01 2.65649617e-01
-9.20994580e-01 4.32729423e-01 9.32361066e-01 8.71833563e-01
7.19678998e-01 -3.82782407e-02 2.08828777e-01 -9.73232627e-01
3.47505242e-01 -2.69604951e-01 -5.08770525e-01 5.15950680e-01
-4.68748093e-01 1.32701993e-01 6.18320346e-01 -6.36595607e-01
-1.47396994e+00 7.22219571e-02 -3.08750719e-01 -2.34285787e-01
-2.33075857e-01 6.16651654e-01 -3.44475508e-01 1.55837322e-02
7.74026960e-02 9.05229151e-01 -3.95542532e-01 -7.04098523e-01
2.96391040e-01 8.81578684e-01 3.84297520e-01 -5.07093251e-01
5.94947636e-01 2.75842160e-01 -2.57337630e-01 -1.14128447e+00
-6.17795348e-01 -4.23802286e-01 -8.75174344e-01 8.63034502e-02
7.15487242e-01 -1.02226853e+00 -2.91560113e-01 7.68164515e-01
-1.12731504e+00 -5.00273883e-01 1.92234203e-01 5.01336515e-01
-3.76390100e-01 1.02375770e+00 -7.51252413e-01 -1.33327639e+00
-3.28670979e-01 -1.15218604e+00 1.24035907e+00 -4.71622571e-02
2.68088952e-02 -1.01576471e+00 -3.20951968e-01 3.27903509e-01
2.73559362e-01 -1.85829192e-01 1.05458295e+00 -8.15758467e-01
-6.58291578e-01 -3.10120523e-01 -1.34967744e-01 4.23588723e-01
7.14783743e-03 -3.85449938e-02 -8.93314898e-01 -1.78606138e-01
9.40629542e-02 2.72497740e-02 1.04911220e+00 3.55032235e-01
1.26121891e+00 -4.21907395e-01 -2.23066986e-01 4.87593949e-01
1.27474701e+00 3.16015154e-01 8.92994463e-01 3.81335676e-01
9.46409941e-01 3.69251758e-01 4.53189790e-01 6.23689175e-01
4.43644583e-01 7.00894833e-01 1.55104265e-01 1.13776527e-01
-3.36780883e-02 -4.09093261e-01 8.41024756e-01 1.21545982e+00
2.38676995e-01 -5.62008381e-01 -9.51991677e-01 2.36516893e-01
-2.10777545e+00 -6.91345632e-01 -3.71381819e-01 2.33136439e+00
6.66651368e-01 3.56166601e-01 -3.45689118e-01 1.18465409e-01
7.26700008e-01 3.26114386e-01 -4.33735728e-01 -5.06942928e-01
-5.18706441e-01 4.73936833e-02 4.34066862e-01 7.15026855e-01
-1.09506392e+00 1.40630460e+00 5.96464777e+00 1.02001762e+00
-1.15536642e+00 -9.73042697e-02 5.51869094e-01 1.90823898e-01
-2.06903994e-01 5.40223271e-02 -1.26663911e+00 6.16757095e-01
9.57288444e-01 3.23050499e-01 4.01701123e-01 6.44000411e-01
4.34087425e-01 1.14018125e-02 -8.60570431e-01 9.95660901e-01
4.89400655e-01 -9.95815814e-01 6.07178986e-01 -2.55062133e-01
5.02645075e-01 4.75706160e-02 1.84792086e-01 5.38431764e-01
1.63998023e-01 -1.03741515e+00 7.01304793e-01 7.35458434e-01
6.00666463e-01 -4.45704520e-01 5.80391705e-01 7.52234638e-01
-1.32355738e+00 -1.27379820e-01 -3.51591289e-01 -5.34489527e-02
6.14585355e-02 2.24677324e-01 -7.32357800e-01 5.39290667e-01
2.46835276e-01 8.24778140e-01 -7.21970558e-01 7.72718549e-01
-3.97497296e-01 1.03296924e+00 -1.86398774e-01 -5.32959811e-02
2.10865393e-01 -2.75265872e-01 5.85759103e-01 1.67855620e+00
1.24406993e-01 -2.01646853e-02 3.81301254e-01 4.23732311e-01
6.83214590e-02 4.80856657e-01 -3.23775470e-01 -3.00933775e-02
2.83851981e-01 8.41009557e-01 -6.09089553e-01 -4.20344174e-01
-7.86832869e-01 1.12400317e+00 1.22906975e-01 5.87401927e-01
-7.08800137e-01 -3.02618481e-02 7.60153085e-02 -1.19889386e-01
5.64986050e-01 -7.17616737e-01 -5.83757699e-01 -1.47015619e+00
1.31244421e-01 -8.29621792e-01 3.99828225e-01 -8.41527581e-01
-1.07527149e+00 5.20634830e-01 -2.88385123e-01 -1.06947017e+00
-3.31784874e-01 -7.03912616e-01 -6.55971229e-01 1.08060491e+00
-1.81701756e+00 -1.47769594e+00 -1.92071274e-02 4.10832226e-01
1.27718377e+00 -3.04395795e-01 6.43258393e-01 2.14726217e-02
-8.90753865e-01 8.10576439e-01 3.53878945e-01 2.76258856e-01
6.36424422e-01 -9.61972713e-01 4.83859867e-01 1.11241210e+00
2.72118092e-01 6.08531892e-01 4.20365244e-01 -8.03573668e-01
-1.46442413e+00 -1.31940794e+00 1.02826095e+00 -4.18695003e-01
5.81677735e-01 -5.61243653e-01 -1.24186897e+00 1.03374243e+00
2.76929997e-02 -4.21464264e-01 7.69589424e-01 3.69145386e-02
-4.21781451e-01 2.86141247e-01 -6.04069114e-01 8.97957623e-01
6.42343938e-01 -5.20350039e-01 -7.29367614e-01 6.94159046e-02
8.46531451e-01 -3.73202026e-01 -4.40607756e-01 3.94234955e-01
3.88396084e-01 -5.54390192e-01 7.03351855e-01 -3.89290154e-01
3.06533605e-01 -3.89559388e-01 -4.99748915e-01 -5.18528819e-01
-2.27656275e-01 -4.81478781e-01 -3.54257196e-01 1.45148337e+00
3.78686041e-01 -5.65528452e-01 6.14229500e-01 6.33980095e-01
-2.52686679e-01 -5.96799672e-01 -7.32535720e-01 -6.13948703e-01
1.47513494e-01 -8.76509786e-01 1.86318070e-01 7.45361328e-01
-5.15191734e-01 4.54479814e-01 -8.61032128e-01 5.18921435e-01
3.36413741e-01 -1.45071805e-01 7.22878397e-01 -9.06707764e-01
-5.16413748e-01 -3.84823322e-01 -2.50997156e-01 -1.78043008e+00
3.45782161e-01 -7.20960796e-01 8.52487460e-02 -1.36663508e+00
1.96873724e-01 -1.88535139e-01 -3.31319124e-01 5.29510021e-01
-2.62778044e-01 -9.78129953e-02 2.36626387e-01 3.32512349e-01
-6.90589368e-01 8.42599332e-01 6.69608057e-01 -1.52273044e-01
-4.25967544e-01 6.44183978e-02 -3.25856656e-01 8.98953497e-01
6.63028657e-01 -2.90224731e-01 -2.79652804e-01 -7.95975327e-01
-6.03577867e-02 -2.68584304e-03 1.91216722e-01 -7.49627590e-01
4.43553030e-01 -1.98072240e-01 6.10611022e-01 -1.03315008e+00
5.00704467e-01 -6.54566050e-01 -1.88697338e-01 1.11134440e-01
-4.82497722e-01 -1.48782462e-01 3.94775659e-01 7.98174798e-01
-1.69396609e-01 -4.73487586e-01 4.93750960e-01 2.20345743e-02
-9.43343520e-01 -3.20914178e-03 -7.21750438e-01 -3.77660930e-01
5.39479017e-01 -4.35070872e-01 -2.71727759e-02 -1.94503337e-01
-6.77275240e-01 1.73939586e-01 2.78605312e-01 5.65399468e-01
8.18321943e-01 -9.06491280e-01 -6.94880903e-01 3.76137257e-01
-2.10026689e-02 2.20937072e-03 2.41273448e-01 7.04297364e-01
-1.19483143e-01 5.69426298e-01 5.14258623e-01 -7.36134171e-01
-1.33370328e+00 5.40031195e-01 4.35894579e-02 -5.17017305e-01
-6.61610484e-01 6.23973727e-01 4.18865681e-01 -5.05959451e-01
5.12407422e-01 -1.22050926e-01 -4.76022691e-01 -3.11997563e-01
5.97365081e-01 1.07553348e-01 -9.35666710e-02 -7.33514786e-01
-4.09417450e-02 8.29832792e-01 -5.77553809e-01 -1.40754148e-01
1.17900956e+00 -6.17829442e-01 -2.75340062e-02 8.71076763e-01
7.36244977e-01 1.68011099e-01 -1.17305982e+00 -6.18018687e-01
2.02569395e-01 -2.57464528e-01 2.00610429e-01 -7.58615255e-01
-4.34650779e-01 1.09007609e+00 2.82365799e-01 -3.91553432e-01
1.19493198e+00 -3.95444810e-01 6.75801337e-01 5.28020322e-01
1.21352822e-01 -9.96681690e-01 6.67906851e-02 9.92644668e-01
7.09033668e-01 -1.12362957e+00 -1.89101651e-01 -4.08267438e-01
-8.88616979e-01 1.28937662e+00 5.32850564e-01 1.21386506e-01
4.73778993e-01 2.14004859e-01 2.09791027e-02 5.09766817e-01
-1.01612270e+00 2.29381956e-02 2.17843711e-01 2.41172791e-01
5.76690257e-01 -1.77069813e-01 -2.25645959e-01 6.90133750e-01
2.37442866e-01 -9.88396406e-02 4.65333283e-01 1.11800134e+00
-5.44968426e-01 -1.28057778e+00 -4.93502855e-01 2.36304387e-01
-3.84446174e-01 -5.67221403e-01 -2.51408547e-01 3.99955481e-01
-3.02029163e-01 9.82433021e-01 -3.34016653e-03 -2.75298357e-01
-5.49105704e-02 4.65914756e-01 2.26126581e-01 -7.27629066e-01
-2.27917358e-01 5.50939322e-01 -1.31789818e-01 3.55582014e-02
-1.06116161e-01 -6.82442069e-01 -1.17215788e+00 5.63750453e-02
-3.23251724e-01 -1.81327492e-01 7.60073900e-01 1.18821788e+00
2.88268477e-01 5.12347937e-01 6.76100731e-01 -6.27841532e-01
-4.27041352e-01 -9.98564780e-01 -2.51474559e-01 -8.27467069e-02
2.45872855e-01 -3.10230315e-01 -1.50261492e-01 4.01999772e-01] | [11.911980628967285, 2.178516387939453] |
667f2460-ad61-4f15-82cc-533a20e71d3f | generating-smooth-pose-sequences-for-diverse | 2108.08422 | null | https://arxiv.org/abs/2108.08422v3 | https://arxiv.org/pdf/2108.08422v3.pdf | Generating Smooth Pose Sequences for Diverse Human Motion Prediction | Recent progress in stochastic motion prediction, i.e., predicting multiple possible future human motions given a single past pose sequence, has led to producing truly diverse future motions and even providing control over the motion of some body parts. However, to achieve this, the state-of-the-art method requires learning several mappings for diversity and a dedicated model for controllable motion prediction. In this paper, we introduce a unified deep generative network for both diverse and controllable motion prediction. To this end, we leverage the intuition that realistic human motions consist of smooth sequences of valid poses, and that, given limited data, learning a pose prior is much more tractable than a motion one. We therefore design a generator that predicts the motion of different body parts sequentially, and introduce a normalizing flow based pose prior, together with a joint angle loss, to achieve motion realism.Our experiments on two standard benchmark datasets, Human3.6M and HumanEva-I, demonstrate that our approach outperforms the state-of-the-art baselines in terms of both sample diversity and accuracy. The code is available at https://github.com/wei-mao-2019/gsps | ['Mathieu Salzmann', 'Miaomiao Liu', 'Wei Mao'] | 2021-08-19 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Mao_Generating_Smooth_Pose_Sequences_for_Diverse_Human_Motion_Prediction_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Mao_Generating_Smooth_Pose_Sequences_for_Diverse_Human_Motion_Prediction_ICCV_2021_paper.pdf | iccv-2021-1 | ['human-pose-forecasting'] | ['computer-vision'] | [ 1.03828507e-02 1.74003035e-01 -2.57486790e-01 -2.29833290e-01
-7.26932764e-01 -5.26685953e-01 6.96248293e-01 -6.16107941e-01
-1.03095576e-01 7.98871994e-01 6.51813626e-01 9.25632194e-02
3.84973645e-01 -5.22802591e-01 -8.28578174e-01 -5.77414930e-01
-1.22127876e-01 5.67095220e-01 1.91588491e-01 -2.79909015e-01
-1.55241340e-01 2.82532245e-01 -1.28706491e+00 -5.75744286e-02
6.41497195e-01 5.18116057e-01 1.95151225e-01 9.09606159e-01
5.57797730e-01 7.64063656e-01 -2.82299936e-01 -5.58675885e-01
3.35325122e-01 -7.68683374e-01 -7.69617438e-01 2.40396723e-01
3.90116751e-01 -5.44379473e-01 -6.82973564e-01 6.64200008e-01
7.00856268e-01 5.23076236e-01 6.63054585e-01 -1.27992165e+00
-4.83890474e-01 3.68555725e-01 -5.01644671e-01 -2.80340046e-01
5.76546252e-01 6.81105494e-01 9.92857158e-01 -6.82347596e-01
1.07232928e+00 1.35031784e+00 5.44834733e-01 1.11710501e+00
-1.21019971e+00 -4.50293988e-01 1.32737264e-01 -7.19109327e-02
-1.20689666e+00 -5.04955828e-01 6.92127764e-01 -5.62485099e-01
6.55131400e-01 2.24573582e-01 8.18981409e-01 1.68716419e+00
4.23691630e-01 9.89477158e-01 3.43531787e-01 -1.07788267e-02
1.55380920e-01 -5.17865419e-01 -5.21640182e-01 5.98401725e-01
2.95842707e-01 4.91115823e-02 -4.29344535e-01 -8.75973627e-02
1.08650839e+00 -2.78015975e-02 -4.55566972e-01 -8.07974994e-01
-1.46210706e+00 7.43481219e-01 1.61402255e-01 -1.31220207e-01
-3.67208183e-01 6.35681033e-01 3.03014070e-01 -1.83024511e-01
1.95746452e-01 4.54881519e-01 -2.64451414e-01 -3.97303075e-01
-9.13596213e-01 9.97333348e-01 7.79604077e-01 1.00784910e+00
3.41303319e-01 1.71881735e-01 -3.81922305e-01 2.39277974e-01
2.97637045e-01 5.51063657e-01 3.86553258e-01 -1.47730494e+00
5.63355029e-01 1.09343544e-01 5.15428603e-01 -9.98483658e-01
-3.48971486e-01 -4.90348160e-01 -8.84932756e-01 1.08154230e-01
5.05380392e-01 -4.65112090e-01 -1.04310501e+00 2.20867562e+00
5.43872416e-01 2.72785753e-01 -4.84720543e-02 1.21432388e+00
2.11673230e-01 7.03094482e-01 8.44942331e-02 1.76599443e-01
9.18485701e-01 -1.11526215e+00 -5.30337036e-01 -3.54685456e-01
4.40817058e-01 -6.38327003e-01 1.10269749e+00 2.45218858e-01
-1.32753670e+00 -6.46440685e-01 -9.49661255e-01 -2.34369516e-01
4.34533805e-01 1.19455606e-01 5.90440750e-01 2.33390749e-01
-8.19287241e-01 7.99553812e-01 -1.42820990e+00 -3.79428118e-01
2.18078822e-01 1.80815950e-01 -3.44374269e-01 -2.68779360e-02
-1.10737097e+00 6.88565135e-01 6.46521971e-02 6.29571965e-03
-1.18362641e+00 -6.30793154e-01 -1.00001144e+00 -6.23055287e-02
3.01903337e-01 -1.41950524e+00 1.31137228e+00 -6.81323349e-01
-1.59617960e+00 5.19024909e-01 -8.57094601e-02 -4.76500779e-01
1.06756234e+00 -6.67813182e-01 -4.51294333e-02 1.48409978e-01
1.35301188e-01 1.14824402e+00 6.62864685e-01 -1.19739556e+00
-3.88876051e-01 -4.18376066e-02 -1.72629897e-02 4.32338893e-01
6.67266697e-02 -3.80825430e-01 -8.33806336e-01 -8.58298123e-01
-1.55771539e-01 -1.50810063e+00 -5.98437428e-01 2.87565768e-01
-5.27307451e-01 1.71725661e-01 4.44836915e-01 -7.63538122e-01
1.24341393e+00 -1.92886901e+00 7.64859915e-01 7.25348201e-03
1.53949961e-01 2.37291586e-02 -1.10496812e-01 4.32886809e-01
1.48758903e-01 -7.71167204e-02 -2.88741410e-01 -7.74106622e-01
1.10772774e-01 2.62826025e-01 -2.83002108e-01 5.27143419e-01
1.42885998e-01 1.10868347e+00 -9.81117308e-01 -2.76378036e-01
1.99461460e-01 7.00254500e-01 -8.89869034e-01 2.63037443e-01
-4.42082763e-01 1.05481160e+00 -4.36879724e-01 3.34637374e-01
2.79313475e-01 -4.20261562e-01 2.84830451e-01 6.33806512e-02
2.08345547e-01 1.77489832e-01 -1.20259154e+00 2.18180561e+00
-2.11294308e-01 3.47911954e-01 -2.05800980e-01 -4.26742285e-01
5.88928342e-01 4.25673366e-01 7.06091940e-01 -2.29366068e-02
1.75066158e-01 1.49285600e-01 -7.14712217e-02 -3.93062502e-01
4.95464385e-01 5.73989153e-02 -1.42049283e-01 2.01972902e-01
-1.44261673e-01 -1.64989218e-01 1.95064455e-01 4.42855805e-02
9.09017265e-01 7.63260543e-01 2.36207515e-01 -3.84670049e-02
2.30007365e-01 -1.10179074e-02 9.43887293e-01 4.70965832e-01
-2.44501963e-01 1.08899534e+00 3.21605295e-01 -4.33028281e-01
-1.38101852e+00 -1.32398713e+00 4.97349173e-01 6.09038651e-01
3.49068075e-01 -4.01820391e-01 -7.23116994e-01 -2.30691165e-01
4.03954759e-02 5.52413821e-01 -4.23979104e-01 -1.82986721e-01
-1.20958459e+00 -4.48114663e-01 3.38019431e-01 7.33195364e-01
3.23123038e-01 -7.92086303e-01 -9.24131513e-01 2.51219362e-01
-4.28103566e-01 -1.13265038e+00 -9.47769523e-01 -5.20489454e-01
-8.32271695e-01 -7.75627792e-01 -1.18344390e+00 -6.29642129e-01
3.77435416e-01 2.99019925e-02 1.34441400e+00 -8.18998888e-02
-2.38780901e-01 2.36216918e-01 -2.09598675e-01 3.83399613e-02
-3.28361630e-01 1.59467757e-01 1.74303532e-01 -1.56895071e-01
-2.45661229e-01 -7.88002849e-01 -1.07354629e+00 1.36264607e-01
-7.44138718e-01 4.98392284e-01 5.33825696e-01 8.16642642e-01
6.76122129e-01 -4.97374266e-01 8.00058767e-02 -4.77462322e-01
2.14712456e-01 -5.00155270e-01 -2.43794858e-01 -5.33658254e-04
-1.08914837e-01 2.37495750e-01 5.84780991e-01 -6.00845575e-01
-1.01459765e+00 4.79898304e-01 -1.89777374e-01 -6.60700202e-01
-3.84955667e-02 1.37411878e-01 -3.36479872e-01 3.46699238e-01
6.54449344e-01 1.72739476e-01 -1.50441928e-02 -3.83856326e-01
6.56471908e-01 -5.82717694e-02 9.45686698e-01 -7.73006439e-01
9.50763822e-01 6.12454712e-01 2.63720393e-01 -5.07064164e-01
-5.50434351e-01 -2.20299214e-01 -6.11276984e-01 -2.24391282e-01
1.22322106e+00 -1.12883198e+00 -5.08140147e-01 4.65839446e-01
-1.19044948e+00 -6.24169171e-01 -2.41158903e-01 6.36076212e-01
-1.04792237e+00 3.53941739e-01 -7.04561353e-01 -6.26012206e-01
-2.94826865e-01 -1.28271592e+00 1.30079770e+00 1.10395864e-01
-8.23620379e-01 -7.57354677e-01 3.48590821e-01 1.05402187e-01
9.23346058e-02 8.55013132e-01 3.62268746e-01 1.34110618e-02
-7.85478354e-01 -9.42170322e-02 5.24378419e-01 7.09634507e-03
4.45170887e-02 -5.86087443e-03 -5.02563536e-01 -4.97354239e-01
-2.72044331e-01 -3.38903069e-01 7.93180823e-01 5.04899621e-01
9.45968866e-01 -4.31416720e-01 -3.88098806e-01 9.02942598e-01
1.10376775e+00 -3.48720551e-02 7.27535248e-01 1.24106564e-01
1.00624740e+00 4.93525624e-01 7.02533185e-01 6.95305705e-01
5.77296674e-01 9.12162125e-01 4.53064114e-01 1.96760505e-01
-3.61037314e-01 -7.21317947e-01 3.99405688e-01 5.17777741e-01
-2.43355617e-01 -5.29918551e-01 -8.66768777e-01 6.45284951e-01
-2.06476331e+00 -1.06775808e+00 1.57920405e-01 2.07439256e+00
7.83000410e-01 9.11838859e-02 4.21219260e-01 -2.25837484e-01
5.76446354e-01 3.29031497e-01 -7.96592355e-01 1.16269544e-01
6.36346787e-02 -4.06010225e-02 3.53777766e-01 6.80035770e-01
-1.11995566e+00 9.76553977e-01 5.69094896e+00 5.11158168e-01
-1.08730137e+00 -1.56006545e-01 6.81504965e-01 -4.64303851e-01
-5.03251314e-01 -6.09948076e-02 -8.19649816e-01 3.98225784e-01
6.92066550e-01 -1.99818075e-01 2.17744261e-01 8.59989822e-01
3.95631582e-01 1.90745920e-01 -1.13983202e+00 9.82685089e-01
-1.48923054e-01 -1.30799305e+00 2.83337981e-01 -8.78482033e-03
1.04462421e+00 -1.62938759e-01 1.78869501e-01 1.67607944e-02
4.49641019e-01 -8.99291754e-01 1.05973411e+00 7.04399347e-01
6.44349694e-01 -7.00240433e-01 2.23559722e-01 4.05064791e-01
-1.31329072e+00 1.77174702e-01 -9.81216580e-02 -7.39987046e-02
6.95564389e-01 2.81496495e-01 -3.31617802e-01 6.09923363e-01
4.01089728e-01 8.03625882e-01 -2.43213564e-01 8.42344522e-01
-4.37465250e-01 4.20248926e-01 -2.41552994e-01 1.81293413e-01
2.13706464e-01 -2.25020070e-02 7.68244565e-01 9.18978512e-01
4.26738352e-01 3.79787460e-02 3.71544242e-01 8.11627150e-01
6.55875280e-02 -2.35125318e-01 -4.26775992e-01 1.31541491e-01
4.66623187e-01 9.22776580e-01 -3.42326343e-01 -1.71393484e-01
-1.71140954e-01 1.28538287e+00 1.38328686e-01 3.55107218e-01
-1.10332239e+00 4.41886522e-02 1.06162906e+00 1.13222487e-01
2.89328456e-01 -7.19092548e-01 -2.92425275e-01 -1.51179302e+00
2.84153312e-01 -8.60134363e-01 1.35009319e-01 -6.34154856e-01
-9.28439319e-01 4.48563904e-01 7.60365874e-02 -1.43467414e+00
-9.74707961e-01 -2.07699209e-01 -6.21382415e-01 8.11955154e-01
-8.16984951e-01 -1.31433630e+00 -1.67609230e-01 3.60867590e-01
6.49126291e-01 1.70763686e-01 5.25821567e-01 1.64996311e-01
-4.01816249e-01 5.76932132e-01 -2.47648880e-01 6.88081607e-02
7.64308631e-01 -1.09653461e+00 1.13910055e+00 1.03157651e+00
6.08496293e-02 5.84074140e-01 1.18937790e+00 -7.81368256e-01
-1.56587195e+00 -1.19750321e+00 7.52794147e-01 -6.93405390e-01
4.46810097e-01 -3.21121156e-01 -6.67435050e-01 8.63012373e-01
-1.05069794e-01 -4.63270955e-03 2.41513819e-01 -4.86922204e-01
5.65819591e-02 2.85303175e-01 -7.74794757e-01 1.19770396e+00
1.43983734e+00 -1.43124029e-01 -2.71575540e-01 1.51326522e-01
8.47267389e-01 -7.77753949e-01 -8.03389609e-01 5.14531910e-01
9.58944678e-01 -8.07755411e-01 1.14256322e+00 -6.38498664e-01
6.50053144e-01 -4.25791532e-01 -1.48248285e-01 -1.13412821e+00
-3.59569579e-01 -1.03261638e+00 -4.16078746e-01 8.10603678e-01
3.08737576e-01 -6.86669797e-02 1.13762867e+00 7.45474517e-01
-1.03717998e-01 -1.01862609e+00 -6.17557287e-01 -9.54743445e-01
2.41585866e-01 -2.52434224e-01 5.34943342e-01 5.81144571e-01
-3.61139417e-01 2.12212875e-01 -1.10180283e+00 1.72068655e-01
6.36721492e-01 7.95775205e-02 1.33328676e+00 -6.40069962e-01
-8.32770944e-01 -2.81580925e-01 -4.58168983e-01 -1.58645666e+00
1.05063029e-01 -3.82239968e-01 1.76689178e-01 -1.54196501e+00
6.95727095e-02 9.05925781e-02 4.30156380e-01 1.46552667e-01
-4.84136015e-01 9.68183652e-02 5.28055429e-01 4.09183979e-01
-4.44955528e-01 9.33939874e-01 1.70751202e+00 1.72859535e-01
-3.25084805e-01 4.28840071e-02 -4.43605393e-01 7.76992261e-01
8.22091043e-01 -2.08641097e-01 -6.42167509e-01 -6.17454171e-01
5.60304858e-02 4.46823269e-01 5.68803906e-01 -1.29285705e+00
1.48812532e-01 -3.16182315e-01 5.46859205e-01 -5.67649543e-01
6.70591593e-01 -2.82072276e-01 6.08140051e-01 8.98016691e-01
-3.81438613e-01 2.04455391e-01 -9.51835811e-02 6.89661682e-01
-2.48164050e-02 3.66962940e-01 7.34754264e-01 -2.13054270e-01
-7.23918140e-01 7.23498762e-01 -9.29400474e-02 3.21684450e-01
8.11196029e-01 -6.55659586e-02 1.35552685e-03 -8.20743322e-01
-8.98355186e-01 3.69016051e-01 8.04276586e-01 6.15008533e-01
4.63369429e-01 -1.58844948e+00 -8.37722957e-01 -6.75712675e-02
-1.30614385e-01 2.14060247e-01 2.54289955e-01 5.92024863e-01
-7.13764250e-01 2.50824422e-01 -3.05185050e-01 -6.77120328e-01
-9.84755456e-01 3.52631897e-01 2.46287316e-01 -2.98834026e-01
-8.09289992e-01 8.17455888e-01 4.63456303e-01 -1.95508614e-01
7.13442788e-02 -2.19867736e-01 1.95181400e-01 -6.33174360e-01
3.37307900e-01 5.37775397e-01 -6.13874793e-01 -7.42246449e-01
-2.89816588e-01 6.22040033e-01 3.06713372e-01 -4.92107511e-01
1.04070818e+00 -2.82596081e-01 4.61189449e-01 2.54488677e-01
1.10766196e+00 -6.78879544e-02 -1.93629885e+00 1.65615782e-01
-1.70549214e-01 -4.89589691e-01 -7.13773727e-01 -3.78049433e-01
-9.80906546e-01 7.38067269e-01 1.90374658e-01 -5.01789749e-01
9.58188474e-01 -4.80361953e-02 1.16347635e+00 2.22915143e-01
4.43889052e-01 -6.84469044e-01 1.87033340e-01 5.71536958e-01
1.02280927e+00 -1.09076071e+00 -5.56629524e-02 -4.88832325e-01
-9.96129632e-01 9.04252112e-01 7.11973429e-01 -4.06144053e-01
4.01015162e-01 3.73903103e-02 -5.78182787e-02 1.62123069e-01
-8.62326682e-01 -9.16019082e-04 5.27414501e-01 4.63374048e-01
5.92662752e-01 1.78827927e-01 -4.05889958e-01 4.34949666e-01
-5.15505612e-01 1.36338651e-01 4.11519021e-01 8.38458478e-01
-3.24415028e-01 -1.10790420e+00 -2.98981518e-01 -7.41561800e-02
-4.91154045e-01 2.10504353e-01 -2.53705949e-01 9.12414849e-01
-6.05190657e-02 6.68268025e-01 -1.32744685e-01 -4.82553929e-01
1.95341080e-01 -1.87035397e-01 6.54076993e-01 -5.81616104e-01
-5.79200238e-02 2.23278567e-01 1.57296345e-01 -9.65759158e-01
-2.68310159e-01 -8.99488807e-01 -1.24402821e+00 -4.51056033e-01
3.15419704e-01 -1.74155176e-01 1.23437189e-01 6.19584620e-01
4.56672221e-01 3.92503351e-01 3.48864913e-01 -1.34846425e+00
-8.59268546e-01 -7.33784497e-01 -3.21144134e-01 7.45895684e-01
3.24140817e-01 -6.95857882e-01 -1.21111505e-01 3.15359563e-01] | [7.28617000579834, -0.24027295410633087] |
309c4c6f-e0b2-4dee-a362-94d5cf70b2bb | graph-networks-as-a-universal-machine | 1812.05055 | null | http://arxiv.org/abs/1812.05055v1 | http://arxiv.org/pdf/1812.05055v1.pdf | Graph Networks as a Universal Machine Learning Framework for Molecules and Crystals | Graph networks are a new machine learning (ML) paradigm that supports both
relational reasoning and combinatorial generalization. Here, we develop, for
the first time, universal MatErials Graph Network (MEGNet) models for accurate
property prediction in both molecules and crystals. We demonstrate that our
MEGNet models significantly outperform prior ML models in 11 out of 13
properties of the QM9 molecule data set. Furthermore, a single-task unified
MEGNet model can accurately predict the internal energy at 0 K and room
temperature, enthalpy and Gibbs free energy, with temperature, pressure and
entropy being global state inputs. Similarly, we show that MEGNet models
trained on $\sim 60,000$ crystals in the Materials Project substantially
outperform prior ML models in the prediction of the formation energies, band
gaps and elastic moduli of crystals, achieving better than DFT accuracy over a
much larger data set. Such MEGNet models are highly interpretable, and
well-established periodic chemical trends can be extracted from the elemental
embeddings. Finally, we demonstrate the transfer learning of elemental
embeddings from a property model trained on a larger data set (formation
energies) to accelerate the training of property models with smaller amounts of
data (band gaps and elastic moduli) | ['Yunxing Zuo', 'Chi Chen', 'Weike Ye', 'Shyue Ping Ong', 'Chen Zheng'] | 2018-12-12 | graph-networks-as-a-universal-machine-1 | null | null | chem-mater-2018-12 | ['formation-energy'] | ['miscellaneous'] | [ 1.71816394e-01 5.02626538e-01 -4.88226384e-01 -2.56933481e-01
-4.29145396e-01 -2.31637076e-01 5.57794571e-01 7.19277799e-01
-4.00867350e-02 1.09495103e+00 3.98063302e-01 -4.33421075e-01
-2.49456644e-01 -1.27755439e+00 -9.73693490e-01 -8.64754021e-01
-3.39034915e-01 7.03265190e-01 -1.72811206e-02 -3.77510548e-01
4.29303735e-01 5.23186088e-01 -1.27657247e+00 -1.10349491e-01
8.88220727e-01 1.16747463e+00 9.77588147e-02 5.70581794e-01
3.24368954e-01 7.39974558e-01 2.07701907e-01 -2.88565665e-01
-1.70191545e-02 -1.01064488e-01 -1.03790426e+00 -5.84646583e-01
4.58237261e-01 -5.51210307e-02 -8.25347960e-01 7.78103173e-01
3.60199630e-01 3.64562541e-01 1.25677979e+00 -7.95476019e-01
-1.45880747e+00 6.03559971e-01 1.76011860e-01 -1.89784601e-01
4.17277753e-01 4.10390496e-01 1.84875381e+00 -8.72986436e-01
9.41659331e-01 8.73488784e-01 6.40783727e-01 5.75082302e-01
-1.62799442e+00 -4.04476643e-01 -3.02690268e-01 3.54282022e-01
-1.27280736e+00 -1.91334456e-01 7.98887610e-01 -3.49542856e-01
2.01255774e+00 -6.68466762e-02 8.90068293e-01 8.97862017e-01
8.23407054e-01 -7.07553327e-02 7.52957165e-01 -3.14914763e-01
3.23570669e-01 -3.18744779e-01 2.10880637e-01 8.79497409e-01
7.44786084e-01 1.18852593e-01 -8.74470353e-01 -3.91285092e-01
6.45016849e-01 -3.67165282e-02 -2.86156893e-01 -4.21876937e-01
-1.10132694e+00 7.03657746e-01 7.66349733e-01 1.25390857e-01
-3.11306864e-01 8.66949379e-01 -1.44131752e-02 2.14423627e-01
5.91797292e-01 1.09532952e+00 -8.90643835e-01 -1.13656573e-01
-3.06775540e-01 4.37711686e-01 1.01845920e+00 9.50002611e-01
1.14965117e+00 2.17279904e-02 5.48440933e-01 2.33004123e-01
5.14566779e-01 4.13401365e-01 3.95357842e-03 -5.64412773e-01
3.73465061e-01 6.79223418e-01 1.83816012e-02 -7.70081520e-01
-7.16214895e-01 -1.78491667e-01 -5.20106256e-01 -1.58799872e-01
2.56198412e-03 3.42683554e-01 -9.64845419e-01 1.66006255e+00
9.65411291e-02 -2.38775626e-01 3.15913230e-01 2.65924424e-01
1.22863162e+00 8.67159605e-01 3.96073729e-01 -1.19619966e-01
8.74720633e-01 -6.49868131e-01 -2.84816861e-01 6.08939398e-03
9.89035308e-01 -3.55504811e-01 8.39718997e-01 2.22352549e-01
-1.15773952e+00 -5.29443324e-01 -1.35985935e+00 -5.57022512e-01
-7.44381189e-01 -5.27582228e-01 1.15051830e+00 3.68478805e-01
-1.01529264e+00 1.54607868e+00 -8.29849541e-01 -1.38423413e-01
3.31034631e-01 9.64925587e-01 -4.55603153e-01 -1.63351819e-01
-1.23599768e+00 1.05048752e+00 5.84487617e-01 -2.95314401e-01
-6.94626451e-01 -1.20611060e+00 -8.40251446e-01 -2.62494739e-02
1.27823502e-01 -6.86257064e-01 6.48897827e-01 -3.43349092e-02
-1.35575962e+00 7.54223168e-01 -2.69457158e-02 -2.76200742e-01
-3.83905321e-01 2.86885351e-01 -7.47833371e-01 1.68109626e-01
-2.88757116e-01 6.54296458e-01 2.29033053e-01 -8.69242251e-01
3.24417174e-01 -8.30969885e-02 -3.23199406e-02 -1.54601768e-01
-4.26750660e-01 -7.13451445e-01 3.74708295e-01 -7.92335123e-02
2.89011449e-01 -6.92968369e-01 -2.39524230e-01 -2.69281924e-01
-4.77616578e-01 -5.62172472e-01 9.89875942e-02 -5.24219155e-01
9.62720692e-01 -1.72659349e+00 5.59024513e-01 4.82499838e-01
8.01593661e-01 -3.83914322e-01 -5.51177748e-02 9.68516231e-01
-3.81478101e-01 4.94912118e-01 -2.06226602e-01 2.01734662e-01
3.06506813e-01 7.59932697e-02 8.98974463e-02 4.62967575e-01
3.83619159e-01 1.47932518e+00 -7.32367396e-01 -1.30685091e-01
1.07178755e-01 3.79301608e-01 -9.85048175e-01 -1.42749222e-02
-9.03419137e-01 1.61495209e-01 -3.34105045e-01 7.22034395e-01
5.45898080e-01 -6.77035034e-01 4.45578188e-01 -4.74667400e-01
-7.99698755e-02 8.87427509e-01 -4.34543878e-01 1.72870696e+00
-3.39752771e-02 3.38374704e-01 -5.12163460e-01 -8.87869716e-01
1.00527263e+00 2.54874885e-01 7.16220737e-01 -8.17803204e-01
-1.00881115e-01 4.48433220e-01 5.04522204e-01 -4.62296247e-01
7.46147871e-01 -5.86428344e-01 -1.69851735e-01 5.36597729e-01
5.16836226e-01 -6.98369086e-01 5.26459627e-02 1.57002240e-01
1.01131845e+00 2.41385773e-01 2.30130509e-01 -6.95475519e-01
2.59477407e-01 2.94796396e-02 1.94752842e-01 3.34506840e-01
2.38586426e-01 2.65666008e-01 5.72001338e-01 -6.69130981e-01
-1.63622379e+00 -1.11960971e+00 -4.02111471e-01 9.62515473e-01
1.49116442e-01 -1.20764601e+00 -3.13070923e-01 -4.37599942e-02
4.20815885e-01 6.16167903e-01 -7.68202782e-01 -6.46148086e-01
-4.17746305e-01 -1.03822255e+00 1.09847821e-02 6.57616258e-01
5.80753796e-02 -1.06042874e+00 4.05698746e-01 4.19722915e-01
4.44137394e-01 -8.66930842e-01 -5.40363118e-02 5.26525557e-01
-8.16892982e-01 -1.20361257e+00 1.94985136e-01 -7.19042778e-01
5.33453047e-01 -3.63558859e-01 1.21096277e+00 3.15628558e-01
-4.75865185e-01 1.94456488e-01 -1.62385982e-02 -9.01472196e-02
-7.54675508e-01 4.07333434e-01 3.77399921e-01 -7.00838983e-01
3.37356001e-01 -9.54814076e-01 -7.45506883e-01 -6.27923757e-02
-6.16952300e-01 1.65541515e-01 3.02789867e-01 6.25565708e-01
7.34853685e-01 -7.93402940e-02 5.49903691e-01 -8.10435176e-01
5.53539097e-01 -4.83314306e-01 -1.33804992e-01 2.52512455e-01
-1.13722110e+00 5.97814083e-01 8.21834922e-01 -1.19052771e-02
-4.72214013e-01 -5.81237257e-01 -3.65669966e-01 3.30374718e-01
2.90338039e-01 6.36799574e-01 -2.28775710e-01 -4.16589528e-01
6.17690623e-01 1.46212578e-01 -2.82171875e-01 -3.07019264e-01
4.84156877e-01 2.41685465e-01 2.48779088e-01 -1.19754219e+00
6.33779049e-01 3.97400148e-02 8.01103294e-01 -7.62643456e-01
-5.17283738e-01 1.30851820e-01 -8.28042686e-01 2.45467499e-01
1.05370045e+00 -7.72090137e-01 -1.23050809e+00 1.83736533e-02
-7.56625772e-01 -3.46583277e-01 -3.21136117e-01 4.08436477e-01
-6.80053055e-01 2.44719401e-01 -7.32971549e-01 -3.29749197e-01
-5.49976528e-01 -9.60042000e-01 8.01225364e-01 -5.55232726e-02
-4.89042789e-01 -1.52750874e+00 2.20805276e-02 9.56459716e-02
4.33083832e-01 5.77362597e-01 1.86896002e+00 -6.19470596e-01
-9.68640089e-01 -1.21835396e-01 -1.94429025e-01 9.40394551e-02
4.53673691e-01 1.20572783e-01 -7.33317256e-01 -2.04546615e-01
-4.34616804e-01 -5.11512160e-01 1.06551158e+00 3.29932630e-01
1.12340415e+00 -1.53677776e-01 -3.95424694e-01 7.30566204e-01
1.61959159e+00 1.16752468e-01 6.79729879e-01 3.81758958e-02
1.29089427e+00 2.42776960e-01 -1.53497264e-01 9.11836028e-02
2.17246592e-01 9.78398696e-02 3.94013703e-01 2.52636969e-01
-2.22900026e-02 -6.63926899e-01 3.15572470e-01 1.16745782e+00
-8.74789298e-01 -2.35243246e-01 -8.99554670e-01 7.16822520e-02
-1.33471036e+00 -8.29483986e-01 -2.24178463e-01 1.93640196e+00
9.50341284e-01 3.97411764e-01 -1.51246905e-01 -7.36828521e-02
1.38686225e-01 4.76568639e-01 -1.05069196e+00 -4.05249238e-01
-3.00080866e-01 1.04833913e+00 9.09480631e-01 6.69909000e-01
-6.79203212e-01 1.05845404e+00 7.53928995e+00 6.15274370e-01
-8.63478482e-01 -6.66238889e-02 5.97812772e-01 6.17730059e-02
-1.16809833e+00 2.69336581e-01 -6.57826066e-01 1.65099904e-01
1.45152688e+00 -1.83320135e-01 6.54318273e-01 5.94733357e-01
-3.00714284e-01 2.03185081e-01 -1.64583874e+00 9.05763745e-01
-2.21813709e-01 -2.11807323e+00 6.10757507e-02 3.89843941e-01
8.35076571e-01 2.14692533e-01 3.55979241e-02 1.25434458e-01
2.74455994e-01 -1.53038299e+00 3.71013165e-01 6.46699667e-01
1.12026989e+00 -6.70240462e-01 9.30664167e-02 -3.49170297e-01
-1.14695764e+00 3.89413476e-01 -7.52983928e-01 -1.78730309e-01
-1.10351421e-01 3.82962853e-01 -9.06144381e-01 7.33716965e-01
2.33746096e-01 1.18056691e+00 -3.68918449e-01 1.90436035e-01
-2.23834798e-01 4.80817467e-01 -3.15859050e-01 -5.17006457e-01
1.60575286e-02 -4.57666427e-01 7.49307275e-02 5.45094550e-01
1.07624210e-01 2.05629960e-01 -3.67737472e-01 1.35183644e+00
-6.61891401e-01 -1.78763606e-02 -5.52298307e-01 -7.60166883e-01
1.92276299e-01 9.11899209e-01 -4.47237968e-01 -1.84326798e-01
-3.10480982e-01 4.08947796e-01 4.10618097e-01 4.54454243e-01
-6.80929542e-01 -2.25256920e-01 8.36796105e-01 2.99030125e-01
3.56669277e-01 -7.15247869e-01 -1.22189425e-01 -1.02351642e+00
-2.02415779e-01 -3.03593785e-01 -1.94241911e-01 -7.67554104e-01
-1.41062593e+00 1.38638154e-01 -1.54716522e-01 -5.37686825e-01
3.88442948e-02 -1.42504561e+00 -4.60455239e-01 8.23548377e-01
-1.42221570e+00 -8.40321660e-01 1.26180947e-01 4.89329584e-02
-3.85365695e-01 -6.04992136e-02 1.06076384e+00 -1.30965084e-01
-4.96194750e-01 2.66867816e-01 4.84376222e-01 -2.48866796e-01
4.29701984e-01 -1.41398406e+00 8.44579339e-01 -3.19881178e-02
-1.73135161e-01 7.40952253e-01 8.70275855e-01 -7.91624367e-01
-2.11159158e+00 -8.84978414e-01 6.86229587e-01 -5.59431076e-01
1.02721477e+00 -7.64685988e-01 -8.72973025e-01 6.85699403e-01
1.66661420e-03 1.22874174e-02 1.11837125e+00 3.96791071e-01
-4.98504072e-01 2.17580408e-01 -1.06363022e+00 4.72401232e-01
1.56378722e+00 -9.60069716e-01 -3.09103221e-01 8.66591573e-01
1.31161344e+00 -3.07231247e-01 -1.84027481e+00 3.25119853e-01
4.53879893e-01 -5.19547343e-01 1.19246578e+00 -1.11592436e+00
8.99804175e-01 2.25572079e-01 -5.95986962e-01 -1.02149618e+00
-6.57948315e-01 -6.56466305e-01 -3.21301162e-01 6.44654870e-01
9.62550044e-01 -9.56581295e-01 8.85815024e-01 1.21086705e+00
-2.88082749e-01 -1.15994501e+00 -8.29024971e-01 -6.51283085e-01
5.24189830e-01 -3.64117175e-01 9.68543291e-01 7.16251075e-01
3.03189635e-01 4.44390446e-01 -6.74372464e-02 -1.59750685e-01
4.62650627e-01 2.44188458e-01 2.19888747e-01 -1.47971344e+00
-4.14149582e-01 -2.45938838e-01 -5.24936557e-01 -9.11204755e-01
2.44450912e-01 -1.57247376e+00 -6.29915118e-01 -1.76021063e+00
2.21501336e-01 -4.61473018e-01 -4.51570511e-01 2.45592132e-01
2.34958500e-01 -1.18421577e-01 -1.71454325e-01 2.08250955e-01
-4.62640136e-01 1.01282859e+00 1.50766337e+00 -2.56309479e-01
-2.81230677e-02 -7.29137301e-01 -9.16568756e-01 2.08323970e-01
7.23765373e-01 -3.38931471e-01 -1.39877766e-01 1.13362692e-01
1.10123575e+00 -1.91362768e-01 1.54127017e-01 -8.98649335e-01
-3.23232673e-02 -2.45175257e-01 6.25567496e-01 -3.70292455e-01
6.00781381e-01 -3.86419445e-01 2.08986476e-01 4.50813293e-01
-2.64100164e-01 -1.52937487e-01 1.92894146e-01 6.56625152e-01
3.79924893e-01 -1.00116417e-01 2.50222623e-01 -2.20259458e-01
-3.73532653e-01 8.35402608e-01 -2.79423818e-02 -1.23299763e-01
7.07657278e-01 -3.43613446e-01 -7.18633413e-01 3.72800627e-04
-8.23497474e-01 1.21638710e-02 9.18777466e-01 -2.82756221e-02
6.89324141e-01 -1.52170622e+00 -2.77467221e-01 1.53027773e-01
1.52920648e-01 1.37821242e-01 1.36840865e-01 5.34934640e-01
-9.25092638e-01 5.93176067e-01 -6.16037622e-02 -1.43934056e-01
-5.18252611e-01 6.71116769e-01 4.03780162e-01 -1.84068412e-01
-3.26093167e-01 8.09579909e-01 4.33096774e-02 -3.82925421e-01
-6.64051354e-01 -8.70882809e-01 4.37436014e-01 -3.18751514e-01
-1.10310324e-01 3.07548821e-01 1.45902485e-01 -3.36498469e-01
-3.79366964e-01 7.57878423e-01 -1.60447165e-01 3.85512590e-01
2.04001141e+00 5.07216156e-01 -5.54182351e-01 3.60598236e-01
1.50328124e+00 -1.82990951e-03 -1.21470737e+00 -1.85825199e-01
-1.64716259e-01 2.14768082e-01 -1.72166258e-01 -3.56627077e-01
-5.72262883e-01 7.32244909e-01 1.47539200e-02 3.20231736e-01
4.07321960e-01 4.56184357e-01 9.07500982e-01 7.85866916e-01
2.38902122e-01 -1.29488897e+00 2.87280619e-01 3.83634925e-01
7.19617546e-01 -1.02395761e+00 5.54211974e-01 -4.20499414e-01
-3.24835591e-02 1.47204530e+00 5.30695260e-01 -2.61530191e-01
8.85371208e-01 -1.73706278e-01 -9.98473585e-01 -8.91555071e-01
-9.04306054e-01 2.32980102e-01 3.95425260e-01 5.64724088e-01
7.75949538e-01 3.59857440e-01 4.69953306e-02 3.88758838e-01
-6.48050070e-01 -4.96398091e-01 4.38740045e-01 8.94852400e-01
-6.71977222e-01 -1.49111414e+00 1.46122083e-01 7.43162394e-01
3.20749299e-04 -5.82185388e-01 -5.79749703e-01 8.66398692e-01
1.53813958e-01 5.02815068e-01 4.88725379e-02 -6.80808961e-01
8.39901343e-02 3.40699524e-01 1.21500778e+00 -7.78907955e-01
-8.24673176e-02 -5.71192265e-01 2.99947321e-01 -3.71097893e-01
-5.67857862e-01 -4.36105013e-01 -1.69098747e+00 -8.40894878e-01
-3.03433508e-01 1.17261082e-01 5.09132206e-01 8.92907679e-01
4.10961717e-01 5.30367374e-01 3.25160205e-01 -9.23531353e-01
-2.41321936e-01 -9.69596386e-01 -8.60288739e-01 4.66326356e-01
2.00150684e-01 -6.14914596e-01 -3.18651050e-01 -3.67774308e-01] | [5.200201034545898, 5.546816825866699] |
20d311df-53eb-43ba-ab47-251382a38ad4 | scalable-and-effective-conductance-based | 2211.12511 | null | https://arxiv.org/abs/2211.12511v1 | https://arxiv.org/pdf/2211.12511v1.pdf | Scalable and Effective Conductance-based Graph Clustering | Conductance-based graph clustering has been recognized as a fundamental operator in numerous graph analysis applications. Despite the significant success of conductance-based graph clustering, existing algorithms are either hard to obtain satisfactory clustering qualities, or have high time and space complexity to achieve provable clustering qualities. To overcome these limitations, we devise a powerful \textit{peeling}-based graph clustering framework \textit{PCon}. We show that many existing solutions can be reduced to our framework. Namely, they first define a score function for each vertex, then iteratively remove the vertex with the smallest score. Finally, they output the result with the smallest conductance during the peeling process. Based on our framework, we propose two novel algorithms \textit{PCon\_core} and \emph{PCon\_de} with linear time and space complexity, which can efficiently and effectively identify clusters from massive graphs with more than a few billion edges. Surprisingly, we prove that \emph{PCon\_de} can identify clusters with near-constant approximation ratio, resulting in an important theoretical improvement over the well-known quadratic Cheeger bound. Empirical results on real-life and synthetic datasets show that our algorithms can achieve 5$\sim$42 times speedup with a high clustering accuracy, while using 1.4$\sim$7.8 times less memory than the baseline algorithms. | ['Tao Jia', 'Rong-Hua Li', 'Longlong Lin'] | 2022-11-22 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [ 1.68892205e-01 3.74425645e-03 -8.03539976e-02 1.07813247e-01
-8.43947172e-01 -9.08991039e-01 -1.55117989e-01 5.59679449e-01
-2.27986634e-01 3.84838074e-01 -6.34040475e-01 -5.48724651e-01
-5.73136687e-01 -1.09682143e+00 -5.79507709e-01 -9.96358454e-01
-4.94041920e-01 6.50967777e-01 6.16157293e-01 1.33575127e-01
2.91968316e-01 7.94856027e-02 -1.04626739e+00 -3.52764368e-01
1.14219224e+00 8.51187527e-01 4.77010310e-02 7.50239909e-01
8.03321823e-02 1.97249129e-01 -2.73803860e-01 -3.96752566e-01
4.39338118e-01 -7.02367246e-01 -7.14792132e-01 1.68040290e-01
-1.85965970e-01 3.02014440e-01 -4.78995353e-01 1.33269250e+00
4.27110881e-01 -4.12286222e-02 3.91417772e-01 -1.53421867e+00
-4.25698251e-01 1.01427341e+00 -1.02835798e+00 6.48831874e-02
3.40443432e-01 -1.48113787e-01 1.18248737e+00 -3.00686777e-01
4.18190122e-01 7.91128039e-01 6.20012045e-01 3.42961729e-01
-1.36000240e+00 -9.30089414e-01 1.45834014e-01 9.22459960e-02
-1.94595385e+00 1.73149686e-02 7.06097305e-01 -5.22597842e-02
4.81885970e-01 7.91539550e-01 5.64542770e-01 5.87134585e-02
-3.49206597e-01 5.54451287e-01 1.09707248e+00 -2.36974135e-01
4.20749903e-01 -2.17151746e-01 1.55176103e-01 7.31658995e-01
7.11828828e-01 -4.02807206e-01 -1.02924302e-01 -3.58708173e-01
4.88308698e-01 -3.72563186e-03 -3.66963863e-01 -4.05753195e-01
-1.03871596e+00 8.39735329e-01 6.09215140e-01 1.74313575e-01
-1.47716329e-03 4.92976695e-01 2.71270961e-01 1.93126082e-01
1.23844393e-01 8.55299681e-02 -1.78080246e-01 -6.29289821e-02
-7.35107183e-01 -9.37018916e-02 9.13468838e-01 1.18122506e+00
8.43368828e-01 -4.44513619e-01 3.33315164e-01 4.43714440e-01
9.65346768e-02 5.98741651e-01 -3.16808611e-01 -8.55679691e-01
5.37345529e-01 9.62534368e-01 -7.91837126e-02 -1.48130858e+00
-4.44308192e-01 -1.98085204e-01 -1.35320699e+00 -2.46444762e-01
3.82961839e-01 -7.08473697e-02 -5.44569016e-01 1.68001199e+00
4.93661106e-01 1.97775677e-01 -1.60748675e-01 7.30862796e-01
2.98743814e-01 7.10890293e-01 -2.10475668e-01 -5.85803211e-01
1.18035042e+00 -8.15050840e-01 -2.04122216e-01 4.89593446e-02
8.38586330e-01 -4.81623918e-01 8.42521667e-01 5.16103625e-01
-1.18938231e+00 1.33087672e-02 -9.55461025e-01 3.62353265e-01
1.60643482e-03 -3.93689364e-01 7.93614388e-01 1.07142389e+00
-1.27743340e+00 4.28553969e-01 -9.33222651e-01 -2.68018484e-01
3.15368235e-01 5.07163644e-01 -1.09410696e-01 -1.72550634e-01
-4.95344132e-01 -8.46692920e-02 4.58197206e-01 -1.25172421e-01
-3.93909633e-01 -4.21203375e-01 -6.46860242e-01 2.03021795e-01
7.74301767e-01 -4.90115225e-01 5.39939582e-01 -4.88494992e-01
-9.02668595e-01 5.85147858e-01 -1.88705146e-01 -3.89549732e-01
2.35912025e-01 4.42763239e-01 -2.34183311e-01 5.14474452e-01
5.12522645e-02 1.52472496e-01 4.76249643e-02 -1.23003483e+00
-5.88567019e-01 -4.93603915e-01 -9.14836526e-02 6.59962147e-02
-7.04930186e-01 -4.00417075e-02 -9.78219450e-01 -2.47540414e-01
6.16401970e-01 -1.13971388e+00 -7.20896900e-01 -2.76625812e-01
-6.60205901e-01 -4.01654005e-01 4.30940121e-01 -2.15295404e-02
1.79129720e+00 -2.05504680e+00 1.68782115e-01 8.99319470e-01
9.18012679e-01 1.91679254e-01 5.41651733e-02 6.77273631e-01
4.15064394e-01 4.45851386e-01 -3.45090121e-01 1.75199673e-01
-2.93235071e-02 6.68773204e-02 2.51739383e-01 6.84473336e-01
-5.86487770e-01 7.66084611e-01 -9.86611009e-01 -4.18649465e-01
1.57230034e-01 6.88253343e-02 -8.10106814e-01 -1.90749303e-01
-1.94209274e-02 1.54521447e-04 -5.54160595e-01 3.89690220e-01
1.10043836e+00 -8.38516116e-01 8.47149968e-01 1.61541373e-01
2.85052955e-01 -1.08359613e-01 -1.47481847e+00 1.24432278e+00
5.81682136e-04 2.33353853e-01 4.74577457e-01 -1.21504176e+00
8.64016831e-01 3.90906557e-02 7.57388055e-01 -4.97876704e-01
2.79413819e-01 3.14664960e-01 -4.45412509e-02 -6.72034100e-02
1.91207692e-01 1.47135302e-01 -3.92513245e-01 7.87962437e-01
-6.49224937e-01 2.68823504e-01 3.75837028e-01 8.23158681e-01
1.66753471e+00 -7.44611979e-01 -3.08737252e-02 -3.11294705e-01
5.17955363e-01 -5.14844246e-02 6.18812263e-01 6.22186840e-01
-2.40497366e-01 4.58125770e-01 8.39911520e-01 -1.05675779e-01
-8.97859335e-01 -9.50384796e-01 3.90423506e-01 8.39164317e-01
7.34169602e-01 -9.22676921e-01 -1.01468909e+00 -4.44448709e-01
-1.41951263e-01 -8.43617469e-02 -4.46246505e-01 -1.04491003e-01
-4.11795825e-01 -9.55912173e-01 4.97326910e-01 4.78198767e-01
3.28876346e-01 -5.34834325e-01 -3.85393500e-01 3.32387894e-01
-1.04294971e-01 -1.13650107e+00 -6.20016932e-01 -4.86440100e-02
-7.36191213e-01 -1.32180977e+00 -2.23495111e-01 -9.10314679e-01
1.00407004e+00 6.51545346e-01 8.24083209e-01 7.82016575e-01
-2.33741924e-01 5.35207018e-02 -5.03947556e-01 2.23380342e-01
-1.40471861e-01 3.02296668e-01 -1.29630342e-01 -8.51712301e-02
5.38795173e-01 -8.64247561e-01 -1.02265763e+00 3.67509276e-01
-8.66270661e-01 5.14445733e-03 4.07135427e-01 4.14108783e-01
7.71687746e-01 5.78798771e-01 4.91063982e-01 -1.10586333e+00
5.86649358e-01 -5.03719211e-01 -8.79269719e-01 7.49783516e-02
-9.41996932e-01 -1.69273823e-01 8.29537690e-01 -1.62078932e-01
-2.10914448e-01 2.44419724e-01 -9.10782441e-02 -2.85750031e-01
2.33247101e-01 4.07617867e-01 -3.21780235e-01 -2.31310904e-01
2.71684974e-01 2.91147143e-01 -1.22034706e-01 -1.15434468e-01
4.11318719e-01 5.22239149e-01 5.11218786e-01 -5.06957233e-01
1.09809172e+00 7.34913826e-01 1.83775112e-01 -3.90948623e-01
-4.53925431e-01 -7.35418200e-01 -4.28295881e-01 -6.01733178e-02
4.88289416e-01 -6.28396988e-01 -1.48379922e+00 2.58195370e-01
-4.88638699e-01 -4.09173638e-01 3.70719373e-01 1.18034877e-01
-1.79405048e-01 8.58711898e-01 -7.01933324e-01 -1.01136947e+00
-5.63462079e-01 -8.08629930e-01 7.09793508e-01 1.66692555e-01
1.17936291e-01 -9.36935544e-01 -2.14578837e-01 3.07071239e-01
1.42795995e-01 6.47081077e-01 1.00243366e+00 -5.12112975e-01
-8.21165264e-01 -1.31229594e-01 -5.25570691e-01 -1.49795592e-01
-2.45412029e-02 -9.54493061e-02 -1.78290367e-01 -7.88421452e-01
-5.09373128e-01 8.66133440e-03 7.93517292e-01 -2.24098243e-04
1.47070956e+00 -4.66874629e-01 -7.95970082e-01 6.68954253e-01
1.81254935e+00 2.71996021e-01 6.32340193e-01 2.68619712e-02
7.39277363e-01 1.36911660e-01 4.15799618e-01 5.37098765e-01
6.25810146e-01 3.01499635e-01 4.62024540e-01 5.45478053e-02
2.95694441e-01 -2.76666194e-01 3.83801684e-02 1.35094273e+00
-1.69106890e-02 -5.27151108e-01 -8.48227859e-01 7.12415695e-01
-1.87247825e+00 -6.31072998e-01 -6.55471504e-01 2.40340352e+00
6.55929863e-01 3.34179163e-01 4.24967229e-01 6.83420479e-01
9.55841720e-01 -2.04105318e-01 -5.77435136e-01 -2.09903792e-01
9.33743715e-02 2.98355967e-01 7.65109837e-01 5.20695627e-01
-7.61597753e-01 9.30012584e-01 5.41167116e+00 1.09938192e+00
-5.70964634e-01 -2.74548400e-02 5.91801345e-01 6.36089081e-03
-4.60670799e-01 3.06859344e-01 -3.15293133e-01 6.33748114e-01
8.91379476e-01 -7.49987245e-01 5.98915040e-01 7.18794346e-01
-5.16728312e-02 -2.26558611e-01 -7.79805183e-01 1.09128332e+00
-9.61572379e-02 -1.19021797e+00 -1.01663671e-01 5.06075919e-01
9.09015000e-01 -3.49063635e-01 -1.25144854e-01 -1.06347576e-01
6.32535219e-01 -1.02189708e+00 6.98394105e-02 -2.01013446e-01
1.00069404e+00 -1.19003987e+00 3.13496143e-01 4.58675474e-01
-1.84831226e+00 -1.32504568e-01 -4.96677369e-01 -4.62043807e-02
3.47745530e-02 8.78743410e-01 -4.90160257e-01 8.73548269e-01
7.62183309e-01 2.38108695e-01 -3.66936207e-01 1.16370332e+00
3.15885954e-02 8.29678535e-01 -6.55447185e-01 -4.58021075e-01
1.87455222e-01 -6.24008060e-01 4.75572199e-01 9.73653615e-01
3.93247306e-01 7.09010005e-01 4.85203803e-01 5.48410356e-01
-4.04710919e-01 4.20009106e-01 -2.03389630e-01 -4.32787687e-02
1.14985371e+00 1.47080553e+00 -1.54663754e+00 -3.32134724e-01
-1.15076266e-01 1.09679985e+00 4.24279511e-01 4.45720404e-02
-1.01757836e+00 -1.03963709e+00 4.23089385e-01 2.63603598e-01
4.12936002e-01 -3.50218415e-01 -3.54295135e-01 -7.18785942e-01
2.44835198e-01 -7.40714133e-01 7.27940798e-01 -3.30456555e-01
-1.04686892e+00 4.88309979e-01 -4.28566098e-01 -8.06633055e-01
5.50687253e-01 -2.06926510e-01 -6.95097685e-01 3.67639810e-01
-9.79657650e-01 -5.99704504e-01 -5.49520850e-01 8.50686133e-01
-1.96095109e-01 4.61302191e-01 6.13042533e-01 4.60853964e-01
-4.43458885e-01 8.26796293e-01 3.83712828e-01 1.56064689e-01
2.31526777e-01 -1.37644100e+00 2.62472987e-01 9.16586995e-01
-1.93937384e-02 4.93353665e-01 4.15139645e-01 -4.76175934e-01
-1.86561275e+00 -1.06833243e+00 5.87908983e-01 -1.71689913e-01
6.12165868e-01 -6.53240144e-01 -7.89614320e-01 3.67751181e-01
-4.56356294e-02 -9.91377011e-02 6.72153473e-01 -4.55602929e-02
-3.78298759e-01 -3.07948023e-01 -1.16642165e+00 6.46490991e-01
1.44758797e+00 -2.12501451e-01 2.14148968e-01 3.97661686e-01
8.40273142e-01 -2.27589086e-01 -1.00780833e+00 2.80178666e-01
8.94326419e-02 -1.02908611e+00 8.27949703e-01 -1.86249074e-02
-1.13466479e-01 -5.43050468e-01 2.51960099e-01 -8.90922904e-01
-4.16920125e-01 -1.31694818e+00 -2.21270975e-02 1.19178367e+00
4.67677772e-01 -8.10171008e-01 1.04500043e+00 3.28362375e-01
3.25144023e-01 -8.55621040e-01 -9.42876935e-01 -8.68689239e-01
2.91124642e-01 -3.02427858e-01 6.02776289e-01 8.69094372e-01
4.26091999e-01 4.25139368e-01 -1.34696096e-01 4.12764728e-01
1.06490767e+00 4.26990092e-01 9.51629758e-01 -1.42104828e+00
-1.36289224e-01 -6.45276129e-01 -4.59417254e-01 -1.26798010e+00
-2.13887259e-01 -1.14156532e+00 -1.09133348e-01 -1.78959703e+00
7.93356180e-01 -1.10616875e+00 -2.51475722e-01 4.16985810e-01
-3.49143028e-01 5.77334881e-01 5.40005751e-02 3.30799222e-02
-1.19720066e+00 2.99057178e-02 1.09392905e+00 2.07732961e-01
-2.74426490e-01 -6.97529092e-02 -1.01542699e+00 4.86387372e-01
9.31944489e-01 -4.49047834e-01 -4.87182498e-01 -1.23066016e-01
6.55415356e-01 2.21768498e-01 7.59108514e-02 -1.02744257e+00
5.91501236e-01 -1.48786023e-01 -9.59308669e-02 -7.33183324e-01
-6.20900691e-02 -7.12402642e-01 3.22763443e-01 7.70078540e-01
-6.62092790e-02 1.35333747e-01 -1.13119133e-01 9.02698874e-01
1.66425258e-01 2.92639136e-01 7.19169497e-01 1.63756102e-01
-2.07000479e-01 6.73269868e-01 -2.03548670e-01 2.46061280e-01
1.59545231e+00 -3.56969386e-01 -3.98239017e-01 -4.34258103e-01
-5.07146835e-01 8.00979793e-01 6.44659877e-01 -2.68313829e-02
5.33419669e-01 -1.15068030e+00 -6.85386717e-01 -1.16343722e-01
1.32286465e-02 8.85401666e-02 2.88231820e-01 1.09935868e+00
-5.89823961e-01 2.55457938e-01 3.70033711e-01 -5.70016742e-01
-1.55181646e+00 1.01644182e+00 -3.69196199e-02 -2.52391458e-01
-6.97717249e-01 9.06325996e-01 9.78761241e-02 -1.42508373e-01
8.27457532e-02 5.15353680e-02 5.37756264e-01 -4.50450659e-01
2.66063899e-01 7.47413039e-01 -2.55950004e-01 -5.12168705e-01
-5.10318816e-01 6.22115970e-01 2.31354628e-02 5.21629937e-02
1.23258054e+00 -5.35976112e-01 -4.59717989e-01 -2.35799968e-01
1.17371142e+00 2.57175475e-01 -8.00979555e-01 -1.55603260e-01
1.76074788e-01 -5.66509902e-01 -3.30934256e-01 -2.99540460e-01
-1.32505202e+00 6.76823616e-01 1.62529588e-01 8.15382838e-01
1.43266261e+00 2.41994485e-01 1.23574805e+00 2.76091874e-01
7.54731119e-01 -1.10848939e+00 -3.55726220e-02 -3.03664822e-02
1.49334352e-02 -7.37975955e-01 -3.49877812e-02 -1.03304231e+00
-3.16472709e-01 7.33392835e-01 5.62922060e-01 -1.81075200e-01
7.33281314e-01 2.90571183e-01 -4.79316235e-01 -3.51820529e-01
-7.22529948e-01 -2.87088066e-01 -3.00582796e-01 4.11745459e-01
-4.63986695e-02 4.34836328e-01 -6.64215088e-01 6.23185396e-01
-1.65864334e-01 -3.38391572e-01 5.51579416e-01 6.26251101e-01
-7.71989942e-01 -1.23693550e+00 -1.01252079e-01 5.09446502e-01
-4.72283542e-01 -1.22400470e-01 -6.20157361e-01 5.68020165e-01
-2.88192630e-01 1.27782357e+00 -1.11367315e-01 -5.89461505e-01
-7.14795738e-02 -4.41669375e-01 3.37949663e-01 -3.14061761e-01
-4.45102513e-01 2.71827698e-01 -2.16429934e-01 -5.37614226e-01
-1.64523110e-01 -3.37024063e-01 -1.73375940e+00 -9.18169618e-01
-5.58339357e-01 6.73271239e-01 1.91860840e-01 5.10858655e-01
5.79555094e-01 1.48626581e-01 1.00853491e+00 -7.02170208e-02
-1.81609303e-01 -4.38538909e-01 -8.00304890e-01 3.12307566e-01
-2.46560171e-01 -7.56329373e-02 -6.69054270e-01 -3.70508105e-01] | [6.892030239105225, 5.1037278175354] |
97aa8592-cb6f-4487-9376-34906ec32e5b | wut-at-semeval-2019-task-9-domain-adversarial | null | null | https://aclanthology.org/S19-2221 | https://aclanthology.org/S19-2221.pdf | WUT at SemEval-2019 Task 9: Domain-Adversarial Neural Networks for Domain Adaptation in Suggestion Mining | We present a system for cross-domain suggestion mining, prepared for the SemEval-2019 Task 9: Suggestion Mining from Online Reviews and Forums (Subtask B). Our submitted solution for this text classification problem explores the idea of treating different suggestions{'} sources as one of the settings of Transfer Learning - Domain Adaptation. Our experiments show that without any labeled target domain examples during training time, we are capable of proposing a system, reaching up to 0.778 in terms of F1 score on test dataset, based on Target Preserving Domain Adversarial Neural Networks. | ['Piotr Andruszkiewicz', 'Mateusz Klimaszewski'] | 2019-06-01 | null | null | null | semeval-2019-6 | ['suggestion-mining'] | ['natural-language-processing'] | [ 2.15616688e-01 5.67079425e-01 -5.57590187e-01 -5.14211357e-01
-9.81698275e-01 -7.56189167e-01 1.13330710e+00 2.40591288e-01
-5.26585340e-01 9.75945055e-01 3.10978740e-02 -8.47178698e-01
-8.90835673e-02 -5.42761326e-01 -6.96381509e-01 -2.21166849e-01
-1.64193418e-02 9.37479913e-01 3.44701171e-01 -7.21859336e-01
6.00094736e-01 -2.63224244e-02 -1.01860535e+00 7.31722414e-01
7.81095088e-01 9.84319866e-01 -4.19083685e-01 4.85123336e-01
-3.14682513e-01 3.44063759e-01 -1.06700349e+00 -8.54179859e-01
4.62345988e-01 -1.28057286e-01 -1.08696985e+00 -1.64507881e-01
5.95788777e-01 -2.36856863e-01 -2.16213055e-02 1.04838538e+00
2.92943239e-01 3.65433156e-01 1.40173519e+00 -1.12489676e+00
-9.06513095e-01 1.15989900e+00 -3.81264746e-01 2.53211260e-01
2.15818673e-01 -3.51409853e-01 1.02858889e+00 -1.17719889e+00
6.56716108e-01 1.22725368e+00 3.60015869e-01 1.00373816e+00
-1.45378959e+00 -9.00011837e-01 1.02423392e-01 -4.64277975e-02
-8.28644037e-01 -3.97068620e-01 7.35237896e-01 -3.01515311e-01
9.61286604e-01 1.97672620e-01 -1.74693778e-01 1.56936598e+00
1.66227832e-01 8.69453251e-01 1.16827106e+00 -6.71367347e-01
5.54767787e-01 9.83698606e-01 1.39135435e-01 1.30643368e-01
1.57559216e-01 1.10815652e-03 -6.53004885e-01 -7.66897261e-01
1.38577431e-01 -5.39071620e-01 1.24685302e-01 -3.71587127e-01
-6.63452029e-01 1.25138974e+00 7.80286938e-02 6.99253529e-02
-6.17844518e-03 -5.70950866e-01 9.71415579e-01 9.96373415e-01
1.33672452e+00 7.33475626e-01 -1.22034740e+00 4.48682040e-01
-5.75761318e-01 5.31767845e-01 1.18400347e+00 1.04469955e+00
4.88913387e-01 -2.68782210e-02 1.71432532e-02 1.29167068e+00
-1.70439836e-02 1.82588339e-01 1.15579116e+00 -6.00599408e-01
8.58512163e-01 2.70074040e-01 3.40430826e-01 -4.41462219e-01
-1.57298639e-01 -1.22064032e-01 -6.76124215e-01 1.80576682e-01
3.48269820e-01 -7.03149974e-01 -4.84658271e-01 1.48667538e+00
1.97014332e-01 -9.27112326e-02 3.77302736e-01 3.43635529e-01
9.17803109e-01 4.42422003e-01 2.22057134e-01 -2.57282674e-01
1.22216535e+00 -1.14116371e+00 -3.11075568e-01 -3.86882126e-01
8.78548324e-01 -5.64188480e-01 1.21149743e+00 8.27296495e-01
-8.89240205e-01 -7.32091308e-01 -1.17606902e+00 3.08678955e-01
-6.92789912e-01 2.41095990e-01 4.23608065e-01 7.35758960e-01
-8.73258412e-01 5.33891201e-01 -9.43339542e-02 -3.99738729e-01
2.89316803e-01 5.34602582e-01 -5.09444535e-01 1.40572309e-01
-1.38546646e+00 8.49974394e-01 3.16986382e-01 -9.42507684e-01
-4.44616586e-01 -8.39484096e-01 -3.72019917e-01 1.01847127e-02
1.96434692e-01 -4.04331297e-01 1.64168739e+00 -1.41742969e+00
-1.65997088e+00 9.94500041e-01 2.84323007e-01 -9.13081169e-01
5.60883701e-01 -4.87225831e-01 -9.52768028e-01 -1.77948728e-01
2.17377350e-01 6.88518286e-01 1.35060263e+00 -1.14078581e+00
-6.79544389e-01 -2.32619986e-01 -2.16835275e-01 1.50225401e-01
-7.95012653e-01 3.63336325e-01 1.19905919e-01 -8.81893933e-01
-6.31150961e-01 -9.51819539e-01 -3.68122250e-01 -5.12312770e-01
-5.25754750e-01 -8.01775753e-01 7.15690613e-01 -4.95536029e-01
7.38921881e-01 -2.08020043e+00 -4.37963605e-01 2.95550883e-01
-5.07158227e-02 5.41688144e-01 -2.95651019e-01 1.70756042e-01
-3.10096502e-01 3.95651847e-01 -2.92130485e-02 -3.69816780e-01
8.94547999e-02 -2.88595915e-01 -1.07297361e+00 2.55657315e-01
2.45394200e-01 4.81691957e-01 -6.84141338e-01 -2.22871512e-01
-1.04474388e-01 -3.41030359e-01 -4.31293517e-01 2.82317698e-01
-4.98583615e-01 3.91896106e-02 -5.66995263e-01 -2.34885584e-03
7.59926856e-01 -9.00245011e-02 2.40748391e-01 4.29439545e-01
4.58777100e-01 9.13134217e-01 -9.63235676e-01 1.53054357e+00
-2.86002815e-01 4.63343382e-01 1.48335109e-02 -1.29679286e+00
1.47084415e+00 2.55974174e-01 4.70633298e-01 -7.28119671e-01
-5.11905141e-02 1.36135072e-01 -8.55830535e-02 1.26665145e-01
7.16779351e-01 -1.77945971e-01 -4.09187436e-01 8.32519352e-01
3.43115360e-01 -8.23372379e-02 -2.85562407e-02 4.99355406e-01
1.28037977e+00 -2.78764874e-01 5.53425074e-01 -4.45092350e-01
6.76169276e-01 2.75663674e-01 -1.15358289e-02 9.13668215e-01
-1.43024415e-01 2.57765353e-01 3.19280237e-01 -3.63895386e-01
-9.14346874e-01 -1.13842070e+00 -1.47947714e-01 1.70240092e+00
-2.22169653e-01 -3.79784584e-01 -5.87697923e-01 -1.68203247e+00
4.00305092e-01 1.01555049e+00 -7.38942742e-01 -2.80321777e-01
-4.16574508e-01 -4.62635815e-01 6.43307090e-01 2.29341224e-01
5.29317819e-02 -1.18985653e+00 2.17423728e-03 2.25050151e-01
3.62782657e-01 -8.39721799e-01 -4.93857145e-01 6.39526129e-01
-1.06865883e+00 -9.25491869e-01 -6.36687517e-01 -8.44721794e-01
4.03968424e-01 2.33034417e-01 1.34513080e+00 -4.31773782e-01
1.01518944e-01 4.28473413e-01 -4.58971769e-01 -7.65732944e-01
-8.93899918e-01 3.12353402e-01 3.80685061e-01 -5.23894310e-01
8.64034176e-01 -4.52909857e-01 -2.18223527e-01 4.29247886e-01
-6.54434204e-01 -3.67546409e-01 6.17808640e-01 1.19434011e+00
2.26913869e-01 -1.65346548e-01 1.47406292e+00 -1.84223950e+00
1.52780592e+00 -8.30786169e-01 -4.32796776e-01 2.26976648e-01
-1.26969683e+00 -4.65799794e-02 1.09034717e+00 -7.07819164e-01
-1.36316478e+00 2.00235490e-02 -1.71241444e-03 -4.86615784e-02
-6.90276802e-01 4.31883186e-02 1.06718108e-01 1.03058197e-01
1.66029656e+00 1.73287332e-01 9.25062001e-02 -6.01587176e-01
7.92913735e-01 9.39008117e-01 5.20830750e-01 -6.60977066e-01
6.25851333e-01 6.67472556e-02 -6.29004955e-01 -3.31092715e-01
-7.65037060e-01 -7.65358210e-01 -5.84024310e-01 4.15937334e-01
9.85126644e-02 -8.07705402e-01 -2.36202613e-01 -9.25393477e-02
-1.06368995e+00 -2.36567974e-01 -3.65123510e-01 4.80158664e-02
-7.41305172e-01 2.54629165e-01 -5.13284981e-01 -4.38661397e-01
-9.54750359e-01 -4.67140675e-01 5.35155535e-01 -8.05250034e-02
-5.03525794e-01 -1.30587077e+00 3.72822374e-01 2.56721437e-01
3.61855865e-01 -5.32707930e-01 1.05478501e+00 -2.13107514e+00
5.94402611e-01 -2.19215870e-01 -6.92613646e-02 7.10257411e-01
1.23351492e-01 -5.59643805e-01 -1.05308354e+00 -6.96940869e-02
1.58044234e-01 -8.03672135e-01 9.00051355e-01 9.03550833e-02
1.55755889e+00 -8.55264843e-01 -3.76024008e-01 -6.52975962e-02
9.48478639e-01 -1.58672128e-02 2.77256101e-01 5.25282621e-01
-1.03667250e-03 9.42790866e-01 1.13584375e+00 5.48444986e-01
-9.65556502e-02 5.70425510e-01 5.09028994e-02 3.14627029e-02
1.61485821e-01 -3.70208234e-01 6.18696094e-01 1.78200275e-01
5.60392022e-01 -4.66816187e-01 -6.76002026e-01 5.09442985e-01
-1.62864625e+00 -8.70100915e-01 7.33968168e-02 1.97006249e+00
1.15269709e+00 5.10306120e-01 4.68672425e-01 -1.55006364e-01
6.56855226e-01 -2.09056437e-01 -7.32768595e-01 -1.15456951e+00
1.12777837e-01 6.42814696e-01 6.66864038e-01 1.75893739e-01
-1.49858725e+00 1.03412783e+00 6.93322086e+00 1.15681446e+00
-9.33230877e-01 2.28078604e-01 9.14246917e-01 2.76803732e-01
-4.13556248e-01 -3.94290596e-01 -9.15801525e-01 4.83226746e-01
1.35561848e+00 -4.34727639e-01 -3.83429579e-03 1.40829611e+00
-3.24793577e-01 1.11102313e-01 -1.04822338e+00 2.65291780e-01
1.89984754e-01 -1.42981744e+00 5.30363917e-02 -5.29444627e-02
8.99399102e-01 3.14822674e-01 4.42922324e-01 7.95274734e-01
9.03752029e-01 -9.60510254e-01 1.75252363e-01 -3.54303986e-01
8.13639820e-01 -7.20878303e-01 4.96468723e-01 6.47827804e-01
-1.36517957e-01 -4.05469052e-02 -7.51368701e-01 1.81620970e-01
-3.33482772e-01 5.24453342e-01 -1.86311316e+00 3.70839983e-01
5.79035461e-01 7.99766839e-01 -3.98529559e-01 6.02025628e-01
-2.76796907e-01 1.17543137e+00 -1.09064206e-01 -3.03221881e-01
-8.10093433e-02 1.65994745e-02 3.59844059e-01 1.53316569e+00
8.05973187e-02 -2.53797978e-01 1.84213877e-01 4.60564762e-01
-5.20224631e-01 5.02768338e-01 -9.11530972e-01 2.90674299e-01
4.11799133e-01 1.18777621e+00 -2.39126727e-01 -5.00542402e-01
-3.35031450e-01 9.45595741e-01 4.31012124e-01 2.61717707e-01
-5.08176208e-01 -2.20986947e-01 4.82200682e-01 2.81868502e-02
2.18901396e-01 2.89774716e-01 -7.64235616e-01 -9.05863404e-01
-2.08066374e-01 -1.24089146e+00 7.24173784e-01 -1.84219941e-01
-2.11429644e+00 4.61717755e-01 1.58454403e-02 -1.42938387e+00
-4.18967664e-01 -1.03506422e+00 -8.62805009e-01 9.51786995e-01
-1.34249365e+00 -1.00580585e+00 6.44251227e-01 8.02842259e-01
9.23387527e-01 -1.01362240e+00 1.20224905e+00 2.34247255e-03
-1.28524229e-02 1.12235081e+00 5.96482515e-01 -2.33849019e-01
1.63382030e+00 -1.65380824e+00 7.72342086e-01 3.05142641e-01
4.47966844e-01 6.35415316e-01 7.32578874e-01 -5.65468311e-01
-6.88282192e-01 -1.12990916e+00 8.11293662e-01 -6.94471657e-01
9.33333695e-01 -4.70397800e-01 -9.56971288e-01 5.55989802e-01
3.40388149e-01 -2.06654176e-01 1.10911262e+00 5.82647800e-01
-6.95573568e-01 -6.54771850e-02 -1.45007133e+00 3.51638973e-01
6.69688523e-01 -3.78538936e-01 -1.01571941e+00 8.11589062e-01
9.15024936e-01 -2.12891966e-01 -7.97354162e-01 1.52375147e-01
7.55104870e-02 -5.29337406e-01 1.14336979e+00 -1.19518769e+00
8.06392550e-01 3.66527706e-01 3.12349617e-01 -1.54449129e+00
-2.31162552e-02 -5.20689189e-01 -2.02266499e-01 1.25571144e+00
8.58197629e-01 -6.12761378e-01 1.04944444e+00 6.17033482e-01
-6.65430203e-02 -3.08960766e-01 -1.10402215e+00 -7.24100232e-01
8.32792282e-01 -4.47717696e-01 2.49330461e-01 1.05729520e+00
4.67503756e-01 8.29449296e-01 -3.78665566e-01 -4.87637699e-01
2.21901178e-01 -2.41387859e-02 1.21212411e+00 -1.35846984e+00
-4.86086816e-01 -3.66439342e-01 3.08326751e-01 -1.20377815e+00
8.02888274e-01 -9.51135576e-01 5.38110174e-02 -9.26074088e-01
-1.71934053e-01 -7.32654393e-01 -6.45530403e-01 5.33248365e-01
8.56269896e-03 -9.74317119e-02 -2.65540957e-01 1.97597429e-01
-6.19644821e-01 4.38110113e-01 7.10343421e-01 -5.21389127e-01
-1.41378284e-01 7.63212621e-01 -1.15245092e+00 5.43055236e-01
1.07064402e+00 -7.94663012e-01 -6.49169564e-01 1.58247292e-01
-1.43421590e-01 -1.01509638e-01 5.40570170e-02 -5.65015972e-01
1.68853179e-01 -2.49643266e-01 2.38087505e-01 -4.37581152e-01
1.97129905e-01 -5.92770576e-01 -6.98767304e-01 3.72204304e-01
-1.09544158e+00 -1.60630479e-01 5.11387527e-01 9.54826474e-01
-1.01212010e-01 -4.23191369e-01 5.14664769e-01 6.16928004e-02
-6.94869876e-01 -2.94741038e-02 -5.53204954e-01 1.08464092e-01
6.89756036e-01 2.89165080e-01 -8.68742764e-01 -3.18736076e-01
-4.64838207e-01 -5.93130179e-02 1.86120853e-01 7.38382518e-01
5.80231488e-01 -8.10876310e-01 -9.18194592e-01 1.12781907e-02
4.86883342e-01 -4.00443196e-01 -2.06508368e-01 1.42000079e-01
2.58283019e-01 6.57685280e-01 -2.50006169e-01 -5.99237606e-02
-1.18165326e+00 7.49919176e-01 1.39655188e-01 -7.36823082e-01
-4.02413964e-01 1.15128005e+00 2.58563280e-01 -9.85730231e-01
3.18873338e-02 2.11375430e-01 -4.74210918e-01 -1.97600096e-01
5.57192743e-01 2.91891813e-01 4.44519222e-01 1.79634422e-01
-1.64952308e-01 -4.34582680e-01 -8.02901328e-01 -4.46467310e-01
1.20552206e+00 1.83991194e-01 2.96583354e-01 7.91717023e-02
9.97527778e-01 2.69950837e-01 -7.78790236e-01 -5.56624651e-01
4.76469725e-01 -1.88078806e-01 -3.39175314e-01 -1.34958661e+00
-3.40033770e-01 7.59444416e-01 5.83392859e-01 4.40412015e-01
9.15044725e-01 7.18781278e-02 4.72460955e-01 9.34737206e-01
1.83326140e-01 -1.38662684e+00 3.79346222e-01 7.11503863e-01
1.07301283e+00 -1.57962322e+00 5.42146005e-02 -1.85051799e-01
-9.70282078e-01 1.24844718e+00 1.01133788e+00 -4.44046885e-01
8.20552468e-01 3.30610313e-02 -5.32439090e-02 1.12079047e-01
-1.05599618e+00 4.95104849e-01 6.10069335e-01 1.05874670e+00
4.73384380e-01 2.05198899e-01 -5.50376058e-01 1.14575970e+00
-1.38715446e-01 -4.70153511e-01 4.85645622e-01 4.77632076e-01
-5.81054449e-01 -1.64473188e+00 3.15450877e-02 8.34196687e-01
-6.25915408e-01 -2.14616641e-01 -9.71615314e-01 6.15892053e-01
-1.30830184e-01 9.70461547e-01 -1.77618414e-01 -5.78464746e-01
1.76820904e-01 2.25262880e-01 -1.54212758e-01 -1.17692959e+00
-1.15781426e+00 -4.40155864e-01 6.47306025e-01 5.72769344e-02
-1.90405264e-01 -7.11842895e-01 -1.03448427e+00 -1.14115067e-01
-5.11664808e-01 6.91224813e-01 7.98454404e-01 9.29059565e-01
6.21491611e-01 1.73801463e-03 8.33428562e-01 -2.92399615e-01
-1.27578461e+00 -1.40519345e+00 -5.75647354e-01 5.80639899e-01
-3.52608338e-02 -3.87138426e-01 -2.93706805e-01 1.45213529e-01] | [10.881022453308105, 7.590798377990723] |
fffecc9b-0225-4cec-958f-d0830442a852 | detecting-bot-behaviour-in-social-media-using | null | null | https://aran.library.nuigalway.ie/handle/10379/15683 | http://aics2019.datascienceinstitute.ie/papers/aics_35.pdf | Detecting Bot Behaviour in Social Media using Digital DNA Compression | A major challenge faced by online social networks such as Facebook and Twitter is the remarkable rise of fake and automated bot accounts over the last few years. Some of these accounts have been reported to engage in undesirable activities such as spamming, political campaigning and spreading falsehood on the platform. We present an approach to detect bot-like behaviour among Twitter accounts by analyzing their past tweeting activity. We build upon an existing technique of analysis of Twitter accounts called Digital DNA. Digital DNA models the behaviour of Twitter accounts by encoding the post history of a user account as a sequence of characters analogous to an actual DNA sequence. In our approach, we employ a lossless compression algorithm on these Digital DNA sequences and use the compression statistics as a measure of predictability in the behaviour of a group of Twitter accounts. We leverage the information conveyed by the compression statistics to visually represent the posting behaviour by a simple two dimensional scatter plot and categorize the user accounts as bots and genuine users by using an off-the-shelf implementation of the logistic regression classification algorithm. | ['Conor Hayes', 'Nivranshu Pasricha'] | 2019-12-05 | null | null | null | 27th-irish-conference-on-artificial | ['twitter-bot-detection'] | ['miscellaneous'] | [ 5.30775249e-01 3.01398009e-01 -2.34612629e-01 -1.02126062e-01
-1.73550740e-01 -9.32959497e-01 1.11022818e+00 6.03843749e-01
-4.83892888e-01 6.13424242e-01 4.59543973e-01 -6.25672877e-01
2.50759870e-01 -9.94136453e-01 -3.57529163e-01 -5.44070303e-01
-8.70682448e-02 3.73437345e-01 3.62072319e-01 -1.22249141e-01
7.87652433e-01 3.95801604e-01 -1.37988257e+00 5.75835645e-01
3.00694644e-01 6.71337247e-01 -3.02653760e-01 1.04377592e+00
-3.12362462e-01 9.43409681e-01 -7.90272832e-01 -8.38628292e-01
4.54928465e-02 -5.73225796e-01 -6.62156641e-01 3.12914327e-02
-1.22706436e-01 -2.98338979e-01 -6.02240622e-01 1.29254031e+00
2.29039788e-02 -4.48680758e-01 7.03823030e-01 -1.10906756e+00
-1.87054202e-01 6.56191945e-01 -5.69993854e-01 4.12920028e-01
6.90745473e-01 1.68209553e-01 6.28435731e-01 -3.54643047e-01
9.12640631e-01 1.49327242e+00 1.02004242e+00 -8.98613483e-02
-1.46052778e+00 -5.70115209e-01 -7.26168036e-01 -4.41812500e-02
-1.15278208e+00 -2.14604214e-01 5.33493400e-01 -9.11803126e-01
4.36953127e-01 5.46967030e-01 7.95771718e-01 1.05200529e+00
3.44391555e-01 4.99114037e-01 1.15881836e+00 -2.34279215e-01
2.02260718e-01 3.36007923e-01 7.89676011e-02 8.18505824e-01
4.98970628e-01 -2.67442405e-01 -3.77453357e-01 -9.17921782e-01
2.31069624e-02 1.45059064e-01 2.56503046e-01 2.30965585e-01
-9.83164608e-01 1.36965287e+00 1.82519764e-01 3.44913632e-01
-2.79569268e-01 1.80643022e-01 7.86536515e-01 4.43478554e-01
8.41956913e-01 3.04388344e-01 1.64127514e-01 -3.39471459e-01
-1.02792943e+00 5.07110238e-01 9.53008831e-01 1.47785231e-01
9.62478280e-01 -4.64098603e-01 2.58334517e-01 4.11867946e-01
2.83925772e-01 2.64561981e-01 9.75238800e-01 -5.95312774e-01
2.68344104e-01 8.28698099e-01 5.45195565e-02 -1.79314923e+00
-1.43174931e-01 -3.08561008e-02 -5.76838434e-01 -2.43022665e-01
7.39868581e-01 2.59988457e-02 -3.48958105e-01 1.10503733e+00
2.94569552e-01 1.01453125e-01 -4.35378790e-01 7.65905231e-02
2.07395628e-01 6.91971481e-01 -1.04745969e-01 -2.83544779e-01
1.33885312e+00 -2.95476198e-01 -6.74734890e-01 9.64933783e-02
8.43732119e-01 -6.40289724e-01 7.72793472e-01 5.71856424e-02
-9.15755272e-01 1.36085212e-01 -9.41577911e-01 2.65236497e-01
-8.35879922e-01 -5.74160516e-01 2.64507800e-01 1.13229442e+00
-5.35867989e-01 9.41962183e-01 -6.83018208e-01 -5.37919521e-01
7.27592707e-01 -3.53142098e-02 -6.90134540e-02 1.90992013e-01
-9.01652038e-01 5.66418171e-01 3.36677313e-01 -6.05023503e-01
-3.34309399e-01 -2.50760943e-01 -4.89196360e-01 -3.30889612e-01
1.10464193e-01 1.18612193e-01 1.15176392e+00 -7.31488585e-01
-1.00516450e+00 1.21017540e+00 -3.75643633e-02 -6.83244646e-01
8.71539474e-01 3.61599356e-01 -4.07038301e-01 2.36743391e-01
5.39445020e-02 -1.32720411e-01 1.07205319e+00 -8.63497138e-01
-4.48860377e-01 -4.85370308e-01 -5.90663373e-01 -5.57095110e-01
-4.57864165e-01 3.03553253e-01 3.78845155e-01 -6.50326073e-01
1.77257955e-01 -1.09550142e+00 -2.04061903e-03 -2.00954020e-01
-7.11315870e-01 -2.68294066e-02 1.26934183e+00 -7.69847155e-01
1.51454973e+00 -2.15916848e+00 -3.62813592e-01 5.95200777e-01
4.78648365e-01 3.95853192e-01 3.90998065e-01 8.78916442e-01
3.65371764e-01 7.40217924e-01 -2.36414269e-01 -3.38827893e-02
-1.42276123e-01 6.65759444e-02 -4.44052815e-01 1.13498902e+00
-1.12810537e-01 6.36121213e-01 -1.26899612e+00 -3.69880289e-01
-2.20747516e-01 2.57993191e-01 -3.40807199e-01 3.58377844e-02
-3.13116878e-01 2.98442304e-01 -6.04800224e-01 4.51592952e-01
6.04695857e-01 -3.97802353e-01 4.96693999e-01 4.24623579e-01
-2.56170511e-01 7.33733118e-01 -4.61271405e-01 6.82851970e-01
1.77107766e-01 1.19307101e+00 -1.63275689e-01 -6.05557740e-01
8.47095609e-01 4.15873826e-02 1.50596678e-01 -2.48233020e-01
4.85830307e-01 2.37229437e-01 -2.21781000e-01 -3.92573029e-01
7.92338848e-01 4.15497757e-02 -2.37873808e-01 1.12683964e+00
-4.49012727e-01 1.19560651e-01 9.27001610e-02 3.16044837e-01
1.53507030e+00 -4.91558015e-01 6.12789512e-01 4.58362028e-02
4.71733212e-01 1.74046472e-01 -2.78414398e-01 8.79761040e-01
-3.83992285e-01 1.49751157e-01 9.11238670e-01 -5.71477056e-01
-1.46087098e+00 -5.79158723e-01 -2.08497532e-02 9.72201765e-01
-1.41399503e-01 -6.59875512e-01 -8.07446957e-01 -5.13449609e-01
2.88184881e-01 5.16611636e-01 -6.97202444e-01 -5.15288413e-02
-6.28469110e-01 -9.94645655e-01 1.09799480e+00 -3.99531871e-01
6.13640606e-01 -8.50622594e-01 -6.56347275e-01 3.07586342e-01
-2.49764562e-01 -9.96921301e-01 -8.53072181e-02 -3.72291923e-01
-7.22372055e-01 -1.00180840e+00 -2.91821212e-01 -2.27743730e-01
4.14659381e-01 3.62407655e-01 6.17083609e-01 3.38667959e-01
-3.59350055e-01 -1.40963078e-01 -3.96521479e-01 -3.46561402e-01
-1.34509254e+00 2.23202147e-02 1.27856359e-01 4.46704268e-01
5.09300768e-01 -6.61177039e-01 -5.25818527e-01 3.51337910e-01
-1.14564741e+00 -3.01666081e-01 1.73939466e-01 3.29344809e-01
-1.81441098e-01 1.16865821e-01 2.03526989e-01 -1.23235166e+00
9.66854692e-01 -1.22725248e+00 -3.60764980e-01 -3.18784833e-01
-4.81023431e-01 -9.49727893e-02 6.63143396e-01 -6.28855646e-01
-4.15481985e-01 -8.12693760e-02 6.74368665e-02 2.05137417e-01
1.66253336e-02 5.27380168e-01 6.62938952e-01 -7.22064823e-02
9.22628224e-01 5.18230975e-01 5.04174232e-01 -3.60351771e-01
2.51279831e-01 1.38548219e+00 4.10988629e-01 2.43037581e-01
9.86069381e-01 7.76362658e-01 -1.41281277e-01 -1.33955085e+00
-4.40612793e-01 -7.83085287e-01 -2.31590748e-01 -4.36290711e-01
5.59755564e-01 -3.39480519e-01 -1.29531431e+00 7.09368348e-01
-1.19630313e+00 1.35658741e-01 2.26730809e-01 -8.86098668e-02
-4.55249310e-01 8.88551950e-01 -8.37246120e-01 -1.03431833e+00
-1.14849140e-03 -6.14228666e-01 7.84692943e-01 -2.68446207e-01
-7.99602687e-01 -8.74901354e-01 3.69049579e-01 1.75713554e-01
5.75158775e-01 7.53468096e-01 9.61917579e-01 -1.05094266e+00
-2.80164719e-01 -9.05466795e-01 -2.19965219e-01 -2.12931663e-01
-7.82091767e-02 8.63891467e-02 -8.65521491e-01 -1.35709479e-01
2.17474595e-01 -1.64654553e-01 5.89448571e-01 -1.33174166e-01
9.23259079e-01 -1.15379751e+00 -4.79085356e-01 2.59914517e-01
1.25631714e+00 -1.40081063e-01 8.02498817e-01 5.93981564e-01
3.21193218e-01 8.46254230e-01 7.84975886e-02 7.06066728e-01
-9.67826471e-02 4.68702078e-01 4.98493522e-01 6.10994458e-01
4.17269439e-01 -7.98883140e-01 4.57900226e-01 6.06637180e-01
1.98160037e-01 -2.82223284e-01 -9.65878129e-01 2.70887077e-01
-1.68475342e+00 -1.34186077e+00 -3.56704354e-01 2.10229778e+00
6.61391258e-01 2.18381360e-01 8.16978872e-01 4.58317608e-01
1.04582763e+00 4.08314764e-01 -2.50140280e-01 -7.26335883e-01
4.12510559e-02 -3.44459832e-01 1.08258784e+00 4.25200045e-01
-7.85512447e-01 5.12405872e-01 6.77938318e+00 8.90551209e-01
-1.04818141e+00 -5.07480558e-03 7.51726329e-01 3.20623934e-01
-3.02306633e-03 -3.61056626e-02 -5.33489168e-01 1.00273204e+00
1.51103687e+00 -4.51226652e-01 5.13084531e-01 6.21211350e-01
4.07465041e-01 -2.64346749e-01 -4.11464483e-01 6.94681883e-01
-2.16230061e-02 -1.69137478e+00 -2.02969223e-01 7.12878466e-01
3.24720204e-01 1.09171048e-01 5.07786497e-02 -3.22707713e-01
3.94502699e-01 -1.01064825e+00 6.75092697e-01 3.76620322e-01
6.11954391e-01 -4.36742634e-01 4.82195824e-01 7.33101547e-01
-6.92413390e-01 -2.51892328e-01 -1.72114268e-01 -3.05688590e-01
1.97261259e-01 5.50071299e-01 -1.55298781e+00 -1.31421119e-01
2.86257625e-01 7.02097714e-01 -7.16065884e-01 7.33049154e-01
3.55955243e-01 9.88821089e-01 -4.11820829e-01 -5.69592357e-01
4.18427020e-01 -3.53619993e-01 8.33799064e-01 1.62040865e+00
2.09146127e-01 -2.57695079e-01 -3.38449806e-01 6.16355240e-01
-1.24493010e-01 -5.37362732e-02 -1.11226487e+00 -8.79442513e-01
3.28078926e-01 9.91098881e-01 -1.00048327e+00 -3.70332420e-01
-5.48984185e-02 7.85499752e-01 8.02775994e-02 -2.06066787e-01
-4.34842318e-01 -3.37652504e-01 4.90324438e-01 8.09213817e-01
1.64831206e-01 -5.08708894e-01 -1.67468965e-01 -8.16753089e-01
-3.30901265e-01 -8.61737132e-01 1.69233650e-01 -3.85557413e-01
-1.33012795e+00 3.33722830e-01 -2.63693810e-01 -1.18729365e+00
-4.37305331e-01 -3.35785121e-01 -5.10643125e-01 4.68378633e-01
-8.39132369e-01 -5.19184649e-01 -5.65747507e-02 2.46038198e-01
1.03377976e-01 -2.39462599e-01 6.21947169e-01 -1.44855484e-01
-1.52529508e-01 2.69317150e-01 4.60067451e-01 3.09047610e-01
1.00069575e-01 -8.63919616e-01 8.15171003e-01 3.59583616e-01
-1.99317724e-01 4.71817672e-01 1.20947051e+00 -9.92143631e-01
-1.17986643e+00 -1.01704884e+00 1.26357734e+00 -5.28831303e-01
1.30352998e+00 -6.25339150e-01 -7.47064173e-01 5.04413128e-01
-1.06917389e-01 -3.93522233e-01 8.04535329e-01 -5.84528983e-01
-5.48927903e-01 6.24606252e-01 -1.55618012e+00 7.11031318e-01
6.71891451e-01 -8.60845983e-01 -4.59050030e-01 7.56042600e-01
4.25926030e-01 2.52660811e-01 -3.49078029e-01 -5.21336496e-01
8.91903102e-01 -1.21213174e+00 6.38798952e-01 -7.76643634e-01
6.53349638e-01 -1.10844769e-01 3.88194504e-03 -7.91122913e-01
-1.14474803e-01 -1.27245033e+00 -8.30587968e-02 8.75490546e-01
5.63513748e-02 -7.83142805e-01 1.17104816e+00 5.41964248e-02
7.91734993e-01 -2.53661960e-01 -8.44567537e-01 -6.17320359e-01
-2.16463909e-01 -1.91023424e-01 3.49997580e-01 1.05037224e+00
4.81940925e-01 5.54783866e-02 -5.29988050e-01 -3.70400935e-01
6.85401022e-01 -5.05539358e-01 9.69155967e-01 -1.19578958e+00
-1.79893672e-01 -4.75925207e-01 -9.99617755e-01 -6.51872635e-01
-1.67999700e-01 -9.12437022e-01 -4.05905604e-01 -6.29533052e-01
1.92449182e-01 -7.78852627e-02 6.16147161e-01 -1.76568121e-01
5.16404808e-01 6.29737079e-01 3.52348015e-02 8.02206695e-01
-3.60800594e-01 -1.24535233e-01 6.33198798e-01 -6.39010668e-02
-1.30824670e-01 1.98954806e-01 -4.35490131e-01 8.74441206e-01
7.36329138e-01 -9.42282081e-01 2.68765658e-01 5.53475857e-01
1.06938612e+00 -1.16016052e-03 3.80876064e-01 -6.04467690e-01
1.84083283e-01 8.66050869e-02 -6.28682524e-02 -4.84542698e-01
2.16436878e-01 -5.83443820e-01 2.54872650e-01 1.07212639e+00
-4.34345871e-01 2.37119108e-01 -2.61260748e-01 1.20314705e+00
8.46392661e-02 -2.97100693e-01 8.60511899e-01 -2.79262334e-01
2.37589121e-01 -2.37634853e-01 -1.23951161e+00 -1.81321144e-01
9.08521175e-01 -4.73275840e-01 -7.11534321e-01 -7.40494251e-01
-4.40877020e-01 -4.58064109e-01 6.68957293e-01 2.05175113e-02
3.19068879e-01 -9.23980832e-01 -7.46267676e-01 2.09948689e-01
-8.70871998e-04 -9.55573738e-01 -3.83656770e-01 8.02450716e-01
-1.01904070e+00 2.54662395e-01 -1.64735213e-01 -4.82970983e-01
-1.54741073e+00 4.09479201e-01 7.66499713e-02 -2.13255838e-01
-4.75047022e-01 2.03401878e-01 -7.64576674e-01 -1.78207129e-01
-1.64994001e-01 2.29743496e-01 -1.57881334e-01 4.03837621e-01
1.00248134e+00 8.96516621e-01 -4.19143252e-02 -1.05729210e+00
-2.05500990e-01 1.09800827e-02 -2.09888756e-01 -1.59364924e-01
1.24195814e+00 -1.20689854e-01 -4.89232868e-01 6.28900349e-01
1.70116842e+00 2.92071879e-01 -6.69733882e-01 -1.86599523e-01
4.57493305e-01 -9.51868474e-01 -3.67699057e-01 -2.29310304e-01
-1.95001796e-01 3.56606215e-01 1.53980672e-01 1.35170782e+00
3.46199214e-01 -1.94791425e-02 9.38300848e-01 2.53832102e-01
1.30604863e-01 -7.74086893e-01 1.40744641e-01 5.14360070e-01
3.38589877e-01 -8.11355889e-01 9.67484489e-02 -5.05007327e-01
-2.40219504e-01 1.18500662e+00 -7.30298638e-01 -2.49022692e-01
7.15133429e-01 7.41447806e-02 -2.69328296e-01 -3.58795315e-01
-3.89772356e-01 2.55507827e-01 -3.37245703e-01 4.65034872e-01
2.75098741e-01 1.40801847e-01 -7.05732882e-01 -1.14902351e-02
-5.94445527e-01 -1.05018035e-01 1.10973251e+00 1.12213504e+00
-8.63121510e-01 -7.45417655e-01 -5.33634067e-01 7.49590039e-01
-1.11951852e+00 6.18131310e-02 -1.01698530e+00 3.68027240e-01
-3.73855084e-01 1.02470672e+00 1.47016391e-01 -8.08328092e-01
-3.75160933e-01 2.38286912e-01 -8.66159946e-02 -2.49283016e-01
-8.05285215e-01 -4.37256932e-01 4.16027695e-01 -3.26972127e-01
-2.88196981e-01 -7.82538772e-01 -7.41034269e-01 -1.21928155e+00
-1.78678378e-01 -8.33818913e-02 1.19725525e+00 8.92330647e-01
1.70015752e-01 -2.77475089e-01 9.09275949e-01 -8.05319428e-01
-6.66235387e-01 -8.89838040e-01 -5.87865472e-01 6.44419611e-01
5.89506567e-01 -2.72677034e-01 -9.11420941e-01 -3.72016616e-02] | [8.171903610229492, 10.209074974060059] |
2e0b3569-6002-41a2-bbf5-cb2b1c56178d | the-first-comprehensive-dataset-with-multiple | 2303.02562 | null | https://arxiv.org/abs/2303.02562v2 | https://arxiv.org/pdf/2303.02562v2.pdf | The First Comprehensive Dataset with Multiple Distortion Types for Visual Just-Noticeable Differences | Recently, with the development of deep learning, a number of Just Noticeable Difference (JND) datasets have been built for JND modeling. However, all the existing JND datasets only label the JND points based on the level of compression distortion. Hence, JND models learned from such datasets can only be used for image/video compression. As known, JND is a major characteristic of the human visual system (HVS), which reflects the maximum visual distortion that the HVS can tolerate. Hence, a generalized JND modeling should take more kinds of distortion types into account. To benefit JND modeling, this work establishes a generalized JND dataset with a coarse-to-fine JND selection, which contains 106 source images and 1,642 JND maps, covering 25 distortion types. To this end, we proposed a coarse JND candidate selection scheme to select the distorted images from the existing Image Quality Assessment (IQA) datasets as JND candidates instead of generating JND maps ourselves. Then, a fine JND selection is carried out on the JND candidates with a crowdsourced subjective assessment. | ['Weisi Lin', 'Yuan Xue', 'Jian Jin', 'Yaxuan Liu'] | 2023-03-05 | null | null | null | null | ['image-quality-assessment'] | ['computer-vision'] | [-2.19500866e-02 -7.40405321e-01 -1.53415039e-01 -4.54574913e-01
-6.11296594e-01 -2.59274542e-01 5.02986491e-01 1.47677302e-01
-1.42394722e-01 4.34416085e-01 5.13678372e-01 -5.36503755e-02
-2.91006211e-02 -9.82689083e-01 -4.25435513e-01 -4.43828017e-01
3.09820682e-01 -8.97653177e-02 1.45538211e-01 -2.34869406e-01
3.04767877e-01 3.66308659e-01 -1.49925101e+00 3.51551801e-01
1.07067752e+00 1.34464061e+00 3.31572831e-01 6.20870173e-01
3.16245444e-02 6.51350617e-01 -1.30510449e+00 -4.37637717e-01
4.72567171e-01 -6.04386091e-01 -4.25205886e-01 2.90046185e-01
3.23611826e-01 -6.71787322e-01 -5.38758159e-01 1.28594732e+00
1.06167746e+00 -4.67890827e-03 6.42636895e-01 -1.37278402e+00
-1.08443260e+00 1.30559787e-01 -4.23621833e-01 4.38971192e-01
4.10489976e-01 4.79939640e-01 6.97612941e-01 -7.01768577e-01
3.28210026e-01 1.37181628e+00 4.60446745e-01 4.00279880e-01
-1.02748060e+00 -5.83404541e-01 -2.21099734e-01 4.34629887e-01
-1.43866193e+00 -2.83890903e-01 1.06046629e+00 -3.42940748e-01
7.55114794e-01 2.66064823e-01 7.44892716e-01 7.92104483e-01
2.74731696e-01 6.77343786e-01 1.15636647e+00 -2.05817625e-01
5.46195507e-01 5.30325882e-02 -5.23345053e-01 1.69861242e-01
1.57831892e-01 8.65156278e-02 -3.03813964e-01 6.61248863e-02
8.32863629e-01 -1.16573116e-02 -2.69139409e-01 -1.04253106e-01
-1.07360947e+00 5.54608881e-01 5.78208804e-01 4.40185443e-02
-2.35189408e-01 -1.16633289e-01 6.35039628e-01 5.64139009e-01
4.12570089e-01 2.87456900e-01 -2.72588700e-01 -5.51223606e-02
-7.47340918e-01 1.92383766e-01 3.92631292e-01 1.01737463e+00
6.41209900e-01 2.52427846e-01 -6.62644327e-01 1.21095824e+00
2.32552439e-01 7.82405138e-01 4.67266202e-01 -1.08315575e+00
5.81957519e-01 7.23171711e-01 2.06096366e-01 -1.58340085e+00
-2.22514570e-01 -5.26477575e-01 -1.25737965e+00 3.85255098e-01
-1.64214492e-01 1.38899788e-01 -1.02879453e+00 1.51620114e+00
-1.98217947e-02 -2.94510931e-01 -1.12154022e-01 1.21357334e+00
9.93003666e-01 7.89681017e-01 4.50053662e-02 -5.46089590e-01
1.02707016e+00 -7.81133533e-01 -8.07411671e-01 -3.29664722e-03
1.17058657e-01 -9.84855533e-01 1.21594095e+00 5.97659647e-01
-1.02968800e+00 -1.16394496e+00 -1.29988217e+00 -7.53591433e-02
-1.52900800e-01 5.94578534e-02 4.31966811e-01 6.64928377e-01
-1.10026312e+00 2.69633591e-01 -1.82760417e-01 -1.27538905e-01
5.64954042e-01 -6.12947308e-02 1.04482584e-01 -3.37199688e-01
-1.42397535e+00 8.20769072e-01 1.74085021e-01 -1.54039441e-02
-1.39023805e+00 -4.88082409e-01 -5.86290240e-01 -1.60281837e-01
7.94396698e-02 -5.91045737e-01 1.01846969e+00 -1.02602506e+00
-1.19270003e+00 8.77983212e-01 5.11902682e-02 -3.64576250e-01
6.66706681e-01 1.75634980e-01 -9.46906388e-01 9.46779922e-02
7.95023963e-02 5.76042354e-01 9.10561264e-01 -1.50053668e+00
-4.04114097e-01 -1.38426840e-01 3.96532565e-01 4.92087662e-01
-2.14093298e-01 1.64484397e-01 -7.57568479e-01 -1.00695300e+00
-2.17592269e-02 -4.50191468e-01 -8.51100460e-02 2.32199863e-01
-5.27534306e-01 -1.59344122e-01 4.72361058e-01 -7.91655481e-01
1.49745739e+00 -2.22631860e+00 -7.45928064e-02 -3.45938765e-02
4.38386917e-01 5.24349213e-01 -2.57151127e-01 1.54721320e-01
2.33091757e-01 3.57872248e-01 -1.59238979e-01 -1.86644509e-01
1.57086685e-01 -3.07112783e-02 -7.12558404e-02 2.01299816e-01
-1.23590045e-01 5.39476871e-01 -8.27976942e-01 -6.36224985e-01
3.40130597e-01 3.64493251e-01 -4.09510225e-01 4.88360971e-01
-1.49558917e-01 3.08832079e-01 -3.09726089e-01 7.24424005e-01
1.10752559e+00 5.11911698e-02 -4.60942030e-01 -6.08687282e-01
4.13143411e-02 -1.23054989e-01 -1.09322846e+00 1.64619255e+00
-4.34741050e-01 5.62911928e-01 -3.64484519e-01 -6.30330622e-01
1.24299812e+00 3.06432545e-01 5.65427959e-01 -1.20673656e+00
1.06418245e-01 4.07654822e-01 2.35006604e-02 -6.80996537e-01
3.27803105e-01 4.16776165e-03 -5.77121861e-02 3.72516990e-01
-2.18571797e-02 -4.35500771e-01 1.91873506e-01 -1.27619162e-01
9.02394593e-01 -3.97643924e-01 1.99387938e-01 -4.80856225e-02
5.25324345e-01 -2.93285668e-01 6.98052466e-01 6.52301908e-01
-6.86217904e-01 9.67422903e-01 1.75767675e-01 -7.60253847e-01
-1.51304007e+00 -1.11349082e+00 -2.15154499e-01 6.35687828e-01
5.56458652e-01 -1.62240475e-01 -8.00027728e-01 -2.33848140e-01
-3.70349765e-01 3.43037248e-01 -2.20930561e-01 -5.42704105e-01
-2.74284720e-01 -7.48705029e-01 6.16120160e-01 3.03042233e-01
1.33360589e+00 -1.21718240e+00 -2.97188669e-01 -5.04502989e-02
-1.59752935e-01 -7.89970160e-01 -6.38340712e-01 -3.74226332e-01
-5.29083431e-01 -8.45999122e-01 -1.11064553e+00 -9.32901740e-01
4.54384863e-01 4.12171930e-01 1.21312809e+00 -1.41738234e-02
-1.98604539e-01 1.30174160e-02 -3.82459074e-01 -4.88178581e-01
-5.58035016e-01 -3.64015132e-01 2.10401669e-01 -6.55043349e-02
4.11074132e-01 -6.42305076e-01 -1.10687137e+00 4.82093513e-01
-9.63171721e-01 1.30204499e-01 6.20478094e-01 3.90699416e-01
9.07063425e-01 7.38065302e-01 5.93677342e-01 -2.20886186e-01
1.00336695e+00 -2.77410746e-01 -2.52350509e-01 2.19515905e-01
-6.93567634e-01 -3.74873638e-01 1.00770926e+00 -4.08343941e-01
-1.10368299e+00 -6.53570652e-01 -1.56947702e-01 -5.69660366e-01
-3.68715346e-01 3.42429787e-01 -7.02031910e-01 1.50578003e-02
7.98217118e-01 6.02153122e-01 -4.20109957e-01 -5.75819492e-01
1.48637384e-01 9.92756069e-01 4.64958251e-01 -1.64436147e-01
6.60331249e-01 1.48117706e-01 -2.29858339e-01 -5.37167490e-01
-5.17089367e-01 -1.03686005e-02 -2.40802959e-01 -1.77030429e-01
8.54063809e-01 -1.08950126e+00 -3.52773577e-01 9.04801190e-01
-1.28087854e+00 -8.71415734e-02 -3.42899442e-01 5.43985784e-01
-5.12650192e-01 2.38064736e-01 -5.52464426e-01 -5.24003148e-01
-2.98187256e-01 -1.35454011e+00 8.22670579e-01 3.84101808e-01
1.59082472e-01 -5.79012096e-01 7.18552098e-02 2.83972561e-01
7.28249848e-01 9.72446501e-02 9.30774689e-01 -1.08168036e-01
-4.91129428e-01 -1.60465091e-01 -4.33520198e-01 8.39638174e-01
3.42758060e-01 -3.37824851e-01 -8.73035610e-01 -2.64976442e-01
3.14761639e-01 -2.95967638e-01 6.33985877e-01 5.47748506e-01
1.52742779e+00 -4.33560699e-01 2.82606572e-01 8.83694828e-01
1.57229662e+00 6.80131435e-01 9.15054560e-01 4.04767394e-01
7.54930079e-01 9.36009064e-02 5.25893688e-01 5.30163050e-01
2.82438904e-01 7.37963200e-01 4.46802974e-01 -3.07629228e-01
-5.65901935e-01 -4.32334960e-01 1.28269255e-01 8.96439254e-01
-2.19671920e-01 -5.68010747e-01 -6.46853745e-01 4.08737719e-01
-1.18379462e+00 -8.78091097e-01 2.81760007e-01 2.09882951e+00
1.10514832e+00 2.83726633e-01 1.12724379e-01 3.64988238e-01
9.73197341e-01 5.60699463e-01 -9.68590915e-01 -5.03844857e-01
-4.36409116e-01 -2.38012642e-01 2.39051521e-01 1.98109642e-01
-9.35815334e-01 3.16395521e-01 6.14818525e+00 1.28969181e+00
-1.05993557e+00 -2.29648761e-02 8.42350185e-01 2.76319869e-02
-5.95932245e-01 -1.28523856e-01 -3.53413910e-01 9.30613577e-01
5.19243598e-01 -4.56690460e-01 6.16230965e-01 8.03820312e-01
5.92468023e-01 6.48669228e-02 -6.94540441e-01 1.44956636e+00
1.53839946e-01 -1.01080167e+00 5.80759645e-01 -1.91818431e-01
1.04405642e+00 -3.37039471e-01 1.64335653e-01 1.41688347e-01
-1.52073711e-01 -1.11070573e+00 7.32898057e-01 5.53143561e-01
1.18802440e+00 -8.46372366e-01 8.71339619e-01 2.40481019e-01
-1.07787883e+00 -6.15061447e-02 -9.46784735e-01 -2.88990606e-03
1.29737601e-01 8.51663291e-01 -2.78712898e-01 4.74054068e-01
8.10185015e-01 7.08512306e-01 -7.66929746e-01 1.33829856e+00
-1.28371865e-01 4.13589954e-01 1.77426502e-01 1.00159623e-01
-3.26661021e-01 2.96034999e-02 4.72688347e-01 1.04543459e+00
5.00075936e-01 8.80354941e-02 3.15648355e-02 8.79227221e-01
-3.21444839e-01 1.47294104e-01 -4.20471400e-01 3.01997215e-01
7.76591063e-01 8.48600209e-01 -3.99573952e-01 -2.10000992e-01
-2.12600172e-01 1.14672720e+00 -3.89134794e-01 4.46546435e-01
-8.86889994e-01 -5.73956192e-01 5.01634359e-01 -4.52036895e-02
-6.41946048e-02 5.36487922e-02 -2.63659954e-01 -1.07708180e+00
2.47881040e-01 -1.14190233e+00 2.15127952e-02 -1.14093125e+00
-1.45072317e+00 8.19485962e-01 -1.31765485e-01 -1.84543204e+00
3.16635877e-01 -2.16823742e-01 -6.88143075e-01 1.00674629e+00
-1.33270669e+00 -8.85607421e-01 -7.35169113e-01 5.62347531e-01
7.96215415e-01 -3.02650511e-01 5.99805534e-01 5.69475830e-01
-3.06312740e-01 7.26898968e-01 -4.88395616e-02 1.59003213e-01
7.21373737e-01 -1.02465272e+00 2.59003162e-01 8.40894461e-01
-3.01831096e-01 3.61233681e-01 7.85428941e-01 -4.80553448e-01
-1.06293511e+00 -1.20102489e+00 6.35001481e-01 -1.44401386e-01
2.93367412e-02 -1.25963181e-01 -6.91218853e-01 8.94706473e-02
2.12045297e-01 1.50567949e-01 3.43873680e-01 -5.31938076e-01
-2.60919273e-01 -7.66652584e-01 -1.37279737e+00 5.20168781e-01
1.14737499e+00 -6.75345063e-01 -2.71958083e-01 1.72960386e-01
1.17558765e+00 -2.29806617e-01 -9.43129420e-01 5.30479014e-01
2.25436896e-01 -1.32804954e+00 1.11532605e+00 -5.28965816e-02
7.95894682e-01 -6.66546464e-01 -3.87166083e-01 -1.56272054e+00
-3.83181870e-01 -2.36392215e-01 -1.46099150e-01 1.37442851e+00
1.23697564e-01 -1.27768099e-01 3.00137758e-01 3.36577028e-01
-2.26836458e-01 -7.07860291e-01 -7.50394523e-01 -7.80983210e-01
-2.43692726e-01 -6.02537036e-01 1.14817035e+00 7.16846108e-01
-5.05421698e-01 8.33266824e-02 -7.05665588e-01 -5.77166304e-02
6.16053939e-01 6.46719337e-02 6.61006033e-01 -7.33055770e-01
-2.67072022e-01 -4.01219010e-01 -6.86210573e-01 -1.12766373e+00
-5.39124250e-01 -6.16090357e-01 4.27450202e-02 -1.82799113e+00
3.94700825e-01 -5.59514046e-01 -5.00608623e-01 1.05624776e-02
-1.00852773e-01 3.62579226e-01 1.41480684e-01 6.71344817e-01
-7.30722189e-01 7.95428991e-01 1.80436778e+00 -5.18468797e-01
-1.51932746e-01 -7.87893236e-02 -7.40108132e-01 6.70201898e-01
8.70855033e-01 -2.29978457e-01 -7.36897826e-01 -8.70818317e-01
3.33365321e-01 1.17390364e-01 3.64845067e-01 -1.32150352e+00
1.42440379e-01 -4.83843476e-01 6.82716310e-01 -5.29195487e-01
6.90870807e-02 -6.47571743e-01 1.30884424e-01 4.05742645e-01
-6.11059248e-01 -1.35624766e-01 -8.51718634e-02 5.04575491e-01
-5.83243310e-01 1.56319112e-01 8.18700850e-01 -2.18110308e-01
-9.80563641e-01 6.20262563e-01 -9.03529301e-02 1.26216605e-01
8.46257567e-01 -3.77944350e-01 -3.96323502e-01 -5.71609378e-01
-3.45791936e-01 -7.50364810e-02 6.63633943e-01 4.08637732e-01
9.73595321e-01 -1.75453329e+00 -8.43014121e-01 2.90866047e-01
3.09058875e-01 2.38049552e-02 6.69489145e-01 4.12050337e-01
-6.62050307e-01 9.49993432e-02 -5.42180002e-01 -3.98997098e-01
-8.53820562e-01 8.47992659e-01 4.29057717e-01 2.26808935e-02
-1.92092955e-01 7.26511061e-01 1.44208923e-01 -1.76432356e-01
2.13566065e-01 -1.82989985e-01 -4.96151716e-01 -2.75513440e-01
6.71791136e-01 5.54895103e-01 5.86473309e-02 -6.63115740e-01
-2.50811458e-01 6.68139458e-01 4.01782066e-01 4.78430130e-02
1.06817961e+00 -6.23970449e-01 -4.36751395e-02 3.12260211e-01
1.50044167e+00 2.79388465e-02 -1.19400692e+00 -2.19336182e-01
-6.33200824e-01 -8.22572589e-01 2.34756265e-02 -1.00375509e+00
-1.28637135e+00 1.01001692e+00 1.09171975e+00 3.62356395e-01
1.77923667e+00 -2.14566261e-01 1.09699059e+00 6.56313598e-02
5.18630862e-01 -1.31655085e+00 3.08187753e-01 1.27782926e-01
1.12514186e+00 -1.38274646e+00 -9.65630710e-02 5.70536917e-03
-7.19126642e-01 8.64512682e-01 7.34286845e-01 -1.99610814e-01
6.56508684e-01 -6.93528354e-02 1.67502627e-01 -6.35726675e-02
-3.44380021e-01 1.55577943e-01 3.82865578e-01 8.73364747e-01
2.31081128e-01 1.01195887e-01 -4.35328394e-01 6.59061253e-01
-2.13035032e-01 2.30469108e-01 4.76746887e-01 4.86504167e-01
-5.57783961e-01 -7.25246310e-01 -2.19231918e-01 4.61497873e-01
-2.29198292e-01 -8.72869045e-02 -8.69066864e-02 2.02831775e-01
8.79802883e-01 1.25936580e+00 -3.07527166e-02 -8.90362501e-01
5.91829956e-01 -7.29354501e-01 1.68262720e-01 -3.53039294e-01
-2.16870420e-02 -2.44070426e-01 -2.92817771e-01 -5.11537611e-01
-5.19335032e-01 5.36253601e-02 -9.00322914e-01 -5.48409104e-01
6.84648976e-02 -1.63885087e-01 3.20684195e-01 8.48036647e-01
2.60726571e-01 5.34910798e-01 1.08007789e+00 -5.89346170e-01
-3.94297570e-01 -9.88062322e-01 -8.01335096e-01 7.91485608e-01
3.20001692e-01 -3.43360990e-01 -4.76410776e-01 1.75508127e-01] | [11.785974502563477, -1.8291007280349731] |
1527365e-811a-4305-bb55-a3ac570f03e2 | dtw-siamesenet-dynamic-time-warped-siamese | 2303.00171 | null | https://arxiv.org/abs/2303.00171v1 | https://arxiv.org/pdf/2303.00171v1.pdf | DTW-SiameseNet: Dynamic Time Warped Siamese Network for Mispronunciation Detection and Correction | Personal Digital Assistants (PDAs) - such as Siri, Alexa and Google Assistant, to name a few - play an increasingly important role to access information and complete tasks spanning multiple domains, and by diverse groups of users. A text-to-speech (TTS) module allows PDAs to interact in a natural, human-like manner, and play a vital role when the interaction involves people with visual impairments or other disabilities. To cater to the needs of a diverse set of users, inclusive TTS is important to recognize and pronounce correctly text in different languages and dialects. Despite great progress in speech synthesis, the pronunciation accuracy of named entities in a multi-lingual setting still has a large room for improvement. Existing approaches to correct named entity (NE) mispronunciations, like retraining Grapheme-to-Phoneme (G2P) models, or maintaining a TTS pronunciation dictionary, require expensive annotation of the ground truth pronunciation, which is also time consuming. In this work, we present a highly-precise, PDA-compatible pronunciation learning framework for the task of TTS mispronunciation detection and correction. In addition, we also propose a novel mispronunciation detection model called DTW-SiameseNet, which employs metric learning with a Siamese architecture for Dynamic Time Warping (DTW) with triplet loss. We demonstrate that a locale-agnostic, privacy-preserving solution to the problem of TTS mispronunciation detection is feasible. We evaluate our approach on a real-world dataset, and a corpus of NE pronunciations of an anonymized audio dataset of person names recorded by participants from 10 different locales. Human evaluation shows our proposed approach improves pronunciation accuracy on average by ~6% compared to strong phoneme-based and audio-based baselines. | ['Srinivas Chappidi', 'Becci Williamson', 'Prabal Vashisht', 'Daniela de la Parra Aguilar', 'Kriti Bhasin', 'Raviteja Anantha'] | 2023-03-01 | null | null | null | null | ['metric-learning', 'metric-learning', 'speech-synthesis', 'dynamic-time-warping'] | ['computer-vision', 'methodology', 'speech', 'time-series'] | [ 6.85381815e-02 -4.64124143e-01 -1.58123132e-02 -4.04627979e-01
-1.51994777e+00 -7.10698128e-01 4.43919599e-01 3.40201077e-03
-5.69784522e-01 7.91555762e-01 4.73357707e-01 -3.07841390e-01
2.48530418e-01 -4.05613631e-01 -5.26989460e-01 -3.47374558e-01
2.41419941e-01 6.15406811e-01 3.11406143e-02 -8.68057311e-02
-2.16667295e-01 4.89513725e-01 -1.30010390e+00 2.18154967e-01
1.16388535e+00 8.56992900e-01 6.85739741e-02 7.78264165e-01
-1.10266946e-01 1.83258414e-01 -8.21654320e-01 -8.40881109e-01
2.55834758e-01 9.32611376e-02 -5.99179268e-01 -4.34781969e-01
9.13720548e-01 -2.44130522e-01 -5.90938687e-01 1.12877452e+00
9.84722018e-01 2.47889310e-01 4.56808388e-01 -1.37523830e+00
-2.75956184e-01 3.57488126e-01 -8.82529840e-02 4.41663675e-02
6.23973310e-01 8.33544135e-02 9.57953811e-01 -8.55220020e-01
2.41209492e-01 1.16781914e+00 7.98785031e-01 7.95246840e-01
-1.29822159e+00 -9.44476843e-01 -2.40215026e-02 2.77892470e-01
-1.85995615e+00 -9.55113649e-01 2.73630321e-01 -8.16819742e-02
9.28144395e-01 6.81097031e-01 1.72483698e-01 1.75291455e+00
-3.59794825e-01 8.98289740e-01 7.82539308e-01 -2.79494137e-01
9.71543342e-02 3.77240986e-01 -1.17706925e-01 3.17155182e-01
-7.89444000e-02 -1.37928918e-01 -9.21534538e-01 -4.85040247e-01
3.93742502e-01 -2.66100138e-01 -5.06065547e-01 -6.26585484e-02
-1.38065994e+00 3.21505070e-01 -1.76465675e-01 1.87920406e-01
-2.22532243e-01 1.22833299e-02 4.75115538e-01 3.08089823e-01
3.86027217e-01 1.42951354e-01 -8.98928642e-01 -6.89186692e-01
-1.07310104e+00 2.84316480e-01 9.86549675e-01 1.24497437e+00
1.71562910e-01 -1.42519558e-02 -3.70842457e-01 1.22487807e+00
1.32311746e-01 9.82489228e-01 6.42609894e-01 -5.23360908e-01
1.05382490e+00 4.16048765e-02 4.30405349e-01 -7.10614204e-01
-7.35442713e-02 -1.94410175e-01 -9.27265704e-01 -6.52403384e-02
7.72460163e-01 -2.93171167e-01 -6.53875470e-01 2.01559997e+00
2.62562335e-01 4.67928410e-01 -1.87609643e-01 5.44194460e-01
4.10755724e-01 4.98293698e-01 1.80036917e-01 1.86440069e-02
1.42586184e+00 -6.28517807e-01 -9.33855891e-01 -1.40730470e-01
4.81689125e-01 -8.29110265e-01 1.28666747e+00 4.04107004e-01
-8.41431797e-01 -3.28536779e-01 -7.40769684e-01 -1.20568953e-01
-5.56600153e-01 2.61714846e-01 2.01436251e-01 1.15508425e+00
-1.04049575e+00 5.41564941e-01 -7.56414831e-01 -5.49863100e-01
3.62318695e-01 6.12490177e-01 -5.53425372e-01 -1.20716030e-02
-1.34361243e+00 8.84145498e-01 4.77065854e-02 -2.40483552e-01
-5.12858987e-01 -9.34398830e-01 -8.12046528e-01 3.92285287e-02
9.33902934e-02 -5.62394917e-01 1.38836896e+00 -5.15670598e-01
-1.52527452e+00 7.04939663e-01 -5.61356366e-01 -5.28692842e-01
7.69140065e-01 -3.29899997e-01 -1.24975097e+00 -2.46017456e-01
1.86857015e-01 2.86059022e-01 1.11689484e+00 -7.21065760e-01
-1.06408930e+00 -4.37210947e-01 -5.84597349e-01 3.20333421e-01
-6.55202150e-01 4.23116028e-01 -5.27697206e-01 -1.02757382e+00
-2.43443757e-01 -9.68065798e-01 1.89334467e-01 -7.54309371e-02
-6.83447480e-01 -1.38682812e-01 6.62973821e-01 -1.32468081e+00
1.42465031e+00 -2.37406111e+00 -1.83225676e-01 3.35047603e-01
5.75617366e-02 5.77742398e-01 -7.29031414e-02 1.95517823e-01
-8.60775169e-03 9.69809201e-03 -3.26023310e-01 -8.72764051e-01
2.75565207e-01 -4.27472107e-02 -4.92655784e-01 3.95178765e-01
-2.28525266e-01 6.35022283e-01 -7.76221871e-01 -2.79471070e-01
2.30189353e-01 5.50703704e-01 -4.39328760e-01 2.50042230e-01
1.52496800e-01 3.91702533e-01 -8.89364928e-02 6.12827778e-01
6.25566840e-01 3.55713248e-01 -5.52023947e-02 -3.45897041e-02
-4.13536467e-02 7.03014493e-01 -1.52882326e+00 1.65015471e+00
-8.05439472e-01 7.10378051e-01 3.79787624e-01 -5.04469454e-01
6.31487310e-01 7.00148046e-01 2.42478251e-01 -4.94086951e-01
-9.72625464e-02 5.60210347e-01 -4.59095538e-01 -3.85889441e-01
4.85435694e-01 2.79736012e-01 -1.73621893e-01 1.57056615e-01
9.92245749e-02 2.48298839e-01 -2.78285265e-01 5.47968149e-02
1.20224667e+00 -3.66277218e-01 2.76933938e-01 1.14390910e-01
6.42247915e-01 -4.86422211e-01 5.04885972e-01 7.41524994e-01
-4.12696242e-01 8.48712564e-01 -2.61894792e-01 -1.89360008e-02
-8.60177815e-01 -1.30389786e+00 -1.72712803e-01 1.26984322e+00
-3.25060606e-01 -5.51182985e-01 -9.80111003e-01 -7.47304022e-01
1.22964885e-02 1.04320335e+00 1.44078270e-01 1.42686674e-02
-6.87407970e-01 -3.76787096e-01 1.15635216e+00 4.42966193e-01
6.41765356e-01 -8.35091233e-01 3.03852141e-01 3.49823147e-01
-5.83909273e-01 -1.45505583e+00 -1.22860897e+00 -1.20171249e-01
-4.30290014e-01 -5.64378679e-01 -9.12594557e-01 -8.73611867e-01
1.62088618e-01 1.92145750e-01 7.43424833e-01 -7.01709270e-01
-9.71601233e-02 5.84838390e-01 -6.59979805e-02 -3.21170837e-01
-4.38480794e-01 3.17173392e-01 7.10461140e-01 3.51386100e-01
7.10083008e-01 -8.28292608e-01 -2.95691311e-01 4.65205520e-01
-2.97322541e-01 -4.32467133e-01 2.30567738e-01 6.90742135e-01
3.74186963e-01 3.80987637e-02 5.74255884e-01 -6.44245684e-01
7.08324552e-01 -3.77639920e-01 -4.28746074e-01 2.96310216e-01
-4.53639567e-01 -1.36555001e-01 8.35861623e-01 -5.80320597e-01
-1.04356325e+00 1.52993113e-01 -5.44736326e-01 -2.88706362e-01
-5.96590757e-01 -4.71283495e-02 -8.75907302e-01 -9.63429064e-02
6.79615855e-01 4.96763885e-01 -3.18475366e-01 -7.31855929e-01
5.43015182e-01 1.42712247e+00 1.21230745e+00 -4.90126848e-01
9.38504696e-01 2.22231612e-01 -7.09283233e-01 -1.17493641e+00
-2.68351883e-01 -8.44853640e-01 -3.92265767e-01 1.96728706e-01
6.65229142e-01 -1.07987678e+00 -7.94529438e-01 8.17542374e-01
-1.39990580e+00 1.11361071e-01 -1.37773052e-01 6.51050687e-01
-3.87980491e-01 4.08753514e-01 -2.73776382e-01 -8.34053695e-01
-4.70631033e-01 -8.36584687e-01 1.05847836e+00 2.63170730e-02
-5.10456324e-01 -7.86885381e-01 -3.35943624e-02 5.76893806e-01
5.73563278e-01 -2.33940288e-01 5.79463899e-01 -1.08062875e+00
-2.98414171e-01 -4.41642851e-01 -7.02248737e-02 4.57134932e-01
3.33541006e-01 -5.46530724e-01 -1.31349850e+00 -3.48134220e-01
-1.05279617e-01 1.66495413e-01 3.85991424e-01 2.08237126e-01
1.11390209e+00 -6.74589932e-01 -3.21948200e-01 7.31706858e-01
8.45380247e-01 6.88107014e-02 6.11380994e-01 4.92389873e-02
9.36042130e-01 4.08816278e-01 3.32638592e-01 5.84021747e-01
5.78425884e-01 1.06391704e+00 -2.62427945e-02 2.33604446e-01
-3.00539285e-01 -5.87020040e-01 5.78057468e-01 8.76882911e-01
1.43991694e-01 -4.93783653e-01 -7.51401544e-01 8.02402854e-01
-1.56870103e+00 -9.16963518e-01 -4.16651517e-02 2.68936133e+00
9.44556355e-01 -2.81206280e-01 3.04898947e-01 2.43158966e-01
1.18248069e+00 -1.83311366e-02 -5.17574549e-01 -1.81901649e-01
-1.12364836e-01 4.64462668e-01 8.34184587e-01 3.53759676e-01
-1.19032276e+00 1.07644606e+00 4.96127987e+00 1.23403633e+00
-1.04072690e+00 2.56693393e-01 1.63501084e-01 -8.33092034e-02
-5.63067384e-02 -4.37808305e-01 -9.39725816e-01 6.85634494e-01
1.29226542e+00 -1.44822046e-01 9.33077335e-01 7.80356348e-01
2.64869094e-01 4.10930932e-01 -1.24641597e+00 1.79803383e+00
1.39628220e-02 -8.79290223e-01 -3.03341091e-01 -1.02663800e-01
3.48977834e-01 3.17943215e-01 2.84845829e-01 1.88131109e-01
2.91636735e-01 -9.46180403e-01 7.71555185e-01 2.04241917e-01
1.35613000e+00 -7.68152356e-01 4.26405311e-01 2.06454724e-01
-1.34444702e+00 4.97645326e-02 -3.35297622e-02 5.05796850e-01
2.66473114e-01 3.42854083e-01 -1.16375434e+00 3.86836231e-01
7.74210870e-01 3.25805336e-01 -2.64404804e-01 1.26780534e+00
-2.24308148e-01 8.30079317e-01 -5.35264790e-01 -9.88752209e-03
-3.60873908e-01 1.56040743e-01 8.98362815e-01 1.42134666e+00
6.75224900e-01 -7.73375016e-03 -3.04445803e-01 4.64718372e-01
-3.47532809e-01 3.77084523e-01 -5.39484739e-01 -1.68067813e-01
1.10493481e+00 8.88661742e-01 2.84101181e-02 -8.05870593e-02
-4.51667696e-01 1.77691281e+00 6.48449659e-02 4.81994987e-01
-6.75225139e-01 -8.09241831e-01 1.35443270e+00 6.24051765e-02
1.74872369e-01 -3.49160343e-01 -2.67633736e-01 -1.37346888e+00
3.55641574e-01 -1.26626003e+00 4.13036376e-01 -3.33689570e-01
-1.40142322e+00 4.90772218e-01 -3.88502002e-01 -1.31960630e+00
-3.65849227e-01 -4.32440877e-01 -5.11068404e-01 1.19009674e+00
-1.23965204e+00 -1.10209143e+00 -4.05407548e-02 1.04093230e+00
4.85330969e-01 -5.30346334e-01 1.19097316e+00 9.67707157e-01
-4.40239042e-01 1.37035120e+00 3.60983551e-01 3.94153267e-01
1.19422007e+00 -1.23062587e+00 9.91583467e-01 9.74396288e-01
5.11939764e-01 7.27245271e-01 6.39927506e-01 -5.71354508e-01
-1.12332976e+00 -1.48105454e+00 1.53936386e+00 -6.79204047e-01
5.01514316e-01 -7.76220500e-01 -8.76713395e-01 7.71890461e-01
-9.46977213e-02 -4.21283506e-02 7.40143359e-01 1.18169062e-01
-4.25371081e-01 -2.21138358e-01 -1.39365745e+00 8.07250261e-01
1.25428951e+00 -1.17085814e+00 -4.70676214e-01 3.72740120e-01
7.77011156e-01 -3.88938546e-01 -6.89743757e-01 -2.35084713e-01
5.07484555e-01 -4.46442753e-01 1.12122512e+00 -6.15889966e-01
-5.72119296e-01 -2.88407356e-01 -4.50483918e-01 -1.50192714e+00
4.15059850e-02 -1.34505177e+00 -1.12149648e-01 1.84901834e+00
4.08504903e-01 -7.84670234e-01 6.50512397e-01 1.02739799e+00
-6.13629669e-02 1.91140130e-01 -1.30191183e+00 -1.01069462e+00
-2.97054470e-01 -1.01223242e+00 8.95791173e-01 1.00820303e+00
-6.19928986e-02 1.45457789e-01 -6.71810806e-01 7.43697107e-01
5.95556021e-01 -7.79933631e-01 7.80731022e-01 -1.09898341e+00
-1.09586433e-01 -5.05638011e-02 -5.05593479e-01 -1.07598376e+00
2.98942655e-01 -1.01064670e+00 6.46547228e-02 -1.11593711e+00
-5.39165676e-01 -5.40949643e-01 -9.09355134e-02 4.40481961e-01
7.84244202e-03 -1.21143579e-01 8.25683847e-02 -1.36109525e-02
-3.20007950e-01 5.73612034e-01 3.60639811e-01 -3.02214265e-01
-4.47123319e-01 5.46679676e-01 -4.51642811e-01 6.46285951e-01
7.02261329e-01 -6.04139745e-01 -1.38798013e-01 -5.59610307e-01
-2.00083613e-01 2.47381162e-03 2.15708554e-01 -1.19027007e+00
6.14585817e-01 1.51717842e-01 -1.01975529e-02 -2.83743143e-01
4.91418302e-01 -8.52985919e-01 4.96065011e-03 1.56779569e-02
-2.74522096e-01 -1.21155024e-01 1.22322530e-01 5.58378100e-01
-9.07193795e-02 6.70126304e-02 6.47879183e-01 1.94866106e-01
-5.26550472e-01 5.58539450e-01 -4.56281036e-01 3.12518269e-01
5.22935331e-01 -4.05819714e-02 -9.32681840e-03 -5.03207564e-01
-5.62137246e-01 5.36016263e-02 1.52415693e-01 6.19622111e-01
4.85197783e-01 -1.49299610e+00 -8.28475654e-01 3.50461841e-01
2.70640850e-01 -3.73643190e-01 1.38791874e-01 4.72886115e-01
-1.45267220e-02 3.85554016e-01 1.25659078e-01 -1.32151932e-01
-1.45934558e+00 1.28166795e-01 2.34373137e-01 1.93538681e-01
-4.51799780e-01 1.17067730e+00 -1.08798742e-01 -8.00506890e-01
7.38048792e-01 -4.31687027e-01 7.22684488e-02 1.36136636e-02
8.00254285e-01 7.71113753e-01 4.71319050e-01 -8.48385513e-01
-6.22884274e-01 2.09182024e-01 -1.34348541e-01 -4.87815142e-01
9.92582977e-01 -4.26257700e-01 2.70594180e-01 1.69204473e-01
1.37745404e+00 5.93336761e-01 -8.47664237e-01 -4.34379339e-01
-3.52582820e-02 -5.67803085e-01 -1.11326188e-01 -8.21482539e-01
-8.20077002e-01 8.23166788e-01 8.51147354e-01 -6.47282973e-02
8.56527090e-01 -2.28999913e-01 1.50047219e+00 3.45408112e-01
5.54940343e-01 -1.30630648e+00 -4.75317210e-01 4.65814769e-01
6.66025758e-01 -1.20172417e+00 -7.26671576e-01 -2.84339458e-01
-7.25251198e-01 7.77211964e-01 2.86467642e-01 4.68176603e-01
4.68013614e-01 1.44215226e-01 8.21646824e-02 4.97243941e-01
-2.96001226e-01 -8.79185796e-02 4.01560873e-01 1.25410986e+00
1.00542195e-01 4.32820052e-01 1.97654054e-01 9.70841825e-01
-6.02369487e-01 -2.67001897e-01 1.14060111e-01 5.48341990e-01
6.84034126e-03 -1.29731452e+00 -5.20185471e-01 3.81534249e-01
-5.25051236e-01 -5.19131720e-01 -2.91375220e-01 2.22643331e-01
5.51841706e-02 1.21534669e+00 -1.56399816e-01 -5.38874805e-01
5.92386246e-01 4.04172361e-01 1.20456517e-01 -4.36217785e-01
-6.56767666e-01 -3.17265242e-01 3.94934803e-01 -5.33381641e-01
2.28873730e-01 -1.01485634e+00 -1.16792941e+00 -5.80491722e-01
-1.38374060e-01 -1.44041017e-01 9.19500470e-01 8.03844869e-01
6.02004945e-01 1.20654695e-01 6.24710917e-01 -3.58753979e-01
-7.45003045e-01 -7.39888251e-01 -7.40629673e-01 4.19052452e-01
4.19309825e-01 -2.12188885e-01 -2.68690556e-01 -4.65273336e-02] | [14.309563636779785, 6.001967430114746] |
dff4f6ad-667e-4210-b834-7d347b8e5261 | learning-towards-the-largest-margins-1 | 2206.11589 | null | https://arxiv.org/abs/2206.11589v1 | https://arxiv.org/pdf/2206.11589v1.pdf | Learning Towards the Largest Margins | One of the main challenges for feature representation in deep learning-based classification is the design of appropriate loss functions that exhibit strong discriminative power. The classical softmax loss does not explicitly encourage discriminative learning of features. A popular direction of research is to incorporate margins in well-established losses in order to enforce extra intra-class compactness and inter-class separability, which, however, were developed through heuristic means, as opposed to rigorous mathematical principles. In this work, we attempt to address this limitation by formulating the principled optimization objective as learning towards the largest margins. Specifically, we firstly define the class margin as the measure of inter-class separability, and the sample margin as the measure of intra-class compactness. Accordingly, to encourage discriminative representation of features, the loss function should promote the largest possible margins for both classes and samples. Furthermore, we derive a generalized margin softmax loss to draw general conclusions for the existing margin-based losses. Not only does this principled framework offer new perspectives to understand and interpret existing margin-based losses, but it also provides new insights that can guide the design of new tools, including sample margin regularization and largest margin softmax loss for the class-balanced case, and zero-centroid regularization for the class-imbalanced case. Experimental results demonstrate the effectiveness of our strategy on a variety of tasks, including visual classification, imbalanced classification, person re-identification, and face verification. | ['Xiangyang Ji', 'Xin Gao', 'Junjun Jiang', 'Deming Zhai', 'Xianming Liu', 'Xiong Zhou'] | 2022-06-23 | learning-towards-the-largest-margins | https://openreview.net/forum?id=hqkhcFHOeKD | https://openreview.net/pdf?id=hqkhcFHOeKD | iclr-2022-4 | ['imbalanced-classification'] | ['miscellaneous'] | [ 1.95681572e-01 9.82427225e-02 -4.64406639e-01 -6.28794551e-01
-4.44865108e-01 -1.53215945e-01 2.63643265e-01 3.19886178e-01
-4.06967640e-01 6.70745969e-01 -2.19366029e-02 -5.57712838e-02
-5.63844085e-01 -6.91224039e-01 -4.28396434e-01 -9.59316254e-01
3.57229039e-02 -4.40211557e-02 -3.30600947e-01 4.12448198e-02
1.15835853e-01 9.23067331e-01 -1.57202876e+00 7.94064552e-02
9.99876261e-01 1.42860031e+00 -2.81006813e-01 4.68195118e-02
1.85023844e-01 4.12848651e-01 -5.70039988e-01 -5.42556345e-01
2.90656447e-01 -5.04778028e-01 -5.81367791e-01 2.29324415e-01
4.43747044e-01 -9.75805521e-02 -1.45453528e-01 9.61204946e-01
6.79519534e-01 1.69460252e-01 8.82858694e-01 -1.57584822e+00
-6.94812179e-01 3.22724253e-01 -6.29555285e-01 2.35382412e-02
-6.76931962e-02 -5.66336252e-02 1.14991045e+00 -8.20319712e-01
1.50080875e-01 1.04839087e+00 7.76088476e-01 5.59222460e-01
-1.14870799e+00 -5.37644923e-01 1.63777798e-01 2.03742087e-01
-1.46708894e+00 -4.67590332e-01 1.14654946e+00 -5.69240570e-01
4.20831025e-01 5.11656642e-01 5.20733714e-01 8.14211309e-01
-7.17138872e-02 9.72526133e-01 9.02584374e-01 -6.30792201e-01
1.70298144e-01 4.95828092e-01 2.12677062e-01 6.26107752e-01
3.29122007e-01 1.67001531e-01 -4.20307100e-01 -1.94778413e-01
6.20639443e-01 1.31694647e-03 -5.36579370e-01 -7.10748374e-01
-6.44384086e-01 1.10879481e+00 3.85466754e-01 3.74723911e-01
-7.49366507e-02 -1.58395126e-01 5.22287488e-01 1.30910814e-01
6.77094281e-01 2.68919587e-01 -2.37454578e-01 1.62671804e-01
-9.46328461e-01 1.46055192e-01 4.92783219e-01 5.00627398e-01
7.03070223e-01 -2.29805354e-02 -4.52166498e-01 1.05188620e+00
4.37565893e-01 1.95262879e-01 4.89323616e-01 -6.80580497e-01
4.60120201e-01 7.17339873e-01 -1.10953726e-01 -1.18192303e+00
-4.29511487e-01 -6.63698494e-01 -1.02597451e+00 4.61042464e-01
6.85121179e-01 -9.63280443e-03 -4.81094837e-01 1.89708436e+00
3.08510244e-01 -2.97816336e-01 -2.13949069e-01 8.46883655e-01
5.39756060e-01 1.92570254e-01 -6.10170551e-02 -2.96651632e-01
1.10509217e+00 -7.09333658e-01 -6.16542101e-01 -9.96242389e-02
6.53379977e-01 -5.35292864e-01 1.22600198e+00 1.02420643e-01
-9.98895228e-01 -5.94671905e-01 -1.11435676e+00 6.94642588e-02
-3.13110232e-01 4.84920502e-01 6.41677737e-01 8.50404084e-01
-6.90496027e-01 5.88079631e-01 -5.70544422e-01 -1.38542101e-01
7.02306986e-01 4.10047412e-01 -4.17243809e-01 1.53739959e-01
-9.84410048e-01 7.11148143e-01 2.11832047e-01 4.55088615e-01
-1.12464853e-01 -7.24309683e-01 -9.69266891e-01 2.47383863e-01
7.76503831e-02 -4.80012000e-01 6.86715186e-01 -1.18614697e+00
-1.33966839e+00 1.07964897e+00 -3.29401754e-02 -3.25323045e-01
6.67871416e-01 1.24556327e-03 -1.57792777e-01 -5.09219766e-02
-2.46547490e-01 3.80462170e-01 8.34744155e-01 -1.10129499e+00
-2.51597136e-01 -6.30383968e-01 -1.70255620e-02 -3.20987217e-02
-9.67967153e-01 -8.70459750e-02 1.57374218e-02 -8.97037268e-01
-9.11292806e-02 -5.44504404e-01 4.22964478e-03 5.48114419e-01
-5.76155961e-01 -4.15991545e-01 6.61320865e-01 -5.34431994e-01
1.27625048e+00 -2.40117931e+00 4.87123877e-02 4.04752523e-01
2.36546576e-01 4.89278376e-01 -2.67753042e-02 1.56397268e-01
-2.22269461e-01 -4.94157225e-02 -5.04909217e-01 -3.93498451e-01
2.61937290e-01 -1.02545016e-01 -2.94101834e-01 7.54146874e-01
4.46083963e-01 8.04721534e-01 -5.25600493e-01 -3.94956857e-01
2.28568420e-01 6.93761587e-01 -7.31074095e-01 5.11665531e-02
2.11162671e-01 1.86633125e-01 -4.34657842e-01 5.59724391e-01
8.86618197e-01 -1.50158092e-01 -3.24485712e-02 -3.93476486e-01
1.14008106e-01 -3.74325663e-02 -1.25178957e+00 1.14499784e+00
-3.39914322e-01 6.47148132e-01 8.67595822e-02 -1.61187804e+00
9.99222636e-01 -6.12534545e-02 5.68962872e-01 -6.52503669e-01
2.82375872e-01 3.00418526e-01 -1.85564727e-01 -4.20777380e-01
1.48889586e-01 -3.50235313e-01 1.13250285e-01 2.52162457e-01
-1.49409696e-01 3.49912554e-01 3.48447100e-03 -2.93079913e-01
4.09340531e-01 -7.95920789e-02 3.20905685e-01 -3.20145279e-01
7.38102496e-01 -6.99985385e-01 6.86893523e-01 3.70473474e-01
-5.03088236e-01 6.12361610e-01 6.55125320e-01 -3.94479305e-01
-7.12150395e-01 -9.14205253e-01 -7.14894533e-01 9.08081889e-01
8.74034539e-02 -1.57809615e-01 -5.99594653e-01 -9.46294546e-01
3.29016536e-01 3.68727475e-01 -6.73475027e-01 -4.71232682e-01
-4.68096882e-01 -1.06309950e+00 5.06852150e-01 6.45543575e-01
3.97257000e-01 -9.12065923e-01 -4.46253717e-01 -6.76336959e-02
-7.62734935e-02 -7.75606871e-01 -5.39653540e-01 3.20756406e-01
-6.80622995e-01 -1.26694810e+00 -8.22323501e-01 -8.02405596e-01
8.94763649e-01 -7.85287768e-02 8.42842519e-01 1.71192273e-01
-4.14863259e-01 2.96357810e-01 -1.73133120e-01 -2.41195261e-01
1.05612807e-03 3.48198079e-02 1.16264850e-01 3.40452522e-01
2.97937036e-01 -4.14107561e-01 -6.92537606e-01 4.07040983e-01
-9.78857994e-01 -4.64322090e-01 3.76800746e-01 1.07472551e+00
4.69085425e-01 6.53384104e-02 9.37119603e-01 -5.74534953e-01
5.80018163e-01 -1.21701211e-01 -4.60874885e-01 3.83853793e-01
-7.51530886e-01 -1.81108378e-02 7.80549288e-01 -4.13257480e-01
-8.01891983e-01 -1.65005147e-01 -2.26242438e-01 -1.78356811e-01
7.29898438e-02 4.53196734e-01 -5.88419795e-01 -2.42455855e-01
5.62326968e-01 2.67760307e-01 1.54162854e-01 -3.45907420e-01
3.42889309e-01 7.87003636e-01 1.32475778e-01 -5.45033991e-01
6.40833914e-01 5.29221237e-01 9.20204222e-02 -9.69726443e-01
-7.58349419e-01 -2.55768836e-01 -6.08259201e-01 -4.77589257e-02
4.84668732e-01 -5.34005404e-01 -7.80778170e-01 6.26087964e-01
-8.06905031e-01 -3.54019642e-01 -7.03739882e-01 4.40732181e-01
-5.81413567e-01 5.79878509e-01 -3.89198691e-01 -9.09323394e-01
-3.79204124e-01 -1.09749520e+00 9.58749771e-01 3.61625433e-01
-2.07946032e-01 -1.18842888e+00 -1.45857111e-01 3.21104586e-01
3.72490883e-01 3.03754091e-01 9.84171569e-01 -6.40040100e-01
-1.25950694e-01 -2.85175264e-01 -4.10072178e-01 8.45865786e-01
2.93221146e-01 -4.48996276e-02 -1.00185347e+00 -5.84549487e-01
2.27723853e-03 -5.74064195e-01 1.08208585e+00 5.75494647e-01
1.52816927e+00 -2.71484762e-01 -2.78008193e-01 8.84832680e-01
1.14314508e+00 -4.28634807e-02 6.55678093e-01 2.14074552e-01
5.01603603e-01 9.14046586e-01 3.91979098e-01 6.37945473e-01
1.23482332e-01 9.06237900e-01 3.38560462e-01 -2.97606617e-01
-1.18268885e-01 1.38048446e-02 1.05948702e-01 4.17018890e-01
1.51971638e-01 -1.13421321e-01 -4.04295176e-01 3.36150974e-01
-1.72306252e+00 -8.64647090e-01 1.70067251e-01 2.60228586e+00
8.33050728e-01 -3.53465676e-02 3.60196054e-01 5.54716408e-01
7.22242236e-01 1.42248318e-01 -5.52163839e-01 -2.72225022e-01
-3.09401304e-01 9.38711986e-02 1.32760286e-01 5.04099846e-01
-1.32153845e+00 4.70597416e-01 5.58337593e+00 1.12503517e+00
-1.31503248e+00 -2.58531988e-01 9.70952332e-01 -1.57341719e-01
-2.29887247e-01 -3.64009410e-01 -1.01745260e+00 6.46857798e-01
3.87151688e-01 -6.84645176e-02 7.15960115e-02 8.44558775e-01
2.57636666e-01 2.02287883e-01 -1.15321410e+00 1.03795266e+00
2.24204943e-01 -1.05017316e+00 -9.81822461e-02 2.50312418e-01
4.47094440e-01 -6.19873464e-01 3.29552144e-01 1.24428593e-01
-5.24743795e-01 -1.07274187e+00 6.74852610e-01 4.14093077e-01
7.06336260e-01 -1.04168665e+00 6.16452932e-01 1.81987554e-01
-1.04918289e+00 -2.46430099e-01 -4.47178960e-01 -3.25474925e-02
-4.00322936e-02 1.02422285e+00 -2.74800271e-01 4.10253167e-01
4.25964981e-01 6.92520022e-01 -4.56753403e-01 1.39367700e+00
3.07440106e-03 3.22305202e-01 -2.65178472e-01 8.02754462e-02
-7.33272135e-02 -3.55172336e-01 4.44232106e-01 1.20573187e+00
9.54224765e-02 -3.71216714e-01 9.08578187e-02 9.29222107e-01
-2.11756781e-01 2.75440633e-01 -3.85059118e-01 1.61239237e-01
4.08465922e-01 1.15748703e+00 -6.51994586e-01 1.15402211e-02
-2.72758365e-01 8.57870758e-01 5.49441993e-01 2.94583857e-01
-8.52346241e-01 -6.85984075e-01 8.72927904e-01 2.59372234e-01
2.68164963e-01 4.08314541e-02 -5.81498682e-01 -1.03726947e+00
4.61107880e-01 -6.53947830e-01 5.16263545e-01 -2.65933201e-02
-1.53386962e+00 4.35146451e-01 -5.28941378e-02 -1.15398920e+00
-1.16975326e-02 -7.99960136e-01 -6.57241940e-01 8.63423526e-01
-1.81909180e+00 -1.16123199e+00 -2.61973470e-01 4.41679627e-01
1.09318011e-01 -1.01222605e-01 5.67327082e-01 6.70002639e-01
-8.66544068e-01 1.28916609e+00 3.11536998e-01 1.81280509e-01
7.44440079e-01 -1.08181810e+00 -3.39482099e-01 6.54083490e-01
-2.11204402e-02 5.53652525e-01 5.03397286e-01 -1.04095489e-01
-9.47272301e-01 -1.02154565e+00 8.87302577e-01 -7.35106468e-02
3.45327258e-01 -3.30895543e-01 -9.43194687e-01 3.51799428e-01
-3.90316784e-01 4.21986639e-01 9.93750036e-01 -7.26701133e-03
-3.87647301e-01 -4.79599953e-01 -1.36032045e+00 3.88004959e-01
8.70163321e-01 -4.20450151e-01 -1.48864418e-01 3.98759454e-01
2.54927903e-01 -6.24781996e-02 -8.55553925e-01 6.85907900e-01
7.99623728e-01 -1.06267893e+00 1.17310071e+00 -5.87273300e-01
2.89951950e-01 -1.90722018e-01 -1.66714609e-01 -1.13052511e+00
-2.80850619e-01 -2.40385801e-01 -1.49390668e-01 1.44438732e+00
2.12024033e-01 -8.47618937e-01 9.45286214e-01 6.80235863e-01
-8.14637467e-02 -1.33201349e+00 -1.14193392e+00 -9.02639270e-01
3.36058825e-01 -3.41596484e-01 3.80788654e-01 9.58961189e-01
8.99408832e-02 -1.88821703e-01 -4.11266923e-01 -1.90186024e-01
6.07685030e-01 1.81420818e-01 4.66722816e-01 -1.32123077e+00
-2.78243005e-01 -8.44177961e-01 -5.87891757e-01 -9.54656363e-01
4.47486281e-01 -9.43592370e-01 -1.88866600e-01 -1.19182849e+00
2.96716541e-01 -6.75879300e-01 -4.50971663e-01 5.67656100e-01
-2.00308904e-01 4.18237746e-01 9.93341506e-02 2.18041707e-02
-2.41940632e-01 1.02506995e+00 1.15405786e+00 -2.44846135e-01
-2.86028385e-01 3.48545820e-01 -1.03945971e+00 7.82122910e-01
6.94956005e-01 -2.17480138e-01 -3.67807388e-01 -4.73035350e-02
1.40016109e-01 -5.05271673e-01 5.09130716e-01 -8.05197477e-01
-6.31597564e-02 -2.34997824e-01 5.78131557e-01 -3.19732964e-01
3.50684166e-01 -7.72958755e-01 -4.72797751e-01 3.59063953e-01
-4.71564353e-01 -4.93471980e-01 1.53289484e-02 5.17110705e-01
-3.30885053e-01 -2.25361183e-01 1.14815247e+00 4.17640865e-01
-2.62531966e-01 3.18961889e-01 -1.07699735e-02 1.17482528e-01
1.10674465e+00 -5.24042964e-01 -1.61003172e-01 -1.54143944e-01
-4.89134163e-01 6.95360601e-02 2.49678507e-01 3.29319835e-01
6.99640930e-01 -1.56266284e+00 -6.92921638e-01 5.36377192e-01
2.04968438e-01 -2.15940103e-01 1.34204909e-01 9.36673701e-01
-1.44919693e-01 4.23512876e-01 -1.27797902e-01 -5.16683102e-01
-1.30318964e+00 3.53160828e-01 6.20472789e-01 -2.45730564e-01
-4.29204524e-01 9.40723717e-01 3.62990081e-01 -3.65259081e-01
5.15140176e-01 -2.66556889e-01 -2.55933017e-01 2.68014550e-01
6.57171905e-01 5.41392624e-01 2.26786077e-01 -6.80586457e-01
-5.41827261e-01 6.56524897e-01 6.62558898e-02 3.98308218e-01
1.30877340e+00 -6.34103194e-02 -1.33411333e-01 1.44536510e-01
1.48408949e+00 -2.13235654e-02 -1.37559175e+00 -1.39744103e-01
-1.88918531e-01 -6.06860995e-01 -3.10407740e-05 -5.53312600e-01
-1.31938040e+00 9.06358719e-01 7.51040220e-01 2.83971488e-01
1.30039585e+00 -4.74117585e-02 6.57820463e-01 8.95721689e-02
7.09412396e-02 -1.04543447e+00 2.04771623e-01 2.21520185e-01
9.67713654e-01 -1.28990352e+00 -1.19673379e-03 -4.28420007e-01
-2.83306658e-01 1.25912237e+00 6.37369812e-01 5.04833125e-02
8.10127676e-01 1.11444607e-01 -1.85201973e-01 2.95059552e-04
-4.76709679e-02 -8.44166949e-02 6.61738455e-01 6.48896635e-01
6.59573793e-01 3.17707881e-02 -5.14301300e-01 8.56753469e-01
-4.80840653e-02 2.33660149e-03 -1.59380641e-02 6.80910468e-01
-3.24195564e-01 -1.18206072e+00 -3.45212460e-01 5.35973787e-01
-3.70259523e-01 1.22168511e-01 -2.64833659e-01 6.98659360e-01
2.87475735e-01 8.02036285e-01 1.22876845e-01 -1.36321962e-01
3.16810071e-01 3.67990918e-02 6.21901393e-01 -2.74129361e-01
-3.40211034e-01 -2.41354406e-01 -3.05068433e-01 -1.96075991e-01
-3.36657852e-01 -4.73584801e-01 -9.02966797e-01 -1.54297754e-01
-6.49910629e-01 6.04910068e-02 4.84389573e-01 1.01363218e+00
2.79183239e-01 2.80175686e-01 8.24359238e-01 -7.81944036e-01
-8.92692864e-01 -7.88436055e-01 -7.73562431e-01 5.65687418e-01
4.35039401e-01 -8.27737331e-01 -6.75651968e-01 -2.15725794e-01] | [9.208586692810059, 3.651777982711792] |
63b63b41-a28b-4e87-ad71-565deab098ad | grounding-language-models-to-images-for | 2301.13823 | null | https://arxiv.org/abs/2301.13823v4 | https://arxiv.org/pdf/2301.13823v4.pdf | Grounding Language Models to Images for Multimodal Inputs and Outputs | We propose an efficient method to ground pretrained text-only language models to the visual domain, enabling them to process arbitrarily interleaved image-and-text data, and generate text interleaved with retrieved images. Our method leverages the abilities of language models learnt from large scale text-only pretraining, such as in-context learning and free-form text generation. We keep the language model frozen, and finetune input and output linear layers to enable cross-modality interactions. This allows our model to process arbitrarily interleaved image-and-text inputs, and generate free-form text interleaved with retrieved images. We achieve strong zero-shot performance on grounded tasks such as contextual image retrieval and multimodal dialogue, and showcase compelling interactive abilities. Our approach works with any off-the-shelf language model and paves the way towards an effective, general solution for leveraging pretrained language models in visually grounded settings. | ['Daniel Fried', 'Ruslan Salakhutdinov', 'Jing Yu Koh'] | 2023-01-31 | null | null | null | null | ['multimodal-generation'] | ['natural-language-processing'] | [ 3.75752479e-01 3.26516509e-01 5.87884299e-02 -2.12295890e-01
-1.14327824e+00 -9.06956851e-01 1.06081891e+00 -4.91165370e-02
-5.87786615e-01 4.32604998e-01 4.58703190e-01 -5.20113170e-01
4.32762504e-01 -7.11993039e-01 -1.04588854e+00 -3.44659001e-01
8.15425627e-03 5.40131867e-01 1.37073338e-01 -2.57294238e-01
6.06712066e-02 3.50136518e-01 -1.65458918e+00 1.13551795e+00
3.65950167e-01 5.33325553e-01 6.16983473e-01 1.13848794e+00
-3.98153067e-01 1.08180010e+00 -2.01673955e-01 -3.43892664e-01
5.36017492e-02 -4.77293700e-01 -8.69764447e-01 1.89062014e-01
6.65004015e-01 -8.72191906e-01 -5.32133043e-01 2.24966154e-01
5.47994912e-01 1.81104168e-01 7.05774844e-01 -1.11759913e+00
-1.17239797e+00 5.67098379e-01 -3.61955106e-01 -2.76130408e-01
7.32349992e-01 7.39103377e-01 9.12910581e-01 -1.28921545e+00
1.02008593e+00 1.58773267e+00 2.18984857e-01 9.04165745e-01
-1.47840691e+00 -5.11206746e-01 5.45747161e-01 -2.73063868e-01
-1.03389835e+00 -8.76175165e-01 2.34938681e-01 -3.95867676e-01
1.27136731e+00 1.45238014e-02 7.69341707e-01 1.49412680e+00
-1.03061557e-01 1.39691556e+00 8.88853014e-01 -5.93131363e-01
-3.19764614e-02 1.41182035e-01 -4.41489637e-01 1.01754081e+00
-2.16876909e-01 8.28582793e-02 -9.83606398e-01 1.19597934e-01
1.10089624e+00 1.78700071e-02 -9.54929963e-02 -6.10595703e-01
-1.55381310e+00 5.53542614e-01 4.79543626e-01 2.91651245e-02
-7.56580681e-02 5.14946520e-01 2.87562042e-01 2.99093306e-01
2.45927170e-01 4.37999755e-01 -1.40820593e-02 1.39047518e-01
-1.14056814e+00 2.41007775e-01 6.02826416e-01 1.22456717e+00
7.24761784e-01 -8.53346139e-02 -5.69243193e-01 7.38922775e-01
3.76693040e-01 9.35083151e-01 2.52031565e-01 -1.01018882e+00
5.19443333e-01 3.39260042e-01 2.37472013e-01 -6.02761388e-01
-1.34474173e-01 2.11483017e-01 -5.14200807e-01 2.97504157e-01
3.62411499e-01 -6.38626441e-02 -1.27059543e+00 1.71257532e+00
7.23182559e-02 -1.87444612e-01 1.51795417e-01 9.12463009e-01
1.04825628e+00 8.95219505e-01 4.12677705e-01 2.39518628e-01
1.32851136e+00 -1.11678588e+00 -3.44814479e-01 -4.31740612e-01
6.90431118e-01 -7.18372583e-01 1.85190630e+00 3.26639533e-01
-1.48789227e+00 -4.33111995e-01 -7.37227082e-01 -7.11529374e-01
-3.70175511e-01 8.61262754e-02 5.91745198e-01 3.77157889e-02
-1.57305050e+00 1.72441959e-01 -7.73555517e-01 -5.69427550e-01
4.16687369e-01 1.01477429e-01 -6.19550645e-01 -1.96081057e-01
-6.95834219e-01 6.08759046e-01 2.68700480e-01 7.69496802e-03
-1.41329598e+00 -4.70079422e-01 -9.42735970e-01 -1.20183110e-01
5.11120558e-01 -1.24053943e+00 1.52574086e+00 -1.09089923e+00
-1.59909463e+00 1.34633732e+00 -1.39140397e-01 -2.34610960e-01
6.70297921e-01 -2.50023872e-01 -2.22670753e-03 6.18332028e-01
-3.12397815e-02 1.46281075e+00 9.70009148e-01 -1.65664816e+00
-2.33032182e-01 -1.38084084e-01 2.22228259e-01 4.94432747e-01
-3.03796828e-01 -1.27047926e-01 -8.44685137e-01 -5.89962244e-01
-2.47081012e-01 -8.61858726e-01 -7.95916915e-02 5.29258609e-01
-2.89664388e-01 1.06630817e-01 7.11364746e-01 -3.93798977e-01
7.35757232e-01 -2.03330994e+00 2.17442796e-01 -1.19692057e-01
2.96087474e-01 -9.24816951e-02 -7.27328837e-01 7.98939705e-01
1.73125520e-01 1.58115596e-01 3.27201560e-02 -8.22299719e-01
1.36739746e-01 1.56869501e-01 -6.93225682e-01 -8.70156065e-02
3.08726043e-01 1.54550672e+00 -1.01102340e+00 -7.75102258e-01
3.83719057e-01 4.79807854e-01 -6.78474844e-01 6.01333678e-01
-8.40743780e-01 3.44878495e-01 -4.15209174e-01 4.55473959e-01
2.24800318e-01 -5.65624356e-01 1.75891250e-01 -2.20151797e-01
5.01847081e-02 3.71752195e-02 -6.39806032e-01 2.27895069e+00
-9.65793192e-01 8.40436518e-01 1.97995335e-01 -3.82514745e-01
4.26089525e-01 3.29831809e-01 7.38547966e-02 -1.05797303e+00
-1.87741339e-01 -1.83808178e-01 -5.35870850e-01 -8.16520393e-01
8.73823047e-01 -1.37961522e-01 -1.77962497e-01 8.98958623e-01
3.65641385e-01 -5.60396969e-01 -8.47835280e-03 9.26034689e-01
7.98409700e-01 7.06613541e-01 -2.23452106e-01 2.44793117e-01
-4.40091752e-02 -6.22325465e-02 -6.38225436e-01 1.11118627e+00
3.44913542e-01 8.24225187e-01 3.68549883e-01 -2.89455503e-01
-1.30030978e+00 -1.41109014e+00 4.17560309e-01 1.77263176e+00
1.31752104e-01 -4.50308710e-01 -4.64801729e-01 -5.57957590e-01
-1.61005288e-01 6.23153210e-01 -5.62450290e-01 -4.12640348e-02
-4.31416422e-01 -1.27658769e-01 6.07913852e-01 5.52466631e-01
3.25885057e-01 -1.30432403e+00 -5.87071478e-01 1.11342736e-01
-2.79824197e-01 -1.24366415e+00 -6.55053377e-01 -8.60961750e-02
-7.95353532e-01 -7.13417649e-01 -9.54822361e-01 -9.06880200e-01
7.66530812e-01 4.29524988e-01 1.56309152e+00 2.57400870e-01
-4.32780176e-01 9.57778633e-01 -2.61477381e-01 4.15043198e-02
-5.26789606e-01 -8.24010223e-02 -4.64683652e-01 -9.53598693e-02
-3.47080767e-01 -4.02792633e-01 -8.29660594e-01 1.10287227e-01
-1.36101234e+00 7.44308650e-01 7.79911399e-01 8.84251893e-01
5.21618783e-01 -8.70553315e-01 3.74356627e-01 -7.66902924e-01
6.96202695e-01 -2.68616766e-01 -3.79071355e-01 5.67981064e-01
-8.19260255e-02 3.92865688e-01 4.23622608e-01 -6.40326798e-01
-1.04254723e+00 7.06761181e-02 4.43259329e-02 -5.04017889e-01
-1.33819655e-01 2.65094370e-01 -9.43645090e-02 1.14557922e-01
8.09105337e-01 3.43562484e-01 -1.15841024e-01 -1.30945861e-01
1.22973406e+00 4.67175961e-01 7.62803972e-01 -8.78928363e-01
6.59479201e-01 6.39362454e-01 -3.58395398e-01 -7.36260951e-01
-5.56839645e-01 -1.91147178e-01 -3.96497697e-01 -2.32167482e-01
1.05746233e+00 -1.20931339e+00 -7.68441319e-01 2.97605366e-01
-1.18470573e+00 -1.08780551e+00 -2.42248014e-01 -1.07908525e-01
-8.00776064e-01 1.25235051e-01 -6.46654427e-01 -8.91566694e-01
-3.61043304e-01 -9.43682790e-01 1.66589260e+00 4.33767699e-02
-1.77949458e-01 -9.22942817e-01 -2.36595765e-01 3.08644116e-01
2.93369383e-01 -9.45023913e-03 8.46324801e-01 -2.39706710e-01
-1.05728936e+00 6.15142025e-02 -4.02422905e-01 -2.29500055e-01
-3.58234227e-01 1.40213210e-03 -1.05877697e+00 -3.64547580e-01
-7.30036080e-01 -1.17951584e+00 1.03187573e+00 -1.00677453e-01
1.14315557e+00 -4.93746102e-01 -4.65252638e-01 5.88379860e-01
1.08087981e+00 -2.76106507e-01 6.06755376e-01 4.22974676e-02
7.54904866e-01 4.10801023e-01 2.42787614e-01 1.27920046e-01
6.92525208e-01 5.67212462e-01 2.66340137e-01 -5.03062189e-01
-3.55878472e-01 -8.87086928e-01 3.11855793e-01 2.50313818e-01
2.38132581e-01 -6.44029140e-01 -9.18173194e-01 7.90246606e-01
-1.88218224e+00 -1.16696608e+00 4.58743542e-01 2.01876163e+00
9.82353091e-01 2.07769242e-03 6.46768883e-02 -6.19682372e-01
2.15728208e-01 2.09373832e-01 -5.63499808e-01 -3.11090112e-01
-1.65767089e-01 1.75821885e-01 7.07826391e-02 7.02010036e-01
-8.74339044e-01 1.43273699e+00 6.80542326e+00 5.48481584e-01
-1.35612381e+00 -6.43946277e-03 6.72813594e-01 -6.34259999e-01
-7.57632613e-01 -2.22575068e-02 -3.62881929e-01 1.49903549e-02
7.77290761e-01 1.56260684e-01 6.21152401e-01 5.29026151e-01
2.53303677e-01 -2.93498516e-01 -1.42187178e+00 1.09276152e+00
1.95715681e-01 -1.44950294e+00 6.34999216e-01 -4.59714830e-02
6.44219398e-01 1.56685468e-02 3.02877873e-01 4.45173472e-01
7.28394866e-01 -1.40373468e+00 9.17743564e-01 6.63274586e-01
1.29746735e+00 -2.66151875e-01 -4.12592106e-02 2.82788485e-01
-9.87914562e-01 1.60556997e-03 8.06710348e-02 1.37814328e-01
3.51881653e-01 -6.56524748e-02 -1.01316142e+00 2.28736669e-01
5.72490990e-01 5.17128527e-01 -6.31234527e-01 5.35931468e-01
-1.09304056e-01 1.68435082e-01 -1.95891246e-01 -1.04436442e-01
3.15188676e-01 2.26859033e-01 2.27285638e-01 1.41767907e+00
1.57619596e-01 1.55220956e-01 3.46012950e-01 8.74384344e-01
-2.70840228e-01 9.75917429e-02 -1.02194202e+00 -4.06742454e-01
4.13601667e-01 1.22327578e+00 -6.65465176e-01 -7.27544069e-01
-4.13987905e-01 1.32512188e+00 4.77007926e-01 8.74475002e-01
-4.90296125e-01 -1.29066542e-01 1.94124922e-01 3.48064274e-01
2.68074930e-01 -5.38614988e-01 -6.12674989e-02 -1.44462621e+00
-1.33082718e-01 -1.09795535e+00 2.50130534e-01 -1.69647706e+00
-1.09052491e+00 6.98833823e-01 1.29665419e-01 -9.53556597e-01
-6.37534797e-01 -6.03636563e-01 -3.61998051e-01 7.23584712e-01
-1.22984874e+00 -1.77656305e+00 -5.44610918e-01 9.03606355e-01
6.24675035e-01 1.47845885e-02 8.59159291e-01 -2.11304054e-01
8.75932798e-02 5.99780917e-01 -2.62916416e-01 1.23861991e-01
9.88244414e-01 -9.93955016e-01 5.40197194e-01 5.60240567e-01
5.74198484e-01 7.03844368e-01 4.77039069e-01 -3.68977696e-01
-1.74588835e+00 -8.13988268e-01 3.88875127e-01 -7.59075522e-01
6.96819961e-01 -9.89941776e-01 -6.23183370e-01 8.08312953e-01
7.00839341e-01 -3.80508378e-02 5.20078599e-01 4.22927625e-02
-6.37893260e-01 1.90692961e-01 -7.68408239e-01 1.23154402e+00
1.34075141e+00 -1.21636748e+00 -5.47149122e-01 3.03333789e-01
9.56692517e-01 -5.79307199e-01 -3.08114231e-01 -1.30195264e-02
8.59936655e-01 -7.05329537e-01 1.26227891e+00 -7.54174232e-01
8.15869510e-01 -9.32109803e-02 -4.77704294e-02 -9.37566459e-01
3.33525464e-02 -8.86835992e-01 -1.24121405e-01 9.37063873e-01
5.89420557e-01 -1.95121348e-01 6.16589069e-01 7.15706289e-01
-8.82819854e-03 -4.55491513e-01 -3.53849292e-01 -2.83896327e-01
1.34820923e-01 -4.02600050e-01 2.94178218e-01 6.62085056e-01
7.71182552e-02 6.22284412e-01 -3.44185233e-01 -1.29317895e-01
3.81787509e-01 1.56179219e-01 1.14310837e+00 -5.63749790e-01
-5.95826507e-01 -3.23154926e-01 1.23239495e-01 -1.50113702e+00
2.30823219e-01 -9.78484929e-01 2.30625510e-01 -1.74374127e+00
3.35424602e-01 -3.48638237e-01 9.43796262e-02 8.17708671e-01
-1.22790717e-01 6.09315276e-01 6.77487969e-01 2.00965926e-01
-9.83754158e-01 4.22671735e-01 1.62683988e+00 -1.79217964e-01
-2.81127334e-01 -6.90813661e-01 -8.04940522e-01 4.85898405e-01
3.17591548e-01 -8.93017463e-03 -8.17011058e-01 -9.35030162e-01
3.86054516e-01 4.11222100e-01 9.71922874e-01 -7.08701015e-01
2.69484580e-01 -3.14410776e-01 6.49995029e-01 -4.53219205e-01
5.29957294e-01 -5.77844739e-01 -1.38457417e-01 -2.20558554e-01
-8.87665927e-01 -2.40042619e-02 4.71015841e-01 5.82667768e-01
1.06207967e-01 2.09481984e-01 3.47191781e-01 -4.83876705e-01
-6.43352449e-01 2.31993601e-01 -4.63168591e-01 2.60311782e-01
6.32632792e-01 -1.70403987e-01 -4.35472667e-01 -8.71062756e-01
-1.11068916e+00 4.22858834e-01 8.39260578e-01 5.77345371e-01
8.42644691e-01 -1.18242168e+00 -5.88614225e-01 1.87988609e-01
4.08828408e-01 3.30670625e-02 2.98388630e-01 2.41111547e-01
-4.16535437e-01 3.71224254e-01 -4.39053290e-02 -8.34738910e-01
-1.15774119e+00 8.10519278e-01 2.14264587e-01 -1.66664511e-01
-6.94225073e-01 8.82615983e-01 7.31751800e-01 -3.01941842e-01
1.82205200e-01 -3.05377424e-01 3.73835653e-01 -1.98501185e-01
5.47389984e-01 -5.29113591e-01 -2.71557450e-01 -1.71035498e-01
-4.52071913e-02 5.36095381e-01 -1.21690750e-01 -9.43101406e-01
1.10644114e+00 -3.63900334e-01 1.79144785e-01 5.31861246e-01
1.15436828e+00 1.90759655e-02 -1.60045600e+00 -2.24796981e-01
-3.83889586e-01 -2.20157728e-01 -1.26065791e-01 -1.19140184e+00
-5.96162081e-01 1.18523490e+00 3.12338382e-01 -1.77673414e-01
1.04307866e+00 2.83062816e-01 6.34546161e-01 8.74340892e-01
4.10393894e-01 -8.98610413e-01 9.34693813e-01 4.43304807e-01
1.14030206e+00 -1.28354168e+00 -1.75462276e-01 1.52607951e-02
-9.11334097e-01 9.18320239e-01 6.29682600e-01 -5.07541820e-02
8.72590914e-02 5.80317855e-01 1.91214591e-01 -2.88882591e-02
-1.35268700e+00 -2.08747566e-01 3.36113423e-01 7.01087594e-01
6.61202192e-01 -1.64765209e-01 4.29253310e-01 1.94412455e-01
6.91735046e-03 1.86195999e-01 3.58248383e-01 9.50481594e-01
-2.37760633e-01 -1.09420264e+00 -3.01929384e-01 5.39279953e-02
8.10572430e-02 -6.25997484e-01 -6.37532473e-01 7.85032094e-01
-3.58597934e-01 8.14503312e-01 1.39620468e-01 -8.09934437e-02
1.23139903e-01 3.55715066e-01 8.11712265e-01 -7.24754453e-01
-5.18422663e-01 1.42423660e-01 1.93334773e-01 -8.14121783e-01
-1.92771405e-01 -2.50722408e-01 -1.37755990e+00 -1.20986737e-01
2.35016614e-01 -2.26933882e-01 5.24396300e-01 8.63723814e-01
6.85899615e-01 3.15183580e-01 1.84917152e-01 -1.27988541e+00
9.20044538e-03 -6.59328878e-01 6.11745156e-02 5.85456431e-01
6.93260610e-01 -1.83700144e-01 5.05984314e-02 5.58922946e-01] | [10.93132495880127, 1.6032183170318604] |
5732a607-902a-4de2-807d-adb28deb877c | contextual-transformer-for-offline-meta | 2211.08016 | null | https://arxiv.org/abs/2211.08016v1 | https://arxiv.org/pdf/2211.08016v1.pdf | Contextual Transformer for Offline Meta Reinforcement Learning | The pretrain-finetuning paradigm in large-scale sequence models has made significant progress in natural language processing and computer vision tasks. However, such a paradigm is still hindered by several challenges in Reinforcement Learning (RL), including the lack of self-supervised pretraining algorithms based on offline data and efficient fine-tuning/prompt-tuning over unseen downstream tasks. In this work, we explore how prompts can improve sequence modeling-based offline reinforcement learning (offline-RL) algorithms. Firstly, we propose prompt tuning for offline RL, where a context vector sequence is concatenated with the input to guide the conditional policy generation. As such, we can pretrain a model on the offline dataset with self-supervised loss and learn a prompt to guide the policy towards desired actions. Secondly, we extend our framework to Meta-RL settings and propose Contextual Meta Transformer (CMT); CMT leverages the context among different tasks as the prompt to improve generalization on unseen tasks. We conduct extensive experiments across three different offline-RL settings: offline single-agent RL on the D4RL dataset, offline Meta-RL on the MuJoCo benchmark, and offline MARL on the SMAC benchmark. Superior results validate the strong performance, and generality of our methods. | ['Yaodong Yang', 'Yali Du', 'Jun Wang', 'Haifeng Zhang', 'Xian Hong Wu Fung', 'Zhaowei Zhang', 'Xidong Feng', 'Ye Li', 'Runji Lin'] | 2022-11-15 | null | null | null | null | ['smac-1', 'smac', 'd4rl'] | ['playing-games', 'playing-games', 'robots'] | [ 3.79126519e-01 -2.07836658e-01 -3.70286345e-01 -3.90598625e-01
-7.43719816e-01 -8.28766346e-01 8.99160802e-01 -9.55249369e-03
-9.53990281e-01 8.39003623e-01 3.45643491e-01 -4.75528538e-01
1.83925137e-01 -4.15539742e-01 -1.00649762e+00 -6.87041521e-01
-3.18999588e-02 4.66740102e-01 1.20324932e-01 -4.73635375e-01
2.96077486e-02 1.40518039e-01 -1.14017379e+00 3.77509505e-01
8.25322986e-01 8.52931440e-01 5.91474712e-01 7.86432743e-01
8.44310448e-02 1.22415698e+00 -4.35807765e-01 -1.49060890e-01
3.35090190e-01 -5.41542232e-01 -8.41525435e-01 -1.88557148e-01
1.51077420e-01 -6.26782477e-01 -5.01461506e-01 7.38088787e-01
6.06358767e-01 5.28553545e-01 5.17728806e-01 -1.22608423e+00
-6.41170204e-01 8.93662155e-01 -4.38786000e-01 2.58691967e-01
2.92495973e-02 8.25792909e-01 1.06005359e+00 -5.29809892e-01
5.17450511e-01 1.52477634e+00 2.86751598e-01 9.04612005e-01
-1.27266562e+00 -5.85487425e-01 6.77961409e-01 3.91624123e-01
-7.27322578e-01 -4.68234122e-01 5.67838788e-01 -3.01120520e-01
1.28411651e+00 -3.23313475e-01 4.16655719e-01 1.78764343e+00
-2.48244760e-04 1.28915846e+00 1.21642685e+00 -4.01899576e-01
4.70948279e-01 -3.78348440e-01 -9.02241096e-02 8.05816293e-01
-4.55240369e-01 8.11300337e-01 -5.35376906e-01 -4.74343486e-02
6.81629598e-01 -8.10910165e-02 -1.20981760e-01 -2.79041290e-01
-1.36832845e+00 8.78610075e-01 3.23510289e-01 -1.00103743e-01
-2.66652256e-01 4.73630220e-01 7.58196533e-01 6.65383577e-01
2.11901158e-01 7.13635981e-01 -8.00618649e-01 -4.28155512e-01
-5.08154094e-01 4.22552794e-01 5.86713731e-01 9.34117496e-01
7.27918983e-01 3.21926355e-01 -6.44540310e-01 9.88136828e-01
-7.62102902e-02 4.00341004e-01 7.87176788e-01 -1.10477149e+00
8.00472140e-01 1.64866373e-01 3.91230397e-02 -2.99894184e-01
-3.14327896e-01 -3.86072457e-01 -4.76753503e-01 1.09272733e-01
3.40320498e-01 -4.51019168e-01 -1.11861646e+00 2.28871584e+00
1.85217485e-01 4.94416684e-01 2.28781998e-01 9.24583912e-01
3.06432575e-01 5.51879764e-01 4.11381274e-01 -1.66784823e-01
9.63967443e-01 -1.62623596e+00 -3.73233736e-01 -5.79844534e-01
8.67985427e-01 -1.98490366e-01 1.40819919e+00 4.26577896e-01
-7.83827782e-01 -7.04332769e-01 -8.45572293e-01 -2.93886214e-02
-7.98415914e-02 1.16608441e-01 5.29459774e-01 -5.60664833e-02
-9.05192673e-01 6.94811165e-01 -1.00537121e+00 -1.98709562e-01
3.66723686e-01 1.83101818e-01 -7.43521675e-02 -1.08130455e-01
-1.36369920e+00 8.75255585e-01 8.02991152e-01 -2.13099942e-02
-1.76307189e+00 -7.99727142e-01 -7.80007064e-01 -3.93460169e-02
9.24236536e-01 -7.09721923e-01 1.79427564e+00 -1.09149373e+00
-1.97095990e+00 4.83316362e-01 7.56882802e-02 -9.48396564e-01
7.15058088e-01 -4.33753461e-01 -1.67308524e-01 1.09560683e-01
-1.15986198e-01 9.58662927e-01 1.09685159e+00 -1.08925796e+00
-8.39547336e-01 -1.11414604e-01 2.17458203e-01 4.30564076e-01
4.56365803e-03 -3.40028524e-01 -2.42251337e-01 -7.96304286e-01
-7.28906512e-01 -1.13413656e+00 -4.38634545e-01 -5.52909791e-01
-1.34754151e-01 -4.45691466e-01 7.32755542e-01 -5.36437333e-01
1.03244197e+00 -2.06266665e+00 3.44028860e-01 -2.13760510e-01
7.75855791e-04 5.18775046e-01 -8.02574873e-01 4.90343034e-01
9.24301222e-02 -2.11348116e-01 -3.25971633e-01 -2.70877808e-01
2.20356092e-01 5.08206606e-01 -6.88851058e-01 1.33135334e-01
3.27871710e-01 1.30194306e+00 -1.21355319e+00 -1.90814063e-01
9.38890651e-02 4.12878133e-02 -7.95879364e-01 5.79984426e-01
-9.42497551e-01 7.27321386e-01 -5.49353719e-01 4.63980943e-01
-3.48750706e-04 -2.44808823e-01 2.89090514e-01 1.14741758e-01
1.01063944e-01 2.89719105e-01 -5.19483626e-01 2.05918503e+00
-6.92181349e-01 2.06640348e-01 -3.84814776e-02 -1.14196885e+00
7.80295670e-01 1.37206882e-01 2.94830978e-01 -1.05546141e+00
-1.67013649e-02 -1.43556213e-02 1.12783737e-01 -5.82572758e-01
4.12673891e-01 -1.34094477e-01 -4.97092083e-02 6.11112833e-01
2.96327770e-01 9.45404768e-02 2.61427879e-01 2.84470081e-01
1.32264674e+00 7.90049374e-01 2.47297704e-01 1.54913485e-01
3.19939315e-01 8.15813392e-02 5.97326636e-01 1.09740520e+00
-5.43108165e-01 3.60670276e-02 3.95395935e-01 -2.26572588e-01
-1.03041875e+00 -8.61937642e-01 4.97028798e-01 1.87515724e+00
-1.10042825e-01 -2.98264086e-01 -6.51196182e-01 -1.22898710e+00
2.56454706e-01 9.24172103e-01 -6.24866903e-01 -5.50996542e-01
-8.56444597e-01 -6.04682446e-01 6.20471418e-01 7.72530794e-01
3.73492211e-01 -1.58958805e+00 -5.86906075e-01 5.09538710e-01
-1.43063828e-01 -1.31497395e+00 -7.28087485e-01 5.87272286e-01
-8.18179131e-01 -9.00958538e-01 -5.03175914e-01 -7.52153516e-01
3.78644168e-01 7.42829442e-02 1.15963662e+00 -5.83083592e-02
-1.08434461e-01 5.37411809e-01 -4.96261984e-01 -7.08062062e-03
-6.41488314e-01 2.60057986e-01 2.74212182e-01 -1.58139303e-01
1.20337121e-01 -5.83077431e-01 -6.00220263e-01 2.52919406e-01
-7.11885214e-01 1.16574503e-01 7.87658334e-01 1.24846816e+00
5.42515934e-01 -4.50721353e-01 1.00284350e+00 -8.52472246e-01
7.45585501e-01 -5.94129980e-01 -6.83532715e-01 4.07714427e-01
-6.66148126e-01 5.71275294e-01 1.11514783e+00 -6.89888239e-01
-1.11994624e+00 2.68732980e-02 -1.33217856e-01 -7.70119131e-01
-1.37337804e-01 3.83204788e-01 4.54485863e-02 3.65031183e-01
6.26853406e-01 4.42850053e-01 1.37267143e-01 -5.28577626e-01
7.30042875e-01 3.64942193e-01 7.27845728e-01 -1.19216061e+00
5.40352464e-01 1.31007671e-01 -3.93383086e-01 -4.39128667e-01
-1.07305634e+00 -3.27631921e-01 -3.03287864e-01 1.09482817e-01
8.06623101e-01 -9.90440190e-01 -9.97441649e-01 3.95863563e-01
-8.03640008e-01 -1.41565681e+00 -2.71792591e-01 3.90262365e-01
-1.11151052e+00 1.45503774e-01 -7.65728176e-01 -6.03014052e-01
-4.06348109e-01 -1.38505292e+00 7.79925287e-01 -1.79989543e-02
7.31968284e-02 -1.15166593e+00 2.76772410e-01 2.69319773e-01
3.12695742e-01 -8.66911635e-02 1.05978525e+00 -7.27627158e-01
-2.91729897e-01 5.66639960e-01 9.53363180e-02 5.38629949e-01
5.27129555e-03 -4.60545272e-01 -8.60298991e-01 -6.67650640e-01
-2.27115795e-01 -1.30001867e+00 1.17673481e+00 5.27772717e-02
1.41273093e+00 -5.13540089e-01 -2.01560229e-01 7.60993361e-01
1.18261337e+00 2.52668858e-01 1.68754458e-01 5.86935043e-01
5.92014313e-01 2.57674813e-01 9.08498347e-01 4.59002912e-01
3.58032107e-01 6.32682800e-01 3.99485379e-01 3.12557817e-01
-1.59265056e-01 -9.76895392e-01 8.52053285e-01 6.88008487e-01
2.11958349e-01 -2.14781910e-01 -6.62634611e-01 2.93309540e-01
-2.18156338e+00 -1.04253697e+00 7.08180845e-01 1.89325452e+00
1.30738020e+00 -1.50443576e-02 3.61703902e-01 -3.95379871e-01
5.11093497e-01 2.55732775e-01 -1.21181071e+00 -3.08321446e-01
5.89841697e-03 1.81597218e-01 4.12940234e-01 5.90547204e-01
-1.17225194e+00 1.46632361e+00 5.62192392e+00 1.03458118e+00
-1.28691196e+00 1.55803978e-01 6.91813052e-01 -2.15581849e-01
-1.72754765e-01 -8.75423104e-03 -1.03907168e+00 4.44297880e-01
9.74840581e-01 6.17300905e-02 9.96505320e-01 8.55370581e-01
3.34503591e-01 1.04781054e-01 -1.45149529e+00 6.76914573e-01
-1.44399852e-01 -1.21298993e+00 2.91604716e-02 -1.72558352e-01
7.45298624e-01 4.39226210e-01 5.59177324e-02 1.15549195e+00
1.00635910e+00 -9.16604519e-01 7.09515214e-01 1.86022669e-01
6.80357575e-01 -6.78653955e-01 6.95309415e-02 7.34224796e-01
-7.65525103e-01 -5.14085233e-01 -3.97339046e-01 1.30195580e-02
9.87323523e-02 -3.07438988e-02 -1.19990981e+00 3.47157955e-01
3.77697796e-01 8.83908033e-01 -4.74942327e-01 5.85928619e-01
-5.97250342e-01 9.37310219e-01 -5.27837798e-02 -1.98593363e-02
7.32588410e-01 -1.80849716e-01 3.85514677e-01 1.18954575e+00
-1.82344005e-01 5.68359978e-02 7.30557919e-01 6.92401528e-01
-2.48852119e-01 -1.60719752e-01 -3.77590269e-01 -3.15877885e-01
4.64525431e-01 1.11805654e+00 -1.83850452e-01 -3.49370509e-01
-1.86867520e-01 9.69959080e-01 1.00293159e+00 6.95846379e-01
-8.33218694e-01 6.88090473e-02 6.58077002e-01 -3.98679346e-01
5.03520250e-01 -3.80726278e-01 1.66390300e-01 -1.23619449e+00
-3.65852445e-01 -1.46100163e+00 5.77139139e-01 -4.24887359e-01
-1.43410528e+00 3.80172104e-01 -6.75118864e-02 -9.38450456e-01
-7.47685909e-01 -5.84161758e-01 -4.91381854e-01 4.64787453e-01
-1.69203341e+00 -1.04314804e+00 1.81884795e-01 6.90298200e-01
9.54257965e-01 -4.82713223e-01 5.67553103e-01 -8.25529993e-02
-6.35209799e-01 8.12664866e-01 1.80557668e-01 2.20203936e-01
8.67394149e-01 -1.24481988e+00 5.21359086e-01 7.25971162e-01
1.47434264e-01 4.59199727e-01 4.38733041e-01 -5.75452745e-01
-1.63562584e+00 -1.29986799e+00 1.58297226e-01 -3.03190261e-01
9.48529780e-01 -4.62141216e-01 -7.45583117e-01 9.16338861e-01
2.14630574e-01 1.28590509e-01 2.51969397e-01 7.68847391e-02
-5.92177689e-01 2.59239990e-02 -7.01804459e-01 7.63732433e-01
1.22210145e+00 -5.07722735e-01 -6.80226028e-01 3.05938035e-01
1.29325426e+00 -4.63963598e-01 -7.83874631e-01 2.57952064e-01
3.44253421e-01 -5.06343067e-01 9.49904263e-01 -1.23759973e+00
5.31364799e-01 -6.27262965e-02 -4.89928424e-02 -1.79369164e+00
-2.65340328e-01 -8.44791114e-01 -3.60198855e-01 9.72108364e-01
3.84164542e-01 -4.88695949e-01 5.72950304e-01 1.62749410e-01
-3.75244021e-01 -9.63402867e-01 -6.26357079e-01 -9.86696959e-01
3.44080210e-01 -4.87883061e-01 5.60504496e-01 8.19481969e-01
-2.34373640e-02 5.66174448e-01 -5.70533931e-01 -2.45394140e-01
5.05753338e-01 1.67101666e-01 8.81958902e-01 -4.78090942e-01
-9.35164690e-01 -4.10045683e-01 3.15359056e-01 -1.49197304e+00
7.48704672e-01 -1.09183848e+00 3.45103770e-01 -1.04316354e+00
1.23730198e-01 -6.52515233e-01 -5.08953273e-01 9.66216505e-01
-4.85328019e-01 -4.12190825e-01 4.20631617e-01 1.34956747e-01
-9.27211761e-01 1.02988970e+00 1.46974969e+00 -1.61328658e-01
-2.63770133e-01 -1.33345500e-01 -4.92582977e-01 4.55226392e-01
8.48544955e-01 -3.85931998e-01 -7.62539089e-01 -6.04614615e-01
-4.44168411e-02 3.53022486e-01 3.24458361e-01 -6.36228085e-01
5.39811440e-02 -5.04632533e-01 1.66794211e-01 -2.48035952e-01
2.02467069e-01 -3.86015356e-01 -7.26582110e-01 5.96865058e-01
-9.64125693e-01 1.44509599e-01 1.67229921e-01 8.74513149e-01
6.05856366e-02 -6.63287565e-02 8.05546224e-01 -2.54531890e-01
-1.12213755e+00 5.18159270e-01 -4.30246741e-02 7.31081367e-01
8.27540338e-01 3.50962818e-01 -5.68460286e-01 -3.08310241e-01
-7.38500535e-01 7.89189041e-01 2.23875493e-01 7.81534612e-01
5.54572046e-01 -1.13532829e+00 -7.33144343e-01 8.62381160e-02
1.82966650e-01 -4.61146720e-02 4.47528996e-02 7.56462991e-01
5.11866659e-02 4.59747702e-01 -1.97219327e-01 -5.27050912e-01
-7.69962072e-01 8.65821242e-01 3.53027552e-01 -5.97975373e-01
-6.85626805e-01 7.88837612e-01 4.61740792e-01 -7.58865654e-01
4.41986144e-01 -2.97371924e-01 -7.05370232e-02 -3.37800562e-01
4.97215003e-01 2.82554150e-01 -9.70859602e-02 -7.26950243e-02
-5.48408143e-02 -1.02593765e-01 -5.22635639e-01 -4.33445454e-01
1.35748422e+00 -1.16235139e-02 4.30155307e-01 3.80910516e-01
1.08629441e+00 -4.92161721e-01 -2.08014321e+00 -5.22712409e-01
1.92087397e-01 -9.64094996e-02 -6.37123361e-02 -1.26854157e+00
-8.99115622e-01 7.92586625e-01 2.89481997e-01 -4.31210577e-01
9.18542802e-01 -1.87066138e-01 9.19768631e-01 8.85292590e-01
4.44130570e-01 -1.30331743e+00 5.67657650e-01 1.03491342e+00
9.13385332e-01 -1.31330204e+00 -4.54089522e-01 4.22292739e-01
-1.13372517e+00 8.58334363e-01 1.02260530e+00 -1.68959469e-01
1.44042939e-01 2.47916788e-01 6.98384717e-02 2.97924906e-01
-1.49390721e+00 -3.64890307e-01 -7.72440955e-02 5.47657788e-01
2.05094367e-01 2.89816596e-02 -3.46144265e-03 5.93448162e-01
-9.86214131e-02 2.40886778e-01 1.92092881e-01 1.03996062e+00
-3.87131572e-01 -1.26287341e+00 -4.31157574e-02 2.78374165e-01
-2.10996211e-01 -1.65411800e-01 -1.88128129e-01 6.07036173e-01
-1.85170472e-01 8.73042226e-01 -1.52935132e-01 -4.97642845e-01
2.07570687e-01 2.27050036e-01 6.30741417e-01 -5.47477067e-01
-8.16975594e-01 3.60240974e-02 9.01814327e-02 -7.96711385e-01
3.99690606e-02 -5.26167214e-01 -1.44335163e+00 7.70979598e-02
1.19911328e-01 2.43505239e-01 4.17608827e-01 1.23112428e+00
4.46005672e-01 5.64275503e-01 7.11078703e-01 -6.90499127e-01
-1.43291068e+00 -9.79357541e-01 -8.21029767e-02 4.88991529e-01
5.78412712e-01 -7.22545862e-01 -2.80036271e-01 3.24013308e-02] | [4.088040351867676, 1.7933942079544067] |
59a3c369-3f53-4d36-9d17-d841dce6622e | risk-sensitive-policy-with-distributional | 2212.14743 | null | https://arxiv.org/abs/2212.14743v1 | https://arxiv.org/pdf/2212.14743v1.pdf | Risk-Sensitive Policy with Distributional Reinforcement Learning | Classical reinforcement learning (RL) techniques are generally concerned with the design of decision-making policies driven by the maximisation of the expected outcome. Nevertheless, this approach does not take into consideration the potential risk associated with the actions taken, which may be critical in certain applications. To address that issue, the present research work introduces a novel methodology based on distributional RL to derive sequential decision-making policies that are sensitive to the risk, the latter being modelled by the tail of the return probability distribution. The core idea is to replace the $Q$ function generally standing at the core of learning schemes in RL by another function taking into account both the expected return and the risk. Named the risk-based utility function $U$, it can be extracted from the random return distribution $Z$ naturally learnt by any distributional RL algorithm. This enables to span the complete potential trade-off between risk minimisation and expected return maximisation, in contrast to fully risk-averse methodologies. Fundamentally, this research yields a truly practical and accessible solution for learning risk-sensitive policies with minimal modification to the distributional RL algorithm, and with an emphasis on the interpretability of the resulting decision-making process. | ['Damien Ernst', 'Thibaut Théate'] | 2022-12-30 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [ 1.47090435e-01 5.85319459e-01 -2.72728711e-01 -3.40793520e-01
-6.49307191e-01 -3.97427410e-01 6.09335721e-01 5.28468311e-01
-1.02014303e+00 9.39244807e-01 3.29654887e-02 -4.54397231e-01
-7.53828108e-01 -1.19813871e+00 -3.59905481e-01 -9.37085450e-01
-9.88628268e-02 3.42468053e-01 -2.28522271e-01 -1.77660570e-01
4.37675953e-01 3.93307239e-01 -1.64616549e+00 -2.63505191e-01
8.65214586e-01 1.27084744e+00 1.46427780e-01 4.49246764e-01
-7.15089440e-02 8.76831830e-01 -6.51997387e-01 -4.35007751e-01
1.99943826e-01 -3.99329215e-01 -6.21407092e-01 -4.04770344e-01
-3.88037324e-01 -2.88510621e-01 4.91248041e-01 1.13371170e+00
5.26687860e-01 3.62751752e-01 1.00355339e+00 -1.06761646e+00
-1.06974589e-02 7.15370953e-01 -2.09325552e-01 7.28341118e-02
2.84360886e-01 1.31973416e-01 1.15112150e+00 -5.84097095e-02
3.05763155e-01 1.35387242e+00 2.22028136e-01 4.95902926e-01
-1.23588669e+00 -3.62328917e-01 3.36416841e-01 -7.10846856e-02
-9.63054717e-01 5.09578399e-02 6.73358798e-01 -4.76326644e-01
6.23464286e-01 1.88671857e-01 5.75463951e-01 1.00726628e+00
5.64633548e-01 7.59905398e-01 1.51527619e+00 -5.26367605e-01
7.81753957e-01 3.78672719e-01 -3.88270736e-01 1.29094973e-01
2.70960122e-01 7.73908854e-01 1.09474756e-01 -4.77120318e-02
4.63628739e-01 -1.65071398e-01 5.05376607e-02 -7.05715537e-01
-4.97085571e-01 1.08013785e+00 1.73137873e-01 2.80223280e-01
-6.35260403e-01 1.97479337e-01 5.04384577e-01 6.16220772e-01
3.41475487e-01 5.03925502e-01 -3.14075202e-01 -1.69572294e-01
-6.97552264e-01 6.17162347e-01 7.56710470e-01 2.92195588e-01
6.33498132e-01 1.04255199e-01 -4.40126359e-01 2.59564012e-01
6.79075480e-01 3.19279432e-01 4.06031489e-01 -7.65428603e-01
4.12406147e-01 5.82940698e-01 4.29015666e-01 -6.96426988e-01
-2.20129818e-01 -5.91745794e-01 -2.78212279e-01 8.93031538e-01
5.03711283e-01 -5.55234373e-01 -1.74137577e-01 1.93146622e+00
4.22103435e-01 -4.85465169e-01 3.25403512e-01 3.90145898e-01
-1.58947527e-01 3.02952796e-01 4.96563822e-01 -6.72697663e-01
9.01376188e-01 -3.32481749e-02 -7.04251826e-01 9.79104172e-03
5.61903417e-01 -2.92154402e-01 1.06746912e+00 4.43784326e-01
-9.56227183e-01 -2.46536419e-01 -1.10216379e+00 5.60826004e-01
-2.35632479e-01 -4.38473523e-01 3.27318430e-01 9.10613894e-01
-7.93932915e-01 9.58737016e-01 -3.52891088e-01 -4.38968949e-02
2.11079061e-01 3.36289853e-01 3.41642916e-01 3.11956495e-01
-1.27903378e+00 1.31232762e+00 6.51958406e-01 4.51317057e-02
-8.92546475e-01 -3.65573615e-01 -6.58287346e-01 1.00503549e-01
8.79122853e-01 -3.39132696e-01 1.43142569e+00 -1.23139679e+00
-1.91997516e+00 4.10399318e-01 5.65733314e-01 -7.29153693e-01
1.26576567e+00 -1.35738835e-01 -8.95272940e-02 -1.26377285e-01
-1.52627319e-01 1.16183661e-01 1.05539680e+00 -1.09437704e+00
-6.69180572e-01 -2.32447326e-01 2.13562310e-01 4.36796874e-01
1.13834701e-01 -1.00780249e-01 7.93039620e-01 -6.27443552e-01
-7.07530797e-01 -4.90867764e-01 -4.99460518e-01 -3.51730704e-01
6.63189143e-02 -3.55297178e-01 3.14119935e-01 -2.72734135e-01
1.24678016e+00 -2.10823250e+00 -2.78750602e-02 3.64788860e-01
-2.83873111e-01 2.76260227e-01 8.76576751e-02 5.62946737e-01
-3.01764179e-02 -6.02858923e-02 -4.40108955e-01 4.84662838e-02
4.99566764e-01 1.81983277e-01 -4.43567991e-01 4.38217729e-01
2.36374229e-01 5.85053802e-01 -1.03019369e+00 -2.70731241e-01
3.88333350e-01 1.98728055e-01 -4.72958535e-01 5.56521356e-01
-5.44681966e-01 1.83865592e-01 -7.85556674e-01 1.25552654e-01
2.89842397e-01 6.80259526e-01 2.41025567e-01 5.31238914e-01
-3.35368246e-01 1.20026127e-01 -1.36428845e+00 9.12676334e-01
-5.04279256e-01 1.85582619e-02 1.48167089e-02 -1.32968581e+00
1.19131505e+00 2.96973288e-01 5.73853076e-01 -6.51401818e-01
3.44442070e-01 2.48903960e-01 -3.71206589e-02 -1.83817312e-01
2.48714045e-01 -8.59041750e-01 -2.43003011e-01 7.51862705e-01
-1.58945426e-01 -3.45304221e-01 -1.16272271e-01 -4.37294990e-01
6.03382707e-01 5.84832489e-01 8.14358354e-01 -5.34721911e-01
6.41339302e-01 -4.08722550e-01 4.71981525e-01 5.91716707e-01
-4.52148139e-01 -1.58325449e-01 1.11425185e+00 -2.87940055e-01
-5.11729300e-01 -1.16784811e+00 -1.72630236e-01 8.70264947e-01
-2.48542070e-01 1.06553577e-01 -8.75137925e-01 -1.05967867e+00
1.45674184e-01 1.31743670e+00 -7.41157830e-01 -4.87742662e-01
-1.66366175e-01 -7.23613679e-01 1.92194849e-01 1.23733319e-01
2.88212478e-01 -1.40426636e+00 -1.31976676e+00 4.20722812e-01
3.34823638e-01 -1.92327648e-01 -3.17533910e-02 4.22590643e-01
-8.79794836e-01 -9.95023608e-01 -6.68227196e-01 2.01322176e-02
2.25934565e-01 -5.96350491e-01 1.02063942e+00 -4.86621737e-01
-9.52721313e-02 7.46006966e-01 -2.36114696e-01 -9.91186321e-01
-5.43297410e-01 -2.75514603e-01 3.31295431e-02 2.65976518e-01
3.41655880e-01 -4.80766356e-01 -5.42738318e-01 5.22994846e-02
-9.95013297e-01 -7.21877337e-01 5.98592520e-01 4.52082425e-01
5.76334000e-01 4.64698374e-01 1.03434587e+00 -7.11136699e-01
1.12173688e+00 -5.46048880e-01 -8.56847525e-01 3.71614099e-01
-1.13386095e+00 6.77360654e-01 6.15769148e-01 -3.36868286e-01
-1.44997275e+00 -2.47478217e-01 -2.28736788e-01 1.54844701e-01
-2.35751644e-01 4.35733765e-01 -5.39244533e-01 2.42585480e-01
5.00462234e-01 1.04921207e-01 2.35467732e-01 -2.95814306e-01
4.61263925e-01 5.73512495e-01 -5.77487126e-02 -8.22908163e-01
4.60399836e-01 1.79476552e-02 3.37809145e-01 -5.66250622e-01
-6.96590722e-01 2.25812681e-02 -4.04705554e-01 -4.78894562e-01
6.68659627e-01 -3.31385702e-01 -9.38808143e-01 1.58859998e-01
-4.25540566e-01 -4.05552626e-01 -1.20772266e+00 5.18814325e-01
-1.42039657e+00 1.23657212e-01 8.18675607e-02 -1.45641553e+00
-2.78760821e-01 -9.84240711e-01 4.71147925e-01 5.99594600e-02
-2.19562531e-01 -1.19849479e+00 2.56223172e-01 -9.60857570e-02
3.89477849e-01 5.36029935e-01 1.27523685e+00 -6.87085509e-01
-1.92674428e-01 -1.42933756e-01 2.19017178e-01 6.86628163e-01
-1.76271889e-02 -2.32563987e-01 -8.15875590e-01 -3.77134413e-01
5.22711396e-01 -3.75219136e-01 5.14897466e-01 5.40912569e-01
6.04530752e-01 -4.44859385e-01 3.29830348e-01 1.14164464e-01
1.74502718e+00 5.34864426e-01 5.41116893e-01 5.97880065e-01
-1.06288097e-03 1.09438109e+00 1.04418755e+00 7.65062630e-01
1.69740468e-01 5.37481308e-01 7.07212150e-01 5.72298825e-01
6.42140090e-01 -4.40156996e-01 7.45410323e-01 -1.20485917e-01
-5.66125624e-02 8.70178714e-02 -6.31978691e-01 1.28102690e-01
-1.90006018e+00 -1.04217029e+00 4.72276837e-01 2.71031404e+00
6.10808551e-01 3.57618570e-01 5.06086826e-01 3.63068461e-01
4.85854387e-01 2.99239218e-01 -6.46034896e-01 -1.14191115e+00
2.95881569e-01 2.06422150e-01 5.48908710e-01 7.16285586e-01
-7.90758908e-01 3.83250207e-01 6.48511028e+00 9.11958456e-01
-9.29266930e-01 -3.42230260e-01 6.45820320e-01 -1.30268425e-01
-5.43662310e-01 -9.50646326e-02 -5.97120881e-01 4.53827858e-01
1.10811818e+00 -4.58608538e-01 2.44992256e-01 9.59287882e-01
4.31852907e-01 -4.34401274e-01 -1.03579700e+00 3.15621525e-01
-5.19844294e-01 -4.89675105e-01 -4.11563106e-02 3.28368753e-01
3.46946061e-01 -6.63402379e-01 2.16748223e-01 5.23317337e-01
4.50755984e-01 -1.08066356e+00 1.03228724e+00 7.97140837e-01
4.94473249e-01 -1.48165929e+00 9.14400518e-01 6.74960613e-01
-7.49283135e-01 -4.15132672e-01 -2.23922715e-01 -3.36434931e-01
-1.48750111e-01 6.15394473e-01 -7.14215040e-01 5.85417211e-01
2.41740555e-01 3.65243495e-01 -1.35093629e-02 6.79845572e-01
-4.24337655e-01 3.69465142e-01 -1.22306891e-01 -4.29205835e-01
5.04539490e-01 -5.41759849e-01 5.47147632e-01 1.00629747e+00
3.55895966e-01 -3.08247983e-01 -1.14664897e-01 9.94727731e-01
4.91304010e-01 3.47283900e-01 -7.85251975e-01 -1.96652394e-02
5.93569353e-02 8.31913531e-01 -5.50484121e-01 8.34522992e-02
-2.61376891e-02 3.02611232e-01 2.82902300e-01 1.20074712e-01
-3.59498292e-01 -2.86559880e-01 6.79978251e-01 4.06278595e-02
2.43968084e-01 7.77341127e-02 -3.75575088e-02 -5.44525921e-01
8.47532824e-02 -7.28878915e-01 5.81726670e-01 9.21969675e-03
-1.25069392e+00 2.91422963e-01 4.20202881e-01 -1.01456499e+00
-7.44408250e-01 -6.51944339e-01 -6.29962623e-01 1.05838883e+00
-1.64465296e+00 -4.65095699e-01 6.80195332e-01 4.22754347e-01
1.63614169e-01 1.79796778e-02 7.00851560e-01 -3.51835102e-01
-5.14131308e-01 3.58179629e-01 2.40311325e-01 -4.53772694e-01
3.48944306e-01 -1.57427502e+00 -2.25651294e-01 5.59820533e-01
-4.50806171e-01 2.20009461e-01 9.26589251e-01 -4.69589114e-01
-9.27816749e-01 -7.47163892e-01 8.51174593e-01 -3.75523150e-01
5.98476112e-01 3.52998637e-02 -4.57590729e-01 2.90285915e-01
-4.53820080e-02 -3.92221093e-01 4.95546967e-01 -1.29276784e-02
-2.81409062e-02 -2.37486675e-01 -1.65422702e+00 5.20828843e-01
4.00912672e-01 -4.12324071e-01 -9.59087074e-01 -4.02765453e-01
4.70806688e-01 2.10459337e-01 -8.24154794e-01 2.82330245e-01
4.96118516e-01 -1.16211605e+00 7.84019947e-01 -4.93275791e-01
1.60739869e-01 4.12116945e-02 -9.71810892e-02 -1.39764965e+00
7.53540769e-02 -7.82764137e-01 -1.19641669e-01 1.15420926e+00
1.82868734e-01 -7.93767989e-01 4.99550045e-01 7.82846272e-01
4.31582332e-01 -9.74327803e-01 -1.30238354e+00 -8.14939976e-01
4.99912053e-01 -5.43790698e-01 5.91002405e-01 4.97750431e-01
-3.08393370e-02 -1.50355697e-01 -2.01251388e-01 -3.06818306e-01
9.62773919e-01 -3.30447033e-02 1.74765587e-01 -1.19664252e+00
-3.84459704e-01 -7.43607402e-01 -1.47510380e-01 -3.31588864e-01
2.24455729e-01 -5.48421323e-01 2.02026457e-01 -1.16894913e+00
-2.91241825e-01 -3.85433376e-01 -7.24194467e-01 -3.30778323e-02
-1.81571603e-01 -7.27383614e-01 3.87160867e-01 -2.58904696e-01
-3.10486585e-01 8.07532787e-01 1.23800123e+00 2.33500615e-01
-4.42381024e-01 6.59694254e-01 -8.71273160e-01 9.29457903e-01
9.29574132e-01 -4.64042127e-01 -8.27430308e-01 2.50764996e-01
5.23880482e-01 2.59245753e-01 1.05553508e-01 -7.30046749e-01
-2.28152186e-01 -7.29421973e-01 -1.01376623e-01 -1.46096170e-01
-1.97202057e-01 -1.04986107e+00 -5.25723845e-02 6.80038214e-01
-7.33456135e-01 2.30030678e-02 -1.26185432e-01 7.13691831e-01
-9.38281938e-02 -6.31620765e-01 8.70323122e-01 -3.34202200e-01
-4.03129607e-01 6.98887790e-03 -6.39557302e-01 1.45477727e-01
1.35463285e+00 -1.18078701e-01 3.34876031e-01 -3.83367807e-01
-6.04015172e-01 2.09739268e-01 1.37956560e-01 1.25630304e-01
3.84444743e-01 -9.86122310e-01 -5.33825815e-01 -1.08320068e-03
-1.15906343e-01 -5.79245901e-03 1.54280871e-01 3.67558688e-01
-1.21055812e-01 3.16400170e-01 -3.84846330e-01 1.30359635e-01
-5.28765678e-01 7.25079536e-01 6.55926764e-01 -6.93859279e-01
-3.30944479e-01 4.31678861e-01 -4.76786271e-02 -2.90738791e-01
4.10516381e-01 -2.98524052e-01 -5.28618336e-01 5.77094674e-01
5.27061880e-01 6.38918817e-01 -1.77837864e-01 -3.01770270e-01
-3.04564908e-02 3.41707557e-01 2.64536709e-01 -5.13286412e-01
1.26947510e+00 -2.04268739e-01 2.14899182e-01 6.55419171e-01
7.89100289e-01 -1.83895305e-01 -1.83368027e+00 -6.69679493e-02
5.95172465e-01 -2.89056599e-01 1.22037448e-01 -1.03315544e+00
-6.52184486e-01 8.09239745e-01 5.50323606e-01 3.61515999e-01
1.15056276e+00 -5.12865126e-01 1.33375466e-01 2.20701754e-01
6.94306254e-01 -1.47207928e+00 -2.39426151e-01 1.75348535e-01
7.88441360e-01 -9.73088920e-01 -6.49091378e-02 1.99960396e-01
-7.91720450e-01 9.75679934e-01 1.95632949e-01 -3.22997272e-01
6.90763533e-01 1.51974156e-01 1.61845121e-03 1.71648711e-01
-6.01159513e-01 -3.32717538e-01 1.00461908e-01 8.33432436e-01
1.99404925e-01 5.44336081e-01 -7.39632607e-01 6.18937194e-01
-1.29847795e-01 9.74998698e-02 2.94553816e-01 1.17844105e+00
-5.75111210e-01 -1.52942145e+00 -2.57902622e-01 2.55289048e-01
-5.58451593e-01 2.06570774e-01 -2.33544677e-01 7.01667130e-01
1.09696396e-01 8.66399646e-01 -2.04201788e-01 -8.15760493e-02
5.78494728e-01 1.71989381e-01 3.75946939e-01 -4.54939306e-01
-6.69502556e-01 -1.13635056e-01 -7.56194293e-02 -5.79254210e-01
-3.68260801e-01 -8.75801861e-01 -8.92102420e-01 -7.06833377e-02
1.02782086e-01 4.18558776e-01 6.20006084e-01 1.17157686e+00
-1.59891859e-01 4.21974093e-01 8.60445440e-01 -7.05997527e-01
-1.55155039e+00 -5.29899061e-01 -9.75935578e-01 1.50939167e-01
5.01761317e-01 -6.83251917e-01 -5.65552175e-01 -7.53570676e-01] | [4.214129447937012, 2.5618948936462402] |
c4404389-1058-4586-9a6e-167f6e500309 | optimizing-nlu-reranking-using-entity | null | null | https://aclanthology.org/2021.naacl-industry.3 | https://aclanthology.org/2021.naacl-industry.3.pdf | Optimizing NLU Reranking Using Entity Resolution Signals in Multi-domain Dialog Systems | In dialog systems, the Natural Language Understanding (NLU) component typically makes the interpretation decision (including domain, intent and slots) for an utterance before the mentioned entities are resolved. This may result in intent classification and slot tagging errors. In this work, we propose to leverage Entity Resolution (ER) features in NLU reranking and introduce a novel loss term based on ER signals to better learn model weights in the reranking framework. In addition, for a multi-domain dialog scenario, we propose a score distribution matching method to ensure scores generated by the NLU reranking models for different domains are properly calibrated. In offline experiments, we demonstrate our proposed approach significantly outperforms the baseline model on both single-domain and cross-domain evaluations. | ['Yang Liu', 'Yue Liu', 'Chengwei Su', 'Han Wang', 'Xin He', 'Shuyan Dong', 'Mohsen Malmir', 'Jiangning Chen', 'Tong Wang'] | 2021-06-01 | null | null | null | naacl-2021-4 | ['entity-resolution'] | ['natural-language-processing'] | [ 1.63393810e-01 4.65601206e-01 -4.19967741e-01 -8.96963060e-01
-9.48960841e-01 -7.80264735e-01 7.00794101e-01 3.88532519e-01
-5.83495796e-01 8.76745880e-01 7.14056492e-01 -1.50106668e-01
1.34041429e-01 -4.37325984e-01 -3.58859301e-01 1.61372572e-01
3.55124742e-01 1.05487287e+00 3.19853842e-01 -3.36710811e-01
2.46133953e-02 -1.70507312e-01 -8.11404526e-01 3.97582859e-01
1.27413392e+00 9.93055582e-01 3.10313255e-01 4.65777934e-01
-5.13637066e-01 8.75590503e-01 -7.51601994e-01 -7.73042679e-01
1.46749988e-01 -1.96415663e-01 -1.04799461e+00 9.83770937e-02
3.15620184e-01 -7.10346997e-01 -2.16407284e-01 9.75692868e-01
2.50977695e-01 5.12996912e-01 7.48981476e-01 -1.10736358e+00
-4.64335412e-01 9.55466092e-01 -3.51871520e-01 -1.73587501e-01
6.35099530e-01 -3.47970784e-01 1.62651646e+00 -9.49491203e-01
5.58505118e-01 1.58457494e+00 3.41375589e-01 7.02179968e-01
-1.19249845e+00 -6.52011693e-01 3.84099364e-01 4.42081951e-02
-9.06422019e-01 -3.84808809e-01 6.89966261e-01 -3.28599721e-01
8.52869749e-01 7.28826830e-03 -3.36247921e-01 1.11506069e+00
-2.74743438e-01 1.04378700e+00 8.32181633e-01 -3.48405451e-01
1.95462152e-01 3.14632982e-01 6.23021603e-01 4.46161807e-01
4.61374819e-02 -2.23681509e-01 -5.86398244e-01 -5.37392020e-01
4.96032953e-01 6.63412809e-02 -1.25874683e-01 -3.78925741e-01
-9.18098807e-01 1.03300738e+00 2.19486982e-01 7.42737874e-02
-2.60117888e-01 -3.65535766e-01 4.40355986e-01 3.42466861e-01
4.73684877e-01 7.32288420e-01 -8.43431175e-01 -2.52245739e-02
-4.93422478e-01 3.18015456e-01 1.02456331e+00 8.85534525e-01
8.38963509e-01 -5.27039886e-01 -5.57810962e-01 1.41733170e+00
5.76626360e-01 3.25567394e-01 6.77789748e-01 -9.20359194e-01
8.54310155e-01 9.04649138e-01 5.62504113e-01 -6.45423830e-01
-5.44007242e-01 -5.74060753e-02 -5.13727248e-01 -4.01455462e-01
4.80501801e-01 -3.94888639e-01 -7.44414210e-01 1.94683683e+00
3.70073080e-01 1.60998970e-01 5.01243949e-01 1.05875242e+00
9.17241335e-01 4.94862467e-01 5.30980706e-01 5.24800122e-02
1.66313326e+00 -9.52827811e-01 -8.22348177e-01 -5.46606660e-01
5.78024149e-01 -7.30455458e-01 1.11087823e+00 -9.56852734e-02
-8.28440070e-01 -4.22625333e-01 -8.96655321e-01 -2.34026641e-01
5.47855161e-02 4.17304903e-01 4.36237305e-01 1.38689458e-01
-5.33069909e-01 3.35855216e-01 -5.29213190e-01 -4.07741457e-01
-2.52323866e-01 3.03835243e-01 -1.70292839e-01 1.27709851e-01
-1.91010582e+00 9.55063283e-01 4.91339177e-01 -4.05416489e-01
-6.31999314e-01 -5.78782439e-01 -9.71434176e-01 3.81365806e-01
4.83523846e-01 -5.70645213e-01 1.84980845e+00 -5.04947484e-01
-1.71414280e+00 6.64807618e-01 -4.08889800e-01 -7.92527020e-01
4.23776865e-01 -5.64394772e-01 -2.81218946e-01 -1.26601577e-01
1.75451234e-01 6.60922408e-01 4.45500344e-01 -1.12358880e+00
-9.32461083e-01 -2.29933128e-01 4.31257963e-01 5.83569884e-01
-2.69966483e-01 -1.56554833e-01 -2.71442980e-01 -4.89552826e-01
1.86912045e-01 -7.71824896e-01 -1.89414993e-01 -5.03334701e-01
-3.70825112e-01 -7.40467131e-01 4.70468551e-01 -7.82822073e-01
1.37781441e+00 -1.94152856e+00 8.32212120e-02 -7.22683817e-02
2.84961075e-01 2.71492839e-01 1.47889019e-03 3.56347620e-01
2.83928424e-01 -6.43191487e-02 -1.50566518e-01 -4.92403507e-01
3.14645559e-01 2.27612019e-01 -6.47665381e-01 -1.65418014e-01
2.01592177e-01 5.25120735e-01 -8.64647388e-01 -4.06628907e-01
8.97210911e-02 -2.58526206e-02 -6.81769848e-01 6.53082430e-01
-5.13354242e-01 4.35802847e-01 -5.86302221e-01 3.14791173e-01
5.77431381e-01 -4.53705788e-01 5.93755960e-01 -3.29073548e-01
3.75365227e-01 9.48246598e-01 -1.06472993e+00 1.68751979e+00
-8.67857397e-01 1.78454354e-01 2.84130406e-02 -5.86678922e-01
9.67374504e-01 4.51239854e-01 2.54574597e-01 -6.69068396e-01
-4.47306922e-03 1.61288157e-01 -1.83655187e-01 -8.83620232e-02
8.80855381e-01 -2.36405078e-02 -5.91215134e-01 4.97245252e-01
2.86029726e-01 4.22837853e-01 -1.46041825e-01 5.01395464e-01
9.88211632e-01 -3.18342149e-01 5.84848762e-01 -1.67934783e-02
6.62519455e-01 2.02266872e-02 7.42709935e-01 7.87034690e-01
-3.13723683e-01 3.14957798e-01 5.08333683e-01 -1.49095044e-01
-6.54355764e-01 -1.13786805e+00 -1.29893020e-01 1.56997263e+00
4.74971086e-01 -2.19936252e-01 -5.44458628e-01 -1.21752834e+00
2.13340953e-01 1.02982986e+00 -1.21552885e-01 -1.10976823e-01
-5.03474355e-01 -3.62135321e-01 4.54360753e-01 5.01995683e-01
6.70472860e-01 -6.73950791e-01 -6.17933199e-02 4.56076771e-01
-5.57313085e-01 -1.48776519e+00 -7.75391579e-01 2.50916719e-01
-7.77542174e-01 -9.17133927e-01 -6.96202219e-01 -7.50708044e-01
4.86699969e-01 6.72829449e-02 1.24465251e+00 -2.10569739e-01
5.86277604e-01 3.33719343e-01 -6.06032610e-01 -4.53731343e-02
-4.79750991e-01 3.26052040e-01 2.38644198e-01 -1.14640102e-01
8.43175948e-01 -1.58108726e-01 -4.87586737e-01 6.26936674e-01
-7.22755671e-01 -9.33575332e-02 4.88554299e-01 1.24357355e+00
1.31056190e-01 -3.99355710e-01 8.36890519e-01 -1.16647196e+00
9.86776114e-01 -6.08432949e-01 -6.06243670e-01 6.71809077e-01
-6.53611183e-01 3.84705245e-01 4.74345088e-01 -4.58033979e-01
-1.54896021e+00 5.42908199e-02 -3.71115394e-02 -1.57920644e-01
-1.32390022e-01 4.46670622e-01 -4.52299237e-01 3.67642909e-01
5.43341339e-01 -1.12482458e-01 -5.35080850e-01 -8.04170012e-01
3.56349707e-01 1.00691438e+00 3.71160239e-01 -8.26375544e-01
6.71662092e-01 -2.11637691e-02 -7.53510118e-01 -5.08741617e-01
-1.18429887e+00 -9.06086862e-01 -5.48506558e-01 1.29252627e-01
8.69156778e-01 -1.26823437e+00 -5.04987001e-01 6.54388964e-02
-1.56933212e+00 -1.00323707e-01 5.70815355e-02 5.57955563e-01
-9.98666584e-02 4.64906275e-01 -6.26231253e-01 -9.82800066e-01
-1.84477329e-01 -1.07421434e+00 1.24047279e+00 5.75583160e-01
-5.77698946e-01 -1.08970058e+00 2.27976903e-01 6.49139881e-01
2.21063897e-01 -5.36548436e-01 8.68826866e-01 -1.61212611e+00
-2.39901826e-01 -3.56688127e-02 -3.96038502e-01 8.41974169e-02
2.16316476e-01 -8.35483015e-01 -1.02247429e+00 -7.25238025e-02
-5.95286265e-02 -5.48062503e-01 8.98731112e-01 5.04079238e-02
5.14709949e-01 -6.25422776e-01 -2.82075524e-01 -6.41068583e-03
9.09242690e-01 2.84432083e-01 1.79423511e-01 1.44038603e-01
4.15169150e-01 8.11730266e-01 1.09097934e+00 6.72709942e-01
8.04922044e-01 9.79893625e-01 -1.88618317e-01 6.13035187e-02
1.57372177e-01 -5.55913627e-01 4.36026186e-01 4.33323592e-01
7.35446393e-01 -4.79510486e-01 -7.81676710e-01 4.72120792e-01
-2.00206637e+00 -6.17665708e-01 4.25847262e-01 2.14505076e+00
1.17098725e+00 4.05975804e-02 -4.13125530e-02 -7.24948108e-01
8.69581163e-01 2.19894037e-01 -7.36425281e-01 -2.52638936e-01
7.65106035e-03 -1.35753781e-01 4.75187153e-01 9.65340436e-01
-1.26138699e+00 1.24381840e+00 5.63160467e+00 5.77726901e-01
-7.37513900e-01 1.90102801e-01 6.64567709e-01 4.23735976e-01
-4.57933217e-01 1.78394839e-01 -1.14711964e+00 4.23967391e-01
6.41486049e-01 -3.81498307e-01 9.01967138e-02 1.08171749e+00
-1.18062697e-01 -3.83310616e-02 -1.19662726e+00 5.97037494e-01
-1.72854975e-01 -9.08018053e-01 2.69566588e-02 -1.41394302e-01
6.31530762e-01 -7.69678503e-02 -2.59475470e-01 7.86181808e-01
1.08038127e+00 -6.43376708e-01 1.69808254e-01 7.66851977e-02
4.85232383e-01 -4.06856835e-01 8.67152572e-01 1.96570605e-01
-1.06196427e+00 3.21365446e-02 -5.13839126e-01 2.09590212e-01
3.50080997e-01 4.35488343e-01 -1.69546902e+00 3.97675216e-01
1.61017552e-01 3.05127174e-01 -3.04761752e-02 5.01167655e-01
-3.20731759e-01 6.27756357e-01 -3.27368267e-02 -1.76130105e-02
9.04923826e-02 3.69406044e-02 6.83855414e-01 1.10983574e+00
1.76958032e-02 3.20364118e-01 5.20092249e-01 6.99406743e-01
-5.36314011e-01 2.07469165e-01 -2.37765491e-01 -4.37935302e-03
8.77762616e-01 1.10169864e+00 -2.01560289e-01 -3.79153877e-01
-5.09888291e-01 1.24161482e+00 5.18782854e-01 3.47665280e-01
-6.88119352e-01 -1.78625867e-01 1.06341136e+00 -2.47195825e-01
2.93852203e-02 -1.92404866e-01 -1.63836405e-01 -1.47667718e+00
-1.64539129e-01 -6.44492149e-01 7.72280514e-01 -2.84072638e-01
-1.67927265e+00 5.22289574e-01 -8.40096676e-04 -1.16360807e+00
-6.83219433e-01 -4.89943266e-01 -5.78405559e-01 7.54612327e-01
-1.74738884e+00 -8.16601396e-01 -1.74592491e-02 2.26776317e-01
8.71650457e-01 -3.05603415e-01 7.23630428e-01 4.04588431e-01
-3.78265381e-01 8.85397136e-01 1.74084213e-02 5.20847619e-01
1.32241464e+00 -1.53822172e+00 3.26255500e-01 4.47650492e-01
1.12826616e-01 8.71410668e-01 8.57678890e-01 -8.00733864e-01
-7.34179914e-01 -8.98345768e-01 1.26102507e+00 -3.89184326e-01
6.95281088e-01 -3.27346116e-01 -1.00698113e+00 7.37087429e-01
2.72297204e-01 -7.57248044e-01 1.06535470e+00 6.28765345e-01
-4.07878786e-01 1.41331762e-01 -1.29351139e+00 5.50231457e-01
7.14616001e-01 -5.11850417e-01 -1.11352623e+00 3.13036263e-01
1.05632102e+00 -6.44173443e-01 -9.55393255e-01 4.36212242e-01
5.23946702e-01 -4.48361754e-01 9.67022777e-01 -8.74071717e-01
2.51515865e-01 -1.34225830e-01 -3.53688598e-01 -1.33488679e+00
-3.87687571e-02 -4.07869846e-01 -2.61003494e-01 1.58895254e+00
8.17932308e-01 -6.18799627e-01 6.99387193e-01 1.26275444e+00
8.37860107e-02 -1.08267128e-01 -1.04069912e+00 -4.75163132e-01
-1.31915241e-01 -2.53720749e-02 7.26397693e-01 9.43939507e-01
4.35996115e-01 9.61206794e-01 -6.25273645e-01 3.54549646e-01
4.25100386e-01 2.51333356e-01 7.65282333e-01 -1.58572364e+00
-3.89261007e-01 -9.17092338e-02 7.12094530e-02 -1.74307096e+00
6.04012370e-01 -6.24898016e-01 2.01389641e-01 -1.38056004e+00
2.69736111e-01 -5.78424394e-01 -3.97982568e-01 3.69175047e-01
-6.35059297e-01 -5.34311414e-01 1.52910903e-01 4.52652633e-01
-1.06636381e+00 8.00839961e-01 7.19804347e-01 -1.66365802e-01
-4.17041570e-01 1.69394702e-01 -7.78978944e-01 6.24698818e-01
6.22043014e-01 -4.79308039e-01 -4.06709969e-01 -4.71900702e-01
-1.78316534e-01 5.25621176e-01 -7.99777359e-02 -6.38746798e-01
4.23192948e-01 -2.13030398e-01 1.49507135e-01 -5.04221678e-01
3.63529652e-01 -9.55730140e-01 -5.52446902e-01 2.76871249e-02
-9.49000955e-01 -2.02302098e-01 -4.90378663e-02 7.00532496e-01
-4.04693067e-01 -3.71276438e-01 6.36137784e-01 9.15707573e-02
-9.47262108e-01 3.41528766e-02 -8.07170346e-02 3.81141722e-01
5.83379805e-01 2.48021692e-01 -3.10836107e-01 -8.08698297e-01
-6.16162181e-01 8.17854524e-01 3.60018402e-01 5.63973010e-01
3.53967309e-01 -1.26980495e+00 -6.31604671e-01 -4.82215472e-02
3.89874458e-01 -3.47284488e-02 4.08367157e-01 3.29330385e-01
1.64690450e-01 7.43952155e-01 1.73970804e-01 -3.62394154e-01
-1.26031137e+00 -1.45630687e-01 2.31283456e-01 -1.04797971e+00
-1.23491973e-01 8.69896114e-01 6.79638505e-01 -1.03295660e+00
4.48346496e-01 -1.80481210e-01 -5.64694464e-01 1.92216486e-01
5.65292120e-01 1.54731467e-01 -1.46216482e-01 -5.58598340e-01
-3.43589127e-01 -6.19231015e-02 -5.08819580e-01 -4.94508535e-01
8.88092518e-01 -4.75984275e-01 4.55567300e-01 1.98470116e-01
9.36681926e-01 1.23783037e-01 -1.20170486e+00 -9.08054054e-01
5.16821802e-01 -2.52001137e-01 -3.01440030e-01 -1.19250965e+00
-3.91953886e-01 5.87681592e-01 4.80271995e-01 7.73704844e-03
7.49106407e-01 2.14165881e-01 1.26549411e+00 6.97538555e-01
3.76684904e-01 -1.31450224e+00 1.12821192e-01 1.06066573e+00
6.55696929e-01 -1.67349422e+00 -3.81868780e-01 -2.93523699e-01
-1.38595676e+00 8.74846637e-01 1.09706855e+00 2.32095063e-01
3.92053843e-01 1.45837711e-02 4.31455523e-01 6.05877079e-02
-8.56735349e-01 -1.32473260e-01 4.96435434e-01 3.57205808e-01
7.19719708e-01 1.36935234e-01 -4.65467066e-01 1.13814485e+00
-2.28878595e-02 -3.54680419e-01 3.20467323e-01 5.86900353e-01
-7.34556139e-01 -1.46939528e+00 -1.62486687e-01 3.42670977e-01
-3.68627042e-01 -2.18193829e-01 -8.00946832e-01 3.01909566e-01
-5.08430839e-01 1.22583187e+00 4.04119678e-02 -6.09758735e-01
4.42562878e-01 3.88938040e-01 -1.62591591e-01 -8.42864811e-01
-5.98297298e-01 -1.09197602e-01 4.92363960e-01 -3.90291244e-01
-1.47527322e-01 -4.97873753e-01 -1.53891027e+00 1.50114521e-01
-4.66056377e-01 6.15157843e-01 3.83106411e-01 9.62205470e-01
5.24177194e-01 2.97348469e-01 8.04026783e-01 -2.59573936e-01
-9.73278344e-01 -1.31097233e+00 -4.66612667e-01 5.84417880e-01
1.99130088e-01 -1.03556538e+00 -3.22328657e-01 -3.37699682e-01] | [12.612105369567871, 7.5748796463012695] |
c61f8d44-b695-46d0-bfc4-3596594962bc | transfer-of-fully-convolutional-policy-value | 2102.12375 | null | https://arxiv.org/abs/2102.12375v1 | https://arxiv.org/pdf/2102.12375v1.pdf | Transfer of Fully Convolutional Policy-Value Networks Between Games and Game Variants | In this paper, we use fully convolutional architectures in AlphaZero-like self-play training setups to facilitate transfer between variants of board games as well as distinct games. We explore how to transfer trained parameters of these architectures based on shared semantics of channels in the state and action representations of the Ludii general game system. We use Ludii's large library of games and game variants for extensive transfer learning evaluations, in zero-shot transfer experiments as well as experiments with additional fine-tuning time. | ['Olivier Teytaud', 'Cameron Browne', 'Matthew Stephenson', 'Eric Piette', 'Vegard Mella', 'Dennis J. N. J. Soemers'] | 2021-02-24 | null | null | null | null | ['board-games'] | ['playing-games'] | [-3.05742532e-01 1.95659339e-01 -5.26699005e-03 -1.00422092e-01
-6.18176281e-01 -8.03330243e-01 6.46802723e-01 -3.67468596e-01
-8.17282081e-01 8.71676564e-01 1.81985632e-01 -4.12646651e-01
9.81967598e-02 -1.11720347e+00 -6.50826454e-01 -1.15202680e-01
-3.66070330e-01 6.69407964e-01 7.67543137e-01 -1.11582661e+00
3.54058370e-02 -4.51053120e-02 -1.44394386e+00 6.09096706e-01
9.18555856e-02 6.74520552e-01 -3.15239257e-03 1.32936811e+00
1.51104942e-01 1.14788592e+00 -7.99548626e-01 -6.73814535e-01
6.08788073e-01 -4.07398760e-01 -1.12293160e+00 -3.01162332e-01
-1.41930491e-01 -6.72420561e-01 -8.89760315e-01 6.63811028e-01
5.92272282e-01 5.97928166e-01 1.88626453e-01 -1.10041308e+00
-3.30884218e-01 7.17254996e-01 3.16510051e-02 6.03533030e-01
3.78250718e-01 7.89066374e-01 1.06029320e+00 -1.80283904e-01
5.60390711e-01 9.21342909e-01 6.98258996e-01 8.47325027e-01
-9.85372186e-01 -8.47419500e-01 -2.03726411e-01 1.63963914e-01
-1.17963982e+00 -6.11305416e-01 2.18149334e-01 -2.54672438e-01
1.67033029e+00 -1.85631692e-01 9.01392043e-01 1.51870060e+00
4.29985583e-01 7.34571815e-01 7.17718363e-01 -2.05345273e-01
2.36470953e-01 -3.24414492e-01 -1.86238382e-02 8.73872817e-01
-1.21973060e-01 4.48227853e-01 -3.73441458e-01 -9.16521996e-02
1.38284302e+00 -5.37984252e-01 1.77673802e-01 -3.18231255e-01
-7.99588203e-01 1.07681632e+00 3.92527819e-01 4.00837250e-02
-1.59478232e-01 7.42557228e-01 9.26406801e-01 8.21659446e-01
2.90569842e-01 7.84919739e-01 -5.12575686e-01 -7.63941765e-01
-1.80610374e-01 6.82356954e-01 9.98591185e-01 8.97105515e-01
8.75114501e-01 2.76801467e-01 -2.02888861e-01 6.00615978e-01
-1.47042111e-01 -5.86376525e-02 8.15729201e-01 -1.01745915e+00
5.29594958e-01 8.51685703e-02 1.11407548e-01 -2.92169392e-01
-5.31917632e-01 -1.14886582e-01 -2.54813880e-01 4.19156313e-01
2.62273788e-01 -6.75632238e-01 -8.14053416e-01 1.88348639e+00
-2.98474818e-01 8.31459403e-01 5.06871164e-01 6.99089527e-01
1.00611472e+00 3.65332514e-01 2.89182365e-01 7.29772329e-01
1.14006829e+00 -9.19393063e-01 -7.79413432e-02 -6.15462959e-01
1.07827032e+00 -1.43502578e-01 1.28926551e+00 2.26423949e-01
-1.38091886e+00 -6.91183627e-01 -1.11582708e+00 1.88154895e-02
-2.69442827e-01 -2.21360654e-01 1.23966372e+00 5.22332728e-01
-1.41817892e+00 1.00920653e+00 -1.18052137e+00 -4.81876284e-01
3.36447835e-01 6.08255208e-01 -5.64843774e-01 3.59685421e-01
-1.78987288e+00 9.46179032e-01 8.06925297e-01 -4.10370559e-01
-1.19559419e+00 -5.65784395e-01 -1.03926575e+00 3.25836182e-01
3.05996120e-01 -7.95295894e-01 1.85796785e+00 -7.69069195e-01
-2.29329300e+00 1.06940711e+00 9.01696742e-01 -7.37415552e-01
1.80703580e-01 -1.22206502e-01 -2.70540267e-01 4.36781440e-03
6.62102476e-02 6.39953852e-01 2.85200387e-01 -1.96711734e-01
-6.57889068e-01 1.08730689e-01 1.15670431e+00 5.21636963e-01
5.09645455e-02 5.64208031e-02 -5.22908211e-01 -1.95695609e-01
-6.82591438e-01 -1.03125238e+00 -3.32121640e-01 -4.48279440e-01
-1.01343513e-01 -4.12829258e-02 2.02625543e-01 -2.28513837e-01
6.61184251e-01 -2.10453343e+00 6.74356744e-02 -1.14961877e-01
1.58152044e-01 3.43836129e-01 -5.32900333e-01 5.74722767e-01
-2.80752748e-01 -1.45419851e-01 4.47384775e-01 -2.08095044e-01
1.97923720e-01 5.98382294e-01 -1.70054287e-01 1.57053158e-01
1.37590393e-01 1.27750289e+00 -9.42812026e-01 1.25785470e-01
3.69635761e-01 -3.14203113e-01 -1.08878374e+00 4.83532637e-01
-2.10107997e-01 3.93440127e-01 -3.73952717e-01 -1.83144554e-01
5.82816750e-02 -6.85683638e-02 5.36809340e-02 2.04450458e-01
1.50629535e-01 8.40343833e-01 -8.79180610e-01 2.47112465e+00
-7.04914927e-01 5.11082947e-01 -2.91071273e-02 -9.22303140e-01
5.11148095e-01 4.64722216e-01 1.59423083e-01 -6.86668932e-01
5.89428544e-01 -2.57473767e-01 3.14653546e-01 -9.66344923e-02
7.36146212e-01 -4.76194948e-01 -6.73954308e-01 6.56791747e-01
9.22032773e-01 -4.32984740e-01 2.87450820e-01 3.54864269e-01
1.60392070e+00 3.58540565e-01 4.92630720e-01 -2.76754230e-01
-9.03606340e-02 4.55642715e-02 3.12076174e-02 1.09582472e+00
-2.56053329e-01 3.72980684e-01 6.74753428e-01 -2.93951958e-01
-8.59505951e-01 -1.09911001e+00 4.51206893e-01 1.66170466e+00
6.67850375e-02 -9.63107407e-01 -5.28429508e-01 -3.85537624e-01
-3.49993765e-01 7.45779276e-01 -8.64173591e-01 -8.37869406e-01
-3.57135683e-01 -3.65802646e-01 1.17193508e+00 8.84306192e-01
6.95201099e-01 -1.43807518e+00 -6.32329881e-01 4.16342914e-01
2.40418389e-01 -8.31028461e-01 -1.52469501e-02 7.44074225e-01
-3.22135657e-01 -1.25369394e+00 -2.46360213e-01 -5.36679387e-01
-4.35423970e-01 1.24517418e-01 1.48481750e+00 -6.63637184e-03
-1.44511878e-01 2.66535968e-01 -4.02928531e-01 -2.54914284e-01
-4.69523907e-01 6.57599032e-01 -1.76616400e-01 -8.31071615e-01
1.87478453e-01 -8.41129303e-01 -3.33635658e-01 1.93100125e-01
-7.67253220e-01 2.85825431e-01 3.12975824e-01 8.75879109e-01
-2.13273048e-01 -2.73666471e-01 9.55792218e-02 -9.23691034e-01
1.05309308e+00 -5.92746258e-01 -7.39962757e-01 -2.07331449e-01
2.65840918e-01 7.60461241e-02 5.96457839e-01 -1.80954844e-01
-8.78717005e-01 -5.44542849e-01 -6.01886451e-01 -3.78776312e-01
-2.23242611e-01 6.09827697e-01 -3.19211893e-02 -7.72345513e-02
1.09826553e+00 1.50822494e-02 -6.83037937e-02 -2.49659926e-01
7.81039238e-01 3.40852648e-01 6.88517690e-01 -9.19482350e-01
5.44891179e-01 -1.99027527e-02 -4.51839536e-01 -4.70432192e-01
-5.70680678e-01 -3.87128860e-01 -4.57675546e-01 1.82616979e-01
9.37906444e-01 -1.26390660e+00 -1.06365919e+00 6.65674090e-01
-7.73818433e-01 -1.38737881e+00 -4.96488541e-01 3.87745589e-01
-1.07215869e+00 -2.22303703e-01 -1.18594348e+00 -3.25464606e-02
-1.86086316e-02 -1.16966379e+00 8.48125517e-01 3.94574344e-01
-2.63030320e-01 -1.20093107e+00 7.67700672e-01 -1.52478203e-01
4.41625208e-01 -3.35815012e-01 5.11713862e-01 -8.77315044e-01
-1.35076642e-01 -3.06733519e-01 4.26228829e-02 1.19606674e-01
-3.82469445e-02 -3.74393225e-01 -8.09558392e-01 -4.32437509e-01
-5.55042982e-01 -1.17279339e+00 9.10293102e-01 1.85531497e-01
8.10320377e-01 1.41272411e-01 -1.18386503e-02 1.05028188e+00
1.12747037e+00 1.91314742e-02 1.02294695e+00 6.11253202e-01
4.60457593e-01 1.48800071e-02 2.69469887e-01 5.58393717e-01
4.48441684e-01 7.20261097e-01 4.32426870e-01 -7.75394812e-02
9.18491483e-02 -2.46473834e-01 4.87994701e-01 4.08306271e-01
-2.90264249e-01 -2.96628684e-01 -7.22755373e-01 3.46409470e-01
-1.85851121e+00 -1.23677325e+00 4.21972394e-01 1.87167120e+00
8.28508198e-01 5.11405706e-01 3.50181818e-01 -4.32256788e-01
3.32085639e-01 3.69907349e-01 -3.55352998e-01 -8.36161077e-01
9.51859057e-02 1.26846826e+00 6.99636579e-01 4.51054364e-01
-1.25552237e+00 1.76841104e+00 7.78100395e+00 8.65680039e-01
-1.02704275e+00 2.98338562e-01 1.59085259e-01 -3.48967910e-01
2.05538094e-01 1.46844223e-01 -4.27911133e-01 1.17465340e-01
1.54429901e+00 -4.21818465e-01 6.24124110e-01 8.54972482e-01
-3.66292685e-01 -4.19854373e-02 -9.05794919e-01 6.96647465e-01
-5.22577882e-01 -1.60123968e+00 -4.78100955e-01 -8.99162665e-02
5.68323553e-01 6.38827443e-01 -1.29134133e-01 1.33468819e+00
1.55526495e+00 -1.10839891e+00 4.94123638e-01 -2.26694290e-02
1.03604031e+00 -8.44806731e-01 6.17400944e-01 1.33926436e-01
-1.09523058e+00 7.15216622e-02 -7.08606422e-01 -9.63083029e-01
-1.72578126e-01 -7.43393719e-01 -7.07573235e-01 5.95817745e-01
4.09555465e-01 8.51296842e-01 -3.85143191e-01 7.16216147e-01
-5.65593719e-01 8.02734435e-01 -2.20850527e-01 1.65425137e-01
6.90349519e-01 1.21180862e-02 7.25555345e-02 9.43053961e-01
-1.40839010e-01 2.92699575e-01 3.19212496e-01 6.02923393e-01
-1.34608477e-01 -3.15113932e-01 -7.83022285e-01 -9.14494023e-02
1.14392221e-01 1.19333136e+00 -4.25813854e-01 -4.31220263e-01
-5.83423257e-01 1.29451418e+00 7.11720884e-01 3.05107802e-01
-1.08570433e+00 -5.46029508e-01 1.42016542e+00 -1.66538022e-02
4.52442050e-01 -2.54413605e-01 2.49248445e-01 -1.51558042e+00
-8.91921818e-01 -6.85167491e-01 4.17094350e-01 -9.49600041e-01
-9.95162606e-01 6.73548222e-01 2.11479038e-01 -1.01770592e+00
-8.71705294e-01 -7.10315943e-01 -1.32798195e+00 9.03919399e-01
-9.28306878e-01 -1.10265946e+00 -1.78752974e-01 1.03607285e+00
2.54695326e-01 -6.16664886e-01 1.20345175e+00 -9.53034777e-03
-5.78259051e-01 7.74891376e-01 2.07516719e-02 4.73058224e-01
6.54263020e-01 -1.26293254e+00 1.29973471e+00 4.74190801e-01
8.36430192e-02 4.36123639e-01 6.13718629e-01 -2.61476755e-01
-8.81333888e-01 -9.96211171e-01 -3.18771750e-01 -3.18718046e-01
1.29114437e+00 -6.01638258e-01 -7.31727660e-01 1.31410456e+00
4.55185413e-01 -1.50175001e-02 7.38054156e-01 5.57809591e-01
-2.67024606e-01 3.80895644e-01 -5.65902472e-01 8.06199789e-01
1.32050395e+00 -8.70059133e-01 -6.49088383e-01 8.29487070e-02
6.41132414e-01 -8.10196579e-01 -6.83588803e-01 -2.38267332e-01
3.37556064e-01 -1.04736519e+00 8.69809628e-01 -1.31543183e+00
6.88077211e-01 2.76705325e-01 -2.47166082e-02 -1.88863301e+00
-6.61540151e-01 -7.14799404e-01 3.88821572e-01 6.55438364e-01
1.52980387e-01 -5.28303325e-01 1.18064201e+00 4.45871383e-01
-4.26652044e-01 -2.93751925e-01 -8.91163409e-01 -6.77128017e-01
5.33113062e-01 -5.24338245e-01 5.54363608e-01 6.38602912e-01
5.07801235e-01 7.02122569e-01 -3.81884962e-01 -2.47067362e-01
-1.67653244e-02 -2.69015938e-01 1.30050063e+00 -9.56493735e-01
-1.04849935e+00 -5.14285266e-01 -9.53652203e-01 -1.02750170e+00
5.51373720e-01 -8.94128263e-01 9.65092257e-02 -1.15900922e+00
3.85716744e-02 -1.82471007e-01 -4.81011420e-01 8.00496697e-01
-2.03251451e-01 4.62906957e-01 2.69014567e-01 1.25897646e-01
-1.11180532e+00 5.96903801e-01 8.96791637e-01 5.23883067e-02
-8.07012916e-02 2.86916848e-02 -8.80889297e-01 6.26869142e-01
1.00557053e+00 -1.77249849e-01 -7.59470344e-01 -6.05004132e-01
1.81134611e-01 3.92667323e-01 4.34027791e-01 -1.40004790e+00
5.80526935e-03 -2.05441847e-01 6.09228760e-02 2.08757341e-01
6.27615094e-01 -2.27318089e-02 -6.65457100e-02 3.59754086e-01
-4.23789978e-01 -1.67676974e-02 7.98939586e-01 3.31008166e-01
-3.10062021e-02 -1.18896544e-01 7.04525530e-01 -5.79722881e-01
-1.26786792e+00 3.16226870e-01 -8.12862158e-01 3.68184179e-01
9.42327023e-01 -2.33548030e-01 -3.65528107e-01 -8.70762765e-01
-1.09034264e+00 1.04587086e-01 6.38914764e-01 4.12050664e-01
1.63448110e-01 -1.18634546e+00 -5.63590825e-01 2.62940377e-01
3.29917163e-01 -5.05514979e-01 4.29695904e-01 3.54674101e-01
-8.41578066e-01 2.94004530e-01 -1.03443182e+00 6.53362051e-02
-8.31086993e-01 1.43488303e-01 6.38351798e-01 -7.27074564e-01
-6.73616350e-01 1.12574196e+00 5.91541886e-01 -7.04887033e-01
-2.74961740e-01 -1.57048911e-01 -1.39709517e-01 -5.95863819e-01
5.17022073e-01 -8.10202807e-02 2.06757113e-01 -1.81294754e-01
-1.55821368e-01 -2.52548128e-01 -1.32410109e-01 -4.92388546e-01
1.21679187e+00 2.47537747e-01 4.95116413e-01 3.88557196e-01
9.42014515e-01 -5.76017916e-01 -1.51163852e+00 -8.06996971e-02
-3.34401757e-01 -9.36494321e-02 6.46518320e-02 -7.00888515e-01
-9.27149355e-01 7.83707440e-01 2.82925576e-01 -1.44036531e-01
6.41904175e-01 -4.32295986e-02 6.59260690e-01 6.40702784e-01
8.13445508e-01 -8.76461029e-01 2.07624584e-01 1.10566115e+00
4.35168982e-01 -9.78462279e-01 -4.17617202e-01 1.70595154e-01
-8.18738282e-01 8.15938771e-01 9.73053455e-01 -9.29358959e-01
5.31437516e-01 4.73713368e-01 5.73149137e-02 -3.57781589e-01
-1.37128663e+00 -5.87468386e-01 -3.02827030e-01 1.01267743e+00
3.93669456e-01 5.85488491e-02 -6.16501598e-03 1.05949187e+00
-4.74718064e-01 1.99279934e-01 8.88030410e-01 1.08536077e+00
-3.13810676e-01 -1.13806808e+00 1.24489747e-01 3.87300909e-01
-3.06194723e-01 -3.31723303e-01 -2.93487310e-01 1.17664826e+00
-2.16438055e-01 6.99979663e-01 4.34993863e-01 -8.50672126e-01
4.23387617e-01 -7.92542323e-02 9.38318968e-01 -1.18575287e+00
-1.09372687e+00 -4.26578611e-01 4.68841314e-01 -9.29010451e-01
1.86806947e-01 -3.27532552e-02 -1.27173901e+00 -7.82241464e-01
-1.20185107e-01 1.52699262e-01 1.50022820e-01 1.03476369e+00
2.59054452e-01 9.20162439e-01 7.40473866e-02 -1.27740884e+00
-6.56789720e-01 -1.22390664e+00 -1.02777278e+00 5.12990236e-01
-4.27728266e-01 -8.50436568e-01 1.14138313e-01 -4.93449628e-01] | [3.6667985916137695, 1.4573943614959717] |
7741aedc-5141-4c3c-af28-d8283deda82d | modular-approach-to-machine-reading | 2210.01750 | null | https://arxiv.org/abs/2210.01750v1 | https://arxiv.org/pdf/2210.01750v1.pdf | Modular Approach to Machine Reading Comprehension: Mixture of Task-Aware Experts | In this work we present a Mixture of Task-Aware Experts Network for Machine Reading Comprehension on a relatively small dataset. We particularly focus on the issue of common-sense learning, enforcing the common ground knowledge by specifically training different expert networks to capture different kinds of relationships between each passage, question and choice triplet. Moreover, we take inspi ration on the recent advancements of multitask and transfer learning by training each network a relevant focused task. By making the mixture-of-networks aware of a specific goal by enforcing a task and a relationship, we achieve state-of-the-art results and reduce over-fitting. | ['Gabriel Bayomi Tinoco Kalejaiye', 'Anusha Kamath', 'Anirudha Rayasam'] | 2022-10-04 | null | null | null | null | ['machine-reading-comprehension', 'common-sense-reasoning'] | ['natural-language-processing', 'reasoning'] | [ 4.01348561e-01 2.08936855e-01 -1.22789718e-01 -5.56997597e-01
-9.18676972e-01 -6.45281255e-01 5.19571424e-01 3.05052161e-01
-7.12452590e-01 6.43471539e-01 3.87981772e-01 -4.53383178e-01
-4.54582930e-01 -6.41207218e-01 -9.26292896e-01 -1.21582128e-01
4.31442916e-01 7.53241479e-01 5.19248903e-01 -4.95541900e-01
2.06212595e-01 -9.87223983e-02 -1.18415165e+00 5.26338339e-01
1.36578023e+00 8.96749973e-01 3.08214277e-01 6.52220964e-01
-8.45709667e-02 1.04437625e+00 -5.03977239e-01 -8.74719203e-01
-9.77387875e-02 -3.74344766e-01 -1.30141270e+00 -9.88479853e-02
9.29238439e-01 7.18186647e-02 -6.75675496e-02 1.01392853e+00
3.37583095e-01 5.33589661e-01 8.78185511e-01 -9.51035321e-01
-8.78720999e-01 8.01583469e-01 -2.12668404e-01 4.10099715e-01
4.42040682e-01 -1.13676228e-01 1.25800967e+00 -6.38812423e-01
3.31286877e-01 1.07506335e+00 6.26228511e-01 5.34150064e-01
-1.20781362e+00 -3.06627423e-01 6.16453409e-01 7.74603128e-01
-1.06090260e+00 -4.27573174e-01 8.16571712e-01 -4.69384372e-01
1.09991241e+00 1.09405935e-01 1.74235001e-01 1.50700653e+00
1.11358456e-01 8.33042860e-01 1.24324381e+00 -8.86373103e-01
7.85303488e-02 -2.77380347e-02 7.75182188e-01 5.84991872e-01
3.29275653e-02 -2.72737205e-01 -6.55783474e-01 7.08103180e-02
4.93203312e-01 -2.42303506e-01 -5.81868052e-01 -1.61189616e-01
-1.15287101e+00 6.81128025e-01 8.07959512e-02 5.17515004e-01
-3.56319159e-01 -2.48061195e-01 2.93620259e-01 6.32160127e-01
4.20050323e-01 7.56392539e-01 -9.61130559e-01 -1.76949762e-02
-8.98203671e-01 2.18146712e-01 1.34890699e+00 8.30891669e-01
7.50338674e-01 -4.12227422e-01 -7.81632185e-01 7.96921790e-01
1.23294666e-01 1.44074202e-01 3.49333823e-01 -1.03711879e+00
7.24820793e-01 4.86321539e-01 2.24828809e-01 -8.03071797e-01
-6.98293984e-01 -6.90667808e-01 -7.44024217e-01 -2.11327925e-01
6.67476654e-01 -3.37035537e-01 -8.58201206e-01 1.95456076e+00
5.70596047e-02 2.60731548e-01 3.33870575e-02 7.47876287e-01
7.06885397e-01 3.02165002e-01 4.38057840e-01 -4.75519821e-02
1.61393714e+00 -1.50862551e+00 -8.82525682e-01 -5.56178629e-01
4.67904806e-01 -5.60652733e-01 1.38259780e+00 4.68326122e-01
-1.20132518e+00 -7.61034489e-01 -8.46552491e-01 -3.62156004e-01
-4.53985065e-01 -1.19362539e-02 2.54214972e-01 1.46001026e-01
-1.04213738e+00 2.90749997e-01 -3.18466902e-01 -2.56782770e-01
1.90169141e-01 -3.31178494e-02 -1.47850245e-01 -2.12777257e-01
-1.64747524e+00 1.46327293e+00 5.61322927e-01 -2.29954123e-01
-8.98498476e-01 -1.03657651e+00 -7.60394096e-01 3.56014222e-01
8.79598677e-01 -1.24017262e+00 1.43345475e+00 -9.49584365e-01
-1.62459207e+00 1.14320910e+00 -1.83225051e-01 -3.21613699e-01
2.52846628e-01 -6.83338284e-01 -6.04387403e-01 1.89126715e-01
1.13976263e-01 2.44711712e-01 7.76877701e-01 -1.13409710e+00
-5.96128106e-01 -2.36126691e-01 3.16116214e-01 2.37397894e-01
-2.57320136e-01 2.51684368e-01 -3.61494184e-01 -5.94057560e-01
-3.40607703e-01 -5.62941909e-01 2.69164927e-02 -3.00193518e-01
-4.22858417e-01 -8.70359898e-01 2.08867788e-01 -6.81914806e-01
1.19204640e+00 -1.80265450e+00 5.74089229e-01 -1.80124030e-01
4.53918368e-01 4.48975205e-01 -4.97628033e-01 3.76455307e-01
4.00118865e-02 8.18271935e-02 -1.04064539e-01 -5.87329805e-01
1.78774923e-01 2.33995184e-01 -1.48580864e-01 -2.58932523e-02
1.81301966e-01 9.25620496e-01 -1.14816070e+00 -4.96904641e-01
-1.67462200e-01 1.80551633e-01 -4.44008768e-01 6.14914536e-01
-6.93997264e-01 4.60415214e-01 -6.70038760e-01 1.23620488e-01
4.07097638e-01 -6.23349547e-01 1.94625288e-01 -2.03767657e-01
8.84256661e-02 7.90283918e-01 -9.26288307e-01 2.21606588e+00
-6.91467464e-01 3.56850445e-01 1.53078929e-01 -1.25822937e+00
6.66607022e-01 5.07191658e-01 5.75612150e-02 -6.92175865e-01
2.28095055e-01 -2.89143398e-02 1.12290934e-01 -8.58421862e-01
3.23574871e-01 -5.12551591e-02 3.36726047e-02 7.13407278e-01
6.88316464e-01 9.52561200e-02 2.32980594e-01 2.18598798e-01
8.93874884e-01 1.80067882e-01 2.95189917e-01 -4.90204364e-01
5.14880657e-01 -1.44230306e-01 2.50040203e-01 1.16983712e+00
-1.98323071e-01 4.91231114e-01 3.38851929e-01 -2.04584286e-01
-7.27555335e-01 -9.37165260e-01 4.61635254e-02 1.99736345e+00
-1.01405725e-01 -2.79360235e-01 -6.65723443e-01 -8.16138983e-01
-2.23519325e-01 1.17998838e+00 -8.90008509e-01 -1.59790009e-01
-6.44585431e-01 -3.81274909e-01 5.10771275e-01 4.96533245e-01
7.78338432e-01 -1.01278710e+00 -6.86966300e-01 1.70798451e-01
-4.97957975e-01 -1.35499465e+00 -4.54561681e-01 2.56614298e-01
-3.82684141e-01 -1.24544382e+00 -5.16827822e-01 -8.95115197e-01
1.47644818e-01 8.89912397e-02 2.10982919e+00 1.34914756e-01
4.38921958e-01 5.66205800e-01 -5.38691103e-01 -6.19983673e-01
-2.44982734e-01 5.93405902e-01 -2.44560167e-01 3.94533016e-02
5.93010247e-01 -6.31436825e-01 -2.66850054e-01 1.50046796e-01
-7.26196051e-01 2.97945254e-02 3.70858908e-01 7.57665992e-01
2.21329227e-01 -9.15394649e-02 5.94114780e-01 -9.91312861e-01
1.05552912e+00 -6.64712965e-01 -1.80361047e-01 1.06872356e+00
-4.93814856e-01 1.79350853e-01 5.54751098e-01 -4.75791544e-01
-1.56526089e+00 -6.38738334e-01 3.72238755e-02 -2.21563026e-01
-4.05279577e-01 6.95129693e-01 -2.09393233e-01 3.48210573e-01
8.13552499e-01 6.95261434e-02 -4.05928314e-01 -5.63086629e-01
6.77730918e-01 2.60700554e-01 5.57038307e-01 -9.96195257e-01
5.89788198e-01 -1.06447503e-01 -3.22992802e-01 -3.29433888e-01
-2.01443696e+00 -4.60289478e-01 -7.98037708e-01 5.38901798e-02
1.21406960e+00 -7.84789562e-01 -5.29098034e-01 5.59853971e-01
-1.37950337e+00 -6.33657217e-01 -2.14330867e-01 2.71925837e-01
-4.41267312e-01 1.20250836e-01 -5.58705032e-01 -3.52000862e-01
-3.70050430e-01 -8.32235217e-01 6.57117486e-01 3.65024418e-01
-2.50972390e-01 -1.63734019e+00 3.45440775e-01 5.62248409e-01
7.23264813e-01 -2.04763755e-01 1.19118893e+00 -1.31116295e+00
-2.86581635e-01 2.00252414e-01 -3.13646406e-01 2.98414409e-01
-1.69881457e-03 -3.90577495e-01 -9.38331068e-01 2.82332785e-02
2.56884634e-01 -7.00944662e-01 1.03406775e+00 2.92327493e-01
1.22282195e+00 -2.12318450e-01 -2.22840682e-01 4.42778617e-01
1.14925945e+00 -2.81891555e-01 3.87720168e-01 3.31840515e-01
7.24142492e-01 8.46026957e-01 1.77191451e-01 -3.33320722e-02
1.01445913e+00 6.16145909e-01 9.66216028e-02 4.57993448e-02
-2.23686859e-01 -2.78575033e-01 -4.94158901e-02 8.77058387e-01
-1.36297107e-01 -5.27124941e-01 -1.01895046e+00 7.88867772e-01
-2.01250529e+00 -8.25289667e-01 -9.32979286e-02 1.79304636e+00
1.50383198e+00 6.71395957e-02 -2.43874677e-02 -3.96990806e-01
5.68391502e-01 2.52842575e-01 -3.71814489e-01 -3.41242939e-01
-2.02286229e-01 4.85666722e-01 9.06967372e-02 7.23680496e-01
-1.04287744e+00 9.74123061e-01 7.04206276e+00 7.62593925e-01
-6.37557507e-01 5.15585542e-01 2.23376751e-01 3.12238693e-01
-4.68385160e-01 -9.67144817e-02 -8.07934105e-01 1.39899001e-01
1.03243995e+00 7.79126026e-03 3.57804835e-01 3.24694425e-01
-1.79424107e-01 -1.36236742e-01 -1.24672842e+00 4.16106999e-01
3.59650314e-01 -1.01970506e+00 1.36487171e-01 -5.91723919e-01
8.59284699e-01 1.69917464e-01 -1.36382654e-01 5.22364259e-01
6.27828717e-01 -1.10217476e+00 4.68338847e-01 1.03666222e+00
3.64256263e-01 -2.27663279e-01 5.18845022e-01 6.57435238e-01
-5.77701926e-01 4.64350618e-02 -2.25653633e-01 -3.47657740e-01
2.27460504e-01 5.88259459e-01 -2.24810109e-01 7.83839524e-01
6.87238574e-01 5.51556826e-01 -6.19077146e-01 8.29372525e-01
-7.42398560e-01 9.22977209e-01 -5.11532500e-02 5.03734984e-02
1.84216753e-01 -6.89704493e-02 4.60148722e-01 1.34121442e+00
-1.72167361e-01 1.60983384e-01 5.20838439e-01 1.15566444e+00
-2.84008235e-01 -3.82780433e-02 -2.16726318e-01 4.35513765e-01
4.67291474e-01 1.09894931e+00 -2.03930717e-02 -6.07399285e-01
-6.73152030e-01 9.12951648e-01 9.64574039e-01 5.27158022e-01
-5.20740509e-01 -2.40991786e-01 3.45894009e-01 -2.03746110e-01
4.22815681e-01 -1.02604143e-01 -4.19631839e-01 -1.56051040e+00
1.37020061e-02 -1.05848563e+00 6.12203419e-01 -8.25956345e-01
-1.77874947e+00 5.02091885e-01 2.80203640e-01 -5.43103576e-01
-1.21119298e-01 -7.75600910e-01 -9.90284264e-01 1.11391878e+00
-1.96299708e+00 -1.19613886e+00 -3.19387019e-01 8.43651354e-01
5.53892374e-01 -1.51976094e-01 1.03040373e+00 1.77319758e-02
-6.37529075e-01 7.12162912e-01 -2.77563989e-01 9.81918722e-02
9.99771953e-01 -1.48307455e+00 1.48337454e-01 7.87837684e-01
2.65004009e-01 8.24062169e-01 8.42550218e-01 -4.12309110e-01
-9.24210846e-01 -8.11565399e-01 1.27101099e+00 -9.89133120e-01
9.49234068e-01 -1.00639418e-01 -1.34688056e+00 1.06141710e+00
9.73964274e-01 -2.79023290e-01 8.85567129e-01 9.19858634e-01
-8.38301539e-01 1.12001620e-01 -8.69497359e-01 3.69642466e-01
1.01962662e+00 -8.25475156e-01 -1.40964031e+00 5.39239109e-01
8.63734484e-01 -5.05135596e-01 -1.02159321e+00 2.84946054e-01
6.89020380e-02 -9.08226311e-01 8.15449953e-01 -1.20350325e+00
6.47852778e-01 1.42409578e-01 1.97335407e-02 -1.83208525e+00
-5.58355629e-01 -5.08561015e-01 -3.26515138e-01 1.14511323e+00
6.88471198e-01 -4.47759509e-01 1.86129138e-01 7.42251396e-01
-3.60184312e-01 -6.48427665e-01 -7.86952257e-01 -5.91940165e-01
6.82579696e-01 -3.24024320e-01 2.67499179e-01 1.17801452e+00
1.85672313e-01 9.42575157e-01 -3.25892091e-01 1.19588621e-01
4.75026160e-01 7.42117837e-02 3.44062328e-01 -1.40820396e+00
-3.84103805e-01 -5.41109383e-01 4.78253663e-01 -1.33464980e+00
5.95859587e-01 -8.20031464e-01 2.29340736e-02 -1.48496377e+00
3.34109217e-01 -1.21222235e-01 -7.77203143e-01 5.68143189e-01
-8.61331224e-01 -2.14360937e-01 9.79461968e-02 -5.55765927e-02
-1.28907168e+00 4.11573946e-01 1.41207862e+00 -7.41498619e-02
9.56541225e-02 -9.36628953e-02 -9.17878449e-01 6.83806360e-01
8.32415164e-01 -3.45856607e-01 -5.15383005e-01 -1.11492217e+00
5.46247900e-01 -7.92877451e-02 3.93474549e-01 -7.18507707e-01
5.74899495e-01 -1.19395643e-01 -2.10074186e-02 -1.87643915e-01
1.91950738e-01 -6.34443343e-01 -5.02755761e-01 7.47326715e-03
-8.65301788e-01 3.30570266e-02 2.14757696e-01 4.65865791e-01
-2.17978463e-01 -4.55579817e-01 4.36545223e-01 -3.80810767e-01
-8.15850914e-01 8.34532175e-03 -1.18405037e-01 8.06910336e-01
5.67050517e-01 2.60280222e-01 -6.27778590e-01 -3.34719211e-01
-7.35713422e-01 5.68759441e-01 1.41647786e-01 5.67591012e-01
2.18518972e-01 -8.38319659e-01 -1.11461699e+00 -1.45578235e-01
2.46830255e-01 3.73767056e-02 3.40693325e-01 7.85979986e-01
2.89934158e-01 6.24537945e-01 -2.80981749e-01 -2.63140410e-01
-6.84530795e-01 6.22821450e-01 5.57561815e-01 -8.25940728e-01
-2.65979469e-01 1.08077526e+00 1.38198301e-01 -7.13047147e-01
1.57982379e-01 -2.27362096e-01 -6.93556368e-01 7.54465386e-02
4.08697724e-01 3.34293455e-01 6.45444170e-02 -1.55487895e-01
-1.26249060e-01 6.34233236e-01 -3.32457125e-01 1.32506058e-01
1.11215019e+00 -2.38371819e-01 -1.86646521e-01 6.43347025e-01
7.39065409e-01 -1.69141710e-01 -1.31984997e+00 -6.99025214e-01
2.14019954e-01 1.36091322e-01 3.35162021e-02 -1.55803561e+00
-7.55487859e-01 8.44312429e-01 -7.87213519e-02 2.66749799e-01
1.05134523e+00 1.91858545e-01 5.44946969e-01 6.27809525e-01
1.45785928e-01 -1.10287642e+00 1.88823834e-01 1.08480644e+00
8.69693160e-01 -1.48532712e+00 -2.92679638e-01 -3.39694738e-01
-6.06587827e-01 9.87283349e-01 9.21534598e-01 -1.63284257e-01
6.62674963e-01 -1.96689591e-01 1.39943227e-01 -4.40116376e-01
-9.38007951e-01 -3.12176436e-01 8.16887438e-01 6.31198525e-01
6.26141250e-01 8.62505883e-02 -2.91690558e-01 8.91809642e-01
-3.91740799e-02 9.62972641e-02 2.73114204e-01 6.95158124e-01
-4.47420925e-01 -1.07646513e+00 -2.67529469e-02 3.98230791e-01
-5.59554458e-01 -4.62585777e-01 -2.68138319e-01 7.30545104e-01
9.40489955e-03 1.17104268e+00 -3.09590325e-02 -1.79141704e-02
5.52661657e-01 4.62767512e-01 7.10911572e-01 -8.07384729e-01
-7.42105782e-01 -5.51283002e-01 3.09730828e-01 -3.06654841e-01
-7.41732359e-01 -3.13884646e-01 -7.31583774e-01 5.64139001e-02
-2.56661415e-01 2.17977777e-01 7.01053813e-02 1.64974582e+00
4.73768979e-01 7.86236763e-01 6.19199872e-02 -2.28099018e-01
-9.55478847e-01 -1.46074820e+00 -3.17594945e-01 7.60845542e-01
3.38604480e-01 -6.14600599e-01 -2.63268054e-01 8.63033235e-02] | [10.87026596069336, 8.32504940032959] |
51f04347-ab1b-4447-95a4-4db8ca8666a8 | piecewise-classifier-mappings-learning-fine | 1805.04288 | null | https://arxiv.org/abs/1805.04288v2 | https://arxiv.org/pdf/1805.04288v2.pdf | Piecewise classifier mappings: Learning fine-grained learners for novel categories with few examples | Humans are capable of learning a new fine-grained concept with very little supervision, \emph{e.g.}, few exemplary images for a species of bird, yet our best deep learning systems need hundreds or thousands of labeled examples. In this paper, we try to reduce this gap by studying the fine-grained image recognition problem in a challenging few-shot learning setting, termed few-shot fine-grained recognition (FSFG). The task of FSFG requires the learning systems to build classifiers for novel fine-grained categories from few examples (only one or less than five). To solve this problem, we propose an end-to-end trainable deep network which is inspired by the state-of-the-art fine-grained recognition model and is tailored for the FSFG task. Specifically, our network consists of a bilinear feature learning module and a classifier mapping module: while the former encodes the discriminative information of an exemplar image into a feature vector, the latter maps the intermediate feature into the decision boundary of the novel category. The key novelty of our model is a "piecewise mappings" function in the classifier mapping module, which generates the decision boundary via learning a set of more attainable sub-classifiers in a more parameter-economic way. We learn the exemplar-to-classifier mapping based on an auxiliary dataset in a meta-learning fashion, which is expected to be able to generalize to novel categories. By conducting comprehensive experiments on three fine-grained datasets, we demonstrate that the proposed method achieves superior performance over the competing baselines. | ['Jianxin Wu', 'Xiu-Shen Wei', 'Chunhua Shen', 'Peng Wang', 'Lingqiao Liu'] | 2018-05-11 | null | null | null | null | ['fine-grained-image-recognition'] | ['computer-vision'] | [ 3.52153718e-01 9.65888351e-02 -1.07818104e-01 -6.97983205e-01
-6.38971686e-01 -5.01219571e-01 7.89755642e-01 -7.47376159e-02
-3.78663987e-01 6.55091882e-01 -3.25632654e-02 1.96857035e-01
-1.92903653e-01 -1.05595219e+00 -9.49961364e-01 -6.16738021e-01
9.84408110e-02 3.71019781e-01 3.16205770e-01 -1.47308260e-01
1.43098429e-01 4.52004820e-01 -2.01940393e+00 5.23364723e-01
8.92180920e-01 1.36180663e+00 2.85606325e-01 4.22376513e-01
-1.18930072e-01 5.31703115e-01 -4.58333015e-01 -3.26466769e-01
5.37300587e-01 -3.71232629e-01 -7.61514604e-01 2.95693785e-01
6.32555962e-01 -1.41249835e-01 -1.50475025e-01 9.78388667e-01
1.54598713e-01 4.76007789e-01 8.37612569e-01 -1.09558451e+00
-8.91301572e-01 4.56318527e-01 -1.69321403e-01 1.99753165e-01
-8.38175565e-02 3.13734978e-01 9.07536447e-01 -1.07552266e+00
4.83045161e-01 1.12562740e+00 7.35624373e-01 7.35508025e-01
-1.14350855e+00 -5.57887912e-01 2.31431186e-01 2.42229491e-01
-1.58542204e+00 -3.65673482e-01 5.66541672e-01 -5.73856652e-01
8.00184309e-01 -5.72184734e-02 4.78655070e-01 1.03366590e+00
7.97468945e-02 3.56827289e-01 1.30100441e+00 -4.33161110e-01
5.30530512e-01 2.81271607e-01 2.46586800e-01 8.06703627e-01
1.00101233e-02 4.96380121e-01 -2.48220727e-01 9.43165943e-02
6.49572968e-01 5.44265985e-01 -3.09855100e-02 -5.32898128e-01
-1.08995140e+00 1.02126050e+00 8.99594128e-01 4.06961083e-01
-2.33050242e-01 1.01634398e-01 1.80178076e-01 4.65074420e-01
5.87406576e-01 5.18488586e-01 -3.82873237e-01 1.26651913e-01
-9.59460855e-01 1.17282361e-01 7.09026873e-01 8.69085431e-01
1.22258544e+00 -2.96246056e-02 -3.95361513e-01 9.69218433e-01
-1.18159197e-01 1.58888787e-01 6.97695971e-01 -5.10465980e-01
8.90999362e-02 5.96844852e-01 1.34067787e-02 -4.93742585e-01
-1.61131069e-01 -6.94142997e-01 -9.30969834e-01 1.55041456e-01
1.54710114e-01 -2.46448591e-02 -1.23452997e+00 1.65423453e+00
3.37189019e-01 5.82540631e-01 -1.39635324e-01 8.57787490e-01
7.31360912e-01 7.21918702e-01 1.13358513e-01 3.09436000e-03
1.09219265e+00 -1.24250066e+00 7.08368560e-03 -2.82734692e-01
3.62977922e-01 -2.01169014e-01 1.43425798e+00 3.75262238e-02
-5.43638647e-01 -9.89915013e-01 -1.26986539e+00 1.89398617e-01
-7.86124408e-01 -1.62907645e-01 4.97107625e-01 5.00629902e-01
-7.73034275e-01 8.16606164e-01 -4.84662890e-01 -3.98282081e-01
7.03457057e-01 9.64515060e-02 -4.05429959e-01 -3.45496029e-01
-1.07150471e+00 8.32159579e-01 6.58237934e-01 -8.26247633e-02
-1.20631862e+00 -7.65385509e-01 -9.75648463e-01 3.32678497e-01
4.33519155e-01 -9.64453936e-01 1.03450990e+00 -1.03366077e+00
-1.42093337e+00 9.39137280e-01 1.94085538e-01 -4.56468314e-01
2.11228594e-01 1.97345465e-01 -3.43254924e-01 1.76651124e-02
2.76920080e-01 6.35087371e-01 1.34652615e+00 -1.13584399e+00
-9.01119828e-01 -3.10948133e-01 2.91229784e-01 -2.10507467e-01
-4.19891775e-01 -3.31159264e-01 -2.75595929e-03 -9.26665246e-01
-4.33702648e-01 -8.08785737e-01 -3.49455953e-01 -8.83561224e-02
-1.02006219e-01 -3.89029264e-01 6.06669605e-01 6.91006109e-02
1.02830386e+00 -2.20322871e+00 1.70307025e-01 -1.39225408e-01
2.73661196e-01 2.65758216e-01 -4.72044021e-01 2.95605630e-01
2.14503035e-02 -2.27591805e-02 -3.04590642e-01 -1.14776090e-01
2.61479765e-01 2.75553793e-01 -5.01697004e-01 2.93997616e-01
4.93227452e-01 1.12358809e+00 -9.49904680e-01 -1.60670757e-01
3.16292703e-01 1.07409038e-01 -4.87573624e-01 4.67928678e-01
-2.55548865e-01 2.28114828e-01 -4.22296345e-01 6.24153733e-01
5.02482295e-01 -4.33955222e-01 -3.52451891e-01 -1.85245693e-01
-1.50697842e-01 -3.83263946e-01 -1.00787866e+00 1.81048262e+00
-8.12485814e-01 2.85752919e-02 -4.16035086e-01 -1.31249797e+00
1.12289703e+00 -9.85536054e-02 -2.12700248e-01 -5.21649241e-01
1.23948723e-01 1.39112607e-01 -1.10691771e-01 -2.51843005e-01
2.15516433e-01 -7.88204014e-01 -5.57004988e-01 5.19233286e-01
6.71866715e-01 -1.73724979e-01 1.69241175e-01 -1.48873314e-01
1.07078016e+00 -1.10698946e-01 6.25038445e-01 -2.72620142e-01
2.87663430e-01 -4.85944413e-02 5.14577746e-01 1.24312091e+00
-2.66325712e-01 5.25422990e-01 -7.24439472e-02 -8.50177884e-01
-1.09037459e+00 -1.26189542e+00 -2.71075517e-01 1.58510256e+00
1.51957154e-01 -1.60330907e-01 -7.36112893e-01 -9.74204242e-01
1.25146240e-01 5.84059894e-01 -1.02999723e+00 -5.72818041e-01
-2.23694533e-01 -4.80551749e-01 2.60128230e-01 6.20843410e-01
5.64377725e-01 -1.26282442e+00 -7.04328716e-01 2.26642221e-01
3.05591077e-01 -8.61338437e-01 -4.77907717e-01 5.33413470e-01
-5.84647715e-01 -7.53966510e-01 -6.65192306e-01 -8.94716620e-01
5.46480060e-01 3.51310164e-01 1.06408429e+00 -1.00493774e-01
-5.63720345e-01 2.33523294e-01 -6.40730143e-01 -2.80867130e-01
-1.78797975e-01 5.67184053e-02 1.56300902e-01 4.54369932e-01
4.10709292e-01 -6.62255585e-01 -5.24837554e-01 1.96994513e-01
-9.77106512e-01 3.61075141e-02 9.28500056e-01 1.45116472e+00
7.82394528e-01 8.82866327e-03 9.53474581e-01 -1.09372640e+00
4.86149758e-01 -5.92685699e-01 -4.99993294e-01 3.32296342e-01
-4.17799711e-01 2.67990828e-01 1.32089567e+00 -5.62087476e-01
-9.83041286e-01 1.12476222e-01 6.32740781e-02 -7.35723853e-01
-5.52434444e-01 4.54138458e-01 -1.80683509e-01 -3.23656052e-01
8.36503804e-01 4.11771506e-01 -4.39661920e-01 -5.81184626e-01
8.51460814e-01 6.67963088e-01 7.43285477e-01 -5.97637713e-01
9.74736571e-01 3.51625890e-01 -2.23441109e-01 -6.75238371e-01
-1.39245176e+00 -4.68983263e-01 -1.06413805e+00 -1.06380936e-02
5.74797094e-01 -8.10101390e-01 -2.59839356e-01 2.44858593e-01
-6.54923916e-01 -3.75175923e-01 -8.36719632e-01 1.47076458e-01
-8.15785706e-01 -9.23673287e-02 -2.75212914e-01 -2.92289555e-01
-2.02519819e-01 -9.29322541e-01 1.28207147e+00 4.40363556e-01
8.97649676e-02 -8.35034847e-01 1.86912164e-01 1.61718890e-01
5.10547757e-01 2.80427784e-01 9.00076926e-01 -7.47303665e-01
-5.65349519e-01 -3.64000261e-01 -3.49653602e-01 5.32327652e-01
2.33468264e-01 -5.84231377e-01 -1.10568762e+00 -5.20782232e-01
-2.05123127e-02 -7.41714835e-01 1.27269506e+00 4.63910103e-02
1.50669456e+00 -4.31881934e-01 -4.55579668e-01 9.17134464e-01
1.47313011e+00 2.83387713e-02 2.68476337e-01 1.71417221e-01
6.34394467e-01 3.70174348e-01 8.07186067e-01 5.77765048e-01
1.90631002e-01 6.31818533e-01 2.12582067e-01 2.02023253e-01
-3.33913952e-01 -3.48813951e-01 -1.01383857e-01 5.58476865e-01
6.47188202e-02 4.08334583e-02 -6.14298046e-01 6.79364622e-01
-1.78593266e+00 -1.08243167e+00 8.56755197e-01 1.98616302e+00
7.21888840e-01 9.41993520e-02 -2.82942895e-02 -1.40740663e-01
7.67494559e-01 1.47470012e-01 -8.41433287e-01 -4.64461952e-01
1.82029024e-01 7.53966630e-01 1.49285020e-02 7.70935863e-02
-1.41518140e+00 1.22311318e+00 5.44045305e+00 1.01528180e+00
-1.16509688e+00 2.10446805e-01 5.56463540e-01 -2.11571045e-02
-3.90753895e-02 -7.15037510e-02 -8.43800664e-01 4.35921013e-01
9.85575020e-01 -3.93813640e-01 5.82163632e-01 1.06017065e+00
-3.20784330e-01 3.69288594e-01 -1.35652161e+00 9.12500501e-01
2.44947478e-01 -1.62713230e+00 3.74148637e-01 -2.13221401e-01
9.42920327e-01 -2.20092628e-02 1.28824055e-01 7.22935140e-01
3.14656377e-01 -1.21557391e+00 6.96376860e-01 8.40025902e-01
1.21240580e+00 -7.38008678e-01 5.37506700e-01 6.38152719e-01
-1.28745246e+00 -4.87512648e-01 -7.88176417e-01 -2.61369139e-01
-2.38331705e-01 4.45097566e-01 -7.12781310e-01 3.66898566e-01
7.71901071e-01 8.57698679e-01 -7.28578031e-01 1.02826607e+00
-7.48430938e-02 4.23936814e-01 -1.16933286e-02 -6.16593361e-02
6.03616059e-01 1.01367481e-01 1.59882918e-01 1.18419337e+00
3.02713364e-01 3.17677468e-01 4.03991222e-01 9.80433106e-01
-3.88795227e-01 -2.32465476e-01 -7.20076621e-01 -7.10567310e-02
4.91418123e-01 1.47548068e+00 -6.27935827e-01 -5.57328463e-01
-4.74282980e-01 1.02253008e+00 8.75825226e-01 8.33117589e-02
-5.34920633e-01 -7.05171049e-01 4.60394233e-01 -6.32234961e-02
7.35930622e-01 3.42073530e-01 -9.83014554e-02 -1.37778378e+00
-1.24633834e-01 -7.81161129e-01 5.20735443e-01 -5.30914664e-01
-1.97898936e+00 8.78050029e-01 -1.25759870e-01 -1.27970707e+00
-5.26584804e-01 -5.73000193e-01 -7.04399526e-01 7.33315408e-01
-1.45243943e+00 -1.51893497e+00 -6.45391643e-01 8.29013526e-01
6.97058260e-01 -3.38601023e-01 1.04348814e+00 1.12575814e-01
-2.13342339e-01 8.00840378e-01 1.75286517e-01 -3.04777212e-02
6.48189485e-01 -1.25533020e+00 4.89533424e-01 7.07164347e-01
2.51719624e-01 4.84649509e-01 3.98121417e-01 -3.72688472e-01
-1.11881101e+00 -1.78855038e+00 5.82652211e-01 -4.43153977e-01
7.22387493e-01 -6.17757857e-01 -9.85167861e-01 7.21596241e-01
-2.99881965e-01 7.83351243e-01 6.39680743e-01 1.10379361e-01
-7.51264751e-01 -3.37086499e-01 -1.15839851e+00 2.60144770e-01
1.23085523e+00 -5.72060645e-01 -1.03434610e+00 1.83536246e-01
8.99481714e-01 5.90434708e-02 -8.58357251e-01 4.88262743e-01
5.39973855e-01 -8.68670225e-01 9.71267760e-01 -8.52712989e-01
3.75413626e-01 -2.91121066e-01 -4.01467443e-01 -1.83336723e+00
-7.73544490e-01 -1.45176992e-01 -1.68564379e-01 9.65556324e-01
3.97214517e-02 -5.98071456e-01 6.17534399e-01 7.71200582e-02
-5.70467830e-01 -8.72241437e-01 -1.02187133e+00 -1.08995187e+00
3.36278319e-01 -1.06984541e-01 8.19373250e-01 8.30651641e-01
-1.25996232e-01 4.99353647e-01 -3.98417652e-01 -6.61488436e-03
8.17545772e-01 8.84907842e-01 7.24252403e-01 -1.40828800e+00
-4.96260315e-01 -4.00121450e-01 -5.62319815e-01 -7.89479136e-01
3.49586457e-01 -1.08769047e+00 3.85015070e-01 -1.22849190e+00
5.59236884e-01 -5.90066910e-01 -5.62319994e-01 6.09387338e-01
-1.75132662e-01 5.08984089e-01 3.76739353e-01 1.36218712e-01
-8.61235976e-01 7.31600821e-01 1.24741817e+00 -5.21543920e-01
-1.48173829e-03 3.49421869e-03 -8.73152018e-01 6.99774325e-01
4.17611271e-01 -4.38058853e-01 -2.36213326e-01 -1.00180998e-01
-3.59454036e-01 -6.76153898e-02 5.06323159e-01 -1.07578719e+00
2.62412041e-01 -2.91329920e-01 5.54875433e-01 -2.62278020e-01
2.79720426e-01 -6.11835241e-01 3.13396528e-02 3.54406834e-01
-2.88805634e-01 -5.53421259e-01 -3.87059501e-03 6.51785195e-01
-3.36534470e-01 -3.30015540e-01 1.16303742e+00 -4.05982614e-01
-1.28822052e+00 7.78407395e-01 1.97641831e-02 1.77967608e-01
1.14851904e+00 -2.63130546e-01 -4.41392690e-01 1.75800905e-01
-8.22649002e-01 -1.46628171e-01 5.06576419e-01 6.41974688e-01
7.33447015e-01 -1.53503060e+00 -6.86074793e-01 4.59562778e-01
7.71317899e-01 -1.30853161e-01 4.40841496e-01 2.50952542e-01
1.29391989e-02 4.89079803e-01 -5.69719136e-01 -5.12397647e-01
-6.13202274e-01 1.01091671e+00 3.85302931e-01 -1.79410979e-01
-6.15691304e-01 9.12435591e-01 7.19978571e-01 -5.68634391e-01
-8.80454481e-02 -2.74072886e-01 -7.93888122e-02 -8.35456252e-02
7.42577970e-01 1.98097289e-01 6.08979575e-02 -5.24670780e-01
-2.22703770e-01 6.49234176e-01 -3.29177290e-01 5.05579114e-01
1.54866409e+00 -4.86439317e-02 1.65465996e-01 6.88848794e-01
1.28453290e+00 -5.02389610e-01 -1.54329240e+00 -4.49374080e-01
-9.65324640e-02 -5.39230943e-01 -1.96710497e-01 -1.02822852e+00
-6.60860837e-01 8.68308246e-01 6.83703601e-01 1.34956539e-01
1.18318319e+00 3.58481914e-01 7.83026874e-01 5.35539091e-01
7.15911984e-01 -8.12123418e-01 2.52466112e-01 7.01716483e-01
7.45588005e-01 -1.41472924e+00 -3.44585419e-01 2.63276678e-02
-3.18746001e-01 1.06148148e+00 6.43414557e-01 -5.18443882e-01
7.83248186e-01 -7.07466761e-03 -2.08368123e-01 -1.88077867e-01
-9.53911483e-01 -3.00090998e-01 5.18167675e-01 6.96862102e-01
-1.09306782e-01 4.01989073e-01 1.27177402e-01 1.15231943e+00
-2.70797938e-01 2.14837611e-01 3.03932488e-01 6.72253013e-01
-8.31301451e-01 -7.67866313e-01 -8.13461281e-03 7.86617577e-01
6.14008941e-02 -1.66961268e-01 -1.61725044e-01 5.88768482e-01
5.64560831e-01 7.29275346e-01 2.31984988e-01 -5.42633772e-01
4.67773616e-01 -4.62572183e-03 5.75611413e-01 -1.21073186e+00
-4.96992588e-01 -3.85162264e-01 -2.70856440e-01 -6.58225417e-01
-3.19723070e-01 -3.34075987e-01 -9.09774184e-01 -1.98133349e-01
-1.00034326e-01 1.57938555e-01 2.21726269e-01 1.25569916e+00
4.08804446e-01 4.92309660e-01 9.20437157e-01 -1.04114091e+00
-7.78478742e-01 -9.63348866e-01 -9.47219729e-01 6.14960849e-01
4.81282055e-01 -9.93178785e-01 -4.10118282e-01 7.61228278e-02] | [9.955596923828125, 2.464423418045044] |
e0bd7439-8ae1-418b-b5e8-048befab401f | pyramid-multi-view-stereo-net-with-self | 1912.03001 | null | https://arxiv.org/abs/1912.03001v2 | https://arxiv.org/pdf/1912.03001v2.pdf | Pyramid Multi-view Stereo Net with Self-adaptive View Aggregation | n this paper, we propose an effective and efficient pyramid multi-view stereo (MVS) net with self-adaptive view aggregation for accurate and complete dense point cloud reconstruction. Different from using mean square variance to generate cost volume in previous deep-learning based MVS methods, our \textbf{VA-MVSNet} incorporates the cost variances in different views with small extra memory consumption by introducing two novel self-adaptive view aggregations: pixel-wise view aggregation and voxel-wise view aggregation. To further boost the robustness and completeness of 3D point cloud reconstruction, we extend VA-MVSNet with pyramid multi-scale images input as \textbf{PVA-MVSNet}, where multi-metric constraints are leveraged to aggregate the reliable depth estimation at the coarser scale to fill in the mismatched regions at the finer scale. Experimental results show that our approach establishes a new state-of-the-art on the \textsl{\textbf{DTU}} dataset with significant improvements in the completeness and overall quality, and has strong generalization by achieving a comparable performance as the state-of-the-art methods on the \textsl{\textbf{Tanks and Temples}} benchmark. Our codebase is at \hyperlink{https://github.com/yhw-yhw/PVAMVSNet}{https://github.com/yhw-yhw/PVAMVSNet} | ['Yu-Wing Tai', 'Zizhuang Wei', 'Runze Zhang', 'Mingyu Ding', 'Yisong Chen', 'Hongwei Yi', 'Guoping Wang'] | 2019-12-06 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/961_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123540732.pdf | eccv-2020-8 | ['3d-point-cloud-reconstruction', 'point-cloud-reconstruction'] | ['computer-vision', 'computer-vision'] | [-1.86620101e-01 -1.24077134e-01 3.09806079e-01 -3.62242758e-01
-1.05633891e+00 -4.70154107e-01 1.81328565e-01 -6.72335401e-02
-1.18876621e-01 5.72949350e-01 1.20370679e-01 -1.09019808e-01
-2.37059042e-01 -1.04014611e+00 -1.02590406e+00 -5.75385630e-01
2.25133240e-01 5.45368016e-01 6.21989787e-01 -2.54726321e-01
2.03120619e-01 5.73957086e-01 -1.61087525e+00 5.84593892e-01
1.01114690e+00 1.36322105e+00 5.17732859e-01 4.59278494e-01
1.45347625e-01 5.19149899e-01 3.70996594e-02 -5.02757728e-01
7.67219782e-01 1.06280327e-01 -5.23219645e-01 -4.04969156e-02
9.98415887e-01 -7.53411829e-01 -4.51000810e-01 1.01942241e+00
7.35877812e-01 -4.13332274e-03 4.14451033e-01 -1.08194149e+00
-5.07182240e-01 1.63959265e-01 -1.05777884e+00 1.26856729e-01
1.84257478e-01 4.49562639e-01 8.74720931e-01 -1.30589390e+00
7.59466648e-01 1.11678278e+00 7.62980819e-01 2.78518319e-01
-1.20301282e+00 -9.12677288e-01 1.83586106e-01 3.22136134e-01
-1.39226139e+00 -2.54188865e-01 8.00286949e-01 -4.17750359e-01
9.79436874e-01 3.23618293e-01 7.48488247e-01 8.57541442e-01
2.81932086e-01 7.23982215e-01 1.19293571e+00 1.11152798e-01
2.10462496e-01 -3.79492849e-01 -1.73888087e-01 8.16369414e-01
1.49388403e-01 1.89817056e-01 -7.19279468e-01 -1.50959179e-01
1.39221752e+00 2.55832702e-01 -5.13885915e-01 -7.69024909e-01
-1.38156044e+00 7.57601082e-01 7.01483190e-01 5.86676761e-04
-4.10715312e-01 2.74410039e-01 3.14333826e-01 1.62308350e-01
7.27090657e-01 4.01566678e-04 -6.66377306e-01 9.08031240e-02
-1.04715502e+00 3.44598264e-01 2.76034683e-01 1.24673951e+00
1.09534609e+00 9.92832035e-02 6.21674815e-03 7.84455657e-01
2.03167960e-01 7.99646974e-01 -1.60861194e-01 -1.62842667e+00
8.36089015e-01 7.01432705e-01 8.15930888e-02 -8.96451652e-01
-4.01715547e-01 -5.36878169e-01 -1.16212487e+00 6.44304097e-01
1.45927906e-01 1.89591616e-01 -8.84735107e-01 1.42110801e+00
5.85365474e-01 8.77167806e-02 -2.75009841e-01 7.60701418e-01
1.09069359e+00 6.12284601e-01 -6.38234317e-01 -1.70693949e-01
1.01983595e+00 -8.82972121e-01 -9.69506577e-02 -9.41542611e-02
2.45255664e-01 -7.58081615e-01 1.18324292e+00 5.29243588e-01
-1.41404867e+00 -5.04650295e-01 -9.16079700e-01 -3.50015670e-01
8.11980963e-02 -1.01747178e-01 3.54168504e-01 6.79187924e-02
-1.32051528e+00 7.29791820e-01 -8.41280282e-01 3.02684084e-02
7.21312881e-01 1.78323567e-01 -5.23539364e-01 -4.79963064e-01
-6.31143630e-01 3.72469008e-01 -1.12931710e-02 -6.10026810e-03
-8.91003430e-01 -1.19004261e+00 -9.62216318e-01 -1.08336188e-01
4.09006774e-01 -1.27847469e+00 8.80729675e-01 -4.16116267e-01
-9.77684200e-01 8.86087716e-01 -1.85431331e-01 -1.98748857e-02
8.71473312e-01 -1.42963156e-01 5.85492402e-02 3.11155617e-01
4.33136046e-01 8.73333752e-01 6.83381259e-01 -1.69558978e+00
-5.95780551e-01 -7.12023854e-01 -1.04869083e-02 2.93188006e-01
2.10972875e-02 -4.17352289e-01 -5.89331388e-01 -6.13746166e-01
4.90090311e-01 -6.66451037e-01 -1.84108183e-01 3.73956114e-01
-3.23700994e-01 3.38312835e-02 6.65129721e-01 -6.89732313e-01
8.27377915e-01 -2.04679799e+00 3.55193198e-01 1.66243553e-01
7.00157762e-01 -8.11030045e-02 -1.83265880e-02 3.79019648e-01
1.62685826e-01 -1.80418342e-01 -3.85266811e-01 -7.54212856e-01
-1.59981757e-01 2.83963203e-01 -5.80777042e-02 5.37631631e-01
-4.27428186e-01 7.95852363e-01 -7.87144899e-01 -4.28067267e-01
7.08582759e-01 6.23759866e-01 -6.87784910e-01 -4.62915190e-02
-2.10776001e-01 7.12661743e-01 -4.11648363e-01 8.22866321e-01
1.16559196e+00 -4.29343373e-01 -3.55222493e-01 -5.40118396e-01
-1.49127990e-01 -2.29636446e-01 -1.41116631e+00 2.12398577e+00
-3.87352914e-01 2.23887146e-01 4.41514254e-01 -5.47106683e-01
6.12354338e-01 -6.48478419e-02 7.90973902e-01 -6.70348585e-01
1.00526184e-01 3.87282521e-01 -5.49036324e-01 -1.40476316e-01
4.05758560e-01 1.02471046e-01 2.82793909e-01 1.32645488e-01
-8.25376734e-02 -4.85915840e-01 -1.51957914e-01 2.50042319e-01
9.59736466e-01 2.71988958e-01 7.36575723e-02 -2.64172286e-01
4.44870800e-01 -9.83087793e-02 6.75531268e-01 4.12756681e-01
1.04155928e-01 1.08368409e+00 7.78945759e-02 -5.24327815e-01
-1.22985995e+00 -1.41942716e+00 -2.31634945e-01 6.55429661e-01
3.55384767e-01 -4.22038674e-01 -4.52608615e-01 -4.72219557e-01
7.77036399e-02 6.41485929e-01 -3.99074793e-01 3.68764728e-01
-5.70031345e-01 -2.15705618e-01 3.31975557e-02 8.02686870e-01
7.37677097e-01 -7.72343576e-01 -5.94264209e-01 -1.75931733e-02
-4.30693567e-01 -1.12713695e+00 -5.16170084e-01 -1.43891811e-01
-1.22345245e+00 -9.28002536e-01 -8.67575169e-01 -3.34323049e-01
6.76145732e-01 3.35005462e-01 1.40062916e+00 4.47081476e-02
-1.23684742e-01 4.82926935e-01 -2.67163604e-01 -2.23383710e-01
4.33241064e-03 -2.79058009e-01 -1.45594310e-03 -2.89416701e-01
-1.22519717e-01 -1.26404715e+00 -1.19595981e+00 4.71475363e-01
-9.77979541e-01 5.43512344e-01 2.43017852e-01 6.69396877e-01
1.27660596e+00 -2.10869759e-01 -9.02738646e-02 -4.94488955e-01
-1.57392129e-01 -1.41383573e-01 -7.90887415e-01 -6.14332147e-02
-5.50652862e-01 -2.82551050e-01 4.25405592e-01 3.01087964e-02
-6.48791134e-01 1.53902946e-02 -5.41151106e-01 -1.10357869e+00
-3.76473255e-02 1.02812827e-01 -1.84148744e-01 -2.73446977e-01
4.49311316e-01 5.47264278e-01 -8.17165524e-02 -5.81212401e-01
3.68275911e-01 1.93816856e-01 3.62490147e-01 -4.09469247e-01
7.67991900e-01 1.01739621e+00 2.47418702e-01 -4.98708814e-01
-7.13981628e-01 -4.12834466e-01 -8.37048054e-01 -4.31461543e-01
9.37480748e-01 -1.19148314e+00 -6.51888609e-01 5.67061722e-01
-1.05392778e+00 -4.16952282e-01 -4.50721532e-01 2.91046023e-01
-8.74826252e-01 5.81879020e-01 -5.99204361e-01 -3.65937442e-01
-6.11162782e-01 -1.42886770e+00 1.59741557e+00 -1.97914153e-01
2.43596956e-01 -8.43850791e-01 -2.24892020e-01 7.73551166e-01
2.04736352e-01 5.27154088e-01 7.91515946e-01 -6.50439113e-02
-1.16680241e+00 1.58326700e-01 -3.47038060e-01 5.51732302e-01
-1.65958658e-01 -2.74283826e-01 -9.04123306e-01 -6.45469546e-01
1.38594911e-01 -1.60606414e-01 7.97079921e-01 7.12891102e-01
1.33746243e+00 -1.77280620e-01 -1.52954295e-01 1.22136521e+00
1.88365090e+00 -1.77592397e-01 8.29043686e-01 3.42518955e-01
1.09546137e+00 2.62582451e-01 5.36131501e-01 6.22962773e-01
8.04578960e-01 7.62910664e-01 9.89957094e-01 -1.19200267e-01
-3.35393876e-01 1.64634697e-02 8.06591883e-02 7.95261562e-01
-5.05139589e-01 -1.21717706e-01 -9.06086862e-01 4.61815089e-01
-1.84659171e+00 -8.84423316e-01 -4.20370072e-01 2.29831743e+00
4.20461535e-01 -7.30447993e-02 -1.99092343e-01 -3.15737166e-02
3.48619431e-01 3.95002127e-01 -7.03346610e-01 7.37785026e-02
-1.64172873e-01 1.62610814e-01 6.77917361e-01 6.65099680e-01
-8.47825289e-01 6.72846079e-01 4.10677481e+00 1.06879079e+00
-9.68199253e-01 5.13630092e-01 6.92269325e-01 -5.37290752e-01
-5.16810596e-01 -2.44169280e-01 -8.08846533e-01 3.37863684e-01
1.50597453e-01 1.45092458e-01 3.58063698e-01 7.01857805e-01
1.74514875e-01 -2.62240320e-01 -1.17598665e+00 1.38692820e+00
1.03698753e-01 -1.78384113e+00 -2.75280438e-02 3.31452519e-01
9.72731113e-01 7.12238967e-01 8.95387903e-02 -7.96492845e-02
2.93802977e-01 -6.72578394e-01 9.25708175e-01 5.28647244e-01
1.01617861e+00 -6.17656589e-01 5.01600027e-01 4.43351805e-01
-1.58134687e+00 -1.71770267e-02 -4.82089788e-01 1.88987210e-01
4.71296519e-01 8.51285040e-01 -4.32670526e-02 1.16652262e+00
1.19406545e+00 8.03551495e-01 -4.56122786e-01 8.84732485e-01
2.04068143e-02 -2.15458032e-02 -3.87764841e-01 6.92412853e-01
7.50094056e-02 -2.29079366e-01 8.89672518e-01 6.93114996e-01
5.87564886e-01 1.83039114e-01 2.35416383e-01 8.92938316e-01
5.69881126e-02 8.20774138e-02 -5.77348888e-01 9.77358639e-01
2.71640062e-01 1.17399955e+00 -6.96468651e-01 -3.32097203e-01
-4.77157891e-01 1.06387281e+00 3.41983527e-01 2.93669105e-01
-7.89302826e-01 2.82534987e-01 6.73812866e-01 7.52099872e-01
6.92137539e-01 -2.93755740e-01 -5.67385614e-01 -1.33717644e+00
4.39999133e-01 -6.92747712e-01 4.46150482e-01 -1.15112340e+00
-1.38486540e+00 7.49189079e-01 2.21729130e-01 -1.67111671e+00
6.34684041e-02 -4.45686251e-01 -3.00206423e-01 9.21667099e-01
-1.50477469e+00 -1.36201501e+00 -6.51186287e-01 1.02361178e+00
7.92328179e-01 -5.76225333e-02 4.16596949e-01 3.70267004e-01
2.71739718e-02 4.07961160e-01 1.83227137e-01 -2.27241516e-01
5.43590963e-01 -9.45756137e-01 4.14262921e-01 7.35241950e-01
-1.69593513e-01 2.58700967e-01 3.75216246e-01 -8.33942592e-01
-1.39230478e+00 -1.25879979e+00 3.50415379e-01 -6.23538792e-01
2.81397074e-01 -3.90861273e-01 -7.75516510e-01 7.11750984e-01
1.25789046e-01 4.91972983e-01 1.91580132e-01 -4.20817316e-01
-3.53501678e-01 -3.59684706e-01 -1.44875431e+00 3.63327622e-01
1.37399781e+00 -3.68014783e-01 -2.48608872e-01 3.67393643e-01
8.90435159e-01 -8.83593500e-01 -1.17559326e+00 9.09286916e-01
4.08847004e-01 -1.65300524e+00 1.39119565e+00 1.82589471e-01
7.38536358e-01 -3.46530378e-01 -6.14644170e-01 -1.08276486e+00
-4.30990517e-01 -1.78231493e-01 -1.62372872e-01 1.00439799e+00
5.90808876e-02 -6.81724250e-01 8.27520251e-01 4.52021658e-01
-6.02911770e-01 -1.24519932e+00 -1.34640813e+00 -7.59264469e-01
1.04962602e-01 -6.34419799e-01 6.18889093e-01 8.58183384e-01
-7.87980616e-01 -1.90616250e-01 -2.60418922e-01 4.31893349e-01
1.19126713e+00 3.21266174e-01 7.89528489e-01 -1.15416455e+00
-3.86557192e-01 -3.33970457e-01 -4.32073385e-01 -1.05932462e+00
-3.92323405e-01 -9.26024973e-01 -3.11665744e-01 -2.13220167e+00
2.32076243e-01 -3.71848673e-01 2.36625578e-02 3.50539625e-01
1.16276011e-01 4.30387110e-01 4.34213489e-01 4.10134375e-01
-4.64157492e-01 6.49637401e-01 1.51432061e+00 1.23856492e-01
3.34870704e-02 -2.96450164e-02 -4.06389087e-01 9.03242409e-01
3.80837709e-01 -2.89172620e-01 -4.54743743e-01 -8.87353778e-01
2.42536217e-01 2.81372309e-01 7.01069474e-01 -1.16043568e+00
1.59681126e-01 -6.57549053e-02 6.49336457e-01 -1.24322891e+00
9.05333281e-01 -9.39161777e-01 4.61316258e-01 2.52663046e-01
3.80277693e-01 4.19332325e-01 2.48522535e-01 5.42831481e-01
3.68358009e-02 2.66512305e-01 8.04127157e-01 -6.52287424e-01
-6.18074417e-01 9.03660119e-01 2.54942030e-01 2.88622566e-02
1.00225675e+00 -5.45423806e-01 -3.62528384e-01 -3.54622841e-01
-6.57422125e-01 4.00559992e-01 1.03526115e+00 1.97881699e-01
9.86195028e-01 -1.29214168e+00 -7.30168104e-01 2.06642196e-01
1.68494657e-01 7.68054724e-01 9.01246130e-01 1.03668118e+00
-7.44421184e-01 4.76124473e-02 -2.52651900e-01 -1.08116174e+00
-1.27435863e+00 4.23539251e-01 3.09111029e-01 -2.96359807e-01
-1.08133984e+00 1.16962934e+00 6.24393702e-01 -5.71469784e-01
2.44986251e-01 -4.95591730e-01 2.61629641e-01 -2.36511230e-01
3.11096340e-01 5.62892497e-01 2.69485712e-01 -6.79898441e-01
-2.84856647e-01 1.01006711e+00 1.67625427e-01 2.40574535e-02
1.73840487e+00 -3.02656621e-01 -2.49109969e-01 2.32947603e-01
1.10899699e+00 5.92134893e-02 -1.58909512e+00 -1.55356273e-01
-7.46305704e-01 -9.21305120e-01 2.23619312e-01 -6.27331972e-01
-1.56172276e+00 7.74996579e-01 8.21009994e-01 -3.10752898e-01
1.27935863e+00 2.17663765e-01 1.00026095e+00 -1.16939861e-02
9.85596478e-01 -7.21697450e-01 8.61705318e-02 4.14330542e-01
1.25670505e+00 -1.26050782e+00 3.78157377e-01 -6.22533381e-01
-6.20198190e-01 8.67913783e-01 7.38846719e-01 -1.53254598e-01
7.89364874e-01 3.27585071e-01 -9.25278887e-02 -5.81983566e-01
-4.34402466e-01 1.45133898e-01 2.29930326e-01 4.29331869e-01
-1.00415153e-02 -1.70973703e-01 -8.09103027e-02 2.48601988e-01
-2.32439846e-01 -1.73373725e-02 2.16963962e-01 9.72958088e-01
-2.25887597e-01 -1.02804279e+00 -3.76202673e-01 6.62967920e-01
-5.84371909e-02 -3.69702399e-01 1.04243979e-01 5.24036348e-01
2.75042951e-01 6.01038456e-01 5.68894781e-02 -4.94427741e-01
4.80050445e-01 -3.35508913e-01 7.30637491e-01 -5.05613506e-01
-5.49085617e-01 2.14391708e-01 -1.57656237e-01 -1.19732535e+00
-3.43250364e-01 -7.82875240e-01 -1.17772877e+00 -5.39011776e-01
3.23608406e-02 -4.46468204e-01 3.36318225e-01 6.22903526e-01
6.29883051e-01 3.75522852e-01 6.35337889e-01 -1.39018631e+00
-3.41131449e-01 -6.76644862e-01 -6.03497028e-01 2.78628439e-01
2.01816097e-01 -7.31660366e-01 -3.15046519e-01 -2.88543403e-01] | [8.835527420043945, -2.8564202785491943] |
1e6a6c29-749c-4d7f-b0d0-e43ad04589f5 | unified-question-answering-in-slovene | 2211.09159 | null | https://arxiv.org/abs/2211.09159v1 | https://arxiv.org/pdf/2211.09159v1.pdf | Unified Question Answering in Slovene | Question answering is one of the most challenging tasks in language understanding. Most approaches are developed for English, while less-resourced languages are much less researched. We adapt a successful English question-answering approach, called UnifiedQA, to the less-resourced Slovene language. Our adaptation uses the encoder-decoder transformer SloT5 and mT5 models to handle four question-answering formats: yes/no, multiple-choice, abstractive, and extractive. We use existing Slovene adaptations of four datasets, and machine translate the MCTest dataset. We show that a general model can answer questions in different formats at least as well as specialized models. The results are further improved using cross-lingual transfer from English. While we produce state-of-the-art results for Slovene, the performance still lags behind English. | ['Marko Robnik-Šikonja', 'Katja Logar'] | 2022-11-16 | null | null | null | null | ['cross-lingual-transfer'] | ['natural-language-processing'] | [-2.05305498e-02 1.22456208e-01 -1.72567084e-01 -4.01380658e-01
-1.59984028e+00 -9.65815663e-01 3.47080112e-01 1.91451870e-02
-7.90582120e-01 8.69100273e-01 5.12523890e-01 -1.01088059e+00
2.50196811e-02 -8.21859896e-01 -6.68793917e-01 3.00615519e-01
5.64407289e-01 8.74229014e-01 4.39308196e-01 -9.13991630e-01
-1.31301656e-01 -2.93037146e-01 -9.46678281e-01 8.97955537e-01
1.03567946e+00 6.83600307e-01 2.37032175e-01 1.07557940e+00
-8.07547390e-01 1.21916294e+00 -4.75998282e-01 -1.00075734e+00
8.67027268e-02 -3.85380983e-01 -1.72196269e+00 -2.92375326e-01
5.60029745e-01 -3.65363806e-01 -1.94669262e-01 4.91940498e-01
3.50586057e-01 -2.45125577e-01 1.73871115e-01 -7.73792624e-01
-1.03415537e+00 6.01612985e-01 9.15581062e-02 2.69771487e-01
1.01317668e+00 5.66500798e-02 1.52642322e+00 -7.55303741e-01
6.03404284e-01 1.23324573e+00 3.88074905e-01 9.10503566e-01
-9.08238292e-01 -3.74051839e-01 1.05367884e-01 4.68786150e-01
-9.68358099e-01 -4.36636537e-01 3.30126911e-01 6.85495809e-02
1.46857178e+00 5.49699664e-01 1.24849021e-01 9.69101489e-01
3.18574533e-02 1.26201916e+00 1.07349145e+00 -6.45687103e-01
-1.88595906e-01 1.99005842e-01 1.16310813e-01 6.99971318e-01
-1.36508301e-01 -5.69912970e-01 -4.59221423e-01 -3.61071192e-02
2.81378061e-01 -4.22206402e-01 -2.56333441e-01 2.73521334e-01
-1.11365819e+00 1.07072961e+00 -1.24928923e-02 2.79318482e-01
-6.11024313e-02 2.05371287e-02 6.63746238e-01 1.08612216e+00
3.72680634e-01 8.05116534e-01 -1.32324183e+00 -3.33963096e-01
-7.94131100e-01 6.50289893e-01 1.45016277e+00 1.14395225e+00
6.58688486e-01 -2.63403058e-01 -2.48532638e-01 9.88879442e-01
8.70497376e-02 5.47609627e-01 5.64549923e-01 -1.06087053e+00
1.12264132e+00 7.25820303e-01 1.50537387e-01 -2.61882126e-01
-4.05600160e-01 -2.56151348e-01 -3.36256742e-01 -6.09801531e-01
7.33246207e-01 -5.54645002e-01 -6.82132661e-01 1.60074997e+00
2.80016184e-01 -5.27175963e-01 5.79731822e-01 6.89534366e-01
1.31091380e+00 1.02553880e+00 1.36019856e-01 3.54435176e-01
2.09268999e+00 -1.41013932e+00 -7.18309820e-01 -7.92391837e-01
9.66018319e-01 -8.63396108e-01 1.49452245e+00 1.00746185e-01
-1.31714833e+00 -4.01075095e-01 -6.09581351e-01 -9.82637107e-01
-5.77931702e-01 7.96199590e-02 5.53543329e-01 7.99079418e-01
-1.12061918e+00 -2.11120516e-01 -4.07652467e-01 -4.81288344e-01
-1.39776185e-01 5.43297045e-02 -3.57642412e-01 -5.47262073e-01
-1.74149287e+00 1.19231987e+00 2.73812264e-01 -2.44433671e-01
-5.09252787e-01 -7.32163966e-01 -1.23953021e+00 -9.10519958e-02
6.65373683e-01 -9.68093455e-01 2.04147911e+00 -7.23275900e-01
-1.69375026e+00 1.12305880e+00 -5.40582061e-01 -6.57713950e-01
1.84078574e-01 -4.86714512e-01 -5.46902239e-01 2.17455775e-01
3.62613469e-01 6.30835295e-01 2.19886839e-01 -6.32332504e-01
-8.92269909e-01 -2.55515218e-01 7.43156493e-01 2.42012158e-01
-3.26287970e-02 4.48159993e-01 -5.86934328e-01 -4.76014525e-01
-8.84174332e-02 -6.67147875e-01 -8.13401677e-03 -2.58729339e-01
-7.77133778e-02 -5.49716830e-01 6.46011591e-01 -1.35633326e+00
1.39268017e+00 -1.62725151e+00 1.59372166e-02 -4.85976994e-01
-2.73518205e-01 2.53478795e-01 -4.64451224e-01 8.58475864e-01
1.98676869e-01 1.90821335e-01 -4.30509955e-01 -1.90983802e-01
2.80245453e-01 5.64059138e-01 -5.53635538e-01 -2.85862386e-01
6.41434968e-01 1.44592857e+00 -8.71817946e-01 -4.32787895e-01
-7.91027769e-02 -1.28167957e-01 -8.91174316e-01 4.43712473e-01
-7.13532269e-01 1.51538804e-01 -6.31236017e-01 6.97758436e-01
2.62943298e-01 -1.81832328e-01 8.61056224e-02 2.19031483e-01
8.43541250e-02 1.13509786e+00 -7.89949834e-01 2.03789544e+00
-1.26262879e+00 5.06321311e-01 2.40534514e-01 -6.94188535e-01
5.75399935e-01 6.24688387e-01 -4.19998765e-02 -1.08546102e+00
-4.03645217e-01 6.59091532e-01 3.36632947e-03 -9.08025384e-01
9.03631449e-01 -2.31688201e-01 -4.63716656e-01 4.00919825e-01
3.74793082e-01 -7.17758834e-01 3.17763090e-01 3.68985772e-01
1.21537971e+00 3.19937527e-01 4.37532932e-01 -2.84439474e-01
7.34394610e-01 2.97802329e-01 1.51789337e-02 7.33034134e-01
3.31011489e-02 4.90695626e-01 5.52293062e-01 -2.59215206e-01
-8.72456968e-01 -8.42325926e-01 -2.01711301e-02 1.55273509e+00
-1.84902042e-01 -7.84715354e-01 -8.36073875e-01 -8.94032240e-01
-2.03274593e-01 1.25845253e+00 -2.82533884e-01 7.85174593e-02
-1.03123021e+00 -1.56930178e-01 8.04239988e-01 3.02934408e-01
6.28203213e-01 -1.11617804e+00 -5.65625787e-01 4.89785790e-01
-7.41042316e-01 -1.61007535e+00 -4.84005302e-01 1.58556193e-01
-5.89987576e-01 -8.29821944e-01 -8.54808986e-01 -9.81917620e-01
-1.78784564e-01 -1.58692926e-01 1.81727767e+00 -5.79163851e-03
3.24718684e-01 7.18121648e-01 -6.56565547e-01 -5.57169437e-01
-4.27685022e-01 7.19615936e-01 -7.23608494e-01 -3.45176488e-01
5.48174739e-01 4.64145243e-02 -4.49052542e-01 2.12960899e-01
-9.96615589e-01 -1.07538275e-01 4.03921515e-01 7.14396417e-01
2.69413054e-01 -7.58458972e-01 9.94878054e-01 -1.35453629e+00
7.62839675e-01 -6.13302112e-01 -2.74800539e-01 4.99892473e-01
-1.34333044e-01 3.39074343e-01 1.13209653e+00 3.16820890e-02
-1.24275470e+00 -3.76390427e-01 -1.00439775e+00 4.43092376e-01
-2.13762745e-01 5.91760576e-01 -3.54728043e-01 2.84152597e-01
5.36812067e-01 1.46776587e-01 -6.34922862e-01 -5.02228796e-01
8.30604792e-01 7.18841314e-01 5.26422977e-01 -8.94221365e-01
5.24762452e-01 1.07223399e-01 -6.74869955e-01 -8.23041081e-01
-1.07189989e+00 -7.39477932e-01 -1.12272516e-01 2.23226517e-01
1.00284612e+00 -9.71871376e-01 -4.33045596e-01 1.19954765e-01
-1.51192987e+00 -4.92188692e-01 -5.39265752e-01 5.50479293e-02
-4.88572806e-01 3.08268487e-01 -1.00801992e+00 -4.85482514e-01
-5.87845981e-01 -1.08721614e+00 1.37644994e+00 2.78803110e-02
-3.40315580e-01 -1.18207800e+00 1.48272872e-01 8.67435217e-01
5.87157488e-01 -4.92386758e-01 1.33872366e+00 -8.75264585e-01
-4.66916412e-01 -1.55109271e-01 -2.39455953e-01 3.53985846e-01
-1.34043664e-01 -6.38950706e-01 -7.06057429e-01 2.48781871e-02
9.76427868e-02 -8.78566444e-01 8.80110085e-01 -9.29635391e-02
9.73737776e-01 -5.01073182e-01 2.39075080e-01 3.19901675e-01
1.26523221e+00 -8.80612954e-02 8.29031587e-01 5.66490829e-01
3.55839163e-01 6.83553755e-01 2.62368709e-01 -1.79310545e-01
1.42675447e+00 4.33969051e-01 1.30085185e-01 -3.61605771e-02
-1.98341101e-01 -5.32640517e-01 5.22388399e-01 1.26941359e+00
4.29417044e-01 -3.25733870e-01 -1.02527046e+00 9.62093413e-01
-1.58064961e+00 -5.15441298e-01 -2.06235737e-01 1.65188086e+00
1.25461006e+00 -1.27411962e-01 -5.94501942e-02 -3.00817728e-01
5.53719923e-02 2.21425802e-01 -3.06900591e-01 -9.45913136e-01
-2.90469915e-01 1.02372622e+00 2.76566297e-01 7.65193403e-01
-9.35609341e-01 1.37849689e+00 6.65156651e+00 9.58504200e-01
-6.35750175e-01 3.50262016e-01 4.88975734e-01 3.45940828e-01
-7.76467860e-01 3.42444748e-01 -7.17701614e-01 -3.35048372e-03
1.34858787e+00 -7.76170269e-02 2.28531241e-01 7.08467007e-01
-3.00771177e-01 -1.22386523e-01 -1.07256675e+00 8.69615853e-01
7.61547163e-02 -1.19453180e+00 1.94729492e-01 -6.12422705e-01
5.38972616e-01 1.30648896e-01 -1.36465535e-01 9.47509408e-01
5.28390646e-01 -1.04825342e+00 6.58485234e-01 2.99284250e-01
8.31035793e-01 -5.41100144e-01 8.82246315e-01 4.70360637e-01
-1.14543986e+00 -1.43296598e-02 -3.63322288e-01 -1.60734639e-01
4.73158449e-01 2.05794036e-01 -4.89094079e-01 8.98438096e-01
6.20480776e-01 2.43464425e-01 -7.48952866e-01 5.20896316e-01
-6.02608204e-01 9.76614177e-01 -3.26296359e-01 -3.63277256e-01
5.63519955e-01 -1.20114401e-01 3.49693775e-01 1.30294251e+00
1.65572628e-01 2.21616000e-01 -3.55326533e-02 4.86349404e-01
-4.06868458e-01 5.05613029e-01 -2.99872786e-01 1.37952611e-01
1.73340421e-02 1.08223498e+00 9.16036253e-04 -4.54946071e-01
-9.97192621e-01 1.26366425e+00 5.47301114e-01 3.14755172e-01
-4.54284310e-01 -6.97334766e-01 3.03109139e-01 -1.76721901e-01
1.51650876e-01 -3.94167453e-01 -4.97892387e-02 -1.57758605e+00
2.55116761e-01 -1.67394805e+00 9.14931297e-01 -8.51970732e-01
-1.32138610e+00 7.07258523e-01 -1.30836755e-01 -5.09274960e-01
-6.44710839e-01 -8.53446007e-01 -2.48217270e-01 9.95392978e-01
-2.00134778e+00 -1.22327089e+00 3.10879499e-01 6.50253773e-01
1.16220462e+00 5.50129078e-02 1.15223587e+00 6.30058885e-01
-1.94515809e-01 5.92332661e-01 -8.81302059e-02 3.03813070e-01
6.81094408e-01 -1.49105883e+00 9.32076812e-01 6.84754014e-01
4.46080446e-01 4.29131806e-01 4.28785264e-01 -1.85754105e-01
-1.63129449e+00 -8.91468406e-01 1.64752269e+00 -8.06165218e-01
1.06303692e+00 -5.93965769e-01 -1.07015264e+00 9.33146894e-01
7.61712790e-01 -3.54270279e-01 6.02018058e-01 2.22984225e-01
-5.19405127e-01 5.85888140e-02 -8.51437867e-01 7.23971188e-01
7.89304137e-01 -1.13094866e+00 -1.14922380e+00 3.38119119e-01
1.27507186e+00 -6.45277262e-01 -8.03818166e-01 1.51368842e-01
3.47732514e-01 -6.43595457e-01 6.71084464e-01 -1.04742646e+00
5.34758210e-01 -4.00073193e-02 -4.20010120e-01 -1.23675692e+00
2.04128951e-01 -6.82483852e-01 -3.26051493e-03 1.03472245e+00
1.15962577e+00 -8.35671246e-01 5.62687814e-01 6.10266745e-01
-2.12139681e-01 -8.43463838e-01 -1.16381884e+00 -4.83247906e-01
8.29929113e-01 -7.31183350e-01 6.95192933e-01 6.59716785e-01
1.00588299e-01 9.55592990e-01 -1.16204292e-01 -2.40737140e-01
-2.33452395e-02 3.01385790e-01 7.53555715e-01 -5.54569423e-01
-4.88831550e-01 -1.06186695e-01 1.91134900e-01 -1.70153093e+00
4.35972869e-01 -1.16059291e+00 -1.31236061e-01 -1.97820938e+00
-8.71514380e-02 -1.33407176e-01 3.32457185e-01 4.98337716e-01
-3.77263606e-01 -4.36323471e-02 1.86113283e-01 -3.15461367e-01
-8.11288118e-01 6.15350842e-01 1.33645546e+00 -5.07642999e-02
2.43755326e-01 -5.66664897e-03 -8.76438439e-01 5.13731480e-01
7.47049630e-01 -1.75569743e-01 -2.63083518e-01 -1.27149081e+00
6.89078748e-01 3.21955979e-01 5.44701852e-02 -6.74402177e-01
2.03428656e-01 2.63296850e-02 -2.00084835e-01 -3.78585190e-01
6.85942024e-02 -6.86702132e-01 -6.36999071e-01 2.11272821e-01
-6.02192402e-01 5.32278836e-01 4.48744893e-01 -7.23817945e-02
-6.30315661e-01 -5.64271092e-01 3.52084905e-01 -4.98302519e-01
-8.05387259e-01 2.70893514e-01 -4.65456158e-01 1.09121895e+00
2.52122313e-01 1.43368050e-01 -5.21360695e-01 -9.60785449e-01
-2.91112125e-01 7.17026234e-01 -2.02250704e-01 5.87124825e-01
4.74224597e-01 -9.38990712e-01 -1.15121281e+00 -1.52524263e-01
3.95164192e-01 -2.77771726e-02 2.62268603e-01 3.73663038e-01
-6.95883751e-01 9.53433812e-01 2.26316869e-01 -1.99162036e-01
-5.76680601e-01 1.87029138e-01 5.23772240e-01 -8.33928943e-01
-2.73554116e-01 7.94119596e-01 -5.51356599e-02 -1.25507414e+00
-3.38804603e-01 -6.26932800e-01 -2.64327556e-01 -1.47515476e-01
4.90975261e-01 1.51135981e-01 4.32860523e-01 -4.09495562e-01
-3.16559047e-01 5.90876698e-01 -4.08383198e-02 -3.87277871e-01
8.39102209e-01 -3.03051293e-01 -2.30760455e-01 2.72560984e-01
1.21340358e+00 1.95382908e-01 -4.54945266e-01 -4.38056290e-01
3.57480466e-01 1.17568433e-01 -4.03919965e-01 -9.51547801e-01
-5.94323277e-01 1.05055892e+00 -2.27001235e-02 2.48581201e-01
1.15423369e+00 2.65684485e-01 1.48899686e+00 8.54769528e-01
3.83780956e-01 -1.04869103e+00 -1.84807196e-01 1.35655904e+00
9.51234698e-01 -1.27919590e+00 -4.71851379e-01 -1.99609324e-01
-6.06338263e-01 9.17076528e-01 6.39296651e-01 1.14860065e-01
3.28417271e-01 5.57784364e-02 4.17729616e-01 -1.47580177e-01
-9.72070515e-01 -5.73488176e-01 3.97545159e-01 4.05495435e-01
7.70566702e-01 -2.52646953e-02 -2.83722311e-01 9.67880487e-01
-7.92673588e-01 -7.42015094e-02 4.16338652e-01 8.71465027e-01
-9.78402719e-02 -1.29034615e+00 -8.20681006e-02 5.22356868e-01
-1.01269484e+00 -6.59686685e-01 -3.05902898e-01 1.10673189e+00
-1.61203235e-01 1.41836882e+00 -1.14693925e-01 3.95863056e-02
8.20385337e-01 4.88204062e-01 4.80559409e-01 -8.68695736e-01
-9.78845358e-01 -5.40850699e-01 6.38936162e-01 -5.02627432e-01
-2.07304463e-01 -2.51350582e-01 -1.16159248e+00 -6.45638183e-02
-2.66436785e-01 5.51498175e-01 5.37161827e-01 1.16334045e+00
2.21920788e-01 2.67362952e-01 9.80049297e-02 1.28081024e-01
-5.47468841e-01 -9.37782526e-01 -1.08802281e-01 2.18046889e-01
5.12858927e-01 1.84676081e-01 -1.58544496e-01 -7.09406510e-02] | [11.373373031616211, 8.316915512084961] |
58560e62-7cab-4dd4-a1d9-4bfbb22d8120 | learning-practically-feasible-policies-for | 2108.13680 | null | https://arxiv.org/abs/2108.13680v3 | https://arxiv.org/pdf/2108.13680v3.pdf | Learning Practically Feasible Policies for Online 3D Bin Packing | We tackle the Online 3D Bin Packing Problem, a challenging yet practically useful variant of the classical Bin Packing Problem. In this problem, the items are delivered to the agent without informing the full sequence information. Agent must directly pack these items into the target bin stably without changing their arrival order, and no further adjustment is permitted. Online 3D-BPP can be naturally formulated as Markov Decision Process (MDP). We adopt deep reinforcement learning, in particular, the on-policy actor-critic framework, to solve this MDP with constrained action space. To learn a practically feasible packing policy, we propose three critical designs. First, we propose an online analysis of packing stability based on a novel stacking tree. It attains a high analysis accuracy while reducing the computational complexity from $O(N^2)$ to $O(N \log N)$, making it especially suited for RL training. Second, we propose a decoupled packing policy learning for different dimensions of placement which enables high-resolution spatial discretization and hence high packing precision. Third, we introduce a reward function that dictates the robot to place items in a far-to-near order and therefore simplifies the collision avoidance in movement planning of the robotic arm. Furthermore, we provide a comprehensive discussion on several key implemental issues. The extensive evaluation demonstrates that our learned policy outperforms the state-of-the-art methods significantly and is practically usable for real-world applications. | ['Kai Xu', 'Hui Huang', 'Xin Xu', 'Chenyang Zhu', 'Hang Zhao'] | 2021-08-31 | null | null | null | null | ['3d-bin-packing'] | ['miscellaneous'] | [-3.17912757e-01 1.81083366e-01 -5.68616211e-01 7.78122768e-02
-4.64693129e-01 -5.53311706e-01 -1.05881214e-01 3.30229908e-01
-5.20934224e-01 9.85534251e-01 -4.37658161e-01 -6.36363566e-01
-4.72354978e-01 -8.62466156e-01 -1.19137537e+00 -9.82200205e-01
-7.58105278e-01 1.20438194e+00 2.78376825e-02 -3.49587858e-01
7.94916749e-02 8.15132439e-01 -1.28936839e+00 -3.36873978e-01
8.90626371e-01 1.46745956e+00 7.82071769e-01 5.44166148e-01
1.54746370e-02 5.37516296e-01 -6.31154835e-01 -1.13476552e-01
6.03440106e-01 -4.03873809e-02 -6.58763707e-01 5.44975638e-01
-4.44520801e-01 -5.48035800e-01 -4.12477970e-01 6.66129768e-01
3.14571053e-01 4.19433087e-01 5.11074245e-01 -1.09389806e+00
-5.88146031e-01 7.32352495e-01 -6.78306937e-01 1.43495917e-01
6.72659874e-02 3.18220079e-01 8.13484490e-01 -9.98719409e-02
3.35638046e-01 1.06028986e+00 7.10868984e-02 2.57024616e-01
-1.18793809e+00 -2.47961760e-01 6.88215673e-01 3.01247984e-01
-8.92778635e-01 3.14908549e-02 4.38404292e-01 -2.14939237e-01
1.04311788e+00 7.99861550e-02 1.09557605e+00 9.11284029e-01
5.26734352e-01 1.03115261e+00 7.74362922e-01 -2.38605902e-01
8.00065875e-01 -5.14580548e-01 -2.15653196e-01 4.75831956e-01
6.50305748e-01 2.66711354e-01 -6.78815916e-02 3.09970617e-01
8.73800933e-01 3.05665612e-01 -2.85375328e-03 -1.06252611e+00
-7.94283152e-01 9.33093607e-01 7.01836050e-01 -1.12258270e-01
-6.26662552e-01 2.92441785e-01 2.55691260e-01 3.05342108e-01
1.23057775e-02 4.98866469e-01 -3.92024428e-01 -2.09964082e-01
-4.88395184e-01 6.79141521e-01 9.52004313e-01 1.53973353e+00
3.78960252e-01 3.06514446e-02 -1.78935766e-01 5.10142982e-01
2.22789809e-01 7.67395735e-01 1.06521146e-02 -9.43818688e-01
9.13860619e-01 1.65362328e-01 8.48491967e-01 -7.58686960e-01
-7.41241038e-01 -3.90293241e-01 -9.70090926e-01 1.93078175e-01
2.39640325e-01 -2.22074434e-01 -9.28234518e-01 1.23429978e+00
3.48400056e-01 -4.45679635e-01 3.90209556e-02 1.16042852e+00
-8.65835175e-02 7.18106925e-01 -2.74525762e-01 -6.01040781e-01
1.19872344e+00 -1.14333582e+00 -8.10318172e-01 -3.94849390e-01
2.45037138e-01 -2.12592661e-01 5.78665853e-01 5.96365631e-01
-1.39710569e+00 -3.20226580e-01 -1.29071331e+00 4.12139952e-01
-3.44765604e-01 -6.09988980e-02 8.82150710e-01 3.64462405e-01
-6.03138685e-01 8.28865528e-01 -1.25393212e+00 1.00054897e-01
3.89870197e-01 7.48702645e-01 2.14346275e-02 -3.28650773e-01
-8.47638607e-01 9.93555605e-01 5.66130996e-01 4.41000044e-01
-1.02473521e+00 -9.49123874e-02 -9.36400414e-01 1.51503116e-01
9.58303630e-01 -5.12446702e-01 1.63127792e+00 -8.32519904e-02
-2.05275989e+00 2.90198565e-01 1.42638341e-01 -8.55773330e-01
5.24559200e-01 -4.02036399e-01 4.35296670e-02 9.30574238e-02
1.87864434e-02 4.68159646e-01 7.96017945e-01 -1.17918730e+00
-8.66713107e-01 -3.43903482e-01 3.18533897e-01 3.48643661e-01
1.20956302e-01 -5.54441929e-01 -4.75537449e-01 -2.81322181e-01
1.26445711e-01 -1.01168394e+00 -7.40540504e-01 -3.71900260e-01
-2.78722703e-01 -1.03680819e-01 2.27715403e-01 -2.37740368e-01
9.40708935e-01 -1.82314515e+00 5.59782684e-01 1.02785818e-01
-1.37655392e-01 6.66196346e-02 8.76324698e-02 6.57000542e-01
4.76055533e-01 -2.87475646e-01 5.90394177e-02 -3.34598958e-01
4.74177331e-01 7.22949684e-01 -4.29099530e-01 6.44462168e-01
3.95242423e-02 8.90229821e-01 -1.00693333e+00 -6.29287809e-02
4.22392130e-01 -2.53853917e-01 -7.47818589e-01 3.63690525e-01
-5.15932977e-01 1.87974334e-01 -6.41807020e-01 6.41286492e-01
9.16538298e-01 -1.96408972e-01 4.41411674e-01 3.64381641e-01
-2.75653303e-01 1.62422404e-01 -1.28015375e+00 1.57400966e+00
-4.81190085e-01 -1.25943229e-01 4.40002948e-01 -1.43692386e+00
9.88863766e-01 -2.77369827e-01 6.05287015e-01 -9.50373650e-01
3.51826578e-01 1.97564170e-01 5.82383806e-03 -3.20391804e-01
6.89078689e-01 7.89505709e-03 -4.43210959e-01 1.26288503e-01
-1.73141673e-01 -2.85423994e-01 5.09787202e-01 -3.82284135e-01
1.10834301e+00 6.78641796e-02 2.97245234e-01 -1.37793928e-01
1.11785509e-01 1.14500679e-01 6.31099701e-01 9.35090482e-01
-1.54320195e-01 -1.88364461e-01 6.19237602e-01 -4.32421744e-01
-9.78326261e-01 -1.04254389e+00 1.19083822e-01 8.68622839e-01
1.00114155e+00 2.11272985e-01 -5.55482328e-01 -4.78682220e-01
6.47971630e-01 6.53577745e-01 -4.76071626e-01 -6.02507591e-02
-7.82788754e-01 -5.57875574e-01 -4.65934157e-01 7.56684601e-01
2.50592023e-01 -1.16755271e+00 -1.10944390e+00 6.91257954e-01
2.03535274e-01 -1.10240090e+00 -3.46090049e-01 1.02049911e+00
-8.25417817e-01 -9.31180239e-01 -6.66669071e-01 -9.84162927e-01
6.17058158e-01 5.08728743e-01 7.41309047e-01 -4.76343513e-01
-9.92354080e-02 1.41191572e-01 -6.20992482e-01 -3.97105724e-01
-8.68941098e-02 2.50887901e-01 2.37929910e-01 -4.72032815e-01
6.71451390e-02 -3.43425721e-01 -6.80999041e-01 5.68647802e-01
-5.33459425e-01 2.43064929e-02 9.73079026e-01 8.42350245e-01
8.50954771e-01 3.20834070e-01 3.99891704e-01 -3.17997932e-01
4.61094052e-01 -2.95961529e-01 -1.24769938e+00 -3.47250216e-02
-2.65231699e-01 1.88268930e-01 1.00288701e+00 -3.62674654e-01
-5.49550772e-01 -5.43286055e-02 8.06671381e-02 -5.96670926e-01
3.35700177e-02 3.40594620e-01 -1.53527141e-01 1.87453672e-01
-9.29498896e-02 6.27792925e-02 1.54256538e-01 -3.67105275e-01
2.08541691e-01 3.31748039e-01 3.66245359e-01 -7.24220037e-01
6.35362387e-01 2.31010467e-01 2.28918552e-01 -3.38133037e-01
-7.13199258e-01 -3.78879249e-01 -5.38639605e-01 -3.12337577e-02
5.78154504e-01 -4.96355534e-01 -1.60833216e+00 1.25943989e-01
-9.90750909e-01 -9.81308877e-01 -4.04060245e-01 3.10153961e-01
-1.30115616e+00 1.57497332e-01 -7.86142290e-01 -1.11793578e+00
-1.43998072e-01 -1.19002581e+00 9.50812459e-01 3.16301852e-01
3.96289855e-01 -5.37431180e-01 -2.44039640e-01 8.41924623e-02
2.39227504e-01 2.83724040e-01 8.64832103e-01 -4.36616272e-01
-8.06217790e-01 -4.37678434e-02 4.06615660e-02 -1.39711738e-01
7.14687556e-02 -4.69363689e-01 4.39606085e-02 -7.03853786e-01
-5.72251864e-02 -2.14378074e-01 4.59422857e-01 7.15459645e-01
1.33664763e+00 -5.20021558e-01 -6.27262771e-01 3.86803895e-01
1.31932926e+00 9.07067597e-01 2.57799834e-01 8.02922249e-01
1.50383607e-01 2.73034334e-01 1.39377034e+00 7.19316840e-01
3.43087614e-01 8.20249081e-01 1.06327116e+00 2.42279783e-01
5.85738122e-01 -3.17292869e-01 2.62730956e-01 6.06863022e-01
6.10322915e-02 -8.01491082e-01 -5.51481009e-01 3.29226017e-01
-2.31328082e+00 -7.57468402e-01 4.57580090e-01 2.25833249e+00
4.79489893e-01 5.30114830e-01 1.96099043e-01 2.81035379e-02
6.87831342e-01 -1.14163915e-02 -9.21388984e-01 -7.45765567e-01
3.13549250e-01 1.12812109e-01 1.17339456e+00 4.78782058e-01
-1.30621743e+00 9.28100705e-01 5.86604881e+00 7.75809228e-01
-6.93188131e-01 -3.57876718e-01 4.31933761e-01 -3.19032788e-01
3.45435292e-01 -4.18653816e-01 -1.09508526e+00 7.22911298e-01
7.54344344e-01 1.55745178e-01 9.24801469e-01 1.10875261e+00
2.97478110e-01 -3.78116280e-01 -1.21795881e+00 9.60467339e-01
-4.06381458e-01 -1.28792167e+00 -2.67315388e-01 4.93088126e-01
5.00798047e-01 -3.11326176e-01 4.25621308e-02 5.30141652e-01
7.00263560e-01 -8.06018114e-01 9.66764152e-01 3.28869551e-01
4.08524692e-01 -1.23100102e+00 8.97854030e-01 5.80802500e-01
-1.05477726e+00 -7.28813529e-01 -7.68507719e-01 -2.21181124e-01
7.23098695e-01 3.81433845e-01 -7.62921095e-01 7.48993099e-01
6.40779555e-01 4.30938512e-01 3.67520183e-01 1.20108187e+00
-1.67308807e-01 -1.41741320e-01 -4.58091766e-01 -6.22233033e-01
6.41518950e-01 -5.23107469e-01 4.32702899e-01 7.68543541e-01
2.71347344e-01 2.80134290e-01 7.45466113e-01 5.55389941e-01
1.88017905e-01 -2.73083359e-01 -3.26541007e-01 2.63089035e-02
5.01864016e-01 8.59186292e-01 -8.70984554e-01 5.62834218e-02
9.25966576e-02 8.63448381e-01 6.28538430e-01 1.18116215e-01
-1.06484556e+00 -3.76590520e-01 6.99229598e-01 -1.40761867e-01
1.17665219e+00 -6.05534136e-01 5.54415733e-02 -6.24635696e-01
1.24451697e-01 -6.00140631e-01 1.19854197e-01 -2.95797497e-01
-1.19215810e+00 2.30160743e-01 -1.64520219e-02 -1.15088475e+00
-3.45980465e-01 -1.04710829e+00 -1.72533855e-01 3.66842389e-01
-1.75246894e+00 -4.67818737e-01 -5.24935648e-02 1.75074220e-01
6.44331396e-01 2.30073999e-03 5.47779262e-01 8.52954946e-03
-7.16138005e-01 4.76260245e-01 4.90500450e-01 -1.65322810e-01
6.36161193e-02 -1.34026158e+00 3.82989645e-01 4.90123093e-01
-4.38129872e-01 5.06427363e-02 8.26828897e-01 -6.39988482e-01
-2.15106273e+00 -1.03497887e+00 -8.54862574e-03 1.29128456e-01
7.09915519e-01 -4.52271968e-01 -4.02505606e-01 5.90384007e-01
-1.29766911e-01 -1.99801371e-01 2.85956204e-01 -1.81061313e-01
4.92208749e-01 -3.92572314e-01 -1.07374489e+00 5.65914512e-01
1.25384402e+00 5.36003530e-01 -3.78952920e-01 4.13460910e-01
1.02433538e+00 -9.57930326e-01 -8.73441160e-01 4.09869552e-01
2.98368871e-01 -6.78984702e-01 7.09069610e-01 -7.26587057e-01
2.42853165e-01 -5.70003837e-02 -2.25931987e-01 -1.49474263e+00
-6.94018602e-01 -1.02503133e+00 -7.83130825e-01 7.28610098e-01
2.44136557e-01 -6.72823727e-01 1.02554166e+00 4.30790603e-01
-4.99155670e-01 -1.37227035e+00 -1.05803502e+00 -1.17864180e+00
1.32932559e-01 1.04262140e-02 7.19267428e-01 1.43386632e-01
3.36598366e-01 -4.79522534e-02 -3.36068392e-01 4.42176998e-01
5.64308941e-01 7.42716074e-01 7.66340554e-01 -6.16247892e-01
-5.72521567e-01 -4.88932222e-01 -1.17689423e-01 -1.94089758e+00
2.34806240e-02 -5.60392678e-01 5.06576061e-01 -1.64053357e+00
-1.46255761e-01 -8.44314158e-01 -1.20048732e-01 2.48791426e-01
3.75522286e-01 -5.48903942e-01 4.03506398e-01 -5.28950244e-02
-9.33384418e-01 9.30296600e-01 1.69667947e+00 -1.58975720e-01
-4.27820385e-01 4.26585704e-01 -5.56504905e-01 3.83576959e-01
9.15128529e-01 -1.11450583e-01 -1.31750762e-01 -5.40417850e-01
1.04019016e-01 6.19277656e-01 -1.58464164e-01 -8.15878630e-01
1.65762559e-01 -5.70562184e-01 3.43487203e-01 -9.49761629e-01
5.09479403e-01 -1.23226941e+00 -2.74562836e-01 8.45754981e-01
3.78598133e-03 1.27448753e-01 3.28540623e-01 9.53337073e-01
3.06016862e-01 -3.43777657e-01 5.98613739e-01 -6.38804212e-02
-5.23308754e-01 6.54391468e-01 -5.41999459e-01 -2.04297408e-01
1.55538487e+00 -1.49171621e-01 -1.44424289e-01 -8.05208758e-02
-7.53375590e-01 1.05487227e+00 3.36018652e-01 2.47536615e-01
3.82087052e-01 -1.21139789e+00 -1.44397303e-01 1.82016313e-01
-4.13337916e-01 5.60283124e-01 2.64927417e-01 6.36149108e-01
-6.70809805e-01 8.30921769e-01 -3.25051963e-01 -6.60279155e-01
-4.24476922e-01 1.30050135e+00 2.19418183e-01 -6.97722375e-01
-7.58382857e-01 5.19998074e-01 -1.87904522e-01 -8.20295811e-02
5.09270251e-01 -7.61411846e-01 -3.10600717e-02 -1.42752072e-02
3.21512789e-01 4.43580091e-01 1.21342838e-01 1.33130178e-01
-1.53824076e-01 2.66036451e-01 -2.86584914e-01 1.87314078e-01
1.60963058e+00 -1.01177074e-01 3.43883365e-01 1.78112447e-01
6.61757290e-01 -5.06424487e-01 -1.78200388e+00 4.98589166e-02
-9.34874639e-02 -5.14607728e-01 -3.13252836e-01 -5.60221434e-01
-8.42504501e-01 5.65331161e-01 1.77619383e-01 5.97220898e-01
7.60423779e-01 -9.89211444e-03 8.82799804e-01 6.33747101e-01
1.15822852e+00 -1.27353513e+00 1.44460604e-01 7.44865358e-01
8.48853230e-01 -9.56163406e-01 3.97669300e-02 -3.34650993e-01
-6.67253435e-01 1.05872142e+00 7.45757222e-01 -6.05436981e-01
2.87361443e-01 4.18739140e-01 -6.12300396e-01 7.49508590e-02
-7.56209731e-01 -3.00817192e-01 -3.40389788e-01 6.54471040e-01
-4.37283099e-01 3.53573412e-01 -1.34324968e-01 6.10809624e-01
-4.39815491e-01 -3.79540980e-01 3.89736563e-01 1.16009712e+00
-8.41562033e-01 -8.87857616e-01 -3.25599581e-01 4.11099225e-01
-2.16184463e-02 5.61842442e-01 2.31944501e-01 9.87171769e-01
-1.76402345e-01 8.52625906e-01 6.16827846e-01 1.12684071e-01
5.72224677e-01 -3.59628469e-01 8.01399648e-01 -5.72658002e-01
-6.04298990e-03 1.48298323e-01 -1.43703714e-01 -6.75616264e-01
-7.90240467e-02 -7.09037006e-01 -1.30709600e+00 -4.51324880e-01
-4.20405120e-01 3.19459260e-01 5.84296107e-01 8.58559668e-01
4.32684541e-01 7.40486681e-01 8.80188584e-01 -1.60387659e+00
-1.00694799e+00 -6.42534673e-01 -9.06319380e-01 -2.02585265e-01
4.18121755e-01 -1.31818986e+00 5.62609918e-03 -6.01847529e-01] | [4.939422130584717, 2.681424856185913] |
a43ad7dd-bb70-4f2d-801b-f70951be0166 | accurate-nuclear-segmentation-with-center | 1907.03951 | null | https://arxiv.org/abs/1907.03951v2 | https://arxiv.org/pdf/1907.03951v2.pdf | Accurate Nuclear Segmentation with Center Vector Encoding | Nuclear segmentation is important and frequently demanded for pathology image analysis, yet is also challenging due to nuclear crowdedness and possible occlusion. In this paper, we present a novel bottom-up method for nuclear segmentation. The concepts of Center Mask and Center Vector are introduced to better depict the relationship between pixels and nuclear instances. The instance differentiation process are thus largely simplified and easier to understand. Experiments demonstrate the effectiveness of Center Vector Encoding, where our method outperforms state-of-the-arts by a clear margin. | ['Jiahui Li', 'Zhiqiang Hu', 'Shuang Yang'] | 2019-07-09 | null | null | null | null | ['nuclear-segmentation'] | ['medical'] | [ 3.63665879e-01 1.54983044e-01 -2.97330737e-01 -1.75394222e-01
-9.39157605e-01 -3.81899118e-01 3.64476621e-01 4.20924395e-01
-3.37677658e-01 7.92617083e-01 1.10925265e-01 2.19378173e-02
5.69902621e-02 -4.05357808e-01 -2.92392910e-01 -1.10900462e+00
6.04214035e-02 3.72366160e-01 3.85649443e-01 6.54590502e-02
-6.07480071e-02 8.67119372e-01 -1.14474213e+00 1.27531305e-01
8.84539545e-01 7.72735178e-01 1.48614436e-01 6.61894262e-01
-3.04352254e-01 5.18619478e-01 -5.75949788e-01 -3.31318259e-01
-4.03167447e-03 -5.07200718e-01 -8.91634166e-01 5.40573657e-01
3.37635189e-01 -1.61987767e-01 -3.49291027e-01 1.27626336e+00
5.14364362e-01 9.72390994e-02 9.14854765e-01 -9.10683572e-01
-2.52594441e-01 6.03666186e-01 -8.72963786e-01 5.63133121e-01
-2.11286694e-01 2.67092913e-01 9.24322784e-01 -7.39873767e-01
8.65744710e-01 9.33629513e-01 4.34508264e-01 5.51338434e-01
-1.19581556e+00 -2.99467862e-01 2.45523989e-01 1.50047928e-01
-1.60042179e+00 -1.27185479e-01 7.44260609e-01 -5.32154977e-01
4.89214480e-01 5.08104026e-01 8.56661975e-01 6.11553073e-01
1.58481002e-01 1.21345007e+00 1.07192934e+00 -2.28564322e-01
2.76678532e-01 -1.91747263e-01 2.67270684e-01 7.67402411e-01
3.22606981e-01 -4.12625909e-01 -3.46744992e-02 -8.68224949e-02
9.17820096e-01 2.13285595e-01 -5.33830822e-01 -4.94194776e-01
-1.26446474e+00 6.48990810e-01 5.30000269e-01 7.22127378e-01
-1.79282144e-01 2.23368123e-01 5.73429167e-01 -4.52216089e-01
3.81271005e-01 1.80837531e-02 1.75557092e-01 1.75356746e-01
-1.33671021e+00 1.52511925e-01 2.65559852e-01 7.01854765e-01
3.07086051e-01 -1.11962873e-02 -6.35762632e-01 7.00561762e-01
2.51979232e-01 2.66899884e-01 5.00463843e-01 -8.31740439e-01
-2.14176714e-01 8.14067423e-01 -2.06814796e-01 -7.91104078e-01
-6.93468034e-01 -9.14532244e-01 -1.01265383e+00 1.66220769e-01
7.35574901e-01 3.59585583e-01 -1.30776167e+00 1.08280265e+00
6.24766588e-01 3.33861887e-01 -2.52509683e-01 9.77088630e-01
8.71450484e-01 2.37764537e-01 1.31022051e-01 -5.00783920e-01
1.61029053e+00 -1.05236959e+00 -1.22558057e+00 -1.04899459e-01
9.43555892e-01 -3.99468124e-01 7.29094803e-01 1.91136107e-01
-1.17958188e+00 -2.02676669e-01 -1.03488529e+00 -2.20842078e-01
-2.46283695e-01 2.46517122e-01 8.07048142e-01 6.27802849e-01
-8.58628035e-01 4.96571213e-01 -1.17062390e+00 -4.32962954e-01
1.02398884e+00 3.74556988e-01 -3.48223835e-01 -5.68206385e-02
-5.48343003e-01 6.97048545e-01 3.70504707e-01 1.52201444e-01
-6.69237435e-01 -7.92400897e-01 -9.00634527e-01 -8.30185413e-02
3.86686295e-01 -4.54230160e-01 1.15484631e+00 -3.34797472e-01
-1.12317133e+00 1.10803664e+00 -2.42136315e-01 -3.65131646e-01
6.35652542e-01 3.51199090e-01 -2.86736786e-01 4.29269820e-01
-4.60693100e-03 8.07319701e-01 5.46459794e-01 -1.27827585e+00
-6.68704152e-01 -5.04461646e-01 -3.91480565e-01 2.01319024e-01
-1.87680364e-01 -3.40039991e-02 -9.76886630e-01 -6.71673834e-01
3.07078749e-01 -6.85170710e-01 -6.21290267e-01 5.06777406e-01
-5.29632866e-01 -7.83171058e-02 8.64883304e-01 -5.82298040e-01
1.35780227e+00 -2.17031145e+00 2.00221147e-02 2.78974861e-01
6.89286590e-01 1.87747255e-01 3.20839435e-01 -2.76056558e-01
-2.54295636e-02 1.16880648e-01 -5.93746722e-01 -2.76647180e-01
-1.29293412e-01 4.56160396e-01 1.45667806e-01 1.06618595e+00
1.24886855e-01 1.06315029e+00 -9.44532156e-01 -1.06604719e+00
2.89253205e-01 6.63928449e-01 -3.71445298e-01 -1.60525367e-01
9.88098606e-02 4.73490894e-01 -3.60056400e-01 1.21934187e+00
7.79613376e-01 -3.28417033e-01 2.36253709e-01 -4.18015301e-01
9.29086059e-02 -4.72225517e-01 -1.10409403e+00 1.68803418e+00
2.66903639e-01 5.43643892e-01 3.09949368e-01 -7.81879783e-01
5.82040370e-01 1.25383630e-01 6.20554388e-01 -3.00988019e-01
6.56618297e-01 1.83147028e-01 2.81786233e-01 -2.08132878e-01
3.74772906e-01 -2.55976379e-01 9.60678980e-02 9.30484477e-03
-1.06169023e-01 -3.41453224e-01 5.69937110e-01 1.29628688e-01
8.74216318e-01 -2.17476115e-01 6.61818862e-01 -3.46704513e-01
6.00177348e-01 -2.90406905e-02 9.75191891e-01 4.33594644e-01
-8.38133752e-01 7.55673707e-01 7.59456396e-01 -1.81016371e-01
-6.68061912e-01 -8.57186973e-01 -2.93517172e-01 3.93946171e-01
2.10905612e-01 -2.99139738e-01 -9.29738700e-01 -1.05428791e+00
-9.64336842e-02 3.89623672e-01 -1.05030775e+00 3.12219113e-01
-7.80237317e-01 -8.18056762e-01 5.18963516e-01 9.07017708e-01
-2.01608390e-02 -4.66796577e-01 -5.35862923e-01 9.44897160e-02
-3.88973169e-02 -1.25205886e+00 -5.50087929e-01 1.18937559e-01
-1.13948250e+00 -9.24200237e-01 -1.27718055e+00 -7.82873929e-01
1.06896126e+00 1.78107694e-01 1.01042306e+00 4.10824090e-01
-8.43133450e-01 -5.21594547e-02 -1.39278278e-01 -3.58383775e-01
-1.82441279e-01 -1.09600425e-01 -2.45947734e-01 3.00256070e-02
-9.97875538e-03 -2.12100297e-01 -8.61679792e-01 2.53469497e-01
-1.32347655e+00 -3.81858051e-02 6.72507405e-01 1.00349545e+00
1.20553684e+00 8.83283764e-02 -6.00527041e-02 -1.09712589e+00
1.55173466e-01 -7.78069720e-02 -6.07930064e-01 1.81095064e-01
-2.34209076e-01 -1.20989993e-01 2.55629212e-01 -2.23167792e-01
-7.12700427e-01 2.99825400e-01 -5.86882383e-02 -4.64513689e-01
-1.50964528e-01 4.89479080e-02 -1.14396401e-01 -3.63520324e-01
3.27230990e-01 -1.51434662e-02 -2.59665418e-02 -1.93704218e-01
4.78245944e-01 2.72422016e-01 7.65630782e-01 -5.43134101e-02
6.12219036e-01 1.13206756e+00 2.41267905e-01 -7.85401285e-01
-6.06456995e-01 -7.03829587e-01 -7.66281962e-01 -3.81401360e-01
9.64615583e-01 -4.67770636e-01 -5.64734161e-01 2.80292839e-01
-9.38917160e-01 -9.46919695e-02 -7.57185578e-01 3.88859391e-01
-3.65458041e-01 4.59531039e-01 -7.82828867e-01 -6.46657526e-01
-2.12455273e-01 -1.54179060e+00 1.01184893e+00 5.36443233e-01
9.28140339e-03 -1.02698493e+00 -1.62596226e-01 4.23710942e-01
6.45108372e-02 3.77788067e-01 9.18446898e-01 -6.66892648e-01
-4.96141970e-01 -2.28184193e-01 -4.38467413e-01 6.07337616e-02
4.70549054e-02 2.69632667e-01 -9.66001391e-01 -2.43362948e-01
-2.40138635e-01 9.48916152e-02 9.01002467e-01 8.02235186e-01
1.48947585e+00 2.96777397e-01 -9.29615438e-01 7.55547225e-01
1.40857494e+00 3.91085930e-02 8.04761291e-01 1.56618785e-02
5.94129443e-01 5.82713723e-01 6.95477962e-01 2.16842443e-01
1.65269092e-01 5.19829750e-01 4.69944626e-01 -7.60399997e-01
-4.92019296e-01 1.55496597e-01 -4.68827397e-01 7.44093716e-01
1.02242112e-01 -2.27487773e-01 -9.04202759e-01 8.28013718e-01
-1.71349895e+00 -6.68559849e-01 -1.92175508e-01 1.68491697e+00
8.09115529e-01 9.10897329e-02 -2.37172972e-02 4.77413893e-01
6.92925334e-01 1.50383666e-01 -5.89883924e-01 -6.93270788e-02
-2.87630796e-01 1.23521797e-01 7.01484323e-01 3.97761822e-01
-1.15803814e+00 7.63855398e-01 7.36038589e+00 1.43540573e+00
-8.50780129e-01 1.56872541e-01 1.10750222e+00 -8.26752186e-02
-3.07550699e-01 -1.88133165e-01 -9.80290592e-01 3.46568286e-01
1.93492219e-01 1.59426957e-01 -2.11421490e-01 5.05864680e-01
-4.33373041e-02 -2.83630550e-01 -1.00301516e+00 1.09699512e+00
2.24601150e-01 -1.58749700e+00 -1.51472986e-01 3.32455844e-01
8.21441114e-01 -3.18847805e-01 -9.58046094e-02 4.95232046e-02
-9.34084281e-02 -1.11120260e+00 3.34665626e-01 3.93872887e-01
8.67830217e-01 -6.96194530e-01 9.37299252e-01 1.34447357e-03
-1.21498334e+00 9.62643698e-02 -3.06785405e-01 5.06028354e-01
3.12263548e-01 9.06923711e-01 -8.33509862e-01 4.00482535e-01
4.36035037e-01 3.59058410e-01 -5.97097576e-01 1.42028725e+00
-1.53402910e-01 3.54163826e-01 -2.33518094e-01 -7.54752979e-02
3.23007345e-01 -1.57478482e-01 5.34194231e-01 1.51261544e+00
1.59285307e-01 3.18982571e-01 2.80670226e-01 7.94313908e-01
-1.46076187e-01 2.78613150e-01 -2.57100254e-01 -7.35665411e-02
7.17424005e-02 1.76497734e+00 -1.47155213e+00 -3.27200741e-01
-2.82543898e-01 1.05552089e+00 1.38059646e-01 2.36801669e-01
-9.18500006e-01 -2.77668208e-01 5.06676614e-01 3.06366801e-01
3.88389587e-01 -7.57836029e-02 -3.65615904e-01 -7.48582363e-01
-2.95066029e-01 -4.72309083e-01 4.65302169e-01 -4.28090811e-01
-1.18405497e+00 3.99306446e-01 -1.51808321e-01 -1.20775497e+00
1.82502747e-01 -7.08103120e-01 -4.66764122e-01 5.50262809e-01
-1.64193428e+00 -1.12369621e+00 -3.70867908e-01 3.85592699e-01
5.52193165e-01 2.29403004e-01 5.46887219e-01 2.86729634e-01
-7.36983120e-01 6.69275105e-01 2.16558456e-01 3.12088579e-01
3.26733291e-01 -1.34304261e+00 -5.71072772e-02 9.19377744e-01
-1.03960030e-01 2.78080672e-01 5.57587743e-01 -6.30962312e-01
-1.20453024e+00 -9.35866356e-01 6.74304605e-01 -2.61836615e-03
6.45657301e-01 -2.28032142e-01 -8.38364065e-01 2.97971159e-01
1.36591136e-01 7.28164375e-01 9.34465587e-01 -3.70270222e-01
1.75623551e-01 -5.95643446e-02 -1.50557208e+00 9.44134831e-01
6.95927382e-01 -2.09392071e-01 -1.53740242e-01 4.98005062e-01
5.75564384e-01 -7.39233494e-01 -9.70277131e-01 4.41749871e-01
2.34691143e-01 -8.00987482e-01 9.47319984e-01 -1.39059156e-01
-7.16990903e-02 -5.97635686e-01 -1.13538429e-01 -1.00979352e+00
-4.33147758e-01 -4.49314535e-01 -2.48431608e-01 1.10435951e+00
3.15738499e-01 -1.71373054e-01 1.19985938e+00 3.23944211e-01
-3.13967794e-01 -1.15419459e+00 -1.18510365e+00 -6.03070438e-01
7.17515498e-03 -3.49679530e-01 3.70123029e-01 8.53045762e-01
5.54946112e-03 3.03210989e-02 1.19989991e-01 6.28116652e-02
7.75011003e-01 -1.08044356e-01 2.39218518e-01 -1.03934753e+00
-1.58233494e-01 -9.14501309e-01 -8.77861321e-01 -8.37966323e-01
-1.62268966e-01 -9.04137611e-01 -9.35573131e-02 -1.68073690e+00
3.59588981e-01 -1.42150819e-01 -3.83637428e-01 2.91674078e-01
-3.43333274e-01 9.40667331e-01 1.42516926e-01 3.07303816e-02
-6.16883814e-01 4.94808793e-01 1.74143553e+00 -4.32386369e-01
4.27990779e-02 -2.79443562e-01 -6.70614541e-01 9.39040005e-01
5.50249100e-01 -3.41442436e-01 -1.14954837e-01 -3.63243930e-02
-2.27775052e-01 -1.64325342e-01 1.93171442e-01 -1.00317454e+00
2.40600958e-01 2.65215300e-02 6.79102719e-01 -1.03046751e+00
4.53213751e-01 -7.10589647e-01 -2.36370236e-01 6.06338918e-01
-1.28323540e-01 -2.44248882e-01 1.49876177e-01 5.52399695e-01
-2.46550873e-01 -2.13658169e-01 1.05683231e+00 -1.29539192e-01
-3.95068198e-01 3.41431141e-01 -2.06825703e-01 4.01917435e-02
1.40462101e+00 -5.42214513e-01 -2.11466834e-01 -2.99471356e-02
-9.15167511e-01 3.46629649e-01 3.69808495e-01 -3.57495517e-01
6.75047338e-01 -1.32908702e+00 -6.50236011e-01 1.71195149e-01
1.19931839e-01 2.57105920e-02 6.16508305e-01 1.42297196e+00
-7.99493074e-01 3.43256027e-01 1.16532035e-01 -1.05849469e+00
-1.44985187e+00 4.47320729e-01 4.49674517e-01 -6.06442928e-01
-9.27112162e-01 1.14588797e+00 5.41794777e-01 -1.97560564e-01
4.73099649e-01 -6.79344893e-01 -4.34072256e-01 2.51089390e-02
7.48477817e-01 4.40728426e-01 1.80770501e-01 -6.58183098e-01
-5.15523911e-01 5.79287231e-01 -2.58896172e-01 1.22499213e-01
8.73681068e-01 3.36446837e-02 -1.31827563e-01 2.56779462e-01
8.80912423e-01 -9.49482471e-02 -1.27060127e+00 -2.90445596e-01
-4.45369743e-02 -5.20908773e-01 4.47336435e-01 -6.64522171e-01
-1.36951733e+00 8.62909615e-01 8.79867435e-01 -9.41439494e-02
1.09701240e+00 2.97409922e-01 8.26051712e-01 -2.12049842e-01
-1.25728518e-01 -1.24831283e+00 -2.99095660e-01 1.41303003e-01
4.66031641e-01 -9.88933682e-01 3.79496306e-01 -7.66237557e-01
-6.97150230e-01 8.34855676e-01 4.07191515e-01 1.00103036e-01
5.31020582e-01 6.76343739e-01 6.63189664e-02 -2.44084105e-01
-3.67387086e-01 -5.58328152e-01 4.56020534e-01 7.77167678e-01
5.00323594e-01 1.38050929e-01 -5.66809773e-01 5.12584925e-01
2.56457001e-01 -3.40459794e-01 3.68570864e-01 9.51683879e-01
-4.90715921e-01 -9.93064940e-01 -6.15121782e-01 5.46972573e-01
-7.02341378e-01 8.94159824e-02 -2.07638532e-01 9.82014716e-01
2.69035041e-01 4.84183222e-01 2.40057424e-01 1.14878111e-01
2.37542287e-01 -3.46354514e-01 6.03362679e-01 -3.41586500e-01
-5.68731725e-01 7.42045581e-01 -3.22799176e-01 -4.97498810e-01
-2.86243707e-01 -8.14316690e-01 -1.83071601e+00 -2.39977948e-02
-4.43355978e-01 -5.04920520e-02 6.88531101e-01 7.84020305e-01
-1.65686794e-02 9.37355578e-01 2.51021981e-01 -7.56039202e-01
-2.55592734e-01 -6.45245910e-01 -1.03325307e+00 2.69436687e-01
9.49918702e-02 -6.07563436e-01 -2.24774666e-02 -1.81953199e-02] | [14.971701622009277, -3.042515277862549] |
dcad0e9a-d38d-492f-9da2-1ed7e46e2099 | learning-an-explicit-hyperparameter | 2107.02378 | null | https://arxiv.org/abs/2107.02378v3 | https://arxiv.org/pdf/2107.02378v3.pdf | Learning an Explicit Hyperparameter Prediction Function Conditioned on Tasks | Meta learning has attracted much attention recently in machine learning community. Contrary to conventional machine learning aiming to learn inherent prediction rules to predict labels for new query data, meta learning aims to learn the learning methodology for machine learning from observed tasks, so as to generalize to new query tasks by leveraging the meta-learned learning methodology. In this study, we interpret such learning methodology as learning an explicit hyper-parameter prediction function shared by all training tasks. Specifically, this function is represented as a parameterized function called meta-learner, mapping from a training/test task to its suitable hyper-parameter setting, extracted from a pre-specified function set called meta learning machine. Such setting guarantees that the meta-learned learning methodology is able to flexibly fit diverse query tasks, instead of only obtaining fixed hyper-parameters by many current meta learning methods, with less adaptability to query task's variations. Such understanding of meta learning also makes it easily succeed from traditional learning theory for analyzing its generalization bounds with general losses/tasks/models. The theory naturally leads to some feasible controlling strategies for ameliorating the quality of the extracted meta-learner, verified to be able to finely ameliorate its generalization capability in some typical meta learning applications, including few-shot regression, few-shot classification and domain generalization. | ['Zongben Xu', 'Deyu Meng', 'Jun Shu'] | 2021-07-06 | null | null | null | null | ['parameter-prediction'] | ['miscellaneous'] | [ 4.62967128e-01 3.45725447e-01 -5.72713971e-01 -4.23696756e-01
-8.69763613e-01 -1.75906643e-01 5.69701552e-01 3.50098759e-01
-3.70444268e-01 8.78333211e-01 -3.62374872e-01 1.71564579e-01
-7.48917162e-01 -9.74423468e-01 -6.80254400e-01 -7.07000256e-01
-4.97878306e-02 4.45669770e-01 1.21146932e-01 -3.93026918e-01
3.04675460e-01 2.61514366e-01 -2.03396940e+00 3.31339180e-01
9.81474400e-01 1.09834230e+00 2.39135832e-01 6.95646167e-01
-3.38917077e-01 5.47115505e-01 -6.00489080e-01 -4.61021513e-01
1.31054163e-01 -4.79306906e-01 -8.70574892e-01 -2.09364924e-03
2.14088827e-01 4.41666842e-01 -3.39951590e-02 1.08205664e+00
3.08591127e-01 5.65707505e-01 1.23776913e+00 -1.32591355e+00
-6.25447690e-01 5.10983765e-01 -2.43226886e-01 3.62065136e-01
7.57213458e-02 2.65496764e-02 9.69311476e-01 -6.98628247e-01
4.35673714e-01 9.49727535e-01 4.40700740e-01 7.40551233e-01
-1.29084063e+00 -4.79180723e-01 -2.63949931e-01 2.32269585e-01
-1.01429498e+00 -1.53891236e-01 7.73938954e-01 -3.70315254e-01
6.04079425e-01 2.28477299e-01 3.77354473e-01 1.30861545e+00
2.06532553e-01 5.84329724e-01 9.47653830e-01 -6.90700889e-01
6.71075702e-01 7.14978337e-01 2.69542605e-01 5.31362176e-01
1.12742849e-01 7.49122128e-02 -3.91157538e-01 -2.58177429e-01
1.15789451e-01 1.78217411e-01 -9.62909088e-02 -6.03920519e-01
-8.22446823e-01 8.26716900e-01 2.78356254e-01 5.20434916e-01
-3.35842311e-01 8.27949122e-02 7.21395731e-01 7.41203427e-01
6.83449686e-01 7.23210275e-01 -6.61576092e-01 3.00037768e-02
-4.61426020e-01 2.02665374e-01 6.95306122e-01 9.29669380e-01
1.27561986e+00 1.87397718e-01 -4.74409401e-01 1.03826034e+00
-2.08938241e-01 1.77812576e-01 1.08321452e+00 -7.61083543e-01
4.48463589e-01 8.47725630e-01 -3.44132259e-02 -5.54362416e-01
-3.91708016e-01 -8.24967563e-01 -7.37383187e-01 2.40617633e-01
1.42773733e-01 -2.16053650e-01 -5.69041431e-01 1.74563742e+00
4.80062962e-01 2.93122411e-01 1.54765829e-01 3.72795641e-01
2.62480348e-01 6.14539921e-01 2.67811388e-01 -6.85987651e-01
1.08526218e+00 -7.09682763e-01 -3.51981729e-01 -8.47608503e-03
9.44634974e-01 -1.12820834e-01 1.47710383e+00 5.18188477e-01
-7.49302089e-01 -7.67779827e-01 -1.10480917e+00 4.30539310e-01
-6.98857844e-01 -2.66046435e-01 4.10876423e-01 6.80933297e-01
-7.06967354e-01 7.53632724e-01 -1.15075201e-01 -4.01009172e-01
3.67828339e-01 2.94688910e-01 -1.27432600e-01 2.35147670e-01
-1.21646488e+00 8.50488484e-01 1.13725162e+00 -3.16643506e-01
-8.25003624e-01 -9.51534808e-01 -5.55259049e-01 2.72315532e-01
7.52827704e-01 -9.54406500e-01 1.34248328e+00 -1.57463348e+00
-1.53155160e+00 9.31162119e-01 1.44230574e-01 -6.49075389e-01
5.18691242e-01 -3.39756832e-02 -5.69999456e-01 -7.99950212e-02
-2.08129570e-01 3.18090826e-01 1.67498696e+00 -1.18458283e+00
-7.93317616e-01 -4.28314686e-01 -1.28310859e-01 8.94346610e-02
-7.67298758e-01 -5.01975417e-01 1.63502082e-01 -6.30031168e-01
-3.77089113e-01 -6.05346084e-01 -5.71383648e-02 -2.35839441e-01
-1.35013983e-01 -3.76458168e-01 8.65199506e-01 1.69695020e-01
1.20349801e+00 -1.96021795e+00 2.25716487e-01 3.43279988e-02
7.31101912e-03 4.81382936e-01 -2.93346852e-01 5.52192986e-01
-7.74481222e-02 -2.17224453e-02 -2.54971474e-01 -3.46024707e-02
-1.78125337e-01 1.54783592e-01 -4.79655445e-01 2.38000080e-01
3.82084772e-02 9.10400867e-01 -9.91009474e-01 -4.43761021e-01
2.75182724e-01 2.08909586e-01 -3.67272764e-01 4.08784777e-01
-7.34844565e-01 8.10520798e-02 -7.90772676e-01 5.19487202e-01
3.20804030e-01 -1.84889600e-01 -1.43777072e-01 4.86326031e-02
2.60837644e-01 -3.74781847e-01 -9.46050942e-01 1.54376268e+00
-1.03767943e+00 2.08327860e-01 -4.01152909e-01 -1.76767576e+00
1.45091784e+00 1.61867097e-01 4.73436981e-01 -5.43274403e-01
2.26378694e-01 1.60550728e-01 -2.98385143e-01 -8.75773430e-01
3.07297468e-01 -3.55073273e-01 -2.27353461e-02 3.32439691e-01
4.24338013e-01 1.35630265e-01 -5.60692810e-02 -3.18867534e-01
8.45227838e-01 -1.81622863e-01 6.36940956e-01 -1.91765308e-01
7.09471405e-01 3.93175185e-02 3.07383835e-01 9.44135487e-01
-1.43372212e-02 1.28440827e-01 1.57117173e-01 -4.83955175e-01
-1.00283217e+00 -7.82218099e-01 -2.50642866e-01 1.61382246e+00
-2.96700411e-02 -1.58510506e-02 -6.25440300e-01 -9.47312891e-01
2.35348195e-01 8.61107171e-01 -8.57768655e-01 -8.65288675e-01
-1.97763279e-01 -6.51719213e-01 4.08589005e-01 9.60318223e-02
4.53178823e-01 -1.33741176e+00 -8.56152534e-01 4.50476974e-01
4.59718198e-01 -5.22177577e-01 -2.20316648e-01 4.85600561e-01
-1.23567939e+00 -1.18328094e+00 -7.02376306e-01 -6.39306962e-01
1.73388124e-01 2.28880271e-01 9.81083810e-01 -4.47003506e-02
-5.72749496e-01 5.84707737e-01 -5.49989223e-01 -5.37739098e-01
-6.44876480e-01 3.75219524e-01 2.76167989e-01 5.29621959e-01
4.01476651e-01 -7.21919119e-01 -3.67952436e-01 1.79437757e-01
-1.22645450e+00 -3.26465398e-01 5.85324585e-01 1.19534147e+00
5.21571934e-01 1.41141266e-01 1.30467927e+00 -1.24076486e+00
6.71448708e-01 -8.87273371e-01 -5.96756756e-01 7.45469928e-01
-1.08871138e+00 4.64700878e-01 9.76067662e-01 -7.21271634e-01
-1.10684752e+00 -1.95492640e-01 3.97558510e-01 -5.01414180e-01
-1.76937923e-01 4.34466928e-01 -2.28219539e-01 -5.64505979e-02
1.07013154e+00 3.69207889e-01 -4.11984697e-02 -4.21948284e-01
5.67852259e-01 8.02689970e-01 4.08199102e-01 -7.69917190e-01
7.52407134e-01 2.72500247e-01 2.45545983e-01 -9.26467299e-01
-1.25874734e+00 -5.42075515e-01 -5.98747551e-01 -2.53879160e-01
6.41290486e-01 -3.02041829e-01 -7.33474553e-01 1.50574669e-01
-7.43902266e-01 -2.99314469e-01 -6.94379389e-01 4.34827268e-01
-1.10922682e+00 -1.62514478e-01 9.46421325e-02 -1.08410919e+00
-3.50123525e-01 -7.21667230e-01 7.72162974e-01 2.23890916e-01
5.32913208e-02 -1.42326951e+00 4.89065379e-01 7.79832527e-02
3.49424243e-01 2.48350784e-01 1.24805772e+00 -1.13392878e+00
-1.75407946e-01 -1.70369297e-01 1.26913235e-01 5.82569540e-01
2.59129763e-01 -2.00095445e-01 -1.30320930e+00 -3.09186846e-01
3.84007782e-01 -6.45754874e-01 9.67959821e-01 2.48482347e-01
1.42540824e+00 -4.99865443e-01 -2.64200330e-01 8.08769226e-01
1.61704481e+00 1.48188651e-01 2.28421435e-01 4.94066536e-01
1.11085087e-01 6.31914020e-01 8.72822285e-01 4.70009506e-01
-2.78126031e-01 6.81287050e-01 4.92335916e-01 6.48819149e-01
4.21375066e-01 -3.63943577e-01 1.39306813e-01 2.11277083e-01
1.44090652e-01 -1.49137098e-02 -7.69759893e-01 1.18443109e-01
-1.97772241e+00 -1.15226245e+00 5.06400228e-01 2.56473756e+00
6.34641171e-01 2.11259201e-01 2.26089895e-01 1.28405809e-01
7.72108614e-01 1.89423651e-01 -8.88325155e-01 -5.46803832e-01
1.10974476e-01 2.24153847e-01 1.97991148e-01 1.84959382e-01
-1.00622153e+00 6.27269864e-01 5.48693609e+00 1.08002269e+00
-1.11993003e+00 1.17109790e-01 4.88999963e-01 2.82448120e-02
-3.30293924e-01 -1.05030768e-01 -7.54349053e-01 2.73731411e-01
1.11943030e+00 -6.16459846e-01 3.26658458e-01 1.26050174e+00
1.00941963e-01 1.90088049e-01 -1.31017435e+00 9.56508815e-01
7.44456649e-02 -1.40340066e+00 4.23866570e-01 7.28955446e-03
7.75163472e-01 -4.29999888e-01 4.35145408e-01 1.02181411e+00
-2.89598972e-01 -8.89598310e-01 2.07594350e-01 9.06007588e-01
6.86408997e-01 -9.00343716e-01 3.62634599e-01 9.47531223e-01
-7.56047785e-01 -6.85208142e-01 -7.67573595e-01 1.54495314e-01
-4.05622840e-01 3.51816654e-01 -9.38224435e-01 4.54029679e-01
2.17550352e-01 5.62259853e-01 -6.46676660e-01 1.13414848e+00
4.17485654e-01 3.90951782e-01 2.48339325e-01 -3.05126280e-01
1.31093994e-01 -1.22441255e-01 7.07667291e-01 9.94416714e-01
4.02300388e-01 -2.73946673e-02 5.19963913e-02 7.42726266e-01
-2.12305248e-01 3.95063311e-01 -9.71905470e-01 8.46732706e-02
5.72963655e-01 1.08221972e+00 -2.19677851e-01 -3.68058830e-01
-1.25743046e-01 8.64938855e-01 6.65264547e-01 5.03908515e-01
-5.25516689e-01 -3.92445862e-01 5.16920745e-01 7.46556818e-02
8.71614814e-02 3.49544227e-01 -1.97439834e-01 -1.14153492e+00
-1.12952068e-01 -6.69276237e-01 7.49618471e-01 -4.14156467e-01
-1.45498240e+00 4.71474975e-01 2.39198044e-01 -1.50021613e+00
-5.98403037e-01 -4.85045850e-01 -8.17892194e-01 6.50624871e-01
-1.39482725e+00 -9.14418817e-01 -1.76564366e-01 6.55389905e-01
8.69311333e-01 -7.86444247e-01 8.16779017e-01 -2.58168280e-01
-3.24879438e-01 8.28017175e-01 3.39722693e-01 -4.64240849e-01
8.22094738e-01 -1.30498219e+00 -2.87504613e-01 1.39431387e-01
2.55358398e-01 4.20420676e-01 8.23487997e-01 -3.30764472e-01
-1.41325259e+00 -1.27372956e+00 4.13554966e-01 -3.68538201e-01
6.19924605e-01 -7.09568784e-02 -1.33694959e+00 4.44527417e-01
-2.64129281e-01 2.05345303e-01 4.84897465e-01 2.43781433e-01
-6.37006104e-01 -6.15790606e-01 -1.40530860e+00 2.67031074e-01
7.49616385e-01 -5.14095664e-01 -8.03873837e-01 3.23515177e-01
8.15853655e-01 1.22446895e-01 -8.78562033e-01 4.51581925e-01
3.98640692e-01 -9.99146223e-01 9.77114558e-01 -1.41458261e+00
2.92220831e-01 2.98630327e-01 -2.31087789e-01 -1.58709109e+00
-1.80099562e-01 -5.77767372e-01 -7.05877006e-01 1.16852844e+00
3.41007948e-01 -5.88607371e-01 8.01956713e-01 3.84641439e-01
-1.21024132e-01 -9.18003500e-01 -1.03829050e+00 -1.21039271e+00
1.29306123e-01 -4.54399437e-01 5.08037567e-01 1.02611077e+00
6.08659685e-02 3.91951919e-01 -3.43473166e-01 -3.30808125e-02
9.87718284e-01 1.99444704e-02 8.21014047e-01 -1.48971319e+00
-6.96327507e-01 -5.71262181e-01 -4.29787517e-01 -2.90755987e-01
5.39057791e-01 -1.12856638e+00 -2.65609980e-01 -8.58934820e-01
7.58289993e-02 -6.25704110e-01 -6.45394862e-01 1.73952073e-01
-1.95283368e-01 -1.86655134e-01 2.84819584e-02 1.81658909e-01
-5.13939083e-01 6.40260816e-01 1.14037895e+00 -4.23398823e-01
-7.28005886e-01 8.32845330e-01 -6.58896565e-01 5.60580671e-01
8.13859940e-01 -6.47064686e-01 -8.60752046e-01 1.68786541e-01
2.89847136e-01 3.20766538e-01 3.61744791e-01 -9.50214565e-01
2.19076678e-01 -5.45468032e-01 1.82406425e-01 2.73958221e-02
2.67081112e-01 -8.23018730e-01 -6.20115213e-02 3.69077146e-01
-8.33017588e-01 -4.55789208e-01 -1.78137690e-01 1.00597060e+00
-2.16436654e-01 -9.03059900e-01 9.50808346e-01 -1.50961027e-01
-8.98439348e-01 4.58453089e-01 5.19004799e-02 4.49556619e-01
1.39072645e+00 -3.17889541e-01 -1.60214663e-01 -9.32617486e-02
-9.39160585e-01 4.02723141e-02 2.02159166e-01 5.28174579e-01
6.51103735e-01 -1.24548793e+00 -5.69443762e-01 7.88394138e-02
6.75734818e-01 -2.92378396e-01 3.74640554e-01 3.67322892e-01
1.36562213e-01 2.08904013e-01 -7.57824555e-02 -5.56913853e-01
-8.56332242e-01 1.15032971e+00 5.79832852e-01 -1.85182646e-01
-6.47289753e-01 4.60503995e-01 -2.94210669e-02 -2.69976228e-01
4.28431153e-01 -1.80568054e-01 -2.03977719e-01 2.92404503e-01
5.87853372e-01 6.70063376e-01 5.38603626e-02 -9.89343524e-02
1.31877765e-01 4.99155641e-01 -4.95769233e-02 5.88598587e-02
1.18962646e+00 2.16878057e-02 3.10222626e-01 1.08509815e+00
1.63591528e+00 -5.05278051e-01 -1.09679091e+00 -4.96581912e-01
6.62518561e-01 -2.31275380e-01 -1.65031165e-01 -6.52278125e-01
-5.83029449e-01 8.79450619e-01 6.98939025e-01 4.46955442e-01
1.02334130e+00 -5.51149435e-02 5.04944921e-01 1.01837134e+00
6.37117982e-01 -1.30738950e+00 4.58047986e-01 3.03214788e-01
6.95259750e-01 -1.46203244e+00 -3.84514928e-01 1.79108158e-01
-5.71226597e-01 1.45766449e+00 5.49508154e-01 -1.71417177e-01
8.67643952e-01 -2.61536270e-01 -3.00088644e-01 -1.01639636e-01
-1.07673705e+00 -1.57928258e-01 5.25764048e-01 8.50373387e-01
5.57849221e-02 -2.76697129e-02 3.87990922e-02 5.50092876e-01
1.20828651e-01 1.96007073e-01 3.72127831e-01 5.82019687e-01
-1.18943942e+00 -1.09597611e+00 -1.22801758e-01 8.07876766e-01
-2.19934783e-03 3.36739480e-01 -3.96384075e-02 8.74232709e-01
1.85441021e-02 5.34050524e-01 -1.18761554e-01 -3.79295677e-01
4.61023152e-01 5.37912607e-01 3.89977813e-01 -1.08884025e+00
-6.45518720e-01 -6.02117240e-01 -3.92632633e-01 -2.55524337e-01
-2.93534338e-01 -2.71362245e-01 -8.67344320e-01 1.79789811e-01
-3.09259653e-01 4.55984801e-01 5.18709838e-01 1.17034233e+00
7.85005912e-02 3.94634336e-01 1.02558386e+00 -4.98800039e-01
-1.31708264e+00 -9.39526200e-01 -8.40146422e-01 4.35035944e-01
4.70836222e-01 -7.60504127e-01 -4.81031746e-01 -1.78290471e-01] | [9.99670124053955, 3.1926109790802] |
9ec10025-508e-4195-a3c4-747e6ed97bcf | using-aspect-extraction-approaches-to | 1804.08666 | null | https://arxiv.org/abs/1804.08666v2 | https://arxiv.org/pdf/1804.08666v2.pdf | Using Aspect Extraction Approaches to Generate Review Summaries and User Profiles | Reviews of products or services on Internet marketplace websites contain a rich amount of information. Users often wish to survey reviews or review snippets from the perspective of a certain aspect, which has resulted in a large body of work on aspect identification and extraction from such corpora. In this work, we evaluate a newly-proposed neural model for aspect extraction on two practical tasks. The first is to extract canonical sentences of various aspects from reviews, and is judged by human evaluators against alternatives. A $k$-means baseline does remarkably well in this setting. The second experiment focuses on the suitability of the recovered aspect distributions to represent users by the reviews they have written. Through a set of review reranking experiments, we find that aspect-based profiles can largely capture notions of user preferences, by showing that divergent users generate markedly different review rankings. | ['Avneesh Saluja', 'Skyler Wharton', 'Christopher Mitcheltree'] | 2018-04-23 | using-aspect-extraction-approaches-to-1 | https://aclanthology.org/N18-3009 | https://aclanthology.org/N18-3009.pdf | naacl-2018-6 | ['aspect-extraction'] | ['natural-language-processing'] | [ 2.47908875e-01 3.43125671e-01 -6.58345878e-01 -8.34276617e-01
-9.12252128e-01 -8.08898926e-01 7.73577750e-01 3.41370076e-01
-4.68464583e-01 3.59454453e-01 5.93687654e-01 -5.29290020e-01
1.46561205e-01 -6.42621934e-01 -2.91131169e-01 -3.01360726e-01
3.61116856e-01 5.97704828e-01 -3.40724468e-01 -4.41777319e-01
5.86712301e-01 1.15895793e-01 -1.46243787e+00 6.19477630e-01
5.37085414e-01 1.06049347e+00 -1.83833987e-01 5.71370959e-01
-3.41891050e-01 1.96265891e-01 -7.94453204e-01 -1.04910576e+00
2.85432488e-01 -3.90318215e-01 -5.09569228e-01 4.86376673e-01
4.41902816e-01 -1.59414113e-01 1.45359993e-01 1.09800529e+00
3.10243696e-01 -2.92692631e-02 9.76989329e-01 -8.72710764e-01
-1.02315402e+00 9.45487201e-01 -6.61192715e-01 -7.59647265e-02
3.01384777e-01 -1.42016187e-01 2.03811312e+00 -1.01744461e+00
5.44819057e-01 1.04334962e+00 3.80244344e-01 3.96206319e-01
-1.08970618e+00 -9.44684148e-02 5.66622794e-01 -1.32689834e-01
-8.43274415e-01 -3.16373318e-01 9.29647684e-01 -1.90166295e-01
1.06892586e+00 2.45918185e-01 7.81555235e-01 6.95468068e-01
1.77357122e-01 1.06510520e+00 9.82316017e-01 -2.90014714e-01
3.06893200e-01 8.69363546e-01 6.69829547e-01 8.74288604e-02
6.35710835e-01 -5.54039836e-01 -5.56672394e-01 -4.30870026e-01
1.40169740e-01 -7.94977620e-02 -1.60356909e-01 -4.68738407e-01
-7.72901237e-01 1.06620443e+00 3.87651511e-02 -2.29057018e-02
-4.39871311e-01 -3.66367310e-01 4.07343715e-01 5.32709658e-01
8.00086856e-01 8.55643272e-01 -9.38543677e-01 -6.99940845e-02
-6.67324841e-01 4.00906980e-01 1.18188667e+00 1.08179891e+00
7.81336486e-01 -2.11357877e-01 1.49052991e-02 1.19088042e+00
4.42091644e-01 5.27932703e-01 5.33126593e-01 -5.70127428e-01
3.27060044e-01 8.00604045e-01 2.20303044e-01 -9.80650902e-01
-1.95491016e-01 -4.28448111e-01 -4.67657715e-01 -1.09380625e-01
-6.97336122e-02 -3.07509482e-01 -8.27616036e-01 1.43662298e+00
-2.55029555e-02 -9.19151962e-01 -1.20128635e-02 5.86739719e-01
5.94898283e-01 3.50073308e-01 -4.24967147e-03 -2.88218707e-01
1.52807701e+00 -8.81952465e-01 -5.83903730e-01 -5.25691032e-01
4.16344404e-01 -9.22046959e-01 1.29950583e+00 6.00322962e-01
-1.14938092e+00 -3.01977009e-01 -1.09838247e+00 6.38374239e-02
-4.90777969e-01 4.85386997e-01 9.96849179e-01 8.26397479e-01
-1.16130662e+00 4.66836095e-01 -2.04769135e-01 -3.04788619e-01
3.63602847e-01 1.82055622e-01 -1.80000424e-01 3.99468392e-02
-1.00319707e+00 8.39770377e-01 -3.71516258e-01 6.42818734e-02
-2.04370081e-01 -3.57513338e-01 -9.64064837e-01 1.13982216e-01
1.30350336e-01 -7.05492496e-01 1.64559722e+00 -1.37838161e+00
-1.17764783e+00 8.37267935e-01 -4.35973585e-01 -1.08959757e-01
-1.21461481e-01 -4.76706624e-01 -6.21846437e-01 -8.82803872e-02
3.21722656e-01 2.83181608e-01 9.75921869e-01 -1.27104640e+00
-7.20083356e-01 -5.32406986e-01 3.29079241e-01 5.03738344e-01
-5.14983237e-01 1.57693803e-01 -5.32226920e-01 -4.05849934e-01
-2.07814977e-01 -1.07099533e+00 -5.42672276e-01 -4.20533776e-01
-5.06343007e-01 -3.79894853e-01 2.07668036e-01 -2.95038193e-01
1.35900187e+00 -1.77318752e+00 -2.82121271e-01 4.41783607e-01
3.06491643e-01 2.59982020e-01 -4.44129199e-01 5.21463394e-01
-7.54161552e-02 5.39572775e-01 -2.03521535e-01 -3.75400305e-01
2.90653199e-01 -2.38174766e-01 -5.48348069e-01 1.06847942e-01
2.29008853e-01 9.65438783e-01 -6.74540579e-01 -1.36708334e-01
-4.55363840e-01 3.52616280e-01 -6.18935943e-01 3.61098088e-02
-3.42640221e-01 -4.40586239e-01 -4.95032430e-01 6.61734998e-01
3.68423551e-01 -4.77485120e-01 2.87193596e-01 -1.18055753e-01
3.81834149e-01 9.00357544e-01 -6.79506600e-01 1.16602135e+00
-5.30409873e-01 7.32575059e-01 -3.40427756e-02 -5.74425220e-01
7.86258996e-01 9.87599269e-02 2.92661309e-01 -5.24242699e-01
1.58836693e-01 6.74550161e-02 6.01216927e-02 -1.09976105e-01
1.11458528e+00 -2.60313123e-01 -3.06918353e-01 1.07105362e+00
-9.07893404e-02 -3.60734940e-01 7.10838377e-01 4.54594582e-01
9.21762228e-01 -1.45125166e-01 6.76779747e-01 -2.47592822e-01
4.34788316e-01 9.29757506e-02 2.55563825e-01 8.07330728e-01
-1.55888200e-01 7.23458827e-01 8.82434666e-01 -5.00059903e-01
-1.05981183e+00 -8.79364729e-01 -5.93638271e-02 1.06509590e+00
-1.58405989e-01 -7.40001917e-01 -5.00602424e-01 -9.73793685e-01
-3.01094297e-02 9.81235147e-01 -6.50292397e-01 -1.21146530e-01
-2.99773872e-01 -8.68500113e-01 -1.53867468e-01 4.96278465e-01
-2.13023871e-01 -1.08267117e+00 -1.23307213e-01 1.06580101e-01
2.51734685e-02 -9.50372517e-01 -6.97925508e-01 1.98169246e-01
-6.69604778e-01 -1.09931195e+00 -6.70309663e-01 -6.90231860e-01
8.21087658e-01 5.04114807e-01 1.99552071e+00 -1.09803438e-01
2.33529255e-01 4.28354234e-01 -4.54297066e-01 -6.30819023e-01
-2.66089946e-01 4.12071347e-01 7.72533566e-02 -5.42431101e-02
1.31158531e+00 -6.46745205e-01 -6.42952263e-01 2.95634687e-01
-9.57859159e-01 -4.53408003e-01 9.25684690e-01 5.43495297e-01
5.22456765e-01 -1.21880315e-01 6.82201743e-01 -1.47739863e+00
1.56471062e+00 -3.53917271e-01 -2.31194973e-01 9.03751478e-02
-1.13656461e+00 -3.78040131e-03 6.12650633e-01 -3.10383856e-01
-8.24097574e-01 -3.11656028e-01 -2.12373525e-01 3.76008779e-01
-5.71122095e-02 8.29516232e-01 -1.65115774e-01 5.56800246e-01
6.91327989e-01 9.50765535e-02 -1.88161835e-01 -3.21233839e-01
6.06856585e-01 7.95749068e-01 -1.01854570e-01 -1.96628585e-01
5.99751711e-01 2.07907990e-01 -5.77868819e-01 -8.35658550e-01
-1.25053465e+00 -7.86915720e-01 -3.35675776e-01 9.37280953e-02
2.88214892e-01 -9.56057012e-01 -3.40196550e-01 1.23685561e-01
-1.03112578e+00 2.01034993e-01 -4.66389894e-01 3.09958279e-01
-3.37527692e-01 2.64896333e-01 -6.27632916e-01 -6.72918558e-01
-5.07889271e-01 -1.06337667e+00 1.06341958e+00 1.73988730e-01
-7.54907727e-01 -7.99408317e-01 3.99296820e-01 2.69699633e-01
4.63850200e-01 -4.27452922e-01 1.29356825e+00 -1.16946661e+00
-4.42642927e-01 -6.43438816e-01 -1.51890060e-02 7.48100042e-01
2.69239604e-01 -7.24723190e-03 -9.18873370e-01 2.73520481e-02
1.84062496e-01 -1.33011952e-01 9.62023079e-01 4.62361306e-01
6.27634287e-01 -3.21738183e-01 -1.09487481e-01 5.13773896e-02
1.14649451e+00 1.00528240e-01 5.46072006e-01 2.55373329e-01
5.55193782e-01 7.83709347e-01 5.61115980e-01 3.41990590e-01
4.95696664e-01 2.29481518e-01 -1.32797092e-01 1.17035046e-01
4.09234971e-01 -3.92152332e-02 4.61593598e-01 1.24841392e+00
3.25322688e-01 -3.59446734e-01 -4.39183772e-01 7.29440093e-01
-1.52154982e+00 -7.49690115e-01 2.80174822e-01 2.00270653e+00
8.14718843e-01 4.77841198e-01 2.41914898e-01 -1.31635547e-01
5.81817031e-01 7.06251025e-01 -6.07655466e-01 -7.83995688e-01
-5.45995794e-02 9.97189656e-02 2.43586510e-01 2.97793150e-01
-9.12758470e-01 6.78829610e-01 6.56358290e+00 4.72327799e-01
-7.20666051e-01 -3.92799675e-01 1.15459371e+00 -1.43497661e-01
-1.08083296e+00 -1.07955083e-01 -1.03776932e+00 7.09059089e-02
9.01804447e-01 -5.12898147e-01 2.25682437e-01 1.24654365e+00
-1.96873788e-02 -1.34986117e-01 -1.34518278e+00 6.79284990e-01
5.31591058e-01 -8.94922078e-01 4.31727082e-01 7.87724629e-02
9.14886475e-01 1.81432888e-02 3.13754708e-01 5.17129004e-01
4.99028444e-01 -8.74908090e-01 1.75264299e-01 4.36789878e-02
4.87056434e-01 -9.33148384e-01 7.72518694e-01 6.44355491e-02
-9.22606885e-01 1.68737933e-01 -6.35729134e-01 -7.22703040e-02
3.99816692e-01 9.70857501e-01 -9.13009405e-01 3.85804214e-02
3.26572716e-01 8.71322811e-01 -7.19804347e-01 7.35906720e-01
-4.20233518e-01 5.54929972e-01 1.35890450e-02 -5.21657646e-01
2.01647401e-01 -5.38393319e-01 4.32013184e-01 1.07092035e+00
1.39553830e-01 -3.30214530e-01 -1.87181950e-01 5.85762501e-01
-4.07877117e-01 4.79311168e-01 -8.82878900e-01 -6.44840539e-01
-1.05397500e-01 1.74933791e+00 -7.47786462e-01 -4.05960649e-01
-6.89354718e-01 5.93174279e-01 1.91885322e-01 5.43938100e-01
-1.22824684e-01 -4.89706337e-01 1.03148043e+00 -4.24220785e-02
5.10108531e-01 -2.32988899e-03 -5.63625455e-01 -1.32001817e+00
3.77072334e-01 -1.27785397e+00 5.37419356e-02 -6.81111753e-01
-1.57257235e+00 1.15105975e+00 -3.48404318e-01 -1.24737322e+00
-6.44278288e-01 -8.18099916e-01 -6.50796831e-01 8.36057425e-01
-1.55587554e+00 -6.22435391e-01 2.37070277e-01 -4.00951803e-02
8.33695769e-01 -3.17571282e-01 6.48992836e-01 1.17601559e-01
-1.00065075e-01 4.17727202e-01 -1.15204908e-01 6.37694597e-02
7.62582004e-01 -1.58086479e+00 1.06843770e+00 6.06741726e-01
5.69355190e-01 1.23542571e+00 8.53874087e-01 -5.78372121e-01
-1.20169866e+00 -6.75900280e-01 1.36277449e+00 -8.43333304e-01
6.38741672e-01 -2.45423675e-01 -5.75575352e-01 6.01654708e-01
5.60812116e-01 -6.51596963e-01 1.28786111e+00 7.32412398e-01
-6.19326949e-01 -1.86799213e-01 -6.79489732e-01 9.65216398e-01
5.38171828e-01 -8.90698910e-01 -6.39640331e-01 2.21799612e-01
7.09559441e-01 3.76593024e-01 -5.02262414e-01 3.01699519e-01
7.19451606e-01 -9.78943110e-01 7.00098932e-01 -8.10884535e-01
8.53115976e-01 -8.70624483e-02 -1.09318949e-01 -1.79763949e+00
-1.00993618e-01 -5.46465933e-01 9.11042094e-02 1.27614522e+00
1.16345763e+00 -4.31617528e-01 9.62632895e-01 1.10116804e+00
1.90780103e-01 -1.17688620e+00 -7.79882222e-02 -4.54627946e-02
3.66577804e-02 -5.08342743e-01 6.39105797e-01 3.76208574e-01
2.66377807e-01 1.17725790e+00 -1.48150459e-01 -3.81010711e-01
2.63070017e-01 3.69898796e-01 7.38265991e-01 -1.19003451e+00
-5.51082015e-01 -7.42833972e-01 4.60379645e-02 -1.26025558e+00
1.23773985e-01 -5.89006662e-01 7.99647123e-02 -1.66100681e+00
4.63725477e-01 -2.80432831e-02 -3.93238217e-01 -6.42151386e-02
-5.13314128e-01 1.46920875e-01 -1.24866843e-01 1.44050941e-01
-8.08486938e-01 5.48667133e-01 1.13210762e+00 -2.87337184e-01
-2.19395995e-01 5.95385969e-01 -1.82120574e+00 9.26310480e-01
8.89958143e-01 -2.48804152e-01 -7.94011772e-01 -2.44320050e-01
1.14416981e+00 -3.14064413e-01 -5.40390968e-01 -2.67467886e-01
-2.09564388e-01 9.10652652e-02 2.59109318e-01 -6.46035850e-01
3.07723016e-01 -6.54944241e-01 -5.60377777e-01 -4.11940366e-01
-8.49567950e-01 7.65893042e-01 -2.85237312e-01 7.05766678e-01
-4.94194865e-01 -3.72582316e-01 2.89070934e-01 -4.25806403e-01
-2.77067989e-01 2.70566136e-01 -6.02306187e-01 2.35756636e-01
2.40259498e-01 7.37324581e-02 -3.29264194e-01 -9.17337954e-01
-3.17692637e-01 -7.27777481e-02 5.42609036e-01 5.62409163e-01
5.44271111e-01 -1.15709972e+00 -5.23322165e-01 1.18629001e-01
4.64859366e-01 -3.98319274e-01 -2.45074332e-01 5.11341393e-01
7.36971423e-02 5.12986004e-01 2.56244272e-01 -1.70907199e-01
-1.17545092e+00 4.42006350e-01 -1.15567438e-01 -6.23551846e-01
-5.26754074e-02 8.01323354e-01 3.69788319e-01 -6.27183378e-01
-4.82411869e-02 -2.83094883e-01 -7.14606464e-01 5.32671988e-01
7.51303673e-01 -1.36598349e-01 1.99403733e-01 -5.77326655e-01
-1.03053056e-01 3.95974219e-01 -7.61191308e-01 -5.05277514e-01
1.34800875e+00 -2.50955939e-01 -1.00835279e-01 5.32761395e-01
1.36063921e+00 3.63380492e-01 -7.87982583e-01 -2.61230290e-01
2.04124987e-01 -2.78503686e-01 -2.90449888e-01 -8.12178791e-01
-1.00937212e+00 6.17361844e-01 -1.84752107e-01 7.04299033e-01
9.48695421e-01 1.17470331e-01 6.50675118e-01 8.81951213e-01
1.74372464e-01 -1.26692331e+00 -1.51246116e-01 6.41669273e-01
5.95577836e-01 -1.30146003e+00 2.44292766e-01 -1.83370665e-01
-1.21905267e+00 6.89960241e-01 4.67188418e-01 -1.36316881e-01
8.83975565e-01 -2.08985414e-02 6.11841381e-01 -4.87227321e-01
-1.08900630e+00 -2.14032263e-01 5.54605663e-01 4.70900595e-01
8.55832279e-01 4.73153181e-02 -6.30336285e-01 8.55103254e-01
-3.86806458e-01 -5.93160987e-01 8.56509924e-01 6.96172237e-01
-4.45821077e-01 -1.25694263e+00 3.42280000e-01 1.07933199e+00
-8.61862779e-01 -6.89452648e-01 -8.58515501e-01 4.37813193e-01
-6.79513872e-01 1.09245563e+00 -2.54183203e-01 -3.18816751e-01
3.21913719e-01 3.54175828e-02 -1.35005325e-01 -9.41619694e-01
-8.46669853e-01 2.74884760e-01 6.48543954e-01 -3.68447810e-01
-4.33287144e-01 -7.54054368e-01 -4.73016381e-01 2.14834422e-01
-3.46105903e-01 6.36155725e-01 8.77783716e-01 8.75342011e-01
4.64997411e-01 1.51289269e-01 8.25072467e-01 -4.66395497e-01
-6.78520322e-01 -9.53007221e-01 -9.24744368e-01 6.21493816e-01
2.25429296e-01 -6.81030676e-02 -6.06138647e-01 -5.00951558e-02] | [11.395294189453125, 6.710931777954102] |
abdac949-e1f8-4156-bb60-b9de92bbaa0d | kafsp-knowledge-aware-fuzzy-semantic-parsing | null | null | https://aclanthology.org/2022.acl-long.35 | https://aclanthology.org/2022.acl-long.35.pdf | KaFSP: Knowledge-Aware Fuzzy Semantic Parsing for Conversational Question Answering over a Large-Scale Knowledge Base | In this paper, we study two issues of semantic parsing approaches to conversational question answering over a large-scale knowledge base: (1) The actions defined in grammar are not sufficient to handle uncertain reasoning common in real-world scenarios. (2) Knowledge base information is not well exploited and incorporated into semantic parsing. To mitigate the two issues, we propose a knowledge-aware fuzzy semantic parsing framework (KaFSP). It defines fuzzy comparison operations in the grammar system for uncertain reasoning based on the fuzzy set theory. In order to enhance the interaction between semantic parsing and knowledge base, we incorporate entity triples from the knowledge base into a knowledge-aware entity disambiguation module. Additionally, we propose a multi-label classification framework to not only capture correlations between entity types and relations but also detect knowledge base information relevant to the current utterance. Both enhancements are based on pre-trained language models. Experiments on a large-scale conversational question answering benchmark demonstrate that the proposed KaFSP achieves significant improvements over previous state-of-the-art models, setting new SOTA results on 8 out of 10 question types, gaining improvements of over 10% F1 or accuracy on 3 question types, and improving overall F1 from 83.01% to 85.33%. The source code of KaFSP is available at https://github.com/tjunlp-lab/KaFSP. | ['Deyi Xiong', 'Junzhuo Li'] | null | null | null | null | acl-2022-5 | ['entity-disambiguation'] | ['natural-language-processing'] | [-7.97015801e-02 5.81048191e-01 -9.27769020e-02 -7.43613303e-01
-8.77413690e-01 -5.46530962e-01 2.85456985e-01 3.23296040e-01
-2.37756744e-01 7.62167752e-01 1.76723942e-01 -2.76895165e-01
-1.87914386e-01 -1.18488359e+00 -7.37778783e-01 -1.96196814e-03
3.71610820e-01 7.34849632e-01 7.87530482e-01 -7.79461265e-01
2.70704269e-01 -2.50319034e-01 -1.73722124e+00 7.99520135e-01
1.42024589e+00 1.28328359e+00 1.62531629e-01 5.11479437e-01
-9.03087497e-01 9.56038475e-01 -6.12428427e-01 -8.65724921e-01
-2.00912923e-01 -9.07319114e-02 -1.43209755e+00 -1.94575638e-01
1.74651116e-01 -2.99969111e-02 -9.47141962e-04 1.12541664e+00
1.33067086e-01 2.80950189e-01 2.70994544e-01 -1.27190137e+00
-7.48586953e-01 9.31410134e-01 1.22238986e-01 7.00524449e-02
7.59676278e-01 -4.12665039e-01 1.21560740e+00 -6.82018816e-01
4.44743723e-01 1.59230494e+00 4.66251612e-01 7.13920355e-01
-4.30023611e-01 -4.68450159e-01 4.02979076e-01 7.16796219e-01
-1.22149909e+00 -1.81313410e-01 7.30043292e-01 -9.01461579e-03
1.17667997e+00 2.38466591e-01 6.22930191e-02 6.71540678e-01
-1.62936494e-01 9.09527957e-01 1.03855753e+00 -6.74320221e-01
3.28706801e-01 2.36980841e-01 6.98210239e-01 8.05204630e-01
-8.72694403e-02 -4.36324328e-01 -5.48706114e-01 -2.62585610e-01
2.21760333e-01 -3.51967961e-01 -1.77578911e-01 1.63050517e-01
-6.19865775e-01 9.57329571e-01 2.68288463e-01 4.14322883e-01
-2.45564923e-01 -1.71074003e-01 3.22927058e-01 3.43776941e-01
2.99034983e-01 5.29014111e-01 -7.45153368e-01 -2.19879076e-01
-2.90004015e-01 3.28311622e-01 1.15806079e+00 1.07749581e+00
8.36623192e-01 -5.51066697e-01 -4.55804020e-02 1.15574050e+00
3.69360059e-01 4.75567251e-01 3.98720682e-01 -1.31279540e+00
7.63627708e-01 1.08933187e+00 1.85311034e-01 -9.10458744e-01
-5.04372001e-01 -2.36473121e-02 -1.24213062e-01 -6.29548907e-01
4.15772289e-01 3.32269706e-02 -6.99435592e-01 1.75355279e+00
5.96326590e-01 2.30480194e-01 4.85841334e-01 6.83115065e-01
1.26713252e+00 5.92955470e-01 2.74842441e-01 5.69356466e-03
1.88761032e+00 -1.10488939e+00 -9.40240622e-01 -3.37930381e-01
6.74383283e-01 -5.83552003e-01 1.24305546e+00 1.43315822e-01
-9.81432498e-01 -3.61964047e-01 -8.89480233e-01 -2.24514201e-01
-7.01031923e-01 -2.18198419e-01 6.81511045e-01 7.57993817e-01
-8.80484641e-01 1.29249126e-01 -4.64498609e-01 -2.08873644e-01
9.65019017e-02 3.23464274e-01 -2.08962709e-03 -3.48924994e-01
-2.04403830e+00 8.28947723e-01 7.14538574e-01 -4.20704894e-02
-2.72015423e-01 -4.86280948e-01 -1.09535921e+00 2.35307172e-01
1.01579678e+00 -7.87938774e-01 1.55785561e+00 -5.06497502e-01
-1.61730778e+00 5.62645733e-01 -4.36851770e-01 -3.92027795e-01
-2.30726190e-02 -2.51807868e-01 -7.53194630e-01 4.65640038e-01
2.62936532e-01 5.93846023e-01 2.87474155e-01 -1.32091534e+00
-1.00163496e+00 -3.64224285e-01 8.37127864e-01 5.28348684e-01
-2.91750193e-01 -3.66565548e-02 -7.50750363e-01 -2.58036166e-01
1.79312259e-01 -8.20791543e-01 7.81319216e-02 -6.77303553e-01
-1.53830811e-01 -7.26187766e-01 6.28286541e-01 -6.83865964e-01
1.28928483e+00 -1.60948575e+00 -1.98197916e-01 1.07240342e-01
-2.09583998e-01 3.74362588e-01 2.29199640e-02 2.08726883e-01
4.71982777e-01 2.08637178e-01 -2.85675317e-01 -8.54609534e-03
1.80819124e-01 5.36702394e-01 -3.66448939e-01 -6.80206120e-01
2.44656757e-01 9.43390131e-01 -1.05158114e+00 -6.43878341e-01
7.91291669e-02 5.75354733e-02 -5.12661815e-01 1.84176028e-01
-6.94411874e-01 5.93248429e-03 -8.96343768e-01 8.94727290e-01
6.81455255e-01 -2.97659904e-01 4.43845749e-01 -8.93806145e-02
4.65729564e-01 6.93771601e-01 -1.23513365e+00 1.68210196e+00
-7.38318622e-01 -2.23565742e-01 6.91173673e-02 -9.11647975e-01
8.24711859e-01 3.63847703e-01 1.02259286e-01 -7.42213130e-01
1.88476786e-01 3.15398574e-01 -3.19991112e-01 -7.49674022e-01
6.95465267e-01 -2.20745236e-01 -4.56785947e-01 2.26042241e-01
1.92459926e-01 -3.44222546e-01 3.55114162e-01 4.68439341e-01
1.00266635e+00 -1.63874462e-01 1.60648853e-01 -4.23890293e-01
1.05405319e+00 2.94107229e-01 7.70969212e-01 6.54275835e-01
-2.98120171e-01 1.59087002e-01 4.52968299e-01 -1.17828794e-01
-1.61241904e-01 -9.50126171e-01 -6.40869290e-02 1.37885296e+00
6.13157868e-01 -5.14916897e-01 -1.00574601e+00 -1.14207387e+00
-1.29593477e-01 1.07863247e+00 -3.72564703e-01 -1.06346197e-01
-6.10855162e-01 -6.01233065e-01 7.07479358e-01 6.15367532e-01
8.74486148e-01 -1.00792503e+00 -1.29586622e-01 2.60544181e-01
-9.65686500e-01 -1.71564734e+00 -3.12301874e-01 -1.83768913e-01
-6.39838338e-01 -1.33758223e+00 1.02346659e-01 -9.43458498e-01
3.86325002e-01 -2.86822557e-03 1.45529413e+00 2.94101536e-01
3.84253234e-01 5.93748868e-01 -8.72558832e-01 -4.54205662e-01
-4.90190566e-01 1.27308458e-01 -2.44833440e-01 -2.81956077e-01
8.40668917e-01 -1.36725500e-01 -3.21693718e-01 5.87069452e-01
-8.74383271e-01 -2.32866883e-01 1.31699771e-01 7.89585590e-01
4.45309550e-01 3.20075393e-01 1.16931868e+00 -1.15704322e+00
7.57671654e-01 -4.97988611e-01 -3.22003782e-01 8.01228583e-01
-6.27675653e-01 1.86908022e-01 6.41951978e-01 2.07672585e-02
-1.69869077e+00 -3.84124190e-01 -4.43467468e-01 1.65154248e-01
-1.52306706e-01 5.89359105e-01 -5.25550783e-01 1.92363366e-01
4.70729947e-01 -1.59585804e-01 -2.83618152e-01 -1.61765918e-01
5.11986732e-01 9.73759472e-01 5.43729901e-01 -1.13989067e+00
2.64732391e-01 3.11476707e-01 -3.89016896e-01 -3.59567583e-01
-1.29153287e+00 -6.15622044e-01 -2.85981864e-01 -1.02283072e-03
8.16200614e-01 -8.55342984e-01 -8.67102444e-01 2.93383777e-01
-9.80324566e-01 -9.57870260e-02 -1.02515727e-01 3.89224559e-01
-3.95911813e-01 5.08167624e-01 -7.51528382e-01 -7.92270124e-01
-4.67886120e-01 -1.01639760e+00 8.87338400e-01 5.57125449e-01
-4.19572629e-02 -1.09750938e+00 -3.26014131e-01 1.34801972e+00
3.30400169e-01 -2.19334081e-01 1.21129096e+00 -1.01581168e+00
-4.24649060e-01 7.06627443e-02 -4.69132513e-02 4.58731443e-01
5.93909696e-02 -3.89473021e-01 -9.82874811e-01 3.46719623e-01
2.16357306e-01 -5.74540973e-01 6.93220258e-01 -1.62097644e-02
9.84055936e-01 -1.54691041e-01 -1.22268453e-01 -1.75713152e-01
1.05582535e+00 4.10380870e-01 5.43814063e-01 3.97254467e-01
3.95043045e-01 9.86337841e-01 1.04693651e+00 1.35792717e-01
1.39142895e+00 6.04366720e-01 2.84082323e-01 6.79318964e-01
-5.00932224e-02 -3.54954660e-01 1.83100566e-01 1.04125047e+00
2.13511616e-01 -2.75983840e-01 -9.36397851e-01 7.45866477e-01
-2.05508447e+00 -5.67246139e-01 -2.59451829e-02 1.74182773e+00
1.04012465e+00 1.54566675e-01 -3.30312252e-01 -3.61665897e-02
9.43122804e-01 -1.20785452e-01 -4.56662625e-01 -5.55042744e-01
-1.74120232e-01 2.50468820e-01 6.85713813e-02 8.94912720e-01
-1.00136101e+00 1.29712510e+00 5.05193043e+00 8.98511350e-01
-5.45188487e-01 2.34688014e-01 3.94103944e-01 3.89022440e-01
-6.18417382e-01 1.98563159e-01 -1.02092290e+00 3.26987535e-01
9.44896460e-01 -1.91737100e-01 3.77252728e-01 8.14626217e-01
-2.42880553e-01 -1.70410976e-01 -6.38795793e-01 6.60845697e-01
8.28052610e-02 -1.28302252e+00 1.16356805e-01 -4.50563967e-01
5.29882908e-01 -2.54858673e-01 -3.40478629e-01 8.90401483e-01
4.02020901e-01 -7.56846488e-01 3.71837735e-01 3.34133178e-01
1.36634141e-01 -8.44371378e-01 1.10846448e+00 4.71695065e-01
-1.21247470e+00 -1.15055628e-01 -3.04544955e-01 7.21732304e-02
2.89608091e-01 5.30972421e-01 -9.02724385e-01 1.11996710e+00
9.28956330e-01 1.73670232e-01 -3.46249670e-01 2.66125888e-01
-6.38089597e-01 6.26373291e-01 -4.60603565e-01 -9.84573662e-02
1.91106230e-01 1.56524237e-02 2.42804229e-01 8.99337649e-01
1.61928266e-01 6.37703538e-01 1.06558166e-01 6.26621485e-01
-3.42937410e-01 2.63364404e-01 5.91954067e-02 1.65265918e-01
8.85037005e-01 1.12800121e+00 -5.61459541e-01 -5.75256109e-01
-5.72759151e-01 6.07067883e-01 3.77548695e-01 8.68188515e-02
-7.96852589e-01 -5.31922400e-01 3.86629760e-01 -2.23678648e-01
3.36378843e-01 1.55726612e-01 -1.96794748e-01 -1.32163453e+00
3.90379637e-01 -8.72201025e-01 9.95042920e-01 -7.40636945e-01
-1.37671328e+00 6.80257022e-01 -4.83329466e-04 -5.66585064e-01
-3.89770448e-01 -5.58350027e-01 -4.63857651e-01 6.15255058e-01
-1.94302988e+00 -1.15955973e+00 -4.03520495e-01 5.13766229e-01
4.53994006e-01 -9.37022362e-03 1.06614435e+00 3.45612139e-01
-2.99247593e-01 6.18221283e-01 -3.98428768e-01 -4.89509404e-02
6.52854323e-01 -1.38591349e+00 2.47991920e-01 6.50868475e-01
-7.43686920e-03 6.66372001e-01 5.08114278e-01 -6.31837130e-01
-1.20485401e+00 -9.63855326e-01 1.21004570e+00 -6.30815268e-01
5.35355568e-01 -5.46022020e-02 -1.30537415e+00 5.95678270e-01
9.99344885e-02 -2.53255665e-01 7.93427050e-01 3.87869060e-01
-5.43041885e-01 5.10866046e-02 -1.55407643e+00 2.73200095e-01
9.36734796e-01 -5.93469262e-01 -1.19621837e+00 2.45471656e-01
1.02570391e+00 -6.80882871e-01 -1.19017684e+00 7.17334867e-01
1.90005586e-01 -8.26797783e-01 8.37129831e-01 -4.06436086e-01
1.87467933e-01 -4.59932178e-01 -5.07953763e-01 -1.18136406e+00
1.80794984e-01 -2.40363374e-01 -2.67225772e-01 1.36202681e+00
7.32367218e-01 -8.32957566e-01 7.49792993e-01 1.02158153e+00
-3.97458196e-01 -7.98657537e-01 -9.88746464e-01 -6.67593598e-01
1.01090841e-01 -5.85458457e-01 9.10277069e-01 1.01090419e+00
4.65966761e-01 4.61570144e-01 2.01523975e-01 5.58342159e-01
2.67218918e-01 3.85642201e-01 2.50239015e-01 -9.44226861e-01
-2.03520834e-01 -1.06436573e-01 -6.27898872e-02 -1.04234600e+00
5.74179232e-01 -8.23591292e-01 1.13706682e-02 -1.83367968e+00
-1.28336236e-01 -6.02397263e-01 -3.54264468e-01 5.55754840e-01
-5.93473554e-01 -2.31747255e-01 1.35904506e-01 -1.24866270e-01
-9.34457839e-01 6.37997210e-01 1.09142816e+00 -8.11344236e-02
-1.05047807e-01 5.46881987e-04 -1.12526488e+00 7.36028075e-01
9.36242580e-01 -2.77708888e-01 -6.77979887e-01 -5.16718447e-01
1.51962101e-01 2.16229752e-01 2.15437599e-02 -6.98667288e-01
3.70820731e-01 -2.23011687e-01 -4.55684155e-01 -3.59420031e-01
4.61743653e-01 -6.82968557e-01 -4.23335642e-01 8.24828818e-02
-3.06748807e-01 -2.15573579e-01 2.19486475e-01 5.01747847e-01
-6.52456939e-01 -5.95364749e-01 4.79130238e-01 -2.22723633e-01
-1.12204957e+00 -2.11726308e-01 -2.60917898e-02 5.57395041e-01
8.74099612e-01 3.16616684e-01 -8.00840497e-01 -4.00853992e-01
-6.06494904e-01 8.64972293e-01 1.39799505e-01 8.45683694e-01
3.97967160e-01 -9.95650053e-01 -3.13337892e-01 -2.44340137e-01
3.89497250e-01 2.53269106e-01 6.50377631e-01 2.95684695e-01
-4.24306899e-01 6.80043280e-01 7.14755729e-02 -2.39799365e-01
-1.14427769e+00 2.94014245e-01 3.60161632e-01 -5.19634843e-01
-7.44853094e-02 9.24539089e-01 -1.16664849e-01 -1.10276484e+00
1.17034167e-01 -4.98010069e-01 -5.75878501e-01 -6.45767674e-02
5.15214741e-01 3.26658815e-01 3.93458337e-01 -4.89378065e-01
-6.21484160e-01 4.85771388e-01 -1.56719685e-01 3.95534076e-02
8.71982157e-01 -4.59923804e-01 -1.75659150e-01 1.71776250e-01
7.55243123e-01 3.79912159e-03 -6.38312042e-01 -6.15133643e-01
3.36204797e-01 -1.17762640e-01 -1.01660021e-01 -1.20328534e+00
-7.43311763e-01 6.58174396e-01 2.06521943e-01 3.23768824e-01
9.71808493e-01 2.68007874e-01 1.27254999e+00 7.88096189e-01
5.47954559e-01 -1.46817148e+00 -3.72107700e-02 9.25676227e-01
6.69326663e-01 -1.34817755e+00 -3.48733395e-01 -1.05433512e+00
-1.01623583e+00 9.22915816e-01 8.95356297e-01 3.44554633e-01
5.93625367e-01 1.71130374e-02 4.08771008e-01 -3.36010486e-01
-8.40935886e-01 -3.73082519e-01 1.68722987e-01 2.58552879e-01
2.25380212e-01 2.52544373e-01 -4.15507346e-01 1.31333554e+00
-3.23291421e-01 2.38160193e-02 4.63318199e-01 1.12059891e+00
-7.79383123e-01 -1.35036957e+00 -1.71265990e-01 2.95189857e-01
-5.26212811e-01 -1.42519385e-01 -3.60056579e-01 4.38327253e-01
5.52018322e-02 1.69728315e+00 -1.61760792e-01 -4.10148829e-01
6.18085861e-01 5.23966849e-01 2.51873344e-01 -7.31563032e-01
-6.53933942e-01 -5.00756085e-01 6.91030741e-01 -4.38835651e-01
-5.40448248e-01 -3.18209589e-01 -1.89108169e+00 5.05553419e-03
-6.01512074e-01 7.49273539e-01 6.20522678e-01 1.32954741e+00
6.43206000e-01 4.53232229e-01 4.32716846e-01 2.72704333e-01
-5.24834931e-01 -8.95354450e-01 -3.87300998e-01 4.25003618e-01
-3.60980153e-01 -8.11506450e-01 -3.56120348e-01 -1.25638127e-01] | [10.655118942260742, 7.980432987213135] |
c1f02615-567d-432b-8e8a-996978376571 | energy-forecasting-in-smart-grid-systems-a | 2011.12598 | null | https://arxiv.org/abs/2011.12598v3 | https://arxiv.org/pdf/2011.12598v3.pdf | Energy Forecasting in Smart Grid Systems: A Review of the State-of-the-art Techniques | Energy forecasting has a vital role to play in smart grid (SG) systems involving various applications such as demand-side management, load shedding, and optimum dispatch. Managing efficient forecasting while ensuring the least possible prediction error is one of the main challenges posed in the grid today, considering the uncertainty and granularity in SG data. This paper presents a comprehensive and application-oriented review of state-of-the-art forecasting methods for SG systems along with recent developments in probabilistic deep learning (PDL) considering different models and architectures. Traditional point forecasting methods including statistical, machine learning (ML), and deep learning (DL) are extensively investigated in terms of their applicability to energy forecasting. In addition, the significance of hybrid and data pre-processing techniques to support forecasting performance is also studied. A comparative case study using the Victorian electricity consumption and American electric power (AEP) datasets is conducted to analyze the performance of point and probabilistic forecasting methods. The analysis demonstrates higher accuracy of the long-short term memory (LSTM) models with appropriate hyper-parameter tuning among point forecasting methods especially when sample sizes are larger and involve nonlinear patterns with long sequences. Furthermore, Bayesian bidirectional LSTM (BLSTM) as a probabilistic method exhibit the highest accuracy in terms of least pinball score and root mean square error (RMSE). | ['ZhaoYang Dong', 'Md. Enamul Haque', 'Md. Apel Mahmud', 'Shama Naz Islam', 'Devinder Kaur'] | 2020-11-25 | null | null | null | null | ['probabilistic-deep-learning'] | ['computer-vision'] | [-5.50781310e-01 -5.83379984e-01 -5.80580719e-02 -4.19794083e-01
-7.88219452e-01 -3.75334680e-01 8.32565367e-01 2.75757074e-01
1.27894431e-01 1.07897997e+00 1.66730881e-01 -6.66328430e-01
-5.57959437e-01 -1.06070149e+00 -2.18755335e-01 -1.43867540e+00
-5.48940957e-01 6.30119920e-01 -4.25962687e-01 -2.08191201e-01
3.93800497e-01 6.91641152e-01 -1.42095292e+00 2.82316059e-01
9.65415955e-01 1.59462094e+00 2.63346225e-01 4.55091536e-01
-9.20839161e-02 7.32511938e-01 -9.51310754e-01 1.46868527e-01
-1.51242673e-01 8.01919326e-02 -4.48213756e-01 -4.84709382e-01
-4.68742698e-01 -3.40546042e-01 1.20052211e-01 8.80557358e-01
9.68505859e-01 5.37326217e-01 7.29775131e-01 -1.53513920e+00
-6.03126943e-01 9.74988937e-01 -5.24942935e-01 6.85449779e-01
2.91741379e-02 9.79238674e-02 5.80653667e-01 -7.49574363e-01
-4.41357255e-01 8.69590759e-01 8.58514428e-01 -8.66890699e-02
-9.21415150e-01 -6.79358423e-01 -1.15044460e-01 1.09671664e+00
-1.18941236e+00 4.82947845e-03 9.25833344e-01 -5.11101246e-01
1.83520555e+00 -6.61879554e-02 6.19323254e-01 9.28941727e-01
8.69619668e-01 5.24155140e-01 1.15288973e+00 -4.47449297e-01
6.59700513e-01 -7.51364455e-02 1.22597881e-01 -3.06132764e-01
7.74230063e-03 6.30866528e-01 -5.00479579e-01 -1.57699808e-01
-6.80937394e-02 -1.80899888e-01 -2.29121819e-02 3.28279048e-01
-8.39697123e-01 1.05126369e+00 7.92133287e-02 8.00232649e-01
-8.99425209e-01 1.70798246e-02 7.88182557e-01 9.01229456e-02
9.32491481e-01 -1.60397574e-01 -9.97584403e-01 -4.47940111e-01
-1.42381489e+00 -1.03347702e-03 9.79213953e-01 5.23137569e-01
-2.46950239e-03 1.21634865e+00 -2.11295918e-01 5.21725237e-01
2.40510941e-01 7.22468913e-01 8.53536963e-01 -3.54196429e-01
2.67970026e-01 8.26445222e-02 2.15376750e-01 -1.04572427e+00
-8.76001775e-01 -7.79327154e-01 -1.49480009e+00 2.93425441e-01
-1.72491536e-01 -2.83025324e-01 -3.75161499e-01 1.20130634e+00
-9.54362974e-02 1.84332460e-01 1.84333235e-01 2.07899705e-01
6.35354340e-01 1.36999917e+00 2.29670510e-01 -4.34606761e-01
1.20591354e+00 -4.71968323e-01 -8.52072001e-01 1.73660964e-01
7.13205099e-01 -5.97492814e-01 2.58080631e-01 6.53912187e-01
-7.28968620e-01 -5.60448229e-01 -1.02972245e+00 4.09918576e-01
-8.41964066e-01 2.00527728e-01 2.78108567e-02 6.10453963e-01
-1.01643348e+00 8.26617301e-01 -1.05110121e+00 -1.38784006e-01
9.62088779e-02 3.16101253e-01 3.30868065e-01 5.67397118e-01
-1.68240321e+00 1.66057575e+00 9.65019584e-01 4.41256732e-01
-5.41635275e-01 -7.99462259e-01 -7.51270115e-01 4.07829732e-01
-3.71125847e-01 -3.24029654e-01 1.27919626e+00 -4.16252851e-01
-1.67213237e+00 -2.44244963e-01 -1.42054725e-02 -1.16664135e+00
1.71176806e-01 -3.11971586e-02 -9.00137067e-01 -3.66004467e-01
-4.43961829e-01 9.46020111e-02 5.67774594e-01 -8.72035444e-01
-1.07710576e+00 -3.23677272e-01 -6.19176984e-01 -6.30021691e-02
-7.81477839e-02 -8.34023058e-02 1.23603904e+00 -7.61338353e-01
-1.98665380e-01 -4.26319838e-01 1.12164319e-01 -1.11889768e+00
-2.25401998e-01 -1.06816399e+00 1.13733160e+00 -1.33095980e+00
1.36855710e+00 -1.59551656e+00 -2.91240573e-01 3.19716036e-01
-5.40749371e-01 2.45079905e-01 4.70634520e-01 9.86033082e-01
-2.64693558e-01 -1.99183479e-01 2.44423170e-02 -2.14219421e-01
4.67860341e-01 4.46556419e-01 -7.27373898e-01 4.84230131e-01
-2.13865742e-01 9.26345885e-01 -7.49868035e-01 -3.65880132e-02
9.31357682e-01 7.60070920e-01 5.43115139e-01 2.01465040e-02
-3.03836286e-01 2.40503043e-01 -3.87003683e-02 4.44921345e-01
7.50469744e-01 -5.47137000e-02 -1.98984495e-03 -4.59120780e-01
-3.35620105e-01 1.41243502e-01 -1.12485981e+00 9.41279113e-01
-7.50460207e-01 6.90027297e-01 -4.90123808e-01 -1.42743838e+00
1.08894706e+00 6.48031414e-01 6.29380941e-01 -7.22837329e-01
3.53285819e-02 3.51578653e-01 5.40216360e-03 -1.93353072e-01
5.46846509e-01 -2.25545451e-01 3.53792667e-01 5.79430223e-01
1.05288349e-01 3.00442763e-02 2.25189209e-01 -4.34064180e-01
2.33585715e-01 5.42346537e-02 3.67879182e-01 -6.38338029e-01
5.03915548e-01 -1.80678144e-01 5.22797346e-01 4.59380120e-01
3.44265252e-02 -1.03265204e-01 2.26859733e-01 -7.73109436e-01
-1.04447782e+00 -5.69855630e-01 -4.67187166e-01 9.42518830e-01
-6.69246376e-01 1.91525564e-01 -2.39443973e-01 -2.11539790e-01
1.58919916e-01 2.10324121e+00 -4.33997631e-01 9.12632793e-03
-4.17786896e-01 -1.51726449e+00 2.19621599e-01 8.51644874e-01
3.50415975e-01 -1.24036372e+00 -7.50925124e-01 6.12972617e-01
3.39794010e-02 -7.24822462e-01 1.81135803e-01 4.96518821e-01
-8.77873063e-01 -6.32661402e-01 -4.83624727e-01 -4.47133690e-01
2.90933773e-02 -3.10694635e-01 1.51266766e+00 -6.42957687e-01
5.08386016e-01 -4.15152870e-02 -2.47501761e-01 -5.43015242e-01
-4.58076805e-01 2.00759739e-01 2.81775355e-01 -5.10120273e-01
8.59996974e-01 -7.06801176e-01 -5.29864848e-01 8.89408514e-02
-6.28506780e-01 -1.91657335e-01 2.41264001e-01 1.05552173e+00
3.81215602e-01 9.72795248e-01 1.24879491e+00 -8.18753093e-02
8.20236146e-01 -9.38420236e-01 -1.07801390e+00 3.05784523e-01
-1.27161765e+00 -2.45631531e-01 9.49195743e-01 -1.00690052e-01
-9.85726297e-01 -4.76593822e-01 -2.99185425e-01 -2.58064747e-01
-2.18890354e-01 8.09822798e-01 1.73242256e-01 1.14273690e-01
1.03252977e-01 8.27403724e-01 -3.80416125e-01 -6.61013067e-01
1.27390206e-01 6.92072093e-01 2.20369995e-01 -3.18424642e-01
3.24123740e-01 1.32034048e-01 1.90129042e-01 -6.62863195e-01
-5.26800215e-01 5.94962873e-02 -7.68582642e-01 -3.38113487e-01
4.99550074e-01 -6.52986586e-01 -1.09954417e+00 9.30428267e-01
-1.08299696e+00 -2.75513232e-01 -3.53164434e-01 5.84649742e-01
-4.03206468e-01 6.09841384e-02 -6.67573869e-01 -1.14414752e+00
-9.86130536e-01 -1.07990444e+00 6.42614841e-01 1.45576477e-01
-3.98449646e-03 -1.68136883e+00 -6.99372515e-02 -2.30884865e-01
1.04586184e+00 5.01429677e-01 1.25410116e+00 -1.18254423e+00
1.17211975e-01 -2.13823646e-01 4.50640097e-02 8.07006359e-01
6.07239567e-02 3.92249934e-02 -8.31950963e-01 -6.02567732e-01
2.84382105e-01 2.32788265e-01 2.49881878e-01 9.86177146e-01
1.15008926e+00 -6.15111947e-01 -5.48127592e-02 2.75451124e-01
1.62533700e+00 8.06842029e-01 1.77501693e-01 6.75795615e-01
4.33395743e-01 3.25194597e-01 1.56025842e-01 8.92001987e-01
6.45548701e-01 2.32808694e-01 2.69676358e-01 4.11719263e-01
3.67802769e-01 5.48664145e-02 4.33400959e-01 1.29925752e+00
2.63201922e-01 -5.09799600e-01 -1.03543353e+00 5.52203357e-01
-1.73753023e+00 -1.19358969e+00 -1.96358964e-01 2.08618903e+00
4.81465369e-01 1.56342953e-01 -7.78733343e-02 6.84059978e-01
5.83084643e-01 2.96817988e-01 -6.32249475e-01 -8.48055840e-01
-4.58569229e-01 1.66225195e-01 6.46686494e-01 3.38718444e-01
-9.10188496e-01 1.18120521e-01 5.87662983e+00 1.46441031e+00
-1.20800340e+00 2.71690547e-01 1.17355907e+00 6.42827228e-02
-1.61943212e-01 -4.69265223e-01 -9.63950276e-01 1.20711625e+00
1.77463639e+00 -5.47755659e-01 2.66252190e-01 8.00260484e-01
8.18992317e-01 -4.60298657e-01 -9.76220727e-01 8.41494918e-01
-6.17108457e-02 -1.48472261e+00 -1.04045846e-01 -7.17599392e-02
1.18114460e+00 5.69933593e-01 5.49760573e-02 3.15168619e-01
3.29158515e-01 -1.11998260e+00 6.64195061e-01 8.34180057e-01
1.54744074e-01 -1.38502407e+00 1.57568336e+00 4.74827409e-01
-1.07017040e+00 -3.58718216e-01 -2.23527744e-01 -8.72109160e-02
6.81286931e-01 1.20561004e+00 -5.24229586e-01 9.85707283e-01
1.11061668e+00 7.20027268e-01 7.79483989e-02 7.85096467e-01
-1.14606328e-01 1.01546109e+00 -7.03345895e-01 -9.45331678e-02
3.80611658e-01 -3.69066447e-01 4.13715065e-01 1.12069964e+00
7.99998999e-01 -1.09976977e-01 1.22016847e-01 5.17075419e-01
4.61096108e-01 -2.57287234e-01 -3.01247656e-01 1.92743152e-01
7.30224252e-01 1.08497894e+00 -4.13818866e-01 -5.54949760e-01
-3.36170405e-01 1.03792362e-01 -4.24810082e-01 3.44945818e-01
-6.02260947e-01 -1.96128070e-01 1.96164951e-01 -2.91146845e-01
3.55208129e-01 -3.69147956e-01 -8.06508958e-01 -5.84712386e-01
-2.93776244e-01 -4.54488248e-01 6.25649631e-01 -8.24600518e-01
-1.88843560e+00 7.30465651e-01 3.32814485e-01 -1.21794200e+00
-9.97657537e-01 -4.71850246e-01 -9.61419582e-01 1.35733831e+00
-1.78961825e+00 -1.18012857e+00 8.81090313e-02 2.70223051e-01
7.85102725e-01 -4.64144140e-01 7.45948911e-01 2.07949281e-01
-6.17213905e-01 1.00692853e-01 1.24434626e+00 -2.66427666e-01
-1.87755749e-01 -1.39922881e+00 4.19856638e-01 7.64403462e-01
-2.80938268e-01 -1.81233793e-01 8.70765090e-01 -4.62674797e-01
-1.19999433e+00 -9.56486106e-01 1.00590861e+00 3.76771949e-02
8.99217546e-01 1.81636095e-01 -1.02581692e+00 5.25122881e-01
8.43302548e-01 -4.02707845e-01 5.93443334e-01 -1.55107945e-01
3.55291128e-01 -3.78328651e-01 -1.36984766e+00 -2.04242140e-01
-2.56516963e-01 -2.53729016e-01 -5.50930738e-01 4.50672507e-01
4.07156013e-02 -2.10856795e-01 -1.46500659e+00 7.72928178e-01
2.78123587e-01 -8.20096791e-01 6.20449424e-01 -1.41898632e-01
-1.91846535e-01 -3.29093486e-01 -3.57276648e-01 -1.98457217e+00
-4.43456203e-01 -4.40482765e-01 -9.21302795e-01 1.37264168e+00
1.10203810e-01 -1.04923987e+00 4.49477643e-01 4.11772490e-01
-1.24190606e-01 -1.05676103e+00 -1.46990967e+00 -9.39777970e-01
4.82260078e-01 -7.53109992e-01 1.15676892e+00 9.76063371e-01
-2.12235805e-02 -2.16789246e-01 -3.32970858e-01 3.72298062e-01
7.53176033e-01 3.83282900e-01 -1.50569631e-02 -1.00571942e+00
3.23885322e-01 -6.55887604e-01 -2.12813124e-01 -3.44811380e-01
2.13719487e-01 -6.07250690e-01 -2.06147105e-01 -1.89959633e+00
-3.22336525e-01 -2.31976748e-01 -8.38921964e-01 4.67231601e-01
3.22234601e-01 -2.95068711e-01 -4.05992046e-02 1.18177518e-01
1.07786313e-01 1.04891324e+00 2.12487325e-01 -1.76191866e-01
9.47753787e-02 4.95713592e-01 1.33582756e-01 5.14744043e-01
1.40537691e+00 -2.53990471e-01 -2.88647950e-01 -3.26228172e-01
2.73811251e-01 8.54972526e-02 9.06275436e-02 -1.03026664e+00
4.64693993e-01 4.11770269e-02 8.07860494e-01 -1.61757910e+00
-9.26862061e-02 -7.96687961e-01 5.32220721e-01 6.90606058e-01
1.19311899e-01 9.18669462e-01 5.13145685e-01 2.42139563e-01
-2.97581762e-01 -1.31038964e-01 6.84699059e-01 9.78108123e-02
-9.20560300e-01 1.83572993e-01 -7.19687045e-01 -3.62573326e-01
1.11050081e+00 -1.30579457e-01 -3.97789776e-01 -4.42135781e-01
-6.58022463e-01 5.42561233e-01 -3.05872172e-01 5.60182452e-01
2.71353483e-01 -1.32451022e+00 -1.07329261e+00 9.40996110e-02
-5.21995783e-01 -3.70026380e-01 5.91639221e-01 8.35907280e-01
-2.70182520e-01 9.42849994e-01 -1.74260903e-02 -7.38352299e-01
-4.16414022e-01 3.74414682e-01 6.94990039e-01 -3.98042917e-01
-5.42210042e-01 4.28177774e-01 -4.52486217e-01 -2.44471431e-01
2.15288043e-01 -5.21615863e-01 -5.76401353e-01 5.81229270e-01
4.33710515e-01 1.18082964e+00 8.22800636e-01 -7.76281416e-01
-2.79580921e-01 2.46896312e-01 3.78187597e-01 1.38712272e-01
1.61304712e+00 -2.11413711e-01 -3.54564458e-01 8.96165967e-01
8.86457741e-01 -8.94535124e-01 -9.47234213e-01 -2.26707272e-02
5.25450647e-01 7.16969073e-02 6.63225532e-01 -1.37061059e+00
-1.24238765e+00 9.28580165e-01 9.82716918e-01 8.12871218e-01
1.14190030e+00 -5.94482780e-01 9.98626351e-01 9.03851446e-03
5.52495956e-01 -1.34715068e+00 -1.03503180e+00 6.24363363e-01
7.42577851e-01 -1.06262600e+00 -8.18273351e-02 8.34494114e-01
-3.56168777e-01 1.46427727e+00 1.07009575e-01 1.79552864e-02
1.39384007e+00 4.66815323e-01 -2.71483958e-01 1.76331580e-01
-1.15801942e+00 3.27338904e-01 2.23987386e-01 6.33363307e-01
1.88664705e-01 4.55491722e-01 -2.27383077e-01 7.35437214e-01
-4.58325058e-01 5.78892007e-02 2.37641841e-01 6.43629432e-01
-2.25645363e-01 -6.07357621e-01 -4.55232561e-01 8.01532686e-01
-6.99324787e-01 -2.86827415e-01 7.88222849e-01 4.32165116e-01
2.57289559e-02 1.40385783e+00 3.35361898e-01 -1.47400916e-01
2.05277614e-02 1.99157134e-01 -1.43230081e-01 8.23369473e-02
-7.64879227e-01 -1.64362907e-01 -1.84974492e-01 -8.22000876e-02
-2.10106596e-01 -6.73214197e-01 -1.04162621e+00 -7.84372687e-01
-5.35530090e-01 6.22031331e-01 1.16057038e+00 1.39529753e+00
3.35196674e-01 6.64836168e-01 8.52094471e-01 -1.22214723e+00
-9.32397723e-01 -1.50741518e+00 -9.40110326e-01 -2.68246979e-01
2.31543630e-01 -5.44890761e-01 -5.74639499e-01 -1.45006552e-01] | [6.157478332519531, 2.8176751136779785] |
25f37fc8-1708-413c-93d1-b03b0a86a2a7 | learning-to-answer-visual-questions-from-web | 2205.05019 | null | https://arxiv.org/abs/2205.05019v2 | https://arxiv.org/pdf/2205.05019v2.pdf | Learning to Answer Visual Questions from Web Videos | Recent methods for visual question answering rely on large-scale annotated datasets. Manual annotation of questions and answers for videos, however, is tedious, expensive and prevents scalability. In this work, we propose to avoid manual annotation and generate a large-scale training dataset for video question answering making use of automatic cross-modal supervision. We leverage a question generation transformer trained on text data and use it to generate question-answer pairs from transcribed video narrations. Given narrated videos, we then automatically generate the HowToVQA69M dataset with 69M video-question-answer triplets. To handle the open vocabulary of diverse answers in this dataset, we propose a training procedure based on a contrastive loss between a video-question multi-modal transformer and an answer transformer. We introduce the zero-shot VideoQA task and the VideoQA feature probe evaluation setting and show excellent results, in particular for rare answers. Furthermore, our method achieves competitive results on MSRVTT-QA, ActivityNet-QA, MSVD-QA and How2QA datasets. We also show that our VideoQA dataset generation approach generalizes to another source of web video and text data. We use our method to generate the WebVidVQA3M dataset from the WebVid dataset, i.e., videos with alt-text annotations, and show its benefits for training VideoQA models. Finally, for a detailed evaluation we introduce iVQA, a new VideoQA dataset with reduced language bias and high-quality manual annotations. Code, datasets and trained models are available at https://antoyang.github.io/just-ask.html | ['Cordelia Schmid', 'Ivan Laptev', 'Josef Sivic', 'Antoine Miech', 'Antoine Yang'] | 2022-05-10 | null | null | null | null | ['video-question-answering'] | ['computer-vision'] | [ 1.86961234e-01 -1.27901718e-01 -2.20852485e-03 -4.13931280e-01
-1.43380368e+00 -1.00297725e+00 5.58937609e-01 -4.07901615e-01
-3.36294204e-01 4.75109547e-01 4.74443346e-01 -2.41435349e-01
2.94176459e-01 -4.79294419e-01 -8.91349733e-01 -3.69110376e-01
3.10747951e-01 3.46277088e-01 5.06891966e-01 -1.85655862e-01
-4.03051525e-02 -3.05320144e-01 -1.64656365e+00 8.72375846e-01
5.47525525e-01 1.23432624e+00 1.44951552e-01 1.16182137e+00
-2.55696863e-01 1.61739326e+00 -6.81429625e-01 -8.33568811e-01
2.99020529e-01 -8.56963098e-01 -1.24452579e+00 3.28811884e-01
1.17421174e+00 -7.78451681e-01 -6.95859373e-01 4.46080387e-01
5.23433030e-01 2.67544746e-01 5.67561626e-01 -1.63807547e+00
-6.66186035e-01 2.88121760e-01 -2.47876793e-01 3.40475827e-01
1.01097882e+00 5.84951282e-01 1.24001288e+00 -1.01406419e+00
1.16046405e+00 1.11774957e+00 4.26334709e-01 7.75254548e-01
-9.92190063e-01 -6.45422459e-01 4.52713892e-02 7.27142453e-01
-1.19534135e+00 -5.54962814e-01 7.91339099e-01 -4.69913244e-01
6.78943217e-01 3.08620602e-01 6.72555208e-01 1.71595430e+00
-2.93883264e-01 1.20083797e+00 7.35206008e-01 -6.64848685e-02
4.21588384e-02 -1.63237065e-01 -4.18422334e-02 9.50212121e-01
-5.17859936e-01 -4.16958660e-01 -8.71688008e-01 2.84292027e-02
4.19482112e-01 -1.97368696e-01 -6.35332406e-01 -4.41983759e-01
-1.36405504e+00 9.84843731e-01 1.37992010e-01 5.09354249e-02
-1.94115490e-01 3.49625736e-01 6.72447860e-01 6.93913817e-01
2.07825854e-01 2.49315679e-01 -3.49644840e-01 -4.14480120e-01
-9.29218590e-01 4.35066730e-01 8.44471097e-01 1.13583505e+00
6.09487951e-01 3.38980672e-03 -5.98048031e-01 6.90942109e-01
4.41559702e-02 7.76579142e-01 2.45569676e-01 -1.64502680e+00
8.35109770e-01 4.49810654e-01 3.03983912e-02 -9.58842456e-01
-1.33718118e-01 3.24183971e-01 -4.71801490e-01 -3.19099069e-01
6.70059741e-01 -4.43269089e-02 -1.00260854e+00 1.58758402e+00
3.71224970e-01 6.19595051e-02 7.78849274e-02 1.16055155e+00
1.54116881e+00 9.25279975e-01 1.36898622e-01 -7.92785138e-02
1.41098773e+00 -1.29117072e+00 -8.65961850e-01 1.66124627e-01
5.94176829e-01 -6.33666456e-01 1.62625980e+00 2.74560273e-01
-1.24390340e+00 -5.78193247e-01 -5.86266220e-01 -6.01214290e-01
-9.90782380e-02 1.84438705e-01 2.04610214e-01 2.99831182e-01
-1.15223849e+00 -7.58150443e-02 -4.66416448e-01 -4.07014817e-01
5.55506170e-01 -5.49764745e-02 -4.97528166e-01 -4.35138553e-01
-1.21103764e+00 4.49966609e-01 1.53505206e-01 -1.03822581e-01
-1.56517112e+00 -7.27063596e-01 -1.07617092e+00 -3.88750166e-01
7.42550731e-01 -8.06861579e-01 1.50664568e+00 -1.41118896e+00
-1.35851991e+00 9.60826933e-01 -2.89338082e-01 -4.47941571e-01
5.05373180e-01 -1.78488493e-01 -3.98610741e-01 1.18229580e+00
2.59431392e-01 1.12846971e+00 1.11468565e+00 -1.05162990e+00
-5.22574186e-01 1.37452595e-02 7.00783193e-01 1.25578508e-01
-2.90082723e-01 -7.96950981e-03 -7.44987369e-01 -5.31331718e-01
-4.60670292e-01 -7.60641575e-01 1.63788810e-01 1.00252926e-01
-1.46693468e-01 -1.39403269e-01 9.42979455e-01 -1.12242925e+00
1.12324274e+00 -2.01766157e+00 3.19032758e-01 -1.51385263e-01
3.66706461e-01 9.55103934e-02 -6.94287539e-01 3.86979192e-01
1.71552554e-01 -6.56055361e-02 -2.23731428e-01 -1.80360198e-01
-5.18244170e-02 2.38327757e-01 -4.56637889e-01 1.79418072e-01
4.41622347e-01 1.24190295e+00 -9.59347844e-01 -8.56978357e-01
4.53100167e-02 3.95434082e-01 -8.24391961e-01 5.82168996e-01
-7.91981041e-01 5.14919341e-01 -3.38077933e-01 8.81433189e-01
2.35471696e-01 -5.10020018e-01 -5.44431657e-02 -5.65636337e-01
2.89667815e-01 2.73172148e-02 -7.51112521e-01 2.09342098e+00
-5.34857273e-01 1.01050866e+00 2.75045447e-03 -7.77013183e-01
5.17190158e-01 6.06760561e-01 4.59026188e-01 -8.85253727e-01
1.97922617e-01 9.19601843e-02 -4.71198767e-01 -1.12157452e+00
5.20115316e-01 1.86771214e-01 -2.08919689e-01 3.69540066e-01
7.18067348e-01 -3.41104090e-01 6.79580331e-01 8.29839468e-01
1.37944758e+00 4.36699510e-01 -5.19694947e-02 2.24846005e-01
6.68779492e-01 2.90398598e-01 1.09421186e-01 5.57200968e-01
-3.82300109e-01 1.00058043e+00 6.94208384e-01 -3.39182377e-01
-1.00532830e+00 -1.07953608e+00 4.37992394e-01 1.21978450e+00
4.13188431e-03 -7.28457034e-01 -7.80999124e-01 -1.08172190e+00
-2.91265398e-01 5.18786967e-01 -5.82936347e-01 1.00240536e-01
-6.68755233e-01 1.10803410e-01 6.91056967e-01 3.73146087e-01
5.60021102e-01 -1.15678942e+00 -5.80696702e-01 -1.06208161e-01
-1.10887718e+00 -1.50984287e+00 -7.69732893e-01 -4.06550229e-01
-4.27352637e-01 -1.48680854e+00 -8.59422445e-01 -9.54939961e-01
3.01205367e-01 3.62381965e-01 1.60513055e+00 1.38086885e-01
-1.14259891e-01 1.27660060e+00 -7.38690019e-01 1.66669369e-01
-3.87573421e-01 3.10217198e-02 -4.05108780e-01 1.71086550e-01
3.31719965e-01 -1.95496768e-01 -7.64445364e-01 4.66224134e-01
-1.07723236e+00 8.16657469e-02 1.37896433e-01 7.29089975e-01
6.51721179e-01 -6.53824151e-01 6.66409254e-01 -4.59546745e-01
2.43872270e-01 -6.14857674e-01 -3.27996641e-01 2.86723346e-01
2.05293939e-01 -1.07533254e-01 5.34125447e-01 -4.98200208e-01
-1.13981950e+00 1.78731173e-01 -3.46218437e-01 -9.96594667e-01
-8.23980123e-02 1.18099026e-01 -3.58822763e-01 1.38098031e-01
6.46979809e-01 1.25360638e-01 -1.46441516e-02 -1.32902503e-01
7.73645282e-01 5.11138082e-01 6.45844162e-01 -4.40290451e-01
8.41288865e-01 6.24529362e-01 -4.08456892e-01 -8.17416608e-01
-1.00121760e+00 -6.30419910e-01 -3.35912079e-01 -6.58242643e-01
1.43517971e+00 -1.15949762e+00 -7.66837478e-01 1.62939385e-01
-1.25680935e+00 -6.90246820e-01 -3.86636436e-01 1.31952956e-01
-8.96688104e-01 3.74283612e-01 -7.00522780e-01 -4.27030653e-01
-3.97050828e-01 -1.03256798e+00 1.19813013e+00 -1.01761103e-01
-1.87447384e-01 -8.51477981e-01 -2.00480614e-02 1.17747688e+00
8.57911929e-02 2.86984473e-01 4.45591807e-01 -6.06345773e-01
-8.92015278e-01 1.36078805e-01 -2.08880901e-01 4.24619257e-01
-2.36332744e-01 -5.55691756e-02 -7.89630473e-01 -6.94226772e-02
-6.54807538e-02 -1.19330311e+00 9.88444805e-01 5.47086969e-02
1.25708735e+00 -4.95455623e-01 1.78503111e-01 4.48777527e-01
1.14688313e+00 -1.42704234e-01 6.69527471e-01 9.53968540e-02
9.30530846e-01 7.14993954e-01 9.55142021e-01 2.73454279e-01
8.71129751e-01 5.35198808e-01 5.05027711e-01 2.22323209e-01
-5.09947836e-01 -3.88447344e-01 7.28905797e-01 9.80454028e-01
2.27260739e-02 -5.76137125e-01 -8.54327261e-01 9.79290724e-01
-1.64836490e+00 -1.40868521e+00 -4.52922612e-01 1.67269444e+00
9.12809610e-01 -3.94761771e-01 3.96235794e-01 -5.24677113e-02
3.39112967e-01 3.88947546e-01 -3.46023053e-01 -2.16241404e-02
-2.42825404e-01 2.03100815e-01 3.22315022e-02 3.41320932e-01
-1.16294658e+00 8.67772639e-01 5.52513552e+00 8.01892281e-01
-6.68268740e-01 5.80172122e-01 4.87473816e-01 -5.15754223e-01
-5.41997671e-01 -1.93980917e-01 -5.02819657e-01 3.84890497e-01
1.12168729e+00 2.46682853e-01 3.85174692e-01 5.11184514e-01
2.30544016e-01 -5.21199927e-02 -1.10682082e+00 1.19784200e+00
7.21343637e-01 -1.46994853e+00 4.12335515e-01 -4.90119666e-01
7.59876072e-01 -8.53861496e-02 -1.65713742e-01 4.87450063e-01
1.28170997e-01 -6.74922168e-01 9.38720167e-01 3.82547408e-01
9.63161707e-01 -4.32436734e-01 5.40861666e-01 -1.40395269e-01
-1.33845186e+00 -1.90221459e-01 -7.73787359e-03 2.37122864e-01
5.68649590e-01 1.26509875e-01 -3.78946632e-01 4.73540515e-01
1.02780604e+00 8.98355603e-01 -8.68906140e-01 6.39227211e-01
-3.74434352e-01 7.87456036e-01 -8.34046528e-02 1.35592952e-01
2.03805387e-01 -2.19948869e-02 4.00834590e-01 1.02223325e+00
2.16350809e-01 2.10684359e-01 6.54539540e-02 5.85445225e-01
-3.17710459e-01 1.37164980e-01 -5.34096420e-01 -9.45310369e-02
1.68449476e-01 1.33586848e+00 -3.41317654e-01 -5.12883127e-01
-7.67105579e-01 1.16319990e+00 1.85116410e-01 5.60288846e-01
-1.18702066e+00 -2.77666092e-01 3.86307091e-01 1.17918320e-01
5.11668861e-01 1.54811181e-02 5.13618767e-01 -1.62078202e+00
1.94838718e-01 -1.44218099e+00 8.18718076e-01 -1.18630326e+00
-1.17874110e+00 5.86185992e-01 7.76441991e-02 -1.42081797e+00
-4.74718183e-01 -5.10801852e-01 -2.66379863e-01 2.92467438e-02
-1.33524513e+00 -1.41888213e+00 -7.98950553e-01 1.23593307e+00
1.04991448e+00 -1.04332976e-01 3.82652849e-01 6.61830902e-01
-1.20718218e-01 5.03164768e-01 -4.36450988e-01 3.85063410e-01
9.94072855e-01 -1.03835571e+00 9.61631984e-02 7.28228867e-01
4.19594765e-01 6.86767569e-04 5.82476854e-01 -3.63565326e-01
-1.63900185e+00 -1.32766938e+00 5.55475771e-01 -9.99610245e-01
9.78780866e-01 -5.30661285e-01 -7.61692166e-01 8.61553609e-01
6.19727433e-01 -6.01267554e-02 6.70599997e-01 -5.42265475e-01
-5.63586891e-01 -5.15602762e-03 -8.94375265e-01 6.26083255e-01
1.25957918e+00 -8.27009916e-01 -6.39512181e-01 4.91611600e-01
1.16803432e+00 -3.69539022e-01 -9.45284009e-01 1.72523320e-01
5.30038774e-01 -8.25234890e-01 9.44799066e-01 -5.78219950e-01
9.13144231e-01 -3.56390089e-01 -4.53738004e-01 -9.15202618e-01
2.70235747e-01 -5.78873336e-01 -3.85565877e-01 1.44527757e+00
3.79159898e-01 5.88562340e-02 6.18088484e-01 2.13067383e-01
-8.81277993e-02 -3.17482203e-01 -8.69686127e-01 -5.45595109e-01
-1.61903232e-01 -6.49747908e-01 2.15584263e-01 8.10590923e-01
-2.55766690e-01 6.61103308e-01 -6.52395010e-01 -2.36283749e-01
4.91069466e-01 1.09311953e-01 1.17649353e+00 -5.99404812e-01
-4.34673339e-01 5.82817905e-02 -3.46343309e-01 -1.21256793e+00
3.86966467e-01 -7.77785063e-01 7.32125342e-02 -1.63284361e+00
2.96098769e-01 2.93895572e-01 3.16771835e-01 3.29113185e-01
-1.07552737e-01 7.97445893e-01 3.80229414e-01 1.53253675e-01
-1.39357555e+00 6.19098067e-01 1.36839616e+00 -3.37377340e-01
-5.49982209e-03 -4.80509043e-01 -1.84147567e-01 5.20737946e-01
6.95215225e-01 -1.98411435e-01 -6.48503125e-01 -6.72855556e-01
3.65932524e-01 3.91547382e-01 7.91222036e-01 -9.52269435e-01
-1.05199240e-01 4.46529351e-02 1.55736521e-01 -6.52624190e-01
5.86705208e-01 -7.49738932e-01 -1.39538096e-02 1.00848183e-01
-4.98614728e-01 1.79087907e-01 8.21435973e-02 6.46432698e-01
-5.20750821e-01 -3.70208137e-02 3.72736245e-01 -1.08602628e-01
-9.77033436e-01 3.94377887e-01 -4.95762408e-01 7.26459682e-01
1.06285238e+00 2.61570956e-03 -6.53961718e-01 -1.03344321e+00
-7.05241740e-01 6.36260867e-01 3.64148945e-01 7.16783702e-01
8.93183708e-01 -1.50491536e+00 -8.39377105e-01 -2.83833385e-01
5.33631325e-01 -2.08916873e-01 6.38877988e-01 9.33883369e-01
-7.06018984e-01 2.64675796e-01 -8.68894160e-02 -7.58057177e-01
-1.36513019e+00 6.56310856e-01 1.58523291e-01 4.56633978e-02
-4.49886024e-01 7.81878829e-01 3.00149053e-01 -4.19697732e-01
1.94516644e-01 -3.20894986e-01 -2.66867042e-01 3.53235275e-01
6.73659623e-01 2.84755737e-01 -2.80018091e-01 -8.74779820e-01
-2.24746570e-01 5.29184103e-01 2.90718526e-01 -2.71501958e-01
1.01223028e+00 -3.06295246e-01 1.57396898e-01 3.62463623e-01
1.51795900e+00 -1.13783695e-01 -1.39597964e+00 6.81548566e-03
-2.16407552e-01 -4.38023835e-01 -4.21743512e-01 -5.89627266e-01
-1.25127077e+00 8.97910416e-01 4.53770906e-01 7.56768212e-02
1.34141016e+00 4.31228817e-01 1.10165536e+00 4.60383385e-01
1.40281260e-01 -9.54577208e-01 9.07945096e-01 3.67975444e-01
1.09523380e+00 -1.48674250e+00 -3.94320816e-01 -1.87022939e-01
-1.03720748e+00 9.24079657e-01 7.56920874e-01 2.36432366e-02
1.75690025e-01 -2.18627974e-01 3.69901747e-01 -2.61944830e-01
-1.11747897e+00 -3.84053886e-01 3.19891453e-01 6.48316681e-01
1.39095888e-01 -3.91341805e-01 -1.08848482e-01 4.97065753e-01
-1.82585374e-01 2.21734554e-01 8.04453969e-01 7.37837017e-01
-1.02933042e-01 -7.89505661e-01 -2.03509673e-01 3.91825706e-01
-5.75223088e-01 -5.50744757e-02 -4.55497324e-01 8.61990035e-01
-1.26863405e-01 1.36262763e+00 1.02175146e-01 -4.69537258e-01
2.63399839e-01 1.40149012e-01 5.01771688e-01 -2.78012246e-01
-5.89135468e-01 -3.20370287e-01 4.81004000e-01 -1.03066409e+00
-8.37452054e-01 -5.83136380e-01 -1.00493109e+00 -1.72470316e-01
1.55546181e-02 9.62817520e-02 2.57844329e-01 7.51747012e-01
4.18767363e-01 3.77570331e-01 3.63519758e-01 -5.91697097e-01
-1.22203328e-01 -6.50393367e-01 6.70683291e-03 8.22314620e-01
4.36903983e-01 -4.01640028e-01 -5.26123106e-01 7.14112699e-01] | [10.4135160446167, 1.0277085304260254] |
c9a3aac8-2584-4d18-a351-472f7869d498 | an-ensemble-deep-learning-approach-for-covid | 2305.10115 | null | https://arxiv.org/abs/2305.10115v1 | https://arxiv.org/pdf/2305.10115v1.pdf | An Ensemble Deep Learning Approach for COVID-19 Severity Prediction Using Chest CT Scans | Chest X-rays have been widely used for COVID-19 screening; however, 3D computed tomography (CT) is a more effective modality. We present our findings on COVID-19 severity prediction from chest CT scans using the STOIC dataset. We developed an ensemble deep learning based model that incorporates multiple neural networks to improve predictions. To address data imbalance, we used slicing functions and data augmentation. We further improved performance using test time data augmentation. Our approach which employs a simple yet effective ensemble of deep learning-based models with strong test time augmentations, achieved results comparable to more complex methods and secured the fourth position in the STOIC2021 COVID-19 AI Challenge. Our code is available on online: at: https://github.com/aleemsidra/stoic2021- baseline-finalphase-main. | ['Kevin McGuinness', "Noel O'Connor", 'Suzanne Little', 'Mayug Maniparambil', 'Sidra Aleem'] | 2023-05-17 | null | null | null | null | ['severity-prediction', 'computed-tomography-ct'] | ['computer-vision', 'methodology'] | [ 7.88670629e-02 -1.42987326e-01 -4.25335646e-01 -6.44414783e-01
-1.13223457e+00 -5.14513910e-01 2.39661410e-01 4.30321127e-01
-5.17387331e-01 6.66136980e-01 2.17512339e-01 -8.99304032e-01
-3.58400971e-01 -7.60943234e-01 -6.65003121e-01 -1.57594815e-01
-3.30892615e-02 1.08028853e+00 7.27117732e-02 1.50459379e-01
-3.33045691e-01 4.28801835e-01 -7.86876798e-01 7.48896182e-01
5.50304174e-01 7.20088005e-01 -2.57616937e-01 1.25664711e+00
4.41363096e-01 9.13496912e-01 -2.15345934e-01 -4.28730220e-01
2.71430612e-01 -2.91236281e-01 -7.92246759e-01 -6.28402412e-01
3.93084258e-01 -6.43007278e-01 -4.79900032e-01 1.12976879e-01
8.86477351e-01 -1.16847895e-01 6.34421587e-01 -1.10300934e+00
-4.29549962e-02 6.06542110e-01 -4.80178863e-01 7.24631190e-01
-1.00082368e-01 4.53970224e-01 7.76027024e-01 -8.94399881e-01
5.12264073e-01 7.08281159e-01 1.29279816e+00 5.77566504e-01
-8.91085804e-01 -8.56892467e-01 -3.74187499e-01 2.09456369e-01
-1.03381240e+00 5.67107759e-02 1.99546039e-01 -6.17626190e-01
1.09258366e+00 6.65177226e-01 8.29665124e-01 1.37623858e+00
3.93118471e-01 7.35438228e-01 8.57197225e-01 -1.85994342e-01
-1.35035321e-01 -2.52033859e-01 5.14806688e-01 6.92871630e-01
4.66983825e-01 1.06788665e-01 2.32387874e-02 -4.98181224e-01
4.99316633e-01 3.68022174e-01 2.29489058e-01 6.74951747e-02
-8.46826077e-01 1.03138912e+00 3.34841013e-01 2.08619237e-01
-4.07429099e-01 2.66786754e-01 7.70436108e-01 2.77472548e-02
5.56981504e-01 5.19493163e-01 -7.61198342e-01 -1.43865868e-01
-1.04286146e+00 4.40261066e-01 3.52893174e-01 1.11137025e-01
-2.68342972e-01 -3.09346765e-01 -6.39959455e-01 7.66989648e-01
1.38044283e-01 5.00432909e-01 4.09347951e-01 -8.42950225e-01
5.46040058e-01 5.72982073e-01 -1.66183561e-01 -4.69479531e-01
-1.07500076e+00 -7.09471881e-01 -8.16328466e-01 -1.29699707e-01
2.31811315e-01 -6.34050071e-01 -1.36220586e+00 1.57153213e+00
1.18821502e-01 1.99412525e-01 -4.95900303e-01 8.36279511e-01
1.05104733e+00 1.04692899e-01 1.62274748e-01 1.31071121e-01
1.27721858e+00 -8.07843208e-01 -4.75193948e-01 4.02707420e-02
9.43165064e-01 -5.57899535e-01 7.97450006e-01 6.21073365e-01
-1.33141005e+00 -2.25428939e-01 -9.11324799e-01 -4.14585806e-02
-3.66143942e-01 8.07317197e-02 6.15263283e-01 1.02975225e+00
-1.16004729e+00 4.98039991e-01 -1.29785979e+00 -1.33921996e-01
7.64572203e-01 5.77648938e-01 -1.07399158e-01 -1.61989927e-01
-1.26884878e+00 8.38807523e-01 3.01695794e-01 -1.82521805e-01
-9.56114352e-01 -1.13893521e+00 -6.37377024e-01 -1.66892171e-01
2.98395693e-01 -1.08830941e+00 1.63659620e+00 -1.28715634e-01
-8.88727903e-01 8.60320508e-01 7.90918991e-02 -6.58410788e-01
7.18392432e-01 -4.10798550e-01 -3.95590186e-01 1.03354931e-01
-1.20294191e-01 5.73701262e-01 1.89119801e-01 -7.00627327e-01
-3.81671160e-01 -4.08760071e-01 -4.64137286e-01 -6.67939782e-02
-1.27242595e-01 3.08471411e-01 -3.65032911e-01 -5.52746296e-01
-3.03112060e-01 -1.05637968e+00 -4.84070331e-01 -3.28811079e-01
-5.10846972e-01 -2.69859154e-02 4.41932023e-01 -1.03499103e+00
1.40529931e+00 -1.68023133e+00 -3.66308451e-01 3.01303983e-01
7.16308177e-01 6.52059793e-01 -3.29609513e-02 1.48439199e-01
-6.93042278e-01 5.14837921e-01 -3.73074025e-01 -3.47720683e-01
-3.74680758e-01 -4.73552011e-02 3.74635339e-01 4.07295316e-01
5.12450933e-01 1.07241583e+00 -5.44101000e-01 -5.00069499e-01
2.95817286e-01 4.47588563e-01 -8.87446344e-01 1.69277996e-01
-1.14318818e-01 4.94349062e-01 -3.16825777e-01 6.99094117e-01
7.15610504e-01 -6.93764746e-01 3.16020846e-02 1.20705500e-01
2.08682474e-02 3.22672486e-01 -6.13428950e-01 1.27598488e+00
-2.93471962e-01 5.14855623e-01 -3.41608018e-01 -7.86134779e-01
3.65857929e-01 4.57393199e-01 9.64277685e-01 -6.03370130e-01
3.64489079e-01 1.21346638e-01 5.32073796e-01 -5.45123041e-01
1.31717086e-01 -4.40138355e-02 4.88925166e-02 5.66536665e-01
-2.35305250e-01 -1.04451306e-01 7.67407492e-02 1.16192274e-01
1.63496530e+00 -3.79890621e-01 2.98884124e-01 8.28306675e-02
1.36688471e-01 2.13972688e-01 3.94188911e-01 9.60358858e-01
-1.83018923e-01 9.61737156e-01 4.41268235e-01 -6.45623565e-01
-1.16279721e+00 -1.06459761e+00 -6.91973805e-01 7.44047880e-01
-7.06259489e-01 -5.20141363e-01 -3.01072985e-01 -7.87131071e-01
4.98025119e-02 7.54901171e-01 -9.86651838e-01 8.65116566e-02
-7.73597240e-01 -1.14775848e+00 1.07971919e+00 9.01662946e-01
5.92823178e-02 -7.41438329e-01 -6.68753505e-01 1.42146513e-01
-1.12754703e-01 -7.74134815e-01 -1.33976996e-01 3.38674337e-01
-7.35493481e-01 -1.25182772e+00 -8.76102328e-01 -5.20612523e-02
1.52719125e-01 -3.36123794e-01 1.41320884e+00 2.89717078e-01
-7.40849137e-01 3.25153381e-01 -2.33329952e-01 -1.07313836e+00
-4.50385153e-01 3.28933299e-01 -7.78303221e-02 -7.08393514e-01
2.92606086e-01 -2.73901224e-01 -9.06048000e-01 -8.99384823e-03
-9.15668786e-01 1.94832280e-01 4.75828648e-01 8.64321411e-01
6.33539200e-01 -5.76263964e-01 3.64838034e-01 -1.10729933e+00
3.95623147e-01 -8.45601201e-01 -3.81284028e-01 -4.88721356e-02
-6.28462732e-01 -3.53413761e-01 3.11870158e-01 5.05213067e-02
-8.74545932e-01 -2.90671349e-01 -6.77123427e-01 -4.20099109e-01
-1.77004129e-01 6.57037556e-01 6.75017178e-01 2.78868973e-01
8.37011278e-01 -3.01786810e-01 -1.96518451e-02 -4.62149709e-01
-2.34125882e-01 5.39128780e-01 1.96312726e-01 -2.47691959e-01
5.22198260e-01 3.12193185e-01 2.78091401e-01 -3.41631591e-01
-9.16141748e-01 -4.26896423e-01 -4.56348181e-01 -5.64684235e-02
1.13423252e+00 -9.63628411e-01 -5.80324829e-01 5.14156938e-01
-9.75131691e-01 -6.54314041e-01 -2.19034091e-01 7.65797913e-01
-1.65752783e-01 -7.04004541e-02 -8.96805584e-01 -5.26117980e-01
-8.27819765e-01 -1.16173899e+00 8.78773749e-01 4.70434614e-02
-4.59800422e-01 -9.77753520e-01 7.33691454e-01 4.29954499e-01
5.15343726e-01 6.40843868e-01 8.28783989e-01 -1.33446145e+00
-1.91789225e-01 -3.05343658e-01 -3.73892218e-01 2.96401024e-01
-1.09583527e-01 1.79987684e-01 -9.32668746e-01 -2.30400130e-01
-3.14985126e-01 -5.29312193e-01 1.05285907e+00 9.92595613e-01
1.88732064e+00 1.90006658e-01 -3.79284114e-01 9.28661346e-01
1.34090018e+00 2.75783837e-01 3.77484679e-01 2.73270220e-01
6.98230028e-01 -8.55872929e-02 2.82870114e-01 7.15998530e-01
3.99085641e-01 3.84541988e-01 5.17926633e-01 -5.78944504e-01
-7.72362724e-02 2.32681841e-01 -3.64377648e-01 5.72315037e-01
-3.42596203e-01 -5.39014220e-01 -1.82181394e+00 6.25055969e-01
-1.36235917e+00 -5.80667615e-01 -5.41119218e-01 1.77050233e+00
6.73784852e-01 3.33377093e-01 3.36162299e-01 -1.52730159e-02
2.98624337e-01 -1.41024992e-01 -7.43795991e-01 -7.06054151e-01
1.10205330e-01 8.72701526e-01 6.38292909e-01 7.73742124e-02
-1.12483394e+00 1.94965348e-01 6.65213776e+00 5.27716398e-01
-1.00338316e+00 5.98353922e-01 1.11474633e+00 -5.85708559e-01
-1.82072401e-01 -5.43384075e-01 -6.07662737e-01 3.68086517e-01
1.29926240e+00 6.37585893e-02 -2.57358462e-01 4.57562119e-01
2.10666835e-01 -3.76642756e-02 -9.35275853e-01 4.99325842e-01
-1.78346127e-01 -1.62253618e+00 -3.71617675e-01 1.05401173e-01
8.34896982e-01 9.35418665e-01 2.90463924e-01 4.15992916e-01
5.13030171e-01 -1.31397283e+00 3.39491144e-02 5.07627010e-01
1.23668134e+00 -6.16497755e-01 1.24587810e+00 -3.36021595e-02
-7.29779720e-01 -6.08041398e-02 2.08682999e-01 1.39425933e-01
-4.97101508e-02 6.62473261e-01 -1.48764277e+00 6.49832070e-01
8.51643026e-01 4.36970681e-01 -8.65131974e-01 1.19054914e+00
2.18371585e-01 1.23765397e+00 -4.73727733e-01 1.41816989e-01
2.38168463e-01 4.76499915e-01 3.45877886e-01 1.44393075e+00
3.15863758e-01 2.22450495e-01 -3.30051854e-02 5.73106170e-01
-1.98559046e-01 1.09048210e-01 -5.62522650e-01 1.80236369e-01
1.77311197e-01 1.15885222e+00 -4.50822681e-01 -4.29162681e-01
-4.50395733e-01 2.45755509e-01 -4.86686602e-02 -1.41540200e-01
-1.32426739e+00 5.39066121e-02 2.29984134e-01 3.01416934e-01
1.37366774e-02 2.07709163e-01 -7.46663511e-01 -8.31072986e-01
-3.90226901e-01 -1.05103290e+00 1.08137929e+00 -6.96834326e-01
-1.16592014e+00 4.85188723e-01 8.61633793e-02 -1.03977299e+00
-5.56112945e-01 -7.12892473e-01 -7.11218834e-01 8.71386528e-01
-1.16243935e+00 -1.08321691e+00 -3.49767059e-01 4.05333340e-01
3.54150236e-01 -1.68205306e-01 9.87516224e-01 4.74959463e-01
-6.91011488e-01 1.03060126e+00 6.55987486e-02 1.80471361e-01
8.51115942e-01 -1.28063035e+00 5.32987356e-01 7.14880288e-01
-4.19472426e-01 2.32342094e-01 3.27383608e-01 -8.86533320e-01
-7.62252569e-01 -1.25324142e+00 6.74889445e-01 -8.73154819e-01
4.84993458e-01 1.65653989e-01 -8.77256334e-01 8.59358430e-01
2.80094296e-01 -1.74033195e-02 1.10209060e+00 1.86329380e-01
3.46983857e-02 1.69843376e-01 -1.15876174e+00 1.97055727e-01
7.72153974e-01 -6.68134019e-02 -3.52443129e-01 4.96832639e-01
7.85644889e-01 -8.88170481e-01 -1.11945343e+00 1.03858912e+00
5.45903623e-01 -8.55128169e-01 9.88162458e-01 -9.47080970e-01
8.24873507e-01 4.05457973e-01 7.77092874e-02 -1.11390865e+00
-3.70124787e-01 -8.57497454e-02 2.53403503e-02 4.56298441e-01
7.65359282e-01 -5.21959305e-01 9.31264579e-01 5.32178104e-01
-1.92670882e-01 -1.13720894e+00 -7.57573366e-01 -3.76665354e-01
4.77269262e-01 -9.04924095e-01 4.93467689e-01 1.00764382e+00
-3.87244880e-01 -1.35210022e-01 -1.83572203e-01 2.01935954e-02
3.27950299e-01 -3.53795260e-01 4.22169924e-01 -1.16078854e+00
-5.74715495e-01 -3.77584070e-01 -1.55526837e-02 5.05664311e-02
-4.30030078e-01 -1.08216560e+00 -4.95795906e-01 -1.57854295e+00
6.79999471e-01 -4.39966589e-01 -7.41103768e-01 7.57369757e-01
-2.61374682e-01 4.40601766e-01 1.39213994e-01 -2.43265748e-01
-4.14310247e-01 -1.17441379e-01 1.16650724e+00 5.97475469e-02
-1.28205582e-01 3.21823478e-01 -4.48434651e-01 7.62091160e-01
1.58673441e+00 -5.85387409e-01 -1.67676166e-01 -6.25175774e-01
5.02867937e-01 4.96574134e-01 3.85389000e-01 -1.02486026e+00
-1.38693467e-01 -4.68743332e-02 9.38013434e-01 -1.28236556e+00
3.42614353e-01 -4.06443655e-01 2.87024230e-01 1.12625849e+00
-4.28927839e-01 6.13766968e-01 7.64508784e-01 2.41645775e-03
2.28241026e-01 -1.73806965e-01 6.87986314e-01 -7.39783421e-02
-1.38157412e-01 4.71821934e-01 -5.87864876e-01 2.51999974e-01
8.31248283e-01 3.12551647e-01 -5.21281362e-01 -1.95359424e-01
-8.64618659e-01 4.47761178e-01 1.76424216e-02 1.07863322e-01
5.62249243e-01 -9.52533007e-01 -1.29032791e+00 2.05383301e-01
5.78088388e-02 1.22469716e-01 4.30975258e-01 1.15357447e+00
-8.70457530e-01 6.08466744e-01 -3.30003202e-01 -8.53454053e-01
-1.46688187e+00 2.75615335e-01 5.83896637e-01 -1.08636272e+00
-3.34998816e-01 9.71102357e-01 -3.05531844e-02 -6.56854630e-01
-9.50862095e-02 -3.22462082e-01 -7.44579285e-02 -1.51060760e-01
3.95948172e-01 5.91956377e-01 3.88490617e-01 -4.39154729e-02
-6.44825637e-01 4.33843248e-02 -3.23901325e-01 -1.19237844e-02
1.71382654e+00 3.69445890e-01 1.45310268e-01 1.18060030e-01
1.19782054e+00 -2.07739368e-01 -5.71364522e-01 4.49323952e-02
-7.02738166e-02 -9.27242562e-02 2.44077995e-01 -1.40970206e+00
-8.87422562e-01 8.51976275e-01 1.02501631e+00 -1.33207098e-01
1.22931683e+00 -3.56960483e-02 7.14571953e-01 2.94541210e-01
-4.09381509e-01 -7.02378750e-01 1.52264208e-01 5.22512376e-01
5.86462796e-01 -1.43087637e+00 1.41456529e-01 -6.49317503e-02
-4.97203767e-01 8.52303445e-01 8.17524254e-01 3.90019789e-02
6.41033173e-01 6.08089030e-01 -6.46209046e-02 -4.69641209e-01
-1.13564157e+00 1.02130570e-01 3.98905516e-01 3.60578805e-01
7.37304509e-01 4.08241570e-01 -3.25227261e-01 9.96181786e-01
-1.56742394e-01 3.84692609e-01 4.95997339e-01 8.33115816e-01
1.41453072e-01 -1.09632230e+00 -4.68933463e-01 1.24566329e+00
-1.08538711e+00 -3.62496227e-01 -3.25917572e-01 1.16606259e+00
4.03233856e-01 5.66201746e-01 2.53954917e-01 -4.67824012e-01
3.29504937e-01 1.56842425e-01 3.07554811e-01 -5.80712199e-01
-8.09932649e-01 -5.86016569e-03 3.91346723e-01 -5.28540730e-01
-2.05954939e-01 -8.90742183e-01 -1.04379845e+00 -4.16810811e-01
-8.61868113e-02 -2.25413010e-01 6.02191746e-01 6.57813013e-01
2.98454285e-01 1.20905674e+00 4.78478163e-01 -2.74394333e-01
-4.42052811e-01 -1.01713908e+00 -3.10341027e-02 1.36236418e-02
4.48444486e-01 -4.11079198e-01 1.21688144e-02 -3.17239404e-01] | [15.328960418701172, -1.8935619592666626] |
5b9c0fde-3133-45da-9a5c-77413c8f116d | leveraging-auxiliary-tasks-with-affinity | 2107.11787 | null | https://arxiv.org/abs/2107.11787v2 | https://arxiv.org/pdf/2107.11787v2.pdf | Leveraging Auxiliary Tasks with Affinity Learning for Weakly Supervised Semantic Segmentation | Semantic segmentation is a challenging task in the absence of densely labelled data. Only relying on class activation maps (CAM) with image-level labels provides deficient segmentation supervision. Prior works thus consider pre-trained models to produce coarse saliency maps to guide the generation of pseudo segmentation labels. However, the commonly used off-line heuristic generation process cannot fully exploit the benefits of these coarse saliency maps. Motivated by the significant inter-task correlation, we propose a novel weakly supervised multi-task framework termed as AuxSegNet, to leverage saliency detection and multi-label image classification as auxiliary tasks to improve the primary task of semantic segmentation using only image-level ground-truth labels. Inspired by their similar structured semantics, we also propose to learn a cross-task global pixel-level affinity map from the saliency and segmentation representations. The learned cross-task affinity can be used to refine saliency predictions and propagate CAM maps to provide improved pseudo labels for both tasks. The mutual boost between pseudo label updating and cross-task affinity learning enables iterative improvements on segmentation performance. Extensive experiments demonstrate the effectiveness of the proposed auxiliary learning network structure and the cross-task affinity learning method. The proposed approach achieves state-of-the-art weakly supervised segmentation performance on the challenging PASCAL VOC 2012 and MS COCO benchmarks. | ['Dan Xu', 'Ferdous Sohel', 'Farid Boussaid', 'Mohammed Bennamoun', 'Wanli Ouyang', 'Lian Xu'] | 2021-07-25 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Xu_Leveraging_Auxiliary_Tasks_With_Affinity_Learning_for_Weakly_Supervised_Semantic_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Xu_Leveraging_Auxiliary_Tasks_With_Affinity_Learning_for_Weakly_Supervised_Semantic_ICCV_2021_paper.pdf | iccv-2021-1 | ['multi-label-image-classification', 'auxiliary-learning'] | ['computer-vision', 'methodology'] | [ 8.65529776e-01 4.06742215e-01 -5.30205429e-01 -6.54024243e-01
-1.06088924e+00 -4.30209070e-01 4.77731705e-01 2.75082052e-01
-4.79457140e-01 7.98388004e-01 -1.54951839e-02 1.40186995e-01
1.78177565e-01 -4.26793754e-01 -1.02046335e+00 -7.43139207e-01
4.42632943e-01 3.95465881e-01 7.77203441e-01 -8.87870342e-02
4.29399282e-01 -1.71609610e-01 -1.71907926e+00 4.45390135e-01
1.45260978e+00 1.10306549e+00 7.43222117e-01 3.49592566e-01
-2.76404053e-01 7.36094952e-01 -2.20654175e-01 -1.18127726e-01
1.00864232e-01 -3.76059353e-01 -1.12363219e+00 3.38639379e-01
2.97981292e-01 1.12743907e-01 4.34510946e-01 1.23983288e+00
2.32018828e-01 1.67461783e-01 4.46068913e-01 -1.08040833e+00
-4.87956733e-01 7.17221618e-01 -8.93210530e-01 1.05001353e-01
-2.73986999e-02 8.02378431e-02 1.15667820e+00 -7.86203265e-01
5.60584724e-01 9.87125099e-01 5.56623518e-01 5.47235548e-01
-1.26012182e+00 -4.05877322e-01 4.87245172e-01 1.74806133e-01
-1.25658762e+00 -2.66991928e-02 1.02402639e+00 -1.54372498e-01
4.27804261e-01 6.17096424e-02 5.12826562e-01 1.00452816e+00
-2.10403010e-01 1.39241731e+00 1.51637292e+00 -4.43516940e-01
2.02475443e-01 3.86901915e-01 2.15693966e-01 9.07685518e-01
9.98814125e-03 -1.46416619e-01 -6.97679996e-01 1.01770863e-01
6.58620358e-01 -1.46734223e-01 -2.23307729e-01 -5.08042157e-01
-1.22247207e+00 7.28661776e-01 8.37605417e-01 3.12751263e-01
-3.54620457e-01 1.41719535e-01 3.43145132e-01 -4.07845110e-01
6.27165437e-01 4.51457739e-01 -6.49327219e-01 3.02587599e-01
-1.30675912e+00 -2.10431702e-02 3.23892981e-01 9.48736608e-01
1.37002492e+00 -2.50411388e-02 -4.48626518e-01 9.52342391e-01
3.83117437e-01 2.78599143e-01 5.02362907e-01 -7.93493271e-01
2.61292875e-01 6.67236924e-01 -4.03380245e-02 -6.81548417e-01
-3.35252374e-01 -8.89217377e-01 -4.27151233e-01 -1.37497801e-02
3.41097236e-01 1.09931991e-01 -1.40650380e+00 1.83132005e+00
5.93882978e-01 5.91728628e-01 -1.28252972e-02 1.04967189e+00
7.23662913e-01 4.43053126e-01 4.42473233e-01 1.59295306e-01
1.17291284e+00 -1.61145699e+00 -5.58063924e-01 -6.59575522e-01
6.51263654e-01 -6.25764132e-01 1.18715310e+00 -4.25773710e-02
-9.83320653e-01 -7.73696482e-01 -1.04837775e+00 -1.40221313e-01
-3.51082772e-01 3.03235471e-01 6.54963911e-01 2.54125953e-01
-1.12391508e+00 3.33885193e-01 -8.77547681e-01 -4.11838926e-02
9.41204369e-01 3.68139297e-01 2.18028098e-01 1.65037587e-02
-1.18873227e+00 6.77549720e-01 8.54574561e-01 1.58974454e-01
-1.08768439e+00 -8.59489143e-01 -9.00706589e-01 -5.36968093e-03
6.64467335e-01 -5.73564827e-01 1.04157078e+00 -1.37288642e+00
-1.28672862e+00 1.08072245e+00 -2.52771020e-01 -4.75738466e-01
3.05524826e-01 -2.61808962e-01 1.77452996e-01 3.21321875e-01
6.97995901e-01 1.49959993e+00 1.02880347e+00 -1.66491938e+00
-1.16417849e+00 -2.11242557e-01 -1.19191483e-01 5.36526859e-01
-7.52350837e-02 -4.97708648e-01 -6.45327270e-01 -7.53260434e-01
1.44709811e-01 -9.47013021e-01 -4.15456533e-01 -3.28219324e-01
-6.62221313e-01 -2.12419137e-01 9.52384710e-01 -4.80821103e-01
8.18740904e-01 -1.80321538e+00 2.75978595e-01 1.04795866e-01
-6.28752038e-02 2.26591691e-01 -1.77676201e-01 -4.06129926e-01
1.31570175e-01 -1.25730097e-01 -7.55866170e-01 -4.84744132e-01
-1.71217918e-01 1.26479566e-01 -2.53043592e-01 2.59383678e-01
5.37077308e-01 1.25268626e+00 -1.16027796e+00 -8.36410582e-01
3.80605519e-01 3.29599231e-01 -3.50649983e-01 3.05282414e-01
-6.15189731e-01 9.07218337e-01 -6.97150052e-01 8.45182300e-01
4.70807999e-01 -5.89874268e-01 -2.17177585e-01 -3.40606570e-01
6.56156763e-02 -6.63278112e-03 -7.55834877e-01 2.31321096e+00
-4.41639125e-01 1.79311216e-01 3.31822336e-02 -1.24595296e+00
8.02022398e-01 -4.90559556e-04 3.07631731e-01 -6.66352987e-01
1.60729349e-01 5.19171536e-01 -3.03103298e-01 -3.74558233e-02
2.99439728e-01 -5.43559045e-02 -1.09414823e-01 3.17611814e-01
3.40085417e-01 -3.96130621e-01 2.45569646e-01 1.97970331e-01
5.58372319e-01 6.63994133e-01 3.26293032e-03 -5.78923523e-01
8.62611532e-01 1.81919709e-01 7.36118972e-01 6.47288620e-01
-2.82451212e-01 7.68219531e-01 1.30563393e-01 -5.24236299e-02
-8.64133477e-01 -9.42572773e-01 -1.89000085e-01 1.56425190e+00
7.14932799e-01 8.95835552e-03 -1.09219754e+00 -1.06247330e+00
-1.88687295e-01 7.17381060e-01 -7.19520092e-01 -1.83055893e-01
-4.31980848e-01 -8.65052342e-01 2.89696544e-01 5.98817110e-01
7.87017345e-01 -1.32324719e+00 -5.77489793e-01 2.14512944e-01
-3.56075227e-01 -1.26196361e+00 -5.93693435e-01 4.32523876e-01
-7.68817544e-01 -9.57460403e-01 -1.01257348e+00 -1.20902646e+00
9.17417765e-01 2.51587570e-01 9.69311476e-01 1.03212528e-01
-1.87345356e-01 1.89554602e-01 -3.52922976e-01 -1.25206336e-01
-2.77976513e-01 3.89603376e-01 -3.99251550e-01 3.00147325e-01
8.27335045e-02 -3.88647467e-01 -6.68311000e-01 3.09038132e-01
-9.93853927e-01 7.04683840e-01 8.20754886e-01 1.05078363e+00
1.12666941e+00 -3.11705887e-01 9.83096421e-01 -1.26063919e+00
6.75846487e-02 -4.66702968e-01 -5.79798162e-01 4.39689785e-01
-6.68556154e-01 3.55085641e-01 3.02291393e-01 -1.80104494e-01
-1.40079093e+00 4.79465604e-01 -1.20634280e-01 -3.41872633e-01
-3.57964307e-01 3.26516271e-01 -6.47508204e-02 -1.76266998e-01
3.46445739e-01 3.22098643e-01 -4.10838783e-01 -3.27922910e-01
6.51533425e-01 4.28719133e-01 8.55988145e-01 -7.01279581e-01
4.48626548e-01 5.67125380e-01 -1.45511985e-01 -3.71960908e-01
-1.76008034e+00 -7.64111638e-01 -9.21064556e-01 -1.78682476e-01
1.32723558e+00 -1.10970426e+00 -1.47411808e-01 5.84662139e-01
-9.91173208e-01 -7.06072807e-01 -2.51221180e-01 7.65510369e-03
-8.20415497e-01 2.82070726e-01 -4.59971458e-01 -4.44227427e-01
-3.24681878e-01 -1.44080460e+00 1.66883028e+00 5.30812263e-01
1.45972818e-01 -1.18537676e+00 -2.65903294e-01 7.96438694e-01
1.77416712e-01 3.11465710e-01 6.35122120e-01 -5.78020573e-01
-7.33590841e-01 2.70397276e-01 -6.75224304e-01 3.64921570e-01
2.84847409e-01 -5.52751839e-01 -1.15269530e+00 -1.40789613e-01
-1.95747674e-01 -8.10747981e-01 1.21120143e+00 5.69481969e-01
1.37247443e+00 1.14141218e-01 -4.72669870e-01 6.47042036e-01
1.36261833e+00 -2.95091540e-01 3.21463905e-02 3.01157653e-01
1.26118410e+00 7.45904624e-01 1.02553916e+00 1.40288725e-01
5.50809085e-01 5.52683234e-01 3.21802676e-01 -4.77925241e-01
-4.77307916e-01 -3.33339453e-01 -7.62129668e-03 6.66882515e-01
3.13114822e-01 1.50060996e-01 -9.61213589e-01 8.51030409e-01
-2.03698921e+00 -2.94019729e-01 -1.15717977e-01 1.83051550e+00
1.29372180e+00 4.18180615e-01 -5.89881092e-02 -2.14509010e-01
1.05030453e+00 1.88896701e-01 -8.69044125e-01 -2.93620070e-03
-2.61694312e-01 2.67320007e-01 6.94045842e-01 6.02198064e-01
-1.44303775e+00 1.56993616e+00 4.89293718e+00 1.20583630e+00
-1.09207940e+00 5.10333836e-01 1.08609128e+00 2.23403335e-01
-3.42760205e-01 5.36721349e-02 -9.35432792e-01 5.46233535e-01
6.38942778e-01 2.02486336e-01 1.52610227e-01 9.09354568e-01
-3.54984440e-02 -4.56863821e-01 -9.53487575e-01 6.94861352e-01
1.39193475e-01 -1.34904563e+00 1.06447227e-01 -2.87771076e-01
1.40216208e+00 1.65882051e-01 1.80784568e-01 2.15923667e-01
3.89920980e-01 -7.22794950e-01 1.01584768e+00 2.58732021e-01
7.49983490e-01 -6.07641399e-01 7.11495638e-01 2.14295730e-01
-1.39046454e+00 4.23751362e-02 -2.80385297e-02 3.68756145e-01
3.83969814e-01 6.07145786e-01 -8.14138055e-01 4.32613313e-01
6.23099566e-01 1.06371808e+00 -8.21489692e-01 7.78624892e-01
-6.17500782e-01 6.43067420e-01 -2.24975407e-01 2.16167912e-01
8.41522455e-01 -1.21014848e-01 2.60384083e-01 1.19726562e+00
-2.11331338e-01 -1.70355201e-01 5.69692373e-01 1.11769497e+00
-7.57741649e-03 -7.40268081e-02 -1.22093726e-02 1.76567450e-01
2.27147236e-01 1.41268802e+00 -1.36349952e+00 -5.10799825e-01
-3.18731546e-01 1.12810647e+00 3.92871797e-01 4.74577427e-01
-9.80969310e-01 -4.81015593e-02 2.42886260e-01 -1.88578740e-01
4.65593696e-01 1.83717698e-01 -8.42447698e-01 -9.33078587e-01
-2.14504987e-01 -4.83569473e-01 3.18097025e-01 -6.82477593e-01
-9.22465622e-01 4.30074364e-01 -3.50382403e-02 -7.88599014e-01
-4.22962159e-02 -8.09101611e-02 -4.87374514e-01 8.27909708e-01
-2.08747339e+00 -1.71220386e+00 -3.61855954e-01 3.94275457e-01
1.03368115e+00 1.61779076e-01 3.98185402e-01 9.54621658e-02
-5.54349661e-01 3.90801340e-01 -1.65357977e-01 -2.55169421e-01
5.54314196e-01 -1.58588672e+00 1.53063774e-01 8.73557806e-01
5.55035323e-02 7.83358291e-02 5.10383248e-01 -8.00973296e-01
-5.73696852e-01 -1.54754841e+00 2.71165937e-01 -2.90120840e-01
4.67225224e-01 -3.79962415e-01 -1.02500343e+00 3.82712126e-01
7.99250677e-02 2.04029366e-01 3.97713959e-01 -1.62940249e-01
-2.28013122e-03 5.03104106e-02 -9.13833797e-01 3.94025594e-01
9.67205346e-01 -5.18299103e-01 -6.74905539e-01 4.44831699e-01
1.05147874e+00 -4.20842022e-01 -3.84407371e-01 7.04479218e-01
6.41086251e-02 -7.29447007e-01 1.05548000e+00 -1.75742403e-01
4.10054833e-01 -4.21001315e-01 1.46141648e-01 -1.10775411e+00
-8.86024684e-02 -2.06290439e-01 2.26770356e-01 1.34896910e+00
5.79293430e-01 -2.64657676e-01 1.01654828e+00 3.18694651e-01
-5.74806452e-01 -9.79313552e-01 -6.97791994e-01 -3.19118202e-01
-2.67218232e-01 -2.63344198e-01 2.40930200e-01 8.08363080e-01
-4.84509975e-01 4.00814980e-01 -1.82995424e-01 2.51426071e-01
8.74980509e-01 4.10096616e-01 3.70213181e-01 -1.03206372e+00
-1.61509454e-01 -4.30332899e-01 -1.84640065e-02 -1.13973618e+00
6.37734652e-01 -1.20007491e+00 6.84478164e-01 -1.46006668e+00
3.60286266e-01 -8.67235422e-01 -6.88527822e-01 6.98536754e-01
-7.64128149e-01 6.25102043e-01 6.09278269e-02 1.87606514e-01
-1.22053969e+00 8.41314852e-01 1.47779799e+00 -2.09141776e-01
-2.17740223e-01 -9.31853503e-02 -7.74990261e-01 8.15841138e-01
5.89910686e-01 -5.36717892e-01 -6.50168598e-01 -3.85171294e-01
-1.30045712e-01 -2.78562427e-01 3.61505866e-01 -1.03419912e+00
1.77683607e-01 -1.20308630e-01 1.08495370e-01 -5.76440990e-01
1.99298665e-01 -4.64053661e-01 -4.43055481e-01 2.61370271e-01
-6.06715918e-01 -7.48371542e-01 1.36842787e-01 8.26605260e-01
-3.23698521e-01 -2.69292176e-01 1.00973344e+00 -2.22829670e-01
-1.13842595e+00 2.61809498e-01 1.33812159e-01 3.19957823e-01
1.06274796e+00 -2.95677066e-01 -1.25849145e-02 1.86960101e-01
-8.81364405e-01 5.89388847e-01 4.37361777e-01 4.98405665e-01
4.48631018e-01 -1.11419022e+00 -5.00535488e-01 1.09118678e-01
1.40840739e-01 5.16220033e-01 2.44161069e-01 9.49024677e-01
-2.38371626e-01 4.05710757e-01 -3.35254639e-01 -1.11837482e+00
-9.45417225e-01 3.92377526e-01 1.00943521e-01 -3.34353536e-01
-2.42660552e-01 1.26213002e+00 7.99365401e-01 -5.17901301e-01
1.50050417e-01 -3.20730865e-01 -1.96371898e-01 -7.98488855e-02
1.94881082e-01 4.04277779e-02 -8.71840492e-02 -7.99998403e-01
-2.49604478e-01 6.64982498e-01 -2.09438875e-01 -9.43059027e-02
1.27918494e+00 -4.77457255e-01 -6.32672608e-02 5.42082727e-01
1.11955035e+00 -5.19444287e-01 -1.71319652e+00 -3.66553605e-01
3.56006622e-01 -2.88366616e-01 2.52041906e-01 -9.03449833e-01
-1.12142062e+00 8.71621728e-01 5.63661098e-01 -2.93004215e-01
1.19508934e+00 2.60077536e-01 1.06986725e+00 1.35641443e-02
4.12872255e-01 -1.33255899e+00 4.44877714e-01 2.21763670e-01
5.14180541e-01 -1.65928233e+00 -2.95399249e-01 -8.84352326e-01
-8.82087946e-01 5.32721102e-01 9.30590987e-01 -8.25363621e-02
3.19009781e-01 -8.16435590e-02 6.42187297e-02 -1.23990253e-01
-3.07862610e-01 -6.04760706e-01 5.38898945e-01 5.18236101e-01
2.30201617e-01 1.39758615e-02 -3.40352178e-01 5.70113480e-01
2.55095005e-01 -1.10113978e-01 2.34456256e-01 9.19198990e-01
-6.30805194e-01 -1.09950590e+00 -2.03089744e-01 5.17721653e-01
-2.92812377e-01 -1.68289632e-01 -2.11003572e-01 2.82539755e-01
3.83760363e-01 7.43106425e-01 -6.07690327e-02 -2.19186749e-02
-1.51578441e-01 9.57090333e-02 2.53970176e-01 -9.97411072e-01
-4.80070114e-01 2.59639502e-01 -2.10065514e-01 -6.23085558e-01
-8.34104300e-01 -5.49056709e-01 -1.65471637e+00 5.16717672e-01
-5.46011567e-01 2.24614203e-01 5.34567773e-01 1.25230539e+00
2.68486917e-01 7.61479855e-01 4.78043407e-01 -1.09590089e+00
-1.74958348e-01 -8.64568055e-01 -4.36774552e-01 7.07393348e-01
3.10225844e-01 -9.13924098e-01 -1.00577056e-01 3.82671952e-01] | [9.671947479248047, 0.6888312101364136] |
97f5c812-d545-4cb4-ad72-6150800a76e0 | defending-substitution-based-profile | 2207.11237 | null | https://arxiv.org/abs/2207.11237v1 | https://arxiv.org/pdf/2207.11237v1.pdf | Defending Substitution-Based Profile Pollution Attacks on Sequential Recommenders | While sequential recommender systems achieve significant improvements on capturing user dynamics, we argue that sequential recommenders are vulnerable against substitution-based profile pollution attacks. To demonstrate our hypothesis, we propose a substitution-based adversarial attack algorithm, which modifies the input sequence by selecting certain vulnerable elements and substituting them with adversarial items. In both untargeted and targeted attack scenarios, we observe significant performance deterioration using the proposed profile pollution algorithm. Motivated by such observations, we design an efficient adversarial defense method called Dirichlet neighborhood sampling. Specifically, we sample item embeddings from a convex hull constructed by multi-hop neighbors to replace the original items in input sequences. During sampling, a Dirichlet distribution is used to approximate the probability distribution in the neighborhood such that the recommender learns to combat local perturbations. Additionally, we design an adversarial training method tailored for sequential recommender systems. In particular, we represent selected items with one-hot encodings and perform gradient ascent on the encodings to search for the worst case linear combination of item embeddings in training. As such, the embedding function learns robust item representations and the trained recommender is resistant to test-time adversarial examples. Extensive experiments show the effectiveness of both our attack and defense methods, which consistently outperform baselines by a significant margin across model architectures and datasets. | ['Dong Wang', 'Lanyu Shang', 'Ziyi Kou', 'Huimin Zeng', 'Zhenrui Yue'] | 2022-07-19 | null | null | null | null | ['adversarial-defense'] | ['adversarial'] | [ 1.70885086e-01 -1.33455291e-01 -1.99177265e-01 3.26929353e-02
-8.04830372e-01 -1.29482007e+00 6.13872528e-01 -4.02163416e-02
-2.70888031e-01 3.71143192e-01 4.85325843e-01 -3.50504160e-01
-2.93040033e-02 -9.14065361e-01 -9.35624361e-01 -7.21589565e-01
-3.38914096e-01 4.93740827e-01 3.26100998e-02 -5.25356233e-01
2.04422265e-01 5.76874793e-01 -1.04185486e+00 3.77394587e-01
7.28658676e-01 5.62742412e-01 -2.99741030e-01 8.42411637e-01
3.67586166e-01 3.31467867e-01 -8.85762155e-01 -5.31107664e-01
6.96131527e-01 -3.49799275e-01 -4.13606226e-01 -2.14252248e-01
2.41967246e-01 -6.37891412e-01 -7.10971832e-01 1.07998669e+00
7.07480252e-01 6.61170661e-01 7.10837066e-01 -1.03634238e+00
-1.08798671e+00 7.50109494e-01 -2.01222330e-01 3.53321761e-01
5.50682247e-01 3.68207216e-01 1.05826139e+00 -8.71886015e-01
2.26343378e-01 1.10193324e+00 6.35803223e-01 8.30607116e-01
-1.40044785e+00 -6.82343781e-01 4.84747052e-01 -1.10346094e-01
-1.27100766e+00 -2.27857396e-01 7.22714186e-01 -1.62450269e-01
5.89431405e-01 5.84258795e-01 4.13480639e-01 1.71148682e+00
1.66265108e-02 8.08996856e-01 5.51782012e-01 -1.21627888e-02
5.47321618e-01 4.09435004e-01 -6.84195012e-02 3.19726199e-01
2.29900256e-01 3.58977139e-01 -2.14728996e-01 -1.13472891e+00
3.18259627e-01 5.37291646e-01 -4.36190814e-01 -4.05935466e-01
-7.17165589e-01 1.07725585e+00 2.50963539e-01 -1.65017560e-01
-4.49240267e-01 1.15859792e-01 4.89629745e-01 5.33148587e-01
2.94592202e-01 8.89283776e-01 -4.39286709e-01 8.60556215e-02
-5.05402088e-01 5.88667095e-01 8.05877805e-01 6.35758281e-01
3.02348703e-01 2.89020509e-01 -3.77694905e-01 6.30617261e-01
6.03942387e-02 7.46438444e-01 5.58868468e-01 -6.16055608e-01
3.19127709e-01 6.53773472e-02 4.49479729e-01 -1.03537476e+00
4.14505117e-02 -4.67067569e-01 -5.19785047e-01 4.00177063e-03
2.54947454e-01 -5.30691504e-01 -5.91622412e-01 1.80976868e+00
5.32643855e-01 8.02452743e-01 1.57077432e-01 8.26916754e-01
-4.57173102e-02 7.92923808e-01 -6.24420568e-02 1.69431716e-02
1.05698621e+00 -6.83847487e-01 -2.08804116e-01 -1.91301107e-02
5.71115315e-01 -4.50520366e-01 1.25095510e+00 3.21279675e-01
-9.26148415e-01 -3.51652473e-01 -1.18707192e+00 6.80626273e-01
-1.62810281e-01 -5.13629675e-01 3.10044345e-02 1.09123874e+00
-4.89649385e-01 7.84561753e-01 -7.26065159e-01 5.99061958e-02
2.62433261e-01 3.78658533e-01 -8.08843002e-02 -1.39770703e-02
-1.55786824e+00 4.89758223e-01 5.53439371e-03 -3.82650018e-01
-1.39108407e+00 -1.11115825e+00 -6.85357809e-01 1.95433915e-01
2.63598233e-01 -5.91205359e-01 1.02602041e+00 -7.41303623e-01
-1.68606758e+00 2.37741638e-02 2.66353339e-01 -8.58284652e-01
3.29161644e-01 -3.63038868e-01 -7.90169895e-01 2.10100129e-01
-4.74124759e-01 -1.44297823e-01 1.34695899e+00 -1.23134923e+00
-4.19001937e-01 -1.08470872e-01 5.07694423e-01 2.30127305e-01
-9.46751893e-01 8.90153795e-02 -3.63263418e-03 -1.05749512e+00
-6.14836454e-01 -1.29060423e+00 -4.91090178e-01 -2.02176020e-01
-3.41299415e-01 2.51617700e-01 8.00621450e-01 -3.39677423e-01
1.44482601e+00 -2.42559290e+00 1.47431687e-01 7.25226343e-01
3.00237481e-02 6.64408982e-01 -4.26336288e-01 7.45335162e-01
-1.71614245e-01 1.90698132e-01 -1.31178424e-01 -2.57866919e-01
1.60640508e-01 2.02332243e-01 -1.17423284e+00 6.66209996e-01
-4.41302732e-02 6.45836532e-01 -1.02925527e+00 5.12581885e-01
-9.78812873e-02 3.97045821e-01 -1.34801424e+00 4.30522233e-01
-4.03050363e-01 2.77701646e-01 -5.58515370e-01 2.89997697e-01
6.56738877e-01 1.09410293e-01 4.34368879e-01 9.04064551e-02
5.71743011e-01 3.33881050e-01 -1.18714917e+00 1.18779755e+00
-6.29363596e-01 7.58640617e-02 -2.57129937e-01 -9.72225785e-01
7.52361774e-01 3.29245806e-01 3.47358882e-01 -1.87001675e-01
-2.20781602e-02 -4.52772193e-02 -1.00648403e-01 -1.17724799e-01
4.25504297e-01 1.33799370e-02 -3.96525055e-01 7.90815473e-01
-4.34561819e-01 2.51729757e-01 -5.71596324e-01 4.25922841e-01
1.34119129e+00 -2.47090340e-01 8.62609521e-02 4.89696627e-03
4.54645902e-01 -3.29276651e-01 4.93082047e-01 1.11872840e+00
-1.24880001e-01 4.04422373e-01 4.84888434e-01 -4.84348863e-01
-9.93258953e-01 -1.17485464e+00 1.66170061e-01 1.39740586e+00
1.45258814e-01 -5.81250787e-01 -6.10313833e-01 -1.36259472e+00
4.36440229e-01 9.62485373e-01 -9.14110303e-01 -8.01036417e-01
-7.40645707e-01 -6.06813252e-01 8.16042721e-01 4.82252181e-01
-4.24591079e-02 -7.51786411e-01 -1.38833776e-01 2.20780984e-01
3.84870589e-01 -6.35681927e-01 -1.10631967e+00 4.22168933e-02
-5.26004851e-01 -9.65361178e-01 -3.82546276e-01 -4.29983854e-01
6.36523008e-01 4.77623254e-01 6.90422773e-01 -5.84998243e-02
2.40740970e-01 3.27550709e-01 -5.08175731e-01 -1.97375249e-02
-5.53067207e-01 -5.73183522e-02 7.26445258e-01 2.78113842e-01
1.76604614e-01 -7.93926537e-01 -8.25966001e-01 3.83656055e-01
-1.04375112e+00 -1.03237295e+00 1.84613839e-01 1.01844025e+00
3.74660641e-01 5.23785204e-02 6.86480761e-01 -1.29007161e+00
9.32393730e-01 -1.12319219e+00 -3.98304701e-01 -4.93282191e-02
-4.55138385e-01 1.79853849e-02 1.50682914e+00 -1.04004896e+00
-6.31906807e-01 -3.69969398e-01 -3.33723068e-01 -8.24505448e-01
-3.05326302e-02 2.17475742e-01 -3.05758327e-01 5.22459950e-03
1.19412827e+00 3.80757958e-01 -1.81197941e-01 -4.64245081e-01
7.18221009e-01 6.39241517e-01 4.58279818e-01 -6.59003198e-01
1.34690511e+00 4.68754679e-01 -5.90560615e-01 -5.17322838e-01
-6.31439686e-01 -1.53941736e-01 8.31159055e-02 2.90808916e-01
3.02585274e-01 -6.73416615e-01 -5.10887980e-01 3.76309492e-02
-6.72156513e-01 -3.35782081e-01 -5.66729069e-01 2.62444079e-01
-4.18890446e-01 4.92293030e-01 -4.46514398e-01 -7.19647646e-01
-4.34134394e-01 -1.03854609e+00 5.90210557e-01 -6.79669380e-02
-3.70360374e-01 -9.73535359e-01 4.90415037e-01 -1.74588487e-02
3.55297536e-01 1.56769827e-01 9.32258964e-01 -1.60237849e+00
-1.87220052e-01 -6.01022482e-01 4.90167826e-01 4.66925740e-01
6.44199103e-02 -1.73294753e-01 -6.83259964e-01 -7.95893431e-01
-4.36174572e-02 -1.86269686e-01 5.00053227e-01 -2.17149153e-01
1.32987940e+00 -8.50333869e-01 -3.07581276e-01 7.90122449e-01
1.08150065e+00 2.45590672e-01 4.26304668e-01 2.32226700e-01
5.43722749e-01 -1.51342433e-02 5.09919822e-01 7.83932090e-01
-1.64828345e-01 7.56323278e-01 4.70406085e-01 5.01365781e-01
3.83868694e-01 -6.82860970e-01 8.10770333e-01 3.30080062e-01
6.17727757e-01 -3.97592515e-01 -4.53601420e-01 4.64089811e-01
-1.58950639e+00 -1.19156146e+00 5.85205913e-01 2.51427746e+00
8.72333944e-01 3.83332074e-01 4.82827783e-01 7.68681150e-03
4.53738421e-01 3.02216470e-01 -8.01741600e-01 -5.97663939e-01
1.61138713e-01 3.36098880e-01 5.77886105e-01 4.74615693e-01
-1.05942011e+00 8.11384737e-01 6.33845282e+00 7.87456453e-01
-9.43241179e-01 1.06789097e-01 1.94639966e-01 -6.12233579e-01
-8.15078020e-01 -5.32330386e-02 -6.65253639e-01 8.79782736e-01
1.28181398e+00 -2.61761189e-01 7.87803769e-01 1.04380977e+00
9.92489606e-02 8.41275871e-01 -9.55214202e-01 3.71484339e-01
1.89093217e-01 -1.38401818e+00 3.33024859e-01 1.46531597e-01
1.01220012e+00 -9.95897278e-02 6.87094510e-01 4.70806092e-01
9.07050610e-01 -7.24507451e-01 4.39021081e-01 2.75269955e-01
5.03180861e-01 -1.22141302e+00 4.01054859e-01 3.58305514e-01
-8.37446213e-01 -5.62892020e-01 -5.31584024e-01 1.61702722e-01
4.17617299e-02 1.97166175e-01 -7.64244080e-01 1.77537277e-01
3.34457695e-01 3.06804627e-01 -7.09538460e-02 8.14613223e-01
-2.02684745e-01 1.05942857e+00 -3.65631163e-01 8.50441679e-02
2.38852248e-01 -1.34672701e-01 8.67530704e-01 1.08279467e+00
2.13790923e-01 4.44296896e-02 4.20960575e-01 3.72022599e-01
-3.43068153e-01 2.24354845e-02 -9.42748129e-01 6.58035278e-04
1.18617022e+00 8.43137562e-01 1.08452821e-02 -2.94773579e-01
-2.44447947e-01 1.25193679e+00 4.22260880e-01 5.49768031e-01
-1.10137558e+00 -4.63393718e-01 1.54149401e+00 1.17296234e-01
6.98623836e-01 -1.25365958e-01 3.58357951e-02 -1.15449870e+00
-2.51814574e-01 -1.47913003e+00 5.59628844e-01 -3.25789638e-02
-1.45865691e+00 4.86314923e-01 -1.49224356e-01 -1.34734893e+00
-3.43934894e-01 -1.12372585e-01 -9.67529655e-01 7.67208636e-01
-9.85008240e-01 -6.45696521e-01 3.98949206e-01 8.70806098e-01
4.03128564e-01 -4.63339329e-01 1.06516349e+00 1.51699349e-01
-8.05032074e-01 1.43633449e+00 6.63041890e-01 5.04446253e-02
7.72059381e-01 -1.17793906e+00 9.69087601e-01 1.00732541e+00
3.31642330e-01 1.25921679e+00 8.59561622e-01 -6.11103952e-01
-1.59132552e+00 -1.48455167e+00 2.05651335e-02 -6.48539126e-01
8.85783494e-01 -5.83553731e-01 -1.04763865e+00 7.29579031e-01
-1.10595308e-01 2.24470243e-01 1.19109881e+00 -5.43889366e-02
-9.52440560e-01 1.16965234e-01 -1.43787158e+00 1.04446363e+00
9.97432530e-01 -7.80941010e-01 -5.26866734e-01 6.37426257e-01
9.88982439e-01 -2.57079661e-01 -8.94210696e-01 6.66050687e-02
5.09389222e-01 -5.25021076e-01 1.26257443e+00 -1.34595418e+00
4.06894721e-02 -2.61481196e-01 -2.96677768e-01 -1.73083305e+00
-6.03058219e-01 -1.22256386e+00 -7.96649992e-01 7.77132392e-01
4.12637174e-01 -7.01777756e-01 8.43639851e-01 4.14777339e-01
-2.35731993e-02 -8.82056952e-01 -5.61412513e-01 -9.43067193e-01
2.97471702e-01 -2.12673932e-01 9.17830050e-01 8.31889510e-01
6.39074296e-02 -1.00292787e-01 -8.12183380e-01 8.55193615e-01
3.88906837e-01 9.36304685e-03 1.03423357e+00 -6.06818914e-01
-9.63457286e-01 -9.97542143e-02 -1.50479808e-01 -1.29429924e+00
3.39273900e-01 -7.18099892e-01 -2.80582160e-01 -7.51908779e-01
-3.54853302e-01 -4.42941129e-01 -5.89929402e-01 1.75159395e-01
-5.18158436e-01 3.70148331e-01 1.04892172e-01 2.59791851e-01
-4.15016741e-01 5.58007777e-01 5.69434643e-01 -7.36057311e-02
-3.45591813e-01 3.79663289e-01 -1.09451020e+00 5.58208764e-01
8.95465851e-01 -6.32196486e-01 -7.06583202e-01 -1.73869208e-01
2.66102582e-01 -1.66913450e-01 -5.93915619e-02 -5.87645769e-01
1.56729948e-03 -3.06136400e-01 1.80270076e-01 -6.76747262e-02
2.92379022e-01 -8.10109794e-01 -3.68495695e-02 5.14875591e-01
-6.17354929e-01 8.82896557e-02 4.30563018e-02 1.12203860e+00
4.27006960e-01 -1.29283279e-01 9.60185230e-01 3.52442741e-01
-1.07658632e-01 5.87287366e-01 -5.81009328e-01 1.59432963e-01
1.15323246e+00 3.07697989e-02 -2.01346636e-01 -4.62494433e-01
-7.12512136e-01 1.70960248e-01 6.28642619e-01 4.37745959e-01
6.63572311e-01 -1.43940723e+00 -5.81180334e-01 5.68118453e-01
1.53778404e-01 -5.47416747e-01 2.56642967e-01 3.31227630e-01
-1.92350745e-01 -1.50411934e-01 1.44180521e-01 1.72823280e-01
-8.72593164e-01 1.19478476e+00 3.11859608e-01 -7.77539536e-02
-6.63217843e-01 9.93285418e-01 9.41350237e-02 -4.67919260e-01
2.67254800e-01 1.85949519e-01 -1.05380908e-01 -1.83851466e-01
9.16593432e-01 3.01585078e-01 -9.30539519e-02 -2.00930730e-01
-3.43685925e-01 -2.21917164e-02 -6.51443839e-01 -2.30855867e-02
1.00265908e+00 5.15006222e-02 5.38893342e-01 -1.67071670e-01
1.32208967e+00 8.02206993e-01 -1.36352813e+00 -4.30593759e-01
-4.13816750e-01 -8.02211583e-01 -3.35273474e-01 -4.98150736e-01
-9.21587944e-01 3.00929517e-01 4.02064651e-01 5.07280469e-01
9.91626143e-01 -4.41064864e-01 1.08902824e+00 4.96973664e-01
1.10451221e-01 -6.68177962e-01 1.57250911e-01 4.60661530e-01
6.39538944e-01 -5.70766151e-01 -2.10136175e-01 5.23899421e-02
-7.86613643e-01 5.43256044e-01 5.00296056e-01 -8.73281658e-01
7.95115292e-01 3.23816180e-01 -1.01466790e-01 2.14182839e-01
-9.37068641e-01 3.21228206e-01 8.30602422e-02 8.44879508e-01
-2.37153351e-01 9.59919915e-02 3.42084914e-02 7.56167889e-01
-1.65131032e-01 -6.50559783e-01 5.93791544e-01 7.64031827e-01
-4.46891665e-01 -1.29904854e+00 -4.05425698e-01 4.09878135e-01
-6.10233963e-01 -2.59414375e-01 -2.59659886e-01 1.44622386e-01
-2.05523387e-01 8.76399934e-01 -7.12703317e-02 -9.42319632e-01
5.77216685e-01 8.93143788e-02 -2.99947290e-03 -7.32312202e-01
-9.65628922e-01 -3.90378386e-01 6.38169143e-03 -6.02338433e-01
6.94649398e-01 -7.42282927e-01 -9.47783530e-01 -5.31817198e-01
-1.79852203e-01 5.07015586e-01 2.15882242e-01 7.16076851e-01
7.27675915e-01 3.13289016e-01 1.47404170e+00 -7.06570745e-01
-1.66369915e+00 -6.58802032e-01 -5.35726726e-01 6.05310857e-01
2.93534279e-01 -5.07780492e-01 -6.05175018e-01 -4.65498835e-01] | [5.812811374664307, 7.7030229568481445] |
d7eb4cdb-837d-4a28-953b-f92607ab2056 | pan-tilt-camera-and-pir-sensor-fusion-based | 1510.07390 | null | http://arxiv.org/abs/1510.07390v1 | http://arxiv.org/pdf/1510.07390v1.pdf | Pan-Tilt Camera and PIR Sensor Fusion Based Moving Object Detection for Mobile Security Robots | One of fundamental issues for security robots is to detect and track people
in the surroundings. The main problems of this task are real-time constraints,
a changing background, varying illumination conditions and a non-rigid shape of
the person to be tracked. In this paper, we propose a solution for tracking
with a pan-tilt camera and a passive infrared range (PIR) sensor to detect the
moving object based on consecutive frame difference. The proposed method is
excellent in real-time performance because it requires only a little memory and
computation. Experiment results show that this method can detect the moving
object such as human efficiently and accurately in non-stationary and complex
indoor environment. | ['MyongSong Choe', 'YongChol Sin', 'GyongIl Ryang'] | 2015-10-26 | null | null | null | null | ['moving-object-detection', 'mobile-security'] | ['computer-vision', 'miscellaneous'] | [ 1.41852200e-01 -6.75301373e-01 6.15817606e-01 -5.66294380e-02
3.16994905e-01 -5.55816293e-01 3.07187587e-01 -3.07922870e-01
-1.02118993e+00 7.02707648e-01 -4.41617638e-01 1.04588598e-01
4.76052351e-02 -6.55060828e-01 -2.09660500e-01 -7.69708633e-01
1.62182778e-01 3.31805378e-01 8.82889211e-01 -2.05165058e-01
4.58306521e-02 7.39671767e-01 -1.53239000e+00 -4.23199683e-01
2.12309808e-01 7.70702302e-01 6.59608662e-01 8.41793239e-01
4.58040297e-01 3.02496403e-01 -7.25319386e-01 2.08837792e-01
5.73120892e-01 -1.95181131e-01 -2.35820517e-01 2.35374212e-01
1.65156275e-01 -3.29486191e-01 -2.75561452e-01 1.12582994e+00
6.94192469e-01 4.62174594e-01 4.49262768e-01 -1.15847754e+00
-1.60767421e-01 -1.46009609e-01 -8.58922601e-01 4.57121313e-01
8.44815612e-01 -8.76070336e-02 -1.89289436e-01 -3.70549619e-01
5.59593678e-01 1.27742815e+00 8.89483511e-01 6.63932145e-01
-5.27566552e-01 -6.47297084e-01 3.43891457e-02 3.69823277e-01
-1.41242826e+00 -2.47115597e-01 5.65107405e-01 -3.79278302e-01
5.66203594e-01 4.37572747e-01 6.88720047e-01 4.59110826e-01
4.72446203e-01 3.49503756e-02 1.07241416e+00 -5.78631997e-01
9.24918354e-02 1.82689786e-01 2.41815925e-01 6.46750569e-01
7.20549524e-01 2.45367035e-01 1.42018333e-01 -4.89853919e-02
8.49608004e-01 5.07719040e-01 -4.25079703e-01 -2.92861402e-01
-1.14790916e+00 2.21078455e-01 2.16846064e-01 4.04886037e-01
-3.11154872e-01 -1.09758317e-01 2.69945003e-02 1.74491808e-01
-1.82929128e-01 -1.28931403e-01 -4.43539828e-01 -1.12832107e-01
-4.00043398e-01 2.80474335e-01 6.01109326e-01 1.01176143e+00
2.01531976e-01 -1.39260858e-01 2.55874038e-01 2.36859918e-01
5.99728107e-01 1.10992694e+00 4.27826613e-01 -7.16811419e-01
1.86428532e-01 3.79406631e-01 7.87407994e-01 -1.27715456e+00
-6.59533203e-01 -7.71572366e-02 -6.66086078e-01 8.40698957e-01
6.36775315e-01 -3.89658481e-01 -6.91857874e-01 1.25031805e+00
8.80378962e-01 1.48031125e-02 2.25773361e-02 1.09054673e+00
6.73472703e-01 8.34546447e-01 -1.21451572e-01 -6.07059002e-01
1.78253746e+00 -7.27061152e-01 -9.66116071e-01 -4.13454801e-01
-6.88120574e-02 -1.11308157e+00 4.07904476e-01 3.60119700e-01
-7.83652484e-01 -8.39165032e-01 -8.92563641e-01 3.37987661e-01
-2.41446450e-01 3.22539151e-01 6.83306903e-02 8.13677430e-01
-5.85386813e-01 2.05693468e-01 -8.61744523e-01 -8.90812039e-01
-4.06453997e-01 6.17035687e-01 -3.26970786e-01 -1.38632745e-01
-8.28247666e-01 1.01540697e+00 5.41606247e-01 3.34155440e-01
-9.96281505e-02 8.42448920e-02 -6.75234199e-01 -3.82060200e-01
2.65653133e-01 -6.82079315e-01 1.00844502e+00 -7.64370203e-01
-1.74101305e+00 7.35715985e-01 -1.98404163e-01 -1.80495068e-01
9.09647346e-01 -4.20704544e-01 -6.32476032e-01 -2.59861723e-02
-1.76369864e-02 -2.27644779e-02 6.50515139e-01 -9.26092207e-01
-9.21383023e-01 -7.62950361e-01 -2.45978415e-01 4.94300097e-01
2.53184378e-01 5.02597451e-01 -3.20090383e-01 -2.55330294e-01
2.47942388e-01 -1.13618112e+00 -3.60109150e-01 2.56992728e-01
8.49185288e-02 -1.40139416e-01 1.55884433e+00 -5.38495421e-01
7.24278867e-01 -2.10070133e+00 -4.40964043e-01 -8.36806968e-02
-3.59554052e-01 6.21284068e-01 5.10076880e-01 -3.81902196e-02
2.68378466e-01 -7.19377577e-01 2.93828309e-01 8.29098523e-02
-4.11595106e-01 -3.47399898e-02 1.13597244e-01 1.02997279e+00
-5.25056183e-01 2.46594220e-01 -7.41449416e-01 -5.94443262e-01
5.19712687e-01 7.95968056e-01 8.30987096e-02 2.28148803e-01
4.34112161e-01 8.58087301e-01 -5.97543359e-01 4.81819212e-01
9.93841887e-01 2.96909958e-01 -2.16217220e-01 -6.00479431e-02
-5.85039496e-01 -3.11641067e-01 -1.99755037e+00 1.03099597e+00
7.40984157e-02 5.29826403e-01 5.14698684e-01 -5.81557095e-01
1.06678438e+00 3.67738426e-01 4.30759847e-01 -5.98330021e-01
5.62952280e-01 -6.53384924e-02 -2.20815092e-01 -5.91662347e-01
5.79009652e-01 -1.77958578e-01 2.34172612e-01 2.74303526e-01
-6.37189567e-01 4.70062822e-01 1.42185643e-01 -4.09783840e-01
1.05392015e+00 2.17608482e-01 6.39298260e-01 -2.12133244e-01
9.34084833e-01 -3.42438854e-02 9.45242167e-01 5.73944688e-01
-6.02782369e-01 2.61142641e-01 -7.44521439e-01 -7.73220778e-01
-8.68766248e-01 -8.28670263e-01 -1.64965075e-02 7.97928452e-01
8.21715236e-01 2.58751720e-01 -4.35372740e-01 -6.36235997e-02
-1.83481708e-01 3.44124660e-02 9.65801417e-04 1.48251355e-01
-1.11929226e+00 -5.20182192e-01 4.93521653e-02 3.02665204e-01
7.71623015e-01 -9.43283200e-01 -1.69621134e+00 3.25168222e-01
-5.78023642e-02 -1.40159571e+00 -3.34369004e-01 -2.46755734e-01
-6.87597871e-01 -1.08236384e+00 -8.27740192e-01 -1.17309177e+00
8.66562665e-01 8.72919738e-01 5.59528649e-01 1.10335685e-01
-5.54124296e-01 4.89011943e-01 -1.85041100e-01 -6.38712704e-01
5.22665717e-02 -7.38180161e-01 3.69618148e-01 -1.61447838e-01
2.54480988e-01 1.01430967e-01 -6.76941812e-01 7.92187095e-01
-3.62379313e-01 -2.21139714e-01 1.81558192e-01 2.32256040e-01
3.69629264e-01 8.33096921e-01 -1.04739353e-01 -2.78256238e-01
2.76718885e-01 1.51183307e-01 -1.23301780e+00 2.40155160e-01
-6.22730739e-02 -4.74139333e-01 3.42583567e-01 -8.01387727e-01
-1.22780478e+00 6.47437990e-01 2.64760882e-01 4.94044870e-02
-3.90180260e-01 -3.36583376e-01 -2.32862324e-01 -2.87659436e-01
4.95849311e-01 2.06255868e-01 -2.76823342e-02 -3.74665111e-01
-8.08915794e-02 7.03965902e-01 8.28444839e-01 -1.61154315e-01
1.14954376e+00 9.04156923e-01 3.14828187e-01 -1.09716225e+00
-2.04507142e-01 -8.85065436e-01 -8.42993915e-01 -5.64731896e-01
1.02040255e+00 -1.01590276e+00 -1.24983859e+00 8.11684906e-01
-1.32792163e+00 1.44775271e-01 1.62943155e-01 8.89128625e-01
-2.19297171e-01 7.44610846e-01 -4.97804105e-01 -1.32592154e+00
-5.78813910e-01 -8.42179537e-01 8.52382064e-01 7.68125474e-01
1.70727745e-01 -7.27761209e-01 1.02740191e-01 2.01875567e-01
3.87443542e-01 5.85619032e-01 -2.03190506e-01 1.48809984e-01
-5.90382993e-01 -5.88496268e-01 4.99376580e-02 -3.09961766e-01
4.11396742e-01 -8.48951414e-02 -6.04400754e-01 -6.41674042e-01
4.96661603e-01 3.70862365e-01 3.69607151e-01 6.03060722e-01
2.03232482e-01 -2.06532627e-01 -7.92083085e-01 2.86649466e-01
1.55639839e+00 8.51484001e-01 4.69849676e-01 6.38837874e-01
6.24901891e-01 3.68044496e-01 1.24322462e+00 4.10745740e-01
1.34010330e-01 9.30888951e-01 6.14511147e-02 1.00664757e-01
1.92756221e-01 3.79118890e-01 5.38706601e-01 3.64850998e-01
-3.63876283e-01 -2.42039695e-01 -6.74836278e-01 8.46381187e-02
-1.91929126e+00 -1.40441632e+00 -5.36163926e-01 2.37155867e+00
2.28456393e-01 6.49475679e-02 2.28689536e-01 3.72723877e-01
1.24258125e+00 -3.89540732e-01 -1.32843569e-01 1.08604394e-01
1.91624746e-01 -3.26833457e-01 7.80605078e-01 5.17126799e-01
-1.29275537e+00 5.90563416e-01 6.61223650e+00 3.06312516e-02
-1.16774797e+00 -2.33477205e-02 -1.85727879e-01 1.78283155e-01
8.30401659e-01 -3.60551834e-01 -1.16910458e+00 6.08958304e-01
4.12444711e-01 -4.88761254e-02 1.09710507e-01 8.56379092e-01
2.06458852e-01 -4.54265177e-01 -6.26060307e-01 1.23857009e+00
2.33772859e-01 -4.29130256e-01 -6.65671229e-01 -2.32348908e-02
3.41050386e-01 -3.57873797e-01 -1.17908441e-01 -6.39229044e-02
4.34472449e-02 -3.58712554e-01 7.22198009e-01 4.99928743e-01
1.54106900e-01 -6.06501400e-01 8.64208341e-01 6.55063212e-01
-1.70176828e+00 -1.90771684e-01 -8.05147767e-01 -5.05947471e-01
3.64418805e-01 2.73870647e-01 -6.54892445e-01 2.28233159e-01
9.92036164e-01 2.27766871e-01 -2.60320902e-01 1.47137511e+00
1.22010738e-01 -1.66116446e-01 -6.42134190e-01 -2.08602533e-01
-9.39522684e-02 -3.14763099e-01 7.09753156e-01 1.19294548e+00
6.43920779e-01 6.87024772e-01 5.23670375e-01 3.73334080e-01
6.72454715e-01 -1.98980749e-01 -6.98902249e-01 6.90647185e-01
2.58500636e-01 1.19592249e+00 -1.07337320e+00 -2.42579088e-01
-3.03700596e-01 9.59504187e-01 -3.48007351e-01 4.76648845e-02
-9.57160175e-01 -4.29345489e-01 3.96078467e-01 3.14138204e-01
3.61992270e-01 -6.70173824e-01 2.13095129e-01 -1.00398052e+00
8.78554210e-02 -4.09273833e-01 5.19508779e-01 -7.94859707e-01
-7.10014403e-01 6.52630329e-01 -1.61901236e-01 -1.54209113e+00
5.92444874e-02 -9.01025653e-01 -5.04302263e-01 6.88982248e-01
-1.25812209e+00 -1.12606454e+00 -7.83896089e-01 1.07893288e+00
5.74400723e-01 -2.20274776e-01 5.29885948e-01 3.19975674e-01
-4.09587532e-01 1.10335447e-01 7.22700804e-02 2.45079383e-01
7.28592694e-01 -9.20059204e-01 -4.68599200e-02 1.25669277e+00
-3.98183912e-01 5.75237989e-01 1.20199704e+00 -9.25334513e-01
-1.58084655e+00 -7.06094503e-01 6.53914154e-01 -5.06053567e-01
6.02262393e-02 -1.75191432e-01 -4.74621415e-01 6.03825390e-01
6.28736913e-02 2.87047267e-01 2.08939627e-01 -4.45527196e-01
3.49974483e-01 -1.78225860e-01 -1.51350617e+00 2.15690464e-01
7.71653473e-01 1.94304124e-01 -7.48972476e-01 1.95514232e-01
2.99828142e-01 -7.59773254e-01 -3.73342842e-01 3.63365084e-01
8.07808459e-01 -8.27497602e-01 1.16223466e+00 2.04432949e-01
-9.84344959e-01 -1.07338595e+00 1.76158305e-02 -6.82753384e-01
-6.76387906e-01 -7.26606548e-01 2.44680330e-01 1.02212107e+00
-3.88572752e-01 -6.64519668e-01 8.01764131e-01 5.80895424e-01
4.17888522e-01 8.82297456e-02 -9.28523481e-01 -1.01899707e+00
-8.50955367e-01 2.10578620e-01 1.61637798e-01 6.47871435e-01
-2.32976720e-01 2.73976028e-01 -6.57363117e-01 7.53661275e-01
8.72818768e-01 -2.37301644e-02 1.04733026e+00 -1.57517445e+00
-5.94004169e-02 2.33272463e-01 -9.19265091e-01 -7.77693570e-01
-3.37767929e-01 1.30460560e-01 2.84185201e-01 -1.63157225e+00
1.50603548e-01 -2.79778242e-01 -9.76532698e-02 -5.20344153e-02
-2.46957883e-01 1.59701809e-01 1.94180101e-01 2.11072251e-01
-7.61112690e-01 1.25214890e-01 8.45931411e-01 -7.02739581e-02
-2.10500017e-01 6.13230109e-01 1.27829283e-01 9.42946613e-01
6.88059628e-01 -6.60974383e-01 -7.94041604e-02 -4.47090715e-01
-2.42772952e-01 -4.82726656e-03 3.40489715e-01 -1.59673035e+00
6.78266168e-01 -2.08565965e-01 8.69552493e-01 -9.69241261e-01
5.69261253e-01 -1.65551901e+00 5.65542102e-01 1.17550719e+00
5.11910677e-01 7.13486850e-01 1.12053826e-01 5.86533308e-01
2.40875363e-01 -3.57794344e-01 1.12683606e+00 -4.03095990e-01
-8.21838200e-01 -2.98848003e-02 -4.22940373e-01 -6.07528925e-01
1.60388279e+00 -6.97343409e-01 -4.40471768e-01 -1.55810758e-01
-4.39065218e-01 -1.00727715e-01 6.09721541e-01 3.48140240e-01
5.31641245e-01 -1.24257052e+00 -4.34977263e-01 2.79471457e-01
-3.38181287e-01 -9.77846757e-02 1.05512522e-01 5.81582665e-01
-1.02623117e+00 4.75672275e-01 -6.15316927e-01 -6.27672613e-01
-2.12697721e+00 8.19109321e-01 4.11446869e-01 1.63166579e-02
-7.81772137e-01 4.97333378e-01 -1.62098631e-01 -2.74455488e-01
2.89857179e-01 -2.77978569e-01 -4.49804962e-01 -4.18841898e-01
9.88324821e-01 7.08935857e-01 -4.30820733e-01 -1.14911294e+00
-6.74643636e-01 1.34304094e+00 3.47315073e-01 -1.64515167e-01
9.60566044e-01 -4.88831073e-01 -9.78895500e-02 3.50321025e-01
6.13953412e-01 2.05980033e-01 -1.16985309e+00 -2.55591989e-01
-1.60859764e-01 -7.69255579e-01 -3.03022325e-01 -5.00096202e-01
-7.82465100e-01 3.53962958e-01 1.37350309e+00 1.00028791e-01
1.09888113e+00 -4.41503465e-01 6.31588578e-01 7.19712019e-01
8.80250633e-01 -1.05776501e+00 -1.05216414e-01 4.92931783e-01
5.27665675e-01 -1.09929967e+00 4.72446859e-01 -5.44103980e-01
-3.13950986e-01 1.10182440e+00 7.39899397e-01 -1.61398545e-01
6.85343385e-01 5.38800836e-01 3.64117414e-01 2.40346923e-01
-1.64094418e-01 -2.21903801e-01 -2.29836795e-02 9.41336155e-01
8.71397182e-02 -5.38944325e-04 -3.59040529e-01 -1.14039324e-01
-1.26476675e-01 4.97130044e-02 4.28684920e-01 1.46551955e+00
-9.54367757e-01 -8.16900134e-01 -1.29835403e+00 -3.18510294e-01
-7.06227243e-01 8.13262820e-01 8.84679779e-02 7.57337093e-01
3.91307056e-01 1.25364935e+00 1.87425427e-02 -2.53949553e-01
4.77886736e-01 -4.05133754e-01 5.94726801e-01 -9.86815318e-02
-3.76920909e-01 2.89414763e-01 -3.05698216e-01 -2.58864850e-01
-6.79472804e-01 -7.94003367e-01 -1.40930033e+00 -2.15075940e-01
-5.51003158e-01 2.27797315e-01 8.44094098e-01 7.27062225e-01
-4.12163995e-02 2.71439523e-01 4.64160413e-01 -8.10187459e-01
-2.22385749e-01 -7.95167625e-01 -5.23009300e-01 3.09751004e-01
2.40085006e-01 -6.92286909e-01 -1.18160710e-01 1.16524898e-01] | [6.907871723175049, -1.8249270915985107] |
36758592-a23d-45a1-800a-bc598797d664 | cta-rnn-channel-and-temporal-wise-attention | 2203.17023 | null | https://arxiv.org/abs/2203.17023v1 | https://arxiv.org/pdf/2203.17023v1.pdf | CTA-RNN: Channel and Temporal-wise Attention RNN Leveraging Pre-trained ASR Embeddings for Speech Emotion Recognition | Previous research has looked into ways to improve speech emotion recognition (SER) by utilizing both acoustic and linguistic cues of speech. However, the potential association between state-of-the-art ASR models and the SER task has yet to be investigated. In this paper, we propose a novel channel and temporal-wise attention RNN (CTA-RNN) architecture based on the intermediate representations of pre-trained ASR models. Specifically, the embeddings of a large-scale pre-trained end-to-end ASR encoder contain both acoustic and linguistic information, as well as the ability to generalize to different speakers, making them well suited for downstream SER task. To further exploit the embeddings from different layers of the ASR encoder, we propose a novel CTA-RNN architecture to capture the emotional salient parts of embeddings in both the channel and temporal directions. We evaluate our approach on two popular benchmark datasets, IEMOCAP and MSP-IMPROV, using both within-corpus and cross-corpus settings. Experimental results show that our proposed method can achieve excellent performance in terms of accuracy and robustness. | ['Pengyuan Zhang', 'Chengxin Chen'] | 2022-03-31 | null | null | null | null | ['cross-corpus'] | ['computer-vision'] | [-1.14059128e-01 -4.17693667e-02 1.71333432e-01 -8.38429332e-01
-9.93347883e-01 -3.54618877e-01 2.04516441e-01 -9.56417620e-02
-5.23595035e-01 1.76236421e-01 5.83418012e-01 -2.29229540e-01
6.04828179e-01 -1.38852954e-01 -5.99525094e-01 -4.78490114e-01
-8.79150108e-02 -2.39491224e-01 -2.36623332e-01 -4.11408156e-01
-2.57551849e-01 4.17380571e-01 -1.31309938e+00 3.67143035e-01
5.73822141e-01 1.48089814e+00 1.69853896e-01 7.27742314e-01
-1.66874919e-02 8.57047260e-01 -5.90874493e-01 -4.74718004e-01
-1.17370449e-01 -3.64511847e-01 -4.96168703e-01 -1.07656740e-01
-2.32022673e-01 5.73315611e-03 -6.22065365e-01 7.34291434e-01
9.17748272e-01 5.60195148e-01 2.15571955e-01 -8.49115789e-01
-1.04120016e+00 6.23999953e-01 -2.45550245e-01 3.29265088e-01
9.88048613e-02 3.41607817e-02 1.28374422e+00 -1.22187233e+00
1.37326345e-01 1.33521795e+00 3.90493274e-01 9.33233082e-01
-8.75289857e-01 -7.68919408e-01 5.23438156e-01 4.55140054e-01
-1.29333389e+00 -9.80045617e-01 1.18763161e+00 7.27453381e-02
1.35859120e+00 1.33296162e-01 3.24141622e-01 1.53786814e+00
-1.22752614e-01 1.20801497e+00 9.48188961e-01 -2.33080834e-01
3.02205682e-01 3.95864666e-01 5.34042753e-02 2.59598404e-01
-8.80783796e-01 1.58633739e-01 -9.40344393e-01 1.57376692e-01
1.59079656e-01 -3.36522132e-01 -4.30552542e-01 3.06509137e-01
-6.79725826e-01 8.13281775e-01 5.01085103e-01 5.23754120e-01
-6.12073898e-01 1.16286881e-01 7.97469139e-01 3.23658258e-01
8.57663810e-01 3.70812088e-01 -7.21355498e-01 -5.14711320e-01
-7.34414160e-01 -4.65426952e-01 3.95553142e-01 5.65375209e-01
4.82810974e-01 5.54062247e-01 -1.63282871e-01 1.41432536e+00
4.60203260e-01 1.51179522e-01 7.47962415e-01 -3.15932602e-01
5.26957750e-01 1.60557166e-01 -4.00161505e-01 -6.17785454e-01
-9.43781212e-02 -5.90498209e-01 -7.83660114e-01 -2.79035658e-01
-4.90843892e-01 -3.29144657e-01 -8.30782890e-01 2.10914540e+00
1.36735886e-01 5.12066066e-01 6.53091729e-01 9.45505440e-01
9.40953493e-01 1.06464541e+00 2.49586299e-01 -1.02460377e-01
1.20877850e+00 -1.24671149e+00 -1.06397855e+00 -5.01014888e-01
5.37005305e-01 -6.23804450e-01 1.17440355e+00 8.37475136e-02
-1.21790791e+00 -6.64866447e-01 -9.98170614e-01 -1.99482679e-01
-4.41297352e-01 3.57617944e-01 1.50421828e-01 5.28842926e-01
-1.09345043e+00 1.64429501e-01 -8.09932351e-01 -8.19105953e-02
3.56203794e-01 1.18887775e-01 -2.24155843e-01 9.52397063e-02
-1.68442154e+00 7.83643544e-01 1.89519465e-01 5.98941922e-01
-8.31349850e-01 -7.90304124e-01 -1.10746980e+00 2.99387604e-01
-6.23908965e-03 -9.91569012e-02 1.24956560e+00 -1.16406584e+00
-2.02217484e+00 4.83895540e-01 -5.34189880e-01 -4.33634162e-01
-1.59720138e-01 -4.14234191e-01 -9.30966973e-01 1.68454066e-01
-5.26117861e-01 6.61963224e-01 6.61427796e-01 -1.13134813e+00
-3.79601866e-01 -3.86441767e-01 -2.58238345e-01 2.97123015e-01
-7.88224936e-01 3.45678389e-01 -3.08344394e-01 -8.76471996e-01
-3.48595828e-01 -7.72915542e-01 -2.04272211e-01 -3.12358409e-01
-2.59648860e-01 -4.70338494e-01 8.40198398e-01 -1.07576728e+00
1.42053390e+00 -2.81815219e+00 8.28210264e-02 3.07652391e-02
-3.44674736e-01 6.29645407e-01 -4.34222192e-01 3.87555540e-01
-3.14725578e-01 2.00003922e-01 -1.37972850e-02 -7.39950418e-01
2.58926123e-01 1.42848864e-01 -5.65136254e-01 1.33829549e-01
8.16521704e-01 9.00109351e-01 -5.79076588e-01 -6.04677461e-02
2.42784873e-01 1.08693123e+00 -4.93465811e-01 6.58560634e-01
6.71436787e-02 4.62340593e-01 -3.16381425e-01 3.92411232e-01
4.69990969e-01 2.70891219e-01 -3.48405056e-02 -1.34845629e-01
3.28547731e-02 9.16275859e-01 -6.56586945e-01 1.79271090e+00
-1.07284379e+00 7.47661829e-01 2.60912567e-01 -1.05227900e+00
1.32045162e+00 8.57751369e-01 1.77818000e-01 -1.00955081e+00
2.18906671e-01 2.87809163e-01 2.56916732e-02 -3.50524187e-01
5.45622647e-01 -5.21554828e-01 -8.36399347e-02 1.80949301e-01
3.02878737e-01 2.58312434e-01 -3.62882733e-01 -2.17512101e-02
1.03555417e+00 -3.28959614e-01 1.50924278e-02 1.47987589e-01
6.66645765e-01 -7.51017272e-01 5.00577092e-01 2.58872271e-01
-4.36139435e-01 4.54161495e-01 1.41068444e-01 -2.67195590e-02
-6.44643962e-01 -8.80732238e-01 -1.06267603e-02 1.44337189e+00
-1.86948419e-01 -2.51556784e-01 -6.37316585e-01 -6.63039088e-01
-3.22475821e-01 8.57238829e-01 -6.23615503e-01 -4.97606844e-01
-4.30943489e-01 -4.78045970e-01 7.22807586e-01 8.01715136e-01
2.61941701e-01 -1.32311928e+00 -2.15871558e-01 5.37768781e-01
-2.79447079e-01 -1.53586149e+00 -5.77811480e-01 4.80571449e-01
-4.57345009e-01 -2.94968396e-01 -5.54217935e-01 -7.96719790e-01
2.15640679e-01 4.29649949e-02 8.86433065e-01 -2.12903887e-01
1.10942135e-02 3.66100490e-01 -7.55792677e-01 -3.28876317e-01
-2.84436703e-01 2.45084194e-03 3.63551825e-03 5.97098529e-01
5.99830270e-01 -7.07552671e-01 -4.49871302e-01 1.30125806e-01
-7.77827084e-01 -4.50352788e-01 7.06868351e-01 8.54942799e-01
4.56265241e-01 -1.39471292e-01 1.09287834e+00 -2.53734380e-01
8.65675092e-01 -5.13313830e-01 1.24944888e-01 7.89173022e-02
-5.60037494e-02 1.43102258e-01 8.65642905e-01 -4.91281986e-01
-1.38503814e+00 -1.33262709e-01 -8.66159141e-01 -7.01287925e-01
-9.60557088e-02 6.52899444e-01 -5.40931404e-01 3.97495478e-01
1.42860919e-01 2.45484367e-01 -3.02635461e-01 -5.08390725e-01
4.00576204e-01 1.29896283e+00 3.99248689e-01 -4.45358485e-01
4.17999119e-01 1.06336311e-01 -7.18438506e-01 -1.07834351e+00
-8.67164016e-01 -6.49554968e-01 -2.59861261e-01 -4.97878119e-02
9.62963462e-01 -1.23655367e+00 -4.95959997e-01 2.70013362e-01
-1.36352527e+00 -2.30507419e-01 -1.79832444e-01 6.15667105e-01
-3.43652457e-01 1.00240543e-01 -8.52739632e-01 -1.19772732e+00
-6.14947081e-01 -1.23913133e+00 1.08722150e+00 7.99264759e-02
-1.34846941e-01 -9.67299581e-01 7.15092570e-02 2.25522697e-01
6.87054455e-01 -3.31004888e-01 7.74115562e-01 -1.11316442e+00
-4.27514426e-02 -1.65827684e-02 -4.56097238e-02 9.42846596e-01
-1.03036508e-01 -3.58499557e-01 -1.64609718e+00 -1.23947501e-01
1.63999364e-01 -6.28651857e-01 1.04401839e+00 1.30556464e-01
1.41072345e+00 -2.95158088e-01 8.27621594e-02 5.14318466e-01
1.07082713e+00 4.44543868e-01 8.42301130e-01 -1.93848144e-02
5.65012813e-01 7.23079383e-01 5.19722283e-01 4.42839831e-01
3.70351225e-01 7.32341230e-01 1.38608769e-01 -1.51786819e-01
-7.86673203e-02 -2.84793317e-01 8.74445021e-01 1.57518923e+00
3.40709031e-01 -3.88352692e-01 -4.87130046e-01 8.76278043e-01
-1.58518624e+00 -7.84960628e-01 4.41464275e-01 1.74520099e+00
8.43393564e-01 -7.97732398e-02 -1.88111931e-01 1.66214660e-01
6.39200926e-01 4.82404500e-01 -4.62503761e-01 -1.09180129e+00
-9.88168269e-02 5.73938727e-01 -2.22012535e-01 3.93751264e-01
-9.24572229e-01 9.90361512e-01 5.10537767e+00 9.94721532e-01
-1.38542926e+00 3.34841281e-01 8.49587262e-01 -2.05664903e-01
-3.56126875e-01 -4.52604681e-01 -5.88878572e-01 3.31808478e-01
1.44733870e+00 1.57274351e-01 4.39571619e-01 7.94044912e-01
3.60693157e-01 6.07271016e-01 -1.07994509e+00 9.94758427e-01
1.60490602e-01 -6.94462657e-01 -3.08105737e-01 -2.53988475e-01
2.91816294e-01 2.65253782e-01 3.56738061e-01 8.42035174e-01
-2.39131898e-02 -1.16115177e+00 7.82055676e-01 1.29312845e-02
7.79603124e-01 -1.01827800e+00 8.53170574e-01 5.27374148e-02
-1.43281376e+00 -5.95660768e-02 -1.22665294e-01 2.00422883e-01
3.28417659e-01 4.07828927e-01 -8.12984765e-01 4.91761088e-01
6.44778967e-01 8.53673697e-01 -1.50080755e-01 3.37611914e-01
-3.07031691e-01 9.88715291e-01 -1.66982576e-01 -1.14431515e-01
4.08437103e-01 1.32393032e-01 3.94486278e-01 1.64415324e+00
3.72208893e-01 1.07368313e-01 -1.55158475e-01 6.83195114e-01
-4.26320881e-01 2.30749458e-01 -3.18829387e-01 -5.29594004e-01
5.24405539e-01 1.18718874e+00 1.41941514e-02 -3.72029543e-02
-5.15286565e-01 1.11244059e+00 6.01255536e-01 7.13228822e-01
-8.89185727e-01 -5.14117122e-01 1.21702874e+00 -4.54703927e-01
6.45866036e-01 -1.71413451e-01 6.86861500e-02 -1.13286352e+00
1.53189912e-01 -8.73641551e-01 1.25994369e-01 -6.93679631e-01
-1.43169284e+00 1.18118262e+00 -7.93698013e-01 -8.01871657e-01
-3.42953838e-02 -7.32511342e-01 -7.90694177e-01 1.02049732e+00
-2.09393811e+00 -1.24070537e+00 1.66889966e-01 5.65901577e-01
8.87566209e-01 -2.85948455e-01 8.96242440e-01 5.22168279e-01
-7.82110274e-01 9.52899873e-01 -2.20112335e-02 2.20589444e-01
7.01822996e-01 -9.89562690e-01 4.33457792e-01 8.20537388e-01
1.80701837e-01 4.67304200e-01 3.76856148e-01 -6.87007532e-02
-1.33908474e+00 -1.31708539e+00 9.43572462e-01 -1.07557826e-01
6.85876906e-01 -7.60951400e-01 -1.21745801e+00 6.72826588e-01
4.59779263e-01 2.94978738e-01 9.72895086e-01 4.48899060e-01
-6.30103171e-01 -2.66398400e-01 -6.58903778e-01 4.38043803e-01
7.25278616e-01 -1.17197907e+00 -4.05003846e-01 -2.52075106e-01
1.02453768e+00 -9.62903127e-02 -8.44679713e-01 3.12690675e-01
4.66077000e-01 -7.66763628e-01 8.62088025e-01 -6.05018795e-01
4.38803732e-01 7.33065456e-02 -6.13966823e-01 -1.75156271e+00
3.93313775e-03 -7.20472574e-01 -1.56401694e-02 1.69828081e+00
6.25495911e-01 -4.74984527e-01 1.04005903e-01 2.81408459e-01
-7.22899914e-01 -1.10815597e+00 -1.12241793e+00 -6.60488486e-01
-1.00504711e-01 -9.82471764e-01 5.31490982e-01 8.03903520e-01
2.02516004e-01 6.05459750e-01 -4.95815754e-01 4.19647902e-01
-1.54379636e-01 -3.87077332e-01 3.07721436e-01 -6.89762294e-01
-2.62818575e-01 -4.18058872e-01 -3.14425141e-01 -1.28863251e+00
7.91950285e-01 -9.18705404e-01 4.51354980e-01 -1.16509318e+00
-1.86361194e-01 -2.82888532e-01 -9.15791512e-01 3.81709307e-01
-3.23866844e-01 -1.50873298e-02 1.60146534e-01 -3.58090550e-01
-6.55927896e-01 1.38969254e+00 7.06645668e-01 -9.45218652e-03
-2.58677602e-01 -4.06123668e-01 -7.05881953e-01 4.13696915e-01
7.58099794e-01 -3.50511163e-01 -2.85682410e-01 -4.79483396e-01
-1.29571274e-01 2.91991960e-02 8.82303417e-02 -7.42688894e-01
7.87837133e-02 1.82318091e-01 1.25452191e-01 -4.21213865e-01
8.79862726e-01 -8.08858097e-01 -4.31324869e-01 -2.06738099e-01
-7.30414093e-01 -1.37664542e-01 5.28433740e-01 6.57364607e-01
-7.86799371e-01 7.94791803e-02 8.97237539e-01 2.46303931e-01
-7.37046361e-01 3.18713069e-01 -3.87275279e-01 1.50841534e-01
6.83505237e-01 3.00542414e-01 -3.85866761e-02 -5.49942613e-01
-5.42293012e-01 1.70315310e-01 -1.39601886e-01 8.25890064e-01
9.57447350e-01 -1.49966669e+00 -6.75066888e-01 3.97252470e-01
3.48684639e-01 -4.83622700e-01 6.43438041e-01 8.07821572e-01
3.81675780e-01 3.63014311e-01 1.80268049e-01 -3.02953511e-01
-1.12007642e+00 5.49383104e-01 5.06047368e-01 -2.30694741e-01
-2.68157750e-01 1.16950846e+00 4.85315353e-01 -6.13490701e-01
4.67890203e-01 -2.89718091e-01 -1.71678871e-01 2.84495689e-02
5.98046660e-01 -2.90228613e-02 1.55470356e-01 -8.48492742e-01
-5.83046913e-01 2.59793460e-01 -8.37381184e-02 -3.48618925e-01
1.52737749e+00 -4.65241164e-01 2.92600363e-01 6.70036674e-01
1.76010561e+00 -4.27414710e-03 -1.05199683e+00 -4.10444349e-01
-5.87504096e-02 -8.40398073e-02 3.86025906e-01 -8.26368868e-01
-1.24792004e+00 1.58762753e+00 5.32719672e-01 -3.73153426e-02
1.34048343e+00 6.76205456e-02 1.29056406e+00 1.01674058e-01
-1.66832596e-01 -1.43602359e+00 4.06844050e-01 7.18993843e-01
1.02265096e+00 -1.20987153e+00 -8.89303803e-01 -3.51035967e-02
-1.10299945e+00 9.89645958e-01 5.46845198e-01 9.55953524e-02
7.30216801e-01 2.94396490e-01 3.36728752e-01 1.04753658e-01
-1.22663438e+00 -2.60371000e-01 4.37199920e-01 3.57894808e-01
7.51050591e-01 2.76112892e-02 1.74171135e-01 1.18531919e+00
-6.95837364e-02 -3.17821115e-01 1.52555764e-01 6.73900962e-01
-3.21010947e-01 -9.25722361e-01 2.15265192e-02 -9.12519023e-02
-7.15192616e-01 -2.70130426e-01 -3.92212808e-01 3.04814517e-01
-3.35126936e-01 1.26579523e+00 9.25723612e-02 -6.73490822e-01
4.65383828e-01 4.51588959e-01 -9.72134247e-02 -4.85659152e-01
-7.54647732e-01 1.97920844e-01 4.15587276e-01 -5.41259706e-01
-3.94583285e-01 -2.71809220e-01 -1.34029734e+00 2.82994419e-01
-3.42535287e-01 3.35074395e-01 9.95387793e-01 1.06108010e+00
7.21354902e-01 8.93723130e-01 1.04819095e+00 -5.99116623e-01
-4.43659782e-01 -1.25854576e+00 -5.88051021e-01 4.52827245e-01
6.90867364e-01 -3.59647661e-01 -4.99808848e-01 -2.26302162e-01] | [13.779556274414062, 5.900847434997559] |
ba4d29df-bdec-42c5-9d13-978325faffb9 | a-survey-on-large-population-systems-and | 2209.03859 | null | https://arxiv.org/abs/2209.03859v1 | https://arxiv.org/pdf/2209.03859v1.pdf | A Survey on Large-Population Systems and Scalable Multi-Agent Reinforcement Learning | The analysis and control of large-population systems is of great interest to diverse areas of research and engineering, ranging from epidemiology over robotic swarms to economics and finance. An increasingly popular and effective approach to realizing sequential decision-making in multi-agent systems is through multi-agent reinforcement learning, as it allows for an automatic and model-free analysis of highly complex systems. However, the key issue of scalability complicates the design of control and reinforcement learning algorithms particularly in systems with large populations of agents. While reinforcement learning has found resounding empirical success in many scenarios with few agents, problems with many agents quickly become intractable and necessitate special consideration. In this survey, we will shed light on current approaches to tractably understanding and analyzing large-population systems, both through multi-agent reinforcement learning and through adjacent areas of research such as mean-field games, collective intelligence, or complex network theory. These classically independent subject areas offer a variety of approaches to understanding or modeling large-population systems, which may be of great use for the formulation of tractable MARL algorithms in the future. Finally, we survey potential areas of application for large-scale control and identify fruitful future applications of learning algorithms in practical systems. We hope that our survey could provide insight and future directions to junior and senior researchers in theoretical and applied sciences alike. | ['Heinz Koeppl', 'Mengguang Li', 'Yannick Eich', 'Ahmed Elshamanhory', 'Gizem Ekinci', 'Anam Tahir', 'Kai Cui'] | 2022-09-08 | null | null | null | null | ['epidemiology'] | ['medical'] | [ 2.88796723e-02 1.26031311e-02 -1.80914000e-01 5.21727443e-01
-1.56073764e-01 -4.15810078e-01 4.51780587e-01 4.49557155e-01
-6.09773874e-01 1.36119580e+00 -6.23468816e-01 -3.09730828e-01
-7.39804268e-01 -9.04395223e-01 -2.77565479e-01 -1.14936888e+00
-7.60944486e-01 6.77325606e-01 6.92927688e-02 -7.14594364e-01
9.48491395e-02 4.31562841e-01 -1.33990180e+00 -8.00497890e-01
8.34468722e-01 6.22941136e-01 7.77054355e-02 9.20223773e-01
4.54567790e-01 7.36902237e-01 -7.37979949e-01 5.73336780e-02
2.06235945e-01 -5.99089205e-01 -4.07828689e-01 2.93486357e-01
-4.22426045e-01 2.13031247e-01 -1.67722568e-01 8.72238338e-01
6.91589236e-01 3.69579852e-01 7.19868660e-01 -1.38558650e+00
-4.02993709e-01 3.31468254e-01 -7.57437468e-01 3.00744295e-01
2.95082927e-02 3.89177680e-01 8.16653371e-01 -1.63742065e-01
2.60440201e-01 1.28079748e+00 4.39771801e-01 5.57757199e-01
-1.05270946e+00 -5.34946263e-01 2.85592556e-01 2.79428642e-02
-1.21504068e+00 -2.35491797e-01 4.89677280e-01 -3.25602919e-01
7.43111432e-01 1.53949529e-01 1.08313692e+00 3.82923573e-01
5.32320440e-01 7.88461685e-01 9.67887998e-01 -5.18173039e-01
6.44181490e-01 -2.81693071e-01 -2.63606071e-01 7.50093937e-01
7.03502476e-01 4.10045773e-01 -1.25612468e-01 -5.46486199e-01
8.90950501e-01 -3.31740268e-02 2.87931934e-02 -5.14106154e-01
-1.20063925e+00 1.38444722e+00 1.97439492e-01 1.61082074e-01
-6.53261304e-01 3.21883261e-01 -6.29909150e-03 5.24175465e-01
4.09794986e-01 9.40207303e-01 -1.95758760e-01 -6.84549883e-02
-3.39825928e-01 8.34882140e-01 1.13471222e+00 6.29201174e-01
6.49881959e-01 3.15194011e-01 5.70292294e-01 6.12740695e-01
2.87395597e-01 8.99768889e-01 -1.22460937e-02 -1.35060418e+00
-1.59373023e-02 4.78439003e-01 4.59199607e-01 -8.00635993e-01
-6.44478500e-01 -3.59506994e-01 -1.20717323e+00 4.99801010e-01
3.98683161e-01 -1.00584924e+00 -1.57335162e-01 1.65709972e+00
4.92625922e-01 4.73477952e-02 3.64683241e-01 6.91523254e-01
-8.05292726e-02 7.35963106e-01 -2.19634518e-01 -9.11773860e-01
1.05486488e+00 -6.93784893e-01 -5.37572801e-01 -4.19340014e-01
3.47272217e-01 -5.84779382e-01 1.90231368e-01 2.53643781e-01
-1.23584974e+00 2.19599321e-01 -7.63282061e-01 9.97477353e-01
-2.64162153e-01 -4.59156305e-01 8.79918456e-01 4.84464973e-01
-1.10632634e+00 4.39409107e-01 -1.13464642e+00 -8.18644583e-01
3.06108773e-01 7.93351710e-01 2.49087080e-01 1.23211831e-01
-1.08408189e+00 9.93952394e-01 -8.69797915e-02 1.13800712e-01
-5.86445928e-01 -2.22877726e-01 -6.79728448e-01 -3.36778983e-02
8.50552678e-01 -8.11968386e-01 1.27224362e+00 -8.08242142e-01
-1.55318320e+00 9.70085189e-02 8.35692286e-02 -5.34975886e-01
3.53739917e-01 4.02389646e-01 -3.71071883e-02 7.62426481e-02
1.57188877e-01 1.60483405e-01 6.27022803e-01 -1.02578378e+00
-8.26624334e-01 -3.36370140e-01 2.45207042e-01 6.34329438e-01
-4.03823853e-01 4.55853008e-02 4.44380403e-01 -4.54587281e-01
-5.46164989e-01 -1.24853063e+00 -1.02998579e+00 -2.37059072e-01
-2.13192943e-02 -3.76510054e-01 6.89144135e-01 2.10788190e-01
8.83011818e-01 -1.45871174e+00 6.60433471e-01 2.20963717e-01
4.32296306e-01 3.10824811e-01 -1.72947481e-01 1.12232924e+00
5.47680140e-01 1.73605144e-01 -3.15466672e-02 2.08693832e-01
-1.01030752e-01 2.13709787e-01 5.41028976e-02 6.40708625e-01
2.38780394e-01 7.33496428e-01 -1.08513820e+00 -3.05059493e-01
2.89746195e-01 1.20631874e-01 -6.16533637e-01 1.17181942e-01
-1.68469146e-01 2.52140313e-01 -1.11348522e+00 3.57586026e-01
1.29417658e-01 -6.01750135e-01 3.88326347e-01 5.99979579e-01
-2.87906855e-01 -5.41222930e-01 -1.27708268e+00 6.22316360e-01
-3.27028126e-01 5.67837715e-01 5.68039358e-01 -1.43953383e+00
4.63636786e-01 2.53295511e-01 9.60426509e-01 -2.95678139e-01
2.98959047e-01 4.89847213e-02 5.62993944e-01 -4.18419182e-01
5.00439882e-01 -3.36626410e-01 -2.65632331e-01 7.88769901e-01
-4.34869230e-01 -4.38418180e-01 6.98000550e-01 1.38901100e-01
1.22274864e+00 -6.20001078e-01 7.02294111e-01 -4.81025338e-01
4.62327003e-01 2.51869857e-01 3.62348646e-01 8.37444663e-01
-5.15608430e-01 -3.97560596e-01 3.72397095e-01 -4.00662094e-01
-8.48436773e-01 -7.34830737e-01 2.19432443e-01 8.28138947e-01
5.14299691e-01 -8.95746946e-02 -6.45299673e-01 1.11169510e-01
3.13427508e-01 -1.21045843e-01 -6.71169937e-01 -1.77092314e-01
-6.85933888e-01 -1.36424434e+00 1.21928193e-01 4.46810164e-02
1.84989482e-01 -1.13681388e+00 -5.93759954e-01 4.99425143e-01
2.64038414e-01 -7.92580545e-01 -1.75442472e-01 1.06183238e-01
-5.35054803e-01 -1.32916105e+00 -1.06777561e+00 -9.23506021e-01
5.65199375e-01 4.09694463e-01 7.93921292e-01 4.77886111e-01
-3.95820200e-01 7.70152271e-01 -1.49459943e-01 -6.51869416e-01
-7.02736318e-01 3.79809171e-01 5.90416193e-01 -1.38352647e-01
3.81800160e-02 -2.72453964e-01 -4.52360302e-01 5.59855282e-01
-7.29543447e-01 -2.74191767e-01 3.79264534e-01 9.91811633e-01
2.69569427e-01 3.64290833e-01 1.07544720e+00 -3.33459109e-01
1.23406470e+00 -5.64202309e-01 -1.01274741e+00 2.93461710e-01
-3.25880319e-01 -6.43564202e-03 8.61653388e-01 -5.95949888e-01
-4.48609203e-01 -3.19739938e-01 2.56274551e-01 6.52020499e-02
6.99782521e-02 6.09662235e-01 3.73799533e-01 -5.06482184e-01
4.49306846e-01 1.78741798e-01 7.54408002e-01 3.17229182e-01
2.09179986e-02 5.90162933e-01 -2.34937653e-01 -6.42248869e-01
8.85268986e-01 2.25003093e-01 4.90670711e-01 -1.43356311e+00
-3.48079175e-01 -1.41934186e-01 -7.53260255e-02 -3.27263027e-01
6.23868465e-01 -6.50834441e-01 -1.38182759e+00 6.66060507e-01
-7.97677517e-01 -7.30390489e-01 -3.31564307e-01 1.87848940e-01
-8.68305445e-01 -6.20735995e-03 -6.84523821e-01 -1.21566081e+00
1.86034709e-01 -1.11095393e+00 7.50388205e-01 7.79511869e-01
1.78171089e-03 -1.59082890e+00 5.31624019e-01 -4.35119914e-03
6.06460512e-01 1.68521255e-01 6.95903122e-01 -3.22876811e-01
-6.58843875e-01 -1.54170886e-01 4.98178810e-01 -2.83061028e-01
2.10725799e-01 2.64784515e-01 -1.58477128e-01 -7.76573181e-01
-2.98604608e-01 -4.75495040e-01 4.30888921e-01 8.40487897e-01
4.31788385e-01 -2.34605834e-01 -5.70192456e-01 -3.16732451e-02
1.38273752e+00 4.25175041e-01 -7.80317858e-02 3.35629135e-01
2.50103742e-01 7.37973154e-01 5.93849361e-01 8.12299252e-01
6.20429873e-01 5.36358297e-01 3.62712711e-01 -2.11383358e-01
5.11189044e-01 2.45241866e-01 2.40245894e-01 8.28308463e-01
-3.37039709e-01 -6.81857526e-01 -9.62886095e-01 1.96825027e-01
-2.17047024e+00 -1.27434874e+00 3.76605809e-01 2.15370154e+00
4.02850062e-01 -2.57198632e-01 5.92413366e-01 -4.73010987e-02
1.06115317e+00 -1.94425546e-02 -8.62781048e-01 -1.64001480e-01
-3.35248083e-01 -2.72225827e-01 4.97622699e-01 6.42162800e-01
-9.09441471e-01 8.94927382e-01 6.70476389e+00 8.01284611e-01
-1.03332424e+00 -2.93740451e-01 6.72083497e-01 -1.08527027e-01
2.09889308e-01 -1.38135061e-01 -7.49069035e-01 1.73411012e-01
8.53990376e-01 -7.71952033e-01 1.01939344e+00 3.89659107e-01
6.87917650e-01 -5.52279651e-01 -5.43670654e-01 7.12742746e-01
-3.76451552e-01 -1.43804622e+00 -4.93450969e-01 6.87098682e-01
1.15075517e+00 1.94947794e-01 8.20798054e-02 2.20656320e-01
8.29485059e-01 -7.94740081e-01 2.20742151e-01 1.47182196e-01
1.93401694e-01 -9.61305499e-01 6.17513180e-01 6.79982722e-01
-1.26840353e+00 -4.48225379e-01 -5.21728873e-01 -8.78386259e-01
3.20177376e-01 2.63123363e-01 -2.93567568e-01 8.32315311e-02
3.23522240e-01 7.88725436e-01 -5.39558008e-02 1.21460426e+00
4.14023519e-01 2.82244563e-01 -6.22462094e-01 -1.10232925e+00
2.87550569e-01 -4.05330956e-01 6.38643682e-01 3.71958077e-01
2.37382144e-01 2.81351149e-01 5.98117590e-01 4.62579191e-01
2.40861863e-01 1.31169427e-02 -8.31595242e-01 -4.02039528e-01
6.43212974e-01 1.24690783e+00 -1.20217156e+00 -2.19528690e-01
-3.63761187e-01 1.80402398e-01 4.05118525e-01 6.01587772e-01
-5.83438218e-01 -4.92979586e-01 9.34558749e-01 5.21399975e-02
2.05044776e-01 -4.51946944e-01 1.39733613e-01 -1.27377963e+00
-5.24791300e-01 -9.51493859e-01 2.54820287e-01 1.82654953e-03
-1.40755153e+00 4.06983316e-01 -4.18544970e-02 -9.62707281e-01
-7.22350836e-01 -5.62787294e-01 -7.62551963e-01 2.56784707e-01
-1.39257801e+00 -7.29347229e-01 2.44313806e-01 4.08186853e-01
3.60365033e-01 -6.20298624e-01 7.44017065e-01 -1.54154330e-01
-9.87477839e-01 7.40971938e-02 7.60212839e-01 -1.65254012e-01
3.43633264e-01 -1.12987876e+00 4.91146594e-02 5.50845981e-01
-2.78828263e-01 2.87314475e-01 9.44153428e-01 -4.89898294e-01
-2.04779315e+00 -8.65107000e-01 1.88430831e-01 -2.21541971e-01
1.17008638e+00 -3.24172199e-01 -3.34535897e-01 2.47002527e-01
4.40587014e-01 -1.49271160e-01 4.10831273e-01 5.69818057e-02
6.70706511e-01 9.76990238e-02 -8.89249742e-01 7.94702530e-01
4.85643059e-01 2.04283446e-01 5.21483123e-02 3.88575733e-01
2.24000394e-01 6.50789291e-02 -8.38994145e-01 2.21609510e-02
3.83415431e-01 -4.84328896e-01 8.39833200e-01 -4.93549347e-01
-1.78885944e-02 -1.91243470e-01 1.19963355e-01 -1.75524426e+00
-3.29216182e-01 -1.26979303e+00 9.54439770e-03 7.32344925e-01
3.40409786e-01 -1.02896786e+00 8.70744407e-01 4.29122776e-01
4.10032362e-01 -1.04240310e+00 -7.64802814e-01 -9.28725660e-01
7.01269209e-01 2.43292317e-01 2.09377855e-01 6.89377010e-01
3.87116343e-01 4.56122041e-01 -2.62560099e-01 1.38619661e-01
8.54927719e-01 1.52262479e-01 7.16525793e-01 -1.15847886e+00
-3.34933370e-01 -7.35594809e-01 -2.73611426e-01 -8.14405262e-01
4.00593638e-01 -3.25866878e-01 1.20776929e-01 -1.40385449e+00
8.25468376e-02 -5.84381282e-01 1.39501840e-01 -4.48213443e-02
-2.01554820e-01 2.51205117e-01 3.08457226e-01 1.46078289e-01
-1.06195724e+00 8.58426452e-01 1.32811642e+00 -2.05189526e-01
-3.85306329e-01 4.53961849e-01 -6.70270860e-01 6.68225944e-01
1.02129245e+00 -4.06436503e-01 -3.32069546e-01 -5.40890694e-02
2.83772975e-01 4.19967085e-01 2.05315694e-01 -8.67647409e-01
4.19655204e-01 -1.04828715e+00 -1.81518152e-01 1.86810121e-01
2.95183033e-01 -5.97928643e-01 -2.84415036e-01 8.97877753e-01
-1.58264384e-01 4.88878429e-01 -3.93913090e-02 6.93256497e-01
-1.07405156e-01 -7.55822137e-02 9.10681605e-01 -2.84079194e-01
-3.10239404e-01 4.80731845e-01 -1.24395585e+00 4.27926362e-01
1.58428276e+00 8.22969601e-02 -4.91387129e-01 -8.40812266e-01
-6.46536112e-01 9.80509162e-01 5.60078859e-01 1.44366488e-01
2.92311162e-01 -8.31688762e-01 -7.54074395e-01 2.23845225e-02
-2.91331172e-01 -3.07655036e-01 7.40904585e-02 6.87959731e-01
-4.38042969e-01 2.01933324e-01 -2.12437481e-01 -4.09571826e-01
-9.09718633e-01 3.53710860e-01 7.04161704e-01 -3.22989494e-01
1.44404322e-01 3.23702276e-01 6.44286200e-02 -1.82232037e-01
-8.37871656e-02 1.86418548e-01 -1.74711883e-01 -1.19042546e-01
5.26779711e-01 6.49353802e-01 -6.24617875e-01 -6.07274532e-01
-2.52674848e-01 7.29319394e-01 3.62324379e-02 -1.87641203e-01
1.56567693e+00 -3.13926339e-01 -2.16316536e-01 1.37564018e-01
4.39407974e-01 -3.51328701e-01 -1.25380230e+00 9.01974738e-02
-1.45027056e-01 2.43814066e-01 -8.10276568e-02 -2.01243773e-01
-9.50986981e-01 7.20473886e-01 9.47621539e-02 1.04581368e+00
9.80443656e-01 1.23634256e-01 2.58785814e-01 9.11804318e-01
7.33845055e-01 -1.02807558e+00 3.14732045e-01 7.02983797e-01
5.79853415e-01 -1.29619503e+00 3.42219174e-02 -1.02640413e-01
-5.26623785e-01 1.11410332e+00 7.10941494e-01 -6.01198077e-01
8.89879644e-01 6.43067420e-01 -1.59152538e-01 5.27735008e-03
-1.18547857e+00 -3.27450663e-01 -3.67144108e-01 8.68533909e-01
2.26902708e-01 1.15900986e-01 -2.29153454e-01 8.40809494e-02
8.40006280e-04 -1.86153382e-01 1.05319273e+00 1.02129722e+00
-7.82861233e-01 -1.36199152e+00 -4.19981539e-01 6.57879710e-01
-2.23039821e-01 2.82321006e-01 -2.59333104e-01 1.01602352e+00
-3.83581042e-01 1.12133300e+00 1.00487754e-01 1.46822557e-01
-1.91446040e-02 -7.04409659e-01 5.71944654e-01 -4.49656993e-01
-3.99569631e-01 3.14337641e-01 -9.23166201e-02 -3.60083282e-02
-6.51897550e-01 -1.07985783e+00 -1.06880772e+00 -7.88319528e-01
-6.58368766e-01 7.00757563e-01 2.51753569e-01 1.09800792e+00
2.30835557e-01 3.41925293e-01 7.51243591e-01 -1.16556275e+00
-7.21433103e-01 -5.86810470e-01 -8.59293222e-01 -4.83248621e-01
5.41854382e-01 -8.86737049e-01 -1.96031049e-01 -3.76195908e-01] | [3.920341968536377, 2.137979030609131] |
574db743-191e-42bd-86cd-ddc3cf6791f2 | exploring-large-language-models-for-classical | 2305.13698 | null | https://arxiv.org/abs/2305.13698v1 | https://arxiv.org/pdf/2305.13698v1.pdf | Exploring Large Language Models for Classical Philology | Recent advances in NLP have led to the creation of powerful language models for many languages including Ancient Greek and Latin. While prior work on Classical languages unanimously uses BERT, in this work we create four language models for Ancient Greek that vary along two dimensions to study their versatility for tasks of interest for Classical languages: we explore (i) encoder-only and encoder-decoder architectures using RoBERTa and T5 as strong model types, and create for each of them (ii) a monolingual Ancient Greek and a multilingual instance that includes Latin and English. We evaluate all models on morphological and syntactic tasks, including lemmatization, which demonstrates the added value of T5's decoding abilities. We further define two probing tasks to investigate the knowledge acquired by models pre-trained on Classical texts. Our experiments provide the first benchmarking analysis of existing models of Ancient Greek. Results show that our models provide significant improvements over the SoTA. The systematic analysis of model types can inform future research in designing language models for Classical languages, including the development of novel generative tasks. We make all our models available as community resources, along with a large curated pre-training corpus for Ancient Greek, to support the creation of a larger, comparable model zoo for Classical Philology. Our models and resources are available at https://github.com/Heidelberg-NLP/ancient-language-models. | ['Anette Frank', 'Frederick Riemenschneider'] | 2023-05-23 | null | null | null | null | ['lemmatization'] | ['natural-language-processing'] | [-1.73643306e-01 3.13237280e-01 -1.84126824e-01 -3.82356286e-01
-1.05503380e+00 -7.66211271e-01 9.58994985e-01 -2.67660916e-02
-8.20149779e-01 6.51460350e-01 8.01083148e-01 -7.71203160e-01
4.61086221e-02 -7.83749521e-01 -4.43583667e-01 -2.68272489e-01
5.72841577e-02 8.47427845e-01 6.82831556e-02 -5.18163323e-01
7.67502412e-02 1.93210497e-01 -9.80943859e-01 4.88738269e-01
9.09843683e-01 1.80921122e-01 3.57694328e-01 4.71954137e-01
-1.58972889e-02 5.44002593e-01 -4.54193771e-01 -7.65010595e-01
1.61476076e-01 -5.38787723e-01 -9.37635243e-01 -5.37842989e-01
-1.03502031e-02 -2.15658993e-01 -3.90829414e-01 4.30969149e-01
7.72907257e-01 -1.84044212e-01 5.15103519e-01 -5.73422909e-01
-1.05264795e+00 1.43841207e+00 1.06836818e-01 3.54486167e-01
3.83876234e-01 3.09816271e-01 1.35241568e+00 -9.66590226e-01
8.95689189e-01 1.34573650e+00 6.77081406e-01 5.69216490e-01
-1.17818284e+00 -3.73194993e-01 -1.46524072e-01 3.98530841e-01
-1.32902038e+00 -7.52224028e-01 5.79623818e-01 -2.21763551e-01
1.35081208e+00 -2.02346942e-03 8.58588934e-01 1.31100512e+00
1.26169190e-01 1.01923716e+00 1.34603727e+00 -6.64928555e-01
-6.26967996e-02 1.35443136e-01 -1.53054953e-01 4.53236789e-01
4.83084694e-02 3.44542801e-01 -5.84128082e-01 4.94694039e-02
5.32850623e-01 -7.09528267e-01 -2.03694358e-01 3.32056284e-01
-1.34881341e+00 9.91612613e-01 1.63922518e-01 8.58934820e-01
-1.55814245e-01 7.12456778e-02 4.09356743e-01 3.86630028e-01
2.86019146e-01 5.18165827e-01 -6.76613390e-01 -5.60882747e-01
-1.17223012e+00 2.48429850e-01 1.14058244e+00 7.99354494e-01
5.39388120e-01 2.37384886e-01 2.37149537e-01 9.95842755e-01
4.52044457e-01 5.94842076e-01 8.71013522e-01 -7.29970098e-01
4.07798320e-01 4.15656924e-01 -5.40623307e-01 -3.59097034e-01
-3.78030509e-01 -2.46005327e-01 -2.27766171e-01 -4.10585433e-01
4.14543122e-01 -2.28985026e-02 -8.86715770e-01 1.74767184e+00
-8.67640506e-03 -3.20059776e-01 3.13099802e-01 4.59186375e-01
1.02769279e+00 7.24558413e-01 1.54711291e-01 9.94783938e-02
1.34271598e+00 -5.72524250e-01 -3.08554053e-01 -5.08953512e-01
1.00088871e+00 -1.03041351e+00 1.26944935e+00 3.12767625e-01
-1.37077427e+00 -3.68393362e-01 -7.81757116e-01 -5.46912789e-01
-4.63061959e-01 1.62952110e-01 6.67380333e-01 7.59763837e-01
-1.15099621e+00 6.60657406e-01 -9.77966726e-01 -5.94695807e-01
1.97593451e-01 2.05084518e-01 -1.13457769e-01 1.15488812e-01
-1.46680951e+00 1.35475683e+00 7.09501863e-01 -8.17966163e-02
-8.54027689e-01 -6.13770425e-01 -8.43316257e-01 -2.93154895e-01
3.28986235e-02 -6.04452491e-01 1.39805055e+00 -9.10591125e-01
-1.46057689e+00 1.23395717e+00 1.63210168e-01 -3.89216810e-01
3.78959268e-01 -2.18675211e-01 -4.91045982e-01 -1.59574598e-01
1.47769824e-01 6.38944685e-01 1.66581869e-01 -9.76899922e-01
-5.29981315e-01 -2.22051904e-01 -9.71584301e-03 5.50328381e-02
-1.34690180e-01 3.16506863e-01 -4.50154394e-01 -8.62130105e-01
-1.61167398e-01 -8.54253054e-01 6.98276758e-02 -4.93410438e-01
-1.80172265e-01 -2.87787467e-01 1.58686683e-01 -1.18604803e+00
1.47159350e+00 -1.90456247e+00 1.85564667e-01 1.05295815e-01
-2.96907067e-01 2.91703492e-01 -2.70062089e-01 9.63730574e-01
6.96091503e-02 3.67203504e-01 -5.08471727e-01 -3.13904732e-01
2.24963456e-01 7.55609274e-01 -2.44392589e-01 2.63905585e-01
2.62862444e-01 1.36527705e+00 -7.98232079e-01 -4.63905454e-01
3.81186381e-02 6.74271345e-01 -7.18353808e-01 -7.87166804e-02
-2.76273400e-01 3.92075926e-01 -1.66184202e-01 5.75172961e-01
9.13044140e-02 6.05481677e-02 4.42059636e-01 2.09889576e-01
-3.94211829e-01 1.18550324e+00 -7.07302153e-01 1.89527249e+00
-6.58985138e-01 3.98582816e-01 -1.52429909e-01 -5.16871810e-01
6.21842206e-01 3.83867413e-01 7.75012299e-02 -9.20428157e-01
3.52162629e-01 8.48663330e-01 6.24917567e-01 -3.95382464e-01
6.04882836e-01 -6.03152394e-01 -2.49800399e-01 6.52453959e-01
3.99944901e-01 -3.47143620e-01 5.92106879e-01 3.40262175e-01
1.05821264e+00 5.87809205e-01 5.91201663e-01 -5.47634125e-01
3.55115980e-01 -6.50779530e-02 3.60267073e-01 4.75775301e-01
2.25169450e-01 5.30033350e-01 1.65526137e-01 -1.43712655e-01
-1.16511834e+00 -1.21423078e+00 -4.50174958e-01 1.15745091e+00
-5.90454519e-01 -9.41227853e-01 -4.62518752e-01 -3.76605868e-01
-8.70094970e-02 1.21325374e+00 -5.99600494e-01 5.27186133e-02
-1.11675513e+00 -8.65199924e-01 1.13406551e+00 5.53511143e-01
2.17937797e-01 -1.40093327e+00 -5.21471500e-01 4.47330326e-01
-2.53409505e-01 -8.91420364e-01 -5.13481535e-02 1.54889271e-01
-6.69600189e-01 -9.24529731e-01 -6.14843965e-01 -7.92095423e-01
-1.27870381e-01 -5.86518347e-01 1.35185456e+00 1.23756707e-01
-4.20169830e-02 4.18358028e-01 -5.19484043e-01 -3.76797080e-01
-9.48001146e-01 4.33974475e-01 -1.56799391e-01 -6.98428690e-01
5.08661747e-01 -7.09530413e-01 -2.38769829e-01 -2.62204021e-01
-1.07980835e+00 -4.50114720e-02 7.22683311e-01 5.02725184e-01
1.96075842e-01 -7.04685330e-01 4.67270494e-01 -9.78107452e-01
5.87830603e-01 -6.54072225e-01 -2.53483772e-01 1.90613821e-01
-4.80895668e-01 2.42762938e-01 6.53844595e-01 -1.72579974e-01
-1.13631141e+00 -4.18563038e-01 -9.08200741e-01 4.93547410e-01
5.12305684e-02 9.10889089e-01 -6.93270341e-02 3.30263853e-01
7.78875947e-01 4.19728637e-01 -3.45539421e-01 -7.77879357e-01
8.53554249e-01 6.99299514e-01 4.59027261e-01 -9.57235157e-01
8.12827110e-01 2.55927861e-01 -3.85676503e-01 -1.27761161e+00
-4.15285856e-01 -1.92678392e-01 -8.97833049e-01 2.35803977e-01
6.38647556e-01 -1.09050524e+00 -1.53442577e-01 1.70111671e-01
-1.00149584e+00 -6.79365158e-01 -5.53039551e-01 4.93394017e-01
-5.80321074e-01 5.11779368e-01 -9.36836958e-01 -4.26960379e-01
-3.53892267e-01 -8.58658612e-01 9.85325098e-01 -1.91024914e-01
-7.72467554e-01 -1.61371815e+00 6.65263414e-01 6.95120543e-02
3.98199558e-01 -3.82665060e-02 1.31525970e+00 -8.57275903e-01
-3.89194995e-01 2.92699158e-01 2.18101397e-01 3.16503197e-01
-1.15919694e-01 -2.76913703e-03 -7.44703174e-01 -3.10062677e-01
-6.40534833e-02 -4.81121659e-01 9.52612519e-01 -5.56643084e-02
1.41234785e-01 -2.22681999e-01 -4.90896218e-02 7.50873983e-01
1.33132863e+00 8.44847187e-02 7.61893094e-01 6.18506730e-01
4.46440786e-01 3.87889177e-01 1.17886346e-02 3.07516813e-01
9.42074656e-01 4.73974824e-01 -2.81093031e-01 1.35754615e-01
-4.88179415e-01 -3.20381254e-01 7.90632784e-01 1.52645218e+00
-1.72183529e-01 -1.31124809e-01 -1.26940227e+00 1.04162729e+00
-1.47391665e+00 -6.63751781e-01 -2.70904507e-02 1.96297109e+00
1.38465381e+00 -5.68495765e-02 1.55062467e-01 -6.83398172e-02
-4.89943922e-02 1.96140721e-01 5.87577559e-02 -5.49013257e-01
-4.21872050e-01 1.04379451e+00 1.08998492e-01 8.15687537e-01
-7.35524178e-01 1.51557386e+00 6.76093340e+00 7.97278106e-01
-9.46682811e-01 2.93566942e-01 7.57325664e-02 -2.14398786e-01
-8.52261603e-01 5.05537450e-01 -9.60988581e-01 4.07867163e-01
1.44566071e+00 -3.26271355e-01 4.26840752e-01 4.40716684e-01
9.13106874e-02 7.31624439e-02 -1.26493943e+00 4.46462333e-01
1.60247132e-01 -1.12256157e+00 2.77332574e-01 1.66662447e-02
5.36900163e-01 5.93123078e-01 -1.59335703e-01 5.99752963e-01
7.55592346e-01 -1.04000771e+00 8.78286004e-01 2.08999634e-01
7.22550869e-01 -5.15731573e-01 3.27131689e-01 2.11313993e-01
-9.76206660e-01 7.81999826e-02 -2.67476678e-01 -9.92510244e-02
5.49655437e-01 2.07855552e-01 -8.49074066e-01 6.08949542e-01
4.00398791e-01 8.85370851e-01 -1.01030219e+00 5.91747642e-01
-8.17580581e-01 1.13173425e+00 -7.01770008e-01 -1.93063878e-02
4.32663739e-01 -1.48582518e-01 6.03432357e-01 1.54735994e+00
3.43251824e-01 8.49750340e-02 7.03142062e-02 8.73600006e-01
1.51541889e-01 6.33584082e-01 -4.53339547e-01 -2.81019181e-01
3.86872828e-01 1.02136719e+00 -6.41712785e-01 -2.38574058e-01
-6.38899505e-01 7.72684932e-01 4.91811603e-01 2.11644515e-01
-6.31450713e-01 -1.19189188e-01 1.30013689e-01 4.12145734e-01
2.22821251e-01 -6.10977232e-01 -3.32057446e-01 -1.28079903e+00
-8.08064342e-02 -1.09728360e+00 5.51288128e-01 -5.28002977e-01
-1.27938652e+00 7.03274667e-01 1.76996842e-01 -4.84977007e-01
-5.44593811e-01 -7.65301704e-01 -3.90158027e-01 8.01512241e-01
-1.47275627e+00 -1.72142994e+00 3.96695793e-01 5.22619188e-01
4.01701599e-01 -3.21899414e-01 9.25849557e-01 4.09077853e-01
-2.19664872e-01 4.82163608e-01 2.11820394e-01 3.55020195e-01
7.85809338e-01 -1.29078496e+00 9.58228946e-01 9.17308271e-01
6.77097023e-01 8.34270477e-01 3.81338835e-01 -7.94031739e-01
-1.01382315e+00 -7.48352170e-01 1.40590990e+00 -7.16669083e-01
1.07962477e+00 -4.66230065e-01 -9.04174566e-01 1.30434072e+00
6.01965129e-01 -1.01537764e+00 9.39017713e-01 2.29407623e-01
-2.18074158e-01 4.04915273e-01 -4.84086424e-01 8.58408630e-01
1.12455344e+00 -5.13985932e-01 -1.11646175e+00 2.26588920e-01
4.13122445e-01 -7.04135373e-02 -1.05663216e+00 1.73840418e-01
5.94939530e-01 -9.59705114e-01 7.03960180e-01 -4.14892912e-01
5.80488920e-01 4.05900069e-02 -9.99719426e-02 -1.47405624e+00
-2.90003836e-01 -6.76886559e-01 2.08310187e-01 1.20774841e+00
6.85728669e-01 -7.37439394e-01 3.32967490e-01 2.25399703e-01
-4.11538064e-01 -5.73473692e-01 -1.16840065e+00 -7.63140202e-01
9.62973475e-01 -6.93985462e-01 2.61434674e-01 7.88124621e-01
1.56411991e-01 7.33499944e-01 -1.15165208e-03 -5.51135898e-01
1.64717153e-01 -7.56202191e-02 5.76875806e-01 -1.03605294e+00
-6.56332433e-01 -5.79199910e-01 -1.63326845e-01 -9.77976382e-01
2.39814356e-01 -1.62113428e+00 -9.17671621e-02 -1.65746772e+00
-6.67655542e-02 -4.53395903e-01 2.06905723e-01 8.37736309e-01
-2.53171641e-02 4.27873969e-01 2.65784025e-01 2.97531933e-01
-1.10220730e-01 7.49109507e-01 7.35609531e-01 4.17180389e-01
-2.54626900e-01 -3.91616285e-01 -7.83509791e-01 7.47722745e-01
9.33140397e-01 -5.52317917e-01 -5.09169698e-02 -7.86906600e-01
6.01122200e-01 -4.11668241e-01 1.36364758e-01 -8.34157288e-01
-7.91730210e-02 1.14318766e-01 1.91718549e-01 -2.61247009e-01
3.05032939e-01 -2.92112648e-01 5.05708791e-02 5.82381606e-01
-6.32524583e-03 2.73676753e-01 3.28884572e-01 -1.95754558e-01
-5.57671674e-02 -3.11738938e-01 7.49673426e-01 -4.93305087e-01
-7.51269400e-01 -6.65108263e-02 -5.71342766e-01 6.01235747e-01
6.57466114e-01 -8.04469883e-02 -2.42468104e-01 -3.28131258e-01
-7.97200203e-01 6.01567850e-02 5.52143633e-01 4.36721444e-01
3.46332580e-01 -1.16564417e+00 -1.22023523e+00 2.59468198e-01
5.92111014e-02 -4.23441678e-01 -2.62095690e-01 7.23239481e-01
-7.49811471e-01 5.00215292e-01 -1.66373804e-01 -2.07823098e-01
-9.33097422e-01 2.94322997e-01 1.27919972e-01 -4.31539655e-01
-6.82356656e-01 6.94025517e-01 -9.03337374e-02 -7.73947716e-01
-3.11888158e-01 -3.27332854e-01 -2.50864457e-02 4.96799126e-02
8.63412395e-02 3.30928147e-01 -1.38832062e-01 -1.01748252e+00
-4.42374468e-01 3.16999465e-01 -5.38354255e-02 -7.06893504e-01
1.54364789e+00 -1.39083350e-02 -2.90760756e-01 8.07950556e-01
9.07140255e-01 4.21965957e-01 -5.14653265e-01 -2.24856868e-01
3.42794538e-01 7.04563633e-02 -3.41428965e-01 -1.11240041e+00
-5.02937078e-01 9.32228506e-01 1.62317064e-02 -2.29216725e-01
8.33045602e-01 2.09262460e-01 8.26115847e-01 3.42170209e-01
1.83646455e-01 -1.23186755e+00 -2.10532799e-01 1.14369047e+00
8.92728388e-01 -7.00450361e-01 3.44776548e-02 4.35153656e-02
-7.09499061e-01 1.08500981e+00 1.57818332e-01 -2.09941998e-01
7.69941032e-01 5.05584717e-01 3.58424455e-01 -2.91498214e-01
-7.73834705e-01 -4.96262282e-01 3.15398306e-01 4.07233357e-01
1.08293831e+00 8.21333602e-02 -9.60013211e-01 8.07070553e-01
-1.13568664e+00 -1.38773277e-01 4.45330590e-01 9.86541629e-01
-2.20994994e-01 -1.85712421e+00 -1.60213351e-01 1.32691473e-01
-5.97620547e-01 -8.36129487e-01 -7.05549538e-01 1.29827213e+00
2.95711279e-01 7.01201320e-01 -1.56700443e-02 -9.82252508e-02
2.54765242e-01 5.45957088e-01 7.97611594e-01 -9.47642088e-01
-9.97442544e-01 2.61115909e-01 4.03290808e-01 -2.05900177e-01
-3.70729893e-01 -8.83834600e-01 -1.25288498e+00 -4.73745257e-01
4.10597809e-02 8.57039094e-02 3.13315958e-01 1.09934783e+00
6.16804212e-02 3.07189524e-01 -1.14188746e-01 -6.73060775e-01
-4.69482005e-01 -1.26916957e+00 -4.78393227e-01 2.41773874e-01
-3.01076710e-01 -1.74606532e-01 -3.13463598e-01 1.60612345e-01] | [10.856406211853027, 9.913775444030762] |
073fc28c-3474-4b83-b6b9-0355cc360c9d | pvnet-a-joint-convolutional-network-of-point | 1808.07659 | null | http://arxiv.org/abs/1808.07659v1 | http://arxiv.org/pdf/1808.07659v1.pdf | PVNet: A Joint Convolutional Network of Point Cloud and Multi-View for 3D Shape Recognition | 3D object recognition has attracted wide research attention in the field of
multimedia and computer vision. With the recent proliferation of deep learning,
various deep models with different representations have achieved the
state-of-the-art performance. Among them, point cloud and multi-view based 3D
shape representations are promising recently, and their corresponding deep
models have shown significant performance on 3D shape recognition. However,
there is little effort concentrating point cloud data and multi-view data for
3D shape representation, which is, in our consideration, beneficial and
compensated to each other. In this paper, we propose the Point-View Network
(PVNet), the first framework integrating both the point cloud and the
multi-view data towards joint 3D shape recognition. More specifically, an
embedding attention fusion scheme is proposed that could employ high-level
features from the multi-view data to model the intrinsic correlation and
discriminability of different structure features from the point cloud data. In
particular, the discriminative descriptions are quantified and leveraged as the
soft attention mask to further refine the structure feature of the 3D shape. We
have evaluated the proposed method on the ModelNet40 dataset for 3D shape
classification and retrieval tasks. Experimental results and comparisons with
state-of-the-art methods demonstrate that our framework can achieve superior
performance. | ['Yue Gao', 'Yifan Feng', 'Rongrong Ji', 'Haoxuan You'] | 2018-08-23 | null | null | null | null | ['3d-shape-retrieval', '3d-shape-recognition', '3d-object-recognition', '3d-shape-representation'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [-3.67626965e-01 -4.64122832e-01 2.61418819e-02 -4.80054528e-01
-6.66172504e-01 -4.41554040e-01 8.02077830e-01 -5.01628146e-02
6.57572150e-02 -1.62275061e-01 1.55621529e-01 9.98636708e-02
-2.10512325e-01 -7.70010769e-01 -5.74928761e-01 -7.98708200e-01
3.40428382e-01 5.53597510e-01 1.34176165e-01 -1.08179957e-01
1.88460469e-01 1.12725484e+00 -1.58054245e+00 3.41417879e-01
4.34554994e-01 1.43921554e+00 2.31861681e-01 2.50141453e-02
-2.10478351e-01 -6.23695068e-02 -2.02389956e-01 -3.05719703e-01
4.32989061e-01 4.01734322e-01 -2.06932455e-01 3.28202069e-01
6.66716635e-01 -5.28924406e-01 -6.87778294e-01 9.52001750e-01
8.92933786e-01 -1.86224788e-01 7.10831285e-01 -1.08972514e+00
-1.21157706e+00 -1.62536770e-01 -9.01338875e-01 9.48685929e-02
2.38640413e-01 1.49663851e-01 1.05065846e+00 -1.59381258e+00
4.49203908e-01 1.58286369e+00 3.33826482e-01 2.60414481e-01
-9.05453682e-01 -5.68470478e-01 3.46464664e-01 2.39246041e-01
-1.44743276e+00 -2.95392305e-01 1.25772905e+00 -3.23884487e-01
1.25144339e+00 4.45970111e-02 8.27420533e-01 9.46292579e-01
6.59893751e-02 1.27640355e+00 7.90295959e-01 -1.20317861e-02
-2.13361979e-01 -1.01531647e-01 -3.12980980e-01 6.55355632e-01
2.52609730e-01 2.48145998e-01 -2.76811779e-01 -2.04256952e-01
1.04813099e+00 6.12989128e-01 -1.95425883e-01 -8.95773649e-01
-1.05172586e+00 6.18259132e-01 7.64447689e-01 3.16194594e-01
-6.16346002e-01 -1.90876126e-02 2.96839654e-01 4.36689984e-03
8.87000084e-01 -1.62689000e-01 -3.56229782e-01 1.44216627e-01
-6.26616716e-01 1.73712507e-01 2.50556022e-01 1.05699253e+00
5.05494654e-01 1.95791185e-01 -8.49947333e-02 1.00316310e+00
8.20626974e-01 8.77724648e-01 5.66952452e-02 -5.35265148e-01
6.41489387e-01 1.25695658e+00 -2.10426673e-01 -1.34980965e+00
-2.54143596e-01 -6.70910001e-01 -1.03359222e+00 6.53789863e-02
-7.43034557e-02 5.90219200e-01 -9.64201212e-01 1.21686780e+00
3.94278646e-01 2.93240935e-01 -1.26201749e-01 1.19338107e+00
1.33587754e+00 5.58729470e-01 -3.10293883e-01 2.76794672e-01
1.20435238e+00 -5.87300181e-01 -2.36848146e-01 6.06923401e-02
2.90516794e-01 -7.56327510e-01 6.51033282e-01 -4.66271341e-02
-1.36782980e+00 -8.18587482e-01 -1.04153931e+00 -1.63047180e-01
-3.21386546e-01 1.50146484e-01 4.30762082e-01 3.66935253e-01
-9.11464691e-01 2.97086567e-01 -9.30198789e-01 -1.75576106e-01
9.02291358e-01 2.63387531e-01 -4.75293994e-01 -4.48498428e-01
-7.88021564e-01 6.31320119e-01 1.17230779e-02 3.22443277e-01
-7.87831187e-01 -6.73041582e-01 -8.86099100e-01 3.50776434e-01
1.23024553e-01 -9.45097506e-01 7.33619273e-01 -1.66151509e-01
-1.06383288e+00 1.18393683e+00 -1.05755836e-01 9.18447003e-02
1.48700595e-01 3.51469740e-02 -2.93049484e-01 2.69641608e-01
-8.93184822e-03 6.51787817e-01 8.45379472e-01 -1.60569394e+00
-4.15734649e-01 -1.00506568e+00 5.47889583e-02 3.80832195e-01
-2.30426311e-01 -2.56723106e-01 -7.46351004e-01 -4.78466213e-01
8.11442554e-01 -7.44925976e-01 9.38452557e-02 2.46039271e-01
-1.42444015e-01 -6.27107441e-01 1.01086450e+00 -2.36238152e-01
5.19224703e-01 -2.38174534e+00 2.81861395e-01 1.10924274e-01
3.24414253e-01 3.77077460e-01 -4.05903965e-01 2.93489754e-01
3.71912308e-02 6.51757121e-02 2.31727827e-02 -4.85931098e-01
1.20089412e-01 2.31156766e-01 -3.48182261e-01 6.18248165e-01
4.13129717e-01 1.32983625e+00 -7.47490287e-01 -1.80507421e-01
4.35584068e-01 9.64788914e-01 -5.20321012e-01 2.30711624e-01
7.87611604e-02 3.51267666e-01 -8.48321080e-01 1.03124559e+00
1.02687430e+00 -5.09116769e-01 -4.02513713e-01 -5.82429588e-01
7.17641711e-02 -5.09189849e-04 -8.53346944e-01 1.88410378e+00
-3.86195630e-01 9.74025875e-02 4.78843637e-02 -9.72793698e-01
1.49591100e+00 2.34483391e-01 7.51321614e-01 -7.10965216e-01
8.82100165e-02 3.83020908e-01 -8.03485215e-02 -3.59842837e-01
2.12664023e-01 -2.60578599e-02 2.56432682e-01 2.39105567e-01
1.35385334e-01 -4.31676179e-01 -5.30628860e-01 -1.90041661e-01
5.20567417e-01 2.50809759e-01 1.61535844e-01 1.90074027e-01
9.34687793e-01 -4.57205176e-01 3.17434609e-01 3.20424169e-01
-2.19896853e-01 8.54090631e-01 8.62174779e-02 -6.81286752e-01
-9.84030545e-01 -1.02096903e+00 -1.85421854e-01 4.70219165e-01
3.88509095e-01 -7.55821541e-02 -1.04225330e-01 -8.12870741e-01
4.65144753e-01 3.29668969e-01 -3.46021831e-01 -2.17041716e-01
-6.11293733e-01 -2.50048995e-01 3.68936986e-01 6.55039787e-01
6.75948024e-01 -8.34955275e-01 -4.30867910e-01 7.93165788e-02
2.86464274e-01 -1.27068210e+00 -5.53323209e-01 -2.78720170e-01
-1.15565336e+00 -9.24829543e-01 -1.04316270e+00 -8.00421119e-01
6.92133605e-01 9.66850817e-01 8.89817238e-01 2.08580971e-01
3.80161852e-02 1.01098359e+00 -4.47453141e-01 -4.29659009e-01
5.34290411e-02 -2.21043348e-01 1.20252818e-02 3.80502820e-01
5.48509419e-01 -8.05916190e-01 -7.28175998e-01 3.52714181e-01
-8.80454421e-01 -1.89405471e-01 7.95433640e-01 9.57970500e-01
1.03678155e+00 -4.35286582e-01 4.10946965e-01 -2.70485282e-01
3.08654964e-01 -3.77314895e-01 -4.72061217e-01 2.29587466e-01
-3.32378089e-01 -2.57651180e-01 4.10946012e-01 -1.21842334e-02
-7.56778121e-01 3.33411470e-02 -4.42600965e-01 -1.31922889e+00
-4.74615991e-01 5.50791085e-01 -6.06869221e-01 -3.34291339e-01
2.01322418e-02 7.87639499e-01 1.84543207e-01 -7.74480343e-01
3.37139487e-01 6.81849778e-01 3.78749222e-02 -3.56680959e-01
7.60364294e-01 8.01164567e-01 2.09449559e-01 -7.98789322e-01
-8.20000887e-01 -6.15996361e-01 -8.09565067e-01 -1.86646402e-01
7.15444744e-01 -1.08332181e+00 -7.73186803e-01 5.83177209e-01
-1.38214457e+00 4.60753083e-01 -7.71372169e-02 3.07213753e-01
-6.85830295e-01 6.31400883e-01 -2.08230510e-01 -7.18050599e-01
-4.71245348e-01 -1.35615909e+00 1.71546292e+00 1.67996839e-01
5.17766178e-01 -8.93897414e-01 -2.59773254e-01 3.76095235e-01
2.93574512e-01 1.08324155e-01 1.07688689e+00 -8.42410088e-01
-9.68218148e-01 -3.97614896e-01 -7.18514442e-01 2.59920597e-01
2.33990639e-01 -3.99958104e-01 -1.02770448e+00 -4.97781128e-01
2.61090815e-01 -1.52518645e-01 9.52153981e-01 4.52305347e-01
1.40264940e+00 1.84864327e-01 -3.83632928e-01 8.53275299e-01
1.26002669e+00 2.36739427e-01 3.71262819e-01 -5.76507039e-02
9.71922219e-01 3.52624953e-01 3.61245185e-01 5.13333321e-01
5.10058284e-01 9.03606832e-01 8.44449222e-01 5.49806952e-02
-1.19519919e-01 -4.24101084e-01 -9.59720910e-02 1.21235847e+00
-2.17820480e-01 -1.20848399e-02 -8.58586490e-01 4.65790719e-01
-1.63764238e+00 -8.74391139e-01 -1.19087454e-02 1.88536370e+00
-1.27432737e-02 -1.79853532e-02 -2.03120664e-01 -5.77946790e-02
4.91448492e-01 6.18143260e-01 -8.06185663e-01 3.01446468e-02
-2.88874418e-01 -3.26816440e-02 -3.94606628e-02 5.02009084e-03
-1.11979520e+00 5.06757557e-01 4.92219257e+00 8.90951812e-01
-1.25637674e+00 -1.51532754e-01 5.48150361e-01 1.80680126e-01
-5.52643657e-01 -3.59444588e-01 -8.11895132e-01 1.36571556e-01
1.05259769e-01 6.31640255e-02 4.77198474e-02 8.19144964e-01
-9.93502606e-03 5.85590839e-01 -1.16136682e+00 1.51334476e+00
3.68625194e-01 -1.32259274e+00 5.85998356e-01 3.22438955e-01
5.79432726e-01 2.83384949e-01 3.17169458e-01 2.23169506e-01
-4.78207618e-01 -7.72561073e-01 5.52105606e-01 6.65490031e-01
8.01140070e-01 -6.22245252e-01 6.74153864e-01 5.57657897e-01
-1.40741313e+00 2.10620556e-02 -6.24346614e-01 2.90671706e-01
9.93642658e-02 4.94819820e-01 -3.99371922e-01 1.00493848e+00
5.48565209e-01 1.44133997e+00 -4.89563912e-01 1.20517337e+00
1.20766640e-01 5.09512797e-02 -2.04756528e-01 -1.65772792e-02
4.85391617e-01 -3.53693306e-01 8.90642941e-01 6.35673583e-01
6.88974380e-01 3.10909688e-01 2.74115086e-01 1.28084922e+00
-1.65072009e-02 4.95826602e-02 -9.86281395e-01 5.74520938e-02
3.07374030e-01 1.17220032e+00 -5.10897636e-01 -2.49165744e-01
-8.59267831e-01 5.42890370e-01 2.50254691e-01 3.10042709e-01
-3.61307085e-01 9.08908099e-02 8.09449971e-01 -1.00860922e-02
8.25941801e-01 -5.11987507e-01 -3.41236621e-01 -1.25367701e+00
1.10419795e-01 -5.81179261e-01 1.75405607e-01 -7.91735947e-01
-1.81646156e+00 5.65984786e-01 -3.15608718e-02 -1.59536743e+00
2.69574404e-01 -8.14260066e-01 -4.28681463e-01 1.14054501e+00
-1.79462838e+00 -1.38329375e+00 -3.28083515e-01 5.08453190e-01
5.88567197e-01 -5.54197967e-01 7.73463190e-01 4.28851634e-01
-1.39649451e-01 5.31716347e-01 -8.41411669e-03 1.18485987e-01
3.06896240e-01 -8.42417955e-01 6.71156824e-01 2.70822197e-01
4.45397288e-01 4.37848926e-01 -1.40962079e-01 -5.61906695e-01
-1.94875109e+00 -1.05380762e+00 6.42051935e-01 -4.77617383e-01
1.99347779e-01 -2.53621817e-01 -1.13131857e+00 2.41981998e-01
-1.72546729e-01 3.90469074e-01 4.75266844e-01 -2.05986902e-01
-5.54572165e-01 -1.80591941e-01 -9.18396473e-01 2.16705725e-01
1.26873624e+00 -6.51666284e-01 -6.06593251e-01 3.03088650e-02
6.25803649e-01 -6.54021382e-01 -9.17936563e-01 7.87683070e-01
5.02476335e-01 -1.11340761e+00 1.48859060e+00 -5.81966281e-01
4.03861433e-01 -2.87004232e-01 -6.06090963e-01 -1.32493079e+00
-5.40018797e-01 1.40258491e-01 -3.34118873e-01 1.01491451e+00
-1.91968642e-02 -5.04435241e-01 8.61415744e-01 2.59881288e-01
-5.00301242e-01 -1.35881770e+00 -1.19034004e+00 -7.28055537e-01
1.58748686e-01 -4.64852303e-01 9.46000159e-01 6.05011761e-01
-6.66401207e-01 3.10002893e-01 -8.04407075e-02 3.27577651e-01
5.60060799e-01 8.39536488e-01 6.57457173e-01 -1.43768954e+00
-2.18099356e-03 -7.43489623e-01 -8.64377558e-01 -1.71277046e+00
2.55885452e-01 -1.45548618e+00 -5.30172527e-01 -1.73753285e+00
2.67476708e-01 -3.85117501e-01 -3.97909045e-01 2.02699691e-01
-3.46535482e-02 1.05833799e-01 5.40555179e-01 2.38558054e-01
-4.92454588e-01 1.22920263e+00 1.88145781e+00 -6.00580215e-01
-8.89881477e-02 1.14285521e-01 -5.55998266e-01 5.74883223e-01
2.65689909e-01 -6.34405240e-02 -1.92905217e-01 -8.63031507e-01
3.38144861e-02 1.59175426e-01 7.61858702e-01 -6.19607747e-01
3.18704486e-01 2.21518174e-01 6.45374894e-01 -1.23281813e+00
8.40110779e-01 -1.25637197e+00 -1.85574204e-01 2.33857796e-01
6.57541677e-02 -8.37764442e-02 2.00319067e-01 8.59427869e-01
-5.46915829e-01 1.75604016e-01 5.04599392e-01 -2.25718841e-01
-6.02593601e-01 1.03363609e+00 3.74399990e-01 -2.59927303e-01
7.31199384e-01 -4.54968780e-01 3.00783515e-02 -2.87738144e-01
-6.35293067e-01 2.06161007e-01 2.22312212e-01 7.79823601e-01
1.31645775e+00 -1.94407845e+00 -8.57926548e-01 7.00099111e-01
3.09853584e-01 1.36785790e-01 5.05588233e-01 7.41886020e-01
1.81953255e-02 7.07735717e-01 -2.14986891e-01 -1.12386417e+00
-9.98130322e-01 5.41451097e-01 3.95348191e-01 9.09534097e-02
-7.67198682e-01 7.11959302e-01 5.34049511e-01 -6.49126470e-01
2.45046064e-01 -3.27702552e-01 -3.17888498e-01 -3.44606228e-02
2.77815163e-01 1.35472789e-01 1.67470098e-01 -9.97179687e-01
-5.22821724e-01 1.22920644e+00 1.13609172e-02 4.49224263e-01
1.78826714e+00 -6.25980720e-02 -5.17934524e-02 3.61190557e-01
1.54979169e+00 -3.68992180e-01 -1.24091899e+00 -6.02320015e-01
-2.09944010e-01 -8.25280070e-01 1.10611066e-01 -4.37905282e-01
-1.34818459e+00 1.31487083e+00 5.82269549e-01 2.15428784e-01
1.04758453e+00 1.41411901e-01 9.18955743e-01 2.91597486e-01
5.56248069e-01 -4.60927784e-01 1.09068759e-01 8.68808925e-01
1.33060193e+00 -1.40459669e+00 3.47361192e-02 -3.79453093e-01
-2.63109416e-01 1.19454384e+00 5.71296632e-01 -3.94773960e-01
1.05893910e+00 -2.93973356e-01 -1.79073244e-01 -6.25383258e-01
-7.20242143e-01 -2.46021509e-01 9.12014306e-01 6.39117181e-01
2.60566652e-01 -2.86325533e-02 2.06500009e-01 7.97451854e-01
3.58184636e-01 -3.59942108e-01 -1.21093392e-01 4.71182764e-01
-2.72773296e-01 -9.72995639e-01 -3.08848828e-01 5.70134223e-01
-7.16218278e-02 1.76856071e-01 -5.56970298e-01 6.69667363e-01
-6.53454736e-02 5.14261365e-01 1.25208482e-01 -4.37460065e-01
8.04264367e-01 9.41352248e-02 5.48975706e-01 -5.07109642e-01
-3.24705690e-01 3.71006936e-01 -4.34205920e-01 -5.39977789e-01
-5.89792252e-01 -6.39145315e-01 -8.91333520e-01 -4.47837412e-02
-1.99607134e-01 -4.17173266e-01 5.77483773e-01 8.11161399e-01
9.17423248e-01 1.43712610e-01 1.02338088e+00 -1.34580696e+00
-7.30363548e-01 -6.20669723e-01 -5.95506907e-01 4.29315835e-01
2.31616423e-01 -1.07318449e+00 -1.75653890e-01 -5.26701748e-01] | [8.149139404296875, -3.850801706314087] |
999c5c79-795d-4af7-974d-7902dbabd790 | precise-aerial-image-matching-based-on-deep | 2107.08768 | null | https://arxiv.org/abs/2107.08768v1 | https://arxiv.org/pdf/2107.08768v1.pdf | Precise Aerial Image Matching based on Deep Homography Estimation | Aerial image registration or matching is a geometric process of aligning two aerial images captured in different environments. Estimating the precise transformation parameters is hindered by various environments such as time, weather, and viewpoints. The characteristics of the aerial images are mainly composed of a straight line owing to building and road. Therefore, the straight lines are distorted when estimating homography parameters directly between two images. In this paper, we propose a deep homography alignment network to precisely match two aerial images by progressively estimating the various transformation parameters. The proposed network is possible to train the matching network with a higher degree of freedom by progressively analyzing the transformation parameters. The precision matching performances have been increased by applying homography transformation. In addition, we introduce a method that can effectively learn the difficult-to-learn homography estimation network. Since there is no published learning data for aerial image registration, in this paper, a pair of images to which random homography transformation is applied within a certain range is used for learning. Hence, we could confirm that the deep homography alignment network shows high precision matching performance compared with conventional works. | ['Seong-Whan Lee', 'Yong-Ju Lee', 'Myeong-Seok Oh'] | 2021-07-19 | null | null | null | null | ['homography-estimation'] | ['computer-vision'] | [ 2.89961070e-01 -3.70366991e-01 8.22170451e-02 -2.97587365e-01
-2.20856413e-01 -5.88550925e-01 4.09990907e-01 -4.48961884e-01
-2.92283654e-01 3.03913146e-01 -2.60390848e-01 -1.96927600e-02
-3.03794652e-01 -7.60286689e-01 -8.30353677e-01 -6.64003968e-01
2.45123193e-01 3.58354926e-01 8.40997621e-02 -4.24387187e-01
1.08645871e-01 5.43633521e-01 -1.20685542e+00 -4.43382889e-01
9.20424581e-01 7.42845714e-01 3.04156572e-01 3.77755076e-01
3.26654732e-01 1.84872508e-01 -4.38978642e-01 -3.15461129e-01
1.04064286e+00 -4.83728319e-01 -6.82066500e-01 2.69731253e-01
1.01611638e+00 -8.93676043e-01 -7.61999011e-01 1.27722025e+00
3.96431118e-01 4.93183322e-02 3.62112224e-01 -1.33573163e+00
-6.65187359e-01 4.21359599e-01 -6.16050124e-01 -2.76305050e-01
4.90361392e-01 1.09316915e-01 8.22902262e-01 -2.74015456e-01
4.90640014e-01 9.77277219e-01 8.95443976e-01 -3.10210753e-02
-9.67374980e-01 -7.31675386e-01 -2.61652209e-02 2.22771451e-01
-1.57602048e+00 -7.48930573e-02 9.44317162e-01 -4.83652771e-01
4.87739831e-01 8.40202495e-02 8.47468317e-01 5.30037522e-01
3.73146921e-01 2.64303982e-01 6.17063642e-01 -3.15400660e-01
-3.99775207e-01 -3.03035378e-01 -3.84300470e-01 7.59113312e-01
1.71159506e-01 3.58774155e-01 1.13102600e-01 1.63962513e-01
1.16593492e+00 4.57250893e-01 -5.96001208e-01 -6.39253259e-01
-1.50981498e+00 4.16238159e-01 7.82777011e-01 3.22068512e-01
-2.11864546e-01 5.02138212e-02 1.62216842e-01 4.85724956e-01
-1.85648918e-01 7.90602922e-01 -1.98805138e-01 3.45978379e-01
-1.00454593e+00 8.91262572e-03 4.71120238e-01 1.42468584e+00
9.92355764e-01 1.58495441e-01 3.05763245e-01 5.31826258e-01
3.13571155e-01 8.23787391e-01 5.67330599e-01 -8.48781645e-01
6.41340375e-01 6.37112141e-01 1.27776742e-01 -1.80294347e+00
-4.89359558e-01 -1.18897960e-01 -1.37419796e+00 -8.72078910e-02
4.45394695e-01 -1.66615605e-01 -4.77105498e-01 1.42247975e+00
1.96976706e-01 4.47258830e-01 6.70713112e-02 1.07503355e+00
6.22520745e-01 7.27775574e-01 -4.11390334e-01 6.28870204e-02
1.05074632e+00 -9.61224556e-01 -7.45965838e-01 -2.65758455e-01
4.60847139e-01 -1.21470451e+00 8.74528766e-01 7.15403035e-02
-8.73502672e-01 -9.27944958e-01 -1.45623410e+00 -1.23588331e-01
-2.18971118e-01 5.93869627e-01 5.29754519e-01 7.89942145e-02
-9.39673483e-01 9.31898654e-01 -6.56665206e-01 -3.93915564e-01
-2.06897650e-02 6.66208565e-01 -6.37809217e-01 5.16993552e-02
-1.29294479e+00 8.89705300e-01 3.90371621e-01 4.73911434e-01
-4.75731879e-01 -5.03661335e-01 -9.04619992e-01 -2.25880183e-02
1.32264048e-01 -8.30527008e-01 8.85772645e-01 -7.90759683e-01
-1.60657477e+00 7.38471389e-01 2.62883782e-01 -1.42880723e-01
6.76570177e-01 -1.93045706e-01 -4.30881888e-01 -1.83851067e-02
6.10526502e-02 5.88314831e-01 9.68072534e-01 -9.97361064e-01
-5.51317513e-01 -3.99328858e-01 1.89958334e-01 3.11011076e-01
-9.85993147e-02 -3.34725976e-01 -2.62679338e-01 -3.30864221e-01
7.07088649e-01 -1.21987355e+00 -8.18371177e-02 1.01195440e-01
-3.91040295e-01 4.19560671e-01 7.97048450e-01 -8.55387807e-01
9.93888736e-01 -1.91973889e+00 1.97996452e-01 2.46611983e-01
7.49382228e-02 2.99751222e-01 -2.85630077e-01 3.31310630e-01
-2.20755190e-01 -1.58270508e-01 -1.45426080e-01 1.75522715e-01
-4.81312051e-02 -7.30093792e-02 -3.83106321e-01 7.67660797e-01
-2.42132515e-01 7.94155538e-01 -8.45822096e-01 -3.53040546e-01
4.66354579e-01 5.04838645e-01 -2.31501833e-01 4.70314324e-01
3.87883514e-01 4.98865932e-01 -4.19944912e-01 4.36186850e-01
9.61429000e-01 -1.00450823e-02 1.49931371e-01 -1.18157375e+00
-2.34549195e-01 -3.51323366e-01 -1.23153877e+00 1.67526472e+00
-5.19868553e-01 8.44707310e-01 -1.78024054e-01 -8.25331450e-01
1.33721602e+00 2.40458712e-01 6.36403799e-01 -3.56415927e-01
5.80402672e-01 -4.95246984e-02 7.73933083e-02 -4.16632801e-01
5.92006505e-01 1.16692811e-01 2.05441728e-01 2.02719554e-01
-4.95259389e-02 -8.47458184e-01 1.06284596e-01 -3.17378432e-01
4.09921348e-01 2.96220511e-01 4.83127028e-01 -3.46450172e-02
6.96024537e-01 -1.36957049e-01 4.23895597e-01 3.20297182e-01
1.78872813e-02 5.92228353e-01 1.26715615e-01 -9.95429277e-01
-1.36585975e+00 -6.93001211e-01 -1.41913518e-01 2.03598097e-01
7.17180014e-01 -2.21739307e-01 -6.47937834e-01 -3.26144956e-02
-1.81748256e-01 -5.22211306e-02 -2.57522076e-01 -3.88899297e-01
-6.69409335e-01 -4.50699270e-01 5.07124305e-01 4.10216451e-01
1.25117171e+00 -5.94282031e-01 -3.04387808e-01 2.99380347e-02
-3.09317648e-01 -1.35099649e+00 -8.58418941e-01 -3.32908839e-01
-1.08406925e+00 -1.29446054e+00 -6.14855945e-01 -1.04627633e+00
1.05245793e+00 6.86810076e-01 7.89344251e-01 2.67146260e-01
3.26227359e-02 -4.29900065e-02 5.22939041e-02 1.16869375e-01
-2.15366676e-01 2.51450032e-01 1.72604159e-01 -1.68271840e-01
2.41314143e-01 -6.71149373e-01 -4.17381018e-01 8.93589497e-01
-8.12436581e-01 1.42600983e-01 5.55792511e-01 7.83167124e-01
6.10350788e-01 3.66115570e-01 -3.36079329e-01 -3.45518619e-01
3.52342904e-01 1.21617898e-01 -1.30735886e+00 5.00934660e-01
-3.25672179e-01 -1.74764246e-01 9.31074798e-01 -5.31477928e-01
-5.29852509e-01 4.43421036e-01 1.46533519e-01 -5.02470493e-01
-2.98567265e-01 2.48031929e-01 -2.70945609e-01 -7.16566205e-01
3.34751487e-01 1.73991457e-01 4.70592603e-02 9.52076316e-02
1.42691344e-01 3.87378901e-01 7.55315304e-01 -2.95456946e-01
1.49682033e+00 2.75600582e-01 2.96265304e-01 -8.21572185e-01
-6.23016775e-01 -3.54261577e-01 -1.38136065e+00 -2.17067435e-01
8.17194462e-01 -1.03323674e+00 -9.66284394e-01 7.94974387e-01
-1.25979054e+00 -2.12715149e-01 1.76628709e-01 8.07023346e-01
-4.26419169e-01 7.32326686e-01 -3.80966872e-01 4.50613759e-02
-2.46219844e-01 -1.43984056e+00 1.00088906e+00 3.86007279e-01
-3.97166423e-02 -1.04422998e+00 9.95718166e-02 1.27865359e-01
2.18936116e-01 2.55343735e-01 4.87735808e-01 -1.24670938e-02
-8.81023288e-01 -4.47494924e-01 -1.67331830e-01 1.70940086e-01
5.79541922e-01 4.81197625e-01 -5.13315678e-01 -3.93812776e-01
2.42759120e-02 6.36633262e-02 2.14140192e-01 3.57786685e-01
9.14576471e-01 -1.68075740e-01 -1.37581646e-01 1.40534198e+00
1.43408012e+00 4.80009735e-01 8.10192287e-01 7.05638945e-01
1.07694674e+00 4.25827295e-01 6.85638130e-01 1.91225752e-01
2.57278234e-01 8.16837013e-01 4.38162029e-01 -4.51126456e-01
3.15402210e-01 -4.76387858e-01 7.04608858e-02 1.00447989e+00
-3.56013566e-01 -6.30676523e-02 -7.35624015e-01 1.01136692e-01
-1.64462721e+00 -9.16387320e-01 -2.62274563e-01 2.24057269e+00
4.95497048e-01 -2.51950443e-01 -4.51775581e-01 -4.35791817e-03
8.53005767e-01 2.60257274e-01 -5.25964677e-01 2.12951988e-01
-1.81975275e-01 -5.05160928e-01 8.59188974e-01 6.27369821e-01
-1.24133027e+00 1.05049551e+00 6.02445936e+00 2.70519763e-01
-1.29050326e+00 -7.44764388e-01 -6.89259768e-02 5.99844217e-01
3.92722823e-02 2.66307384e-01 -7.96377480e-01 3.40257019e-01
1.04237482e-01 -1.31326124e-01 6.36487246e-01 7.30363488e-01
-7.06924126e-02 2.45277017e-01 -1.00307906e+00 1.23894441e+00
1.37501061e-01 -1.05189252e+00 2.17543319e-01 -1.13521889e-02
9.86585438e-01 -3.43188107e-01 -3.45676802e-02 -2.32618779e-01
2.82040775e-01 -6.19667351e-01 1.16418980e-01 6.18203878e-01
7.99862862e-01 -4.65300202e-01 1.12353277e+00 2.06071779e-01
-1.35320556e+00 1.41634300e-01 -9.64716434e-01 1.17733389e-01
2.26886049e-01 2.60037333e-01 -6.99670136e-01 9.19893324e-01
4.39240098e-01 1.05716515e+00 -6.14805400e-01 1.03121877e+00
-4.30957794e-01 9.30238143e-02 -2.68163174e-01 4.93278295e-01
2.24881053e-01 -1.09471714e+00 2.25377470e-01 3.77500236e-01
9.18791473e-01 2.03494638e-01 2.70058453e-01 7.44140744e-01
-7.49964938e-02 1.21792391e-01 -1.00587273e+00 9.09357443e-02
4.06646132e-01 1.36296642e+00 -5.10231316e-01 -9.87309888e-02
-2.43413746e-01 9.49968040e-01 -2.94947803e-01 2.54254252e-01
-8.33879352e-01 -4.29760039e-01 3.74222785e-01 -7.21784830e-02
-5.86059988e-02 -4.53036368e-01 1.56381741e-01 -1.34136009e+00
-2.80204296e-01 -8.45284045e-01 -6.63434789e-02 -1.16451228e+00
-9.51535583e-01 8.88324916e-01 1.29495323e-01 -1.79001772e+00
-1.45721421e-01 -6.12835109e-01 -7.63751805e-01 8.42534304e-01
-1.23960447e+00 -1.25057864e+00 -1.10057998e+00 6.20325565e-01
1.77416548e-01 -5.27031243e-01 6.52095973e-01 3.69675517e-01
-5.28022289e-01 4.97965932e-01 1.02557600e-01 3.51046681e-01
9.30413961e-01 -9.48963106e-01 5.00707686e-01 1.01357090e+00
9.43635404e-03 7.16760099e-01 5.45904338e-01 -6.03127241e-01
-1.24391353e+00 -7.89429843e-01 7.28452563e-01 -1.52354911e-01
6.75079226e-01 -3.31546064e-03 -8.78231227e-01 9.60562289e-01
2.81497449e-01 -1.62293449e-01 4.56877619e-01 -3.23770911e-01
4.19819355e-02 -5.36864281e-01 -8.80493104e-01 6.90781057e-01
9.04845357e-01 -5.01623750e-01 -5.05283177e-01 4.58445638e-01
5.81982672e-01 -7.97727942e-01 -1.09251177e+00 4.14705873e-01
6.87370718e-01 -8.95445943e-01 1.13878429e+00 -1.66596293e-01
5.54897666e-01 -7.14486420e-01 -1.76311672e-01 -1.52324224e+00
-5.03951848e-01 -6.38799906e-01 5.35185814e-01 1.15795231e+00
-6.67841360e-02 -7.78045237e-01 5.45884430e-01 4.79908049e-01
4.64023575e-02 -2.39299104e-01 -2.56493121e-01 -6.27625406e-01
-3.34567755e-01 4.20938343e-01 1.34914827e+00 1.30079424e+00
-4.33223009e-01 3.08892757e-01 -7.07366168e-01 7.47587979e-01
5.36994040e-01 4.34243977e-01 1.22099662e+00 -1.12085414e+00
4.06558812e-02 -2.40560710e-01 -7.17487454e-01 -1.36701548e+00
3.30981851e-01 -6.46408379e-01 5.82493562e-03 -1.17635286e+00
-2.72459626e-01 -2.91417599e-01 3.21955085e-01 2.37700060e-01
6.62270188e-02 1.23980835e-01 9.97409672e-02 4.99598354e-01
2.33798791e-02 6.89905286e-01 1.78716373e+00 -3.42365205e-01
-2.16464072e-01 2.52864987e-01 -1.50656700e-01 8.72283816e-01
1.01506126e+00 -1.59035072e-01 -5.50797820e-01 -8.52412999e-01
1.39018506e-01 1.36153311e-01 2.16306731e-01 -1.12813067e+00
6.80065513e-01 -1.11483403e-01 5.35502374e-01 -7.79668093e-01
6.38465285e-02 -1.37121415e+00 5.75974643e-01 3.85067075e-01
-1.27524689e-01 6.07515812e-01 1.06938146e-01 9.16745793e-03
-5.29532254e-01 -3.42839122e-01 6.94012642e-01 -7.69136101e-02
-6.68960273e-01 6.55938208e-01 2.28321925e-01 -2.90269971e-01
8.99188936e-01 -5.58848202e-01 -1.69022307e-01 -4.99935627e-01
-2.89304465e-01 2.10143626e-01 7.61213958e-01 1.88275531e-01
6.80243909e-01 -1.55221891e+00 -4.04982269e-01 6.80938840e-01
-1.36583462e-01 2.82920897e-01 3.92540485e-01 6.94295704e-01
-1.10665655e+00 3.37821484e-01 -8.47323060e-01 -6.59757853e-01
-1.31629503e+00 4.07170862e-01 6.60095930e-01 -1.74382813e-02
-4.97435063e-01 1.82795539e-01 2.08009735e-01 -9.13298309e-01
-7.98862725e-02 -4.70880806e-01 -1.98385522e-01 -9.38132107e-02
2.07251459e-01 4.35700893e-01 -2.34182224e-01 -8.97475898e-01
-2.81296317e-02 1.48710227e+00 1.08816214e-01 4.73399550e-01
1.17387426e+00 -2.61456639e-01 -2.30409309e-01 -9.47176144e-02
1.28448939e+00 -3.01720887e-01 -1.36331379e+00 -1.47741348e-01
-5.51091671e-01 -7.80558825e-01 -5.44702224e-02 -1.29177615e-01
-1.40120208e+00 9.20025826e-01 6.90201938e-01 6.05875812e-03
1.12148929e+00 -6.14474058e-01 8.24954629e-01 8.38795722e-01
2.88906902e-01 -8.93453956e-01 -1.14637069e-01 5.73720753e-01
8.32449317e-01 -1.34302557e+00 4.06940430e-01 -4.05956537e-01
-3.81667495e-01 1.52126718e+00 8.59081686e-01 -3.74070823e-01
4.65973288e-01 -3.77503969e-02 3.56078327e-01 -1.02491871e-01
5.30573688e-02 9.60982442e-02 4.87513721e-01 6.79206133e-01
2.38256887e-01 -2.87954986e-01 1.37717173e-01 -1.61493197e-01
-7.44428158e-01 -3.15758795e-01 5.87235212e-01 3.83109778e-01
-1.40441284e-01 -1.11688602e+00 -4.35055196e-01 -6.09790981e-02
1.12379558e-01 1.35048823e-02 -3.21032554e-01 1.14980960e+00
-3.57497297e-02 4.58187819e-01 6.22884594e-02 -6.49968147e-01
6.85346007e-01 -6.54537797e-01 6.00654662e-01 3.60468440e-02
-2.88953513e-01 -1.46597952e-01 -5.67369401e-01 -4.98627067e-01
-6.99208617e-01 -2.85222203e-01 -8.32545638e-01 -5.08120656e-01
-5.02846599e-01 -8.29194412e-02 4.68282044e-01 8.45489085e-01
-2.63452046e-02 4.29498374e-01 8.23834062e-01 -1.02833223e+00
-6.40799046e-01 -5.19493878e-01 -6.26979828e-01 3.47311020e-01
2.69890487e-01 -4.99415934e-01 -3.82997066e-01 3.63027394e-01] | [8.642317771911621, -2.1907479763031006] |
7c32bc01-6f33-468d-a69a-ff2b6a4a4e1d | automated-classification-of-sleep-stages-and | 1809.08443 | null | http://arxiv.org/abs/1809.08443v1 | http://arxiv.org/pdf/1809.08443v1.pdf | Automated Classification of Sleep Stages and EEG Artifacts in Mice with Deep Learning | Sleep scoring is a necessary and time-consuming task in sleep studies. In
animal models (such as mice) or in humans, automating this tedious process
promises to facilitate long-term studies and to promote sleep biology as a
data-driven field. We introduce a deep neural network model that is able to
predict different states of consciousness (Wake, Non-REM, REM) in mice from EEG
and EMG recordings with excellent scoring results for out-of-sample data.
Predictions are made on epochs of 4 seconds length, and epochs are classified
as artifact-free or not. The model architecture draws on recent advances in
deep learning and in convolutional neural networks research. In contrast to
previous approaches towards automated sleep scoring, our model does not rely on
manually defined features of the data but learns predictive features
automatically. We expect deep learning models like ours to become widely
applied in different fields, automating many repetitive cognitive tasks that
were previously difficult to tackle. | ['Justus T. C. Schwabedal', 'Moritz D. Brandt', 'Stephan Bialonski', 'Daniel Sippel'] | 2018-09-22 | null | null | null | null | ['sleep-stage-detection'] | ['medical'] | [ 1.50593266e-01 -4.10878122e-01 1.47270158e-01 -5.49191117e-01
-2.73301482e-01 -4.35973078e-01 2.60915965e-01 1.18186735e-01
-8.63890588e-01 1.00681329e+00 -3.93724404e-02 -4.04102623e-01
-1.22004583e-01 -3.17879915e-01 -3.25552762e-01 -6.84066236e-01
-2.89137185e-01 3.67094636e-01 2.07767621e-01 -1.56927973e-01
2.23590851e-01 4.71000463e-01 -1.45668387e+00 1.35536700e-01
5.90800285e-01 1.06128252e+00 4.00746822e-01 6.44819736e-01
5.21266386e-02 5.60741723e-01 -6.81537986e-01 4.80369255e-02
6.37026737e-03 -4.72114682e-01 -8.05629015e-01 -3.56550932e-01
-1.71098322e-01 1.25133365e-01 1.65131271e-01 7.85235584e-01
5.34696996e-01 7.32469559e-02 2.56356925e-01 -9.35813844e-01
-3.90471250e-01 3.02072037e-02 4.89297546e-02 8.73637557e-01
1.57440186e-01 3.04100186e-01 6.32883847e-01 -3.38837147e-01
1.50779903e-01 2.94637293e-01 9.17669415e-01 6.89760864e-01
-1.66610873e+00 -8.06175351e-01 -4.57209915e-01 6.01414680e-01
-1.13699019e+00 -6.79141045e-01 4.71877724e-01 -5.58931172e-01
1.16640139e+00 2.24177614e-01 1.28979242e+00 1.22380495e+00
8.16376030e-01 1.19180180e-01 1.33247876e+00 -9.31647047e-02
6.89634085e-01 -2.61533797e-01 5.25411144e-02 4.54543263e-01
1.42082036e-01 -9.29758027e-02 -6.99023128e-01 1.18168354e-01
4.62068319e-01 3.77537102e-01 -1.39565930e-01 -1.25945151e-01
-1.15756524e+00 6.94772959e-01 2.86680877e-01 4.88603711e-01
-5.68269849e-01 2.30442390e-01 2.53562272e-01 3.52462530e-01
4.00845259e-01 6.25212848e-01 -8.05730700e-01 -5.55687547e-01
-1.85556269e+00 1.06618442e-01 5.36855996e-01 3.29608232e-01
7.24266469e-01 3.11890040e-02 2.20848508e-02 7.85421491e-01
-2.57987771e-02 5.59698306e-02 1.20378232e+00 -1.04452848e+00
-3.06639344e-01 7.17585802e-01 2.12609872e-01 -2.82733530e-01
-1.07240212e+00 -3.14904809e-01 -8.82800043e-01 4.64409381e-01
2.53061086e-01 4.62551899e-02 -8.61837924e-01 1.65174174e+00
-2.11047292e-01 8.20757076e-02 -2.63929188e-01 7.30772257e-01
7.49654651e-01 1.87739909e-01 1.31313011e-01 -3.30695182e-01
1.34305465e+00 -4.82057154e-01 -5.85678101e-01 -4.57443565e-01
3.90920669e-01 -3.78218770e-01 1.06296217e+00 8.64455938e-01
-1.11582351e+00 -3.58111560e-01 -1.13911974e+00 -2.00376421e-01
-6.58772945e-01 9.93111916e-03 6.25156283e-01 4.29756880e-01
-1.53504121e+00 9.45030034e-01 -1.27664042e+00 -6.74222827e-01
7.95571685e-01 1.06856215e+00 -5.38189054e-01 6.07709110e-01
-7.34384775e-01 1.18618333e+00 4.16719913e-02 1.72575668e-01
-1.21962023e+00 -3.83876234e-01 -5.28010130e-01 2.08473310e-01
-2.36835644e-01 -8.18313360e-01 1.22230875e+00 -6.63676679e-01
-1.39889109e+00 1.21666634e+00 -4.75807935e-01 -7.96312451e-01
-1.13160326e-03 -1.16532810e-01 -3.70928854e-01 -6.17359811e-03
-6.25940859e-02 4.94255781e-01 5.99639952e-01 -5.94659865e-01
-1.88175976e-01 -4.39961672e-01 -1.26657799e-01 -4.52606201e-01
-2.41345391e-01 2.37681165e-01 1.81256965e-01 -2.64148384e-01
-1.57247081e-01 -9.85767722e-01 -3.07222664e-01 -1.18578626e-02
-9.92448851e-02 -2.16982618e-01 4.12535191e-01 -5.26775360e-01
1.00463009e+00 -1.81257224e+00 1.50136814e-01 -2.90835410e-01
5.91575146e-01 4.72107321e-01 7.21288323e-02 3.05997044e-01
-2.01668844e-01 4.18035686e-02 -2.34236777e-01 -4.51725096e-01
-1.43118858e-01 2.50859797e-01 1.91053137e-01 5.77850759e-01
1.35776699e-01 8.84674728e-01 -7.87036538e-01 -2.60359019e-01
1.96883842e-01 4.23757672e-01 -3.32008988e-01 4.19282854e-01
7.34471157e-03 5.03772259e-01 4.38619480e-02 5.17962754e-01
8.68388042e-02 -4.14017856e-01 -1.64442435e-02 1.32935405e-01
-4.78762478e-01 6.94052160e-01 -1.65160134e-01 1.84128988e+00
-6.85044825e-01 9.01891887e-01 -9.01520401e-02 -1.14331234e+00
6.53985560e-01 2.79282779e-01 4.97051388e-01 -9.04465258e-01
3.97917688e-01 -2.14409661e-02 2.79819548e-01 -6.16710544e-01
1.22348003e-01 -4.87523735e-01 2.20830515e-01 4.76572156e-01
5.15506268e-01 -1.21330522e-01 3.58949125e-01 -2.96868831e-01
1.64551353e+00 5.94880618e-02 5.56197762e-01 -4.76162970e-01
1.93203673e-01 -1.75711721e-01 7.44265974e-01 3.31394762e-01
-5.43218017e-01 5.67455351e-01 5.02515137e-01 -7.30918825e-01
-8.22363794e-01 -9.82907295e-01 -1.52195662e-01 9.22107697e-01
-2.28561714e-01 -6.38592422e-01 -7.64047861e-01 -3.95103484e-01
-5.54845691e-01 5.87164819e-01 -8.67223263e-01 -4.20372933e-01
-2.48260900e-01 -9.60534036e-01 2.48334050e-01 4.86608446e-01
-2.18599886e-02 -1.43675113e+00 -1.10032868e+00 4.35319155e-01
-4.19755001e-03 -8.85960460e-01 2.07368240e-01 1.20858729e+00
-8.50402892e-01 -1.08824682e+00 -5.10050774e-01 -6.21243775e-01
3.67648065e-01 8.08036923e-02 1.24113810e+00 1.70432851e-01
-5.70941806e-01 -1.29752636e-01 -2.46265665e-01 -7.60017455e-01
6.54148459e-02 -8.41825828e-02 3.96288365e-01 -3.58855963e-01
9.12591040e-01 -1.36972666e+00 -1.01631653e+00 1.53335199e-01
-8.08577120e-01 -8.56466666e-02 6.39908254e-01 8.40937614e-01
6.20511651e-01 -2.01010823e-01 6.78299248e-01 -4.75007534e-01
6.48926258e-01 -6.46656036e-01 -4.66456622e-01 -1.27229095e-01
-7.76262999e-01 -4.77393949e-03 9.23765719e-01 -3.15312117e-01
-2.57664412e-01 1.79653484e-02 -4.21658456e-01 -2.40171537e-01
-5.12182474e-01 2.39970863e-01 2.35695362e-01 -1.48860306e-01
6.35894239e-01 5.37790358e-01 3.26109454e-02 -4.02215540e-01
-3.84892404e-01 5.16811609e-01 2.76495636e-01 2.85830926e-02
5.39229155e-01 4.54326063e-01 1.10853307e-01 -8.74598265e-01
-9.42441225e-01 -4.62540984e-01 -8.16444099e-01 -4.85794358e-02
1.19544637e+00 -6.92198455e-01 -9.97628331e-01 2.22530827e-01
-7.80936182e-01 -8.40994239e-01 -1.67712584e-01 4.63201344e-01
-7.99162626e-01 4.41705510e-02 -3.96414787e-01 -6.04381919e-01
-6.78789794e-01 -1.08223164e+00 8.80337536e-01 3.63268226e-01
-8.00053239e-01 -9.12639916e-01 6.28684700e-01 2.92041004e-01
5.86979628e-01 -7.98979402e-02 7.80128419e-01 -9.01753306e-01
-2.03658745e-01 -9.13686082e-02 3.14031541e-01 4.02942449e-01
1.33239463e-01 -2.79602259e-01 -1.15048635e+00 -4.63373475e-02
3.05781752e-01 -5.91037154e-01 7.20546305e-01 3.60854149e-01
1.43854463e+00 -1.08516783e-01 -5.24680242e-02 7.84029126e-01
1.21049583e+00 2.86088586e-01 6.68233633e-01 6.29938662e-01
-1.13715110e-02 2.36075789e-01 -1.59687791e-02 1.63776964e-01
3.26078057e-01 5.73777378e-01 6.37112558e-01 -1.07501402e-01
2.21053854e-01 4.67455626e-01 4.16797191e-01 5.91288209e-01
-3.03265154e-01 3.90420184e-02 -8.33789051e-01 7.06856132e-01
-1.51087451e+00 -1.13034368e+00 -1.57075629e-01 2.17281055e+00
1.00653648e+00 2.39277542e-01 3.29434901e-01 2.27686703e-01
1.06423162e-01 -1.65780798e-01 -5.61129451e-01 -6.93423271e-01
1.00991331e-01 8.71430099e-01 5.14059216e-02 -1.38695106e-01
-7.62925863e-01 6.65737867e-01 7.25207567e+00 2.56315529e-01
-1.25103986e+00 2.85448879e-01 3.16521645e-01 -7.84832299e-01
2.61506468e-01 -1.78547159e-01 -4.46206450e-01 8.49699974e-01
1.57910609e+00 -1.84272692e-01 1.00110400e+00 6.79661453e-01
6.89118087e-01 -3.74184370e-01 -1.33051574e+00 9.55764711e-01
-1.20255798e-01 -1.34557354e+00 -7.38924682e-01 1.76983401e-01
2.76752979e-01 5.70090294e-01 -1.61746070e-01 3.66084754e-01
2.69568950e-01 -1.37207937e+00 6.02173984e-01 7.82362521e-01
6.45807862e-01 -3.60136211e-01 9.42272186e-01 4.89164412e-01
-7.32471704e-01 -1.56719849e-01 -5.11081219e-01 -2.94724703e-01
-7.49319419e-02 6.35268152e-01 -7.03091860e-01 -1.41000494e-01
1.26842427e+00 6.67563498e-01 -9.90556836e-01 1.34509623e+00
-1.46378607e-01 8.93787563e-01 -1.69107124e-01 -1.80836022e-01
-7.89222196e-02 -1.99764594e-01 7.25542530e-02 1.07817781e+00
4.89701599e-01 -1.99996307e-01 -3.88039827e-01 1.10112369e+00
-4.66793217e-03 -3.79166275e-01 -4.14393365e-01 -1.91579014e-01
7.49013126e-02 1.67314589e+00 -8.35971713e-01 8.35985541e-02
-2.85366029e-01 8.64786983e-01 5.49810767e-01 6.12893887e-02
-5.80645978e-01 -3.84714931e-01 7.41609275e-01 2.40754828e-01
2.28890002e-01 -4.02766377e-01 -4.27511007e-01 -1.05605853e+00
-6.61341473e-02 -4.14504617e-01 -7.12213852e-03 -1.02677071e+00
-9.73121345e-01 6.71273768e-01 -4.23596680e-01 -1.04866397e+00
-2.72906214e-01 -5.51287830e-01 -1.23592472e+00 6.94620490e-01
-1.47603071e+00 -8.28848660e-01 -2.33248964e-01 4.90428358e-01
5.39722323e-01 -3.37065570e-02 1.15598547e+00 2.31816098e-01
-6.27749562e-01 1.44293562e-01 9.18524563e-02 -1.35572478e-01
6.13649845e-01 -1.41424143e+00 8.31612423e-02 6.21202052e-01
3.03129584e-01 1.00478971e+00 8.90686274e-01 -8.97344723e-02
-9.19548869e-01 -1.00377274e+00 1.06572711e+00 -5.88348687e-01
8.40159893e-01 -5.16462147e-01 -6.90623283e-01 5.59757054e-01
2.59582251e-01 7.05943303e-03 1.42939878e+00 3.26525390e-01
2.49434546e-01 -3.42690080e-01 -9.51067448e-01 4.19621497e-01
7.92336047e-01 -3.99740607e-01 -8.47464502e-01 3.35302979e-01
1.05342962e-01 1.24343403e-01 -7.24776328e-01 -2.45043561e-02
6.53610170e-01 -1.32959735e+00 7.59934187e-01 -7.08333254e-01
4.12683696e-01 -2.27601796e-01 2.71840334e-01 -1.44536495e+00
-5.41809618e-01 -6.34604752e-01 -2.50485837e-01 8.72000039e-01
2.69178987e-01 -2.90202588e-01 8.69189620e-01 3.42106432e-01
-5.25533438e-01 -1.00531256e+00 -1.15546203e+00 -8.01214874e-01
6.91705346e-02 -2.83506304e-01 4.46080387e-01 3.46258104e-01
2.09462747e-01 3.83782446e-01 -1.58375949e-01 -2.39751711e-01
3.14242959e-01 3.23234834e-02 4.88194317e-01 -1.67396462e+00
1.44409925e-01 -4.55761492e-01 -6.51617825e-01 -2.94218540e-01
2.75831372e-01 -8.50171089e-01 2.76759118e-01 -1.71643901e+00
3.34612459e-01 -4.75443602e-02 -8.16291511e-01 8.87817204e-01
1.04606777e-01 7.78113902e-01 -3.16598475e-01 9.02885348e-02
-9.96740103e-01 5.19174218e-01 5.50454378e-01 2.51860619e-02
-6.55433610e-02 7.10768402e-02 -6.97414398e-01 9.03477132e-01
1.20934546e+00 -7.48684883e-01 -1.76778808e-01 -1.92187890e-01
5.58645904e-01 -2.42763817e-01 6.45955086e-01 -1.60129726e+00
1.88640133e-01 -2.14634687e-01 8.11319053e-01 -2.85185277e-01
6.49285495e-01 -7.44997501e-01 2.69708484e-01 3.12888622e-01
-2.35477671e-01 6.89645261e-02 1.26101434e-01 3.44001263e-01
1.03297830e-01 -1.96638212e-01 8.97768676e-01 -3.94185722e-01
-3.42741549e-01 1.83644384e-01 -9.10224020e-01 -1.02831736e-01
8.62617552e-01 -5.15151858e-01 -1.99564755e-01 -5.78054041e-02
-1.20240211e+00 6.80807158e-02 6.68742061e-01 1.28194720e-01
3.96636963e-01 -9.15345609e-01 -1.04005314e-01 2.90235966e-01
5.15891835e-02 -2.58834571e-01 2.70545274e-01 1.23056078e+00
-3.52124840e-01 4.74789202e-01 -7.17852294e-01 -4.77120042e-01
-1.28794241e+00 3.91584814e-01 2.59965003e-01 -2.73200780e-01
-4.63564008e-01 6.75659478e-01 -1.57430038e-01 -5.06495009e-04
-2.79652737e-02 -4.30340022e-01 -5.94422877e-01 -7.95097575e-02
5.91539443e-01 2.08911359e-01 5.09940326e-01 -3.08072448e-01
-3.69498521e-01 2.13685006e-01 1.10972092e-01 2.89177865e-01
2.02472377e+00 1.80400498e-02 -3.33590001e-01 8.22602749e-01
9.36552465e-01 -1.83384642e-01 -1.09860313e+00 4.36286718e-01
4.11059484e-02 1.76126495e-01 2.80475229e-01 -9.54187334e-01
-7.92919397e-01 1.21765804e+00 7.97750771e-01 4.73741412e-01
1.29388440e+00 -1.41845509e-01 1.02022195e+00 5.81772745e-01
3.84156317e-01 -1.19171834e+00 1.71627663e-02 2.98926950e-01
6.40787959e-01 -1.19674385e+00 1.19724892e-01 5.91598868e-01
-2.14080885e-01 1.21080422e+00 3.77711445e-01 -1.34126231e-01
5.37764788e-01 1.86629996e-01 -6.28684014e-02 -4.90410328e-01
-1.08863401e+00 -3.67399991e-01 1.24031365e-01 7.65313745e-01
5.81754625e-01 -2.11540028e-01 -4.01618868e-01 1.16480184e+00
-6.26713872e-01 5.90161622e-01 7.59852171e-01 9.08160329e-01
-6.36753619e-01 -1.10356486e+00 2.32222695e-02 9.95636761e-01
-8.56819808e-01 -2.54770696e-01 -3.71550560e-01 3.90610546e-01
3.72130394e-01 1.09142017e+00 -1.70678049e-02 -5.48137307e-01
-9.05627683e-02 4.36130196e-01 3.86260897e-01 -8.99940073e-01
-7.74154127e-01 -2.38785043e-01 -7.44661540e-02 -7.42848635e-01
-5.50940573e-01 -6.11730635e-01 -1.17375910e+00 -1.33150890e-01
-1.06208764e-01 2.32993498e-01 8.13161790e-01 1.09228361e+00
5.19411325e-01 6.59652472e-01 2.76861221e-01 -1.15388691e+00
-1.19736500e-03 -1.01630449e+00 -6.04070842e-01 1.35929152e-01
6.89060569e-01 -7.45911300e-01 -5.00599861e-01 4.01279271e-01] | [13.478666305541992, 3.524874687194824] |
9049e17c-b6c5-4066-908d-a34efc9f4338 | calibrated-nonparametric-scan-statistics-for | 2206.12786 | null | https://arxiv.org/abs/2206.12786v1 | https://arxiv.org/pdf/2206.12786v1.pdf | Calibrated Nonparametric Scan Statistics for Anomalous Pattern Detection in Graphs | We propose a new approach, the calibrated nonparametric scan statistic (CNSS), for more accurate detection of anomalous patterns in large-scale, real-world graphs. Scan statistics identify connected subgraphs that are interesting or unexpected through maximization of a likelihood ratio statistic; in particular, nonparametric scan statistics (NPSSs) identify subgraphs with a higher than expected proportion of individually significant nodes. However, we show that recently proposed NPSS methods are miscalibrated, failing to account for the maximization of the statistic over the multiplicity of subgraphs. This results in both reduced detection power for subtle signals, and low precision of the detected subgraph even for stronger signals. Thus we develop a new statistical approach to recalibrate NPSSs, correctly adjusting for multiple hypothesis testing and taking the underlying graph structure into account. While the recalibration, based on randomization testing, is computationally expensive, we propose both an efficient (approximate) algorithm and new, closed-form lower bounds (on the expected maximum proportion of significant nodes for subgraphs of a given size, under the null hypothesis of no anomalous patterns). These advances, along with the integration of recent core-tree decomposition methods, enable CNSS to scale to large real-world graphs, with substantial improvement in the accuracy of detected subgraphs. Extensive experiments on both semi-synthetic and real-world datasets are demonstrated to validate the effectiveness of our proposed methods, in comparison with state-of-the-art counterparts. | ['Feng Chen', 'Daniel B. Neill', 'Chunpai Wang'] | 2022-06-26 | null | null | null | null | ['tree-decomposition'] | ['graphs'] | [ 5.15046120e-01 1.98241338e-01 -1.89884901e-01 -8.88123512e-02
-4.22061831e-01 -6.39383733e-01 3.45470518e-01 2.85561740e-01
1.16426684e-01 8.56836379e-01 -2.52557278e-01 -5.88570476e-01
-5.56702733e-01 -1.07043672e+00 -8.37985754e-01 -8.03688705e-01
-7.17998683e-01 5.44255316e-01 7.59004176e-01 2.40725756e-01
1.69736013e-01 5.23605466e-01 -1.28016162e+00 -2.05798358e-01
1.00281405e+00 6.74783528e-01 -2.86820859e-01 5.79738915e-01
3.34817767e-01 1.74159765e-01 -4.44111615e-01 -3.09861183e-01
2.78074533e-01 -5.01782894e-01 -4.85630929e-01 3.14505786e-01
3.57706219e-01 9.62305218e-02 -1.70759037e-01 1.34588575e+00
2.45267078e-01 -1.81599706e-01 6.54701233e-01 -1.35692155e+00
3.81787866e-02 7.73116052e-01 -1.30365229e+00 4.97431636e-01
3.64856929e-01 1.87628955e-01 1.22757995e+00 -6.09855115e-01
5.08737922e-01 1.12404835e+00 6.87265217e-01 -3.16147208e-01
-1.60453558e+00 -7.80254900e-01 1.46463156e-01 -6.83048368e-02
-1.67554462e+00 -1.03285499e-01 7.09862351e-01 -3.08480024e-01
4.82856363e-01 6.27772391e-01 5.55405438e-01 8.20234716e-01
1.87881455e-01 1.82913378e-01 1.11323953e+00 -2.19009981e-01
2.83826113e-01 -9.97524932e-02 1.96316957e-01 8.75357211e-01
1.15737510e+00 1.32304892e-01 -3.46271008e-01 -7.21698821e-01
6.40340567e-01 -2.88413167e-01 -2.75078148e-01 -6.45891666e-01
-1.21394098e+00 7.67016232e-01 7.39128441e-02 2.44267598e-01
-5.66159070e-01 -1.35052890e-01 3.83213490e-01 1.25943393e-01
5.65686345e-01 1.88769102e-01 -3.89367968e-01 1.09222889e-01
-1.15899336e+00 1.55016586e-01 8.30036342e-01 7.88277090e-01
7.67923295e-01 -6.11747913e-02 -3.19313824e-01 4.17809874e-01
1.37811109e-01 6.08149350e-01 -1.83215346e-02 -3.43365878e-01
3.37117821e-01 9.71227705e-01 -8.53765979e-02 -1.48427260e+00
-4.71936405e-01 -1.06503630e+00 -1.21381068e+00 -2.00147018e-01
4.98187929e-01 -7.59851653e-03 -6.27784312e-01 1.90367520e+00
5.77908099e-01 5.63128710e-01 -3.78201991e-01 5.66480577e-01
2.07806364e-01 2.96124160e-01 -2.59007514e-01 -5.38891554e-01
1.27825785e+00 -3.46710056e-01 -2.50328034e-01 5.23472801e-02
5.46737432e-01 -3.32724929e-01 9.76451635e-01 3.93884778e-01
-6.49794519e-01 -5.51534966e-02 -9.27969098e-01 8.86825204e-01
-3.89890037e-02 -8.30232799e-02 5.41924357e-01 1.03255749e+00
-9.94271278e-01 4.73165512e-01 -7.05643773e-01 -5.69069803e-01
9.83666852e-02 3.56872648e-01 -2.63028741e-01 -1.71618573e-02
-8.14861059e-01 6.53042197e-02 4.10743833e-01 1.34787317e-02
-7.00866103e-01 -5.19891381e-01 -6.58489227e-01 3.13302070e-01
9.10634696e-01 -4.37899411e-01 4.72339839e-01 -7.11544931e-01
-7.69689083e-01 6.33565187e-01 -2.88444042e-01 -5.04808664e-01
4.10746515e-01 3.48877877e-01 -4.72529978e-01 1.87151402e-01
3.27348650e-01 -1.03117399e-01 6.41484618e-01 -9.96455550e-01
-3.60317886e-01 -4.07581568e-01 -2.42926732e-01 -4.19808090e-01
-1.55136809e-01 -5.33428639e-02 -3.68915468e-01 -5.86604893e-01
4.15348738e-01 -8.01043034e-01 -3.38983595e-01 -4.40450877e-01
-1.03025210e+00 -1.27359644e-01 4.99649614e-01 -5.19109726e-01
1.35488498e+00 -2.10618567e+00 -1.16459534e-01 1.21828771e+00
4.85556841e-01 -1.60145193e-01 -1.14889599e-01 4.44103748e-01
-2.16448903e-01 2.55009830e-01 -4.68718112e-01 2.26627260e-01
-3.03541809e-01 1.18773691e-01 -1.17052086e-01 8.70965898e-01
2.66565174e-01 4.11518961e-01 -8.59436870e-01 -3.77137005e-01
-4.99147810e-02 -2.75097460e-01 -5.35712421e-01 3.67897898e-02
1.06654882e-01 2.88137227e-01 -4.72110659e-01 6.95005059e-01
1.09417009e+00 -6.12895310e-01 6.21051490e-01 4.60194647e-02
-1.48904368e-01 2.63051651e-02 -1.61422098e+00 5.90321541e-01
2.83706546e-01 1.01859137e-01 9.75205824e-02 -1.28688657e+00
9.43122685e-01 -8.44257921e-02 4.08393830e-01 -3.64937395e-01
-5.48852002e-03 2.68185496e-01 1.69445425e-01 -1.25800699e-01
6.67083338e-02 1.75876006e-01 -8.80932733e-02 3.89682412e-01
-1.89280272e-01 4.90080118e-01 6.18323028e-01 4.87018853e-01
1.82308280e+00 -4.75556254e-01 7.64435768e-01 -7.04136908e-01
6.98356032e-01 -9.16414261e-02 7.21256733e-01 1.02971375e+00
-4.39405739e-02 3.12950373e-01 1.20691848e+00 -1.36033028e-01
-8.16781819e-01 -1.16583240e+00 -1.21314868e-01 6.95403397e-01
1.70808896e-01 -3.81942034e-01 -6.47288442e-01 -7.81462073e-01
1.89362377e-01 4.89157856e-01 -6.19101107e-01 -1.77847907e-01
-3.72248411e-01 -1.27442431e+00 5.20312667e-01 1.05549254e-01
1.84866071e-01 -6.49262130e-01 -2.57823288e-01 1.04768880e-01
-3.12841311e-02 -1.28542721e+00 -3.80852044e-01 3.99294756e-02
-9.95982170e-01 -1.81245852e+00 -2.57771224e-01 -2.67795533e-01
1.02667761e+00 3.97953600e-01 1.07735634e+00 2.57582366e-01
-3.27314407e-01 2.67205715e-01 -3.53141189e-01 2.41726533e-01
-4.17409509e-01 -6.81423908e-03 1.31239921e-01 3.01730156e-01
2.41251918e-03 -9.96157944e-01 -3.53239983e-01 6.57724142e-01
-8.06746244e-01 -2.17457384e-01 9.62746203e-01 7.94532776e-01
7.17874765e-01 5.00243068e-01 3.71824712e-01 -1.22126818e+00
4.19268489e-01 -8.77634525e-01 -9.10951614e-01 1.94345564e-01
-8.89675617e-01 2.42434576e-01 5.03781617e-01 -3.55197906e-01
-5.59291482e-01 -2.34917477e-01 4.21337068e-01 -2.14903891e-01
-1.25426212e-02 7.68401980e-01 -2.97014751e-02 -2.18458667e-01
5.64713180e-01 3.96807313e-01 -1.02741793e-01 -3.69383782e-01
-2.83846054e-02 9.83078778e-02 5.28738976e-01 -5.24713576e-01
1.09788740e+00 4.30594236e-01 5.65465987e-01 -8.28612626e-01
-2.52202839e-01 -6.63578749e-01 -5.09844303e-01 -1.41612083e-01
1.58109441e-01 -7.40001440e-01 -6.26788318e-01 2.33078077e-01
-7.36415386e-01 1.69302717e-01 3.09051592e-02 3.70330811e-01
-1.87754661e-01 8.67913544e-01 -7.09334835e-02 -9.23707724e-01
-1.85911581e-01 -7.67024696e-01 8.39522362e-01 -1.39814213e-01
-2.34030500e-01 -7.04240441e-01 2.42692515e-01 -4.42058593e-02
3.00769340e-02 8.67328942e-01 1.16369736e+00 -9.98950243e-01
-6.27018869e-01 -4.77396131e-01 -5.79191983e-01 -1.89043283e-01
1.50711630e-02 4.28955108e-01 -3.67357194e-01 -4.94418949e-01
-4.75163609e-01 2.98329443e-01 7.09322393e-01 4.51776654e-01
1.14661908e+00 -4.92548048e-01 -6.75467551e-01 3.36270988e-01
1.35375202e+00 -3.48216623e-01 3.80967021e-01 -2.94759292e-02
3.07545036e-01 3.95168126e-01 6.00110233e-01 7.97599018e-01
7.99581110e-02 6.01724327e-01 6.36618793e-01 -1.36456594e-01
8.89905319e-02 -3.03084075e-01 3.14232618e-01 6.23398006e-01
-1.90156966e-01 -4.43431526e-01 -8.00389886e-01 5.43969691e-01
-1.82849455e+00 -8.01072359e-01 -7.96475887e-01 2.68247533e+00
4.81630147e-01 4.41110939e-01 4.80619967e-01 3.00245434e-01
1.23627079e+00 5.25728352e-02 -5.99236131e-01 -7.62398839e-02
-3.47297966e-01 2.00887412e-01 8.79022419e-01 1.56341761e-01
-8.70955408e-01 4.20783460e-01 6.55292988e+00 9.51720357e-01
-7.14215159e-01 -4.71801944e-02 5.85198581e-01 2.62254238e-01
-3.81119996e-01 2.49050155e-01 -4.44270223e-01 3.97680610e-01
9.94246960e-01 -3.72563004e-01 3.05921376e-01 7.96054840e-01
4.68576491e-01 -3.30021828e-01 -6.52710497e-01 5.09639144e-01
-1.59708530e-01 -9.59338546e-01 -8.14224184e-02 5.21703660e-01
7.40721107e-01 -7.63883302e-03 -3.23344380e-01 3.64551730e-02
2.82428831e-01 -5.12221456e-01 5.25864184e-01 2.40880534e-01
7.00314105e-01 -7.93116927e-01 8.46354485e-01 3.62271339e-01
-1.39671659e+00 2.15872936e-02 -2.18723387e-01 3.21963988e-02
-1.52524531e-01 1.36425233e+00 -1.06170237e+00 7.79297173e-01
6.33493841e-01 5.02029657e-01 -8.76530647e-01 1.18690908e+00
-2.19680384e-01 1.09212744e+00 -6.91477358e-01 -1.23237059e-01
-3.43426317e-02 -1.94865257e-01 1.03403389e+00 9.51712251e-01
4.24979717e-01 -1.61738098e-01 1.87505171e-01 9.14329290e-01
9.68905911e-02 2.59395927e-01 -5.91549814e-01 -6.32507056e-02
5.98009586e-01 1.29165220e+00 -1.26795900e+00 -2.21761256e-01
-2.64788270e-01 4.61187929e-01 2.11587146e-01 3.02370369e-01
-7.76959181e-01 -2.50729471e-01 3.05237979e-01 5.90828061e-01
3.56986761e-01 -2.06827223e-01 -2.92766094e-01 -9.84731555e-01
2.66456693e-01 -9.23517525e-01 7.62265503e-01 -1.33664399e-01
-1.15624034e+00 2.31308818e-01 2.04354584e-01 -1.11566734e+00
-2.40173399e-01 -4.16334510e-01 -9.73908424e-01 3.42999429e-01
-1.09683597e+00 -6.73179209e-01 -2.88754493e-01 5.52362204e-01
-2.47316137e-02 1.36355668e-01 4.33585167e-01 9.06183645e-02
-6.82727754e-01 7.11967587e-01 6.76503107e-02 -1.68177694e-01
4.04655963e-01 -1.11149192e+00 3.50786060e-01 1.31555128e+00
1.29166814e-02 4.32419360e-01 1.05820823e+00 -9.92715120e-01
-1.29969251e+00 -9.88503814e-01 4.65029895e-01 2.07024828e-01
8.05212080e-01 -5.09903848e-01 -8.05469930e-01 4.29206580e-01
-2.89808989e-01 4.52490263e-02 4.67016071e-01 5.54495275e-01
-3.44392717e-01 -9.13078710e-02 -1.34752762e+00 4.44724441e-01
1.28719330e+00 1.22956328e-01 -1.24187358e-02 3.54336202e-01
5.06544471e-01 1.01903453e-01 -7.24725366e-01 8.74503791e-01
3.22436839e-01 -1.27291548e+00 7.21188605e-01 -3.53024483e-01
-1.49407536e-01 -5.10434628e-01 8.05610940e-02 -8.79538417e-01
-4.52210218e-01 -7.96791434e-01 -1.47490352e-01 1.10167718e+00
5.15663147e-01 -9.26988125e-01 8.70107830e-01 9.92727801e-02
2.49159932e-01 -6.30813241e-01 -1.24179816e+00 -1.02175999e+00
-7.16852665e-01 -2.74847180e-01 4.37260956e-01 9.00502324e-01
2.37334464e-02 1.84154715e-02 -3.71777982e-01 7.48104572e-01
1.03320563e+00 2.19104677e-01 1.00342369e+00 -1.54475343e+00
-5.05391002e-01 -2.61582792e-01 -8.95201802e-01 -4.97774571e-01
-1.24814376e-01 -5.89558899e-01 -2.55923033e-01 -8.91499639e-01
6.42032802e-01 -1.52373999e-01 -1.35300756e-01 4.15677488e-01
-3.25903684e-01 2.04700857e-01 -5.51113784e-01 1.72204375e-02
-5.60037732e-01 3.53583097e-01 8.38464379e-01 1.78747118e-01
-2.18358338e-01 2.64952391e-01 -6.77439928e-01 5.69953859e-01
6.49218619e-01 -7.55074918e-01 -1.56682745e-01 6.32483363e-01
4.20699865e-01 5.82853742e-02 5.11970818e-01 -9.68833804e-01
8.96685496e-02 -6.49431199e-02 2.40525499e-01 -8.02543223e-01
-2.91614860e-01 -5.75572729e-01 3.71858090e-01 8.14757049e-01
1.39222994e-01 2.29595646e-01 8.71936008e-02 1.05302525e+00
-1.22131594e-01 3.48573402e-02 6.89443827e-01 2.99220175e-01
-2.82664388e-01 1.94076091e-01 -3.67305398e-01 5.45644239e-02
1.11805725e+00 -2.32772663e-01 -2.61508226e-01 -4.10724610e-01
-4.70819116e-01 2.42656752e-01 3.41953903e-01 6.91109430e-03
4.34260100e-01 -1.06011319e+00 -8.02844167e-01 3.23319942e-01
1.91097617e-01 -2.70662755e-01 2.37902611e-01 1.34916651e+00
-2.23340899e-01 1.05504252e-01 1.52092367e-01 -7.59607971e-01
-1.40440536e+00 5.35048962e-01 -5.54914512e-02 -6.72611594e-01
-5.22694111e-01 3.94662380e-01 1.32330105e-01 -3.17806870e-01
-2.91838795e-01 -3.83950919e-01 1.52054623e-01 -2.71974057e-01
1.51548043e-01 6.55512094e-01 3.36355151e-04 -3.61011952e-01
-5.43765128e-01 4.81249660e-01 9.73675549e-02 2.59235144e-01
1.19793653e+00 -1.06214307e-01 -4.49528843e-01 1.89389605e-02
7.38383055e-01 6.17884457e-01 -7.56675839e-01 -3.52698088e-01
1.96660653e-01 -5.31751573e-01 -1.39578611e-01 -4.34565902e-01
-1.04121482e+00 1.70972705e-01 2.32313648e-01 6.45591736e-01
1.28493202e+00 1.74049407e-01 1.57858968e-01 1.93336174e-01
5.19555509e-01 -5.23531914e-01 -2.21815124e-01 1.52733997e-01
5.53511381e-01 -9.13679600e-01 3.03314090e-01 -1.14371705e+00
-2.20367223e-01 1.09880519e+00 4.25238550e-01 -3.10869455e-01
5.97696841e-01 1.24911621e-01 -7.68071711e-01 -4.11031574e-01
-6.72719836e-01 -1.00546874e-01 1.54659435e-01 3.03205311e-01
5.22861024e-03 2.09064916e-01 -5.79682648e-01 6.06228232e-01
-2.89030895e-02 -4.09545213e-01 5.77244163e-01 4.41112399e-01
-4.95162070e-01 -7.85464287e-01 -4.55491543e-01 7.45417118e-01
-3.79976392e-01 -9.78579372e-02 -5.80475152e-01 1.02669132e+00
-1.26232058e-01 9.10554767e-01 2.91324891e-02 -3.57584745e-01
2.70766377e-01 -2.58941114e-01 1.51600972e-01 -3.71605009e-01
-7.87480175e-02 1.56979159e-01 2.92648256e-01 -6.54213846e-01
-2.09453285e-01 -7.59063482e-01 -7.82902062e-01 -4.24081683e-01
-5.27666152e-01 3.55333477e-01 1.64943427e-01 9.07845318e-01
5.61034918e-01 3.70299041e-01 9.27003622e-01 -3.12422454e-01
-6.13186002e-01 -9.06589568e-01 -9.82831180e-01 1.42024830e-01
-6.05545305e-02 -8.07619154e-01 -7.76180863e-01 -5.20208001e-01] | [7.015446186065674, 5.208978652954102] |
f88eae4e-38ba-4c6a-8a6e-25714527a4a2 | boosting-deep-ctr-prediction-with-a-plug-and | null | null | https://aclanthology.org/2022.coling-1.249 | https://aclanthology.org/2022.coling-1.249.pdf | Boosting Deep CTR Prediction with a Plug-and-Play Pre-trainer for News Recommendation | Understanding news content is critical to improving the quality of news recommendation. To achieve this goal, recent studies have attempted to apply pre-trained language models (PLMs) such as BERT for semantic-enhanced news recommendation. Despite their great success in offline evaluation, it is still a challenge to apply such large PLMs in real-time ranking model due to the stringent requirement in inference and updating time. To bridge this gap, we propose a plug-and-play pre-trainer, namely PREC, to learn both user and news encoders through multi-task pre-training. Instead of directly leveraging sophisticated PLMs for end-to-end inference, we focus on how to use the derived user and item representations to boost the performance of conventional lightweight models for click-through-rate prediction. This enables efficient online inference as well as compatibility to conventional models, which would significantly ease the practical deployment. We validate the effectiveness of PREC through both offline evaluation on public datasets and online A/B testing in an industrial application. | ['Xiaoming Wu', 'Quanyu Dai', 'Jieming Zhu', 'Qijiong Liu'] | null | null | null | null | coling-2022-10 | ['click-through-rate-prediction'] | ['miscellaneous'] | [ 3.04647200e-02 -1.16441071e-01 -3.71195853e-01 -5.71150839e-01
-1.06771827e+00 -3.73960227e-01 5.91752946e-01 2.82746460e-03
-5.06789744e-01 6.27956271e-01 3.76060963e-01 -5.87089658e-01
-8.68652016e-02 -6.47817135e-01 -9.76294577e-01 -5.30898422e-02
1.39846904e-02 3.63333732e-01 1.65660173e-01 -3.38882744e-01
7.00405389e-02 -1.83580592e-01 -1.64001203e+00 5.62434554e-01
8.53481710e-01 1.06404638e+00 5.43917239e-01 7.29053319e-01
-9.19808075e-02 8.09281766e-01 -2.72061467e-01 -7.96754122e-01
1.77179351e-01 -1.90312430e-01 -4.33740526e-01 -2.54713297e-01
2.42608041e-01 -8.61678123e-01 -2.79159337e-01 7.92686045e-01
7.16405869e-01 4.53534782e-01 2.58827925e-01 -8.28411758e-01
-8.54414999e-01 1.00907052e+00 -3.18149954e-01 1.39096707e-01
2.28056744e-01 -4.01393563e-01 1.48063672e+00 -8.58069062e-01
1.69724613e-01 1.04772758e+00 5.83500504e-01 3.91418874e-01
-8.85210276e-01 -5.25853753e-01 5.81932783e-01 1.53734103e-01
-9.27674055e-01 -6.11912906e-01 4.61554497e-01 -1.49072409e-01
9.96987402e-01 2.43932709e-01 3.14387828e-01 1.33005011e+00
-1.07897617e-01 1.27625513e+00 7.07249522e-01 -2.57970810e-01
1.09686874e-01 2.97398776e-01 1.45239815e-01 5.62784910e-01
1.98046908e-01 -1.00418337e-01 -6.23134732e-01 -1.36290584e-02
5.85818112e-01 2.91785747e-01 -1.06128313e-01 -1.93500221e-02
-9.01136816e-01 8.42668772e-01 2.63250053e-01 -9.56926048e-02
-3.23510885e-01 3.67614597e-01 5.18605232e-01 3.10872883e-01
9.07538354e-01 4.21608657e-01 -7.95485318e-01 -4.09570754e-01
-1.05088019e+00 2.81030327e-01 9.82726932e-01 8.77403021e-01
2.33726799e-01 -6.72070831e-02 -2.87777305e-01 1.12339258e+00
4.23202574e-01 3.41419548e-01 5.44989526e-01 -4.62307036e-01
5.14787674e-01 1.59637824e-01 2.54998296e-01 -5.73399246e-01
-1.61475360e-01 -1.10523820e+00 -3.61253440e-01 -4.59368795e-01
1.79606587e-01 -1.49485201e-01 -7.66289830e-01 1.62772298e+00
1.91331878e-01 4.03796494e-01 -1.09127410e-01 1.03738940e+00
5.37811399e-01 7.75990665e-01 8.96328762e-02 5.51734045e-02
1.45369256e+00 -1.49534070e+00 -6.01970673e-01 -3.97651196e-01
7.16578603e-01 -7.96899319e-01 1.45774245e+00 4.04370070e-01
-1.03398442e+00 -5.67847013e-01 -9.62656677e-01 -3.01667362e-01
-1.73244968e-01 6.05393171e-01 8.78986716e-01 5.19456625e-01
-6.14486456e-01 5.39853811e-01 -9.92779553e-01 -1.50669724e-01
1.00457475e-01 2.18579277e-01 1.84837937e-01 -1.14508763e-01
-1.42831087e+00 6.56023681e-01 1.69343665e-01 1.83395907e-01
-8.59088540e-01 -8.55280936e-01 -5.05112410e-01 4.19069588e-01
5.97990334e-01 -8.11830103e-01 1.73013067e+00 -8.66838515e-01
-2.06217456e+00 1.33951753e-01 4.18684352e-03 -5.69083095e-01
4.35145646e-01 -9.37926054e-01 -6.00620270e-01 -3.13069969e-01
-1.39763713e-01 8.17596987e-02 6.51582897e-01 -1.03455925e+00
-8.63076627e-01 -1.00501314e-01 6.25556052e-01 1.51675880e-01
-7.47898102e-01 1.50687784e-01 -9.09411907e-01 -8.52921367e-01
-4.76200610e-01 -7.44558513e-01 -2.97329396e-01 -2.48402163e-01
-2.56235301e-01 -1.95337236e-01 4.36099738e-01 -1.00368166e+00
1.36014044e+00 -2.06251478e+00 -3.48974466e-01 8.02744254e-02
-6.71259090e-02 4.41536129e-01 -3.49056274e-01 4.48654920e-01
5.11555612e-01 -2.72510350e-01 4.16579694e-01 -5.81386447e-01
3.04951161e-01 1.12722009e-01 -4.99799550e-01 9.66744646e-02
-7.19920918e-02 8.42320085e-01 -1.01519895e+00 -6.07269481e-02
9.26673636e-02 6.12519801e-01 -1.15084577e+00 4.80269134e-01
-4.99002367e-01 1.54593438e-01 -6.44716144e-01 3.45789254e-01
3.27523202e-01 -6.51342869e-01 3.30096245e-01 -1.68938309e-01
1.77465081e-01 1.10516632e+00 -1.10131741e+00 1.56411922e+00
-1.16526043e+00 7.12180361e-02 4.56747487e-02 -1.02044070e+00
5.71207464e-01 2.25849867e-01 2.35745251e-01 -1.12913549e+00
8.40295032e-02 1.27299860e-01 -2.45923162e-01 -3.64667773e-01
8.86148453e-01 2.52454370e-01 1.05293624e-01 5.56429088e-01
3.61193568e-02 5.91626525e-01 1.88656896e-01 2.32586399e-01
8.35074663e-01 4.21634763e-01 7.99823180e-02 2.57746894e-02
2.11161628e-01 -4.06555116e-01 4.30980325e-01 8.62120628e-01
4.54810798e-01 3.58583778e-01 1.00179516e-01 -1.09697841e-01
-8.78427982e-01 -8.60603213e-01 2.19376106e-02 2.04317021e+00
3.38917202e-03 -8.05767834e-01 -5.16417086e-01 -7.49763846e-01
1.54160624e-02 1.04637063e+00 -1.74009904e-01 -7.15424940e-02
-4.02714670e-01 -7.72546887e-01 1.90718800e-01 6.29281461e-01
2.49826729e-01 -3.89508247e-01 -1.65523201e-01 6.22169554e-01
-1.71591550e-01 -1.23370802e+00 -5.40127158e-01 2.96987258e-02
-7.94188261e-01 -5.41009545e-01 -6.92658126e-01 -4.13163066e-01
4.48978156e-01 5.55115104e-01 1.02793598e+00 -8.99897218e-02
2.59074628e-01 2.15925336e-01 -7.26107180e-01 -3.49444658e-01
-1.80720672e-01 4.22423482e-01 4.31123078e-02 1.70505908e-03
8.42453614e-02 -6.41459167e-01 -9.81284916e-01 5.33745944e-01
-7.78253794e-01 4.31089759e-01 8.13054204e-01 8.55576813e-01
4.06249255e-01 -2.32665420e-01 8.48145843e-01 -1.24313664e+00
7.61143208e-01 -5.95447183e-01 -6.99748337e-01 4.79913324e-01
-9.16402340e-01 7.56331831e-02 9.82363045e-01 -5.32856882e-01
-1.14160883e+00 -3.93665761e-01 -6.60709500e-01 1.01400331e-01
5.15829980e-01 1.07641554e+00 2.26266250e-01 3.10127914e-01
4.10531372e-01 1.76779404e-01 -2.62700826e-01 -9.24670875e-01
6.35932088e-01 9.54404533e-01 3.23940217e-01 -6.75248027e-01
5.55142462e-01 8.86830837e-02 -6.88126504e-01 -3.35469782e-01
-1.54783297e+00 -6.06586754e-01 1.04299687e-01 2.20366139e-02
3.61270517e-01 -1.46508253e+00 -5.91340780e-01 -1.61270514e-01
-7.72407651e-01 -3.36902499e-01 4.03253734e-02 8.40620399e-01
-3.35308522e-01 2.03806221e-01 -8.74817312e-01 -8.22742105e-01
-7.54956245e-01 -1.03589034e+00 1.10848272e+00 -3.45296115e-02
1.13480322e-01 -1.01226413e+00 -4.43512723e-02 6.79875195e-01
8.36422324e-01 -6.49299979e-01 5.99383414e-01 -8.16464543e-01
-4.86619294e-01 -4.68078613e-01 -3.98267061e-01 5.23300588e-01
-1.17680423e-01 -2.85160422e-01 -1.02146041e+00 -5.71852505e-01
-3.34552497e-01 -5.26501656e-01 8.30001950e-01 1.92797244e-01
1.53102720e+00 -4.50962752e-01 -1.57218367e-01 5.60438871e-01
1.14296317e+00 -2.59379894e-01 3.94031346e-01 3.48751485e-01
5.43934822e-01 4.76889648e-02 7.56144047e-01 6.44469917e-01
7.23395109e-01 1.02863681e+00 3.49270329e-02 3.07242312e-02
-1.48226544e-01 -7.85106122e-01 7.80855119e-01 1.28955555e+00
-2.06810907e-01 -4.90896255e-01 -3.81076455e-01 3.12823743e-01
-2.00791693e+00 -7.02180088e-01 2.78210700e-01 2.27450919e+00
1.02310169e+00 3.11017752e-01 4.69068065e-02 -4.47603017e-01
4.44209516e-01 1.15700319e-01 -5.12717068e-01 -1.23300746e-01
3.67283911e-01 1.51014656e-01 5.03315508e-01 4.52728271e-01
-8.96859050e-01 8.93534660e-01 5.68243790e+00 1.15139782e+00
-1.22999561e+00 4.20154244e-01 5.45849562e-01 -3.47343743e-01
-5.00699043e-01 3.61001343e-02 -1.27492416e+00 6.15699530e-01
1.45023000e+00 -9.60361734e-02 6.72352374e-01 1.27689540e+00
4.02065396e-01 2.65142351e-01 -1.00713563e+00 7.54420459e-01
4.51384149e-02 -1.23568392e+00 1.09098442e-01 -6.22229688e-02
6.48581564e-01 3.25055361e-01 -2.01541875e-02 9.81542587e-01
4.92844582e-01 -5.38766623e-01 7.68217206e-01 3.08837146e-01
7.16461122e-01 -4.94361997e-01 7.44612098e-01 5.29883087e-01
-9.13035214e-01 -1.51812270e-01 -5.42079091e-01 -1.65473968e-01
5.00809610e-01 9.19408262e-01 -8.76807153e-01 6.49683654e-01
3.80793482e-01 6.73028648e-01 -2.44170904e-01 1.03487110e+00
-2.17804596e-01 1.03144121e+00 -3.93729508e-01 -2.14082584e-01
7.27144256e-02 -4.16986033e-04 6.11588806e-02 1.28841126e+00
5.36567271e-01 -2.64941782e-01 2.94679880e-01 4.15376127e-01
-3.70462686e-01 3.22790593e-01 1.57294601e-01 -3.64060640e-01
3.22219968e-01 1.35940981e+00 -2.78431267e-01 -3.98736715e-01
-9.05450642e-01 9.37149107e-01 5.68923235e-01 4.46480304e-01
-1.26526618e+00 -1.99255198e-01 6.55552208e-01 2.80128807e-01
4.85386282e-01 -3.39786500e-01 8.32170844e-02 -1.54482925e+00
-3.83644141e-02 -9.27022278e-01 3.08291435e-01 -4.46930289e-01
-1.37608290e+00 5.24691939e-01 -2.44543046e-01 -1.07087755e+00
-2.66931087e-01 -5.30621231e-01 -3.50929260e-01 6.70831084e-01
-1.96468353e+00 -1.25998902e+00 5.89662977e-03 3.08365136e-01
6.60830200e-01 -5.41079463e-03 6.93679810e-01 9.36075270e-01
-6.77997530e-01 1.09029257e+00 4.86531913e-01 -1.42169535e-01
8.73601079e-01 -1.01668692e+00 5.93074799e-01 6.30764425e-01
4.25454706e-01 8.48413706e-01 6.50008321e-01 -3.54446769e-01
-1.59381115e+00 -1.29943776e+00 7.06045687e-01 -2.58687407e-01
7.81515241e-01 -7.12023735e-01 -5.75056314e-01 7.00022459e-01
-1.31053701e-01 -1.40610054e-01 9.18170750e-01 7.88566828e-01
-3.92863482e-01 -3.55589718e-01 -4.81759101e-01 6.89208150e-01
1.04337502e+00 -6.67372346e-01 -4.35908943e-01 5.02643466e-01
8.60761642e-01 -4.85048622e-01 -9.56677735e-01 3.54508877e-01
8.51855934e-01 -5.12103677e-01 1.05648434e+00 -8.40596676e-01
6.98320031e-01 -1.52318820e-01 -1.78156018e-01 -1.38697886e+00
-3.80590975e-01 -5.03383934e-01 -5.88850021e-01 1.16783524e+00
7.34209418e-01 -5.69938302e-01 6.72625721e-01 5.83233297e-01
-3.39185417e-01 -9.10525680e-01 -4.17373717e-01 -6.61978781e-01
-4.71336782e-01 -6.24128699e-01 4.35863853e-01 6.60854459e-01
-4.07055207e-02 5.87881804e-01 -7.86885440e-01 1.33844078e-01
2.32225940e-01 3.05367261e-01 9.35736835e-01 -9.61076975e-01
-1.09067154e+00 -7.65396133e-02 2.24468246e-01 -1.96606982e+00
-3.11279297e-03 -9.22367334e-01 7.80379623e-02 -1.64059961e+00
2.91331679e-01 -6.37121379e-01 -7.66059339e-01 3.94853145e-01
-4.01746899e-01 1.47819370e-01 -3.61838192e-02 6.68966025e-02
-1.16507936e+00 7.90909648e-01 1.19396603e+00 2.78014332e-01
-9.74049978e-03 3.62652391e-01 -1.05322397e+00 4.77130562e-01
5.07138193e-01 -4.07698482e-01 -6.85231388e-01 -8.02718103e-01
6.96376860e-01 7.17914999e-02 4.09222953e-02 -6.68196440e-01
2.04918921e-01 3.28119099e-02 6.12051077e-02 -3.07374954e-01
3.15455973e-01 -7.07338333e-01 -2.20926583e-01 -1.11907413e-02
-6.59159660e-01 -5.48675917e-02 -1.24681018e-01 8.27069581e-01
-2.19880551e-01 -1.80336460e-01 2.57096976e-01 1.16374895e-01
-6.44636452e-01 4.05927598e-01 -4.46770713e-02 -8.77889171e-02
5.77169001e-01 3.09981972e-01 -4.36910957e-01 -7.11363137e-01
-4.49663103e-01 3.05111259e-01 -7.32965395e-02 5.20092487e-01
3.51789474e-01 -1.11520517e+00 -5.31107962e-01 7.18110576e-02
3.23559672e-01 -3.43617380e-01 2.99111813e-01 9.21169996e-01
-2.17324987e-01 4.95818496e-01 2.87160307e-01 -2.95907352e-03
-9.45438564e-01 3.33246350e-01 4.66026403e-02 -6.55779421e-01
-5.88490725e-01 8.33481669e-01 3.98599915e-02 -4.45383757e-01
4.98343140e-01 -2.77062863e-01 -5.83178662e-02 -2.58462578e-01
8.06492448e-01 1.30079597e-01 2.86035240e-01 2.23781765e-01
6.14211969e-02 -9.02263969e-02 -5.44857025e-01 -3.36999632e-02
1.55247486e+00 -2.77890623e-01 3.30619931e-01 3.60638738e-01
1.12880731e+00 3.46108943e-01 -1.22161746e+00 -4.73371387e-01
-3.49902064e-02 -5.44937372e-01 5.64004958e-01 -1.04566789e+00
-9.00497854e-01 6.77967846e-01 2.99776763e-01 5.33624738e-02
8.23378980e-01 -2.03033522e-01 1.47148860e+00 7.20039546e-01
5.14543056e-01 -1.29500151e+00 -1.33931544e-02 5.30353189e-01
5.50585568e-01 -1.23905158e+00 -3.36815196e-04 -2.96250373e-01
-6.86788380e-01 9.73730624e-01 4.10014927e-01 -8.05394053e-02
7.24416196e-01 3.03201377e-01 -6.18864372e-02 1.33758515e-01
-1.12637079e+00 -1.90481946e-01 5.26894093e-01 -1.08700112e-01
9.90479708e-01 4.91379984e-02 -5.16100883e-01 1.12265873e+00
-1.24974310e-01 4.48059738e-01 1.06331393e-01 7.12039649e-01
-4.22915101e-01 -1.21711898e+00 3.16620559e-01 7.91037083e-01
-7.45368123e-01 -3.96403402e-01 3.57683182e-01 3.15644652e-01
-4.30461645e-01 8.39700520e-01 -1.46515667e-01 -5.38652956e-01
2.59196907e-01 -3.51127870e-02 2.52147555e-01 -6.56587601e-01
-4.79833186e-01 1.55469865e-01 7.71654785e-01 -5.30503809e-01
-2.22817540e-01 -1.54428139e-01 -8.23146939e-01 -2.75903076e-01
-7.08638489e-01 4.29272354e-01 9.32425380e-01 9.59593236e-01
6.40340805e-01 7.75520384e-01 5.68879008e-01 -5.43386102e-01
-1.25959826e+00 -1.12062609e+00 -4.38250184e-01 4.00970668e-01
-6.27144519e-03 -6.46328092e-01 -8.56175050e-02 -2.06858128e-01] | [10.258197784423828, 5.769360065460205] |
6878764f-995c-4ddd-a396-aff43c0cbd61 | wide-activation-for-efficient-and-accurate | 1808.08718 | null | http://arxiv.org/abs/1808.08718v2 | http://arxiv.org/pdf/1808.08718v2.pdf | Wide Activation for Efficient and Accurate Image Super-Resolution | Keras-based implementation of WDSR, EDSR and SRGAN for single image super-resolution | ['Zhaowen Wang', 'Jianchao Yang', 'Xinchao Wang', 'Ning Xu', 'Yuchen Fan', 'Thomas Huang', 'Jiahui Yu'] | 2018-08-27 | null | null | null | null | ['multi-frame-super-resolution'] | ['computer-vision'] | [ 9.23140109e-01 4.60193425e-01 -1.05345547e-02 1.44847766e-01
-5.14781237e-01 -5.50331652e-01 5.84267557e-01 -1.37631786e+00
-3.43372732e-01 1.07992470e+00 3.71504486e-01 -4.64029759e-01
1.52319968e-01 -1.02053404e+00 -2.44257256e-01 -8.00235689e-01
-1.86576188e-01 2.47913137e-01 1.08373642e+00 -8.26467097e-01
3.62827748e-01 4.78425801e-01 -7.57620037e-01 3.28490555e-01
2.79656023e-01 7.56072462e-01 8.28854918e-01 1.43307066e+00
3.13710958e-01 1.18025446e+00 -9.37720895e-01 -4.85996045e-02
6.99623466e-01 -6.59816504e-01 -9.42489684e-01 -8.98547247e-02
7.45218039e-01 -9.04833734e-01 -1.14747345e+00 1.06777203e+00
6.94140613e-01 -2.74801046e-01 9.83646214e-02 -2.49451026e-01
-1.33329892e+00 9.85230505e-01 -8.84843469e-01 8.35816979e-01
2.44710460e-01 3.21121752e-01 2.66325206e-01 -4.23668832e-01
1.22516632e+00 1.15121639e+00 5.50696194e-01 1.10532224e+00
-1.35877395e+00 -9.45716500e-01 -1.74292719e+00 5.31934872e-02
-1.36887097e+00 -8.91666174e-01 -1.65155634e-01 1.06392717e+00
1.37055469e+00 6.48779154e-01 5.41950427e-02 5.48517346e-01
6.58418894e-01 -2.46068016e-01 2.51292419e+00 -2.97222286e-01
-1.89696312e-01 3.25966895e-01 -1.02315983e-02 4.73099679e-01
7.68537343e-01 4.01290417e-01 -6.06900632e-01 2.54327115e-02
1.78718340e+00 -4.18232262e-01 -8.36132094e-02 6.91366911e-01
-1.02454233e+00 5.71209371e-01 5.63894868e-01 3.40832204e-01
-4.41114306e-01 7.53421724e-01 -2.27916971e-01 1.24616337e+00
1.44591197e-01 4.94287789e-01 2.40076974e-01 3.48332018e-01
-9.44332659e-01 -3.69962871e-01 -1.26854420e-01 1.14447057e+00
6.18498266e-01 8.22439253e-01 -1.04678042e-01 -1.91193838e-02
2.11133048e-01 1.05611789e+00 4.13797468e-01 -1.68229878e+00
-3.62848431e-01 8.28563236e-03 3.10213178e-01 -4.22867328e-01
1.71419367e-01 5.00275977e-02 -9.15814400e-01 4.04419422e-01
-6.61000967e-01 8.89825821e-02 -1.36220813e+00 7.42002904e-01
4.23462659e-01 9.17926967e-01 9.91040289e-01 9.73854601e-01
1.30388463e+00 8.98251891e-01 -4.69345778e-01 -1.45526543e-01
1.91588056e+00 -8.28559875e-01 -6.86922848e-01 6.11301437e-02
-1.55167744e-01 -5.55476725e-01 1.22008212e-01 -2.04911321e-01
-1.49856532e+00 -3.02594393e-01 -1.41004360e+00 -5.26795626e-01
1.34029686e-01 -2.61234522e-01 3.90825272e-01 1.13382721e+00
-2.05694413e+00 -6.00272976e-03 -3.54960501e-01 -1.56945124e-01
5.22386611e-01 7.06954181e-01 -6.78612649e-01 -1.60908639e-01
-1.18529177e+00 8.94946516e-01 6.61608338e-01 -8.81896913e-01
-8.63604605e-01 -6.07368886e-01 -1.76754862e-01 -2.74541765e-01
-1.11132618e-02 -6.15703821e-01 6.31722212e-01 -8.34904984e-03
-2.16764760e+00 1.72335124e+00 6.74446821e-02 -1.28406549e+00
6.01137765e-02 8.27761054e-01 -6.50023580e-01 9.79355454e-01
-6.88232601e-01 3.24062854e-01 1.12082350e+00 -1.31111574e+00
-7.62289584e-01 -6.36584461e-01 -8.18453655e-02 -1.30285397e-01
3.55100513e-01 1.15597677e+00 8.66265118e-01 -4.08902206e-02
-1.81966890e-02 -4.42167878e-01 -4.27756488e-01 -7.08041549e-01
-4.04793173e-01 6.53796673e-01 1.96633255e+00 -9.99682426e-01
5.26629508e-01 -2.03316426e+00 -3.63278121e-01 -1.17154554e-01
6.16077662e-01 8.57656717e-01 -5.28768301e-01 4.96862411e-01
3.56524855e-01 9.97267216e-02 7.43080303e-02 -2.46858597e-01
-6.67870939e-01 3.78408611e-01 -8.25731814e-01 7.82187223e-01
-3.31486106e-01 1.09681773e+00 -1.97212204e-01 -6.00517690e-01
2.54306972e-01 8.79727006e-01 2.92593300e-01 2.36555010e-01
4.79383469e-01 1.07852125e+00 -3.71328264e-01 2.75879323e-01
1.56370580e+00 -3.83318722e-01 4.56920713e-02 5.15281707e-02
-5.22736192e-01 1.07630678e-01 -1.37316287e+00 1.28625917e+00
-3.18499953e-01 6.24903679e-01 3.65988731e-01 -7.58183420e-01
9.61410582e-01 7.50937819e-01 1.59970894e-01 -1.00000429e+00
-1.40242338e-01 5.41075647e-01 -6.95828378e-01 -9.17848647e-02
1.13184285e+00 -1.41232252e-01 7.87421316e-02 8.14965606e-01
1.57776579e-01 -2.94033557e-01 -5.50242245e-01 4.17135805e-01
1.51167715e+00 -5.99029481e-01 2.43771851e-01 -2.66951114e-01
9.27211821e-01 2.17812777e-01 1.54301897e-01 9.48665082e-01
1.14210241e-01 6.61862135e-01 -6.36783302e-01 -5.18923581e-01
-2.10844970e+00 -1.10524964e+00 -1.74376175e-01 3.59224617e-01
8.87093604e-01 1.47827521e-01 -7.86161184e-01 2.72237420e-01
-8.83597732e-01 6.99681103e-01 3.56654286e-01 2.22263098e-01
-1.16064596e+00 -6.01310551e-01 1.18110776e+00 -3.72288525e-01
1.64949763e+00 -1.07150626e+00 -1.09025908e+00 1.78653851e-01
2.17856586e-01 -2.00695252e+00 -2.15438500e-01 -6.88216031e-01
-9.14369404e-01 -9.78512764e-01 -6.66843832e-01 -1.12871730e+00
4.00065392e-01 1.12185562e+00 9.99897540e-01 4.01875943e-01
-8.07692289e-01 1.08421594e-01 -2.19998300e-01 1.78186461e-01
-1.29518735e+00 -7.12278366e-01 2.50881701e-03 -4.63903874e-01
-1.11973546e-02 -1.19649005e+00 -1.34319782e+00 3.14620525e-01
-8.72917116e-01 8.82672518e-02 3.83177102e-01 3.80285621e-01
4.92465287e-01 2.57992864e-01 1.61669478e-01 -9.00897443e-01
2.98532456e-01 3.35055947e-01 -1.04692781e+00 2.30557963e-01
-7.66652048e-01 -3.14640880e-01 2.39656031e-01 -2.38630548e-01
-1.30709827e+00 -4.72573966e-01 2.29997545e-01 -1.68448821e-01
-6.06220476e-02 -7.48472810e-01 1.07481241e+00 -1.48488724e+00
6.24071360e-01 1.28706026e+00 2.76994109e-01 1.60462052e-01
6.38742387e-01 1.50938308e+00 9.41605389e-01 4.17386174e-01
1.65264714e+00 1.14552236e+00 7.41078556e-02 -1.21194136e+00
6.00914210e-02 -7.72420242e-02 -8.41487125e-02 1.10661119e-01
9.71028507e-01 -1.62401462e+00 -6.90152109e-01 8.50433588e-01
-1.20975733e+00 -4.93679166e-01 -9.07388270e-01 -7.50702694e-02
-7.67551661e-01 3.88967305e-01 -1.27979589e+00 -9.99222159e-01
-1.24443448e+00 -6.78473711e-01 1.24863958e+00 9.79068518e-01
6.25228524e-01 -2.09631354e-01 3.53763103e-01 1.06118655e+00
1.35916376e+00 5.58983803e-01 2.76164740e-01 2.51751430e-02
-1.68461645e+00 2.55797189e-02 -6.61291718e-01 2.02432677e-01
1.70613542e-01 -6.95883811e-01 -4.00041223e-01 -6.28970921e-01
3.56448591e-01 -2.54165500e-01 1.01811492e+00 5.30198038e-01
-4.15258966e-02 -9.79539275e-01 2.59133857e-02 1.12901580e+00
2.45701504e+00 -5.72891459e-02 1.65942192e+00 3.78777057e-01
6.32518828e-02 -4.82576191e-01 -4.26468216e-02 2.30086416e-01
5.34216464e-01 9.08440948e-01 3.39284390e-01 -1.85276166e-01
-9.38664019e-01 1.54250666e-01 8.45005393e-01 7.72173822e-01
-5.92162549e-01 -1.59933150e-01 3.23864341e-01 1.13711759e-01
-9.23453748e-01 -1.61605310e+00 -5.28585374e-01 1.70499563e+00
1.05552673e+00 -8.27926815e-01 9.72209200e-02 -2.25298449e-01
9.72791851e-01 6.96118772e-01 -1.41874656e-01 -4.21629459e-01
-1.02725875e+00 1.04860377e+00 1.67439616e+00 4.51583505e-01
-1.58085644e-01 1.71137166e+00 8.02628994e+00 9.93263423e-01
-6.58664167e-01 1.30272448e+00 2.66373098e-01 5.07083476e-01
-7.20864773e-01 1.91455111e-01 -9.70625699e-01 2.65576184e-01
1.42688262e+00 -4.76515740e-01 1.43694413e+00 -7.51802698e-02
-1.96215868e-01 -6.10132962e-02 2.67098159e-01 1.13342476e+00
-1.17805697e-01 -2.38123560e+00 -3.74392152e-01 4.53688711e-01
1.28992498e+00 5.21445632e-01 4.02807266e-01 -5.70342779e-01
1.28579676e+00 -1.00811970e+00 1.97462022e-01 -2.95823008e-01
1.73433506e+00 -5.67425072e-01 4.47967917e-01 1.17590338e-01
-7.04845846e-01 -3.45924675e-01 -1.10041010e+00 6.94287777e-01
1.38143003e-01 1.79711692e-02 -6.96239233e-01 5.92137396e-01
6.40274286e-01 9.58304331e-02 -6.54006302e-02 2.35763878e-01
-3.72129902e-02 4.83423024e-01 -4.47832465e-01 7.45095611e-01
-4.13969764e-03 -4.05812711e-01 1.34381044e+00 5.59978545e-01
7.31948256e-01 7.43461013e-01 -5.00120401e-01 1.21282208e+00
-6.03243224e-02 -7.27423489e-01 -8.01739931e-01 -5.03447615e-02
5.51606894e-01 1.13770449e+00 -5.59937119e-01 -1.75315678e-01
3.48681360e-01 1.29166496e+00 -4.06971484e-01 6.49973676e-02
-1.59332410e-01 -1.03331082e-01 7.38584697e-01 3.73847574e-01
7.36132383e-01 -6.03121340e-01 1.42118439e-01 -1.03440332e+00
-8.22449863e-01 -1.03216410e+00 5.33919968e-02 -9.59450006e-01
-4.77073491e-01 1.18305576e+00 -5.95461309e-01 -7.41732597e-01
2.29895011e-01 1.41622379e-01 -5.12803435e-01 7.83058524e-01
-2.36213589e+00 -1.61665082e+00 -2.97927558e-01 1.23045242e+00
-2.15498433e-02 -7.13865221e-01 8.59517038e-01 -1.46745250e-01
4.48279470e-01 2.10207567e-01 6.02007389e-01 -3.56187850e-01
5.00447810e-01 -7.36052454e-01 8.60484838e-01 1.29260838e+00
-1.53817674e-02 -1.39335431e-02 1.02869618e+00 -6.24269009e-01
-2.32371998e+00 -6.86692357e-01 1.79668605e-01 -5.86491704e-01
3.50311399e-01 -1.81677062e-02 -3.78522009e-01 1.12782753e+00
1.03527689e+00 1.66867197e-01 -6.62576593e-03 -1.73949742e+00
-3.30589682e-01 -3.09536576e-01 -2.44367313e+00 3.52215111e-01
1.17469895e+00 -9.80589092e-01 -9.03107300e-02 5.76255798e-01
1.40663362e+00 -8.28371584e-01 -1.35464299e+00 -2.74183005e-01
-1.67703461e-02 -1.08245480e+00 1.62387240e+00 -6.99335858e-02
3.79451931e-01 -4.65052009e-01 -5.58188319e-01 -8.25348437e-01
-4.10125613e-01 -1.74031878e+00 -2.74423480e-01 1.04710913e+00
-4.99251276e-01 -1.44650948e+00 5.18396378e-01 -3.12447727e-01
1.97931632e-01 6.93819225e-01 -1.74144542e+00 -8.23457539e-01
-5.96969664e-01 8.88846934e-01 1.07556617e+00 6.02446496e-01
-6.19568467e-01 1.45793930e-01 -1.28341150e+00 7.43034959e-01
1.58511937e+00 7.54841492e-02 1.14460409e+00 -5.04416108e-01
2.27487430e-01 3.64883989e-01 -6.83484197e-01 -4.95855689e-01
-2.45311007e-01 -5.37036836e-01 -5.18137991e-01 -9.36838865e-01
1.55662954e-01 -3.57040554e-01 -2.01125383e-01 -4.88799065e-02
5.82063854e-01 1.52881289e+00 3.56430084e-01 6.77509427e-01
-4.19564724e-01 -3.28967333e-01 1.58593929e+00 -2.02637672e-01
4.22376692e-02 -3.94495994e-01 -7.57625997e-01 -7.65444338e-01
5.93450069e-01 -8.26994538e-01 -2.98522174e-01 -1.03114389e-01
2.22791180e-01 5.35558164e-01 6.36588991e-01 -7.30389535e-01
3.98512006e-01 -3.42613310e-01 -5.64734042e-01 1.26285374e-01
2.52313703e-01 -8.43557775e-01 8.37511420e-01 6.11692369e-01
-1.13715574e-01 3.07621788e-02 -5.97033739e-01 4.81709599e-01
2.53043342e-02 3.57856035e-01 1.39470673e+00 -6.84516490e-01
-9.03565586e-01 2.45964110e-01 -2.08731264e-01 -1.96076378e-01
9.80707169e-01 -1.04001844e+00 -1.21039677e+00 -3.18188280e-01
-3.62517387e-02 -3.77692193e-01 9.62026596e-01 2.41297305e-01
1.30112112e+00 -1.17665923e+00 -1.64695752e+00 4.61040854e-01
-7.82856703e-01 -2.41614774e-01 3.01699519e-01 6.66723192e-01
-1.06293666e+00 3.69204909e-01 -8.98747027e-01 -3.60705033e-02
-1.89809048e+00 2.81988591e-01 4.47178155e-01 -5.98024845e-01
-1.61978292e+00 8.29488337e-01 -6.24843776e-01 3.40250283e-01
-8.08397353e-01 6.84068918e-01 -3.31396133e-01 -1.14999640e+00
1.45371509e+00 4.75771934e-01 -6.64788425e-01 -9.47095156e-01
-1.70492887e-01 9.57770884e-01 -2.34770685e-01 1.83057487e-01
1.81846476e+00 -8.98348749e-01 -8.91311646e-01 -7.44742811e-01
3.59252542e-01 -8.83501694e-02 -7.47295618e-01 -4.56464171e-01
-3.64142239e-01 -7.48231471e-01 6.70203984e-01 -6.18329525e-01
-1.36480093e+00 1.63939670e-02 9.91303980e-01 1.57796666e-01
1.58097565e+00 5.36973635e-03 1.81509161e+00 1.86984520e-02
8.02152872e-01 -8.51724386e-01 -1.01426885e-01 -8.38347748e-02
6.36638403e-01 -9.89801824e-01 2.36056387e-01 -3.82724196e-01
-5.87424338e-01 9.44021046e-01 -2.61823952e-01 -8.40987563e-01
1.72117174e-01 7.51601934e-01 -6.12007491e-02 -3.92909020e-01
-8.06497037e-01 3.63008499e-01 -7.02598155e-01 9.62678850e-01
-7.91857481e-01 1.02472804e-01 -2.38737360e-01 -7.69004107e-01
-3.08126003e-01 4.52924341e-01 1.51403236e+00 5.12599766e-01
-3.34283084e-01 -7.37032950e-01 -1.72353327e-01 -3.89644444e-01
-1.00797963e+00 -2.80150682e-01 1.03990741e-01 3.93266045e-02
-6.88266695e-01 1.12699199e+00 2.28709310e-01 -7.37598598e-01
-1.66972242e-02 -7.95304537e-01 6.23304367e-01 -1.43154502e-01
-7.44544029e-01 -1.72763959e-01 1.58789113e-01 -8.25625658e-01
-1.04811299e+00 1.94372311e-02 -1.37017095e+00 -1.10710192e+00
-2.84954935e-01 -2.07910448e-01 5.84629714e-01 5.63253999e-01
6.10249758e-01 -7.86101520e-02 1.38170803e+00 -4.66171168e-02
-6.94630802e-01 -3.33314002e-01 -1.30244982e+00 -7.54323974e-02
8.10425162e-01 4.53263193e-01 -4.20655340e-01 -1.77609101e-01] | [4.337944030761719, 8.03076457977295] |
e255df31-f060-427d-b3fb-a1bbd5f3c117 | pfedsim-similarity-aware-model-aggregation | 2305.15706 | null | https://arxiv.org/abs/2305.15706v1 | https://arxiv.org/pdf/2305.15706v1.pdf | pFedSim: Similarity-Aware Model Aggregation Towards Personalized Federated Learning | The federated learning (FL) paradigm emerges to preserve data privacy during model training by only exposing clients' model parameters rather than original data. One of the biggest challenges in FL lies in the non-IID (not identical and independently distributed) data (a.k.a., data heterogeneity) distributed on clients. To address this challenge, various personalized FL (pFL) methods are proposed such as similarity-based aggregation and model decoupling. The former one aggregates models from clients of a similar data distribution. The later one decouples a neural network (NN) model into a feature extractor and a classifier. Personalization is captured by classifiers which are obtained by local training. To advance pFL, we propose a novel pFedSim (pFL based on model similarity) algorithm in this work by combining these two kinds of methods. More specifically, we decouple a NN model into a personalized feature extractor, obtained by aggregating models from similar clients, and a classifier, which is obtained by local training and used to estimate client similarity. Compared with the state-of-the-art baselines, the advantages of pFedSim include: 1) significantly improved model accuracy; 2) low communication and computation overhead; 3) a low risk of privacy leakage; 4) no requirement for any external public information. To demonstrate the superiority of pFedSim, extensive experiments are conducted on real datasets. The results validate the superb performance of our algorithm which can significantly outperform baselines under various heterogeneous data settings. | ['Shui Yu', 'Jessie Hui Wang', 'Gang Liu', 'Yipeng Zhou', 'Jiahao Tan'] | 2023-05-25 | null | null | null | null | ['personalized-federated-learning'] | ['methodology'] | [ 1.34755746e-02 -1.67804375e-01 -3.99304360e-01 -5.80796778e-01
-8.47530842e-01 -7.38659024e-01 5.65279603e-01 -2.50464268e-02
-3.30818236e-01 5.47911823e-01 1.43348649e-01 -1.82655215e-01
-2.13644624e-01 -6.61608338e-01 -8.65684032e-01 -1.05180502e+00
1.73894361e-01 3.32089365e-01 1.83705568e-01 2.61780143e-01
-9.86233652e-02 5.35025954e-01 -1.35207927e+00 5.00689149e-01
8.15110266e-01 1.50943625e+00 -2.66895980e-01 -2.03269087e-02
-5.81020601e-02 8.52836430e-01 -2.28032157e-01 -7.04775751e-01
7.07855821e-01 -1.27240330e-01 -6.19778574e-01 -1.80062965e-01
2.43267864e-01 -7.25055218e-01 -4.12288606e-01 1.07755566e+00
3.78693014e-01 -9.96572375e-02 1.94671825e-01 -1.73792744e+00
-6.87621474e-01 7.21450627e-01 -3.46000761e-01 -3.29963535e-01
-1.54854819e-01 2.03978911e-01 8.74521375e-01 -5.99750340e-01
4.33198929e-01 1.03607011e+00 7.48476803e-01 6.62795603e-01
-1.42816746e+00 -9.39357460e-01 1.22963846e-01 3.44741225e-01
-1.49567914e+00 -5.72212636e-01 6.94131911e-01 -9.11895260e-02
4.26947951e-01 7.15786934e-01 1.42771915e-01 1.12135637e+00
-1.08608902e-01 1.03701937e+00 1.07214701e+00 2.57256906e-02
6.61238968e-01 7.07984328e-01 2.53619224e-01 2.15218097e-01
3.33477110e-01 1.44109264e-01 -4.53333408e-01 -1.04296529e+00
2.67782927e-01 6.18437529e-01 -4.49580938e-01 -7.89732754e-01
-6.88452244e-01 7.23305762e-01 4.05207187e-01 -1.11610830e-01
-4.04811680e-01 -5.72825111e-02 6.06339097e-01 3.73598516e-01
1.34182498e-01 -2.83634901e-01 -6.93374932e-01 2.71290839e-01
-5.38344979e-01 3.55781764e-01 9.16423917e-01 1.08539307e+00
9.56916809e-01 -4.63012099e-01 -1.92774355e-01 5.26040554e-01
3.00725371e-01 2.94656992e-01 6.10194027e-01 -8.81666183e-01
6.47966623e-01 8.69575918e-01 1.20381005e-01 -8.21214020e-01
1.65223524e-01 -3.80173117e-01 -9.27243173e-01 2.04206109e-02
3.82854827e-02 -1.43561810e-01 -1.06551789e-01 2.05320740e+00
6.88441515e-01 2.55212486e-01 3.24307919e-01 7.16871083e-01
2.47473568e-01 3.60536009e-01 -6.26968741e-02 -1.98541313e-01
1.17692399e+00 -1.06173503e+00 -3.92015070e-01 3.26307952e-01
6.55918598e-01 -1.88317165e-01 8.39066923e-01 3.27898651e-01
-7.56292939e-01 -7.85420462e-02 -8.02707851e-01 5.98848239e-02
-3.08340013e-01 -1.81576774e-01 5.92053294e-01 8.23627114e-01
-8.28562915e-01 3.88919324e-01 -7.67446101e-01 -1.34224385e-01
9.52220023e-01 6.51322007e-01 -7.48229384e-01 -5.32168224e-02
-9.76100147e-01 2.61226863e-01 2.59389222e-01 -2.82662541e-01
-8.55016887e-01 -9.82141018e-01 -5.29067099e-01 3.58692467e-01
3.42114836e-01 -6.86530888e-01 1.37525356e+00 -8.66804540e-01
-1.35608637e+00 4.83101189e-01 -8.12363029e-02 -6.61183476e-01
8.65157604e-01 5.64808547e-02 -2.72261620e-01 -1.00094937e-01
-3.98250371e-01 -1.94479480e-01 4.90061611e-01 -1.53512037e+00
-1.08736622e+00 -7.12856293e-01 -8.60150382e-02 -2.06455529e-01
-9.02716994e-01 2.03899622e-01 -4.05439615e-01 -2.76789665e-01
-8.60616267e-02 -7.55048871e-01 -5.74817732e-02 3.75597745e-01
-4.51121688e-01 -2.18347043e-01 1.16112828e+00 -5.56564391e-01
1.28967607e+00 -2.15912676e+00 -3.74060154e-01 4.94020075e-01
4.13565874e-01 5.10726035e-01 -1.12318642e-01 6.60499871e-01
1.14559747e-01 1.05715752e-01 -1.88418090e-01 -7.78321087e-01
3.45411092e-01 2.20631599e-01 -5.86914957e-01 5.95295310e-01
-3.35492134e-01 8.54043841e-01 -4.98041302e-01 -4.64178443e-01
-1.00817144e-01 3.86796892e-01 -5.34996331e-01 4.92518663e-01
-1.21137194e-01 3.15873891e-01 -6.11374378e-01 6.06235087e-01
1.30249476e+00 -3.79186392e-01 3.90895784e-01 -3.66522640e-01
1.01081856e-01 6.29456788e-02 -1.28211617e+00 1.23697948e+00
-3.35654825e-01 -2.45639697e-01 5.05792141e-01 -8.35965633e-01
7.55005240e-01 5.20603418e-01 5.37708461e-01 -4.95361626e-01
1.13481104e-01 1.36717886e-01 -4.87572968e-01 -3.59878391e-01
-1.72617316e-01 1.80272877e-01 -5.72252907e-02 1.03581965e+00
-2.28574827e-01 8.80873442e-01 -5.00387788e-01 4.49467637e-02
1.21086061e+00 -3.06370437e-01 5.61000817e-02 -1.60766780e-01
7.40286291e-01 -4.73188758e-01 1.05511403e+00 7.49102652e-01
-3.22183818e-01 3.32981467e-01 3.87470305e-01 -5.74391127e-01
-6.65178835e-01 -9.73225772e-01 -6.90650614e-03 1.09974408e+00
1.76253632e-01 -4.22364801e-01 -8.83586466e-01 -1.06570554e+00
5.51912725e-01 6.29402339e-01 -7.23496616e-01 -2.91389912e-01
-4.48589325e-01 -5.76403201e-01 5.94987452e-01 4.44899619e-01
6.64162159e-01 -4.51667845e-01 -3.78497988e-01 8.00860673e-02
-1.81767659e-03 -8.86431873e-01 -6.67258620e-01 -2.37899013e-02
-6.44916654e-01 -9.12229061e-01 -1.05843894e-01 -4.39706087e-01
6.54608667e-01 4.63144451e-01 5.67657769e-01 -6.19955696e-02
1.39431447e-01 2.18067467e-01 -1.20623387e-01 -3.45309049e-01
-2.81321794e-01 1.20893627e-01 1.61262706e-01 8.82737219e-01
5.47240973e-01 -9.87525642e-01 -7.52301812e-01 4.72742081e-01
-1.18010211e+00 -6.55003861e-02 6.07675970e-01 7.73123801e-01
6.04936540e-01 -8.76324773e-02 5.12615800e-01 -1.22768021e+00
6.09710872e-01 -8.36742520e-01 -6.62989855e-01 7.80587256e-01
-1.01623917e+00 2.09287722e-02 1.14895546e+00 -5.77306569e-01
-1.23354053e+00 1.86312616e-01 3.78417611e-01 -6.19399130e-01
-9.75376219e-02 2.38322932e-02 -7.30025053e-01 -2.58766741e-01
3.65236044e-01 5.03878057e-01 3.89855385e-01 -9.28860605e-01
2.52846807e-01 1.24790347e+00 5.82411528e-01 -6.07137620e-01
8.93095613e-01 7.03089654e-01 -1.99076220e-01 1.21493340e-01
-2.56946951e-01 -2.18788877e-01 -1.48090079e-01 3.27891201e-01
1.64890483e-01 -8.98415923e-01 -1.31888425e+00 7.46378541e-01
-9.36076343e-01 -5.73226400e-02 -1.99290678e-01 3.17208856e-01
-3.28454465e-01 6.00781381e-01 -6.33012891e-01 -8.79181325e-01
-8.95535290e-01 -9.19021428e-01 5.47036529e-01 4.27396387e-01
1.62625432e-01 -7.37377763e-01 -4.24521156e-02 5.86441696e-01
8.00498962e-01 2.43157282e-01 9.41554248e-01 -1.24304402e+00
-6.23562336e-01 -5.27900696e-01 -3.16878557e-01 5.03892422e-01
3.16108614e-01 -3.23622555e-01 -1.14770710e+00 -6.19159818e-01
3.66743952e-01 -3.08023840e-01 2.60683060e-01 -4.22477633e-01
1.44303584e+00 -1.04443383e+00 -3.55380863e-01 8.80232573e-01
1.53143156e+00 1.03356853e-01 2.26429895e-01 2.42329404e-01
7.06333518e-01 4.69615191e-01 2.17115879e-01 9.71529126e-01
6.14265442e-01 6.95009828e-01 6.93344712e-01 9.56163183e-02
3.40293914e-01 -4.32510525e-01 2.77009696e-01 5.31105936e-01
5.46378553e-01 -2.24859323e-02 -5.34838855e-01 5.06864429e-01
-2.21181965e+00 -8.58694196e-01 1.98410064e-01 2.37239408e+00
8.97886276e-01 -4.93249953e-01 2.83614159e-01 -1.05593115e-01
7.29872882e-01 -1.22532696e-01 -1.10022068e+00 -2.60127574e-01
-7.97621831e-02 -7.83441141e-02 6.46926463e-01 1.54071106e-02
-8.60575616e-01 4.07821238e-01 4.40375900e+00 7.94618428e-01
-1.10487485e+00 4.13152754e-01 5.61894655e-01 -3.36072296e-01
-4.50555742e-01 2.18808696e-01 -7.12331295e-01 9.20261204e-01
8.33328724e-01 -5.63768208e-01 7.61185706e-01 1.29029262e+00
-1.58348262e-01 4.84984457e-01 -1.48134613e+00 9.51885998e-01
-2.62210250e-01 -1.43693900e+00 2.07930163e-01 2.80104280e-01
6.53676689e-01 2.01880544e-01 1.36079356e-01 1.45353973e-01
5.30551970e-01 -6.64290905e-01 5.89293838e-01 6.43946528e-01
5.97915530e-01 -1.00201893e+00 6.87039673e-01 6.70996130e-01
-1.01069725e+00 -6.00258648e-01 -3.63381892e-01 4.44531798e-01
-2.42744744e-01 3.04086924e-01 -4.56064135e-01 8.83857727e-01
9.09411073e-01 2.25344107e-01 -3.80261987e-01 8.63552511e-01
3.02129537e-01 5.92799008e-01 -6.02595568e-01 2.57277727e-01
7.10045965e-03 -1.90864742e-01 1.60422519e-01 9.17216480e-01
2.55278535e-02 -2.45744716e-02 -1.36893829e-02 8.74910057e-01
-3.94105017e-01 2.19866231e-01 -2.10418075e-01 4.99923706e-01
1.10697794e+00 1.18274021e+00 2.99211890e-01 -2.54745394e-01
-5.70379734e-01 9.96499717e-01 5.21015763e-01 3.73115987e-01
-6.08720541e-01 -4.08248276e-01 1.14363599e+00 -3.04418448e-02
3.31338048e-01 4.05881435e-01 -2.17158511e-01 -1.16280973e+00
5.51801920e-01 -1.07562339e+00 7.79139578e-01 2.48394511e-03
-1.79523587e+00 6.50814593e-01 -2.66255736e-01 -1.25807369e+00
-8.75509381e-02 2.20018439e-02 -6.54999554e-01 9.64783430e-01
-1.49249983e+00 -1.48826253e+00 -3.66509050e-01 1.10712910e+00
-3.12765718e-01 -2.67511308e-01 9.71784711e-01 3.71037722e-01
-8.92601132e-01 1.49785089e+00 7.52223313e-01 1.75390448e-02
6.72800243e-01 -7.01569080e-01 2.81589299e-01 7.16262341e-01
-2.87207037e-01 8.89800012e-01 1.29709199e-01 -3.61667991e-01
-1.84463561e+00 -1.49741244e+00 8.02106738e-01 -3.08528394e-01
3.97849977e-01 -6.06955707e-01 -1.15932453e+00 9.13519681e-01
-2.97540069e-01 5.41816473e-01 1.06808448e+00 -4.47755665e-01
-9.41803753e-01 -8.97331536e-01 -1.89013612e+00 3.72574300e-01
9.43747878e-01 -5.56769013e-01 -4.73413356e-02 2.21412137e-01
7.49562681e-01 -5.06617799e-02 -1.08801150e+00 2.90255666e-01
6.72746062e-01 -1.26256323e+00 7.15534687e-01 -9.02577937e-01
-3.19887519e-01 -2.16849521e-01 -4.44879919e-01 -6.27274156e-01
-2.43327036e-01 -1.13708389e+00 -6.02677345e-01 1.70520318e+00
1.98445227e-02 -1.40355015e+00 9.55011249e-01 1.54535031e+00
3.15524340e-01 -8.91151130e-01 -1.10802543e+00 -9.50941384e-01
-5.96445054e-02 -9.23679620e-02 1.60491896e+00 1.00703895e+00
-6.10208437e-02 -3.65053624e-01 -4.87743884e-01 3.88646126e-01
1.00471592e+00 4.48401660e-01 1.17610347e+00 -1.22289491e+00
-4.55089420e-01 -2.34089226e-01 -1.97939947e-01 -7.43850231e-01
1.99365139e-01 -8.94703329e-01 -5.15183091e-01 -9.46174502e-01
5.13957858e-01 -7.30215490e-01 -7.36149311e-01 9.53844786e-01
1.17943622e-01 -3.02547067e-01 2.26468608e-01 6.11368775e-01
-5.79403341e-01 6.14616573e-01 4.31033731e-01 2.25738347e-01
-1.44903779e-01 3.20315927e-01 -1.10034001e+00 2.32662156e-01
8.20617735e-01 -7.14856327e-01 -4.82005328e-01 -4.17952359e-01
-4.89122003e-01 -1.08366236e-01 4.45630670e-01 -8.11164141e-01
6.35735869e-01 -3.71679604e-01 -1.76288653e-03 -2.22183108e-01
-3.95376422e-02 -1.40126777e+00 7.83882141e-01 5.88336825e-01
-3.12990278e-01 -2.06732467e-01 -2.87231207e-01 6.59074903e-01
-2.82593146e-02 1.84278965e-01 7.43962109e-01 1.85388058e-01
-1.11769751e-01 7.69343019e-01 2.93074042e-01 -3.51311445e-01
1.27323592e+00 -1.69953685e-02 -6.30031884e-01 -3.25806737e-01
-1.56620026e-01 3.91923308e-01 8.50660503e-01 3.04986894e-01
1.51890829e-01 -1.41575527e+00 -7.05501020e-01 5.17607212e-01
2.37379223e-01 1.92826949e-02 4.49369371e-01 8.03557456e-01
-4.33487147e-02 3.21664393e-01 3.91620537e-03 -2.18062907e-01
-1.31718814e+00 9.16326046e-01 2.54234672e-01 -2.73307353e-01
-4.90940094e-01 5.65938771e-01 3.77686292e-01 -7.32488990e-01
6.96935415e-01 -3.72704566e-02 4.87788796e-01 -3.39845568e-01
8.85614097e-01 5.00996947e-01 1.83705479e-01 -6.39635205e-01
-5.64047873e-01 1.25781968e-01 -4.99252319e-01 2.47940421e-01
1.59983706e+00 -2.68392622e-01 -3.04038286e-01 -1.60247281e-01
1.62252188e+00 -8.19746181e-02 -1.51076186e+00 -9.06976998e-01
-1.66048050e-01 -7.30490327e-01 -1.37473062e-01 -9.03168321e-01
-1.34563684e+00 4.80115592e-01 7.45618224e-01 5.90563081e-02
1.29057634e+00 -1.16028212e-01 1.25197279e+00 2.22638980e-01
6.12790942e-01 -8.06327641e-01 -5.13663828e-01 -1.35035992e-01
7.38460243e-01 -9.02372658e-01 -2.28496552e-01 -1.37274235e-01
-5.96883476e-01 8.23345482e-01 5.68540871e-01 1.49109244e-01
6.85344994e-01 2.43846998e-01 9.94242728e-02 2.79295534e-01
-1.05341685e+00 5.55842638e-01 -2.05604240e-01 6.03801727e-01
-5.85959256e-01 -5.11219688e-02 5.56195574e-03 1.46745324e+00
1.18150435e-01 2.34538525e-01 -4.51927669e-02 1.18415129e+00
2.47958884e-01 -1.37952685e+00 -2.02863410e-01 4.57262248e-01
-5.52200794e-01 3.84217173e-01 -4.03844059e-01 2.80111551e-01
1.25332907e-01 8.44236553e-01 -1.92564592e-01 -7.12140024e-01
2.80828774e-01 7.60124475e-02 -2.08763495e-01 -5.56626692e-02
-1.13476217e+00 -3.86200130e-01 -3.18605870e-01 -8.86080563e-01
2.29629919e-01 -7.78835118e-01 -1.12755001e+00 -6.21638954e-01
-3.74174893e-01 3.72888327e-01 7.00550258e-01 5.79438031e-01
1.15382099e+00 -4.29347634e-01 1.43098819e+00 -3.90034527e-01
-1.40847683e+00 -5.61028779e-01 -6.19375527e-01 6.95737720e-01
3.68939757e-01 -1.90922812e-01 -6.60268903e-01 -1.33396700e-01] | [5.85145902633667, 6.441000938415527] |
204b86e1-ddb6-4fdd-9b30-b76c3676db11 | holistic-image-manipulation-detection-using | 2104.05693 | null | https://arxiv.org/abs/2104.05693v1 | https://arxiv.org/pdf/2104.05693v1.pdf | Holistic Image Manipulation Detection using Pixel Co-occurrence Matrices | Digital image forensics aims to detect images that have been digitally manipulated. Realistic image forgeries involve a combination of splicing, resampling, region removal, smoothing and other manipulation methods. While most detection methods in literature focus on detecting a particular type of manipulation, it is challenging to identify doctored images that involve a host of manipulations. In this paper, we propose a novel approach to holistically detect tampered images using a combination of pixel co-occurrence matrices and deep learning. We extract horizontal and vertical co-occurrence matrices on three color channels in the pixel domain and train a model using a deep convolutional neural network (CNN) framework. Our method is agnostic to the type of manipulation and classifies an image as tampered or untampered. We train and validate our model on a dataset of more than 86,000 images. Experimental results show that our approach is promising and achieves more than 0.99 area under the curve (AUC) evaluation metric on the training and validation subsets. Further, our approach also generalizes well and achieves around 0.81 AUC on an unseen test dataset comprising more than 19,740 images released as part of the Media Forensics Challenge (MFC) 2020. Our score was highest among all other teams that participated in the challenge, at the time of announcement of the challenge results. | ['B. S. Manjunath', 'Shivkumar Chandrasekaran', 'Tajuddin Manhar Mohammed', 'Michael Goebel', 'Lakshmanan Nataraj'] | 2021-04-12 | null | null | null | null | ['image-manipulation-detection', 'image-forensics'] | ['computer-vision', 'computer-vision'] | [ 5.03000379e-01 -4.33824599e-01 1.49695590e-01 -3.31231020e-02
-1.06903636e+00 -6.56803668e-01 5.96736550e-01 3.09469908e-01
-5.44134617e-01 2.33985677e-01 -1.14451669e-01 -5.59462845e-01
4.27757263e-01 -7.39670277e-01 -9.20636594e-01 -4.61050421e-01
-2.15750575e-01 -1.47390276e-01 2.74133563e-01 9.06012058e-02
6.86299503e-01 7.01811194e-01 -1.23761570e+00 8.15909863e-01
4.28045541e-01 9.83239770e-01 -3.17992628e-01 8.59693408e-01
3.07814926e-01 9.62005138e-01 -1.10511065e+00 -8.64234746e-01
3.99064451e-01 -2.14389607e-01 -7.07599998e-01 3.26652288e-01
6.77911997e-01 -7.02148080e-01 -5.99288464e-01 1.19184065e+00
4.80234921e-01 -3.42312753e-01 3.32397640e-01 -1.14936161e+00
-7.44569480e-01 1.43805936e-01 -8.85588586e-01 8.78169477e-01
1.83098793e-01 5.01488268e-01 4.53097790e-01 -8.43915820e-01
6.69124007e-01 1.03798044e+00 9.64808941e-01 1.32638782e-01
-9.75215673e-01 -7.21743643e-01 -5.26371360e-01 5.70721447e-01
-1.29578960e+00 -3.60892981e-01 7.93180883e-01 -4.18903381e-01
6.86922133e-01 6.44375533e-02 2.04813078e-01 1.17129040e+00
3.18923324e-01 8.38602364e-01 1.36440098e+00 -5.07482052e-01
7.88063183e-02 1.83883752e-03 4.95575294e-02 6.46004200e-01
4.73995268e-01 3.33549678e-02 -5.16905844e-01 -4.32576060e-01
4.83516127e-01 -1.53527046e-02 -1.85415134e-01 1.08546861e-01
-1.10863543e+00 9.46361542e-01 2.87298486e-02 2.66882330e-01
-4.49273109e-01 1.55277133e-01 7.68363595e-01 3.31912816e-01
1.81377128e-01 5.43121815e-01 3.20863314e-02 -1.39331132e-01
-1.26388037e+00 2.66558826e-01 4.90226686e-01 4.41947609e-01
3.69799376e-01 -6.24123663e-02 -9.82400924e-02 6.04666889e-01
-3.19561273e-01 4.33341384e-01 3.05217862e-01 -9.77572918e-01
5.89902401e-01 3.26926231e-01 -2.27593184e-02 -1.54270816e+00
-1.42174304e-01 -2.28252128e-01 -6.61300123e-01 4.29274440e-01
5.52892506e-01 -3.24890725e-02 -9.05126989e-01 1.00546920e+00
1.81351095e-01 2.35831425e-01 -2.16973275e-01 7.22791553e-01
5.83677948e-01 2.80188978e-01 -2.01114770e-02 8.02869797e-02
1.59990561e+00 -5.89730859e-01 -5.80313504e-01 -1.83371101e-02
4.78768408e-01 -8.91788244e-01 6.63911521e-01 9.21397269e-01
-9.29980576e-01 -4.79098499e-01 -1.24656570e+00 -1.24700349e-02
-6.74491823e-01 1.74820900e-03 3.46504480e-01 1.19171548e+00
-7.46806324e-01 7.84538925e-01 -5.25866032e-01 -1.86755173e-02
9.84469712e-01 8.60265121e-02 -5.69110036e-01 -2.77386814e-01
-9.38751578e-01 6.71860337e-01 3.24399829e-01 -1.51942000e-01
-9.66222882e-01 -4.84979510e-01 -5.46373188e-01 -1.46514356e-01
3.08927655e-01 8.65737125e-02 8.31069529e-01 -9.10252452e-01
-6.38229191e-01 1.31674373e+00 2.96726733e-01 -8.21708798e-01
8.85369360e-01 -2.04260014e-02 -7.85816193e-01 7.24247098e-01
2.73445547e-01 3.47077638e-01 1.20126975e+00 -1.18601966e+00
-7.92438805e-01 -3.89410526e-01 -1.05558194e-01 -6.15764737e-01
-5.05297422e-01 5.09710014e-01 -5.03844440e-01 -9.26102340e-01
-2.27638438e-01 -6.10708296e-01 1.76235452e-01 -6.05090931e-02
-4.68029916e-01 2.20850348e-01 1.35910904e+00 -1.12915289e+00
1.12879109e+00 -2.23751688e+00 -7.72184551e-01 4.93845642e-02
4.28556383e-01 7.32449055e-01 -6.00937530e-02 4.41333145e-01
-2.40175605e-01 2.52731889e-01 -5.34769952e-01 -2.72751093e-01
-1.65411830e-01 -2.59987503e-01 -5.48312843e-01 1.17859185e+00
3.46639037e-01 7.84247220e-01 -9.17784095e-01 -5.29276311e-01
3.15053254e-01 4.47044313e-01 -1.30127877e-01 -1.45990014e-01
1.84187472e-01 7.00303912e-03 -6.71420544e-02 1.10795844e+00
1.12495697e+00 -8.16019252e-02 8.59284773e-02 -2.89201051e-01
1.76087037e-01 -1.52102917e-01 -9.03649807e-01 1.25802815e+00
-2.26935986e-02 1.15631688e+00 1.18728288e-01 -1.07565355e+00
7.05289185e-01 1.87911436e-01 4.33520049e-01 -9.43570971e-01
2.67401904e-01 2.37231925e-01 -2.13234618e-01 -8.42836618e-01
7.15531766e-01 1.63826253e-02 -1.06718957e-01 7.94524968e-01
-1.81996990e-02 3.37240994e-01 2.22450033e-01 1.83058679e-01
1.59214497e+00 -2.77380228e-01 5.90185709e-02 1.22337133e-01
3.87973666e-01 -7.63284266e-02 3.78998637e-01 1.04372466e+00
-7.22963333e-01 8.44806910e-01 6.31688356e-01 -6.28791809e-01
-1.17641604e+00 -8.54675412e-01 -1.05678037e-01 6.43530250e-01
1.14662401e-01 -2.72529602e-01 -9.11894977e-01 -1.00967860e+00
5.69127835e-02 6.18847430e-01 -8.32441568e-01 -1.21440344e-01
-8.63135874e-01 -8.52794409e-01 1.33948588e+00 3.35433960e-01
8.31829548e-01 -1.13879502e+00 -8.77612591e-01 5.07825101e-03
-2.42599770e-01 -1.50077331e+00 -4.94944841e-01 -2.44661689e-01
-4.49746728e-01 -1.61016822e+00 -5.48232794e-01 -5.06080866e-01
4.05222535e-01 4.90887374e-01 8.61044466e-01 3.55868876e-01
-8.44402313e-01 3.39233488e-01 -4.07767683e-01 -2.94036895e-01
-5.67801297e-01 -3.42467695e-01 -3.92009079e-01 3.33953381e-01
4.84007567e-01 -2.46723264e-01 -6.44808173e-01 7.97728226e-02
-1.32563090e+00 -4.77101833e-01 5.49272299e-01 6.63278878e-01
1.89625740e-01 3.90302271e-01 3.54292393e-01 -8.48466337e-01
7.40029752e-01 -5.28449833e-01 -3.71346235e-01 2.48760924e-01
-3.36884439e-01 -5.18184423e-01 4.60314870e-01 -4.22295809e-01
-6.43511295e-01 2.10839212e-02 8.84187818e-02 -7.78432071e-01
-3.11891884e-01 4.44730818e-02 1.39625549e-01 -2.58551866e-01
7.61315763e-01 4.51457918e-01 -3.08016986e-02 -4.12352264e-01
2.27615491e-01 9.10049438e-01 1.02575409e+00 -2.96830505e-01
8.40005279e-01 9.52136338e-01 3.01825926e-02 -1.00265467e+00
-2.31046736e-01 -5.88941276e-01 -5.71025014e-01 -3.02013278e-01
6.61849201e-01 -6.47441864e-01 -8.26937616e-01 9.94228721e-01
-1.17246079e+00 5.15522584e-02 2.01857522e-01 6.45062104e-02
-1.81872174e-01 1.27825773e+00 -7.20967472e-01 -1.00506306e+00
-4.85802889e-01 -1.06738913e+00 1.14257121e+00 -1.63665816e-01
3.74188134e-03 -7.75786638e-01 -3.80960912e-01 7.89257705e-01
3.78907949e-01 9.56604242e-01 6.23096287e-01 -6.64097548e-01
-5.00671387e-01 -8.44286680e-01 -4.70421463e-01 6.27440155e-01
-3.65788862e-02 -3.46220024e-02 -1.10965943e+00 -3.77325833e-01
1.44039989e-01 -2.81679541e-01 1.22019804e+00 -5.55517823e-02
1.55392134e+00 -1.75985396e-01 -1.49028972e-01 4.47050929e-01
1.36754692e+00 1.76094279e-01 1.04458463e+00 7.30059326e-01
5.74847698e-01 5.43914258e-01 3.09091777e-01 3.69751513e-01
-7.08946306e-03 5.50036490e-01 7.39744663e-01 1.96021497e-02
-1.79572403e-01 2.21008062e-03 4.13396120e-01 -4.96590547e-02
2.71282434e-01 -2.03611091e-01 -1.07044911e+00 7.74541020e-01
-1.31353092e+00 -1.28692758e+00 -2.90613353e-01 2.24715376e+00
5.20442903e-01 1.86580792e-01 3.59533548e-01 4.41068023e-01
1.28707194e+00 2.19111443e-01 -4.17093515e-01 -4.70318258e-01
-3.92403096e-01 5.34292161e-01 9.54058230e-01 2.08131177e-03
-1.59894252e+00 7.68800974e-01 6.50908899e+00 1.17700291e+00
-1.26637650e+00 2.33296245e-01 9.56206143e-01 -2.71210391e-02
4.11223084e-01 -3.26451242e-01 -3.80283773e-01 8.48979235e-01
9.62363839e-01 3.62716228e-01 2.28794813e-01 5.71035087e-01
1.35630250e-01 -1.40159011e-01 -5.21164417e-01 1.10099542e+00
5.72044909e-01 -1.63826263e+00 -2.10185528e-01 3.51219028e-01
5.23741126e-01 -1.74712375e-01 3.61457556e-01 4.40989342e-03
-3.21754217e-02 -1.25412273e+00 8.83435071e-01 1.82042539e-01
8.48878980e-01 -8.82941544e-01 7.29850113e-01 2.27606982e-01
-6.30834758e-01 -2.76782542e-01 -1.59560323e-01 2.15209931e-01
8.02455246e-02 6.41565800e-01 -8.37958038e-01 3.65183502e-01
7.64605582e-01 4.86896187e-01 -8.34078014e-01 1.05691242e+00
1.81463994e-02 8.26954603e-01 5.91449533e-03 5.31560540e-01
2.47441173e-01 2.12054700e-01 3.78270954e-01 1.66226864e+00
1.62377298e-01 -2.75447279e-01 -2.59649038e-01 7.17781723e-01
-3.76390398e-01 -3.21328938e-02 -4.47596997e-01 -1.81637138e-01
4.32359308e-01 1.22052360e+00 -1.05314505e+00 -4.87663507e-01
-2.59868115e-01 1.13333333e+00 -1.58105250e-02 3.57394367e-02
-9.21695471e-01 -5.54285407e-01 2.98748702e-01 2.13717356e-01
6.63789451e-01 -1.04034282e-02 -3.42274785e-01 -1.00018537e+00
2.65000671e-01 -1.21290720e+00 4.55439627e-01 -4.77871954e-01
-1.34599197e+00 3.41589153e-01 -3.00990939e-01 -1.16982424e+00
1.38491943e-01 -7.29950070e-01 -7.29049742e-01 4.77373004e-01
-1.49096024e+00 -1.10657918e+00 -2.67958015e-01 5.09513080e-01
3.64114285e-01 -2.82909006e-01 4.37718332e-01 6.06755435e-01
-5.61510324e-01 7.33590007e-01 7.59928897e-02 7.90800571e-01
7.87117183e-01 -1.04086936e+00 6.59668744e-01 1.24380648e+00
-1.71609037e-02 5.54233372e-01 5.07221460e-01 -7.38959908e-01
-1.39103687e+00 -1.14537752e+00 5.54202735e-01 -4.65958923e-01
6.32560968e-01 -9.55211520e-02 -8.70692432e-01 3.94787252e-01
2.66801775e-01 2.26719037e-01 6.14903271e-01 -5.52371979e-01
-8.20939541e-01 2.23889828e-01 -1.47967219e+00 -2.83310767e-02
4.98067677e-01 -7.00187683e-01 -3.84301454e-01 4.25412953e-01
-6.53620344e-03 -2.85934716e-01 -6.71950936e-01 2.39892066e-01
5.56817889e-01 -1.33166850e+00 1.20302081e+00 -5.36954403e-01
6.99505448e-01 -1.86490938e-01 -1.93800285e-01 -5.45187175e-01
4.15720940e-02 -6.26034141e-01 -3.07496428e-01 9.51319516e-01
-1.63227007e-01 -4.13512796e-01 7.79469192e-01 2.07819398e-02
6.42893910e-02 -6.32995784e-01 -9.93379354e-01 -8.90280068e-01
2.37227440e-01 -5.28274715e-01 4.72408742e-01 1.25627458e+00
-3.23842376e-01 -4.53375876e-01 -6.23867035e-01 2.33122006e-01
1.01068604e+00 -3.29910219e-02 8.46382618e-01 -6.54613853e-01
-3.44118536e-01 -5.60323358e-01 -7.31573403e-01 -3.99963677e-01
-1.59181416e-01 -5.86397648e-01 -3.24256778e-01 -1.03130126e+00
3.89587253e-01 -2.90287882e-02 -1.17813937e-01 3.89727086e-01
-2.52375543e-01 9.39115882e-01 2.56782383e-01 4.09183055e-01
-5.46805859e-01 -2.72906303e-01 8.51284146e-01 -3.39386642e-01
4.96468484e-01 -2.28142321e-01 -6.37831509e-01 5.16968250e-01
8.71254444e-01 -5.46264648e-01 1.89819202e-01 -2.64532119e-01
-1.79989219e-01 -1.70766950e-01 8.64429176e-01 -1.10554767e+00
1.65594816e-02 9.03206468e-02 7.55373359e-01 -7.54149735e-01
3.11118186e-01 -3.68730605e-01 -8.57273340e-02 7.34502614e-01
-1.59852371e-01 8.02961439e-02 2.70130843e-01 5.82174778e-01
-1.27927735e-01 -3.26243728e-01 1.00308073e+00 -2.35758901e-01
-6.93011105e-01 9.55824926e-03 -4.28172827e-01 -5.86819090e-02
1.11183500e+00 -6.55925095e-01 -6.08877361e-01 -2.30753064e-01
-3.79282236e-01 -3.87159199e-01 5.81072330e-01 3.70138288e-01
7.32037067e-01 -1.00292873e+00 -7.88583875e-01 9.38147455e-02
8.09596479e-02 -7.12070823e-01 3.44332933e-01 6.58801198e-01
-1.04511404e+00 8.46830085e-02 -3.24635744e-01 -4.39654440e-01
-1.59344327e+00 7.46480405e-01 2.65859991e-01 -2.52291739e-01
-8.33313525e-01 4.83579308e-01 -3.86429280e-01 2.19439119e-02
8.81994050e-03 2.50153303e-01 7.21743703e-02 3.06383651e-02
9.78800833e-01 7.54475951e-01 1.98511556e-01 -1.06226039e+00
-2.82830626e-01 2.18975574e-01 -2.24421605e-01 1.48906708e-01
1.27987301e+00 2.72819281e-01 -7.15305731e-02 -2.70293742e-01
1.61008859e+00 3.30396593e-02 -1.05634332e+00 -1.11660101e-01
4.32746746e-02 -9.29931939e-01 1.58836737e-01 -6.42172098e-01
-1.30701923e+00 9.96392667e-01 8.35557282e-01 3.79881829e-01
1.07590413e+00 -2.74480104e-01 1.18138456e+00 1.39075831e-01
3.11467089e-02 -1.19389904e+00 4.08051312e-01 4.45825532e-02
5.07404089e-01 -1.12639868e+00 2.82548219e-01 -2.31547430e-01
-4.31461006e-01 1.30165613e+00 9.37267765e-02 -3.80698264e-01
3.56092751e-01 2.65498221e-01 9.91936550e-02 -4.19668436e-01
-2.29726866e-01 3.32337469e-02 -1.82992399e-01 7.26006985e-01
1.49689108e-01 -4.43138182e-02 -2.55644232e-01 6.63937926e-02
-2.09020507e-02 -2.61560418e-02 7.07688272e-01 1.29088032e+00
-4.56795156e-01 -9.53708410e-01 -9.76466238e-01 4.73007768e-01
-1.08349192e+00 1.75227150e-01 -5.62858045e-01 7.63921738e-01
4.94701624e-01 1.15335727e+00 -4.48666401e-02 -2.76429087e-01
7.89165348e-02 -2.34888390e-01 4.20268476e-01 -1.54223800e-01
-7.98405588e-01 -2.66737878e-01 -8.47484618e-02 -6.42493606e-01
-1.04385547e-01 -9.26293373e-01 -7.94165373e-01 -6.25908852e-01
-2.02913448e-01 -2.59574026e-01 8.18961561e-01 7.86749363e-01
3.26736748e-01 2.01067984e-01 6.14449859e-01 -6.29115582e-01
-5.66709995e-01 -8.12076032e-01 -5.81958771e-01 9.78130102e-01
3.79509985e-01 -4.69565541e-01 -3.76086086e-01 3.61173630e-01] | [12.402562141418457, 1.023018479347229] |
ba07f3f2-15b6-452c-a75c-6a9ae519e2d2 | monitoring-term-drift-based-on-semantic | 1502.01753 | null | http://arxiv.org/abs/1502.01753v1 | http://arxiv.org/pdf/1502.01753v1.pdf | Monitoring Term Drift Based on Semantic Consistency in an Evolving Vector Field | Based on the Aristotelian concept of potentiality vs. actuality allowing for
the study of energy and dynamics in language, we propose a field approach to
lexical analysis. Falling back on the distributional hypothesis to
statistically model word meaning, we used evolving fields as a metaphor to
express time-dependent changes in a vector space model by a combination of
random indexing and evolving self-organizing maps (ESOM). To monitor semantic
drifts within the observation period, an experiment was carried out on the term
space of a collection of 12.8 million Amazon book reviews. For evaluation, the
semantic consistency of ESOM term clusters was compared with their respective
neighbourhoods in WordNet, and contrasted with distances among term vectors by
random indexing. We found that at 0.05 level of significance, the terms in the
clusters showed a high level of semantic consistency. Tracking the drift of
distributional patterns in the term space across time periods, we found that
consistency decreased, but not at a statistically significant level. Our method
is highly scalable, with interpretations in philosophy. | ['Ioannis Kompatsiaris', 'Sándor Darányi', 'Efstratios Kontopoulos', 'Theodoros Moysiadis', 'Peter Wittek'] | 2015-02-05 | null | null | null | null | ['lexical-analysis'] | ['natural-language-processing'] | [-2.51153708e-01 -2.81995654e-01 -1.62804633e-01 -2.86196202e-01
2.62652218e-01 -8.43809724e-01 9.70760822e-01 6.81850076e-01
-7.76953101e-01 1.94373325e-01 4.78039771e-01 -2.32221559e-01
-4.86367524e-01 -8.29479337e-01 -2.24363506e-01 -5.64258933e-01
-3.59525353e-01 2.49029696e-01 2.21025541e-01 -4.10631657e-01
7.10880458e-01 1.58308923e-01 -1.71200943e+00 2.02151969e-01
6.62383914e-01 6.66905701e-01 2.49034688e-01 3.69022846e-01
-7.36535966e-01 5.04262686e-01 -8.54564965e-01 -1.97175875e-01
-4.62337863e-03 -3.24746311e-01 -7.20655024e-01 -2.87851274e-01
1.23445429e-01 6.54194355e-01 -2.63142198e-01 1.40177822e+00
1.56996235e-01 3.79119366e-01 7.25981593e-01 -1.20845926e+00
-1.18256617e+00 8.43997419e-01 -2.10112110e-01 6.76172256e-01
4.25923496e-01 -3.92821640e-01 1.23376942e+00 -8.50366116e-01
1.14266384e+00 1.38022792e+00 7.17825234e-01 2.66044647e-01
-1.22109926e+00 -2.80699193e-01 2.12642223e-01 5.77370763e-01
-1.60809612e+00 -1.68121736e-02 8.98984671e-01 -6.94601655e-01
1.21673465e+00 3.12754512e-01 1.02274501e+00 8.67609382e-01
6.84752941e-01 8.08310211e-02 8.79213870e-01 -7.97568977e-01
7.52740264e-01 1.97929457e-01 4.06653821e-01 2.67368704e-01
-7.82857239e-02 -3.40801537e-01 -8.69364202e-01 -3.37086082e-01
1.22023225e-01 -4.33776937e-02 4.36693393e-02 -3.16167921e-01
-1.14775062e+00 1.04352820e+00 3.43913376e-01 1.01524115e+00
-4.80916023e-01 -7.04729557e-02 5.36677718e-01 3.40961486e-01
9.36807930e-01 4.25612420e-01 -4.76883084e-01 -3.54919434e-01
-7.75405526e-01 3.37891072e-01 5.07960021e-01 3.44557583e-01
3.28723311e-01 -3.97521019e-01 1.31605923e-01 8.52636814e-01
3.50539863e-01 2.69836813e-01 1.18133962e+00 -7.64329791e-01
-1.71231076e-01 7.67404974e-01 -1.70907438e-01 -1.44379330e+00
-4.50306714e-01 -8.86481702e-02 -2.08368853e-01 -1.64354444e-01
1.31796941e-01 2.99247444e-01 -7.21682608e-01 1.91297019e+00
3.12216550e-01 -6.30578473e-02 -2.21840829e-01 4.22300398e-01
1.47831291e-01 7.57412136e-01 2.17211857e-01 -5.65510690e-01
1.42714787e+00 -2.36993745e-01 -1.13634634e+00 1.61772519e-01
7.15508640e-01 -4.43361521e-01 1.39418387e+00 2.01697275e-01
-7.99198985e-01 -3.34442198e-01 -1.04017758e+00 1.55899227e-01
-9.18389320e-01 -1.21324849e+00 4.74811614e-01 5.17451227e-01
-1.33243585e+00 7.19748259e-01 -8.76703560e-01 -8.30547869e-01
6.45474419e-02 -3.19431394e-01 5.57376221e-02 4.45306569e-01
-1.47446954e+00 1.05809259e+00 5.27122736e-01 -3.45272958e-01
-8.75641778e-02 -6.77722216e-01 -6.47862077e-01 1.80012211e-02
-1.63110629e-01 -2.92963624e-01 7.38740921e-01 -7.30414450e-01
-9.97741342e-01 9.95854437e-01 -2.83881575e-01 -5.18516243e-01
-3.55537347e-02 1.83598563e-01 -9.76359010e-01 -9.98807773e-02
1.12328492e-01 3.05913240e-01 4.92633879e-01 -8.79848123e-01
-4.41673130e-01 -6.02101564e-01 -3.61298561e-01 2.23545015e-01
-1.08906817e+00 2.30210900e-01 7.87666291e-02 -7.25530565e-01
2.99188107e-01 -7.93470144e-01 7.09428564e-02 -4.03033853e-01
2.32350320e-01 -9.29498196e-01 5.59731901e-01 -4.57505733e-01
1.89911091e+00 -2.48193955e+00 1.57146484e-01 3.49622071e-01
3.33910793e-01 -4.66664672e-01 2.81019479e-01 5.27687788e-01
-1.19145304e-01 3.37473601e-01 -1.89559832e-01 4.84523103e-02
1.69683486e-01 1.71849370e-01 -2.80426800e-01 6.29913032e-01
-3.69528145e-01 8.22150767e-01 -1.21911800e+00 -3.98288608e-01
3.13396007e-02 4.07473445e-01 -4.16204959e-01 -5.79191923e-01
-2.58936226e-01 -3.44085842e-01 1.64919626e-02 4.37673122e-01
1.21618114e-01 -1.83748811e-01 2.06657887e-01 1.56309232e-01
-5.28112411e-01 3.40678185e-01 -8.93943310e-01 1.88735008e+00
-2.77868688e-01 1.15809774e+00 -5.06237090e-01 -9.01687682e-01
9.59742129e-01 -7.60220876e-03 6.79838657e-01 -1.03448582e+00
-1.28695846e-01 1.32735953e-01 1.80021092e-01 -1.49190679e-01
9.84959841e-01 -3.11234087e-01 -2.76088685e-01 4.21473205e-01
8.25732797e-02 6.91621453e-02 2.68566698e-01 2.70280808e-01
1.22209799e+00 -2.61044413e-01 -3.44855227e-02 -1.07292283e+00
1.40407607e-01 1.87568188e-01 2.39244685e-01 2.69393116e-01
-1.97600737e-01 2.37594381e-01 1.17877297e-01 -6.73005760e-01
-9.36597466e-01 -1.35611963e+00 -7.76144505e-01 1.35650861e+00
3.09206843e-01 -1.07596254e+00 -6.30486786e-01 -1.05229385e-01
1.15115069e-01 1.31622922e+00 -9.77496266e-01 -5.18328488e-01
-2.71872938e-01 -9.66944695e-01 1.78253546e-01 3.79992202e-02
-8.10934976e-02 -1.30305600e+00 -1.14102519e+00 2.18913183e-01
-1.78908333e-02 -4.20944363e-01 -2.89388269e-01 2.70454258e-01
-5.80358267e-01 -5.19392431e-01 -3.63760352e-01 -5.39994895e-01
3.17741275e-01 -1.90643985e-02 1.41271389e+00 -1.94909617e-01
-4.71539021e-01 5.43198228e-01 -4.05695975e-01 -6.56713247e-01
-2.28073642e-01 -1.10203840e-01 5.45252919e-01 -2.70884156e-01
9.68414903e-01 -5.42409003e-01 -6.27101719e-01 5.61988130e-02
-1.14364409e+00 -6.63155913e-01 -4.72372115e-01 5.42142332e-01
2.67351806e-01 5.08122146e-01 2.11736441e-01 -3.73437464e-01
1.20271492e+00 -8.26171458e-01 -3.73402894e-01 2.59778500e-01
-1.30360353e+00 2.10051477e-01 7.37127289e-02 -6.32797360e-01
-6.15685284e-01 -6.29207075e-01 6.73497438e-01 -2.41774887e-01
4.20586884e-01 6.61789000e-01 4.49722916e-01 5.76976657e-01
9.66050327e-01 6.53577521e-02 -2.61209995e-01 -2.66870052e-01
8.03105056e-01 8.35955918e-01 5.41941762e-01 -3.55082661e-01
4.03560400e-01 5.62736452e-01 -5.03052175e-01 -9.46727872e-01
-4.30680126e-01 -4.83002871e-01 -5.96670151e-01 -3.09923351e-01
8.84642422e-01 -4.42482859e-01 -3.69036674e-01 2.94283424e-02
-1.00724828e+00 -1.68637503e-02 -7.32087791e-01 4.35584277e-01
-2.06455037e-01 2.37465158e-01 -3.05909663e-01 -7.15967357e-01
-2.51171052e-01 -4.58433062e-01 6.10604823e-01 -1.76738612e-02
-1.05973029e+00 -1.50006890e+00 7.52250493e-01 -3.94004971e-01
5.91144681e-01 1.17630102e-01 1.20397747e+00 -7.69292593e-01
4.22294647e-01 -1.02710743e-02 5.26207864e-01 -1.07381046e-01
3.62357199e-01 2.42264330e-01 -7.46982574e-01 -1.74408436e-01
4.97102559e-01 3.94534588e-01 6.35916173e-01 4.46042180e-01
8.78601670e-01 -2.58946270e-01 -4.29087192e-01 2.23712578e-01
1.25547707e+00 6.89723015e-01 4.58852947e-01 8.63115668e-01
3.79285693e-01 7.69274652e-01 2.43729711e-01 6.01537466e-01
4.26209301e-01 6.37838066e-01 -7.39693940e-02 4.80511427e-01
3.20905566e-01 -2.08999753e-01 3.87757063e-01 1.25932074e+00
1.61031857e-01 -2.69514501e-01 -1.37567925e+00 9.75975811e-01
-1.68951845e+00 -1.13832188e+00 -1.50388870e-02 2.22676826e+00
8.53914678e-01 3.25643182e-01 1.63665459e-01 8.08077753e-02
8.47818613e-01 2.49281794e-01 -5.13804853e-01 -7.48311698e-01
-2.88937151e-01 3.35656963e-02 3.44247878e-01 5.71962118e-01
-5.76869845e-01 7.42708325e-01 7.02928877e+00 6.52062654e-01
-1.02837312e+00 9.13862735e-02 2.93555677e-01 -3.48864943e-01
-7.88806617e-01 -2.66061034e-02 -1.90408051e-01 9.37761724e-01
1.54664421e+00 -9.46175039e-01 4.56127375e-01 5.59135020e-01
-1.10739857e-01 3.40243056e-02 -8.55087519e-01 1.15260398e+00
2.32067928e-01 -1.13809025e+00 -1.23055570e-01 1.32773310e-01
5.71079612e-01 1.66811824e-01 2.79966332e-02 -4.29353025e-03
3.39263350e-01 -9.65572059e-01 9.76578653e-01 5.11704922e-01
4.16085839e-01 -6.94486380e-01 3.54134947e-01 2.47909963e-01
-1.10080147e+00 -6.84295222e-02 -4.66944009e-01 -3.26104939e-01
1.58202723e-01 7.31241524e-01 -4.49411213e-01 1.66848525e-01
1.12868476e+00 8.17526281e-01 -6.63818479e-01 3.76694888e-01
2.82410055e-01 4.39649403e-01 -3.78154814e-01 -2.38011241e-01
4.44670916e-02 -3.91222656e-01 4.67961907e-01 1.20897067e+00
3.37148935e-01 -3.45593125e-01 -4.04758900e-01 7.79316425e-01
1.32279158e-01 3.84113669e-01 -6.28249764e-01 -3.14036697e-01
9.12269831e-01 1.01849198e+00 -1.26729119e+00 -4.16684121e-01
-1.51510209e-01 9.82789814e-01 -1.95385423e-02 6.02824688e-02
-6.81182861e-01 -5.00034332e-01 6.41568601e-01 3.40280175e-01
5.26758246e-02 -2.96184301e-01 -4.65069234e-01 -1.10031545e+00
1.53111026e-01 -2.57378995e-01 4.18114603e-01 -5.94225228e-01
-1.51861751e+00 6.49594426e-01 3.25579494e-01 -7.74493337e-01
-2.80983150e-01 -4.04144078e-01 -4.39159960e-01 7.52725661e-01
-6.35070086e-01 -1.35448009e-01 -3.87453549e-02 5.20587087e-01
4.81245995e-01 -1.01208597e-01 9.30790484e-01 -6.11194335e-02
-1.06346413e-01 1.04806095e-01 3.75197679e-01 -2.97209144e-01
4.98775870e-01 -1.54376435e+00 8.14301789e-01 4.81946975e-01
7.29664028e-01 1.09777701e+00 9.99207973e-01 -7.40039766e-01
-1.01127446e+00 -3.74370486e-01 1.32305849e+00 -8.49624515e-01
1.33328199e+00 -5.72537124e-01 -1.10346794e+00 4.13717300e-01
4.82691139e-01 -2.20840812e-01 7.77970731e-01 2.26627871e-01
-3.74903977e-01 3.00933123e-01 -9.13515091e-01 6.89548016e-01
1.31281209e+00 -8.05230260e-01 -1.17244768e+00 4.86004144e-01
9.11177278e-01 2.55337566e-01 -1.11340916e+00 1.96858272e-01
6.68610752e-01 -6.05415225e-01 9.28578079e-01 -5.05789578e-01
9.72164944e-02 -2.28626192e-01 -4.10065591e-01 -1.53142440e+00
-6.48938775e-01 -4.37975198e-01 -1.16707936e-01 1.29304314e+00
1.49001375e-01 -6.04977190e-01 3.40297401e-01 5.63419342e-01
2.84865826e-01 -4.13718849e-01 -1.09815705e+00 -8.87356818e-01
4.86771882e-01 -4.78043854e-01 4.53284770e-01 1.35368228e+00
7.20709503e-01 4.67195094e-01 4.53589916e-01 -4.03417856e-01
4.76433307e-01 -1.99353591e-01 -2.53912777e-01 -1.60996544e+00
1.93647772e-01 -7.72217274e-01 -9.06561077e-01 -2.07541764e-01
2.71777391e-01 -1.27310503e+00 -5.88843375e-02 -1.36503065e+00
9.82048735e-02 -1.20794073e-01 -8.58607590e-01 4.06637341e-02
-4.46489491e-02 -1.57042623e-01 2.03092024e-01 5.58229148e-01
-6.42188430e-01 4.69010234e-01 2.24069059e-01 -1.54803887e-01
-3.95370185e-01 -8.75461459e-01 -5.13366163e-01 7.23981440e-01
7.99339831e-01 -5.86222231e-01 -6.72595680e-01 -2.90731519e-01
7.53480971e-01 -5.75234652e-01 -3.30598652e-03 -7.24175572e-01
2.58144855e-01 -7.12914020e-02 2.33835086e-01 -4.28803504e-01
-1.39774188e-01 -6.65575027e-01 4.15832072e-01 5.31518996e-01
-5.64086616e-01 8.81082535e-01 2.46878386e-01 4.55322295e-01
-1.94361135e-01 -2.05366127e-02 3.50731075e-01 -1.20877527e-01
-9.10694718e-01 -3.34338367e-01 -4.90509897e-01 2.37206623e-01
1.01882887e+00 -8.91016573e-02 -2.52159327e-01 -1.57629419e-02
-5.70771635e-01 3.08299810e-02 8.53560746e-01 9.49451387e-01
3.44772547e-01 -1.49498987e+00 -4.09252286e-01 -6.48109317e-02
2.74765104e-01 -5.51311970e-01 9.57932845e-02 3.29377383e-01
-2.24276245e-01 3.28838736e-01 -1.23811578e-02 -7.90792644e-01
-9.98031974e-01 9.12107646e-01 1.44359261e-01 1.51457503e-01
-7.23483980e-01 6.17430031e-01 -8.05982724e-02 -2.50795275e-01
1.02617085e-01 -3.62482369e-01 -3.40658337e-01 6.76612377e-01
6.23496771e-01 4.63455141e-01 1.37423247e-01 -6.79938138e-01
-8.01164985e-01 6.61173761e-01 1.56274155e-01 -5.25565267e-01
1.33099747e+00 -4.06258315e-01 -5.30512571e-01 1.59473407e+00
1.26327813e+00 -8.76872316e-02 -6.36015534e-01 -1.21979304e-01
9.11031008e-01 -2.73927331e-01 1.35276288e-01 -6.38537943e-01
-4.32321876e-01 3.48039806e-01 9.97541964e-01 9.17751610e-01
7.44588435e-01 2.23736346e-01 3.38347852e-01 1.79736972e-01
3.52977723e-01 -1.44160163e+00 -1.47984445e-01 5.51616669e-01
5.39042711e-01 -6.29504085e-01 -5.92010766e-02 2.98507988e-01
-3.62507910e-01 9.26368773e-01 2.00247876e-02 1.94467586e-02
9.75464165e-01 1.27142236e-01 1.57737628e-01 -7.40886807e-01
-8.87637019e-01 3.02138388e-01 2.36886308e-01 1.33212775e-01
6.17423117e-01 1.50285453e-01 -7.18268096e-01 9.34893563e-02
-7.57244885e-01 -2.78768986e-01 2.85313815e-01 1.11191475e+00
-4.51241046e-01 -8.46652329e-01 -3.43929082e-01 2.14516297e-01
-4.15987074e-01 -3.36226344e-01 -5.03906012e-01 5.21312892e-01
5.96585646e-02 8.80479336e-01 9.41407323e-01 -5.21304727e-01
2.79248536e-01 6.36251271e-01 2.36334339e-01 -3.80796492e-01
-3.84985685e-01 -2.27051437e-01 -4.11195934e-01 -4.93117988e-01
-5.31721950e-01 -9.39595580e-01 -1.36799252e+00 -2.15688586e-01
-7.49229342e-02 5.70809543e-01 1.06076932e+00 7.76686370e-01
3.39581072e-01 5.65408647e-01 5.40112197e-01 -3.02686274e-01
-3.43441069e-01 -1.15040541e+00 -7.09767997e-01 8.95459652e-01
-7.80945420e-02 -7.52278566e-01 -7.63501942e-01 1.31723464e-01] | [10.204349517822266, 8.95897388458252] |
33ec88aa-565f-4ffb-82ab-c05c4ba1035b | optimized-participation-of-multiple-fusion | 1805.12270 | null | http://arxiv.org/abs/1805.12270v1 | http://arxiv.org/pdf/1805.12270v1.pdf | Optimized Participation of Multiple Fusion Functions in Consensus Creation: An Evolutionary Approach | Recent studies show that ensemble methods enhance the stability and
robustness of unsupervised learning. These approaches are successfully utilized
to construct multiple clustering and combine them into a one representative
consensus clustering of an improved quality. The quality of the consensus
clustering is directly depended on fusion functions used in combination. In
this article, the hierarchical clustering ensemble techniques are extended by
introducing a new evolutionary fusion function. In the proposed method,
multiple hierarchical clustering methods are generated via bagging. Thereafter,
the consensus clustering is obtained using the search capability of genetic
algorithm among different aggregated clustering methods made by different
fusion functions. Putting some popular data sets to empirical study, the
quality of the proposed method is compared with regular clustering ensembles.
Experimental results demonstrate the accuracy improvement of the aggregated
clustering results. | ['Elaheh Rashedi', 'Abdolreza Mirzaei'] | 2018-05-31 | null | null | null | null | ['clustering-ensemble'] | ['graphs'] | [-2.82249957e-01 -5.95336258e-01 2.93803602e-01 -2.55665869e-01
-5.76695085e-01 -5.06681740e-01 3.71766716e-01 2.57187903e-01
-2.68738002e-01 8.48094225e-01 -9.50237215e-02 4.33552891e-01
-7.25502729e-01 -6.92010820e-01 -1.31681710e-01 -1.57799947e+00
2.03100473e-01 2.78324008e-01 -7.52047598e-02 5.09680957e-02
5.05212009e-01 1.96674958e-01 -2.15104413e+00 1.64795592e-01
1.67924547e+00 5.75827301e-01 1.61483169e-01 2.34076306e-01
-4.53714319e-02 2.05340758e-01 -8.77226055e-01 -1.59219261e-02
1.64666742e-01 -6.94497943e-01 -3.09838746e-02 2.84977347e-01
-3.13326925e-01 4.21718150e-01 4.71908271e-01 8.78431261e-01
6.54392242e-01 4.46069717e-01 8.81912529e-01 -1.40835941e+00
-3.24144661e-01 9.72753823e-01 -6.24930561e-01 -2.27380231e-01
5.86637445e-02 -2.14122534e-01 6.15804970e-01 -6.65713012e-01
7.34865069e-02 1.28152037e+00 6.09027863e-01 2.26219535e-01
-1.23027146e+00 -9.56632912e-01 -9.72353816e-02 3.08035433e-01
-1.76819098e+00 -2.96522260e-01 8.48229289e-01 -3.73540580e-01
3.11836153e-01 2.21047670e-01 5.33939421e-01 4.81313914e-01
2.77462453e-01 3.06877792e-01 1.46488190e+00 -6.72472954e-01
4.66088116e-01 1.97636366e-01 5.06390274e-01 5.65912545e-01
7.65988469e-01 -1.04443260e-01 -2.74279803e-01 -1.90602422e-01
-1.43330013e-02 5.92269003e-02 -2.97654688e-01 -2.17303157e-01
-8.85175884e-01 7.61201501e-01 3.22646111e-01 8.66614223e-01
-3.37313205e-01 -2.04753369e-01 1.91815853e-01 1.66701879e-02
4.16277647e-01 3.15616190e-01 -1.49237975e-01 2.22203255e-01
-1.08419824e+00 -3.86544503e-02 4.75580513e-01 4.86169994e-01
8.32445681e-01 1.77694947e-01 1.26505822e-01 8.70460272e-01
3.52795273e-01 5.75979471e-01 8.13900948e-01 -8.31485212e-01
-1.49305716e-01 1.12754714e+00 6.49419241e-03 -1.29052114e+00
-4.26508486e-01 -4.07424599e-01 -1.11798823e+00 2.39282519e-01
-3.22899558e-02 -4.19286191e-01 -5.85271776e-01 1.46251822e+00
4.28787977e-01 2.69144863e-01 3.75252813e-01 4.15606916e-01
7.51886845e-01 8.39450598e-01 -4.67706360e-02 -8.51427913e-01
8.12602997e-01 -7.07642913e-01 -1.08095050e+00 6.61801636e-01
3.64852339e-01 -5.69997072e-01 2.03022167e-01 8.63963068e-01
-8.56371880e-01 -1.09612751e+00 -1.30322802e+00 1.00898182e+00
-4.64677244e-01 4.11907911e-01 2.60679573e-01 1.02418578e+00
-1.16436875e+00 5.79352677e-01 -6.90436065e-01 -3.49822849e-01
-1.53349295e-01 6.40239656e-01 -2.16707632e-01 1.77572161e-01
-9.50161934e-01 7.24268258e-01 1.18390667e+00 3.65896299e-02
-1.30984515e-01 -1.68232426e-01 -2.90226191e-01 -7.11869225e-02
4.47488539e-02 -7.18752384e-01 4.13013011e-01 -1.11234295e+00
-1.30766904e+00 1.39045328e-01 -2.10654721e-01 5.94290625e-03
1.23830453e-01 1.15877673e-01 -6.96555734e-01 4.32907604e-02
-3.07920892e-02 2.66122609e-01 7.90287793e-01 -1.87380362e+00
-8.94189119e-01 -5.55876434e-01 -7.54960954e-01 3.60712051e-01
-4.98669773e-01 -1.97534040e-01 -7.48626292e-02 -4.34388220e-01
4.53291625e-01 -8.57657135e-01 -2.18691528e-01 -1.18933654e+00
-1.65238157e-02 -6.77125990e-01 9.75157797e-01 -5.37909985e-01
1.70724297e+00 -2.11518812e+00 7.06880569e-01 7.50004113e-01
3.70779610e-03 9.36164483e-02 3.38172823e-01 5.29218554e-01
-6.84710976e-04 4.00780171e-01 -4.39333051e-01 -1.39635265e-01
-9.15080234e-02 2.63940662e-01 2.72524863e-01 2.20509723e-01
-3.18734229e-01 2.43567601e-01 -6.64997160e-01 -9.26104248e-01
3.45012754e-01 3.72727662e-01 -3.54409605e-01 3.10472757e-01
3.64708960e-01 5.50166607e-01 -4.06729668e-01 4.90696281e-01
7.51656771e-01 1.09043747e-01 5.57385564e-01 -3.51734608e-01
-1.50355935e-01 -9.08008754e-01 -1.67484379e+00 1.38309109e+00
-4.56580035e-02 -1.69504285e-01 3.77782919e-02 -1.15966952e+00
1.48554182e+00 3.81659746e-01 8.49794507e-01 1.04560010e-01
3.76029074e-01 3.68529081e-01 2.87886560e-01 -3.18913758e-01
4.07329679e-01 -1.77425351e-02 -3.20120305e-02 5.20155430e-01
5.36857396e-02 8.46367050e-03 4.23477978e-01 7.75084272e-02
4.65206474e-01 9.31631327e-02 4.06163722e-01 -4.84670341e-01
8.52292180e-01 -8.26543793e-02 7.65817165e-01 5.97277761e-01
1.68505222e-01 2.90004194e-01 -2.15915456e-01 2.28839889e-02
-9.60388184e-01 -9.47893977e-01 -2.00425550e-01 8.03203404e-01
2.37277478e-01 -4.06708598e-01 -9.47393000e-01 -5.56257963e-01
-5.99374063e-02 6.36675656e-01 -4.02404130e-01 -3.20745826e-01
-4.08624828e-01 -1.38825548e+00 5.30771375e-01 2.97159374e-01
8.94724011e-01 -8.18738103e-01 -1.96504533e-01 3.51680934e-01
-2.99610823e-01 -3.71439844e-01 1.14822142e-01 7.47163370e-02
-1.11502612e+00 -9.91033137e-01 -2.34362945e-01 -6.30962551e-01
6.42524898e-01 3.41992974e-01 7.39142418e-01 6.49192274e-01
1.84451997e-01 3.74844283e-01 -8.95633161e-01 -2.43018106e-01
-7.27192819e-01 4.68368642e-02 5.33244729e-01 3.19409847e-01
4.00213659e-01 -8.28479350e-01 -2.18352139e-01 4.12609488e-01
-9.12155867e-01 -3.57595682e-01 4.58647609e-01 8.95353377e-01
2.51452208e-01 1.09936488e+00 1.03337514e+00 -2.26500288e-01
9.29457068e-01 -4.41448748e-01 -3.91785800e-01 4.79203343e-01
-1.05681264e+00 1.83106050e-01 7.74111748e-01 -1.86828658e-01
-1.52593601e+00 9.88897532e-02 3.27335447e-01 -4.39209402e-01
-3.20604682e-01 4.74944502e-01 -4.04222161e-01 1.49839036e-02
4.02283162e-01 4.25475299e-01 2.86155194e-01 -3.73755038e-01
2.98404425e-01 9.13106441e-01 4.26036358e-01 -8.75544488e-01
6.50797963e-01 2.12451667e-01 -1.85241804e-01 -4.89721209e-01
-1.44903511e-01 -4.09063578e-01 -1.11900997e+00 -6.79253519e-01
9.06765521e-01 -6.84316695e-01 -7.03584075e-01 6.48476183e-01
-8.55380297e-01 5.21331728e-01 2.51660109e-01 5.42123079e-01
-2.45377690e-01 5.40087223e-01 -9.69684571e-02 -1.12358510e+00
-3.64374965e-01 -1.21492028e+00 6.21923864e-01 4.86904830e-01
5.08003607e-02 -9.25668180e-01 7.50601441e-02 3.06940436e-01
-1.27690015e-02 4.49673533e-01 7.70729542e-01 -7.00060427e-01
-1.24256708e-01 -5.69401421e-02 4.53788400e-01 5.55198193e-01
4.54625070e-01 6.79169238e-01 -7.27967203e-01 -4.55200702e-01
9.27040204e-02 1.79996118e-01 9.13987935e-01 5.43782234e-01
1.01764190e+00 -5.47817908e-02 -7.06204176e-01 3.53894264e-01
1.61882317e+00 9.76336360e-01 4.41625506e-01 2.50137269e-01
5.70830405e-01 6.47571623e-01 6.42864585e-01 4.57501262e-01
2.95534544e-02 3.21099877e-01 -5.69993667e-02 3.76298368e-01
4.26123649e-01 2.22346574e-01 3.47493708e-01 1.55697978e+00
-6.51000261e-01 -3.84438932e-01 -7.54353344e-01 1.86808869e-01
-2.18582654e+00 -1.41610551e+00 -3.01419675e-01 2.09486008e+00
6.35451555e-01 -1.51478633e-01 7.17871115e-02 6.51669025e-01
1.25454164e+00 -3.07460070e-01 -2.22034529e-01 -3.59907061e-01
-3.63833338e-01 3.72753851e-02 2.85154060e-02 3.37855339e-01
-1.16401637e+00 4.69034702e-01 6.18835926e+00 1.17525625e+00
-7.56489754e-01 -4.61273305e-02 4.40989047e-01 1.79899946e-01
-1.15510195e-01 3.43154594e-02 -6.67896986e-01 9.06775534e-01
7.96561003e-01 -3.32752854e-01 2.86996573e-01 3.74913663e-01
2.56614894e-01 -3.74177545e-01 -5.31319201e-01 1.04246080e+00
1.76253408e-01 -8.22423458e-01 8.11418518e-02 -2.37372261e-03
1.21286702e+00 -7.19802856e-01 -1.54810369e-01 2.11515073e-02
7.41948962e-01 -9.05489445e-01 9.94083136e-02 9.85842943e-01
8.13701972e-02 -1.34985423e+00 1.03524268e+00 5.41766047e-01
-1.30816805e+00 -4.22551215e-01 -1.32280156e-01 3.86679210e-02
8.26567262e-02 8.10850978e-01 -5.11076748e-01 1.45017993e+00
9.10446286e-01 6.87878728e-01 -9.26667690e-01 9.32014406e-01
5.88925257e-02 6.97157443e-01 -4.11078066e-01 -1.31722569e-01
-8.92977864e-02 -8.78095567e-01 5.17232776e-01 8.83476496e-01
7.17461169e-01 4.23300505e-01 2.35971957e-01 6.12129569e-01
4.16419774e-01 4.36762571e-01 -5.22626638e-01 2.47369960e-01
1.10376310e+00 1.35752511e+00 -9.84377444e-01 -5.51866531e-01
1.53934479e-01 5.32935381e-01 -3.24283205e-02 1.45273551e-01
-8.46152306e-01 -2.36094460e-01 5.95995560e-02 -5.43823421e-01
2.34203488e-01 -1.74150810e-01 -3.66829187e-01 -8.73326421e-01
-3.61197352e-01 -1.07816100e+00 6.79299057e-01 -7.51795948e-01
-1.49852896e+00 5.72148025e-01 2.68615156e-01 -1.33435249e+00
-1.05123967e-01 -2.07650661e-01 -7.03378379e-01 5.51347256e-01
-2.72925973e-01 -8.11266541e-01 -5.02886772e-01 6.42902970e-01
2.42251560e-01 -6.84338033e-01 6.93778992e-01 3.86882387e-02
-8.66258323e-01 3.76097411e-01 8.23717177e-01 -3.01142186e-01
8.62114847e-01 -1.20616794e+00 -8.37951481e-01 9.32149053e-01
-1.88036650e-01 7.11265504e-01 5.84165096e-01 -8.87086987e-01
-7.27477133e-01 -7.92804182e-01 4.07607824e-01 -2.84261703e-01
9.20423269e-02 2.13101849e-01 -9.12710309e-01 2.21287027e-01
7.19386339e-01 -8.90852749e-01 9.33872998e-01 2.53139194e-02
-5.94653562e-03 -4.95157748e-01 -1.44983399e+00 3.70549351e-01
4.88354415e-01 2.23581433e-01 -8.24200392e-01 -2.77704567e-01
5.18209577e-01 3.16249579e-01 -1.32289803e+00 6.95772111e-01
4.51947212e-01 -1.29953480e+00 7.09284723e-01 -1.91859566e-02
1.28149629e-01 -8.60953033e-01 -1.69215620e-01 -1.67353499e+00
-6.40416622e-01 -2.83901393e-02 2.86459446e-01 1.65728211e+00
2.91285608e-02 -7.93749690e-01 1.97259739e-01 1.62044302e-01
4.11903001e-02 -3.66032183e-01 -7.05807507e-01 -6.51441455e-01
-1.43229337e-02 2.85602808e-01 7.81632304e-01 1.12117445e+00
1.99980006e-01 3.47624362e-01 -1.16792589e-01 1.02127753e-01
9.09269273e-01 1.60715327e-01 6.31932735e-01 -1.67581284e+00
5.32608852e-02 -6.03764057e-01 -3.48611981e-01 5.49282916e-02
1.54717460e-01 -7.87864149e-01 -5.26515581e-03 -1.48366904e+00
2.97396153e-01 -4.76461112e-01 -5.85316598e-01 2.16781259e-01
-6.59448922e-01 4.74066362e-02 2.42281556e-01 5.29196501e-01
-5.14075696e-01 5.39684594e-01 9.14570212e-01 -9.30035952e-03
-5.04182935e-01 -7.91813508e-02 -7.16459990e-01 6.84442997e-01
1.18188167e+00 -3.86232108e-01 -5.14545202e-01 2.34156102e-01
-3.01095068e-01 -1.00988261e-01 -1.75262049e-01 -1.51984465e+00
4.16780710e-01 8.91148206e-03 5.20387292e-01 -9.57251012e-01
-2.24059541e-03 -1.00600111e+00 8.00636888e-01 5.60502410e-01
3.74829993e-02 3.02340090e-01 1.01570897e-01 6.31861746e-01
-3.13135505e-01 -3.02916199e-01 8.20044398e-01 -4.06432487e-02
-5.69127202e-01 -2.21754029e-01 -4.56980914e-01 -6.51188493e-01
1.36294878e+00 -5.96112251e-01 8.98922309e-02 -1.58151254e-01
-9.15214181e-01 2.46156067e-01 4.90837008e-01 2.46735513e-01
3.91399384e-01 -1.41941905e+00 -9.44375455e-01 -2.71955281e-02
-2.68838197e-01 -1.30594477e-01 2.99506366e-01 9.27947700e-01
-2.18148455e-01 1.29341915e-01 -3.98835301e-01 -7.79190779e-01
-1.57713211e+00 6.54822052e-01 2.78262705e-01 -1.29087836e-01
2.60036755e-02 3.92275959e-01 -3.07138622e-01 -5.52307844e-01
3.73105868e-04 2.16644689e-01 -7.42309213e-01 3.81503791e-01
8.59862044e-02 8.91007185e-01 2.05150731e-02 -6.89451694e-01
-3.48948151e-01 8.08039010e-01 4.16763723e-01 9.70194209e-03
1.18897569e+00 -2.62813538e-01 -7.87942946e-01 4.99163955e-01
7.48524129e-01 9.51410457e-02 -5.98091006e-01 1.47917569e-01
1.23570263e-02 -6.92007840e-02 -2.42818817e-01 -8.00886452e-01
-1.11059391e+00 3.68192822e-01 8.40772510e-01 3.83784980e-01
1.56566632e+00 -4.35602367e-01 1.28613859e-01 2.04506427e-01
4.90356445e-01 -1.28075206e+00 -8.39482024e-02 -4.86835949e-02
5.37400186e-01 -1.07154977e+00 2.45624721e-01 -1.67916760e-01
-4.91204292e-01 1.21731675e+00 8.72798145e-01 -1.19720817e-01
6.46092057e-01 3.50442141e-01 -8.00821111e-02 9.32081342e-02
-5.99147797e-01 -1.65360961e-02 2.47170776e-01 5.44620991e-01
3.54475737e-01 1.73733681e-01 -1.22056770e+00 8.64483416e-01
-1.66339085e-01 -2.07316995e-01 3.68695825e-01 6.56921983e-01
-8.97581756e-01 -1.28893042e+00 -1.02634108e+00 3.31885278e-01
-3.50012444e-02 2.27729484e-01 -5.09870946e-01 6.78549647e-01
7.12564170e-01 1.61247408e+00 4.81392667e-02 -9.79954839e-01
-7.55936280e-02 3.88608187e-01 3.33392471e-01 -9.02789831e-02
-8.54758680e-01 2.75961250e-01 -2.30363116e-01 8.64865910e-03
-1.18653619e+00 -6.58192694e-01 -1.31135881e+00 -4.34353471e-01
-7.48722255e-01 1.04040778e+00 4.38223243e-01 8.76341522e-01
2.20302194e-01 5.87176204e-01 1.03787172e+00 -7.97962904e-01
-2.11423159e-01 -9.94081616e-01 -7.54243791e-01 4.53493625e-01
-3.35605025e-01 -8.87908220e-01 -7.94356883e-01 2.52040803e-01] | [7.638338088989258, 4.5386810302734375] |
3af1d008-218a-4f3c-bcbe-4dbdde266341 | multi-view-multi-person-3d-pose-estimation | 2104.02273 | null | https://arxiv.org/abs/2104.02273v1 | https://arxiv.org/pdf/2104.02273v1.pdf | Multi-View Multi-Person 3D Pose Estimation with Plane Sweep Stereo | Existing approaches for multi-view multi-person 3D pose estimation explicitly establish cross-view correspondences to group 2D pose detections from multiple camera views and solve for the 3D pose estimation for each person. Establishing cross-view correspondences is challenging in multi-person scenes, and incorrect correspondences will lead to sub-optimal performance for the multi-stage pipeline. In this work, we present our multi-view 3D pose estimation approach based on plane sweep stereo to jointly address the cross-view fusion and 3D pose reconstruction in a single shot. Specifically, we propose to perform depth regression for each joint of each 2D pose in a target camera view. Cross-view consistency constraints are implicitly enforced by multiple reference camera views via the plane sweep algorithm to facilitate accurate depth regression. We adopt a coarse-to-fine scheme to first regress the person-level depth followed by a per-person joint-level relative depth estimation. 3D poses are obtained from a simple back-projection given the estimated depths. We evaluate our approach on benchmark datasets where it outperforms previous state-of-the-arts while being remarkably efficient. Our code is available at https://github.com/jiahaoLjh/PlaneSweepPose. | ['Gim Hee Lee', 'Jiahao Lin'] | 2021-04-06 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Lin_Multi-View_Multi-Person_3D_Pose_Estimation_With_Plane_Sweep_Stereo_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Lin_Multi-View_Multi-Person_3D_Pose_Estimation_With_Plane_Sweep_Stereo_CVPR_2021_paper.pdf | cvpr-2021-1 | ['3d-pose-estimation', '3d-multi-person-pose-estimation'] | ['computer-vision', 'computer-vision'] | [-1.53833956e-01 -2.18846172e-01 1.48707807e-01 -4.36101168e-01
-1.24834716e+00 -6.37354016e-01 5.80217183e-01 -7.57159069e-02
-4.18646783e-01 3.37811649e-01 4.73055661e-01 5.10970354e-01
2.04806149e-01 -5.57166338e-01 -6.72429919e-01 -2.86236584e-01
4.29211020e-01 8.43950033e-01 1.64162040e-01 1.05772547e-01
1.57753944e-01 4.14381117e-01 -1.36366284e+00 1.60550922e-01
3.34911317e-01 5.10232389e-01 -1.21367633e-01 1.01179385e+00
5.81644416e-01 1.86908707e-01 -2.35803455e-01 -6.20986044e-01
5.42510748e-01 -2.07981944e-01 -4.70908999e-01 7.43610620e-01
1.18401170e+00 -7.33783722e-01 -2.76365072e-01 7.29872823e-01
7.62692273e-01 1.05630942e-01 3.76540482e-01 -1.11159909e+00
2.29085609e-01 -4.46542829e-01 -1.13225126e+00 7.22016841e-02
1.25616622e+00 1.01134352e-01 8.52496743e-01 -1.24611223e+00
7.07922101e-01 1.49156570e+00 7.51698792e-01 5.52098572e-01
-1.18872893e+00 -4.63458061e-01 3.61391336e-01 -1.05344251e-01
-1.55888796e+00 -5.35532176e-01 7.34827876e-01 -6.35373175e-01
9.20614839e-01 1.20037794e-01 1.09424531e+00 9.88463640e-01
1.42771125e-01 6.02940440e-01 9.87262487e-01 -1.94433898e-01
-2.44050518e-01 -2.03327432e-01 -8.95926356e-02 9.26378012e-01
3.28674763e-01 1.14798620e-01 -8.47013950e-01 -1.32164583e-01
1.16415906e+00 3.45689029e-01 -5.59152216e-02 -7.07760036e-01
-1.35216296e+00 5.95266283e-01 2.96363860e-01 -4.00283694e-01
-3.11540335e-01 2.57441133e-01 1.73401862e-01 -1.52333423e-01
6.30810142e-01 -3.15057449e-02 -4.08695132e-01 -1.54725313e-01
-8.28567982e-01 8.44853818e-01 4.10278916e-01 1.05911195e+00
7.34533548e-01 -6.30331278e-01 6.21362366e-02 6.30465448e-01
4.33661669e-01 5.61713338e-01 -2.43642166e-01 -1.33705556e+00
7.71608114e-01 5.99375248e-01 2.32941940e-01 -8.94116402e-01
-5.11126339e-01 -3.07675421e-01 -4.82448637e-01 -4.65287641e-02
7.49786973e-01 -1.47121921e-01 -6.38087571e-01 1.36420143e+00
9.35640574e-01 8.12697113e-02 -4.35235411e-01 1.17665136e+00
5.86971819e-01 3.05238098e-01 -3.78866613e-01 9.17981192e-02
1.60652232e+00 -9.96692300e-01 -2.64580488e-01 -7.13526726e-01
2.61598855e-01 -7.04541326e-01 6.74214065e-01 5.88206530e-01
-1.41619134e+00 -4.94550586e-01 -7.44799972e-01 -3.77504438e-01
1.45181984e-01 2.71512836e-01 2.28165731e-01 4.86132890e-01
-7.30579972e-01 2.29140565e-01 -9.37841058e-01 -5.88765085e-01
1.12575285e-01 3.59157771e-01 -6.91873968e-01 -3.34571719e-01
-5.39007068e-01 9.23043311e-01 -3.99888400e-03 1.35796443e-01
-7.53720641e-01 -6.94829226e-01 -9.85904992e-01 -4.64741081e-01
7.22917438e-01 -1.39428294e+00 1.24757564e+00 -2.73251712e-01
-1.12416947e+00 1.29376900e+00 -5.12089252e-01 1.86599061e-01
8.59711528e-01 -7.95795321e-01 1.76177457e-01 3.17612290e-01
4.42002147e-01 6.66094184e-01 5.44079661e-01 -1.42939460e+00
-6.58141613e-01 -9.90583599e-01 1.90767482e-01 8.30566883e-01
1.74846455e-01 4.66900766e-02 -1.24969268e+00 -2.92708129e-01
6.46854520e-01 -1.14372218e+00 -3.47807765e-01 3.16875577e-01
-6.10009551e-01 1.25312775e-01 2.67218322e-01 -8.93948793e-01
7.19646215e-01 -1.78978968e+00 5.66668391e-01 7.03400671e-02
4.30610150e-01 -5.14966905e-01 3.64166379e-01 2.24954277e-01
1.79261595e-01 -4.21953052e-01 1.78216353e-01 -1.06918693e+00
-2.53362536e-01 -1.76903129e-01 4.65965718e-01 1.05959451e+00
-1.32245049e-01 6.22442305e-01 -7.59125710e-01 -6.16827011e-01
5.61053157e-01 6.67286575e-01 -8.53863657e-01 4.30967152e-01
2.71655202e-01 7.72618055e-01 -3.55561227e-01 7.17940629e-01
6.74214423e-01 -3.28960776e-01 8.90235454e-02 -4.18070555e-01
-1.36063591e-01 1.01635307e-01 -1.43625307e+00 2.24453902e+00
-2.72665948e-01 8.50357115e-02 1.48352876e-01 -3.45794946e-01
4.91956472e-01 2.20928833e-01 8.13948989e-01 -1.76333204e-01
1.15566418e-01 2.23656539e-02 -5.55629849e-01 -1.20604791e-01
3.81111234e-01 -1.85512051e-01 -3.03544790e-01 1.65519521e-01
-2.52152253e-02 -2.37442881e-01 1.14132548e-02 1.73517406e-01
7.23324120e-01 4.96709764e-01 4.86107200e-01 1.54426515e-01
5.63524365e-01 -2.38720953e-01 5.29537797e-01 3.14143777e-01
-1.02319077e-01 1.15899146e+00 1.36227891e-01 -5.20951092e-01
-1.20850635e+00 -1.38459754e+00 2.08290339e-01 6.29837930e-01
2.67071426e-01 -7.35938072e-01 -6.93956971e-01 -5.92478216e-01
3.10900770e-02 9.28109884e-02 -4.03065354e-01 3.96112680e-01
-7.17062593e-01 -4.00234938e-01 1.42544821e-01 5.20767331e-01
4.76102501e-01 -2.09102795e-01 -7.22051382e-01 -1.16537765e-01
-5.92044175e-01 -1.51014817e+00 -9.88513708e-01 -2.44363666e-01
-8.45519185e-01 -1.25506735e+00 -1.01963103e+00 -3.20847392e-01
9.09607172e-01 5.90787470e-01 1.09520268e+00 -1.05330311e-01
-2.20104113e-01 8.16248417e-01 -3.25191617e-02 3.05255074e-02
2.11953789e-01 -1.33615747e-01 4.39765215e-01 -1.25422344e-01
3.35851222e-01 -6.02422118e-01 -9.31580365e-01 4.59024429e-01
-1.68463271e-02 3.80463570e-01 2.32068464e-01 5.03164709e-01
8.69134724e-01 -3.64282399e-01 -2.83853263e-01 -6.73878610e-01
-7.04985410e-02 -2.97504850e-02 -5.93240738e-01 3.39933634e-02
-7.18197227e-02 -4.55025673e-01 3.78259527e-03 -3.75597551e-02
-9.45093930e-01 5.51966667e-01 -7.56102353e-02 -6.24208510e-01
-2.74435043e-01 -8.66102800e-02 -4.07297194e-01 1.01125345e-01
4.24547404e-01 8.01780149e-02 -1.26474947e-01 -4.82211381e-01
2.67699391e-01 3.26331019e-01 6.99882090e-01 -6.34503901e-01
9.58617687e-01 8.03283632e-01 -1.21643879e-02 -6.47397995e-01
-1.27911603e+00 -9.30390596e-01 -1.31789279e+00 -6.59471273e-01
1.18472052e+00 -1.60707879e+00 -8.82160246e-01 4.31073278e-01
-1.30288446e+00 4.82240282e-02 1.84187159e-01 5.95594525e-01
-7.00080454e-01 5.36838830e-01 -5.06060123e-01 -7.89718211e-01
-2.29228422e-01 -1.14704049e+00 1.83488309e+00 1.72156151e-02
-4.23465878e-01 -8.06926131e-01 2.01688826e-01 9.79724705e-01
-4.65995163e-01 4.55869377e-01 -2.00123508e-02 4.17252295e-02
-7.17001617e-01 -3.62475395e-01 3.56525555e-02 -1.76061884e-01
-6.96624666e-02 -3.74966770e-01 -9.54228997e-01 -5.47108591e-01
-1.12573482e-01 -1.51455760e-01 4.96954501e-01 6.56905115e-01
7.05074131e-01 3.57860923e-02 -3.10007125e-01 9.19325292e-01
1.31568003e+00 -5.24060071e-01 3.27935100e-01 5.28290868e-01
1.11825240e+00 6.69277728e-01 7.44118750e-01 7.93593705e-01
9.83912885e-01 1.19761384e+00 2.27359787e-01 1.44108208e-02
-1.66266888e-01 -5.50117135e-01 4.18634951e-01 6.12871826e-01
-3.98642987e-01 6.33983165e-02 -8.21160197e-01 3.01021755e-01
-1.68616474e+00 -8.40707719e-01 -3.88186246e-01 2.45777249e+00
3.40767473e-01 1.31216764e-01 6.73664093e-01 -1.53240278e-01
6.60386384e-01 5.03027886e-02 -4.84693378e-01 4.19438809e-01
2.15852708e-01 -3.29900235e-01 4.94545132e-01 7.33363688e-01
-1.10775781e+00 6.57986581e-01 5.70456219e+00 1.19246751e-01
-4.22146946e-01 2.10132092e-01 3.41819972e-01 -6.98781908e-01
-6.00945279e-02 1.10671613e-02 -1.25718737e+00 1.49598569e-01
4.19692546e-01 3.41383785e-01 1.92144930e-01 5.54916263e-01
2.94595182e-01 -5.26677370e-01 -1.33051944e+00 1.63448405e+00
3.93378615e-01 -1.01046813e+00 -2.32592434e-01 4.54912394e-01
7.10655153e-01 -1.09907247e-01 -2.78480232e-01 -1.52658209e-01
1.68595407e-02 -4.63892967e-01 8.47742915e-01 5.68186164e-01
7.41824210e-01 -8.83637011e-01 3.61248851e-01 5.94784498e-01
-1.51284897e+00 1.31241307e-01 -1.03606932e-01 -1.91299766e-01
6.09028995e-01 5.59969664e-01 -4.02203918e-01 7.43944108e-01
8.49219203e-01 9.72835481e-01 -5.76837957e-01 7.63905287e-01
-1.76200211e-01 -3.73384982e-01 -4.30934936e-01 5.13412654e-01
-2.92803109e-01 -2.67612159e-01 6.35380089e-01 9.47648883e-01
4.51589763e-01 1.38988525e-01 5.17658710e-01 5.84535003e-01
2.75133342e-01 -2.88238019e-01 -4.27722126e-01 7.17815757e-01
3.46949726e-01 1.17723131e+00 -6.58684611e-01 -2.93589801e-01
-6.00616455e-01 1.42640984e+00 3.90765786e-01 8.32226351e-02
-9.12917852e-01 4.63906288e-01 8.41597378e-01 5.40377140e-01
1.52517796e-01 -4.96144801e-01 -2.79787034e-01 -1.70864046e+00
3.90379518e-01 -7.44357109e-01 5.62256813e-01 -8.65415215e-01
-1.21874034e+00 1.89361513e-01 3.32167447e-01 -1.50492167e+00
-4.44111824e-01 -3.85606021e-01 -2.31114358e-01 8.93686116e-01
-1.12677073e+00 -1.40740991e+00 -6.22543514e-01 7.53026485e-01
8.29662085e-01 1.22684531e-01 5.87126255e-01 3.43879968e-01
-4.72443938e-01 5.24807453e-01 -6.12545550e-01 1.09222926e-01
9.10126507e-01 -1.15287280e+00 5.72012365e-01 9.86151338e-01
3.38311382e-02 6.10618651e-01 6.63017213e-01 -8.59685302e-01
-1.73609805e+00 -7.92977273e-01 8.04458082e-01 -1.20836377e+00
5.28988801e-02 -5.59363961e-01 -2.42730603e-01 9.73452747e-01
-3.41189981e-01 2.01534167e-01 5.04064798e-01 2.69083828e-01
-2.96219915e-01 8.86114594e-03 -1.07091117e+00 4.56544369e-01
1.28650582e+00 -4.72434610e-01 -3.95395517e-01 2.21346945e-01
2.56513596e-01 -9.83539224e-01 -8.83298159e-01 -2.28729658e-02
8.54951560e-01 -1.23481905e+00 1.51757872e+00 -6.61343858e-02
4.11917597e-01 -5.48304439e-01 -3.04011375e-01 -9.61747348e-01
-2.42411435e-01 -3.20925951e-01 -2.59120733e-01 9.18692112e-01
-5.02443388e-02 -3.10860366e-01 1.11319470e+00 7.64706612e-01
1.12091705e-01 -6.91221356e-01 -9.25094903e-01 -4.88456458e-01
-3.58744204e-01 -3.91747087e-01 8.22050348e-02 4.58584160e-01
-9.63554829e-02 5.19371510e-01 -6.40614152e-01 5.96861064e-01
1.21767163e+00 1.57516804e-02 1.41120100e+00 -1.15789449e+00
-5.82044125e-01 9.00139138e-02 -4.04432386e-01 -1.53034353e+00
-1.85895488e-01 -4.59205598e-01 -6.32463992e-02 -1.57695293e+00
7.37794578e-01 1.68101847e-01 5.25158763e-01 1.37768574e-02
-3.06059450e-01 4.12998527e-01 3.58142406e-01 2.69913256e-01
-6.65243924e-01 3.08722526e-01 1.29779947e+00 2.38181457e-01
-4.83034216e-02 1.31177649e-01 -6.29884422e-01 1.05161941e+00
3.50382417e-01 -4.11015064e-01 -2.53525972e-01 -6.27317667e-01
1.74670711e-01 4.31489140e-01 8.02109301e-01 -1.12633038e+00
2.99649894e-01 -8.41277093e-02 9.60288107e-01 -1.12743139e+00
1.17002201e+00 -6.50856674e-01 3.57977450e-01 4.03810143e-01
1.67909879e-02 3.52352798e-01 -1.89854726e-01 6.52359664e-01
2.07774147e-01 2.30340913e-01 7.79870272e-01 -7.02332318e-01
-4.23912168e-01 6.29431307e-01 1.99350089e-01 7.67463744e-02
1.00726366e+00 -6.13623619e-01 1.56915233e-01 -5.50141096e-01
-8.19483578e-01 3.09601188e-01 8.92611682e-01 3.52418393e-01
8.91549587e-01 -1.35866761e+00 -8.22492123e-01 9.32869315e-02
1.27495557e-01 3.53666067e-01 4.72819030e-01 9.55290139e-01
-5.07404327e-01 3.03954363e-01 1.58303585e-02 -1.04431522e+00
-1.69062006e+00 1.63255170e-01 6.06195569e-01 -3.02928537e-01
-8.02303195e-01 9.11478400e-01 4.23053086e-01 -6.74249411e-01
1.10618565e-02 1.07662082e-01 1.30087510e-01 -9.53553244e-02
6.45414412e-01 6.54733896e-01 -7.04323426e-02 -1.04644394e+00
-6.64550781e-01 1.28988683e+00 -5.79952747e-02 -5.14860034e-01
1.35112476e+00 -6.21046126e-01 1.64635137e-01 4.34886128e-01
1.28001463e+00 1.96970508e-01 -1.65161741e+00 -1.37461707e-01
-6.40534163e-01 -9.61595118e-01 -1.25614092e-01 -4.22400236e-01
-9.24314141e-01 7.16141880e-01 4.86614048e-01 -6.34154379e-01
8.42785478e-01 2.62486845e-01 6.83559060e-01 7.50933066e-02
6.05369687e-01 -1.01985538e+00 3.09038967e-01 3.48065585e-01
8.25133443e-01 -1.31666088e+00 6.93590045e-01 -8.89048100e-01
-4.88959521e-01 1.00415111e+00 9.94021893e-01 -2.78864652e-01
4.02106971e-01 1.78193495e-01 -1.71095476e-01 -4.33230877e-01
-4.54038441e-01 -3.50308679e-02 3.97821873e-01 5.03170133e-01
4.92107272e-01 -1.21839307e-02 2.51761079e-01 1.85924530e-01
-2.78421879e-01 -2.12992862e-01 4.43557054e-01 8.02578032e-01
-1.70240641e-01 -1.05014575e+00 -9.25756097e-01 1.67161986e-01
-2.84220248e-01 2.45580509e-01 -3.61832649e-01 6.73459530e-01
1.22774713e-01 8.30776513e-01 4.83020060e-02 -4.53153402e-01
7.22880483e-01 -1.59455866e-01 1.08469486e+00 -8.20175767e-01
-4.36904907e-01 6.35347068e-01 2.34666675e-01 -9.17642176e-01
-5.55214107e-01 -1.33948898e+00 -9.37966645e-01 -4.32289571e-01
-4.90600243e-02 -4.84108925e-01 3.47672105e-01 8.96026552e-01
1.19063184e-01 1.93424970e-01 4.77370054e-01 -1.47734714e+00
-2.52046764e-01 -6.58773839e-01 -4.36351478e-01 4.30730224e-01
4.16614473e-01 -8.86805773e-01 -2.80488193e-01 1.19735613e-01] | [7.043490409851074, -0.9693453311920166] |
aaf15698-45fb-49a3-a46d-e64c49b14b02 | codet5-open-code-large-language-models-for | 2305.07922 | null | https://arxiv.org/abs/2305.07922v2 | https://arxiv.org/pdf/2305.07922v2.pdf | CodeT5+: Open Code Large Language Models for Code Understanding and Generation | Large language models (LLMs) pretrained on vast source code have achieved prominent progress in code intelligence. However, existing code LLMs have two main limitations in terms of architecture and pretraining tasks. First, they often adopt a specific architecture (encoder-only or decoder-only) or rely on a unified encoder-decoder network for different downstream tasks. The former paradigm is limited by inflexibility in applications while in the latter, the model is treated as a single system for all tasks, leading to suboptimal performance on a subset of tasks. Secondly, they often employ a limited set of pretraining objectives which might not be relevant to some downstream tasks and hence result in substantial performance degrade. To address these limitations, we propose ``CodeT5+'', a family of encoder-decoder LLMs for code in which component modules can be flexibly combined to suit a wide range of downstream code tasks. Such flexibility is enabled by our proposed mixture of pretraining objectives to mitigate the pretrain-finetune discrepancy. These objectives cover span denoising, contrastive learning, text-code matching, and causal LM pretraining tasks, on both unimodal and bimodal multilingual code corpora. Furthermore, we propose to initialize CodeT5+ with frozen off-the-shelf LLMs without training from scratch to efficiently scale up our models, and explore instruction-tuning to align with natural language instructions. We extensively evaluate CodeT5+ on over 20 code-related benchmarks in different settings, including zero-shot, finetuning, and instruction-tuning. We observe state-of-the-art (SoTA) model performance on various code-related tasks, such as code generation and completion, math programming, and text-to-code retrieval tasks. Particularly, our instruction-tuned CodeT5+ 16B achieves new SoTA results on HumanEval code generation task against other open code LLMs. | ['Steven C. H. Hoi', 'Junnan Li', 'Nghi D. Q. Bui', 'Akhilesh Deepak Gotmare', 'Hung Le', 'Yue Wang'] | 2023-05-13 | null | null | null | null | ['code-search', 'code-search', 'arithmetic-reasoning'] | ['computer-code', 'computer-vision', 'reasoning'] | [ 1.63689807e-01 -2.33956844e-01 -4.23692137e-01 -3.35280985e-01
-1.05645132e+00 -5.81884027e-01 4.74189311e-01 1.15425447e-02
-1.64842337e-01 2.05317020e-01 1.62387013e-01 -8.73693764e-01
2.28990823e-01 -5.72050750e-01 -9.65653300e-01 -2.83021957e-01
5.64244092e-02 3.38677853e-01 3.34065408e-02 -4.69466180e-01
2.68221885e-01 -1.89059794e-01 -1.50158930e+00 7.62402415e-01
1.46844935e+00 7.04095304e-01 7.60617137e-01 7.40841091e-01
-4.81921047e-01 1.03226113e+00 -3.96874040e-01 -7.13068843e-01
6.62365183e-02 -2.66482979e-01 -6.66966856e-01 -2.69560516e-01
4.66133684e-01 -2.12405160e-01 -2.01757222e-01 1.10380137e+00
4.00239110e-01 -2.25410044e-01 6.63483620e-01 -1.00380468e+00
-1.09668601e+00 9.90354896e-01 -6.83622301e-01 -1.01406559e-01
3.97312969e-01 2.39902034e-01 9.55583096e-01 -1.14241171e+00
2.90705085e-01 1.17980075e+00 8.31699193e-01 6.27072573e-01
-1.47312498e+00 -6.54432178e-01 1.03925347e-01 -7.85286203e-02
-1.46899545e+00 -5.78414977e-01 3.97775114e-01 -7.52832353e-01
1.43813002e+00 -3.81174944e-02 -2.75899787e-02 1.16378856e+00
3.88685435e-01 8.88190866e-01 6.58732593e-01 -3.57560396e-01
-4.24748249e-02 7.50915557e-02 -1.61279887e-02 1.04765439e+00
-1.17507234e-01 -4.48016934e-02 -2.67025739e-01 -2.37140879e-01
3.15533966e-01 -5.84407598e-02 -2.53907800e-01 -3.54651392e-01
-1.26986980e+00 9.39327717e-01 2.98382670e-01 1.88548118e-01
2.56838977e-01 3.61978024e-01 8.47103357e-01 7.17483342e-01
1.33907199e-01 3.70451748e-01 -7.58655012e-01 -4.23205167e-01
-1.03054821e+00 1.15497917e-01 8.95256162e-01 1.34497130e+00
9.54858243e-01 3.13210934e-01 -3.25330824e-01 1.19686961e+00
3.12158793e-01 6.03710115e-01 9.36256588e-01 -6.74150169e-01
1.21459246e+00 6.21048152e-01 -6.77048504e-01 -5.81467092e-01
-1.31682113e-01 -4.85560775e-01 -8.50757837e-01 1.95178036e-02
1.34546384e-01 -1.73719540e-01 -8.70511532e-01 1.85072470e+00
-4.57366496e-01 -1.23111596e-02 1.20035060e-01 4.83390480e-01
9.37049508e-01 6.69145823e-01 -7.42439553e-02 2.08989501e-01
1.00258899e+00 -1.46178412e+00 -2.47045636e-01 -6.79839194e-01
1.13220990e+00 -1.02704787e+00 1.54905760e+00 3.25225621e-01
-1.06781733e+00 -8.85228515e-01 -1.06140542e+00 -2.93462485e-01
-2.43603885e-01 4.70196456e-01 6.53265893e-01 5.39113045e-01
-1.10720396e+00 3.73090863e-01 -7.97878385e-01 -1.51254147e-01
9.54388380e-02 3.84134889e-01 -1.50931984e-01 -2.39258185e-01
-9.30008590e-01 7.18262851e-01 4.24335003e-01 -2.99396396e-01
-1.18636036e+00 -6.74962878e-01 -1.14780223e+00 2.84835160e-01
1.59602433e-01 -6.93192363e-01 1.34827185e+00 -9.49918389e-01
-1.46923351e+00 8.64641070e-01 -1.12425424e-01 -4.12047446e-01
2.81252086e-01 -2.04547808e-01 -4.58105564e-01 -3.58755529e-01
1.45233810e-01 8.39677811e-01 1.01828527e+00 -1.09468257e+00
-3.06406111e-01 2.82775551e-01 9.83659402e-02 -5.01732081e-02
-6.32687986e-01 -7.10216016e-02 -6.59490645e-01 -8.78619552e-01
-2.80163229e-01 -9.67225671e-01 1.62741058e-02 -4.22393054e-01
-2.60232091e-01 -7.43047744e-02 6.82432771e-01 -5.13289154e-01
1.61086881e+00 -2.41446090e+00 3.54705215e-01 -1.98948190e-01
4.13061120e-04 2.70705849e-01 -6.85010791e-01 4.74902809e-01
-2.60622621e-01 1.16886362e-01 -6.82057440e-01 -5.51558077e-01
1.30843654e-01 3.76570910e-01 -5.84583759e-01 1.00450143e-01
2.39860818e-01 1.03952324e+00 -7.60879338e-01 -5.00302732e-01
-6.51777759e-02 1.02614999e-01 -1.28634751e+00 3.41949791e-01
-4.66914535e-01 1.54113755e-01 -2.63702780e-01 6.34912193e-01
3.91619742e-01 -2.87691146e-01 -1.55325696e-01 3.22666287e-01
-3.68151553e-02 4.71779704e-01 -7.28809834e-01 2.46489739e+00
-1.16394818e+00 6.68611050e-01 1.18918426e-01 -9.46887136e-01
6.78246021e-01 2.33579695e-01 1.06589265e-01 -5.48826814e-01
-9.46651846e-02 4.71357644e-01 2.79727697e-01 -7.65406311e-01
4.20278847e-01 2.36871719e-01 -3.68929207e-01 5.13583243e-01
2.41038412e-01 -4.27429676e-01 4.40637946e-01 2.26432621e-01
1.34689319e+00 2.87512332e-01 -1.22192660e-02 -2.72645980e-01
7.71849036e-01 -1.63334146e-01 2.00174093e-01 8.49127710e-01
1.26864463e-01 7.84086168e-01 4.38328147e-01 -8.37273300e-02
-8.19763362e-01 -9.27035093e-01 -3.05275738e-01 1.67300594e+00
-2.39304498e-01 -7.63736546e-01 -7.65601635e-01 -7.61393845e-01
1.18667940e-02 7.28656769e-01 -3.52543712e-01 -4.44685429e-01
-6.81207716e-01 -8.10469747e-01 9.50772047e-01 4.73364115e-01
5.01069188e-01 -8.46114576e-01 -2.67840475e-01 2.30079591e-01
-2.94477403e-01 -8.93329144e-01 -9.60134447e-01 5.66706479e-01
-7.82726288e-01 -8.75335276e-01 -4.18932766e-01 -1.02276790e+00
7.18121350e-01 1.02150023e-01 1.57775152e+00 3.55053455e-01
-2.74577420e-02 2.37055376e-01 -3.56571168e-01 1.30630180e-01
-1.11109602e+00 5.77038586e-01 -1.85077474e-01 -3.61245155e-01
1.29808784e-01 -6.61944568e-01 -2.42696166e-01 9.06483829e-02
-1.02437639e+00 1.09872550e-01 1.01144075e+00 1.14037836e+00
2.46957447e-02 -2.37253115e-01 4.84649658e-01 -8.13406408e-01
9.12686467e-01 -7.74319112e-01 -5.42233229e-01 4.44207072e-01
-6.92778230e-01 5.56883574e-01 9.37654078e-01 -6.30095303e-01
-1.05513561e+00 3.14827524e-02 -4.76736695e-01 -4.37641829e-01
7.83787742e-02 8.62118006e-01 -9.03640240e-02 -1.72830462e-01
9.23182428e-01 4.70124543e-01 -1.48750618e-01 -4.46012676e-01
4.93777812e-01 7.67405272e-01 6.53098941e-01 -1.23634899e+00
7.30681479e-01 -1.47949338e-01 -4.88719702e-01 -2.27280229e-01
-5.32633364e-01 -2.22217411e-01 -4.62253690e-01 3.42818439e-01
6.65192604e-01 -1.19175541e+00 -4.64586139e-01 3.39505285e-01
-1.26465261e+00 -7.03404248e-01 2.57536232e-01 1.40846476e-01
-4.64410156e-01 4.33412939e-01 -9.58882332e-01 -9.02005583e-02
-3.83702636e-01 -1.97959626e+00 1.21619689e+00 -2.02467769e-01
-2.87322491e-01 -1.06984830e+00 7.33019784e-02 2.32360095e-01
7.28923500e-01 -3.40159565e-01 1.42612636e+00 -4.92151231e-01
-4.67961192e-01 9.54724625e-02 -2.03816965e-01 5.54167211e-01
-3.88053171e-02 7.59284198e-02 -8.71450424e-01 -4.79345977e-01
-4.33018208e-02 -8.69133353e-01 1.13648474e+00 1.63798295e-02
1.36884224e+00 -2.97993809e-01 -2.18835101e-01 1.06526649e+00
1.35281444e+00 -1.81974284e-02 4.20852154e-01 1.18764929e-01
8.62739801e-01 1.78869247e-01 1.20886169e-01 3.19199324e-01
5.37614226e-01 6.84188008e-01 5.48155844e-01 2.54193544e-01
-1.70646057e-01 -4.01536882e-01 1.00027359e+00 1.49390495e+00
3.96944970e-01 -2.70832647e-02 -1.14113557e+00 4.65721458e-01
-1.74678421e+00 -4.53402340e-01 -2.14685962e-01 2.04467058e+00
1.24091148e+00 2.11202323e-01 -3.24743032e-01 -3.40454459e-01
3.85014653e-01 3.54823209e-02 -4.69540864e-01 -6.29544973e-01
5.86760566e-02 2.10415423e-01 4.27117407e-01 3.73923302e-01
-9.74349320e-01 9.34220731e-01 5.65794325e+00 1.29213929e+00
-1.15556097e+00 4.43298817e-01 4.99990076e-01 1.24672174e-01
-4.90276873e-01 1.65496081e-01 -7.22266197e-01 5.53644955e-01
1.17869401e+00 -4.10958529e-02 7.78648973e-01 1.21715379e+00
-2.73235559e-01 1.14503510e-01 -1.50327098e+00 1.01200080e+00
1.40670002e-01 -1.23192704e+00 3.47604305e-02 -1.46922007e-01
1.06581593e+00 5.17448425e-01 2.33928159e-01 1.09926057e+00
4.82615501e-01 -1.03279245e+00 9.96140301e-01 3.71040218e-02
1.16831064e+00 -5.52611172e-01 4.58929956e-01 5.14508128e-01
-1.37061191e+00 -2.95387208e-01 -3.12196136e-01 -1.18597373e-01
-1.61472678e-01 4.00646448e-01 -4.46608365e-01 2.99494743e-01
4.45441127e-01 7.57737100e-01 -9.55362380e-01 8.03814471e-01
-2.76875407e-01 5.77720344e-01 1.21065103e-01 2.46208444e-01
2.94219762e-01 1.08850680e-01 1.63783923e-01 1.61367047e+00
5.80921590e-01 -4.50432271e-01 4.34606820e-01 1.27762973e+00
-4.03684407e-01 7.47137889e-03 -6.19089186e-01 -4.95920107e-02
4.42562878e-01 1.13584852e+00 -3.31306159e-01 -2.83385992e-01
-8.99302185e-01 1.04946864e+00 6.30382001e-01 4.12257403e-01
-1.03503716e+00 -5.42191982e-01 6.70833468e-01 -4.85768430e-02
2.44110897e-01 -3.30799580e-01 -2.83674896e-01 -1.58144963e+00
-1.02166697e-01 -1.33677053e+00 2.51718163e-01 -5.35267770e-01
-1.15052164e+00 6.86227739e-01 -6.33523315e-02 -1.19612610e+00
-4.38308984e-01 -7.13895559e-01 -6.59918249e-01 7.72480011e-01
-1.43870151e+00 -1.00382769e+00 -8.34720805e-02 5.75720251e-01
8.52419674e-01 -4.99465734e-01 7.35054910e-01 7.75252879e-01
-5.28844655e-01 9.36623752e-01 2.79135495e-01 2.92218000e-01
8.98116410e-01 -1.22356451e+00 6.45781040e-01 9.94397879e-01
7.64301121e-02 9.70640600e-01 3.83963168e-01 -4.54526246e-01
-1.62135971e+00 -1.27352941e+00 6.73643053e-01 -5.30796170e-01
9.25517380e-01 -7.06939518e-01 -9.64305818e-01 7.02105999e-01
1.98185995e-01 9.06585343e-03 4.19091761e-01 1.33621812e-01
-5.84742725e-01 4.45953757e-03 -4.87722546e-01 6.01985514e-01
1.01794732e+00 -8.66293132e-01 -5.25293410e-01 2.52368867e-01
9.57485795e-01 -7.68712401e-01 -8.81940722e-01 3.86871010e-01
1.30523518e-01 -1.01221991e+00 9.99745667e-01 -3.13669235e-01
1.13117135e+00 -1.49803221e-01 -2.14986220e-01 -1.31954730e+00
-1.55538321e-01 -5.47977328e-01 -1.65592045e-01 1.45073128e+00
6.01443052e-01 -3.55979234e-01 2.58482158e-01 4.00282264e-01
-7.80564427e-01 -7.41782188e-01 -6.45927370e-01 -7.36597478e-01
5.31868935e-01 -7.80021310e-01 6.10036016e-01 8.50365639e-01
1.86935052e-01 3.50349605e-01 -2.75227875e-01 -1.83140919e-01
1.12458497e-01 2.82281816e-01 8.25769067e-01 -7.23282695e-01
-8.21001530e-01 -8.41140628e-01 2.61906832e-01 -1.54403770e+00
6.87277377e-01 -1.51693594e+00 2.41498038e-01 -1.00947857e+00
4.24732208e-01 -5.19168735e-01 6.81364685e-02 7.22863853e-01
-2.70111918e-01 -1.47754088e-01 1.20390713e-01 4.13287491e-01
-5.22923589e-01 6.27902627e-01 9.47451115e-01 -4.65645909e-01
-8.74962136e-02 -1.21882565e-01 -5.98508298e-01 5.44597685e-01
4.54732805e-01 -5.30240953e-01 -6.70463681e-01 -1.05556571e+00
5.31368315e-01 3.25015068e-01 1.07391536e-01 -1.05455828e+00
1.55235797e-01 2.75950711e-02 -1.47467539e-01 -1.17279790e-01
-1.27173692e-01 -5.64548790e-01 -1.64906517e-01 5.74107111e-01
-4.71963257e-01 3.76560688e-01 3.79141986e-01 3.84953529e-01
-4.28426176e-01 -7.03852475e-01 6.58040166e-01 -2.54512697e-01
-7.90534258e-01 1.07836448e-01 -3.90269995e-01 4.38229978e-01
4.98902380e-01 1.05363414e-01 -4.65867162e-01 -2.15983883e-01
-1.94805533e-01 2.64294744e-01 5.11892378e-01 7.43807852e-01
5.20197570e-01 -1.27151525e+00 -7.20282674e-01 4.08856601e-01
5.12873292e-01 -3.68333384e-02 -1.13195311e-02 8.44295740e-01
-5.73276997e-01 6.21754587e-01 1.14096843e-01 -6.88850641e-01
-8.42632353e-01 7.41457939e-01 3.24635208e-01 -5.57541788e-01
-3.17761064e-01 9.28119898e-01 4.90430683e-01 -9.08460200e-01
2.42865607e-01 -8.01097333e-01 1.61569864e-01 -3.85936558e-01
2.69325107e-01 -5.28330989e-02 2.94122249e-01 -3.96364957e-01
-1.31452024e-01 5.10209441e-01 -5.13572730e-02 3.11647356e-01
1.04089952e+00 -4.36928682e-02 -4.25952375e-01 1.98830500e-01
1.50332320e+00 1.63219422e-01 -1.26730800e+00 -2.70403147e-01
1.52507693e-01 -1.67328864e-01 -2.41148561e-01 -5.89193225e-01
-1.16114295e+00 1.08302176e+00 2.34092295e-01 -2.34292522e-01
1.07768703e+00 -1.41826034e-01 8.95157516e-01 6.45975530e-01
3.33488047e-01 -8.73652995e-01 3.10610026e-01 1.05142558e+00
8.00522029e-01 -1.52106357e+00 -3.95642608e-01 1.83552913e-02
-4.76087779e-01 1.30060017e+00 8.98566484e-01 1.76992595e-01
4.40983355e-01 7.10148156e-01 -1.74400404e-01 1.18517481e-01
-1.11718118e+00 -8.57836828e-02 3.96243542e-01 2.32606634e-01
1.01551735e+00 -1.57536373e-01 -6.47769719e-02 6.40548944e-01
-1.88412294e-01 -1.37721449e-01 4.12945926e-01 8.57805192e-01
-2.40858927e-01 -1.40627348e+00 -3.30015987e-01 5.18138587e-01
-3.18949699e-01 -7.57709742e-01 1.19722381e-01 5.41425347e-01
2.91942477e-01 8.56663585e-01 -2.72853285e-01 -4.95918304e-01
4.52962071e-02 2.10624859e-01 3.61109138e-01 -9.97131288e-01
-8.85325968e-01 -8.63859504e-02 -1.40330806e-01 -3.49409193e-01
5.02521954e-02 -3.80509019e-01 -1.31220567e+00 -3.64430904e-01
-2.65616745e-01 -3.96950245e-02 4.49354142e-01 9.08877850e-01
5.81163287e-01 7.86963582e-01 1.94587171e-01 -7.44655252e-01
-7.15776324e-01 -1.02057707e+00 1.71722658e-02 5.27425945e-01
2.50276655e-01 -3.96440148e-01 -1.54949605e-01 2.90433794e-01] | [7.700126647949219, 7.90936279296875] |
718aa02a-6f7f-41af-84b5-e9000ae0650d | a-first-look-at-llm-powered-generative-news | 2305.06566 | null | https://arxiv.org/abs/2305.06566v2 | https://arxiv.org/pdf/2305.06566v2.pdf | A First Look at LLM-Powered Generative News Recommendation | Personalized news recommendation systems have become essential tools for users to navigate the vast amount of online news content, yet existing news recommenders face significant challenges such as the cold-start problem, user profile modeling, and news content understanding. Previous works have typically followed an inflexible routine to address a particular challenge through model design, but are limited in their ability to understand news content and capture user interests. In this paper, we introduce GENRE, an LLM-powered generative news recommendation framework, which leverages pretrained semantic knowledge from large language models to enrich news data. Our aim is to provide a flexible and unified solution for news recommendation by moving from model design to prompt design. We showcase the use of GENRE for personalized news generation, user profiling, and news summarization. Extensive experiments with various popular recommendation models demonstrate the effectiveness of GENRE. We will publish our code and data for other researchers to reproduce our work. | ['Xiao-Ming Wu', 'Tetsuya Sakai', 'Nuo Chen', 'Qijiong Liu'] | 2023-05-11 | null | null | null | null | ['news-generation'] | ['natural-language-processing'] | [-2.31850445e-02 6.27869442e-02 -5.40760815e-01 -5.69534481e-01
-8.82961631e-01 -4.99416441e-01 7.20786333e-01 -1.22586772e-01
-5.22946753e-02 6.15615487e-01 1.33151543e+00 -1.57008767e-01
-1.54905140e-01 -6.07354224e-01 -4.63351607e-01 -8.96247178e-02
2.69340098e-01 5.95590770e-01 1.39658093e-01 -6.91609502e-01
4.80611682e-01 -2.21709624e-01 -1.67689538e+00 7.49808967e-01
7.97626913e-01 7.52958417e-01 3.99642050e-01 7.39178777e-01
-5.12619913e-01 1.01302421e+00 -4.38622564e-01 -4.75855082e-01
-1.77431837e-01 -5.33948183e-01 -9.47027564e-01 -8.85555372e-02
4.21236791e-02 -3.22225839e-01 -2.12183923e-01 6.39768004e-01
7.56031454e-01 7.39709079e-01 5.03893793e-01 -7.49629557e-01
-1.21060002e+00 1.48899817e+00 5.04567437e-02 3.04796994e-01
5.67511618e-01 -7.62845397e-01 1.29983592e+00 -7.96347678e-01
6.92036867e-01 9.63531256e-01 6.68667793e-01 7.94788599e-01
-8.75554442e-01 -1.31119058e-01 6.94451988e-01 8.17234665e-02
-8.83267939e-01 -5.63575506e-01 6.48224056e-01 -4.92416590e-01
8.90884876e-01 5.81381083e-01 6.94295824e-01 1.62954426e+00
-7.83392340e-02 9.94826019e-01 5.62462807e-01 -1.65173128e-01
2.15002730e-01 2.62088686e-01 4.91521150e-01 1.91228583e-01
-2.68638171e-02 -2.35331908e-01 -5.50146580e-01 -3.89325708e-01
3.29729557e-01 4.01491553e-01 -2.27102324e-01 1.12639122e-01
-8.06708932e-01 1.06128502e+00 -6.03629053e-02 1.61151439e-01
-3.85193110e-01 -2.13425472e-01 2.35728294e-01 3.62763226e-01
1.00997698e+00 1.03551400e+00 -6.18225992e-01 -3.91011745e-01
-7.17739880e-01 6.35278761e-01 1.28223717e+00 1.12380958e+00
1.97918266e-01 -9.78804305e-02 -5.09619355e-01 1.26720452e+00
3.89242560e-01 2.41860256e-01 8.63502026e-01 -6.14997387e-01
-5.63647039e-02 2.36251727e-01 3.29185992e-01 -8.04962158e-01
-5.26612580e-01 -8.38187158e-01 -1.76065102e-01 -8.11780035e-01
-7.11185560e-02 -5.87584198e-01 -6.55040681e-01 1.38445902e+00
2.42715418e-01 1.97306618e-01 2.66416967e-01 7.26800859e-01
1.24468982e+00 1.05480909e+00 1.85955167e-01 -2.48496681e-01
1.33154261e+00 -1.35480285e+00 -6.46419287e-01 -3.52584720e-01
6.67151690e-01 -1.01908731e+00 1.21615458e+00 2.73263544e-01
-1.00875056e+00 -2.74265409e-01 -3.94923866e-01 -1.69873029e-01
-2.08041608e-01 2.02077955e-01 7.67521381e-01 4.78402227e-01
-7.84067750e-01 5.01821637e-01 -4.87289935e-01 -6.36315525e-01
-4.14507054e-02 -1.24180064e-01 3.27773899e-01 -1.37060389e-01
-1.30439007e+00 5.12436271e-01 1.19298093e-01 -3.31041366e-01
-5.66334426e-01 -8.07340026e-01 -4.63738680e-01 1.99218795e-01
6.22741580e-01 -1.01513815e+00 1.77841389e+00 -8.22311997e-01
-1.87979138e+00 8.25729370e-02 -6.30152449e-02 -2.71688432e-01
7.86209106e-02 -7.30716169e-01 -6.67741418e-01 -6.06183529e-01
6.80286903e-03 -2.82648653e-02 4.45776075e-01 -1.12247598e+00
-9.11495566e-01 -9.03138611e-03 1.31711036e-01 4.33266133e-01
-4.25234735e-01 4.05137360e-01 -6.32425845e-01 -1.04253685e+00
-2.91872293e-01 -7.71495938e-01 -4.57332313e-01 -1.10782206e+00
-5.37228703e-01 -1.55488670e-01 3.64307225e-01 -5.38711786e-01
1.55460453e+00 -1.96213126e+00 -3.33896726e-02 2.44807936e-02
2.41913274e-02 -5.07228114e-02 -1.15461089e-01 7.46391177e-01
5.30270636e-01 7.22803921e-03 5.80985546e-01 -4.22046393e-01
2.61426568e-01 -4.28854339e-02 -8.54640663e-01 -2.24965811e-01
-7.68747032e-01 8.96103978e-01 -1.02817988e+00 1.07138148e-02
-3.96470428e-01 5.35735250e-01 -1.10804534e+00 3.24271798e-01
-6.99148476e-01 3.47823143e-01 -8.78457487e-01 2.46129438e-01
-9.77255926e-02 -5.91962814e-01 1.62079170e-01 1.37047460e-02
-1.36673138e-01 8.46809328e-01 -9.38544691e-01 1.44371343e+00
-5.62901437e-01 6.81825876e-02 -2.91868329e-01 -4.04757529e-01
8.87063682e-01 1.98513925e-01 5.15312135e-01 -6.04143143e-01
2.98678041e-01 -1.51468843e-01 -3.71727496e-01 -6.04764223e-01
1.29075444e+00 8.17095861e-02 -1.89424500e-01 9.67889428e-01
-4.16115001e-02 4.57689434e-01 2.03361943e-01 3.92133266e-01
7.45369971e-01 1.14433825e-01 1.96254417e-01 -2.23583117e-01
-1.25445247e-01 4.09673117e-02 4.17045474e-01 1.13614058e+00
5.17190576e-01 5.53905904e-01 1.79976866e-01 -3.30051154e-01
-7.50460565e-01 -5.56578696e-01 1.19160526e-01 2.21678543e+00
-3.99086326e-02 -1.09588623e+00 -5.26857793e-01 -7.21088171e-01
-2.65406549e-01 1.18244696e+00 -5.38396060e-01 -9.24018621e-02
-3.65880430e-01 -1.01203287e+00 -8.93045813e-02 5.70100904e-01
-2.71186680e-01 -9.05767024e-01 6.93415403e-02 5.30659914e-01
-3.95415276e-01 -7.70933390e-01 -8.39747250e-01 -1.98664606e-01
-5.70110261e-01 -5.60814440e-01 -6.47281647e-01 -6.05995357e-01
4.40786242e-01 5.94820857e-01 1.18883991e+00 -1.29543617e-01
2.86538780e-01 7.11773515e-01 -1.12902188e+00 -5.10811508e-01
-3.64100426e-01 5.26981473e-01 1.36680692e-01 8.04230794e-02
1.62163898e-01 -6.02134049e-01 -6.33603036e-01 4.37721878e-01
-7.74537265e-01 2.73994803e-01 2.42970422e-01 5.49568355e-01
4.99239296e-01 -2.60711759e-01 9.14285183e-01 -1.72904682e+00
1.22680366e+00 -1.12704110e+00 -1.06410116e-01 3.48161280e-01
-9.41645265e-01 -6.57942295e-02 8.66491020e-01 -4.81749624e-01
-1.46764135e+00 -4.45768535e-01 -6.89581156e-01 3.48990232e-01
-2.74695829e-03 1.02520037e+00 1.41688317e-01 4.78388488e-01
8.66452932e-01 3.44712406e-01 -6.05204463e-01 -1.09583080e+00
7.87631035e-01 7.08544850e-01 3.11960459e-01 -6.81478739e-01
1.95703626e-01 2.46999085e-01 -9.85567987e-01 -5.28482080e-01
-1.56165612e+00 -6.71850026e-01 2.54807234e-01 -5.70224188e-02
4.04429495e-01 -9.20793533e-01 -6.05561845e-02 -4.39960025e-02
-6.98595345e-01 -2.80826479e-01 -5.24063051e-01 6.67950749e-01
-4.24539030e-01 1.38527811e-01 -8.59133184e-01 -6.00958526e-01
-6.44812763e-01 -7.71086514e-01 7.79643774e-01 2.83591360e-01
-4.18510258e-01 -1.20531237e+00 4.77937102e-01 5.22376537e-01
1.02439797e+00 -3.98811966e-01 6.43810153e-01 -1.40949488e+00
-1.87556371e-01 -3.46578449e-01 1.95434198e-01 -4.97017056e-02
1.45817269e-02 -2.08371673e-02 -8.07534873e-01 -5.65149002e-02
-1.90501690e-01 -1.81845278e-01 7.96624899e-01 4.78649735e-01
1.12952447e+00 -7.90991366e-01 -3.53610367e-01 5.67790985e-01
7.88019300e-01 2.77779728e-01 3.44385356e-01 3.84737581e-01
7.93717682e-01 4.72039044e-01 4.23367321e-01 8.61335695e-01
8.68941426e-01 6.70312226e-01 -2.52061427e-01 2.59318143e-01
8.61495081e-03 -6.93265080e-01 3.73238236e-01 1.19840682e+00
-1.90602168e-01 -7.15761483e-01 -5.01185417e-01 2.30825886e-01
-2.10309982e+00 -1.19726229e+00 1.65080592e-01 1.75413978e+00
7.40644276e-01 -8.69327113e-02 3.78201127e-01 -5.88739634e-01
4.44873542e-01 1.41639411e-01 -1.81226000e-01 -1.69878826e-01
1.55750006e-01 -3.52264822e-01 1.74888939e-01 5.40891111e-01
-8.34691942e-01 1.07440495e+00 6.79244661e+00 6.82608664e-01
-9.79156733e-01 2.59459376e-01 3.97680283e-01 -4.39891756e-01
-8.90412509e-01 8.49886015e-02 -1.25843775e+00 6.62728548e-01
1.00725758e+00 -5.70165992e-01 3.85191381e-01 1.32343674e+00
3.16724360e-01 3.12889665e-01 -8.79040956e-01 5.87091982e-01
4.07271236e-01 -1.64668798e+00 1.57031968e-01 -1.81628004e-01
1.13979065e+00 2.35283643e-01 6.65125474e-02 7.43348897e-01
8.97769570e-01 -5.13362110e-01 6.40411079e-01 7.80499697e-01
1.50252119e-01 -5.43401659e-01 5.55218160e-01 4.20543104e-01
-6.65924132e-01 -3.39730173e-01 -3.91398758e-01 -1.91028371e-01
6.69960320e-01 5.78060687e-01 -7.28563488e-01 3.01582456e-01
5.94261289e-01 8.43730927e-01 -1.90021992e-01 1.13750589e+00
-5.78558892e-02 8.39137971e-01 1.68444961e-02 -3.26458752e-01
4.74700369e-02 -2.53281333e-02 5.87534547e-01 1.35757482e+00
6.27693355e-01 1.96060926e-01 4.80256915e-01 2.94589102e-01
-2.68200070e-01 7.33845949e-01 -1.63093314e-01 -1.46265760e-01
4.99969184e-01 1.14220321e+00 -6.25334144e-01 -4.30504918e-01
-5.07848561e-01 7.26599395e-01 1.45517752e-01 3.49576384e-01
-6.27101421e-01 -9.42327008e-02 6.40670836e-01 3.39564115e-01
3.33900154e-01 2.24747360e-01 -7.88562372e-02 -1.55284703e+00
-5.47157884e-01 -9.42893624e-01 5.43119133e-01 -6.70302451e-01
-1.51203310e+00 8.54860544e-01 -5.08592948e-02 -7.91661084e-01
-4.91510272e-01 -6.25935718e-02 -6.55610383e-01 4.13883686e-01
-1.17204702e+00 -1.02214408e+00 -1.92726981e-02 3.54622602e-01
8.72162879e-01 -3.74787480e-01 7.65502095e-01 5.14499426e-01
-5.40633500e-01 4.84520257e-01 6.89818323e-01 -3.19422156e-01
7.36332476e-01 -1.04891014e+00 6.26995087e-01 7.21181512e-01
3.19429755e-01 8.73238206e-01 9.97579694e-01 -8.35816443e-01
-1.36116219e+00 -1.25726116e+00 1.01421297e+00 -8.37226808e-01
6.49369776e-01 -2.89728820e-01 -6.62250876e-01 1.00775707e+00
1.04276352e-02 -5.80220997e-01 1.37813079e+00 8.27713907e-01
-3.20739627e-01 7.59018734e-02 -5.97807169e-01 7.29963303e-01
1.12098980e+00 -1.89124957e-01 -6.75997972e-01 5.62877953e-01
9.92869020e-01 -4.62738037e-01 -8.22088540e-01 -3.93005200e-02
5.89215755e-01 -6.27805293e-01 7.82264352e-01 -9.97348607e-01
4.34335530e-01 -7.19368905e-02 -1.19380273e-01 -1.66751778e+00
-8.46781611e-01 -9.00985956e-01 -4.67743814e-01 1.46751177e+00
6.53803229e-01 -3.67908776e-01 5.36700606e-01 7.03074694e-01
-6.09292984e-01 -7.40598857e-01 2.27489807e-02 -3.53345484e-01
-4.19278860e-01 -7.45123148e-01 7.96781301e-01 1.06811571e+00
3.51823717e-01 8.63272667e-01 -1.05415702e+00 -1.16357982e-01
5.96679188e-02 4.05739278e-01 7.93733358e-01 -1.25264549e+00
-7.17248023e-01 -5.00160336e-01 3.70345861e-01 -1.48384631e+00
-1.20677352e-01 -8.88133466e-01 6.52542338e-02 -1.96863818e+00
4.75623786e-01 -4.64406610e-01 -1.96624339e-01 2.00506762e-01
-2.70163417e-01 -3.45188752e-02 -5.22364639e-02 3.32444757e-01
-1.14168465e+00 5.05799592e-01 1.11563623e+00 3.66510093e-01
-4.51931506e-01 5.54251075e-01 -1.65369904e+00 6.89619005e-01
6.80915475e-01 -5.59217870e-01 -9.20558870e-01 -5.77607214e-01
9.20882583e-01 -1.31615758e-01 -2.66906768e-01 -4.68289137e-01
2.73556441e-01 -3.58770639e-01 1.24156088e-01 -2.10705325e-01
6.68413043e-02 -3.25881958e-01 2.74952561e-01 -3.37937653e-01
-9.23201799e-01 -2.31691208e-02 -2.54992396e-01 7.84884334e-01
5.16074374e-02 -1.57852262e-01 1.39236465e-01 -1.62326291e-01
-8.78918290e-01 5.29619932e-01 -6.47964537e-01 3.51633847e-01
5.68261802e-01 2.88011789e-01 -5.68530262e-01 -9.60958838e-01
-9.34406698e-01 2.13555068e-01 2.89283216e-01 8.96951377e-01
4.51567888e-01 -1.17845297e+00 -7.56924033e-01 1.60252396e-02
3.13571632e-01 -5.02881765e-01 6.62726223e-01 5.23625612e-01
7.33326972e-02 2.94708967e-01 2.10986868e-01 2.49001026e-01
-9.59672332e-01 5.50331235e-01 -7.14688376e-02 -2.58144468e-01
-8.23800027e-01 1.05591953e+00 6.44685403e-02 -2.91643351e-01
3.34068477e-01 -8.31791386e-02 -7.82327831e-01 1.90155447e-01
1.16936350e+00 3.51767540e-01 -1.18940622e-01 -3.61945570e-01
1.56826735e-01 8.94666687e-02 -5.33219397e-01 9.88099203e-02
1.69577038e+00 -5.27359664e-01 2.96157479e-01 3.80783170e-01
8.14827681e-01 3.99892360e-01 -1.00616550e+00 -5.88149309e-01
3.79153341e-02 -3.30798656e-01 2.71797389e-01 -9.83955026e-01
-8.01149130e-01 9.78840888e-02 4.86703254e-02 6.13329649e-01
7.40376055e-01 4.71466899e-01 1.18676531e+00 5.06776392e-01
1.65544778e-01 -1.21929967e+00 -1.53305367e-01 9.39790905e-01
6.85768902e-01 -8.96758139e-01 -1.21653140e-01 -7.88545981e-02
-1.14578128e+00 8.62349868e-01 5.38810790e-01 2.41396442e-01
8.41318071e-01 3.36430311e-01 2.02631682e-01 -2.21221924e-01
-1.18304038e+00 -2.67488837e-01 6.38480544e-01 2.89860219e-01
9.03445125e-01 -2.29429398e-02 -3.29825044e-01 1.51100683e+00
-4.62508053e-01 1.43156871e-01 5.83252013e-01 8.54598165e-01
-8.87227178e-01 -1.16496515e+00 8.55725184e-02 7.61445522e-01
-8.25220883e-01 -2.38573581e-01 -2.99602062e-01 8.63292292e-02
-2.67275035e-01 9.67316568e-01 -1.41382843e-01 -6.70076966e-01
4.20531452e-01 8.91600847e-02 2.26120390e-02 -1.06152701e+00
-5.04108906e-01 3.10669512e-01 5.79742670e-01 -4.08131689e-01
-1.70410514e-01 -6.27222776e-01 -8.03735614e-01 -3.97339642e-01
-2.13846594e-01 6.56147480e-01 6.49598718e-01 1.01486421e+00
8.19945455e-01 5.71523845e-01 5.18129826e-01 -6.14008069e-01
-4.63664770e-01 -9.74765778e-01 -6.01276159e-01 2.69904315e-01
-1.54261123e-02 -4.18232352e-01 -3.99603769e-02 2.46424191e-02] | [10.328009605407715, 5.828691005706787] |
2b9da343-3b4c-4c38-ac09-a30ba9ebeb6a | weighted-low-rank-matrix-approximation-and | 2109.11057 | null | https://arxiv.org/abs/2109.11057v1 | https://arxiv.org/pdf/2109.11057v1.pdf | Weighted Low Rank Matrix Approximation and Acceleration | Low-rank matrix approximation is one of the central concepts in machine learning, with applications in dimension reduction, de-noising, multivariate statistical methodology, and many more. A recent extension to LRMA is called low-rank matrix completion (LRMC). It solves the LRMA problem when some observations are missing and is especially useful for recommender systems. In this paper, we consider an element-wise weighted generalization of LRMA. The resulting weighted low-rank matrix approximation technique therefore covers LRMC as a special case with binary weights. WLRMA has many applications. For example, it is an essential component of GLM optimization algorithms, where an exponential family is used to model the entries of a matrix, and the matrix of natural parameters admits a low-rank structure. We propose an algorithm for solving the weighted problem, as well as two acceleration techniques. Further, we develop a non-SVD modification of the proposed algorithm that is able to handle extremely high-dimensional data. We compare the performance of all the methods on a small simulation example as well as a real-data application. | ['Trevor Hastie', 'Elena Tuzhilina'] | 2021-09-22 | null | null | null | null | ['low-rank-matrix-completion'] | ['methodology'] | [ 3.26756567e-01 -1.98519126e-01 -1.36653841e-01 -1.62791505e-01
-6.21551812e-01 -2.41907790e-01 4.19909716e-01 -6.95269033e-02
-4.67596650e-01 5.98382711e-01 4.12244231e-01 -4.65707958e-01
-6.57863855e-01 -4.57322955e-01 -5.56533396e-01 -9.28943276e-01
-3.63152504e-01 4.23454434e-01 -2.66397804e-01 -3.67602825e-01
1.60141408e-01 3.86535466e-01 -1.44980323e+00 1.83223803e-02
1.06789339e+00 6.63074851e-01 2.79468894e-01 4.87200648e-01
8.77668187e-02 3.61750424e-01 -2.04970292e-03 -9.41899493e-02
4.69873279e-01 -1.25802770e-01 -4.70321804e-01 3.62019658e-01
3.76459241e-01 1.08725742e-01 -3.25620174e-01 9.82064545e-01
3.75643849e-01 6.31754994e-01 7.59913504e-01 -1.14823067e+00
-2.21699610e-01 5.48178971e-01 -9.85407412e-01 8.76579527e-03
2.65040666e-01 -6.59937918e-01 9.58918631e-01 -1.46151388e+00
2.57808387e-01 1.51893747e+00 6.22335076e-01 3.65325585e-02
-1.57319856e+00 -4.28009659e-01 -6.73926994e-03 3.59426528e-01
-1.52396274e+00 -3.58542889e-01 5.60182929e-01 -5.64103663e-01
2.11038202e-01 5.97211182e-01 2.46725202e-01 6.26070023e-01
1.72818094e-01 7.59020627e-01 1.17110324e+00 -5.15076816e-01
2.49676391e-01 -3.12860385e-02 4.66853023e-01 4.10391897e-01
7.40393996e-01 -2.13997975e-01 -2.24585310e-01 -8.00494134e-01
4.37511325e-01 2.85880238e-01 -3.69492263e-01 -6.12906456e-01
-1.32012105e+00 1.04322588e+00 -1.31907433e-01 2.24368438e-01
-2.86708921e-01 -1.64376482e-01 1.05278371e-02 3.34575176e-01
4.36757982e-01 6.99380785e-02 -8.07024688e-02 1.78461164e-01
-8.35359693e-01 3.03002357e-01 8.59365761e-01 7.88273990e-01
6.63485765e-01 2.25179538e-01 -1.09874494e-01 1.14075875e+00
4.35789227e-01 6.50880575e-01 4.63446409e-01 -7.39445388e-01
6.77612841e-01 2.80945957e-01 1.96132138e-01 -1.27130139e+00
-6.80282950e-01 -6.79463387e-01 -1.40924799e+00 -9.22618210e-02
4.26860482e-01 5.67051284e-02 -1.68021053e-01 1.67913806e+00
4.20425057e-01 5.33092439e-01 1.21832095e-01 8.21898580e-01
2.74495125e-01 8.87586772e-01 -4.87597525e-01 -8.00723255e-01
1.18736422e+00 -5.66869497e-01 -8.35657895e-01 -1.50073200e-01
5.50414860e-01 -8.96815538e-01 9.17574227e-01 9.26269531e-01
-7.77342975e-01 -3.40218604e-01 -1.02473724e+00 1.01914868e-01
4.96151261e-02 4.45442557e-01 7.69214034e-01 6.23485923e-01
-8.13266098e-01 6.26733661e-01 -4.10471648e-01 -3.22538555e-01
-2.00638145e-01 5.03461778e-01 -6.17114425e-01 -4.88023520e-01
-1.12634265e+00 5.71411014e-01 2.30596453e-01 4.66287464e-01
-3.15807313e-01 -5.35743177e-01 -8.53331387e-01 -2.09160551e-01
4.34348553e-01 -2.56582201e-01 7.97016621e-01 -4.08512741e-01
-1.12583280e+00 3.33144963e-01 -2.32937202e-01 -2.66815782e-01
2.15334132e-01 -3.04663599e-01 -7.81656861e-01 -3.45335811e-01
-1.88182056e-01 -3.97356898e-01 1.25840855e+00 -8.78462613e-01
-5.88751376e-01 -5.08108616e-01 -2.66275495e-01 -2.27555707e-02
-4.71084356e-01 -2.30034292e-01 -2.17684031e-01 -9.68338549e-01
4.85199809e-01 -1.07027960e+00 -7.12715209e-01 -4.47559476e-01
-1.28222346e-01 -5.52370921e-02 5.52181065e-01 -6.88292623e-01
1.60980392e+00 -2.34332204e+00 6.85906410e-01 7.19646633e-01
1.78927511e-01 1.55163661e-01 -4.03868169e-01 7.73449957e-01
-3.66538763e-01 -2.52388507e-01 -3.69764090e-01 -2.27706835e-01
-6.94264472e-02 8.71676430e-02 -5.10360122e-01 8.85930061e-01
-2.08675534e-01 4.90487218e-02 -7.95808017e-01 -1.17528640e-01
-5.01913717e-03 3.24984699e-01 -5.78493714e-01 2.86367466e-03
2.68910855e-01 2.00031951e-01 -2.19045505e-01 2.92295426e-01
9.64557707e-01 -2.84503344e-02 4.34417546e-01 -5.22484779e-01
-2.55048454e-01 -4.47547227e-01 -2.17056990e+00 1.44944048e+00
-5.71711481e-01 3.97668868e-01 3.76245201e-01 -1.07568276e+00
8.94825220e-01 1.33298516e-01 5.93073666e-01 -1.93390269e-02
1.60174053e-02 3.37611020e-01 1.33833721e-01 -2.65673429e-01
7.56130457e-01 1.12065189e-01 1.37232915e-02 4.57345456e-01
-3.44124675e-01 3.18781197e-01 5.82457304e-01 3.69209081e-01
9.39864755e-01 -3.36800843e-01 8.01341057e-01 -7.03912854e-01
1.06507659e+00 -5.20324051e-01 8.32732856e-01 6.38780534e-01
5.16070068e-01 4.02245462e-01 4.55242932e-01 -2.25262538e-01
-7.92635977e-01 -7.44889081e-01 -3.39682728e-01 8.88760090e-01
-1.10240608e-01 -6.67036653e-01 -3.66684854e-01 -6.17298968e-02
5.84451370e-02 6.34601831e-01 -3.07149649e-01 -2.39303499e-01
-3.61721843e-01 -1.20275342e+00 -1.39915213e-01 9.96945947e-02
1.05038390e-01 -1.60501599e-01 2.80806303e-01 4.26419884e-01
-1.60396978e-01 -9.65823650e-01 -6.71097398e-01 -1.47072017e-01
-1.15589857e+00 -1.12377393e+00 -6.27578914e-01 -5.60245991e-01
7.24020839e-01 8.02479863e-01 6.98738933e-01 -1.53021559e-01
-1.50225058e-01 3.78769368e-01 -3.62905890e-01 -8.18803832e-02
-2.12370113e-01 -1.37423381e-01 8.42251837e-01 8.78651202e-01
1.08450055e-01 -7.35754788e-01 -3.84834439e-01 5.86740196e-01
-1.22508824e+00 -8.62070248e-02 8.01177621e-01 9.25644696e-01
7.62405217e-01 2.22036242e-01 5.73736966e-01 -1.24103451e+00
1.00699413e+00 -6.67269588e-01 -8.31674695e-01 3.01928282e-01
-6.63935125e-01 2.80716777e-01 7.46561885e-01 -6.69507444e-01
-8.48340154e-01 2.81276405e-01 1.72434911e-01 -2.53154397e-01
5.04377604e-01 8.54675591e-01 -5.17494500e-01 -8.58430713e-02
5.70579469e-01 1.77968726e-01 1.04666971e-01 -9.87087488e-01
5.85765839e-01 7.10716546e-01 3.14222932e-01 -6.63324952e-01
1.20030761e+00 4.09240276e-01 5.93426585e-01 -1.30483854e+00
-7.43987083e-01 -8.23368609e-01 -5.70750773e-01 1.07535295e-01
2.28518322e-01 -9.20536041e-01 -7.89711773e-01 1.28421202e-01
-7.12231815e-01 9.69240889e-02 -5.99898398e-02 1.11156607e+00
-5.12947440e-01 6.78880990e-01 -2.96738625e-01 -8.88437033e-01
-1.09146155e-01 -8.19734573e-01 4.60869789e-01 -1.85236245e-01
6.30761087e-02 -8.24854970e-01 4.82897669e-01 1.96104180e-02
2.72895116e-02 9.71909687e-02 9.78222728e-01 -6.70290828e-01
-6.76369593e-02 -4.46577460e-01 9.29761827e-02 3.87668222e-01
-1.97013970e-02 -3.05302203e-01 -4.25333023e-01 -7.96925843e-01
2.62356728e-01 1.12155899e-01 7.62597680e-01 2.35650390e-01
1.02727568e+00 -3.79048198e-01 -1.71605140e-01 5.29370368e-01
1.28456354e+00 -2.48860255e-01 4.91553813e-01 -4.29566242e-02
8.67218375e-01 6.18400931e-01 1.03414500e+00 8.52880359e-01
1.03178255e-01 9.05450523e-01 1.28205851e-01 3.60084213e-02
5.89578390e-01 4.97985296e-02 5.65218925e-01 1.53973413e+00
-1.89914718e-01 8.20179433e-02 -4.91749346e-01 1.10659957e-01
-2.30263710e+00 -9.13579345e-01 -8.27461481e-01 2.85667682e+00
5.24370968e-01 -3.75595719e-01 1.76131323e-01 4.45507675e-01
7.63598800e-01 1.37606055e-01 -3.23487908e-01 -3.96070302e-01
-1.97295561e-01 1.29493520e-01 7.04890490e-01 7.48683453e-01
-1.07009912e+00 4.80881989e-01 5.86232662e+00 1.13225448e+00
-3.36620748e-01 6.27349690e-02 4.02939357e-02 1.25744447e-01
-1.36183202e-01 1.08519763e-01 -8.53445530e-01 2.05999181e-01
7.84759223e-01 -2.60635853e-01 6.47532046e-01 8.38672042e-01
5.39646327e-01 -5.17728589e-02 -1.09986103e+00 1.45726407e+00
2.73478448e-01 -7.83218205e-01 5.13911434e-02 4.68347192e-01
8.21427345e-01 -4.84077364e-01 3.70143615e-02 2.99219340e-01
-9.34276506e-02 -7.14350402e-01 1.47015736e-01 6.88318729e-01
8.27057183e-01 -1.02837288e+00 6.89519823e-01 5.01435459e-01
-1.24567926e+00 -1.82393551e-01 -1.00126052e+00 -1.45501226e-01
1.13265879e-01 1.10403550e+00 -3.89655620e-01 6.70339406e-01
1.51876658e-01 8.70279849e-01 -4.45397288e-01 1.24698603e+00
2.15072647e-01 5.75649321e-01 -3.58474940e-01 2.73183703e-01
-1.12888299e-01 -9.71709013e-01 6.84150219e-01 9.91125882e-01
6.69313312e-01 3.52290869e-01 3.61958295e-01 1.35523841e-01
6.04365394e-02 7.07086802e-01 -5.64680517e-01 1.17941186e-01
1.63811564e-01 1.47922647e+00 -3.50946426e-01 -5.60683496e-02
-5.71821749e-01 7.36180663e-01 1.08346112e-01 4.06223595e-01
-3.74010801e-01 -3.11524600e-01 8.61688972e-01 -1.86908673e-02
1.29663289e-01 -5.05803525e-01 2.66339511e-01 -1.52660346e+00
-6.58141822e-02 -1.27737606e+00 5.65820158e-01 -1.48741513e-01
-1.25588214e+00 3.38910192e-01 1.37432376e-02 -1.75356913e+00
-2.87924767e-01 -6.68059409e-01 -2.96566278e-01 7.04721212e-01
-1.10742450e+00 -5.80658138e-01 -7.52989501e-02 7.68556356e-01
3.31875622e-01 -4.37480867e-01 7.88254678e-01 6.83605254e-01
-7.82739401e-01 3.98075849e-01 9.71808791e-01 -3.84298116e-01
7.33829975e-01 -1.16373122e+00 1.30403386e-02 1.11801636e+00
2.89864868e-01 8.58727634e-01 9.24787402e-01 -5.83132148e-01
-1.94840312e+00 -1.15683639e+00 5.93201816e-01 5.82940970e-03
7.30658472e-01 -4.85268146e-01 -8.74213874e-01 5.91696203e-01
-3.78160447e-01 -7.41358772e-02 1.03132248e+00 3.82400930e-01
-2.21272364e-01 -4.41960663e-01 -9.53113914e-01 6.95065200e-01
9.11378622e-01 -1.43681794e-01 -4.33420956e-01 6.00074530e-01
4.27728504e-01 -1.11865617e-01 -1.03455484e+00 3.23216647e-01
5.59075534e-01 -6.43195450e-01 1.15916812e+00 -7.20004320e-01
-1.39960617e-01 -6.82229996e-01 -5.15571117e-01 -1.32406127e+00
-7.42531061e-01 -9.66906250e-01 -5.95521271e-01 1.15445745e+00
1.46598190e-01 -5.39138794e-01 4.85443383e-01 3.66945535e-01
2.24071741e-01 -5.92562973e-01 -7.84850121e-01 -9.88148570e-01
-3.67895275e-01 -5.74237049e-01 3.39574158e-01 7.93552160e-01
-5.95864356e-02 5.13840139e-01 -1.21883774e+00 2.30881259e-01
1.12217474e+00 1.77453279e-01 1.00086057e+00 -1.46269679e+00
-7.39558816e-01 3.45834121e-02 -4.33353961e-01 -1.02960956e+00
9.79142860e-02 -9.98052120e-01 -1.43312231e-01 -1.10609293e+00
7.09302947e-02 -2.94504941e-01 -2.72290677e-01 -1.89678371e-01
-4.00015973e-02 1.36531636e-01 1.48896500e-01 2.97152698e-01
-3.64078492e-01 6.09999418e-01 7.54779100e-01 -4.24978957e-02
-4.35288429e-01 6.63691819e-01 -6.14366055e-01 7.71869719e-01
5.98964095e-01 -3.96940827e-01 -6.40374303e-01 -1.32338712e-02
4.50025916e-01 2.16837618e-02 -2.23547786e-01 -8.60959709e-01
8.84438679e-02 -2.63581157e-01 7.75928721e-02 -6.83674753e-01
4.19357240e-01 -1.01887572e+00 4.61451054e-01 3.19463402e-01
-2.06266388e-01 9.19074118e-02 -2.69140810e-01 8.95156384e-01
-8.89019370e-02 -4.86370206e-01 6.69068694e-01 2.21376076e-01
-5.52319407e-01 4.46333319e-01 -4.88076985e-01 7.45705655e-03
7.17615962e-01 1.25471920e-01 4.94231796e-03 -5.41477025e-01
-8.29322934e-01 2.50754118e-01 1.22268125e-01 2.23412126e-01
7.36367881e-01 -1.71196711e+00 -8.57533157e-01 2.22203970e-01
2.42966130e-01 -3.41291338e-01 3.87868792e-01 1.23447156e+00
-2.47291118e-01 4.22883391e-01 1.78001925e-01 -3.73936296e-01
-1.27308786e+00 7.36139655e-01 -4.06534463e-01 -2.72004515e-01
-4.25719053e-01 2.47247323e-01 2.79564500e-01 -3.07159692e-01
3.44405085e-01 -1.82217494e-01 -5.31907201e-01 1.76953867e-01
1.08158863e+00 8.66936564e-01 1.04916617e-02 -7.83938944e-01
-1.60439506e-01 8.11616242e-01 -3.62114534e-02 -9.62671712e-02
1.45064056e+00 -3.60199213e-01 -6.53969049e-01 7.70560682e-01
1.21383762e+00 3.98212522e-01 -6.85404658e-01 -7.81593382e-01
2.54381448e-01 -6.42359376e-01 3.16660814e-02 1.24212496e-01
-7.76065171e-01 5.84266007e-01 5.63751400e-01 -3.17414179e-02
1.07638013e+00 -4.97887909e-01 4.80919808e-01 9.67186809e-01
5.36396503e-01 -1.02998066e+00 -2.78223664e-01 5.35345137e-01
1.10434461e+00 -1.04082537e+00 5.30261397e-01 -4.95063692e-01
-4.55445170e-01 1.09764063e+00 1.87110543e-01 -2.42501244e-01
9.21698749e-01 -2.27345288e-01 -4.73633081e-01 3.45800430e-01
-5.92206359e-01 -2.36448929e-01 5.57569265e-01 3.65546823e-01
3.06728035e-01 2.04069391e-01 -1.15077007e+00 7.36729264e-01
1.95333697e-02 -2.71472514e-01 8.27763498e-01 5.47166228e-01
-4.74092066e-01 -1.46064711e+00 -8.80935311e-01 7.76512027e-01
-2.85882086e-01 -9.54551026e-02 2.73176227e-02 5.15944958e-01
-2.57159352e-01 1.10343134e+00 -3.73469561e-01 -5.96835196e-01
3.17670763e-01 -1.72875732e-01 2.04397291e-01 -6.17422402e-01
-4.83044470e-03 3.02063376e-01 7.33214477e-03 -5.92719257e-01
-3.53500783e-01 -9.62554097e-01 -8.02206933e-01 -3.79521906e-01
-2.68381476e-01 4.65021819e-01 7.44051993e-01 7.82493651e-01
1.95939094e-01 1.85090601e-02 1.05918252e+00 -8.05871487e-01
-8.59879315e-01 -9.66718554e-01 -1.25628734e+00 2.61292070e-01
2.14202985e-01 -7.35908806e-01 -4.24748361e-01 -3.32966655e-01] | [7.1392927169799805, 4.572071075439453] |
8a624536-f40f-40e0-ba91-cf2a867a098a | a-simple-and-effective-model-for-multi-hop | null | null | https://openreview.net/forum?id=IV5YUaQ4pzG | https://openreview.net/pdf?id=IV5YUaQ4pzG | A Simple and Effective Model for Multi-Hop Question Generation | Previous research on automated question generation has almost exclusively focused on generating factoid questions whose answers can be extracted from a single document. However, there is an increasing interest in developing systems that are capable of more complex multi-hop question generation (QG), where answering the question requires reasoning over multiple documents. In this work, we propose a simple and effective approach based on the transformer model for multi-hop QG. Our approach consists of specialized input representations, a supporting sentence classification objective, and training data weighting. Prior work on multi-hop QG considers the simplified setting of shorter documents and also advocates the use of entity-based graph structures as essential ingredients in model design. On the contrary, we showcase that our model can scale to the challenging setting of longer documents as input, does not rely on graph structures, and substantially outperforms the state-of-the-art approaches as measured by automated metrics and human evaluation. | ['Anonymous'] | 2021-09-17 | null | null | null | acl-arr-september-2021-9 | ['sentence-classification'] | ['natural-language-processing'] | [ 3.75479639e-01 8.72913420e-01 2.09274963e-01 -2.36189827e-01
-1.04109251e+00 -6.60390615e-01 9.95710254e-01 6.53204739e-01
-2.04288885e-01 7.89323509e-01 4.46245164e-01 -7.31156468e-01
-3.03248137e-01 -1.07473123e+00 -4.44087535e-01 8.86198953e-02
1.77565277e-01 7.16041386e-01 5.95538795e-01 -7.77335703e-01
3.90999317e-01 4.88576293e-02 -1.38460732e+00 3.74310970e-01
1.23054886e+00 7.48998940e-01 -8.35768282e-02 1.00396562e+00
-5.76582670e-01 1.08086038e+00 -8.47131908e-01 -9.57616210e-01
-2.05484089e-02 -8.06529760e-01 -1.46615601e+00 7.93095902e-02
5.88382900e-01 -2.28420705e-01 -5.41121408e-05 5.64511180e-01
4.57546145e-01 1.17952749e-01 5.98978519e-01 -1.27923775e+00
-8.23562324e-01 8.16678941e-01 1.30620942e-01 3.27083319e-02
8.52284551e-01 -1.03973329e-01 1.53209472e+00 -7.32120931e-01
7.08822548e-01 1.17365289e+00 6.17654920e-01 6.89269364e-01
-1.32438099e+00 1.25253886e-01 1.73740715e-01 3.96046251e-01
-8.39688897e-01 -1.91160157e-01 6.22424603e-01 -1.89776033e-01
1.23346961e+00 4.53670830e-01 4.20399129e-01 8.21991146e-01
1.84483193e-02 7.65961885e-01 8.70297074e-01 -8.30754638e-01
2.58181006e-01 3.17877829e-01 3.80001724e-01 7.32958436e-01
5.04213274e-01 -5.40625989e-01 -2.09979728e-01 -3.23415846e-01
1.32767573e-01 -4.69706893e-01 -3.44696939e-01 -3.15002710e-01
-1.07182086e+00 1.13152289e+00 1.59282446e-01 4.56668317e-01
-3.39186132e-01 3.32417712e-02 1.90721557e-01 5.81349075e-01
4.32550400e-01 1.08080149e+00 -4.74576354e-01 -4.66612750e-04
-1.07486963e+00 7.40738571e-01 1.43264914e+00 1.18342185e+00
5.68898320e-01 -3.30747753e-01 -7.11582601e-01 5.10976851e-01
2.62558848e-01 1.14106938e-01 3.80315483e-01 -9.52988505e-01
7.15394199e-01 9.32256997e-01 3.52064997e-01 -9.37482417e-01
-3.81088823e-01 -3.02223086e-01 -3.30734581e-01 -2.36709103e-01
5.14356971e-01 -4.05366004e-01 -6.60447896e-01 1.64729583e+00
3.57061833e-01 -4.64396447e-01 3.25798571e-01 5.36891937e-01
1.04785943e+00 4.50209111e-01 -1.57305703e-01 8.37820321e-02
1.44657123e+00 -1.06111920e+00 -5.98244369e-01 -2.32896477e-01
8.31846654e-01 -6.34296715e-01 1.20268404e+00 1.47705987e-01
-1.22628677e+00 -3.27378988e-01 -9.22909975e-01 -5.25254488e-01
-6.60366297e-01 7.28736147e-02 5.51266491e-01 7.45051861e-01
-1.43157065e+00 5.09639323e-01 -1.19927734e-01 -7.07556844e-01
5.59726134e-02 1.33379996e-01 -9.94539261e-02 -4.07708108e-01
-1.42792511e+00 1.07946229e+00 2.94730663e-01 -1.55722156e-01
-3.02853554e-01 -5.84288359e-01 -8.59125197e-01 1.92745954e-01
7.59549737e-01 -1.33482146e+00 1.38679719e+00 -4.84034240e-01
-1.38480747e+00 4.20213342e-01 3.84622328e-02 -4.75996882e-01
3.55504781e-01 -5.96957505e-02 -1.35092482e-01 4.44396496e-01
6.02599308e-02 4.92650449e-01 7.63742030e-01 -1.15450966e+00
-3.70202094e-01 -9.83415693e-02 8.87072623e-01 1.91107124e-01
-2.76087552e-01 -4.30007987e-02 -4.53779437e-02 -5.31544447e-01
-2.91784197e-01 -7.46310592e-01 -3.79004657e-01 -3.24587196e-01
-5.32911181e-01 -6.03706181e-01 5.06985486e-01 -7.10819304e-01
1.31276584e+00 -1.48456359e+00 9.01879072e-02 -3.84809449e-02
2.62638509e-01 1.76704928e-01 -3.27730477e-01 1.08368647e+00
2.36045539e-01 4.26710218e-01 -2.62102038e-01 -2.95826852e-01
3.69802952e-01 7.55020976e-02 -4.24168110e-01 -3.65439683e-01
5.12361228e-01 1.14230049e+00 -1.10683787e+00 -7.11547017e-01
-3.73248726e-01 9.34176967e-02 -7.24385262e-01 2.97331572e-01
-5.45338988e-01 -2.09877074e-01 -4.94700491e-01 2.95456350e-01
2.65704423e-01 -5.48885167e-01 2.39624366e-01 4.76804078e-02
1.80318996e-01 6.82608068e-01 -9.80539382e-01 1.59023297e+00
-5.33330619e-01 3.52425158e-01 -3.98321897e-02 -7.38148451e-01
8.57219279e-01 4.02245939e-01 1.00556925e-01 -3.06193084e-01
-5.37698716e-02 2.73793876e-01 -6.98372647e-02 -5.94729066e-01
1.01839674e+00 -9.55967903e-02 -1.03139505e-01 8.05713713e-01
3.54059249e-01 -6.62689388e-01 8.58806849e-01 7.68531799e-01
1.35990262e+00 2.21432343e-01 2.33619481e-01 -1.61320299e-01
6.16091669e-01 1.72781199e-01 -1.41375273e-01 1.06631994e+00
2.98536718e-01 7.33500242e-01 7.70281911e-01 1.26690283e-01
-8.18343282e-01 -7.46733665e-01 3.67166072e-01 8.99769247e-01
-1.79269299e-01 -8.83736789e-01 -9.95054305e-01 -1.18930328e+00
7.07918257e-02 1.16461384e+00 -5.73305070e-01 -1.27905041e-01
-5.37557900e-01 -4.70019281e-01 5.29285550e-01 3.49049747e-01
7.77936205e-02 -1.06185615e+00 -5.82933843e-01 5.18592894e-01
-4.04555947e-01 -1.22017097e+00 -3.92479271e-01 -2.46502727e-01
-9.25499141e-01 -1.16164255e+00 -7.72290766e-01 -5.00371516e-01
6.28822923e-01 2.54862756e-01 1.65814221e+00 3.20033342e-01
1.97838969e-03 8.72147799e-01 -8.10998559e-01 -4.93420452e-01
-5.12876987e-01 5.76702356e-01 -6.07327402e-01 -4.19992469e-02
1.73135087e-01 -5.03602207e-01 -6.27983332e-01 -4.37837355e-02
-1.28133225e+00 1.59477949e-01 6.59655333e-01 6.49479032e-01
8.32481980e-02 -3.69317919e-01 1.15514362e+00 -1.25846493e+00
1.34521675e+00 -6.61832273e-01 -2.17055112e-01 7.60602593e-01
-9.09568787e-01 4.75982249e-01 9.14919674e-01 2.92539131e-02
-9.74498987e-01 -5.48740089e-01 -2.23985374e-01 2.91005880e-01
1.30573928e-01 7.01374948e-01 -7.21245408e-02 1.24980360e-01
7.51443446e-01 8.27005804e-02 -1.40566900e-01 -3.22853774e-01
9.35101748e-01 4.73580301e-01 7.23425448e-02 -6.05065405e-01
9.14575636e-01 1.09977946e-01 1.57453775e-01 -5.80946445e-01
-9.05095637e-01 -3.36690933e-01 -4.08138752e-01 -1.91168994e-01
6.34601772e-01 -3.15006793e-01 -3.75519574e-01 -9.47615057e-02
-1.29488611e+00 -1.53381392e-01 -6.83641076e-01 -5.45850918e-02
-5.48579812e-01 6.11747265e-01 -4.25845623e-01 -7.97607183e-01
-6.87149107e-01 -6.38820648e-01 1.08575904e+00 8.57297927e-02
-4.20028031e-01 -1.23731875e+00 4.14157920e-02 6.47378206e-01
6.51555300e-01 1.43798575e-01 1.38685012e+00 -9.19456065e-01
-7.35724568e-01 -3.73776466e-01 -1.43480077e-01 3.04445088e-01
1.09229192e-01 -2.11246818e-01 -7.05074310e-01 -5.59597611e-02
-9.91417468e-02 -4.82761949e-01 7.48767197e-01 -1.74679935e-01
8.23292196e-01 -6.83462560e-01 1.02589078e-01 -2.38697425e-01
1.26336682e+00 -4.00399148e-01 5.21811724e-01 7.54614249e-02
4.92086053e-01 1.16302347e+00 3.96909684e-01 9.50432047e-02
1.17569506e+00 6.32352412e-01 3.80158842e-01 9.48747471e-02
-3.24667990e-01 -4.34232354e-01 1.80021048e-01 8.94812047e-01
4.36746217e-02 -4.81680036e-01 -7.96667159e-01 8.57815862e-01
-1.86614442e+00 -9.84525084e-01 -3.23441327e-01 1.92002642e+00
8.37444246e-01 -2.69236658e-02 1.61669388e-01 4.17734444e-01
2.84085214e-01 1.83459789e-01 -1.20195024e-01 -5.13575554e-01
1.48321716e-02 4.98147845e-01 3.86777371e-02 5.94379485e-01
-5.86113036e-01 6.82349741e-01 6.46131945e+00 4.81261730e-01
-5.00461936e-01 -6.87101781e-02 2.81538755e-01 1.52529538e-01
-1.03364646e+00 2.93479353e-01 -7.08563209e-01 7.42758512e-02
9.97516692e-01 -5.38637936e-01 2.96951085e-01 5.47745585e-01
-1.22281276e-01 -1.04991727e-01 -1.14565504e+00 3.98327112e-01
4.59881634e-01 -1.30523944e+00 4.42047864e-01 -1.16653055e-01
5.93147695e-01 -5.38305640e-01 -2.89069176e-01 4.86308306e-01
3.44928890e-01 -8.00500035e-01 6.24625981e-01 6.41771257e-01
4.08572346e-01 -4.01367098e-01 8.59519243e-01 5.36034584e-01
-9.62931752e-01 -1.53132007e-01 -1.17698021e-01 -1.35279512e-02
2.40713924e-01 5.92343211e-01 -1.16394973e+00 1.15605748e+00
2.69012719e-01 1.63155124e-01 -1.05579031e+00 7.93865502e-01
-4.38219845e-01 6.96259439e-01 1.10649935e-03 -5.50681055e-01
2.80982554e-01 -1.57918222e-02 2.74394333e-01 1.24446046e+00
4.10630316e-01 9.08536166e-02 8.06554034e-02 6.92998528e-01
-1.45391494e-01 4.87673908e-01 -6.28421962e-01 -2.05094486e-01
1.47686467e-01 1.46447706e+00 -5.39330602e-01 -4.54459339e-01
-5.52289128e-01 7.07049489e-01 5.70354819e-01 2.44747370e-01
-3.65109950e-01 -7.86804676e-01 -5.20390868e-02 3.67041975e-01
4.22783256e-01 -2.48819962e-01 -1.10056251e-01 -1.30201054e+00
3.63492548e-01 -1.06979430e+00 6.24143004e-01 -8.04482341e-01
-1.34411800e+00 8.10283720e-01 4.64788228e-02 -8.63119423e-01
-8.81941020e-01 -4.36950028e-01 -6.01281643e-01 8.43775392e-01
-1.82802463e+00 -1.22101545e+00 -2.08475322e-01 4.48735744e-01
2.68893361e-01 1.90099731e-01 9.82416868e-01 2.90241931e-02
-1.58565849e-01 4.84701961e-01 -4.14177597e-01 -1.41421244e-01
4.25173819e-01 -1.72322845e+00 5.58589637e-01 9.40331638e-01
3.52622837e-01 6.38200283e-01 7.89684355e-01 -4.26959336e-01
-1.42278063e+00 -1.05465484e+00 1.65472555e+00 -7.34916866e-01
7.46308923e-01 -4.50180501e-01 -8.89070392e-01 4.28191394e-01
5.47416210e-01 -4.53273088e-01 7.95937240e-01 1.37007594e-01
-2.53210485e-01 5.47071956e-02 -1.10734379e+00 6.80568457e-01
9.25410271e-01 -5.26680827e-01 -9.66697693e-01 5.23477018e-01
8.19979489e-01 -5.11328503e-02 -8.63477647e-01 2.40669459e-01
5.65295257e-02 -8.20918620e-01 5.65030634e-01 -8.72417688e-01
4.82505113e-01 -3.21239561e-01 6.26651943e-02 -1.30975175e+00
-1.33031324e-01 -7.89642155e-01 -3.93294483e-01 1.25816739e+00
9.06620622e-01 -6.14785254e-01 6.24720156e-01 7.40958512e-01
-2.00989038e-01 -1.09629810e+00 -6.67090595e-01 -5.45852602e-01
7.37959445e-02 -1.99451849e-01 8.24193239e-01 5.66902995e-01
1.76500022e-01 7.88288057e-01 -1.51167065e-01 -1.58783227e-01
3.50936890e-01 4.02666062e-01 7.61621058e-01 -1.32105923e+00
-4.63301301e-01 -3.88844073e-01 2.71038478e-03 -1.05424547e+00
4.03477461e-04 -9.66354847e-01 7.29431771e-03 -2.44491816e+00
-9.50654596e-02 -1.14171676e-01 5.06848060e-02 2.65179366e-01
-6.80865347e-01 -5.34679182e-02 3.78463835e-01 -2.12965339e-01
-7.76327670e-01 6.20159090e-01 1.29856062e+00 -5.10986382e-03
1.93377391e-01 1.18305050e-02 -1.26824760e+00 3.42339128e-01
8.47308517e-01 -3.74026507e-01 -8.73229980e-01 -4.78483498e-01
6.84238911e-01 1.80309027e-01 2.69927889e-01 -8.51653874e-01
3.15706700e-01 -1.83552038e-02 -1.74464181e-01 -2.29715019e-01
1.99005142e-01 -5.25335193e-01 -3.13437194e-01 1.44107893e-01
-6.36644721e-01 3.12785566e-01 -1.29302993e-01 3.98316473e-01
-2.59284794e-01 -6.96016848e-01 1.45587668e-01 -2.92378515e-01
-1.55937746e-01 1.47617742e-01 -2.87199616e-01 5.11650681e-01
6.50002837e-01 -9.05629322e-02 -5.33351660e-01 -7.62223899e-01
-3.81764352e-01 3.72406423e-01 1.78365096e-01 5.06478429e-01
5.97704291e-01 -1.14360726e+00 -9.96719062e-01 -3.17238122e-01
2.48750165e-01 -2.51990497e-01 -4.86456184e-03 4.92073715e-01
-3.11423600e-01 7.32146382e-01 2.44877383e-01 4.68848683e-02
-9.89263415e-01 3.95738304e-01 1.99252844e-01 -9.37106669e-01
-3.50198209e-01 5.51756740e-01 -3.45488280e-01 -6.18777633e-01
-2.68963784e-01 -4.47043478e-01 -5.75237393e-01 3.06297600e-01
4.30240750e-01 3.57560456e-01 3.59016389e-01 -2.33212695e-01
-5.19243814e-02 1.99741125e-01 -1.61843732e-01 -3.35732669e-01
1.10636091e+00 -8.39903764e-03 -1.20192245e-01 2.00663507e-01
6.99557185e-01 1.23814195e-01 -5.97014546e-01 -2.23871306e-01
4.68724519e-01 -1.64734423e-01 -3.72761548e-01 -9.43467677e-01
-4.98534471e-01 7.11499631e-01 -2.54096746e-01 9.83565092e-01
1.02743316e+00 1.74100369e-01 9.15893495e-01 7.14279115e-01
3.80553037e-01 -9.10997272e-01 3.16975325e-01 6.91444576e-01
1.21091616e+00 -8.86531174e-01 -4.98442948e-02 -5.02804339e-01
-5.55262387e-01 1.13390696e+00 4.80356902e-01 9.13662836e-02
2.76092857e-01 -1.91217467e-01 2.25981101e-02 -2.94942796e-01
-1.13493025e+00 -4.56816971e-01 4.06705976e-01 5.94791412e-01
6.21896148e-01 -2.88614810e-01 -5.72993517e-01 7.08286226e-01
-5.79848588e-01 1.42685622e-01 9.41498399e-01 1.13677490e+00
-3.85669142e-01 -1.54950893e+00 -2.62185205e-02 7.77478635e-01
-4.85011905e-01 -3.78409773e-01 -8.21196795e-01 7.49460399e-01
-1.79832399e-01 1.49745274e+00 -4.82408732e-01 -1.34463236e-01
6.52922928e-01 4.44006711e-01 6.93051100e-01 -1.07283711e+00
-1.04347944e+00 -7.95281649e-01 6.18856907e-01 -2.90748239e-01
-3.20416659e-01 -5.12560487e-01 -9.08089519e-01 3.72901969e-02
-4.00584936e-01 6.35496914e-01 5.79360664e-01 9.37204540e-01
7.31733263e-01 4.20694947e-01 4.80627179e-01 -1.92769349e-01
-8.80012214e-01 -1.16100931e+00 -3.45740139e-01 4.78098243e-01
2.75356770e-01 -2.33440012e-01 -2.31120661e-01 -7.06769302e-02] | [11.439347267150879, 8.19583797454834] |
43a9e5b9-b55c-4add-9348-20eb878e0154 | token-transformer-can-class-token-help-window | 2211.06083 | null | https://arxiv.org/abs/2211.06083v2 | https://arxiv.org/pdf/2211.06083v2.pdf | Token Transformer: Can class token help window-based transformer build better long-range interactions? | Compared with the vanilla transformer, the window-based transformer offers a better trade-off between accuracy and efficiency. Although the window-based transformer has made great progress, its long-range modeling capabilities are limited due to the size of the local window and the window connection scheme. To address this problem, we propose a novel Token Transformer (TT). The core mechanism of TT is the addition of a Class (CLS) token for summarizing window information in each local window. We refer to this type of token interaction as CLS Attention. These CLS tokens will interact spatially with the tokens in each window to enable long-range modeling. In order to preserve the hierarchical design of the window-based transformer, we designed Feature Inheritance Module (FIM) in each phase of TT to deliver the local window information from the previous phase to the CLS token in the next phase. In addition, we have designed a Spatial-Channel Feedforward Network (SCFFN) in TT, which can mix CLS tokens and embedded tokens on the spatial domain and channel domain without additional parameters. Extensive experiments have shown that our TT achieves competitive results with low parameters in image classification and downstream tasks. | ['Xuesong Yin', 'Yuanqi Chang', 'Jiawei Mao'] | 2022-11-11 | null | null | null | null | ['long-range-modeling'] | ['natural-language-processing'] | [ 3.62746492e-02 -9.33907107e-02 -3.02904725e-01 -4.07638639e-01
-5.20392835e-01 -6.74038678e-02 4.07047182e-01 -1.79819744e-02
-2.45925769e-01 4.00967509e-01 1.34510741e-01 -4.67764348e-01
2.43809864e-01 -8.55872035e-01 -5.74410796e-01 -7.54407763e-01
-1.53984046e-02 -3.84013236e-01 8.84936094e-01 1.94090661e-02
2.67153203e-01 3.21080714e-01 -1.34450340e+00 7.73582518e-01
8.83068144e-01 1.56545699e+00 9.11068618e-01 3.03318769e-01
-6.08009577e-01 7.34875321e-01 -5.77166021e-01 -1.49901980e-03
1.27191558e-01 -3.10499817e-01 -6.00332618e-01 -3.64066571e-01
-1.10972226e-01 -3.90251160e-01 -4.14211124e-01 6.00107551e-01
5.66152513e-01 -5.72328232e-02 3.10134947e-01 -1.36367977e+00
-4.76176709e-01 9.91612196e-01 -5.40081084e-01 2.89356261e-01
-1.00402012e-01 2.81975102e-02 1.12617254e+00 -1.17752385e+00
3.47147167e-01 1.07453024e+00 5.16843438e-01 3.84310424e-01
-1.00476682e+00 -1.07414591e+00 6.57561302e-01 3.09473515e-01
-1.29935610e+00 -5.11019766e-01 6.90925539e-01 -2.91910410e-01
1.00368226e+00 4.36678566e-02 8.84968996e-01 6.95527911e-01
3.76747519e-01 1.15932715e+00 7.21408367e-01 -4.20545936e-01
1.08678676e-01 2.79471993e-01 2.59713560e-01 3.75313818e-01
-4.48793471e-01 -1.18972883e-01 -9.41854000e-01 1.14784539e-01
1.06571639e+00 2.55777866e-01 -5.92951402e-02 -8.39858130e-02
-1.04379630e+00 7.21155107e-01 8.44780505e-01 3.17683846e-01
-2.26368710e-01 3.90584677e-01 4.15516585e-01 3.13457727e-01
4.09736127e-01 -8.45007002e-02 -4.91281599e-01 3.97998281e-02
-9.82321203e-01 1.05193974e-02 2.91717649e-01 1.13788617e+00
6.37604117e-01 5.22397682e-02 -7.58002162e-01 8.69719803e-01
4.55193788e-01 1.82149515e-01 6.51913762e-01 -3.65518123e-01
6.53045654e-01 6.64817989e-01 -1.98036388e-01 -5.41678071e-01
-3.05117637e-01 -6.84786320e-01 -7.24496484e-01 2.63407733e-02
8.86046365e-02 1.43281877e-01 -1.03170156e+00 1.73408175e+00
1.71945274e-01 2.53976852e-01 -2.46299192e-01 6.06444716e-01
8.74137163e-01 8.09980929e-01 2.57493794e-01 9.92381796e-02
1.51177037e+00 -1.03823495e+00 -5.91301799e-01 -1.29059821e-01
4.85802412e-01 -8.13287020e-01 1.17926586e+00 2.03567613e-02
-1.04675019e+00 -6.93481207e-01 -1.12186778e+00 -1.58183664e-01
-4.30441886e-01 2.60195673e-01 5.22992492e-01 3.50910336e-01
-9.65983272e-01 4.14614737e-01 -8.00227523e-01 1.11487126e-02
5.43551743e-01 4.46950257e-01 -6.48239106e-02 2.49668092e-01
-1.42184794e+00 6.05587363e-01 1.15319133e-01 5.32305479e-01
-7.90152907e-01 -8.77464294e-01 -7.34402120e-01 6.45864189e-01
7.29194432e-02 -3.39661539e-01 1.30703509e+00 -8.15061927e-01
-1.37783110e+00 3.07891488e-01 -5.51925480e-01 -4.97156858e-01
2.59354025e-01 1.55846134e-03 -2.25905672e-01 -5.97461918e-03
9.77808461e-02 8.72202635e-01 9.41575408e-01 -8.91500413e-01
-1.00223827e+00 -1.77521199e-01 6.02389947e-02 5.68407699e-02
-4.97327328e-01 -4.43467870e-02 -5.42228401e-01 -8.95459473e-01
2.12958992e-01 -5.21179438e-01 -1.08128473e-01 5.63261248e-02
-5.04674852e-01 -3.68649125e-01 1.11366975e+00 -2.55954891e-01
1.70210040e+00 -2.55313349e+00 -2.36332878e-01 2.01555461e-01
3.71635288e-01 3.52512896e-01 -9.01530534e-02 5.16818941e-01
-1.16154499e-01 2.92345453e-02 9.13917646e-02 -5.41501522e-01
-2.92681605e-02 9.89707839e-03 -5.77105343e-01 -2.47962531e-02
2.66146362e-01 8.35081577e-01 -4.98059064e-01 -5.85824788e-01
1.09709106e-01 4.48188394e-01 -6.26309335e-01 1.57350153e-01
-2.59796064e-02 5.14788106e-02 -6.14052355e-01 7.13633716e-01
7.20735550e-01 -2.28519529e-01 -1.95263743e-01 -3.82963538e-01
-4.33662444e-01 6.30282462e-01 -1.03952765e+00 1.54550624e+00
-5.85859656e-01 6.35930657e-01 5.63895591e-02 -8.73888433e-01
9.25833285e-01 4.39135462e-01 3.80679518e-01 -1.00953603e+00
-4.11875360e-02 8.62323493e-02 -1.02270059e-01 -3.10091555e-01
6.38507724e-01 1.97089780e-02 8.43243077e-02 1.89597800e-01
-2.03362387e-02 5.40599644e-01 -1.77898005e-01 2.89043784e-01
9.20024812e-01 5.61211854e-02 -1.41553670e-01 -1.32767290e-01
3.92276734e-01 -2.09219843e-01 8.30386221e-01 6.05805755e-01
-2.18336940e-01 4.98986125e-01 5.27094364e-01 -2.58050561e-01
-7.52254426e-01 -1.16608274e+00 -3.33543926e-01 1.23785090e+00
2.05299675e-01 -6.58141971e-01 -6.63637519e-01 -4.63001847e-01
-5.02981851e-03 2.19528139e-01 -5.25063515e-01 -2.78946459e-01
-3.74845117e-01 -4.50724304e-01 5.26422858e-01 1.00902855e+00
8.84492457e-01 -1.00241244e+00 -7.28448689e-01 4.96532828e-01
-1.53124616e-01 -9.12766635e-01 -9.54571307e-01 4.50698286e-01
-7.22855270e-01 -7.84078956e-01 -6.47898018e-01 -9.86256957e-01
4.86240894e-01 4.25191134e-01 4.97801036e-01 -1.26009837e-01
-2.12109312e-01 -2.82706410e-01 -5.65982878e-01 -4.18064475e-01
2.53207624e-01 3.80998909e-01 -3.74135464e-01 2.22004414e-01
1.77733585e-01 -6.20053232e-01 -9.50370133e-01 5.29639721e-01
-8.87976110e-01 1.92295745e-01 7.66516149e-01 1.00521433e+00
4.18924063e-01 -8.56758282e-02 7.13391304e-01 -7.06996620e-01
5.79772055e-01 -2.84931213e-01 -5.40127814e-01 4.61486839e-02
-4.65319216e-01 3.07037365e-02 5.99181533e-01 -6.31332934e-01
-9.09704745e-01 8.30193311e-02 -2.29550555e-01 -3.33234936e-01
3.59745383e-01 5.52930713e-01 -3.11139494e-01 1.58451200e-01
1.00942448e-01 2.96469778e-01 -3.44199538e-02 -6.27792001e-01
1.29840359e-01 8.95449996e-01 1.23993911e-01 -3.05262387e-01
4.30829972e-01 3.64826620e-01 -4.23562557e-01 -6.07331336e-01
-4.36317742e-01 -2.64276713e-01 -4.65797991e-01 -1.72072247e-01
8.04979801e-01 -9.59701896e-01 -9.44298506e-01 7.66569436e-01
-1.13086152e+00 -4.64546263e-01 -2.92258531e-01 3.55851501e-01
-1.51088849e-01 -9.20958817e-02 -6.85306311e-01 -7.44625390e-01
-2.94364035e-01 -1.37696874e+00 9.41803634e-01 3.60187858e-01
9.40839052e-02 -6.91088498e-01 -6.52950466e-01 -7.50065297e-02
8.38977039e-01 -2.66386092e-01 1.05731153e+00 -4.54405040e-01
-6.74510062e-01 -1.01651289e-01 -5.32221556e-01 3.40648413e-01
-2.36670151e-02 -1.85379252e-01 -1.31163204e+00 -4.55118157e-02
-6.02127649e-02 5.49120493e-02 1.19251525e+00 6.61631405e-01
1.33167756e+00 -5.37502356e-02 -8.00791442e-01 5.70896029e-01
1.28452897e+00 5.23313165e-01 5.98605990e-01 2.91684002e-01
5.58580816e-01 2.17045218e-01 4.94810849e-01 4.45322901e-01
5.31755328e-01 7.63694942e-01 3.42303425e-01 -4.13333982e-01
-3.08737874e-01 -4.25801307e-01 6.28695369e-01 5.84859729e-01
2.84773052e-01 -3.13880146e-02 -5.50508559e-01 5.47693968e-01
-1.89910448e+00 -9.25494134e-01 1.17703252e-01 2.15760350e+00
7.08957195e-01 4.91566688e-01 -1.03701279e-01 1.86231524e-01
7.04022050e-01 1.95856452e-01 -5.08020818e-01 -4.14146453e-01
1.28083631e-01 4.40120213e-02 5.29155970e-01 3.70287091e-01
-7.78630257e-01 9.79442894e-01 6.33162498e+00 1.20206594e+00
-1.39929247e+00 1.08614005e-01 4.80545014e-01 -3.50312963e-02
-3.29520285e-01 8.22164938e-02 -1.17221975e+00 7.23237216e-01
5.76843083e-01 9.55128893e-02 5.00739776e-02 6.68040216e-01
4.27630961e-01 -3.59628499e-01 -1.10613406e+00 7.72217333e-01
-4.66143519e-01 -1.45797062e+00 9.41123962e-02 1.23011589e-01
-1.08543504e-02 3.51781212e-02 2.19879791e-01 2.96900898e-01
-1.89222991e-02 -8.76788735e-01 1.05768299e+00 2.58198917e-01
8.54126096e-01 -7.45747328e-01 4.33209002e-01 2.44457975e-01
-1.96342814e+00 -4.28326070e-01 -3.96290600e-01 4.33249725e-03
4.93932776e-02 7.41824746e-01 -6.55827880e-01 3.97584409e-01
8.57951760e-01 6.90638542e-01 -2.69786477e-01 1.11360693e+00
-4.30599898e-02 5.12615144e-01 -2.91986018e-01 9.16327983e-02
3.08271796e-01 7.55516440e-02 1.88094512e-01 1.23073339e+00
1.91528618e-01 -1.32957801e-01 1.83125317e-01 9.21887577e-01
-6.00368865e-02 -1.45514637e-01 -3.38011354e-01 3.00908148e-01
1.03631365e+00 1.09087241e+00 -6.79269850e-01 -4.18306887e-01
-5.38860500e-01 7.30951965e-01 2.02172026e-01 4.36739415e-01
-9.08445239e-01 -9.06624019e-01 7.30166137e-01 3.40922624e-01
6.51762724e-01 -1.52804241e-01 -3.69563133e-01 -8.38205934e-01
3.46221060e-01 -2.30888560e-01 2.26897776e-01 -6.15661383e-01
-9.97916460e-01 7.81003058e-01 -8.02831948e-02 -1.34508777e+00
3.41239810e-01 -3.01984221e-01 -9.23198640e-01 1.14318180e+00
-1.87658679e+00 -1.37462354e+00 -9.22431797e-02 8.09608638e-01
6.62714362e-01 1.60822123e-02 6.77412510e-01 5.46286047e-01
-6.80793941e-01 9.53522503e-01 -4.20182869e-02 3.46220970e-01
7.52677977e-01 -9.83226120e-01 2.90027320e-01 7.58800328e-01
-3.50076616e-01 7.36503839e-01 2.20604971e-01 -2.22013682e-01
-1.16224647e+00 -1.06587434e+00 9.92334664e-01 2.50003844e-01
5.20794570e-01 -7.31528282e-01 -7.88126707e-01 7.25976884e-01
1.02617137e-01 1.36246175e-01 6.18333399e-01 1.15970500e-01
-5.44784904e-01 -6.42999232e-01 -8.90337110e-01 4.40348059e-01
8.01277995e-01 -5.22130847e-01 -3.39618146e-01 -9.45691243e-02
9.34255302e-01 -4.38705713e-01 -7.03070879e-01 9.90702212e-02
8.28285694e-01 -8.34823847e-01 8.22543144e-01 -5.89442775e-02
3.38906050e-01 -3.94472480e-01 -1.26214996e-01 -1.26757324e+00
-3.27731848e-01 -4.96438771e-01 1.88800752e-01 1.61077297e+00
7.02082217e-01 -8.17489326e-01 6.10483646e-01 4.21831280e-01
-4.53954309e-01 -1.11107075e+00 -1.08722699e+00 -7.63188601e-01
-1.57582670e-01 -5.93338013e-01 8.17126215e-01 4.10469979e-01
3.65442872e-01 3.34726840e-01 -1.67344257e-01 6.31358176e-02
2.48747438e-01 1.24984406e-01 3.17170918e-01 -1.06326377e+00
-1.38045356e-01 -6.54235005e-01 -1.28215879e-01 -1.59485388e+00
-1.97538048e-01 -9.25531328e-01 1.45705253e-01 -1.56489801e+00
1.95292756e-01 -1.08006477e+00 -4.85935271e-01 7.09680378e-01
-1.12735465e-01 -1.83351617e-02 3.99833947e-01 2.51146704e-01
-2.36732289e-01 5.71495235e-01 1.35134935e+00 -7.85571188e-02
-4.90122825e-01 1.54207885e-01 -5.94696403e-01 3.58586878e-01
6.85632527e-01 -4.82814431e-01 -5.68862915e-01 -5.94427466e-01
-1.27416492e-01 6.36401847e-02 2.36740083e-01 -8.45518649e-01
6.36355877e-01 1.22575566e-01 6.25045598e-01 -8.22428465e-01
4.29213703e-01 -7.98123896e-01 -3.37968707e-01 4.00601834e-01
-4.66513693e-01 1.26587197e-01 1.79834381e-01 3.77174288e-01
-4.84875143e-01 2.22575948e-01 7.68692374e-01 3.85953896e-02
-5.80460787e-01 3.88668656e-01 -6.24974847e-01 -3.12256962e-01
1.00929284e+00 -4.86363441e-01 -9.92479920e-02 1.91135369e-02
-7.95260370e-01 5.74458599e-01 -4.54571173e-02 4.67088401e-01
8.58998537e-01 -1.33890176e+00 -2.26069406e-01 6.36965990e-01
3.00783794e-02 2.89988995e-01 2.32856780e-01 9.10553873e-01
-8.90786126e-02 5.14164090e-01 -1.99204192e-01 -6.23273075e-01
-1.12077653e+00 2.92779982e-01 3.54288727e-01 -3.47440153e-01
-8.87815177e-01 1.00939119e+00 5.02650499e-01 1.06585383e-01
6.22586548e-01 -7.50927329e-01 -2.20853627e-01 2.40555808e-01
6.58446372e-01 1.04054265e-01 2.17359960e-02 -2.46252567e-01
-3.99308532e-01 4.58403528e-01 -3.14611167e-01 -2.03644186e-01
1.22654688e+00 -2.49543682e-01 -1.65325981e-02 5.26748955e-01
1.07580388e+00 -1.93223029e-01 -1.31596756e+00 -3.64257544e-01
-1.99081555e-01 -3.13443154e-01 2.20206767e-01 -5.76063633e-01
-1.35460472e+00 1.07233918e+00 4.92138088e-01 2.36021802e-01
1.35726249e+00 -8.79958197e-02 1.15393162e+00 -1.71938390e-01
1.29764780e-01 -1.21752346e+00 -6.55510202e-02 5.63505888e-01
4.70551312e-01 -7.34966755e-01 -5.43706179e-01 -5.89196205e-01
-5.90497851e-01 1.01676464e+00 7.13227928e-01 2.22599488e-02
1.08991861e+00 5.62892735e-01 -2.65866593e-02 -3.11912578e-02
-1.03790772e+00 -8.89431462e-02 -6.40292913e-02 5.09302378e-01
5.55073857e-01 -1.18885763e-01 -1.26433715e-01 7.79644489e-01
5.00106737e-02 1.94173604e-01 1.87766030e-01 9.77214634e-01
-7.15260625e-01 -1.33120215e+00 -1.42257497e-01 4.17153865e-01
-3.86854559e-01 -3.53107721e-01 1.94515869e-01 6.02973402e-01
2.53386915e-01 9.70883131e-01 3.55753154e-01 -5.78042030e-01
3.87822539e-01 1.08154550e-01 7.01451749e-02 -5.12462020e-01
-8.30064178e-01 3.49677712e-01 -2.68622428e-01 -6.72386408e-01
-1.50267556e-01 -3.58654916e-01 -1.54735768e+00 -1.69396147e-01
-5.01834571e-01 3.02679300e-01 6.53403401e-01 9.45459425e-01
5.46437263e-01 9.67272222e-01 8.17056119e-01 -7.17542827e-01
-3.04820776e-01 -9.59467709e-01 -8.27686787e-01 -2.00244069e-01
5.54494262e-01 -6.03291929e-01 6.44422844e-02 -5.48591465e-02] | [9.707976341247559, 1.0887463092803955] |
9c0071bd-d26d-4f86-9807-44b6871b1fea | riga-rotation-invariant-and-globally-aware | 2209.13252 | null | https://arxiv.org/abs/2209.13252v1 | https://arxiv.org/pdf/2209.13252v1.pdf | RIGA: Rotation-Invariant and Globally-Aware Descriptors for Point Cloud Registration | Successful point cloud registration relies on accurate correspondences established upon powerful descriptors. However, existing neural descriptors either leverage a rotation-variant backbone whose performance declines under large rotations, or encode local geometry that is less distinctive. To address this issue, we introduce RIGA to learn descriptors that are Rotation-Invariant by design and Globally-Aware. From the Point Pair Features (PPFs) of sparse local regions, rotation-invariant local geometry is encoded into geometric descriptors. Global awareness of 3D structures and geometric context is subsequently incorporated, both in a rotation-invariant fashion. More specifically, 3D structures of the whole frame are first represented by our global PPF signatures, from which structural descriptors are learned to help geometric descriptors sense the 3D world beyond local regions. Geometric context from the whole scene is then globally aggregated into descriptors. Finally, the description of sparse regions is interpolated to dense point descriptors, from which correspondences are extracted for registration. To validate our approach, we conduct extensive experiments on both object- and scene-level data. With large rotations, RIGA surpasses the state-of-the-art methods by a margin of 8\degree in terms of the Relative Rotation Error on ModelNet40 and improves the Feature Matching Recall by at least 5 percentage points on 3DLoMatch. | ['Slobodan Ilic', 'Benjamin Busam', 'Kai Wang', 'Ivan Shugurov', 'Mahdi Saleh', 'Zheng Qin', 'Ji Hou', 'Hao Yu'] | 2022-09-27 | null | null | null | null | ['point-cloud-registration'] | ['computer-vision'] | [-1.92434669e-01 -3.10991734e-01 -3.85778457e-01 -4.60692644e-01
-9.89070237e-01 -7.36468732e-01 7.90878117e-01 4.47463691e-01
-1.08313657e-01 1.83209717e-01 1.13703281e-01 4.01141286e-01
-2.17929602e-01 -8.81090164e-01 -8.30549121e-01 -5.91911554e-01
-2.47289483e-02 4.39435214e-01 3.17846745e-01 -3.56453568e-01
4.13533568e-01 1.26804149e+00 -1.45340228e+00 -6.29427582e-02
5.33908784e-01 1.19644046e+00 -3.87160629e-02 -9.85168368e-02
-4.06364873e-02 3.65542322e-01 -1.99126273e-01 1.35869220e-01
5.37491441e-01 1.05219796e-01 -5.90695262e-01 -5.93369231e-02
1.09709418e+00 -2.89018989e-01 -5.59831023e-01 8.06658983e-01
2.90798098e-01 2.17281595e-01 6.22232974e-01 -7.66187787e-01
-5.54736614e-01 -4.23291652e-03 -6.07660174e-01 -1.14281423e-01
6.43763244e-01 1.73499689e-01 1.16621232e+00 -1.36304331e+00
9.30052757e-01 1.24572790e+00 7.56867170e-01 8.34290460e-02
-1.21002579e+00 -7.39268482e-01 2.80201703e-01 -1.73110604e-01
-1.86013341e+00 -3.60871136e-01 1.18962717e+00 -3.75592887e-01
1.25590515e+00 9.52308774e-02 7.34281242e-01 5.30427754e-01
1.45158380e-01 2.81150997e-01 6.84931934e-01 -7.68791139e-02
1.10696107e-01 -4.79114175e-01 3.00538857e-02 7.65369594e-01
2.98617482e-01 3.10350269e-01 -7.58937836e-01 -1.12177491e-01
1.24373341e+00 4.34065878e-01 -7.62130916e-02 -8.37096930e-01
-1.41585553e+00 7.65186727e-01 1.18376112e+00 2.36566126e-01
-5.68613350e-01 3.23977768e-01 2.18530353e-02 1.35228634e-01
4.83938605e-01 6.05759740e-01 -3.27478081e-01 1.17602071e-03
-8.76624882e-01 4.12036449e-01 5.29285192e-01 1.22840488e+00
1.37605107e+00 -7.23011270e-02 7.47557655e-02 7.17864513e-01
3.82515669e-01 8.95665944e-01 3.76279056e-02 -1.02040362e+00
3.53810310e-01 1.01170468e+00 -2.18093604e-01 -1.48025954e+00
-3.80203962e-01 -4.19746727e-01 -7.95587599e-01 1.37876227e-01
1.61700502e-01 5.29668689e-01 -9.56804514e-01 1.50189710e+00
3.70236248e-01 2.82577693e-01 -4.08338219e-01 9.95960534e-01
8.54818821e-01 2.79977620e-01 -1.64658114e-01 2.76039094e-01
1.40767932e+00 -4.77330148e-01 -1.85856242e-02 -1.33890733e-01
6.60027742e-01 -7.99023330e-01 5.15466928e-01 -1.27268210e-01
-9.44192886e-01 -6.80211246e-01 -9.61028874e-01 -3.09197247e-01
-2.32272893e-01 -2.34519601e-01 5.56899667e-01 2.26158202e-01
-1.07351744e+00 5.55425942e-01 -9.07879889e-01 -4.09005970e-01
6.47136867e-01 5.10489285e-01 -7.26840198e-01 -4.11346793e-01
-7.26345181e-01 8.04766595e-01 8.21265504e-02 -2.57594079e-01
-6.60106719e-01 -1.06223166e+00 -1.22317445e+00 -4.68167514e-02
-5.46324141e-02 -8.86116147e-01 6.68151259e-01 -4.47454065e-01
-1.24381971e+00 1.01819730e+00 -3.58533651e-01 -1.21970579e-01
2.30898440e-01 -1.62475780e-01 -6.84887990e-02 4.08730596e-01
2.37903237e-01 8.51491034e-01 7.03875065e-01 -1.37338698e+00
-4.24743354e-01 -6.11182332e-01 1.52826250e-01 2.57556498e-01
2.35744625e-01 -1.52653411e-01 -5.47539234e-01 -5.79940259e-01
8.34816813e-01 -9.96065021e-01 -1.45462245e-01 4.01783377e-01
-4.21878248e-02 -2.27692381e-01 8.88019979e-01 -2.25563705e-01
7.03199625e-01 -2.43892002e+00 -1.15358531e-02 6.09398484e-01
4.20033216e-01 -1.78200498e-01 -4.04917330e-01 3.49051923e-01
-1.77737042e-01 -1.50877297e-01 -7.00031519e-02 -3.05881888e-01
-1.03701435e-01 1.45659342e-01 -3.79015237e-01 8.35388422e-01
5.41639268e-01 1.00189114e+00 -9.90486026e-01 -3.34257811e-01
5.52922070e-01 6.91019118e-01 -7.57311642e-01 -1.68722495e-02
1.21471263e-01 5.39353371e-01 -8.96836400e-01 9.36769128e-01
8.20028782e-01 -2.20888659e-01 -3.39983225e-01 -4.98489976e-01
-1.97081357e-01 3.27575833e-01 -1.02518022e+00 2.24997926e+00
-4.46904004e-01 4.82722789e-01 -2.92869866e-01 -8.47846150e-01
1.38652205e+00 4.27071517e-03 8.85581613e-01 -7.44069755e-01
-2.57776957e-02 3.26125264e-01 -4.76081342e-01 2.16135919e-01
5.26061535e-01 1.04572445e-01 -8.59554261e-02 1.05675220e-01
1.10936798e-02 -5.36269665e-01 -4.13517833e-01 9.50673744e-02
1.17220175e+00 2.24550858e-01 4.22926962e-01 -2.77648866e-01
4.02008384e-01 9.47249830e-02 4.32578504e-01 6.58750534e-01
9.99031812e-02 1.05151033e+00 -1.28413975e-01 -7.80687213e-01
-9.99497652e-01 -1.31930375e+00 -4.44029927e-01 6.69082880e-01
6.45563900e-01 -6.23341203e-01 -3.09409797e-01 -4.51702654e-01
3.87209803e-01 1.00598507e-01 -5.18472672e-01 -1.97349846e-01
-8.76504242e-01 -5.20729646e-02 1.95016921e-01 6.48460865e-01
4.74476814e-01 -7.56484389e-01 -5.30320525e-01 2.00914502e-01
2.10364297e-01 -1.30851746e+00 -5.23283064e-01 3.24756652e-02
-1.05317342e+00 -9.88769293e-01 -5.49115717e-01 -6.70315862e-01
8.88819933e-01 7.12452471e-01 1.20499444e+00 2.31649429e-01
-1.91231921e-01 3.13991636e-01 -3.78604174e-01 -1.17668333e-02
1.92359790e-01 1.34506166e-01 1.30074531e-01 -1.51076868e-01
5.16432226e-01 -8.32822204e-01 -7.57058740e-01 4.76549119e-01
-4.58990484e-01 -2.31221139e-01 7.08293974e-01 6.30974650e-01
1.00370407e+00 -5.04443228e-01 1.18205249e-01 -3.39972466e-01
-9.24571678e-02 -3.51861596e-01 -6.57404184e-01 -7.52085373e-02
-1.75909400e-01 1.55141190e-01 3.17175955e-01 -3.04934919e-01
-4.40047175e-01 4.89995360e-01 1.37573242e-01 -7.79632568e-01
-2.94429094e-01 3.99892777e-01 -1.04113318e-01 -5.99391520e-01
6.72376633e-01 2.25647599e-01 -1.43544197e-01 -5.13059676e-01
4.89964992e-01 2.15765238e-01 5.40569067e-01 -8.96617353e-01
1.30100513e+00 7.51297951e-01 2.99012065e-01 -7.76087344e-01
-6.77358091e-01 -7.06460118e-01 -1.10692048e+00 4.83097285e-02
8.03011179e-01 -1.32412386e+00 -6.01909399e-01 3.44735265e-01
-1.26116753e+00 1.10143371e-01 -4.92169112e-01 6.25649571e-01
-6.52697384e-01 2.88759410e-01 -2.95386970e-01 -1.62394047e-01
-1.41742721e-01 -1.21589792e+00 1.68272686e+00 1.20272443e-01
-2.94965714e-01 -8.29212010e-01 1.17515653e-01 -4.17661071e-02
3.82500231e-01 5.32204628e-01 7.74184346e-01 -4.40512806e-01
-1.18859756e+00 -4.19063061e-01 -5.47261536e-01 -3.69017683e-02
2.80310899e-01 -1.18203208e-01 -9.27791655e-01 -3.42710584e-01
-1.97087124e-01 2.54287181e-04 7.08970428e-01 2.38319486e-01
1.00308156e+00 1.61877155e-01 -4.80538249e-01 1.09549963e+00
1.44827783e+00 -1.85693711e-01 3.90912235e-01 2.93253392e-01
1.05606294e+00 1.84325039e-01 5.42991400e-01 4.98735696e-01
4.83742386e-01 1.02517998e+00 5.55956781e-01 -1.99536994e-01
-3.12298000e-01 -4.66329485e-01 1.78554088e-01 9.50069487e-01
-3.52636933e-01 4.52212453e-01 -1.00033033e+00 4.38191414e-01
-1.58666074e+00 -8.26237142e-01 1.20727524e-01 2.24537015e+00
6.95950806e-01 -7.47102648e-02 -1.55881181e-01 -2.00968564e-01
5.02843082e-01 4.68252897e-01 -4.45758820e-01 6.49700612e-02
-2.86669523e-01 4.50272381e-01 6.51287138e-01 3.27229202e-01
-1.07622004e+00 1.18353665e+00 5.78109646e+00 6.78537428e-01
-1.24288726e+00 -1.71315879e-01 1.87465638e-01 -9.32286493e-03
-3.32782328e-01 2.99305439e-01 -7.96850681e-01 1.92770036e-03
4.65914428e-01 -8.83976892e-02 3.23578596e-01 9.89348292e-01
-8.39211717e-02 2.34861970e-01 -1.39355326e+00 1.33975899e+00
1.33777440e-01 -1.48684978e+00 2.76543528e-01 3.90293926e-01
8.53044271e-01 4.88696009e-01 -2.74058692e-02 1.46108329e-01
1.88935414e-01 -1.04152119e+00 7.28189290e-01 5.90608656e-01
1.02371883e+00 -7.71195889e-01 5.15788734e-01 5.45553048e-04
-1.65215099e+00 2.49845117e-01 -6.11065924e-01 4.22432832e-02
-5.51935099e-03 4.46050078e-01 -5.97778916e-01 8.77584815e-01
5.88118255e-01 1.31989503e+00 -5.44976830e-01 1.00838363e+00
-7.03238845e-02 6.48691729e-02 -6.62210166e-01 5.14880896e-01
4.26977187e-01 -1.31691709e-01 4.80012089e-01 9.80113864e-01
4.55379605e-01 2.55599260e-01 4.22353566e-01 9.76455867e-01
-2.01375842e-01 6.10527769e-02 -8.32690895e-01 4.24298674e-01
8.53394389e-01 1.21502745e+00 -6.12430274e-01 -1.33067131e-01
-4.48906392e-01 6.65656209e-01 3.22194040e-01 3.38846475e-01
-5.35592556e-01 -2.43437558e-01 1.09314883e+00 2.99266398e-01
4.18679357e-01 -7.60555625e-01 -2.60579586e-01 -1.23494339e+00
9.77151990e-02 -5.47084212e-01 -1.41972397e-02 -6.60070360e-01
-1.29840660e+00 5.53222656e-01 1.63875823e-03 -1.63385367e+00
-1.89981431e-01 -4.24088001e-01 -4.72737312e-01 9.73053992e-01
-1.70150924e+00 -1.36262727e+00 -8.76594424e-01 7.11706996e-01
2.72005647e-01 -1.78578198e-01 9.14910257e-01 1.65531799e-01
-2.50883754e-02 5.64156473e-01 -1.15888603e-01 3.60598087e-01
6.39459908e-01 -9.07606006e-01 7.40616262e-01 4.33234930e-01
4.66586083e-01 1.02315867e+00 1.67941034e-01 -5.86346447e-01
-1.68391383e+00 -1.10304403e+00 7.54112124e-01 -8.17839742e-01
5.40552557e-01 -3.42010736e-01 -9.98145223e-01 3.70770454e-01
-5.30245245e-01 7.85141110e-01 2.64959186e-01 1.18048429e-01
-9.93736744e-01 -2.21389040e-01 -1.02200508e+00 2.86806315e-01
1.41729677e+00 -1.11533940e+00 -5.79618394e-01 1.78629681e-01
6.62523270e-01 -6.58430099e-01 -1.30846024e+00 4.40163493e-01
5.60002685e-01 -7.25422740e-01 1.39442801e+00 -2.98909098e-01
1.62373066e-01 -5.04850388e-01 -5.34843564e-01 -1.10664880e+00
-5.48738897e-01 -4.44601685e-01 1.55872613e-01 8.96675706e-01
-8.78174528e-02 -4.48283553e-01 7.14449406e-01 2.77844071e-01
-2.92496890e-01 -5.85269272e-01 -1.19451976e+00 -9.64040577e-01
1.09577402e-01 -5.06481051e-01 9.74632144e-01 1.22540176e+00
-2.69982308e-01 6.45198822e-02 1.92785427e-01 3.06594670e-01
5.76507449e-01 3.66689146e-01 9.05265629e-01 -1.38865495e+00
3.33050489e-01 -5.68841696e-01 -1.18439376e+00 -1.47158933e+00
3.19894284e-01 -1.09465075e+00 -8.55143890e-02 -1.15862656e+00
3.49285938e-02 -7.36902118e-01 -4.23536211e-01 5.65429449e-01
1.34050891e-01 5.31186938e-01 3.39074135e-01 5.90264022e-01
-5.95485747e-01 6.92385316e-01 1.09709585e+00 -1.30880654e-01
-1.52467117e-01 -4.91215199e-01 -3.39083344e-01 6.28035784e-01
3.62176687e-01 -4.99310076e-01 -2.96837091e-03 -5.51441073e-01
-9.86483470e-02 -1.75819471e-01 6.14328921e-01 -1.21810532e+00
2.71401554e-01 -1.49114817e-01 4.85586643e-01 -7.60409057e-01
6.07104242e-01 -1.08374000e+00 1.76711738e-01 1.58829004e-01
-2.55612358e-02 1.76547393e-01 -3.74098867e-02 4.95618135e-01
-4.11852986e-01 4.01951909e-01 7.89819002e-01 1.09392926e-01
-6.48932993e-01 8.42361033e-01 4.35116649e-01 -5.82796670e-02
8.54437232e-01 -6.64219677e-01 -1.72578961e-01 -2.96087354e-01
-2.74389118e-01 -1.68520004e-01 1.06916690e+00 5.21699548e-01
8.36947381e-01 -1.62714469e+00 -6.09995782e-01 5.95655382e-01
6.10259295e-01 5.23824930e-01 8.92368853e-02 7.88633227e-01
-6.85890675e-01 5.89044094e-01 -2.76243180e-01 -1.13813257e+00
-9.77881849e-01 3.82224262e-01 3.44184220e-01 2.68400103e-01
-9.38633859e-01 5.68466961e-01 5.30424595e-01 -3.60727221e-01
-4.89538722e-02 -6.47512436e-01 8.41238201e-02 -7.86424950e-02
2.88109362e-01 -1.97057668e-02 1.24925613e-01 -1.28795791e+00
-7.45709777e-01 1.46790910e+00 -4.30642292e-02 2.64291078e-01
1.48272991e+00 9.95217413e-02 -6.08100370e-02 2.23263875e-02
1.57340789e+00 2.41828158e-01 -1.42182899e+00 -7.97218204e-01
6.92020729e-03 -8.32110882e-01 1.95371345e-01 -3.94890010e-01
-1.28191531e+00 7.11770415e-01 5.54210186e-01 -2.97473252e-01
8.54574025e-01 4.13245171e-01 6.85366333e-01 4.39410061e-01
7.81581521e-01 -4.88377452e-01 -4.18196805e-03 7.50673473e-01
1.02616596e+00 -1.23697078e+00 2.57765830e-01 -4.70736027e-01
-2.65642136e-01 1.19490194e+00 2.79295057e-01 -6.27194226e-01
7.90722907e-01 -2.08800212e-02 -4.95392829e-02 -6.82330430e-01
-2.87564635e-01 -1.61808148e-01 7.42036343e-01 6.63467169e-01
2.77384639e-01 -3.62202860e-02 4.37049896e-01 1.59010276e-01
-3.87643397e-01 -2.60384202e-01 -1.64011851e-01 8.28997016e-01
-4.12359715e-01 -1.00275970e+00 -3.42603832e-01 8.61536041e-02
7.16882199e-02 -1.38955703e-02 -4.36924636e-01 8.91416728e-01
-3.35752331e-02 4.85437840e-01 4.70387250e-01 -4.50333863e-01
5.88666618e-01 -3.26399654e-01 5.12816191e-01 -7.03046739e-01
-4.50524777e-01 2.61548907e-02 -2.90871024e-01 -1.16752136e+00
-5.39839149e-01 -7.19921291e-01 -1.38426459e+00 -3.61561030e-01
-1.31078273e-01 -1.55421823e-01 7.09596515e-01 8.93196642e-01
6.42220557e-01 7.15198442e-02 8.06488514e-01 -1.29742849e+00
-3.04904014e-01 -5.75929642e-01 -4.39592063e-01 6.25550747e-01
4.14645106e-01 -8.61910462e-01 -2.94395983e-01 -1.46297455e-01] | [7.774737358093262, -2.9232728481292725] |
ae7c912f-cf4a-4a33-b793-141a38eba133 | towards-minimizing-efforts-for-morphing | 2305.18216 | null | https://arxiv.org/abs/2305.18216v1 | https://arxiv.org/pdf/2305.18216v1.pdf | Towards minimizing efforts for Morphing Attacks -- Deep embeddings for morphing pair selection and improved Morphing Attack Detection | Face Morphing Attacks pose a threat to the security of identity documents, especially with respect to a subsequent access control process, because it enables both individuals involved to exploit the same document. In this study, face embeddings serve two purposes: pre-selecting images for large-scale Morphing Attack generation and detecting potential Morphing Attacks. We build upon previous embedding studies in both use cases using the MagFace model. For the first objective, we employ an pre-selection algorithm that pairs individuals based on face embedding similarity. We quantify the attack potential of differently morphed face images to compare the usability of pre-selection in automatically generating numerous successful Morphing Attacks. Regarding the second objective, we compare embeddings from two state-of-the-art face recognition systems in terms of their ability to detect Morphing Attacks. Our findings demonstrate that ArcFace and MagFace provide valuable face embeddings for image pre-selection. Both open-source and COTS face recognition systems are susceptible to generated attacks, particularly when pre-selection is based on embeddings rather than random pairing which was only constrained by soft biometrics. More accurate face recognition systems exhibit greater vulnerability to attacks, with COTS systems being the most susceptible. Additionally, MagFace embeddings serve as a robust alternative for detecting morphed face images compared to the previously used ArcFace embeddings. The results endorse the advantages of face embeddings in more effective image pre-selection for face morphing and accurate detection of morphed face images. This is supported by extensive analysis of various designed attacks. The MagFace model proves to be a powerful alternative to the commonly used ArcFace model for both objectives, pre-selection and attack detection. | ['Christoph Busch', 'Juan Tapia', 'Kiran Raja', 'Roman Kessler'] | 2023-05-29 | null | null | null | null | ['face-recognition'] | ['computer-vision'] | [ 1.23348914e-01 -1.98476270e-01 2.28053227e-01 -2.90629506e-01
-5.09383857e-01 -9.61640060e-01 7.89697230e-01 9.44703817e-02
-3.56829375e-01 1.25078544e-01 -1.69521198e-01 -4.22620624e-01
-1.24057829e-01 -9.40994203e-01 -4.12239194e-01 -4.05573517e-01
-3.14777583e-01 2.27108523e-01 -2.90931165e-01 -3.74728918e-01
4.10133749e-01 1.17361033e+00 -1.87763238e+00 2.92555809e-01
2.77974755e-01 9.38047647e-01 -5.58997929e-01 5.92857242e-01
1.40235186e-01 -1.97929606e-01 -8.66398513e-01 -9.19930041e-01
7.31855631e-01 -1.41179353e-01 -5.06609917e-01 -1.77131608e-01
1.06761169e+00 -5.15264034e-01 -2.76599824e-01 1.12542069e+00
7.87582278e-01 -2.79454112e-01 6.00841641e-01 -1.48747253e+00
-7.60947287e-01 1.86299190e-01 -5.36556780e-01 2.68990278e-01
8.20447206e-01 4.83789563e-01 5.52587807e-01 -1.04775512e+00
4.36978132e-01 1.55126655e+00 6.51032627e-01 8.09287310e-01
-1.39500117e+00 -1.27330637e+00 -5.30503452e-01 7.27725551e-02
-1.74151814e+00 -8.85251641e-01 6.06895804e-01 -4.97984320e-01
7.87410975e-01 4.75342661e-01 3.56625855e-01 1.08704948e+00
1.08293518e-01 1.19220084e-02 1.32726812e+00 -5.61337590e-01
8.48912522e-02 5.95352948e-01 -8.77160206e-02 8.45191479e-01
5.05155265e-01 3.83025974e-01 -6.09528542e-01 -7.39367902e-01
5.86832106e-01 -2.90690929e-01 3.50934966e-03 1.59473389e-01
-5.32388747e-01 8.26174080e-01 -1.33213336e-02 3.32055598e-01
-2.56002903e-01 -1.87881842e-01 3.40194464e-01 5.60765624e-01
7.71256089e-02 6.83572054e-01 4.98526134e-02 2.20230557e-02
-1.14294612e+00 7.91280568e-02 6.88850582e-01 5.70845187e-01
8.06637943e-01 1.76404357e-01 -1.06672272e-01 6.94729567e-01
5.63773036e-01 5.67016304e-01 3.56651366e-01 -3.74468088e-01
3.21192473e-01 6.17821574e-01 -1.96183696e-01 -1.39578497e+00
1.86541438e-01 2.86660939e-01 -3.30293894e-01 7.56517231e-01
3.21696073e-01 -3.01157013e-02 -9.86119330e-01 1.58296120e+00
3.70406866e-01 9.11769420e-02 -3.72821428e-02 3.36271852e-01
3.89067233e-01 3.06161821e-01 2.33366042e-01 1.25164881e-01
1.78507483e+00 -9.28055495e-02 -4.35445160e-01 1.04675062e-01
2.95315325e-01 -1.20205200e+00 1.04499376e+00 2.49757603e-01
-8.41455758e-01 -5.84692955e-01 -1.46344101e+00 5.13579249e-01
-9.20339584e-01 6.85398132e-02 4.11441654e-01 1.99462521e+00
-1.33331597e+00 5.29820681e-01 -3.10506642e-01 -4.90522593e-01
6.89036071e-01 9.44524825e-01 -1.01850939e+00 2.58373562e-02
-1.09564793e+00 9.87180054e-01 -1.16875712e-02 -6.98601454e-02
-8.08359444e-01 -6.43850267e-01 -8.59070659e-01 -1.28858805e-01
-2.45433912e-01 -1.00005597e-01 6.28296018e-01 -8.60862911e-01
-1.36385131e+00 1.36063838e+00 6.60686120e-02 -2.80649871e-01
3.77178431e-01 9.12594274e-02 -9.89476502e-01 4.43398595e-01
6.25349060e-02 5.30606866e-01 1.58831787e+00 -1.04877901e+00
-1.12017587e-01 -6.23833477e-01 -5.99413700e-02 -2.29476556e-01
-9.36734974e-01 6.86217725e-01 1.37727529e-01 -7.06902027e-01
-4.05639470e-01 -8.25000286e-01 3.50475103e-01 5.42101488e-02
-2.76846319e-01 9.97620299e-02 1.51156867e+00 -8.02937806e-01
1.15791881e+00 -2.16581273e+00 -4.78526086e-01 7.92932391e-01
1.20217010e-01 8.99953485e-01 -2.45464429e-01 6.84131622e-01
-4.53734696e-01 7.16624379e-01 1.25525266e-01 -3.12731624e-01
9.35500860e-02 -2.48169541e-01 -3.73745650e-01 7.96769321e-01
5.33056259e-01 5.40220737e-01 -3.40460271e-01 -4.46046531e-01
3.47035259e-01 7.27258623e-01 -3.85692090e-01 2.86624193e-01
4.02741671e-01 -5.16754985e-02 -1.25821918e-01 1.08704460e+00
1.00266099e+00 7.30554879e-01 8.45968500e-02 -2.27352366e-01
1.71824321e-01 -1.75919056e-01 -1.43357921e+00 5.55607319e-01
-3.27816606e-01 5.85488737e-01 3.56773853e-01 -2.89338440e-01
1.08191717e+00 4.07289326e-01 1.93271607e-01 -3.11185628e-01
3.62360775e-01 2.73931712e-01 2.88757123e-02 -3.22924376e-01
4.74849999e-01 5.82293533e-02 1.92189485e-01 8.21604252e-01
3.88470381e-01 1.09253339e-01 1.42359376e-01 1.41820222e-01
9.16637540e-01 -3.71807605e-01 1.10918611e-01 -4.23022717e-01
9.07579005e-01 -6.23873651e-01 -3.72916535e-02 5.63486516e-01
-4.16615158e-01 4.70343381e-01 3.16131145e-01 -2.64373720e-01
-7.13431180e-01 -1.05844259e+00 -3.33037585e-01 7.59352744e-01
-1.36098817e-01 -6.06769264e-01 -1.05048430e+00 -9.72464979e-01
3.29690665e-01 3.12343359e-01 -7.73120403e-01 -3.81493241e-01
-4.66526806e-01 -4.89672542e-01 1.39405668e+00 1.86395288e-01
2.84752667e-01 -1.02507567e+00 -4.43748713e-01 -2.24505767e-01
4.20337677e-01 -8.73444080e-01 -6.92261994e-01 -3.91641974e-01
-3.51958811e-01 -1.45901716e+00 -4.94980276e-01 -5.68258286e-01
9.42794323e-01 1.35696068e-01 6.27835095e-01 5.11506796e-01
-1.02727187e+00 9.41205740e-01 -1.47657692e-01 -4.02828455e-01
-9.28246498e-01 -4.03290480e-01 5.83499968e-01 5.31352341e-01
7.29578614e-01 -3.33375871e-01 -5.51174343e-01 6.41556740e-01
-1.04091048e+00 -9.57301438e-01 4.31160420e-01 6.31110251e-01
-4.21713404e-02 2.04695046e-01 4.49020237e-01 -6.88925147e-01
1.01029611e+00 -3.62066418e-01 -4.12128448e-01 3.68397892e-01
-8.63541722e-01 -1.90360397e-01 4.37873483e-01 -6.07327700e-01
-6.64672613e-01 -3.14151496e-02 -2.14258924e-01 -4.27819043e-01
-3.47560972e-01 -9.29520205e-02 -4.27058131e-01 -1.09727156e+00
7.88398802e-01 -6.45288092e-04 5.90916336e-01 -2.23742738e-01
8.99005681e-02 8.80896211e-01 2.24689335e-01 -5.72589755e-01
1.38413441e+00 3.11507374e-01 4.21625143e-03 -1.18347907e+00
5.44775128e-01 -2.34022722e-01 -6.02968574e-01 -4.48959410e-01
6.51097953e-01 -4.93649542e-01 -9.88369405e-01 6.66524708e-01
-9.47139621e-01 3.91246885e-01 1.69037566e-01 8.53934959e-02
-7.86525756e-02 8.08556557e-01 -4.59294647e-01 -1.04106879e+00
-4.33020502e-01 -1.44064116e+00 1.07611704e+00 1.83006242e-01
-4.13675249e-01 -7.91806698e-01 -2.65192799e-02 4.04836476e-01
6.79553628e-01 3.93615812e-01 7.97169387e-01 -7.86420763e-01
3.53554077e-02 -6.88079238e-01 -7.38338754e-02 5.25471628e-01
4.60627109e-01 5.18098414e-01 -1.25875592e+00 -5.72671175e-01
-1.32691965e-01 -1.68566182e-01 4.07692492e-01 -3.03710312e-01
7.24708796e-01 -3.20407480e-01 -1.21776514e-01 4.31132734e-01
1.36574376e+00 2.15498954e-01 9.22339261e-01 3.02564770e-01
3.86214614e-01 9.15395677e-01 3.48224729e-01 3.51680845e-01
-3.39725494e-01 8.54443431e-01 4.05766666e-01 -5.57648130e-02
-5.61126061e-02 -1.97692648e-01 7.10118175e-01 -2.42038757e-01
9.07254592e-02 1.70260474e-01 -8.74111414e-01 8.76271948e-02
-1.01192820e+00 -1.12172091e+00 3.35633069e-01 2.53701234e+00
6.83846891e-01 -1.68573976e-01 3.82877827e-01 3.06413978e-01
9.58094120e-01 4.79091518e-02 2.19193157e-02 -7.82029331e-01
1.76122487e-02 8.52208614e-01 6.01343095e-01 5.03296554e-01
-1.18739247e+00 8.95466387e-01 6.28768349e+00 7.04057574e-01
-1.16787255e+00 2.95673180e-02 5.97548664e-01 -1.63889788e-02
1.00236326e-01 -1.32344484e-01 -1.20724487e+00 6.03101373e-01
1.03601909e+00 -6.39519542e-02 4.77997333e-01 7.12092698e-01
-1.06116652e-01 1.39122039e-01 -1.06031466e+00 1.13522112e+00
4.72904891e-01 -1.22634900e+00 3.85375887e-01 5.73860347e-01
1.79895058e-01 -7.79359221e-01 5.22446156e-01 6.80889655e-03
-5.24483360e-02 -1.34021235e+00 6.48883820e-01 -4.77347374e-02
1.09871471e+00 -9.37645793e-01 6.34243369e-01 -5.19252121e-01
-1.23562539e+00 -1.03691570e-01 -3.55162293e-01 3.55468303e-01
-5.26659265e-02 -5.22002727e-02 -1.09637117e+00 2.65264422e-01
5.31391084e-01 -1.16666980e-01 -1.02319038e+00 5.38238227e-01
9.38070863e-02 4.43997949e-01 -4.98949856e-01 3.49079937e-01
2.06409581e-02 -1.44793257e-01 5.31873047e-01 1.31763518e+00
3.96608323e-01 -3.13163638e-01 -1.80480987e-01 8.63848329e-01
-9.36030596e-02 2.55201668e-01 -9.26317275e-01 -4.87502098e-01
8.08333278e-01 1.43287694e+00 -5.66465437e-01 4.90214303e-02
-1.87506020e-01 9.51601326e-01 -1.16620965e-01 2.51509175e-02
-5.34933567e-01 -6.31782353e-01 1.16812193e+00 3.98672193e-01
4.36404645e-02 -1.27812341e-01 1.34932455e-02 -6.78611875e-01
4.32493091e-02 -1.42475462e+00 4.96493042e-01 -9.19534490e-02
-1.15087080e+00 8.28289390e-01 -1.67776886e-02 -1.03470063e+00
-2.63522863e-01 -9.96853650e-01 -8.49754810e-01 1.18085301e+00
-1.00487089e+00 -1.59958148e+00 -7.23339990e-02 7.59149313e-01
4.86892164e-02 -8.02695513e-01 1.16141689e+00 5.02641737e-01
-6.36552513e-01 1.28888440e+00 -5.33330023e-01 3.38471234e-01
8.00884247e-01 -8.80764246e-01 4.43373859e-01 1.13442266e+00
3.22594583e-01 1.27888143e+00 4.01422828e-01 -7.43103743e-01
-1.72910070e+00 -7.37259984e-01 4.01628554e-01 -5.71797073e-01
4.10311580e-01 -6.47261620e-01 -5.82855523e-01 3.32702398e-01
1.35154292e-01 -9.52939764e-02 1.09137571e+00 -1.86078966e-01
-8.50610435e-01 -5.44224195e-02 -1.87316155e+00 6.16165102e-01
5.05115569e-01 -1.05164731e+00 -1.87086523e-01 2.06331369e-02
-1.06316552e-01 2.71623075e-01 -9.22433317e-01 1.77911893e-01
8.92170846e-01 -9.92377579e-01 1.18926144e+00 -5.40021122e-01
2.03814227e-02 -2.23856196e-01 -2.47323066e-01 -6.98309302e-01
-7.38043413e-02 -9.75533545e-01 -1.57847498e-02 1.80910683e+00
3.21211070e-01 -9.79370773e-01 9.59606528e-01 9.22827542e-01
5.96613824e-01 -3.17932576e-01 -9.82728362e-01 -9.32446718e-01
-2.23206550e-01 -3.75043638e-02 8.75199974e-01 1.04052937e+00
-1.58748284e-01 -3.58357668e-01 -2.84506112e-01 5.96969724e-01
8.29463899e-01 -6.51288986e-01 9.13533151e-01 -1.07927811e+00
-2.00445458e-01 -6.68738127e-01 -1.13435757e+00 1.92124084e-01
2.37195030e-01 -6.65803492e-01 -5.87658763e-01 -4.67287689e-01
-2.01417580e-01 -3.21735889e-01 -5.21561690e-02 5.39375961e-01
3.80640328e-02 8.99607241e-01 2.59831458e-01 2.68912297e-02
4.73487884e-01 7.13267177e-02 3.19991231e-01 -2.31058210e-01
-5.39762005e-02 -4.23431396e-02 -8.38807344e-01 3.57528746e-01
7.52600074e-01 -3.72676760e-01 -2.08374500e-01 1.42321825e-01
-1.56466261e-01 -6.32608593e-01 3.70754331e-01 -1.09918785e+00
-4.52390909e-02 8.90485272e-02 5.42580843e-01 -2.99303439e-02
4.06076372e-01 -9.06939328e-01 1.82383955e-01 5.18113852e-01
1.77929148e-01 4.58367795e-01 7.14617610e-01 2.21481100e-01
-4.60249968e-02 -3.99127871e-01 1.03223205e+00 -4.45925556e-02
-6.44111037e-01 3.94970953e-01 -5.27531683e-01 -4.36587602e-01
1.40084493e+00 -8.16615403e-01 -2.35026523e-01 -2.47956604e-01
-6.05966747e-01 -3.80591393e-01 6.75263286e-01 5.75124264e-01
9.18562770e-01 -1.24803841e+00 -7.11257041e-01 9.95112598e-01
1.26596078e-01 -1.24252903e+00 -9.15826261e-02 3.19381624e-01
-7.08450317e-01 -5.14287874e-02 -7.00860262e-01 -2.32625276e-01
-2.14647079e+00 5.00244796e-01 2.91034281e-01 3.26382279e-01
-3.19745392e-02 8.04959655e-01 -2.37973139e-01 -2.65240699e-01
4.97335941e-02 5.83925247e-01 -1.66590288e-01 2.43175611e-01
1.02422535e+00 4.56953853e-01 3.43893826e-01 -1.11048150e+00
-6.15383506e-01 4.43467885e-01 -4.39384669e-01 -2.75636107e-01
8.51404488e-01 2.09858656e-01 -4.18868929e-01 -4.87789541e-01
1.38738263e+00 4.79745597e-01 -5.89607120e-01 4.36917931e-01
-3.63431089e-02 -1.06993806e+00 -1.72164410e-01 -5.68271458e-01
-9.14517403e-01 7.70847797e-01 1.06717694e+00 3.30722213e-01
8.79348874e-01 -4.44087714e-01 4.67903256e-01 2.81828512e-02
4.00100708e-01 -9.17600334e-01 1.49420545e-01 -1.58829708e-02
7.86840498e-01 -9.94128823e-01 -1.80959255e-02 -5.34771502e-01
-2.58177608e-01 1.29621637e+00 7.53916740e-01 9.25758258e-02
7.20354855e-01 3.36346656e-01 2.03118578e-01 -4.34076071e-01
-2.66153999e-02 2.69204736e-01 2.61827707e-01 1.01625419e+00
3.21188390e-01 -1.08486362e-01 -9.35628936e-02 1.14123277e-01
-3.42685640e-01 -6.21512592e-01 4.04072434e-01 1.00633681e+00
-6.48964345e-02 -1.65929043e+00 -8.90936196e-01 4.94609237e-01
-5.99957108e-01 -1.53717408e-02 -1.01425111e+00 6.30004406e-01
1.08251959e-01 1.01444900e+00 -2.42570385e-01 -7.87382662e-01
2.02422917e-01 3.35951954e-01 6.37925267e-01 -6.48067415e-01
-1.17502499e+00 -6.49570704e-01 4.96456288e-02 -4.93458182e-01
1.99135393e-02 -5.91898859e-01 -4.05981928e-01 -6.92192256e-01
-4.52237964e-01 4.24424261e-02 8.42152655e-01 5.55440247e-01
6.99503660e-01 -2.65126914e-01 8.97103906e-01 -1.11183310e+00
-8.52306187e-01 -6.81837499e-01 -6.98156059e-01 5.90334773e-01
2.98348684e-02 -7.82115281e-01 -3.52913707e-01 -7.18184710e-02] | [13.0278959274292, 1.0629799365997314] |
68d81bb5-6f21-4c49-b86b-08a99adee71b | hausanlp-at-semeval-2023-task-12-leveraging | 2304.13634 | null | https://arxiv.org/abs/2304.13634v1 | https://arxiv.org/pdf/2304.13634v1.pdf | HausaNLP at SemEval-2023 Task 12: Leveraging African Low Resource TweetData for Sentiment Analysis | We present the findings of SemEval-2023 Task 12, a shared task on sentiment analysis for low-resource African languages using Twitter dataset. The task featured three subtasks; subtask A is monolingual sentiment classification with 12 tracks which are all monolingual languages, subtask B is multilingual sentiment classification using the tracks in subtask A and subtask C is a zero-shot sentiment classification. We present the results and findings of subtask A, subtask B and subtask C. We also release the code on github. Our goal is to leverage low-resource tweet data using pre-trained Afro-xlmr-large, AfriBERTa-Large, Bert-base-arabic-camelbert-da-sentiment (Arabic-camelbert), Multilingual-BERT (mBERT) and BERT models for sentiment analysis of 14 African languages. The datasets for these subtasks consists of a gold standard multi-class labeled Twitter datasets from these languages. Our results demonstrate that Afro-xlmr-large model performed better compared to the other models in most of the languages datasets. Similarly, Nigerian languages: Hausa, Igbo, and Yoruba achieved better performance compared to other languages and this can be attributed to the higher volume of data present in the languages. | ['Abdulmalik Yusuf Jamoh', 'Abdulkadir Abdullahi', 'Mahmoud Said Ahmad', 'Sanah Abdullahi Muaz', 'Murja Sani Gadanya', 'Saminu Mohammad Aliyu', 'Shamsuddeen Umaru Adamu', 'Musa Bello', 'Nur Bala Rabiu', 'Aliyu Yusuf', 'Aliyu Rabiu Shuaibu', 'Amina Abubakar Imam', 'Ahmad Mustapha Wali', 'Falalu Ibrahim Lawan', 'Saheed Abdullahi Salahudeen'] | 2023-04-26 | null | null | null | null | ['twitter-sentiment-analysis'] | ['natural-language-processing'] | [-5.20056605e-01 -4.57262039e-01 -2.19765425e-01 -4.94772553e-01
-9.77445483e-01 -8.77797782e-01 9.16929126e-01 4.91303176e-01
-5.85164428e-01 8.18644047e-01 5.68098605e-01 -5.76578438e-01
4.55882758e-01 -4.29869592e-01 -3.45652997e-01 -3.44062358e-01
5.70924534e-03 4.91675258e-01 -1.76137999e-01 -1.28594518e+00
4.13353354e-01 -2.80604213e-02 -9.93181944e-01 7.87763417e-01
4.66183364e-01 7.21288860e-01 -2.06816241e-01 8.37554574e-01
-2.57646739e-01 1.26880038e+00 -3.42725694e-01 -7.46495903e-01
1.81915283e-01 -1.08301699e-01 -1.08742535e+00 -4.03732926e-01
3.68098974e-01 4.26302254e-01 5.70640743e-01 7.25294530e-01
5.38219333e-01 -9.64051485e-02 7.97793150e-01 -1.16182613e+00
-7.31440365e-01 5.77749729e-01 -1.00448608e+00 1.74725279e-01
5.17500579e-01 -4.65829194e-01 1.09904981e+00 -1.16439390e+00
8.36627960e-01 1.56180763e+00 1.11495566e+00 8.36859047e-02
-4.45841044e-01 -9.40653920e-01 2.89189786e-01 -3.36982042e-01
-9.32751775e-01 -4.82463568e-01 8.92931372e-02 -7.26575732e-01
1.26481593e+00 2.38187045e-01 3.93581539e-01 8.80936444e-01
7.15962350e-01 5.77199340e-01 2.00742435e+00 -4.35783207e-01
-1.63309917e-01 6.82535946e-01 4.33337599e-01 5.70895135e-01
6.09597517e-03 -5.85054874e-01 -8.95139635e-01 -8.79935920e-02
-4.51830119e-01 4.05192450e-02 3.41722786e-01 6.06393516e-01
-1.08039725e+00 1.18949592e+00 3.67354840e-01 4.78163779e-01
-1.59458578e-01 -3.77976745e-01 9.04105425e-01 8.98569167e-01
1.32028091e+00 1.56837404e-01 -1.21689630e+00 -7.83196688e-02
-8.03131223e-01 2.13322058e-01 1.09124982e+00 8.46765220e-01
9.27144170e-01 -2.12614406e-02 1.93782017e-01 1.06727839e+00
6.44507051e-01 8.92282248e-01 6.54574275e-01 -4.99957390e-02
7.84671366e-01 7.02950597e-01 7.92676732e-02 -9.39094484e-01
-6.45557642e-01 1.92181717e-04 -2.99253404e-01 -2.55525321e-01
2.48635098e-01 -7.85734713e-01 -1.06456113e+00 1.16369843e+00
2.57666826e-01 -5.69445133e-01 5.88679373e-01 6.03591740e-01
1.08566260e+00 7.70231903e-01 1.85086921e-01 3.17609645e-02
1.62721729e+00 -1.20771492e+00 -6.36280060e-01 -3.93623352e-01
1.11696708e+00 -1.52484655e+00 1.23603094e+00 4.37605619e-01
-3.98889720e-01 -1.21610470e-01 -7.60738790e-01 -4.11375947e-02
-1.35874736e+00 1.60233870e-01 6.81025088e-01 1.02020442e+00
-1.15334630e+00 -1.99500546e-01 -5.40100574e-01 -8.68169844e-01
1.81589052e-01 2.51383066e-01 -6.45717621e-01 4.95487936e-02
-1.17752719e+00 9.76045489e-01 5.18324338e-02 -4.16419096e-02
-6.38992846e-01 -5.25383532e-01 -9.51526463e-01 -6.95538759e-01
-2.05915064e-01 3.93804997e-01 1.21822560e+00 -1.40202951e+00
-1.14997506e+00 1.38095999e+00 -4.39645231e-01 -6.53975308e-02
2.11337358e-01 -2.38009244e-01 -9.02170002e-01 -2.64962196e-01
9.36965883e-01 3.24738383e-01 3.61679703e-01 -1.04212236e+00
-1.12774360e+00 -7.08505690e-01 4.52910177e-02 2.62235582e-01
-4.39444780e-01 9.14129972e-01 -1.09808650e-02 -6.27614856e-01
-1.62831217e-01 -1.21458805e+00 -6.75639138e-02 -1.18181360e+00
-3.94741535e-01 -7.87562281e-02 1.07737613e+00 -7.24625170e-01
1.03606057e+00 -1.86581576e+00 -4.28988546e-01 -2.04185694e-02
-2.03149185e-01 -1.37583017e-01 3.08624506e-02 9.73337591e-01
-1.31181106e-01 4.43787634e-01 2.16200918e-01 -3.70165914e-01
-1.35898784e-01 -1.47018880e-01 -3.25980306e-01 6.41842842e-01
6.71024472e-02 5.85500658e-01 -8.60349655e-01 -3.31744999e-01
-1.53047442e-01 2.86442637e-01 -3.35190684e-01 -4.56313014e-01
-2.84555890e-02 4.17197376e-01 -4.01437044e-01 1.16412210e+00
5.35485506e-01 -9.44758728e-02 1.02788158e-01 -1.36339292e-01
-6.46997213e-01 2.27719352e-01 -6.25379086e-01 1.20078671e+00
-7.59053826e-01 8.59936357e-01 1.84074834e-01 -3.12247872e-01
9.25975323e-01 3.72193754e-01 3.46200794e-01 -7.99712598e-01
3.44599456e-01 8.87185872e-01 -1.96087331e-01 -2.43570000e-01
8.39640498e-01 -3.68392259e-01 -7.62830615e-01 7.32178986e-01
-7.96460211e-02 -2.79340208e-01 4.87405181e-01 3.86880934e-01
4.57932711e-01 8.00120309e-02 2.57438481e-01 -7.84326851e-01
8.26245248e-01 5.10851800e-01 3.02014947e-01 3.10382664e-01
-4.23742622e-01 4.05351549e-01 3.65932107e-01 -6.76587105e-01
-6.93508148e-01 -3.36473465e-01 -2.85368919e-01 2.04548097e+00
-4.14447844e-01 -6.56490743e-01 -3.27319235e-01 -8.22069764e-01
-1.84026256e-01 3.92418355e-01 -9.14520442e-01 7.75406778e-01
-2.82252043e-01 -1.44984722e+00 5.41481197e-01 -7.73753896e-02
5.33158004e-01 -1.05171192e+00 -2.62830943e-01 -1.10580601e-01
-2.77105659e-01 -1.06024957e+00 -3.37955862e-01 4.65616286e-01
-3.15911055e-01 -9.16232407e-01 -5.58687449e-01 -1.02027273e+00
4.70405459e-01 3.04338001e-02 1.03460586e+00 -3.09561014e-01
1.92679554e-01 1.44712269e-01 -7.96638429e-01 -1.25234437e+00
-1.78752273e-01 4.30789739e-01 -4.42647412e-02 -2.85978094e-02
9.83537018e-01 3.81577104e-01 -1.84306592e-01 2.59984471e-02
-5.57559729e-01 -9.96627212e-02 9.68998522e-02 4.05443043e-01
3.02096635e-01 -2.64620245e-01 8.87937903e-01 -1.39641905e+00
5.64075530e-01 -1.05113542e+00 -1.34954881e-02 -3.30705866e-02
-4.18837428e-01 -6.92402899e-01 6.68043375e-01 1.12986207e-01
-1.00322223e+00 -1.73373893e-01 -2.02258721e-01 4.19675827e-01
2.28288904e-01 1.06533694e+00 6.31477892e-01 1.48022637e-01
6.51727378e-01 -3.50495577e-01 -2.75312781e-01 -1.59888133e-01
7.14058802e-02 1.34511876e+00 -2.05967650e-01 -5.90113640e-01
2.48945817e-01 7.41505623e-01 -4.48810935e-01 -1.04926741e+00
-1.29688430e+00 -7.60178864e-01 -5.26644409e-01 -3.45397294e-01
1.00884938e+00 -1.59983838e+00 -3.01752210e-01 9.20713842e-01
-5.51594913e-01 -4.52788353e-01 2.39432782e-01 2.62911379e-01
1.43062353e-01 -3.55979234e-01 -1.04435182e+00 -9.01937187e-01
-7.72925794e-01 -1.27448440e+00 1.20506644e+00 -9.18529835e-03
-5.20344913e-01 -1.47711778e+00 4.33399409e-01 4.85458940e-01
4.51712161e-01 4.93977636e-01 6.85841680e-01 -1.33019948e+00
5.45403481e-01 -1.55972540e-01 -8.60414729e-02 2.08862126e-01
4.08464551e-01 2.81862617e-01 -9.62811947e-01 -6.30041420e-01
-7.09733814e-02 -1.15282583e+00 6.05454862e-01 1.55060917e-01
5.48772849e-02 -2.33152464e-01 6.87370449e-02 4.96335104e-02
1.55287492e+00 1.60021540e-02 -8.62392336e-02 1.02410126e+00
7.14040160e-01 8.47958028e-01 1.00449145e+00 3.64588469e-01
1.20537126e+00 9.44291335e-03 -1.62136510e-01 -1.85054466e-01
3.21000665e-01 1.47485375e-01 1.00303757e+00 1.66305304e+00
2.03253135e-01 -6.18413799e-02 -1.48742664e+00 9.20332491e-01
-1.73433352e+00 -4.72386450e-01 -6.78467453e-01 1.66770685e+00
9.21191931e-01 8.21017195e-03 2.81103969e-01 -5.25877774e-02
1.88882843e-01 2.84470439e-01 1.44884698e-02 -1.16882861e+00
-4.50929374e-01 3.49795520e-01 6.74122214e-01 6.82112038e-01
-1.19032907e+00 1.46601677e+00 5.61995173e+00 6.22571111e-01
-1.48516667e+00 6.12051666e-01 1.02059352e+00 -1.63607493e-01
-2.02564791e-01 9.92206708e-02 -1.12459028e+00 3.45569342e-01
1.41706836e+00 1.20482735e-01 -3.63915786e-02 4.48454380e-01
2.77900368e-01 -4.00505513e-01 -4.73398268e-01 3.82918268e-01
3.66772622e-01 -9.90637660e-01 -4.14671726e-04 -2.75709808e-01
1.29145777e+00 1.05304360e+00 1.93645939e-01 7.00375497e-01
6.55833662e-01 -9.98767555e-01 1.00147700e+00 -1.93618182e-02
8.30456257e-01 -1.05195260e+00 9.41335857e-01 1.53219655e-01
-1.13713562e+00 2.15213802e-02 -1.93429530e-01 -2.37418145e-01
-7.79970288e-02 3.50767523e-01 -3.64272177e-01 5.84899366e-01
1.34149921e+00 1.31773818e+00 -8.36929321e-01 -5.89186624e-02
9.05238837e-02 7.28549957e-01 -8.34967643e-02 -2.30926141e-01
9.37293053e-01 -3.62052917e-01 1.59935057e-01 1.87503374e+00
2.62231559e-01 -3.47048908e-01 3.66801709e-01 -2.80915082e-01
-6.92166612e-02 8.33853364e-01 -6.66002631e-01 -2.94092208e-01
1.68995529e-01 1.68128777e+00 -9.69044805e-01 -2.82682806e-01
-8.34588230e-01 4.09022093e-01 1.40357062e-01 1.80080354e-01
-3.07620734e-01 -5.33895969e-01 2.31034547e-01 -4.22314070e-02
-1.17004506e-01 1.40130386e-01 -1.47685468e-01 -1.34143519e+00
-6.04417562e-01 -1.58002090e+00 5.55231571e-01 -7.56958842e-01
-1.32237816e+00 9.27759528e-01 -2.59199470e-01 -8.98935258e-01
-1.12988185e-02 -9.29457247e-01 -3.24605763e-01 1.11403871e+00
-1.82984042e+00 -1.90857112e+00 1.30733505e-01 8.86473656e-01
7.50395954e-01 -7.73532271e-01 9.27106738e-01 5.41000247e-01
-5.50186038e-01 1.95031136e-01 1.26840308e-01 2.02055365e-01
1.27711678e+00 -1.29575658e+00 1.26334101e-01 6.13204300e-01
-1.86041847e-01 7.15746343e-01 5.16389430e-01 -8.74627531e-01
-1.22252381e+00 -1.19035327e+00 1.54955852e+00 -8.49626541e-01
1.45653725e+00 -4.69063729e-01 -2.23032102e-01 1.18467760e+00
7.04657972e-01 -4.94622648e-01 1.19457686e+00 3.89982402e-01
-3.54800582e-01 2.68820841e-02 -1.08530390e+00 4.52165067e-01
-5.25039881e-02 -8.17200959e-01 -3.19608152e-01 7.53159046e-01
3.12706381e-01 -3.50941509e-01 -1.03818452e+00 2.00731561e-01
6.89490557e-01 -8.23264480e-01 2.33322248e-01 -5.76782465e-01
8.61080050e-01 -1.39559343e-01 -6.68092489e-01 -1.36112583e+00
3.17929864e-01 -4.55568314e-01 7.77175248e-01 1.27506852e+00
9.89306867e-01 -9.82664406e-01 4.13880199e-01 -5.31131737e-02
-1.50187584e-02 -5.21026611e-01 -4.79617089e-01 -2.08979864e-02
6.36868238e-01 -5.69118381e-01 3.14565271e-01 1.57207561e+00
-1.08441591e-01 8.17897856e-01 -2.33435288e-01 -2.10447773e-01
4.10170667e-02 1.50856316e-01 8.91005456e-01 -9.93514657e-01
3.74630272e-01 -1.94781914e-01 1.58402443e-01 -1.35641634e-01
3.13357055e-01 -1.09034920e+00 -3.68842900e-01 -1.40503109e+00
2.07497299e-01 -6.54297650e-01 -1.48915917e-01 6.83162689e-01
-1.09753489e-01 8.03810954e-01 1.77957654e-01 3.83014262e-01
-7.11708903e-01 -6.80681840e-02 8.76484752e-01 -1.73789840e-02
-1.41111724e-02 -2.63701499e-01 -1.22427440e+00 9.54298913e-01
5.96051574e-01 -6.67531192e-01 -2.18986824e-01 -3.81402135e-01
1.27815557e+00 -4.90364820e-01 -4.15460467e-01 -3.66379052e-01
-1.70062669e-02 -7.66874617e-03 2.72414267e-01 -8.77221763e-01
1.14957966e-01 -6.51928306e-01 -2.43892893e-01 2.69276291e-01
-1.43306941e-01 7.69831240e-01 3.59772235e-01 -1.71316117e-02
-5.06791770e-01 -1.78490542e-02 6.66654229e-01 -2.38417104e-01
-5.89117765e-01 1.04699127e-01 -7.51427770e-01 3.72113258e-01
9.09474850e-01 6.35322705e-02 -6.15696132e-01 -2.78931975e-01
-6.47695780e-01 5.17321944e-01 3.39794368e-01 7.20578849e-01
-3.11848503e-02 -1.01775682e+00 -1.11706090e+00 1.98119543e-02
4.49078500e-01 -4.06243593e-01 -2.81839609e-01 1.15645671e+00
-8.84589732e-01 6.60738528e-01 -2.28371352e-01 -4.38104004e-01
-1.13284373e+00 -3.59854400e-01 1.95179999e-01 -5.85081756e-01
3.47656876e-01 8.94156158e-01 -9.86498445e-02 -1.50919068e+00
-4.58783865e-01 7.34179318e-02 -7.92645156e-01 8.91642034e-01
2.48122945e-01 3.01818967e-01 3.51689100e-01 -1.63430643e+00
-5.76322973e-01 6.03052437e-01 -4.29792345e-01 -3.36786568e-01
1.47979259e+00 -3.55519742e-01 -7.43875146e-01 1.17763186e+00
1.36488712e+00 6.22081041e-01 -1.58093721e-01 1.69544846e-01
2.66142130e-01 -7.58332089e-02 3.09175928e-04 -1.13564920e+00
-9.50582564e-01 5.68374872e-01 4.73357230e-01 2.35869244e-01
8.95204902e-01 -3.89440000e-01 5.23891747e-01 1.27497151e-01
5.28818130e-01 -1.39113390e+00 -1.21488616e-01 1.27358985e+00
6.70382142e-01 -1.74811292e+00 1.16245493e-01 1.58658713e-01
-1.30823815e+00 9.07320917e-01 4.66200680e-01 -2.70994380e-02
1.24594891e+00 1.50590092e-01 9.26170170e-01 -6.33346975e-01
-8.35670531e-01 -3.73883992e-02 1.42448291e-01 1.41835451e-01
1.37741530e+00 2.74216533e-01 -4.49526131e-01 5.73811531e-01
-8.20695639e-01 -2.89295524e-01 6.66879058e-01 1.29314196e+00
-4.73324284e-02 -9.22173500e-01 -3.72168005e-01 6.86724961e-01
-1.52650440e+00 -4.22086060e-01 -6.86213493e-01 1.09565938e+00
3.32336389e-02 1.44138253e+00 -4.78148423e-02 -2.38426074e-01
1.31910533e-01 -1.08484931e-01 -2.87184566e-01 -9.61027682e-01
-1.68675125e+00 3.83235872e-01 5.24153173e-01 -1.60868913e-01
-8.24711919e-01 -7.58275449e-01 -1.13381088e+00 -6.20088696e-01
-2.81168014e-01 4.38958585e-01 1.05037034e+00 9.74635243e-01
1.34710804e-01 1.14757992e-01 7.51421332e-01 -5.86178243e-01
1.42215729e-01 -1.27712250e+00 -6.21478260e-01 2.63657987e-01
5.27417243e-01 -1.49502769e-01 -3.71939540e-01 6.44975156e-02] | [11.176915168762207, 6.968197345733643] |
b52e9938-a321-4824-ad5f-dd4ba771f4aa | r-tuning-regularized-prompt-tuning-in-open | 2303.05122 | null | https://arxiv.org/abs/2303.05122v1 | https://arxiv.org/pdf/2303.05122v1.pdf | R-Tuning: Regularized Prompt Tuning in Open-Set Scenarios | In realistic open-set scenarios where labels of a part of testing data are totally unknown, current prompt methods on vision-language (VL) models always predict the unknown classes as the downstream training classes. The exhibited label bias causes difficulty in the open set recognition (OSR), by which an image should be correctly predicted as one of the known classes or the unknown one. To learn prompts in open-set scenarios, we propose the Regularized prompt Tuning (R-Tuning) to mitigate the label bias. It introduces open words from the WordNet to extend the range of words forming the prompt texts from only closed-set label words to more. Thus, prompts are tuned in a simulated open-set scenario. Besides, inspired by the observation that classifying directly on large datasets causes a much higher false positive rate than on small datasets, we propose the Combinatorial Tuning and Testing (CTT) strategy for improving performance. CTT decomposes R-Tuning on large datasets as multiple independent group-wise tuning on fewer classes, then makes comprehensive predictions by selecting the optimal sub-prompt. For fair comparisons, we construct new baselines for OSR based on VL models, especially for prompt methods. Our method achieves the best results on datasets with various scales. Extensive ablation studies validate the effectiveness of our method. | ['Junchi Yan', 'Qi Tian', 'Min Cao', 'Xiaopeng Zhang', 'Ning Liao'] | 2023-03-09 | null | null | null | null | ['open-set-learning'] | ['miscellaneous'] | [ 4.83359307e-01 1.39728621e-01 -4.53136951e-01 -6.36414707e-01
-1.04102993e+00 -7.75810719e-01 4.70199406e-01 -2.94495612e-01
-3.76537323e-01 5.78157961e-01 1.23324715e-01 -4.37751412e-01
1.66575745e-01 -5.55107832e-01 -6.79900467e-01 -5.55389822e-01
4.13844138e-01 6.31824672e-01 3.63188833e-01 -4.48055975e-02
1.48984104e-01 8.37503374e-02 -1.63728166e+00 5.10380745e-01
7.75300086e-01 1.12999749e+00 3.82935822e-01 4.46769655e-01
-3.88068706e-01 6.70069635e-01 -6.44483984e-01 -4.22533602e-01
5.67379177e-01 -6.23080209e-02 -7.63718247e-01 3.75282824e-01
9.50253487e-01 -3.94249499e-01 -2.24065900e-01 1.03107715e+00
5.60550451e-01 2.72511113e-02 9.39685106e-01 -1.30288529e+00
-1.00861526e+00 5.13807118e-01 -5.79869270e-01 -2.20602322e-02
1.74408436e-01 4.07712758e-01 1.26343906e+00 -1.21615529e+00
4.02377963e-01 1.20523953e+00 5.77880621e-01 8.65225494e-01
-1.22217715e+00 -9.64888394e-01 6.36994839e-01 3.19881141e-01
-1.54782248e+00 -3.47988427e-01 4.26097363e-01 -5.09536743e-01
6.23195112e-01 3.72893810e-01 2.20873460e-01 1.56698680e+00
-2.65985101e-01 9.08482254e-01 1.29308879e+00 -3.72441828e-01
1.91759124e-01 3.59534174e-01 7.30233490e-01 4.61660951e-01
2.55579352e-01 4.02745515e-01 -5.00096440e-01 -8.81370977e-02
5.27989149e-01 -5.76524660e-02 -5.07475019e-01 6.36850446e-02
-1.33820891e+00 8.13419521e-01 2.77283639e-01 -2.92639852e-01
-5.98911643e-02 -7.81976208e-02 2.48747677e-01 2.83031136e-01
5.69646418e-01 3.81028980e-01 -9.05302942e-01 2.71341801e-01
-7.75970101e-01 -1.28554836e-01 6.82555199e-01 1.16958427e+00
8.34773183e-01 -2.62046784e-01 -9.59226966e-01 1.12020385e+00
4.52174783e-01 6.07979715e-01 5.02542615e-01 -5.65408528e-01
3.62466067e-01 6.29392803e-01 1.70744490e-02 -4.56735194e-01
-2.08381444e-01 -6.82371616e-01 -5.63705862e-01 -2.95377165e-01
4.14278328e-01 -6.39965758e-02 -1.42374444e+00 1.82561350e+00
1.68483436e-01 4.67745930e-01 8.07064399e-02 1.03279078e+00
1.24023855e+00 7.06961632e-01 1.42564356e-01 -4.01464909e-01
1.40425718e+00 -1.16560864e+00 -6.30165219e-01 -5.07843256e-01
7.47418880e-01 -7.46366620e-01 1.50323737e+00 4.55714703e-01
-7.01906756e-02 -6.13305092e-01 -8.25441241e-01 3.53436582e-02
-2.56613016e-01 4.56553668e-01 5.09664237e-01 4.82688725e-01
-1.01151597e+00 1.49666946e-02 -1.39617205e-01 -2.47273445e-01
5.74485183e-01 1.82966173e-01 -1.30181670e-01 -4.37898487e-01
-1.15359557e+00 6.13537848e-01 4.03540820e-01 -6.73051551e-02
-1.16234708e+00 -7.87615299e-01 -6.82715356e-01 -6.46331981e-02
9.47222233e-01 -4.45996910e-01 1.25551307e+00 -9.17516172e-01
-1.05920804e+00 1.08321416e+00 -2.84397930e-01 -2.64301389e-01
5.45606434e-01 -1.70533299e-01 -3.34317356e-01 -2.12510079e-01
2.55879551e-01 1.12764907e+00 1.11050165e+00 -1.48840845e+00
-4.83634323e-01 -2.86038429e-01 2.10757390e-01 1.20355517e-01
-4.78087455e-01 -1.45560697e-01 -4.86214846e-01 -7.11470842e-01
1.92170173e-01 -8.76319289e-01 -2.24993363e-01 7.31705204e-02
-6.57966614e-01 -6.53689682e-01 5.22733867e-01 -2.25931868e-01
1.31581545e+00 -2.24004817e+00 -4.40000176e-01 -1.10087857e-01
4.66331601e-01 3.82565469e-01 -5.13461590e-01 -4.30907533e-02
-2.35310465e-01 -9.01759043e-03 4.11840864e-02 -1.48061037e-01
-8.61011669e-02 5.90580106e-01 -9.70123768e-01 5.53388782e-02
3.05434018e-01 9.76004183e-01 -7.75919259e-01 -5.63884199e-01
-1.43512517e-01 -9.17029008e-02 -3.70417446e-01 3.58805776e-01
-5.64770699e-01 3.18304390e-01 -4.39073741e-01 6.46730900e-01
5.26156425e-01 -7.24041998e-01 -3.28876317e-01 -2.78413594e-01
8.49765986e-02 2.84155514e-02 -1.21602857e+00 1.21748960e+00
-2.60855943e-01 4.67900962e-01 -5.77867150e-01 -7.61830389e-01
1.17371714e+00 1.10957928e-01 7.29156425e-03 -4.96867329e-01
1.18249059e-01 1.48136079e-01 -9.89366993e-02 -6.30856276e-01
3.33125025e-01 -1.08779453e-01 8.58979151e-02 2.66538590e-01
1.99979350e-01 1.89695448e-01 1.05868340e-01 3.15811634e-01
9.79118705e-01 7.08083510e-02 2.65926212e-01 -1.19643524e-01
5.02610445e-01 1.02377012e-01 8.58709157e-01 1.10222459e+00
-5.20221531e-01 7.51667619e-01 3.94247621e-01 -4.14629489e-01
-7.20764816e-01 -1.00005257e+00 -4.65628356e-01 1.41509914e+00
3.55279684e-01 -4.38923270e-01 -5.07477164e-01 -1.13559675e+00
-1.24374129e-01 7.55525649e-01 -6.98448598e-01 -1.62451372e-01
-1.15125719e-02 -5.94826877e-01 4.83235329e-01 5.16432583e-01
2.36246362e-01 -9.55939233e-01 -2.22791821e-01 -2.15757072e-01
-4.01015520e-01 -1.33068776e+00 -7.81536102e-01 5.20361602e-01
-4.83786315e-01 -1.22503090e+00 -8.06189477e-01 -9.20904040e-01
8.32797885e-01 4.27051425e-01 1.10297692e+00 -1.21711008e-01
-1.79057151e-01 3.93144369e-01 -6.36677265e-01 -5.97475529e-01
-2.58323550e-01 -2.90573746e-01 1.88826695e-01 3.87306720e-01
6.65072262e-01 1.66745007e-01 -4.95846033e-01 6.46084785e-01
-6.18002594e-01 2.42590874e-01 5.44592917e-01 1.15134358e+00
9.02077079e-01 -3.46638650e-01 7.06729174e-01 -1.07429802e+00
3.40793908e-01 -4.43855047e-01 -5.02101839e-01 6.37725234e-01
-8.34367812e-01 2.65521593e-02 5.54973066e-01 -1.20547140e+00
-7.74034858e-01 4.97980863e-02 8.88858140e-02 -7.07550406e-01
-3.96555066e-01 9.62968618e-02 -1.18055329e-01 -1.81983560e-02
8.74331355e-01 3.66766900e-01 -3.41894984e-01 -2.86556751e-01
4.95446801e-01 1.04088426e+00 4.20669496e-01 -5.91728568e-01
7.26539731e-01 3.59373271e-01 -4.26995873e-01 -6.80162489e-01
-1.63661575e+00 -7.43363619e-01 -3.15151155e-01 -3.46387684e-01
7.29264021e-01 -1.19536018e+00 -2.71338284e-01 3.74024779e-01
-1.24617672e+00 -4.07023638e-01 -4.10715878e-01 3.20249379e-01
-3.89493793e-01 2.42328823e-01 -2.45197758e-01 -6.82973683e-01
-1.80531189e-01 -1.30001831e+00 1.38077343e+00 2.61900425e-01
-1.90585643e-01 -6.35035336e-01 -9.30262133e-02 5.08350194e-01
-6.28925934e-02 -5.21110058e-01 7.46982098e-01 -1.28497219e+00
-5.45257628e-01 1.92289203e-02 -5.48262179e-01 5.72838426e-01
-1.32122666e-01 -3.35038990e-01 -1.27464354e+00 -2.43699104e-01
-1.33578420e-01 -8.09955478e-01 1.11612964e+00 3.10142338e-01
1.37512434e+00 -1.36792183e-01 -3.55381370e-01 5.70883751e-01
1.30140984e+00 1.41886203e-02 4.50257570e-01 9.58060622e-02
8.15337300e-01 5.51182926e-01 1.03842831e+00 4.23293203e-01
2.57848859e-01 4.86172289e-01 3.63590419e-01 -7.53984749e-02
-3.62474442e-01 -3.96235406e-01 4.67595190e-01 4.83571142e-01
4.96120691e-01 -4.31629449e-01 -9.23858762e-01 5.34677088e-01
-1.67289257e+00 -4.32256490e-01 -3.40810597e-01 2.31122112e+00
1.04565990e+00 1.94560215e-01 -2.87405580e-01 -9.93667841e-02
9.10052299e-01 6.30502105e-02 -7.94468045e-01 1.07781872e-01
-2.24354774e-01 -8.22434127e-02 6.00667000e-01 3.46065074e-01
-1.01672924e+00 1.20017684e+00 6.20885468e+00 1.32102358e+00
-1.16078568e+00 3.21116716e-01 8.22408319e-01 -6.15873784e-02
-3.91026288e-01 1.19490325e-01 -1.45400679e+00 3.96732450e-01
5.99501431e-01 1.43657967e-01 2.09762201e-01 8.32517922e-01
1.31637052e-01 -2.62945406e-02 -1.43823600e+00 1.29267812e+00
3.44972372e-01 -1.15453303e+00 4.11300540e-01 8.81158933e-03
7.64161289e-01 1.89946771e-01 3.37552242e-02 7.55079150e-01
2.71646619e-01 -8.16653609e-01 7.85275280e-01 3.00148219e-01
1.29450059e+00 -1.12294987e-01 5.99504173e-01 6.57303572e-01
-7.87326694e-01 -3.88889492e-01 -5.55942059e-01 6.27221167e-02
-6.84532449e-02 5.86452067e-01 -8.94272327e-01 9.01954919e-02
5.50179005e-01 7.10844457e-01 -8.36722374e-01 9.26709056e-01
-3.81578088e-01 1.00362861e+00 -1.84776768e-01 -4.98223528e-02
1.81846917e-01 -7.33336210e-02 4.71510500e-01 9.03483152e-01
-1.47480220e-01 2.44139582e-01 5.52963853e-01 8.87397051e-01
-5.68444794e-03 5.77537119e-02 -5.63782811e-01 6.02875650e-02
4.86198664e-01 1.11163378e+00 -7.09863424e-01 -3.86591941e-01
-5.46839416e-01 7.14801788e-01 4.10358071e-01 5.23637235e-01
-8.92151117e-01 2.99548730e-02 3.76538128e-01 2.66334284e-02
2.49872446e-01 2.30205178e-01 -3.88366550e-01 -1.17899346e+00
4.29030657e-02 -7.07748652e-01 6.26462817e-01 -8.88247073e-01
-1.68082798e+00 5.91131032e-01 -1.57971933e-01 -1.51143456e+00
2.66090870e-01 -8.89856279e-01 -3.49729955e-01 5.95219731e-01
-1.85091639e+00 -1.07759392e+00 -3.92617553e-01 6.33717716e-01
1.12382317e+00 -1.87193871e-01 7.05946147e-01 2.08644673e-01
-6.20156586e-01 9.52831507e-01 -1.55579165e-01 1.31787792e-01
1.23736477e+00 -1.13634086e+00 1.04350038e-01 7.77760088e-01
4.50679481e-01 4.69355375e-01 5.11827886e-01 -6.80139303e-01
-9.14851010e-01 -1.44987524e+00 7.17637241e-01 -6.56622827e-01
7.43859947e-01 -4.63131219e-01 -9.41296339e-01 6.93865836e-01
-3.57801259e-01 4.52845097e-01 8.43789458e-01 1.28954083e-01
-8.27707469e-01 -6.11373596e-02 -7.11097062e-01 6.20322406e-01
1.12719631e+00 -5.08903384e-01 -7.89440632e-01 6.47171259e-01
1.29034865e+00 -2.97133923e-01 -4.67882037e-01 5.76639593e-01
2.34686300e-01 -5.61840653e-01 7.80539334e-01 -7.31294394e-01
5.64480484e-01 -3.32835346e-01 -4.56792295e-01 -1.10545802e+00
-2.54783928e-01 -2.89865315e-01 -1.35361731e-01 1.31192446e+00
5.62842786e-01 -6.99128926e-01 5.83375037e-01 3.73277903e-01
-2.16059372e-01 -9.60572302e-01 -7.27430761e-01 -9.00766611e-01
-1.24424040e-01 -7.50179052e-01 3.21508557e-01 1.05598307e+00
-4.77051646e-01 6.82428002e-01 -3.07937920e-01 3.76390815e-01
6.27539098e-01 1.27077818e-01 5.79166770e-01 -1.36095023e+00
-3.96379650e-01 -4.76979353e-02 -1.76936090e-01 -1.50143278e+00
3.63548636e-01 -9.33895946e-01 2.76715666e-01 -1.14331222e+00
5.75340748e-01 -7.89328694e-01 -3.63535166e-01 7.77134418e-01
-5.28107584e-01 4.12698776e-01 6.04715720e-02 4.04280216e-01
-9.10824656e-01 6.12971902e-01 1.43369889e+00 -3.26730669e-01
-1.05393203e-02 4.60141078e-02 -8.93478096e-01 6.62219048e-01
2.89602965e-01 -4.66079354e-01 -5.77659965e-01 -3.04661661e-01
2.73575246e-01 -1.37578035e-02 3.79465640e-01 -5.96613228e-01
1.10814206e-01 -2.72888720e-01 1.61176443e-01 -7.59556532e-01
1.08110815e-01 -7.97119796e-01 -4.37406152e-01 1.40819848e-01
-8.40759158e-01 -6.40202224e-01 -3.34735848e-02 8.90959442e-01
3.61143537e-02 -3.02140355e-01 6.53679311e-01 1.36844009e-01
-1.12078488e+00 6.09720409e-01 1.55146107e-01 4.04019624e-01
1.14505792e+00 -4.47700590e-01 -6.85845137e-01 -7.11069554e-02
-8.07809651e-01 6.77202344e-01 3.64779904e-02 7.70642340e-01
7.04549313e-01 -1.10512400e+00 -7.37563789e-01 2.72062540e-01
8.89134288e-01 2.89916009e-01 1.48003891e-01 6.69193029e-01
6.39742762e-02 3.79513204e-01 4.12449747e-01 -9.50604320e-01
-1.33268487e+00 6.52962208e-01 1.99431628e-01 -1.72765985e-01
-4.66412336e-01 1.13877940e+00 8.74134421e-01 -6.81257427e-01
6.00925207e-01 -9.74558741e-02 -5.30989885e-01 1.18040174e-01
6.89880550e-01 2.48826016e-02 -6.72498122e-02 -5.03247797e-01
-1.22598507e-01 5.02728462e-01 -3.90210986e-01 2.36313462e-01
9.14904535e-01 -1.93343073e-01 1.21493198e-01 7.69215882e-01
9.15927649e-01 -2.72592217e-01 -1.47153032e+00 -6.54441953e-01
-2.96337932e-01 -4.01078165e-01 1.12445056e-01 -1.16103947e+00
-6.54292047e-01 7.54768312e-01 7.22215712e-01 -1.67593956e-01
1.00082624e+00 3.51129174e-01 5.91780007e-01 3.91609997e-01
5.07334232e-01 -8.89042497e-01 2.33592123e-01 7.16046810e-01
7.21872866e-01 -1.74213612e+00 -4.13237959e-01 -7.88271010e-01
-8.51842701e-01 9.12926316e-01 1.17783487e+00 2.15092838e-01
4.67176616e-01 1.05362311e-01 3.50321621e-01 -9.20406058e-02
-1.03381789e+00 -3.94122928e-01 3.73227865e-01 6.52922928e-01
4.53259014e-02 2.77225763e-01 -2.61639267e-01 1.00182915e+00
2.70265434e-03 9.19934828e-03 5.07471740e-01 3.62455070e-01
-5.01701415e-01 -8.00713181e-01 -5.36999226e-01 8.85106742e-01
-4.40574139e-02 -4.19450134e-01 -3.63458157e-01 2.79086828e-01
2.94173211e-01 1.08613122e+00 6.41522110e-02 -5.84293902e-01
2.34910578e-01 1.42873302e-01 2.63174236e-01 -1.01068485e+00
-2.48939171e-01 -1.37646750e-01 -3.69385816e-02 -5.38472235e-01
-1.26910349e-02 -3.07865322e-01 -1.06699562e+00 3.95376295e-01
-8.32597256e-01 -1.00605838e-01 2.64990717e-01 1.29976499e+00
1.99453592e-01 5.86548507e-01 7.45124102e-01 -3.20433021e-01
-1.14297724e+00 -9.95331526e-01 -4.90658909e-01 4.53762352e-01
3.53391558e-01 -8.71247888e-01 -4.99938160e-01 1.05279140e-01] | [9.942891120910645, 1.9193633794784546] |
144f8a2d-e6d7-4bb9-b904-971b6c48a45e | giraffe-using-deep-reinforcement-learning-to | 1509.01549 | null | http://arxiv.org/abs/1509.01549v2 | http://arxiv.org/pdf/1509.01549v2.pdf | Giraffe: Using Deep Reinforcement Learning to Play Chess | This report presents Giraffe, a chess engine that uses self-play to discover
all its domain-specific knowledge, with minimal hand-crafted knowledge given by
the programmer. Unlike previous attempts using machine learning only to perform
parameter-tuning on hand-crafted evaluation functions, Giraffe's learning
system also performs automatic feature extraction and pattern recognition. The
trained evaluation function performs comparably to the evaluation functions of
state-of-the-art chess engines - all of which containing thousands of lines of
carefully hand-crafted pattern recognizers, tuned over many years by both
computer chess experts and human chess masters. Giraffe is the most successful
attempt thus far at using end-to-end machine learning to play chess. | ['Matthew Lai'] | 2015-09-04 | null | null | null | null | ['game-of-chess'] | ['playing-games'] | [-5.29190660e-01 -2.36439884e-01 -8.92157927e-02 -2.66688138e-01
-6.85709894e-01 -1.19442475e+00 3.57754737e-01 1.49827033e-01
-5.19487500e-01 7.17398822e-01 -4.33161706e-01 -9.47016120e-01
-1.49882510e-01 -9.40700054e-01 -5.32461703e-01 2.80245114e-02
-1.92051664e-01 7.08562493e-01 8.04654002e-01 -8.23249936e-01
7.68656969e-01 2.76150823e-01 -1.50092041e+00 8.12359571e-01
4.49113369e-01 7.55641699e-01 -1.90719590e-01 1.18310523e+00
2.01990888e-01 1.17537344e+00 -9.64461625e-01 -7.40553081e-01
5.96656919e-01 -7.05946833e-02 -9.33631182e-01 -5.08745134e-01
2.58807451e-01 2.99255289e-02 -2.66683519e-01 6.29797101e-01
2.86693037e-01 -9.54085961e-02 4.82244343e-01 -1.14030910e+00
-3.10876101e-01 5.14044702e-01 -1.59812972e-01 5.29681444e-01
8.26214671e-01 6.61281824e-01 1.17640269e+00 -4.46015358e-01
2.48460203e-01 5.96655965e-01 1.02734578e+00 2.89416671e-01
-9.07062769e-01 -8.41711342e-01 -5.50701261e-01 2.29565501e-01
-1.47496939e+00 -4.23687622e-02 3.41347754e-01 -6.82802618e-01
1.68867922e+00 3.28439146e-01 9.59998548e-01 4.48818386e-01
4.49952722e-01 8.17356110e-01 1.14359736e+00 -6.91062808e-01
3.83892000e-01 1.16976015e-01 6.45501390e-02 1.02530169e+00
5.69331683e-02 5.88735521e-01 -4.40345526e-01 -2.98049599e-01
9.96574640e-01 -6.83240294e-01 2.67839938e-01 -3.08375508e-01
-1.08970761e+00 6.38137758e-01 -1.19127668e-01 4.07242686e-01
-1.06583491e-01 -2.95396503e-02 7.00674474e-01 8.32792640e-01
-6.69370964e-02 1.61644697e+00 -1.09928834e+00 -8.97379160e-01
-1.48061168e+00 5.58973372e-01 1.35089922e+00 8.77725780e-01
5.28349161e-01 1.20747462e-01 9.12503749e-02 3.35495383e-01
-1.80125147e-01 -3.51242907e-02 6.50600493e-01 -6.97644174e-01
3.91883552e-01 9.72642601e-01 2.43013099e-01 -9.85745609e-01
-2.88727760e-01 -5.17733991e-01 -1.66831225e-01 9.96933103e-01
6.45283818e-01 -5.44391930e-01 -5.14582813e-01 9.53245342e-01
-8.77046585e-03 -9.58671942e-02 1.95976064e-01 6.89197242e-01
6.58494473e-01 4.14277464e-01 -1.34908304e-01 2.96405345e-01
1.26671886e+00 -1.22631872e+00 3.35957468e-01 -5.51827140e-02
5.72844386e-01 -8.32795322e-01 1.21827710e+00 9.58137929e-01
-1.36206818e+00 -6.98580682e-01 -1.58563614e+00 2.75286853e-01
-4.54955578e-01 2.37218559e-01 9.61604655e-01 9.50418353e-01
-7.92846680e-01 7.96311796e-01 -7.13083625e-01 2.05017135e-01
1.96040437e-01 8.27803493e-01 -2.26835668e-01 4.43327993e-01
-1.00049150e+00 9.88954663e-01 7.64253199e-01 -6.53870523e-01
-9.69921172e-01 -8.98865402e-01 -5.54858983e-01 8.59503448e-02
6.64727449e-01 -7.69542396e-01 1.79461479e+00 -1.01246023e+00
-1.70304608e+00 1.00564170e+00 4.13330376e-01 -5.11621833e-01
4.33877945e-01 1.37125820e-01 -4.78460073e-01 -1.99550048e-01
6.68086251e-03 8.16590488e-02 9.07443464e-01 -5.81208169e-01
-1.03307962e+00 1.37896791e-01 4.75981891e-01 3.53781059e-02
-7.46884048e-02 2.54871368e-01 -3.61916751e-01 -6.24958277e-01
-3.97698134e-01 -4.80880558e-01 -1.42432541e-01 -6.80876791e-01
-2.52162963e-01 -3.62357914e-01 2.38691941e-01 -4.35290068e-01
1.50686932e+00 -1.82828450e+00 -1.80210933e-01 3.79697621e-01
1.66157782e-01 6.98881030e-01 -1.02755301e-01 4.01060849e-01
-3.61505896e-01 -1.36893839e-01 3.51118982e-01 5.73747814e-01
2.17897952e-01 -1.98802814e-01 8.66475236e-03 1.93939254e-01
-8.31904858e-02 8.94436359e-01 -1.20207977e+00 -3.23561043e-01
9.87482518e-02 -4.00352180e-01 -6.58896208e-01 5.31448066e-01
-5.54957628e-01 6.31286949e-02 -5.25174320e-01 8.19853365e-01
6.06185459e-02 -2.48738125e-01 4.89519984e-02 1.97815776e-01
-3.46147418e-01 4.79418099e-01 -1.18503356e+00 1.44118536e+00
-5.04097044e-01 5.10521054e-01 -4.23526317e-01 -7.18150973e-01
1.07507765e+00 1.97734877e-01 1.02274060e-01 -6.18290424e-01
1.44589424e-01 2.60296136e-01 2.99034506e-01 -5.43490350e-01
6.45122170e-01 -1.49738103e-01 -3.82069081e-01 7.14567840e-01
1.72564447e-01 -3.70203733e-01 6.38164759e-01 1.88479677e-01
1.67361200e+00 1.68124199e-01 7.70266473e-01 -2.21998394e-01
6.27362192e-01 8.38794947e-01 1.73915714e-01 9.71576810e-01
1.76174253e-01 2.98881501e-01 4.52004522e-01 -1.21712339e+00
-1.58256066e+00 -5.93873322e-01 2.63274670e-01 1.60128975e+00
-1.77500710e-01 -9.43553329e-01 -9.35520947e-01 -5.89075327e-01
-1.63612798e-01 8.02805603e-01 -4.91893500e-01 -1.21508609e-03
-4.15093005e-01 -6.50649190e-01 9.87875104e-01 7.23272443e-01
6.60624504e-01 -1.28764713e+00 -1.02986324e+00 4.40413088e-01
6.28881991e-01 -2.98969179e-01 -3.24014723e-01 2.43891597e-01
-4.69733894e-01 -1.53310013e+00 -3.03908378e-01 -8.12152982e-01
2.23674133e-01 -2.63976753e-01 1.69181240e+00 4.35358495e-01
-6.77573442e-01 4.39364642e-01 -2.97054082e-01 -3.70854825e-01
-6.81343198e-01 6.85459614e-01 -2.36418962e-01 -7.86972523e-01
6.50753558e-01 -7.40298510e-01 -2.17014477e-01 4.86181200e-01
-4.13070381e-01 6.64643273e-02 8.67222250e-01 9.92587328e-01
1.66793466e-01 7.22332656e-01 -6.14047013e-02 -9.53435540e-01
9.10951436e-01 -1.74040794e-01 -1.10824335e+00 5.19788742e-01
-6.72301769e-01 1.11412935e-01 9.21071768e-01 -4.36818331e-01
-6.83366716e-01 1.43029511e-01 -5.57113737e-02 4.52332804e-03
-4.66728628e-01 3.63075733e-01 -1.35155441e-02 -5.71838021e-01
1.29886985e+00 3.50650132e-01 -3.82856667e-01 -2.10672483e-01
9.37622041e-02 4.62940335e-01 9.57853854e-01 -1.09412313e+00
9.32089269e-01 -5.62634230e-01 -1.85338095e-01 -4.55648422e-01
-5.70992351e-01 -5.33022881e-01 -6.44042373e-01 -2.19619293e-02
2.63332069e-01 -7.29927123e-01 -1.17378855e+00 5.54457963e-01
-8.79356921e-01 -6.45950317e-01 -3.39417249e-01 -7.28999451e-02
-8.05599391e-01 5.62947337e-03 -5.20290494e-01 -6.08242750e-01
-3.21586251e-01 -8.28906476e-01 6.57181561e-01 3.79848093e-01
-6.84469283e-01 -9.26412523e-01 6.71137154e-01 4.72284913e-01
3.13669741e-01 8.93164799e-02 1.06501937e+00 -9.49377954e-01
-2.81506270e-01 -6.19785428e-01 -4.15513851e-02 2.41135776e-01
-2.29322329e-01 1.16176084e-02 -5.70551693e-01 -2.62204379e-01
-3.56300771e-01 -4.99969065e-01 2.83598572e-01 -1.93924993e-01
8.03264618e-01 -2.46625841e-01 -8.53830129e-02 6.04210258e-01
1.35465562e+00 4.21602309e-01 7.05986857e-01 1.17323983e+00
2.22602963e-01 -1.50169078e-02 5.36326945e-01 5.59784710e-01
1.86197057e-01 4.57194686e-01 2.39997916e-02 1.07993215e-01
1.30577981e-01 -4.12411660e-01 1.02157697e-01 2.39045292e-01
-5.64620614e-01 5.54137453e-02 -1.25631583e+00 5.23312747e-01
-1.70746040e+00 -8.94990742e-01 2.34221309e-01 1.92572904e+00
1.11225915e+00 5.99366844e-01 6.12703502e-01 4.23763931e-01
3.91903162e-01 -1.77782908e-01 -5.49762428e-01 -6.13607168e-01
2.47932032e-01 7.91668653e-01 5.67212880e-01 2.47176692e-01
-1.18783033e+00 1.27880204e+00 7.95132637e+00 1.23346388e+00
-8.05803001e-01 -4.11005139e-01 2.45715469e-01 -8.57780948e-02
2.28338853e-01 2.24964455e-01 -7.88701594e-01 2.52284437e-01
9.96906042e-01 -2.75247008e-01 8.60582948e-01 1.18469703e+00
-4.08087254e-01 -2.31184110e-01 -1.26954591e+00 9.18250322e-01
-5.41302972e-02 -1.63271236e+00 -3.50200951e-01 -3.16243708e-01
7.42780507e-01 -4.18220669e-01 1.06365375e-01 7.59214938e-01
9.58193898e-01 -1.36113262e+00 7.47290015e-01 3.86192888e-01
6.94745898e-01 -9.23887908e-01 6.60921991e-01 4.60565448e-01
-1.02681386e+00 -2.18598321e-01 -3.03598851e-01 -6.69549048e-01
-5.05096495e-01 8.11138153e-02 -1.29132009e+00 2.40478724e-01
4.08083469e-01 3.13006908e-01 -9.58565176e-01 1.27198708e+00
-4.88848984e-01 8.91307533e-01 -2.60173976e-01 -4.97868866e-01
4.80587274e-01 3.24237615e-01 3.90427887e-01 1.43679810e+00
-9.33990553e-02 3.62700611e-01 1.45515591e-01 5.70477903e-01
3.59190732e-01 6.14164770e-03 1.95224676e-02 -2.48862594e-01
3.10822457e-01 9.97561872e-01 -6.71301484e-01 -5.09245515e-01
-2.99747407e-01 6.98529601e-01 3.73109698e-01 5.28019480e-02
-5.21903396e-01 -7.98456311e-01 2.77864665e-01 4.31268841e-01
6.80875003e-01 -2.74097413e-01 -5.60435951e-01 -9.63681221e-01
-3.05631876e-01 -1.51089096e+00 4.74457115e-01 -7.52957046e-01
-1.30493736e+00 6.19927049e-01 -1.35797888e-01 -9.72534180e-01
-6.85414672e-01 -1.07068610e+00 -1.08961880e+00 7.18359172e-01
-6.60084009e-01 -1.02569449e+00 1.83013856e-01 5.85764945e-01
4.37373817e-01 -1.30091596e+00 8.98769915e-01 -8.98084939e-02
-2.86876619e-01 9.52102065e-01 -5.08866943e-02 5.11000454e-01
3.92744809e-01 -1.58038437e+00 7.24846482e-01 3.99071217e-01
4.19169217e-01 6.70102060e-01 9.64732647e-01 -5.57997108e-01
-1.60381603e+00 -5.50789356e-01 4.59209412e-01 -5.51269054e-01
8.57700467e-01 -5.84196299e-02 -6.01505935e-01 6.54654741e-01
1.86174572e-01 -2.12835968e-01 8.53532791e-01 5.72490454e-01
-6.50632918e-01 -2.06086989e-02 -1.01283622e+00 4.02805150e-01
5.50174952e-01 -4.69853133e-01 -1.28953636e+00 2.34006152e-01
-3.82985845e-02 -8.05706859e-01 -8.25110853e-01 1.99855715e-02
8.56556416e-01 -1.16873074e+00 9.56824899e-01 -1.29837000e+00
4.89198804e-01 -4.61991131e-01 6.65407702e-02 -1.22699463e+00
-5.83192587e-01 -8.01686585e-01 2.31829658e-01 7.11759627e-01
8.29734445e-01 -1.66206360e-01 1.28014004e+00 6.67572856e-01
-4.00903858e-02 -8.56136501e-01 -3.95451158e-01 -9.20264363e-01
2.52302676e-01 -4.65455353e-01 8.19096208e-01 9.16570127e-01
6.03126526e-01 4.90878701e-01 -2.40088701e-01 -3.33378389e-02
3.03241014e-01 5.13358831e-01 1.30096447e+00 -1.15407312e+00
-1.18198740e+00 -6.31116748e-01 -7.60230362e-01 -8.62954080e-01
-7.78315142e-02 -7.07485616e-01 -4.72926795e-02 -7.72962153e-01
3.41781136e-03 -3.72709364e-01 -2.14527830e-01 9.33194697e-01
2.06213146e-01 2.04490080e-01 -3.32746990e-02 -5.49692363e-02
-1.22312820e+00 -3.59612942e-01 1.01943839e+00 -1.86544389e-01
-2.58794099e-01 2.82366306e-01 -7.68552780e-01 1.03460312e+00
7.88242936e-01 -4.80768174e-01 -2.91560590e-01 -1.15448788e-01
6.75723553e-01 3.63774300e-02 1.04770347e-01 -1.40410674e+00
8.31835687e-01 -4.54020798e-01 6.36446953e-01 -2.59211928e-01
-9.18903649e-02 -4.03715789e-01 1.39134511e-01 3.51553053e-01
-2.30404884e-01 1.89438954e-01 5.79775929e-01 -1.29237160e-01
-1.84118956e-01 -7.12848425e-01 5.87415874e-01 -7.04897106e-01
-1.17254913e+00 -1.79438442e-02 -7.69323170e-01 4.11719263e-01
1.24768531e+00 -3.16727698e-01 -6.93578869e-02 -1.71894789e-01
-4.97406095e-01 -7.97836259e-02 4.86244678e-01 -9.77656096e-02
3.47699463e-01 -6.89231396e-01 -6.70612752e-01 3.79509896e-01
-9.53911319e-02 -3.07906002e-01 -1.01859830e-01 9.26748216e-02
-1.05235946e+00 4.48046416e-01 -6.22417212e-01 -7.91714564e-02
-1.28934443e+00 6.20540857e-01 3.47591221e-01 -7.51121879e-01
-4.30025488e-01 1.06234050e+00 -5.05958259e-01 -4.04361784e-01
1.23814745e-02 -1.38453528e-01 1.92288980e-01 -4.77820933e-01
8.11858177e-01 3.10645670e-01 3.20059180e-01 1.10137373e-01
-2.48864353e-01 4.49098259e-01 -1.97271600e-01 -9.47675481e-02
1.41564476e+00 8.58181000e-01 1.25832945e-01 1.19134493e-01
5.48952758e-01 1.63922772e-01 -9.22526240e-01 1.97112232e-01
1.36752278e-01 -4.14511532e-01 -1.75839197e-02 -1.60615683e+00
-6.94775283e-01 6.34564817e-01 -7.86363613e-03 2.38264110e-02
9.87222552e-01 -3.93438697e-01 6.90609992e-01 1.02169836e+00
8.65589142e-01 -1.03738725e+00 2.14699686e-01 7.12631166e-01
5.71351171e-01 -9.37394559e-01 2.41089657e-01 5.29582798e-02
-6.26486599e-01 1.49523914e+00 6.50187731e-01 -7.36203015e-01
4.33366895e-01 8.88320446e-01 -1.73665419e-01 -2.30708405e-01
-9.67934251e-01 1.68740541e-01 3.71944129e-01 5.37487566e-01
2.57441491e-01 1.58503816e-01 2.52943151e-02 1.24955046e+00
-1.05754554e+00 4.52801585e-01 2.67583162e-01 1.19325888e+00
-6.53232217e-01 -1.42277968e+00 -6.05039537e-01 5.16508043e-01
-2.74378896e-01 -3.61159563e-01 -5.29798865e-01 1.06634104e+00
4.90784079e-01 5.29388607e-01 -1.82732075e-01 -8.42008471e-01
5.72332978e-01 1.56958237e-01 9.01878417e-01 -6.48009479e-01
-1.41800404e+00 -4.24274266e-01 3.67931783e-01 -3.94946247e-01
1.84212014e-01 -5.12277961e-01 -9.31491137e-01 -6.63048089e-01
3.75043526e-02 7.09818780e-01 2.84617245e-01 9.87065852e-01
-2.65043695e-02 3.25248003e-01 3.70759189e-01 -3.89402032e-01
-6.25882626e-01 -5.73814929e-01 -7.63110876e-01 -4.69078077e-03
-1.26105040e-01 -4.57331449e-01 1.54376328e-02 5.39780129e-03] | [3.461700677871704, 1.4481619596481323] |
cf0c51e3-516e-4419-81cd-97b7e93c5ad0 | learning-to-exploit-the-sequence-specific | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Qin_Learning_To_Exploit_the_Sequence-Specific_Prior_Knowledge_for_Image_Processing_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Qin_Learning_To_Exploit_the_Sequence-Specific_Prior_Knowledge_for_Image_Processing_CVPR_2023_paper.pdf | Learning To Exploit the Sequence-Specific Prior Knowledge for Image Processing Pipelines Optimization | The hardware image signal processing (ISP) pipeline is the intermediate layer between the imaging sensor and the downstream application, processing the sensor signal into an RGB image. The ISP is less programmable and consists of a series of processing modules. Each processing module handles a subtask and contains a set of tunable hyperparameters. A large number of hyperparameters form a complex mapping with the ISP output. The industry typically relies on manual and time-consuming hyperparameter tuning by image experts, biased towards human perception. Recently, several automatic ISP hyperparameter optimization methods using downstream evaluation metrics come into sight. However, existing methods for ISP tuning treat the high-dimensional parameter space as a global space for optimization and prediction all at once without inducing the structure knowledge of ISP. To this end, we propose a sequential ISP hyperparameter prediction framework that utilizes the sequential relationship within ISP modules and the similarity among parameters to guide the model sequence process. We validate the proposed method on object detection, image segmentation, and image quality tasks. | ['Weiming Hu', 'Bing Li', 'Wentao Ma', 'Juan Wang', 'Weihua Xiong', 'Longfei Han', 'Haina Qin'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['hyperparameter-optimization'] | ['methodology'] | [ 4.85694915e-01 -3.11476678e-01 1.90371778e-02 -5.70672929e-01
-7.50650346e-01 -7.03209817e-01 6.96714297e-02 -1.06387705e-01
-4.40226972e-01 -7.57618546e-02 -2.39375040e-01 -2.02810004e-01
-1.27596289e-01 -4.95398223e-01 -5.06832600e-01 -9.22599077e-01
2.92457819e-01 2.68081933e-01 5.46500206e-01 1.18402079e-01
4.49115574e-01 4.38994437e-01 -1.37881029e+00 -4.83944044e-02
8.41804326e-01 1.54380178e+00 3.45493048e-01 8.61849606e-01
8.80406275e-02 1.18710726e-01 -4.83770221e-01 -2.27492258e-01
6.02749109e-01 -2.15397581e-01 -3.64732742e-01 3.04854363e-01
-1.05637141e-01 -9.72725078e-02 2.53024250e-01 1.27395940e+00
6.33984685e-01 -1.14185564e-01 3.96946192e-01 -1.45934749e+00
-2.42602274e-01 6.83597863e-01 -5.83124459e-01 7.72409737e-02
-6.72318488e-02 7.97678530e-01 7.84665585e-01 -7.05880880e-01
-2.35179663e-02 9.31467891e-01 5.03340900e-01 -2.26055197e-02
-1.12667477e+00 -7.23648429e-01 1.47206292e-01 3.12924892e-01
-1.33990765e+00 -2.05056265e-01 8.75468731e-01 -5.81354082e-01
7.18622565e-01 1.42254233e-01 7.83897161e-01 6.18509650e-01
1.85081705e-01 4.76793438e-01 8.44612718e-01 -9.20908228e-02
5.01980960e-01 2.39448398e-01 4.48391259e-01 4.42565262e-01
1.07026882e-01 -1.42849371e-01 -5.13086200e-01 -7.58165270e-02
8.77277851e-01 -3.34023356e-01 -3.32201868e-01 -2.83638239e-01
-9.84314144e-01 3.83491218e-01 3.52277577e-01 -1.42616928e-01
-3.63028407e-01 2.07648352e-01 1.70206562e-01 9.58399847e-02
-3.44802976e-01 5.09372950e-01 -7.20006883e-01 -1.35207847e-01
-6.69426143e-01 -2.61441857e-01 6.58674598e-01 1.05628312e+00
9.08332348e-01 -2.43832439e-01 -6.01389036e-02 7.12363482e-01
5.84414840e-01 3.18911165e-01 5.56777775e-01 -9.59244490e-01
4.74434346e-01 9.15626168e-01 1.82302266e-01 -1.08813524e+00
-6.55150771e-01 -3.42821807e-01 -6.38847709e-01 9.29507539e-02
3.02000970e-01 -1.52629152e-01 -8.72654974e-01 1.41203916e+00
5.62903583e-01 2.15712145e-01 -2.72243589e-01 1.33029079e+00
4.72104371e-01 7.76487291e-01 2.36679539e-01 -2.20033020e-01
1.48103023e+00 -8.68247509e-01 -4.08985734e-01 -3.69768441e-01
4.65182215e-02 -8.90826941e-01 1.57694209e+00 6.04040027e-01
-1.05031645e+00 -6.78196728e-01 -1.40846372e+00 7.25056529e-02
-1.00137673e-01 4.79559064e-01 5.09369731e-01 7.50519812e-01
-8.71076405e-01 3.25246811e-01 -9.25677538e-01 -1.26055062e-01
1.22231178e-01 7.26785898e-01 3.29209059e-01 3.21852714e-01
-8.68744731e-01 4.82925922e-01 5.00096023e-01 5.88193059e-01
-6.89985216e-01 -7.80246139e-01 -4.42594439e-01 -1.99059839e-04
4.33140934e-01 -7.71697640e-01 1.18412256e+00 -1.03576863e+00
-2.21567130e+00 6.17845416e-01 2.82240450e-01 -4.39879782e-02
4.63529408e-01 -1.64834827e-01 -2.56564558e-01 1.23381605e-02
-4.21315491e-01 4.87055421e-01 1.36938035e+00 -1.16374516e+00
-6.21394038e-01 -4.58475292e-01 -2.14854032e-01 3.41224521e-01
-2.90403098e-01 1.20438963e-01 -1.11887121e+00 -3.21029574e-01
4.12426919e-01 -9.08559084e-01 -3.47748220e-01 -1.20198384e-01
-5.42041421e-01 1.16006568e-01 5.79049170e-01 -2.54976720e-01
1.30778468e+00 -2.32074451e+00 1.15137026e-01 5.95432639e-01
-1.19340211e-01 1.48722231e-01 -1.47319809e-02 -1.62751496e-01
5.94571605e-02 2.41388138e-02 -1.69591993e-01 3.79192531e-02
9.61745754e-02 -1.21858701e-01 -1.33833230e-01 4.99655366e-01
2.16286927e-01 6.82393491e-01 -6.31295264e-01 -7.03236520e-01
3.11409414e-01 2.79223055e-01 -6.24727190e-01 4.35538799e-01
-2.36518741e-01 5.50383806e-01 -7.68680751e-01 7.26022899e-01
7.13409901e-01 -5.81355631e-01 1.18953578e-01 -8.47322702e-01
-2.12780491e-01 -4.19546515e-02 -1.55982411e+00 1.60441196e+00
-2.44841725e-01 1.91487461e-01 3.45746756e-01 -7.83235490e-01
9.58326399e-01 1.64166074e-02 6.48444057e-01 -3.68544847e-01
7.85259247e-01 1.37604266e-01 -8.44334960e-02 -8.05153251e-01
3.52247477e-01 2.81892866e-01 -8.11586380e-02 4.01907891e-01
-1.28845841e-01 -3.16361994e-01 -1.73328875e-03 -4.45250005e-01
9.28388000e-01 2.44536519e-01 2.29589805e-01 2.49061361e-02
5.90665817e-01 1.84389606e-01 7.85811663e-01 6.16900086e-01
-3.90530854e-01 5.96599996e-01 4.86121684e-01 -1.40998483e-01
-1.03357685e+00 -1.02790248e+00 -1.93925098e-01 9.92230654e-01
5.24279773e-01 -2.01014742e-01 -8.38647783e-01 -2.84299143e-02
-2.75729716e-01 2.13095516e-01 -2.28863999e-01 -2.32255861e-01
-5.88121772e-01 -1.11102796e+00 2.47665614e-01 4.35843766e-01
7.05568612e-01 -7.85810828e-01 -1.00173151e+00 1.75408930e-01
2.25091249e-01 -1.36227262e+00 -6.03209078e-01 3.01715881e-01
-8.97666454e-01 -1.02593398e+00 -1.17485784e-01 -6.14732444e-01
6.32606685e-01 -8.29911456e-02 9.08071160e-01 -4.66707051e-02
-5.48960745e-01 2.49778807e-01 -1.96895257e-01 -3.31844836e-01
-1.74761340e-01 -9.16389450e-02 -1.65405497e-01 3.68863523e-01
1.27844170e-01 -6.05404496e-01 -9.91979063e-01 7.46343851e-01
-8.69967937e-01 2.29227558e-01 8.51578653e-01 5.32901704e-01
1.04395366e+00 1.19548492e-01 5.40162548e-02 -5.38976133e-01
4.71417695e-01 -1.74800694e-01 -1.34617472e+00 3.79461110e-01
-5.34835637e-01 9.25826654e-02 6.18807852e-01 -6.38708711e-01
-9.38637853e-01 7.72651970e-01 3.37712854e-01 -5.29471219e-01
6.14724271e-02 2.82628864e-01 -6.63739860e-01 -2.37793028e-01
4.89541918e-01 -2.50865277e-02 -1.04989186e-01 -2.59517759e-01
4.59051251e-01 6.91377342e-01 6.89236581e-01 -4.39707041e-01
6.66434646e-01 1.34565145e-01 -6.77418010e-03 -5.94777465e-01
-4.59083408e-01 -5.02300858e-01 -5.55185497e-01 -5.12675941e-01
9.70338106e-01 -6.76236868e-01 -1.13544047e+00 6.01771235e-01
-9.99905407e-01 -2.75335342e-01 1.16743799e-02 3.89115512e-01
-4.86278087e-01 1.26125172e-01 -2.98964411e-01 -5.55604279e-01
-2.77379215e-01 -1.76223063e+00 1.13602352e+00 5.48831999e-01
1.25556499e-01 -3.70486498e-01 -2.42057890e-01 4.96295810e-01
1.88276008e-01 -7.84066468e-02 8.70570958e-01 -1.42398685e-01
-9.17890012e-01 -1.59737304e-01 -3.37000161e-01 2.58957446e-01
-1.11690767e-01 3.82914901e-01 -9.96615231e-01 3.00097242e-02
2.76191980e-01 -5.17835803e-02 4.44332570e-01 5.97648263e-01
1.62268710e+00 -1.33008575e-02 -8.27258080e-03 1.07412302e+00
1.56386936e+00 3.04760635e-01 4.65663105e-01 4.51453865e-01
6.78053141e-01 4.62618798e-01 6.77626550e-01 5.93371093e-01
2.56949067e-01 6.89625144e-01 4.93005306e-01 1.03870109e-01
2.09272340e-01 4.85922992e-02 2.60454535e-01 6.85788333e-01
2.05688894e-01 -1.75433889e-01 -1.16518843e+00 -6.75198901e-03
-1.75461411e+00 -3.59879315e-01 -4.98768128e-02 2.12715936e+00
9.28381324e-01 2.11522400e-01 5.19461147e-02 9.18315873e-02
8.21450830e-01 -1.98361471e-01 -1.03609776e+00 -9.17994678e-02
1.59564093e-01 -2.23928630e-01 8.31726253e-01 2.66780347e-01
-1.03807545e+00 8.52622449e-01 6.07473373e+00 4.73048449e-01
-1.39495206e+00 -3.61158222e-01 6.38460338e-01 -1.37054548e-01
-4.20950074e-03 4.62696739e-02 -7.91062772e-01 5.97921848e-01
7.16150880e-01 7.29981214e-02 4.99469608e-01 7.44973898e-01
5.21419764e-01 -3.00372392e-01 -1.05287158e+00 1.52383542e+00
-3.38238150e-01 -9.46659088e-01 -5.78356497e-02 -1.84802711e-01
4.45591211e-01 -1.25745267e-01 3.71832728e-01 -2.65178680e-01
1.44736677e-01 -8.38694632e-01 8.04932952e-01 5.28189063e-01
3.49366993e-01 -5.74839652e-01 4.05398935e-01 1.72654107e-01
-1.11881959e+00 -5.40147483e-01 -3.22589874e-01 4.46012408e-01
1.27228856e-01 5.50358713e-01 -8.15265954e-01 1.48142532e-01
8.97106171e-01 2.65673876e-01 -6.60198987e-01 1.28761709e+00
-2.75108576e-01 5.65950871e-01 -4.66370344e-01 -5.36983125e-02
-3.48430835e-02 -5.75303674e-01 5.20537496e-01 9.86700296e-01
2.17879027e-01 3.91994774e-01 1.74762264e-01 8.22453141e-01
2.65187770e-01 1.26651466e-01 1.06309891e-01 -6.56471923e-02
5.84329605e-01 1.54090869e+00 -9.05412436e-01 1.66028365e-02
-1.98187307e-01 5.95314920e-01 -9.75928009e-02 3.59553635e-01
-1.09056616e+00 -2.31401309e-01 6.55625403e-01 -2.04951480e-01
2.46148646e-01 -3.24288547e-01 -8.67001355e-01 -7.86054313e-01
1.51938662e-01 -9.25360739e-01 3.16878170e-01 -8.40659916e-01
-9.72519100e-01 4.54820931e-01 -3.60602178e-02 -1.50712705e+00
-5.81602082e-02 -8.66001308e-01 -5.39834857e-01 5.89332283e-01
-1.40853882e+00 -8.74171138e-01 -8.11769009e-01 6.91442728e-01
4.20714200e-01 9.32928473e-02 2.86647260e-01 1.67138234e-01
-1.29057193e+00 5.54658413e-01 -2.15177462e-01 -1.95837706e-01
5.71587205e-01 -9.26635265e-01 -7.89723173e-02 9.53327477e-01
-4.74401295e-01 3.94998670e-01 8.09893191e-01 -2.99949676e-01
-1.79577136e+00 -9.69181061e-01 -2.50231335e-03 -5.53390794e-02
9.28123236e-01 -2.66655594e-01 -7.53888071e-01 2.59058654e-01
-1.21266693e-01 -7.00806454e-02 5.66712737e-01 -3.43873411e-01
-7.38796964e-02 -5.66183448e-01 -8.76995385e-01 6.16423368e-01
7.41040945e-01 -2.36130118e-01 -4.30402979e-02 1.47864968e-01
6.49792492e-01 -4.64977831e-01 -9.43170190e-01 3.90542388e-01
4.59479690e-01 -7.08261728e-01 1.01111066e+00 -4.16262895e-02
8.87125954e-02 -7.49272287e-01 -5.31575792e-02 -8.73367846e-01
-1.95215344e-01 -9.15601194e-01 3.01806629e-01 1.08062184e+00
5.39273977e-01 -3.26139152e-01 7.54451931e-01 1.07753289e+00
-2.61207968e-01 -6.45024121e-01 -6.18049324e-01 -4.62770313e-01
-6.12488091e-01 -8.05479586e-01 7.90267348e-01 2.98227727e-01
-2.50337958e-01 3.49754721e-01 1.50605857e-01 8.26875865e-01
8.44148099e-01 2.21532628e-01 7.11990833e-01 -9.27617788e-01
-6.07749343e-01 -7.15355217e-01 -4.98740882e-01 -1.03008723e+00
-4.33120310e-01 -4.47935015e-01 4.05151099e-01 -9.02912736e-01
-3.24135870e-02 -6.02963805e-01 -3.32680404e-01 3.00351501e-01
-2.43684292e-01 7.22204149e-02 9.86962691e-02 4.65857089e-01
-7.18241572e-01 2.24425703e-01 1.19431460e+00 -8.51164460e-02
-6.81511998e-01 1.70118064e-01 -5.14650583e-01 8.83337915e-01
9.12272274e-01 -2.96068549e-01 -7.01689601e-01 -6.60569012e-01
4.89431292e-01 -5.60690798e-02 4.03795749e-01 -9.82370317e-01
6.04592383e-01 -3.96395981e-01 3.99056315e-01 -4.50015098e-01
2.15730682e-01 -1.08803594e+00 2.91825950e-01 1.37022153e-01
-1.56078592e-01 -2.01214314e-03 -1.59995873e-02 4.78471071e-01
-1.18354820e-01 -1.14320047e-01 1.14822245e+00 1.18159063e-01
-8.43184531e-01 3.43033880e-01 -1.24430254e-01 -1.03886396e-01
1.12375379e+00 -5.99697113e-01 1.22867357e-02 5.70202060e-02
-7.52136707e-01 4.91869420e-01 3.42321932e-01 2.03233719e-01
5.65719664e-01 -8.95500481e-01 -1.83654189e-01 5.38047731e-01
-9.93898325e-03 3.21127445e-01 1.32624075e-01 9.15648639e-01
-4.57138062e-01 6.33018138e-03 -1.46158889e-01 -1.06733990e+00
-1.07513905e+00 4.33517456e-01 5.99656165e-01 -4.65389974e-02
-4.00975138e-01 8.91794801e-01 4.55102623e-02 -6.98571503e-02
5.54606259e-01 -6.54353797e-01 -3.01241845e-01 2.05431581e-02
3.92921358e-01 3.69253844e-01 1.63183600e-01 -3.20390582e-01
-2.07760587e-01 1.02835178e+00 3.68928999e-01 -1.37796685e-01
1.38616872e+00 -5.22005618e-01 -1.04465477e-01 3.71259272e-01
1.23576307e+00 -5.73343754e-01 -1.65002370e+00 4.35021985e-03
1.17028356e-02 -3.24040502e-01 3.80010098e-01 -6.77769780e-01
-1.20016742e+00 5.14258564e-01 9.71335590e-01 3.18016484e-02
1.74870360e+00 -7.86222890e-02 6.10167384e-01 2.83454120e-01
2.62495577e-01 -1.50910246e+00 1.18139781e-01 2.22373396e-01
7.68621922e-01 -1.16907454e+00 -1.09670591e-02 -6.25632763e-01
-7.49263227e-01 1.13271320e+00 7.58499086e-01 -6.43173698e-03
9.34311092e-01 5.32170475e-01 2.27842450e-01 -1.83547899e-01
-5.69400966e-01 1.72131494e-01 3.68536502e-01 4.16278988e-01
-4.02964838e-02 -1.33258149e-01 -9.62995887e-02 8.10009539e-01
-3.99132073e-01 -8.41701054e-04 9.05788615e-02 6.38800263e-01
-6.11454248e-01 -9.79149640e-01 -5.31010389e-01 2.50371903e-01
-1.38366312e-01 2.10357249e-01 -9.24912170e-02 5.25741160e-01
1.78377077e-01 8.57539356e-01 5.72747141e-02 -5.86429656e-01
5.49610376e-01 -1.73284918e-01 2.53175765e-01 -2.18222767e-01
-8.21420312e-01 2.24439770e-01 -3.72759134e-01 -1.04697978e+00
-1.19752213e-01 -6.07660413e-01 -1.45795405e+00 1.54167175e-01
-2.73927599e-01 -3.33041161e-01 1.09689593e+00 7.34950125e-01
3.82603794e-01 4.71973240e-01 8.53374660e-01 -7.92501211e-01
-7.50074029e-01 -5.36012590e-01 -4.84563529e-01 2.39950806e-01
1.21217325e-01 -4.18514937e-01 -1.94348410e-01 2.35711798e-01] | [9.488543510437012, -0.3373841941356659] |
430ded44-6ca8-4625-96a3-c39477a76893 | constructing-multilingual-code-search-dataset | 2306.15604 | null | https://arxiv.org/abs/2306.15604v1 | https://arxiv.org/pdf/2306.15604v1.pdf | Constructing Multilingual Code Search Dataset Using Neural Machine Translation | Code search is a task to find programming codes that semantically match the given natural language queries. Even though some of the existing datasets for this task are multilingual on the programming language side, their query data are only in English. In this research, we create a multilingual code search dataset in four natural and four programming languages using a neural machine translation model. Using our dataset, we pre-train and fine-tune the Transformer-based models and then evaluate them on multiple code search test sets. Our results show that the model pre-trained with all natural and programming language data has performed best in most cases. By applying back-translation data filtering to our dataset, we demonstrate that the translation quality affects the model's performance to a certain extent, but the data size matters more. | ['Hitomi Yanaka', 'Shuai Lu', 'Nan Duan', 'Ryo Sekizawa'] | 2023-06-27 | null | null | null | null | ['code-search', 'code-search', 'machine-translation'] | ['computer-code', 'computer-vision', 'natural-language-processing'] | [-2.59426147e-01 -4.23479766e-01 -7.20514715e-01 -3.53520393e-01
-9.00991261e-01 -7.14966655e-01 5.09023726e-01 1.82316408e-01
-4.29024518e-01 1.12936750e-01 1.05235577e-01 -8.73480260e-01
1.87247694e-01 -7.66427815e-01 -8.06580663e-01 1.95822045e-01
1.88349530e-01 4.76851076e-01 4.34282511e-01 -4.58691925e-01
3.69110167e-01 -1.79928184e-01 -1.37916780e+00 5.25062919e-01
1.50271893e+00 3.65537614e-01 4.63911057e-01 4.68725771e-01
-4.78238285e-01 1.12870467e+00 -2.16907218e-01 -7.87176669e-01
3.15912217e-01 -2.06946120e-01 -1.11504161e+00 -7.76800573e-01
2.89693862e-01 7.42480308e-02 -9.66762453e-02 1.43638337e+00
2.31458098e-01 -4.10475999e-01 2.43013963e-01 -1.17533576e+00
-1.06123853e+00 7.99846172e-01 -2.98776418e-01 4.08561006e-02
7.10834026e-01 -4.91225384e-02 1.21985304e+00 -1.05467772e+00
8.70970547e-01 9.33090329e-01 8.38151753e-01 4.46609795e-01
-1.27593005e+00 -7.06974030e-01 -2.87488937e-01 -6.67453632e-02
-1.55003822e+00 -2.09398732e-01 6.36928320e-01 -7.67873824e-01
1.46304643e+00 1.57956973e-01 2.45031640e-01 7.75445342e-01
4.07402188e-01 6.53403580e-01 9.24998164e-01 -7.82093763e-01
-3.06786984e-01 4.08967108e-01 1.38508350e-01 1.05264378e+00
-1.15297377e-01 7.91814774e-02 -1.59819067e-01 -5.57713330e-01
2.35289901e-01 -7.55499899e-02 -1.83731243e-01 -4.41925228e-01
-1.20180368e+00 1.04065609e+00 3.83201122e-01 5.86817920e-01
1.45365223e-01 1.32840529e-01 7.97721982e-01 9.08252835e-01
2.57569671e-01 7.95127332e-01 -7.59772539e-01 -2.72563577e-01
-1.00201118e+00 1.20327942e-01 1.08641422e+00 1.42520893e+00
1.07826054e+00 -3.80515277e-01 -2.81578931e-03 1.29818225e+00
2.66386658e-01 5.82353652e-01 8.22437048e-01 -6.24290228e-01
8.29750597e-01 9.73871410e-01 -2.88227439e-01 -1.00276101e+00
-1.15501896e-01 -2.99335033e-01 -8.70072469e-02 -2.59651452e-01
3.39843452e-01 2.43380606e-01 -5.02114713e-01 1.75183272e+00
-3.21486026e-01 -7.52191722e-01 6.14050254e-02 6.57715976e-01
7.23023772e-01 4.67764616e-01 -1.84619296e-02 3.03453028e-01
1.24785411e+00 -1.30823529e+00 -2.15493932e-01 -4.82939631e-01
1.11119258e+00 -1.10429049e+00 1.71064019e+00 1.54779404e-01
-7.85950363e-01 -5.00575602e-01 -9.72551823e-01 -3.53596091e-01
-4.12431389e-01 3.77662033e-01 5.47839701e-01 5.88568747e-01
-1.23856795e+00 1.09551288e-01 -6.87917471e-01 -7.22495675e-01
-2.83792764e-01 3.27370949e-02 -2.75710762e-01 -3.43551219e-01
-1.18221974e+00 1.08471763e+00 3.58114004e-01 -6.08878493e-01
-7.23228097e-01 -5.25676429e-01 -6.67900920e-01 1.00510940e-01
1.55763537e-01 -4.40356553e-01 1.35520267e+00 -1.22772324e+00
-1.00266302e+00 1.24141049e+00 -2.76548892e-01 -2.21399397e-01
3.47622126e-01 2.64468528e-02 -6.02734089e-01 -2.87631154e-01
5.50500274e-01 5.25689006e-01 1.33062065e-01 -9.93669689e-01
-6.05847061e-01 -4.02243286e-02 4.60060120e-01 -2.23557666e-01
-3.48253250e-01 6.91355228e-01 -1.03292406e+00 -7.07184613e-01
-1.05232626e-01 -1.14631069e+00 -1.92172378e-01 -1.51416257e-01
-4.81693633e-02 -2.96656162e-01 2.36408621e-01 -7.48802304e-01
1.57965183e+00 -2.19977784e+00 2.28756547e-01 2.25329965e-01
-9.37927607e-03 -1.44803002e-01 -3.67286712e-01 7.52954900e-01
6.90142587e-02 3.13953340e-01 -1.46431044e-01 8.47173035e-02
1.78635240e-01 1.49370790e-01 -3.98744881e-01 1.45331219e-01
-3.80531773e-02 1.00059521e+00 -7.70319402e-01 -6.76024795e-01
-4.75776911e-01 5.16038500e-02 -1.15436292e+00 1.69217497e-01
-4.57373559e-01 -4.52871434e-03 -4.61147100e-01 7.66402364e-01
3.34035963e-01 -3.88857543e-01 1.67654648e-01 2.02430770e-01
-2.41853803e-01 4.64482814e-01 -5.11004508e-01 2.17910457e+00
-1.08213520e+00 7.42041707e-01 -2.75459439e-01 -5.45644522e-01
8.49917650e-01 1.46143571e-01 3.13906372e-01 -1.22585273e+00
-3.48516226e-01 7.85174847e-01 1.07539602e-01 -8.29512477e-01
6.25027955e-01 2.83659697e-01 -4.39335555e-01 5.88207364e-01
-7.87147656e-02 -1.58999145e-01 4.33818251e-01 1.80190206e-01
1.23113644e+00 -1.85563453e-02 1.48067221e-01 -6.91928983e-01
6.82610035e-01 7.28387296e-01 4.05008852e-01 8.65837216e-01
1.28799632e-01 2.77813137e-01 4.02017117e-01 -4.40459818e-01
-1.06482065e+00 -6.45015538e-01 -1.85873479e-01 1.49510336e+00
1.01597123e-01 -8.44686449e-01 -7.00988412e-01 -7.20460713e-01
-8.26088041e-02 6.43586278e-01 -2.16808870e-01 -1.22281708e-01
-5.66933036e-01 -3.35501522e-01 9.90335166e-01 1.24620952e-01
3.37014914e-01 -8.79385591e-01 -5.95947862e-01 2.11605266e-01
-4.32700545e-01 -8.67949843e-01 -8.91971588e-01 1.30737275e-01
-4.62459534e-01 -1.22538614e+00 -2.90691644e-01 -1.31474435e+00
5.59654832e-01 8.65826383e-02 1.61219680e+00 6.53482556e-01
-9.71523747e-02 7.07984194e-02 -6.00161970e-01 -1.10926241e-01
-8.98167849e-01 5.58413982e-01 -4.17988569e-01 -7.86895394e-01
7.71595001e-01 -2.25703165e-01 -7.12261200e-02 3.53465617e-01
-8.34220171e-01 2.92706843e-02 5.42749286e-01 7.86504805e-01
4.86304909e-02 -1.60728365e-01 1.41990855e-01 -9.77645218e-01
8.68400931e-01 -5.50550640e-01 -9.24294114e-01 6.16970062e-01
-1.05637646e+00 4.46972579e-01 6.56250060e-01 -4.25348341e-01
-7.02291310e-01 -9.85078886e-03 -3.18920106e-01 -1.83537483e-01
3.28505039e-01 1.32112741e+00 2.16405809e-01 -3.50010544e-01
1.07946837e+00 3.64686608e-01 -3.43774319e-01 -5.32498479e-01
1.32249519e-01 6.99907422e-01 3.45530808e-01 -1.11397314e+00
6.92604959e-01 -1.68054942e-02 -7.28108168e-01 -1.40360653e-01
-2.30408028e-01 -3.93936604e-01 -4.12875324e-01 1.21986412e-01
7.11903334e-01 -1.03947020e+00 -1.18000582e-01 5.76826967e-02
-1.31395674e+00 -3.27947855e-01 3.08788419e-01 3.85595798e-01
-3.10952187e-01 2.16980815e-01 -7.25742102e-01 -1.52001053e-01
-2.89389133e-01 -1.92390811e+00 9.40831900e-01 -4.09862459e-01
-2.83827186e-01 -7.52064526e-01 5.99246144e-01 1.59080416e-01
7.86867201e-01 -3.22637558e-01 1.57191634e+00 -6.63310111e-01
-6.42518461e-01 -1.69966474e-01 -3.18279952e-01 -1.05360210e-01
-6.14455231e-02 2.17920795e-01 -4.78869468e-01 -3.72914165e-01
-2.02150643e-01 -4.45083529e-01 4.77922827e-01 -2.63469338e-01
9.52400088e-01 -4.04698581e-01 -3.27702910e-01 6.81845009e-01
1.90989387e+00 2.41178691e-01 4.07528132e-01 6.09685481e-01
6.13620758e-01 5.23772180e-01 3.67140919e-01 7.48639647e-03
8.30709279e-01 8.21367741e-01 2.12894514e-01 6.09171353e-02
1.29959032e-01 -5.53955615e-01 4.86886740e-01 1.35696685e+00
4.33965236e-01 1.53375044e-01 -1.60113657e+00 7.23209977e-01
-1.72433877e+00 -5.94427347e-01 -3.48368473e-02 1.90514827e+00
1.37178087e+00 -4.94717434e-02 -5.84110580e-02 -5.29800594e-01
3.61010611e-01 -3.16306680e-01 -2.40698710e-01 -5.10081351e-01
-1.11833820e-02 1.81711733e-01 5.78878343e-01 5.47214329e-01
-7.42352486e-01 1.03137350e+00 6.57296419e+00 6.74452662e-01
-1.33571196e+00 3.95790726e-01 6.14497811e-02 2.64399678e-01
-7.83740461e-01 5.67502558e-01 -3.51377249e-01 5.26559591e-01
8.60987902e-01 -5.67746639e-01 9.19928193e-01 1.22815311e+00
-2.48823926e-01 1.38739437e-01 -1.37933481e+00 8.77429307e-01
3.92834991e-02 -1.28431463e+00 -8.27246755e-02 -2.17171386e-01
8.35068047e-01 8.45355272e-01 -3.07587713e-01 8.65818739e-01
8.21150303e-01 -9.89507139e-01 1.16810024e+00 2.75355160e-01
8.87288213e-01 -3.86855751e-01 4.85403061e-01 5.62408805e-01
-1.19508910e+00 -2.47669056e-01 -2.35681936e-01 1.76035658e-01
-4.00500834e-01 4.89146620e-01 -3.94417197e-01 3.54131311e-01
8.65822852e-01 5.81320465e-01 -1.09954488e+00 8.90835643e-01
-1.82675898e-01 3.42609853e-01 -4.61676717e-02 -1.45730168e-01
1.39143527e-01 -3.25112492e-02 2.19401479e-01 1.46034932e+00
5.83596289e-01 -5.13399780e-01 4.92141157e-01 1.41296673e+00
-2.24531204e-01 5.05528748e-01 -8.19558680e-01 -1.19384021e-01
4.89164531e-01 8.79739702e-01 -1.96733952e-01 -2.39069521e-01
-1.20990777e+00 8.26141179e-01 4.82285470e-01 3.82452488e-01
-8.93682718e-01 -8.02421212e-01 2.63860971e-01 -1.08231381e-01
-7.76615962e-02 -1.85413212e-01 -8.04398954e-02 -1.62814009e+00
4.68908548e-01 -1.50567603e+00 4.01887894e-01 -9.06143546e-01
-1.12167108e+00 8.95928741e-01 -9.24839526e-02 -1.09243894e+00
-4.60864663e-01 -4.04072702e-01 -3.85094851e-01 9.79067266e-01
-1.54025209e+00 -1.20507503e+00 2.01698113e-02 4.41013217e-01
4.86563772e-01 -5.60287476e-01 8.50902557e-01 9.76100445e-01
-1.91224039e-01 8.64513099e-01 2.36385494e-01 5.36172748e-01
8.21352124e-01 -9.46018159e-01 5.43790400e-01 9.68171537e-01
2.83381850e-01 1.42157125e+00 6.38218939e-01 -6.20373249e-01
-1.72759199e+00 -9.55768883e-01 1.36698496e+00 -5.96056342e-01
1.04015648e+00 -4.05688316e-01 -9.18109655e-01 6.69103742e-01
4.30489987e-01 -1.03563413e-01 3.94601464e-01 1.75371945e-01
-8.55347455e-01 1.61614880e-01 -8.63991201e-01 4.94852692e-01
9.02645767e-01 -1.35595047e+00 -6.16177022e-01 3.07373196e-01
8.48204195e-01 -5.10835588e-01 -1.06253612e+00 2.41662487e-01
6.35864079e-01 -7.67050624e-01 7.16561973e-01 -6.11047208e-01
8.62078369e-01 -4.28885937e-01 -5.12415349e-01 -1.21378863e+00
-1.21141970e-01 -1.45279914e-01 5.98330140e-01 1.12835562e+00
9.14113164e-01 -6.23907447e-01 3.38231146e-01 6.19679213e-01
-4.40242290e-02 -4.32484835e-01 -7.66463459e-01 -8.75554681e-01
5.24838150e-01 -3.95486563e-01 8.20080459e-01 1.35266006e+00
4.89595056e-01 2.95755565e-01 7.82059133e-02 -5.42967282e-02
-5.05354591e-02 6.39915943e-01 8.86561275e-01 -1.06620383e+00
-4.98637944e-01 -5.57479739e-01 3.56146954e-02 -9.80190635e-01
6.06277764e-01 -1.67680919e+00 8.66496353e-04 -1.13386762e+00
6.19200826e-01 -6.98468685e-01 7.24126324e-02 7.42881298e-01
4.15919423e-02 -2.49749720e-01 1.26325697e-01 6.84338570e-01
-4.16065723e-01 1.79717392e-01 6.16193533e-01 -5.71713507e-01
1.92385272e-03 -2.91704923e-01 -6.75872684e-01 3.78832281e-01
5.18472135e-01 -8.05351377e-01 -4.03583974e-01 -1.25626969e+00
9.34012353e-01 1.63355768e-01 1.28787801e-01 -8.03072214e-01
5.05434990e-01 -2.28258193e-01 -4.70975518e-01 -1.21084601e-01
-4.94504809e-01 -1.06042171e+00 3.66781592e-01 6.58717811e-01
-7.12937951e-01 7.87538469e-01 2.14819595e-01 9.48544517e-02
-4.87114102e-01 -5.57046533e-01 6.85364842e-01 -2.72876382e-01
-7.31413186e-01 8.30894336e-02 -3.98662120e-01 4.20003116e-01
5.36333382e-01 1.90627828e-01 -5.49007952e-01 -1.95462536e-02
-2.47205421e-02 2.76605308e-01 1.12498367e+00 8.68773818e-01
1.07786477e-01 -1.45254660e+00 -5.11156023e-01 3.62427056e-01
8.81770074e-01 -6.03417158e-01 -6.74328089e-01 6.63692355e-01
-7.39898562e-01 7.37532139e-01 -1.86629698e-01 -6.29502892e-01
-1.02664351e+00 7.50193000e-01 4.21332598e-01 -2.94023782e-01
-1.02635458e-01 4.71646428e-01 -2.71966815e-01 -1.24858320e+00
4.67385799e-02 -5.89912653e-01 1.74360001e-03 -5.03550708e-01
1.22862728e-02 -1.67348951e-01 1.23303562e-01 -6.71611845e-01
-4.67171907e-01 6.50539756e-01 1.09530553e-01 -5.73424697e-02
1.12087131e+00 1.28949404e-01 -7.88830340e-01 1.25111639e-01
1.61012971e+00 4.62456554e-01 -2.73025334e-01 -5.73198617e-01
5.79108059e-01 -5.71767271e-01 -2.24818155e-01 -6.32932603e-01
-1.15028441e+00 6.21507585e-01 5.08077204e-01 6.87302202e-02
9.57531631e-01 1.77526444e-01 4.65554386e-01 8.26626778e-01
8.62088740e-01 -1.01740766e+00 -1.01149358e-01 1.02861845e+00
6.80975497e-01 -1.33590353e+00 -4.04639870e-01 -1.40953645e-01
-4.04907227e-01 1.07416332e+00 7.56342947e-01 1.91259921e-01
6.46514714e-01 3.47723365e-01 2.43384972e-01 -3.16959411e-01
-8.98682892e-01 4.16004807e-02 2.21169963e-01 1.72473997e-01
8.49433303e-01 -2.68374175e-01 -4.81941760e-01 4.40886289e-01
-3.14253598e-01 2.01735243e-01 3.39620024e-01 1.05318654e+00
-1.15982078e-01 -1.51845062e+00 -8.47194940e-02 5.43716073e-01
-5.70382535e-01 -6.56936765e-01 -3.76215100e-01 5.77301204e-01
-7.54655898e-02 8.22162151e-01 -3.20211440e-01 -6.20066524e-01
4.27037686e-01 7.82984719e-02 1.37059867e-01 -8.43816221e-01
-9.24373984e-01 -4.12689209e-01 -9.70346201e-03 -6.87054813e-01
-3.14773142e-01 -4.63135809e-01 -1.09527016e+00 -3.88807207e-01
-4.26906377e-01 4.79357749e-01 6.70819938e-01 6.77611053e-01
5.05310774e-01 1.33711159e-01 3.99780214e-01 4.31287438e-02
-5.49167633e-01 -6.67827308e-01 1.94290519e-01 3.82500768e-01
1.78216040e-01 -3.26657385e-01 -1.20687760e-01 1.91926360e-01] | [7.595117092132568, 7.986221790313721] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.