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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
440b9b5b-1b94-4424-8ac1-4bd6812e09bd | instruct-neuraltalker-editing-audio-driven | 2306.10813 | null | https://arxiv.org/abs/2306.10813v1 | https://arxiv.org/pdf/2306.10813v1.pdf | Instruct-NeuralTalker: Editing Audio-Driven Talking Radiance Fields with Instructions | Recent neural talking radiance field methods have shown great success in photorealistic audio-driven talking face synthesis. In this paper, we propose a novel interactive framework that utilizes human instructions to edit such implicit neural representations to achieve real-time personalized talking face generation. Given a short speech video, we first build an efficient talking radiance field, and then apply the latest conditional diffusion model for image editing based on the given instructions and guiding implicit representation optimization towards the editing target. To ensure audio-lip synchronization during the editing process, we propose an iterative dataset updating strategy and utilize a lip-edge loss to constrain changes in the lip region. We also introduce a lightweight refinement network for complementing image details and achieving controllable detail generation in the final rendered image. Our method also enables real-time rendering at up to 30FPS on consumer hardware. Multiple metrics and user verification show that our approach provides a significant improvement in rendering quality compared to state-of-the-art methods. | ['Bo Yan', 'Weimin Tan', 'Reian He', 'Yuqi Sun'] | 2023-06-19 | null | null | null | null | ['talking-face-generation', 'face-generation'] | ['computer-vision', 'computer-vision'] | [ 5.66261351e-01 1.90709636e-01 1.68490022e-01 -5.52520454e-01
-7.54038870e-01 -3.34244341e-01 5.21052182e-01 -4.47901934e-01
-3.09185032e-02 3.68943900e-01 4.53977972e-01 -4.99823876e-03
1.93095252e-01 -6.31151438e-01 -6.04996979e-01 -4.78302568e-01
3.65328938e-01 -4.58854809e-03 -6.23347179e-04 -3.00093919e-01
3.31385463e-01 6.68701351e-01 -1.85904288e+00 5.09712040e-01
7.09451914e-01 1.09066248e+00 3.18921000e-01 9.13653553e-01
-1.48062155e-01 8.11785102e-01 -6.52952850e-01 -3.24398845e-01
2.87415862e-01 -4.74903852e-01 -3.78409564e-01 2.22556025e-01
5.18100083e-01 -7.09306121e-01 -3.40126306e-01 7.00913787e-01
9.85104322e-01 1.27514675e-01 3.46563578e-01 -9.54190075e-01
-4.40894037e-01 5.14641225e-01 -6.44987702e-01 -1.75161988e-01
6.02917194e-01 3.11331660e-01 5.84022701e-01 -8.88180017e-01
5.55315912e-01 1.33045483e+00 5.57702243e-01 1.00415027e+00
-1.34438717e+00 -1.02880180e+00 1.58979386e-01 6.23149909e-02
-1.51363730e+00 -1.13476408e+00 1.07240117e+00 -2.09887162e-01
6.11544192e-01 4.82801825e-01 7.05500126e-01 7.41414964e-01
4.39002290e-02 4.30028737e-01 8.56118083e-01 -5.00053585e-01
1.59772068e-01 1.88472733e-01 -4.85658854e-01 7.66482949e-01
-6.60588503e-01 2.88510155e-02 -9.73384440e-01 -5.66055551e-02
1.03940821e+00 -5.56692421e-01 -6.11307323e-01 -4.47190441e-02
-8.98552775e-01 4.63006258e-01 1.83208942e-01 -2.13363320e-02
-2.68240005e-01 5.06798983e-01 2.25342453e-01 -8.52643475e-02
7.01344669e-01 1.39384583e-01 -1.58194974e-02 -2.20914930e-01
-1.25009823e+00 1.89388528e-01 4.88144934e-01 8.14439952e-01
5.41091561e-01 3.71446282e-01 -3.62436056e-01 1.01558566e+00
2.29425207e-01 3.64834815e-01 1.78618863e-01 -1.59154713e+00
1.95764259e-01 1.07116565e-01 4.59792204e-02 -8.97925973e-01
-6.63564354e-02 7.88471568e-03 -6.70333266e-01 5.79561472e-01
-1.39014022e-02 -1.23976089e-01 -8.36172581e-01 1.83202541e+00
5.39775252e-01 4.76439565e-01 -2.52039224e-01 9.33807194e-01
8.12524140e-01 8.96954536e-01 -9.94285569e-02 -3.27685684e-01
1.20043826e+00 -6.52807295e-01 -1.00111198e+00 2.70344138e-01
1.34927332e-01 -8.36071253e-01 1.30481482e+00 6.21231377e-01
-1.45748961e+00 -6.96585238e-01 -9.70704675e-01 -4.43728387e-01
3.47047925e-01 2.80309111e-01 6.66668713e-01 9.45554137e-01
-1.39802778e+00 5.18242478e-01 -5.87562680e-01 9.92698222e-02
3.19651455e-01 6.17875695e-01 -2.47879345e-02 1.39907166e-01
-7.42315590e-01 3.20188999e-01 -2.76578516e-01 1.10406324e-03
-9.59983230e-01 -1.35518467e+00 -7.71058500e-01 7.50752315e-02
1.59269840e-01 -8.07719171e-01 1.36046374e+00 -1.12236702e+00
-2.53024840e+00 7.46405661e-01 -5.08391261e-01 -2.78613210e-01
5.19467533e-01 -1.82394698e-01 -2.05621287e-01 3.39351594e-01
-4.35938209e-01 1.22280717e+00 1.19964969e+00 -1.45276141e+00
-4.53463554e-01 -3.30056921e-02 1.90330118e-01 4.81565058e-01
-4.75939691e-01 1.43244714e-01 -7.99004734e-01 -8.19418967e-01
-1.44836903e-02 -7.01094449e-01 -4.38540522e-03 6.31676674e-01
-1.79728210e-01 4.43243794e-02 1.11274338e+00 -6.11277938e-01
1.07191277e+00 -2.32417345e+00 7.47177973e-02 8.51851627e-02
3.73377830e-01 1.63921535e-01 -1.24705665e-01 -1.06873438e-01
-1.02061573e-02 -8.46417248e-02 -9.74024311e-02 -8.75996053e-01
-1.55038998e-01 -1.31885543e-01 -5.29913068e-01 3.93464088e-01
-6.15345053e-02 5.27433515e-01 -4.95868355e-01 -4.78373498e-01
4.90541965e-01 1.34384215e+00 -1.05678499e+00 2.68020451e-01
-2.85953075e-01 7.75280237e-01 -4.34500426e-02 4.54011708e-01
8.06218624e-01 2.50657946e-01 1.03915177e-01 -4.34170216e-01
-2.23888576e-01 3.27974141e-01 -1.12041938e+00 2.04336262e+00
-1.04833376e+00 1.03526890e+00 5.28158605e-01 -2.11927250e-01
9.59356248e-01 4.22665209e-01 5.15909851e-01 -6.39973819e-01
2.93029219e-01 -1.66973367e-01 -4.87018794e-01 -1.62575036e-01
6.92130625e-01 -4.99823131e-02 5.04750669e-01 3.77134532e-01
-1.16206951e-01 -6.92599714e-01 -2.97710449e-01 1.80990651e-01
6.31622612e-01 3.35755676e-01 -3.92283946e-01 -2.17393249e-01
7.31401861e-01 -7.68086851e-01 3.98430943e-01 1.60587937e-01
1.69340208e-01 9.05004144e-01 3.30629170e-01 -2.71378756e-01
-8.65089715e-01 -9.55961108e-01 9.33287889e-02 1.22652924e+00
4.49498892e-02 -6.58356786e-01 -1.24301565e+00 1.21187558e-02
-5.11572123e-01 8.62109244e-01 -4.73173499e-01 1.32208750e-01
-6.96814418e-01 -7.62698725e-02 4.84222174e-01 2.20073834e-01
6.28628373e-01 -9.52376604e-01 -6.62760317e-01 1.18146472e-01
-2.60815382e-01 -1.03218734e+00 -8.72215867e-01 -5.02371132e-01
-5.65880477e-01 -4.53542531e-01 -9.66303647e-01 -6.94109619e-01
6.56416714e-01 1.11048132e-01 7.77476251e-01 -2.92833839e-02
-5.63732028e-01 4.68611270e-01 2.97564089e-01 -1.74498498e-01
-5.58438838e-01 -2.25675046e-01 8.41609463e-02 4.06081021e-01
-3.88555616e-01 -7.90358126e-01 -9.56581831e-01 2.97448665e-01
-7.54469991e-01 4.54580635e-01 -9.77537632e-02 4.44127440e-01
6.68361425e-01 -1.25718996e-01 2.55127698e-01 -4.02828187e-01
7.38966584e-01 2.47247607e-01 -8.77825797e-01 -2.19401736e-02
-2.81872213e-01 -7.16126189e-02 6.06292546e-01 -7.04776824e-01
-1.63620043e+00 2.13009343e-01 -2.95159370e-01 -6.97356105e-01
1.57223344e-01 -2.67582804e-01 -2.67208993e-01 -3.73819858e-01
5.56585789e-01 -1.50642367e-02 -1.03920817e-01 -2.72672385e-01
6.64900243e-01 7.41348326e-01 7.69586623e-01 -5.62040091e-01
5.11649847e-01 7.61339664e-01 -9.75367054e-02 -1.10138321e+00
-2.14593396e-01 2.67752819e-02 -2.85386443e-01 -6.22516155e-01
8.88386369e-01 -9.57890093e-01 -1.34532070e+00 5.23623943e-01
-1.30827343e+00 -7.10641801e-01 -4.55056489e-01 2.91448534e-01
-7.99505115e-01 1.43319339e-01 -6.03344321e-01 -9.82050538e-01
-5.08626759e-01 -1.48249435e+00 1.31461406e+00 2.83200920e-01
-1.96182206e-01 -5.41833580e-01 -3.97147164e-02 3.61591965e-01
6.63397789e-01 -2.01812945e-02 5.13153017e-01 6.66052043e-01
-8.35576236e-01 2.31256709e-01 -1.95593953e-01 1.33227427e-02
2.31282532e-01 3.04862738e-01 -1.41982377e+00 -1.13295056e-01
3.75256017e-02 -1.02283031e-01 5.95127404e-01 6.08726263e-01
1.52980733e+00 -3.05853039e-01 -2.10510761e-01 1.03792334e+00
1.00298536e+00 2.20633492e-01 8.02321851e-01 -3.47983569e-01
6.86339617e-01 7.66810238e-01 3.74233276e-01 7.36143470e-01
2.74416089e-01 9.21354055e-01 1.48138806e-01 -1.85632885e-01
-6.31798804e-01 -3.86942595e-01 3.38176429e-01 5.98370612e-01
4.96884547e-02 -2.18501121e-01 -4.04653758e-01 3.04305315e-01
-1.21592391e+00 -8.53918076e-01 2.50223368e-01 2.28169179e+00
9.46664393e-01 -1.01250768e-01 -6.97641969e-02 1.67251110e-01
7.05818772e-01 2.29077041e-01 -5.10817230e-01 -6.99371159e-01
1.40098453e-01 5.01278698e-01 2.67613977e-01 1.05267859e+00
-5.54502666e-01 1.11230516e+00 6.33596468e+00 9.99384046e-01
-1.67388201e+00 1.00537091e-01 9.07841802e-01 -5.04310608e-01
-7.63552606e-01 -2.83855885e-01 -6.84055746e-01 5.87089285e-02
8.57451677e-01 -2.47411385e-01 7.72381067e-01 5.93052208e-01
7.83868313e-01 -1.14010526e-02 -8.91447365e-01 1.32904482e+00
2.96436846e-01 -1.55986059e+00 1.58273533e-01 5.51548004e-02
5.90935171e-01 -5.06917357e-01 4.61326957e-01 -2.68807530e-01
4.80765924e-02 -9.86275017e-01 8.30003321e-01 4.43401515e-01
1.49428689e+00 -1.05900466e+00 -3.14585626e-01 -2.02440806e-02
-1.30523455e+00 1.63320646e-01 -1.14187576e-01 3.40899110e-01
4.10875082e-01 3.90417933e-01 -8.13890159e-01 5.56271821e-02
6.73248112e-01 2.67371118e-01 -1.12249017e-01 5.60823083e-01
-7.26157427e-02 4.82018322e-01 -4.33866352e-01 2.88214296e-01
-3.14782202e-01 -4.78575416e-02 6.22803569e-01 1.05636907e+00
3.54531288e-01 3.38754863e-01 -3.57178599e-01 9.77150142e-01
-3.05832058e-01 2.62349278e-01 -4.63971376e-01 2.45257035e-01
2.23680571e-01 1.24144745e+00 -5.78096390e-01 -1.37124524e-01
-1.61174051e-02 1.35247302e+00 -8.86681676e-02 2.80060560e-01
-1.11806846e+00 -3.55782062e-01 8.95084262e-01 3.50258052e-01
1.32219777e-01 -3.02827269e-01 -2.34543487e-01 -7.99193323e-01
-8.08308572e-02 -7.39223063e-01 -3.02255332e-01 -1.12851977e+00
-3.02431643e-01 8.03488433e-01 -2.43070960e-01 -8.94440770e-01
-2.91071653e-01 -2.45012641e-01 -4.68686223e-01 8.56183648e-01
-1.43105352e+00 -1.08221805e+00 -5.26010334e-01 8.93805325e-01
8.12856734e-01 -1.68801583e-02 7.53477156e-01 5.00440776e-01
-1.93460494e-01 9.09193039e-01 -4.30566788e-01 -3.15175384e-01
7.40901947e-01 -5.73716581e-01 4.13727880e-01 5.75884044e-01
2.39413884e-02 5.25357246e-01 6.56507611e-01 -4.31413054e-01
-1.34502804e+00 -8.80983770e-01 4.61603314e-01 8.36729035e-02
6.28573522e-02 -6.79095089e-01 -5.89124322e-01 3.27558577e-01
4.19236481e-01 -1.53836906e-01 5.59310973e-01 -4.65356648e-01
-3.26636523e-01 -5.15881658e-01 -1.38972604e+00 1.06679380e+00
1.11097240e+00 -6.59458518e-01 2.19024599e-01 2.45214775e-01
8.84884119e-01 -6.41030014e-01 -4.88229364e-01 3.09609532e-01
7.63265610e-01 -1.10179794e+00 9.24660802e-01 1.42875552e-01
2.91708052e-01 -3.50842923e-01 -2.23816875e-02 -1.06322348e+00
-4.77025472e-02 -1.54000032e+00 5.49156852e-02 1.39464974e+00
2.31309921e-01 -3.47148538e-01 9.41652477e-01 6.42091632e-01
-5.26600182e-02 -4.93608743e-01 -9.27214503e-01 -5.46701290e-02
-3.92960370e-01 -7.45117486e-01 6.46898627e-01 6.22730732e-01
-1.48393691e-01 6.00845888e-02 -5.35049915e-01 2.31827170e-01
6.27587497e-01 -1.89582393e-01 9.66789424e-01 -8.03951502e-01
-3.35342944e-01 -3.28057855e-01 -2.57989138e-01 -1.41694009e+00
2.77814299e-01 -4.33673710e-01 9.33267623e-02 -1.19585025e+00
-1.91726968e-01 -4.85626519e-01 2.52981037e-01 1.59011200e-01
1.38748601e-01 6.41248226e-01 3.26620728e-01 -1.39509022e-01
6.27838373e-02 6.68515265e-01 1.29343545e+00 -5.11144735e-02
-5.89428544e-01 -6.99638352e-02 -5.86977482e-01 7.10813582e-01
6.49352133e-01 -6.37719184e-02 -8.32864821e-01 -6.84379935e-01
-9.73950326e-02 2.91424602e-01 2.91396409e-01 -1.23093951e+00
2.85976827e-01 9.44002531e-03 2.10345238e-01 -4.68452930e-01
1.13247585e+00 -7.22597361e-01 2.02911779e-01 1.15952574e-01
-6.59167707e-01 -1.39165223e-01 4.46592778e-01 2.45453343e-01
1.29416987e-01 1.72145575e-01 1.08564985e+00 2.20982775e-01
-2.12058455e-01 3.75990152e-01 -3.81068200e-01 -2.21280947e-01
8.74678731e-01 -2.19670057e-01 1.26751095e-01 -1.09631312e+00
-6.48871183e-01 -2.87269771e-01 3.81547481e-01 3.92815679e-01
8.88761640e-01 -1.12534642e+00 -6.84967339e-01 4.42811519e-01
-3.62179667e-01 -2.10475594e-01 4.58173722e-01 3.91921759e-01
-7.32378721e-01 5.46413250e-02 4.64745425e-02 -8.53839159e-01
-1.66105819e+00 2.70422250e-01 6.22912705e-01 2.82620817e-01
-7.43574023e-01 1.16997087e+00 5.55954218e-01 -2.87416819e-02
4.41728473e-01 -2.64155924e-01 7.85479397e-02 -2.01923564e-01
8.09266448e-01 2.19344974e-01 5.24792634e-02 -6.22491777e-01
-6.94151968e-02 8.54802012e-01 7.73383230e-02 -7.51994014e-01
1.04775345e+00 -3.72065544e-01 1.39732480e-01 2.42503107e-01
1.15087795e+00 6.24150813e-01 -1.68329012e+00 1.57541737e-01
-8.75265539e-01 -7.10026443e-01 3.58891100e-01 -6.39822781e-01
-1.40845501e+00 1.03831041e+00 8.61061990e-01 -3.73827696e-01
1.39282572e+00 -3.23174447e-01 7.80412078e-01 -6.47283047e-02
2.33752161e-01 -9.18351233e-01 1.14363998e-01 7.85857514e-02
1.05811131e+00 -6.30284131e-01 -1.04420736e-01 -9.05862510e-01
-5.51531732e-01 1.06165731e+00 4.33718234e-01 1.79080784e-01
5.88141978e-01 7.52660155e-01 3.16995651e-01 1.43279240e-01
-6.54425442e-01 2.33955324e-01 2.32193723e-01 6.67588115e-01
5.54372668e-01 -2.47492716e-01 -6.77752960e-03 -9.25709903e-02
-5.08293211e-01 3.07193071e-01 3.11088920e-01 4.22546953e-01
-1.78519353e-01 -8.87126625e-01 -3.62822086e-01 -1.32017419e-01
-5.57366192e-01 -2.00653717e-01 -1.94732666e-01 3.17677259e-01
-2.70825066e-02 1.07653236e+00 2.54221261e-01 -2.56995648e-01
2.22855851e-01 -1.28041938e-01 7.40153372e-01 -3.09444368e-01
-7.01616168e-01 3.78928006e-01 -7.94217810e-02 -9.32760000e-01
-2.21870989e-01 -2.08211273e-01 -1.29681277e+00 -5.22106349e-01
-1.92022949e-01 -1.08085014e-01 1.00078547e+00 4.27144051e-01
7.36409009e-01 6.81669891e-01 8.42635095e-01 -1.40504944e+00
1.77658811e-01 -5.90858638e-01 -2.34901890e-01 -1.47955455e-02
3.22908312e-01 -3.99809450e-01 -2.47692883e-01 3.01576495e-01] | [13.11867618560791, -0.44803112745285034] |
df2e54cc-995c-4a79-b7d7-e11290836d0c | ebsr-feature-enhanced-burst-super-resolution | null | null | https://openaccess.thecvf.com/content/CVPR2021W/NTIRE/html/Luo_EBSR_Feature_Enhanced_Burst_Super-Resolution_With_Deformable_Alignment_CVPRW_2021_paper.html | https://openaccess.thecvf.com/content/CVPR2021W/NTIRE/papers/Luo_EBSR_Feature_Enhanced_Burst_Super-Resolution_With_Deformable_Alignment_CVPRW_2021_paper.pdf | EBSR: Feature Enhanced Burst Super-Resolution With Deformable Alignment | We propose a novel architecture to handle the problem of multi-frame super-resolution (MFSR). The proposed framework is known as Enhanced Burst Super-Resolution (EBSR), which divides the MFSR problem into three parts: alignment, fusion, and reconstruction. We propose a Feature Enhanced Pyramid Cascading and Deformable convolution (FEPCD) module to align multiple low-resolution burst images in the feature level. And then the aligned features are fused by a Cross Non-Local Fusion (CNLF) module. Finally, the SR image is reconstructed by the Long Range Concatenation Network (LRCN). In addition, we build a cascading residual pathway structure (CR) to improve the performance. We conduct several experiments to analyze and demonstrate these modules. Our EBSR model won the champion in the real track and second place in the synthetic track in the NTIRE21 Burst Super-Resolution Challenge. | ['Shuaicheng Liu', 'Jian Sun', 'Haoqiang Fan', 'Lanpeng Jia', 'Youwei Li', 'Xuan Mo', 'Lei Yu', 'Ziwei Luo'] | 2021-06-15 | null | null | null | proceedings-of-the-ieee-cvf-conference-on | ['multi-frame-super-resolution', 'burst-image-super-resolution'] | ['computer-vision', 'computer-vision'] | [ 4.79322523e-01 -2.38483697e-01 3.05181503e-01 -3.36055845e-01
-1.25878775e+00 -1.36109293e-01 4.61377442e-01 -5.98793507e-01
-3.39378566e-01 9.02848005e-01 5.52947998e-01 5.18077433e-01
-4.54415753e-03 -6.96350574e-01 -8.73534322e-01 -5.99437594e-01
8.31767619e-02 -1.12880498e-01 5.35712600e-01 -4.52863783e-01
7.15034157e-02 5.75456500e-01 -1.58619452e+00 9.33475137e-01
8.77162218e-01 8.09372008e-01 4.71393347e-01 6.94297850e-01
9.47626084e-02 9.21589136e-01 -3.49809349e-01 -1.07174538e-01
2.82256067e-01 -5.67134380e-01 -9.17739272e-01 -1.47410423e-01
6.77769363e-01 -4.56497312e-01 -1.90478563e-01 1.08780050e+00
8.08006346e-01 2.11227909e-01 -9.08904523e-02 -6.70525789e-01
-7.20673203e-01 7.26665020e-01 -8.74230146e-01 8.72016966e-01
4.63041037e-01 -1.23917021e-01 6.92407250e-01 -1.39840746e+00
1.01946378e+00 1.51284111e+00 7.30708957e-01 4.26190138e-01
-1.22764087e+00 -7.69273400e-01 -1.58910602e-01 2.27615774e-01
-1.43853772e+00 -6.71651125e-01 3.41989398e-01 -1.54775023e-01
7.72264302e-01 1.74926803e-01 4.42921072e-01 1.17246866e+00
8.93228576e-02 4.49722528e-01 1.25412166e+00 2.24498902e-02
-2.01766148e-01 -5.54164231e-01 3.77161987e-02 6.50728643e-01
-1.13510189e-03 4.08665597e-01 -1.01129007e+00 7.93426558e-02
1.57920849e+00 -1.88012555e-01 -3.86018753e-01 4.11171496e-01
-1.58227682e+00 3.70067775e-01 6.49340153e-01 5.53336084e-01
-4.46479678e-01 2.57263780e-02 2.75899172e-02 3.63232136e-01
5.10510743e-01 2.08844647e-01 -1.94943532e-01 2.87883162e-01
-1.22551000e+00 4.33094829e-01 8.98647234e-02 6.03714049e-01
5.39195597e-01 6.92795515e-02 -4.84598696e-01 8.85133982e-01
-1.82777569e-01 -8.22958164e-03 3.98743331e-01 -1.47358072e+00
3.59360188e-01 -2.65861675e-02 4.31203574e-01 -9.67699945e-01
-5.07382751e-01 -7.71920443e-01 -1.10456347e+00 2.94802040e-01
1.02892987e-01 1.97173189e-02 -7.45378137e-01 1.88649428e+00
3.10632408e-01 1.00887501e+00 1.95702985e-01 1.37309146e+00
1.31577039e+00 5.74535847e-01 -2.44302526e-01 -5.43605804e-01
1.36844480e+00 -1.17517531e+00 -8.85076761e-01 2.73427039e-01
-9.33591798e-02 -9.26546991e-01 6.60874903e-01 2.43264437e-01
-1.64002335e+00 -1.09024656e+00 -1.07424402e+00 -5.21335840e-01
2.08190382e-01 1.14879899e-01 3.02711308e-01 -2.37386748e-01
-1.26350498e+00 1.03070700e+00 -8.29617381e-01 -5.92624135e-02
5.06508589e-01 1.92444921e-01 -5.46623588e-01 -1.48528501e-01
-1.28793943e+00 7.26760447e-01 1.77874103e-01 1.85570464e-01
-8.43916178e-01 -8.85344326e-01 -6.42880440e-01 -3.62314507e-02
-4.50332016e-02 -1.06535566e+00 9.16920662e-01 -8.53179514e-01
-1.34708405e+00 7.68131733e-01 -3.67082804e-01 -5.20559847e-01
3.46672028e-01 -4.30812478e-01 -7.44268477e-01 1.74103767e-01
2.05456764e-01 6.92475915e-01 8.71704102e-01 -1.16939867e+00
-8.71991277e-01 -1.77899063e-01 -6.21579662e-02 3.37593257e-01
4.84589785e-01 5.06591499e-01 -2.91690528e-01 -9.98270214e-01
2.67417371e-01 -3.60604435e-01 -3.10247481e-01 -3.35738152e-01
-1.11241698e-01 5.60647659e-02 6.73288882e-01 -1.10309649e+00
1.23935556e+00 -2.24843550e+00 7.32167065e-01 -4.05299872e-01
4.55930263e-01 1.49449036e-01 -4.37620819e-01 -5.73988631e-02
-3.66597265e-01 -2.60143206e-02 -2.97875255e-01 -5.46917319e-01
-6.78784192e-01 3.45153250e-02 -4.84666824e-01 2.24007010e-01
3.52999955e-01 8.87075305e-01 -9.04082239e-01 -4.44316715e-01
-2.04370782e-01 1.00109422e+00 -4.27801192e-01 2.09452242e-01
2.13859349e-01 1.06461847e+00 -2.61205882e-01 5.07694960e-01
9.96567786e-01 -4.55395907e-01 -2.24129260e-01 -9.20106411e-01
-6.23807490e-01 2.64340378e-02 -1.42261684e+00 2.27903485e+00
-1.53557882e-01 2.52527475e-01 2.00081021e-01 -3.39294940e-01
9.15803909e-01 2.29751483e-01 5.39823532e-01 -7.73635268e-01
-2.89524943e-01 2.53833383e-01 -3.51515979e-01 -2.87123322e-01
9.71244514e-01 -3.69115882e-02 2.20626667e-01 3.06684256e-01
1.99869826e-01 3.35969120e-01 4.50431928e-02 3.12842309e-01
1.06461775e+00 5.36503077e-01 5.96393272e-03 -1.08583970e-02
6.44332349e-01 -2.81780064e-01 8.48035932e-01 3.79765034e-01
-9.46277902e-02 1.21959019e+00 2.59202700e-02 -5.48467994e-01
-1.37645638e+00 -1.24501026e+00 1.30785555e-02 7.71945119e-01
3.89555573e-01 -5.11693597e-01 -7.11668909e-01 -2.68181503e-01
-5.44264317e-01 1.85183138e-01 -4.80399370e-01 8.53085816e-02
-9.32082117e-01 -8.07651997e-01 4.91103232e-01 5.57104528e-01
1.03452432e+00 -1.05632675e+00 -5.05347431e-01 2.80738264e-01
-7.14969397e-01 -1.50593030e+00 -5.19721985e-01 -3.57436478e-01
-6.40611947e-01 -7.09241390e-01 -6.57933772e-01 -5.52599847e-01
1.95570409e-01 5.11136949e-01 1.02506793e+00 -1.84124950e-02
-2.33656853e-01 -2.05183372e-01 -2.63536602e-01 3.46823603e-01
-1.29499987e-01 -1.52577028e-01 1.05350472e-01 2.05554068e-01
-3.03708583e-01 -9.69938874e-01 -7.10711658e-01 3.07942212e-01
-7.72448301e-01 4.71266210e-01 6.89678133e-01 8.12127829e-01
1.39250946e+00 8.04891065e-03 6.45434678e-01 -5.63900411e-01
6.20374560e-01 -3.11186641e-01 -4.51543212e-01 2.31127769e-01
-1.08732648e-01 1.57094467e-02 4.40074444e-01 -3.35081607e-01
-1.37497604e+00 2.26891655e-02 -1.52693629e-01 -6.85861290e-01
9.15053636e-02 1.25832260e-01 -3.12562548e-02 -2.71928310e-01
6.18575454e-01 3.10678542e-01 -2.21069604e-01 -9.12884176e-01
4.88106489e-01 2.27916241e-01 1.21084714e+00 -3.43902141e-01
8.97807658e-01 7.16990530e-01 4.96160937e-03 -3.00862819e-01
-1.29202688e+00 -6.74962401e-02 -7.44199157e-01 -4.22880240e-02
1.15803468e+00 -1.40263700e+00 -3.64886969e-01 4.46037740e-01
-1.27896893e+00 -7.87924081e-02 -4.82852638e-01 4.11469042e-01
-7.30987489e-01 2.69543350e-01 -8.98155510e-01 -3.13593149e-01
-5.12869120e-01 -1.05974829e+00 1.20858371e+00 6.45067930e-01
2.22261876e-01 -2.44348392e-01 1.94330826e-01 4.25357550e-01
7.86084652e-01 4.97511387e-01 1.38645604e-01 -3.15775685e-02
-9.59359825e-01 5.60021341e-01 -7.80790091e-01 2.81625777e-01
-6.25490258e-03 -1.89222917e-01 -8.12076569e-01 -3.82369906e-01
7.21682385e-02 -2.12354630e-01 1.12929237e+00 3.44837487e-01
1.04119623e+00 5.23467511e-02 9.31862928e-03 1.21955562e+00
1.52255750e+00 -3.34145904e-01 9.63304579e-01 2.92580366e-01
7.56982744e-01 2.12266505e-01 7.76966095e-01 2.22051367e-01
4.68863994e-01 9.91384208e-01 2.81607807e-01 -3.62290829e-01
-7.65558183e-01 -4.18427680e-03 5.27453542e-01 8.26860249e-01
-8.22764814e-01 2.00341463e-01 -2.72025257e-01 2.49062032e-01
-1.93262184e+00 -1.48104644e+00 -2.50388771e-01 1.79158223e+00
8.11992228e-01 -1.90652758e-01 1.13536483e-02 -2.30776817e-01
1.00084925e+00 2.98409671e-01 -3.25500309e-01 -9.80295017e-02
-8.30976069e-01 4.85988170e-01 1.51191860e-01 6.32618666e-01
-1.10652888e+00 1.04638314e+00 6.17646360e+00 9.40604031e-01
-9.33235526e-01 4.69766676e-01 4.93074000e-01 -6.79819360e-02
-1.70179501e-01 -1.12934992e-01 -9.37612593e-01 2.81890601e-01
8.86382878e-01 -1.26266584e-01 8.29583466e-01 1.98288918e-01
2.94161260e-01 7.16148317e-02 -5.45684695e-01 1.11081481e+00
-8.06033164e-02 -1.78830993e+00 1.43515304e-01 -2.94221550e-01
7.07045794e-01 2.45887592e-01 1.09884344e-01 2.83645783e-02
2.56745994e-01 -1.29670393e+00 6.37149453e-01 1.08373213e+00
1.21865332e+00 -8.71691883e-01 4.53797698e-01 9.65073854e-02
-1.68975496e+00 -1.41887888e-01 -4.79802877e-01 3.90329510e-01
5.79281032e-01 5.13345897e-01 -3.59967425e-02 1.22424734e+00
1.08911204e+00 9.41908062e-01 -3.29624802e-01 8.37818444e-01
-3.22631076e-02 -2.14233082e-02 -8.64821598e-02 1.24550736e+00
-9.21799764e-02 -8.86100456e-02 9.20044124e-01 1.16836023e+00
4.40756321e-01 4.66075271e-01 -2.39223372e-02 1.02810037e+00
-1.21357024e-01 -4.31114972e-01 1.21695243e-01 3.02064359e-01
4.78747547e-01 1.66134501e+00 -4.32636648e-01 -2.89941519e-01
-3.58577579e-01 1.33937657e+00 5.17159283e-01 3.18938881e-01
-1.03952980e+00 -3.95680442e-02 6.64792120e-01 -4.98717539e-02
4.74148035e-01 -1.86544001e-01 -7.03583881e-02 -1.42067790e+00
-5.99577511e-03 -9.69332814e-01 4.27033454e-01 -1.20376444e+00
-1.19305515e+00 1.08171213e+00 -2.52566934e-01 -1.28578925e+00
7.24176541e-02 2.28148982e-01 -2.98340589e-01 1.22419119e+00
-1.92342699e+00 -1.34935915e+00 -7.38778949e-01 9.81957257e-01
7.09624231e-01 1.01048402e-01 6.13559723e-01 6.56800866e-01
-5.47719240e-01 3.73696953e-01 -2.94800788e-01 -1.19972147e-01
8.90872180e-01 -8.42068553e-01 5.34752488e-01 1.09522080e+00
-1.08609445e-01 3.96525770e-01 5.71402252e-01 -7.31743157e-01
-9.16457713e-01 -1.52470434e+00 7.65282035e-01 -2.37165704e-01
3.45232487e-01 7.78416395e-02 -1.18316102e+00 6.33100927e-01
1.03087150e-01 5.86775303e-01 1.64078414e-01 -4.82023001e-01
-5.36519229e-01 5.12085073e-02 -1.32107246e+00 2.48832926e-01
1.53551042e+00 -4.45262313e-01 -5.18553853e-01 8.26878846e-02
1.32639933e+00 -7.67258406e-01 -1.38221598e+00 7.03702033e-01
4.98844981e-01 -1.24750888e+00 1.37467384e+00 -4.87613916e-01
8.31383884e-01 -8.34032476e-01 -5.31362772e-01 -1.03938711e+00
-8.49214435e-01 -7.87514925e-01 -2.61733681e-01 1.20723951e+00
4.93756160e-02 -2.03150988e-01 2.89546132e-01 -1.92352563e-01
-3.24014127e-01 -7.38800287e-01 -1.01562119e+00 -6.16440833e-01
-3.13692629e-01 2.72344023e-01 8.07620406e-01 1.08610439e+00
-6.96275473e-01 4.17092353e-01 -6.05661094e-01 5.49440145e-01
1.03779125e+00 3.03774983e-01 4.82643783e-01 -8.99708867e-01
-3.66258025e-01 -1.03918880e-01 -2.11841706e-02 -8.81082416e-01
-2.51631916e-01 -8.81939709e-01 -3.04165691e-01 -1.49163985e+00
4.69128817e-01 -1.98850721e-01 -4.34086740e-01 1.01722099e-01
-3.34383786e-01 5.62092304e-01 2.20560804e-01 5.38878322e-01
-8.63600254e-01 3.83425415e-01 1.61104798e+00 4.85410631e-01
-2.32216790e-01 -3.37833375e-01 -7.20953107e-01 5.80538034e-01
4.47995275e-01 -1.59239829e-01 7.18510300e-02 -5.76014400e-01
1.09332956e-01 6.52604043e-01 6.67867362e-01 -1.06946933e+00
3.09417397e-01 -5.33142537e-02 6.14661455e-01 -7.70063221e-01
4.94487703e-01 -3.42377901e-01 6.13149524e-01 7.05564842e-02
-3.27520639e-01 3.56628090e-01 -1.06037762e-02 3.59932840e-01
-3.68986309e-01 5.95404446e-01 1.19962323e+00 -1.97849870e-01
-7.25006759e-01 5.59941053e-01 3.05004686e-01 -1.54932678e-01
5.89874446e-01 -1.03062376e-01 -8.29334497e-01 4.10973243e-02
-1.23659182e+00 2.01496184e-01 4.73453909e-01 3.07877749e-01
9.61470127e-01 -1.49540353e+00 -1.32528102e+00 2.31425688e-01
-4.21722174e-01 2.81329900e-01 7.51444280e-01 1.09104669e+00
-2.84024656e-01 -1.95397988e-01 -6.40322089e-01 -4.93471920e-01
-1.27945971e+00 4.64005798e-01 4.98415142e-01 -7.27865100e-01
-1.01319253e+00 1.09278953e+00 1.27008125e-01 6.02595741e-03
-2.29416385e-01 1.40277296e-01 -6.44466937e-01 -7.51170889e-02
1.11732662e+00 5.34697056e-01 -1.43211007e-01 -9.42166865e-01
-8.43026638e-02 8.49559903e-01 -1.24873549e-01 -1.98743805e-01
1.78875256e+00 -4.46168244e-01 -3.01863611e-01 1.00866988e-01
7.96419978e-01 6.96611628e-02 -1.31322944e+00 -5.08214772e-01
-2.84734517e-01 -4.40950602e-01 6.75835907e-02 -8.30738783e-01
-1.28240192e+00 4.47920144e-01 8.08722675e-01 -3.00159186e-01
1.52442813e+00 -1.69339195e-01 1.13736951e+00 -2.59585708e-01
4.03941333e-01 -6.46968961e-01 1.38354823e-01 5.44889152e-01
1.31824338e+00 -7.61852682e-01 7.78619871e-02 -6.36335433e-01
-5.71452618e-01 1.08903325e+00 6.37417018e-01 -5.63722908e-01
3.12701702e-01 5.19797087e-01 -3.43451053e-01 -1.15037806e-01
-9.54992831e-01 -2.16473445e-01 2.03127965e-01 5.22653818e-01
3.39320302e-01 -2.85387009e-01 -5.02884209e-01 9.64794397e-01
-1.75131798e-01 4.74566907e-01 5.33824205e-01 4.76083457e-01
-5.74129701e-01 -1.04693782e+00 -4.23769057e-01 -1.06920853e-01
-5.96789300e-01 -2.29807049e-01 -2.17000348e-03 3.14881235e-01
5.55455804e-01 5.92230320e-01 1.19397186e-01 -7.02975273e-01
3.54059666e-01 -3.44662756e-01 7.08484054e-01 -3.84078532e-01
-7.77095675e-01 3.14478189e-01 6.30410835e-02 -1.29900253e+00
-8.54356289e-01 -4.38943565e-01 -1.37199545e+00 -2.11219907e-01
-3.29263806e-02 -8.13912228e-02 2.98730105e-01 6.87214196e-01
7.27719784e-01 9.08035219e-01 6.46921992e-01 -1.10690784e+00
-1.17811330e-01 -8.03837597e-01 -3.90728444e-01 2.42834479e-01
5.39180696e-01 -5.39472163e-01 -1.31718874e-01 2.09554955e-01] | [11.001164436340332, -1.8890049457550049] |
88582fc5-afaa-4795-a932-824f049e87c3 | augmented-transformer-achieves-97-and-85-for | 2003.02804 | null | https://arxiv.org/abs/2003.02804v2 | https://arxiv.org/pdf/2003.02804v2.pdf | State-of-the-Art Augmented NLP Transformer models for direct and single-step retrosynthesis | We investigated the effect of different training scenarios on predicting the (retro)synthesis of chemical compounds using a text-like representation of chemical reactions (SMILES) and Natural Language Processing neural network Transformer architecture. We showed that data augmentation, which is a powerful method used in image processing, eliminated the effect of data memorization by neural networks, and improved their performance for the prediction of new sequences. This effect was observed when augmentation was used simultaneously for input and the target data simultaneously. The top-5 accuracy was 84.8% for the prediction of the largest fragment (thus identifying principal transformation for classical retro-synthesis) for the USPTO-50k test dataset and was achieved by a combination of SMILES augmentation and a beam search algorithm. The same approach provided significantly better results for the prediction of direct reactions from the single-step USPTO-MIT test set. Our model achieved 90.6% top-1 and 96.1% top-5 accuracy for its challenging mixed set and 97% top-5 accuracy for the USPTO-MIT separated set. It also significantly improved results for USPTO-full set single-step retrosynthesis for both top-1 and top-10 accuracies. The appearance frequency of the most abundantly generated SMILES was well correlated with the prediction outcome and can be used as a measure of the quality of reaction prediction. | ['Igor V. Tetko', 'Pavel Karpov', 'Guillaume Godin', 'Ruud Van Deursen'] | 2020-03-05 | null | null | null | null | ['retrosynthesis'] | ['medical'] | [ 7.79197872e-01 8.64383355e-02 -3.06361914e-01 1.76891744e-01
-7.92730212e-01 -7.80143142e-01 1.10375869e+00 5.37523925e-01
-6.81079149e-01 8.26909125e-01 -1.54600784e-01 -4.97401953e-01
1.71052366e-01 -9.78037357e-01 -9.47138190e-01 -1.05291343e+00
7.51052946e-02 5.29558063e-01 2.85053849e-01 -3.75081867e-01
3.82951468e-01 8.99748445e-01 -1.29271817e+00 9.01975095e-01
6.77465439e-01 1.04535222e+00 5.11746347e-01 7.07105637e-01
-7.40378052e-02 8.96449208e-01 -5.25118172e-01 -4.36618507e-01
1.62729755e-01 -3.63473982e-01 -7.31392860e-01 -5.46204031e-01
2.98452288e-01 3.86555284e-01 -2.38956347e-01 8.07161570e-01
6.38916552e-01 1.58013090e-01 1.00636792e+00 -7.96376169e-01
-4.23252463e-01 8.57083499e-01 -2.51729161e-01 8.99568666e-03
4.50545639e-01 1.17635794e-01 8.28984737e-01 -1.41243935e+00
1.07369852e+00 9.13924634e-01 3.90772343e-01 3.38251770e-01
-1.34478676e+00 -9.58956301e-01 -3.30430716e-01 4.66812879e-01
-1.03952920e+00 -2.59683967e-01 3.37394089e-01 -5.67047060e-01
1.37303686e+00 2.94080347e-01 5.29570222e-01 1.23544049e+00
5.04044831e-01 4.80008602e-01 1.01133096e+00 -5.27653694e-01
9.86019596e-02 1.50224194e-01 -5.88870980e-02 5.26741862e-01
-2.16513634e-01 4.67604429e-01 -4.95584637e-01 2.20245227e-01
4.17158633e-01 -1.82327643e-01 -2.69235283e-01 1.92730263e-01
-1.39910352e+00 7.71906972e-01 7.54121423e-01 6.38233066e-01
-4.80743229e-01 -2.88347334e-01 4.90196109e-01 1.44687265e-01
3.36446404e-01 1.20020425e+00 -7.78722644e-01 1.89522341e-01
-8.14447701e-01 7.23168701e-02 5.72507560e-01 6.84866369e-01
3.77698869e-01 2.03190774e-01 -3.69557559e-01 7.11620092e-01
-5.32081544e-01 2.13121787e-01 6.85948610e-01 -4.04571414e-01
4.16850775e-01 4.80647951e-01 -1.12794444e-01 -5.39892375e-01
-8.93176675e-01 -7.68773794e-01 -1.01439333e+00 7.94352666e-02
6.93387747e-01 1.92883715e-01 -1.30636370e+00 1.78563356e+00
1.12706393e-01 -4.47781235e-01 3.70511293e-01 4.17859644e-01
8.41596842e-01 1.34328020e+00 9.95078385e-02 -7.19591677e-01
1.24350107e+00 -9.92594004e-01 -5.26961803e-01 1.22087881e-01
9.49726641e-01 -9.65937138e-01 7.70681083e-01 1.01276016e+00
-1.11996973e+00 -7.26451278e-01 -1.19652629e+00 -7.70289302e-02
-1.10008800e+00 2.34163016e-01 4.99119997e-01 4.98774946e-01
-6.45845294e-01 1.23004425e+00 -2.44391441e-01 -1.50750309e-01
1.83096886e-01 6.57285690e-01 -6.86800122e-01 -7.92950541e-02
-1.22145784e+00 1.16954458e+00 1.07785952e+00 -2.10804731e-01
-9.69048381e-01 -1.14099050e+00 -4.37322944e-01 1.89569145e-01
5.75177372e-01 -3.36890519e-01 9.98491526e-01 -5.34900963e-01
-1.44323647e+00 6.76962316e-01 5.89961596e-02 -7.11243272e-01
2.35354125e-01 9.52586830e-02 -5.88570774e-01 7.05093145e-02
-1.40916064e-01 9.87642705e-01 4.88888592e-01 -8.10200155e-01
-4.97558385e-01 -3.15033644e-01 -4.44614947e-01 -1.40256256e-01
-1.62754536e-01 -1.34665489e-01 7.85617307e-02 -6.40375912e-01
8.71502534e-02 -9.90125120e-01 -1.96981952e-01 -6.39274478e-01
-4.68590409e-01 -1.80709735e-01 3.54065448e-01 -7.97785103e-01
8.61569583e-01 -1.84458661e+00 2.95855850e-01 3.57954085e-01
3.60256433e-02 4.99567300e-01 -4.36944067e-01 7.04172432e-01
-1.31610775e+00 2.68972635e-01 7.33116120e-02 3.59785199e-01
-3.73759598e-01 -6.16149485e-01 -3.90645415e-01 1.15646631e-01
2.75255919e-01 6.92812204e-01 -9.23896432e-01 -2.13487707e-02
2.82929868e-01 3.69052857e-01 -4.49840605e-01 1.58630148e-01
-3.60643893e-01 1.71350181e-01 1.34592280e-01 5.15587330e-01
4.57733810e-01 -1.24765687e-01 2.69211650e-01 -7.47864306e-01
-4.26810265e-01 2.53987730e-01 -5.98445058e-01 1.50808597e+00
-4.63094771e-01 3.68814558e-01 -8.80858600e-01 -7.25265503e-01
1.12059820e+00 5.15795529e-01 5.40872514e-01 -1.17944860e+00
6.13620020e-02 2.96111673e-01 4.81892705e-01 -2.89143026e-02
4.90993857e-01 -4.09600824e-01 2.00230062e-01 -4.25461419e-02
3.77764583e-01 -1.82156190e-01 7.23574340e-01 2.25213051e-01
7.81516314e-01 5.25700338e-02 6.18917406e-01 -2.61791021e-01
8.11793506e-01 1.30090684e-01 2.50514746e-01 5.44951797e-01
5.23409784e-01 4.07869935e-01 5.12902141e-01 -7.61434078e-01
-1.29525852e+00 -7.60633469e-01 2.79879253e-02 1.14990497e+00
-5.44535756e-01 -8.02221060e-01 -5.53671837e-01 -5.52365363e-01
-5.51036954e-01 1.18812084e+00 -5.71539521e-01 -6.17394626e-01
-2.78327316e-01 -1.06214738e+00 4.93396610e-01 4.19357508e-01
1.25470653e-01 -1.00855553e+00 3.05737078e-01 3.75823736e-01
-1.23462312e-01 -1.11136413e+00 -2.12772146e-01 9.45082307e-01
-6.09487236e-01 -1.26562834e+00 -6.81248784e-01 -5.14189482e-01
2.53289968e-01 -1.16286851e-01 6.75433636e-01 -3.14237654e-01
-2.38210082e-01 -4.47503209e-01 -6.62759915e-02 -6.30736947e-01
-1.17805302e+00 2.16490656e-01 -5.81546240e-02 -3.16295445e-01
-1.86630458e-01 -3.17730248e-01 -3.69811624e-01 1.34615481e-01
-8.92725527e-01 3.45032692e-01 6.86998129e-01 1.10923672e+00
6.82462037e-01 -3.67892534e-02 4.52268183e-01 -6.66617930e-01
2.45123819e-01 -1.26396164e-01 -7.54171431e-01 3.15720707e-01
-8.57634664e-01 4.75399375e-01 1.08581591e+00 -6.15686238e-01
-7.82166839e-01 4.51727748e-01 -4.19175655e-01 -1.92096233e-01
9.94603150e-03 4.96313602e-01 -1.23456284e-01 -2.56528296e-02
9.95039880e-01 5.43289185e-01 -1.66658059e-01 -3.12544435e-01
2.33580098e-01 2.11720049e-01 5.69068611e-01 -4.10431534e-01
4.26404119e-01 -1.02044530e-01 6.78223073e-01 -9.80811119e-01
-6.87894285e-01 -3.77442867e-01 -6.13052785e-01 8.27041790e-02
1.04978144e+00 -8.09988201e-01 -1.00242221e+00 4.75812048e-01
-1.30581331e+00 -3.01292032e-01 -8.93313140e-02 3.97217751e-01
-4.75410074e-01 4.75151271e-01 -6.02927089e-01 -4.97711182e-01
-7.62115777e-01 -1.45770609e+00 5.34730732e-01 -2.91293859e-01
-2.93721497e-01 -5.68225145e-01 -5.23219891e-02 2.15785414e-01
-6.88407347e-02 3.67964894e-01 1.39100969e+00 -1.04575574e+00
-3.21030587e-01 -2.24549651e-01 -8.72490332e-02 1.49325997e-01
-3.34105194e-02 1.26125231e-01 -1.08045745e+00 -1.07077897e-01
-4.64136899e-01 -2.96500832e-01 1.15500331e+00 2.05063656e-01
1.19804001e+00 -6.38325810e-02 -2.34333873e-01 2.96919703e-01
1.27962160e+00 8.11076283e-01 9.54574227e-01 2.95565069e-01
5.72552145e-01 3.83109838e-01 6.31643116e-01 4.92699705e-02
-4.92493838e-01 9.10254240e-01 6.72507644e-01 -2.21795961e-01
-9.65298116e-02 -3.97271186e-01 5.44149280e-01 2.28914946e-01
-3.39539498e-01 -4.62834984e-01 -8.10593426e-01 -5.83126359e-02
-1.32558274e+00 -1.17074764e+00 -4.04143453e-01 2.35249019e+00
8.77162099e-01 3.38329911e-01 3.82739186e-01 6.64955199e-01
6.52595639e-01 -2.41715595e-01 -3.58323246e-01 -5.30028760e-01
-2.77878523e-01 6.57545984e-01 6.68619871e-01 2.93957651e-01
-9.71392334e-01 1.14627135e+00 6.58241177e+00 1.34556675e+00
-1.31072462e+00 -1.93756849e-01 8.17356408e-01 -7.13715777e-02
2.80093163e-01 -3.00064474e-01 -8.28714132e-01 3.95586401e-01
1.36681366e+00 3.77636664e-02 4.06487167e-01 6.57251894e-01
1.40788242e-01 7.02936053e-02 -1.33626461e+00 8.66329312e-01
-2.61418015e-01 -2.10252380e+00 3.85168910e-01 2.06376627e-01
5.41432202e-01 -4.58598807e-02 -6.59080893e-02 1.58968985e-01
-7.67920166e-02 -1.00468493e+00 9.15916562e-01 1.49029940e-01
1.06600296e+00 -9.40164804e-01 5.88426411e-01 2.96671301e-01
-9.80401456e-01 -2.49647886e-01 -1.30043447e-01 2.37078249e-01
-2.67772675e-01 5.43804049e-01 -1.51203072e+00 6.68038249e-01
1.51372358e-01 7.51307487e-01 -5.48038006e-01 6.68252468e-01
-9.35063660e-02 3.30842972e-01 -3.51595134e-01 -1.30613834e-01
3.88734609e-01 -2.18692526e-01 2.01518580e-01 1.45534742e+00
4.20118123e-01 -1.60293683e-01 -8.41959640e-02 6.39466584e-01
-5.12247123e-02 2.79551506e-01 -5.36070943e-01 -4.77629602e-01
2.08973259e-01 1.09054041e+00 -8.76997113e-01 -4.73735839e-01
1.46935791e-01 7.21447825e-01 8.78857374e-02 1.76118419e-01
-8.16086352e-01 -2.52095550e-01 7.85127506e-02 5.20909913e-02
2.35296622e-01 1.83478773e-01 -9.52422172e-02 -7.07828939e-01
-5.03949881e-01 -1.05090964e+00 7.25169837e-01 -9.30379629e-01
-4.77911949e-01 6.21693790e-01 -3.18900853e-01 -9.65339184e-01
-1.08759940e-01 -1.04172397e+00 -4.26218450e-01 6.21855974e-01
-1.17820442e+00 -8.93592179e-01 8.29601735e-02 3.22316140e-01
6.60286009e-01 -4.16579723e-01 1.08266258e+00 2.74167478e-01
-6.71229362e-01 4.70532656e-01 2.30073303e-01 -3.68120581e-01
8.32718253e-01 -1.23728013e+00 1.28779829e-01 7.29231536e-01
3.25391352e-01 4.10509497e-01 8.29211354e-01 -4.03956443e-01
-1.11815977e+00 -1.07214153e+00 8.61552417e-01 -5.38001433e-02
5.30579269e-01 -3.43846440e-01 -5.36571085e-01 1.84201941e-01
7.51357945e-03 -6.02005541e-01 7.46754646e-01 -2.00929433e-01
-5.54272711e-01 -6.33980930e-02 -1.08119714e+00 6.88656032e-01
7.66491771e-01 -3.76987904e-01 -2.38088012e-01 8.43025923e-01
9.14045036e-01 -3.46606821e-01 -1.25594449e+00 2.79781640e-01
4.45312321e-01 -7.25674331e-01 1.40781462e+00 -8.53418827e-01
9.87068355e-01 -1.06643014e-01 -1.06996872e-01 -1.29088640e+00
-7.26644993e-01 -3.50333542e-01 8.34840685e-02 6.21967852e-01
9.37583983e-01 -3.81320745e-01 6.38008177e-01 6.50981143e-02
-4.07262802e-01 -6.59009814e-01 -6.09697998e-01 -7.33908772e-01
1.21726014e-01 -2.05039591e-01 2.86311179e-01 7.53152668e-01
4.67787422e-02 7.60815859e-01 -2.94119120e-01 -8.87372196e-02
4.26653437e-02 -1.49581358e-01 4.36779648e-01 -9.75805938e-01
-1.86482698e-01 -5.03879666e-01 -2.00569257e-01 -4.51820254e-01
-1.42043214e-02 -1.41189754e+00 -3.10614914e-01 -1.05003107e+00
1.55652881e-01 1.90034077e-01 -5.21969676e-01 5.84261537e-01
1.17499299e-01 2.24984780e-01 4.11084712e-01 -1.32709533e-01
-8.89593677e-04 3.09751391e-01 1.18438351e+00 -3.21343601e-01
-1.92086279e-01 -1.38528287e-01 -5.40220261e-01 3.98877561e-01
7.78265595e-01 -4.85026836e-01 -2.75804430e-01 4.58043158e-01
6.35142148e-01 3.15057695e-01 1.06600448e-01 -1.07630157e+00
5.57987355e-02 -2.06910819e-01 7.98095345e-01 -1.17408609e+00
5.46190560e-01 -6.69103205e-01 5.10541081e-01 1.09361231e+00
-6.35897756e-01 -1.35266498e-01 4.06203687e-01 3.79248112e-01
6.64189309e-02 -2.38975585e-01 8.42836261e-01 -2.93569446e-01
-4.91779417e-01 2.16039240e-01 -6.44727707e-01 -5.71043730e-01
9.65256512e-01 -2.25865021e-01 -3.08742553e-01 -1.02997132e-01
-8.83547068e-01 -3.54015559e-01 1.05949432e-01 3.27554584e-01
3.84044349e-01 -9.54645514e-01 -4.34587449e-01 1.98012143e-02
7.85983130e-02 -2.81351954e-01 -2.38476153e-02 7.76027143e-01
-5.03700316e-01 8.63097906e-01 -3.43413413e-01 -3.33004355e-01
-1.48534977e+00 1.07873976e+00 3.91011715e-01 -8.06058228e-01
-1.01073645e-03 6.88969433e-01 4.86985922e-01 -2.00965464e-01
3.29069495e-02 -5.65053105e-01 -2.77148515e-01 3.06248158e-01
6.35339081e-01 3.52356672e-01 6.16386056e-01 -2.01690719e-01
-1.45093009e-01 2.06775934e-01 -5.30058086e-01 3.10370594e-01
1.52279603e+00 7.60450721e-01 -3.52278650e-02 2.28513256e-01
1.26871705e+00 -3.99243206e-01 -6.13497615e-01 8.05268660e-02
-9.27058863e-04 1.74542725e-01 1.54211193e-01 -1.35678864e+00
-8.74641120e-01 1.01103961e+00 5.92285156e-01 -1.94289327e-01
9.52665210e-01 -3.32691878e-01 4.70947832e-01 8.76556635e-01
-4.02248800e-02 -9.27058220e-01 2.75166422e-01 6.63717091e-01
1.13262177e+00 -9.78678584e-01 1.54573321e-01 -6.88819587e-01
-3.45292300e-01 1.62637162e+00 3.51112217e-01 5.01251221e-01
5.85548915e-02 -1.43101616e-02 -6.47215664e-01 -1.21373795e-01
-7.55641103e-01 2.12641969e-01 2.45047912e-01 4.59143311e-01
5.20935416e-01 -4.73899767e-02 -3.48047733e-01 5.50502300e-01
-2.55421430e-01 -1.84475444e-02 4.90152508e-01 5.52213013e-01
-3.42324466e-01 -1.20978045e+00 -3.82437348e-01 1.70497730e-01
-4.35685188e-01 -5.51967859e-01 -6.74733758e-01 8.67016435e-01
1.40580952e-01 5.49036860e-01 -2.03867033e-01 -6.07910991e-01
1.87610492e-01 6.56847000e-01 7.38129795e-01 -3.83045673e-01
-9.21329021e-01 8.17626268e-02 3.88558507e-01 -4.50472027e-01
-3.80851701e-02 -2.78508157e-01 -1.18449366e+00 -1.60414457e-01
-4.92565513e-01 2.82680631e-01 7.06178606e-01 7.78817832e-01
1.53693736e-01 6.35493279e-01 6.89145505e-01 -8.46338749e-01
-4.68880355e-01 -9.59984303e-01 -2.81160444e-01 2.75296271e-01
-1.36556029e-01 -4.94346887e-01 -7.20555112e-02 1.70028135e-01] | [4.490431308746338, 6.105625152587891] |
798ada7a-f90a-41cd-b330-13662c097474 | adaptive-transfer-learning-for-plant | 2201.05261 | null | https://arxiv.org/abs/2201.05261v1 | https://arxiv.org/pdf/2201.05261v1.pdf | Adaptive Transfer Learning for Plant Phenotyping | Plant phenotyping (Guo et al. 2021; Pieruschka et al. 2019) focuses on studying the diverse traits of plants related to the plants' growth. To be more specific, by accurately measuring the plant's anatomical, ontogenetical, physiological and biochemical properties, it allows identifying the crucial factors of plants' growth in different environments. One commonly used approach is to predict the plant's traits using hyperspectral reflectance (Yendrek et al. 2017; Wang et al. 2021). However, the data distributions of the hyperspectral reflectance data in plant phenotyping might vary in different environments for different plants. That is, it would be computationally expansive to learn the machine learning models separately for one plant in different environments. To solve this problem, we focus on studying the knowledge transferability of modern machine learning models in plant phenotyping. More specifically, this work aims to answer the following questions. (1) How is the performance of conventional machine learning models, e.g., partial least squares regression (PLSR), Gaussian process regression (GPR) and multi-layer perceptron (MLP), affected by the number of annotated samples for plant phenotyping? (2) Whether could the neural network based transfer learning models improve the performance of plant phenotyping? (3) Could the neural network based transfer learning be improved by using infinite-width hidden layers for plant phenotyping? | ['Jingrui He', 'Kaiyu Guan', 'Sheng Wang', 'Elizabeth A. Ainsworth', 'Jun Wu'] | 2022-01-14 | null | null | null | null | ['plant-phenotyping', 'gpr', 'gpr'] | ['computer-vision', 'computer-vision', 'miscellaneous'] | [ 3.82825911e-01 -1.25424862e-01 -1.82780817e-01 1.12673670e-01
6.31061941e-02 -6.67742014e-01 -2.86010485e-02 3.71956944e-01
3.03771764e-01 7.66938925e-01 -5.95271230e-01 -7.46949494e-01
-4.99206603e-01 -1.06483650e+00 -4.21378911e-01 -1.00305140e+00
1.97301388e-01 2.84454465e-01 5.22607975e-02 -2.11461395e-01
2.62743458e-02 9.31588471e-01 -1.47846901e+00 5.74626401e-03
1.13685822e+00 1.02375901e+00 8.29016685e-01 5.49140573e-01
-1.37651443e-01 2.87251502e-01 -2.17317134e-01 2.92751133e-01
-1.02624714e-01 -1.42273605e-01 -5.97322047e-01 3.65124404e-01
-3.26539040e-01 6.24246821e-02 3.62132713e-02 1.15277421e+00
2.74826497e-01 -2.17145652e-01 6.25416517e-01 -1.06179357e+00
-1.15679789e+00 5.51200449e-01 -8.06116045e-01 -4.21345800e-01
-1.71665132e-01 3.25299144e-01 6.67919815e-01 -4.49410290e-01
-6.97157951e-03 9.06962693e-01 4.98662025e-01 2.19158232e-01
-1.61018980e+00 -4.55579251e-01 1.02536725e-02 4.85644162e-01
-1.31362295e+00 1.53739616e-01 6.87671840e-01 -8.05202186e-01
8.60611796e-02 2.97600389e-01 5.98919809e-01 7.62794793e-01
5.50182834e-02 5.62889218e-01 1.20036387e+00 -3.34914148e-01
1.22542255e-01 7.85164610e-02 -2.13007573e-02 3.08020115e-01
4.38763440e-01 2.01196104e-01 9.62592363e-02 -1.81753822e-02
7.59031117e-01 1.13284476e-02 -5.12476861e-01 -3.24172884e-01
-7.55428672e-01 6.94511116e-01 3.66315275e-01 4.38793838e-01
-6.22863710e-01 -3.34938824e-01 1.84928507e-01 7.58224055e-02
1.26716271e-01 7.76064098e-01 -1.13244200e+00 1.68102592e-01
-7.12754726e-01 -1.93733856e-01 7.97368646e-01 6.71730042e-01
8.58194113e-01 2.99296141e-01 1.22178547e-01 1.13639879e+00
3.19923520e-01 5.58970392e-01 4.73985940e-01 -7.20333755e-01
-2.71163821e-01 5.12412608e-01 9.95227247e-02 -9.57979500e-01
-4.18835819e-01 -2.49874547e-01 -1.08269417e+00 2.77406484e-01
5.93777835e-01 -3.28216583e-01 -5.13449788e-01 1.40651023e+00
-2.77528204e-02 7.65442550e-02 8.05689692e-02 5.79796195e-01
5.89366496e-01 8.04393470e-01 3.89355302e-01 -3.63343924e-01
1.31456900e+00 -6.59969866e-01 -3.46677899e-01 -2.88622946e-01
4.05808002e-01 -7.63434172e-01 9.15668190e-01 5.82981467e-01
-5.76164663e-01 -7.44016886e-01 -7.68806934e-01 4.39466536e-01
-4.74362582e-01 7.48961985e-01 8.49581420e-01 8.62234056e-01
-5.36556184e-01 9.23152924e-01 -7.53107429e-01 -5.06279588e-01
3.72816473e-01 2.78005779e-01 -3.82718712e-01 2.21281406e-02
-8.68366480e-01 9.15793240e-01 7.19329715e-01 4.09302622e-01
-6.47683144e-01 -6.93532109e-01 -3.98194462e-01 4.76209581e-01
5.63238740e-01 -1.55647188e-01 8.54687691e-01 -1.05831134e+00
-2.04459882e+00 6.98397040e-01 -6.33160248e-02 1.29277632e-01
-2.41584495e-01 1.39067948e-01 -4.90216643e-01 -1.37241110e-01
-2.89976925e-01 3.08111936e-01 5.59267282e-01 -1.44136882e+00
-4.28400010e-01 -7.67847121e-01 -3.47960323e-01 -2.95910925e-01
-3.75761211e-01 -4.29467373e-02 3.01955581e-01 -1.90969899e-01
4.41674858e-01 -1.17181253e+00 -2.06360161e-01 2.02682242e-01
-2.90202320e-01 1.86293915e-01 1.00502479e+00 -1.01609707e+00
3.99745226e-01 -2.24554086e+00 8.93349722e-02 -8.85038376e-02
-9.96430367e-02 7.38659859e-01 -3.71086359e-01 3.44596416e-01
-3.27201664e-01 3.05983216e-01 -2.73747802e-01 6.22817338e-01
-4.08769488e-01 2.23094851e-01 1.33365598e-02 3.12929660e-01
5.01035035e-01 7.52135575e-01 -7.01778173e-01 -5.21199685e-03
3.82705122e-01 3.51865828e-01 5.71279302e-02 1.06278129e-01
-1.01333104e-01 7.62496591e-01 -4.68172848e-01 9.70808089e-01
8.88204277e-01 -1.02995194e-01 2.44846359e-01 -1.80213928e-01
-5.02368212e-01 -3.87292415e-01 -8.86207879e-01 8.72809172e-01
-3.70393366e-01 4.76509809e-01 3.19279045e-01 -1.33783376e+00
1.44169915e+00 5.03709495e-01 6.73499465e-01 1.33473808e-02
-8.50378349e-02 3.52367520e-01 1.94770053e-01 -4.24146324e-01
2.95922831e-02 -8.30444843e-02 6.22557402e-01 -3.13457727e-01
-4.65701707e-02 -2.10756496e-01 -1.00009598e-01 -8.66737843e-01
7.08071530e-01 2.39911273e-01 4.36746836e-01 -2.59869307e-01
7.30486572e-01 6.74644634e-02 7.69546688e-01 1.64274514e-01
-4.45967197e-01 3.18904638e-01 7.32556641e-01 1.45491036e-02
-8.40390086e-01 -9.19344902e-01 -4.14397359e-01 9.26549137e-01
-9.27659050e-02 3.36080968e-01 -2.19713345e-01 -4.92819473e-02
1.32716879e-01 7.72493541e-01 -3.07543695e-01 -2.97942430e-01
7.31016174e-02 -1.18366814e+00 4.47020531e-01 5.09725392e-01
4.85347122e-01 -1.13462806e+00 -3.47494751e-01 2.22248822e-01
4.45137098e-02 -1.04285574e+00 5.83332419e-01 6.12971485e-01
-1.06347954e+00 -1.02392566e+00 -6.73072755e-01 -5.79715490e-01
4.00904298e-01 2.43341938e-01 6.19070351e-01 -2.58541226e-01
-2.52737671e-01 -2.40851372e-01 -6.30415738e-01 -7.47193038e-01
-4.27282035e-01 1.96195796e-01 -2.77506888e-01 -9.00585875e-02
3.68941516e-01 -8.64470065e-01 -2.83245772e-01 3.86457115e-01
-6.11810744e-01 -1.55682206e-01 8.81064415e-01 8.75504017e-01
7.32618153e-01 6.08589530e-01 3.74505430e-01 -8.44882607e-01
3.75848681e-01 -5.58773339e-01 -9.64323997e-01 5.71303964e-01
-5.08072913e-01 -3.14035803e-01 9.91281033e-01 -5.70642352e-01
-8.66313994e-01 2.58000225e-01 -4.14012596e-02 -1.83975380e-02
-8.03240359e-01 8.48375022e-01 -5.77558160e-01 -1.94696471e-01
5.98043799e-01 3.65382850e-01 -6.24690810e-03 -3.95053953e-01
7.04647899e-02 5.05397439e-01 3.59650224e-01 -7.10011542e-01
8.94262850e-01 9.80046690e-02 4.65579927e-01 -1.14733529e+00
-6.55579805e-01 -4.14279521e-01 -8.47165942e-01 -1.59220055e-01
1.02199650e+00 -5.16595423e-01 -1.16160488e+00 5.90995491e-01
-8.00502062e-01 -5.10900557e-01 -1.12977766e-01 7.31183350e-01
-7.05655098e-01 4.49192554e-01 -5.24328589e-01 -9.36494291e-01
-1.94442272e-01 -1.18926752e+00 6.43337071e-01 5.34167528e-01
4.93785553e-02 -9.31319833e-01 -6.06561936e-02 9.46444646e-02
4.38305140e-01 3.84871721e-01 1.36980784e+00 -4.15221125e-01
-1.53038532e-01 -2.02020586e-01 -5.88721693e-01 5.92007697e-01
4.23608154e-01 8.45587730e-01 -1.14407444e+00 -1.92971509e-02
1.35482669e-01 -2.30365902e-01 3.31469536e-01 8.03137004e-01
1.53608525e+00 9.71974060e-02 -8.90245438e-02 6.21832430e-01
1.58348179e+00 5.66147745e-01 7.37700045e-01 2.94289708e-01
6.40054345e-01 1.13505161e+00 5.43690026e-01 2.79355109e-01
-2.49802113e-01 2.37293422e-01 7.12953806e-01 -2.98933964e-02
3.93245012e-01 -1.08935490e-01 1.72381476e-01 5.50302148e-01
-3.00577044e-01 -1.33153990e-01 -1.03278732e+00 2.37517804e-01
-1.79978824e+00 -9.94793832e-01 -6.88823938e-01 2.11295533e+00
5.02097607e-01 -3.64701331e-01 -2.13501319e-01 3.87223005e-01
9.75901484e-01 -4.84998338e-02 -6.89331591e-01 -3.86667281e-01
-4.40023661e-01 1.23813137e-01 8.28193963e-01 -7.70727322e-02
-8.42655897e-01 7.98738480e-01 5.84046221e+00 4.97611701e-01
-1.42304969e+00 -2.80327767e-01 7.55928695e-01 5.90823889e-01
1.72864377e-01 4.25424069e-01 -6.23362720e-01 2.43339956e-01
9.20810819e-01 -1.13476515e-02 6.32310510e-01 9.00558114e-01
5.47705829e-01 -1.54168740e-01 -6.97556615e-01 5.96665561e-01
-3.94845724e-01 -7.78407574e-01 -4.70052958e-01 1.51628703e-01
6.69398725e-01 -1.57251716e-01 -8.49889778e-03 3.15491974e-01
3.16368371e-01 -9.95899439e-01 3.06825675e-02 4.91972506e-01
4.64088947e-01 -4.34718788e-01 6.80182517e-01 5.53521156e-01
-1.08603036e+00 -3.18728805e-01 -8.86490226e-01 6.97768852e-02
-1.92229539e-01 8.59242558e-01 -6.70553029e-01 7.78893173e-01
5.48570037e-01 5.21057904e-01 -7.01282680e-01 1.07023346e+00
-4.90145981e-01 9.03434038e-01 -6.07328042e-02 8.30567256e-02
-6.25961497e-02 -6.13063991e-01 1.69347778e-01 5.49505949e-01
7.50945032e-01 -1.36345491e-01 8.13551247e-02 1.11681795e+00
4.96073514e-01 2.99678147e-01 -2.55737066e-01 -5.33574760e-01
3.99713904e-01 1.50574279e+00 -7.23633230e-01 2.48239636e-01
-3.36730659e-01 6.86572671e-01 -1.17915489e-01 4.21485782e-01
-4.24226731e-01 -2.40694918e-02 4.13442791e-01 1.19797759e-01
4.56922024e-01 -2.06434816e-01 -5.71471691e-01 -7.32370019e-01
-3.51598710e-01 -6.59523010e-01 4.63995151e-02 -1.07026386e+00
-1.38719261e+00 1.13684721e-01 -2.26146176e-01 -6.90183342e-01
3.05762112e-01 -1.17711008e+00 -6.27359986e-01 1.36999559e+00
-1.58869517e+00 -1.16463172e+00 -6.18747830e-01 1.27074584e-01
1.13683864e-01 -1.47413254e-01 1.19867980e+00 4.47733402e-02
-8.70310545e-01 -6.89976290e-02 5.57743013e-01 -1.27554893e-01
5.61422467e-01 -1.21042383e+00 -1.66087478e-01 5.86583674e-01
-3.03885549e-01 1.08064879e-02 5.82497358e-01 -5.19207299e-01
-1.18307233e+00 -1.13137960e+00 5.94307482e-01 1.93512872e-01
9.11589205e-01 3.97556275e-01 -1.26684260e+00 5.21277666e-01
-3.19752157e-01 5.19006792e-03 9.38423991e-01 3.78793001e-01
-4.02516015e-02 -2.14442372e-01 -1.12570167e+00 4.89631832e-01
2.44652987e-01 -3.60639095e-01 1.62319809e-01 3.65175545e-01
3.17246288e-01 9.55520421e-02 -1.30603218e+00 6.97609127e-01
5.50760686e-01 -5.14227450e-01 7.32445240e-01 -6.21221721e-01
6.15168452e-01 -3.70673954e-01 -2.40081280e-01 -1.46875298e+00
-9.32166934e-01 -1.50835350e-01 2.58935750e-01 1.27635741e+00
3.94030839e-01 -7.39006400e-01 6.15521610e-01 6.09736264e-01
5.68267331e-02 -4.72086757e-01 -2.49976560e-01 -8.54207098e-01
3.72248203e-01 -9.31545645e-02 6.11338973e-01 9.55370426e-01
-2.57396817e-01 5.61091416e-02 -1.30502477e-01 7.24601984e-01
1.70770615e-01 2.35722326e-02 6.02625787e-01 -1.94477820e+00
-5.36103904e-01 -7.03404963e-01 -3.34448159e-01 -3.45211089e-01
3.00438702e-01 -6.35219395e-01 -4.94076274e-02 -1.36981153e+00
-3.70469168e-02 -4.60896432e-01 -1.06622957e-01 5.18671453e-01
-4.08848852e-01 -5.21745741e-01 1.21707454e-01 1.61590219e-01
4.68901843e-01 4.66019154e-01 1.36699319e+00 -3.27112861e-02
-3.42998296e-01 5.53279161e-01 -7.21559286e-01 7.10789025e-01
1.32636833e+00 -2.36705795e-01 -3.35820138e-01 -2.32863799e-01
2.19288226e-02 1.63163826e-01 5.58916688e-01 -8.84081304e-01
-1.14736699e-01 -7.52563059e-01 4.23427343e-01 -4.55414355e-01
1.62044913e-01 -9.54099774e-01 3.60369831e-01 6.08792305e-01
-4.82058264e-02 -3.73597145e-02 5.10756016e-01 4.32788670e-01
-6.57822266e-02 -9.05056655e-01 7.84866273e-01 -7.36958832e-02
-6.08666420e-01 3.35110694e-01 -7.01206088e-01 -6.05794907e-01
1.18970132e+00 -2.86905795e-01 -3.77121806e-01 -5.72017878e-02
-7.59012640e-01 2.95870355e-03 2.87404001e-01 7.24629462e-02
9.50803906e-02 -9.07556653e-01 -8.02218318e-01 -3.22591066e-02
1.07001916e-01 -1.22348718e-01 3.26767325e-01 9.12670970e-01
-8.48720372e-01 4.53585416e-01 -7.42329717e-01 -5.23423254e-01
-1.16393387e+00 6.79098487e-01 2.63242245e-01 -1.23751804e-01
-9.51703545e-03 4.76245284e-01 1.07142173e-01 -6.35605514e-01
-3.67181569e-01 -6.66517988e-02 -4.15449083e-01 -5.71001656e-02
9.95147601e-02 5.27621984e-01 -1.08881190e-01 -4.08824563e-01
7.24424720e-02 2.32439831e-01 3.28875095e-01 4.05536652e-01
1.42601728e+00 2.34748900e-01 -3.77398044e-01 8.92964005e-01
7.66900420e-01 -2.86629856e-01 -1.09789133e+00 3.81575823e-02
1.48489848e-01 -3.10093701e-01 8.21529478e-02 -7.72976637e-01
-1.01546216e+00 9.23702776e-01 5.96777499e-01 6.52610958e-01
1.27481496e+00 -3.70030165e-01 2.42420390e-01 4.48408723e-01
5.15769795e-02 -9.26922023e-01 -5.93389273e-01 4.60415214e-01
7.33760238e-01 -1.26613688e+00 -2.36262172e-01 -7.96821594e-01
-4.99555111e-01 1.54132867e+00 5.54121137e-01 2.82285899e-01
8.46549273e-01 9.39482301e-02 -8.26644897e-02 3.62295881e-02
-5.84712863e-01 -3.21417153e-01 7.87505601e-03 1.04849088e+00
6.86047852e-01 5.69750309e-01 -2.02070057e-01 3.67139518e-01
-1.57533243e-01 7.67889321e-02 4.25930947e-01 4.74499971e-01
-6.00612998e-01 -1.15264535e+00 -7.26708889e-01 5.10509253e-01
-1.29963960e-02 7.57306144e-02 -4.87654209e-01 5.20943522e-01
1.82189763e-01 1.04062104e+00 -4.17933255e-01 -1.19286537e-01
2.48957813e-01 2.03705579e-01 3.40465665e-01 -8.77263725e-01
-2.14600548e-01 -2.10194271e-02 -3.12668592e-01 1.42217919e-01
-3.27380896e-01 -8.80598426e-01 -6.62462592e-01 -4.00377065e-01
-7.57753432e-01 -1.92526475e-01 8.77548814e-01 9.23642278e-01
9.61679034e-03 7.48177707e-01 7.79788315e-01 -5.23313046e-01
-6.83561385e-01 -1.03116906e+00 -9.87217069e-01 -7.17739090e-02
-4.10465121e-01 -5.63020706e-01 -2.06780821e-01 8.07526782e-02] | [9.184743881225586, -1.566736102104187] |
cd34e8ef-3fb4-4fa7-b6cb-456dbdd3b76a | uav-crowd-violent-and-non-violent-crowd | 2208.06702 | null | https://arxiv.org/abs/2208.06702v1 | https://arxiv.org/pdf/2208.06702v1.pdf | UAV-CROWD: Violent and non-violent crowd activity simulator from the perspective of UAV | Unmanned Aerial Vehicle (UAV) has gained significant traction in the recent years, particularly the context of surveillance. However, video datasets that capture violent and non-violent human activity from aerial point-of-view is scarce. To address this issue, we propose a novel, baseline simulator which is capable of generating sequences of photo-realistic synthetic images of crowds engaging in various activities that can be categorized as violent or non-violent. The crowd groups are annotated with bounding boxes that are automatically computed using semantic segmentation. Our simulator is capable of generating large, randomized urban environments and is able to maintain an average of 25 frames per second on a mid-range computer with 150 concurrent crowd agents interacting with each other. We also show that when synthetic data from the proposed simulator is augmented with real world data, binary video classification accuracy is improved by 5% on average across two different models. | ['Mayamin Hamid Raha', 'Shahriar Ali Bijoy', 'Tonmoay Deb', 'Mahieyin Rahmun'] | 2022-08-13 | null | null | null | null | ['video-classification'] | ['computer-vision'] | [ 2.51239210e-01 -8.62348229e-02 5.94068229e-01 -1.64194591e-02
-1.84190258e-01 -1.00865257e+00 8.29157770e-01 -2.32855812e-01
-5.74929476e-01 8.88059795e-01 -1.37443259e-01 -1.19894736e-01
2.76484847e-01 -7.15308905e-01 -5.90817928e-01 -5.92391193e-01
-3.34599733e-01 3.37108076e-01 7.05783129e-01 -2.68465191e-01
-2.29424343e-01 6.99094236e-01 -1.78560472e+00 1.51882231e-01
6.08185172e-01 8.47178698e-01 2.79668793e-02 1.30976093e+00
8.06178749e-01 9.95322049e-01 -1.03479052e+00 -3.87686223e-01
8.56447041e-01 -2.92389512e-01 -5.46227336e-01 8.36870432e-01
5.27807951e-01 -8.35783005e-01 -4.09037769e-01 7.95425594e-01
3.58927161e-01 3.61090213e-01 5.21104336e-01 -1.66258013e+00
1.95534691e-01 -1.78725600e-01 -6.51597738e-01 3.06712389e-01
8.01864982e-01 8.60114276e-01 3.05848956e-01 -3.30219179e-01
7.60486662e-01 1.15969694e+00 4.77752954e-01 6.09073699e-01
-7.85272300e-01 -5.50222993e-01 6.98345825e-02 -2.17310473e-01
-1.48230088e+00 -2.15665668e-01 3.38314176e-01 -6.64541006e-01
7.51175106e-01 1.91298112e-01 1.19434726e+00 1.23182631e+00
2.53445327e-01 4.19818670e-01 7.47215509e-01 7.61817023e-03
4.34427917e-01 -1.77546144e-01 -6.88507736e-01 7.26899564e-01
5.61760008e-01 1.60484642e-01 -2.01443881e-01 -5.15697241e-01
8.63692999e-01 -2.39180345e-02 -4.48487282e-01 -3.93566519e-01
-1.42941141e+00 6.93090916e-01 9.36700925e-02 -1.66628137e-01
-6.40989721e-01 1.13688223e-01 2.99098819e-01 -5.23545966e-02
4.10535663e-01 4.11238343e-01 -2.22249329e-02 -8.16269815e-02
-1.13040209e+00 8.40229332e-01 7.05608666e-01 1.22537553e+00
3.32623303e-01 3.55820209e-01 -1.12160347e-01 5.19435443e-02
-3.71306809e-03 9.45895076e-01 1.30680710e-01 -1.52051640e+00
3.57297093e-01 6.61918879e-01 8.71498883e-01 -1.09419191e+00
-2.54595995e-01 2.45375082e-01 -6.94449425e-01 5.33071458e-01
2.54768252e-01 -8.59133840e-01 -7.07469821e-01 1.27128661e+00
6.34126663e-01 4.94534254e-01 3.61566961e-01 9.76781607e-01
5.55145264e-01 7.77319252e-01 7.26858303e-02 -5.21161735e-01
1.19122553e+00 -7.82183290e-01 -3.83962214e-01 -2.99936056e-01
3.99541229e-01 -4.71970975e-01 6.30460322e-01 3.99079531e-01
-1.17615616e+00 -3.25913578e-01 -8.87249231e-01 6.42530024e-01
-7.30305165e-02 -4.25482094e-02 2.69478887e-01 6.21155798e-01
-1.22019780e+00 2.25679547e-01 -9.20435190e-01 -4.49944884e-01
3.70236039e-01 3.05130363e-01 -6.20784044e-01 -4.55607288e-02
-8.60656559e-01 5.80213308e-01 1.46627590e-01 -3.02538842e-01
-1.55126476e+00 -2.26066932e-01 -9.23518002e-01 -3.96900743e-01
4.20999020e-01 -8.49821985e-01 1.24744594e+00 -1.19038486e+00
-1.23399127e+00 7.72069871e-01 -2.21989281e-03 -6.96107686e-01
1.05852759e+00 -1.16995268e-01 -1.70382693e-01 6.64153337e-01
1.55007049e-01 8.15990925e-01 8.08229089e-01 -1.34537435e+00
-8.58892560e-01 -3.46589774e-01 5.33569872e-01 5.11354506e-01
1.32513404e-01 3.19177508e-01 -1.23340242e-01 -5.88893652e-01
-7.10920036e-01 -1.39017272e+00 -4.38788772e-01 8.54213014e-02
-2.17061549e-01 3.44549537e-01 1.10532796e+00 -3.62422138e-01
7.76731253e-01 -1.78015029e+00 1.20894596e-01 -2.75044292e-01
2.20334604e-01 5.86652577e-01 -6.06663935e-02 2.81620830e-01
3.28467578e-01 1.24025367e-01 -3.86173755e-01 -1.58144042e-01
-5.43781698e-01 8.66995901e-02 -2.15409189e-01 5.40828466e-01
8.41890275e-02 4.96992439e-01 -1.22493994e+00 -5.21688998e-01
4.12782937e-01 2.92850226e-01 -5.55568814e-01 5.41813552e-01
-1.52661845e-01 7.57299960e-01 -2.36446857e-01 7.66296625e-01
7.49110878e-01 1.54963225e-01 1.58960432e-01 2.75469184e-01
-1.19957320e-01 -7.73975611e-01 -1.02850497e+00 1.09244657e+00
6.68444932e-02 8.62476647e-01 3.16169381e-01 -2.40244016e-01
4.15135533e-01 5.07627666e-01 5.87588787e-01 2.77680039e-01
5.09784818e-01 -2.46829152e-01 -1.26089334e-01 -4.60732013e-01
9.24110591e-01 6.14885166e-02 -8.40793699e-02 1.96708813e-01
-2.28650659e-01 -6.32113934e-01 5.26658893e-01 2.01833814e-01
1.42670119e+00 -9.67269018e-02 4.79352206e-01 1.02224305e-01
3.96250099e-01 6.50158465e-01 4.00627941e-01 5.39753437e-01
-6.33803606e-01 6.29686832e-01 3.01852673e-01 -6.60853088e-01
-1.08913052e+00 -1.01948500e+00 4.09810185e-01 5.21335363e-01
5.64820945e-01 -1.69694006e-01 -1.20041728e+00 -5.90035915e-01
-3.85270953e-01 6.42478108e-01 -4.37522680e-01 1.57923982e-01
-4.56934988e-01 -6.42915606e-01 7.61657178e-01 2.10447252e-01
8.04601252e-01 -1.06343067e+00 -1.53372300e+00 9.03437883e-02
-3.68422776e-01 -1.81753540e+00 -1.94498762e-01 -6.61347866e-01
-5.83084106e-01 -1.54858708e+00 -9.78864491e-01 -3.61651391e-01
8.68026972e-01 9.49404240e-01 1.03822732e+00 1.24399938e-01
-3.79558355e-01 9.23188627e-01 -5.85209787e-01 -5.70767939e-01
-4.71195310e-01 -4.63377953e-01 3.93490911e-01 1.14877209e-01
9.92751941e-02 -1.23182558e-01 -6.56467319e-01 6.38546824e-01
-1.23933172e+00 3.37878242e-02 -2.18179300e-02 2.86561072e-01
3.39558512e-01 1.33716837e-01 -2.85187274e-01 -2.54733860e-01
4.58070129e-01 -5.49769104e-01 -8.21123421e-01 -3.20301689e-02
4.56181526e-01 -8.04060519e-01 8.53842735e-01 -6.92952335e-01
-7.53723145e-01 3.24483365e-01 6.65062487e-01 -7.45383680e-01
-6.71302080e-01 -1.83612302e-01 1.31625995e-01 -2.65920222e-01
6.13202929e-01 3.57412472e-02 -2.31231347e-01 6.59177184e-01
-1.36810048e-02 7.36431181e-01 4.57483321e-01 -2.10560977e-01
1.03816164e+00 8.80876005e-01 5.26797995e-02 -1.40720260e+00
-3.11692148e-01 -2.00799361e-01 -6.84308112e-01 -8.54084074e-01
1.13214123e+00 -1.30269825e+00 -7.59504914e-01 7.71060050e-01
-1.27281880e+00 -6.53693676e-01 -1.77728906e-01 5.53271294e-01
-6.55707896e-01 4.54263657e-01 -3.26228976e-01 -1.16393018e+00
-6.62795007e-02 -1.38752890e+00 1.39957058e+00 3.70867729e-01
-3.07209849e-01 -6.24567330e-01 -8.50333869e-02 3.58699143e-01
4.33491170e-02 9.17829692e-01 4.47297236e-03 -2.27988228e-01
-7.91175783e-01 -3.74662220e-01 9.31068882e-02 2.02897325e-01
-6.32688776e-02 2.93393612e-01 -5.85880220e-01 -5.18120527e-01
-2.88192600e-01 -2.67874599e-01 1.67668045e-01 4.97742116e-01
5.64124405e-01 -2.73974419e-01 -3.03031236e-01 3.13771933e-01
1.14660406e+00 4.77605402e-01 5.75321555e-01 6.88148439e-02
4.76792246e-01 6.27721488e-01 1.00270987e+00 8.49754274e-01
3.37318331e-01 6.93423867e-01 8.22487116e-01 -6.19071245e-04
3.43938410e-01 8.13173968e-03 5.75028777e-01 -1.89488128e-01
-7.64332652e-01 -8.84631813e-01 -1.06792307e+00 6.24652028e-01
-1.67256284e+00 -1.42823994e+00 -1.96784034e-01 2.29880381e+00
1.74527336e-02 -2.46362552e-01 5.25032759e-01 -4.56612818e-02
9.31842923e-01 2.67933309e-01 -3.68932307e-01 2.51881890e-02
-1.15810754e-03 -4.66218442e-01 7.10545301e-01 5.42892814e-01
-1.34001911e+00 1.19124818e+00 6.35250473e+00 2.15254173e-01
-9.10647988e-01 -3.29427898e-01 4.62671012e-01 -3.95754308e-01
4.15159851e-01 -1.98129326e-01 -4.91704524e-01 4.65552598e-01
8.58317494e-01 -3.45472842e-01 4.80710655e-01 8.47824156e-01
6.24984801e-01 -6.20833814e-01 -5.47830284e-01 9.79955971e-01
3.80152315e-01 -1.06204951e+00 3.00573081e-01 2.76566803e-01
1.02939558e+00 -6.35059699e-02 -1.45648420e-01 -2.25544628e-02
7.39752173e-01 -9.13457692e-01 9.12535489e-01 3.32748383e-01
7.60361552e-01 -9.07579482e-01 9.46212709e-01 6.68067992e-01
-1.34913719e+00 -1.10830754e-01 -3.84361148e-01 -2.90515423e-01
5.69075465e-01 -1.69661447e-01 -8.55094314e-01 4.03980047e-01
7.82178640e-01 4.83758569e-01 -5.03668129e-01 9.42632437e-01
6.84506819e-02 3.03647965e-01 -1.65374577e-01 -6.74091280e-02
4.96861666e-01 -2.90098727e-01 9.59475756e-01 8.25904429e-01
7.56803215e-01 6.14722490e-01 6.09078884e-01 2.67705858e-01
2.36626193e-01 -2.93899715e-01 -1.34005058e+00 1.25478446e-01
3.22102010e-01 1.05550408e+00 -8.71811092e-01 -5.06243467e-01
-8.90316665e-02 1.22881699e+00 -3.59082609e-01 3.20837289e-01
-1.32859814e+00 -2.28314281e-01 9.65335965e-01 4.52823251e-01
1.01982459e-01 -5.68273067e-01 5.29364467e-01 -1.12234759e+00
-1.18549004e-01 -8.71751606e-01 -4.14391570e-02 -1.18794167e+00
-6.04277849e-01 1.15848136e+00 4.21574414e-01 -1.68760765e+00
-4.63715971e-01 -5.08712411e-01 -4.24027473e-01 3.50040704e-01
-8.58790994e-01 -1.11278522e+00 -1.12158132e+00 6.40785277e-01
8.76753032e-01 -4.90297467e-01 6.21766984e-01 -3.03152651e-01
-3.05889249e-01 -1.21439129e-01 -5.43275774e-01 2.10007250e-01
3.83104295e-01 -7.90043712e-01 5.27937889e-01 1.12667131e+00
-1.81000397e-01 1.18671864e-01 9.85178769e-01 -8.62417996e-01
-9.85993981e-01 -1.65248501e+00 2.57912725e-01 -7.89416015e-01
3.65333974e-01 -6.85449019e-02 -2.45321080e-01 7.05510080e-01
3.07935655e-01 3.07324857e-01 4.75732177e-01 -1.21097219e+00
3.16284448e-01 3.00965697e-01 -1.53078663e+00 9.16060030e-01
1.17738497e+00 1.46191707e-03 -2.51834214e-01 3.86162579e-01
5.07311642e-01 -5.15153944e-01 -4.50675577e-01 3.57278615e-01
3.27605814e-01 -1.46794486e+00 8.95830750e-01 -4.45241660e-01
4.03217614e-01 -6.34532213e-01 -2.83109128e-01 -1.42370284e+00
7.51042813e-02 -8.15588117e-01 2.53366828e-01 6.39908195e-01
-9.07190219e-02 -3.59490126e-01 7.80334234e-01 6.05360150e-01
2.44121566e-01 -2.27634445e-01 -6.87354445e-01 -9.44099426e-01
-3.88025612e-01 -3.06061387e-01 2.87408501e-01 5.16878366e-01
-2.84882098e-01 9.29844659e-03 -7.16102540e-01 5.29431522e-01
5.24512529e-01 -4.28277940e-01 1.34377122e+00 -1.35941005e+00
4.31278199e-02 8.70023072e-02 -8.77249539e-01 -6.90282583e-01
1.67918220e-01 -1.36570036e-01 6.77094832e-02 -1.39955163e+00
9.72241461e-02 3.63587499e-01 8.05065453e-01 5.67537546e-02
2.51959153e-02 4.62885231e-01 3.62009674e-01 9.12077501e-02
-8.38716149e-01 3.37502778e-01 1.05070877e+00 2.16741650e-03
-1.11618020e-01 4.13078554e-02 -6.02796003e-02 1.15503550e+00
7.36166775e-01 -3.07325900e-01 -5.46700239e-01 -2.50297070e-01
-3.30047250e-01 5.64258337e-01 7.65552580e-01 -1.67048407e+00
-3.70345376e-02 -4.31303740e-01 3.42614859e-01 -2.65905023e-01
6.43817067e-01 -1.06391191e+00 4.30554152e-01 6.26176357e-01
6.44086674e-02 5.55154443e-01 2.89123625e-01 8.46790254e-01
-1.70135558e-01 5.79642132e-02 9.65026140e-01 -5.93517959e-01
-6.05861425e-01 5.35749793e-01 -1.08768010e+00 3.22806269e-01
1.98120701e+00 -3.72026950e-01 -1.45198449e-01 -7.35504329e-01
-3.33558202e-01 6.44202530e-02 1.22548258e+00 3.00298631e-01
6.72293782e-01 -8.64283204e-01 -8.10825706e-01 9.92649943e-02
1.18596315e-01 1.33941993e-02 3.79485905e-01 4.68968093e-01
-1.31220388e+00 2.51121759e-01 -2.78891355e-01 -6.01150751e-01
-1.85278487e+00 4.02597576e-01 2.37163067e-01 -1.13654852e-01
-2.85724819e-01 5.29884279e-01 1.53159901e-01 -1.30730063e-01
-7.43717030e-02 -8.45395178e-02 -1.92423999e-01 -1.56436339e-01
8.40412796e-01 6.46970570e-01 -4.74410892e-01 -1.41749990e+00
-4.03491288e-01 5.90618134e-01 5.09158850e-01 -3.81218046e-01
9.47963774e-01 5.66999009e-03 3.16411108e-01 -2.17531566e-02
3.87151569e-01 2.68572476e-02 -1.77277040e+00 4.24601912e-01
-5.48870564e-01 -6.98164940e-01 -2.65870333e-01 -2.37685949e-01
-8.23801875e-01 5.28056502e-01 4.99060780e-01 2.73241818e-01
1.03285360e+00 -3.33722919e-01 7.60953844e-01 5.20991027e-01
1.18278515e+00 -7.30339885e-01 1.41095594e-01 4.42460060e-01
8.34002912e-01 -1.26202393e+00 -1.56019494e-01 -4.37779546e-01
-1.24427319e+00 8.13648880e-01 7.26518452e-01 -4.18113619e-01
8.99961144e-02 3.31125468e-01 1.50682837e-01 3.61853950e-02
-6.72318757e-01 -3.72592032e-01 -2.31131911e-01 1.12252855e+00
-1.87026769e-01 2.06437394e-01 1.92072198e-01 1.34359837e-01
-1.81367204e-01 2.80489996e-02 1.22824633e+00 1.11571562e+00
-5.09228587e-01 -3.50444138e-01 -7.02527106e-01 9.60645750e-02
-4.00380522e-01 1.99979335e-01 -6.53110147e-01 9.01042581e-01
1.54385135e-01 1.46759856e+00 1.40066862e-01 -4.76953536e-01
3.79361480e-01 -4.66441453e-01 3.82038236e-01 -4.89139915e-01
-5.93299627e-01 -2.90813267e-01 1.40381500e-01 -5.68156183e-01
-6.39282286e-01 -6.00073040e-01 -9.07742023e-01 -4.35991794e-01
1.10942960e-01 1.43620029e-01 2.04553008e-01 6.36172354e-01
2.25249350e-01 2.13653177e-01 6.27942383e-01 -1.78735638e+00
-4.00911346e-02 -6.77894175e-01 -4.63637352e-01 3.55260313e-01
2.56285429e-01 -8.21549833e-01 -6.60696387e-01 4.82860774e-01] | [7.412619113922119, -1.3567650318145752] |
043a2992-cd29-47a7-9c5c-4486f32125df | person-recognition-using-facial-micro | 2306.13907 | null | https://arxiv.org/abs/2306.13907v1 | https://arxiv.org/pdf/2306.13907v1.pdf | Person Recognition using Facial Micro-Expressions with Deep Learning | This study investigates the efficacy of facial micro-expressions as a soft biometric for enhancing person recognition, aiming to broaden the understanding of the subject and its potential applications. We propose a deep learning approach designed to capture spatial semantics and motion at a fine temporal resolution. Experiments on three widely-used micro-expression databases demonstrate a notable increase in identification accuracy compared to existing benchmarks, highlighting the potential of integrating facial micro-expressions for improved person recognition across various fields. | ['David Mendlovic', 'Mor-Avi Azulay', 'Khen Cohen', 'Yuval Ringel', 'Tuval Kay'] | 2023-06-24 | null | null | null | null | ['person-recognition'] | ['computer-vision'] | [-2.19042245e-02 -5.75793564e-01 -4.55805302e-01 -6.41024530e-01
-2.55534500e-01 -1.59003735e-01 5.82509875e-01 -7.71227777e-01
-3.64380062e-01 5.55069089e-01 3.38826299e-01 4.99550372e-01
4.36337898e-03 -5.56141734e-01 -1.46508887e-01 -9.13558125e-01
-1.06714360e-01 -2.75546134e-01 -6.25392437e-01 -2.01100543e-01
-1.77626699e-01 8.58491123e-01 -1.69927359e+00 2.43327439e-01
1.25277862e-01 1.36222768e+00 -7.65467823e-01 3.40518117e-01
1.91503748e-01 6.68668985e-01 -6.90451503e-01 -9.27167654e-01
4.99051325e-02 -9.37315598e-02 -6.16257906e-01 -5.04981205e-02
7.66625345e-01 -4.39405411e-01 -6.52140856e-01 1.11954236e+00
8.52187872e-01 8.24014023e-02 3.81129652e-01 -1.02572250e+00
-1.01646459e+00 -1.46295279e-01 -6.55589342e-01 4.64205950e-01
9.99928474e-01 2.55346060e-01 5.35213530e-01 -7.99890339e-01
5.60147285e-01 1.26125681e+00 1.00298536e+00 1.00076306e+00
-7.31215179e-01 -7.59128869e-01 -3.08513463e-01 5.15015006e-01
-1.72576416e+00 -7.74714828e-01 7.12091088e-01 -3.04847687e-01
1.10002697e+00 1.87608004e-01 5.99122107e-01 1.69516599e+00
7.52336229e-04 6.93825841e-01 1.19766021e+00 -3.37599725e-01
-1.30801290e-01 -8.43114406e-02 1.39539212e-01 7.16902018e-01
-1.83455460e-02 6.44857407e-01 -1.03265560e+00 5.77805340e-02
8.67936015e-01 -3.09633035e-02 4.59797308e-02 3.26274008e-01
-8.88152719e-01 1.20628081e-01 2.41070449e-01 5.38524151e-01
-7.47259557e-01 3.90511334e-01 5.83959341e-01 2.00996488e-01
4.99463737e-01 3.03268433e-04 -9.04398598e-03 -8.78214657e-01
-8.95168841e-01 2.28969127e-01 3.47805560e-01 3.35673451e-01
5.33136129e-01 5.66698074e-01 -3.81992012e-01 9.76085961e-01
9.87101868e-02 8.83449495e-01 2.65872538e-01 -9.35047626e-01
-6.52282760e-02 6.87541187e-01 6.31453618e-02 -1.22129691e+00
-2.45388806e-01 -2.11031944e-01 -9.47254002e-01 2.68861681e-01
3.59688461e-01 -1.44050151e-01 -6.25872910e-01 1.84789038e+00
1.45545691e-01 6.35068774e-01 -1.70327067e-01 1.04786348e+00
8.64929378e-01 1.97391123e-01 4.54515219e-01 7.60565251e-02
1.28465068e+00 -3.90328348e-01 -8.83712411e-01 7.02497661e-02
3.84578049e-01 -1.76919460e-01 8.96390498e-01 3.84967178e-01
-8.24126303e-01 -7.69654691e-01 -7.99545467e-01 2.42335781e-01
-4.54450786e-01 3.61182660e-01 9.38359737e-01 1.39880764e+00
-1.13943994e+00 4.18787807e-01 -6.44489825e-01 -3.68557513e-01
7.94677913e-01 7.36222982e-01 -8.07947934e-01 9.47029963e-02
-1.30868340e+00 9.38131213e-01 -2.78245807e-01 3.91923606e-01
-3.79469931e-01 -7.16024816e-01 -9.28296268e-01 -2.51950532e-01
-2.66176403e-01 -5.64287722e-01 8.10807765e-01 -1.18218505e+00
-1.81741726e+00 1.37989557e+00 -6.71909153e-01 -2.33458862e-01
3.66564095e-01 -3.67618352e-01 -9.15191174e-01 1.59637287e-01
-1.84219807e-01 4.60095972e-01 9.91901875e-01 -4.42202747e-01
-1.52153850e-01 -9.90989745e-01 -2.12304890e-01 7.30173439e-02
-7.33333945e-01 5.51991224e-01 -4.74020094e-01 -8.20944846e-01
-6.71598613e-01 -8.13348234e-01 2.27962002e-01 1.78936857e-03
-1.15019746e-01 -4.67765778e-01 1.14271688e+00 -7.88240671e-01
1.03533399e+00 -2.22530627e+00 1.62761822e-01 3.50718170e-01
1.42600238e-01 5.90325058e-01 -4.94480580e-02 -2.09378228e-01
-7.57168091e-05 -7.40633458e-02 1.53598964e-01 -5.70680737e-01
6.45679384e-02 -8.86061601e-03 3.00014089e-03 6.78919256e-01
-1.86452121e-02 1.51003814e+00 -4.11738336e-01 -1.06751174e-01
4.51457024e-01 1.01039982e+00 6.64835721e-02 -1.42882522e-02
4.43356007e-01 4.35242385e-01 -2.52136320e-01 1.25133681e+00
7.30158329e-01 1.09326228e-01 -3.82520378e-01 -2.26990566e-01
2.76073337e-01 -2.67416030e-01 -9.42511261e-01 1.32698286e+00
-1.40956834e-01 9.50361192e-01 1.00235499e-01 -7.54415274e-01
9.98764515e-01 2.93273568e-01 9.16341662e-01 -1.09132338e+00
2.82931119e-01 -1.68384567e-01 -5.51072299e-01 -5.17158151e-01
3.54494333e-01 -2.42205709e-01 7.99269527e-02 2.01304853e-01
8.31396133e-02 8.19207489e-01 -3.62271637e-01 -5.37290573e-01
6.59504771e-01 1.29774839e-01 7.47662708e-02 -1.13552496e-01
8.55048060e-01 -5.64952135e-01 2.51493156e-01 4.48913932e-01
-9.27119255e-01 -5.67808412e-02 -1.16934337e-01 -6.59291208e-01
-6.78779960e-01 -1.03490138e+00 -2.66072273e-01 1.30491912e+00
1.53782636e-01 -2.25629166e-01 -8.92750740e-01 -4.88517106e-01
2.98366427e-01 -1.48504572e-02 -1.13177478e+00 -1.39641091e-01
-4.24886644e-01 -8.51491809e-01 1.52159417e+00 1.01313770e+00
1.00376821e+00 -1.06519973e+00 -6.49156511e-01 -2.76733279e-01
-9.83261988e-02 -1.49502957e+00 -3.27334523e-01 -7.53627241e-01
-4.04420108e-01 -7.60097802e-01 -1.09380376e+00 -5.40104270e-01
2.17680812e-01 3.55329588e-02 8.33061635e-01 -2.26873793e-02
-6.97852850e-01 9.15946603e-01 -4.10280451e-02 -2.71061242e-01
1.34719551e-01 -2.91852266e-01 7.78586924e-01 5.63965559e-01
1.23968005e+00 -4.11008775e-01 -5.06092191e-01 2.47222453e-01
-3.94257098e-01 -6.33882523e-01 2.36100763e-01 6.04944706e-01
1.41846702e-01 -4.27229494e-01 5.55833876e-01 -2.78817248e-02
8.06499541e-01 -8.67102295e-02 -1.72800764e-01 2.38383874e-01
-4.06420380e-01 -3.41271073e-01 8.50674883e-02 -4.16203499e-01
-1.37983608e+00 -1.75059617e-01 -3.77715945e-01 -4.07336891e-01
-5.12518048e-01 3.01332712e-01 -7.10781962e-02 -8.20004225e-01
5.65194428e-01 5.57066381e-01 4.04302090e-01 -1.30265266e-01
2.49902010e-01 6.62700355e-01 1.03147352e+00 -6.45425618e-01
4.00947779e-01 7.52612233e-01 2.15634793e-01 -1.48451054e+00
-1.83351308e-01 -6.66397929e-01 -8.23860228e-01 -4.55897182e-01
9.15369034e-01 -9.49176013e-01 -1.35461473e+00 1.01001561e+00
-9.11730051e-01 -1.96953356e-01 -9.95642841e-02 2.71693558e-01
-3.69266957e-01 5.89474320e-01 -8.52761149e-01 -1.03094387e+00
-6.24922872e-01 -8.04158688e-01 1.47432506e+00 4.70570594e-01
-5.60536861e-01 -1.09278679e+00 2.06423417e-01 3.74157459e-01
6.29069686e-01 2.45280728e-01 -2.34232098e-02 -3.07218749e-02
-1.62353620e-01 -4.58944023e-01 -3.50116223e-01 1.67681322e-01
1.72177568e-01 1.48958135e-02 -1.36260772e+00 -1.26667798e-01
-4.46913570e-01 -4.61272299e-01 6.32147729e-01 5.01967847e-01
1.15833426e+00 8.92732479e-03 -2.60075092e-01 1.00396979e+00
8.80411565e-01 7.20595196e-02 1.03448880e+00 3.53647918e-01
6.39385760e-01 7.86284745e-01 7.52848610e-02 7.40816593e-01
2.64766067e-01 1.21061242e+00 -5.46581112e-02 -6.92426711e-02
-7.04463497e-02 1.55559823e-01 5.40734351e-01 -2.43051890e-02
-7.61133969e-01 4.12985593e-01 -8.80663335e-01 3.02909672e-01
-1.46676147e+00 -1.53148723e+00 3.02924037e-01 1.69633698e+00
4.98421699e-01 -6.82369351e-01 5.68562031e-01 5.29376268e-02
4.81034815e-01 4.37975615e-01 -5.96086621e-01 -3.51487041e-01
-7.50020862e-01 4.89558727e-01 1.02941841e-01 2.08807543e-01
-1.32808018e+00 1.06630945e+00 7.71883535e+00 4.23525363e-01
-1.64446294e+00 -7.49546438e-02 6.44032240e-01 -1.10553555e-01
3.97844881e-01 -9.96078253e-01 -1.01860297e+00 3.26900780e-01
1.05274272e+00 -1.52141109e-01 3.57329696e-01 8.60435069e-01
2.67743498e-01 3.32389534e-01 -8.83377910e-01 1.71520925e+00
4.19564575e-01 -1.49613559e+00 -5.97099066e-02 3.14897239e-01
5.72552860e-01 -2.89719313e-01 5.72552204e-01 2.24438414e-01
-2.71620095e-01 -1.59912860e+00 3.92993808e-01 7.62728453e-01
1.19708097e+00 -5.99227548e-01 8.59640896e-01 -1.85930341e-01
-1.34909880e+00 -3.63949388e-02 -7.73046613e-02 -3.64388674e-01
1.48801312e-01 -6.75905496e-02 -4.03071165e-01 3.95830184e-01
1.01439834e+00 9.96989191e-01 -5.75176597e-01 4.97885674e-01
1.05771974e-01 4.24433380e-01 -3.19772422e-01 -1.33469105e-01
-5.11083566e-03 -5.53131364e-02 4.07708943e-01 1.90035653e+00
2.02839240e-01 1.33648321e-01 -3.18582624e-01 1.04383421e+00
-1.69073101e-02 -5.37534133e-02 -7.07697690e-01 -1.37791291e-01
2.41120815e-01 9.68228400e-01 1.09289259e-01 -2.06095994e-01
-4.27717149e-01 1.44214904e+00 -7.07807690e-02 5.83728254e-01
-7.51232028e-01 -1.46170437e-01 1.44097686e+00 -3.58051658e-02
-1.76329926e-01 -3.17409486e-01 -6.59931898e-01 -1.22530305e+00
1.40192717e-01 -1.00743806e+00 4.42429334e-01 -6.32824957e-01
-1.26047266e+00 5.51321626e-01 -2.34714463e-01 -7.41207659e-01
-5.76544344e-01 -9.52992857e-01 -4.00369465e-01 1.16682971e+00
-1.14238858e+00 -1.57238972e+00 -8.41692507e-01 9.50869560e-01
6.56242892e-02 -6.41538501e-01 1.08059549e+00 4.58017111e-01
-7.10291207e-01 1.32414377e+00 -3.02752048e-01 5.29742658e-01
6.34240568e-01 -7.73210764e-01 4.07954067e-01 8.53050828e-01
1.89390510e-01 8.52728724e-01 3.13413024e-01 -2.92922974e-01
-1.70711374e+00 -6.92056119e-01 5.25352478e-01 -8.64386857e-01
2.77419221e-02 8.22990667e-03 -7.68041551e-01 5.82948506e-01
3.17488089e-02 2.45511636e-01 1.12179971e+00 1.80065244e-01
-6.29326463e-01 -2.89189577e-01 -1.36639237e+00 5.08079827e-01
1.22962832e+00 -1.25859714e+00 -2.13546008e-01 -4.23150390e-01
-2.99741924e-01 -1.65777385e-01 -1.07549167e+00 6.44066572e-01
1.31700218e+00 -9.56333935e-01 1.55075169e+00 -8.37647855e-01
1.26510605e-01 4.79039922e-02 -3.11322331e-01 -8.96532595e-01
-4.99458104e-01 -6.93335235e-01 -6.71312094e-01 9.80580568e-01
-3.28714639e-01 -4.79949623e-01 1.38717580e+00 1.02382064e+00
5.64469218e-01 -6.75264180e-01 -1.29109728e+00 -7.32587099e-01
-9.37772170e-02 -6.69739246e-01 6.17690027e-01 1.02944970e+00
2.30701610e-01 -7.10518286e-02 -6.82009518e-01 4.22899351e-02
8.43974590e-01 -3.83958727e-01 8.75763595e-01 -1.02846920e+00
1.48315594e-01 -9.04291928e-01 -1.13725269e+00 -8.73853445e-01
7.24855304e-01 -5.51356554e-01 -6.14392042e-01 -8.44827294e-01
1.91132620e-01 1.88457161e-01 -4.13331479e-01 4.31175977e-01
-7.57637769e-02 8.72401893e-01 -4.96966240e-04 -5.60055748e-02
-4.44223225e-01 5.96170425e-01 8.20713162e-01 -3.60138029e-01
-8.36657882e-02 -1.11808881e-01 -7.13902295e-01 5.73259771e-01
3.04542631e-01 6.38112724e-01 -2.33801305e-02 -1.54301733e-01
-3.16453546e-01 -2.14362875e-01 7.59644210e-01 -1.06651390e+00
2.43112311e-01 -1.94366574e-01 9.90167201e-01 -2.56834980e-02
7.78848886e-01 -4.36242461e-01 1.95800159e-02 1.91616490e-01
-2.86668062e-01 6.06333837e-02 5.54678321e-01 3.98034990e-01
-3.39544922e-01 5.99250317e-01 8.19213510e-01 2.56682307e-01
-1.42236531e+00 7.60832429e-01 -2.40761995e-01 -4.52446491e-01
1.10104489e+00 -8.45229685e-01 -1.93309516e-01 -6.33061469e-01
-8.37738216e-01 -3.60976160e-01 2.69819170e-01 7.01613724e-01
8.79093766e-01 -1.69235003e+00 -6.31680667e-01 6.19307399e-01
2.35897109e-01 -1.11835468e+00 4.04117376e-01 7.81397641e-01
-1.99442461e-01 6.00031555e-01 -7.35078156e-01 -7.11413860e-01
-1.75240350e+00 2.94605680e-02 9.70588446e-01 1.16156884e-01
-5.90537429e-01 1.09242761e+00 -8.96053538e-02 -1.82632774e-01
5.24849534e-01 1.84494525e-01 -3.32740217e-01 -1.53811648e-01
1.16209579e+00 7.13516414e-01 -1.50164083e-01 -1.43718481e+00
-7.93830812e-01 8.48640382e-01 3.28637064e-01 -1.42738953e-01
1.17211354e+00 -2.10597292e-01 -2.31462643e-01 1.05999999e-01
1.28713918e+00 -2.09351569e-01 -1.21337545e+00 -3.54494333e-01
2.07389563e-01 -6.80001795e-01 -1.70236215e-01 -8.18948805e-01
-8.19279492e-01 9.74193454e-01 1.16473961e+00 -3.66778880e-01
1.28542423e+00 -1.59997061e-01 7.73774445e-01 4.98358846e-01
4.65492904e-01 -1.03971171e+00 2.18203023e-01 4.89824682e-01
8.66999149e-01 -1.20430732e+00 -3.19725782e-01 -2.08511844e-01
-7.26009548e-01 1.32373583e+00 6.94371223e-01 1.00816108e-01
7.66166270e-01 4.82170016e-01 2.23517299e-01 -2.65462935e-01
-9.95423943e-02 -1.58767328e-01 6.69217229e-01 1.00979972e+00
5.61001122e-01 2.47834235e-01 3.42660397e-01 7.93180227e-01
-8.93150270e-02 5.72725952e-01 -1.94266111e-01 5.04997432e-01
-7.72826141e-03 -9.04203951e-01 -3.58984470e-01 1.85543969e-01
-8.36693883e-01 4.25283819e-01 -7.61788487e-01 3.52320939e-01
9.48113482e-03 5.47781587e-01 -9.96411666e-02 -7.83964932e-01
3.13571215e-01 3.86340022e-01 7.84469962e-01 -1.66604982e-03
-4.40995127e-01 -4.19203639e-01 7.35236034e-02 -1.00883842e+00
-8.24269950e-01 -6.97128654e-01 -8.07894170e-01 -8.45306039e-01
4.73727494e-01 -4.60855454e-01 3.58893037e-01 1.08939314e+00
6.91783547e-01 1.19643733e-01 2.41248444e-01 -9.42440867e-01
-5.59750795e-01 -8.58887076e-01 -5.71990252e-01 9.64714706e-01
4.64222521e-01 -8.28066289e-01 2.80123711e-01 1.88135087e-01] | [13.622872352600098, 1.812815546989441] |
c84d01ce-0bf1-4b7a-9111-69782f666875 | a-side-by-side-comparison-of-transformers-for | 2307.03378 | null | https://arxiv.org/abs/2307.03378v1 | https://arxiv.org/pdf/2307.03378v1.pdf | A Side-by-side Comparison of Transformers for English Implicit Discourse Relation Classification | Though discourse parsing can help multiple NLP fields, there has been no wide language model search done on implicit discourse relation classification. This hinders researchers from fully utilizing public-available models in discourse analysis. This work is a straightforward, fine-tuned discourse performance comparison of seven pre-trained language models. We use PDTB-3, a popular discourse relation annotated dataset. Through our model search, we raise SOTA to 0.671 ACC and obtain novel observations. Some are contrary to what has been reported before (Shi and Demberg, 2019b), that sentence-level pre-training objectives (NSP, SBO, SOP) generally fail to produce the best performing model for implicit discourse relation classification. Counterintuitively, similar-sized PLMs with MLM and full attention led to better performance. | ['Jason Hyung-Jong Lee', 'BongSeok Yang', 'Bruce W. Lee'] | 2023-07-07 | null | null | null | null | ['discourse-parsing', 'relation-classification', 'implicit-discourse-relation-classification'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 5.05414963e-01 1.17672431e+00 -8.43960345e-01 -1.70122892e-01
-9.91407216e-01 -6.21166408e-01 1.25925016e+00 4.86421317e-01
-3.69730413e-01 1.14594364e+00 9.69357729e-01 -1.00019991e+00
-8.13732147e-02 -4.57906127e-01 -4.31017816e-01 -3.46811116e-01
1.14655532e-01 4.53432411e-01 2.05747411e-01 -3.36644620e-01
6.40610397e-01 -1.99652568e-01 -1.09184015e+00 9.26780820e-01
8.27057064e-01 4.70957398e-01 1.08343355e-01 9.57273841e-01
-1.20339580e-01 1.97745609e+00 -9.59988356e-01 -4.87928748e-01
-5.07845342e-01 -5.44964790e-01 -1.68689513e+00 -1.80204719e-01
2.41158307e-01 -2.15558723e-01 -2.17042997e-01 5.73974907e-01
3.03594559e-01 1.63099870e-01 7.87088752e-01 -7.53448129e-01
-9.43140626e-01 1.17968810e+00 -3.41526777e-01 6.57437742e-01
7.22060442e-01 9.55672469e-03 1.33264768e+00 -4.09203887e-01
1.16616702e+00 1.67038167e+00 7.17695475e-01 5.76349795e-01
-1.21234930e+00 -2.23101266e-02 4.87922400e-01 4.23138738e-01
-5.86565852e-01 -6.40632153e-01 7.09594429e-01 -5.40709078e-01
1.66090298e+00 7.66756773e-01 4.14583266e-01 1.55113041e+00
-7.46491104e-02 8.44926417e-01 1.29358149e+00 -7.20417440e-01
-2.02024400e-01 9.62634310e-02 6.40183389e-01 1.67714283e-01
1.15684010e-01 -3.56548727e-01 -7.06697226e-01 -8.08596015e-02
1.70647055e-01 -1.01283562e+00 -2.99133241e-01 5.06654859e-01
-1.13753796e+00 8.71701658e-01 1.37675852e-01 6.60400689e-01
-2.02018648e-01 4.01002914e-03 7.58180082e-01 4.63384032e-01
7.71870136e-01 8.31224978e-01 -4.84955788e-01 -6.76421821e-01
-4.71813440e-01 5.03496826e-01 1.11386716e+00 7.66690552e-01
1.03973664e-01 -1.30679563e-01 -4.18671668e-01 7.20204532e-01
4.20708567e-01 -7.35937804e-02 3.53424907e-01 -1.40080571e+00
9.83888984e-01 7.42640495e-01 -6.92849085e-02 -1.19808710e+00
-4.39849108e-01 -2.58472804e-02 -3.75780791e-01 -2.74688780e-01
6.39998138e-01 -1.90382257e-01 -2.67611027e-01 1.57425964e+00
-4.91183437e-02 -2.34356150e-01 4.02067900e-01 6.60284281e-01
1.32515240e+00 8.20362568e-01 5.69041848e-01 -5.63399196e-01
1.39171040e+00 -1.13695991e+00 -1.04992092e+00 -4.97392863e-01
1.31423891e+00 -9.00585771e-01 1.15086186e+00 1.56203806e-01
-1.33700800e+00 -2.71161467e-01 -1.05634236e+00 -6.59431398e-01
-1.25261039e-01 -1.71515614e-01 7.87209511e-01 5.46564996e-01
-8.59504223e-01 3.12474400e-01 -6.02239966e-01 -1.93758711e-01
5.31518698e-01 2.62320936e-02 -4.03194949e-02 3.91295373e-01
-1.31720686e+00 1.72996545e+00 7.40420878e-01 8.08598325e-02
-4.44306254e-01 -5.78370333e-01 -8.76029074e-01 -1.86576024e-01
6.56066060e-01 -5.76044202e-01 1.50636387e+00 -9.69571769e-01
-1.63977969e+00 1.28854740e+00 -5.38196385e-01 -9.77187216e-01
2.24357337e-01 -4.91347820e-01 -9.35083032e-02 -9.57387015e-02
2.64886525e-02 3.36741775e-01 2.65948653e-01 -1.19530773e+00
-4.23682243e-01 1.04899310e-01 5.05824745e-01 4.91591543e-01
-1.16115205e-01 5.58297873e-01 4.26472455e-01 -4.78217483e-01
5.63409599e-03 -4.22956765e-01 -2.41889134e-02 -7.75792301e-01
-5.83334208e-01 -8.04921985e-01 6.95005715e-01 -1.09890258e+00
1.70081151e+00 -1.57947218e+00 2.43606254e-01 -4.41720366e-01
3.52489084e-01 4.67720658e-01 1.98093444e-01 4.60363775e-01
-8.98807719e-02 6.59068227e-01 -2.33819008e-01 -3.70580941e-01
1.10797003e-01 3.00700873e-01 -3.87702584e-01 3.49833965e-01
5.33641040e-01 1.20212305e+00 -9.87932384e-01 -7.84162164e-01
9.01871026e-02 1.98483542e-01 -4.37006414e-01 3.67580168e-02
-6.57224000e-01 4.73139375e-01 -2.28849947e-01 4.23212558e-01
6.24262169e-02 -4.88033354e-01 6.46965444e-01 2.49694511e-01
-2.91764796e-01 1.58517599e+00 -4.61982220e-01 1.20680141e+00
-1.96379200e-01 1.48638141e+00 -2.83971839e-02 -1.28349876e+00
6.85822368e-01 7.28500247e-01 1.01039648e-01 -7.34012008e-01
3.54833901e-01 1.97527006e-01 7.65386045e-01 -9.23563600e-01
7.81174064e-01 -1.06845856e-01 -5.71098775e-02 3.58169287e-01
-4.98573333e-02 8.85483027e-02 3.86525422e-01 7.70904869e-02
9.96612191e-01 1.34074673e-01 5.38946867e-01 -5.47506571e-01
7.61737585e-01 3.84760588e-01 4.07024831e-01 6.31763279e-01
1.53283672e-02 4.50233698e-01 1.09395957e+00 -4.90674525e-02
-9.86258030e-01 -5.59624612e-01 -3.24464023e-01 1.32318604e+00
-1.93539917e-01 -6.24227583e-01 -7.39417851e-01 -4.48125422e-01
-3.59139621e-01 1.17437875e+00 -6.57718241e-01 3.53614956e-01
-1.36285806e+00 -9.00154889e-01 7.39105225e-01 3.88154805e-01
3.08889955e-01 -1.37230957e+00 -4.56833154e-01 3.09560955e-01
-4.47520316e-01 -9.95608211e-01 2.00132102e-01 5.54772168e-02
-8.18427742e-01 -1.25049150e+00 -3.45244080e-01 -9.59718823e-01
7.10084662e-02 -2.94927150e-01 1.41974497e+00 2.64932364e-01
3.92930895e-01 6.58632889e-02 -4.40917611e-01 -4.50024545e-01
-1.06180775e+00 4.94080424e-01 -5.57478130e-01 -7.67155290e-01
4.63022113e-01 -1.96101934e-01 5.56648150e-02 -3.72595191e-01
-2.33539060e-01 2.06709579e-01 7.81890005e-02 9.29384410e-01
1.34976894e-01 -5.75736701e-01 1.01155376e+00 -1.25984156e+00
1.11343205e+00 -6.53447807e-01 -6.16986118e-02 1.99192002e-01
-6.39101386e-01 -3.02369267e-01 4.62031290e-02 -6.49359584e-01
-1.41019011e+00 -8.92217577e-01 -3.23458046e-01 5.87704897e-01
-9.90414992e-02 8.33953202e-01 -1.72741693e-02 5.66596627e-01
8.92604530e-01 -3.85586351e-01 -2.89325938e-02 -3.18303645e-01
3.73171002e-01 5.08303404e-01 2.10989028e-01 -6.62910044e-01
2.24037632e-01 -7.37467185e-02 -3.00836414e-01 -1.13135099e+00
-1.30680084e+00 -1.27537042e-01 -5.74338913e-01 -7.62065053e-02
1.26665807e+00 -7.12831616e-01 -9.41596568e-01 1.02064118e-01
-1.59077084e+00 -8.96880031e-01 -2.96919763e-01 3.10447723e-01
-6.53537929e-01 4.09484684e-01 -1.01110232e+00 -1.22409701e+00
-2.83312231e-01 -9.01180267e-01 4.49943602e-01 8.20370913e-02
-1.11516488e+00 -1.48706734e+00 9.11868960e-02 7.24652648e-01
1.51599333e-01 5.31356633e-01 1.05728078e+00 -8.27537954e-01
-3.59417439e-01 5.66735804e-01 -1.55051321e-01 1.41340241e-01
3.63913514e-02 6.80096447e-02 -1.03200376e+00 1.66862413e-01
1.41944438e-01 -6.49779797e-01 7.52553046e-01 4.12087262e-01
6.61867857e-01 -8.19220841e-01 -3.01964521e-01 2.11646091e-02
6.86409652e-01 2.18485624e-01 6.93898737e-01 9.54844415e-01
5.15625656e-01 8.30988348e-01 7.00096011e-01 -1.47972763e-01
7.99824953e-01 4.09803301e-01 8.05211067e-02 1.32999822e-01
-3.64617646e-01 -2.01178398e-02 5.27125299e-01 9.67968762e-01
-4.37785387e-01 -5.47602057e-01 -1.20323181e+00 7.43510544e-01
-1.93612218e+00 -1.19767821e+00 -6.48680389e-01 1.37928760e+00
1.26849508e+00 5.35330772e-01 -2.90932618e-02 2.49827862e-01
3.89507532e-01 7.64258087e-01 -3.99736455e-03 -7.94587255e-01
-6.07160211e-01 8.60661864e-02 -3.19698192e-02 1.04107308e+00
-1.21556985e+00 1.04261279e+00 6.64999294e+00 6.86740339e-01
-7.01250434e-01 3.54129821e-01 8.26824069e-01 -2.27313936e-02
-4.54644799e-01 2.06536636e-01 -6.70105040e-01 2.77514815e-01
1.28101647e+00 -3.49393994e-01 -5.22423945e-02 5.32254934e-01
2.05081161e-02 -3.02957952e-01 -1.33826566e+00 4.76301998e-01
4.99581993e-02 -1.79783809e+00 -2.75438398e-01 -6.75880611e-02
7.76162446e-01 -2.13279560e-01 -1.38318554e-01 7.16087461e-01
3.52330267e-01 -1.42879915e+00 7.63924420e-01 2.54324913e-01
2.23500744e-01 -1.78854331e-01 7.97957182e-01 6.41296446e-01
-5.87558091e-01 -7.69371390e-02 -7.85216317e-02 -8.51235926e-01
6.07093692e-01 -5.11563122e-02 -1.09241951e+00 3.45551074e-01
3.45430970e-01 5.95960081e-01 -4.26763922e-01 1.76302731e-01
-5.12202501e-01 1.23362648e+00 -1.64287940e-01 -3.69680613e-01
4.07847524e-01 1.94928199e-01 9.09744620e-01 1.48204768e+00
-4.13581610e-01 4.37963516e-01 -1.13125689e-01 6.80532336e-01
-3.52356918e-02 -7.61501025e-03 -4.89039123e-01 -2.05663219e-01
4.56586748e-01 7.36121595e-01 -4.63403314e-01 -4.64069247e-01
-3.58257920e-01 1.11956634e-01 5.34505486e-01 3.02391291e-01
-6.82984591e-01 4.25323427e-01 3.83108079e-01 2.02073127e-01
-1.69708624e-01 -2.40156278e-01 -8.32267702e-01 -7.74998844e-01
-5.90447076e-02 -9.97608125e-01 4.23600972e-01 -4.77219641e-01
-1.30881953e+00 4.79162425e-01 2.38742441e-01 -4.68263239e-01
-6.48709774e-01 -6.66852653e-01 -6.40785396e-01 1.04944623e+00
-1.28813338e+00 -1.19157910e+00 1.42324954e-01 -2.18397364e-01
8.75091255e-01 -1.92476641e-02 9.90516543e-01 1.55460075e-01
-3.74437630e-01 3.88387382e-01 -4.39742118e-01 3.47455233e-01
5.47014952e-01 -1.23622966e+00 4.42782909e-01 5.25836766e-01
-1.45899162e-01 8.16137433e-01 9.11276460e-01 -7.51083851e-01
-7.78674483e-01 -5.26227832e-01 1.41384363e+00 -1.05086386e+00
1.01812088e+00 -1.21488959e-01 -1.26993752e+00 1.18898833e+00
1.07565129e+00 -8.79404664e-01 6.87964320e-01 6.36077523e-01
6.03435412e-02 7.78312624e-01 -6.99841976e-01 7.15348184e-01
1.06859279e+00 -5.10801196e-01 -1.44522429e+00 3.79785657e-01
1.05301201e+00 -8.16695631e-01 -1.23448205e+00 5.24275482e-01
2.21660450e-01 -6.64443195e-01 1.00369394e+00 -1.15368629e+00
1.07660711e+00 2.64893860e-01 -1.38904974e-01 -7.15311825e-01
-1.34773046e-01 -4.78935659e-01 -6.54042780e-01 1.61538279e+00
1.00730145e+00 -6.96394205e-01 6.35576069e-01 6.76906765e-01
-4.80542928e-01 -1.09550011e+00 -8.68778408e-01 -5.28558075e-01
7.27696538e-01 -4.23319638e-01 1.05335154e-01 1.46735227e+00
6.03380263e-01 9.92900014e-01 4.23776619e-02 -3.29061598e-01
7.51265138e-02 4.77644335e-03 5.14515460e-01 -1.12828696e+00
-2.21346125e-01 -9.57647622e-01 2.68410802e-01 -1.22364080e+00
6.45874023e-01 -9.46389198e-01 -2.55943984e-01 -1.97606778e+00
8.43837559e-02 -5.88877141e-01 3.52776945e-01 3.76132578e-01
-4.11333561e-01 -1.66793182e-01 2.51649767e-01 8.37248117e-02
-6.27050817e-01 4.51142639e-01 1.22648084e+00 -1.59012720e-01
-4.24394399e-01 -1.18303187e-01 -8.39686394e-01 1.11598969e+00
1.10668397e+00 -3.40672046e-01 -4.42013115e-01 -5.83630741e-01
2.71308899e-01 7.99960494e-02 1.44565850e-01 -2.78216869e-01
-6.85546920e-03 -4.68975127e-01 1.60246361e-02 -7.20278919e-01
5.44613063e-01 3.31347715e-03 -2.29894966e-01 3.61087173e-01
-8.70346725e-01 -1.01120956e-01 3.59962136e-01 -1.15091717e-02
-3.20440233e-01 -6.95694983e-01 1.61366656e-01 -2.30031565e-01
-5.28995156e-01 -5.74930310e-01 -8.28239858e-01 6.59122586e-01
6.56704724e-01 -2.12807819e-01 -1.17445481e+00 -3.50416601e-01
-7.25200295e-01 3.42199773e-01 1.47293046e-01 5.18130898e-01
9.48687717e-02 -9.81636345e-01 -1.01800489e+00 -6.76680267e-01
-3.61723721e-01 4.54345435e-01 -3.11672199e-03 1.12751055e+00
-5.04775047e-01 7.75983691e-01 3.08335632e-01 -5.45408964e-01
-1.50520205e+00 4.10665452e-01 8.31727833e-02 -7.66668618e-01
-6.24988675e-01 1.03569901e+00 -7.17828050e-02 -3.83933663e-01
3.43416631e-02 -4.14823979e-01 -6.21974766e-01 3.85755152e-01
3.75167519e-01 5.67523777e-01 -2.57079661e-01 -7.86213934e-01
-3.07365865e-01 -1.59760326e-01 7.61362985e-02 -2.84349382e-01
1.31599724e+00 -1.96477830e-01 -3.16594839e-01 9.75928605e-01
8.51377010e-01 1.11139596e-01 -8.78250301e-01 -8.43315646e-02
7.68638492e-01 7.04288557e-02 -2.32864797e-01 -7.20750093e-01
-3.07504032e-02 7.70301878e-01 -3.63139719e-01 8.29824388e-01
5.55845678e-01 2.15080991e-01 3.27605933e-01 5.21242082e-01
-2.30654225e-01 -1.25336480e+00 -1.00935791e-02 1.15512288e+00
1.20294976e+00 -1.24183464e+00 4.04751390e-01 -5.90663493e-01
-7.45223701e-01 8.07939887e-01 9.41329837e-01 1.82127561e-02
2.48411402e-01 3.05628031e-01 8.23238268e-02 -3.52987111e-01
-1.19257939e+00 7.55038112e-02 2.49212727e-01 5.79911709e-01
1.17859411e+00 1.91593543e-02 -8.08709502e-01 7.10380137e-01
-6.67695165e-01 -5.13152003e-01 5.85476279e-01 7.73684084e-01
-4.11840826e-01 -1.05696893e+00 -3.72396857e-01 4.15319979e-01
-7.31636167e-01 -2.47515604e-01 -7.24714935e-01 1.29814792e+00
-2.70412117e-01 1.10628200e+00 3.12884390e-01 6.04314245e-02
8.46376792e-02 2.77252436e-01 5.61813712e-01 -8.08672309e-01
-1.00018454e+00 -2.57577002e-01 1.38918388e+00 -1.36395857e-01
-1.00468171e+00 -8.55854869e-01 -1.26443148e+00 -5.42234659e-01
-3.90477061e-01 -4.95928340e-02 3.95766273e-03 1.17939925e+00
-1.26741514e-01 9.38825130e-01 -2.55029470e-01 -5.74386895e-01
-2.34962970e-01 -1.69315290e+00 2.61520475e-01 2.85025597e-01
3.99701178e-01 -6.67634368e-01 -2.51047462e-01 2.98298687e-01] | [10.864924430847168, 9.336297988891602] |
dd8667ef-9c10-4df0-9e8e-c99164893d72 | bio-inspired-gait-imitation-of-hexapod-robot | 2004.05450 | null | https://arxiv.org/abs/2004.05450v1 | https://arxiv.org/pdf/2004.05450v1.pdf | Bio-inspired Gait Imitation of Hexapod Robot Using Event-Based Vision Sensor and Spiking Neural Network | Learning how to walk is a sophisticated neurological task for most animals. In order to walk, the brain must synthesize multiple cortices, neural circuits, and diverse sensory inputs. Some animals, like humans, imitate surrounding individuals to speed up their learning. When humans watch their peers, visual data is processed through a visual cortex in the brain. This complex problem of imitation-based learning forms associations between visual data and muscle actuation through Central Pattern Generation (CPG). Reproducing this imitation phenomenon on low power, energy-constrained robots that are learning to walk remains challenging and unexplored. We propose a bio-inspired feed-forward approach based on neuromorphic computing and event-based vision to address the gait imitation problem. The proposed method trains a "student" hexapod to walk by watching an "expert" hexapod moving its legs. The student processes the flow of Dynamic Vision Sensor (DVS) data with a one-layer Spiking Neural Network (SNN). The SNN of the student successfully imitates the expert within a small convergence time of ten iterations and exhibits energy efficiency at the sub-microjoule level. | ['Yan Fang', 'Arijit Raychowdhury', 'Justin Ting', 'Ashwin Sanjay Lele'] | 2020-04-11 | null | null | null | null | ['event-based-vision'] | ['computer-vision'] | [ 2.11831465e-01 -5.96670583e-02 2.25020602e-01 1.77844867e-01
5.54582059e-01 -3.17909598e-01 3.68396580e-01 -4.91616547e-01
-7.93159306e-01 9.63233829e-01 -6.15558863e-01 4.57492024e-01
1.01651631e-01 -8.18680167e-01 -1.17525232e+00 -1.02657044e+00
1.20429546e-01 6.21945821e-02 6.79476321e-01 -6.30550608e-02
5.06865442e-01 2.26066574e-01 -1.89351892e+00 -1.41101137e-01
8.54662955e-01 6.46824181e-01 8.90638411e-01 7.90410221e-01
3.87432694e-01 7.60152221e-01 -1.66204080e-01 2.49058634e-01
2.03293920e-01 -1.00624776e+00 -3.93021286e-01 -2.13351533e-01
-4.10746783e-01 1.68007597e-01 -4.14212048e-01 9.88180757e-01
6.07894182e-01 1.19302347e-01 8.04597259e-01 -1.45841622e+00
-6.55529797e-01 4.41976577e-01 -1.75721869e-01 9.91998389e-02
2.38612816e-02 8.03531885e-01 3.34918380e-01 -5.29556036e-01
1.01301825e+00 8.01060200e-01 5.84976912e-01 1.11980832e+00
-1.15626788e+00 -5.47750592e-01 -5.01038671e-01 5.68856657e-01
-1.11467826e+00 -3.10680538e-01 5.46919823e-01 -2.76234925e-01
1.29369187e+00 -5.79992175e-01 1.62592244e+00 1.36966598e+00
1.17156124e+00 4.15137053e-01 1.09846950e+00 -2.44342715e-01
9.87976313e-01 -5.43888152e-01 -5.11680961e-01 1.01271284e+00
5.97133875e-01 1.33883581e-01 -1.13941741e+00 3.62835824e-01
1.00770187e+00 -1.08935677e-01 -1.38332620e-01 -3.88434917e-01
-1.34860826e+00 4.20476377e-01 8.06227565e-01 2.22283736e-01
-7.38663971e-01 1.08281338e+00 -9.09859240e-02 3.22981536e-01
-8.23676348e-01 4.66175139e-01 -1.37948647e-01 -3.80845129e-01
-6.06362760e-01 4.13559936e-02 1.01660120e+00 4.28136081e-01
5.06122828e-01 3.76141191e-01 2.82647043e-01 4.93758053e-01
6.48067534e-01 8.85201514e-01 8.22965324e-01 -1.47888160e+00
-2.03411728e-01 8.09501290e-01 -2.03832045e-01 -7.11531043e-01
-2.84187675e-01 -1.87685732e-02 -9.12293375e-01 9.39593196e-01
3.50413561e-01 -2.55868256e-01 -7.47377753e-01 1.80040014e+00
2.62642801e-01 1.48776591e-01 2.45102331e-01 1.10515809e+00
3.23127568e-01 7.39065111e-01 -1.11209594e-01 -1.33502841e-01
1.09305453e+00 -9.21513915e-01 -2.05575079e-01 -5.76000750e-01
1.03518754e-01 7.58501440e-02 7.00269878e-01 4.62436765e-01
-1.12523425e+00 -3.15726638e-01 -1.25269413e+00 2.39882290e-01
-2.01048732e-01 -4.16935347e-02 2.83833388e-02 -5.73609360e-02
-1.08378506e+00 1.11335170e+00 -1.12485373e+00 -1.11123633e+00
3.36280912e-01 5.35560429e-01 -2.35005677e-01 2.21522257e-01
-7.45271444e-01 1.01148605e+00 2.56328881e-01 -2.31074303e-01
-1.28787327e+00 -3.53699893e-01 -3.74782056e-01 -6.71438053e-02
-1.92634597e-01 -1.39972854e+00 1.08062184e+00 -8.03418636e-01
-2.09033871e+00 8.27696025e-01 1.61412537e-01 -4.88703370e-01
8.18026438e-02 3.89780641e-01 2.83168077e-01 4.02555972e-01
-5.73783405e-02 1.12125361e+00 1.01487505e+00 -9.08105195e-01
-2.57886171e-01 -4.79177892e-01 -7.11860240e-01 2.79045284e-01
-2.02839583e-01 -5.39298534e-01 7.52764046e-02 -3.04568350e-01
2.01397553e-01 -1.15223181e+00 -1.10879332e-01 6.17226422e-01
1.24974169e-01 -2.15273246e-01 7.60954320e-01 -1.10627003e-01
3.84186000e-01 -1.96853983e+00 7.01612830e-01 -1.03349417e-01
-1.73640456e-02 -2.46750209e-02 -1.97775200e-01 6.18493617e-01
5.98618209e-01 -4.73482966e-01 -4.45920378e-01 3.07538301e-01
-2.15029359e-01 8.18467975e-01 2.38247924e-02 4.53090191e-01
1.25558421e-01 1.11052132e+00 -9.26733553e-01 -5.64588785e-01
-1.74836218e-01 4.50618833e-01 -5.77978671e-01 3.32994193e-01
-2.08623737e-01 7.89339244e-01 -4.54668462e-01 6.12029076e-01
4.45723794e-02 -4.44231182e-01 2.95831412e-01 2.25766048e-01
-3.94276679e-01 -3.20505291e-01 -7.69010246e-01 1.98010468e+00
-3.26926410e-01 6.10937119e-01 3.79602104e-01 -1.30921328e+00
1.01374495e+00 1.83772221e-02 3.23705643e-01 -1.12578917e+00
4.07647431e-01 5.97909987e-01 4.06891763e-01 -8.61180782e-01
-4.92993265e-01 1.25261080e-02 4.19977270e-02 4.51817453e-01
4.35215712e-01 -6.08822107e-01 1.95663556e-01 -2.00713068e-01
1.85476243e+00 5.92228889e-01 3.78337681e-01 -1.25755683e-01
8.89462605e-02 2.30737463e-01 6.45183444e-01 4.38065022e-01
-2.96338409e-01 9.54234526e-02 -3.88179757e-02 -1.38298750e-01
-1.26935863e+00 -1.49396229e+00 3.46972346e-01 9.15361226e-01
2.99742252e-01 3.99867892e-01 -9.73598540e-01 3.61700207e-01
-2.92570703e-02 2.15582103e-01 -4.15567189e-01 -4.93190110e-01
-6.38854563e-01 -3.81718278e-01 6.67356670e-01 4.94994283e-01
9.80272174e-01 -2.02298999e+00 -2.01863742e+00 8.04820836e-01
8.54724795e-02 -8.13907802e-01 2.82570273e-02 6.91677153e-01
-9.08548832e-01 -9.85254586e-01 -7.51401365e-01 -1.16406143e+00
7.47283399e-01 -1.71335161e-01 5.20719767e-01 -2.75340118e-02
-7.42418587e-01 5.37550628e-01 -2.22412661e-01 -3.64164829e-01
-1.55117691e-01 -2.28231475e-01 3.00495565e-01 -4.85124677e-01
6.27662763e-02 -1.29696786e+00 -7.87352383e-01 -6.98792981e-03
-5.32199264e-01 1.83991671e-01 9.78103817e-01 8.96529078e-01
7.14226246e-01 -3.08022171e-01 4.69695926e-01 3.09407532e-01
4.66866702e-01 -5.36152005e-01 -4.24823642e-01 1.71007559e-01
-4.39316392e-01 4.91150290e-01 1.03549623e+00 -7.32228398e-01
-5.19934237e-01 6.46730006e-01 3.16950589e-01 -2.68634915e-01
-1.92912649e-02 1.05386218e-02 2.93133259e-01 -4.94261086e-01
7.49499381e-01 9.71902966e-01 3.32892001e-01 1.50571793e-01
1.74252538e-03 3.61689806e-01 8.86406839e-01 -4.11135077e-01
5.59104562e-01 4.88657922e-01 3.98111671e-01 -9.04304028e-01
3.14500421e-01 1.41550392e-01 -5.12589276e-01 -8.37805390e-01
1.11756694e+00 -4.74622428e-01 -1.37397790e+00 9.11760867e-01
-1.15442133e+00 -8.65890324e-01 -5.05109668e-01 5.09865582e-01
-1.09036291e+00 1.02476422e-02 -7.23337829e-01 -5.48672199e-01
-4.29316014e-01 -5.93306601e-01 5.67588031e-01 9.25570011e-01
-2.31310159e-01 -4.30922210e-01 7.04098046e-01 -5.01584560e-02
4.94663626e-01 3.59177768e-01 7.73357689e-01 1.83259562e-01
-6.16980791e-01 4.48435783e-01 2.68590450e-01 -5.40182739e-02
-2.71454304e-01 1.61912546e-01 -5.26275396e-01 -2.71309286e-01
-1.66321844e-01 -7.40425289e-01 8.01424801e-01 4.52819020e-01
7.97574222e-01 -3.30381006e-01 -5.05124331e-01 3.34887773e-01
1.68058765e+00 5.03546059e-01 3.48409861e-01 1.28014579e-01
1.25182599e-01 6.67926252e-01 -1.44052044e-01 4.61647481e-01
3.60489368e-01 1.60705671e-02 6.99260354e-01 6.26713932e-01
-2.79785186e-01 -3.32862079e-01 8.15637410e-01 1.36413383e+00
-4.13384646e-01 -9.38252881e-02 -9.07607019e-01 6.31716788e-01
-1.90442848e+00 -1.03059733e+00 4.78797257e-02 1.78753161e+00
7.76541471e-01 -1.65122256e-01 -9.68006551e-02 2.84369856e-01
5.00166774e-01 -6.67633832e-01 -1.30884886e+00 -5.44337571e-01
-1.04708523e-02 5.24260402e-01 3.42503846e-01 -1.46477684e-01
-1.74713597e-01 7.13627696e-01 5.53484011e+00 4.61300835e-02
-1.33033597e+00 -1.49857759e-01 -2.28707597e-01 -8.66794661e-02
2.31319964e-01 -1.22825839e-01 -2.57475257e-01 9.33377981e-01
1.08368957e+00 -3.47616464e-01 1.17285967e+00 8.16139519e-01
3.39480370e-01 -7.22051442e-01 -1.11887515e+00 1.16442752e+00
-9.89024788e-02 -1.24873388e+00 -2.88855493e-01 1.41087055e-01
6.34900212e-01 2.82833189e-01 -1.96406022e-01 2.22811997e-02
3.73019964e-01 -1.05514121e+00 5.97583890e-01 1.00078940e+00
3.93984497e-01 -3.39197338e-01 8.08309689e-02 8.38256419e-01
-1.38531017e+00 -6.78754270e-01 -4.40881401e-01 -4.13194537e-01
5.69190877e-03 2.63876826e-01 -3.69340181e-01 -2.23791599e-01
1.05716097e+00 6.92830980e-01 -2.83727288e-01 1.20794296e+00
-3.71035874e-01 4.25636560e-01 -5.20052195e-01 -8.12371552e-01
1.23342872e-01 -3.53163898e-01 6.64313436e-01 5.74094296e-01
8.41437221e-01 2.81164855e-01 -3.96210790e-01 1.16828442e+00
-1.07564814e-01 -2.64493942e-01 -1.03091979e+00 -2.26183712e-01
5.63842654e-01 1.17540646e+00 -1.01145625e+00 -1.65317059e-01
1.18226305e-01 1.18525076e+00 2.81899273e-01 1.99405566e-01
-6.33368731e-01 -6.56408608e-01 3.82189900e-01 -8.39035437e-02
3.63365620e-01 -5.12618303e-01 -2.56919682e-01 -7.42177367e-01
-6.47850856e-02 -2.52598971e-01 -3.03448349e-01 -1.10008645e+00
-9.99881804e-01 -6.14202507e-02 -6.70094967e-01 -1.11694682e+00
-3.57673377e-01 -6.23672366e-01 -9.99053478e-01 5.85994497e-02
-8.24047506e-01 -5.36042273e-01 -6.51846826e-01 7.57025421e-01
4.47488606e-01 -1.07318431e-01 6.59493506e-01 -1.53868169e-01
-4.13121939e-01 -7.20378608e-02 8.85524005e-02 -1.27183627e-02
3.56776893e-01 -9.24601972e-01 -1.28011793e-01 7.66432703e-01
-1.18254550e-01 2.43880451e-02 6.77040517e-01 -5.97679198e-01
-1.96395731e+00 -9.50187385e-01 6.52444303e-01 3.29607248e-01
4.59147453e-01 -5.24057597e-02 -7.46784449e-01 3.30118239e-02
5.24163008e-01 6.91811889e-02 2.15488911e-01 -1.16665602e+00
2.09449425e-01 -2.39153549e-01 -1.46831262e+00 7.27587640e-01
1.41826916e+00 -4.32960801e-02 -9.39972460e-01 -2.20376849e-01
1.66926727e-01 2.83027709e-01 -4.81552899e-01 9.41568837e-02
8.83194745e-01 -6.80308402e-01 6.00047946e-01 -2.10598901e-01
5.30811429e-01 -5.86773217e-01 7.29039684e-02 -1.60568595e+00
-1.21347308e-01 -4.22441155e-01 -3.19304824e-01 3.67885560e-01
-1.74084641e-02 -6.71761811e-01 7.46297121e-01 -7.88527504e-02
-3.08858901e-02 -5.83774745e-01 -1.37498987e+00 -1.04867578e+00
1.16537169e-01 3.83796990e-01 -3.18736106e-01 4.93240237e-01
5.33815861e-01 2.85087824e-01 -1.32596828e-02 -6.13139980e-02
1.10641885e+00 1.41903386e-02 3.39886189e-01 -1.19800174e+00
-2.78154761e-01 -2.97383428e-01 -4.93262917e-01 -6.82541251e-01
2.15913281e-02 -1.11395383e+00 7.01376915e-01 -1.86851776e+00
1.01759478e-01 2.92520434e-01 9.07152370e-02 7.34557867e-01
6.92551792e-01 2.39040196e-01 1.36189908e-01 1.99810907e-01
-5.39228559e-01 6.27362370e-01 1.39508152e+00 1.22076191e-01
-8.87045637e-02 -5.15689254e-01 1.17698699e-01 6.86816633e-01
1.03390241e+00 -8.22957158e-01 -3.54821831e-01 -4.51817036e-01
4.08464134e-01 2.99458057e-01 8.23395967e-01 -1.94788396e+00
1.21574807e+00 -2.15952113e-01 6.74942255e-01 -7.15397000e-02
2.53704190e-01 -7.01109767e-01 2.81026751e-01 1.86657143e+00
-3.14044990e-02 2.21808806e-01 -3.37277710e-01 6.95464671e-01
1.21802866e-01 -3.59941013e-02 1.18894744e+00 -4.62757826e-01
-9.22789156e-01 -1.53026775e-01 -1.53859198e+00 1.39225453e-01
1.39085913e+00 -6.16458058e-01 -3.58936101e-01 1.32454321e-01
-6.29779518e-01 3.49316835e-01 5.04150271e-01 6.93880394e-02
9.51794267e-01 -1.28637695e+00 -2.56294042e-01 2.93668449e-01
-2.29528189e-01 -2.77994961e-01 9.81781855e-02 8.02663565e-01
-7.41864562e-01 2.36971676e-03 -1.36553061e+00 -7.24190533e-01
-7.70931065e-01 7.67122284e-02 4.16155338e-01 1.23056434e-01
-5.88077247e-01 7.64966786e-01 -4.93734956e-01 -3.63299012e-01
-1.24605104e-01 -3.15311611e-01 5.79545908e-02 -2.78729230e-01
-1.31425187e-01 5.66959321e-01 -6.73516631e-01 -4.57842201e-02
-5.73739588e-01 1.02712357e+00 9.48370934e-01 -4.25896794e-01
1.57030094e+00 4.08607125e-02 -1.84458867e-01 4.70126957e-01
6.97606564e-01 -8.65200400e-01 -1.43515265e+00 3.79517078e-01
-1.27656788e-01 3.27313274e-01 -5.46006978e-01 -7.28721976e-01
-8.44300330e-01 1.09944451e+00 6.10548794e-01 -2.13873550e-01
1.09216917e+00 -1.74095824e-01 9.18128192e-01 1.00729096e+00
8.53192985e-01 -1.77626109e+00 9.37420070e-01 5.59948742e-01
8.34464550e-01 -8.19913745e-01 -4.44347709e-01 3.24611217e-01
-4.38017756e-01 1.33603811e+00 8.94459963e-01 -9.52535152e-01
5.31283021e-01 5.67371190e-01 -4.38451052e-01 1.30576212e-02
-1.26357758e+00 -3.36954487e-03 -4.88437563e-01 8.34671974e-01
-3.11914772e-01 -1.93124026e-01 -4.34021384e-01 2.65380293e-01
-9.67876464e-02 6.57648861e-01 6.57080114e-01 1.27442122e+00
-1.13737249e+00 -6.90607250e-01 -7.46488869e-02 1.95648134e-01
2.09354013e-01 4.47064072e-01 -3.80826682e-01 5.51130176e-01
4.19835120e-01 5.68078458e-01 1.73863173e-01 -4.16139662e-01
9.27554443e-02 1.31709293e-01 9.10168350e-01 -3.62665504e-01
-4.16753232e-01 -6.09647930e-01 -6.90009177e-01 -7.36442983e-01
-5.33294916e-01 -5.80719113e-01 -2.13889265e+00 -2.17713907e-01
6.83072031e-01 -9.91183817e-02 9.91313398e-01 7.72460461e-01
3.98838758e-01 4.89950091e-01 5.86857319e-01 -1.06159616e+00
-4.97303307e-01 -6.33142889e-01 -4.42917436e-01 -1.02520771e-01
-8.07721391e-02 -6.84614420e-01 -4.44579005e-01 4.44691330e-01] | [8.198049545288086, 2.3616535663604736] |
e9d52eba-db07-4ecb-b368-e1f95c155e19 | detecting-adversarial-examples-in-batches-a | 2206.08738 | null | https://arxiv.org/abs/2206.08738v1 | https://arxiv.org/pdf/2206.08738v1.pdf | Detecting Adversarial Examples in Batches -- a geometrical approach | Many deep learning methods have successfully solved complex tasks in computer vision and speech recognition applications. Nonetheless, the robustness of these models has been found to be vulnerable to perturbed inputs or adversarial examples, which are imperceptible to the human eye, but lead the model to erroneous output decisions. In this study, we adapt and introduce two geometric metrics, density and coverage, and evaluate their use in detecting adversarial samples in batches of unseen data. We empirically study these metrics using MNIST and two real-world biomedical datasets from MedMNIST, subjected to two different adversarial attacks. Our experiments show promising results for both metrics to detect adversarial examples. We believe that his work can lay the ground for further study on these metrics' use in deployed machine learning systems to monitor for possible attacks by adversarial examples or related pathologies such as dataset shift. | ['Peter Steinbach', 'Danush Kumar Venkatesh'] | 2022-06-17 | null | null | null | null | ['adversarial-attack-detection', 'adversarial-attack-detection'] | ['computer-vision', 'knowledge-base'] | [ 5.15693724e-01 3.00894082e-01 4.61885601e-01 -3.29975009e-01
-7.12823808e-01 -9.60640728e-01 8.84015620e-01 3.96968096e-01
-5.73618412e-01 9.03171301e-01 -3.38304162e-01 -3.36573780e-01
3.30640152e-02 -6.65594876e-01 -8.53771210e-01 -7.66876757e-01
-4.78056252e-01 4.46095258e-01 4.06774253e-01 2.86605097e-02
4.58233133e-02 8.78418684e-01 -1.16700733e+00 4.89221752e-01
2.73412347e-01 7.28596330e-01 -6.10076308e-01 9.37012851e-01
4.38062072e-01 8.29699576e-01 -1.20282197e+00 -7.14216650e-01
4.62113082e-01 -1.17954120e-01 -8.11587870e-01 -1.93753451e-01
6.09174252e-01 -2.06302822e-01 -4.87320036e-01 1.30061245e+00
7.83152640e-01 -1.72496393e-01 6.08581781e-01 -1.28166580e+00
-4.79145795e-01 6.16632104e-01 -1.22670338e-01 7.47389436e-01
2.35326558e-01 6.47076786e-01 3.30871671e-01 -4.04279470e-01
6.35518909e-01 1.28625476e+00 7.39624560e-01 9.47444916e-01
-1.19274104e+00 -6.65690362e-01 -3.80584896e-01 4.39060554e-02
-1.13736248e+00 -6.54902875e-01 5.86429596e-01 -5.49657166e-01
6.74002767e-01 5.70441425e-01 -9.58456099e-03 1.82289171e+00
5.64291239e-01 4.84266043e-01 1.19948828e+00 -2.34169051e-01
5.22821069e-01 5.25578618e-01 -2.82740239e-02 4.52120960e-01
2.94470549e-01 6.98050618e-01 -2.92563170e-01 -5.21744132e-01
4.06771839e-01 -2.30967924e-01 -3.22050273e-01 8.23514536e-02
-1.04471791e+00 8.82480562e-01 4.95313674e-01 4.03758973e-01
-2.64075309e-01 7.87423700e-02 6.19712591e-01 6.53723657e-01
5.08298755e-01 9.03059244e-01 -5.06907165e-01 1.71029180e-01
-6.07428908e-01 2.55278170e-01 7.84014225e-01 6.65325224e-01
1.57070369e-01 2.43194997e-01 -3.04165900e-01 6.26125455e-01
9.29877162e-02 4.51232374e-01 5.73876977e-01 -5.11533856e-01
1.97834566e-01 2.14058474e-01 -9.33101326e-02 -1.09323180e+00
-4.45797026e-01 -4.84639317e-01 -6.97729886e-01 3.70552182e-01
6.19960427e-01 -3.34684044e-01 -1.01686049e+00 1.51465952e+00
2.02553421e-01 3.53729546e-01 1.84857100e-01 6.74502373e-01
8.09855342e-01 1.75406680e-01 2.10850313e-01 1.38202727e-01
1.13485169e+00 -4.26340371e-01 -5.22077799e-01 -3.64366680e-01
6.26022041e-01 -8.04554164e-01 7.96340704e-01 5.36344826e-01
-1.04308879e+00 -3.57925713e-01 -1.14302719e+00 4.52893078e-01
-6.65247977e-01 -4.00153548e-01 2.89680272e-01 1.18014073e+00
-6.84459269e-01 9.48492110e-01 -8.94618154e-01 -3.86898190e-01
8.04094732e-01 4.94149357e-01 -3.90391171e-01 1.31889479e-02
-1.43555832e+00 9.49080288e-01 3.14497471e-01 6.07737936e-02
-1.28206265e+00 -6.06010497e-01 -6.50098681e-01 -3.57473552e-01
-7.68121406e-02 -4.68146473e-01 1.09809041e+00 -9.75443542e-01
-9.41419721e-01 1.36475825e+00 4.85107452e-01 -9.92952466e-01
9.77943838e-01 -3.72473407e-03 -8.73978257e-01 1.31017223e-01
-3.45682502e-01 4.32146132e-01 9.57143784e-01 -1.21728015e+00
-2.25027949e-01 -4.22386199e-01 9.85472575e-02 -4.93846059e-01
-4.91749734e-01 1.70552552e-01 2.34946892e-01 -9.55049396e-01
-3.16670150e-01 -9.15024698e-01 -2.88656026e-01 7.09926188e-02
-7.16333270e-01 2.04454437e-01 7.77803361e-01 -3.15998644e-01
7.55856872e-01 -2.11491060e+00 -4.16679829e-01 2.80030131e-01
2.86752522e-01 8.20774913e-01 -2.94040918e-01 2.67842233e-01
-9.63887647e-02 3.97043139e-01 -3.69559050e-01 -1.36385635e-01
-1.97749361e-01 3.52414101e-01 -5.30557275e-01 8.95545721e-01
4.31241721e-01 8.76845181e-01 -9.50722516e-01 -6.34135753e-02
1.56175941e-01 5.37041247e-01 -2.62475371e-01 2.76298344e-01
-1.17233731e-01 5.36492407e-01 -4.27588046e-01 5.94824314e-01
6.36968195e-01 1.78390443e-01 -2.41434082e-01 -3.28356437e-02
7.08200216e-01 1.70920547e-02 -8.40164244e-01 8.14275980e-01
-1.60297379e-01 8.13130140e-01 -1.35094985e-01 -1.03880668e+00
8.39389026e-01 3.78658056e-01 -1.04643278e-01 -4.26703572e-01
5.44806302e-01 1.52013943e-01 2.63041526e-01 -5.69410205e-01
2.26361211e-02 -2.58586019e-01 -9.44108814e-02 1.91219226e-01
8.24124441e-02 -1.01971487e-03 -2.19276786e-01 4.51351367e-02
1.56524980e+00 -5.70782602e-01 4.09116149e-02 -3.44038367e-01
7.02631235e-01 -1.77294966e-02 3.15372080e-01 1.14749908e+00
-5.91456294e-01 6.62595689e-01 4.46498990e-01 -8.29743207e-01
-1.05510366e+00 -1.27268028e+00 -3.83431137e-01 6.16125941e-01
-1.09800681e-01 2.84922481e-01 -9.71893847e-01 -1.10551822e+00
1.28366962e-01 7.60239780e-01 -9.78513479e-01 -7.07860231e-01
-5.00306487e-01 -8.96661758e-01 1.38598931e+00 5.03629088e-01
6.94937259e-02 -1.39967227e+00 -5.89368939e-01 1.81216866e-01
4.63526487e-01 -1.40114343e+00 -2.19989598e-01 3.17951113e-01
-6.59088492e-01 -1.41485333e+00 -3.92227203e-01 -4.82912123e-01
7.87051380e-01 -4.07069057e-01 1.27040184e+00 5.30825555e-01
-8.85602117e-01 4.52645451e-01 -5.28678060e-01 -9.41497982e-01
-1.34263325e+00 -2.29249224e-01 3.81333649e-01 3.05103417e-02
5.43260634e-01 -3.77141625e-01 -6.13336921e-01 4.81549591e-01
-1.50968635e+00 -8.71002257e-01 2.58564025e-01 6.62628531e-01
2.63725728e-01 -2.10340768e-02 7.09879398e-01 -1.28341103e+00
8.82631183e-01 -6.39308929e-01 -4.43254530e-01 -2.23731510e-02
-3.95154357e-01 -2.30751708e-02 1.00692356e+00 -7.28925765e-01
-2.89662272e-01 -1.26631603e-01 -4.66995806e-01 -7.14914024e-01
-6.61814392e-01 6.33228198e-02 -5.48228323e-02 -3.39083374e-01
1.21782947e+00 -3.35974693e-02 -1.68448940e-01 -1.64288357e-01
-3.96335982e-02 6.39193416e-01 6.69165611e-01 -3.09488922e-01
1.11525059e+00 5.32665610e-01 -1.24855824e-02 -8.80661547e-01
-5.65889120e-01 -8.34699199e-02 -4.15294677e-01 -1.01701215e-01
5.69016874e-01 -5.16492367e-01 -4.43229377e-01 7.01669931e-01
-1.11588109e+00 -1.94211394e-01 -2.60300815e-01 1.21992826e-01
-2.99783260e-01 3.63915473e-01 -6.59663498e-01 -7.08958924e-01
-3.76691461e-01 -1.22037721e+00 9.01708066e-01 1.44941518e-02
-3.78433675e-01 -1.33486426e+00 5.38827777e-02 1.89716786e-01
4.41807300e-01 8.92798245e-01 7.32354105e-01 -1.51861453e+00
-7.01546371e-02 -7.89180219e-01 3.46535414e-01 7.10730016e-01
1.04804054e-01 7.51692876e-02 -1.52469742e+00 -5.92253149e-01
1.11545496e-01 -5.79410791e-01 7.84761846e-01 -1.10180117e-02
1.20091331e+00 -6.09825194e-01 -3.79844755e-01 6.13612592e-01
1.22426260e+00 2.17451572e-01 8.09483826e-01 3.88818204e-01
4.05461013e-01 6.22149467e-01 3.00042927e-01 1.56295910e-01
-5.30962408e-01 3.89819145e-01 7.40418196e-01 -3.02491754e-01
9.53328684e-02 8.90522897e-02 3.72308552e-01 7.62405843e-02
3.50114375e-01 -6.21144354e-01 -1.02360618e+00 4.49618429e-01
-1.32048702e+00 -1.02851963e+00 -3.94557714e-02 2.15404105e+00
7.98710883e-01 5.81874669e-01 -3.86475362e-02 4.17622358e-01
6.91629946e-01 -6.01158589e-02 -6.95891201e-01 -6.41288102e-01
-2.12335259e-01 4.60512400e-01 6.55054927e-01 2.46810868e-01
-1.39220774e+00 8.46130013e-01 6.72636700e+00 5.79868793e-01
-1.14025795e+00 9.82217025e-03 9.99708593e-01 -1.87958583e-01
4.05222364e-02 -6.12877131e-01 -4.21105355e-01 5.30782163e-01
1.08817136e+00 1.25497356e-01 2.63072461e-01 7.59223521e-01
8.48308802e-02 2.69155443e-01 -1.32253242e+00 5.48515677e-01
6.87831268e-02 -1.24256670e+00 3.60043608e-02 -7.27524310e-02
6.51805997e-01 1.88165858e-01 3.60988915e-01 1.89799935e-01
6.82664573e-01 -1.38742769e+00 5.28002977e-01 3.23748231e-01
4.17409718e-01 -8.77442479e-01 1.10354328e+00 3.27734083e-01
-3.70795727e-01 6.40597790e-02 -3.63167167e-01 2.67756194e-01
-2.08660945e-01 5.90584815e-01 -1.29749894e+00 1.40614554e-01
6.73555613e-01 1.02835804e-01 -7.99573064e-01 7.99553931e-01
-7.67184645e-02 7.61010885e-01 -2.15854838e-01 5.39220273e-02
2.77680516e-01 6.15774393e-01 8.59502435e-01 1.23776865e+00
-1.90740481e-01 -3.82940918e-01 -7.51544163e-02 8.38513315e-01
-2.04102233e-01 -2.31863275e-01 -9.25373912e-01 -1.05613098e-01
5.10868549e-01 1.07441163e+00 -7.88643658e-01 -3.86544429e-02
8.18908885e-02 1.12015963e+00 5.99446595e-02 2.61849105e-01
-8.99079800e-01 -3.55313212e-01 9.42841470e-01 1.51711285e-01
3.43771689e-02 2.30684176e-01 -2.24370584e-01 -7.98715353e-01
3.61052006e-02 -1.26456296e+00 3.68318021e-01 -2.71253139e-01
-1.69252205e+00 1.15473378e+00 -2.48544037e-01 -1.15462601e+00
-2.36701876e-01 -9.61664557e-01 -7.49914348e-01 7.08197415e-01
-1.07006478e+00 -9.50166047e-01 1.00142017e-01 7.15545297e-01
3.26071233e-01 -5.69942355e-01 9.25351262e-01 3.39042991e-02
-5.46928704e-01 1.22907388e+00 1.29684776e-01 5.34012794e-01
6.49339259e-01 -1.10673320e+00 9.15645123e-01 9.65256155e-01
3.20443511e-01 5.74049413e-01 1.19213784e+00 -4.95266020e-01
-9.45276797e-01 -1.41340625e+00 3.26669544e-01 -9.69662130e-01
7.10052729e-01 -5.31750798e-01 -1.00876224e+00 6.59163475e-01
-2.33532134e-02 3.39904398e-01 7.42179096e-01 -3.14230621e-01
-4.39839512e-01 2.31737301e-01 -1.81729245e+00 4.24185097e-01
6.83983147e-01 -4.30472046e-01 -6.14878595e-01 5.63010395e-01
5.69674909e-01 -3.73961568e-01 -1.04680347e+00 7.29970455e-01
3.08527529e-01 -1.08805633e+00 1.09850442e+00 -1.12248707e+00
2.90419549e-01 -8.75293091e-02 -5.17557040e-02 -1.42528534e+00
-1.10183218e-02 -6.46995187e-01 -1.02990024e-01 1.12444019e+00
5.72565258e-01 -7.55868316e-01 6.00639164e-01 4.92094219e-01
9.94856283e-02 -7.79944897e-01 -1.01079893e+00 -8.47261846e-01
3.66321981e-01 -5.86162984e-01 6.11469567e-01 1.19373095e+00
-4.27728981e-01 -5.16876765e-02 -1.42223179e-01 6.34630501e-01
4.17950690e-01 -8.42234313e-01 7.34449625e-01 -1.09404194e+00
-2.02793986e-01 -3.39253157e-01 -9.95963991e-01 6.70541823e-02
-5.43448608e-03 -7.80669987e-01 -8.78631994e-02 -8.06626678e-01
-2.82924235e-01 -4.18514222e-01 -4.27512288e-01 4.20875520e-01
-2.63214588e-01 7.64478087e-01 6.20943047e-02 3.60202566e-02
-1.43898100e-01 2.42946520e-02 6.92417979e-01 -4.36662108e-01
1.96125850e-01 2.89654523e-01 -5.24400651e-01 9.35953438e-01
9.73895848e-01 -1.01306391e+00 -2.49540925e-01 -3.48723441e-01
4.78748903e-02 -5.42255223e-01 6.50914311e-01 -1.20294845e+00
-1.02306739e-03 9.31907669e-02 6.40757799e-01 -1.10866539e-01
-2.23051496e-02 -1.01878655e+00 -1.03843644e-01 7.37455547e-01
-5.08784175e-01 6.69622570e-02 5.86653888e-01 5.42409360e-01
9.32028070e-02 -2.54793972e-01 1.21829486e+00 -1.37751237e-01
-2.53894508e-01 3.76744717e-01 -4.37404513e-01 1.96608201e-01
1.42592239e+00 -1.57230482e-01 -4.78902370e-01 -1.39233351e-01
-9.49495733e-01 5.03336377e-02 3.44786137e-01 6.62727416e-01
7.29098380e-01 -1.00617445e+00 -9.53571498e-01 4.92613673e-01
1.10375702e-01 -3.60848546e-01 5.70083261e-02 4.95802820e-01
-6.98773086e-01 9.69028994e-02 -1.24942593e-01 -6.52943909e-01
-1.51831627e+00 9.64658439e-01 6.77746713e-01 -2.49659583e-01
-2.95143783e-01 1.03661346e+00 -6.83370084e-02 -5.05439401e-01
4.87614095e-01 -1.72426417e-01 2.44135931e-02 -2.47121304e-01
7.19865203e-01 2.38700047e-01 3.89330000e-01 -4.90992844e-01
-5.96334815e-01 -1.24368399e-01 -4.47426558e-01 2.72340685e-01
1.11587965e+00 3.97062510e-01 2.46667072e-01 2.36196458e-01
1.28146100e+00 -4.01152112e-02 -8.88606727e-01 -1.88568071e-01
1.31217211e-01 -2.77344555e-01 -2.09112465e-01 -1.01711285e+00
-1.15779245e+00 8.61069977e-01 9.91811931e-01 6.92919433e-01
9.96710241e-01 1.76849104e-02 6.89013481e-01 3.49946529e-01
2.24832788e-01 -5.72584093e-01 5.91543652e-02 2.20859259e-01
7.74585664e-01 -1.11624050e+00 -7.16094077e-02 -8.81078988e-02
-4.63730514e-01 9.64338601e-01 5.54855108e-01 -2.84646273e-01
5.99725485e-01 6.43860102e-01 5.68700910e-01 -1.77344471e-01
-6.33556366e-01 2.93372035e-01 2.68154174e-01 9.44850981e-01
2.18460709e-01 -8.02241080e-03 1.66473463e-01 2.48106778e-01
-4.36894476e-01 -2.66767323e-01 5.89799643e-01 8.34306657e-01
-1.71146855e-01 -9.77390409e-01 -6.78177357e-01 5.68822801e-01
-1.05474412e+00 -1.76020488e-01 -8.60014975e-01 7.79913843e-01
1.43599182e-01 9.93807793e-01 4.33350615e-02 -6.27516806e-01
5.50557792e-01 -5.81268184e-02 3.50682586e-01 -4.23400879e-01
-1.10510755e+00 -5.40053368e-01 -5.03406413e-02 -4.54591990e-01
-1.14530243e-01 -5.76547027e-01 -7.88161635e-01 -2.72481948e-01
-2.72182792e-01 -1.40664056e-01 3.76771003e-01 1.04737699e+00
2.13974491e-01 7.24059105e-01 7.40156710e-01 -5.33845425e-01
-8.91880929e-01 -1.12296939e+00 -4.20554727e-01 7.76853621e-01
6.96795464e-01 -3.70476335e-01 -7.93075502e-01 1.02891251e-01] | [5.655708312988281, 7.79729700088501] |
3c726f66-8a5c-4217-8fb7-d4da50ebafe6 | domain-adaptive-cascade-r-cnn-for-mitosis | 2109.00965 | null | https://arxiv.org/abs/2109.00965v2 | https://arxiv.org/pdf/2109.00965v2.pdf | Domain Adaptive Cascade R-CNN for MItosis DOmain Generalization (MIDOG) Challenge | We present a summary of the domain adaptive cascade R-CNN method for mitosis detection of digital histopathology images. By comprehensive data augmentation and adapting existing popular detection architecture, our proposed method has achieved an F1 score of 0.7500 on the preliminary test set in MItosis DOmain Generalization (MIDOG) Challenge at MICCAI 2021. | ['Jingxin Liu', 'Lian Liu', 'Xiao Mu', 'Ying Cheng', 'Xi Long'] | 2021-09-01 | null | null | null | null | ['mitosis-detection'] | ['medical'] | [ 4.24679369e-01 2.13983461e-01 -5.83421350e-01 -1.38630271e-01
-1.07465839e+00 -1.79423988e-01 4.75322902e-01 4.75066155e-01
-8.47734869e-01 1.11456311e+00 1.18912019e-01 -4.04285252e-01
1.75843894e-01 -3.47270459e-01 -1.52923584e-01 -1.25337803e+00
-1.52051687e-01 4.62134212e-01 3.45069170e-01 -3.28818828e-01
-3.70462164e-02 7.21168935e-01 -7.23436952e-01 5.59449375e-01
2.24213272e-01 9.51268315e-01 -2.67568916e-01 1.31545734e+00
5.84884286e-01 6.70382619e-01 -7.93144524e-01 -1.88415140e-01
-3.56227368e-01 -2.30638534e-01 -1.14893317e+00 -6.11108122e-03
2.18123719e-01 -3.82968009e-01 -6.38830185e-01 9.40611660e-01
1.04745436e+00 -2.84920722e-01 1.15189373e+00 -7.49066889e-01
-4.94697690e-01 6.51636541e-01 -8.04429829e-01 1.41437101e+00
-3.72851372e-01 1.14416577e-01 4.33291346e-01 -8.03251624e-01
1.21115804e+00 7.43305445e-01 8.96300077e-01 1.06604576e+00
-1.25037646e+00 -7.15203941e-01 -4.45254177e-01 1.16990045e-01
-1.72768331e+00 -2.01697171e-01 4.80007231e-01 -1.87540755e-01
1.33290076e+00 -1.62959144e-01 4.22128499e-01 1.40550172e+00
6.66733921e-01 1.05570102e+00 9.24149990e-01 3.18990126e-02
1.86528832e-01 -1.48154140e-01 4.31410521e-01 3.72123271e-01
3.40679973e-01 -4.35874686e-02 -5.54248035e-01 -2.06249177e-01
1.10558808e+00 -2.40891695e-01 -1.40233055e-01 2.73899287e-01
-1.25266290e+00 7.85786748e-01 8.21259856e-01 4.19187158e-01
-8.38675946e-02 1.12549685e-01 9.46670175e-01 4.59546387e-01
8.17423940e-01 1.90811753e-02 -4.27426845e-01 4.64196354e-01
-7.87291288e-01 1.32805528e-02 1.76315621e-01 6.23765290e-01
-2.07004234e-01 -1.54308811e-01 -4.69894230e-01 4.37522709e-01
6.46660328e-02 1.57019466e-01 1.02045059e+00 -2.79424667e-01
-5.18166684e-02 8.11090231e-01 -4.36780334e-01 -2.30955154e-01
-9.35592413e-01 -9.83564317e-01 -1.78185618e+00 -4.01997231e-02
6.07513964e-01 -2.19753440e-02 -1.56962204e+00 1.09649706e+00
2.86759824e-01 3.70403558e-01 4.31654155e-01 6.98933601e-01
1.63909137e+00 1.55114323e-01 5.80418289e-01 -1.40247479e-01
1.33029592e+00 -6.28694534e-01 -5.53906560e-01 1.73208907e-01
1.14295626e+00 -3.95691514e-01 5.94413400e-01 3.01117003e-01
-7.28432238e-01 -2.39907697e-01 -1.22361004e+00 -4.90123808e-01
-7.21361279e-01 2.98162639e-01 6.83397830e-01 6.08711243e-01
-1.51849055e+00 2.32948750e-01 -1.08965623e+00 -5.64041555e-01
1.20136559e+00 9.57962453e-01 -6.09391391e-01 -1.51226029e-01
-7.66723394e-01 5.50353110e-01 4.51863348e-01 -2.36700416e-01
-1.31566346e+00 -9.84483004e-01 -4.23956990e-01 -4.52779710e-01
-5.25604427e-01 -1.00129294e+00 1.33670306e+00 -4.41354185e-01
-1.04643250e+00 1.55612493e+00 -1.30063206e-01 -1.02428508e+00
3.40318114e-01 4.16675985e-01 -3.05814177e-01 5.19956827e-01
-5.68472594e-02 1.40232956e+00 1.61066175e-01 -5.26371062e-01
-7.48228848e-01 -4.46214795e-01 -3.85464251e-01 4.77636204e-04
-4.43342894e-01 -1.10189110e-01 -4.31573577e-02 -9.90240037e-01
-3.32914926e-02 -8.25158298e-01 -6.62267804e-01 7.78188258e-02
-2.26713240e-01 -6.29298687e-01 7.66005099e-01 -6.09522223e-01
5.91263950e-01 -1.90704036e+00 7.46258274e-02 1.08837314e-01
5.01299560e-01 1.97135314e-01 -9.13112760e-02 -5.22549152e-01
-3.81232679e-01 1.03865638e-01 -1.30896524e-01 -4.53334808e-01
-7.57456243e-01 -1.09639801e-01 2.10854053e-01 9.20825779e-01
3.05153608e-01 1.15435123e+00 -9.13288176e-01 -7.31860638e-01
-2.92301208e-01 7.91109324e-01 -3.92169595e-01 -1.27070382e-01
7.58257583e-02 4.65077996e-01 3.57387736e-02 1.37604594e+00
6.48916125e-01 -7.18931794e-01 3.31938900e-02 1.37382932e-02
6.71426177e-01 -8.68216753e-02 -4.15352136e-01 1.74518514e+00
3.13384801e-01 7.73959935e-01 -1.89761207e-01 -8.46871734e-01
7.41023719e-01 5.51466465e-01 4.75634307e-01 -4.65264261e-01
5.73324025e-01 1.78965509e-01 1.64898083e-01 6.36584684e-02
1.71051696e-01 -2.20230207e-01 -7.05350935e-02 -5.27113192e-02
6.49513543e-01 2.81547517e-01 1.39986813e-01 2.35347256e-01
1.56912577e+00 -5.82371891e-01 8.10518920e-01 -5.11488378e-01
7.95835137e-01 2.29841918e-01 5.43065190e-01 6.68888032e-01
-6.49998784e-01 6.84009731e-01 5.20016074e-01 -7.31362462e-01
-1.00365639e+00 -9.33311999e-01 -4.57655072e-01 1.01571035e+00
-2.29293287e-01 1.23865351e-01 -3.66122454e-01 -1.12640667e+00
-6.17532469e-02 -2.34134778e-01 -1.34168303e+00 -2.73110420e-01
-6.60450637e-01 -1.49939978e+00 1.20062506e+00 1.01286530e+00
7.29735017e-01 -9.20543194e-01 -2.24052407e-02 1.53755769e-01
8.23576003e-02 -1.03287435e+00 -3.77914816e-01 5.99270225e-01
-1.38015354e+00 -1.09604716e+00 -1.23201251e+00 -1.49216402e+00
1.03091228e+00 1.71094507e-01 1.13204181e+00 1.88703999e-01
-9.61227179e-01 -1.19257607e-01 -1.68042257e-01 -6.51991010e-01
-5.97474217e-01 3.95687610e-01 -2.46976465e-01 -6.57108665e-01
8.28100920e-01 -2.08942041e-01 -9.04682755e-01 4.93342464e-04
-8.81296337e-01 -2.88440734e-01 1.04253757e+00 1.18356812e+00
1.23103726e+00 -7.92957935e-03 8.45897734e-01 -1.02923608e+00
4.37985182e-01 -5.77873170e-01 -4.46684398e-02 -1.42407373e-01
-3.31562430e-01 -5.88391066e-01 3.78817350e-01 -7.24120319e-01
-7.26518691e-01 4.83433157e-01 -2.00348511e-01 -3.04310769e-01
-1.43240035e-01 3.88078839e-01 3.57381701e-01 -4.95153248e-01
1.19311571e+00 3.53423566e-01 2.64592655e-02 -7.10232481e-02
-1.92993775e-01 6.03031516e-01 1.11149919e+00 2.61947811e-01
3.41021657e-01 7.94782102e-01 5.27588487e-01 -7.14123011e-01
-4.94937003e-01 -8.32871735e-01 -7.48678446e-01 -5.58502749e-02
7.20100820e-01 -1.33418167e+00 -6.23436332e-01 9.66480196e-01
-7.92238414e-01 -4.73518968e-01 -1.02125928e-01 1.37472644e-01
-2.93905497e-01 5.89771681e-02 -1.26631057e+00 -4.90626739e-03
-9.65215504e-01 -8.17025900e-01 1.18784499e+00 5.46194911e-01
-3.26690733e-01 -1.13959014e+00 2.87376791e-01 3.91410321e-01
4.37718034e-01 5.49381256e-01 9.53198612e-01 -1.08460903e+00
-9.83834825e-03 -5.77513278e-01 -2.51671106e-01 -4.28353203e-03
1.39086649e-01 -8.56301114e-02 -1.05049276e+00 -8.58190894e-01
-5.61373830e-01 -3.64856988e-01 1.58456254e+00 8.20127726e-01
1.44005132e+00 3.89645882e-02 -1.09209073e+00 6.12481177e-01
1.48341835e+00 -5.32578751e-02 6.68421686e-01 4.92911845e-01
9.11178440e-02 2.14444138e-02 2.73036510e-01 4.13691700e-02
1.59793779e-01 -2.54026074e-02 2.24001810e-01 -7.99473524e-01
-5.92891335e-01 2.51137614e-01 -2.24818990e-01 1.05633736e-01
1.90200750e-02 -2.58245170e-01 -1.38413143e+00 1.07362032e+00
-1.36932218e+00 -3.60709190e-01 5.21073118e-02 1.39547396e+00
1.09996605e+00 4.06248629e-01 4.05964255e-01 4.91940707e-01
7.96391428e-01 -4.88252163e-01 -7.57314801e-01 -1.29683092e-01
-3.99913847e-01 5.24210930e-01 8.54567170e-01 1.09688781e-01
-1.80704117e+00 8.93234909e-01 7.43322229e+00 1.01307762e+00
-9.51722085e-01 2.75064558e-01 1.34761012e+00 -2.04122588e-02
6.44672871e-01 -8.40253890e-01 -9.65217352e-01 -8.33146572e-02
1.20171773e+00 -4.78208840e-01 -4.10416096e-01 6.99278235e-01
-2.38223001e-01 2.23028976e-02 -1.05202234e+00 7.30797350e-01
-2.74924608e-03 -1.81869340e+00 2.92897612e-01 3.55631590e-01
9.17601228e-01 2.72985131e-01 5.72910786e-01 4.17903155e-01
2.90430754e-01 -1.55908489e+00 -5.78759670e-01 1.97704524e-01
1.09301054e+00 -1.06761956e+00 1.71251893e+00 1.81536581e-02
-8.21978271e-01 -4.79313023e-02 -4.60551500e-01 3.92484903e-01
-7.11951315e-01 5.18778384e-01 -1.92694759e+00 7.74751976e-02
6.59721911e-01 8.70092213e-01 -9.26207960e-01 1.39653540e+00
1.79738492e-01 8.92150998e-01 -2.32019722e-01 -1.41690984e-01
1.35387823e-01 1.06665087e+00 4.03549403e-01 1.83493233e+00
-1.69881210e-01 2.25840226e-01 -1.71656638e-01 1.87675357e-01
-6.90425098e-01 -8.79810154e-02 -2.95708507e-01 1.65932715e-01
1.74512893e-01 1.48987210e+00 -1.55555546e+00 -3.07005286e-01
4.44592200e-02 4.23983306e-01 1.66051257e-02 1.04286551e-01
-4.92108047e-01 -3.00177634e-01 5.14197528e-01 2.22210120e-02
2.78866053e-01 1.98313817e-01 -6.02577627e-01 -4.44419682e-01
-7.71654606e-01 -8.50766718e-01 1.06412446e+00 -2.26987928e-01
-1.56988382e+00 4.67823625e-01 -3.49401057e-01 -1.19973993e+00
2.47877061e-01 -7.64146149e-01 -5.29848754e-01 3.93526256e-01
-1.67177057e+00 -1.42084503e+00 -3.01161945e-01 6.84705794e-01
5.15319407e-01 -4.80382830e-01 9.14120138e-01 2.15365291e-01
-5.01863599e-01 1.32941759e+00 4.75060567e-02 5.79377830e-01
7.96880305e-01 -1.48540580e+00 6.90255105e-01 4.31811512e-01
-5.98962009e-01 4.41240400e-01 4.74922359e-01 -6.93413615e-01
-7.57775724e-01 -1.59009516e+00 5.99976301e-01 -2.90528804e-01
5.18486559e-01 -9.66969207e-02 -9.07920301e-01 5.91496348e-01
1.00585066e-01 3.82418871e-01 1.21769667e+00 -3.20980161e-01
-1.65893033e-01 1.17187552e-01 -1.77670395e+00 3.53198111e-01
6.84301019e-01 -2.67444938e-01 -6.31366670e-02 7.48354197e-01
7.38376558e-01 -7.19441414e-01 -1.33775616e+00 1.02350914e+00
7.56346679e-04 -1.05305694e-01 1.09556508e+00 -8.65627646e-01
3.50566983e-01 -2.10547134e-01 2.21321627e-01 -8.57751369e-01
-6.03649259e-01 -2.95463413e-01 -1.18760839e-01 7.19440460e-01
4.70553637e-01 -2.57621527e-01 1.60930073e+00 -3.83430719e-01
-3.00049651e-02 -1.15416992e+00 -1.67105103e+00 -4.17642236e-01
7.50444114e-01 2.96135694e-01 9.63007882e-02 6.03762507e-01
9.85300913e-02 3.26324195e-01 2.14526266e-01 3.13159049e-01
7.02125132e-01 -5.54463387e-01 3.86918485e-01 -1.17404151e+00
9.39642861e-02 -6.75682008e-01 -1.13254893e+00 -6.32938743e-01
2.09457595e-02 -1.25891781e+00 -3.26811016e-01 -1.10796738e+00
7.56508827e-01 2.93266982e-01 -1.00278163e+00 4.83298272e-01
-8.14652890e-02 1.12087834e+00 -6.81661606e-01 1.31518885e-01
-9.65904355e-01 -2.10952297e-01 1.25703049e+00 -8.42771769e-01
-6.95453063e-02 -3.30645256e-02 -8.61028433e-01 4.47997659e-01
9.15630758e-01 -6.03983879e-01 -4.96771373e-02 1.59810498e-01
-1.56577259e-01 1.20410249e-02 3.52551907e-01 -1.24766243e+00
5.97674012e-01 2.22640932e-01 1.33315134e+00 -1.00347626e+00
1.31439626e-01 -2.76626758e-02 -3.79513949e-01 8.57028008e-01
-5.85159302e-01 -4.17825341e-01 5.61485469e-01 8.33120823e-01
-6.42652065e-02 3.89132112e-01 1.11979795e+00 2.51159910e-02
-5.16170025e-01 4.29809302e-01 -8.43631327e-01 -2.65863329e-01
1.10375822e+00 -5.02739608e-01 -9.57599819e-01 2.82533675e-01
-1.08261192e+00 4.15598780e-01 2.25822687e-01 3.90710607e-02
7.93426692e-01 -1.17349172e+00 -1.11496449e+00 4.31631953e-02
2.94497877e-01 2.09115446e-01 2.22124472e-01 8.49548876e-01
-6.45010889e-01 6.82751060e-01 -3.79471391e-01 -9.47697520e-01
-1.75133121e+00 2.89523095e-01 7.65661001e-01 -8.36479545e-01
-4.62363243e-01 1.72267711e+00 1.62662953e-01 5.83474175e-04
4.02575433e-01 -5.04499197e-01 -6.45723939e-01 -2.44534209e-01
6.69277668e-01 3.67804438e-01 5.66854417e-01 -4.32527155e-01
-7.48946905e-01 7.05514848e-02 -9.24665093e-01 5.54741919e-01
1.29617739e+00 3.86578381e-01 -6.49638101e-03 -1.01180740e-01
1.36039269e+00 -9.17868257e-01 -7.35569179e-01 -2.70450085e-01
-1.36250339e-03 4.61083800e-01 3.07811975e-01 -1.16977692e+00
-1.07230031e+00 6.14318192e-01 1.37601566e+00 -3.77065510e-01
1.36667418e+00 3.03434610e-01 7.91246831e-01 7.06524074e-01
2.58319196e-03 -8.37413192e-01 2.61186779e-01 4.04721290e-01
5.18854678e-01 -1.27280986e+00 1.99199900e-01 -3.66027564e-01
-3.91549826e-01 1.25912929e+00 6.09606743e-01 -3.50808114e-01
6.65824115e-01 6.15607321e-01 1.87412664e-01 -3.31297666e-01
-1.12262666e+00 -8.93892944e-02 5.58650680e-02 1.06084013e+00
5.24929106e-01 8.95140246e-02 -9.43841487e-02 7.68803120e-01
-3.08137364e-03 3.43381822e-01 8.06312919e-01 8.83033693e-01
-4.71040100e-01 -5.56061625e-01 -3.45450006e-02 4.67962235e-01
-1.00891042e+00 2.95193307e-02 -8.92970204e-01 9.58011687e-01
-1.33773342e-01 5.73144138e-01 1.56135276e-01 -1.60132200e-01
5.55620119e-02 -2.58724213e-01 3.93196374e-01 -5.64287961e-01
-6.75975502e-01 1.06926493e-01 -2.51246184e-01 2.43952405e-02
-6.79431200e-01 -4.68942940e-01 -1.47483170e+00 -1.03450790e-01
-9.66022313e-02 -2.75878012e-01 2.78547525e-01 5.46948552e-01
2.08929166e-01 8.11304510e-01 3.20144475e-01 -6.05821133e-01
-1.73921913e-01 -1.24557829e+00 -6.64272904e-01 -1.15269907e-01
8.07976186e-01 -5.18870234e-01 -3.22788060e-01 3.83209884e-01] | [15.1071195602417, -3.109243631362915] |
8b217f22-2258-4569-8870-2fbb623f8ea1 | time-series-as-images-vision-transformer-for | 2303.12799 | null | https://arxiv.org/abs/2303.12799v1 | https://arxiv.org/pdf/2303.12799v1.pdf | Time Series as Images: Vision Transformer for Irregularly Sampled Time Series | Irregularly sampled time series are becoming increasingly prevalent in various domains, especially in medical applications. Although different highly-customized methods have been proposed to tackle irregularity, how to effectively model their complicated dynamics and high sparsity is still an open problem. This paper studies the problem from a whole new perspective: transforming irregularly sampled time series into line graph images and adapting powerful vision transformers to perform time series classification in the same way as image classification. Our approach largely simplifies algorithm designs without assuming prior knowledge and can be potentially extended as a general-purpose framework. Despite its simplicity, we show that it substantially outperforms state-of-the-art specialized algorithms on several popular healthcare and human activity datasets. Especially in the challenging leave-sensors-out setting where a subset of variables is masked during testing, the performance improvement is up to 54.0\% in absolute F1 score points. Our code and data are available at \url{https://github.com/Leezekun/ViTST}. | ['Xifeng Yan', 'Shiyang Li', 'Zekun Li'] | 2023-03-01 | null | null | null | null | ['time-series-classification'] | ['time-series'] | [ 4.16863710e-01 -1.26511276e-01 -3.56722116e-01 -1.90129176e-01
-6.61345541e-01 -5.62995970e-01 5.39950550e-01 1.94847479e-01
-2.76371926e-01 5.98366916e-01 -1.09183028e-01 -2.19109952e-01
-2.28382125e-01 -4.60992128e-01 -6.18829668e-01 -8.07666183e-01
-3.89603317e-01 3.91099334e-01 -1.12260310e-02 -4.93890420e-02
-6.50306121e-02 2.22505078e-01 -1.30384040e+00 -9.00065228e-02
9.00353312e-01 9.42323625e-01 -3.21860388e-02 6.73293829e-01
3.20714176e-01 6.44417465e-01 -4.05366242e-01 -1.57441437e-01
1.80281386e-01 -4.31131124e-01 -4.52775747e-01 3.45097899e-01
2.33821288e-01 1.63048908e-01 -3.92589152e-01 9.68137920e-01
5.29345334e-01 -1.69649720e-02 3.64243358e-01 -1.12887704e+00
-4.48826790e-01 2.36271784e-01 -8.60617578e-01 5.71299434e-01
2.95791596e-01 1.44305259e-01 8.71426165e-01 -5.33383012e-01
6.09466314e-01 5.91837406e-01 8.17272604e-01 1.29017025e-01
-1.49412560e+00 -5.88539362e-01 7.81096891e-02 3.09171170e-01
-1.28918433e+00 -3.28832269e-01 1.02681363e+00 -6.04646325e-01
7.08615065e-01 3.64153892e-01 7.14925528e-01 1.37370133e+00
1.66864857e-01 8.57778549e-01 1.15659046e+00 -1.59698486e-01
1.23875484e-01 -3.44087064e-01 3.52568388e-01 6.74382806e-01
4.16424632e-01 -1.15226701e-01 -4.49710131e-01 -2.22189069e-01
8.77227783e-01 2.07259774e-01 -4.12205100e-01 -3.95048082e-01
-1.64766860e+00 7.92554557e-01 1.24107480e-01 4.77030307e-01
-5.12951672e-01 5.76008298e-02 4.01912779e-01 4.03394014e-01
5.61113358e-01 3.65007907e-01 -4.24612045e-01 -4.22871143e-01
-9.08046663e-01 2.00528368e-01 7.07642317e-01 6.15803421e-01
3.62591565e-01 1.54769957e-01 -6.22638986e-02 8.15843821e-01
-2.85632461e-01 4.76302505e-01 6.31154299e-01 -8.37235093e-01
3.32590401e-01 4.87497509e-01 -1.34258151e-01 -1.08287859e+00
-7.23458171e-01 -9.24482703e-01 -1.24723649e+00 -2.88698345e-01
5.63854039e-01 -1.79697841e-01 -7.70439386e-01 1.58653212e+00
4.67391312e-01 6.39946938e-01 -4.40953463e-01 7.95224547e-01
5.31382263e-01 3.58330429e-01 -2.07761034e-01 -6.47615552e-01
1.42380488e+00 -8.24580669e-01 -8.28010499e-01 -2.22263351e-01
4.56069678e-01 -7.28321970e-01 1.02068758e+00 6.95252717e-01
-8.24638546e-01 -4.16976124e-01 -9.13364470e-01 1.50270760e-02
-6.25652680e-03 8.48458186e-02 7.66659319e-01 3.37973207e-01
-8.03614199e-01 6.68895900e-01 -1.35642099e+00 -3.02639127e-01
5.92181087e-01 2.82703400e-01 -3.16860884e-01 3.46652456e-02
-8.93625915e-01 4.56550181e-01 -1.39370278e-01 -1.47789279e-02
-6.50068581e-01 -8.35253000e-01 -7.91232884e-01 -3.04781199e-01
6.31034315e-01 -7.45243251e-01 1.25123668e+00 -7.45915413e-01
-1.16746318e+00 8.69450867e-01 -1.99106783e-01 -6.55593216e-01
7.62062609e-01 -2.69416392e-01 -3.83220702e-01 1.31763160e-01
1.49399444e-01 -8.39442313e-02 1.01452720e+00 -6.76654518e-01
-2.88982332e-01 -5.68170488e-01 -2.18727559e-01 -3.04107498e-02
-3.72400343e-01 -2.94651657e-01 -6.42045021e-01 -1.12728918e+00
1.35714695e-01 -1.18597722e+00 -3.92408758e-01 4.92067672e-02
-3.06018412e-01 -5.13971746e-02 6.29589677e-01 -5.42697608e-01
1.52278209e+00 -1.96338785e+00 1.99321136e-01 9.93922800e-02
3.77896786e-01 1.62735537e-01 6.58439994e-02 4.69757676e-01
-2.41643026e-01 -2.73590505e-01 -6.04219854e-01 -3.86553794e-01
-2.85687655e-01 1.55445874e-01 -1.32892281e-01 8.24707448e-01
2.82183848e-02 9.41802919e-01 -9.58347678e-01 -2.84736156e-01
2.78052866e-01 5.75504899e-01 -3.30155045e-01 -1.36961490e-01
-3.91854309e-02 8.68855774e-01 -5.64983010e-01 7.47345567e-01
3.09713811e-01 -7.31419921e-01 2.18886510e-01 7.57887075e-03
1.65524274e-01 -2.31380135e-01 -1.14588201e+00 1.92930043e+00
-2.04285949e-01 6.10637188e-01 -1.34992391e-01 -1.54603302e+00
6.84868217e-01 2.82486767e-01 9.48322117e-01 -7.74816513e-01
1.34450853e-01 9.84432325e-02 1.88608689e-03 -7.39506721e-01
1.22056259e-02 -6.35763630e-03 -9.22856554e-02 1.74433738e-01
-2.80219585e-01 2.01988295e-02 2.85614938e-01 -2.06875801e-01
1.37027085e+00 7.88906217e-02 6.98550940e-01 -2.21624017e-01
3.75160635e-01 8.12635273e-02 5.43419600e-01 6.73126817e-01
-2.38344833e-01 8.03670585e-01 5.36123633e-01 -6.03903651e-01
-7.61325061e-01 -9.21127081e-01 -3.37969601e-01 7.47334838e-01
-1.18166335e-01 -3.43952537e-01 -5.55844069e-01 -5.18318236e-01
1.22986935e-01 3.33427310e-01 -7.74142385e-01 9.11460146e-02
-6.18984938e-01 -1.07138824e+00 5.10882318e-01 4.45565164e-01
2.19231978e-01 -1.00157714e+00 -6.39448822e-01 2.36343682e-01
-3.47274959e-01 -1.32986498e+00 -5.28127372e-01 1.71017833e-02
-1.11334312e+00 -1.27469182e+00 -8.77924025e-01 -5.72317004e-01
5.90191305e-01 3.70169491e-01 1.15596378e+00 -4.06067297e-02
-7.46380031e-01 3.82960111e-01 -3.09670538e-01 -4.83694911e-01
-5.22558913e-02 -3.93489283e-03 1.30269991e-03 3.56633395e-01
3.15656930e-01 -8.69651735e-01 -9.15536046e-01 2.50171840e-01
-8.99073899e-01 -3.55704650e-02 3.75287831e-01 8.98157775e-01
9.29098427e-01 -4.12142836e-02 4.83652711e-01 -1.07058585e+00
4.87248093e-01 -5.86016417e-01 -5.82405150e-01 -6.17022179e-02
-6.68874979e-01 -1.35936916e-01 6.77582622e-01 -6.57000065e-01
-3.27092290e-01 2.43417412e-01 5.49699068e-02 -6.88087285e-01
-1.60257801e-01 6.38666987e-01 4.81154501e-01 1.27457842e-01
7.17510223e-01 3.97567093e-01 2.21319139e-01 -5.10379314e-01
1.21911809e-01 2.56315887e-01 5.35795808e-01 -4.03403759e-01
8.12016904e-01 7.81806111e-01 2.60792643e-01 -1.10504270e+00
-7.57511318e-01 -6.56897962e-01 -4.07362342e-01 -9.02728587e-02
6.82549596e-01 -1.04164672e+00 -5.62213659e-01 5.46696067e-01
-6.26919210e-01 -5.11223674e-01 -3.45603526e-01 5.47461092e-01
-6.59390628e-01 5.83140016e-01 -3.64735216e-01 -5.54667830e-01
-3.63483340e-01 -9.29019690e-01 1.22638965e+00 -1.48579041e-02
-4.05971020e-01 -1.14490759e+00 3.18400800e-01 4.97642994e-01
1.87608466e-01 8.13582122e-01 5.53897083e-01 -4.93217438e-01
-2.56237924e-01 -3.40150714e-01 1.48828849e-01 8.34389180e-02
3.69017988e-01 -2.47137204e-01 -8.53719473e-01 -4.43471402e-01
2.67253637e-01 -1.01535358e-01 8.68618488e-01 7.10767150e-01
1.37292707e+00 -9.16393772e-02 -2.82197446e-01 6.86411262e-01
1.31123698e+00 7.16084093e-02 4.59945679e-01 1.58180088e-01
7.47485876e-01 4.03362632e-01 5.10330856e-01 6.93716645e-01
3.53517234e-01 8.46223116e-01 2.50268042e-01 -1.16238348e-01
-5.19291610e-02 1.67700961e-01 2.22264364e-01 8.44626248e-01
-2.69877464e-01 1.42948832e-02 -1.02147973e+00 6.62850559e-01
-2.07658982e+00 -1.17034769e+00 -3.03163677e-01 2.19567871e+00
9.11968827e-01 2.04859041e-02 4.36880708e-01 4.63503927e-01
5.33654511e-01 3.40429991e-01 -6.22764051e-01 1.86544254e-01
4.45886999e-02 2.73248017e-01 3.10083091e-01 2.47822464e-01
-1.32779181e+00 3.87333721e-01 5.56393051e+00 7.63920605e-01
-1.31077445e+00 1.98110521e-01 6.84863925e-01 -2.83092678e-01
8.02396759e-02 -3.29052150e-01 -3.33633244e-01 4.95323807e-01
8.28169525e-01 -2.32670859e-01 4.81239200e-01 5.47692180e-01
5.02064705e-01 -4.48318496e-02 -9.35347676e-01 1.37819672e+00
5.43960650e-03 -1.14191651e+00 -3.61603290e-01 1.37758642e-01
7.45509207e-01 1.96975678e-01 2.22104743e-01 7.72803947e-02
-2.10866734e-01 -9.06026900e-01 3.69198948e-01 5.37107408e-01
7.55273461e-01 -3.13467830e-01 3.53291959e-01 4.05971706e-01
-1.20650673e+00 5.73929809e-02 5.95404245e-02 -2.62145400e-01
1.95733029e-02 1.03048658e+00 -5.38695335e-01 7.75485754e-01
7.11250842e-01 1.06579673e+00 -5.91960728e-01 1.08861566e+00
8.77579823e-02 8.58511746e-01 -4.07399356e-01 2.39630088e-01
-6.29015937e-02 -2.41192669e-01 6.87181890e-01 9.87068355e-01
4.19087172e-01 -1.13685001e-02 3.34446669e-01 3.20289671e-01
2.04889357e-01 1.34805769e-01 -7.68055975e-01 -1.61341712e-01
7.86501989e-02 1.22184801e+00 -8.82307649e-01 -1.56870216e-01
-7.06952572e-01 9.78337646e-01 2.40364835e-01 4.06452984e-01
-1.10451627e+00 -1.46561712e-01 5.86142063e-01 3.70150566e-01
4.83385265e-01 -4.04735088e-01 -2.45127290e-01 -1.49054754e+00
4.80998635e-01 -1.05705845e+00 6.90105081e-01 -4.14916039e-01
-1.25136173e+00 4.67163384e-01 -2.28010323e-02 -1.65421236e+00
-2.56208032e-01 -4.10358518e-01 -4.22217071e-01 3.04132640e-01
-1.42292023e+00 -9.62398350e-01 -6.14173055e-01 8.63642573e-01
7.16903627e-01 -2.80167069e-02 8.45903277e-01 5.12589097e-01
-7.41486847e-01 4.49566394e-01 2.47263610e-01 1.42581284e-01
6.07882857e-01 -1.13465118e+00 2.74395317e-01 9.41722512e-01
4.99425948e-01 5.76790512e-01 8.35340381e-01 -5.29314041e-01
-1.63492608e+00 -1.18387294e+00 7.10952461e-01 -5.13878405e-01
9.53302622e-01 -2.60490984e-01 -9.63283956e-01 6.00699306e-01
-3.75193991e-02 4.29175675e-01 6.58724010e-01 1.09235875e-01
-3.33281726e-01 -1.60388470e-01 -6.83922291e-01 6.45416617e-01
1.27681375e+00 -3.86564821e-01 -4.13304627e-01 6.75079346e-01
2.85746306e-01 -5.86439133e-01 -9.69826579e-01 5.40619016e-01
4.51140940e-01 -9.10544217e-01 9.88790452e-01 -4.63635713e-01
2.20587090e-01 -3.33123177e-01 6.82373121e-02 -1.07301855e+00
-2.49167204e-01 -1.04007518e+00 -5.41211784e-01 6.77589357e-01
1.40073225e-01 -8.59431148e-01 6.77334070e-01 -4.92112190e-02
7.62480497e-02 -9.89921629e-01 -1.05935025e+00 -9.05902267e-01
-2.96667933e-01 -5.78560352e-01 9.02991295e-02 1.17048097e+00
-1.20739952e-01 2.94860929e-01 -4.66348380e-01 8.01906213e-02
7.08629608e-01 2.64689773e-01 7.41799712e-01 -1.39733946e+00
-5.15134990e-01 -4.77862775e-01 -5.09535789e-01 -7.10571051e-01
-1.09082438e-01 -9.45018649e-01 -3.24813724e-01 -1.31840897e+00
7.81405345e-02 -1.03242904e-01 -3.53121042e-01 4.75427657e-01
-2.11110890e-01 4.74249154e-01 -3.90536338e-02 3.40138406e-01
-5.11132121e-01 3.83462757e-01 1.30065835e+00 -7.06049949e-02
-2.48803854e-01 2.71843702e-01 -6.44016027e-01 6.64913416e-01
9.04680789e-01 -4.10679072e-01 -6.86313808e-01 -2.37096012e-01
-6.88055996e-03 3.31159562e-01 6.34659052e-01 -1.17619312e+00
9.76744294e-02 -1.02217272e-01 2.02386692e-01 -2.89348632e-01
2.82342494e-01 -7.86664784e-01 5.21635234e-01 6.65585399e-01
-8.04842710e-02 4.53399599e-01 2.87143558e-01 8.87591600e-01
-1.63859501e-01 1.86059535e-01 6.30220890e-01 -7.20543414e-03
-4.83880907e-01 5.38330734e-01 -5.76973595e-02 3.20950389e-01
1.21904600e+00 -1.85174555e-01 -2.88564682e-01 -5.43289006e-01
-9.64538574e-01 1.45246938e-01 3.10177267e-01 4.75557417e-01
3.27145278e-01 -1.18039715e+00 -7.30510294e-01 2.43339241e-01
2.56718785e-01 -1.00078821e-01 4.08098251e-01 1.44178700e+00
-4.07747656e-01 2.97418892e-01 -1.42020802e-03 -9.60744262e-01
-1.18649840e+00 6.13050580e-01 1.59285948e-01 -4.83681887e-01
-1.02452123e+00 4.16559845e-01 2.89013028e-01 6.40900061e-02
2.27091521e-01 -4.72449183e-01 -1.43361241e-01 2.21608311e-01
4.61451501e-01 4.14812714e-01 1.53415769e-01 -4.72983718e-01
-4.47228879e-01 9.00520384e-01 9.32866707e-02 1.59749508e-01
1.49895644e+00 4.24174741e-02 8.36682841e-02 6.92358911e-01
1.23840261e+00 -5.91967739e-02 -1.10986578e+00 -3.37106019e-01
4.19203676e-02 -1.76499590e-01 -1.51343524e-01 -4.00378853e-01
-1.09624195e+00 7.20185041e-01 5.80735207e-01 5.20570636e-01
1.40923309e+00 -1.30864695e-01 6.26028895e-01 2.79930860e-01
3.28974456e-01 -8.10671568e-01 1.96844861e-01 3.36531073e-01
9.86924529e-01 -1.37690759e+00 1.91774130e-01 -5.60264051e-01
-7.20344961e-01 8.69803607e-01 8.70839953e-02 -3.21396083e-01
7.68747032e-01 2.06528783e-01 8.99828598e-02 -3.11995089e-01
-6.44225121e-01 -2.57085890e-01 3.78727734e-01 5.17341375e-01
5.31627595e-01 5.67417182e-02 -4.50965613e-01 4.56514150e-01
-2.33094290e-01 2.68956989e-01 4.28993553e-01 8.12004030e-01
1.24119751e-01 -9.82339263e-01 -2.20922217e-01 6.88720822e-01
-7.50804782e-01 1.62114963e-01 1.43960565e-01 8.38546038e-01
-2.37707302e-01 8.07320893e-01 -5.28013967e-02 -1.44609690e-01
4.35487479e-01 7.24152476e-02 4.09481108e-01 -4.96561855e-01
-4.90999132e-01 2.91569918e-01 7.64170811e-02 -8.27202320e-01
-6.24167442e-01 -1.00677860e+00 -1.11978805e+00 -8.02465454e-02
2.65659969e-02 -1.49335817e-01 2.87063539e-01 7.14050591e-01
4.98699814e-01 7.01576173e-01 5.76548636e-01 -6.57714128e-01
-5.79790175e-01 -6.69555962e-01 -4.81600732e-01 6.46069288e-01
7.04857230e-01 -7.07079053e-01 -2.81772792e-01 3.02932322e-01] | [7.315160274505615, 3.2367212772369385] |
9f928017-5292-4716-a697-42b37f13bb75 | beyond-tree-structure-models-a-new-occlusion | null | null | http://openaccess.thecvf.com/content_iccv_2015/html/Fu_Beyond_Tree_Structure_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Fu_Beyond_Tree_Structure_ICCV_2015_paper.pdf | Beyond Tree Structure Models: A New Occlusion Aware Graphical Model for Human Pose Estimation | Occlusion is a main challenge for human pose estimation, which is largely ignored in popular tree structure models. The tree structure model is simple and convenient for exact inference, but short in modeling the occlusion coherence especially in the case of self-occlusion. We propose an occlusion aware graphical model which is able to model both self-occlusion and occlusion by the other objects simultaneously. The proposed model structure can encode the interactions between human body parts and objects, and hence enable it to learn occlusion coherence from data discriminatively. We evaluate our model on several public benchmarks for human pose estimation including challenging subsets featuring significant occlusion. The experimental results show that our method obtains comparable accuracy with the state-of-the-arts, and achieves promising performance in 2D human pose estimation with occlusion. | ['Junge Zhang', 'Lianrui Fu', 'Kaiqi Huang'] | 2015-12-01 | null | null | null | iccv-2015-12 | ['2d-human-pose-estimation'] | ['computer-vision'] | [-3.44063997e-01 1.01713493e-01 -3.63087803e-01 -4.39756870e-01
-4.41215962e-01 -1.26241416e-01 3.65473986e-01 -2.00258642e-01
-9.71062630e-02 5.88783026e-01 7.00117469e-01 4.80136305e-01
2.00664014e-01 -3.78367007e-01 -6.27207279e-01 -4.63871717e-01
-1.63751677e-01 1.06631899e+00 1.89379513e-01 -9.29810703e-02
-5.64703941e-01 2.29231477e-01 -1.53961885e+00 2.25842949e-02
6.45528018e-01 1.00229156e+00 -1.19291462e-01 3.37782741e-01
3.68788362e-01 5.77288330e-01 -6.08463109e-01 -6.11013532e-01
2.45018676e-01 8.29081535e-02 -5.54188907e-01 2.14804202e-01
9.82661009e-01 -4.37444121e-01 -9.37261760e-01 4.21507150e-01
8.55161786e-01 -1.65089052e-02 7.18985796e-01 -1.12618446e+00
2.15391926e-02 1.53516188e-01 -8.82337689e-01 -5.34130819e-03
8.84322822e-01 4.70372848e-02 1.00060999e+00 -8.95919025e-01
9.30865467e-01 1.92958331e+00 8.10310066e-01 2.92386264e-01
-1.29296052e+00 -5.54586828e-01 5.46571732e-01 1.11817539e-01
-1.45556271e+00 -2.87724048e-01 7.12110639e-01 -5.50845623e-01
6.10463619e-01 3.76480013e-01 1.19521785e+00 1.31135416e+00
4.01012093e-01 1.46288431e+00 8.87064576e-01 -1.43722042e-01
-2.30238289e-01 -4.04614478e-01 3.74024063e-01 8.62633884e-01
6.27731740e-01 1.81626573e-01 -9.10730958e-01 -3.04542840e-01
8.02934885e-01 -7.03423619e-02 -6.29920065e-02 -9.73932266e-01
-1.02924550e+00 5.63734591e-01 6.75462663e-01 -2.94864267e-01
-9.03801620e-02 4.52489555e-01 6.15780413e-01 -1.95696607e-01
5.65479279e-01 -1.00961201e-01 -4.87449259e-01 2.25721940e-01
-6.72408283e-01 1.02459896e+00 7.67120183e-01 1.28296435e+00
2.71778941e-01 -1.91888809e-01 -8.27575862e-01 7.08257079e-01
6.12584114e-01 6.74581051e-01 -5.94516397e-02 -6.55657589e-01
6.94462657e-01 4.76010621e-01 1.52273312e-01 -1.28230882e+00
-8.69363725e-01 -8.28299940e-01 -8.09561670e-01 -2.11189836e-01
5.28036654e-01 1.12230413e-01 -9.34778154e-01 1.81433105e+00
8.92073572e-01 -9.81441513e-02 -7.49779701e-01 1.02847636e+00
1.50248504e+00 1.13749228e-01 4.19758350e-01 -6.31366447e-02
1.94225991e+00 -1.25341594e+00 -1.09879541e+00 -6.71176910e-01
3.93490106e-01 -7.90453434e-01 5.84232748e-01 2.52570450e-01
-8.66406024e-01 -9.16421831e-01 -8.36254954e-01 -4.74668741e-01
1.89759135e-01 3.78326923e-01 1.10880876e+00 6.20849192e-01
-2.73579270e-01 3.65723729e-01 -9.55752254e-01 -1.75297394e-01
5.54693878e-01 3.88065696e-01 -5.62873125e-01 -2.27287635e-01
-1.08968651e+00 8.96511734e-01 2.26478264e-01 4.10856068e-01
-7.04275906e-01 -4.67559457e-01 -1.28418744e+00 -3.12658191e-01
4.74201679e-01 -1.40191102e+00 9.91266727e-01 6.68917149e-02
-1.29005826e+00 1.00922692e+00 -4.12852466e-01 -2.28904426e-01
1.08339584e+00 -1.11838627e+00 2.62224805e-02 -1.90336153e-01
1.10318668e-01 8.85317683e-01 7.84407020e-01 -1.11743414e+00
-2.56083667e-01 -7.69780338e-01 -1.20666973e-01 5.15333116e-01
1.79374721e-02 -1.20925851e-01 -1.22701728e+00 -8.69772255e-01
5.48793674e-01 -1.33617938e+00 -3.60775411e-01 2.60434449e-01
-7.02676058e-01 -3.74446392e-01 8.42214882e-01 -8.78532350e-01
1.37547266e+00 -1.56304622e+00 4.19476330e-01 8.07937086e-02
5.56564450e-01 9.04376153e-03 1.54114604e-01 1.68882236e-01
2.34306902e-01 -5.82597911e-01 2.14853674e-01 -9.71659780e-01
3.53960544e-01 4.68130887e-01 -1.25181139e-03 8.42827201e-01
-1.49323389e-01 1.41894388e+00 -7.13415205e-01 -1.10759509e+00
4.66887653e-01 6.16443753e-01 -7.51891971e-01 3.97318810e-01
-9.33341235e-02 8.56115818e-01 -6.10223293e-01 8.95879447e-01
7.73792267e-01 -2.58053303e-01 2.39892006e-01 -6.02701783e-01
3.70093822e-01 3.12846661e-01 -1.30547559e+00 1.97770524e+00
1.35254696e-01 3.71725112e-01 -3.44316363e-01 -2.66567260e-01
7.16542244e-01 3.31388474e-01 5.79471409e-01 -3.49561423e-01
4.63557571e-01 -2.76167870e-01 -2.78408408e-01 -4.26909566e-01
2.06910983e-01 4.16323990e-01 -2.60315776e-01 -1.34507120e-01
-3.48306037e-02 -6.86396286e-02 1.78267762e-01 2.76275910e-03
7.64971733e-01 6.72329187e-01 4.52216864e-01 -2.32626781e-01
1.89643592e-01 -3.81498754e-01 8.24012041e-01 6.60178483e-01
-4.80310589e-01 6.80169106e-01 1.82857469e-01 -9.07372475e-01
-6.90265656e-01 -1.26255035e+00 -4.18214858e-01 1.06222141e+00
3.76898885e-01 -9.46556449e-01 -6.17379069e-01 -5.95910251e-01
3.10879767e-01 -1.36544585e-01 -8.54283869e-01 5.24656661e-02
-8.75962377e-01 -7.47380018e-01 3.28968406e-01 7.48057783e-01
5.45407295e-01 -7.85483778e-01 -5.02961993e-01 1.32281125e-01
-9.11267400e-01 -1.43946493e+00 -5.51629543e-01 -1.47111654e-01
-9.55153525e-01 -1.06827199e+00 -7.36237764e-01 -6.39835894e-01
4.91014093e-01 4.13199961e-02 1.52486086e+00 -1.08675480e-01
-7.16046870e-01 2.76562661e-01 7.09048361e-02 -4.09082055e-01
4.20374960e-01 1.68878719e-01 1.75025180e-01 -2.99194634e-01
2.43870229e-01 -3.48120302e-01 -9.60695624e-01 6.88991606e-01
-1.97036445e-01 1.29776031e-01 3.52142543e-01 7.32733130e-01
7.69855320e-01 -3.05079073e-01 -1.70713246e-01 -8.47456396e-01
-1.10787079e-01 5.70328487e-03 -2.16257006e-01 1.28057569e-01
-2.91917864e-02 1.60725378e-02 -3.06262404e-01 -5.12467980e-01
-1.04222834e+00 6.02873862e-01 -1.97441354e-01 -3.40412378e-01
-1.07699931e-01 1.38833299e-01 -5.62260866e-01 -2.59919286e-01
3.91290486e-01 -4.25510764e-01 -2.58000582e-01 -8.95007253e-01
2.26503536e-01 -1.14304870e-01 7.44357228e-01 -9.54150796e-01
7.90909231e-01 6.79665923e-01 4.02862817e-01 -6.44246638e-01
-1.27017117e+00 -7.22253323e-01 -1.04195595e+00 -3.50744665e-01
7.88130820e-01 -1.42815650e+00 -7.61379421e-01 4.60316509e-01
-1.32564342e+00 1.75502092e-01 -1.03431553e-01 4.71842766e-01
-6.10128582e-01 5.48738956e-01 -8.17561150e-01 -6.75309360e-01
-2.07072392e-01 -1.11866903e+00 1.72626495e+00 -5.52373379e-02
-7.09628582e-01 -7.70386159e-01 2.98477501e-01 5.82085669e-01
-3.17697525e-01 7.43883669e-01 3.24920118e-01 8.97692442e-02
-5.65406263e-01 -5.82198679e-01 -2.48204842e-02 -3.17557812e-01
-1.17607661e-01 -3.50538135e-01 -9.86935496e-01 -4.96447504e-01
-2.85767853e-01 -5.69069028e-01 9.00969684e-01 8.20582032e-01
1.02967834e+00 -4.44073416e-02 -7.76380956e-01 7.06750512e-01
7.46345103e-01 -8.71626735e-01 9.03621435e-01 -3.98929454e-02
1.16808629e+00 7.23996222e-01 9.60976064e-01 6.65018380e-01
6.78757787e-01 1.37913656e+00 3.31653118e-01 -3.09686869e-01
-4.43934977e-01 -6.59181058e-01 2.11950578e-02 7.72400796e-01
-4.90631998e-01 -2.76891887e-02 -5.65152764e-01 1.91186696e-01
-2.16569972e+00 -7.35879838e-01 -5.14410496e-01 1.99534535e+00
8.00847709e-01 1.39249966e-01 3.91099632e-01 -2.22340580e-02
5.65361023e-01 4.41440612e-01 -2.66618460e-01 3.56270045e-01
-9.37993973e-02 1.52744606e-01 2.44079739e-01 4.45572495e-01
-1.80882215e+00 1.07639372e+00 7.21570826e+00 8.73694122e-01
-2.31898740e-01 1.04498573e-01 2.63480902e-01 -2.16556698e-01
2.62654841e-01 -2.54244208e-01 -1.29571056e+00 1.88960895e-01
-7.00731799e-02 2.86219090e-01 -4.20368016e-01 9.77486908e-01
-6.49598241e-02 -8.40904266e-02 -1.30747068e+00 1.22237134e+00
6.28087297e-02 -5.14236093e-01 -4.32566777e-02 2.25668281e-01
8.36828589e-01 -5.10108352e-01 9.81981307e-02 2.63933629e-01
9.70162675e-02 -9.39512312e-01 8.91828954e-01 4.48432386e-01
4.54207778e-01 -4.65169609e-01 6.85233831e-01 2.09491655e-01
-1.82253373e+00 1.98635072e-01 -4.08708066e-01 -2.78869569e-01
4.81484383e-01 7.00576603e-01 -2.09802240e-01 4.43255216e-01
7.90567100e-01 7.94106126e-01 -7.80811548e-01 1.24988317e+00
-5.35725057e-01 1.15513682e-01 -5.35158515e-01 3.44169378e-01
-2.84912258e-01 1.30955875e-01 5.85525930e-01 1.14683747e+00
-2.66945362e-01 -4.07198146e-02 7.93500721e-01 2.83068150e-01
6.87606335e-02 1.74171865e-01 -3.59095275e-01 6.64232135e-01
2.14802623e-01 9.79584217e-01 -5.46659350e-01 -3.51101786e-01
-1.28316522e-01 9.26158369e-01 6.04588568e-01 2.06388697e-01
-1.14741492e+00 2.64406204e-01 7.10926592e-01 3.02542418e-01
2.96072274e-01 -4.30755526e-01 -3.00787628e-01 -1.21712625e+00
3.75335783e-01 -9.39808011e-01 5.97139239e-01 -4.01477456e-01
-1.22889173e+00 3.69171172e-01 2.82287240e-01 -1.27926421e+00
-2.03471720e-01 -5.65984309e-01 5.48402593e-02 4.84741420e-01
-8.53891373e-01 -1.66994965e+00 -5.79658866e-01 4.87543941e-01
4.23155457e-01 4.02593791e-01 7.94539809e-01 5.78034997e-01
-5.78797698e-01 8.62846255e-01 -5.34036815e-01 1.75366148e-01
9.27127540e-01 -1.02827537e+00 6.01332903e-01 4.34687942e-01
3.55242968e-01 7.78539300e-01 1.14130557e+00 -1.10313165e+00
-1.23365402e+00 -9.34832156e-01 1.16275728e+00 -9.86732543e-01
2.11096220e-02 -7.24551558e-01 -4.24696058e-01 7.79101849e-01
-3.55621040e-01 3.08229566e-01 6.13227308e-01 7.36972809e-01
-6.02485597e-01 2.50332087e-01 -7.71113157e-01 6.27821982e-01
1.75473356e+00 -2.20483378e-01 -7.54345477e-01 5.21986544e-01
6.02397621e-01 -1.13053799e+00 -8.71550322e-01 9.79590535e-01
1.26998687e+00 -7.53647506e-01 1.68802607e+00 -6.88264847e-01
2.94766039e-01 -2.04677314e-01 -1.57196432e-01 -6.63745880e-01
-6.92784250e-01 -4.55263436e-01 -8.84769380e-01 8.18674326e-01
-1.82259247e-01 -5.44655286e-02 1.18302643e+00 7.43741930e-01
2.24689797e-01 -7.10266292e-01 -9.00564194e-01 -8.89460802e-01
-3.38318229e-01 -2.38558695e-01 2.73778647e-01 3.46550584e-01
-3.75588238e-01 3.98125738e-01 -1.05547190e+00 1.37695134e-01
1.14757669e+00 2.13925928e-01 1.45543349e+00 -1.48203731e+00
-5.16292870e-01 -2.89915819e-02 -6.93020642e-01 -1.69807482e+00
2.33060256e-01 -3.99944693e-01 7.34572411e-02 -1.38709521e+00
5.45012057e-01 -2.37394363e-01 1.54440120e-01 2.75949121e-01
-6.30631089e-01 6.05893731e-01 1.82003975e-01 3.32731724e-01
-9.99459028e-01 8.84936333e-01 1.65037251e+00 -2.38400817e-01
-8.14547017e-02 2.92321593e-01 -2.83531547e-01 1.08258343e+00
6.70950487e-02 -5.95888078e-01 -1.30084157e-01 -4.32990283e-01
-3.83649282e-02 -2.12403670e-01 6.89742565e-01 -1.24417555e+00
6.59204833e-03 1.69710219e-01 9.35182452e-01 -1.23143840e+00
7.62917578e-01 -7.11802423e-01 4.69109356e-01 7.31211483e-01
-1.44671395e-01 -1.69007450e-01 5.47412001e-02 7.99507856e-01
-5.09940535e-02 4.09369677e-01 5.15555322e-01 1.39481795e-03
-4.62732583e-01 9.08485472e-01 2.84767419e-01 3.69700879e-01
5.71565032e-01 -2.46594497e-03 1.18496411e-01 -3.82650793e-01
-9.92298126e-01 4.15828079e-01 1.84976220e-01 6.21267200e-01
3.27835172e-01 -1.82726777e+00 -7.47105539e-01 -7.54598305e-02
4.42255467e-01 -3.47305350e-02 3.06799650e-01 7.27396250e-01
-4.14357245e-01 6.16182745e-01 -1.38059795e-01 -1.02194607e+00
-1.62189996e+00 5.25378764e-01 4.88226078e-02 -7.66762197e-01
-8.47285688e-01 9.28156912e-01 7.89092302e-01 -3.71116638e-01
6.96650624e-01 -1.41659886e-01 -6.14739396e-02 -8.06809738e-02
5.04826427e-01 5.51810205e-01 -1.92529485e-01 -9.95728016e-01
-5.45453072e-01 9.79276896e-01 1.07227020e-01 2.07579091e-01
8.15179944e-01 -1.91648126e-01 6.34602457e-02 3.06455344e-01
1.19689739e+00 -4.32300381e-02 -1.25969625e+00 -6.22620404e-01
-3.57318014e-01 -8.70051920e-01 -2.97313124e-01 -3.94415706e-01
-9.36346233e-01 9.38292265e-01 5.95486224e-01 -5.46683669e-01
5.70371151e-01 2.73040414e-01 8.04865777e-01 3.68654013e-01
8.21955800e-01 -1.08306050e+00 2.75244176e-01 5.16670167e-01
1.22225952e+00 -1.30158651e+00 7.83377647e-01 -1.30778277e+00
-3.44554245e-01 5.44983923e-01 8.64431262e-01 -1.82681426e-01
7.79602826e-01 1.20936908e-01 -1.76624075e-01 -4.68239367e-01
-4.53550965e-01 -4.60950285e-01 1.10293925e+00 6.00694597e-01
5.70009589e-01 3.34132791e-01 -5.04465044e-01 6.61184371e-01
-3.71212959e-01 -1.89204946e-01 -7.81356394e-01 8.53340805e-01
-1.87705576e-01 -1.23693442e+00 -6.36534095e-01 2.46584684e-01
-2.82458872e-01 3.50454330e-01 -4.53579009e-01 9.44905818e-01
2.77869135e-01 5.93544126e-01 -2.60832429e-01 -2.03300208e-01
5.24266958e-01 -1.14571735e-01 1.07437384e+00 -4.92858946e-01
-3.47907960e-01 4.14343119e-01 3.27995598e-01 -9.51775551e-01
-3.60050797e-01 -8.68728459e-01 -6.10131860e-01 -3.53661954e-01
-5.05331814e-01 -3.58163714e-01 1.88513875e-01 9.11325574e-01
9.64930281e-02 6.18672550e-01 1.19873464e-01 -1.29188740e+00
-2.36454770e-01 -9.14600372e-01 -4.66085911e-01 7.97544301e-01
-8.20887685e-02 -1.56009221e+00 3.44118178e-02 -4.07066606e-02] | [7.003200531005859, -0.8866998553276062] |
438ff477-339a-46ba-ba13-19f3b6c8ff57 | global-optimization-networks | 2202.01277 | null | https://arxiv.org/abs/2202.01277v1 | https://arxiv.org/pdf/2202.01277v1.pdf | Global Optimization Networks | We consider the problem of estimating a good maximizer of a black-box function given noisy examples. To solve such problems, we propose to fit a new type of function which we call a global optimization network (GON), defined as any composition of an invertible function and a unimodal function, whose unique global maximizer can be inferred in $\mathcal{O}(D)$ time. In this paper, we show how to construct invertible and unimodal functions by using linear inequality constraints on lattice models. We also extend to \emph{conditional} GONs that find a global maximizer conditioned on specified inputs of other dimensions. Experiments show the GON maximizers are statistically significantly better predictions than those produced by convex fits, GPR, or DNNs, and are more reasonable predictions for real-world problems. | ['Maya Gupta', 'Olexander Mangylov', 'Erez Louidor', 'Sen Zhao'] | 2022-02-02 | null | null | null | null | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [-7.47086108e-02 6.52763486e-01 -7.69676864e-02 -8.55087459e-01
-1.20340323e+00 -5.23755908e-01 -3.02833226e-02 -4.22795504e-01
-1.09972715e-01 1.21333289e+00 1.18602589e-01 -2.64483869e-01
-4.53425944e-01 -8.15652311e-01 -1.19271922e+00 -6.51086807e-01
-1.71000898e-01 1.11099088e+00 -3.51983070e-01 1.13793798e-01
1.56341434e-01 2.45273918e-01 -1.02241254e+00 1.82060063e-01
9.39173222e-01 1.14880300e+00 2.78866082e-01 5.23960829e-01
4.67305034e-02 4.82296556e-01 -4.36611593e-01 -5.11447251e-01
5.80743611e-01 -3.37993294e-01 -5.85422873e-01 -5.56838699e-02
6.14647090e-01 -4.83552694e-01 -1.19277447e-01 1.31072402e+00
3.38245690e-01 2.74520695e-01 8.88299167e-01 -1.11766291e+00
-8.70697558e-01 9.11644638e-01 -2.42942035e-01 -1.70380563e-01
1.19350804e-02 5.89490160e-02 1.56267428e+00 -7.58337259e-01
4.81206238e-01 1.57340801e+00 6.25916660e-01 3.64129961e-01
-1.68178415e+00 -6.30173445e-01 2.23684996e-01 -3.08059037e-01
-1.37245929e+00 -3.24951679e-01 3.64747137e-01 -3.00210297e-01
1.12168491e+00 2.91263610e-01 1.32551372e-01 8.61740828e-01
2.17870086e-01 8.71150374e-01 8.76459420e-01 -1.74704313e-01
3.65675509e-01 2.74843164e-02 -6.23792782e-02 8.57358992e-01
1.61317274e-01 2.59382930e-02 -3.44969779e-01 -3.68262798e-01
1.00743401e+00 -1.85733661e-01 -3.42826009e-01 -7.59665146e-02
-8.87102962e-01 8.69598389e-01 4.24176216e-01 -7.34259784e-02
-3.13869029e-01 4.43794996e-01 -2.98805892e-01 3.79058897e-01
6.38136983e-01 1.13174367e+00 -7.12731898e-01 -1.24768496e-01
-9.28160250e-01 4.65997308e-01 1.24897468e+00 1.16190803e+00
8.37074578e-01 9.89068523e-02 -1.28730178e-01 8.67579281e-01
2.92229474e-01 6.83351099e-01 -9.98201966e-02 -1.19655287e+00
8.44894588e-01 4.67366964e-01 4.48946238e-01 -9.33131933e-01
-8.24023724e-01 -5.18351257e-01 -7.49786019e-01 -5.38240001e-02
7.46236324e-01 -4.75318581e-01 -9.29398119e-01 1.93158507e+00
-2.32070107e-02 8.25863332e-03 -1.72552198e-01 9.28145468e-01
3.65060866e-01 1.00090075e+00 -3.18789035e-01 -2.13472888e-01
6.32516801e-01 -7.73138762e-01 -4.60823506e-01 -5.38272560e-01
4.94037122e-01 -3.76762956e-01 8.29350948e-01 7.06194699e-01
-1.06271815e+00 -2.78545290e-01 -9.51419890e-01 -1.63890123e-01
6.61416072e-03 2.11574778e-01 6.98782027e-01 6.16078734e-01
-1.19144189e+00 8.86487365e-01 -7.97571659e-01 2.07402214e-01
2.15085730e-01 8.23165536e-01 -1.28702790e-01 -5.02153486e-02
-8.95331144e-01 8.02766025e-01 4.73307699e-01 1.72821686e-01
-1.18412590e+00 -5.94534695e-01 -7.54802823e-01 2.38243103e-01
6.80894613e-01 -5.13575733e-01 1.13243282e+00 -7.08007812e-01
-1.36019015e+00 5.69141805e-01 -2.54987329e-01 -6.25426114e-01
2.54721403e-01 -8.17160979e-02 -1.65128782e-01 -3.34975958e-01
-5.23428172e-02 6.85254872e-01 8.64259899e-01 -1.03335106e+00
-4.80012923e-01 -4.92278874e-01 3.19743864e-02 6.80911839e-02
-2.65174769e-02 -3.12765479e-01 -2.84657031e-01 -3.51895064e-01
3.80217433e-01 -8.24629247e-01 -6.10087872e-01 -1.55675441e-01
-1.00718427e+00 -3.49651605e-01 5.15353195e-02 -9.08702552e-01
1.48095334e+00 -1.80171108e+00 3.83031696e-01 7.83322096e-01
4.92234319e-01 -4.26110655e-01 -2.20924497e-01 6.37778565e-02
1.40297294e-01 4.26316828e-01 -4.24809009e-01 -4.84837145e-01
2.77049869e-01 3.78969908e-01 -3.59989405e-01 6.49479985e-01
1.24041259e-01 9.10079777e-01 -5.49444616e-01 -6.17211014e-02
-1.55446067e-01 4.35559824e-02 -8.18692029e-01 1.18174508e-01
-7.92755663e-01 2.06779391e-01 -6.06011212e-01 7.65083313e-01
7.60580897e-01 -4.96052027e-01 4.52321351e-01 2.24955499e-01
1.46748856e-01 4.62140888e-01 -1.48411214e+00 1.17699301e+00
-4.05301094e-01 3.48273873e-01 2.06045777e-01 -1.36507297e+00
1.51865304e+00 2.73853298e-02 4.51599121e-01 -5.23264527e-01
1.77947551e-01 4.44089234e-01 -1.58845901e-01 -2.60380715e-01
3.53210121e-01 -2.24049136e-01 -2.01784298e-01 4.82291251e-01
3.49721789e-01 -1.09472044e-01 3.07254735e-02 -2.55229622e-01
1.03685975e+00 -4.87445630e-02 2.17947811e-01 -5.37122846e-01
-1.28546700e-01 -3.20522547e-01 7.51749873e-01 1.19449651e+00
2.18052894e-01 8.22148621e-01 1.00654256e+00 -4.50116247e-01
-1.29957402e+00 -1.13004005e+00 -2.61404127e-01 1.07165468e+00
-1.39218897e-01 -3.87304246e-01 -7.92785108e-01 -4.52179015e-01
3.32014233e-01 8.36200774e-01 -4.15320605e-01 1.67027459e-01
-5.81032336e-01 -1.06141436e+00 3.21823180e-01 5.39948404e-01
2.96468675e-01 -9.20885205e-01 1.56957656e-01 3.56410861e-01
-1.80945456e-01 -8.12900782e-01 -6.31256163e-01 6.31335258e-01
-1.02331281e+00 -6.52602911e-01 -6.99086905e-01 -6.94012582e-01
8.11580539e-01 -4.91662085e-01 1.39268351e+00 -2.49317035e-01
2.51845960e-02 5.82111441e-02 2.70307988e-01 -1.64360672e-01
-1.86161205e-01 2.26034641e-01 3.27289939e-01 6.22569211e-03
2.53213644e-01 -7.47085571e-01 -2.73575336e-01 5.26666522e-01
-4.66301113e-01 -2.16081113e-01 4.86989230e-01 8.25663924e-01
1.04087603e+00 1.07277103e-01 6.93489194e-01 -9.78919685e-01
8.39389682e-01 -7.19693482e-01 -1.17765331e+00 5.38006723e-01
-6.49657488e-01 8.29909146e-01 9.75931287e-01 -2.94565856e-01
-8.65728557e-01 2.84582287e-01 -4.78331074e-02 -3.73514742e-01
1.87148467e-01 5.86689293e-01 -3.45730126e-01 9.58445817e-02
9.80756104e-01 -3.94001193e-02 -1.53824002e-01 -7.73279011e-01
3.94132167e-01 6.17321312e-01 4.84531492e-01 -1.14309275e+00
4.95952338e-01 1.14978455e-01 7.75101185e-02 -7.37497985e-01
-1.00990820e+00 -2.23897353e-01 -2.84620196e-01 1.25541940e-01
8.08392823e-01 -6.53640628e-01 -1.11199844e+00 -1.01026788e-03
-1.13804805e+00 -4.48571771e-01 -2.40621686e-01 3.47998708e-01
-9.07034159e-01 -7.26005360e-02 -2.45662615e-01 -1.22482407e+00
-1.60634071e-01 -1.03852034e+00 9.53998268e-01 1.82214424e-01
-2.52217382e-01 -1.07047832e+00 1.18457563e-02 2.71232039e-01
2.64598519e-01 1.10472418e-01 9.74177897e-01 -7.64256775e-01
-1.08278859e+00 -1.26875952e-01 -2.46517658e-01 7.17662811e-01
-3.05663407e-01 -1.82141632e-01 -7.81967878e-01 -1.06042199e-01
1.33585602e-01 -3.51319194e-01 9.83666599e-01 9.53084767e-01
1.37448573e+00 -8.17305386e-01 -3.38640094e-01 1.05602896e+00
1.37018871e+00 3.03233981e-01 5.88686407e-01 -6.19672425e-02
4.68443274e-01 3.41958135e-01 3.01504701e-01 6.25612259e-01
2.23823622e-01 3.19628894e-01 5.30794144e-01 2.94014633e-01
5.30323863e-01 -6.54617310e-01 3.11461300e-01 3.65961552e-01
-1.37586072e-01 -5.49852848e-01 -8.46627831e-01 3.24891895e-01
-2.14849567e+00 -8.37802768e-01 2.35589474e-01 2.30197477e+00
8.91583562e-01 1.52888969e-01 -3.25011685e-02 -6.46209002e-01
7.18618214e-01 -1.30117284e-02 -9.58316505e-01 -2.60044038e-01
-3.36835951e-01 4.87933069e-01 8.10008049e-01 9.49569702e-01
-1.00919032e+00 7.37171590e-01 7.75370836e+00 1.04164803e+00
-4.88687545e-01 -2.31872693e-01 1.22980237e+00 -2.12433189e-01
-6.63085639e-01 1.04224391e-03 -1.29845643e+00 3.72938097e-01
9.03108358e-01 8.34693015e-02 1.11603403e+00 1.10833693e+00
2.02337787e-01 7.56585831e-03 -1.34129357e+00 9.87936616e-01
-2.03475788e-01 -1.35223234e+00 -4.71040010e-01 1.96603298e-01
1.12003231e+00 -1.06120091e-02 3.47242117e-01 4.94456381e-01
8.31596851e-01 -1.68104553e+00 2.68403471e-01 9.00520802e-01
6.36413872e-01 -7.58687556e-01 3.78996521e-01 7.82898664e-01
-7.58925676e-01 -1.67803243e-01 -9.02641952e-01 -8.97426978e-02
1.21138722e-01 7.60897696e-01 -1.03215706e+00 -1.96833555e-02
4.33118165e-01 5.08811295e-01 -1.09766789e-01 1.08493841e+00
-2.09655121e-01 7.25934207e-01 -1.12286985e+00 -3.59388649e-01
2.53430694e-01 -5.53823650e-01 4.58251506e-01 8.98844421e-01
5.99300683e-01 7.48235136e-02 1.96833029e-01 1.60676312e+00
-4.18941081e-01 8.18824843e-02 -5.83647728e-01 -1.14799060e-01
3.60806108e-01 1.02692568e+00 -4.06447768e-01 1.93243120e-02
-1.90724374e-03 5.66779852e-01 7.00049639e-01 4.85771000e-01
-6.80490851e-01 -5.34117818e-02 7.07828224e-01 2.80069821e-02
2.95909792e-01 -1.59105346e-01 -7.19674349e-01 -1.19736362e+00
2.50650525e-01 -7.65049815e-01 3.02227855e-01 -6.32061124e-01
-1.55561256e+00 3.12266350e-01 -1.25989065e-01 -9.10262346e-01
-4.12145376e-01 -9.14436638e-01 -3.68076801e-01 9.28682983e-01
-9.12130713e-01 -5.38343608e-01 4.10299450e-01 4.05916780e-01
2.83434391e-02 -2.41608635e-01 7.20164478e-01 -8.26661885e-02
-3.22473228e-01 6.61466181e-01 5.37266254e-01 -6.20101094e-02
2.34510899e-01 -1.64293885e+00 4.08709884e-01 5.20004094e-01
3.49936485e-01 7.61876047e-01 7.75817513e-01 -6.72031462e-01
-1.18341970e+00 -8.54951620e-01 1.01920843e+00 -4.96147245e-01
6.07634008e-01 -3.86025667e-01 -7.07975090e-01 1.02266455e+00
-2.40446374e-01 -5.48218302e-02 3.57008755e-01 6.09888494e-01
-2.18169287e-01 -1.15592495e-01 -1.39726949e+00 4.16339278e-01
1.09982216e+00 -3.66877407e-01 -2.08810464e-01 8.00263166e-01
7.84069180e-01 -6.12562418e-01 -1.02127802e+00 2.42730156e-01
3.58475715e-01 -7.84830213e-01 1.02294278e+00 -1.02722907e+00
4.59756315e-01 5.18230647e-02 -5.64843774e-01 -1.38980341e+00
-1.88611522e-01 -9.87748682e-01 -1.95738226e-01 7.51684964e-01
1.06843424e+00 -6.99075997e-01 8.57037902e-01 1.05327857e+00
-9.80430618e-02 -1.05101347e+00 -1.26731253e+00 -8.42974186e-01
2.36992210e-01 -6.95627153e-01 5.39095640e-01 5.57521224e-01
1.25577033e-01 2.57265985e-01 -8.86857212e-01 2.14711845e-01
6.02990746e-01 1.10832527e-01 5.63091159e-01 -1.06694984e+00
-8.47311437e-01 -3.34907115e-01 -1.71633020e-01 -1.88080215e+00
1.75275460e-01 -1.12894177e+00 3.81231993e-01 -1.25866354e+00
6.75260499e-02 -7.48469174e-01 -1.52888909e-01 4.87416416e-01
3.00479457e-02 -2.78266162e-01 6.32132813e-02 -4.98506725e-02
-6.85288787e-01 4.92062122e-01 1.36737406e+00 2.18474902e-02
-4.24188733e-01 5.90962827e-01 -1.00092268e+00 6.93582535e-01
7.02313244e-01 -6.88983977e-01 -1.99461117e-01 -4.01170939e-01
7.18610764e-01 6.21115625e-01 1.53980762e-01 -6.59754813e-01
2.94482172e-01 -3.34287167e-01 3.54480684e-01 -4.90350962e-01
6.14729106e-01 -6.95149720e-01 -8.10317248e-02 -1.01092853e-01
-5.29779732e-01 -1.38904434e-02 -1.70692936e-01 5.12095809e-01
-4.91312109e-02 -5.46770811e-01 6.45020068e-01 -1.22357771e-01
-2.09589809e-01 4.61258113e-01 1.09601421e-02 4.70647693e-01
6.88196361e-01 2.49469548e-01 -4.07877177e-01 -8.76416504e-01
-1.10336077e+00 5.07512033e-01 -1.92916617e-01 -5.28411977e-02
6.98194802e-01 -1.16678607e+00 -6.96236014e-01 1.17572509e-01
-4.28482473e-01 2.46289536e-01 -1.88875981e-02 5.19673645e-01
-4.63273853e-01 6.33795857e-01 3.67228895e-01 -3.82291496e-01
-4.56337810e-01 2.51709729e-01 7.50052154e-01 -5.32901466e-01
-3.26541841e-01 1.23023880e+00 9.10302103e-02 -9.73728061e-01
4.12160397e-01 -7.10325062e-01 3.23560238e-01 -2.07248479e-01
2.48188898e-01 4.48995948e-01 -2.73439378e-01 -1.10686891e-01
-4.02475260e-02 2.04970539e-01 2.68636823e-01 -1.01919286e-01
1.58383560e+00 2.22448006e-01 -2.50357658e-01 2.71204352e-01
1.34811735e+00 -3.03003490e-01 -1.56520283e+00 -3.48024160e-01
-1.13192778e-02 -3.67429048e-01 -1.62683815e-01 -8.56272876e-01
-1.12532902e+00 6.43729985e-01 2.05985010e-01 4.53040630e-01
7.42154419e-01 2.37988025e-01 4.68389899e-01 9.76224244e-01
3.74135315e-01 -1.16627932e+00 -2.29418933e-01 6.62568152e-01
1.05498862e+00 -1.25824142e+00 -1.82835504e-01 -1.04586221e-01
-6.40756905e-01 1.34951341e+00 4.95053262e-01 -5.18227041e-01
1.01151657e+00 5.49525738e-01 -4.00732398e-01 -1.15735315e-01
-8.15783978e-01 1.86598394e-02 5.41712284e-01 5.36580265e-01
1.36842299e-02 4.02206481e-01 1.53975546e-01 1.05502141e+00
-6.15648568e-01 -9.87037718e-02 7.03246817e-02 2.62787282e-01
-9.02621508e-01 -8.02257836e-01 -3.49100262e-01 9.05316174e-01
-4.25233960e-01 -3.55436713e-01 -2.64530063e-01 6.23374820e-01
3.15933339e-02 7.74424255e-01 3.05977240e-02 -5.01786113e-01
7.14142323e-02 1.57841504e-01 3.77849489e-01 -6.54466987e-01
-1.88373506e-01 4.07666415e-02 2.40626961e-01 -5.09453356e-01
2.49715328e-01 -6.12730384e-01 -1.13656342e+00 -3.92966717e-01
-2.78099328e-01 -8.07176530e-02 5.17142773e-01 1.15104258e+00
5.98020777e-02 1.30029544e-01 4.75856721e-01 -5.87963283e-01
-1.17507005e+00 -9.27007020e-01 -9.04389977e-01 -1.96882337e-01
3.43320101e-01 -4.01272148e-01 -5.65873325e-01 -3.89197737e-01] | [6.967605113983154, 3.9902615547180176] |
116ffa4a-4228-407e-87b4-1abd9c0a8e94 | bibl-amr-parsing-and-generation-with | null | null | https://aclanthology.org/2022.coling-1.485 | https://aclanthology.org/2022.coling-1.485.pdf | BiBL: AMR Parsing and Generation with Bidirectional Bayesian Learning | Abstract Meaning Representation (AMR) offers a unified semantic representation for natural language sentences. Thus transformation between AMR and text yields two transition tasks in opposite directions, i.e., Text-to-AMR parsing and AMR-to-Text generation. Existing AMR studies only focus on one-side improvements despite the duality of the two tasks, and their improvements are greatly attributed to the inclusion of large extra training data or complex structure modifications which harm the inference speed. Instead, we propose data-efficient Bidirectional Bayesian learning (BiBL) to facilitate bidirectional information transition by adopting a single-stage multitasking strategy so that the resulting model may enjoy much lighter training at the same time. Evaluation on benchmark datasets shows that our proposed BiBL outperforms strong previous seq2seq refinements without the help of extra data which is indispensable in existing counterpart models. We release the codes of BiBL at: https://github.com/KHAKhazeus/BiBL. | ['Hai Zhao', 'Zuchao Li', 'Ziming Cheng'] | null | null | null | null | coling-2022-10 | ['amr-parsing'] | ['natural-language-processing'] | [ 5.43963850e-01 5.49139082e-01 -2.29204431e-01 -5.50079048e-01
-1.23140109e+00 -5.28974712e-01 8.70878994e-01 -6.49397299e-02
-3.99915367e-01 9.03573990e-01 6.58174455e-01 -6.05026901e-01
1.81392655e-01 -7.50321448e-01 -7.37459123e-01 -4.58244354e-01
6.31710768e-01 5.38170934e-01 1.32211819e-01 -4.21492010e-01
2.30842993e-01 -2.98119962e-01 -9.62234557e-01 5.25539696e-01
7.90248036e-01 6.52673542e-01 6.05121315e-01 8.54298770e-01
-6.75061405e-01 7.24631846e-01 -3.35944951e-01 -6.23574674e-01
1.39605984e-01 -4.45401996e-01 -1.12464559e+00 -4.60557818e-01
5.76526858e-02 -2.72819966e-01 -3.30203921e-01 9.07242656e-01
5.53517818e-01 1.78911433e-01 4.10083711e-01 -1.01661980e+00
-8.32069755e-01 1.13918746e+00 -6.09094501e-01 -2.49985959e-02
5.15026033e-01 -7.51443729e-02 1.28560448e+00 -9.43984509e-01
3.85720640e-01 1.47367048e+00 4.44836318e-01 9.61508274e-01
-1.21435249e+00 -5.64231515e-01 3.35691780e-01 6.29835427e-02
-9.86605585e-01 -7.25862563e-01 7.20697880e-01 -1.07708484e-01
1.17587626e+00 3.55703562e-01 2.27893159e-01 1.56142592e+00
2.59264827e-01 1.00682700e+00 9.55551684e-01 -4.35312986e-01
1.71266198e-01 -4.64186780e-02 2.07941875e-01 5.80504179e-01
2.11043283e-01 -4.45651472e-01 -7.29556322e-01 -1.03848994e-01
5.15304923e-01 -1.16210751e-01 -7.78548568e-02 1.38666764e-01
-1.34511948e+00 7.25119531e-01 3.64139080e-02 1.89495370e-01
-1.56174600e-01 3.91379356e-01 4.41991329e-01 3.53411019e-01
6.29340053e-01 9.91392955e-02 -6.23515785e-01 -3.77251476e-01
-5.74544966e-01 2.22940773e-01 6.59612954e-01 1.30629516e+00
6.19291246e-01 5.46238609e-02 -5.48820198e-01 1.00812984e+00
3.99579048e-01 6.55735016e-01 5.86412430e-01 -7.97755837e-01
9.76094365e-01 4.10360068e-01 9.03323516e-02 -5.64390361e-01
-2.07230017e-01 -2.19285443e-01 -8.53563428e-01 -5.85512400e-01
3.27681124e-01 -3.28785747e-01 -8.52275729e-01 1.82293880e+00
2.59579211e-01 -2.63305660e-02 3.33938926e-01 5.92119157e-01
9.44348395e-01 9.00379419e-01 7.53197819e-02 2.51429468e-01
1.60496533e+00 -9.44892108e-01 -8.74911726e-01 -6.60391152e-01
1.08225799e+00 -7.80748963e-01 1.32927465e+00 2.47041389e-01
-1.17061400e+00 -3.24624330e-01 -6.44193590e-01 -4.83969152e-01
-3.26758772e-01 2.28684798e-01 7.63154507e-01 5.62583864e-01
-9.21299279e-01 2.42161915e-01 -7.82955170e-01 -1.76285341e-01
2.09291294e-01 6.16790466e-02 -1.95618600e-01 -2.10274622e-01
-1.42518198e+00 6.12414062e-01 3.88437003e-01 3.07196498e-01
-5.83244264e-01 -6.62805676e-01 -1.07089329e+00 -7.27289915e-02
5.44321537e-01 -1.06321704e+00 1.54522157e+00 -7.01883018e-01
-1.81174517e+00 5.80712140e-01 -5.34942031e-01 -3.18919510e-01
6.16052508e-01 -5.98352373e-01 7.52699524e-02 -2.68856943e-01
1.68201044e-01 9.14294541e-01 5.88729799e-01 -9.25535679e-01
-5.35096526e-01 -3.08809042e-01 3.54513712e-02 4.14233357e-01
-3.37772280e-01 -8.93348604e-02 -6.10392034e-01 -7.73355186e-01
-8.28078715e-04 -8.48563969e-01 -2.29730785e-01 -6.39229178e-01
-6.44650578e-01 -5.80125391e-01 2.89021492e-01 -8.18153024e-01
1.14349639e+00 -1.83785713e+00 1.15494013e-01 -4.38738078e-01
-2.11991798e-02 9.05324668e-02 -5.12118816e-01 5.96940160e-01
3.71969379e-02 2.31508344e-01 -3.62788260e-01 -7.40867019e-01
1.01587437e-01 2.62463182e-01 -6.13687515e-01 5.60947917e-02
2.77214706e-01 1.18300653e+00 -1.00833380e+00 -4.19202685e-01
-4.91224341e-02 2.24464759e-01 -5.96947074e-01 2.30577633e-01
-4.45969015e-01 5.42539358e-01 -6.70518339e-01 2.60742545e-01
4.64306742e-01 -2.19682530e-01 2.47917593e-01 4.53628823e-02
1.11216724e-01 1.03041315e+00 -9.54304993e-01 2.24050069e+00
-7.80068696e-01 3.56582463e-01 -1.20621733e-01 -1.10703075e+00
9.88355339e-01 4.95386660e-01 -7.47174323e-02 -8.01749945e-01
5.66616356e-02 7.71347387e-03 -1.71505749e-01 -2.07955509e-01
7.33518481e-01 -2.21292868e-01 -3.33894610e-01 4.86216754e-01
-7.51707479e-02 -1.37299255e-01 4.67128754e-02 3.56438935e-01
1.07308471e+00 5.18690169e-01 2.47112483e-01 -2.15460971e-01
2.57151872e-01 -1.89114124e-01 8.29546869e-01 1.09843349e+00
2.38251209e-01 5.00226319e-01 4.11320031e-01 -2.20128447e-01
-8.79759073e-01 -9.68257010e-01 4.54273522e-02 1.34410882e+00
-3.91968526e-02 -5.55827260e-01 -5.77228606e-01 -7.58423030e-01
-3.04794103e-01 1.29125607e+00 -4.85402882e-01 -1.35929629e-01
-7.33679175e-01 -8.49052668e-01 6.15184069e-01 6.62937999e-01
3.43423277e-01 -1.00920641e+00 -3.83391500e-01 1.25064522e-01
-6.09304607e-01 -1.37577939e+00 -4.81548756e-01 5.34411743e-02
-9.80659664e-01 -3.65466833e-01 -7.51556098e-01 -5.35056472e-01
4.00320113e-01 2.64040023e-01 1.13003540e+00 -1.79990873e-01
-1.01821788e-01 1.61671579e-01 -6.61239445e-01 -4.33760762e-01
-5.49338579e-01 3.37388068e-01 -2.73496687e-01 -1.81525990e-01
2.60456949e-01 -4.46269184e-01 -4.58196700e-01 -8.63305628e-02
-8.22534800e-01 7.77211666e-01 8.05748641e-01 9.24723029e-01
2.02635467e-01 -5.23408771e-01 9.00982022e-01 -1.14640176e+00
4.83768076e-01 -4.26292717e-01 -3.57271940e-01 3.18590164e-01
-4.77712095e-01 3.59027267e-01 6.51898444e-01 -1.18716374e-01
-1.65673530e+00 -2.56030858e-01 -4.31837201e-01 1.75705001e-01
-3.35931391e-01 4.44306642e-01 -1.19549461e-01 7.87036598e-01
2.24509925e-01 5.10677099e-01 -3.26834381e-01 -5.82194924e-01
7.94573963e-01 8.39452326e-01 1.76429391e-01 -7.74218261e-01
6.06693983e-01 3.19238693e-01 -2.99160182e-01 -7.20289469e-01
-1.02921271e+00 -3.62818509e-01 -4.31933880e-01 1.69845462e-01
8.44806731e-01 -1.06649292e+00 -3.70459676e-01 2.38681287e-01
-1.44832754e+00 -6.05223000e-01 -1.34767964e-01 3.63827765e-01
-5.14770985e-01 4.47879404e-01 -7.11982965e-01 -7.93388546e-01
-6.72507346e-01 -8.56672823e-01 1.35223591e+00 -1.65443018e-01
-3.63151222e-01 -1.05288398e+00 2.33842563e-02 8.66962790e-01
4.31733936e-01 -4.05835688e-01 1.02548313e+00 -7.90165365e-01
-4.84072179e-01 -2.49464121e-02 -2.92142987e-01 1.91075295e-01
2.35245883e-01 -4.46308523e-01 -1.00614560e+00 -1.64760888e-01
-1.89157635e-01 -4.00601745e-01 1.01855147e+00 1.94644704e-01
1.08630538e+00 -4.42904055e-01 -2.45272249e-01 4.17470485e-01
1.07667184e+00 1.44475475e-01 6.06674731e-01 1.62783191e-01
8.90834391e-01 6.37473464e-01 5.54433644e-01 3.97928864e-01
7.40095913e-01 7.26635814e-01 1.49684167e-02 -1.55801877e-01
-2.73307174e-01 -4.23312485e-01 7.09227324e-01 1.31450748e+00
1.95212901e-01 -3.39929700e-01 -9.16338086e-01 5.07034957e-01
-2.01138806e+00 -5.16302288e-01 -1.99527830e-01 1.95567274e+00
1.21087742e+00 1.16698310e-01 -2.53904939e-01 -3.71661961e-01
5.04985869e-01 2.91464061e-01 -2.66848713e-01 -4.96590078e-01
1.96368545e-01 1.23941399e-01 4.47435290e-01 7.08331525e-01
-6.95048392e-01 1.30599236e+00 5.79061604e+00 1.10456753e+00
-8.19052041e-01 3.36661488e-01 6.71397626e-01 -1.34860575e-01
-8.36254299e-01 2.57456422e-01 -1.11242068e+00 3.85216832e-01
1.05467188e+00 -3.87361757e-02 3.08813363e-01 5.93334734e-01
3.73215646e-01 1.94873456e-02 -1.03027022e+00 8.45208645e-01
-1.35036796e-01 -1.20531833e+00 5.51192343e-01 -2.68690646e-01
3.87169003e-01 9.38684344e-02 -2.48664662e-01 4.84023064e-01
7.22801447e-01 -7.59259403e-01 7.11124420e-01 4.65498120e-01
6.33072674e-01 -4.54043448e-01 8.11304271e-01 4.55766708e-01
-1.04667842e+00 1.38453450e-02 -4.46908891e-01 -2.10683152e-01
3.69940370e-01 6.57817721e-01 -8.12579572e-01 9.95870113e-01
4.76392597e-01 6.49053037e-01 -4.83498424e-01 2.59677947e-01
-4.46661323e-01 9.06476855e-01 -5.40822260e-02 -1.25771284e-01
1.99564353e-01 -3.85692626e-01 6.51073396e-01 1.44139957e+00
4.08034533e-01 -1.70253411e-01 5.65307736e-02 1.05780923e+00
-1.93701312e-01 1.20215878e-01 -6.97772145e-01 -3.34651917e-02
4.41677630e-01 1.20630336e+00 -5.54323256e-01 -5.35459161e-01
-6.37397885e-01 1.09216273e+00 3.82115453e-01 2.80285060e-01
-7.55238175e-01 -3.32750320e-01 4.65493947e-01 -1.31582126e-01
1.53686017e-01 -1.51492760e-01 -5.47271907e-01 -1.26804042e+00
8.85714665e-02 -7.18794882e-01 4.00600076e-01 -7.99328148e-01
-1.15564120e+00 3.99693877e-01 9.54089388e-02 -7.03305125e-01
-2.52647877e-01 -4.35488582e-01 -5.14427662e-01 9.54169869e-01
-1.61542618e+00 -1.30268371e+00 -2.54574008e-02 3.31927806e-01
1.05832338e+00 3.45196091e-02 8.60769689e-01 1.56675860e-01
-7.77753770e-01 6.34384751e-01 -4.79117259e-02 2.54499793e-01
7.58979976e-01 -1.50648701e+00 9.98020411e-01 9.65103447e-01
7.93004856e-02 6.75591409e-01 5.83256364e-01 -6.73223674e-01
-1.55069077e+00 -1.22531652e+00 1.08228970e+00 -6.97250724e-01
7.06690311e-01 -8.86236370e-01 -7.44348466e-01 9.74673688e-01
4.38502729e-01 -4.35099214e-01 6.37412429e-01 4.01064247e-01
-4.78120446e-01 -8.01986754e-02 -4.32451099e-01 9.17544067e-01
1.17600477e+00 -4.52352196e-01 -7.78413236e-01 2.06330106e-01
1.22811675e+00 -1.60396308e-01 -6.15771174e-01 3.70396793e-01
3.94953489e-01 -6.70738697e-01 6.82053924e-01 -6.80808365e-01
5.52551806e-01 -1.80314660e-01 -2.93808222e-01 -1.14792860e+00
-1.36748657e-01 -7.89695501e-01 1.12621851e-01 1.49295592e+00
6.92491651e-01 -8.80333900e-01 5.69425881e-01 4.77627039e-01
-3.48191053e-01 -6.06521547e-01 -9.53205645e-01 -7.77897000e-01
2.17027158e-01 -6.31859720e-01 4.22150999e-01 8.48646641e-01
2.63369158e-02 1.03381646e+00 -4.60699856e-01 -1.32299453e-01
5.14014661e-01 2.10813493e-01 8.06250811e-01 -8.21474373e-01
-5.13110876e-01 -2.57694125e-01 2.94392139e-01 -1.50991929e+00
1.95068896e-01 -1.22407329e+00 3.63550961e-01 -1.79170215e+00
4.48416173e-01 -4.83362764e-01 -1.14148289e-01 8.40731919e-01
-5.68170249e-01 -7.54783154e-02 9.83159170e-02 -4.78612185e-02
-5.82983851e-01 9.48810935e-01 1.19922411e+00 8.30608383e-02
-1.02644637e-01 -6.27215207e-02 -1.00089979e+00 5.13308764e-01
9.29566085e-01 -7.48237371e-01 -4.76285577e-01 -8.35559726e-01
5.53768814e-01 1.39021084e-01 1.54103652e-01 -4.29680049e-01
-2.96658706e-02 -6.71938583e-02 -6.94845105e-03 -3.58348727e-01
3.10927361e-01 -2.45525420e-01 -1.10595949e-01 2.52527088e-01
-7.11074889e-01 1.33911267e-01 1.93333447e-01 6.93271101e-01
2.62079947e-02 -3.44883174e-01 4.82807755e-01 -2.72336811e-01
-4.11364496e-01 -5.99511638e-02 -5.92495084e-01 2.15776488e-01
4.95007753e-01 4.03801426e-02 -4.32006985e-01 -4.61065412e-01
-3.88258725e-01 1.11252211e-01 8.36784299e-03 7.06952274e-01
4.14474785e-01 -9.87335682e-01 -9.23682690e-01 -1.46868899e-01
5.78886941e-02 3.79213393e-01 1.91362575e-01 8.41202676e-01
2.28433330e-02 7.51815617e-01 2.73203880e-01 -3.29457819e-01
-1.11733139e+00 1.47207588e-01 -1.31878167e-01 -4.55494285e-01
-8.33303750e-01 7.29240000e-01 5.48193157e-01 -7.66888976e-01
5.78397401e-02 -3.60254556e-01 -7.42026642e-02 -9.49658528e-02
4.32871699e-01 2.60198116e-01 7.72057101e-02 -1.67101264e-01
-9.06645954e-02 1.83222607e-01 -3.31873089e-01 -1.78770006e-01
1.27013278e+00 -3.30571949e-01 -1.94949761e-01 6.43323958e-01
7.78028548e-01 -8.87078270e-02 -9.56994295e-01 -2.87621051e-01
3.06031734e-01 -2.65632302e-01 1.14778094e-01 -7.47991025e-01
-6.12860560e-01 9.71150696e-01 1.18489586e-01 1.59277767e-02
8.61456573e-01 1.09110020e-01 1.13742578e+00 6.02563500e-01
1.32304803e-01 -9.71813500e-01 1.15897946e-01 8.26174140e-01
8.68376613e-01 -1.04486287e+00 -2.79149264e-01 -4.73834664e-01
-8.05995107e-01 8.96564484e-01 4.13170040e-01 2.47855112e-01
2.96147257e-01 2.89990455e-01 1.57840569e-02 1.08549051e-01
-1.24490416e+00 -1.21819027e-01 5.10263592e-02 3.97541642e-01
7.50478089e-01 1.15949534e-01 -3.61370802e-01 8.36744905e-01
-2.68500298e-01 -1.67614803e-01 4.87919360e-01 7.99626768e-01
-2.35703200e-01 -1.41901004e+00 -2.44539175e-02 4.26588207e-01
-5.50649405e-01 -6.33898258e-01 -2.19278976e-01 4.53127027e-01
-5.22692144e-01 9.77371573e-01 5.15419338e-03 -1.10721149e-01
1.15215756e-01 4.63281125e-01 3.98337513e-01 -8.32840502e-01
-3.18847805e-01 1.05323918e-01 4.96529341e-01 -5.38700640e-01
-1.70195207e-01 -5.48867285e-01 -1.29195786e+00 -1.94435686e-01
-2.57919759e-01 7.30715692e-02 5.62892497e-01 1.14538610e+00
7.25212812e-01 6.35203779e-01 3.67571592e-01 -5.75876951e-01
-7.13834524e-01 -1.34133041e+00 -2.02645108e-01 3.84041846e-01
1.01347283e-01 -2.68472880e-01 -3.18009406e-02 1.86799049e-01] | [10.966828346252441, 8.949389457702637] |
1b870cb7-271a-49da-b1ee-764c8c914875 | aerial-pass-panoramic-annular-scene | 2105.07209 | null | https://arxiv.org/abs/2105.07209v1 | https://arxiv.org/pdf/2105.07209v1.pdf | Aerial-PASS: Panoramic Annular Scene Segmentation in Drone Videos | Aerial pixel-wise scene perception of the surrounding environment is an important task for UAVs (Unmanned Aerial Vehicles). Previous research works mainly adopt conventional pinhole cameras or fisheye cameras as the imaging device. However, these imaging systems cannot achieve large Field of View (FoV), small size, and lightweight at the same time. To this end, we design a UAV system with a Panoramic Annular Lens (PAL), which has the characteristics of small size, low weight, and a 360-degree annular FoV. A lightweight panoramic annular semantic segmentation neural network model is designed to achieve high-accuracy and real-time scene parsing. In addition, we present the first drone-perspective panoramic scene segmentation dataset Aerial-PASS, with annotated labels of track, field, and others. A comprehensive variety of experiments shows that the designed system performs satisfactorily in aerial panoramic scene parsing. In particular, our proposed model strikes an excellent trade-off between segmentation performance and inference speed suitable, validated on both public street-scene and our established aerial-scene datasets. | ['Jian Bai', 'Kaiwei Wang', 'Xiangdong Zhou', 'Kaikai Wu', 'Kailun Yang', 'Jia Wang', 'Lei Sun'] | 2021-05-15 | null | null | null | null | ['scene-parsing', 'scene-segmentation'] | ['computer-vision', 'computer-vision'] | [ 3.15620124e-01 -3.89230102e-01 2.26571057e-02 -5.28284550e-01
2.37616114e-02 -9.54084933e-01 2.81024247e-01 -4.24155623e-01
-3.08984309e-01 2.88296133e-01 -5.07756114e-01 -4.18473810e-01
-2.08173349e-01 -1.06403184e+00 -7.58914828e-01 -3.61893862e-01
2.83338159e-01 -1.78230792e-01 7.24857748e-01 -9.18148905e-02
5.15819862e-02 5.40084243e-01 -1.45692933e+00 -2.97321260e-01
8.17987025e-01 1.18118310e+00 8.74121666e-01 7.89345086e-01
3.60771596e-01 4.95580673e-01 -2.89487749e-01 -4.13300782e-01
7.95246661e-01 -7.68048391e-02 -6.51440918e-01 5.86375415e-01
8.61215472e-01 -9.03090179e-01 -3.09679955e-01 1.38458610e+00
2.99321525e-02 -1.14774331e-01 1.42936930e-01 -1.18496394e+00
-3.71737748e-01 3.67242634e-01 -7.56324887e-01 5.99841960e-02
1.13556378e-01 1.52288571e-01 6.80324197e-01 -2.34729648e-01
3.09924543e-01 9.12596762e-01 7.95097053e-01 1.67086437e-01
-3.59279096e-01 -6.04280472e-01 3.43756676e-01 -3.83853823e-01
-1.11831844e+00 -4.49635126e-02 5.66023469e-01 -4.05883789e-01
5.08071244e-01 3.18239182e-01 8.80647242e-01 5.67077816e-01
4.95423555e-01 4.91790831e-01 1.08947730e+00 -1.34186953e-01
1.06660150e-01 2.28043050e-02 4.43487093e-02 1.22679389e+00
3.99249285e-01 3.07754218e-03 6.52286857e-02 2.86069363e-01
1.15113366e+00 5.41053355e-01 -4.95934933e-01 -4.66686934e-01
-1.21965981e+00 3.53340000e-01 4.98522907e-01 -1.04807958e-01
-1.29353568e-01 7.37210289e-02 7.22905844e-02 -7.72575885e-02
1.92706242e-01 2.90726662e-01 -7.07955778e-01 3.61135751e-01
-9.35383022e-01 -5.54043241e-02 4.24797505e-01 1.30650711e+00
4.78899002e-01 1.39521152e-01 3.29192013e-01 4.43166316e-01
4.24654692e-01 9.08515096e-01 1.91912815e-01 -1.05934179e+00
4.57108349e-01 8.03468883e-01 6.55691400e-02 -1.08716774e+00
-5.15157223e-01 2.64123501e-03 -7.75461793e-01 2.71119863e-01
2.25186840e-01 -5.36280751e-01 -9.68396485e-01 1.20029914e+00
4.69444752e-01 -1.12958230e-01 1.82542801e-01 1.18062758e+00
7.99564481e-01 8.74596298e-01 -2.42345463e-02 -9.92847830e-02
1.67920530e+00 -1.13187408e+00 -4.78459895e-01 -6.29508853e-01
1.20426364e-01 -9.18052793e-01 9.99471545e-01 4.61441904e-01
-9.81405318e-01 -7.52985597e-01 -1.10428047e+00 -8.08539540e-02
-4.07912314e-01 8.91666770e-01 8.18777144e-01 7.00867534e-01
-8.74121785e-01 1.74067885e-01 -6.29332185e-01 -6.13926888e-01
1.28507063e-01 2.43477196e-01 -4.19445813e-01 -1.25391826e-01
-8.62576187e-01 3.55498791e-01 6.76703155e-01 2.38454610e-01
-8.81426394e-01 -5.07731616e-01 -9.04786170e-01 -1.07243657e-02
6.69343412e-01 -7.12616324e-01 1.20652544e+00 -8.90066445e-01
-1.64464033e+00 7.78258145e-01 4.91805732e-01 -5.01229703e-01
2.75524288e-01 -4.51693594e-01 -2.33749226e-01 4.07823592e-01
-1.44492919e-02 9.14402902e-01 8.48633289e-01 -1.06783700e+00
-1.08718371e+00 -5.51909089e-01 5.92728734e-01 6.04451120e-01
-2.26613179e-01 1.02963857e-01 -5.63758492e-01 -4.09884542e-01
2.08931997e-01 -9.37973619e-01 -3.49228382e-01 9.26518887e-02
-4.80799943e-01 2.41063088e-01 1.12649286e+00 -5.91431439e-01
9.39417779e-01 -2.03753734e+00 -2.07831249e-01 -2.11961299e-01
-2.72201654e-02 5.73179781e-01 3.14864576e-01 6.53857226e-03
2.55534619e-01 -2.47347489e-01 -4.23392594e-01 1.96755171e-01
-4.46802557e-01 -3.46561782e-02 -3.54282379e-01 3.58740807e-01
-1.88895196e-01 5.16589940e-01 -7.75140822e-01 -7.15490699e-01
6.65544808e-01 -9.75373909e-02 -4.21848506e-01 3.14048916e-01
-3.02108973e-01 8.39835480e-02 -6.02720678e-01 1.11207163e+00
9.23010409e-01 -2.58200523e-02 8.34829956e-02 -4.63966697e-01
-4.69663113e-01 -6.64520800e-01 -1.10322702e+00 1.58879340e+00
-3.41332912e-01 6.52612269e-01 3.75163376e-01 -4.16069239e-01
1.11252224e+00 -1.30505949e-01 1.13437466e-01 -1.78241894e-01
3.77488464e-01 -9.61623862e-02 -4.58371937e-01 -7.01490343e-01
9.67015386e-01 3.73539507e-01 -1.69392899e-01 -6.25310317e-02
-5.12467995e-02 -6.64444983e-01 1.81646481e-01 -4.49110568e-02
5.71140826e-01 3.78556460e-01 5.00884652e-01 -2.64053524e-01
5.17004550e-01 4.29220468e-01 4.61978018e-01 2.74899185e-01
-1.18396588e-01 6.90661132e-01 2.44262367e-01 -6.09127700e-01
-9.52095866e-01 -6.96114659e-01 -9.09594074e-02 6.22056127e-01
9.56599176e-01 -2.80866295e-01 -1.18078005e+00 -5.86327553e-01
-4.95282084e-01 2.68266022e-01 -1.10009961e-01 2.54956245e-01
-3.27478260e-01 -6.02469623e-01 6.17864251e-01 5.50457656e-01
1.34506023e+00 -6.29582405e-01 -1.46518123e+00 -1.85363278e-01
-1.05504140e-01 -1.69799721e+00 -3.35165650e-01 -8.50192457e-02
-8.74825060e-01 -1.59338927e+00 -5.64038336e-01 -9.01357949e-01
6.30173802e-01 1.04516053e+00 8.33734512e-01 -3.52097422e-01
-2.72999763e-01 4.39035952e-01 -2.23550528e-01 -2.29924917e-01
1.61892071e-01 -1.07728325e-01 -6.05702363e-02 -1.01264901e-01
2.18636617e-01 -1.14308171e-01 -7.13471770e-01 5.73577702e-01
-1.06460130e+00 3.21972638e-01 7.12822437e-01 3.51409376e-01
6.62165642e-01 4.93326187e-01 -3.33388388e-01 -6.15005493e-01
1.72147110e-01 -2.32850045e-01 -1.44449234e+00 5.50227463e-01
-3.01166207e-01 -7.63248444e-01 1.02310574e+00 3.62708904e-02
-1.16874957e+00 5.14949143e-01 1.06650360e-01 -5.11983514e-01
-6.31783903e-01 4.72559817e-02 -3.46219808e-01 -2.62548029e-01
3.10778409e-01 2.42993787e-01 -1.63781166e-01 -1.85015991e-01
2.87752688e-01 8.94851387e-01 7.44473934e-01 -8.36695507e-02
6.39199078e-01 5.65253735e-01 9.65790525e-02 -1.17942035e+00
-9.10483241e-01 -4.48044628e-01 -9.26758766e-01 -4.72017378e-01
1.35262585e+00 -1.26227868e+00 -5.78502655e-01 7.97587395e-01
-1.09106433e+00 -3.96946520e-01 1.72227621e-01 6.42012894e-01
-3.70391786e-01 5.78323841e-01 -3.46103847e-01 -6.01921916e-01
-4.23023850e-01 -1.31723213e+00 1.31349242e+00 6.92466080e-01
4.06968325e-01 -7.34461665e-01 -4.53070432e-01 6.32395208e-01
1.97560191e-02 4.08305496e-01 4.07916695e-01 7.75454864e-02
-1.01554263e+00 -1.39000192e-01 -6.21924877e-01 5.00299335e-01
-1.13968767e-01 4.25909191e-01 -7.85671532e-01 1.66300461e-01
-4.91328202e-02 -1.06419042e-01 6.55572474e-01 6.89846873e-01
1.33648038e+00 1.72685366e-02 -3.72852266e-01 1.02906418e+00
1.72834718e+00 5.11510551e-01 3.68907452e-01 1.85551405e-01
8.16200674e-01 4.50883240e-01 1.23541427e+00 3.61395746e-01
4.50218886e-01 4.79304701e-01 7.79188275e-01 -1.40403137e-01
1.90912217e-01 -3.44132572e-01 4.07066971e-01 4.91047055e-01
-3.81928496e-02 -5.75230718e-01 -8.65885615e-01 2.93963939e-01
-1.61026359e+00 -6.52055144e-01 -4.12583411e-01 2.01606250e+00
1.85510129e-01 -4.93707741e-03 -2.28836074e-01 -1.41894713e-01
7.36995816e-01 3.33609521e-01 -3.99181515e-01 -1.42018184e-01
-5.02457134e-02 -5.00342429e-01 1.07173538e+00 3.23503703e-01
-1.70391166e+00 1.33523297e+00 5.63341951e+00 8.16876829e-01
-1.17148590e+00 -2.08065197e-01 4.60141480e-01 4.10826623e-01
3.22410524e-01 -8.85111615e-02 -1.11201501e+00 4.23204839e-01
4.18463647e-01 4.16496903e-01 2.96175659e-01 1.24431229e+00
-4.93794084e-02 -4.03203338e-01 -4.85453516e-01 1.01008773e+00
2.15628460e-01 -1.00421393e+00 2.90766750e-02 -1.00333370e-01
5.70424855e-01 -3.64811881e-03 -5.33547774e-02 -2.69892752e-01
2.16391191e-01 -5.23595154e-01 6.91317558e-01 2.61547506e-01
8.85577023e-01 -4.69476253e-01 7.65301704e-01 4.96030003e-01
-1.39871407e+00 -3.16993237e-01 -7.32639909e-01 -6.19827770e-02
3.05687964e-01 3.42253238e-01 -4.63792801e-01 6.75431371e-01
1.00723195e+00 7.69764483e-01 -7.43350327e-01 9.66738343e-01
-1.28829017e-01 3.05782497e-01 -4.88086402e-01 9.43100974e-02
6.52188420e-01 -7.13282466e-01 4.27297175e-01 8.88809085e-01
5.88571370e-01 4.64440286e-01 4.75485712e-01 5.38225532e-01
9.74963382e-02 -1.20949082e-01 -1.06041348e+00 -1.14238011e-02
3.15589696e-01 1.58713889e+00 -1.16527534e+00 -3.06559145e-01
-4.71394062e-01 8.90866756e-01 -2.77073413e-01 -4.57957387e-03
-1.02251112e+00 -4.33483452e-01 2.99802125e-01 -1.14566432e-02
2.05652505e-01 -4.31607515e-01 -5.91652654e-02 -1.10947478e+00
-2.97963530e-01 -5.45462072e-01 9.56899449e-02 -1.29770958e+00
-6.53719723e-01 8.62152576e-01 2.48828098e-01 -1.37143898e+00
4.39918578e-01 -9.89309788e-01 -4.09601241e-01 2.49357775e-01
-1.52013028e+00 -1.59401000e+00 -9.95086789e-01 7.14187682e-01
1.06936038e+00 -2.78428555e-01 5.78997731e-01 1.25325501e-01
-8.43348622e-01 -1.62369162e-01 -1.47545695e-01 2.37113371e-01
4.28184718e-01 -1.09520614e+00 3.09833795e-01 1.46855092e+00
-4.85500954e-02 3.11497837e-01 2.90737480e-01 -6.41164839e-01
-1.38377106e+00 -1.53901589e+00 1.47442237e-01 -3.07956845e-01
2.35137135e-01 -4.72413488e-02 -1.30465537e-01 7.90856183e-01
3.16559047e-01 5.86342700e-02 2.63889849e-01 -5.64866245e-01
2.45384827e-01 -2.46023178e-01 -1.04100740e+00 6.71757221e-01
9.76615608e-01 6.08224384e-02 -4.08819735e-01 5.22278130e-01
9.45418477e-01 -7.71838069e-01 -6.52275920e-01 6.62539005e-01
6.54485583e-01 -1.24814284e+00 1.08573687e+00 6.13584043e-03
5.65802515e-01 -5.74209154e-01 -3.87192100e-01 -9.51627672e-01
-1.32403532e-02 -3.81069243e-01 3.29139709e-01 1.10259318e+00
-9.16678756e-02 -5.48139751e-01 6.11556590e-01 3.75312388e-01
-4.24389929e-01 -6.25076354e-01 -9.16180015e-02 -4.93417263e-01
-6.13150179e-01 -3.87949608e-02 5.65297306e-01 5.33317208e-01
-5.76246202e-01 3.08520764e-01 -4.60525692e-01 6.88450873e-01
6.21897697e-01 5.35775185e-01 9.39075649e-01 -1.21565306e+00
-1.14732236e-02 -3.65965106e-02 -2.65554160e-01 -1.43494225e+00
-1.25120297e-01 -2.38319471e-01 4.41691875e-02 -1.45938432e+00
-1.18169539e-01 -4.37805176e-01 3.85448933e-01 2.14778021e-01
2.18212940e-02 3.97389591e-01 1.89818054e-01 1.95697453e-02
-6.74493253e-01 2.07804218e-01 1.35428214e+00 6.02724962e-02
-9.49944258e-02 2.42633238e-01 -3.36335957e-01 1.34406734e+00
9.98514950e-01 -1.38388753e-01 -7.16445506e-01 -8.54796112e-01
-5.18214181e-02 1.60933211e-01 4.74502563e-01 -1.23203027e+00
3.90616000e-01 -3.73353273e-01 5.39185047e-01 -8.05937707e-01
5.20447493e-01 -1.22680175e+00 8.43766406e-02 5.39520442e-01
1.07088186e-01 2.16401786e-01 5.68589568e-02 6.77746773e-01
-2.98186600e-01 -3.03375781e-01 7.91676521e-01 -4.92411643e-01
-1.34961045e+00 5.13945460e-01 -2.74322182e-01 -1.50243685e-01
1.54856694e+00 -4.90489632e-01 -4.22154337e-01 1.63934499e-01
1.02900267e-01 2.74247020e-01 7.87965298e-01 3.64171922e-01
7.51338780e-01 -6.15900993e-01 -7.16987029e-02 6.20691419e-01
-7.13349581e-02 4.48097229e-01 4.26253647e-01 6.82502210e-01
-1.46482754e+00 7.24285424e-01 -4.87151831e-01 -6.70222580e-01
-1.50837624e+00 5.31025112e-01 2.52628744e-01 2.37144306e-01
-7.12080598e-01 7.59456158e-01 3.50501209e-01 -7.15687871e-01
3.08512542e-02 -8.24060321e-01 -3.59263420e-01 -1.87315866e-01
2.82686889e-01 3.27190101e-01 -2.19786718e-01 -5.74518263e-01
-1.00373626e-01 1.20952010e+00 4.23332006e-01 3.27064365e-01
9.64617193e-01 -4.67487812e-01 -1.60777524e-01 -5.61667942e-02
6.25740051e-01 9.34852939e-03 -1.53421295e+00 1.23690024e-01
-6.16940498e-01 -7.19889343e-01 3.16500872e-01 -5.41318059e-01
-1.08609939e+00 8.62070382e-01 5.53793371e-01 2.25270987e-01
1.23486388e+00 -5.53313673e-01 1.12346053e+00 5.64719379e-01
6.52877927e-01 -8.87277722e-01 -3.75475734e-01 3.51572990e-01
2.83594996e-01 -1.30931008e+00 1.39698103e-01 -7.65832603e-01
-8.26651216e-01 1.50147820e+00 9.85368192e-01 -2.41899580e-01
5.10058045e-01 2.54290432e-01 7.01911002e-02 -2.81044930e-01
-1.71441168e-01 -2.46441707e-01 6.74927309e-02 4.39861476e-01
-1.90179124e-01 1.79844186e-01 1.51736990e-01 4.85094726e-01
-3.60545009e-01 -3.72452736e-02 6.86069906e-01 7.25227356e-01
-7.10444093e-01 -4.50407028e-01 -4.22442764e-01 1.89745799e-01
-4.34250414e-01 1.05965706e-02 -2.99599111e-01 1.04629540e+00
4.18685377e-01 8.81021380e-01 1.70374393e-01 -3.23567688e-01
2.87423551e-01 -5.02505243e-01 3.14410716e-01 -4.11681920e-01
-3.00464928e-01 -8.25978890e-02 -1.83934718e-01 -6.21115863e-01
-6.55946672e-01 -2.28688642e-01 -7.75332570e-01 -1.44354239e-01
-4.93765324e-01 -8.13075826e-02 7.01782703e-01 7.36041427e-01
3.90384607e-02 3.93595070e-01 5.78546524e-01 -7.30500400e-01
-2.20880508e-01 -5.10070920e-01 -7.00372756e-01 -3.40758830e-01
1.23041250e-01 -4.01155263e-01 2.61076819e-02 5.18807113e-01] | [8.722941398620605, -0.8853453993797302] |
ef0f1123-4fac-4f84-a9a2-2a72632c57d8 | an-efficient-and-scalable-deep-learning | 2011.09577 | null | https://arxiv.org/abs/2011.09577v3 | https://arxiv.org/pdf/2011.09577v3.pdf | An Efficient and Scalable Deep Learning Approach for Road Damage Detection | Pavement condition evaluation is essential to time the preventative or rehabilitative actions and control distress propagation. Failing to conduct timely evaluations can lead to severe structural and financial loss of the infrastructure and complete reconstructions. Automated computer-aided surveying measures can provide a database of road damage patterns and their locations. This database can be utilized for timely road repairs to gain the minimum cost of maintenance and the asphalt's maximum durability. This paper introduces a deep learning-based surveying scheme to analyze the image-based distress data in real-time. A database consisting of a diverse population of crack distress types such as longitudinal, transverse, and alligator cracks, photographed using mobile-device is used. Then, a family of efficient and scalable models that are tuned for pavement crack detection is trained, and various augmentation policies are explored. Proposed models, resulted in F1-scores, ranging from 52% to 56%, and average inference time from 178-10 images per second. Finally, the performance of the object detectors are examined, and error analysis is reported against various images. The source code is available at https://github.com/mahdi65/roadDamageDetection2020. | ['Hassan Zargarzadeh', 'Amir R. Kashani', 'M-Mahdi Naddaf-Sh', 'Sadra Naddaf-sh'] | 2020-11-18 | null | null | null | null | ['road-damage-detection'] | ['computer-vision'] | [-6.10773563e-02 -1.90620765e-01 2.80320505e-03 -3.20201993e-01
-1.03954756e+00 -2.82172263e-01 -9.78582799e-02 5.72548956e-02
-2.41464123e-01 5.73213935e-01 -6.61861822e-02 -4.09575522e-01
-1.10751025e-01 -1.24518311e+00 -5.76306224e-01 -8.34686875e-01
-7.21749142e-02 2.08931882e-02 4.71015334e-01 -1.38218001e-01
6.60908401e-01 8.55473638e-01 -1.76068521e+00 7.22946301e-02
9.26096857e-01 1.07017481e+00 4.70803350e-01 8.63802731e-01
2.74937570e-01 5.91598809e-01 -4.55119491e-01 -3.22836071e-01
-1.54591411e-01 3.99020463e-01 -5.34351766e-01 -1.02765843e-01
6.03388548e-01 -1.13720167e+00 -6.20113254e-01 8.98496449e-01
7.82197058e-01 1.08931772e-01 8.00775886e-01 -1.04028821e+00
-5.28405130e-01 4.57272977e-02 -5.93670070e-01 5.25949836e-01
6.72905222e-02 2.86094397e-01 7.62543559e-01 -1.33803785e+00
-8.38260651e-02 1.16823697e+00 7.38267303e-01 3.45564097e-01
-5.07121742e-01 -5.78949571e-01 -1.75071076e-01 6.85159683e-01
-1.37618887e+00 -4.13017929e-01 8.01919699e-01 -5.39510369e-01
6.92266762e-01 1.40693158e-01 5.28536260e-01 7.65256405e-01
2.52828091e-01 6.51812553e-01 7.20202684e-01 -1.99267358e-01
1.61339611e-01 -2.37673134e-01 2.66139656e-01 1.11318052e+00
5.64213514e-01 -1.00706868e-01 -1.83612138e-01 2.07211912e-01
7.10248828e-01 1.13692269e-01 -1.94467381e-01 4.61295307e-01
-4.19289470e-01 6.19913220e-01 4.74072248e-01 -2.90749967e-01
-4.62505966e-01 2.29042113e-01 3.93006474e-01 -2.70612855e-02
4.73434031e-01 -1.39214814e-01 -1.05095536e-01 -1.06738418e-01
-5.41121423e-01 2.41820350e-01 2.77079314e-01 6.29144490e-01
8.08162332e-01 3.77196640e-01 1.40088573e-01 1.16491556e+00
5.60628235e-01 9.96491194e-01 -1.58596650e-01 -1.20057428e+00
7.43038654e-01 5.54156065e-01 8.36468264e-02 -1.31596589e+00
-4.83066976e-01 7.12746456e-02 -7.88169742e-01 3.85862976e-01
2.94625037e-03 -4.17878360e-01 -9.20293987e-01 9.97458935e-01
4.77535516e-01 3.14999014e-01 -3.81220698e-01 8.02072942e-01
7.49028981e-01 8.21022689e-01 9.09706652e-02 2.69785494e-01
1.19952059e+00 -6.01021826e-01 -5.74805200e-01 -2.95765936e-01
4.75898832e-01 -5.14105022e-01 1.34675491e+00 3.48701030e-01
-9.68449354e-01 -3.51583511e-01 -1.11676466e+00 1.00801475e-02
-3.41808587e-01 4.90368664e-01 2.26065442e-01 5.56941330e-01
-7.89657056e-01 5.00490665e-01 -8.99688721e-01 -2.51903176e-01
1.06758058e+00 -1.31707247e-02 -1.49627447e-01 -5.02946198e-01
-1.01378345e+00 1.04346311e+00 -2.87694871e-01 8.42086434e-01
-1.48051310e+00 -6.12874746e-01 -6.72118664e-01 6.22321218e-02
1.28661960e-01 -2.80699786e-02 9.89490747e-01 2.71373898e-01
-8.98189783e-01 6.85037613e-01 1.43751517e-01 -1.38262004e-01
3.39912683e-01 -7.06209958e-01 -4.80413973e-01 4.77916479e-01
9.74706709e-02 1.25107303e-01 5.21238208e-01 -1.23535991e+00
-8.12995374e-01 -3.15737605e-01 -1.02912366e-01 5.99870756e-02
-5.40883482e-01 -7.58387847e-03 -1.60310000e-01 -3.95114273e-01
-3.30880284e-02 -7.66728759e-01 -1.44836053e-01 4.67716783e-01
-5.77618837e-01 -4.00931425e-02 1.09466434e+00 -1.08519244e+00
1.27681708e+00 -1.87222970e+00 -4.29450065e-01 1.43046916e-01
-9.16838273e-02 4.84172225e-01 4.85837013e-02 7.91052222e-01
2.43835643e-01 1.38861716e-01 -7.70317852e-01 -2.91812599e-01
-2.56478578e-01 3.96071672e-01 -1.07169226e-01 6.89696431e-01
3.90173584e-01 2.54784644e-01 -4.60887074e-01 -5.95707178e-01
3.25906426e-01 2.46082410e-01 -3.10292453e-01 3.69863331e-01
3.39231193e-01 -4.84171994e-02 -5.25886595e-01 1.40035915e+00
7.98690677e-01 3.16783488e-01 -5.54410398e-01 -3.03729206e-01
-1.68382704e-01 -4.86821868e-02 -1.20867956e+00 1.05031598e+00
-3.83800030e-01 8.57741773e-01 1.65570647e-01 -1.12757099e+00
9.92400646e-01 3.36686552e-01 3.70657265e-01 -5.59067309e-01
8.40898976e-02 2.54875720e-01 -6.10112786e-01 -1.41815972e+00
3.94802034e-01 3.06948572e-01 3.17848325e-02 2.14921743e-01
-5.80865026e-01 -1.15128485e-02 -1.43400514e-02 5.44698723e-02
1.23081887e+00 -3.80987585e-01 -4.61895823e-01 -4.64458950e-02
2.85168022e-01 8.11729804e-02 3.74463409e-01 3.78541023e-01
-2.98108637e-01 6.11288548e-01 1.56361505e-01 -4.77301180e-01
-1.12960124e+00 -1.23941493e+00 -2.52850026e-01 7.37671912e-01
2.11480200e-01 6.15040600e-01 -8.10040116e-01 -3.85187060e-01
2.67421484e-01 5.28739333e-01 -5.30997396e-01 -5.20353079e-01
-6.59542382e-01 -7.14762747e-01 8.89764845e-01 9.42852616e-01
8.19950819e-01 -9.77703691e-01 -5.79841435e-01 2.17067264e-02
-2.69241691e-01 -7.70495057e-01 -8.63954723e-02 -4.24550325e-01
-9.98096466e-01 -1.44577205e+00 -6.29774094e-01 -9.55905855e-01
7.11721718e-01 4.87265617e-01 6.03006780e-01 5.82349181e-01
-6.64634466e-01 4.59440172e-01 -1.73515469e-01 -4.24979329e-01
-9.39490870e-02 -1.99247926e-01 -1.54625818e-01 -1.74081866e-02
9.47793014e-03 -4.89972383e-01 -1.03086615e+00 2.52626687e-01
-6.76220119e-01 -4.68534499e-01 6.65800393e-01 6.17284238e-01
3.28279853e-01 2.85764243e-02 9.25534546e-01 -4.76598471e-01
5.60167849e-01 -8.85792196e-01 -4.44321841e-01 1.23624958e-01
-7.24081933e-01 -6.09003067e-01 6.40794113e-02 -1.40259519e-01
-1.38035202e+00 -8.82524103e-02 -5.44663191e-01 -2.91211307e-01
-4.13403422e-01 7.62064517e-01 -1.08517982e-01 1.46415308e-01
5.91222167e-01 -1.14734434e-01 -8.03209990e-02 -6.56263828e-01
2.50037536e-02 9.34338391e-01 9.13673759e-01 -6.22884333e-01
7.31287062e-01 5.13465285e-01 -1.06130026e-01 -1.20365191e+00
-6.62388444e-01 -4.33539957e-01 -4.63153452e-01 -1.00496113e+00
7.30466008e-01 -9.72728312e-01 -4.78391588e-01 1.19766557e+00
-1.13719678e+00 -3.26362878e-01 2.11847723e-01 3.51167828e-01
1.64361317e-02 4.45707381e-01 -9.36650813e-01 -1.29138100e+00
-5.18116891e-01 -7.23804653e-01 9.43402648e-01 4.17380154e-01
5.51302969e-01 -6.46989346e-01 1.46428663e-02 7.21319854e-01
2.05266848e-01 4.58488435e-01 8.14650357e-01 4.42877226e-02
-6.01306498e-01 -5.89081347e-01 -4.35306102e-01 7.99497545e-01
7.24060014e-02 4.31479454e-01 -1.04391634e+00 -4.10092846e-02
-4.85052884e-01 -2.88337618e-01 9.03706610e-01 4.50460345e-01
1.16249907e+00 -2.85886437e-01 -2.69090205e-01 6.21825829e-02
1.45720840e+00 3.86405885e-01 1.11401486e+00 1.32309794e-01
9.33902502e-01 7.73784578e-01 9.36285794e-01 5.73436141e-01
7.30875373e-01 9.44223329e-02 1.01338744e+00 -2.49511689e-01
-3.93716693e-02 -1.23736024e-01 1.74568519e-01 7.23700345e-01
-2.86936373e-01 -4.15689468e-01 -1.25812650e+00 1.16017246e+00
-1.41690445e+00 -1.19227624e+00 -7.72735000e-01 2.00866246e+00
4.66604948e-01 -1.38197290e-02 -3.32691669e-02 4.19451982e-01
1.01067924e+00 -6.16561994e-02 -7.30454922e-01 -1.18049838e-01
1.28340274e-01 1.83631619e-03 5.98782599e-01 4.09831017e-01
-9.05692697e-01 6.44112706e-01 5.95874786e+00 5.41150510e-01
-9.25593019e-01 1.13021061e-01 6.21871114e-01 1.01961449e-01
-1.74160436e-01 -1.73500776e-01 -5.20152390e-01 5.50613940e-01
8.84077072e-01 5.27103245e-01 2.39423569e-02 8.91664267e-01
7.32604206e-01 -3.56337517e-01 -3.87459934e-01 6.08361900e-01
-2.99796045e-01 -1.27006996e+00 -3.67733449e-01 -5.93257509e-02
3.61271709e-01 2.85394818e-01 -5.89053407e-02 -1.54760608e-03
9.30872113e-02 -6.74835324e-01 6.79765821e-01 8.39352906e-01
7.79307485e-01 -8.64531279e-01 7.46875823e-01 1.40157223e-01
-1.31944478e+00 -5.96426904e-01 -3.50492209e-01 -2.72177011e-02
4.80766565e-01 6.27404749e-01 -6.19784892e-01 1.27207175e-01
1.14801860e+00 6.81704581e-01 -4.05359477e-01 1.24542522e+00
-2.73516417e-01 1.19714725e+00 -2.74007529e-01 1.46630146e-02
-1.10122385e-02 -7.31179118e-02 2.69946277e-01 1.19078898e+00
5.54524243e-01 2.62439907e-01 -1.88369304e-02 3.90746742e-01
5.33188395e-02 -2.32423738e-01 -5.61894774e-01 4.50390786e-01
1.14236832e+00 1.09515345e+00 -3.47078979e-01 -1.54617488e-01
-2.83074260e-01 4.43049043e-01 2.24584967e-01 3.07166845e-01
-1.15977240e+00 -6.64169729e-01 7.06655741e-01 4.49580073e-01
1.62777200e-01 -3.95661443e-01 -4.74236995e-01 -6.23724163e-01
1.02788880e-01 -3.43971282e-01 3.70902389e-01 -1.03721797e+00
-1.20460010e+00 4.78884578e-03 1.49057791e-01 -1.08787012e+00
6.42485321e-01 -5.00587702e-01 -1.17615581e+00 4.93231654e-01
-1.63963962e+00 -1.15171266e+00 -7.00479448e-01 4.95165229e-01
7.71616578e-01 3.47306356e-02 2.52553821e-01 1.14279974e+00
-1.33406258e+00 4.57470685e-01 8.07503387e-02 3.26895654e-01
3.60946089e-01 -8.22907448e-01 2.88487941e-01 1.16236925e+00
-5.36478460e-01 -2.40318273e-04 4.71858203e-01 -7.76377261e-01
-1.25327909e+00 -1.35339212e+00 4.12572652e-01 -1.47131935e-01
3.77953559e-01 2.10122809e-01 -1.30689502e+00 2.89289296e-01
-1.46304056e-01 -1.78773981e-02 1.45285949e-01 -3.34400028e-01
4.37994935e-02 -4.16262150e-01 -1.29871988e+00 4.62228894e-01
8.93537998e-01 -4.89834666e-01 -1.27143577e-01 3.21342021e-01
4.21141118e-01 -3.20633143e-01 -9.51511383e-01 2.77096272e-01
7.30035782e-01 -6.98342025e-01 1.04468489e+00 -2.56234437e-01
8.79290044e-01 -2.10770909e-02 -3.03594023e-01 -7.94568777e-01
-1.27834961e-01 1.56142786e-01 -2.06742316e-01 1.40623283e+00
4.10797358e-01 -6.66842997e-01 8.64104509e-01 7.03714132e-01
-9.83132422e-01 -1.03635633e+00 -1.00744796e+00 -3.44403088e-01
2.55999472e-02 -7.34499514e-01 4.09325987e-01 6.41981602e-01
-5.68175435e-01 -1.23815112e-01 -2.93764591e-01 1.01234746e+00
6.80267870e-01 -4.63302732e-01 4.83697236e-01 -1.15660763e+00
4.71398354e-01 -9.96834263e-02 -3.34522963e-01 -7.35082805e-01
-3.24323118e-01 -2.43037730e-01 4.24642593e-01 -1.97028744e+00
-1.52255028e-01 -6.00945532e-01 -1.96550488e-01 7.27940261e-01
-7.84209073e-02 7.31841698e-02 -4.43194896e-01 1.78364426e-01
1.26347631e-01 6.67377889e-01 1.22740090e+00 -2.95680940e-01
2.96750247e-01 2.21400186e-01 -3.09466422e-01 8.04053485e-01
1.05408955e+00 -6.12732172e-01 -4.20559138e-01 -7.93245077e-01
2.42398456e-02 2.38132104e-01 9.69529748e-01 -1.25897658e+00
3.29864860e-01 -3.83383751e-01 -1.05429664e-01 -9.43054676e-01
4.81682599e-01 -7.06238747e-01 -1.71246514e-01 6.33051634e-01
-9.92714390e-02 9.48641896e-02 4.20333818e-02 8.58066320e-01
-5.55802807e-02 -4.53842491e-01 9.11829829e-01 1.71689361e-01
-8.18147421e-01 4.13061559e-01 -8.10118914e-01 -1.26715869e-01
1.02426708e+00 -4.46463108e-01 -6.29617214e-01 -1.72703326e-01
-2.99822956e-01 3.34781080e-01 -3.32329497e-02 2.46113241e-01
1.17921782e+00 -1.22123873e+00 -8.54345977e-01 -1.20924316e-01
-3.52254882e-02 3.16112936e-01 9.67922628e-01 7.68401623e-01
-1.05810356e+00 -2.75070935e-01 -2.75960177e-01 -5.71088493e-01
-1.18596041e+00 -1.04976602e-01 3.78518194e-01 2.70700216e-01
-5.98235726e-01 8.68628800e-01 -4.91393268e-01 -1.34962276e-01
1.45428568e-01 -4.97459888e-01 -3.82747710e-01 1.27980217e-01
3.69689226e-01 1.23655248e+00 2.01743513e-01 -5.54391623e-01
-3.53116006e-01 6.77028060e-01 1.90442190e-01 1.07720986e-01
1.51888478e+00 -3.60599875e-01 2.19406888e-01 1.93942599e-02
1.14846909e+00 -3.58780950e-01 -1.51710105e+00 7.97288865e-02
-1.58517122e-01 -5.55473387e-01 4.01964754e-01 -5.60422480e-01
-1.57483327e+00 1.10881543e+00 1.16984546e+00 1.06868640e-01
1.15708458e+00 -1.13529906e-01 1.35960734e+00 5.76338172e-01
1.65353402e-01 -1.30000520e+00 3.62742484e-01 9.37213153e-02
9.64830577e-01 -1.50164473e+00 5.55381551e-02 -2.15976119e-01
-6.47761703e-01 9.76042092e-01 8.13816667e-01 -3.76375735e-01
7.61774004e-01 3.28669548e-01 4.16579098e-02 -7.17300117e-01
-4.29032505e-01 2.47440338e-02 -1.23837881e-01 6.78431749e-01
-1.76681608e-01 -5.13124391e-02 1.65273473e-01 3.66666019e-01
4.30386603e-01 -3.23556930e-01 5.90964496e-01 1.21782732e+00
-1.14589357e+00 -3.61272275e-01 -5.29068053e-01 7.12251246e-01
-3.19656223e-01 2.01501250e-01 1.84244111e-01 7.57267594e-01
5.40052503e-02 1.24447811e+00 -8.12644809e-02 -4.41472322e-01
7.56156087e-01 -4.25765067e-01 -1.42088339e-01 -4.17223752e-01
-3.30261253e-02 -3.99138600e-01 4.37028319e-01 -2.95426279e-01
-1.21650234e-01 -7.10099399e-01 -1.37747717e+00 -4.08015728e-01
-6.20739222e-01 -2.74235606e-01 7.63575196e-01 8.90392184e-01
1.82194099e-01 4.05764997e-01 1.05931687e+00 -9.19206858e-01
-4.56379533e-01 -9.32086349e-01 -6.00831747e-01 -4.71305698e-02
4.80100155e-01 -9.98625994e-01 -3.38765144e-01 3.09152871e-01] | [7.43552303314209, 1.2182921171188354] |
4c7dd664-38cd-4c99-b559-d249af9054da | application-of-deep-q-network-in-portfolio | 2003.06365 | null | https://arxiv.org/abs/2003.06365v1 | https://arxiv.org/pdf/2003.06365v1.pdf | Application of Deep Q-Network in Portfolio Management | Machine Learning algorithms and Neural Networks are widely applied to many different areas such as stock market prediction, face recognition and population analysis. This paper will introduce a strategy based on the classic Deep Reinforcement Learning algorithm, Deep Q-Network, for portfolio management in stock market. It is a type of deep neural network which is optimized by Q Learning. To make the DQN adapt to financial market, we first discretize the action space which is defined as the weight of portfolio in different assets so that portfolio management becomes a problem that Deep Q-Network can solve. Next, we combine the Convolutional Neural Network and dueling Q-net to enhance the recognition ability of the algorithm. Experimentally, we chose five lowrelevant American stocks to test the model. The result demonstrates that the DQN based strategy outperforms the ten other traditional strategies. The profit of DQN algorithm is 30% more than the profit of other strategies. Moreover, the Sharpe ratio associated with Max Drawdown demonstrates that the risk of policy made with DQN is the lowest. | ['Jionglong Su', 'Zhengyong Jiang', 'Yuan Gao', 'Yi Hu', 'Ziming Gao'] | 2020-03-13 | null | null | null | null | ['stock-market-prediction'] | ['time-series'] | [-6.72234774e-01 1.42872175e-02 -3.03407818e-01 1.77836586e-02
6.69910386e-02 -3.90875459e-01 1.31459072e-01 -4.10009027e-01
-5.68133295e-01 1.05434167e+00 -5.64847440e-02 -4.35221463e-01
-5.14938116e-01 -1.36184895e+00 -4.94840473e-01 -5.53093016e-01
-3.29163402e-01 4.08963382e-01 7.80931264e-02 -5.77240705e-01
5.34131348e-01 4.02437955e-01 -1.04602647e+00 -4.82030883e-02
6.10608459e-01 1.57760763e+00 -3.23763043e-01 1.89690590e-01
-5.07744730e-01 1.04541659e+00 -9.56604600e-01 -8.22707593e-01
9.12436187e-01 -3.43094885e-01 -2.35067949e-01 -3.72046053e-01
-2.04903424e-01 -9.38721299e-01 -3.27268153e-01 1.09209073e+00
7.60180235e-01 5.60979061e-02 6.64664805e-01 -1.28263068e+00
-6.71436727e-01 6.72344625e-01 -6.56126738e-01 4.17367578e-01
-1.92696840e-01 1.66244432e-01 8.20280671e-01 -3.81697357e-01
1.45828232e-01 1.29633629e+00 7.32407868e-01 5.68956971e-01
-5.69110155e-01 -1.16627228e+00 -1.08541086e-01 9.19070393e-02
-8.57534111e-01 4.21637058e-01 5.77855647e-01 -4.56155121e-01
9.00048256e-01 -6.89938515e-02 1.05814469e+00 5.90554655e-01
8.03453565e-01 6.33912086e-01 9.10315394e-01 -1.73353493e-01
3.77656013e-01 4.13266681e-02 -2.85456538e-01 3.51743042e-01
4.02894974e-01 8.87085855e-01 -2.72867888e-01 -2.55905807e-01
1.07203460e+00 2.90601134e-01 5.24642646e-01 -8.77634659e-02
-3.99768412e-01 1.27498603e+00 2.78081864e-01 3.81573401e-02
-7.49247789e-01 3.40113640e-01 1.81101471e-01 6.41765475e-01
2.49964267e-01 6.68977797e-01 -6.03594363e-01 -5.57475649e-02
-7.30737805e-01 7.50479043e-01 7.82249331e-01 3.62732381e-01
1.54175103e-01 6.83285952e-01 -3.71096551e-01 4.42479908e-01
2.50284135e-01 5.61559796e-01 7.67381966e-01 -1.26134050e+00
2.82008141e-01 6.97795868e-01 2.23133773e-01 -9.73838508e-01
-5.07427275e-01 -4.63488430e-01 -6.36623263e-01 6.35452151e-01
3.43339562e-01 -8.97802591e-01 -4.66985762e-01 1.28615117e+00
2.21027173e-02 2.29066148e-01 1.73487037e-01 4.03994560e-01
5.15192330e-01 6.97733641e-01 -5.64653799e-02 -3.99316996e-01
1.07095945e+00 -6.20900989e-01 -8.16252708e-01 1.55696616e-01
-9.56586599e-02 -2.64469713e-01 2.43578762e-01 5.98116457e-01
-1.07526505e+00 -4.01842594e-01 -1.06961775e+00 9.28785205e-01
-3.95055354e-01 -4.12453949e-01 7.29584575e-01 1.01607645e+00
-8.33637238e-01 1.04327571e+00 -1.99887380e-01 2.07760736e-01
7.32092977e-01 7.79948592e-01 3.46344143e-01 6.41214252e-01
-1.84381509e+00 8.45940530e-01 8.01270127e-01 -1.55295447e-01
-7.57816970e-01 -5.47438264e-01 -3.18102866e-01 3.76173347e-01
4.18170780e-01 -3.83150041e-01 1.51663446e+00 -1.45466173e+00
-1.84252107e+00 1.91754356e-01 1.00554836e+00 -9.29410398e-01
6.72891259e-01 -1.10499710e-01 -4.67744082e-01 4.14735638e-02
-2.53863871e-01 5.92997313e-01 9.10400391e-01 -2.66307950e-01
-9.28804815e-01 -1.94367200e-01 1.78807989e-01 2.35083342e-01
-6.45090699e-01 1.52608991e-01 3.64170730e-01 -1.06121004e+00
-5.72977126e-01 -5.34904242e-01 -3.27829689e-01 -1.98671713e-01
1.65096432e-01 -4.95771199e-01 3.50367546e-01 -6.17210031e-01
1.23162758e+00 -1.99924874e+00 -4.32657033e-01 4.21259224e-01
2.44047716e-02 4.09756869e-01 -6.95691258e-02 3.39664996e-01
-2.04677805e-01 1.58121318e-01 1.09320264e-02 6.09569788e-01
2.30918035e-01 9.28357802e-03 -2.58969605e-01 1.06051154e-02
3.64444792e-01 1.07593918e+00 -7.15631247e-01 -8.13699439e-02
-2.16740295e-01 -2.49922678e-01 -4.43190217e-01 3.94587010e-01
-3.94686490e-01 -1.37557253e-01 -6.80401146e-01 6.10653579e-01
7.92372346e-01 1.34712895e-02 -1.02954507e-02 5.10841727e-01
-5.87295964e-02 -3.22859287e-01 -1.45364022e+00 6.74042404e-01
2.29039043e-01 2.92224228e-01 -4.92016196e-01 -1.12117910e+00
1.34861636e+00 3.20838928e-01 6.97718263e-01 -1.23317862e+00
2.43510306e-01 3.19779098e-01 2.29540095e-01 -5.02618372e-01
2.88136989e-01 -6.83766961e-01 6.91819191e-02 7.10684061e-01
-1.38274193e-01 2.05786303e-01 -1.14222966e-01 -3.65883976e-01
9.31937099e-01 2.57541668e-02 2.83320785e-01 -3.54784966e-01
2.05680266e-01 -2.97925174e-01 9.32955384e-01 6.86178148e-01
-5.36248803e-01 3.03354170e-02 1.02933145e+00 -8.23639214e-01
-8.42448473e-01 -1.02542186e+00 2.25651674e-02 8.28640878e-01
-1.44563287e-01 5.08956611e-01 -8.19492936e-01 -7.04636931e-01
7.41259515e-01 4.79239464e-01 -6.19660378e-01 -3.45934898e-01
-3.75482440e-01 -9.75186706e-01 4.69399631e-01 6.67121291e-01
1.00995934e+00 -1.73799717e+00 -7.84795761e-01 5.04394412e-01
7.39222527e-01 -2.37753168e-01 -1.89733028e-01 -1.23003557e-01
-8.96088183e-01 -1.02811098e+00 -1.11257112e+00 -4.91090983e-01
-1.60484365e-03 -6.14011347e-01 1.14837825e+00 -3.15946676e-02
1.51453868e-01 9.17838514e-02 -1.74179092e-01 -1.04611349e+00
-9.61650312e-02 -6.29230309e-03 1.68617427e-01 -1.77085195e-02
6.04428411e-01 -1.77270442e-01 -8.34896326e-01 2.31045298e-02
-8.74959052e-01 -7.01979220e-01 5.99807024e-01 9.00261343e-01
1.16580606e-01 4.56976771e-01 1.48405516e+00 -7.16470301e-01
1.07938206e+00 -6.37395263e-01 -1.18769455e+00 3.25042903e-01
-1.05035067e+00 1.00988351e-01 3.66805911e-01 -4.61964577e-01
-9.18799102e-01 -3.35849792e-01 -1.04480579e-01 -2.43500456e-01
6.76887810e-01 3.96957546e-01 -1.17621534e-01 -6.54596835e-02
2.39184231e-01 7.43368566e-02 4.74186748e-01 -1.82454914e-01
-2.25329529e-02 5.96062064e-01 7.53594935e-02 -1.48044556e-01
6.72671556e-01 -8.38729888e-02 1.56080320e-01 -1.57546699e-01
-3.91681612e-01 4.04193372e-01 -1.26168460e-01 -3.58228803e-01
7.79514074e-01 -5.08501351e-01 -1.26067376e+00 8.86974633e-01
-7.53009796e-01 -2.85191506e-01 -4.83315766e-01 4.08109963e-01
-6.26348495e-01 -7.72446319e-02 -6.04221940e-01 -1.12335598e+00
-6.09785676e-01 -7.37768173e-01 4.97756265e-02 7.85740614e-01
2.35182941e-01 -1.03558755e+00 2.13898554e-01 -3.98887545e-02
5.45153141e-01 5.48915923e-01 7.61458993e-01 -8.66934121e-01
-4.61999148e-01 -8.29832405e-02 -6.88153058e-02 6.14893854e-01
2.72061359e-02 -1.55849591e-01 -7.06541359e-01 -3.25842321e-01
1.77047834e-01 -3.18007171e-01 7.28836358e-01 6.69889808e-01
1.16685259e+00 -3.47316802e-01 2.80343473e-01 5.46150327e-01
1.55493605e+00 1.20035183e+00 7.74725854e-01 1.03286099e+00
1.51436597e-01 6.65037751e-01 7.97389328e-01 7.48374283e-01
7.22132996e-02 1.90174133e-01 5.17698348e-01 4.03528661e-02
7.74654686e-01 -8.20678547e-02 4.55471665e-01 1.39188394e-01
4.35098447e-02 8.62870589e-02 -7.92612493e-01 1.81072718e-03
-1.93677080e+00 -1.19605196e+00 8.06437731e-01 2.01497436e+00
6.68864906e-01 7.08001792e-01 7.30454028e-01 -5.11036180e-02
5.48546493e-01 -1.07093446e-01 -1.08948886e+00 -6.87267900e-01
-1.42588645e-01 4.80388343e-01 8.28097045e-01 1.74339227e-02
-1.00329530e+00 6.32792532e-01 7.22086859e+00 9.85728502e-01
-1.06018758e+00 -2.81199008e-01 1.15316176e+00 -2.60641873e-01
-3.46316338e-01 -5.13187468e-01 -7.89396703e-01 6.62147641e-01
1.00775695e+00 -5.64589262e-01 3.55389565e-01 7.39376366e-01
1.14573725e-01 1.64543226e-01 -5.82350433e-01 7.84770370e-01
-3.35312247e-01 -1.49976468e+00 1.33904675e-02 1.43327489e-01
8.17928016e-01 -3.76009673e-01 5.50062478e-01 7.92012155e-01
4.12966698e-01 -1.26589322e+00 7.39790678e-01 8.94546807e-01
3.15405697e-01 -1.44325757e+00 1.17122257e+00 3.22060525e-01
-8.42243314e-01 -7.93090522e-01 -5.28331995e-01 -4.30518180e-01
-2.04442158e-01 1.36987433e-01 -3.63291591e-01 3.86587173e-01
7.95337379e-01 3.78513992e-01 -3.26645911e-01 1.08053088e+00
1.02413177e-01 4.71022159e-01 -7.20513538e-02 -6.90650523e-01
5.16343653e-01 -4.24445122e-01 1.76173389e-01 6.01353586e-01
3.75121176e-01 2.96456873e-01 1.58396587e-02 8.38445306e-01
-1.93635568e-01 8.48337114e-02 -6.66512430e-01 -1.69051781e-01
4.18499053e-01 6.90438926e-01 -4.82540101e-01 -2.30517685e-01
-3.35795462e-01 4.12295491e-01 -8.45893696e-02 1.71829790e-01
-7.53582716e-01 -6.16039991e-01 4.53222632e-01 4.62315418e-03
4.64196622e-01 3.05879146e-01 -3.05860996e-01 -5.50018847e-01
-1.39922962e-01 -9.23207343e-01 5.66700041e-01 -3.26894432e-01
-1.41741538e+00 2.94270575e-01 -4.12289687e-02 -1.33293509e+00
-5.50078273e-01 -9.53499734e-01 -7.95986176e-01 9.59602714e-01
-1.70300674e+00 -2.92891443e-01 4.14884061e-01 4.00353491e-01
1.34987816e-01 -1.24304152e+00 4.32604253e-01 2.51745224e-01
-7.69893408e-01 8.96191478e-01 6.56721830e-01 4.62489486e-01
1.71549961e-01 -1.31888270e+00 2.72340298e-01 4.48162287e-01
-4.44594085e-01 7.05176070e-02 1.56766668e-01 -8.69189143e-01
-7.83086777e-01 -1.01193452e+00 3.05075645e-01 -5.35013229e-02
6.77762270e-01 2.35976115e-01 -6.51238143e-01 1.79301590e-01
5.09200752e-01 -3.98949444e-01 6.07185602e-01 -3.69389236e-01
1.94976777e-01 -8.53766561e-01 -1.38847077e+00 3.64319950e-01
3.25825900e-01 4.90188338e-02 -6.23738945e-01 -3.23927812e-02
8.87602806e-01 2.64126901e-03 -1.08884001e+00 2.53943235e-01
7.68376291e-01 -1.11545372e+00 9.78159070e-01 -9.46755052e-01
3.90832841e-01 3.10736716e-01 2.39559874e-01 -1.07703912e+00
-4.28317249e-01 -8.83036077e-01 -2.93564290e-01 1.10295570e+00
3.65508378e-01 -1.00704718e+00 1.16411936e+00 7.09853709e-01
5.29382586e-01 -1.05857837e+00 -1.17894185e+00 -1.10770977e+00
6.74746573e-01 3.35668445e-01 1.34650564e+00 7.87390113e-01
-3.19787472e-01 -1.02617584e-01 -3.30843925e-01 -4.92507488e-01
7.32257426e-01 -4.16682512e-02 2.05051824e-01 -1.22627246e+00
-3.76113534e-01 -8.95332038e-01 -2.19662488e-01 -3.39296162e-01
8.26165155e-02 -2.98006892e-01 -4.78702903e-01 -1.05921614e+00
-9.62926075e-02 -2.00630739e-01 -1.05986285e+00 3.69500667e-01
-1.14182457e-01 -2.66181499e-01 2.46644959e-01 -9.90176350e-02
-1.63182482e-01 6.47040606e-01 1.42115808e+00 -3.89995903e-01
-2.81344444e-01 4.01166856e-01 -7.78887987e-01 5.81202626e-01
1.17491281e+00 -4.47541744e-01 -3.45229834e-01 6.06760941e-02
4.26000237e-01 5.10676622e-01 -9.63786170e-02 -8.66904020e-01
-1.30593657e-01 -7.55833983e-01 8.21915090e-01 -4.49776381e-01
-2.37154886e-01 -6.73140228e-01 1.46595493e-01 1.22756088e+00
-7.65303075e-02 5.54153740e-01 2.88725019e-01 2.64151603e-01
-2.16546565e-01 -5.91902375e-01 7.58966327e-01 -3.61007363e-01
-6.76825583e-01 6.68217361e-01 -3.67486000e-01 8.97789747e-02
1.40336621e+00 -3.50248843e-01 -7.47169629e-02 -3.99830312e-01
-3.45038086e-01 5.88429213e-01 -2.22068608e-01 3.17178756e-01
7.84510493e-01 -1.68450153e+00 -7.39985108e-01 2.17320219e-01
-4.64928180e-01 -4.46293890e-01 1.43688601e-02 -2.07184553e-02
-1.00384486e+00 2.19885036e-01 -9.77518499e-01 3.40454847e-01
-4.14470822e-01 5.13980508e-01 1.16185427e+00 -5.29142916e-01
-1.61248758e-01 7.92847276e-01 1.04224086e-02 -3.01081866e-01
3.27745885e-01 -3.06039490e-02 -9.59714293e-01 3.86318028e-01
6.48573041e-01 6.98567390e-01 -2.25123033e-01 1.39096648e-01
-1.10065646e-01 4.53395158e-01 -1.08686574e-02 -2.85522431e-01
1.55412245e+00 4.91581917e-01 -1.68576077e-01 6.31141737e-02
6.21959090e-01 -6.43632770e-01 -1.52857590e+00 1.49277851e-01
4.92467284e-01 -2.59809524e-01 -2.24480499e-02 -8.44236791e-01
-1.43805707e+00 6.61922872e-01 1.10852289e+00 4.77770954e-01
1.27430165e+00 -8.75469625e-01 9.35154855e-01 4.79898751e-01
1.06545024e-01 -1.55588830e+00 3.68315965e-01 4.74051118e-01
7.36835897e-01 -9.58320737e-01 -2.67876163e-02 5.74904025e-01
-5.25024474e-01 1.22643006e+00 8.16792786e-01 -7.60255396e-01
1.10344124e+00 1.68316811e-01 2.31090128e-01 -2.56504919e-02
-6.04888797e-01 1.82528496e-01 2.12404653e-02 5.22410750e-01
-2.46375620e-01 1.56426728e-01 -4.23436016e-01 1.05364490e+00
-2.28955016e-01 3.51743758e-01 5.61332285e-01 7.25203753e-01
-5.93086183e-01 -1.19065559e+00 -4.49862063e-01 9.06840026e-01
-7.22460449e-01 6.93968916e-03 -2.79903919e-01 7.52257645e-01
3.03205848e-01 4.80392873e-01 5.06751895e-01 -3.41829598e-01
5.90406060e-01 -1.91113546e-01 2.19082579e-01 -2.01552019e-01
-1.11694431e+00 -4.89942990e-02 -4.75521296e-01 -3.56767505e-01
-1.31018505e-01 -7.07484484e-01 -1.23043430e+00 -4.95049983e-01
2.49534212e-02 1.70886189e-01 7.69302100e-02 7.98014998e-01
-3.45688388e-02 5.40710926e-01 1.05896819e+00 -1.39889076e-01
-1.32573330e+00 -7.61187911e-01 -1.30277073e+00 6.40368983e-02
2.99661130e-01 -8.26150894e-01 -3.68363783e-02 -7.04687238e-01] | [4.473752975463867, 3.9802324771881104] |
c513703b-cdf4-43da-a5f4-2d201cea71b0 | ernie-layout-layout-knowledge-enhanced-pre | 2210.06155 | null | https://arxiv.org/abs/2210.06155v2 | https://arxiv.org/pdf/2210.06155v2.pdf | ERNIE-Layout: Layout Knowledge Enhanced Pre-training for Visually-rich Document Understanding | Recent years have witnessed the rise and success of pre-training techniques in visually-rich document understanding. However, most existing methods lack the systematic mining and utilization of layout-centered knowledge, leading to sub-optimal performances. In this paper, we propose ERNIE-Layout, a novel document pre-training solution with layout knowledge enhancement in the whole workflow, to learn better representations that combine the features from text, layout, and image. Specifically, we first rearrange input sequences in the serialization stage, and then present a correlative pre-training task, reading order prediction, to learn the proper reading order of documents. To improve the layout awareness of the model, we integrate a spatial-aware disentangled attention into the multi-modal transformer and a replaced regions prediction task into the pre-training phase. Experimental results show that ERNIE-Layout achieves superior performance on various downstream tasks, setting new state-of-the-art on key information extraction, document image classification, and document question answering datasets. The code and models are publicly available at http://github.com/PaddlePaddle/PaddleNLP/tree/develop/model_zoo/ernie-layout. | ['Haifeng Wang', 'Hua Wu', 'Hao Tian', 'Yu Sun', 'Shikun Feng', 'Yin Zhang', 'Yongfeng Chen', 'Weichong Yin', 'Teng Hu', 'Zhengjie Huang', 'Zhenyu Zhang', 'Bin Luo', 'Wenjin Wang', 'Yinxu Pan', 'Qiming Peng'] | 2022-10-12 | null | null | null | null | ['document-image-classification', 'semantic-entity-labeling', 'key-information-extraction'] | ['computer-vision', 'natural-language-processing', 'natural-language-processing'] | [ 3.51637810e-01 -2.78754741e-01 -9.41498280e-02 -3.87549788e-01
-8.30068409e-01 -8.06168318e-01 6.41851306e-01 2.73718625e-01
-2.15967387e-01 1.70781583e-01 4.13339823e-01 -6.44649386e-01
-4.73515958e-01 -6.03814363e-01 -7.40965724e-01 -5.34349620e-01
3.33162636e-01 3.61079961e-01 9.31848660e-02 1.29970923e-01
6.29156530e-01 2.97413319e-01 -1.36428428e+00 1.02949965e+00
1.00748765e+00 7.74617910e-01 7.97097445e-01 9.92320538e-01
-4.76065487e-01 9.50204372e-01 -3.85627687e-01 -2.08474576e-01
-1.12470947e-01 -3.60968739e-01 -9.57822740e-01 1.03230163e-01
5.40160716e-01 -3.98540169e-01 -3.81527275e-01 6.90777063e-01
5.70231855e-01 1.79417972e-02 5.27391374e-01 -8.66565526e-01
-1.24127972e+00 6.06114447e-01 -9.20771420e-01 3.46462607e-01
2.89533049e-01 1.11434184e-01 1.18320441e+00 -1.11470616e+00
5.91965437e-01 1.17933691e+00 1.86937347e-01 2.88961887e-01
-9.63137805e-01 -4.89960402e-01 6.80395782e-01 6.99696243e-01
-1.17316735e+00 -4.60519344e-01 8.96090150e-01 -4.94175732e-01
9.21168268e-01 3.70372564e-01 4.67969269e-01 1.14372122e+00
-1.42586470e-01 1.47627819e+00 1.02321732e+00 -7.80000687e-01
-2.40260497e-01 1.32415548e-01 5.74247777e-01 9.26962495e-01
1.03771821e-01 -1.54302388e-01 -6.37221158e-01 4.99544114e-01
5.23212492e-01 1.10736035e-01 -4.52858955e-01 -4.11708057e-01
-1.04827058e+00 4.57486212e-01 5.40335715e-01 4.23689246e-01
-6.95003346e-02 -1.75586075e-01 4.11703765e-01 -1.09121822e-01
2.33649284e-01 4.97831315e-01 -4.76332128e-01 2.91944351e-02
-8.98790836e-01 1.55528128e-01 3.32388133e-01 1.02256227e+00
5.06248355e-01 -4.84227926e-01 -6.50884390e-01 8.40399265e-01
3.69088411e-01 3.61182839e-01 3.84558201e-01 -4.00976539e-01
1.04463565e+00 8.72435331e-01 -2.52524763e-01 -9.21204746e-01
-5.81104934e-01 -5.98725259e-01 -7.98084021e-01 -3.42261285e-01
4.67156798e-01 2.34973684e-01 -1.00675631e+00 1.28435040e+00
1.99666932e-01 -1.83292121e-01 -1.49483860e-01 9.04877841e-01
8.67874503e-01 7.46119440e-01 -6.31586313e-02 2.72582948e-01
1.68267429e+00 -1.57354510e+00 -6.29767239e-01 -4.38689619e-01
8.05705130e-01 -8.75935972e-01 1.52723658e+00 6.12584412e-01
-9.41276073e-01 -7.75589347e-01 -1.02306962e+00 -6.76671982e-01
-6.94239438e-01 6.91727519e-01 5.83888412e-01 3.64025354e-01
-6.10362232e-01 2.31598765e-01 -5.96989572e-01 -2.86810637e-01
7.95862257e-01 4.10237983e-02 -2.45065987e-01 -7.43413448e-01
-7.17914343e-01 6.49996877e-01 4.68503743e-01 4.06608850e-01
-5.32113433e-01 -7.32720673e-01 -6.27293348e-01 2.05845401e-01
6.59401119e-01 -7.01739788e-01 1.04389966e+00 -5.22731006e-01
-1.19696486e+00 7.32770503e-01 -3.26609671e-01 9.37343687e-02
2.93983012e-01 -5.62339962e-01 -2.42163524e-01 1.22092254e-01
-3.13536904e-04 6.48561418e-01 7.06270814e-01 -1.60329700e+00
-4.92201209e-01 -6.78718865e-01 5.07819373e-03 3.18244815e-01
-5.23057401e-01 -1.80017516e-01 -1.07295918e+00 -4.31231052e-01
-8.27764645e-02 -4.91937369e-01 2.20003739e-01 -2.32665628e-01
-8.95016909e-01 -1.42044991e-01 9.04733062e-01 -1.06308889e+00
1.41748381e+00 -2.04696226e+00 2.96788841e-01 1.49844855e-01
9.82788652e-02 2.99751014e-01 -6.25021458e-01 5.36754549e-01
-1.26203284e-01 1.22460775e-01 1.12962022e-01 -5.25798976e-01
6.05139323e-02 -3.93215194e-02 -3.23912084e-01 8.74779522e-02
1.80116951e-01 1.33564746e+00 -7.05217540e-01 -5.45774817e-01
3.54776412e-01 3.70310903e-01 -5.06436706e-01 2.02781156e-01
-5.01866162e-01 3.90815794e-01 -5.13337135e-01 6.89149082e-01
7.66135097e-01 -7.31594563e-01 3.31954449e-01 -4.90406126e-01
-1.69672728e-01 4.20221895e-01 -8.76621008e-01 2.13013268e+00
-4.82864946e-01 7.33165979e-01 -1.89823672e-01 -9.98474836e-01
6.40124142e-01 -2.03741983e-01 9.67798159e-02 -1.27252781e+00
8.91632512e-02 -2.39131257e-01 -8.48011151e-02 -8.43000352e-01
5.72703838e-01 7.41335213e-01 1.72211677e-01 5.60284257e-01
-2.96613537e-02 3.57687920e-01 4.26661104e-01 4.45105195e-01
9.49788809e-01 4.57648665e-01 8.23356211e-02 8.68029892e-03
5.29926598e-01 2.95498013e-03 -1.51353031e-01 6.90220058e-01
2.84624696e-01 5.42628050e-01 6.05674267e-01 -1.21911019e-01
-9.45366085e-01 -7.97788024e-01 8.40637386e-02 1.49091482e+00
2.25557148e-01 -7.29775190e-01 -6.65932119e-01 -6.97740495e-01
-4.40828614e-02 7.17427671e-01 -6.95998073e-01 -1.26998713e-02
-7.06629515e-01 -6.22557163e-01 5.03187835e-01 8.61342430e-01
5.73362768e-01 -9.27275240e-01 -4.92920369e-01 -1.78066224e-01
-2.28479266e-01 -1.09420311e+00 -4.92259204e-01 3.18185240e-01
-6.27847791e-01 -1.18823659e+00 -5.88711679e-01 -8.22581291e-01
9.67438877e-01 4.96917605e-01 9.19544101e-01 3.81995976e-01
-5.45738161e-01 2.99607247e-01 -4.88214940e-01 -2.92014748e-01
2.06769973e-01 4.65946645e-01 -8.05399358e-01 -2.21945077e-01
2.57184058e-01 -1.91603214e-01 -8.16894889e-01 7.16180056e-02
-8.28017712e-01 6.55166268e-01 9.61803555e-01 9.08690691e-01
5.32578588e-01 -2.80346759e-02 1.46791399e-01 -9.37123299e-01
6.98639512e-01 -2.16523096e-01 -5.36982596e-01 7.84715056e-01
-6.15422785e-01 2.45258331e-01 6.90043926e-01 -2.15122357e-01
-1.21303606e+00 -1.50354967e-01 -2.61384863e-02 -2.55139530e-01
-3.87167305e-01 6.93096638e-01 -6.44532740e-01 4.81303334e-01
3.08463573e-01 4.53729302e-01 -4.21711713e-01 -9.12481427e-01
7.34927058e-01 6.15267634e-01 5.83112895e-01 -7.04131067e-01
7.33741343e-01 3.39374661e-01 -3.03794354e-01 -6.74906194e-01
-1.20924032e+00 -6.32968426e-01 -9.84848082e-01 -9.79143679e-02
7.75115550e-01 -6.18021131e-01 -7.56456435e-01 4.41398174e-01
-1.26661026e+00 -3.77839893e-01 -4.80362587e-02 -6.54538050e-02
-1.57744929e-01 4.37113285e-01 -3.25056434e-01 -6.75161541e-01
-3.63696337e-01 -1.01232588e+00 1.23569214e+00 3.65829676e-01
9.82949808e-02 -8.24375093e-01 -6.87450171e-02 8.33683074e-01
9.50715244e-02 -2.90670693e-01 1.39723611e+00 -6.25396192e-01
-1.07239342e+00 9.26791728e-02 -8.87568057e-01 1.63146198e-01
5.28293811e-02 -8.36872607e-02 -1.07050252e+00 -1.58584654e-01
-5.91618657e-01 -2.71123558e-01 1.30356348e+00 2.89934069e-01
1.78791988e+00 -3.38809758e-01 -5.20073652e-01 7.97586441e-01
1.43570387e+00 2.39623904e-01 6.75048470e-01 5.85157156e-01
1.19557905e+00 7.69036889e-01 4.76651162e-01 4.97350127e-01
7.68205285e-01 4.81581837e-01 3.49783152e-01 -2.04836711e-01
-3.60787362e-01 -5.53755581e-01 3.99466902e-02 7.53615618e-01
2.00892538e-01 -6.95097327e-01 -1.13144720e+00 4.89910722e-01
-1.89424479e+00 -9.15996969e-01 -8.81375894e-02 1.65364432e+00
5.89075804e-01 4.75978106e-02 -1.57075837e-01 2.49730051e-01
3.95277739e-01 3.17262411e-01 -6.18993640e-01 -1.95930734e-01
-1.40318826e-01 3.87272872e-02 2.70416021e-01 3.90951693e-01
-1.13500583e+00 1.12288499e+00 4.57306051e+00 8.08013618e-01
-9.15970445e-01 -6.68237433e-02 5.46088994e-01 -7.70386308e-02
-5.04532993e-01 5.29803010e-03 -1.03392422e+00 3.13694417e-01
5.21309018e-01 2.20033675e-01 3.85405689e-01 4.20511931e-01
2.35588759e-01 -2.88997203e-01 -1.20255959e+00 8.30896020e-01
3.47281992e-01 -1.53027642e+00 3.51453841e-01 2.06631655e-03
3.64571869e-01 -3.85262221e-01 3.23881596e-01 2.23668069e-01
-4.63901460e-02 -1.18320763e+00 7.64711201e-01 7.67130673e-01
6.17407084e-01 -6.46510482e-01 3.56493920e-01 4.96548653e-01
-1.27620804e+00 -2.93778777e-01 -1.23708688e-01 2.85246611e-01
9.04190093e-02 4.60021406e-01 -6.05907142e-01 8.99024904e-01
6.98711097e-01 8.86670589e-01 -1.07512772e+00 1.02622640e+00
-4.14841890e-01 4.06662315e-01 2.21398473e-02 -9.09630209e-02
2.93386072e-01 -6.74713254e-02 1.27431005e-01 1.42652488e+00
8.53590146e-02 -3.79497977e-03 -4.31711152e-02 8.25665951e-01
-4.43201996e-02 2.33481124e-01 -2.13229328e-01 -2.97987163e-01
3.04782003e-01 1.23355925e+00 -7.07471073e-01 -1.59448951e-01
-5.79841554e-01 1.02422833e+00 4.97048169e-01 4.16461527e-01
-7.03545153e-01 -4.71147567e-01 1.68652609e-01 -1.78228058e-02
6.18638813e-01 -2.99027979e-01 -4.63600099e-01 -1.15163040e+00
3.08896989e-01 -1.02721691e+00 3.78997058e-01 -9.85518515e-01
-9.77532744e-01 3.74461114e-01 -7.20213205e-02 -8.07766080e-01
1.94316003e-02 -1.00281823e+00 -5.16374052e-01 7.60225594e-01
-1.72245133e+00 -1.68809474e+00 -5.27246594e-01 4.57947880e-01
8.98115396e-01 -7.63678700e-02 7.07880914e-01 2.93345153e-01
-1.00155270e+00 6.75332785e-01 2.92648017e-01 3.17979276e-01
6.57608688e-01 -1.19874346e+00 3.07060927e-01 9.61172640e-01
4.78422761e-01 7.14098990e-01 1.75003812e-01 -4.18904692e-01
-1.60947192e+00 -8.53096187e-01 6.71951771e-01 -4.86283213e-01
5.18186748e-01 -7.81683922e-01 -1.00068188e+00 6.01022124e-01
4.40599591e-01 -3.55072886e-01 7.21009433e-01 3.63510251e-01
-7.24135637e-01 2.06826888e-02 -3.43253195e-01 6.91296518e-01
1.09084165e+00 -6.75807893e-01 -5.11032522e-01 3.22377861e-01
5.38809419e-01 -3.51073533e-01 -6.06036365e-01 1.91429079e-01
7.25630283e-01 -8.17611516e-01 1.10226464e+00 -5.70898652e-01
7.31660962e-01 -2.01688915e-01 -2.22947691e-02 -9.94826913e-01
-4.02740538e-01 -1.82174906e-01 -3.22919458e-01 1.37977111e+00
4.68980104e-01 -2.97975447e-03 7.29642868e-01 -1.18419891e-02
-2.15886220e-01 -1.11482430e+00 -3.88655961e-01 -4.98408318e-01
-3.33964042e-02 -4.40738618e-01 5.68928480e-01 6.04346812e-01
5.84520847e-02 7.77481496e-01 -3.04613948e-01 3.28361273e-01
3.76801729e-01 4.46798176e-01 8.23192060e-01 -9.65486348e-01
-4.63808000e-01 -6.10457003e-01 2.90064007e-01 -1.74437237e+00
-6.91840844e-03 -9.82678473e-01 -8.63647014e-02 -2.17400670e+00
5.04554808e-01 -1.29739180e-01 -3.76309782e-01 6.15145445e-01
-3.16112757e-01 -3.48051131e-01 4.25231248e-01 1.50069043e-01
-1.00597978e+00 4.65835363e-01 1.58279169e+00 -4.60193634e-01
-1.32201239e-01 -3.00719410e-01 -9.82022226e-01 5.42810082e-01
6.41553164e-01 -7.37810433e-02 -6.41094565e-01 -9.65926409e-01
2.74627179e-01 -2.30106950e-01 5.64031720e-01 -7.56259501e-01
4.52204317e-01 1.31383106e-01 9.23804343e-01 -1.29912972e+00
2.00204432e-01 -6.77577138e-01 -6.21382356e-01 8.94819871e-02
-6.52264714e-01 1.39018729e-01 4.58862424e-01 4.63722676e-01
-1.19329423e-01 -3.91546607e-01 3.91837925e-01 6.12046383e-03
-1.06569731e+00 7.83037990e-02 -2.18624510e-02 -4.72742878e-02
6.76525712e-01 -1.65073112e-01 -9.95467901e-01 6.26122728e-02
-4.15833622e-01 4.51178461e-01 1.44278333e-01 6.82266474e-01
8.35965037e-01 -7.61041284e-01 -3.97580296e-01 2.80508727e-01
3.72591108e-01 1.48826569e-01 6.10185266e-01 7.52473116e-01
-3.70609492e-01 8.95648122e-01 -8.53557065e-02 -4.94297564e-01
-1.31279349e+00 8.05029154e-01 8.88227224e-02 -7.15595126e-01
-6.97406948e-01 9.97031212e-01 3.49076688e-01 -3.05560291e-01
4.21253622e-01 -4.26129103e-01 -3.89324605e-01 1.49117440e-01
6.23599470e-01 2.06583947e-01 1.52311876e-01 -2.30039880e-01
-4.19515699e-01 6.84316993e-01 -6.42114460e-01 8.28969702e-02
1.20988405e+00 -2.54682183e-01 2.40024235e-02 1.94770843e-01
1.23934138e+00 -4.44296859e-02 -1.17315352e+00 -1.73132464e-01
3.47554535e-01 -4.50672686e-01 9.35606137e-02 -1.20156610e+00
-8.62246215e-01 1.14950967e+00 4.81609374e-01 -6.44939616e-02
1.34486353e+00 1.09379642e-01 5.55879354e-01 5.33257067e-01
-2.12999299e-01 -9.68150139e-01 4.16181684e-01 6.76864386e-01
1.00246835e+00 -1.20407474e+00 2.33470201e-01 -4.38470989e-01
-5.48520446e-01 9.91796970e-01 8.87948394e-01 1.98016182e-01
5.44503450e-01 2.79434502e-01 -8.45784321e-02 -2.68532604e-01
-9.26449299e-01 -3.06036115e-01 7.35516906e-01 4.40612525e-01
5.76444149e-01 -2.26344749e-01 1.92007676e-01 8.93459916e-01
-2.20042709e-02 -2.06866011e-01 -4.50526364e-02 1.00424874e+00
-2.67197967e-01 -1.24765861e+00 -2.76447445e-01 5.29377520e-01
-5.80039457e-04 -4.13809299e-01 -4.24807101e-01 8.45101476e-01
1.73771232e-01 7.42557943e-01 1.22048356e-01 -3.12261313e-01
5.50171912e-01 1.88387990e-01 7.17737913e-01 -5.78994989e-01
-3.25401425e-01 1.36556417e-01 -5.73549569e-02 -5.48701406e-01
-1.47166386e-01 -4.82064813e-01 -1.17416608e+00 -5.39097823e-02
-4.27258879e-01 -1.06596485e-01 6.60402000e-01 1.00574815e+00
6.13906622e-01 9.70848143e-01 2.46612400e-01 -6.21598423e-01
-2.08090723e-01 -9.40498412e-01 -2.31353849e-01 1.54253140e-01
2.36344591e-01 -4.80590850e-01 5.52938282e-02 2.25679442e-01] | [11.557393074035645, 2.3175253868103027] |
77d550ab-60fd-44cc-b617-ea9997d0e871 | unsupervised-constrative-person-re | 2010.07608 | null | https://arxiv.org/abs/2010.07608v2 | https://arxiv.org/pdf/2010.07608v2.pdf | Fully Unsupervised Person Re-identification viaSelective Contrastive Learning | Person re-identification (ReID) aims at searching the same identity person among images captured by various cameras. Unsupervised person ReID attracts a lot of attention recently, due to it works without intensive manual annotation and thus shows great potential of adapting to new conditions. Representation learning plays a critical role in unsupervised person ReID. In this work, we propose a novel selective contrastive learning framework for unsupervised feature learning. Specifically, different from traditional contrastive learning strategies, we propose to use multiple positives and adaptively sampled negatives for defining the contrastive loss, enabling to learn a feature embedding model with stronger identity discriminative representation. Moreover, we propose to jointly leverage global and local features to construct three dynamic dictionaries, among which the global and local memory banks are used for pairwise similarity computation and the mixture memory bank are used for contrastive loss definition. Experimental results demonstrate the superiority of our method in unsupervised person ReID compared with the state-of-the-arts. | ['Xianming Liu', 'Junjun Jiang', 'Deming Zhai', 'Bo Pang'] | 2020-10-15 | null | null | null | null | ['unsupervised-person-re-identification'] | ['computer-vision'] | [ 1.58033833e-01 -6.12727821e-01 -1.06832065e-01 -5.53071201e-01
-4.51684088e-01 -3.07882518e-01 7.33865499e-01 2.40092441e-01
-8.14273238e-01 5.51582098e-01 2.22390503e-01 4.84232545e-01
-3.69850844e-01 -5.66320479e-01 -3.01548719e-01 -9.50077236e-01
-1.11139249e-02 3.76756608e-01 -2.30776578e-01 1.74864873e-01
2.20221534e-01 5.48713923e-01 -1.64602792e+00 -3.89305443e-01
8.57250869e-01 8.03961098e-01 5.70108779e-02 1.31854162e-01
1.09615829e-03 4.69737977e-01 -2.90704280e-01 -7.29834616e-01
5.27909994e-01 -4.37928617e-01 -4.71512496e-01 -2.69015040e-02
7.02649891e-01 -2.93848127e-01 -5.80553174e-01 1.19808424e+00
9.42991078e-01 6.37924433e-01 8.03596556e-01 -1.06421304e+00
-7.12523222e-01 3.40764463e-01 -8.71157944e-01 6.10166728e-01
5.43395698e-01 4.83415388e-02 9.20499861e-01 -9.80555475e-01
3.99486989e-01 1.30935204e+00 6.55784190e-01 6.87853098e-01
-1.29871905e+00 -1.16085982e+00 3.43283474e-01 6.26756608e-01
-1.94068134e+00 -6.79234326e-01 9.95961666e-01 -3.69744241e-01
4.83320028e-01 1.85212240e-01 5.46741366e-01 1.03238642e+00
-2.82927513e-01 8.18348944e-01 1.08441103e+00 -4.12764370e-01
-5.89937158e-02 2.88784921e-01 1.89504743e-01 6.58130825e-01
2.64288306e-01 2.39615634e-01 -7.21156180e-01 -2.28385970e-01
6.53983176e-01 2.87555486e-01 -2.81664729e-01 -4.52862859e-01
-1.07496047e+00 5.72641313e-01 3.92940491e-01 2.30834112e-01
-2.76074052e-01 -4.33898680e-02 4.76270854e-01 2.44467840e-01
3.39513749e-01 1.06908515e-01 4.07265238e-02 -5.01722954e-02
-8.26131701e-01 2.57973015e-01 3.59642178e-01 9.16265249e-01
9.11495507e-01 -2.65632987e-01 -9.32657868e-02 1.13051963e+00
1.29580602e-01 3.98838162e-01 8.39563251e-01 -5.30301213e-01
3.20231587e-01 6.67342961e-01 -4.61986624e-02 -1.20936227e+00
-3.70745212e-01 -5.21269858e-01 -1.13375366e+00 -2.48878017e-01
2.39628166e-01 6.54432252e-02 -4.68389899e-01 1.74572253e+00
4.81861442e-01 6.43913329e-01 -1.07289694e-01 8.26110363e-01
5.32320678e-01 3.90284538e-01 2.54447550e-01 -1.41380936e-01
1.26505685e+00 -6.61577523e-01 -5.56393683e-01 -7.22992420e-02
1.92309305e-01 -6.33328080e-01 5.90047359e-01 1.14068203e-01
-8.94162774e-01 -8.69137704e-01 -9.73936379e-01 2.92834360e-02
-4.07128572e-01 1.64145548e-02 4.93100703e-01 6.88377619e-01
-8.50055873e-01 3.06235820e-01 -5.65499902e-01 -4.65070337e-01
4.37576145e-01 6.80884540e-01 -5.97447157e-01 -1.95009455e-01
-9.81934547e-01 6.64161921e-01 4.25221980e-01 1.78371549e-01
-4.68130827e-01 -4.72641438e-01 -1.07707119e+00 9.77879539e-02
8.80126879e-02 -8.09451163e-01 6.23196661e-01 -8.65678668e-01
-1.50160170e+00 1.07216609e+00 -3.29247266e-01 -4.18629706e-01
6.60726786e-01 -2.08733812e-01 -4.29264665e-01 6.79767430e-02
2.54434139e-01 5.38653731e-01 1.11455381e+00 -1.35729873e+00
-7.28512824e-01 -6.04442894e-01 -1.99947596e-01 3.07061940e-01
-8.77989411e-01 1.77069858e-01 -6.14353597e-01 -8.56526017e-01
6.61342219e-02 -8.01261246e-01 -2.82194708e-02 -2.04035252e-01
-1.85249090e-01 -5.51642835e-01 6.18058324e-01 -6.37878358e-01
1.15213275e+00 -2.11606359e+00 1.87693179e-01 5.64164162e-01
2.39285111e-01 1.43414751e-01 -2.83188373e-02 1.21713318e-01
-5.13172746e-02 -2.25425422e-01 -9.01703387e-02 -7.58224785e-01
7.17940927e-02 -1.26491651e-01 8.55568871e-02 8.57632935e-01
5.13656214e-02 6.64477527e-01 -9.82426703e-01 -6.95841014e-01
3.54860723e-01 6.10965967e-01 -3.54593247e-01 2.69792020e-01
7.13773549e-01 7.09505260e-01 -5.41785479e-01 5.30232131e-01
8.76724422e-01 -3.27162072e-02 8.83865133e-02 -1.89062878e-01
-1.11125246e-01 -1.54196605e-01 -1.39490616e+00 1.69759262e+00
-2.65073508e-01 1.20646186e-01 -1.69965863e-01 -1.12928486e+00
1.11045206e+00 1.40014095e-02 4.64872867e-01 -7.91232526e-01
2.84834236e-01 -6.84777554e-03 -3.15881521e-01 -2.86274105e-01
4.61639553e-01 -3.54185216e-02 -1.18812986e-01 3.56185466e-01
1.61247417e-01 5.36430061e-01 1.11304715e-01 -5.54090291e-02
6.12200439e-01 -8.37103054e-02 3.28633606e-01 -3.06455016e-01
1.17280817e+00 -5.99753857e-01 8.71160746e-01 7.08139360e-01
-4.05983061e-01 4.10658389e-01 -1.63943365e-01 -5.32579660e-01
-7.51482904e-01 -1.12432671e+00 -2.33072385e-01 1.04057276e+00
7.84522951e-01 -2.71379799e-01 -6.26861632e-01 -7.45901644e-01
1.22733541e-01 2.38403618e-01 -5.60033321e-01 -2.77273208e-01
-8.68873656e-01 -6.79240644e-01 4.51378345e-01 5.81196010e-01
8.37785244e-01 -7.29423106e-01 -1.22875452e-01 3.57893370e-02
-1.03507280e-01 -8.81855667e-01 -1.00350332e+00 -3.90453935e-01
-5.38347006e-01 -1.09209859e+00 -9.76138055e-01 -1.15015304e+00
8.80781651e-01 6.30656004e-01 5.92117965e-01 -4.76784036e-02
-4.07697171e-01 7.87766039e-01 -2.01488152e-01 -1.01729050e-01
8.74842182e-02 1.11544877e-01 6.86210454e-01 6.84919596e-01
5.67223310e-01 -7.11265147e-01 -8.14186811e-01 4.19124246e-01
-7.49011397e-01 -2.70361841e-01 6.33741677e-01 1.11988223e+00
6.46019280e-01 9.31932479e-02 8.17618906e-01 -7.17470348e-01
5.79760313e-01 -4.21772659e-01 -5.85249007e-01 2.50733197e-01
-7.15283930e-01 -7.49182180e-02 7.18933403e-01 -4.60379422e-01
-1.23571873e+00 2.29505692e-02 2.05032453e-01 -3.92995089e-01
-5.48360907e-02 9.55833569e-02 -4.49927717e-01 -2.80706942e-01
2.35842615e-01 6.18618548e-01 -8.30832273e-02 -5.66407859e-01
3.07031721e-01 5.92699766e-01 8.01634431e-01 -6.57272398e-01
1.25157964e+00 6.38097703e-01 -2.49597043e-01 -6.04392290e-01
-3.25633198e-01 -1.00740695e+00 -8.42592895e-01 -2.34120101e-01
7.22831905e-01 -1.15131557e+00 -1.03221512e+00 6.33921564e-01
-7.85136163e-01 3.49878162e-01 -1.70689493e-01 5.65082014e-01
-3.28351378e-01 6.77339256e-01 -3.73029411e-01 -8.62574339e-01
-5.96434832e-01 -7.84162283e-01 8.75082970e-01 7.36495197e-01
-1.03874788e-01 -9.03260708e-01 1.05203770e-01 4.22301203e-01
2.51924902e-01 1.08344257e-02 5.05389690e-01 -6.92549229e-01
-3.37621987e-01 -4.22732860e-01 -4.34818298e-01 3.16454589e-01
4.67253447e-01 -7.18056440e-01 -1.06363499e+00 -7.66348839e-01
-1.68823004e-01 -1.50158003e-01 9.37193632e-01 -6.16383515e-02
1.20968771e+00 -3.63239646e-01 -4.80099499e-01 8.51963580e-01
1.25520611e+00 3.48393782e-03 3.37249756e-01 4.24602509e-01
1.02446461e+00 5.31447232e-01 4.11238819e-01 6.50932848e-01
6.25761628e-01 8.22216868e-01 -1.55515373e-01 1.36332005e-01
8.32851827e-02 -3.77388000e-01 2.33006284e-01 8.67137015e-01
-2.87283838e-01 1.60077557e-01 -6.00358725e-01 6.26066208e-01
-1.88162315e+00 -1.24457383e+00 5.53737462e-01 2.56695914e+00
7.75458634e-01 -1.30757123e-01 2.94429362e-01 4.42751199e-02
1.07293367e+00 1.96448281e-01 -6.37501538e-01 2.08405450e-01
-2.22422063e-01 1.30465537e-01 4.90504324e-01 1.33845404e-01
-1.39861178e+00 7.96143532e-01 4.89244747e+00 9.13425326e-01
-8.64404917e-01 8.54818150e-02 2.75240600e-01 1.20044731e-01
1.46836583e-02 -1.65768102e-01 -8.74574184e-01 6.72830760e-01
4.29280102e-01 -3.67805451e-01 2.91302621e-01 7.01837897e-01
1.34197757e-01 1.93927690e-01 -1.19090033e+00 1.65144944e+00
4.29348201e-01 -8.79871428e-01 2.37045854e-01 -4.87267449e-02
7.56402850e-01 -5.70297539e-01 3.15345198e-01 1.99179515e-01
-1.56927742e-02 -7.28845000e-01 5.78182697e-01 7.77584016e-01
6.10667348e-01 -1.08698773e+00 8.15044820e-01 1.43060654e-01
-1.48861635e+00 -1.56311572e-01 -4.50874776e-01 2.27228954e-01
4.97551374e-02 1.63141236e-01 -4.68913347e-01 6.14315450e-01
6.97250366e-01 1.01017404e+00 -5.97639561e-01 1.27506101e+00
3.72355543e-02 1.87644362e-01 -1.29312843e-01 1.98606029e-01
-1.91174001e-01 -2.73665220e-01 5.78222156e-01 1.41511500e+00
6.18108250e-02 1.16056703e-01 4.54100758e-01 6.38990164e-01
-3.82434011e-01 3.28841716e-01 -3.31792861e-01 3.35351348e-01
6.77212596e-01 1.10468781e+00 -2.87954479e-01 -2.34526187e-01
-4.42485690e-01 1.45204425e+00 4.14709955e-01 4.00215238e-01
-5.73973656e-01 -3.54525030e-01 8.90232027e-01 -1.18685395e-01
2.87296832e-01 -2.25667730e-01 1.62982821e-01 -1.34920764e+00
9.40200686e-02 -6.76835179e-01 7.41629183e-01 2.47153286e-02
-1.94012547e+00 2.03237191e-01 9.07513648e-02 -1.40503824e+00
-9.33809578e-02 -2.22649097e-01 -6.09191597e-01 1.02117240e+00
-1.87126398e+00 -1.44795978e+00 -4.96724755e-01 9.57333207e-01
4.05452728e-01 -4.26987469e-01 7.46498644e-01 6.36263549e-01
-8.56955528e-01 1.34151518e+00 2.60055512e-01 5.02318799e-01
9.99876022e-01 -1.04125321e+00 2.24868372e-01 7.97884464e-01
-3.82686332e-02 1.03610837e+00 4.31270152e-01 -4.31138784e-01
-1.58788610e+00 -1.10620856e+00 7.82340050e-01 -1.57397300e-01
3.65711898e-01 -2.58804649e-01 -8.35144222e-01 4.10196900e-01
-1.37291774e-01 -6.22250279e-03 1.03395081e+00 5.14642633e-02
-6.12680554e-01 -6.03132188e-01 -1.30666709e+00 3.71854275e-01
1.22108042e+00 -6.73428059e-01 -5.02178848e-01 4.01742235e-02
1.27492055e-01 -8.51093903e-02 -7.95036376e-01 2.02853039e-01
8.20795774e-01 -7.90865362e-01 1.21923399e+00 -2.30408192e-01
-2.83839345e-01 -3.65696341e-01 1.34768084e-01 -9.79559958e-01
-6.40619278e-01 -5.78878522e-01 -2.95589548e-02 1.73378873e+00
-2.69484669e-01 -1.03916192e+00 7.72781074e-01 6.36300802e-01
2.27998912e-01 -3.24410141e-01 -1.05584943e+00 -1.03334963e+00
-2.21989721e-01 1.73406005e-01 6.09972358e-01 1.04219842e+00
-9.29496139e-02 1.80690482e-01 -5.66984177e-01 3.20630372e-01
1.07776427e+00 6.46445602e-02 9.83706236e-01 -1.37128389e+00
-2.86507577e-01 -5.53676665e-01 -1.07687473e+00 -1.16855407e+00
6.13362312e-01 -1.01767135e+00 -1.21518686e-01 -1.06525552e+00
6.59149945e-01 -6.11000001e-01 -8.04823756e-01 1.63305759e-01
-5.07351220e-01 2.48490185e-01 1.74684793e-01 5.53356051e-01
-6.78645432e-01 8.48107100e-01 6.64871871e-01 -5.82539260e-01
-2.79177487e-01 5.97938243e-03 -8.31451118e-01 6.03071928e-01
5.72997987e-01 -3.42296839e-01 -2.89924353e-01 -3.46739680e-01
-2.62369603e-01 -6.76104248e-01 5.09955823e-01 -1.13692391e+00
5.37878633e-01 1.60054147e-01 7.68339455e-01 -2.56070614e-01
3.18078190e-01 -7.68674195e-01 -4.71850000e-02 1.33741260e-01
-2.33478621e-01 1.46451384e-01 1.74581222e-02 8.37090611e-01
-2.84909338e-01 -1.20628327e-01 8.61869514e-01 -5.18396087e-02
-9.90283310e-01 6.86934233e-01 8.12126845e-02 -1.11084148e-01
7.76868105e-01 -2.42155880e-01 5.30984299e-03 -1.35994852e-01
-4.72257107e-01 4.46397036e-01 5.32435417e-01 5.29768944e-01
8.55786264e-01 -1.39652979e+00 -8.76264811e-01 2.94733346e-01
4.71230388e-01 -2.71422148e-01 4.16997582e-01 6.92842841e-01
-6.59407452e-02 1.15875080e-02 -1.50723442e-01 -3.60754073e-01
-1.35586083e+00 5.71024060e-01 1.80385396e-01 -9.22612846e-02
-6.86929941e-01 1.00311673e+00 3.13342005e-01 -2.95531213e-01
3.37629527e-01 4.32702750e-01 -4.25139904e-01 1.89757183e-01
9.22555089e-01 6.00025356e-01 -2.41181880e-01 -1.15925622e+00
-5.33483684e-01 9.24970627e-01 -4.20411766e-01 1.26099050e-01
1.23704743e+00 -4.33980912e-01 -1.21223010e-01 2.57628679e-01
1.43184900e+00 1.04248270e-01 -1.04327738e+00 -6.94476008e-01
1.34406596e-01 -7.63897538e-01 -2.33603463e-01 -2.18330517e-01
-8.48655641e-01 5.35735905e-01 1.12179160e+00 -3.07001263e-01
1.16099679e+00 -1.05429195e-01 9.83934641e-01 4.33381885e-01
5.94656229e-01 -1.12523377e+00 -4.53744270e-02 1.27732888e-01
5.08490622e-01 -1.41800213e+00 1.66899070e-01 -1.77341849e-01
-2.84864634e-01 8.64367783e-01 5.19493878e-01 -2.22918466e-01
4.99213368e-01 -3.12963933e-01 -9.17685479e-02 1.04951717e-01
1.44001126e-01 -2.50949860e-01 4.49993402e-01 7.14828014e-01
1.39054641e-01 3.77704129e-02 -3.46367478e-01 6.57985270e-01
-1.46714598e-01 -2.63363123e-01 -7.91249871e-02 7.14211583e-01
-1.06737636e-01 -1.36259639e+00 -3.10489208e-01 4.12547141e-01
-1.25404567e-01 2.54305396e-02 -2.65681207e-01 5.42964458e-01
3.68109643e-01 9.27336514e-01 -7.54307508e-02 -5.04119098e-01
3.06242764e-01 -4.71487548e-03 6.00230396e-01 -4.51864541e-01
-5.76238751e-01 -2.76385993e-01 -4.17319238e-01 -2.49619439e-01
-6.34647191e-01 -1.07132113e+00 -1.01011872e+00 -2.47499183e-01
-4.26246643e-01 1.42242193e-01 4.05595392e-01 9.22841311e-01
2.46733367e-01 2.75667589e-02 9.59398985e-01 -1.05899298e+00
-4.90466982e-01 -7.36501455e-01 -5.59905052e-01 7.15244353e-01
4.20374304e-01 -8.19943368e-01 -2.94479609e-01 2.09933445e-01] | [14.746077537536621, 1.0358006954193115] |
d8d40ca1-e379-46f3-a2cd-32724aac37a3 | sparsification-and-filtering-for-spatial | 2203.03991 | null | https://arxiv.org/abs/2203.03991v1 | https://arxiv.org/pdf/2203.03991v1.pdf | Sparsification and Filtering for Spatial-temporal GNN in Multivariate Time-series | We propose an end-to-end architecture for multivariate time-series prediction that integrates a spatial-temporal graph neural network with a matrix filtering module. This module generates filtered (inverse) correlation graphs from multivariate time series before inputting them into a GNN. In contrast with existing sparsification methods adopted in graph neural network, our model explicitly leverage time-series filtering to overcome the low signal-to-noise ratio typical of complex systems data. We present a set of experiments, where we predict future sales from a synthetic time-series sales dataset. The proposed spatial-temporal graph neural network displays superior performances with respect to baseline approaches, with no graphical information, and with fully connected, disconnected graphs and unfiltered graphs. | ['Tomaso Aste', 'Yuanrong Wang'] | 2022-03-08 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [ 2.57829398e-01 3.20398837e-01 1.19779728e-01 -2.38717973e-01
1.06684707e-01 -4.37832475e-01 7.13972151e-01 3.56935829e-01
5.19927964e-02 4.08539116e-01 7.95179084e-02 -8.15074801e-01
-7.12527335e-01 -1.05295312e+00 -8.39378953e-01 -1.50605157e-01
-1.07648921e+00 4.02114868e-01 -2.34333187e-01 -5.49999833e-01
-1.35776162e-01 5.42904079e-01 -1.21123588e+00 1.98987126e-01
7.97710836e-01 1.19435859e+00 -2.57084638e-01 7.68553019e-01
1.72671050e-01 1.19634974e+00 -4.58458304e-01 -4.84359443e-01
6.00301027e-01 -2.03434542e-01 -1.55792028e-01 1.18064217e-01
3.79859984e-01 1.40150860e-01 -1.08591509e+00 8.97921205e-01
1.18836991e-01 5.05304575e-01 5.24112821e-01 -1.34906900e+00
-1.18828154e+00 9.55608130e-01 -5.48168242e-01 4.01682287e-01
3.27997804e-01 2.44495958e-01 9.06729579e-01 -6.37369037e-01
8.83103967e-01 1.19784629e+00 1.01467538e+00 -4.11620587e-01
-2.09835958e+00 -5.65791428e-01 5.85304797e-01 1.11828551e-01
-1.23350084e+00 1.93694472e-01 1.35732555e+00 -4.64818329e-01
1.50963175e+00 1.12269446e-01 1.07325888e+00 1.15451729e+00
7.11689055e-01 3.67524505e-01 8.17896008e-01 4.91536371e-02
3.28038707e-02 -6.75214350e-01 2.01388553e-01 4.68733132e-01
9.66949835e-02 2.90996850e-01 -5.14480174e-01 -1.69350252e-01
8.48592818e-01 4.74073440e-01 4.81383540e-02 -5.78667581e-01
-1.20122004e+00 6.96018636e-01 9.76160467e-01 2.60921329e-01
-8.60612214e-01 2.39870161e-01 4.77989137e-01 8.49140942e-01
7.67352164e-01 3.09002459e-01 -3.18861306e-01 2.71727830e-01
-1.02716482e+00 8.20765197e-02 1.10567689e+00 9.72224832e-01
5.06748259e-01 7.12642848e-01 5.37984213e-03 2.88716257e-01
1.06202215e-01 4.74603057e-01 3.54590267e-01 -3.46206933e-01
7.56357372e-01 8.54668617e-01 -4.50480044e-01 -1.88421965e+00
-8.26686442e-01 -7.96188951e-01 -1.57521796e+00 8.56546834e-02
3.05996180e-01 -8.09697285e-02 -1.06286180e+00 1.42282963e+00
-5.49064763e-02 6.72350943e-01 -1.14312492e-01 6.01377487e-01
8.00020933e-01 5.47842443e-01 -1.76992714e-01 -3.07478726e-01
6.61071181e-01 -6.58673882e-01 -7.67431855e-01 -1.64448753e-01
1.89764112e-01 -4.64469135e-01 7.79525816e-01 6.46704614e-01
-9.72138822e-01 -6.08921468e-01 -1.12346876e+00 1.50845915e-01
-6.19491696e-01 1.81788579e-02 9.04525876e-01 1.47995383e-01
-1.20351529e+00 1.30145967e+00 -8.95809174e-01 -1.13665238e-01
9.98147726e-02 5.41382432e-01 -4.30471569e-01 3.60579818e-01
-1.19647920e+00 5.76615632e-01 4.25789267e-01 5.36900222e-01
-4.48873729e-01 -5.80018103e-01 -9.88819957e-01 4.83100340e-02
3.37464571e-01 -7.22678959e-01 6.71771049e-01 -9.94996786e-01
-1.36656761e+00 1.50549129e-01 3.00480753e-01 -1.18864512e+00
4.96040076e-01 1.30279660e-01 -1.13452029e+00 -1.62268192e-01
-2.39138305e-01 -3.64941470e-02 1.25322759e+00 -8.71798158e-01
2.56799832e-02 -2.66960442e-01 -1.66036069e-01 -3.31327915e-01
4.77682427e-02 -4.46002990e-01 -8.24675262e-02 -9.64107513e-01
4.03426141e-01 -7.32254207e-01 -8.40733647e-01 -5.00748217e-01
-7.20689654e-01 1.89286798e-01 9.85774636e-01 -8.19171190e-01
1.50402141e+00 -1.95857871e+00 1.74739867e-01 1.00766671e+00
6.44139349e-01 -1.23809218e-01 -3.96647453e-01 9.55772340e-01
-7.54047751e-01 -3.00529808e-01 3.84573750e-02 -3.47178102e-01
-1.78997833e-02 5.94884269e-02 -4.70200092e-01 4.04647022e-01
4.48074192e-01 1.17081809e+00 -9.66552436e-01 2.68554121e-01
4.15476561e-01 6.60424352e-01 -1.82651863e-01 -1.56399254e-02
-2.84212738e-01 3.07110429e-01 1.73950440e-03 5.36602497e-01
6.01561129e-01 -7.49183059e-01 6.66223228e-01 -4.00538146e-01
1.27813593e-01 1.96342900e-01 -1.31385839e+00 1.57543051e+00
-2.31025457e-01 5.15893042e-01 2.89367978e-03 -1.22418523e+00
1.28681874e+00 -1.29341751e-01 7.19554663e-01 -9.65179026e-01
2.29122818e-01 -1.26403347e-01 7.85425529e-02 -9.74859148e-02
5.47457933e-01 2.18873993e-01 -1.29881939e-02 2.68024951e-01
2.99813956e-01 3.88372950e-02 5.34607351e-01 5.11429429e-01
1.47934186e+00 -1.07394559e-02 1.86287865e-01 -1.94997191e-01
3.87986422e-01 -9.63931009e-02 1.24862827e-01 3.70526701e-01
2.76898861e-01 3.03007931e-01 9.27229464e-01 -8.59615147e-01
-1.06023014e+00 -1.33978474e+00 4.96761292e-01 8.93187344e-01
-2.82400668e-01 -7.18647897e-01 -4.95014340e-02 -5.13515532e-01
4.70709980e-01 6.05388224e-01 -8.45882416e-01 -8.93198624e-02
-6.23892009e-01 -3.35700095e-01 9.12642404e-02 4.99035478e-01
-1.35436594e-01 -9.56215143e-01 1.67150274e-01 5.08742213e-01
4.74804372e-01 -8.67464125e-01 -4.44466352e-01 4.93155867e-01
-1.09279442e+00 -1.20606709e+00 -1.30339265e-01 -6.44282341e-01
6.58029616e-01 2.04076916e-01 1.45217133e+00 -3.14068735e-01
-1.83159858e-01 3.46318215e-01 -8.42300057e-02 -2.52435386e-01
-3.12632620e-01 -4.51610833e-02 -5.24054235e-03 2.47483596e-01
2.15599239e-01 -1.45924783e+00 -6.59883797e-01 6.91958815e-02
-8.50729167e-01 3.46027277e-02 5.46032429e-01 8.89466584e-01
7.39117444e-01 2.02032521e-01 4.84326214e-01 -8.29797328e-01
1.27702677e+00 -6.38998568e-01 -1.01994681e+00 2.05124840e-02
-8.50664139e-01 6.77419128e-03 1.27071166e+00 -7.29514480e-01
-2.16014013e-01 3.86057109e-01 4.94574815e-01 -8.85336757e-01
3.22561771e-01 1.10028017e+00 4.06439215e-01 -2.53678448e-02
8.58395517e-01 9.14196894e-02 5.54630280e-01 -2.36332402e-01
7.85250008e-01 -1.65255532e-01 6.07007027e-01 1.65250748e-02
1.00451052e+00 2.51277387e-01 4.26650673e-01 -6.77610219e-01
-3.24118473e-02 -2.59366482e-01 -9.35844064e-01 -3.41588944e-01
2.57316351e-01 -9.05658960e-01 -1.07941902e+00 5.90857565e-02
-8.21965456e-01 -4.97798890e-01 -3.37836236e-01 5.32398880e-01
-4.95724261e-01 3.19250464e-01 -8.05505872e-01 -6.80221677e-01
-5.51488221e-01 -4.13574755e-01 8.87145400e-01 -3.02357912e-01
-2.85183072e-01 -1.26557481e+00 -3.65337431e-02 -3.58572423e-01
5.44631004e-01 1.00105321e+00 7.93640375e-01 -6.47208214e-01
-6.49314702e-01 -4.91009980e-01 -1.88789651e-01 3.13749760e-01
-1.14466637e-01 1.70489565e-01 -3.94225031e-01 -3.08837205e-01
-3.03212613e-01 1.45568699e-01 9.61706460e-01 6.35712922e-01
9.66091156e-01 -5.61847389e-01 -2.57170647e-01 6.72787189e-01
1.35241175e+00 1.12553269e-01 3.86203200e-01 1.83053408e-02
1.05446160e+00 4.96249557e-01 1.18745804e-01 4.68563199e-01
5.08629739e-01 -5.13240835e-03 5.37254751e-01 -1.59599856e-01
2.38379370e-02 -5.44076324e-01 3.61660033e-01 1.14955986e+00
-2.24675685e-01 -2.06180632e-01 -9.22773719e-01 4.03316170e-01
-2.27787232e+00 -1.14166784e+00 -4.07885998e-01 1.96327806e+00
1.41173363e-01 5.64622760e-01 4.70329642e-01 2.54678279e-01
4.88241702e-01 6.22550488e-01 -7.75149822e-01 -3.05169076e-01
-2.32364893e-01 3.49391907e-01 7.62771606e-01 3.01919281e-01
-1.14224637e+00 5.47128856e-01 6.92248297e+00 3.23207110e-01
-1.33787119e+00 -2.98410803e-01 3.63761395e-01 -1.39869928e-01
-4.10035491e-01 -2.26692349e-01 8.31620544e-02 -4.08745883e-03
1.39010882e+00 -5.75569332e-01 9.85322714e-01 8.89924347e-01
2.92211831e-01 6.20957017e-01 -1.18726170e+00 9.73716557e-01
-1.07759424e-01 -1.43884337e+00 -2.80740801e-02 -4.08068076e-02
5.50814450e-01 4.36716259e-01 2.47369364e-01 4.41054165e-01
8.30014825e-01 -1.31940365e+00 6.20228231e-01 8.92954290e-01
4.48799700e-01 -8.02227318e-01 2.97149569e-01 2.09048212e-01
-1.77197850e+00 -3.08828145e-01 -1.12648375e-01 -4.31110889e-01
4.82535958e-01 8.71622086e-01 -1.04026890e+00 9.73125458e-01
3.73300970e-01 1.32907522e+00 -6.41021967e-01 6.83911741e-01
-8.77107456e-02 4.47823137e-01 -5.43842435e-01 -6.28795847e-02
2.60756344e-01 -6.90550148e-01 4.27104324e-01 1.06071937e+00
4.09831822e-01 -1.98522881e-01 3.06417286e-01 9.86372888e-01
4.68206108e-02 2.75790673e-02 -1.17376530e+00 -6.10318780e-01
2.15367526e-02 1.28213930e+00 -8.86137784e-01 -2.20341459e-01
-6.00899518e-01 6.37673020e-01 3.96642387e-01 7.63952553e-01
-5.76659143e-01 -4.44260895e-01 1.45803124e-01 3.37595493e-01
4.49991822e-01 -7.03638792e-01 -2.19682395e-01 -1.14378119e+00
1.67025462e-01 -8.96536887e-01 6.29619539e-01 -6.29042625e-01
-1.83161521e+00 8.65624726e-01 -2.01925263e-01 -1.48174691e+00
-6.55658722e-01 -7.63021290e-01 -6.30576968e-01 8.42448831e-01
-9.64551032e-01 -1.48728621e+00 -1.31695867e-01 8.32170784e-01
-1.01662360e-01 -3.44922483e-01 6.42866254e-01 1.48835078e-01
-3.28035176e-01 2.02077299e-01 1.46577686e-01 1.44019380e-01
6.24431930e-02 -1.43781412e+00 1.09786975e+00 9.82639670e-01
2.54752398e-01 8.19339216e-01 7.39200890e-01 -1.17420495e+00
-1.98114049e+00 -1.39665461e+00 5.80313206e-01 -1.79737970e-01
1.55356741e+00 -6.30329430e-01 -8.88094306e-01 1.04268348e+00
4.13220823e-01 2.40546942e-01 3.45286876e-01 5.06850064e-01
-4.53082144e-01 -3.28343153e-01 -6.98937893e-01 6.72686815e-01
1.18593681e+00 -6.48789704e-01 -4.04313177e-01 4.53220606e-01
7.39593863e-01 -2.32833341e-01 -1.31153226e+00 3.22385132e-01
5.29336333e-01 -8.58437240e-01 1.14435494e+00 -6.71577096e-01
1.69164464e-01 -3.74652475e-01 1.11796603e-01 -1.70703268e+00
-7.28750885e-01 -1.18809891e+00 -5.57816267e-01 5.77610075e-01
6.46020174e-01 -8.14209402e-01 8.23596239e-01 1.97153956e-01
-3.45405079e-02 -5.45791030e-01 -5.99125266e-01 -9.57282841e-01
-4.15823489e-01 -6.64325833e-01 7.27515996e-01 1.09203434e+00
1.92230478e-01 6.93151176e-01 -4.67025459e-01 1.56988025e-01
6.34903550e-01 3.76156807e-01 1.00858736e+00 -1.46973121e+00
-4.88158703e-01 -5.89555860e-01 -6.45577908e-01 -7.01670825e-01
1.11400165e-01 -1.14256489e+00 -4.63922709e-01 -1.53400528e+00
-7.61615753e-01 3.06981891e-01 -5.29879332e-01 1.89354300e-01
3.12975764e-01 -9.20689013e-03 2.67133921e-01 -1.67352781e-01
-4.31854218e-01 5.69051862e-01 1.29235828e+00 -3.11181724e-01
-5.33372104e-01 4.22360599e-02 -2.62261569e-01 4.03557241e-01
4.92063135e-01 -1.61912158e-01 -7.81759739e-01 4.46813814e-02
6.07346833e-01 5.15823066e-01 2.78812438e-01 -8.70717883e-01
4.10890013e-01 1.28070787e-01 5.26023686e-01 -1.02977300e+00
3.48695725e-01 -1.07982290e+00 8.19660187e-01 4.65112537e-01
2.34131590e-02 9.05429125e-01 3.68095696e-01 1.10837138e+00
-3.61786336e-01 7.83542514e-01 -6.55893460e-02 1.72155946e-02
-4.97739047e-01 4.07739639e-01 -2.00834081e-01 -6.76147282e-01
9.45082486e-01 -2.12482914e-01 -4.41641062e-01 -6.93449140e-01
-1.38118577e+00 2.87615657e-01 1.11949090e-02 5.08374572e-01
5.97035348e-01 -1.50709414e+00 -4.61935133e-01 6.09886408e-01
2.71141566e-02 -3.14779401e-01 2.90830612e-01 1.14202011e+00
-4.24355447e-01 2.10884944e-01 -3.76558036e-01 -5.46552062e-01
-8.10891390e-01 1.14887869e+00 3.85624953e-02 -6.62035704e-01
-8.33046138e-01 3.73822987e-01 -4.17851567e-01 -5.67848027e-01
2.72847623e-01 -9.32226300e-01 -2.06504121e-01 2.00577304e-01
1.66532502e-01 3.32412004e-01 1.85087055e-01 -5.74470460e-01
-1.60655141e-01 1.78613409e-01 1.89255372e-01 3.30604643e-01
1.78655565e+00 1.28290117e-01 -2.53748119e-01 6.71760499e-01
9.79590595e-01 -3.20507824e-01 -1.01374590e+00 -2.74724960e-01
1.39592648e-01 -1.36873290e-01 -1.57571390e-01 -5.28438330e-01
-1.23064828e+00 4.59702462e-01 3.66388500e-01 1.11131442e+00
1.32171917e+00 -5.27912498e-01 5.18569648e-01 6.81417167e-01
2.53771782e-01 -9.16011572e-01 8.02802667e-02 7.01132476e-01
1.21729815e+00 -9.00109649e-01 3.24341476e-01 -4.17612463e-01
-5.53204179e-01 1.37650323e+00 2.00443238e-01 -9.95640755e-01
1.18152440e+00 3.02025348e-01 -2.40974739e-01 -6.48162842e-01
-1.10469830e+00 -2.97800303e-01 7.86988080e-01 6.87600434e-01
3.44906688e-01 2.70202279e-01 -1.77580446e-01 4.30242032e-01
-5.75422227e-01 -6.50899634e-02 2.52636015e-01 5.02493083e-01
2.35650361e-01 -7.88037777e-01 -1.35877386e-01 8.96838665e-01
-3.63099054e-02 -9.69462916e-02 -5.60937405e-01 1.07248974e+00
-4.59194928e-01 8.84406865e-01 3.36170018e-01 -9.74481463e-01
8.34878325e-01 -1.12755343e-01 1.30908221e-01 -2.28391632e-01
-1.02072930e+00 3.44005078e-01 3.35640788e-01 -9.33916390e-01
-2.36793473e-01 -4.01257128e-01 -9.32803810e-01 -6.10502303e-01
-3.40145268e-02 -3.00086617e-01 3.21024776e-01 3.12580943e-01
7.65128314e-01 1.00217056e+00 4.71115887e-01 -1.15096521e+00
-2.15576217e-01 -1.16157544e+00 -9.61846054e-01 7.09925413e-01
4.39770043e-01 -3.10910344e-01 -1.60183400e-01 -7.82355741e-02] | [6.84735631942749, 2.922158718109131] |
de0f4bf9-f42a-470b-ae12-1a9b6c94dbee | discriminative-feature-learning-for | 1811.09791 | null | http://arxiv.org/abs/1811.09791v1 | http://arxiv.org/pdf/1811.09791v1.pdf | Discriminative Feature Learning for Unsupervised Video Summarization | In this paper, we address the problem of unsupervised video summarization
that automatically extracts key-shots from an input video. Specifically, we
tackle two critical issues based on our empirical observations: (i) Ineffective
feature learning due to flat distributions of output importance scores for each
frame, and (ii) training difficulty when dealing with long-length video inputs.
To alleviate the first problem, we propose a simple yet effective
regularization loss term called variance loss. The proposed variance loss
allows a network to predict output scores for each frame with high discrepancy
which enables effective feature learning and significantly improves model
performance. For the second problem, we design a novel two-stream network named
Chunk and Stride Network (CSNet) that utilizes local (chunk) and global
(stride) temporal view on the video features. Our CSNet gives better
summarization results for long-length videos compared to the existing methods.
In addition, we introduce an attention mechanism to handle the dynamic
information in videos. We demonstrate the effectiveness of the proposed methods
by conducting extensive ablation studies and show that our final model achieves
new state-of-the-art results on two benchmark datasets. | ['Yunjae Jung', 'Dahun Kim', 'Sanghyun Woo', 'In So Kweon', 'Donghyeon Cho'] | 2018-11-24 | null | null | null | null | ['unsupervised-video-summarization', 'supervised-video-summarization'] | ['computer-vision', 'computer-vision'] | [ 3.55700016e-01 -2.53778875e-01 -4.28011447e-01 -2.23944947e-01
-8.38221967e-01 -9.26374942e-02 2.15350181e-01 -8.04996490e-02
-4.89450037e-01 6.99320436e-01 7.17882812e-01 1.49682939e-01
-1.82661451e-02 -2.14072213e-01 -7.35825062e-01 -6.66435540e-01
-2.41779476e-01 -3.45986277e-01 4.96635854e-01 1.04161829e-01
4.21949059e-01 9.04179364e-02 -1.64306867e+00 3.99678081e-01
1.08180225e+00 1.10064042e+00 4.50529158e-01 6.80682957e-01
2.85510179e-02 1.18591261e+00 -6.51857734e-01 -7.90240318e-02
3.27167988e-01 -8.17712307e-01 -5.97252667e-01 3.10014606e-01
5.48228920e-01 -7.54645288e-01 -5.49830258e-01 1.02477956e+00
5.95547259e-01 3.76749516e-01 6.06881917e-01 -1.22353828e+00
-3.54066849e-01 6.96228743e-01 -8.84133756e-01 5.66163063e-01
3.51607889e-01 1.93679348e-01 1.16835332e+00 -9.01648462e-01
5.35272658e-01 1.08527803e+00 6.64760530e-01 4.11947221e-01
-7.92630672e-01 -4.57609117e-01 4.19848114e-01 5.80233097e-01
-1.19433117e+00 -6.66649222e-01 8.15539122e-01 -3.08356017e-01
8.28068495e-01 6.02737367e-02 5.56711078e-01 9.51933861e-01
2.95023173e-01 1.20042944e+00 3.99801105e-01 -1.17626704e-01
7.25552067e-02 -2.56296217e-01 1.05440758e-01 7.93189347e-01
7.99262449e-02 -1.86910212e-01 -6.91573918e-01 -1.71838980e-02
6.41917288e-01 2.02023357e-01 -5.94439328e-01 -1.97769672e-01
-1.21522081e+00 6.24348044e-01 1.80233568e-01 5.48085794e-02
-4.68774289e-01 2.64722645e-01 7.63918400e-01 3.87168318e-01
5.02141058e-01 2.22528532e-01 -4.19350624e-01 -3.25736463e-01
-1.15989041e+00 1.75516501e-01 5.64268708e-01 1.01878655e+00
5.59184194e-01 3.25993538e-01 -5.37344873e-01 8.71801734e-01
8.81020352e-02 2.57989228e-01 6.26504421e-01 -1.17203498e+00
6.88332140e-01 2.73680806e-01 -6.07354045e-02 -1.20013070e+00
-2.25948155e-01 -3.61965716e-01 -9.47613597e-01 -2.48318166e-01
4.48120162e-02 -3.32087040e-01 -7.88508713e-01 1.84236836e+00
-2.70654615e-02 4.79057193e-01 1.05422579e-01 1.01412487e+00
1.06725109e+00 9.58939135e-01 -7.17283189e-02 -7.67853796e-01
9.99613404e-01 -1.50764263e+00 -8.67368042e-01 1.96371209e-02
4.61127877e-01 -5.45912921e-01 1.04580355e+00 2.61233807e-01
-1.21960187e+00 -6.12165689e-01 -1.11898899e+00 1.06577918e-01
3.50444257e-01 1.83699146e-01 3.70443970e-01 -2.90827807e-02
-9.98988509e-01 7.36695528e-01 -7.78034031e-01 -3.81621033e-01
4.77633655e-01 3.28562289e-01 -1.91557258e-01 -2.00799759e-02
-1.07608199e+00 2.57587671e-01 3.96917224e-01 -7.96089172e-02
-7.88973927e-01 -6.34470761e-01 -1.01943696e+00 3.53207976e-01
6.37675345e-01 -7.04326034e-01 1.22734225e+00 -1.36550665e+00
-1.50646996e+00 2.28262991e-01 -4.05693352e-01 -4.98220444e-01
4.20545101e-01 -5.31540036e-01 -1.62661076e-01 6.28221273e-01
1.25418842e-01 6.14949167e-01 1.11779916e+00 -1.03056014e+00
-8.38355362e-01 -4.30563875e-02 -3.50687020e-02 4.17997360e-01
-6.51957631e-01 2.76505202e-03 -6.51079595e-01 -1.01915240e+00
-9.96039063e-02 -5.88491201e-01 -1.33896053e-01 -1.69177145e-01
-3.07390928e-01 -1.70906112e-01 8.05202365e-01 -6.99996710e-01
1.76348746e+00 -2.24314666e+00 3.35605949e-01 -3.07077765e-01
2.84115493e-01 5.17668843e-01 -2.97558546e-01 3.03828359e-01
8.14414471e-02 2.39964388e-03 -1.47949323e-01 -3.02314550e-01
-3.19562703e-01 -8.27008765e-03 -3.53804290e-01 2.02757537e-01
2.00818151e-01 7.68240392e-01 -9.96701479e-01 -7.15024352e-01
-9.28224400e-02 1.89729705e-01 -8.27003121e-01 4.37953472e-01
-1.35043999e-02 2.93259829e-01 -4.74278033e-01 4.62455571e-01
5.52208781e-01 -2.93279350e-01 -2.56289989e-02 -5.99860370e-01
-8.14056396e-02 7.92954117e-02 -1.03550613e+00 1.86557889e+00
-2.66010817e-02 5.96545756e-01 -2.04581454e-01 -1.09362793e+00
5.50200343e-01 3.70366752e-01 5.91219246e-01 -5.04401684e-01
2.13432476e-01 -4.96201068e-02 -6.87754676e-02 -8.37633848e-01
6.45191431e-01 2.40971521e-01 1.14835694e-01 4.87104684e-01
3.40595603e-01 4.41117525e-01 4.82714862e-01 4.01926488e-01
1.15577853e+00 2.11127147e-01 4.57429081e-01 -2.06411749e-01
6.58557594e-01 -3.43584538e-01 1.08551955e+00 7.98070848e-01
-4.86615688e-01 8.96705389e-01 8.14875245e-01 -4.48751807e-01
-1.00835729e+00 -7.25376546e-01 4.47902232e-01 1.22935617e+00
3.98323894e-01 -8.08983147e-01 -7.53999114e-01 -8.63746881e-01
-2.03289658e-01 2.67332524e-01 -4.49783742e-01 -3.52705985e-01
-7.37482131e-01 -4.47825193e-01 2.94816196e-01 7.30022669e-01
6.71606183e-01 -1.02036071e+00 -5.55998325e-01 1.53767437e-01
-4.74198192e-01 -1.12374592e+00 -1.03032386e+00 -2.59875983e-01
-1.00809586e+00 -9.61401343e-01 -7.65046895e-01 -8.04952204e-01
6.01094425e-01 6.87687457e-01 9.60463643e-01 1.20271474e-01
-8.18195269e-02 2.86748290e-01 -6.58769608e-01 -7.66672418e-02
-5.79257086e-02 2.33242914e-01 9.34831053e-02 6.92876130e-02
1.52308702e-01 -5.40153146e-01 -8.41518760e-01 2.75828809e-01
-1.03870797e+00 2.71268994e-01 7.45361745e-01 8.52032840e-01
5.66595614e-01 2.02177931e-02 9.84293461e-01 -6.66236937e-01
6.78436637e-01 -4.99214679e-01 -1.67779401e-01 2.56960988e-01
-4.02299553e-01 7.95931891e-02 9.68181729e-01 -3.87368679e-01
-9.38873470e-01 -6.51884601e-02 1.36919599e-02 -6.62031710e-01
1.50356770e-01 5.54874718e-01 -5.24540283e-02 2.83440650e-01
3.31309795e-01 5.11426926e-01 9.91969928e-02 -3.38395715e-01
6.58090413e-02 5.43081999e-01 6.07499003e-01 -4.19490337e-01
5.37929237e-01 3.77144098e-01 -2.23249838e-01 -9.27425325e-01
-1.05102456e+00 -4.89763260e-01 -4.87249017e-01 -1.97522640e-01
7.16042638e-01 -1.17409539e+00 -5.45646489e-01 5.75790048e-01
-1.00675166e+00 -2.01160610e-01 -2.66856700e-01 6.28377736e-01
-6.80586934e-01 8.45035017e-01 -7.45194793e-01 -5.40227771e-01
-7.25557506e-01 -1.12876058e+00 8.43276441e-01 3.40242714e-01
-4.05758023e-02 -7.55951703e-01 -5.38641624e-02 6.21636584e-02
3.17044050e-01 9.28314850e-02 6.45065248e-01 -4.46831405e-01
-4.69098449e-01 1.27973631e-01 -3.60094309e-01 7.16196477e-01
2.24561706e-01 1.50238201e-01 -5.26399374e-01 -5.15764177e-01
1.00234970e-01 -4.41463053e-01 1.36699831e+00 5.26448250e-01
1.38624609e+00 -5.05306661e-01 5.33500873e-03 7.93826103e-01
1.26365995e+00 -3.51028293e-02 6.35055721e-01 1.58919811e-01
7.97512650e-01 2.79305339e-01 8.31632257e-01 7.99234092e-01
2.97513306e-01 4.50041592e-01 2.13391542e-01 9.14945155e-02
-9.92030650e-02 -2.47161776e-01 6.32262349e-01 1.21378541e+00
-2.81838000e-01 -3.71062011e-01 -3.77106577e-01 6.33918822e-01
-2.33525324e+00 -1.14035499e+00 2.04678252e-01 2.11889625e+00
6.88240767e-01 2.63459146e-01 3.25802237e-01 -1.76925197e-01
8.16065013e-01 5.62261820e-01 -6.46189690e-01 -1.10444084e-01
-1.26964882e-01 -2.64425009e-01 3.72863710e-01 1.78152129e-01
-1.23505902e+00 8.99561524e-01 6.33851719e+00 1.00532031e+00
-1.08573294e+00 2.45163124e-02 5.43410301e-01 -3.84620607e-01
-2.79567186e-02 -2.42531642e-01 -6.07683063e-01 8.03077757e-01
6.79016352e-01 -4.52594012e-01 2.21416816e-01 7.04995036e-01
4.57927793e-01 -1.57931760e-01 -1.08853376e+00 1.16185784e+00
5.21089375e-01 -1.31537306e+00 3.54877681e-01 -3.11362118e-01
8.61496389e-01 -5.26021980e-02 -1.66664794e-01 3.51914853e-01
-2.33921766e-01 -6.44000113e-01 6.32523417e-01 4.27396446e-01
8.64520073e-01 -9.05978203e-01 8.09455931e-01 2.17801958e-01
-1.38918996e+00 -2.40351692e-01 -5.15309870e-01 -2.55931877e-02
2.52817452e-01 3.91181111e-01 -2.52717823e-01 6.07191086e-01
4.98855054e-01 1.25912273e+00 -4.38748449e-01 1.24868071e+00
-2.27069899e-01 7.55439281e-01 -6.29711896e-02 2.66505152e-01
4.12384003e-01 -9.65212733e-02 7.09458530e-01 1.33633959e+00
4.36220169e-01 1.86672717e-01 4.11106467e-01 4.63314116e-01
-2.67411381e-01 8.07863399e-02 -4.22544479e-01 3.28149833e-02
3.31143558e-01 1.16272652e+00 -5.37756205e-01 -5.31475902e-01
-6.20995104e-01 9.97523010e-01 3.40210229e-01 4.82179523e-01
-9.15863574e-01 -6.32851899e-01 5.73425949e-01 5.05790040e-02
5.83407462e-01 -8.80791694e-02 1.82406977e-01 -1.56306100e+00
1.80016011e-01 -8.50115895e-01 4.86354858e-01 -5.71028829e-01
-1.07128215e+00 5.53104401e-01 -5.07592782e-02 -1.48743045e+00
-3.08080614e-01 -1.03601202e-01 -8.60277355e-01 3.02457690e-01
-1.61916983e+00 -7.16078103e-01 -4.59933519e-01 5.84735036e-01
1.16532731e+00 -4.12417680e-01 2.85542935e-01 3.68652910e-01
-7.88156211e-01 7.52880216e-01 1.49807230e-01 1.43075883e-01
9.29811001e-01 -1.04514921e+00 2.46748820e-01 9.76901412e-01
-2.07270041e-01 3.74802023e-01 6.39037848e-01 -5.41001320e-01
-1.39064217e+00 -1.18799293e+00 5.32782853e-01 1.24329686e-01
4.60547835e-01 -3.24771064e-03 -8.90795708e-01 5.77452958e-01
3.12194824e-01 -6.20146841e-02 6.56688094e-01 -2.61096507e-01
-1.83268398e-01 -2.53675371e-01 -8.24582160e-01 5.82257628e-01
1.12719154e+00 -1.44164860e-01 -7.45330155e-01 1.21045575e-01
1.10057163e+00 -2.28981569e-01 -5.91033697e-01 5.86556792e-01
5.79580963e-01 -1.15155053e+00 6.57385826e-01 -6.66434824e-01
9.28739667e-01 -2.05962181e-01 -8.89995694e-02 -1.32997239e+00
-4.50769603e-01 -8.90104413e-01 -6.14596665e-01 1.09527862e+00
1.77375168e-01 -2.05939323e-01 7.36370265e-01 2.00886384e-01
-3.29568893e-01 -1.04964936e+00 -5.16420841e-01 -7.78036594e-01
-4.28071260e-01 -8.96024853e-02 3.06229979e-01 6.36531174e-01
5.44148944e-02 5.20780742e-01 -8.90419424e-01 -1.75795689e-01
6.58562183e-01 1.99329674e-01 7.77334571e-01 -8.51092696e-01
-3.00549030e-01 -3.90808463e-01 -4.37082231e-01 -1.58447945e+00
2.10653991e-01 -5.18350124e-01 1.06394827e-01 -1.49138212e+00
7.77371585e-01 1.12654321e-01 -5.25779009e-01 1.41115993e-01
-4.72329587e-01 1.00208230e-01 4.15091515e-01 4.12304163e-01
-1.37789345e+00 8.75486195e-01 1.27218878e+00 6.62830919e-02
-3.04474533e-01 -1.00730948e-01 -7.83638060e-01 8.16675484e-01
6.24436140e-01 -3.52449149e-01 -6.15371168e-01 -5.85272133e-01
2.69855466e-02 2.06266314e-01 4.06860895e-02 -1.11854231e+00
3.07065994e-01 -2.20606908e-01 2.23055527e-01 -8.08475137e-01
7.78006017e-02 -5.56371093e-01 -3.59210461e-01 2.97367960e-01
-4.89564449e-01 3.94156501e-02 -9.91857946e-02 7.60274231e-01
-4.92046684e-01 -1.83375210e-01 7.58541524e-01 -7.79138952e-02
-8.79941463e-01 5.03617942e-01 -2.38799855e-01 2.84368515e-01
8.96016121e-01 -2.81219721e-01 -3.38134974e-01 -6.65179074e-01
-4.93098050e-01 5.87747991e-01 4.53034788e-01 4.31003809e-01
8.35003078e-01 -1.47380364e+00 -9.15860236e-01 1.87485024e-01
4.32423651e-02 -3.50514278e-02 5.30802071e-01 9.94423985e-01
-6.03854895e-01 2.40612894e-01 -2.70483792e-01 -6.22467816e-01
-1.28343952e+00 5.54141104e-01 -8.73264447e-02 -2.80264974e-01
-7.36060083e-01 7.15398431e-01 5.34168839e-01 1.97384983e-01
4.75064307e-01 -2.63283640e-01 -3.67656946e-01 1.77063853e-01
8.50142956e-01 4.71639186e-01 -3.39093953e-01 -6.63238645e-01
-1.28059968e-01 5.14190555e-01 -4.82938230e-01 2.96758741e-01
1.50181234e+00 -3.61147404e-01 5.06561399e-02 3.83516639e-01
1.41949117e+00 -2.32370928e-01 -1.71690333e+00 -4.59711373e-01
-2.76073456e-01 -6.33386374e-01 -7.56165832e-02 -2.81032354e-01
-1.22991467e+00 6.06160402e-01 3.07943344e-01 3.95098552e-02
1.36655319e+00 -2.92099535e-01 1.30448997e+00 4.16963398e-01
-6.78763911e-03 -1.41027904e+00 4.23742741e-01 6.18457019e-01
7.21462846e-01 -1.27479529e+00 2.56229430e-01 -2.15419516e-01
-8.63397658e-01 1.06168926e+00 6.77663147e-01 -2.15624407e-01
4.08771753e-01 -3.79390791e-02 -2.70710766e-01 3.97337675e-02
-1.05554724e+00 -8.45106617e-02 3.10127318e-01 1.86518371e-01
3.39458972e-01 -3.58945936e-01 -5.68153203e-01 6.81657016e-01
1.67119056e-01 1.66207060e-01 7.52144158e-01 9.86057758e-01
-8.13606799e-01 -6.85108244e-01 -4.88993805e-03 7.04511642e-01
-7.46320903e-01 -7.63977170e-02 -6.88071251e-02 4.77743894e-01
-2.67569959e-01 7.69853055e-01 1.45860165e-01 -4.74653363e-01
3.10727417e-01 -1.41267791e-01 3.25539291e-01 -5.01289070e-01
-2.76130736e-01 3.93118560e-01 -1.47717863e-01 -7.51534581e-01
-6.33502126e-01 -5.62392354e-01 -1.14027393e+00 -2.63689965e-01
-2.96504170e-01 1.74405783e-01 2.25272924e-01 8.42941999e-01
6.40666604e-01 7.95836449e-01 9.46715355e-01 -1.04649699e+00
-7.36989617e-01 -9.84976590e-01 -6.53628647e-01 7.18303561e-01
4.70776767e-01 -5.13981998e-01 -4.90370572e-01 3.06816190e-01] | [10.37368106842041, 0.4380369484424591] |
ba8fc417-4014-4650-98ba-095d30323bf2 | what-underlies-rapid-learning-and-systematic | 2107.06994 | null | https://arxiv.org/abs/2107.06994v2 | https://arxiv.org/pdf/2107.06994v2.pdf | Systematic human learning and generalization from a brief tutorial with explanatory feedback | Neural networks have long been used to model human intelligence, capturing elements of behavior and cognition, and their neural basis. Recent advancements in deep learning have enabled neural network models to reach and even surpass human levels of intelligence in many respects, yet unlike humans, their ability to learn new tasks quickly remains a challenge. People can reason not only in familiar domains, but can also rapidly learn to reason through novel problems and situations, raising the question of how well modern neural network models capture human intelligence and in which ways they diverge. In this work, we explore this gap by investigating human adults' ability to learn an abstract reasoning task based on Sudoku from a brief instructional tutorial with explanatory feedback for incorrect responses using a narrow range of training examples. We find that participants who master the task do so within a small number of trials and generalize well to puzzles outside of the training range. We also find that most of those who master the task can describe a valid solution strategy, and such participants perform better on transfer puzzles than those whose strategy descriptions are vague or incomplete. Interestingly, fewer than half of our human participants were successful in acquiring a valid solution strategy, and this ability is associated with high school mathematics education. We consider the challenges these findings pose for building computational models that capture all aspects of our findings and point toward a possible role for learning to engage in explanation-based reasoning to support rapid learning and generalization. | ['Andrew J. Nam', 'James L. McClelland'] | 2021-07-10 | null | null | null | null | ['systematic-generalization', 'high-school-mathematics'] | ['reasoning', 'reasoning'] | [ 1.75001547e-01 2.06716791e-01 9.31403339e-02 -3.15446734e-01
-2.17567503e-01 -5.99955499e-01 6.93986192e-02 3.69813144e-01
-5.10005593e-01 6.15666330e-01 1.86311632e-01 -3.76030862e-01
-7.70087957e-01 -8.86963844e-01 -4.33447927e-01 -4.01601009e-02
9.79967266e-02 7.57824123e-01 -6.35997877e-02 -4.49221730e-01
4.10668045e-01 4.15648609e-01 -1.41915619e+00 4.10815060e-01
1.52713883e+00 2.11418957e-01 2.89556801e-01 4.88153100e-01
-1.18644461e-01 1.13217890e+00 -6.88800335e-01 -6.06368959e-01
2.84253638e-02 -5.95387697e-01 -1.14750361e+00 -2.29531050e-01
5.26663303e-01 -8.39816988e-01 -3.94620568e-01 6.89794600e-01
1.72952101e-01 5.41854024e-01 6.19192779e-01 -8.43672872e-01
-1.18420279e+00 6.70728564e-01 -1.91753730e-02 4.39052194e-01
7.84064829e-01 6.02205157e-01 6.18856072e-01 -4.01369065e-01
3.14252943e-01 9.34207201e-01 8.44291091e-01 8.81958663e-01
-1.22100961e+00 -8.76811445e-01 1.94118842e-01 3.91273350e-01
-7.54791737e-01 -1.06987685e-01 4.32042181e-01 -4.67677832e-01
1.27128530e+00 -1.58575073e-01 1.48275733e+00 9.38847661e-01
-6.47883713e-02 6.47530198e-01 1.36339307e+00 -6.24280795e-02
1.42168745e-01 -2.69967318e-01 5.48807085e-01 6.32759690e-01
7.75166094e-01 4.34114546e-01 -7.53012836e-01 3.28238964e-01
1.05596948e+00 1.54031590e-01 -2.21403778e-01 -6.49939198e-03
-1.04201472e+00 7.54730761e-01 6.49460673e-01 5.09247780e-01
-5.56038260e-01 2.59709414e-02 8.87136012e-02 6.86297178e-01
-3.79257888e-01 1.20788527e+00 -6.00690484e-01 -5.21388888e-01
-6.10146463e-01 5.07224143e-01 8.45380366e-01 5.40678799e-01
3.60739797e-01 4.77424473e-01 4.33047079e-02 8.41996610e-01
-6.92646354e-02 3.65964502e-01 5.86820245e-01 -1.23665440e+00
6.09814644e-01 8.58529687e-01 -1.52529910e-01 -9.58043337e-01
-6.88586473e-01 -7.98346341e-01 -5.08291185e-01 3.96784157e-01
8.65651965e-01 -1.24548435e-01 -6.59268200e-01 2.05686688e+00
-6.46102726e-02 -1.33031249e-01 4.93182540e-02 7.71443188e-01
7.79188514e-01 3.32993418e-01 3.92540365e-01 -1.32643446e-01
1.24834347e+00 -6.95045114e-01 -2.68227726e-01 -8.73469949e-01
6.59662068e-01 -1.10383056e-01 1.15869677e+00 8.83681238e-01
-1.73210061e+00 -8.21331739e-01 -1.19892383e+00 -2.39511445e-01
-3.17746401e-01 -6.81354284e-01 1.25545394e+00 8.48244429e-01
-1.25156832e+00 8.08470845e-01 -4.66691881e-01 -2.56772488e-01
7.06050217e-01 6.16547823e-01 -4.43948746e-01 -6.19618118e-01
-1.05591929e+00 1.07307804e+00 7.08880603e-01 -2.40924940e-01
-6.95579350e-01 -1.10004663e+00 -8.47089946e-01 6.83201730e-01
1.55779749e-01 -1.09457123e+00 1.52777958e+00 -9.20927227e-01
-1.36201680e+00 6.80807114e-01 9.79781374e-02 -1.18173771e-01
-1.27741903e-01 1.04982024e-02 -3.51889478e-03 3.16344857e-01
9.68330353e-02 7.04213798e-01 2.29268327e-01 -6.20734751e-01
-4.76786643e-01 -4.61851746e-01 4.93203908e-01 3.65881234e-01
-1.47266150e-01 -2.02498376e-01 3.18624109e-01 -6.94157660e-01
5.62449872e-01 -5.45133829e-01 1.59112439e-01 -2.59666622e-01
4.67471510e-01 -5.95268190e-01 -1.56942040e-01 -7.72854328e-01
8.32742214e-01 -1.56563950e+00 1.51710957e-01 1.44864723e-01
4.90995705e-01 2.74034560e-01 -2.59056896e-01 2.87422657e-01
-1.70573950e-01 3.55486393e-01 5.08236773e-02 2.37344176e-01
2.78691947e-01 5.01805954e-02 3.06226965e-02 -6.55788481e-02
-3.81504148e-02 1.30632472e+00 -1.12256944e+00 1.71101049e-01
-1.45521790e-01 -3.45419422e-02 -9.83322144e-01 1.83586895e-01
-2.57414728e-01 4.58024472e-01 -2.56676339e-02 3.13210279e-01
4.25744772e-01 -2.61528760e-01 2.68545479e-01 6.80661559e-01
2.56374985e-01 7.73440540e-01 -1.01672268e+00 1.45606935e+00
-3.73831004e-01 7.60976672e-01 -1.69973969e-01 -1.10606027e+00
6.15380108e-01 3.64523619e-01 -2.61541396e-01 -9.18651879e-01
8.41298774e-02 9.88890678e-02 1.16095972e+00 -4.69705284e-01
-1.23599274e-02 -4.27025676e-01 3.65470618e-01 9.56801176e-01
2.13085432e-02 -5.28191090e-01 3.66087645e-01 1.05222508e-01
1.16820014e+00 -1.27268016e-01 3.98066431e-01 -1.46546468e-01
2.34124035e-01 1.24200970e-01 5.26441872e-01 1.25952220e+00
-2.48845324e-01 1.93793982e-01 4.61288989e-01 -6.92523420e-01
-8.86046946e-01 -1.05540097e+00 4.20593798e-01 1.16072893e+00
-4.13246572e-01 -3.21920514e-02 -7.49259770e-01 -2.32474908e-01
-2.70619392e-01 1.04345381e+00 -4.90065902e-01 -5.17411053e-01
-6.99977398e-01 -3.07889014e-01 3.17154467e-01 8.48405480e-01
1.00381076e+00 -1.53869224e+00 -7.17377782e-01 2.46160492e-01
5.96254366e-03 -7.53014326e-01 -3.26214842e-02 -1.11791119e-01
-1.18971038e+00 -1.13354552e+00 -4.58205760e-01 -9.49124336e-01
6.90356612e-01 5.24911225e-01 1.35636294e+00 1.06040502e+00
-1.90533087e-01 8.03243697e-01 -7.71935582e-02 -3.49600583e-01
-1.67232037e-01 9.21268761e-02 7.82236829e-02 -1.14261508e+00
6.71503067e-01 -8.39799047e-01 -5.19855738e-01 8.81344080e-02
-6.34538233e-01 5.05674124e-01 7.38968134e-01 7.79211998e-01
-2.39067823e-01 3.39164615e-01 8.04056883e-01 -5.11904001e-01
9.71619427e-01 -6.51704669e-01 -2.46541649e-01 4.96248975e-02
-6.20314062e-01 -5.95367863e-04 4.83233780e-01 -8.46940875e-01
-1.11041963e+00 -5.73235333e-01 -1.58303589e-01 3.40794474e-01
-4.40190703e-01 8.26194048e-01 1.50223628e-01 -1.94132790e-01
7.96330273e-01 4.10394847e-01 -1.27231658e-01 -1.40115455e-01
-6.67932406e-02 -4.40208018e-02 4.09453273e-01 -1.24616241e+00
9.43173707e-01 -1.60394982e-01 -2.54076779e-01 -6.56213880e-01
-1.00070488e+00 2.32792512e-01 -3.06031644e-01 -5.06421588e-02
8.39925408e-01 -7.21129954e-01 -1.56073189e+00 5.46119988e-01
-9.78933454e-01 -9.77003217e-01 -1.40675083e-01 7.44748116e-01
-5.72832525e-01 -1.59767047e-01 -7.87786007e-01 -4.88848597e-01
1.94784552e-02 -1.10787928e+00 -1.07066683e-01 8.60285819e-01
-8.81206632e-01 -1.04399502e+00 -1.84767693e-01 8.39195788e-01
6.30150855e-01 -1.27615914e-01 1.64721262e+00 -1.01643062e+00
-7.23338544e-01 1.42756402e-01 -2.84617931e-01 -1.09257936e-01
-3.12433898e-01 -7.41147161e-01 -4.03293520e-01 -8.46213773e-02
3.05396616e-01 -6.94145143e-01 7.12243140e-01 2.98353493e-01
1.01763046e+00 -2.42358342e-01 9.30506140e-02 6.34704292e-01
9.68088031e-01 4.98752207e-01 6.24908686e-01 4.32483882e-01
2.79131383e-01 6.44853830e-01 -1.17837839e-01 -2.44430378e-01
8.92448485e-01 4.00186032e-01 -1.86593458e-01 6.23219073e-01
-1.62377715e-01 -4.20933336e-01 2.00804725e-01 5.79288423e-01
-1.90916210e-01 2.03688934e-01 -1.26773322e+00 5.39109051e-01
-1.43011427e+00 -1.26280510e+00 3.40298228e-02 2.09075427e+00
1.20480394e+00 3.48260850e-01 2.13179648e-01 2.84340203e-01
1.45588517e-01 -2.55678236e-01 -9.24460471e-01 -6.15684688e-01
1.99237987e-01 7.51784861e-01 -2.76949704e-01 6.86276197e-01
-2.02054039e-01 1.10367191e+00 7.13488865e+00 3.06823045e-01
-5.68291962e-01 -1.55295342e-01 6.17083967e-01 2.72160340e-02
-4.09474641e-01 -3.85078073e-01 -3.96647424e-01 -1.31438836e-01
1.17810941e+00 -4.65750359e-02 1.07645273e+00 4.69816715e-01
-1.70020789e-01 -2.45491371e-01 -1.30744195e+00 8.13382804e-01
2.18931362e-02 -1.13411283e+00 -6.68237731e-02 -1.98681578e-01
7.74836719e-01 -6.29238069e-01 3.07833523e-01 9.93986428e-01
5.49776196e-01 -1.76606798e+00 4.06393200e-01 3.90710980e-01
2.22969919e-01 -6.01381004e-01 3.97146553e-01 6.57443225e-01
-5.28257132e-01 -3.25206250e-01 -3.40278149e-01 -1.38050711e+00
-4.62495059e-01 -2.93374866e-01 -9.22351778e-01 -3.50757003e-01
5.18428802e-01 2.56348044e-01 -6.90627933e-01 1.06929803e+00
-7.89713383e-01 5.05863607e-01 -1.37571305e-01 -4.19894606e-01
-1.43258125e-01 -2.68012332e-03 -8.79434124e-03 3.67495090e-01
4.87590224e-01 1.10959947e+00 -1.30972221e-01 1.13892686e+00
9.13473293e-02 -1.41689867e-01 -4.21419799e-01 -2.45935753e-01
7.19373882e-01 6.60180151e-01 -8.72032344e-01 -1.91983774e-01
-4.75997657e-01 9.02056277e-01 7.82045722e-01 5.89946508e-01
-3.06410998e-01 8.27210955e-03 7.64451563e-01 1.39081210e-01
-3.48743647e-02 -6.82217181e-01 -9.97090816e-01 -1.01770985e+00
-1.77481011e-01 -1.36268032e+00 2.51912415e-01 -1.15813971e+00
-1.10906756e+00 -1.25713021e-01 4.02695723e-02 -2.72819757e-01
-3.16268682e-01 -7.68844962e-01 -9.06106889e-01 9.33490217e-01
-7.98915863e-01 -6.52251244e-01 -4.94005650e-01 5.42627156e-01
7.09785104e-01 -3.16576242e-01 8.31475437e-01 -3.77168566e-01
-4.48085874e-01 5.44254959e-01 -4.43836391e-01 3.76491487e-01
1.02401689e-01 -1.27569997e+00 5.94862163e-01 5.03373265e-01
8.60514864e-02 1.29698515e+00 5.00862062e-01 -8.73141766e-01
-1.17576766e+00 -5.24414219e-02 6.15209579e-01 -3.86111230e-01
2.22541824e-01 -2.06795141e-01 -1.23543823e+00 8.70280683e-01
3.03341210e-01 -6.67664349e-01 9.00450289e-01 5.11547148e-01
-4.24637258e-01 2.29669899e-01 -1.32866168e+00 1.04848623e+00
1.58745575e+00 -3.02152216e-01 -1.50792527e+00 2.76429355e-01
6.89019203e-01 -3.66057277e-01 -7.83462226e-01 -8.26122053e-03
6.02298200e-01 -1.31733346e+00 1.23769462e+00 -9.65683520e-01
7.09569454e-01 1.36973634e-01 4.21813577e-01 -1.73337984e+00
-7.90566087e-01 -4.75784451e-01 -9.86120328e-02 7.33793139e-01
1.64997131e-01 -6.99984014e-01 9.56107914e-01 1.22552049e+00
-1.32957354e-01 -5.68602383e-01 -6.83952987e-01 -5.25760055e-01
7.38258362e-01 -4.98235583e-01 6.95214748e-01 9.58136499e-01
4.01189029e-01 1.44113228e-01 4.34939891e-01 3.38943638e-02
6.50506556e-01 -1.91875651e-01 4.99645591e-01 -1.77114344e+00
-2.29884401e-01 -9.33254421e-01 -1.22884974e-01 -8.89221489e-01
4.38972294e-01 -9.07736421e-01 -4.16088492e-01 -1.77549541e+00
1.88655138e-01 -3.22984278e-01 1.22059867e-01 5.08731008e-01
-4.31428522e-01 -1.11168638e-01 3.88841122e-01 -2.16014832e-01
-4.21340406e-01 7.40334168e-02 1.27280235e+00 1.24594547e-01
-1.76169977e-01 -9.57501233e-02 -1.50175154e+00 9.85216141e-01
1.18225968e+00 -2.11938411e-01 -7.46511400e-01 -6.56715631e-01
7.33260870e-01 1.63818434e-01 3.26870352e-01 -1.36112487e+00
4.01895106e-01 -6.61065340e-01 1.21913350e+00 -2.38463841e-03
1.28218248e-01 -6.30362272e-01 -9.53480154e-02 8.73982489e-01
-3.13193381e-01 3.66850197e-01 5.95313251e-01 -9.89450589e-02
3.72228354e-01 -6.60725534e-01 6.86510861e-01 -4.79428262e-01
-7.36122847e-01 -7.76454583e-02 -6.87801003e-01 5.25492728e-01
7.68414676e-01 -5.91683686e-01 -6.84013665e-01 -8.95523369e-01
-8.55407715e-01 2.57056296e-01 3.82628232e-01 3.15954924e-01
9.71632957e-01 -1.29986417e+00 -5.86100221e-01 3.41835350e-01
-4.04797494e-01 -1.41151622e-01 4.23761010e-01 6.36018455e-01
-8.39621246e-01 6.58645511e-01 -4.31635261e-01 1.74459517e-01
-6.19670630e-01 5.38153887e-01 5.06435394e-01 -2.16730937e-01
-5.45681536e-01 1.08577406e+00 2.91761398e-01 -4.53047156e-01
1.29629895e-01 -4.07058299e-01 -2.55202919e-01 -1.61317885e-01
1.05270362e+00 3.95162523e-01 -4.13140595e-01 7.28301033e-02
1.03694126e-01 4.61336792e-01 -2.85944462e-01 -8.05907249e-02
1.52922976e+00 1.40564948e-01 5.46520390e-02 1.61547512e-01
4.80938673e-01 -2.66337365e-01 -1.05084276e+00 -1.59012854e-01
-5.48933148e-02 -3.59999478e-01 -4.55104262e-01 -1.43766749e+00
-7.42917061e-01 9.09483075e-01 5.69047667e-02 -4.50190566e-02
9.47019219e-01 -1.49077430e-01 5.35457790e-01 8.10155213e-01
3.01954091e-01 -8.84785056e-01 5.09590328e-01 9.87426221e-01
9.52520609e-01 -1.29325640e+00 8.72209668e-02 1.25502571e-02
-2.42889777e-01 1.24825108e+00 1.17966640e+00 -2.44241521e-01
-4.65104245e-02 -2.31710792e-01 -4.32550430e-01 -3.27717751e-01
-8.79597247e-01 -1.54597253e-01 3.05271715e-01 1.12965834e+00
6.49784625e-01 -1.70662642e-01 -3.28936815e-01 1.23821723e+00
-9.06743526e-01 1.07369423e-01 7.53237128e-01 6.90895796e-01
-8.93966258e-01 -6.66540504e-01 -6.27875328e-01 5.47640264e-01
-4.66208495e-02 -4.15033847e-01 -3.37823570e-01 9.36682403e-01
1.09755799e-01 9.05500829e-01 1.19542286e-01 -1.61864385e-01
2.56774217e-01 5.83364546e-01 1.01572919e+00 -9.04597342e-01
-9.45055783e-01 -8.70955765e-01 -1.40524939e-01 -3.28115970e-01
2.17181042e-01 -9.32287276e-01 -1.43911529e+00 -7.62716115e-01
4.91703063e-01 -1.11203626e-01 4.10916954e-02 1.32010460e+00
-1.57796338e-01 6.39016330e-01 -4.78343993e-01 -5.81055880e-01
-5.76077044e-01 -1.08811641e+00 -3.45794439e-01 3.20028812e-01
1.70474648e-01 -6.62277579e-01 -3.37029219e-01 -4.52155262e-01] | [9.525124549865723, 7.265414237976074] |
aed5982b-c4df-4957-bcc9-ee3b5dd2a885 | hyperef-spectral-hypergraph-coarsening-by | 2210.14813 | null | https://arxiv.org/abs/2210.14813v2 | https://arxiv.org/pdf/2210.14813v2.pdf | HyperEF: Spectral Hypergraph Coarsening by Effective-Resistance Clustering | This paper introduces a scalable algorithmic framework (HyperEF) for spectral coarsening (decomposition) of large-scale hypergraphs by exploiting hyperedge effective resistances. Motivated by the latest theoretical framework for low-resistance-diameter decomposition of simple graphs, HyperEF aims at decomposing large hypergraphs into multiple node clusters with only a few inter-cluster hyperedges. The key component in HyperEF is a nearly-linear time algorithm for estimating hyperedge effective resistances, which allows incorporating the latest diffusion-based non-linear quadratic operators defined on hypergraphs. To achieve good runtime scalability, HyperEF searches within the Krylov subspace (or approximate eigensubspace) for identifying the nearly-optimal vectors for approximating the hyperedge effective resistances. In addition, a node weight propagation scheme for multilevel spectral hypergraph decomposition has been introduced for achieving even greater node coarsening ratios. When compared with state-of-the-art hypergraph partitioning (clustering) methods, extensive experiment results on real-world VLSI designs show that HyperEF can more effectively coarsen (decompose) hypergraphs without losing key structural (spectral) properties of the original hypergraphs, while achieving over $70\times$ runtime speedups over hMetis and $20\times$ speedups over HyperSF. | ['Zhuo Feng', 'Ali Aghdaei'] | 2022-10-26 | null | null | null | null | ['hypergraph-partitioning'] | ['graphs'] | [-1.85344853e-02 6.49567246e-01 -4.48226184e-01 3.48513603e-01
-5.66654563e-01 -4.54387039e-01 -2.14729339e-01 1.01729974e-01
2.33574808e-01 5.63823760e-01 1.75253730e-02 -3.75174880e-01
-5.59768617e-01 -1.18821371e+00 -4.49237972e-01 -8.65396440e-01
-4.74828869e-01 6.79191291e-01 3.89976561e-01 -2.17172548e-01
1.87683374e-01 4.74180758e-01 -1.02036703e+00 7.23241046e-02
1.04227757e+00 4.89795893e-01 -2.21850321e-01 5.93891799e-01
-1.01606093e-01 1.39340088e-01 -2.24269167e-01 -5.26610851e-01
4.02299643e-01 -2.82869160e-01 -9.15918827e-01 5.15313089e-01
3.85897517e-01 3.36607754e-01 -9.05867696e-01 1.38534653e+00
2.01517239e-01 -2.08594710e-01 5.81978321e-01 -1.21242917e+00
-4.67906862e-01 1.33649182e+00 -1.33509445e+00 -7.89613798e-02
3.26683968e-01 -1.28474191e-01 1.08973706e+00 -5.06987333e-01
5.16278684e-01 1.18453467e+00 7.87128985e-01 1.01737872e-01
-1.83563840e+00 -7.32566059e-01 -1.35828584e-01 -9.32536926e-03
-2.20131540e+00 -1.31670132e-01 7.15593994e-01 -2.04368681e-01
1.05161250e+00 6.36789560e-01 8.95501614e-01 2.28130549e-01
3.69348288e-01 1.73197553e-01 7.85788834e-01 -3.16824108e-01
2.52662212e-01 4.10725214e-02 6.80410266e-01 8.42438698e-01
9.80904758e-01 -2.13625237e-01 -2.11266696e-01 -5.44955611e-01
9.98384595e-01 -3.28534663e-01 -5.12490451e-01 -7.02797830e-01
-1.02872109e+00 6.23627901e-01 7.70812809e-01 3.36213082e-01
-3.44943047e-01 3.95432711e-01 4.27737504e-01 1.20466225e-01
-5.39703630e-02 2.24350214e-01 1.18147291e-01 2.77450502e-01
-9.46810722e-01 -1.35483563e-01 1.12504554e+00 1.39040792e+00
1.08079159e+00 2.93418676e-01 1.33411020e-01 5.33253253e-01
1.08552903e-01 6.91176355e-01 -4.43584740e-01 -7.94650853e-01
5.51166117e-01 1.14776278e+00 -3.76298636e-01 -1.39380479e+00
-6.94817305e-01 -5.84730983e-01 -1.57028425e+00 -4.25761819e-01
-2.13957369e-01 1.56063929e-01 -8.31253350e-01 1.39345479e+00
4.34430957e-01 3.32138121e-01 -3.58288623e-02 7.25132465e-01
4.68727738e-01 9.42561507e-01 -2.91672707e-01 -7.66955197e-01
1.28454578e+00 -6.04145110e-01 -5.45889616e-01 1.59303308e-01
5.75879693e-01 -4.73100334e-01 5.47178626e-01 4.28826600e-01
-1.34498072e+00 -3.61751825e-01 -1.27364361e+00 3.55501562e-01
5.25960959e-02 -3.00756067e-01 7.65128672e-01 1.31197631e+00
-1.42056286e+00 2.97764391e-01 -7.69274831e-01 -1.52183294e-01
-2.59799045e-02 7.20943928e-01 -4.73509431e-02 -1.92417815e-01
-9.98121440e-01 1.37503251e-01 5.24825394e-01 5.30887693e-02
-5.10287762e-01 -1.14429021e+00 -8.37059975e-01 4.52152520e-01
5.94872236e-01 -7.88985431e-01 2.12281615e-01 -2.92163789e-01
-1.01717961e+00 5.45750558e-01 9.94344652e-02 -2.57634491e-01
-2.12699622e-01 5.67570329e-01 -5.41735888e-01 4.83602792e-01
-1.16546072e-01 7.98074603e-02 4.85428512e-01 -1.41012156e+00
-6.99732453e-02 -5.72267294e-01 -1.44167259e-01 7.15945214e-02
-8.20209801e-01 -5.84117293e-01 -9.12517965e-01 -2.47744322e-01
5.33301532e-01 -1.22988725e+00 -7.67493010e-01 -1.04950976e+00
-9.51149106e-01 5.53881116e-02 4.21217144e-01 -1.71483070e-01
2.03050137e+00 -2.01822090e+00 6.77594304e-01 1.23660314e+00
1.05053186e+00 2.68847179e-02 1.38724251e-02 9.30589497e-01
-1.83665842e-01 8.88152048e-02 -2.84666102e-02 5.39370894e-01
4.05567640e-04 -1.03289209e-01 5.71184186e-03 6.87225103e-01
-7.00062454e-01 7.16674149e-01 -7.09520638e-01 -2.74372846e-01
1.33986205e-01 4.96033847e-01 -6.98736906e-01 -5.59523702e-01
4.00128305e-01 -4.06085730e-01 -6.19396567e-01 4.18019861e-01
1.31727612e+00 -6.81181490e-01 9.90246117e-01 -7.49065101e-01
1.04891852e-01 -3.89321059e-01 -1.80707014e+00 1.44254899e+00
-2.56287545e-01 -1.28768682e-02 6.36825144e-01 -8.47918272e-01
7.55150557e-01 1.09816633e-01 9.45859313e-01 -4.10528034e-01
3.89016084e-02 1.12008616e-01 -1.97443649e-01 9.48358700e-02
5.26578784e-01 1.71811670e-01 -4.22078073e-01 4.06148106e-01
-2.90767640e-01 -8.24108906e-03 6.05343580e-01 1.18561161e+00
1.62811375e+00 -9.06010926e-01 -4.30027619e-02 -1.24948704e+00
6.06126845e-01 2.02874333e-01 5.39386451e-01 4.01982099e-01
3.03426474e-01 1.58373624e-01 8.07386458e-01 -1.27403975e-01
-1.08573210e+00 -1.18938982e+00 -2.42524907e-01 3.31149489e-01
6.85148716e-01 -1.11900902e+00 -1.14014244e+00 -1.47953093e-01
1.69277087e-01 2.13119894e-01 -4.56328750e-01 -2.65707761e-01
-3.41238767e-01 -1.20919430e+00 4.69147265e-01 6.36256695e-01
1.52517751e-01 -8.31365511e-02 -7.10987821e-02 2.99738377e-01
3.72663699e-02 -1.03139079e+00 -5.44375420e-01 5.39813004e-02
-8.62306416e-01 -1.13907480e+00 -3.02737117e-01 -7.40341604e-01
1.04306829e+00 6.87250495e-01 9.99722362e-01 3.13705385e-01
-7.37436950e-01 3.88014972e-01 -1.35636538e-01 6.64626658e-01
-2.29377612e-01 2.82864660e-01 2.20462903e-01 -1.80625409e-01
5.05842566e-02 -7.12410867e-01 -6.40408993e-01 3.80733609e-01
-8.32366407e-01 1.05658904e-01 6.22992277e-01 5.86477458e-01
7.43611872e-01 8.35405469e-01 2.66771972e-01 -1.30381894e+00
3.59118611e-01 -4.18249518e-01 -8.39874983e-01 2.45595336e-01
-7.87434459e-01 1.99618146e-01 5.42998672e-01 -1.21145070e-01
-7.89793730e-01 -5.88941807e-03 1.16297744e-01 -3.60066801e-01
6.52039766e-01 4.17366952e-01 -2.95527011e-01 -4.97040570e-01
5.64838767e-01 -1.58551395e-01 -5.67459226e-01 4.05864138e-03
5.47383010e-01 2.18958557e-01 2.95813203e-01 -5.98797619e-01
1.20183277e+00 5.83332241e-01 5.91138303e-01 -9.65325594e-01
-1.18638858e-01 -6.93582773e-01 -5.84661424e-01 -2.06277847e-01
6.91183329e-01 -9.47436273e-01 -1.26626194e+00 -6.86440840e-02
-5.33960581e-01 -4.02634926e-02 -2.74425801e-02 9.62896124e-02
-2.60567576e-01 4.96004552e-01 -1.03626108e+00 -5.62273681e-01
-2.03000665e-01 -1.07641208e+00 9.34240103e-01 1.33195743e-02
-7.21571073e-02 -1.03117335e+00 5.25371805e-02 3.90025467e-01
2.56832242e-01 2.76455164e-01 1.43262315e+00 -3.27055901e-02
-9.76660848e-01 -5.10720052e-02 -4.20754284e-01 -3.38695139e-01
-2.64577121e-01 -3.71356569e-02 -2.91997582e-01 -7.99248874e-01
-3.18169713e-01 2.20550150e-02 8.82994533e-01 4.94714409e-01
8.13077033e-01 -1.94950685e-01 -1.03072250e+00 8.04060042e-01
2.08465219e+00 -1.33729249e-01 9.48259413e-01 -2.22722009e-01
1.13716352e+00 8.31449926e-02 1.45760953e-01 4.53282356e-01
2.87606150e-01 4.86640573e-01 1.12835221e-01 -2.35095695e-01
-1.77187875e-01 -1.62915453e-01 1.18106026e-02 1.31339800e+00
-2.98565589e-02 -2.32514098e-01 -9.26614404e-01 5.23050606e-01
-1.77622449e+00 -6.52459443e-01 -9.58548009e-01 2.19060636e+00
6.31307423e-01 2.60150135e-01 2.14093044e-01 3.65190476e-01
1.06277299e+00 4.63258214e-02 -5.42713761e-01 -3.40168417e-01
1.50732830e-01 9.11006704e-02 9.97209728e-01 6.91065609e-01
-5.11654854e-01 8.30429494e-01 6.37798977e+00 1.28076816e+00
-3.65130901e-01 1.11743892e-02 5.36278188e-01 2.23837197e-02
-7.41999447e-01 1.04293704e-01 -9.75559115e-01 -5.80729991e-02
9.56132472e-01 -5.99775851e-01 4.74553138e-01 6.94402337e-01
-2.64556617e-01 -7.72189647e-02 -7.54376590e-01 9.93293643e-01
-9.06222686e-03 -1.54912925e+00 -2.68156058e-03 5.89491427e-01
1.34132683e+00 -4.35400993e-01 -1.40573516e-01 5.52014522e-02
4.84433711e-01 -6.99005306e-01 -3.65168676e-02 -1.42058015e-01
1.01160407e+00 -1.38531804e+00 3.43318731e-01 -6.42926097e-02
-1.98122013e+00 -1.80732787e-01 -7.04580784e-01 5.11099994e-01
2.25815654e-01 1.15170002e+00 -4.45892274e-01 1.02604043e+00
5.21984160e-01 2.60607302e-01 -3.38134646e-01 8.38450551e-01
3.60551357e-01 4.39127117e-01 -3.13659012e-01 1.77211508e-01
5.06422594e-02 -5.70709944e-01 6.42144322e-01 1.18656647e+00
2.09395990e-01 6.11218512e-01 4.57308859e-01 8.63437176e-01
-2.73777634e-01 1.86994344e-01 -4.06332791e-01 -6.01614527e-02
6.96195006e-01 1.69156861e+00 -1.33839345e+00 -2.22505525e-01
-2.99316853e-01 7.45366871e-01 7.14218169e-02 5.80645025e-01
-9.56102610e-01 -7.51862288e-01 4.95716751e-01 6.51050270e-01
2.37469122e-01 -3.98281455e-01 -5.58097243e-01 -7.24715233e-01
-2.97420412e-01 -8.56807172e-01 4.08805847e-01 -2.44369596e-01
-9.80309010e-01 7.52293646e-01 -6.33574948e-02 -7.54996061e-01
3.85807455e-01 -3.04050624e-01 -2.72731602e-01 5.49609840e-01
-7.03616619e-01 -8.25320721e-01 -2.43037239e-01 6.30133688e-01
-2.24522385e-03 2.76573032e-01 3.61720026e-01 2.78556854e-01
-5.53365529e-01 5.53022802e-01 1.27268136e-01 -2.25322559e-01
2.43076980e-01 -1.16602135e+00 3.29277515e-01 7.57287145e-01
-1.86965987e-01 7.67293155e-01 6.07703090e-01 -1.00236762e+00
-2.49405694e+00 -9.25829470e-01 1.96124598e-01 2.37107594e-02
1.05457246e+00 -5.69559276e-01 -9.66306210e-01 4.00880784e-01
3.64170462e-01 -4.05529350e-01 5.77792108e-01 4.02269095e-01
-2.70307183e-01 -3.84592295e-01 -1.01407778e+00 5.81874669e-01
1.40570235e+00 -2.49249592e-01 2.41636693e-01 4.05669451e-01
7.65710831e-01 -4.05842453e-01 -1.26015055e+00 6.61639214e-01
1.06374063e-01 -9.09056962e-01 1.20126212e+00 2.01055389e-02
2.75873821e-02 -4.14453179e-01 4.43533901e-03 -1.10315073e+00
-1.00059259e+00 -1.10434413e+00 -1.89977303e-01 1.07941175e+00
4.11723197e-01 -4.21106905e-01 1.13723552e+00 5.26465476e-01
-1.55906547e-02 -6.53993428e-01 -5.85977793e-01 -8.01576853e-01
-2.35428423e-01 -3.44945155e-02 3.55307013e-01 9.81446803e-01
6.73000097e-01 7.64359295e-01 -1.02711424e-01 7.18477845e-01
1.19358873e+00 3.15323204e-01 5.30971885e-01 -1.21774995e+00
-3.67709547e-01 -5.42840242e-01 -6.75954163e-01 -9.24052179e-01
1.72918335e-01 -1.12299061e+00 -4.98650938e-01 -1.65794468e+00
8.74441624e-01 -5.05151451e-01 1.55420043e-02 1.80239171e-01
-1.19621232e-01 6.23629153e-01 -1.03574045e-01 -2.35105604e-01
-7.24765897e-01 2.58533001e-01 1.29457045e+00 -9.78130847e-02
-4.40467089e-01 -2.32218698e-01 -7.32054710e-01 3.38602215e-01
4.05835569e-01 -1.73886180e-01 -7.48712659e-01 -9.59157124e-02
6.52569115e-01 6.84906662e-01 -1.69800729e-01 -1.08418155e+00
4.83259469e-01 1.69235822e-02 1.97911225e-02 -7.51791179e-01
1.24378473e-01 -8.55427921e-01 8.27799678e-01 7.04928041e-01
5.37202135e-02 1.54157698e-01 4.49207425e-01 8.32576513e-01
9.34471413e-02 4.53893207e-02 8.38174164e-01 2.50934273e-01
-4.23211485e-01 3.90029669e-01 -3.63548994e-01 -6.15417175e-02
1.46069777e+00 -4.57930565e-01 -4.43299294e-01 -5.82609214e-02
-7.57621109e-01 5.20919800e-01 5.59415817e-01 -4.03183326e-02
5.56792200e-01 -1.32555997e+00 -6.63039505e-01 1.85168967e-01
-1.86102301e-01 -3.40178162e-01 8.62606585e-01 9.05658066e-01
-6.41338468e-01 5.21361589e-01 1.56175226e-01 -6.58254087e-01
-1.38151050e+00 9.39227760e-01 -6.23220066e-03 -4.41201240e-01
-7.89002657e-01 9.39422190e-01 5.66326559e-01 4.81343046e-02
1.14502534e-02 1.08996056e-01 4.23823953e-01 -2.48241857e-01
3.36934596e-01 9.47567403e-01 1.92139357e-01 -4.95713353e-01
-3.51962388e-01 7.69911706e-01 -2.31428936e-01 1.55147612e-01
1.11512327e+00 -3.12545836e-01 -5.86766958e-01 -1.42435297e-01
1.05796814e+00 3.98356229e-01 -5.66182315e-01 -6.72711954e-02
-1.96627960e-01 -5.48840225e-01 2.79261023e-01 -2.02949151e-01
-1.45524490e+00 4.54012245e-01 1.64843380e-01 7.17950284e-01
1.29713416e+00 1.30085364e-01 8.42731774e-01 2.12983534e-01
8.19452465e-01 -1.13140059e+00 -1.94236021e-02 3.80522124e-02
4.26936001e-01 -1.99868977e-01 3.56370747e-01 -1.54239643e+00
-4.44239587e-01 9.40641165e-01 7.24170268e-01 -8.77005458e-02
8.07358086e-01 9.03059304e-01 -6.14491165e-01 -6.65959358e-01
-5.60624778e-01 -3.51733640e-02 1.47597507e-01 4.52511847e-01
1.35219350e-01 3.33715379e-01 -3.37901533e-01 3.34210426e-01
-1.21179724e-03 -7.86858320e-01 6.76151276e-01 3.25510263e-01
-6.76598728e-01 -1.17776704e+00 -5.17283022e-01 4.89351034e-01
7.44927675e-02 -2.84015387e-01 -2.33831555e-01 6.93393290e-01
-4.88614619e-01 7.65753388e-01 -1.31689280e-01 -6.20372713e-01
2.77795106e-01 -5.37879467e-01 6.47583306e-01 -5.78196347e-01
-3.39657605e-01 5.80097616e-01 9.70968530e-02 -7.71417975e-01
1.94217935e-01 -2.93977916e-01 -1.60393476e+00 -8.07761192e-01
-5.00609696e-01 4.85012174e-01 2.22915798e-01 3.61240000e-01
4.23048377e-01 9.38241243e-01 9.11379099e-01 -6.24837399e-01
-2.09290445e-01 -2.47401968e-01 -1.23558533e+00 6.76611625e-03
-3.18525195e-01 -3.97286862e-01 -2.51678556e-01 -3.80650818e-01] | [7.02017068862915, 5.213601589202881] |
858fdeb7-c064-460e-99cb-ed6bc603fd38 | arbitrary-shape-text-detection-via | 2208.12419 | null | https://arxiv.org/abs/2208.12419v1 | https://arxiv.org/pdf/2208.12419v1.pdf | Arbitrary Shape Text Detection via Segmentation with Probability Maps | Arbitrary shape text detection is a challenging task due to the significantly varied sizes and aspect ratios, arbitrary orientations or shapes, inaccurate annotations, etc. Due to the scalability of pixel-level prediction, segmentation-based methods can adapt to various shape texts and hence attracted considerable attention recently. However, accurate pixel-level annotations of texts are formidable, and the existing datasets for scene text detection only provide coarse-grained boundary annotations. Consequently, numerous misclassified text pixels or background pixels inside annotations always exist, degrading the performance of segmentation-based text detection methods. Generally speaking, whether a pixel belongs to text or not is highly related to the distance with the adjacent annotation boundary. With this observation, in this paper, we propose an innovative and robust segmentation-based detection method via probability maps for accurately detecting text instances. To be concrete, we adopt a Sigmoid Alpha Function (SAF) to transfer the distances between boundaries and their inside pixels to a probability map. However, one probability map can not cover complex probability distributions well because of the uncertainty of coarse-grained text boundary annotations. Therefore, we adopt a group of probability maps computed by a series of Sigmoid Alpha Functions to describe the possible probability distributions. In addition, we propose an iterative model to learn to predict and assimilate probability maps for providing enough information to reconstruct text instances. Finally, simple region growth algorithms are adopted to aggregate probability maps to complete text instances. Experimental results demonstrate that our method achieves state-of-the-art performance in terms of detection accuracy on several benchmarks. | ['Xu-Cheng Yin', 'Jie-Bo Hou', 'Lei Chen', 'Xiaobin Zhu', 'Shi-Xue Zhang'] | 2022-08-26 | null | null | null | null | ['scene-text-detection'] | ['computer-vision'] | [ 3.67580175e-01 -2.98425853e-01 1.03205025e-01 -1.88787535e-01
-7.03334033e-01 -4.14959311e-01 4.39191014e-01 3.66167665e-01
-1.86516076e-01 4.36286002e-01 -3.63604248e-01 -1.08346932e-01
1.74557626e-01 -7.82716870e-01 -6.19810581e-01 -8.48508179e-01
6.09140217e-01 6.97449446e-01 9.39245224e-01 8.14447999e-02
3.09972465e-01 4.05012816e-01 -1.41689110e+00 2.31183767e-01
1.33866608e+00 9.18613970e-01 3.85111690e-01 3.85363281e-01
-6.45723581e-01 1.71973512e-01 -7.52089739e-01 -3.10628295e-01
-3.18957120e-02 -2.66603142e-01 -2.40306884e-01 4.00410980e-01
2.83411801e-01 -3.59521657e-01 -2.05058739e-01 1.27446365e+00
1.07767731e-01 1.44993290e-01 9.41938460e-01 -9.32912767e-01
-4.05202687e-01 7.30278134e-01 -9.49135602e-01 -6.57697162e-03
1.82918862e-01 8.28590319e-02 8.78333509e-01 -1.03028393e+00
2.19932303e-01 1.05430722e+00 7.67157793e-01 1.69600010e-01
-8.93712699e-01 -5.11330247e-01 5.15226960e-01 5.16578853e-02
-1.77187157e+00 7.67240077e-02 7.52542615e-01 -4.60041821e-01
3.47825319e-01 4.15721178e-01 4.69296843e-01 6.74544156e-01
1.15181379e-01 1.06336141e+00 9.80140448e-01 -2.85471201e-01
4.88403179e-02 2.97606677e-01 -4.00660113e-02 6.26162827e-01
3.71570230e-01 -5.80167592e-01 -3.20536703e-01 -1.10692764e-02
7.74188221e-01 7.07927421e-02 -3.58788937e-01 -2.15992585e-01
-1.28547609e+00 3.98221940e-01 4.65356141e-01 4.12561834e-01
-1.03959464e-01 -2.14802220e-01 2.42529765e-01 -4.40060109e-01
4.35151517e-01 -9.34719741e-02 -3.03030610e-01 8.90246853e-02
-1.16886079e+00 8.89832601e-02 4.55914825e-01 8.44540477e-01
8.19523811e-01 -4.14455086e-02 -3.94011170e-01 8.52375925e-01
3.13879400e-01 6.30739689e-01 2.93482661e-01 -2.12097540e-01
6.47810757e-01 1.00890636e+00 1.85418844e-01 -1.15142870e+00
-4.43379521e-01 -2.98866272e-01 -9.87217784e-01 -1.01500385e-01
7.69660354e-01 9.46298242e-02 -9.48240340e-01 9.22017992e-01
4.77199614e-01 7.22144246e-02 -3.80492628e-01 1.07419014e+00
5.81943870e-01 8.68401289e-01 -2.90239826e-02 -2.70481348e-01
1.36149764e+00 -5.93555033e-01 -6.94747984e-01 -3.10941070e-01
4.47374672e-01 -8.76832068e-01 1.24500608e+00 4.10072327e-01
-6.68025017e-01 -5.21535575e-01 -8.80795181e-01 1.90764695e-01
-2.95326978e-01 6.16901100e-01 1.66464686e-01 6.47413850e-01
-5.59138417e-01 1.49884388e-01 -8.58925045e-01 -1.74771026e-01
4.67028379e-01 1.05633661e-01 2.44670704e-01 -1.39584895e-02
-9.50968504e-01 3.29788446e-01 7.40580857e-01 3.08105677e-01
-2.77669638e-01 -2.74619401e-01 -5.37607253e-01 1.05145819e-01
5.68323791e-01 -2.33027413e-01 8.21451008e-01 -7.93755412e-01
-1.11960816e+00 6.00126028e-01 -1.33966491e-01 -5.37134819e-02
8.66205812e-01 4.38856930e-02 -2.11315855e-01 6.82738945e-02
1.56124800e-01 5.61007619e-01 1.11071169e+00 -1.36752927e+00
-9.52225268e-01 -2.70215362e-01 -3.70168209e-01 3.08079839e-01
-5.20613372e-01 -9.49616954e-02 -7.72131264e-01 -8.34033072e-01
5.61166346e-01 -5.44875562e-01 1.41635817e-02 1.29850864e-01
-7.53248453e-01 -3.64583939e-01 1.16463244e+00 -6.86152637e-01
1.53968179e+00 -2.06952906e+00 -6.64248392e-02 2.11009249e-01
1.26038268e-01 6.64791092e-02 4.32933807e-01 -1.25807598e-01
5.63101828e-01 2.36997589e-01 -5.41071236e-01 -1.47077203e-01
1.20619431e-01 -9.10066534e-03 -2.25665361e-01 3.60045791e-01
1.30495995e-01 5.81812322e-01 -6.51896894e-01 -1.03598726e+00
5.93112469e-01 5.25424659e-01 -1.15574837e-01 -3.04715615e-02
-5.64395726e-01 3.63111496e-01 -8.38526726e-01 9.57330108e-01
1.02891755e+00 -2.19344988e-01 -7.71702155e-02 -3.74880135e-01
-1.47390351e-01 -4.23397213e-01 -1.30624032e+00 1.13090599e+00
-9.39612091e-02 5.65527558e-01 9.11203027e-02 -8.78167093e-01
1.25347757e+00 -1.14419963e-02 3.34286362e-01 -2.29364187e-01
3.80192131e-01 3.68628293e-01 -2.17350587e-01 -2.47176513e-01
7.02232063e-01 1.18431309e-02 -7.89142996e-02 1.80730611e-01
-5.97796023e-01 -3.47322285e-01 1.89878270e-01 -8.38361830e-02
7.83572495e-01 1.79264992e-01 8.51517320e-02 -7.03643709e-02
7.18986452e-01 1.19648241e-01 5.95562220e-01 6.70266032e-01
-2.01958969e-01 9.37062562e-01 4.40823942e-01 -4.23789799e-01
-9.68415797e-01 -9.70403016e-01 -6.45554602e-01 9.61880863e-01
5.41547537e-01 -1.37057900e-01 -1.06733859e+00 -6.41844332e-01
-1.36967748e-01 5.86907268e-01 -3.46611232e-01 2.23840207e-01
-6.01900041e-01 -1.10014176e+00 4.29746658e-01 4.71253812e-01
8.08306217e-01 -1.10136652e+00 -3.66960526e-01 1.71983689e-01
-3.46565634e-01 -1.20965064e+00 -7.05914855e-01 2.58452166e-02
-7.97613144e-01 -9.56391454e-01 -9.31589663e-01 -8.06995809e-01
8.73205364e-01 3.18827659e-01 6.76818192e-01 1.16397224e-01
-2.93603897e-01 -4.12117690e-02 -4.74687994e-01 -1.46571234e-01
-4.35864091e-01 8.55100341e-03 -2.88814545e-01 2.94504791e-01
2.85001010e-01 -7.98832625e-03 -5.73435545e-01 7.55992472e-01
-1.05807173e+00 8.19197446e-02 6.72744274e-01 6.95443988e-01
1.01924086e+00 5.97710967e-01 2.29448825e-01 -7.74127543e-01
4.44727242e-01 -1.46686211e-01 -8.21926534e-01 4.55329269e-01
-3.59670699e-01 -1.96861759e-01 7.08550632e-01 -5.82855821e-01
-1.26839948e+00 3.25771481e-01 9.61405039e-02 -4.69174534e-01
-4.29397047e-01 2.12766781e-01 -3.16055924e-01 9.87711027e-02
4.40447569e-01 6.76789641e-01 -4.24029291e-01 -1.97478548e-01
1.27658665e-01 9.47123528e-01 4.82400894e-01 -5.94550431e-01
7.19397843e-01 5.07638752e-01 -1.85179472e-01 -9.14478898e-01
-7.74906397e-01 -5.34620583e-01 -7.70441949e-01 -3.58321607e-01
8.33753288e-01 -6.63750410e-01 -4.32523817e-01 8.55699480e-01
-1.06390870e+00 -2.83615470e-01 9.04654935e-02 1.34742498e-01
-1.44697770e-01 7.28397250e-01 -5.45068324e-01 -8.76012981e-01
-2.15308473e-01 -1.20410645e+00 1.51371729e+00 6.05643928e-01
1.76186770e-01 -8.78469348e-01 -4.86420065e-01 4.54520702e-01
6.66758493e-02 1.85207147e-02 7.22691894e-01 -3.77908856e-01
-8.53957891e-01 -4.38110054e-01 -7.01515257e-01 1.22003583e-02
1.38644770e-01 3.61234039e-01 -7.64424145e-01 -4.46604379e-02
-1.67545170e-01 1.01914898e-01 8.47984910e-01 6.29868388e-01
1.51517248e+00 -3.22213508e-02 -7.38644898e-01 4.19847459e-01
1.19491351e+00 2.67975871e-02 4.24603254e-01 3.31527621e-01
1.02465951e+00 3.93209606e-01 9.78862703e-01 7.69830227e-01
2.28335530e-01 6.15114450e-01 3.38294089e-01 5.09169288e-02
-2.22033411e-02 -1.97301552e-01 9.49631780e-02 5.59864223e-01
1.21001467e-01 -6.57561779e-01 -9.95025218e-01 3.54563922e-01
-1.90790057e+00 -5.09484589e-01 -6.96817696e-01 2.18801355e+00
8.02407742e-01 5.75572670e-01 2.39267782e-03 2.14401588e-01
1.39454842e+00 1.29594803e-02 -7.13652670e-01 8.65878984e-02
-2.48523161e-01 -2.94691056e-01 4.98717904e-01 2.59431064e-01
-1.24070883e+00 1.07059026e+00 5.19106674e+00 1.43002093e+00
-9.46924865e-01 -1.49515033e-01 8.37683201e-01 3.85895759e-01
-2.04087466e-01 -2.69747525e-01 -1.18644893e+00 7.10053086e-01
2.34653857e-02 8.97666067e-02 7.09731951e-02 5.99649072e-01
2.83425152e-01 -3.57650071e-01 -5.21762073e-01 9.32096004e-01
6.12963326e-02 -8.21082592e-01 1.78545713e-01 -2.55153716e-01
9.73975301e-01 -3.27351987e-01 -3.89518067e-02 1.00650869e-01
7.04571530e-02 -7.73348987e-01 7.87684321e-01 4.68272626e-01
8.64681900e-01 -5.62942386e-01 7.87970543e-01 7.11570859e-01
-1.53136706e+00 1.58785120e-01 -6.01236701e-01 2.78638065e-01
5.44140162e-03 9.02231336e-01 -9.71000195e-01 2.61217684e-01
6.95806146e-01 5.52378356e-01 -5.80119789e-01 1.18247926e+00
-3.43050957e-01 6.26148224e-01 -7.18922436e-01 -3.96606952e-01
1.30395174e-01 -2.64023036e-01 4.78561044e-01 1.11502671e+00
4.04778689e-01 5.49277514e-02 5.32798171e-01 1.15468693e+00
-7.42436573e-02 3.71131957e-01 -7.20844492e-02 2.69630671e-01
7.01046169e-01 1.39334249e+00 -1.52336371e+00 -2.37738013e-01
-3.93140376e-01 8.32664192e-01 4.77360487e-02 3.84397209e-01
-9.57932234e-01 -3.38127762e-01 -9.91353393e-02 2.38232270e-01
3.91993374e-01 1.64999231e-03 -7.60661423e-01 -9.92217302e-01
2.96319723e-01 -4.40948248e-01 2.93925613e-01 -7.36000896e-01
-1.12741768e+00 4.26322877e-01 -1.42876804e-01 -1.30640543e+00
4.42938149e-01 -5.70067108e-01 -6.89132094e-01 6.71491206e-01
-1.26027453e+00 -1.15091133e+00 -7.38906205e-01 2.77060509e-01
9.57323015e-01 2.41584852e-01 2.59431601e-01 -2.42387019e-02
-9.32578444e-01 6.39674604e-01 3.96953821e-01 2.67110467e-01
6.15362644e-01 -1.19336748e+00 9.80370194e-02 8.52583230e-01
4.27716002e-02 -3.27024683e-02 6.43519521e-01 -8.68505359e-01
-9.86621857e-01 -1.20066273e+00 3.34728211e-01 -3.20830941e-01
6.72235608e-01 -2.92438745e-01 -1.33158743e+00 4.71952170e-01
-3.73980135e-01 2.36289456e-01 -6.51813895e-02 -2.01828659e-01
2.42909007e-02 -4.40298282e-02 -1.07055700e+00 6.73493803e-01
6.75550163e-01 -1.76946297e-01 -3.74542505e-01 4.35273260e-01
3.92123014e-01 -6.39756441e-01 -6.68930531e-01 4.48705435e-01
3.28371435e-01 -1.04909325e+00 7.61632740e-01 5.06054401e-01
1.51552841e-01 -6.07018948e-01 1.12097263e-01 -8.23253334e-01
6.66315900e-03 -2.05789059e-01 1.77706052e-02 1.38017368e+00
3.26096207e-01 -5.39684713e-01 1.15101051e+00 3.85000706e-01
-1.71197191e-01 -7.89099693e-01 -1.02549791e+00 -4.79579687e-01
7.29344785e-02 -4.03308958e-01 5.41539550e-01 7.46750534e-01
-1.00138322e-01 -8.28502029e-02 1.32771479e-02 4.31100696e-01
6.60143614e-01 3.17811936e-01 5.85408092e-01 -1.31936336e+00
-1.01568349e-01 -7.87218571e-01 -1.87000021e-01 -1.39455473e+00
-8.21854845e-02 -5.81891179e-01 5.51986814e-01 -1.57334876e+00
3.47045392e-01 -9.79880333e-01 -6.27257973e-02 3.48919243e-01
-7.46135175e-01 2.41724163e-01 -1.36740893e-01 5.19547820e-01
-7.85842001e-01 5.46282709e-01 1.43094015e+00 -3.58623356e-01
-2.97470987e-01 2.65072048e-01 -1.65265977e-01 9.36047375e-01
8.36682379e-01 -3.62952620e-01 -5.79723865e-02 -2.52343625e-01
2.50051379e-01 5.30456705e-03 2.28901148e-01 -1.06929410e+00
3.30689400e-01 -2.11031422e-01 5.40698767e-01 -1.09482133e+00
9.22711417e-02 -7.82925725e-01 -2.17564166e-01 1.96156234e-01
6.00356758e-02 -5.78638256e-01 2.07600534e-01 7.19483018e-01
-1.50516927e-01 -4.49805796e-01 7.28501856e-01 2.72490494e-02
-6.18798435e-01 2.04530001e-01 -4.60487396e-01 5.97965531e-02
1.05753636e+00 -6.18433416e-01 -1.15027480e-01 6.93775192e-02
-5.04402578e-01 3.37112635e-01 7.32551873e-01 1.51489556e-01
6.76778853e-01 -9.90981221e-01 -7.03554451e-01 1.34723052e-01
1.21927388e-01 6.97449028e-01 3.12464237e-01 8.71517837e-01
-6.97012067e-01 2.37043560e-01 3.23766351e-01 -1.14870667e+00
-1.16312480e+00 3.69832069e-01 2.83420831e-01 -2.20077857e-01
-6.74353480e-01 7.27726340e-01 4.66762543e-01 -4.07964230e-01
1.98598787e-01 -6.06798589e-01 -1.72188699e-01 -3.24836969e-02
2.83184379e-01 3.64283234e-01 -5.48173115e-02 -5.86637318e-01
-1.66285545e-01 9.28981543e-01 -2.07788214e-01 4.74625766e-01
5.92719555e-01 -2.87277341e-01 1.85078561e-01 4.83230442e-01
4.39032584e-01 1.95536744e-02 -1.62143469e+00 -2.56642103e-01
-2.09719658e-01 -4.50199664e-01 1.49849340e-01 -6.58409297e-01
-9.91441727e-01 8.86969507e-01 3.45296860e-01 4.21199083e-01
1.05381775e+00 1.09545797e-01 8.02302897e-01 2.43944436e-01
1.14917412e-01 -1.30020905e+00 4.57407013e-02 2.82452881e-01
3.80194962e-01 -1.36529195e+00 1.76114678e-01 -8.58106077e-01
-5.44292927e-01 1.36192250e+00 8.52754951e-01 1.72067121e-01
3.83225173e-01 4.24249500e-01 1.72788382e-03 3.89426537e-02
-5.71730658e-02 -6.69718981e-02 2.78590173e-01 4.42415267e-01
2.79780596e-01 2.43307978e-01 -2.06057087e-01 5.38233697e-01
1.01534590e-01 -3.86795551e-01 4.08254057e-01 5.33750296e-01
-9.09819961e-01 -7.10792243e-01 -1.02177715e+00 7.77650774e-01
-3.37619781e-01 -9.13206767e-03 -3.32065284e-01 7.41064310e-01
1.19402245e-01 7.42940366e-01 2.09505171e-01 -4.06958684e-02
1.78450301e-01 -1.19487837e-01 2.52864599e-01 -4.41115201e-01
-2.10829750e-01 4.17936265e-01 -4.20307726e-01 1.48036927e-01
-7.34782740e-02 -8.89480710e-01 -1.81092179e+00 -2.01521695e-01
-8.69087517e-01 -1.05326779e-01 5.25371373e-01 1.02899861e+00
-1.09483272e-01 6.10300422e-01 5.18090189e-01 -8.38576913e-01
-3.65525663e-01 -1.01695597e+00 -6.92871988e-01 2.52080560e-01
-2.34466299e-01 -6.23524070e-01 -5.36587596e-01 9.21335220e-02] | [12.0801420211792, 2.2722771167755127] |
8e9fb8b9-8b6d-47c0-84b9-74a009ed6bd9 | towards-cover-song-detection-with-siamese | 2005.10294 | null | https://arxiv.org/abs/2005.10294v1 | https://arxiv.org/pdf/2005.10294v1.pdf | Towards Cover Song Detection with Siamese Convolutional Neural Networks | A cover song, by definition, is a new performance or recording of a previously recorded, commercially released song. It may be by the original artist themselves or a different artist altogether and can vary from the original in unpredictable ways including key, arrangement, instrumentation, timbre and more. In this work we propose a novel approach to learning audio representations for the task of cover song detection. We train a neural architecture on tens of thousands of cover-song audio clips and test it on a held out set. We obtain a mean precision@1 of 65% over mini-batches, ten times better than random guessing. Our results indicate that Siamese network configurations show promise for approaching the cover song identification problem. | ['Marko Stamenovic'] | 2020-05-20 | null | null | null | null | ['cover-song-identification'] | ['music'] | [ 5.81406713e-01 -3.42697024e-01 6.62841797e-02 -9.47680231e-03
-1.40588152e+00 -1.09480202e+00 1.38940349e-01 -5.45692742e-02
-1.49240747e-01 8.01733375e-01 1.45397350e-01 1.62919879e-01
4.53877784e-02 -3.91340584e-01 -1.10188007e+00 -5.95439613e-01
-5.96979499e-01 4.92223591e-01 -1.25533089e-01 -8.69407356e-02
1.89896032e-01 3.96162510e-01 -1.56882691e+00 5.63573778e-01
-1.91713721e-01 1.09251690e+00 -5.56452833e-02 1.02831447e+00
3.13643873e-01 5.55759072e-01 -1.61072350e+00 -1.68787017e-01
3.84313822e-01 -6.50875986e-01 -5.36128461e-01 1.01364277e-01
6.20713830e-01 -3.93720746e-01 -2.61972278e-01 7.82241940e-01
4.72240150e-01 -1.60526618e-01 4.43499058e-01 -7.50430107e-01
-4.39169556e-01 1.26812804e+00 -3.48113775e-01 4.91707325e-01
6.19831383e-01 1.90089010e-02 1.11759436e+00 -4.84483182e-01
3.69987845e-01 6.37757838e-01 9.45670605e-01 1.17080897e-01
-1.49371457e+00 -1.06046855e+00 -1.99163452e-01 -1.43586561e-01
-1.84505284e+00 -6.22201741e-01 8.46710145e-01 -4.48651016e-01
7.83724844e-01 6.34093285e-01 7.97238767e-01 1.17745733e+00
1.10798895e-01 7.07575321e-01 6.03625357e-01 -3.84487599e-01
2.86557794e-01 1.71663135e-01 -3.79907429e-01 1.80094063e-01
1.50830135e-01 1.39995351e-01 -6.89124048e-01 -5.93998671e-01
4.10986900e-01 -2.09504113e-01 -3.75348359e-01 7.28662163e-02
-1.30825222e+00 6.20078623e-01 8.63836110e-02 3.39965820e-01
-2.51956701e-01 7.61002600e-01 4.78706717e-01 7.43037760e-01
2.67065287e-01 1.14310753e+00 -5.63694000e-01 -3.77989382e-01
-1.39005673e+00 3.47192794e-01 9.93233085e-01 7.95891941e-01
4.62504417e-01 6.62004709e-01 9.29955170e-02 6.36475980e-01
-3.99315566e-01 4.83391166e-01 7.72042215e-01 -8.24097335e-01
3.43753934e-01 -1.68513492e-01 2.47997895e-01 -8.99342060e-01
-1.34715080e-01 -7.82258153e-01 -6.40406013e-01 -1.18359521e-01
2.40974262e-01 -2.12397456e-01 -6.93676233e-01 1.39002895e+00
-2.39139840e-01 5.12602210e-01 -3.83116543e-01 5.85766375e-01
5.49772263e-01 6.73434854e-01 -4.91178781e-01 -3.11960816e-01
8.40496957e-01 -6.96469009e-01 -5.05359650e-01 -1.06562644e-01
-9.64122638e-03 -9.27912295e-01 7.04380631e-01 9.49212849e-01
-8.40173662e-01 -6.06360376e-01 -1.40816379e+00 8.93050313e-01
-2.27969512e-01 1.79417327e-01 3.40857387e-01 8.32938313e-01
-7.48846769e-01 9.30875659e-01 -3.87345374e-01 3.04294556e-01
4.22889560e-01 5.32789230e-01 -1.82185844e-01 4.05955344e-01
-1.22542858e+00 1.09015785e-01 6.03734612e-01 -4.49270830e-02
-1.26351178e+00 -7.55457163e-01 -3.77266467e-01 1.35821611e-01
4.82074350e-01 1.92974769e-02 1.39749181e+00 -1.25616121e+00
-1.25622249e+00 7.26488531e-01 2.78514743e-01 -7.83537567e-01
1.86832249e-01 -2.54385382e-01 -9.09838676e-01 7.58122355e-02
-3.85566987e-02 1.34604335e-01 1.49837673e+00 -8.79700959e-01
-4.09897000e-01 1.57185812e-02 -1.34860024e-01 -2.45664820e-01
-2.04524383e-01 -1.52103161e-03 -7.47343227e-02 -1.14877725e+00
-2.28904009e-01 -1.11715567e+00 2.02063248e-01 -5.95422864e-01
-7.21168995e-01 3.78998935e-01 4.47516620e-01 -5.12248874e-01
1.47789526e+00 -2.51257229e+00 -1.73659503e-01 3.34326178e-01
-1.27520040e-01 1.79402694e-01 -1.16568968e-01 6.46094501e-01
-3.94663900e-01 3.80923659e-01 6.80714287e-03 -3.95939648e-02
-1.26263320e-01 -8.25091675e-02 -8.77143204e-01 5.62035620e-01
-1.09970081e-03 5.55291116e-01 -5.54830074e-01 1.32273242e-01
-2.98118055e-01 4.61087257e-01 -3.91537428e-01 8.41676518e-02
-3.20278764e-01 3.19510341e-01 -7.18813390e-02 8.99747670e-01
2.74129868e-01 -5.98390251e-02 1.38227165e-01 1.01221092e-01
2.24286780e-01 4.13704634e-01 -1.49811089e+00 1.83467972e+00
-1.88543469e-01 1.00248909e+00 -2.84740385e-02 -7.16137767e-01
9.64267910e-01 5.68247914e-01 4.79743481e-01 4.63145413e-02
2.22960874e-01 1.86966553e-01 -5.10093337e-03 -5.86141795e-02
6.13854468e-01 -1.26047388e-01 -4.11224455e-01 8.74054015e-01
3.00049949e-02 -2.45587736e-01 -1.00579880e-01 -2.50077546e-01
1.16922486e+00 -2.95941800e-01 4.87184078e-01 1.42136037e-01
8.01027268e-02 -1.37004718e-01 3.09749484e-01 1.12928092e+00
6.85411543e-02 8.79343390e-01 3.37381184e-01 -6.63525224e-01
-1.01299000e+00 -9.15690541e-01 -2.25670382e-01 1.39628696e+00
-2.45987669e-01 -5.88159382e-01 -5.13133883e-01 -2.89639533e-01
1.99139401e-01 3.45622241e-01 -6.70475721e-01 -1.36213914e-01
-5.99089026e-01 -3.81471246e-01 1.10471022e+00 4.14146364e-01
-6.45683892e-03 -1.33327949e+00 -6.27506435e-01 5.39499640e-01
1.54121727e-01 -7.52864838e-01 -8.54782522e-01 4.50094074e-01
-5.74762225e-01 -7.51997828e-01 -5.06537437e-01 -5.93954861e-01
3.27984244e-02 1.65543649e-02 1.32539654e+00 -1.92924368e-03
-3.76689225e-01 4.99995910e-02 -4.63490754e-01 -5.68372428e-01
-5.46950638e-01 4.06723738e-01 5.18639207e-01 1.15297884e-01
1.84870541e-01 -9.07188475e-01 -1.86869711e-01 5.68320714e-02
-9.53077674e-01 -5.76383412e-01 4.35334682e-01 9.15621400e-01
8.40431690e-01 6.37233555e-01 4.42967266e-01 -9.37719107e-01
7.01714039e-01 -3.92020017e-01 -4.62930679e-01 -6.40934845e-03
-3.18217993e-01 -1.54858187e-01 6.62824392e-01 -9.88455117e-01
-3.22861830e-03 4.33331937e-01 1.66935697e-01 -7.95686066e-01
-3.51464987e-01 3.09252679e-01 -4.88877371e-02 -7.87250623e-02
8.49225819e-01 6.33717358e-01 -4.02084827e-01 -8.01787078e-01
1.53480917e-01 7.62957275e-01 7.49361575e-01 -4.85905617e-01
1.02131724e+00 2.96505660e-01 -5.31498373e-01 -6.92502201e-01
-1.00391841e+00 -2.32805610e-01 -6.07393444e-01 -7.40643814e-02
1.97099581e-01 -1.11281180e+00 -4.78060186e-01 3.24285507e-01
-8.99438262e-01 -4.69223969e-02 -5.29403985e-01 2.76132524e-01
-4.17254716e-01 -1.72309175e-01 -2.58678824e-01 -8.36450279e-01
-3.98034304e-01 -7.49347687e-01 1.21015775e+00 -4.27178219e-02
-5.31869471e-01 -4.08563226e-01 3.68969142e-01 1.02467053e-01
3.41518432e-01 5.52410603e-01 4.08644408e-01 -9.96896684e-01
-5.44208944e-01 -8.56663644e-01 4.10326689e-01 5.10202050e-01
3.70379180e-01 -7.42317215e-02 -1.19843030e+00 -7.03501999e-01
-1.04605937e-02 -3.77225518e-01 8.73113155e-01 1.67772174e-01
1.31121850e+00 -7.90234089e-01 -3.05987056e-02 7.59869099e-01
1.46932578e+00 4.60308194e-01 4.22492027e-01 3.06511700e-01
5.00576913e-01 -7.81851038e-02 2.14641318e-01 7.11139500e-01
-5.37527800e-01 6.75204754e-01 3.65477085e-01 2.87435234e-01
6.43377081e-02 -4.76533979e-01 4.89948869e-01 7.23911881e-01
-1.64429434e-02 -2.51742750e-01 -5.80554366e-01 5.46972096e-01
-1.11131060e+00 -1.24514842e+00 5.90114176e-01 2.45286608e+00
8.30361247e-01 4.76002127e-01 5.32985985e-01 6.25151753e-01
7.83026099e-01 1.80066615e-01 -8.06977272e-01 -4.52918649e-01
-2.12232158e-01 4.05429423e-01 6.52656496e-01 1.11612178e-01
-1.32325101e+00 5.67099571e-01 7.93931341e+00 9.90222931e-01
-1.16702139e+00 -1.37532637e-01 4.46761936e-01 -7.10901380e-01
-2.32855342e-02 -3.41415346e-01 -5.57602763e-01 6.13714099e-01
1.08430934e+00 -2.52914637e-01 1.08607268e+00 7.67333448e-01
-2.69623280e-01 2.23822787e-01 -1.23414576e+00 1.13599455e+00
4.65568691e-01 -1.38567650e+00 -1.06402397e-01 2.40545496e-01
8.17232668e-01 1.76317513e-01 3.07174623e-01 2.59428889e-01
-1.17807582e-01 -1.32314456e+00 9.30851579e-01 2.86328346e-01
1.21219730e+00 -9.49464977e-01 6.52772486e-01 2.16847256e-01
-1.17980373e+00 -7.65803382e-02 -8.17634389e-02 -1.45539761e-01
-1.81542948e-01 1.79292053e-01 -1.15111041e+00 1.22138768e-01
7.13837624e-01 5.78979969e-01 -5.32768011e-01 1.09158015e+00
9.69325826e-02 1.16950047e+00 -4.80247438e-01 -1.17854066e-02
-3.76994796e-02 3.90534490e-01 8.96109521e-01 1.31095552e+00
3.03088635e-01 -3.85142565e-01 1.68315694e-01 6.90501153e-01
-3.46071720e-01 -7.24156499e-02 -6.78476453e-01 -4.52909946e-01
6.59141779e-01 8.85488272e-01 -4.64328110e-01 -7.51203969e-02
2.84759641e-01 8.89063120e-01 -2.11525410e-01 1.95554137e-01
-9.06743944e-01 -6.15369081e-01 6.80879056e-01 3.10419053e-01
9.86643195e-01 6.71867728e-02 1.92262262e-01 -9.69286144e-01
2.70559974e-02 -1.17469561e+00 2.29774684e-01 -6.53734326e-01
-1.24730432e+00 1.02614236e+00 -3.36659253e-01 -1.53487098e+00
-6.74122274e-01 -3.35586190e-01 -4.03789401e-01 5.50200939e-01
-6.55937433e-01 -6.70592964e-01 2.18645155e-01 2.74758995e-01
4.46835130e-01 -8.36756289e-01 1.20200050e+00 1.90326035e-01
-2.60215908e-01 8.19911480e-01 6.28859222e-01 3.56125981e-01
5.47894835e-01 -1.12310672e+00 4.90828216e-01 3.73208195e-01
1.13905180e+00 3.16160053e-01 1.07205832e+00 -2.65002221e-01
-1.26635659e+00 -9.94454026e-01 4.28747356e-01 -5.59804797e-01
8.19012642e-01 -5.59825718e-01 -8.29846323e-01 6.77046895e-01
-5.38014546e-02 -3.43435317e-01 1.13563406e+00 1.76088512e-01
-7.26425886e-01 -4.40759569e-01 -8.79449189e-01 -2.71523651e-02
6.04525387e-01 -9.12073076e-01 -6.41320825e-01 3.42598081e-01
8.91492963e-01 -5.35684526e-01 -7.95454800e-01 -7.24442257e-03
9.17936742e-01 -8.33392978e-01 1.01017678e+00 -7.17808485e-01
3.20159346e-01 -4.07901198e-01 -5.08762002e-01 -1.23936021e+00
-2.28934661e-01 -1.01277852e+00 -4.15923715e-01 1.22009635e+00
4.50747728e-01 -2.10147947e-01 8.46848667e-01 -6.21509962e-02
2.86062837e-01 -4.90362912e-01 -1.03702724e+00 -1.07335961e+00
-2.28298292e-01 -8.33755255e-01 1.02771854e+00 7.94110835e-01
-2.25514755e-01 3.31101477e-01 -1.03174865e+00 1.19206794e-01
5.54076195e-01 5.11582077e-01 7.22486317e-01 -1.23309243e+00
-1.11449456e+00 -4.27919865e-01 -6.77911997e-01 -6.89760506e-01
-2.12565064e-02 -6.07290745e-01 8.05812031e-02 -4.09169376e-01
-1.02331698e-01 -2.06956014e-01 -6.99336529e-01 4.28648591e-01
3.89134943e-01 9.45986748e-01 1.19496219e-01 4.06546712e-01
-6.38639033e-01 1.27534643e-01 6.23044848e-01 -6.80076182e-01
-4.45854545e-01 5.29552460e-01 -8.20552826e-01 4.72152174e-01
8.90460014e-01 -1.01996231e+00 -1.51528478e-01 -3.59221905e-01
3.68839234e-01 1.54439077e-01 1.10068902e-01 -1.35977006e+00
-8.34547356e-02 1.00605354e-01 4.42385286e-01 -5.39637208e-01
4.99961376e-01 -7.10462332e-01 6.28914952e-01 2.90034980e-01
-5.91652572e-01 -1.97916761e-01 4.28152442e-01 9.16931808e-01
-3.27417493e-01 -2.38988310e-01 3.23139757e-01 -1.53628206e-02
-8.95823017e-02 -5.47068613e-03 -4.88172024e-01 5.80032542e-02
5.55507958e-01 -3.05443019e-01 1.57658756e-01 -6.45231783e-01
-8.94955993e-01 -4.70784575e-01 2.18003914e-01 5.28765380e-01
4.35469002e-01 -1.38027418e+00 -9.21546459e-01 3.62402409e-01
1.83489993e-01 -4.62035865e-01 -5.89346401e-02 2.15938419e-01
-4.02022153e-01 3.25413674e-01 -1.46770906e-02 -4.12661523e-01
-1.35960746e+00 5.14290214e-01 2.91619956e-01 -4.09868583e-02
-4.81488138e-01 1.02676499e+00 -3.15634638e-01 1.12188719e-01
5.51984727e-01 -1.14792950e-01 2.43857149e-02 2.88997814e-02
9.38030601e-01 4.43358272e-02 3.81287858e-02 -6.80285573e-01
-3.96227181e-01 3.50059330e-01 -2.18007900e-02 -1.11757070e-01
1.31307650e+00 2.39575252e-01 2.19468802e-01 1.13314188e+00
1.25569415e+00 3.29445451e-01 -9.45066988e-01 -1.19994551e-01
-3.92871261e-01 -5.49630761e-01 -2.17086464e-01 -9.29025590e-01
-1.07580614e+00 2.94661462e-01 6.56100154e-01 6.45842671e-01
1.16988635e+00 1.03754722e-01 7.32628345e-01 7.09926486e-01
6.35676146e-01 -8.68542433e-01 -1.38448864e-01 1.65649489e-01
1.03871500e+00 -8.09516907e-01 1.41192183e-01 4.71183628e-01
-4.51213807e-01 9.32641089e-01 -7.60636404e-02 -4.28015709e-01
7.36814499e-01 3.96505356e-01 4.43532970e-03 3.21340896e-02
-7.32929885e-01 2.04860523e-01 5.39194703e-01 4.56452876e-01
3.19366783e-01 3.43578905e-01 5.34749448e-01 8.28520656e-01
-7.65070856e-01 -2.13680640e-01 5.09815454e-01 6.34398699e-01
-4.70739335e-01 -7.99472034e-01 -6.00235164e-01 6.58027589e-01
-7.90801346e-01 -1.94766849e-01 -7.28966951e-01 6.00044966e-01
4.60484624e-01 7.80446291e-01 1.88156754e-01 -1.03519750e+00
1.80346817e-01 1.49666355e-03 3.42644483e-01 -7.01272964e-01
-9.46719587e-01 3.41333807e-01 2.77244039e-02 -2.96985954e-01
-2.58427739e-01 -8.15298855e-01 -7.50005841e-01 -2.76815742e-01
-3.80725533e-01 4.30158794e-01 5.26488423e-01 7.59870648e-01
1.96966872e-01 6.34827673e-01 9.62520599e-01 -8.73652279e-01
-5.27224422e-01 -1.09736979e+00 -1.16057742e+00 1.28584430e-01
7.06776142e-01 -2.48514578e-01 -6.64880991e-01 3.49398851e-01] | [15.734756469726562, 5.325026512145996] |
a0529885-8335-47fc-9f2a-d57019f3bd92 | audio-visual-speech-and-gesture-recognition | null | null | https://www.mdpi.com/1424-8220/23/4/2284 | https://www.mdpi.com/1424-8220/23/4/2284/pdf?version=1676649264 | Audio-Visual Speech and Gesture Recognition by Sensors of Mobile Devices | Audio-visual speech recognition (AVSR) is one of the most promising solutions for reliable speech recognition, particularly when audio is corrupted by noise. Additional visual information can be used for both automatic lip-reading and gesture recognition. Hand gestures are a form of non-verbal communication and can be used as a very important part of modern human–computer interaction systems. Currently, audio and video modalities are easily accessible by sensors of mobile devices. However, there is no out-of-the-box solution for automatic audio-visual speech and gesture recognition. This study introduces two deep neural network-based model architectures: one for AVSR and one for gesture recognition. The main novelty regarding audio-visual speech recognition lies in fine-tuning strategies for both visual and acoustic features and in the proposed end-to-end model, which considers three modality fusion approaches: prediction-level, feature-level, and model-level. The main novelty in gesture recognition lies in a unique set of spatio-temporal features, including those that consider lip articulation information. As there are no available datasets for the combined task, we evaluated our methods on two different large-scale corpora—LRW and AUTSL—and outperformed existing methods on both audio-visual speech recognition and gesture recognition tasks. We achieved AVSR accuracy for the LRW dataset equal to 98.76% and gesture recognition rate for the AUTSL dataset equal to 98.56%. The results obtained demonstrate not only the high performance of the proposed methodology, but also the fundamental possibility of recognizing audio-visual speech and gestures by sensors of mobile devices. | ['Elena Ryumina', 'Denis Ivanko', 'Dmitry Ryumin'] | 2023-02-17 | null | null | null | sensors-2023-2 | ['sign-language-recognition', 'gesture-recognition', 'audio-visual-speech-recognition'] | ['computer-vision', 'computer-vision', 'speech'] | [ 2.31869057e-01 -2.92693108e-01 -3.09500694e-01 -1.69139266e-01
-1.17344701e+00 -2.50855148e-01 7.53928304e-01 -2.28465438e-01
-5.94123900e-01 2.94527560e-01 2.72172004e-01 -1.84412763e-01
-3.77167240e-02 -2.11845309e-01 -2.90082157e-01 -9.97527719e-01
2.20509455e-01 3.04104954e-01 1.52982771e-01 1.32942870e-01
4.91457395e-02 8.48986685e-01 -2.23489618e+00 4.66828853e-01
3.19665104e-01 1.39378500e+00 3.79224449e-01 8.81987929e-01
-3.70335668e-01 3.22009683e-01 -4.43287402e-01 -4.01420183e-02
-1.42639533e-01 -1.97747290e-01 -3.17104191e-01 1.24894127e-01
2.16039076e-01 -2.37838134e-01 -2.94352621e-01 6.75927997e-01
9.73788142e-01 2.37012237e-01 6.83532894e-01 -1.35074735e+00
-1.48810834e-01 2.36620098e-01 -4.12102938e-01 -1.43775046e-01
5.39656043e-01 4.40125704e-01 6.42785788e-01 -1.12852716e+00
3.93553287e-01 1.28141809e+00 2.52764732e-01 8.99296045e-01
-8.57507944e-01 -5.58678687e-01 -7.93745369e-02 6.40020967e-01
-1.46838284e+00 -1.04555154e+00 7.48197019e-01 -4.34995055e-01
1.07988703e+00 5.30176044e-01 5.90752661e-01 1.20314395e+00
-2.38804027e-01 1.04099572e+00 9.78765249e-01 -5.81332386e-01
1.66612506e-01 9.11579579e-02 -2.18996629e-02 3.22207928e-01
-2.71843940e-01 3.79487425e-01 -9.55908656e-01 -3.50863226e-02
4.50979948e-01 4.30703815e-03 -2.58876115e-01 -2.68269449e-01
-1.15649116e+00 5.09454250e-01 -1.64076257e-02 6.43607736e-01
-4.83399540e-01 -2.98739541e-02 4.53694791e-01 1.50540188e-01
9.66894850e-02 -4.70603943e-01 -2.38808632e-01 -5.66710889e-01
-1.06384766e+00 -2.46577576e-01 4.54084516e-01 5.30216396e-01
1.53252140e-01 2.81806052e-01 -2.92118609e-01 1.21049726e+00
8.58017623e-01 8.11853945e-01 9.60717440e-01 -4.13668811e-01
4.89689410e-01 3.24380308e-01 -1.87414438e-01 -5.74937999e-01
-4.04684275e-01 1.57331690e-01 -8.16823244e-01 5.33742309e-01
5.37846148e-01 2.53824711e-01 -1.08694994e+00 1.55443537e+00
3.79181802e-01 1.78711340e-01 1.28068686e-01 1.01510775e+00
1.39839208e+00 5.17489910e-01 3.22450191e-01 -3.94841373e-01
1.54815209e+00 -8.79186988e-01 -8.94812703e-01 2.99730804e-02
2.05702603e-01 -8.85134935e-01 1.12159729e+00 3.47113431e-01
-1.07383287e+00 -6.20600402e-01 -8.79590213e-01 6.09445572e-03
-2.86859840e-01 5.71352661e-01 1.21464439e-01 8.59428525e-01
-1.17087698e+00 -1.53968304e-01 -7.24083483e-01 -5.79229414e-01
1.58227712e-01 5.17974675e-01 -6.79090798e-01 2.13295266e-01
-7.64916718e-01 5.94210625e-01 6.88759610e-02 3.18603545e-01
-8.11491966e-01 -3.68407741e-02 -8.05990458e-01 1.32393092e-01
2.68694967e-01 -2.33671576e-01 1.26501369e+00 -8.99458826e-01
-1.93798029e+00 8.97118211e-01 -7.05019712e-01 -2.52167255e-01
5.90764761e-01 1.62719890e-01 -6.94222093e-01 3.56137812e-01
-5.81018031e-01 6.59764290e-01 1.33157635e+00 -1.08201861e+00
-5.41443050e-01 -5.26819825e-01 -6.50297165e-01 3.29842627e-01
-3.39125127e-01 2.45280057e-01 -5.32089412e-01 -4.12175059e-01
2.18307659e-01 -5.88763952e-01 5.01326501e-01 1.69978291e-01
-1.39004529e-01 -4.79846358e-01 1.05835342e+00 -8.52173030e-01
1.01986396e+00 -2.29611349e+00 -4.19122502e-02 1.92754760e-01
-9.91304144e-02 9.56812501e-01 -3.64309520e-01 3.20722163e-01
7.17061087e-02 -5.26975244e-02 6.48431405e-02 -6.63956821e-01
5.33406325e-02 -3.31616029e-02 -1.79182187e-01 3.20322871e-01
-6.11217506e-02 9.09872711e-01 -3.65006596e-01 -4.79579568e-01
7.88667917e-01 9.88151431e-01 3.11274081e-02 3.04445088e-01
2.16896273e-02 4.44759995e-01 -1.69424564e-01 9.64994252e-01
4.98989969e-01 3.22774470e-01 4.37301546e-02 -7.73356855e-02
-7.83122107e-02 2.33796611e-01 -1.31692863e+00 1.46261251e+00
-6.77650392e-01 8.23233545e-01 4.45348144e-01 -8.02013814e-01
1.06379640e+00 7.84177899e-01 3.79915476e-01 -7.51345277e-01
2.55174011e-01 4.75975931e-01 -1.89246237e-01 -8.75760674e-01
2.57530510e-01 -3.59615944e-02 3.57368350e-01 3.05572659e-01
1.31346807e-01 5.26480600e-02 -2.36945629e-01 -3.18940103e-01
6.01319313e-01 -1.83539107e-01 2.65115738e-01 3.26283813e-01
1.07477796e+00 -5.89556277e-01 8.49518478e-02 4.98307705e-01
-3.48715007e-01 6.90035582e-01 6.82092011e-02 -7.69465938e-02
-5.46148717e-01 -1.01456594e+00 -2.63914205e-02 9.87786889e-01
-1.46331161e-01 -1.18693195e-01 -6.49539769e-01 -4.84873295e-01
-6.26928061e-02 3.37502956e-01 -1.77687764e-01 2.18338937e-01
-2.27720931e-01 -1.18488893e-01 8.07210386e-01 5.05136549e-01
4.04300809e-01 -1.50781250e+00 -4.86505836e-01 4.68115136e-02
-1.77996546e-01 -1.21106052e+00 -3.38535458e-01 -1.79602370e-01
-5.72781980e-01 -1.03103590e+00 -1.01070499e+00 -8.92570436e-01
1.87361911e-01 2.79762268e-01 3.62368613e-01 7.07621723e-02
-3.04882437e-01 7.12124586e-01 -3.12828869e-01 -3.82228255e-01
-5.86753726e-01 -1.41110584e-01 2.17781261e-01 4.66457307e-01
5.73454320e-01 -3.61888051e-01 -3.75324070e-01 5.40479124e-01
-7.27813244e-01 -2.76465267e-01 6.60549641e-01 8.11856747e-01
5.81929624e-01 -4.58282143e-01 5.54148376e-01 2.91364163e-01
6.58032358e-01 -4.73696217e-02 -5.23958087e-01 3.80397081e-01
-3.26321393e-01 -2.64937937e-01 1.66129082e-01 -7.22377300e-01
-9.58078921e-01 3.44723314e-01 -8.19483399e-01 -5.40886343e-01
-7.41265714e-01 2.67469853e-01 -4.85445350e-01 -6.65667504e-02
1.85473874e-01 5.27441204e-01 4.63601917e-01 -8.14852118e-01
1.76248655e-01 1.48526943e+00 5.52501500e-01 -1.92073643e-01
2.82717139e-01 3.64874035e-01 -1.07633315e-01 -1.57603741e+00
3.00690919e-01 -8.05807531e-01 -5.41827202e-01 -4.36763585e-01
8.32094729e-01 -7.68795609e-01 -1.27178562e+00 1.07602465e+00
-1.09292996e+00 -2.10872695e-01 -2.03189075e-01 7.44194567e-01
-5.56825936e-01 4.63454068e-01 -2.24157110e-01 -1.47700930e+00
-4.12591696e-01 -1.57342017e+00 1.41144633e+00 9.52765569e-02
-5.89446314e-02 -6.02325559e-01 -1.92804679e-01 6.10197425e-01
4.89324629e-01 -2.38196850e-01 5.01198709e-01 -7.66690254e-01
-2.91712463e-01 -2.94039607e-01 -2.32269242e-01 4.29488748e-01
7.70900771e-02 -1.05680218e-02 -1.43794417e+00 -1.55518264e-01
-1.77224234e-01 -4.37265515e-01 9.96712983e-01 5.19003928e-01
8.88655365e-01 -6.87360689e-02 -2.03919694e-01 2.77889937e-01
9.84639823e-01 6.00405276e-01 7.03346848e-01 -1.73091918e-01
5.37934601e-01 7.25655854e-01 5.39493084e-01 2.14333475e-01
1.66533515e-01 1.29068828e+00 2.98859626e-01 -7.52058299e-03
-5.72066963e-01 -2.51772940e-01 5.53446114e-01 7.96160042e-01
-8.52568746e-02 -3.16093504e-01 -9.41824496e-01 5.02002895e-01
-1.73840940e+00 -9.95770574e-01 -6.14035726e-02 2.36850405e+00
3.64468366e-01 -3.98435652e-01 4.27764028e-01 7.22144961e-01
7.52448738e-01 1.30047634e-01 -3.63783032e-01 -4.24729675e-01
-6.04388192e-02 3.38564217e-01 -8.75637904e-02 6.83491826e-01
-1.05348718e+00 7.78891683e-01 5.40150213e+00 1.06742465e+00
-1.55158746e+00 1.92144185e-01 1.97417051e-01 -3.03644538e-01
9.44506451e-02 -6.70523345e-01 -8.36819649e-01 3.92118365e-01
9.70696509e-01 4.55441862e-01 3.88421148e-01 7.08101094e-01
4.83329594e-01 -1.76751152e-01 -9.07309413e-01 1.54006279e+00
2.88902462e-01 -1.03645253e+00 2.01864745e-02 1.36841372e-01
-5.61059751e-02 -6.29329961e-03 1.78370640e-01 3.94861354e-03
-5.50175488e-01 -1.15439045e+00 7.94579387e-01 4.29233521e-01
1.22766793e+00 -5.50679624e-01 6.12238944e-01 4.12043303e-01
-1.53530359e+00 -1.62807897e-01 1.42999321e-01 2.26028189e-01
2.99715608e-01 8.85102805e-03 -8.46758544e-01 1.90218180e-01
6.53504670e-01 3.50911409e-01 -1.46576300e-01 1.03529561e+00
-3.19728494e-01 4.83789295e-01 -4.76952225e-01 -3.97699863e-01
-7.55026340e-02 2.99669325e-01 6.41038239e-01 1.22497809e+00
5.29840767e-01 -1.11704215e-01 -1.50545090e-01 3.81377757e-01
1.17958255e-01 3.17831963e-01 -6.25554979e-01 -1.10471599e-01
5.63874006e-01 1.08482575e+00 -5.34693182e-01 -6.66863769e-02
-3.54902655e-01 5.49689770e-01 -2.83811063e-01 3.71778935e-01
-4.82266814e-01 -2.82215297e-01 8.32547367e-01 3.39610055e-02
3.98786455e-01 -3.47237200e-01 -1.69157013e-01 -9.24754262e-01
2.72525489e-01 -1.04224980e+00 1.33823946e-01 -7.41298914e-01
-7.94471622e-01 6.48496449e-01 -1.31043881e-01 -1.25119293e+00
-7.41878808e-01 -8.57468963e-01 -3.90850127e-01 9.77466345e-01
-1.46133590e+00 -1.40190983e+00 -4.17279124e-01 9.04512882e-01
8.06046128e-01 -5.22444785e-01 9.64218438e-01 2.93669432e-01
-2.71442652e-01 8.89553070e-01 1.38023525e-01 5.48034795e-02
4.23135698e-01 -5.25140524e-01 1.58138245e-01 6.77137792e-01
4.97459471e-01 3.23244870e-01 2.47858226e-01 -2.87127674e-01
-1.48295474e+00 -5.38602948e-01 1.13248658e+00 -8.42499286e-02
3.58550638e-01 -3.95305544e-01 -8.23849857e-01 1.48211822e-01
5.03933337e-03 -1.28350541e-01 6.68872118e-01 -2.60368824e-01
-1.60868317e-01 -1.39605343e-01 -1.31859446e+00 3.63475293e-01
8.46427798e-01 -8.62363279e-01 -4.18103606e-01 -6.77261204e-02
3.16542149e-01 -2.11753845e-01 -4.34794039e-01 3.11106890e-01
1.10846996e+00 -9.08426523e-01 1.03716338e+00 -1.80456534e-01
-1.94486067e-01 -2.98276216e-01 -3.72147828e-01 -9.35974956e-01
4.47525203e-01 -5.48401237e-01 -3.19664150e-01 1.44733346e+00
3.08558017e-01 -7.19834268e-01 6.69440508e-01 2.98739284e-01
1.57891333e-01 -5.83677232e-01 -1.56598890e+00 -7.86404431e-01
-4.26976174e-01 -9.60841775e-01 5.90453923e-01 2.66154826e-01
5.26680201e-02 4.90311198e-02 -2.89665520e-01 -1.03533253e-01
4.50668901e-01 -2.92440265e-01 7.41726518e-01 -1.22290969e+00
-1.63273707e-01 -6.48695886e-01 -7.25095093e-01 -9.66022551e-01
2.55805254e-01 -6.35753870e-01 9.25801918e-02 -1.52021444e+00
-8.91657844e-02 -1.06282540e-01 -1.10548906e-01 5.41475356e-01
4.01692182e-01 3.63443702e-01 4.06015664e-01 2.05095947e-01
-1.04415528e-01 5.70376694e-01 8.43149304e-01 -3.20269436e-01
-4.16263551e-01 3.38834345e-01 -1.95082147e-02 6.33065343e-01
6.08860493e-01 1.44338456e-03 -3.30312222e-01 -2.12843210e-01
-5.56971550e-01 3.01778793e-01 5.29527009e-01 -8.82508636e-01
3.54153484e-01 1.03816725e-01 9.41673480e-03 -7.40266144e-01
9.72108245e-01 -8.78144562e-01 -3.02413404e-02 3.24556470e-01
-2.75764227e-01 -3.66151750e-01 2.56423891e-01 3.98668230e-01
-4.40102041e-01 -4.46301699e-02 7.51387715e-01 3.21633816e-01
-8.76978159e-01 -2.32812893e-02 -7.60842741e-01 -4.28901434e-01
1.00336790e+00 -4.68632847e-01 -2.55651206e-01 -5.49079776e-01
-1.04652333e+00 -2.31911421e-01 3.66116203e-02 7.72926867e-01
1.05360055e+00 -1.27491391e+00 -5.21188319e-01 7.11922050e-01
1.79710582e-01 -4.30077404e-01 4.35746670e-01 1.04472959e+00
-1.83740839e-01 7.37350821e-01 -1.71062440e-01 -9.25947845e-01
-2.08210754e+00 3.42491388e-01 2.90570408e-01 1.24934852e-01
-4.48514938e-01 6.91574633e-01 -2.06479207e-01 -2.19283894e-01
9.99709964e-01 -2.93544829e-01 -4.60107356e-01 2.12113023e-01
8.64362895e-01 4.60194379e-01 3.17567289e-01 -1.26012325e+00
-6.04338825e-01 8.23616922e-01 3.77522498e-01 -5.46557486e-01
1.11175442e+00 -1.84442639e-01 2.79541731e-01 5.08863032e-01
1.15094531e+00 -1.37405410e-01 -8.70562553e-01 -1.75338894e-01
-2.04802856e-01 -3.56343180e-01 1.75895780e-01 -9.66857910e-01
-9.32463646e-01 1.59060228e+00 1.13191044e+00 3.70280184e-02
1.26261699e+00 7.40187168e-02 6.96275413e-01 2.46384546e-01
3.91232044e-01 -9.90672112e-01 -3.74634899e-02 2.40455836e-01
1.16493821e+00 -1.21393371e+00 -4.92864758e-01 -1.99478298e-01
-6.75893903e-01 1.05526555e+00 1.80964693e-01 6.75570190e-01
6.19524956e-01 3.67578298e-01 3.78515571e-01 2.68864721e-01
-5.89625478e-01 -5.30109763e-01 7.49200940e-01 9.25308108e-01
4.03800815e-01 1.27418473e-01 -3.06593090e-01 5.86911857e-01
1.72656730e-01 8.87978449e-02 -1.37179434e-01 7.46338010e-01
-3.88882130e-01 -1.05129075e+00 -6.80363178e-01 1.61456168e-01
-2.96175838e-01 5.84374182e-02 -4.94798899e-01 6.88399136e-01
-1.04106076e-01 1.39916217e+00 -1.31336331e-01 -7.04568446e-01
3.15618664e-01 5.08276701e-01 3.93950582e-01 -1.34100944e-01
-5.47290802e-01 3.94224793e-01 -4.73397374e-02 -5.75261712e-01
-4.86509711e-01 -8.50540757e-01 -1.22440147e+00 -5.80272302e-02
-2.87730664e-01 -1.14417449e-01 1.26468635e+00 1.02378631e+00
3.51941079e-01 2.30453849e-01 4.09038007e-01 -1.33261442e+00
-5.59164762e-01 -1.09355330e+00 -5.82625031e-01 1.90551758e-01
5.78341961e-01 -8.12883079e-01 -4.27409202e-01 -1.43780693e-01] | [14.313926696777344, 5.044588565826416] |
da83c443-1e6c-43fb-a0b4-a4f061dcdf4c | a-qualitative-investigation-of-optical-flow | 2204.08791 | null | https://arxiv.org/abs/2204.08791v1 | https://arxiv.org/pdf/2204.08791v1.pdf | A qualitative investigation of optical flow algorithms for video denoising | A good optical flow estimation is crucial in many video analysis and restoration algorithms employed in application fields like media industry, industrial inspection and automotive. In this work, we investigate how well optical flow algorithms perform qualitatively when integrated into a state of the art video denoising algorithm. Both classic optical flow algorithms (e.g. TV-L1) as well as recent deep learning based algorithm (like RAFT or BMBC) will be taken into account. For the qualitative investigation, we will employ realistic content with challenging characteristic (noisy content, large motion etc.) instead of the standard images used in most publications. | ['Hannes Fassold'] | 2022-04-19 | null | null | null | null | ['video-denoising'] | ['computer-vision'] | [-1.21304236e-01 -4.97231096e-01 2.85717577e-01 3.80736552e-02
-9.36413705e-02 -3.39123487e-01 5.18157244e-01 -1.58748373e-01
-3.26137722e-01 1.27828455e+00 2.83065110e-01 -1.64045542e-01
-2.05100432e-01 -4.72439826e-01 -6.86693192e-01 -8.55737925e-01
-1.41965926e-01 -1.72022596e-01 3.15383941e-01 -3.70344311e-01
5.54720521e-01 5.02598226e-01 -1.41569805e+00 5.02483454e-03
6.74639463e-01 9.71944690e-01 1.13286830e-01 7.68881917e-01
8.71744007e-03 1.13599408e+00 -3.57231647e-01 -4.68648493e-01
4.18561727e-01 -5.74482083e-01 -9.78538990e-01 2.75940359e-01
6.72612906e-01 -6.45181537e-01 -5.88427782e-01 1.14737022e+00
3.24079752e-01 4.44744140e-01 7.06062794e-01 -1.03590631e+00
-4.81475532e-01 1.54286087e-01 -6.28802419e-01 8.21135521e-01
5.46962023e-01 3.99121910e-01 5.03377318e-01 -7.84160078e-01
1.09288633e+00 1.27499557e+00 5.10576546e-01 3.34264398e-01
-1.01619673e+00 -2.88066387e-01 -1.47014499e-01 9.58632171e-01
-8.77406716e-01 -6.45959318e-01 1.04625440e+00 -5.54856241e-01
5.25505602e-01 1.24851596e-02 6.96798027e-01 1.06744063e+00
7.28427947e-01 8.92165482e-01 1.06038165e+00 -2.44439945e-01
1.29023716e-01 -1.48898080e-01 1.67702571e-01 5.38978636e-01
1.67714462e-01 1.70113236e-01 -4.67228293e-01 2.50019044e-01
7.64754534e-01 -2.97971457e-01 -4.80539262e-01 -2.44377762e-01
-1.17276144e+00 6.59916341e-01 3.21430176e-01 5.35603583e-01
-4.80618596e-01 3.30688179e-01 6.46788359e-01 6.44895911e-01
3.46551716e-01 -1.85256135e-02 -7.72797316e-02 -3.11635643e-01
-1.05790222e+00 4.32032585e-01 6.30822778e-01 7.70000041e-01
9.21897054e-01 5.94100773e-01 -1.49400085e-01 4.77239043e-01
3.54620248e-01 1.87720433e-01 4.45221990e-01 -1.51991260e+00
9.97647569e-02 -1.28597274e-01 3.09708506e-01 -1.18219256e+00
-1.98053122e-01 -4.57189590e-01 -1.16160190e+00 4.52116489e-01
5.27585387e-01 -1.07185937e-01 -5.99277556e-01 1.32636893e+00
2.57219523e-01 7.35476553e-01 1.03723727e-01 1.09437656e+00
8.46181989e-01 6.60648704e-01 -2.61811614e-01 -5.35211861e-01
1.01434898e+00 -9.08405781e-01 -1.11142182e+00 1.33873582e-01
5.04996181e-01 -1.33539546e+00 4.15474534e-01 9.22516704e-01
-1.37218440e+00 -9.89030004e-01 -9.96212959e-01 -1.88668102e-01
-1.04639582e-01 -4.04481202e-01 5.81294715e-01 5.42644620e-01
-1.30552399e+00 9.43369448e-01 -6.82035685e-01 -4.79419231e-01
4.78982419e-01 1.14589920e-02 -6.51271880e-01 -3.76658082e-01
-1.22003734e+00 8.98980856e-01 1.98218644e-01 4.79124755e-01
-1.16902101e+00 -5.53145051e-01 -7.57645786e-01 -3.24528575e-01
6.18673638e-02 -9.64055598e-01 1.08427477e+00 -1.07838798e+00
-1.45925283e+00 8.04539263e-01 -1.59270346e-01 -5.23355007e-01
9.39307332e-01 -3.71266365e-01 -2.74852037e-01 4.08041924e-01
-2.00659037e-01 6.08550668e-01 1.21338296e+00 -1.21606588e+00
-4.68635798e-01 -1.59552563e-02 2.26270065e-01 9.11335088e-03
1.61204129e-01 6.29855990e-02 2.38706768e-01 -6.13909900e-01
-2.27906406e-02 -5.40616572e-01 -2.25715861e-02 1.95505306e-01
-2.77108461e-01 1.79228466e-02 1.25525224e+00 -9.67289448e-01
8.69216442e-01 -1.75868106e+00 3.78610998e-01 -1.80578500e-01
1.32350832e-01 5.01613975e-01 3.79252955e-02 5.89353740e-01
-1.97908536e-01 -2.53289223e-01 -1.91764578e-01 -3.31276923e-01
-3.83939594e-01 2.99335748e-01 4.60335873e-02 1.07770407e+00
5.65753952e-02 6.35686338e-01 -9.19238448e-01 -6.30644619e-01
8.45536768e-01 6.39988482e-01 -6.03915691e-01 7.62703419e-02
1.80782825e-01 9.98543084e-01 -2.84321994e-01 5.28041184e-01
9.78551745e-01 4.41022128e-01 -5.39180577e-01 -7.46720910e-01
-4.01134402e-01 -2.42340684e-01 -1.33669102e+00 1.72867346e+00
-3.10597688e-01 9.61423874e-01 4.69362736e-01 -1.43110168e+00
6.87671244e-01 4.11798567e-01 7.93341160e-01 -4.88635987e-01
4.59002018e-01 1.93283886e-01 2.11540043e-01 -9.42663789e-01
5.82403600e-01 -2.70615339e-01 8.21066022e-01 2.59308785e-01
4.09854680e-01 5.21219932e-02 7.34024704e-01 1.06041498e-01
1.03393364e+00 2.79972017e-01 1.84472963e-01 -5.64434528e-01
1.19165075e+00 -1.69291258e-01 6.13508403e-01 5.60798764e-01
-9.05882001e-01 7.18757868e-01 4.45716530e-01 -6.14335120e-01
-1.32453310e+00 -6.01509809e-01 -2.43146256e-01 3.47807854e-01
1.75243750e-01 6.32871548e-03 -6.70448303e-01 -1.93867356e-01
-2.39496753e-01 8.71820375e-02 -2.96030074e-01 -8.36765319e-02
-9.82053161e-01 -5.87602675e-01 1.66894555e-01 1.78427443e-01
7.77745187e-01 -1.40333521e+00 -4.51318860e-01 5.20757198e-01
-1.64615333e-01 -1.37911963e+00 -2.16324627e-01 -3.12630624e-01
-9.46526408e-01 -1.10053682e+00 -9.39152539e-01 -9.20524895e-01
1.57066375e-01 2.87578672e-01 1.31750202e+00 2.21773148e-01
-3.93645555e-01 3.85317892e-01 -4.33424175e-01 2.59653151e-01
-6.98848009e-01 -3.24145824e-01 4.33498994e-02 3.53637934e-01
-1.73116904e-02 -7.95540035e-01 -1.02704549e+00 1.16881631e-01
-1.21903396e+00 -2.62057364e-01 3.46413851e-01 9.32577193e-01
8.73248726e-02 8.99632499e-02 2.19776809e-01 -7.46902287e-01
7.21296668e-01 -3.59622538e-01 -4.44153637e-01 -1.59607634e-01
-2.10268185e-01 -2.02584043e-01 8.28497648e-01 -1.35547981e-01
-1.06332433e+00 -2.08161056e-01 -4.52048570e-01 -6.92809939e-01
-3.69084328e-01 3.77020001e-01 1.78412482e-01 -5.27962685e-01
4.49415177e-01 4.40005183e-01 9.43987668e-02 -3.42514992e-01
8.98349509e-02 4.54039812e-01 6.81131601e-01 -3.04387301e-01
8.87184799e-01 7.21733034e-01 3.39479327e-01 -1.17332959e+00
-3.83069456e-01 -5.35295069e-01 -7.54007220e-01 -6.63308740e-01
8.71532559e-01 -5.89480042e-01 -8.07876945e-01 8.58242154e-01
-1.43227601e+00 -1.07377199e-02 -5.25143780e-02 7.23593354e-01
-8.57653260e-01 8.46275032e-01 -9.90390301e-01 -8.17437410e-01
-2.42806777e-01 -1.39646494e+00 7.41649866e-01 3.70058298e-01
8.68675262e-02 -1.41329908e+00 2.68088520e-01 3.34541082e-01
4.93896931e-01 4.11620438e-01 4.62152511e-01 1.45563051e-01
-6.92712665e-01 2.28304207e-01 -9.50855613e-02 7.78492987e-01
1.13739103e-01 5.09094000e-01 -9.34415936e-01 -4.01534796e-01
3.42288941e-01 -1.73908278e-01 1.02809012e+00 9.11737204e-01
6.81318402e-01 1.94537684e-01 2.66337752e-01 7.48980522e-01
1.73368204e+00 2.87847996e-01 1.08853614e+00 2.67726183e-01
7.54718065e-01 8.69244874e-01 4.28763688e-01 2.93216944e-01
-2.36464590e-01 4.10432398e-01 6.37813747e-01 1.37958974e-01
-3.10611755e-01 3.24129432e-01 3.69774789e-01 1.06486428e+00
-4.09292430e-01 -3.80184144e-01 -4.93950218e-01 5.14927447e-01
-1.85692644e+00 -1.24164331e+00 -5.29724240e-01 1.92774034e+00
2.90053129e-01 1.92838803e-01 -1.38731241e-01 7.85895765e-01
6.94893181e-01 4.83622104e-01 -1.71921536e-01 -6.88652754e-01
-3.31778556e-01 1.28016517e-01 3.15103561e-01 6.74702883e-01
-1.28376889e+00 6.39323473e-01 6.03658867e+00 7.78038502e-01
-1.07163072e+00 2.33953804e-01 5.65176606e-01 4.29448396e-01
9.43751633e-02 -9.97416675e-02 -3.40615839e-01 4.86720920e-01
6.26993060e-01 3.27005126e-02 4.44599688e-01 1.94688782e-01
7.53444791e-01 -3.57964307e-01 -7.72568226e-01 1.33763444e+00
1.84175335e-02 -1.32804036e+00 7.16107432e-03 -1.30588531e-01
8.90860200e-01 -3.47417980e-01 -5.82464896e-02 -1.25694409e-01
-2.63365120e-01 -7.36934602e-01 7.59632230e-01 7.27055132e-01
2.05906570e-01 -5.50735295e-01 1.06221485e+00 -6.30434826e-02
-9.57102239e-01 -5.47582423e-03 -4.07431632e-01 -2.35213548e-01
8.65063608e-01 7.01811612e-01 1.45830020e-01 8.84171009e-01
6.74178958e-01 1.48895633e+00 -2.81785607e-01 1.42211246e+00
7.94223100e-02 5.89146316e-01 -1.20576052e-02 3.57576221e-01
6.00288868e-01 -6.43869281e-01 8.36186349e-01 1.05475855e+00
2.69475490e-01 -1.61594078e-01 -6.08365424e-02 8.08629453e-01
4.26883996e-02 1.07110962e-01 -6.83800399e-01 1.99340999e-01
-2.96828657e-01 1.28587639e+00 -7.14832723e-01 -3.67652833e-01
-5.15812993e-01 7.53181279e-01 -3.28944206e-01 4.79388416e-01
-4.31737721e-01 -3.76298390e-02 9.45912123e-01 1.23923168e-01
1.06092744e-01 -4.73857462e-01 2.73798257e-01 -1.30379999e+00
-1.12621062e-01 -5.72007298e-01 1.88318398e-02 -9.42199469e-01
-1.39849114e+00 3.95574123e-01 -1.11453660e-01 -1.62844825e+00
-1.98182523e-01 -8.42469215e-01 -9.74292099e-01 6.77299261e-01
-1.79939246e+00 -7.84633458e-01 -5.16243517e-01 8.18047643e-01
7.65809596e-01 -1.70522556e-01 4.07156125e-02 7.26946712e-01
-6.60477698e-01 -5.35812303e-02 2.63729185e-01 5.82875274e-02
7.44751513e-01 -1.13183856e+00 1.24934487e-01 1.26841605e+00
-8.86361599e-02 3.34075779e-01 1.28088355e+00 -3.60638231e-01
-1.58216274e+00 -7.04499900e-01 4.41708028e-01 2.79885586e-02
8.17492664e-01 2.34937727e-01 -8.77616704e-01 4.41031933e-01
8.26052189e-01 4.21537876e-01 -2.49438006e-02 -8.35151732e-01
4.29532558e-01 -3.40064019e-01 -1.15043104e+00 3.34626108e-01
8.49543989e-01 -3.17526013e-01 -2.98478663e-01 2.45303646e-01
1.95035785e-01 -2.91298598e-01 -7.95270324e-01 3.08344156e-01
4.37299103e-01 -1.58648396e+00 1.05596960e+00 -6.03589118e-01
5.79499781e-01 -3.02921712e-01 1.54592007e-01 -1.32102156e+00
-3.34812641e-01 -1.08047807e+00 -2.89023876e-01 1.03732967e+00
-4.34095323e-01 -4.12285566e-01 7.23345637e-01 -8.88763815e-02
-2.76423305e-01 -3.94482911e-01 -9.36551630e-01 -5.55695534e-01
7.35103190e-02 -3.77545685e-01 -2.06433833e-01 8.11209977e-01
-5.54635584e-01 -8.16940069e-02 -7.18460679e-01 -2.33855024e-01
8.56060207e-01 -2.51477957e-01 4.93760526e-01 -1.16467404e+00
-2.04571355e-02 -5.19126356e-01 -1.07745612e+00 -9.30700064e-01
1.19138464e-01 -2.97522426e-01 -1.61781490e-01 -1.36129761e+00
-2.53974438e-01 -1.49043826e-02 -4.15995091e-01 -4.60077941e-01
6.79931194e-02 6.50373757e-01 1.14019148e-01 2.63853725e-02
-3.54543924e-01 6.20672226e-01 1.83109355e+00 -3.27536941e-01
1.73293456e-01 -4.04734910e-02 -1.09715641e-01 6.21726692e-01
6.84921622e-01 -1.93860620e-01 -3.21551919e-01 -3.59733701e-01
1.74177200e-01 1.93639755e-01 4.13184434e-01 -1.25733972e+00
1.31146282e-01 3.47404703e-02 2.73669690e-01 -3.70712936e-01
2.72695124e-01 -7.10493624e-01 8.29296038e-02 6.79791391e-01
4.59448397e-02 3.33427727e-01 -9.28044990e-02 5.24567544e-01
-8.05092096e-01 -6.39105082e-01 1.00447941e+00 -4.60933894e-01
-9.10385847e-01 2.69141346e-01 -6.19209886e-01 -6.51072115e-02
8.98865283e-01 -4.04215604e-01 -3.07740808e-01 -6.76246107e-01
-7.16699719e-01 -2.01684788e-01 2.24981055e-01 8.71407390e-02
5.95537364e-01 -1.17107654e+00 -8.84169161e-01 1.04959562e-01
-4.47958857e-01 -3.23481172e-01 4.32413727e-01 1.38659239e+00
-1.43708837e+00 3.83941323e-01 -7.59581327e-01 -5.78129709e-01
-8.74438643e-01 6.20991766e-01 4.40865189e-01 7.28536472e-02
-5.46040416e-01 5.98799706e-01 -3.04861460e-02 1.97936609e-01
1.15479775e-01 -3.14388126e-01 -4.77902442e-01 3.32463346e-02
5.21907270e-01 9.00133073e-01 1.49681926e-01 -1.03099501e+00
-1.73334852e-01 8.90091360e-01 2.58858532e-01 9.25596505e-02
1.22266614e+00 -4.74408776e-01 -2.28860617e-01 3.49951595e-01
1.30570889e+00 -2.84023613e-01 -1.41063023e+00 1.43256113e-01
-1.83025181e-01 -5.96254170e-01 4.42493409e-01 2.42002904e-02
-1.54932439e+00 1.10249138e+00 7.48434901e-01 5.08828342e-01
1.15015972e+00 -6.12016201e-01 1.16824985e+00 1.78973436e-01
4.37540561e-01 -8.06198835e-01 -1.07459210e-01 3.20575893e-01
7.91357160e-01 -1.23689032e+00 1.41018853e-01 -3.33964080e-01
-2.67469376e-01 1.43589377e+00 3.46674919e-01 -6.32767737e-01
9.71500993e-01 2.21901000e-01 6.04964346e-02 1.88197315e-01
-6.63536251e-01 -3.43647093e-01 3.13776806e-02 6.17605925e-01
5.70497036e-01 -6.96266711e-01 -6.97044730e-01 -4.90652978e-01
2.15351105e-01 2.77823895e-01 9.95905876e-01 9.89775121e-01
-3.42656285e-01 -9.65993106e-01 -5.23075879e-01 2.45757356e-01
-7.15672672e-01 2.51230121e-01 1.28213212e-01 6.10101104e-01
2.60011166e-01 1.00006068e+00 -8.63832906e-02 -1.79404452e-01
1.02477856e-01 -2.64277875e-01 7.49635041e-01 1.37611136e-01
-7.44982719e-01 1.44946679e-01 -1.16718702e-01 -7.92448938e-01
-1.26202297e+00 -5.91611803e-01 -6.54467583e-01 -6.88055158e-01
-5.04712425e-02 -1.41611338e-01 3.52820545e-01 1.14462900e+00
-3.91375631e-01 5.77084601e-01 5.62042177e-01 -1.26830637e+00
8.36888552e-02 -1.17408121e+00 -7.78787255e-01 7.43528664e-01
7.11555839e-01 -8.37667048e-01 -5.62090397e-01 5.06536067e-01] | [10.973615646362305, -1.9529390335083008] |
d3c09810-0658-4c98-ba83-fefde82ba8b7 | one-eye-is-all-you-need-lightweight-ensembles | 2211.11936 | null | https://arxiv.org/abs/2211.11936v1 | https://arxiv.org/pdf/2211.11936v1.pdf | One Eye is All You Need: Lightweight Ensembles for Gaze Estimation with Single Encoders | Gaze estimation has grown rapidly in accuracy in recent years. However, these models often fail to take advantage of different computer vision (CV) algorithms and techniques (such as small ResNet and Inception networks and ensemble models) that have been shown to improve results for other CV problems. Additionally, most current gaze estimation models require the use of either both eyes or an entire face, whereas real-world data may not always have both eyes in high resolution. Thus, we propose a gaze estimation model that implements the ResNet and Inception model architectures and makes predictions using only one eye image. Furthermore, we propose an ensemble calibration network that uses the predictions from several individual architectures for subject-specific predictions. With the use of lightweight architectures, we achieve high performance on the GazeCapture dataset with very low model parameter counts. When using two eyes as input, we achieve a prediction error of 1.591 cm on the test set without calibration and 1.439 cm with an ensemble calibration model. With just one eye as input, we still achieve an average prediction error of 2.312 cm on the test set without calibration and 1.951 cm with an ensemble calibration model. We also notice significantly lower errors on the right eye images in the test set, which could be important in the design of future gaze estimation-based tools. | ['Rohan Kalahasty', 'Lakshmi Sritan Motati', 'Rishi Athavale'] | 2022-11-22 | null | null | null | null | ['gaze-estimation'] | ['computer-vision'] | [-1.10733761e-02 2.34478459e-01 2.68121749e-01 -4.12501156e-01
5.18413596e-02 -2.73303658e-01 3.02844107e-01 -3.19589704e-01
-5.31459510e-01 6.14789546e-01 -6.28405437e-02 -2.34430432e-01
7.01241987e-03 -3.35872084e-01 -5.49621284e-01 -4.24712986e-01
3.06430638e-01 1.38735343e-02 2.56007165e-01 -2.07647830e-01
3.88892144e-01 1.46484882e-01 -2.14352870e+00 -8.94673094e-02
9.04857993e-01 1.16653323e+00 6.74026310e-02 8.42127204e-01
3.26868325e-01 5.34958839e-01 -6.91885352e-01 -3.18676025e-01
3.52831751e-01 -4.15901870e-01 -4.31760401e-01 -2.01009497e-01
1.27310932e+00 -3.31411064e-01 1.57928184e-01 8.02032650e-01
8.09453189e-01 2.90911105e-02 4.39320147e-01 -1.36705077e+00
-6.83778644e-01 1.03675224e-01 -1.01487577e+00 2.00185508e-01
3.59135658e-01 4.35220659e-01 7.57471323e-01 -7.65322328e-01
5.40228546e-01 8.63261640e-01 7.92205334e-01 8.23236883e-01
-1.44582784e+00 -1.07010126e+00 3.88555862e-02 2.69777238e-01
-1.40822935e+00 -7.08309948e-01 3.97520930e-01 -4.09994602e-01
9.65251267e-01 1.44497782e-01 7.24554002e-01 8.01216125e-01
2.86392003e-01 3.91981542e-01 1.33712649e+00 -5.59660554e-01
-1.16426885e-01 4.11126614e-01 1.92546919e-01 4.19076294e-01
3.40378404e-01 2.78646260e-01 -6.81607962e-01 3.56913328e-01
5.80146551e-01 -2.04340190e-01 -4.94133800e-01 -1.93918142e-02
-1.00849676e+00 6.03519857e-01 8.35447669e-01 3.94957997e-02
-1.76396474e-01 5.37110157e-02 -1.55125290e-01 1.97090507e-01
6.37122333e-01 5.51399112e-01 -2.18685225e-01 -6.57735765e-02
-1.17911923e+00 7.89051428e-02 6.39138997e-01 9.71696556e-01
8.00331831e-01 -4.42623831e-02 5.18435203e-02 7.56896079e-01
3.21046710e-01 7.59036243e-01 3.60099792e-01 -1.09396303e+00
1.28321931e-01 6.54761851e-01 1.79770157e-01 -1.00846839e+00
-7.84615695e-01 -4.06595320e-01 -4.87188816e-01 7.28552163e-01
6.35937154e-01 -4.02488649e-01 -1.04535294e+00 1.73641849e+00
9.68080759e-02 1.80897042e-01 -3.00699770e-01 8.48408937e-01
7.94227779e-01 1.37067139e-01 6.97423145e-02 -1.23437047e-01
1.21630430e+00 -9.24694777e-01 -4.82904017e-01 -5.06020308e-01
6.43779218e-01 -7.33168840e-01 1.06125271e+00 4.30405557e-01
-1.15072262e+00 -6.36432588e-01 -1.13827598e+00 -1.40989169e-01
-2.36056656e-01 1.66100621e-01 3.13911974e-01 9.26638663e-01
-1.68743610e+00 4.10109192e-01 -4.51763749e-01 -6.26328886e-01
3.67819130e-01 7.93777168e-01 -3.64668399e-01 6.80381805e-02
-5.26491821e-01 1.33373201e+00 9.55544226e-03 6.80385455e-02
-6.25729784e-02 -6.87423170e-01 -7.14218438e-01 1.53737277e-01
9.01822560e-03 -6.80802107e-01 1.20118272e+00 -1.46599376e+00
-1.36050725e+00 7.61847258e-01 -5.10318339e-01 -4.08374041e-01
3.68423641e-01 -2.62933373e-01 -4.39527959e-01 -1.27724335e-01
-4.15758252e-01 1.27065027e+00 8.64704132e-01 -1.06041062e+00
-7.54259646e-01 -4.82245117e-01 8.51115286e-02 1.63468525e-01
-4.08650041e-01 5.92889972e-02 -1.68584287e-01 5.76646365e-02
-7.69622773e-02 -1.24424148e+00 3.06006789e-01 2.95958787e-01
-1.88806057e-01 -3.06489114e-02 7.22330332e-01 -5.71005821e-01
1.26733828e+00 -2.12850857e+00 -1.47945702e-01 1.68902069e-01
7.38722682e-01 2.96313077e-01 -1.88835531e-01 -1.71411708e-01
-3.40082705e-01 1.34730399e-01 2.35305473e-01 -6.48185790e-01
-4.24615860e-01 -3.23664516e-01 1.63898706e-01 3.37724566e-01
5.78206331e-02 7.93606877e-01 -5.06435692e-01 -4.25229043e-01
2.15935230e-01 6.23383462e-01 -6.80507183e-01 -3.39242630e-02
6.71968684e-02 3.20716083e-01 3.54153931e-01 3.00160438e-01
7.48970270e-01 -6.18127465e-01 -1.57486126e-01 -4.99769039e-02
-2.90250778e-01 3.28114070e-02 -8.01742196e-01 1.26217341e+00
-4.23767030e-01 1.48709011e+00 -2.53186911e-01 2.85417102e-02
7.91290760e-01 5.32452874e-02 2.11343616e-01 -7.58969486e-01
3.58874887e-01 -1.12413261e-02 6.06022239e-01 -2.14461789e-01
6.22445643e-01 2.25068837e-01 6.86618268e-01 7.10526347e-01
5.96177988e-02 2.66528010e-01 1.58906266e-01 -8.16780627e-02
7.98496246e-01 8.66160691e-02 9.38836113e-02 -1.49542883e-01
3.83046716e-01 -3.10223073e-01 2.24375010e-01 4.74592358e-01
-3.51690918e-01 9.52034473e-01 3.58779788e-01 -4.04132009e-01
-9.97682035e-01 -5.95814824e-01 -2.69609571e-01 1.22110796e+00
1.56865507e-01 -4.17062968e-01 -1.07378447e+00 -4.80248272e-01
-1.78552419e-01 8.91808450e-01 -9.22316790e-01 -8.40798989e-02
-1.77803323e-01 -5.22786796e-01 2.99843699e-01 4.92419153e-01
4.87219453e-01 -1.04456222e+00 -8.04714739e-01 -3.53604406e-01
2.74314685e-03 -7.86526382e-01 -3.81565273e-01 -4.06146407e-01
-5.95313907e-01 -1.22307599e+00 -8.60662639e-01 -4.67666626e-01
7.30291605e-01 2.90240407e-01 1.17749131e+00 2.93580890e-01
7.79040158e-02 3.47408295e-01 8.55871942e-03 -8.27764809e-01
-2.94565298e-02 2.78715760e-01 1.86462343e-01 8.76464844e-02
8.84916663e-01 -3.37675035e-01 -8.84079278e-01 6.18337929e-01
-2.17802718e-01 3.12709570e-01 3.01523298e-01 6.15696907e-01
3.97953615e-02 -6.89607084e-01 2.80328572e-01 -6.84374928e-01
4.89454269e-01 -3.35747927e-01 -8.51665735e-01 1.96952224e-01
-1.18830013e+00 -7.50135332e-02 -3.21273529e-03 -5.81596017e-01
-9.56453443e-01 -1.56787321e-01 2.07498550e-01 -6.03410363e-01
-3.57173383e-02 1.27655402e-01 5.52972257e-01 -4.67106879e-01
1.32319057e+00 7.81264901e-03 4.27165747e-01 -2.32292309e-01
-3.13340314e-02 9.59367931e-01 1.73170179e-01 2.14550212e-01
4.90365893e-01 1.75390780e-01 -2.10939888e-02 -7.57764757e-01
-8.41186464e-01 -1.83682680e-01 -5.66809177e-01 -5.38898885e-01
7.67294884e-01 -9.14565980e-01 -1.35240066e+00 6.38523698e-01
-9.18758631e-01 -3.39636147e-01 6.94103092e-02 4.64185327e-01
-3.75280231e-01 -1.82226941e-01 -2.47880779e-02 -8.49537075e-01
-5.08282244e-01 -1.14170587e+00 8.33291590e-01 7.93225288e-01
-4.86721247e-01 -8.60231042e-01 -2.22283322e-02 2.47720927e-01
7.30786562e-01 -7.34720379e-02 2.83618778e-01 -4.79160547e-01
-5.63255668e-01 -2.03706384e-01 -5.21301389e-01 3.48036766e-01
-1.00347335e-02 4.44753349e-01 -1.38980901e+00 -4.09396529e-01
-3.56370300e-01 -1.22753859e-01 6.49138510e-01 6.17876589e-01
1.05090213e+00 1.84202209e-01 -4.83929902e-01 6.63978398e-01
1.16344655e+00 -7.83803910e-02 8.36340904e-01 1.85670257e-01
6.92659974e-01 6.91460192e-01 2.03628987e-01 7.57368430e-02
6.50950551e-01 4.27682132e-01 4.13365841e-01 4.91569377e-03
-4.01310116e-01 -3.50257047e-02 1.32930622e-01 2.31235519e-01
-6.26062751e-01 -3.38219583e-01 -1.14619577e+00 2.33097285e-01
-1.46356833e+00 -9.75626945e-01 -3.24629158e-01 2.42164660e+00
5.94915450e-01 1.24785192e-01 2.30729714e-01 -1.38550624e-01
7.99652457e-01 -1.78872094e-01 -8.33653152e-01 -3.22388709e-01
7.86316618e-02 1.84897542e-01 5.41972458e-01 2.54658878e-01
-7.64049709e-01 5.94463587e-01 7.47980547e+00 2.35665105e-02
-1.58084822e+00 -9.46139619e-02 4.79346067e-01 -5.77161133e-01
9.10314322e-02 -1.75362393e-01 -1.15441024e+00 7.93473601e-01
1.33511591e+00 -1.58126965e-01 5.52943945e-01 7.74757266e-01
3.27352956e-02 -5.82917750e-01 -1.25043321e+00 1.24141479e+00
4.88028437e-01 -1.15699136e+00 -6.05216503e-01 3.76631379e-01
7.02010393e-01 3.87222767e-01 3.05107772e-01 1.13927007e-01
9.64441523e-03 -1.24080920e+00 5.60435951e-01 7.80095100e-01
1.23339736e+00 -4.51510638e-01 7.56851971e-01 2.79082924e-01
-6.13639832e-01 -2.46665031e-01 -2.43680611e-01 -2.43540689e-01
-2.68901587e-01 -5.72862029e-02 -9.79602337e-01 -1.13694414e-01
1.01832592e+00 8.70982647e-01 -1.10915112e+00 1.40717411e+00
-1.32942453e-01 4.89882141e-01 -6.25236094e-01 -1.83267340e-01
-3.87464583e-01 1.19721726e-01 3.08524400e-01 5.75092733e-01
3.13955218e-01 -1.37765497e-01 -7.66428113e-01 8.00502777e-01
-2.16837958e-01 -7.16101378e-02 -4.97107238e-01 4.69828755e-01
5.66978753e-01 1.14993477e+00 -1.74488559e-01 2.54516229e-02
-5.24620950e-01 5.02771139e-01 3.94272774e-01 5.68845570e-01
-6.57167971e-01 -3.59636992e-01 7.43178248e-01 4.53412056e-01
1.18610762e-01 7.52394125e-02 -5.42255163e-01 -1.00031734e+00
-2.42247149e-01 -6.03781700e-01 -6.09763386e-03 -1.56402886e+00
-1.00701189e+00 7.22738504e-01 -2.00464964e-01 -1.32957959e+00
-5.16564608e-01 -7.07276642e-01 -6.09189749e-01 1.34643698e+00
-1.29896915e+00 -1.07678092e+00 -9.07720566e-01 5.06561756e-01
1.48430780e-01 -1.56217009e-01 7.85197198e-01 -3.23562324e-02
-7.96859562e-01 9.59693611e-01 -1.00659877e-01 7.33084325e-03
1.18841553e+00 -1.12932336e+00 3.44044000e-01 6.80406213e-01
-9.93417799e-02 7.92855680e-01 7.40481198e-01 -2.66857088e-01
-6.48029149e-01 -6.23214185e-01 9.09352541e-01 -9.39757586e-01
2.29497030e-01 -1.63443640e-01 -8.22826445e-01 1.03245759e+00
6.36908770e-01 1.43673673e-01 7.53020406e-01 6.06176853e-01
-2.71521747e-01 -1.99860975e-01 -1.12478530e+00 7.28200614e-01
9.19449151e-01 -2.50189483e-01 -3.70666146e-01 -2.37491690e-02
2.77603328e-01 -5.48945248e-01 -6.53581262e-01 1.39731184e-01
9.94928241e-01 -1.28588712e+00 5.69768250e-01 -2.66153365e-01
2.95074195e-01 -1.77944496e-01 3.26344967e-01 -1.25369775e+00
-3.19438130e-01 -3.67702812e-01 -2.00246498e-01 8.93091023e-01
6.43009663e-01 -1.07210708e+00 8.08595955e-01 1.08158624e+00
3.42642069e-01 -5.55707514e-01 -5.93597412e-01 -3.44459027e-01
-4.23623882e-02 -2.64135212e-01 5.59346795e-01 5.33030510e-01
-2.01232880e-01 5.34894884e-01 -2.40662351e-01 -1.05354451e-01
6.82856262e-01 -1.03594497e-01 1.07163227e+00 -1.72253871e+00
2.00454265e-01 -4.36193347e-01 -6.34873688e-01 -7.78071821e-01
1.14003070e-01 -2.94938982e-01 -1.17908828e-01 -1.16929865e+00
7.06700459e-02 -3.96812707e-01 -3.25661808e-01 5.28338730e-01
-2.82708168e-01 5.97233057e-01 3.92063946e-01 2.97688276e-01
-3.07127208e-01 1.27800405e-01 1.11227369e+00 2.45911911e-01
-2.10266113e-01 1.14200473e-01 -9.05693352e-01 7.33426571e-01
7.57639289e-01 -3.25305820e-01 -5.22959113e-01 -4.71147031e-01
4.50692117e-01 -3.56238216e-01 3.35420042e-01 -1.37714076e+00
5.79718947e-01 1.23220466e-01 9.30873036e-01 -4.69029546e-01
5.04618168e-01 -7.98734307e-01 1.26748055e-01 1.20343871e-01
-1.76276267e-01 9.27921608e-02 5.75966477e-01 3.32068115e-01
3.09554785e-01 -8.06767121e-02 9.65901256e-01 3.18516195e-01
-6.85671747e-01 7.30313882e-02 9.62279066e-02 -8.04291070e-02
1.13389099e+00 -8.66825700e-01 -6.24696314e-01 -5.41256368e-01
-7.36552179e-01 3.52952093e-01 1.01279688e+00 5.49720764e-01
3.39240193e-01 -8.62271190e-01 -6.30006433e-01 5.91732562e-01
2.06851006e-01 -2.17851207e-01 8.76029506e-02 1.05640662e+00
-4.69459593e-01 2.32114956e-01 -5.66330135e-01 -1.00551558e+00
-1.58732569e+00 2.84020245e-01 6.07808173e-01 5.17364144e-01
-9.42225382e-02 9.44468081e-01 1.49030000e-01 -2.92410981e-02
-5.14943674e-02 -4.93471650e-03 -5.40157199e-01 1.93718672e-01
8.44444811e-01 4.31320876e-01 1.21967439e-02 -8.72150600e-01
-2.42600948e-01 7.71852374e-01 -2.04251677e-01 9.92239714e-02
1.12993610e+00 -4.61846322e-01 4.66516651e-02 3.02465469e-01
8.06504011e-01 2.90704351e-02 -1.43849576e+00 -1.02863312e-01
-3.35717112e-01 -5.57489395e-01 2.43824020e-01 -1.09267640e+00
-1.06460381e+00 1.01671457e+00 1.03497291e+00 6.68747351e-02
1.41137362e+00 -2.91842073e-01 2.29929253e-01 1.13434702e-01
1.82188198e-01 -8.74078035e-01 -2.85080016e-01 4.00551051e-01
8.36728692e-01 -1.50300622e+00 1.82643458e-02 -1.03057101e-01
-6.42742395e-01 1.00288308e+00 1.17660689e+00 2.83108857e-02
6.82233393e-01 1.86071582e-02 2.75581896e-01 -1.78423285e-01
-9.85866189e-01 -3.84421200e-01 6.64533198e-01 7.04754293e-01
5.09968579e-01 -2.21002981e-01 1.79915458e-01 -7.95329269e-03
-5.54603636e-01 2.99116850e-01 6.11814141e-01 5.20065904e-01
-3.89605075e-01 -6.39845550e-01 -3.45146298e-01 8.39666367e-01
-2.29927376e-01 -2.39324927e-01 -3.36931854e-01 8.62892687e-01
7.97054246e-02 1.05355501e+00 6.42667353e-01 -6.84103072e-01
3.13673109e-01 2.92219400e-01 5.61008573e-01 -6.62193358e-01
-6.66454434e-01 -2.41146132e-01 -2.46719494e-01 -6.38288677e-01
-4.12326336e-01 -5.06770492e-01 -7.71068394e-01 -8.39798927e-01
-7.69372165e-01 -4.20636266e-01 7.96753585e-01 7.23082960e-01
8.72421205e-01 3.77724051e-01 2.60796577e-01 -1.09408605e+00
-1.65948153e-01 -1.49920237e+00 -4.56631154e-01 -4.16252110e-03
6.35511100e-01 -7.88016140e-01 -4.78709310e-01 1.48538738e-01] | [14.124768257141113, 0.0789031982421875] |
6622953c-3182-43f4-a439-e88726eb5740 | plug-and-play-pseudo-label-correction-network | 2206.06607 | null | https://arxiv.org/abs/2206.06607v1 | https://arxiv.org/pdf/2206.06607v1.pdf | Plug-and-Play Pseudo Label Correction Network for Unsupervised Person Re-identification | Clustering-based methods, which alternate between the generation of pseudo labels and the optimization of the feature extraction network, play a dominant role in both unsupervised learning (USL) and unsupervised domain adaptive (UDA) person re-identification (Re-ID). To alleviate the adverse effect of noisy pseudo labels, the existing methods either abandon unreliable labels or refine the pseudo labels via mutual learning or label propagation. However, a great many erroneous labels are still accumulated because these methods mostly adopt traditional unsupervised clustering algorithms which rely on certain assumptions on data distribution and fail to capture the distribution of complex real-world data. In this paper, we propose the plug-and-play graph-based pseudo label correction network (GLC) to refine the pseudo labels in the manner of supervised clustering. GLC is trained to perceive the varying data distribution at each epoch of the self-training with the supervision of initial pseudo labels generated by any clustering method. It can learn to rectify the initial noisy labels by means of the relationship constraints between samples on the k Nearest Neighbor (kNN) graph and early-stop training strategy. Specifically, GLC learns to aggregate node features from neighbors and predict whether the nodes should be linked on the graph. Besides, GLC is optimized with 'early stop' before the noisy labels are severely memorized to prevent overfitting to noisy pseudo labels. Consequently, GLC improves the quality of pseudo labels though the supervision signals contain some noise, leading to better Re-ID performance. Extensive experiments in USL and UDA person Re-ID on Market-1501 and MSMT17 show that our method is widely compatible with various clustering-based methods and promotes the state-of-the-art performance consistently. | ['Jinqiao Wang', 'Ming Tang', 'Guibo Zhu', 'Haiyun Guo', 'Kuan Zhu', 'Tianyi Yan'] | 2022-06-14 | null | null | null | null | ['unsupervised-person-re-identification'] | ['computer-vision'] | [ 1.14554942e-01 -7.27666309e-03 -1.37869403e-01 -7.49451935e-01
-2.27179229e-01 -3.50311100e-01 6.29371881e-01 1.60387844e-01
-4.74548548e-01 5.90823293e-01 -9.24247205e-02 1.64234504e-01
-4.83722389e-01 -8.04492056e-01 -2.91983306e-01 -8.72380912e-01
2.41968960e-01 7.62661219e-01 1.19763911e-02 6.37307093e-02
-1.71771541e-01 1.41482130e-01 -1.53369212e+00 -2.35262848e-02
1.09463155e+00 6.00252330e-01 -9.13384184e-02 2.57890582e-01
-1.53402701e-01 5.63078523e-01 -5.70607662e-01 -5.01230717e-01
2.56865084e-01 -5.26217520e-01 -7.28105843e-01 3.78305823e-01
-6.32669311e-03 1.69149980e-01 -2.84836173e-01 1.46680963e+00
5.56014419e-01 2.51024753e-01 8.08744788e-01 -1.54284418e+00
-7.09701955e-01 8.08283329e-01 -6.10088170e-01 -1.55782148e-01
2.86439091e-01 -2.35816967e-02 6.97557569e-01 -6.52629077e-01
4.42606956e-01 1.23206723e+00 9.03982759e-01 9.65821028e-01
-1.48236871e+00 -8.23742747e-01 3.26491684e-01 3.01807374e-01
-1.84845638e+00 -3.30434263e-01 1.01558018e+00 -4.53641266e-01
2.68646598e-01 2.73888558e-01 4.30410355e-01 1.06739736e+00
-4.68612403e-01 3.96776795e-01 1.13081956e+00 -4.30495918e-01
3.44176441e-01 3.93982321e-01 4.75606889e-01 6.04878426e-01
2.76928574e-01 8.91105309e-02 -4.91244495e-01 -7.44255856e-02
4.58912909e-01 1.33450672e-01 -5.59278429e-02 -2.24934176e-01
-1.02969992e+00 5.21620691e-01 4.14424598e-01 3.20385754e-01
-1.41226396e-01 -1.73820823e-01 2.03075677e-01 2.60920614e-01
3.95879120e-01 1.94583610e-01 -2.10699096e-01 2.60931998e-01
-7.90662587e-01 -1.07640199e-01 5.68998218e-01 9.08599138e-01
1.09331727e+00 -2.18874961e-01 -5.58280163e-02 1.02822924e+00
5.49520075e-01 4.01390702e-01 6.19944096e-01 -7.67843664e-01
1.74575478e-01 1.04057658e+00 -1.32394768e-02 -1.30718601e+00
-6.08735859e-01 -7.16705501e-01 -1.27882242e+00 2.70472281e-02
5.08269012e-01 -2.68484920e-01 -1.07905960e+00 1.77826416e+00
4.82324302e-01 5.71826041e-01 -3.00638266e-02 9.26766932e-01
8.76437306e-01 3.09769303e-01 1.89248234e-01 -3.82276654e-01
9.46786821e-01 -6.67916656e-01 -7.91300476e-01 -1.29964232e-01
7.76605904e-01 -3.92381966e-01 8.28395069e-01 3.48584086e-01
-3.31812531e-01 -9.71874714e-01 -9.19336081e-01 4.84062105e-01
-4.60362226e-01 1.77157909e-01 4.91090357e-01 8.88573110e-01
-1.09295368e+00 6.22734129e-01 -5.27708650e-01 -4.37496006e-01
3.36432636e-01 6.07993841e-01 -4.30970550e-01 -1.54154047e-01
-1.30724049e+00 3.73249710e-01 5.47232032e-01 3.59594077e-01
-4.21444952e-01 -4.04076338e-01 -6.03857517e-01 -2.51935571e-01
4.53192711e-01 -3.61896247e-01 6.63970947e-01 -1.20390213e+00
-1.30225551e+00 8.81155610e-01 -1.49787813e-01 -1.80967525e-01
4.19523120e-01 2.50281841e-01 -8.92874360e-01 -1.86323851e-01
3.63658279e-01 6.90129876e-01 8.48189771e-01 -1.58159161e+00
-7.69605994e-01 -4.47313517e-01 -4.56274450e-01 2.66911596e-01
-4.91870672e-01 -4.46852893e-01 -6.03062212e-01 -6.01107419e-01
4.45921361e-01 -9.74698842e-01 -2.31536731e-01 -5.48254371e-01
-6.68681860e-01 -6.27687573e-01 7.05168426e-01 -4.01452571e-01
1.24654043e+00 -2.14529133e+00 -4.95213754e-02 8.72970641e-01
3.88951004e-01 1.94024891e-01 -8.88444558e-02 -6.06523603e-02
-2.70845652e-01 1.83365688e-01 -1.57551497e-01 -5.42324960e-01
-6.14701100e-02 4.16694194e-01 2.58799642e-01 5.74175358e-01
-2.11194977e-01 6.79247856e-01 -1.09578526e+00 -6.08180225e-01
1.88617200e-01 2.61575252e-01 -2.60991514e-01 1.92998692e-01
7.62203708e-02 8.97159994e-01 -3.04803729e-01 5.42612672e-01
7.05739200e-01 -3.27076972e-01 4.37039256e-01 -3.37374538e-01
2.76211351e-01 -1.45881176e-01 -1.61127985e+00 1.44871771e+00
1.48302242e-01 1.61957905e-01 -9.10550728e-02 -1.20037293e+00
1.07808101e+00 1.24002367e-01 6.83158696e-01 -5.67587912e-01
2.56963968e-01 -1.22270379e-02 -1.55353531e-01 -3.65456462e-01
1.40419602e-01 1.84564963e-01 -7.93915913e-02 4.02300775e-01
-1.32582830e-02 5.68154454e-01 1.39018223e-01 3.02606821e-01
7.99348176e-01 -1.30909786e-01 -5.75301843e-03 -1.54091522e-01
7.47098923e-01 -1.54458120e-01 9.13230121e-01 8.74625444e-01
-2.43557706e-01 5.19522071e-01 1.04183689e-01 -1.91942275e-01
-7.12751806e-01 -9.58891273e-01 -1.70346841e-01 1.06758201e+00
4.88603920e-01 -4.78877664e-01 -8.44005287e-01 -1.02714992e+00
-9.37549248e-02 5.14976561e-01 -5.86511552e-01 -3.22302490e-01
-2.38157049e-01 -9.86749291e-01 4.91353303e-01 2.79126912e-01
7.85045683e-01 -8.23888958e-01 4.05956566e-01 1.55124068e-01
-3.07548076e-01 -8.69595051e-01 -5.61693132e-01 1.01374887e-01
-5.11836112e-01 -1.24706328e+00 -3.48955870e-01 -1.00833249e+00
1.26311255e+00 2.47576550e-01 6.94557250e-01 3.26796055e-01
9.04511735e-02 3.88499349e-01 -5.97667873e-01 1.91309541e-01
-4.78934765e-01 1.66960910e-01 5.34557104e-01 5.61001658e-01
8.94995511e-01 -7.47454762e-01 -3.22157174e-01 6.76089764e-01
-5.42255163e-01 -8.31923112e-02 3.47444087e-01 8.46635640e-01
6.01223767e-01 9.31563377e-01 7.89457202e-01 -1.31531024e+00
5.45958698e-01 -5.10489643e-01 -3.63583416e-01 3.21872234e-01
-1.23547852e+00 6.95174411e-02 6.74752891e-01 -7.22428560e-01
-1.00864828e+00 3.10484380e-01 5.14827445e-02 -2.85004884e-01
-4.83078361e-01 3.77598345e-01 -5.39358616e-01 -7.75417313e-02
7.49826431e-01 1.73416227e-01 -1.47506028e-01 -5.03264368e-01
4.28349912e-01 9.86705840e-01 8.80807877e-01 -3.24852526e-01
1.11956763e+00 3.54806095e-01 -1.22622855e-01 -4.05659586e-01
-7.93230355e-01 -7.52571046e-01 -9.84875619e-01 -4.93361950e-01
7.97482550e-01 -8.90173376e-01 -8.20122719e-01 7.92728662e-01
-6.70976937e-01 -1.96665838e-01 -2.07245931e-01 3.85024846e-01
1.08302347e-01 5.58215857e-01 -4.07545149e-01 -8.10376227e-01
-3.57251391e-02 -9.00706649e-01 5.50060868e-01 6.34618461e-01
-2.29840502e-01 -1.02006841e+00 -2.64876813e-01 3.63348156e-01
-1.60864107e-02 5.17410645e-03 7.80771017e-01 -8.49370837e-01
-1.27731219e-01 -1.44249022e-01 -2.99893171e-01 4.12633002e-01
5.60045421e-01 -3.89239252e-01 -1.00009716e+00 -4.92943347e-01
-2.32136577e-01 -4.17659096e-02 7.02362776e-01 1.32042632e-01
9.18364227e-01 -2.83966839e-01 -5.55094719e-01 5.72635174e-01
1.20400071e+00 1.36718571e-01 2.98047781e-01 1.93214357e-01
1.12777627e+00 7.56845951e-01 4.13187325e-01 4.21177298e-01
6.11526072e-01 4.96040255e-01 1.49677664e-01 -2.30259020e-02
-2.54427820e-01 -5.64657390e-01 1.93776950e-01 9.67719078e-01
4.24425565e-02 -5.86794168e-02 -8.48285377e-01 3.57739359e-01
-1.96762836e+00 -6.67460501e-01 -4.44203347e-01 2.26850629e+00
9.43908334e-01 2.53964782e-01 2.30154932e-01 4.49639499e-01
1.27310896e+00 -1.52166456e-01 -6.27553761e-01 3.80769044e-01
-2.09037259e-01 -2.97185838e-01 6.51810288e-01 5.92385232e-01
-1.25860977e+00 9.27081943e-01 5.35389996e+00 9.16158199e-01
-7.27145314e-01 1.23535201e-01 7.36790299e-01 3.57092857e-01
-3.50333154e-02 -3.11311223e-02 -9.20540571e-01 8.11458290e-01
6.68972611e-01 1.44801363e-01 7.51967609e-01 7.20656812e-01
1.07651107e-01 -1.26241997e-01 -1.10970843e+00 1.38360512e+00
1.12510547e-01 -8.31025600e-01 -9.34902579e-02 -5.93199991e-02
1.03479385e+00 -4.04257536e-01 6.19496393e-04 2.56168365e-01
6.53509915e-01 -8.72629464e-01 3.91118407e-01 7.66373932e-01
8.17683935e-01 -8.43847871e-01 9.18540001e-01 4.41150635e-01
-1.22217977e+00 -1.75047845e-01 -3.39973450e-01 2.42548585e-02
-9.12801623e-02 8.05667400e-01 -7.46287286e-01 5.94580352e-01
8.17056239e-01 9.16816473e-01 -9.79331374e-01 7.66047835e-01
-4.46964473e-01 8.13162804e-01 -4.16634887e-01 1.65773854e-01
-2.04152837e-01 -3.58470529e-01 4.16263282e-01 9.15853500e-01
-6.45548254e-02 5.50725311e-02 4.58084881e-01 8.01479816e-01
-6.81034848e-02 -7.02618575e-03 -1.31181777e-01 2.06325799e-01
8.21470439e-01 1.17163098e+00 -8.36594045e-01 -2.88140714e-01
-1.58962518e-01 1.03777361e+00 3.06654006e-01 6.54427171e-01
-6.61299109e-01 -1.92207813e-01 2.50408858e-01 2.09633231e-01
-1.57585591e-01 -1.27357349e-03 -3.98539037e-01 -8.96903574e-01
-1.23423412e-01 -6.03836596e-01 5.80701292e-01 -4.53581721e-01
-1.81440818e+00 4.28130925e-01 -8.65394250e-02 -1.24631667e+00
-2.11084411e-01 -1.31486565e-01 -4.16714340e-01 6.02631390e-01
-1.24615335e+00 -9.71265972e-01 -5.17298400e-01 8.25119734e-01
1.04223669e-01 -3.51703435e-01 5.93659282e-01 5.99946856e-01
-8.46412838e-01 9.75971639e-01 2.40408182e-01 5.10420978e-01
9.65530634e-01 -1.29243076e+00 -1.20725207e-01 8.93267035e-01
1.06627494e-01 5.55467665e-01 5.10465801e-01 -1.12778068e+00
-7.55350053e-01 -1.49923778e+00 7.64769971e-01 -2.19913870e-01
3.14562142e-01 -4.58966583e-01 -8.08038533e-01 6.00729227e-01
-3.66507381e-01 -1.84364077e-02 7.09869325e-01 2.49055624e-01
-2.35216901e-01 -5.36031663e-01 -1.25072169e+00 3.95586997e-01
1.20755577e+00 -4.59121525e-01 -3.11877340e-01 4.50088471e-01
5.32566965e-01 -1.54051427e-02 -8.26171100e-01 3.30156267e-01
4.25341986e-02 -7.84574568e-01 8.01418185e-01 -2.24478364e-01
-4.23092991e-01 -8.19531143e-01 2.24119887e-01 -1.30720365e+00
-6.75799906e-01 -5.82103252e-01 2.02984586e-01 2.02849340e+00
3.28060865e-01 -7.26359308e-01 9.83332396e-01 8.40453207e-01
2.40332589e-01 -1.23726897e-01 -9.01222467e-01 -6.93008840e-01
-5.57304323e-01 -3.55318844e-01 6.98267460e-01 1.43198085e+00
-2.79109925e-03 3.64304423e-01 -5.29899955e-01 4.39756453e-01
1.06995475e+00 -3.40881735e-01 7.47081518e-01 -1.62010145e+00
-5.67997359e-02 -1.99288622e-01 -3.97165239e-01 -8.89487982e-01
2.97727674e-01 -1.21004963e+00 1.73894599e-01 -1.18859458e+00
1.85398996e-01 -9.88918424e-01 -4.41198200e-01 6.09908342e-01
-3.58779222e-01 3.00486326e-01 -5.63178509e-02 6.16698980e-01
-1.00640965e+00 3.13704312e-01 8.95807564e-01 -3.09100300e-01
-5.12425661e-01 1.09680034e-01 -6.32912517e-01 6.41768336e-01
7.31030166e-01 -6.92457616e-01 -4.86838251e-01 -8.11322108e-02
5.92383742e-02 -4.09951836e-01 3.28972518e-01 -1.14293396e+00
7.50388682e-01 2.41191220e-02 6.09552622e-01 -4.67762828e-01
-1.11435816e-01 -9.27876115e-01 4.27276731e-01 1.31371647e-01
-4.18730378e-01 -2.21377268e-01 -5.60186088e-01 8.29334497e-01
-2.64074728e-02 -2.01368049e-01 8.15398812e-01 -6.77756313e-03
-8.77987802e-01 2.46750355e-01 -2.37758815e-01 -1.38853878e-01
9.03930306e-01 -3.67330283e-01 -8.37518722e-02 -3.59826744e-01
-1.13778603e+00 6.08131826e-01 4.72678840e-01 2.75469989e-01
4.88002598e-01 -1.52281415e+00 -5.45574605e-01 5.57889879e-01
9.05586686e-03 2.02022836e-01 3.27742755e-01 6.04471922e-01
7.68100470e-02 -1.01341726e-02 1.45756096e-01 -7.55956292e-01
-1.06296062e+00 6.46560609e-01 3.17188293e-01 -2.57096440e-01
-4.44429368e-01 7.55776763e-01 -1.72386363e-01 -7.76427627e-01
5.69010258e-01 3.37601751e-01 -5.94738364e-01 -2.23569255e-02
3.46832186e-01 5.39241970e-01 -2.81063914e-01 -9.51173425e-01
-3.63011956e-01 5.59749424e-01 -6.97152168e-02 2.64669091e-01
9.53993320e-01 -6.47919118e-01 -2.34085396e-01 3.98202837e-01
9.91204262e-01 -1.84652179e-01 -1.09305978e+00 -4.61786300e-01
2.35560805e-01 -2.67020524e-01 -1.66739658e-01 -8.33810925e-01
-1.27836823e+00 2.66706198e-01 9.72312212e-01 2.29994670e-01
1.05302513e+00 -1.94775313e-02 6.15286231e-01 3.16508114e-01
3.75528008e-01 -1.49004185e+00 6.98143616e-02 8.91471878e-02
1.59537792e-01 -1.28481245e+00 -3.03130671e-02 -6.36929393e-01
-5.39389074e-01 7.23197222e-01 8.40099454e-01 5.19358255e-02
7.86803842e-01 2.79435068e-02 2.18264416e-01 -1.40221063e-02
-1.19918957e-01 -2.70871133e-01 2.63898998e-01 1.08317494e+00
-1.98132455e-01 2.41697773e-01 -1.41662747e-01 8.93588901e-01
-1.43859357e-01 -2.11110428e-01 1.88870132e-01 4.04636085e-01
-1.92836016e-01 -1.34622073e+00 -5.58202267e-01 5.49130261e-01
4.75359857e-02 1.91195518e-01 -4.22212481e-01 5.01548827e-01
7.84277022e-01 1.46067119e+00 -1.42980620e-01 -9.40457404e-01
1.60818204e-01 1.40315652e-01 3.23735774e-02 -5.53809643e-01
-4.06626910e-01 2.97177657e-02 -8.85125101e-02 -2.46230096e-01
-7.97130346e-01 -6.98836327e-01 -1.44600630e+00 -3.57418388e-01
-5.46455562e-01 3.71286690e-01 1.95319116e-01 1.07270265e+00
3.23572487e-01 3.29033375e-01 8.56364191e-01 -4.07159001e-01
-1.48008615e-01 -9.89993453e-01 -9.97806847e-01 9.94341195e-01
-9.25043449e-02 -7.50815570e-01 -4.77145225e-01 2.88383693e-01] | [14.849331855773926, 1.1271573305130005] |
e2a36e68-a498-4b89-90c3-9d494b4c3fbe | joint-blind-room-acoustic-characterization | 2010.11167 | null | https://arxiv.org/abs/2010.11167v1 | https://arxiv.org/pdf/2010.11167v1.pdf | Joint Blind Room Acoustic Characterization From Speech And Music Signals Using Convolutional Recurrent Neural Networks | Acoustic environment characterization opens doors for sound reproduction innovations, smart EQing, speech enhancement, hearing aids, and forensics. Reverberation time, clarity, and direct-to-reverberant ratio are acoustic parameters that have been defined to describe reverberant environments. They are closely related to speech intelligibility and sound quality. As explained in the ISO3382 standard, they can be derived from a room measurement called the Room Impulse Response (RIR). However, measuring RIRs requires specific equipment and intrusive sound to be played. The recent audio combined with machine learning suggests that one could estimate those parameters blindly using speech or music signals. We follow these advances and propose a robust end-to-end method to achieve blind joint acoustic parameter estimation using speech and/or music signals. Our results indicate that convolutional recurrent neural networks perform best for this task, and including music in training also helps improve inference from speech. | ['Milos Cernak', 'Paul Callens'] | 2020-10-21 | null | null | null | null | ['room-impulse-response'] | ['audio'] | [-1.18713699e-01 -6.78405225e-01 5.56108952e-01 -2.27021381e-01
-1.06173599e+00 -5.63916445e-01 1.34776488e-01 -3.33214760e-01
-2.89227426e-01 2.65008420e-01 7.85678446e-01 -6.20217264e-01
-9.47168618e-02 -3.04324865e-01 -2.61147320e-01 -7.09345341e-01
-9.85051617e-02 -3.65307122e-01 -2.37971455e-01 -1.23980023e-01
4.52011526e-02 4.08084482e-01 -1.83014989e+00 1.04378955e-02
6.52456164e-01 1.01945794e+00 3.21736097e-01 1.25173426e+00
1.30998090e-01 5.03416657e-01 -9.47527349e-01 -1.40975997e-01
7.07841590e-02 -4.37434703e-01 -2.68991053e-01 -4.94428366e-01
-2.05984898e-02 -5.41809142e-01 -2.34764665e-01 1.04954231e+00
1.27174366e+00 2.82185286e-01 6.06310546e-01 -6.12477839e-01
-6.81132257e-01 7.53134310e-01 4.19844128e-02 2.10912198e-01
6.66054189e-01 4.05461900e-02 7.04900563e-01 -8.29145253e-01
-3.13790977e-01 1.12229073e+00 8.10207188e-01 5.03530622e-01
-6.95112526e-01 -9.11147654e-01 -4.22973841e-01 3.48383933e-01
-1.16322684e+00 -1.06513941e+00 6.86965883e-01 -2.33046725e-01
9.45091665e-01 7.18377352e-01 3.80634099e-01 1.24129856e+00
-2.08584562e-01 4.84457225e-01 9.29803133e-01 -8.19235563e-01
2.48789638e-01 1.04198568e-01 -1.69365834e-02 2.98025846e-01
-2.11466223e-01 4.95909393e-01 -4.23344433e-01 -1.63449883e-01
5.76658785e-01 -3.35392475e-01 -8.20502043e-01 5.09848893e-01
-1.04588187e+00 3.88888627e-01 1.60317361e-01 4.12949443e-01
-8.43178183e-02 2.28791490e-01 3.81054878e-01 5.07935703e-01
3.16864371e-01 6.58590317e-01 -2.72843689e-01 -3.93818974e-01
-7.46476650e-01 -1.96869925e-01 8.86725187e-01 5.13903797e-01
1.08283557e-01 4.29343849e-01 -1.93193614e-01 1.58387530e+00
6.24179304e-01 8.24011385e-01 5.21127403e-01 -9.32466865e-01
3.25385243e-01 -5.82361341e-01 2.96614587e-01 -7.00886607e-01
-5.14426887e-01 -7.12325513e-01 -8.32171500e-01 2.18825817e-01
3.27074677e-01 -1.35395244e-01 -8.80411625e-01 1.56343997e+00
3.43711413e-02 3.22758257e-01 6.34962246e-02 1.08507323e+00
9.22445178e-01 6.43468201e-01 -2.16452867e-01 -2.79874235e-01
1.30771065e+00 -9.04057384e-01 -1.20808446e+00 -8.33193585e-02
1.15035318e-01 -1.24320400e+00 1.18748581e+00 8.25859964e-01
-9.51440454e-01 -7.73316205e-01 -1.17286134e+00 2.12075964e-01
-2.75566757e-01 6.07315153e-02 2.51406014e-01 1.59567845e+00
-1.00390255e+00 3.70565832e-01 -7.89380550e-01 2.05689281e-01
-1.69876501e-01 -3.45518775e-02 1.09655480e-03 1.27967834e-01
-1.19774604e+00 5.96154690e-01 -6.40238702e-01 5.49157262e-01
-9.96086180e-01 -5.48796535e-01 -7.81315863e-01 1.45232314e-02
-1.87593415e-01 -7.11541414e-01 1.46264827e+00 -3.94688845e-01
-2.13078237e+00 3.75328839e-01 -2.96973586e-01 -4.04123574e-01
2.12731004e-01 -5.96569717e-01 -1.20293736e+00 -3.16097885e-02
-3.45206261e-01 -1.69664279e-01 1.24631417e+00 -1.19858718e+00
-8.12010467e-02 -7.90436864e-02 -4.07176852e-01 8.98972526e-02
-4.33647424e-01 6.44168317e-01 -2.97924042e-01 -7.10872293e-01
5.34443200e-01 -5.86346328e-01 -4.74522524e-02 -3.73003215e-01
-5.97501755e-01 2.66265661e-01 3.89667124e-01 -1.22891104e+00
1.12601578e+00 -2.24862146e+00 -4.65096086e-01 2.45032802e-01
3.70171033e-02 6.02107942e-01 -1.37768194e-01 1.24185301e-01
-1.72801331e-01 -1.34185441e-02 2.73566367e-03 -5.97835898e-01
2.24621803e-01 -5.73011398e-01 -5.75612426e-01 5.84852099e-01
-3.36782366e-01 3.42555225e-01 -6.83735430e-01 -1.16696700e-01
5.41562557e-01 9.69954789e-01 -4.42926496e-01 5.56088984e-01
5.70692599e-01 5.26723981e-01 1.54845591e-03 6.42153859e-01
8.43833864e-01 6.01662338e-01 -4.51774389e-01 -2.27105558e-01
-2.61715129e-02 6.73493922e-01 -1.14895642e+00 1.36377883e+00
-1.05203474e+00 9.44636941e-01 5.55591643e-01 -5.46489418e-01
1.15136302e+00 7.00116992e-01 -1.78158328e-01 -8.00214827e-01
1.46045953e-01 4.66771334e-01 4.60203029e-02 -9.20310318e-01
3.36583018e-01 -1.14454761e-01 4.94391948e-01 3.40177208e-01
-2.10612550e-01 -4.22706127e-01 -5.27978122e-01 -4.37777996e-01
9.64514792e-01 -1.85833097e-01 6.34301547e-03 1.93836793e-01
6.41617537e-01 -1.11766684e+00 1.56899899e-01 7.29246259e-01
-2.74512440e-01 1.01756239e+00 -2.61770517e-01 1.84891343e-01
-7.71819115e-01 -1.31009126e+00 -2.42700756e-01 1.25104630e+00
-1.30910367e-01 -1.50100380e-01 -8.73219132e-01 2.63895184e-01
-4.44083512e-01 8.42769563e-01 -1.35870278e-01 -1.66997582e-01
-5.18455505e-01 -3.40477794e-01 9.04669642e-01 4.32274580e-01
2.03007072e-01 -8.50378811e-01 -6.66819140e-02 1.83296770e-01
-3.88141185e-01 -9.65072453e-01 -6.97560668e-01 1.62237749e-01
-3.74317616e-01 -4.32684481e-01 -1.02754855e+00 -6.69042349e-01
8.06972533e-02 3.72309268e-01 7.60646403e-01 -2.27134869e-01
-4.03324276e-01 6.06575787e-01 -4.86620605e-01 -5.37955582e-01
-8.10711801e-01 -4.68606502e-01 5.70149660e-01 1.60087477e-02
-1.04086474e-01 -9.71721828e-01 -9.00947511e-01 6.43639326e-01
-4.99475420e-01 -3.51737171e-01 2.43054196e-01 4.27963465e-01
1.92806527e-01 9.12336856e-02 8.07618499e-01 5.04154600e-02
9.76727128e-01 5.61785847e-02 -3.55841935e-01 1.06397860e-01
-3.65847290e-01 -2.23707587e-01 8.17131102e-01 -4.96054709e-01
-1.13614130e+00 -5.63006043e-01 -7.23410368e-01 -3.73436362e-01
-4.20987248e-01 2.10512459e-01 -5.33492506e-01 4.70765978e-02
8.10023367e-01 1.07666217e-01 -1.95546180e-01 -7.26086080e-01
2.08784804e-01 1.45170033e+00 8.76980364e-01 -1.67574912e-01
6.34351373e-01 1.52945444e-01 -3.70299160e-01 -1.05548680e+00
-3.93224686e-01 -7.93801725e-01 9.92741436e-02 -2.74002224e-01
5.98161101e-01 -8.54082346e-01 -9.19056714e-01 6.50111198e-01
-1.11162651e+00 -1.77189097e-01 1.62179708e-01 1.14590144e+00
-4.63428289e-01 5.64287424e-01 -5.57234049e-01 -1.58676100e+00
-6.73057020e-01 -1.18659401e+00 7.13318825e-01 9.47653223e-03
-3.10354561e-01 -6.47561729e-01 1.13046058e-01 6.46260560e-01
1.00421393e+00 -4.51889813e-01 5.27095616e-01 -3.23846817e-01
-8.67999643e-02 -4.36273634e-01 1.52844876e-01 8.37924063e-01
4.33211625e-01 -2.17011362e-01 -1.80124092e+00 -1.78063691e-01
5.15922546e-01 1.76546462e-02 7.72815287e-01 8.48804653e-01
1.21335983e+00 -3.00568521e-01 1.20416351e-01 9.05226707e-01
8.67552519e-01 3.84891957e-01 9.64914262e-01 2.36619748e-02
4.58246320e-01 4.97311413e-01 1.53074771e-01 4.52223688e-01
-1.24231502e-01 5.26282489e-01 3.23223561e-01 -1.72521934e-01
-5.89800119e-01 -1.29009515e-01 6.37670517e-01 1.24674118e+00
-5.95339090e-02 -3.13267469e-01 -7.18044877e-01 3.48669499e-01
-8.57913852e-01 -9.43742871e-01 -1.01771178e-02 2.52588511e+00
7.65513062e-01 -2.12209970e-01 -6.25283569e-02 6.81777418e-01
7.48310506e-01 6.38837367e-02 -2.60127544e-01 -5.93660414e-01
-1.10390894e-01 4.37478751e-01 3.97207230e-01 8.22901607e-01
-8.07591140e-01 4.04385448e-01 6.83727121e+00 6.71559215e-01
-1.27891505e+00 1.21199921e-01 3.31172556e-01 -4.46491949e-02
-3.49520475e-01 -4.85183239e-01 -3.02792996e-01 1.17823228e-01
1.23078692e+00 2.57475823e-01 9.56960320e-01 5.89374065e-01
5.29075444e-01 1.34025723e-01 -1.04201615e+00 1.38856816e+00
1.12032250e-01 -6.07245207e-01 -3.87967348e-01 -2.47831210e-01
1.89031065e-01 -2.80326549e-02 7.11438775e-01 5.15423231e-02
-6.03015311e-02 -1.31805944e+00 6.39810443e-01 5.86905122e-01
7.96961010e-01 -6.47589862e-01 4.84908581e-01 -3.60265262e-02
-1.14277220e+00 -1.10474840e-01 -3.39239895e-01 1.79887235e-01
1.58909649e-01 6.62715614e-01 -1.23542774e+00 3.32721055e-01
8.21739852e-01 7.13192374e-02 -1.94056258e-01 1.58034861e+00
-2.58595198e-01 1.08017790e+00 -2.37628132e-01 -7.82118067e-02
-3.78838658e-01 -8.51598531e-02 8.34527254e-01 1.32443762e+00
7.00550258e-01 -2.04976112e-01 -6.05110943e-01 8.55058670e-01
1.77550167e-01 9.69222412e-02 -2.08951607e-01 3.71140689e-02
6.04713738e-01 1.01657462e+00 -3.36272061e-01 2.48491809e-01
-1.09666251e-01 8.59215081e-01 -7.36161649e-01 8.16312790e-01
-7.98559427e-01 -9.50630665e-01 8.65996182e-01 -2.52200335e-01
2.01388970e-02 -2.61930048e-01 -2.88506687e-01 -6.60104692e-01
-5.00439629e-02 -1.00185215e+00 -2.51365781e-01 -1.02765727e+00
-1.06121016e+00 8.72869253e-01 -5.64051509e-01 -1.55138886e+00
-3.36732477e-01 -8.23764384e-01 -7.38985717e-01 1.25188911e+00
-1.43835211e+00 -6.86065376e-01 -8.03564489e-02 6.24997735e-01
3.54714215e-01 -3.10642034e-01 1.22538710e+00 5.13741910e-01
-4.63723391e-01 1.03049040e+00 4.29259241e-01 2.30184853e-01
8.42319548e-01 -1.06433797e+00 6.09533906e-01 7.85322309e-01
1.12409264e-01 8.37256849e-01 1.29543936e+00 1.22224027e-02
-1.48465860e+00 -6.65068150e-01 5.05132914e-01 -2.70182312e-01
4.41672891e-01 -3.01491231e-01 -8.00069392e-01 3.17383818e-02
-1.06048703e-01 -1.12946860e-01 9.80440795e-01 3.50811839e-01
-7.21926391e-01 -3.85912597e-01 -8.41519177e-01 5.62366068e-01
7.72982240e-01 -1.12486899e+00 -4.51555431e-01 1.46243393e-01
9.96955633e-01 -2.25139335e-01 -4.58074272e-01 7.12720724e-03
7.86724925e-01 -1.13618016e+00 1.28746629e+00 3.68990824e-02
-1.21854685e-01 -3.94405216e-01 -4.49160844e-01 -1.48836911e+00
-4.79445979e-03 -1.22390020e+00 6.42541274e-02 1.38299692e+00
6.42824590e-01 -7.78013527e-01 1.36769190e-01 3.89663666e-01
-4.52229857e-01 4.39895019e-02 -8.93428385e-01 -1.03593731e+00
-8.78455490e-02 -1.17313564e+00 8.05045307e-01 5.44932008e-01
2.52371281e-02 1.44324556e-01 -6.12223506e-01 5.62168241e-01
5.88207483e-01 -2.27928489e-01 6.18177176e-01 -9.65588510e-01
-5.93056023e-01 -4.33151335e-01 -1.05589673e-01 -1.15745842e+00
-5.15080616e-02 -4.18497682e-01 3.83865356e-01 -1.62381077e+00
-5.57894766e-01 -3.60103250e-01 -4.36017245e-01 -4.59274575e-02
-1.37622997e-01 6.25240281e-02 -1.50570720e-01 -2.08053023e-01
-9.98459533e-02 5.64751089e-01 9.99668956e-01 -1.72432646e-01
-5.37126303e-01 6.60259068e-01 -5.14303565e-01 7.09487736e-01
6.87161565e-01 -1.74595043e-01 -2.27970466e-01 -5.05198836e-01
2.28251323e-01 3.60906005e-01 3.63083780e-01 -1.27334166e+00
4.14433211e-01 3.26719075e-01 6.43839315e-02 -2.97948599e-01
7.59074807e-01 -8.28858078e-01 1.10989973e-01 1.30153134e-01
-3.75136852e-01 -5.23427844e-01 1.08597167e-01 4.36423540e-01
-2.82636672e-01 -1.79134965e-01 7.31754720e-01 2.12516040e-01
2.64952090e-02 -1.63639307e-01 -5.03605902e-01 -3.43194366e-01
-1.47185195e-03 -2.64532715e-01 -2.38709703e-01 -9.90215421e-01
-6.37877643e-01 -5.09496272e-01 -1.97361648e-01 4.50558066e-01
7.94593990e-01 -1.04135513e+00 -8.47138345e-01 4.18782413e-01
-1.64792448e-01 -5.83171785e-01 5.45994043e-01 7.28156984e-01
-4.20835465e-01 4.14435774e-01 2.25482762e-01 -4.43832308e-01
-1.32764721e+00 4.43691373e-01 7.73495555e-01 5.11988163e-01
-4.11870360e-01 1.13553238e+00 1.10914484e-01 -2.88886428e-01
7.66882002e-01 -4.57192928e-01 -3.60067338e-01 -2.19947979e-01
1.10869205e+00 7.48986483e-01 4.62512285e-01 -3.06180924e-01
-2.61205822e-01 4.88243312e-01 3.56741399e-01 -6.97060168e-01
1.05795836e+00 -5.08821070e-01 -7.22017512e-02 6.24331713e-01
1.44083512e+00 5.38820148e-01 -6.07217193e-01 -1.15704186e-01
-3.14756870e-01 -5.09306908e-01 4.10065979e-01 -9.67045188e-01
-6.72427356e-01 1.05655789e+00 1.20864391e+00 4.77725536e-01
1.32697654e+00 -2.29624137e-01 8.11883807e-01 5.12730241e-01
2.17199743e-01 -1.02147436e+00 7.80861601e-02 4.58213717e-01
1.09872019e+00 -9.81409907e-01 -4.53228801e-01 -1.27579406e-01
-2.04457209e-01 1.12504113e+00 5.67335412e-02 4.80633944e-01
9.17342603e-01 5.76096416e-01 5.93303800e-01 5.11212587e-01
-1.41569450e-01 -1.07528023e-01 4.72957224e-01 9.81008887e-01
6.44473791e-01 2.35759839e-01 3.99759889e-01 8.84440184e-01
-1.00334287e+00 -5.19843578e-01 1.70046210e-01 1.66062400e-01
-7.18539298e-01 -7.34816551e-01 -1.17346060e+00 1.31073043e-01
-6.90509140e-01 -4.55362499e-01 -5.69197610e-02 -1.71801016e-01
-2.26865068e-01 1.86177528e+00 -4.53053974e-02 -5.83289742e-01
3.40840518e-01 1.65765956e-01 3.43394876e-01 -1.73351586e-01
-4.49420333e-01 4.12340492e-01 1.05525441e-01 -3.24841261e-01
-1.74388558e-01 -3.37023109e-01 -1.02898526e+00 -1.71321511e-01
-6.39022291e-01 2.36138508e-01 1.24415815e+00 7.70275712e-01
-7.96156935e-03 8.32990646e-01 1.05127001e+00 -8.12601328e-01
-5.33446610e-01 -1.33726799e+00 -8.55673373e-01 -9.10492763e-02
1.02793097e+00 -6.03241101e-02 -9.21274900e-01 -7.73635805e-02] | [15.102478981018066, 5.7777099609375] |
4a89ba4e-bd94-4e2f-8cde-d66a8bf6d025 | instruct-fingpt-financial-sentiment-analysis | 2306.12659 | null | https://arxiv.org/abs/2306.12659v1 | https://arxiv.org/pdf/2306.12659v1.pdf | Instruct-FinGPT: Financial Sentiment Analysis by Instruction Tuning of General-Purpose Large Language Models | Sentiment analysis is a vital tool for uncovering insights from financial articles, news, and social media, shaping our understanding of market movements. Despite the impressive capabilities of large language models (LLMs) in financial natural language processing (NLP), they still struggle with accurately interpreting numerical values and grasping financial context, limiting their effectiveness in predicting financial sentiment. In this paper, we introduce a simple yet effective instruction tuning approach to address these issues. By transforming a small portion of supervised financial sentiment analysis data into instruction data and fine-tuning a general-purpose LLM with this method, we achieve remarkable advancements in financial sentiment analysis. In the experiment, our approach outperforms state-of-the-art supervised sentiment analysis models, as well as widely used LLMs like ChatGPT and LLaMAs, particularly in scenarios where numerical understanding and contextual comprehension are vital. | ['Xiao-Yang Liu', 'Hongyang Yang', 'Boyu Zhang'] | 2023-06-22 | null | null | null | null | ['sentiment-analysis'] | ['natural-language-processing'] | [-2.44687125e-01 -1.49400964e-01 -4.43444729e-01 -5.39278924e-01
-5.55935860e-01 -9.77732301e-01 5.26855111e-01 8.23061526e-01
-4.43325669e-01 5.35049856e-01 4.40089196e-01 -1.02432716e+00
5.18374920e-01 -9.36796188e-01 -6.84391856e-01 -6.35015126e-03
6.47214279e-02 5.76243959e-02 1.00484248e-02 -7.75355220e-01
1.04902232e+00 9.14385095e-02 -1.34972548e+00 6.74562454e-01
7.91945815e-01 1.00773764e+00 -1.23805739e-01 8.98355544e-01
-1.06916857e+00 1.88123047e+00 -7.69783616e-01 -7.26658762e-01
-7.35128075e-02 -2.11564466e-01 -8.72152984e-01 -3.48508507e-01
1.76938236e-01 -2.98311383e-01 3.80029440e-01 9.89992857e-01
1.01614252e-01 -1.40714020e-01 3.70543212e-01 -1.01423883e+00
-8.90394628e-01 1.13408339e+00 -8.20234001e-01 4.96260494e-01
6.00159407e-01 -4.30353284e-02 1.34547150e+00 -9.14875329e-01
3.53574187e-01 1.19105661e+00 5.86801469e-01 2.78076053e-01
-4.93603170e-01 -6.93956792e-01 4.63787049e-01 -4.94556762e-02
-4.53711867e-01 -1.92877039e-01 7.17229903e-01 -5.37397563e-01
1.49238014e+00 1.84472397e-01 6.94285452e-01 6.87710404e-01
6.55056655e-01 1.10592127e+00 1.02390766e+00 -6.42483056e-01
2.81095237e-01 3.02980661e-01 4.54254359e-01 6.40492558e-01
1.37397628e-02 -3.89384240e-01 -1.14444447e+00 7.27793947e-03
3.14978808e-01 2.60194447e-02 2.26540089e-01 2.19770312e-01
-1.22346973e+00 1.16705227e+00 -5.70294783e-02 3.50291640e-01
-2.47470334e-01 4.67555337e-02 7.58552432e-01 6.85477555e-01
8.57812703e-01 8.84761810e-01 -1.09530604e+00 -7.11611986e-01
-9.02391851e-01 4.09674376e-01 1.33670795e+00 7.19923913e-01
3.34735930e-01 7.20406920e-02 3.12571645e-01 3.74972194e-01
3.42908442e-01 4.00237352e-01 8.58871400e-01 -5.92569590e-01
9.67921674e-01 8.87093484e-01 2.38928981e-02 -1.12181020e+00
-4.59956706e-01 -2.81868160e-01 -1.75973594e-01 1.94151029e-02
5.32357454e-01 -3.18769932e-01 -3.83863389e-01 1.18936491e+00
8.81312191e-02 -2.18613178e-01 4.95229363e-01 5.32866538e-01
6.59131646e-01 8.66091192e-01 3.04447502e-01 -1.05317920e-01
1.55928099e+00 -1.10978115e+00 -6.38635457e-01 -7.22065210e-01
1.16250789e+00 -1.17699635e+00 1.40780783e+00 7.64975131e-01
-1.11773491e+00 -3.84181619e-01 -1.06271732e+00 -2.51271546e-01
-7.28824079e-01 -2.87296385e-01 1.13681412e+00 7.53756821e-01
-6.34230733e-01 5.76115906e-01 -7.92647719e-01 8.39639977e-02
2.26657629e-01 1.08001448e-01 1.31733775e-01 1.43393025e-01
-1.19458449e+00 6.72231317e-01 1.12569213e-01 -7.22427890e-02
-1.12032056e-01 -1.02171636e+00 -1.06113005e+00 5.62742017e-02
5.02023160e-01 -1.51487455e-01 1.84143698e+00 -9.17956650e-01
-1.75980830e+00 7.23398685e-01 -1.79407850e-01 -6.68168008e-01
3.89438868e-01 -6.95380092e-01 -3.51389825e-01 -4.73037064e-02
9.03008357e-02 3.51855129e-01 4.86716747e-01 -5.71927905e-01
-6.84091091e-01 -1.61950231e-01 2.51706332e-01 5.64015470e-02
-7.24625230e-01 5.33805907e-01 -6.79624826e-02 -7.89276063e-01
-8.57221857e-02 -3.92875373e-01 -4.92481291e-01 -6.47639275e-01
-1.65289998e-01 -2.93516487e-01 4.91001904e-01 -5.83872437e-01
1.47409844e+00 -1.88227487e+00 -2.77222574e-01 7.99649507e-02
1.36734366e-01 7.37860287e-03 -6.96515199e-03 5.66651583e-01
-1.10715002e-01 4.47140545e-01 1.25214100e-01 -4.33451355e-01
2.68508554e-01 7.18277544e-02 -8.60912323e-01 -4.32625674e-02
3.80152851e-01 1.11785960e+00 -1.01090991e+00 -4.69486207e-01
1.75412834e-01 1.84283275e-02 -8.82049024e-01 1.98728919e-01
-6.65829480e-01 1.92423224e-01 -8.77176523e-01 9.39699888e-01
9.06024724e-02 -3.59806597e-01 1.84816852e-01 3.44317138e-01
-3.96845847e-01 6.66219711e-01 -9.01488841e-01 1.37037778e+00
-6.94726110e-01 5.72135210e-01 -3.16867709e-01 -1.11855769e+00
9.30735886e-01 3.78657840e-02 3.20493460e-01 -9.41936374e-01
3.85180265e-01 2.70777881e-01 -2.48683304e-01 -6.28049612e-01
9.78655756e-01 -2.24419996e-01 -5.44939935e-01 9.33702469e-01
-2.43305489e-01 -4.22179639e-01 5.82551837e-01 3.93557638e-01
6.34390175e-01 1.06828548e-01 5.67825913e-01 -2.64361262e-01
6.35430455e-01 4.07420039e-01 2.00687781e-01 5.13165355e-01
-7.55145848e-02 1.83944747e-01 1.03312886e+00 -7.01204896e-01
-6.37227893e-01 -3.94985348e-01 1.77501634e-01 1.64226305e+00
-3.30409110e-01 -8.53989601e-01 -8.03267479e-01 -5.69466829e-01
-3.09108384e-02 8.18649530e-01 -4.05551314e-01 2.80565858e-01
-6.55362129e-01 -9.57980633e-01 1.28388152e-01 8.07909310e-01
3.54069114e-01 -1.18697000e+00 -7.59990394e-01 3.90053749e-01
-1.21371657e-01 -1.25207114e+00 -3.40438157e-01 4.32721436e-01
-8.32860291e-01 -1.04611742e+00 -2.90386558e-01 -6.74091339e-01
4.13151771e-01 -3.91344167e-02 1.57672882e+00 1.83855277e-02
1.82821900e-02 8.78981873e-02 -4.65257615e-01 -9.57892716e-01
-7.44387925e-01 3.21011603e-01 -1.29017234e-01 -4.90679145e-01
8.42904329e-01 -1.00323640e-01 -2.34105945e-01 -1.95437223e-01
-1.02531970e+00 6.68509491e-03 3.60757113e-01 4.14350927e-01
2.56019741e-01 -3.34450440e-03 6.93348408e-01 -1.44007468e+00
1.10861421e+00 -5.26767254e-01 -7.12382019e-01 2.19209328e-01
-7.23292589e-01 2.18581066e-01 1.16712677e+00 -3.46149623e-01
-8.24386537e-01 -3.68005544e-01 -9.28491503e-02 3.36069077e-01
1.40747413e-01 1.35564971e+00 3.29355568e-01 5.26235290e-02
4.83417839e-01 -1.19814359e-01 -1.55464515e-01 -3.02950531e-01
3.01050156e-01 4.60354626e-01 3.52144659e-01 -7.43913591e-01
4.17751551e-01 -6.60362095e-03 -2.88866907e-01 -4.84852821e-01
-1.35852778e+00 -4.80759025e-01 -6.53890431e-01 -9.89094526e-02
7.79497445e-01 -8.99008989e-01 -9.08046246e-01 6.65848672e-01
-8.26256812e-01 -2.98604161e-01 -2.64063299e-01 3.56006324e-01
-2.70437002e-01 4.17470664e-01 -1.14897013e+00 -6.98721945e-01
-3.26454878e-01 -1.07810915e+00 7.52325356e-01 2.73058265e-01
-6.60390079e-01 -1.45178986e+00 4.08926569e-02 7.33519793e-01
4.68512625e-01 2.80202389e-01 1.08265793e+00 -9.88465250e-01
-2.90558547e-01 -4.86016244e-01 -8.81610066e-02 3.75499219e-01
-1.70634799e-02 2.38768905e-01 -7.56207108e-01 2.06063166e-01
2.20951751e-01 -8.11070383e-01 7.11704195e-01 1.45005614e-01
1.00195169e+00 -4.11289811e-01 2.39745826e-01 1.64418638e-01
1.22474110e+00 1.94989249e-01 3.61853205e-02 8.09232414e-01
5.31966686e-01 7.66782403e-01 8.06995749e-01 6.17971480e-01
7.76394129e-01 -2.79934019e-01 2.17271030e-01 2.28313338e-02
6.42297328e-01 -2.58074462e-01 8.09543431e-01 1.50142562e+00
1.29999533e-01 -7.33125284e-02 -1.25507665e+00 3.74336034e-01
-1.86295390e+00 -4.92243558e-01 -6.59748614e-02 1.52579975e+00
1.30899739e+00 7.47403562e-01 -1.84279606e-01 2.32292175e-01
-1.11328624e-01 4.58001107e-01 -5.39996326e-01 -1.01639032e+00
-7.78619200e-02 4.44201559e-01 5.11177957e-01 4.60007876e-01
-1.03208828e+00 1.09866774e+00 6.77169800e+00 6.21237278e-01
-1.22712243e+00 -3.68943185e-01 9.40184951e-01 -2.20308378e-02
-5.23868740e-01 -1.62154704e-01 -9.82028782e-01 8.99760276e-02
1.29073155e+00 -2.96743602e-01 1.66545585e-01 1.06698227e+00
2.15052053e-01 -2.70179629e-01 -1.14211500e+00 5.04960299e-01
1.12907715e-01 -1.58085763e+00 1.31928235e-01 -3.19136232e-01
7.68085480e-01 -7.99547136e-02 3.13600183e-01 4.60606724e-01
4.92844313e-01 -1.08480906e+00 1.03008819e+00 2.48338714e-01
2.18678132e-01 -9.08978641e-01 9.95081782e-01 5.11256337e-01
-1.11122060e+00 -1.93428501e-01 -2.01145560e-01 -8.39975953e-01
3.04157864e-02 5.33361852e-01 -5.56066096e-01 2.54370511e-01
6.42365217e-01 1.05398893e+00 -7.19331861e-01 -4.01104847e-03
-3.00147682e-01 8.16907108e-01 -1.29392684e-01 -5.25495589e-01
6.70690656e-01 -1.89425394e-01 -1.92017958e-01 1.30172265e+00
-2.26103589e-02 4.06251669e-01 2.26135239e-01 6.49156153e-01
-1.34755582e-01 5.91126621e-01 -2.27911815e-01 -6.78519011e-01
-8.89690146e-02 1.09726906e+00 -1.04949677e+00 -6.56291068e-01
-9.94009495e-01 4.06110495e-01 2.46946797e-01 1.77150011e-01
-6.54120028e-01 -3.65456402e-01 6.07458830e-01 -6.49954528e-02
1.85234979e-01 -4.59368497e-01 -7.56313622e-01 -1.54363251e+00
1.32530585e-01 -1.24829960e+00 4.56779569e-01 -6.94354832e-01
-1.24995339e+00 4.68321919e-01 -1.56566918e-01 -9.09072101e-01
-6.06786847e-01 -1.05805731e+00 -7.67813981e-01 6.23362184e-01
-1.95118654e+00 -6.94222093e-01 2.70157933e-01 4.08965409e-01
7.77285933e-01 -3.77278924e-01 7.69079685e-01 -5.73880263e-02
-5.07793486e-01 3.47729832e-01 -3.96355540e-02 2.93032438e-01
7.37987936e-01 -1.42037499e+00 8.18663001e-01 8.31207037e-01
2.12850645e-01 8.80876362e-01 8.62688780e-01 -4.75335002e-01
-1.54699039e+00 -6.89956129e-01 1.19789624e+00 -7.17123032e-01
1.50115502e+00 -4.62647259e-01 -8.19481015e-01 5.50119638e-01
3.00962389e-01 -3.50695789e-01 1.10613108e+00 1.92531303e-01
-5.68622291e-01 3.26927722e-01 -7.38540232e-01 7.00356901e-01
3.75907987e-01 -7.09924638e-01 -9.24021184e-01 3.21931750e-01
1.10005987e+00 -6.10706687e-01 -9.43702281e-01 -1.69664733e-02
4.47961599e-01 -9.34045732e-01 9.04194236e-01 -1.04079711e+00
1.34760273e+00 2.20324010e-01 -6.19942807e-02 -1.06955957e+00
5.79475999e-01 -6.97594583e-01 -1.89492598e-01 1.12155461e+00
7.45452344e-01 -6.86110258e-01 9.84438062e-01 7.85743117e-01
7.15745464e-02 -9.01136816e-01 -1.89266533e-01 -2.18912959e-01
5.74858725e-01 -1.02150536e+00 6.48485482e-01 1.20957553e+00
6.20590031e-01 1.79562360e-01 6.07302524e-02 -2.63584346e-01
3.08423579e-01 5.63997567e-01 6.44489467e-01 -7.60221183e-01
-3.45758677e-01 -5.51004887e-01 2.45703720e-02 -1.09358490e+00
3.15413415e-01 -5.53069353e-01 -3.07459235e-01 -1.20409846e+00
-9.07758474e-02 -9.49161649e-02 -3.76194537e-01 4.52885091e-01
-1.28329113e-01 6.18406907e-02 2.72939473e-01 5.79619482e-02
-6.34690404e-01 2.55538933e-02 1.17970908e+00 -5.90008758e-02
-1.33763388e-01 -3.60889323e-02 -1.12364829e+00 1.16108561e+00
9.33441520e-01 -3.56749386e-01 -1.72431946e-01 -3.02546024e-01
1.16799462e+00 1.73722163e-01 -1.54476121e-01 -4.39940363e-01
4.42312628e-01 -5.41028440e-01 3.18433881e-01 -7.62707531e-01
-3.58916789e-01 -3.96226406e-01 -9.78620529e-01 3.69825691e-01
-5.90295851e-01 5.14789581e-01 4.77488846e-01 1.24308474e-01
-8.13113332e-01 -3.76170307e-01 9.94782820e-02 -4.21433359e-01
-1.01001477e+00 -9.80055034e-02 -6.36567235e-01 3.81958365e-01
7.40980327e-01 -3.15167531e-02 -4.48359191e-01 -3.62950176e-01
-3.39715630e-01 4.95565236e-01 1.74660787e-01 6.75248921e-01
4.07513112e-01 -6.80790842e-01 -5.26295781e-01 4.27837223e-01
-3.28279771e-02 1.03625804e-01 -1.68457612e-01 5.45955241e-01
-9.20006633e-01 6.93063855e-01 -4.42161784e-03 -2.11827651e-01
-1.00242889e+00 4.54663336e-01 -1.21285468e-01 -8.05545628e-01
-1.81523263e-01 1.00679743e+00 7.17955753e-02 -4.65262443e-01
1.80650994e-01 -1.36066139e+00 -4.67879713e-01 3.27814430e-01
7.47765601e-01 -1.15578689e-01 1.63835406e-01 -1.49487138e-01
-1.21100008e-01 7.23155558e-01 -2.13675827e-01 4.84160446e-02
1.61998260e+00 -4.70264666e-02 -3.60390067e-01 9.01513875e-01
1.08290374e+00 4.68882710e-01 -9.89139736e-01 -2.21886992e-01
5.82230747e-01 -7.92169049e-02 -1.15086779e-01 -6.61535442e-01
-9.24315453e-01 9.14917350e-01 -4.32088703e-01 5.54750443e-01
9.81193125e-01 -1.99436247e-01 1.14052534e+00 7.03582406e-01
2.96310961e-01 -1.20530450e+00 2.92668819e-01 9.81643140e-01
4.29762363e-01 -1.59408045e+00 1.13427572e-01 -2.94145644e-01
-9.34953094e-01 1.65946257e+00 3.78977239e-01 -1.19117402e-01
7.60736108e-01 8.12683642e-01 5.19198060e-01 -3.84940714e-01
-1.03657937e+00 4.01819766e-01 9.03677717e-02 -1.39739573e-01
1.06510615e+00 -1.51240140e-01 5.89423254e-02 1.04339671e+00
-1.04513907e+00 9.50643793e-02 8.16908658e-01 1.30202961e+00
-5.08891106e-01 -1.08088708e+00 -3.54408890e-01 3.02125067e-01
-1.15420425e+00 -6.51414156e-01 -3.53063822e-01 7.88829684e-01
-3.87253344e-01 9.67890382e-01 2.15703338e-01 -1.29809052e-01
3.53507876e-01 2.86186188e-01 1.62268039e-02 -8.30129325e-01
-1.25294447e+00 -1.43937886e-01 1.80416867e-01 -6.33093238e-01
-4.61081237e-01 -7.49728858e-01 -1.54617584e+00 -6.28355324e-01
-6.00116886e-02 2.88991690e-01 6.68981493e-01 1.30654776e+00
-1.42135903e-01 5.67971766e-01 4.56633210e-01 -2.93050885e-01
-8.50582957e-01 -7.11803854e-01 -4.53521460e-01 1.52806386e-01
4.12202835e-01 1.86397247e-02 -3.88197452e-01 3.99641812e-01] | [11.105313301086426, 7.212611675262451] |
3eab743c-847f-480c-a722-44984dae07ae | functional-causal-bayesian-optimization | 2306.06409 | null | https://arxiv.org/abs/2306.06409v1 | https://arxiv.org/pdf/2306.06409v1.pdf | Functional Causal Bayesian Optimization | We propose functional causal Bayesian optimization (fCBO), a method for finding interventions that optimize a target variable in a known causal graph. fCBO extends the CBO family of methods to enable functional interventions, which set a variable to be a deterministic function of other variables in the graph. fCBO models the unknown objectives with Gaussian processes whose inputs are defined in a reproducing kernel Hilbert space, thus allowing to compute distances among vector-valued functions. In turn, this enables to sequentially select functions to explore by maximizing an expected improvement acquisition functional while keeping the typical computational tractability of standard BO settings. We introduce graphical criteria that establish when considering functional interventions allows attaining better target effects, and conditions under which selected interventions are also optimal for conditional target effects. We demonstrate the benefits of the method in a synthetic and in a real-world causal graph. | ['Silvia Chiappa', 'Alexis Bellot', 'Virginia Aglietti', 'Limor Gultchin'] | 2023-06-10 | null | null | null | null | ['gaussian-processes', 'bayesian-optimization'] | ['methodology', 'methodology'] | [ 3.98112297e-01 4.18342859e-01 -4.67325449e-01 -4.72087339e-02
-5.89529037e-01 -4.56333846e-01 7.26825476e-01 2.64040321e-01
-1.59008563e-01 9.80967820e-01 4.48496193e-01 -3.24353606e-01
-9.30032253e-01 -1.03709090e+00 -1.01493454e+00 -9.17153597e-01
-6.75887108e-01 4.28797007e-01 -1.94555670e-01 3.10635418e-01
2.05282822e-01 6.18466377e-01 -9.88456607e-01 -3.35185021e-01
1.05239737e+00 3.49013209e-01 2.57304996e-01 7.86541462e-01
5.27573824e-01 6.78233981e-01 -4.08253938e-01 -2.65132904e-01
-3.28482948e-02 -5.38049757e-01 -7.40949631e-01 -4.35080715e-02
-6.48630708e-02 1.29357621e-01 -2.42800102e-01 1.15157950e+00
3.50411326e-01 3.50644916e-01 1.19530618e+00 -1.28817570e+00
-7.87746847e-01 8.30544055e-01 -1.30500495e-01 -4.83621322e-02
4.41282600e-01 2.32425660e-01 1.34320593e+00 -5.22697747e-01
3.56465340e-01 1.87283325e+00 4.02462512e-01 3.44195604e-01
-2.09671211e+00 -2.91178674e-01 -3.54948523e-03 -4.31826542e-04
-1.14450741e+00 -4.28029895e-01 4.42987859e-01 -8.17408860e-01
4.75610703e-01 3.74533892e-01 6.83818579e-01 1.15595806e+00
6.83117211e-01 5.08247852e-01 1.12655401e+00 -4.81577516e-01
5.29151440e-01 -1.73215702e-01 2.38615960e-01 7.57948399e-01
3.79940301e-01 6.96687996e-01 -7.10601687e-01 -5.87212622e-01
6.32552385e-01 -1.61618054e-01 -5.90839446e-01 -6.21579051e-01
-1.15549433e+00 1.29943693e+00 2.93552995e-01 3.13131325e-02
-7.73443401e-01 8.56819451e-01 -1.84940860e-01 8.41702670e-02
2.67746568e-01 6.88180149e-01 -2.74064481e-01 1.98110834e-01
-3.28569859e-01 3.99978191e-01 8.20595443e-01 8.69015157e-01
5.71656108e-01 -1.30414307e-01 -7.46935904e-01 2.25013912e-01
6.16538823e-01 8.00148368e-01 -3.92305434e-01 -9.79387164e-01
2.66278893e-01 1.94643736e-01 4.03188765e-01 -8.93470883e-01
-4.06613469e-01 -3.18910956e-01 -5.22309542e-01 1.09847113e-01
6.78961694e-01 -2.38358781e-01 -6.77278399e-01 1.97653770e+00
3.42223257e-01 1.63970873e-01 -2.81135201e-01 5.67356467e-01
-4.78680283e-02 7.20271051e-01 3.20782125e-01 -6.61986768e-01
9.88266706e-01 -2.15197459e-01 -7.97268689e-01 1.32073238e-01
4.06667739e-01 -2.66606092e-01 1.35959184e+00 4.18877810e-01
-8.18939090e-01 -3.18404511e-02 -8.14994991e-01 6.46847963e-01
1.18646398e-02 -3.06888759e-01 6.74386621e-01 9.35024023e-01
-1.13170171e+00 8.89889717e-01 -8.65257442e-01 -2.20390614e-02
5.23776770e-01 4.40677315e-01 -8.46863985e-02 -1.22090466e-02
-1.29429662e+00 8.36480737e-01 3.02884370e-01 -9.58656296e-02
-1.88369441e+00 -1.15766406e+00 -7.38070190e-01 4.20427918e-01
7.77659953e-01 -8.07257414e-01 9.92495894e-01 -6.49908960e-01
-1.47640717e+00 1.37061939e-01 -1.40117392e-01 -4.82405007e-01
3.32459539e-01 -2.27199465e-01 -1.58705592e-01 4.43358980e-02
1.59798592e-01 1.44104317e-01 1.12488401e+00 -1.16765070e+00
-3.74510676e-01 -4.19505388e-01 2.15481773e-01 -4.70696017e-02
-1.55998364e-01 -3.39523181e-02 -1.01751462e-01 -3.39684397e-01
-4.40998763e-01 -8.30868602e-01 -5.94269872e-01 -3.75349104e-01
-9.11848426e-01 -3.95948261e-01 2.43468657e-01 -5.25654674e-01
1.31435633e+00 -1.75985920e+00 7.00283170e-01 5.58600187e-01
2.90411055e-01 -6.88418865e-01 5.17415553e-02 4.18098718e-01
-1.14854507e-01 4.40804332e-01 -4.72436666e-01 2.02418834e-01
2.04350621e-01 -4.33927700e-02 -1.15688235e-01 9.78497028e-01
3.55624259e-01 7.90153801e-01 -1.07511699e+00 -4.02469546e-01
1.76024307e-02 1.60146683e-01 -6.61526024e-01 2.67779380e-01
-4.28006440e-01 6.83159649e-01 -8.02833438e-01 3.09835941e-01
6.12554885e-02 -8.01946968e-02 1.87607870e-01 2.86425650e-01
-3.90541134e-03 -2.46939398e-02 -1.24949396e+00 9.89335895e-01
-3.63559365e-01 4.60661888e-01 -1.14882685e-01 -1.23682082e+00
6.08016849e-01 3.21826518e-01 5.87079227e-01 -1.44286202e-02
1.06699295e-01 -2.27539316e-01 9.56845656e-02 -4.35110658e-01
-2.49225795e-01 -3.78760576e-01 -2.11678922e-01 5.63539490e-02
1.02226086e-01 -2.10623771e-01 1.72444973e-02 1.28381774e-01
1.46427166e+00 4.36852463e-02 5.16000628e-01 -8.86817276e-01
2.48373553e-01 -5.91844507e-02 4.83148009e-01 1.29394078e+00
-1.45960063e-01 -1.47578437e-02 1.17102182e+00 2.55185992e-01
-7.29248703e-01 -1.43357289e+00 -2.03817159e-01 8.31353784e-01
-4.95699607e-02 -4.33761850e-02 -6.63661480e-01 -8.00697505e-01
2.98601747e-01 1.31901407e+00 -1.04186678e+00 -2.78783649e-01
-3.78433049e-01 -7.87104249e-01 2.69824564e-01 6.88097253e-02
-1.72420949e-01 -7.95698643e-01 -4.88496840e-01 2.33043939e-01
2.96145350e-01 -1.98216781e-01 -6.13519907e-01 2.84709126e-01
-8.06522131e-01 -1.48268843e+00 -4.02358204e-01 -5.03292531e-02
5.78859389e-01 -3.40964288e-01 9.85013247e-01 -5.59790015e-01
-1.63254172e-01 7.61301577e-01 1.10196754e-01 -3.89649779e-01
-4.74403828e-01 -4.09932494e-01 -1.76343303e-02 9.61624831e-02
-2.68729627e-01 -3.10855329e-01 -4.33701545e-01 9.69231427e-02
-7.23581135e-01 -3.39786053e-01 2.40230158e-01 1.12775981e+00
4.78948295e-01 3.00072700e-01 6.46257758e-01 -9.23683047e-01
9.13857698e-01 -7.95808315e-01 -1.12453902e+00 5.22263050e-01
-8.88194740e-01 4.93555754e-01 3.39883566e-01 -5.28412223e-01
-1.14231360e+00 4.16020071e-03 6.18036687e-01 -1.11059882e-01
1.89219221e-01 7.90958464e-01 -4.83254373e-01 3.68207276e-01
8.81232142e-01 -4.00286317e-01 -1.80442140e-01 -2.05351904e-01
7.58790076e-01 3.04854624e-02 1.78650349e-01 -9.84086633e-01
5.36836445e-01 3.43169361e-01 5.89225650e-01 -5.34205198e-01
-6.86153471e-01 -1.15792468e-01 -4.83533084e-01 -5.45058608e-01
1.06217253e+00 -3.17444056e-01 -1.33533168e+00 -1.88949451e-01
-9.71260369e-01 -5.84783614e-01 -3.47072244e-01 8.03788483e-01
-9.26499248e-01 -2.73304522e-01 -2.26586834e-02 -1.31921184e+00
2.65301347e-01 -1.25910568e+00 8.11611116e-01 1.54403567e-01
-2.91642547e-01 -1.42322195e+00 9.48187560e-02 -1.98483989e-01
-2.51404345e-02 5.43694854e-01 1.24059391e+00 -4.15972501e-01
-6.00385129e-01 2.33179089e-02 2.03355819e-01 5.64265903e-03
1.10604152e-01 1.28288269e-01 -5.37415743e-01 -1.79232270e-01
5.82308583e-02 2.35470846e-01 7.04779327e-01 1.24018443e+00
9.59290802e-01 -5.66000462e-01 -7.23402023e-01 2.70612895e-01
1.33841300e+00 6.89552009e-01 3.10221344e-01 1.86232682e-02
6.76345944e-01 7.62286365e-01 5.97543240e-01 3.81290108e-01
-6.27457649e-02 6.85237765e-01 4.83527452e-01 2.48077214e-01
4.26215798e-01 -3.66958559e-01 5.06642222e-01 -2.70497240e-02
7.87345693e-02 -3.61362696e-01 -9.80607450e-01 6.90694273e-01
-2.00147676e+00 -1.00195348e+00 -3.31579149e-01 2.58787489e+00
1.02246082e+00 5.34483744e-03 2.47795999e-01 -4.07469302e-01
1.11045313e+00 -2.89214045e-01 -5.99999607e-01 -1.51572302e-01
1.24692358e-01 5.47626376e-01 1.04926205e+00 1.02235103e+00
-1.06657684e+00 5.67070961e-01 7.17429256e+00 7.24581480e-01
-5.05765617e-01 2.67637163e-01 6.81780934e-01 -8.95107239e-02
-5.78111291e-01 4.02396739e-01 -6.07988834e-01 2.68215209e-01
1.13171208e+00 -5.88462830e-01 7.86390066e-01 6.71035588e-01
8.61011446e-01 -1.07564233e-01 -1.46630573e+00 2.46612504e-01
-6.07041776e-01 -1.24705601e+00 -4.60882038e-01 3.42386752e-01
8.07599068e-01 -5.34680068e-01 -2.66799368e-02 1.31363750e-01
1.15738010e+00 -1.48344707e+00 6.63663507e-01 1.06559265e+00
5.56100249e-01 -1.00656188e+00 2.05851182e-01 1.71508208e-01
-6.50566280e-01 -3.39933336e-01 -1.41865000e-01 2.25150973e-01
2.00274482e-01 9.08941269e-01 -1.09302855e+00 4.38336581e-01
4.05095220e-01 5.29472589e-01 -5.83350733e-02 1.01856220e+00
-6.09992802e-01 1.13618588e+00 -1.07662268e-01 -2.92499781e-01
7.52716139e-02 -4.74404722e-01 1.04224455e+00 1.06978381e+00
3.23956758e-01 -1.84244126e-01 1.06023878e-01 1.32216775e+00
2.02427268e-01 8.00498500e-02 -9.19566274e-01 -3.13725740e-01
6.52786672e-01 6.82550132e-01 -5.46042204e-01 -1.71487316e-01
-3.06913313e-02 5.26411533e-01 1.50509432e-01 6.94932461e-01
-9.95351791e-01 -2.12683201e-01 6.50400043e-01 -1.00318514e-01
2.42761895e-02 -2.42890026e-02 -4.67799872e-01 -6.87729597e-01
-5.12734175e-01 -7.06749260e-01 7.00694978e-01 -4.00778025e-01
-1.13339162e+00 -3.21795225e-01 6.58820391e-01 -5.93788862e-01
-3.84432077e-02 -4.38914478e-01 -4.85531539e-01 1.10452664e+00
-5.68408728e-01 -7.07745969e-01 3.32979113e-01 5.98002911e-01
2.02619433e-01 1.78286076e-01 4.03551340e-01 -2.40211010e-01
-6.08565688e-01 6.92609176e-02 2.39326969e-01 -4.70176280e-01
3.70322257e-01 -1.72505605e+00 -1.90169573e-01 8.88168037e-01
-6.11116290e-02 7.83157706e-01 1.07035053e+00 -1.00880492e+00
-1.58967614e+00 -9.86409962e-01 4.34984267e-01 -4.48820949e-01
9.65069592e-01 -2.71891505e-01 -4.84692395e-01 7.51421511e-01
1.52907642e-02 -4.40520972e-01 1.50947362e-01 5.25468469e-01
9.07172635e-02 -4.86738496e-02 -1.13555884e+00 1.02479136e+00
1.07654941e+00 -2.53190219e-01 -3.80790859e-01 5.77557504e-01
9.74980533e-01 9.17789638e-02 -1.13322759e+00 3.37807387e-01
3.02585840e-01 -5.10768592e-01 1.03543437e+00 -1.03722000e+00
3.85105669e-01 -2.28932321e-01 -1.62151650e-01 -1.76517618e+00
-4.81233627e-01 -9.72275436e-01 -6.27691522e-02 1.04970670e+00
4.42107648e-01 -9.03326750e-01 3.03759277e-01 6.15860462e-01
8.86923447e-03 -5.23688793e-01 -9.98736918e-01 -7.82345831e-01
-1.54101653e-02 -6.43325746e-01 5.22305846e-01 8.25395286e-01
1.24687999e-01 3.89854491e-01 -3.92593652e-01 5.28772354e-01
9.80872273e-01 -2.18631923e-01 4.80235636e-01 -1.16435182e+00
-7.33512998e-01 -6.28372490e-01 -2.36239627e-01 -5.79936862e-01
4.39002573e-01 -8.80931556e-01 3.41620833e-01 -1.27705824e+00
3.38963747e-01 -4.37409222e-01 -1.23065569e-01 2.36867279e-01
-7.10424244e-01 -7.69034266e-01 3.56295556e-02 -1.60744071e-01
5.52602001e-02 7.45797873e-01 1.18755329e+00 -2.11949378e-01
-5.24384260e-01 2.84942985e-01 -7.24490285e-01 5.34262717e-01
5.72594702e-01 -8.47109079e-01 -6.88334763e-01 4.15428013e-01
1.11521870e-01 6.65690958e-01 5.53500533e-01 -3.61302853e-01
-1.29597500e-01 -9.55047071e-01 1.57256126e-01 -1.63045809e-01
-6.58096448e-02 -6.47626579e-01 6.38478875e-01 8.21318924e-01
-7.65733778e-01 -2.25807682e-01 -1.39731422e-01 1.09093511e+00
3.50948602e-01 -5.63042223e-01 7.90148914e-01 3.24408621e-01
-1.34142578e-01 1.60841376e-01 -5.79016984e-01 -8.12000558e-02
1.14555383e+00 3.41486722e-01 -2.02043563e-01 -6.26702070e-01
-9.58521962e-01 3.88251275e-01 7.75764436e-02 -3.64035591e-02
4.20768827e-01 -1.26076460e+00 -7.55133927e-01 -5.55297852e-01
-1.80869564e-01 -3.84027183e-01 -1.11940093e-01 1.04401851e+00
-9.49928090e-02 4.23548520e-01 3.86011541e-01 -6.08318627e-01
-9.68034267e-01 8.40764463e-01 5.68116069e-01 -3.77114415e-01
-3.73991251e-01 8.06796789e-01 6.81237459e-01 -1.55595839e-01
2.70427181e-03 -2.32543558e-01 -1.22249566e-01 -1.17235407e-01
1.80936843e-01 6.78752065e-01 -3.84078026e-01 1.74888335e-02
-2.06681758e-01 -1.82879955e-01 7.56289423e-01 -6.97662890e-01
1.27431035e+00 -6.11871993e-03 -1.87077492e-01 6.25623226e-01
9.07182038e-01 2.24437848e-01 -1.46798193e+00 5.15809581e-02
4.85644400e-01 -5.61584055e-01 1.91340059e-01 -6.74322486e-01
-6.27811372e-01 4.09342498e-01 4.63275492e-01 5.67953050e-01
9.43265319e-01 2.28850573e-01 -3.95789206e-01 1.14472784e-01
1.66961044e-01 -7.00408816e-01 2.23208386e-02 6.45939410e-02
1.10459650e+00 -6.59680963e-01 6.16601370e-02 -5.32742143e-01
-4.26469207e-01 1.01980865e+00 2.63255145e-02 -2.12640375e-01
8.51104915e-01 6.29236996e-02 -7.48712242e-01 -4.72145259e-01
-7.90830791e-01 -2.79144019e-01 5.10289371e-01 7.45849788e-01
2.87265956e-01 6.02693915e-01 -5.55110216e-01 3.24610084e-01
7.67282173e-02 -2.96524107e-01 3.74710977e-01 5.32880008e-01
-2.87396997e-01 -8.77132118e-01 -8.10159326e-01 6.97450101e-01
-3.75594586e-01 -2.06208751e-01 -1.67320982e-01 8.81482005e-01
-3.43269110e-01 1.25687325e+00 -1.63327515e-01 3.60389873e-02
4.49231803e-01 3.00098751e-02 5.52489936e-01 -6.31957412e-01
-2.55850881e-01 2.32296348e-01 3.82421345e-01 -6.74924135e-01
-5.76539896e-02 -1.02674305e+00 -9.36165750e-01 -1.01248547e-01
-6.48532450e-01 1.62105128e-01 4.70246255e-01 1.04618359e+00
-1.25769358e-02 8.26156616e-01 7.62006164e-01 -5.45942068e-01
-7.97691226e-01 -9.08232987e-01 -6.98369622e-01 -5.50972484e-03
2.75142848e-01 -1.06582475e+00 -3.90412271e-01 2.11174846e-01] | [7.727949142456055, 5.295323848724365] |
7094d3a4-7308-4906-a63d-8d987f93583d | aisfg-abundant-information-slot-filling | null | null | https://aclanthology.org/2022.naacl-main.308 | https://aclanthology.org/2022.naacl-main.308.pdf | AISFG: Abundant Information Slot Filling Generator | As an essential component of task-oriented dialogue systems, slot filling requires enormous labeled training data in a certain domain. However, in most cases, there is little or no target domain training data is available in the training stage. Thus, cross-domain slot filling has to cope with the data scarcity problem by zero/few-shot learning. Previous researches on zero/few-shot cross-domain slot filling focus on slot descriptions and examples while ignoring the slot type ambiguity and example ambiguity issues. To address these problems, we propose Abundant Information Slot Filling Generator (AISFG), a generative model with a novel query template that incorporates domain descriptions, slot descriptions, and examples with context. Experimental results show that our model outperforms state-of-the-art approaches in zero/few-shot slot filling task. | ['LiWen Wang', 'Zhongbao Zhang', 'Junda Ye', 'Yang Yan'] | null | null | null | null | naacl-2022-7 | ['slot-filling', 'task-oriented-dialogue-systems'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.33362496e-01 5.63460112e-01 -5.55605292e-01 -3.84051144e-01
-7.63455451e-01 -1.12286560e-01 6.71719253e-01 1.47263795e-01
-4.76943970e-01 1.29499352e+00 1.51888683e-01 -2.21782655e-01
1.01767533e-01 -9.54087913e-01 1.59338545e-02 -2.50230849e-01
4.15510923e-01 1.16929555e+00 7.30780184e-01 -8.00354004e-01
2.03486115e-01 -5.82343459e-01 -1.31144166e+00 2.56867498e-01
1.41713059e+00 6.30330384e-01 7.37432122e-01 1.34818241e-01
-1.39277196e+00 3.14122528e-01 -8.02928269e-01 -2.33706564e-01
-2.24043205e-02 -6.40261710e-01 -9.14308429e-01 4.45194125e-01
-3.79818767e-01 -3.50780100e-01 -3.53319764e-01 9.23448801e-01
6.53351903e-01 3.72526675e-01 3.42337042e-01 -1.17790174e+00
-4.60648805e-01 7.19179749e-01 -3.11447471e-01 1.93833768e-01
3.40987027e-01 -1.04623452e-01 9.12995517e-01 -9.50254738e-01
9.34715390e-01 1.20025349e+00 3.38367492e-01 1.00819719e+00
-9.41948831e-01 -5.11553049e-01 1.09152347e-01 1.07689112e-01
-9.94893372e-01 -3.26009810e-01 8.92101645e-01 -1.04945846e-01
9.53432679e-01 -1.56383917e-01 5.29545784e-01 1.00053513e+00
-5.56839824e-01 1.20043075e+00 8.75018895e-01 -6.87232196e-01
4.92142469e-01 5.78926504e-01 2.83973753e-01 2.46770471e-01
4.33277749e-02 -2.36671895e-01 -5.03687501e-01 -2.77042538e-01
7.67269135e-01 6.39700741e-02 -3.52640860e-02 -5.52356601e-01
-7.40654647e-01 1.10027874e+00 -2.33226910e-01 4.31556523e-01
-1.86974198e-01 -7.44526684e-01 7.85149693e-01 3.39069337e-01
7.80971229e-01 5.15761793e-01 -6.49520218e-01 -3.80426168e-01
-9.26759481e-01 5.11590123e-01 9.88210022e-01 1.32840693e+00
9.70100462e-01 5.54934070e-02 -4.64265496e-01 1.39249146e+00
4.10425663e-02 1.65156454e-01 1.01655209e+00 -5.05006194e-01
8.46081197e-01 7.81267762e-01 3.19854975e-01 -2.08927020e-01
-3.38636130e-01 -2.09014371e-01 -4.62820888e-01 -4.51507956e-01
2.30046585e-01 -3.49089861e-01 -9.59087610e-01 1.67380905e+00
6.91195071e-01 1.05208412e-01 6.30200863e-01 8.58351469e-01
1.17176270e+00 6.63119853e-01 5.09788990e-01 -5.45618832e-01
1.56849086e+00 -1.12022460e+00 -1.20005894e+00 -7.48589337e-01
6.94836497e-01 -5.97388446e-01 1.36527956e+00 -2.15950474e-01
-8.42798054e-01 -5.43630838e-01 -8.39932680e-01 -1.53739154e-01
-4.28365082e-01 -3.33064720e-02 8.63244176e-01 6.66489005e-01
-2.51881450e-01 3.39479238e-01 -3.22978288e-01 -5.30147016e-01
2.59995908e-01 -8.35741386e-02 -2.27638707e-01 -4.41777617e-01
-1.84808612e+00 6.74469531e-01 8.44877839e-01 -6.96993828e-01
-5.81318140e-01 -5.88040113e-01 -1.28393531e+00 1.18795924e-01
9.67643142e-01 -4.31998074e-01 1.63492703e+00 -3.51146400e-01
-1.39304221e+00 6.78116202e-01 -2.79177934e-01 -5.99391043e-01
3.26690018e-01 -6.97275102e-02 -4.34901386e-01 -1.51856035e-01
5.94414175e-01 7.28705823e-01 5.57930887e-01 -9.71991599e-01
-7.89475262e-01 -2.46335506e-01 1.72935680e-01 7.63983250e-01
-2.44927347e-01 -3.79692078e-01 -4.42155570e-01 -3.68716657e-01
2.70900667e-01 -4.32434559e-01 -4.72731948e-01 -4.79030132e-01
-3.14882249e-01 -6.81775987e-01 9.97603714e-01 -5.38054407e-01
1.31301033e+00 -2.19402981e+00 -3.95203382e-01 -4.21526670e-01
-1.18346721e-01 5.76156020e-01 3.59866731e-02 6.68255508e-01
3.08834404e-01 -1.69749200e-01 -2.87925422e-01 -4.01988536e-01
1.64923772e-01 5.87579906e-01 -4.24938649e-01 -3.42618227e-01
2.04192430e-01 7.29125679e-01 -1.23789501e+00 -9.43983972e-01
3.18289459e-01 -6.91176429e-02 -4.40634191e-01 4.43523884e-01
-7.64152110e-01 2.26829529e-01 -6.50862515e-01 6.10582590e-01
6.44288063e-01 -2.47742265e-01 4.73659247e-01 3.16573501e-01
1.32851213e-01 7.77635753e-01 -1.05396628e+00 2.20265818e+00
-4.43940282e-01 1.59174532e-01 -3.37058097e-01 -1.09790778e+00
1.03909063e+00 6.81576371e-01 2.91411608e-01 -9.48307455e-01
-2.98637152e-02 4.64651048e-01 -1.83973297e-01 -3.67639899e-01
1.01579964e+00 -5.56878388e-01 -4.50831503e-01 4.64337319e-01
5.06769776e-01 -2.94868857e-01 6.49133205e-01 2.92954057e-01
9.01791334e-01 2.24635303e-01 6.16946995e-01 2.12231740e-01
1.92129090e-01 5.69041312e-01 9.40280497e-01 6.10804737e-01
-3.90186578e-01 4.30707633e-01 3.96584630e-01 -1.24621995e-01
-1.27776814e+00 -7.10227847e-01 -1.43005878e-01 1.27164662e+00
3.74807358e-01 -4.72695947e-01 -8.33651125e-01 -6.65600598e-01
-2.00788230e-01 8.19852889e-01 -3.03760976e-01 -1.07633695e-01
-2.67154276e-01 -4.62038279e-01 3.17872494e-01 1.55007064e-01
8.55011165e-01 -1.48159349e+00 -5.20566881e-01 6.98027968e-01
-3.44279051e-01 -1.10251939e+00 -3.65551442e-01 3.08105022e-01
-8.33509266e-01 -1.00780988e+00 -8.76131594e-01 -8.91336620e-01
5.01091897e-01 6.32173598e-01 1.34036219e+00 -1.40687615e-01
-3.28319281e-01 -1.10393032e-01 -7.74295390e-01 -3.22315276e-01
-3.49133343e-01 2.03168198e-01 -1.38921708e-01 -7.61286914e-01
1.04699337e+00 -4.51917022e-01 -2.82213748e-01 3.66091728e-01
-9.41980839e-01 2.97457457e-01 4.66576993e-01 1.45567465e+00
3.88673782e-01 1.97028041e-01 8.86800766e-01 -1.42349470e+00
1.03565919e+00 -7.77121961e-01 -2.06371725e-01 3.65616560e-01
-5.88096917e-01 4.77569737e-02 3.64460081e-01 -4.84644562e-01
-1.44784570e+00 -6.43010437e-02 -6.56741709e-02 -7.23974556e-02
-2.95116782e-01 5.01795709e-01 -4.37549949e-01 5.79399049e-01
8.22771013e-01 5.24203956e-01 -8.83934945e-02 -6.06779575e-01
4.21990603e-01 7.90949523e-01 3.03386539e-01 -6.46624327e-01
4.58410680e-01 -1.05788475e-02 -7.24700630e-01 -8.39825392e-01
-1.00508618e+00 -9.44172978e-01 -4.57930028e-01 3.55471551e-01
5.47026753e-01 -8.53176951e-01 7.39269853e-02 9.08961892e-02
-1.03381324e+00 -2.28479728e-01 -7.28887856e-01 1.01843573e-01
-6.25509441e-01 4.66268241e-01 -3.89323533e-01 -1.17843950e+00
-3.50577682e-01 -8.03839862e-01 1.02024329e+00 7.05114603e-01
-1.85708091e-01 -8.90591681e-01 1.59888655e-01 3.92971665e-01
3.96481037e-01 -4.52405602e-01 9.34552252e-01 -1.28709698e+00
-2.77279347e-01 -2.78219804e-02 -1.24151371e-01 -1.01670824e-01
2.16578990e-01 -9.77770567e-01 -6.96399510e-01 1.42928779e-01
3.66219580e-02 -7.45409548e-01 5.90961039e-01 7.16397464e-02
6.64789557e-01 -1.27297297e-01 -3.89555812e-01 -1.74355403e-01
1.10595644e+00 4.68736649e-01 5.52881181e-01 2.95438737e-01
4.69400994e-02 7.17051983e-01 1.49847603e+00 8.07350397e-01
4.28393215e-01 6.67958796e-01 -1.66315258e-01 1.62024349e-02
6.94563910e-02 -7.16261387e-01 -3.70734721e-01 4.01092917e-01
6.15949690e-01 -3.45039293e-02 -7.93258369e-01 9.46844161e-01
-2.00814700e+00 -8.39844763e-01 4.54868764e-01 1.97961056e+00
1.32919371e+00 4.19457644e-01 5.35664372e-02 1.88081175e-01
9.93951440e-01 2.04311162e-01 -8.25081468e-01 2.02868432e-02
-3.06202248e-02 2.47383028e-01 7.17742145e-02 2.47119978e-01
-1.10188448e+00 1.57226527e+00 5.64490938e+00 1.22530007e+00
-5.23547590e-01 3.30670267e-01 4.34006870e-01 4.05514866e-01
-5.04824221e-01 2.32676312e-01 -1.06980360e+00 5.50559402e-01
5.90205193e-01 -5.31997979e-01 9.60810483e-02 1.51841426e+00
-2.11486727e-01 -3.37967843e-01 -5.39452493e-01 7.27562010e-01
-2.06777513e-01 -1.17893207e+00 -1.58539876e-01 -1.46552131e-01
5.10001540e-01 -3.97077590e-01 -4.52407092e-01 1.09737611e+00
5.11523306e-01 -5.74679136e-01 2.13631038e-02 -1.80766180e-01
9.72999990e-01 -4.83570307e-01 7.16527402e-01 7.94715822e-01
-9.81264174e-01 7.34050721e-02 -9.12976801e-01 -1.03828676e-01
5.20170629e-01 5.89390457e-01 -1.42035115e+00 4.83420342e-01
1.72143504e-01 2.83768594e-01 1.03307433e-01 9.61571932e-01
-1.99599773e-01 3.85020465e-01 -1.07747622e-01 -1.04196347e-01
4.04882759e-01 -1.05868287e-01 2.86197931e-01 8.99330914e-01
3.72383475e-01 4.57013845e-01 6.29177928e-01 7.34950244e-01
1.01864271e-01 3.78965795e-01 -6.22550666e-01 -3.44810933e-01
1.02944100e+00 1.06385386e+00 -6.87058091e-01 -6.95790291e-01
-4.83385623e-01 1.02893090e+00 1.07167318e-01 9.52041522e-02
-3.26641440e-01 -6.15338445e-01 7.04509258e-01 -2.79548336e-02
8.83824751e-02 -4.59911935e-02 -3.62232894e-01 -1.18974423e+00
4.45220210e-02 -5.90388775e-01 4.44310009e-01 -5.32174289e-01
-1.14586198e+00 4.31792378e-01 5.72526604e-02 -1.30277491e+00
-8.33843291e-01 -1.94285527e-01 -5.95972717e-01 9.69206989e-01
-1.71219468e+00 -7.86027193e-01 -1.38359159e-01 4.00568157e-01
1.42221403e+00 -5.46132088e-01 1.08057141e+00 2.23831832e-01
-2.26366475e-01 3.20073307e-01 -1.12691205e-02 2.11926728e-01
6.26211643e-01 -1.21442473e+00 8.37243199e-01 4.43262070e-01
-2.94698961e-02 6.55068815e-01 8.50367606e-01 -9.44475770e-01
-7.41686285e-01 -7.18916714e-01 1.18699348e+00 1.86845973e-01
3.27153713e-01 -3.43676716e-01 -1.23385954e+00 2.34923661e-01
8.04456100e-02 -1.46649763e-01 9.80034590e-01 3.46787691e-01
-8.23791772e-02 5.23771882e-01 -1.32774067e+00 5.38310289e-01
1.04134274e+00 -4.98639256e-01 -1.30941534e+00 5.64853728e-01
1.15480781e+00 -7.21115112e-01 -3.52829933e-01 2.06677213e-01
-4.04238980e-03 -7.28106439e-01 7.55815089e-01 -7.18330324e-01
2.27968976e-01 -5.01834676e-02 2.11937893e-02 -1.47348487e+00
1.11832224e-01 -4.62614328e-01 -1.39058605e-01 1.27352273e+00
4.11492288e-01 -2.60671318e-01 1.11081100e+00 8.92085373e-01
-3.97544593e-01 -4.63291377e-01 -1.09944069e+00 -6.75141215e-01
-1.48361295e-01 -1.80679917e-01 7.99454570e-01 1.07948363e+00
6.29035354e-01 7.21084714e-01 -5.81457138e-01 -5.56543648e-01
2.24163830e-01 2.17299715e-01 7.56433308e-01 -1.35398352e+00
-2.42832914e-01 1.49851516e-01 1.42143056e-01 -1.21272397e+00
2.22637504e-01 -3.32944214e-01 2.34602034e-01 -1.81788158e+00
7.45420307e-02 -9.20903146e-01 -2.96902489e-02 5.44779122e-01
-5.28432190e-01 -3.79601747e-01 1.94415357e-03 2.02423707e-01
-9.26069021e-01 8.64948213e-01 1.20373118e+00 2.04714015e-02
-4.88514245e-01 1.85342729e-02 -7.38724947e-01 2.43436337e-01
6.93006217e-01 -2.95254081e-01 -8.65011573e-01 4.11476381e-03
-2.72289723e-01 6.51738346e-01 -3.90855998e-01 -7.78810203e-01
2.70341456e-01 -4.02609944e-01 -6.17595203e-02 -5.96196115e-01
6.23030782e-01 -5.59811354e-01 -3.28485638e-01 1.34342492e-01
-3.44213098e-01 -5.29696405e-01 7.90454820e-02 5.52404702e-01
-4.49354023e-01 -6.45075023e-01 4.57886577e-01 -7.47388363e-01
-1.38249910e+00 2.86898971e-01 -2.08724514e-01 5.48316061e-01
7.65674829e-01 -1.98382050e-01 -4.66312438e-01 -2.13939548e-01
-7.32874751e-01 5.78618050e-01 4.08141345e-01 6.09989941e-01
4.86096710e-01 -1.43941510e+00 -3.27386320e-01 3.44858646e-01
5.59569597e-01 3.97644877e-01 7.14586258e-01 4.82810363e-02
1.90678373e-01 7.45298028e-01 -3.24693561e-01 -2.34195963e-01
-9.28557217e-01 4.78786826e-01 -3.76169942e-02 -5.98375380e-01
-5.05767286e-01 6.53512418e-01 6.56035170e-02 -8.77435863e-01
1.66598260e-01 2.78729856e-01 -5.52132726e-01 1.33641139e-01
7.16435134e-01 -2.39697039e-01 -2.34402150e-01 -2.46781841e-01
2.11981833e-01 -1.58456832e-01 -2.84014165e-01 -3.28466117e-01
1.03967786e+00 -3.11790407e-01 3.56950402e-01 5.27361810e-01
5.02885044e-01 -5.90648413e-01 -1.18362391e+00 -1.00948906e+00
3.24416220e-01 -7.43461907e-01 -3.15506220e-01 -7.55909204e-01
-2.61583954e-01 9.50394869e-01 3.32717091e-01 3.44075292e-01
7.63100207e-01 2.57202163e-02 1.19085300e+00 4.63634014e-01
8.13021958e-01 -1.71150196e+00 9.88827571e-02 7.93424070e-01
3.34053963e-01 -1.58820260e+00 -3.03659201e-01 -6.06732070e-01
-1.08601224e+00 7.34336913e-01 1.16937256e+00 3.40719283e-01
3.34031075e-01 -1.52950376e-01 1.48468569e-01 -5.45283593e-02
-7.94491887e-01 -5.93391061e-01 -2.01624632e-01 8.23297441e-01
6.01595402e-01 -4.50169072e-02 -7.90876806e-01 9.54458892e-01
-8.39771405e-02 8.96615088e-02 2.37236872e-01 1.33017445e+00
-1.08935153e+00 -1.71364498e+00 -4.29687016e-02 4.51977044e-01
-2.08595976e-01 -3.27114850e-01 -4.25120033e-02 6.76920772e-01
2.32662689e-02 8.90695214e-01 3.75140272e-02 -1.20196581e-01
1.54497862e-01 7.35360444e-01 -5.31434827e-02 -1.34249949e+00
-1.47929778e-02 1.63774297e-01 3.25840592e-01 -1.31976843e-01
-1.62672281e-01 -4.66563493e-01 -1.45707846e+00 1.41181007e-01
-5.98833323e-01 9.28793252e-01 6.38506770e-01 1.22688532e+00
3.51166695e-01 4.64429855e-01 3.28413963e-01 -5.82220733e-01
-6.63694680e-01 -1.40226769e+00 -9.47191656e-01 4.95032549e-01
-1.59359872e-01 -1.11447775e+00 1.15824774e-01 -4.44371849e-01] | [12.610681533813477, 7.427403926849365] |
f81e5d4d-43e7-4016-a175-7c633a6d8b28 | grounded-situation-recognition-with | 2111.10135 | null | https://arxiv.org/abs/2111.10135v1 | https://arxiv.org/pdf/2111.10135v1.pdf | Grounded Situation Recognition with Transformers | Grounded Situation Recognition (GSR) is the task that not only classifies a salient action (verb), but also predicts entities (nouns) associated with semantic roles and their locations in the given image. Inspired by the remarkable success of Transformers in vision tasks, we propose a GSR model based on a Transformer encoder-decoder architecture. The attention mechanism of our model enables accurate verb classification by capturing high-level semantic feature of an image effectively, and allows the model to flexibly deal with the complicated and image-dependent relations between entities for improved noun classification and localization. Our model is the first Transformer architecture for GSR, and achieves the state of the art in every evaluation metric on the SWiG benchmark. Our code is available at https://github.com/jhcho99/gsrtr . | ['Suha Kwak', 'Hyeonjun Lee', 'Youngseok Yoon', 'Junhyeong Cho'] | 2021-11-19 | null | null | null | null | ['grounded-situation-recognition', 'situation-recognition'] | ['computer-vision', 'computer-vision'] | [ 2.39792049e-01 1.53181478e-01 -7.96421841e-02 -4.84908253e-01
-6.63324893e-01 -3.11935753e-01 6.85928106e-01 8.89525786e-02
-4.62019503e-01 3.86350125e-01 7.26215065e-01 -1.28789812e-01
1.71573967e-01 -8.56170297e-01 -8.01975131e-01 -4.17721599e-01
1.49508953e-01 4.08675969e-01 4.29697305e-01 -4.07292873e-01
8.34287107e-02 3.17034185e-01 -1.34196270e+00 7.13219941e-01
7.15096831e-01 1.01876783e+00 8.07110310e-01 2.79621631e-01
1.28017738e-01 1.43564069e+00 -3.09451610e-01 -6.73758149e-01
-9.83095616e-02 -2.01732457e-01 -9.30448353e-01 -1.30033707e-02
2.34078676e-01 -3.71037930e-01 -1.01670980e+00 1.13903832e+00
3.23681712e-01 3.75927761e-02 4.69109595e-01 -1.19683480e+00
-1.24483013e+00 8.18816781e-01 -3.98375362e-01 4.06693548e-01
4.35285211e-01 5.94895557e-02 1.43880916e+00 -1.28694010e+00
7.51698256e-01 1.29846549e+00 2.11178377e-01 4.41136092e-01
-7.88114071e-01 -4.70919669e-01 3.67877811e-01 7.41844833e-01
-1.63343525e+00 -4.75015551e-01 6.55269086e-01 -2.20746487e-01
1.26771200e+00 2.06657052e-02 9.14893925e-01 1.23917854e+00
2.93567985e-01 1.23830867e+00 8.49074423e-01 -9.03506875e-02
6.91403821e-02 -3.13826412e-01 3.33770216e-02 9.31761801e-01
6.73703402e-02 -2.21392274e-01 -8.57237160e-01 3.88146460e-01
8.21746111e-01 1.71320230e-01 -3.74775112e-01 -3.26703608e-01
-1.59028518e+00 7.86900818e-01 9.90815997e-01 2.12498590e-01
-6.34080708e-01 5.28878272e-01 3.72689158e-01 -3.81845534e-02
2.22750917e-01 4.30234462e-01 -2.99490899e-01 -1.09014489e-01
-5.59669316e-01 -2.29261033e-02 3.36131871e-01 1.12802911e+00
5.25022686e-01 -1.77875191e-01 -5.26126564e-01 7.05973327e-01
5.01309335e-01 5.66078484e-01 4.39255774e-01 -9.30558681e-01
5.90713680e-01 8.83952558e-01 -1.44478977e-01 -9.78798628e-01
-2.09349900e-01 -2.65945941e-01 -6.41237676e-01 -3.01335514e-01
1.48194924e-01 4.83951122e-01 -1.16890049e+00 1.70871711e+00
9.51453820e-02 -4.01656665e-02 1.28110245e-01 1.12921393e+00
1.31661522e+00 7.09358931e-01 3.32066387e-01 2.84848481e-01
1.79858363e+00 -1.03908730e+00 -6.68110371e-01 -6.39406383e-01
2.53474474e-01 -3.35950613e-01 1.24956620e+00 9.94222015e-02
-9.57100093e-01 -1.91822842e-01 -8.02294850e-01 -8.19863796e-01
-4.71101671e-01 2.00640276e-01 7.54796922e-01 -1.62438482e-01
-1.07496095e+00 2.05788642e-01 -8.59847367e-01 -5.51632285e-01
7.78902233e-01 1.41389638e-01 -6.01859689e-01 -2.51497805e-01
-1.42475343e+00 1.04584742e+00 3.53392869e-01 4.23245668e-01
-1.17414796e+00 -3.64398241e-01 -1.15100205e+00 1.80558428e-01
3.96618605e-01 -5.72461963e-01 1.21277678e+00 -8.31707358e-01
-9.90493953e-01 1.20648742e+00 -4.06064093e-01 -6.05711281e-01
2.11936072e-01 -8.72515216e-02 -2.64807284e-01 4.10875142e-01
3.68499935e-01 1.06601274e+00 6.51819706e-01 -8.82838845e-01
-7.21705317e-01 -5.00935972e-01 4.55687106e-01 5.01068473e-01
-7.82631040e-02 2.49471635e-01 -6.58199012e-01 -7.11109459e-01
3.11363041e-01 -6.15780830e-01 -1.16848581e-01 -4.01348099e-02
-3.79535586e-01 -4.55163598e-01 4.86242205e-01 -8.21337640e-01
9.54952598e-01 -2.32609057e+00 3.58264983e-01 -3.00086439e-01
4.91344810e-01 -9.72526446e-02 -2.87935108e-01 4.20560390e-01
-2.25930437e-01 -1.28754646e-01 -1.53053969e-01 -1.49863645e-01
1.49635211e-01 1.86407000e-01 -6.45225585e-01 4.64841217e-01
2.37304464e-01 1.46519780e+00 -1.13451898e+00 -4.88070816e-01
2.75241822e-01 5.45418620e-01 -3.69819313e-01 6.84397295e-02
-1.72667935e-01 2.44779810e-01 -5.58777153e-01 8.49432468e-01
1.09077215e-01 -5.61712742e-01 1.00228608e-01 -6.75407410e-01
6.61794171e-02 6.24982655e-01 -6.71362996e-01 1.81698120e+00
-3.53997022e-01 8.94762218e-01 -3.32226515e-01 -9.94963169e-01
8.10304523e-01 3.38172764e-01 3.16941917e-01 -1.10344398e+00
7.76414722e-02 8.86292830e-02 -1.29670188e-01 -4.32693571e-01
5.97502828e-01 1.63243800e-01 -3.99409741e-01 1.35991815e-02
3.79735351e-01 8.78334194e-02 6.20953068e-02 4.10331100e-01
1.26133978e+00 2.26152420e-01 6.59893155e-01 -1.26419276e-01
2.17688784e-01 7.20305592e-02 6.30646050e-01 5.71448326e-01
-3.45621973e-01 5.80236018e-01 4.23638612e-01 -6.32466972e-01
-7.16708064e-01 -1.19005048e+00 9.94899124e-02 1.07228613e+00
6.04174852e-01 -5.88264465e-01 -3.02991629e-01 -7.05980003e-01
-9.43551883e-02 7.97482371e-01 -7.55954087e-01 -2.86136359e-01
-5.28397858e-01 -3.66693050e-01 3.70344967e-01 8.73748899e-01
8.31453443e-01 -1.49130535e+00 -7.84219384e-01 4.32469919e-02
-7.07330823e-01 -1.42583144e+00 -4.87882435e-01 2.47202918e-01
-4.53745902e-01 -1.06791615e+00 -3.24775487e-01 -9.60225582e-01
7.28111327e-01 2.69261628e-01 1.35270655e+00 -1.44650280e-01
-2.46516496e-01 3.16631645e-01 -5.94040215e-01 -3.29908460e-01
-2.43975855e-02 -1.22846186e-01 -3.38300616e-01 5.52968448e-03
5.46302795e-01 -3.12953800e-01 -7.13442147e-01 3.23030829e-01
-6.38309360e-01 5.36620557e-01 5.14151812e-01 7.56660163e-01
8.62404168e-01 -2.24515460e-02 4.57801640e-01 -6.94905400e-01
2.53217965e-01 -2.50813544e-01 -5.87329626e-01 4.87660229e-01
-1.46010950e-01 4.86628823e-02 3.18859339e-01 -7.83661678e-02
-9.34393644e-01 2.44006112e-01 -2.32168674e-01 -4.26518589e-01
-1.13285542e-01 5.28519750e-01 -3.42255414e-01 1.96422070e-01
4.08133686e-01 4.55618054e-01 -5.86307049e-01 -1.83491066e-01
6.02604032e-01 5.82258224e-01 5.75586081e-01 -3.94754678e-01
4.56755728e-01 6.12679720e-01 -2.34433562e-01 -5.61983585e-01
-1.14622080e+00 -4.19686377e-01 -5.61442137e-01 -1.91347554e-01
1.08287024e+00 -1.42138195e+00 -5.97435713e-01 4.34691012e-01
-1.23980916e+00 -5.14983118e-01 -5.05706608e-01 3.46911550e-01
-6.96483731e-01 -7.45093673e-02 -7.05235183e-01 -3.69705975e-01
-3.42339009e-01 -1.23505926e+00 1.43692911e+00 9.50279608e-02
-6.68611899e-02 -7.10195720e-01 -4.04063225e-01 5.25199056e-01
1.12254426e-01 8.61071944e-02 5.79806209e-01 -4.21582282e-01
-1.00243628e+00 -6.08345568e-02 -5.41502774e-01 -4.99471128e-02
1.49036378e-01 -3.92115682e-01 -8.22796047e-01 -1.22473061e-01
-1.87644616e-01 -3.21515828e-01 1.22947288e+00 2.49864176e-01
1.23729432e+00 -2.11778387e-01 -4.06388998e-01 6.13366961e-01
1.34595728e+00 1.31427497e-01 9.89334822e-01 2.69801408e-01
9.39566135e-01 2.46481866e-01 8.27745438e-01 1.62466049e-01
1.12767100e+00 7.70468235e-01 7.52071857e-01 -1.66033357e-01
-3.82605314e-01 -5.80078542e-01 4.37981844e-01 3.57281089e-01
-8.87756199e-02 -3.70859772e-01 -1.04931843e+00 7.81906664e-01
-2.13852715e+00 -1.15155888e+00 2.62508839e-02 1.69607997e+00
7.46630907e-01 -1.01518452e-01 -2.07372695e-01 -3.11857760e-01
7.19368219e-01 3.92191380e-01 -5.16375661e-01 -8.52216780e-02
-1.16482303e-01 7.35838562e-02 4.54698473e-01 2.24596575e-01
-1.30618787e+00 1.45358336e+00 5.84674501e+00 6.55254483e-01
-9.72617745e-01 3.07494074e-01 3.98948908e-01 5.75115755e-02
-1.84999064e-01 -1.37200534e-01 -7.36441255e-01 2.78704554e-01
6.06790483e-01 -1.83933854e-01 5.91534138e-01 6.62106454e-01
7.35719576e-02 -7.72807673e-02 -1.10252309e+00 1.10850275e+00
2.41452485e-01 -1.45000601e+00 2.47516498e-01 -7.17869177e-02
2.89342284e-01 4.27349627e-01 1.34033067e-02 2.87763417e-01
4.41782534e-01 -9.90244210e-01 1.17703414e+00 3.98712754e-01
7.68246949e-01 -4.32955474e-01 4.69722122e-01 7.11723603e-03
-1.52119553e+00 -2.71251589e-01 -3.83675069e-01 1.28835127e-01
1.28151044e-01 1.97450519e-01 -7.49826074e-01 3.18346143e-01
8.21638823e-01 1.26845884e+00 -7.57611334e-01 8.52266192e-01
-8.42853725e-01 5.28536379e-01 -9.31668282e-03 5.45215830e-02
1.03335418e-01 3.29740681e-02 5.96517861e-01 1.08140004e+00
1.58733025e-01 2.85285622e-01 -2.41056364e-02 8.82432461e-01
-4.23090696e-01 -3.66620347e-02 -7.34524786e-01 -7.65315443e-02
6.16285503e-01 1.35377085e+00 -9.32888031e-01 -2.06526309e-01
-4.53916997e-01 1.23358893e+00 6.71221793e-01 3.65357339e-01
-1.14933181e+00 -2.63541073e-01 5.91372073e-01 1.26685917e-01
6.11205935e-01 -1.91652134e-01 -7.50916079e-02 -1.52495754e+00
1.02875076e-01 -8.70549262e-01 5.35433471e-01 -1.37693691e+00
-9.22521412e-01 6.96811318e-01 -1.80585891e-01 -1.09951246e+00
1.17465839e-01 -7.30894983e-01 -2.99337059e-01 4.40911144e-01
-1.59295547e+00 -1.62428081e+00 -3.78701925e-01 8.20084333e-01
6.13009214e-01 -5.99096566e-02 7.71458745e-01 3.28132838e-01
-4.72704440e-01 3.34130555e-01 -4.95830357e-01 5.66191256e-01
4.35748011e-01 -1.32887626e+00 6.37822449e-01 1.04850364e+00
5.18800914e-01 5.88329852e-01 4.03943211e-01 -6.21514976e-01
-1.42132366e+00 -1.39214528e+00 1.20470524e+00 -7.34930098e-01
7.95560658e-01 -7.14592755e-01 -6.03577673e-01 1.10911107e+00
1.79625690e-01 2.42948398e-01 2.43259773e-01 -9.83091351e-03
-6.63779497e-01 -1.09202132e-01 -9.29568529e-01 7.96423972e-01
1.53541458e+00 -9.09620941e-01 -7.80476868e-01 4.71973181e-01
9.33327615e-01 -7.01660335e-01 -7.32720673e-01 2.19276428e-01
3.18576574e-01 -5.62511802e-01 1.16416788e+00 -4.33088034e-01
6.24972761e-01 -4.84578937e-01 -6.16923869e-01 -1.10841465e+00
-6.17117405e-01 -1.26670763e-01 -2.24513829e-01 9.84920144e-01
4.47385848e-01 -5.01267314e-01 4.25307900e-01 4.01788563e-01
-2.37086162e-01 -7.31600046e-01 -1.02720535e+00 -3.96545053e-01
-4.17590767e-01 -2.92336524e-01 6.28935397e-01 8.88761640e-01
-6.99855611e-02 5.60125351e-01 -1.21994674e-01 4.70949173e-01
5.41774213e-01 3.99705946e-01 7.73815736e-02 -8.52714896e-01
-2.37032548e-02 -1.51594520e-01 -7.96348393e-01 -1.01673186e+00
3.19192559e-01 -1.18498969e+00 1.76533878e-01 -2.22161674e+00
4.89676625e-01 -1.08625904e-01 -3.21734548e-01 1.14141870e+00
-1.14855893e-01 4.78116393e-01 3.70850235e-01 1.25466794e-01
-1.12820852e+00 7.31811941e-01 1.30647135e+00 -3.10119331e-01
2.42777869e-01 -4.58007991e-01 -9.87091601e-01 7.40217090e-01
6.41495228e-01 -3.45010132e-01 -3.77523512e-01 -9.04086530e-01
3.08643937e-01 -3.43281366e-02 8.96494389e-01 -7.71177948e-01
2.70258576e-01 -8.59168395e-02 4.34128344e-01 -5.87633967e-01
4.03822124e-01 -9.03960645e-01 -4.45671044e-02 1.43390998e-01
-4.54071045e-01 2.09071845e-01 -6.39792606e-02 4.50546980e-01
-3.39937389e-01 1.51603743e-01 6.33651018e-01 -1.76609725e-01
-1.44346166e+00 5.19877136e-01 -1.58468306e-01 1.76204801e-01
8.93063307e-01 -1.47875892e-02 -5.50674975e-01 -2.98229337e-01
-8.45683694e-01 3.45582217e-01 4.31613654e-01 6.91856027e-01
1.07467413e+00 -1.39720368e+00 -6.19975865e-01 -2.78340769e-03
4.75993484e-01 5.26194684e-02 3.91500033e-02 6.76175833e-01
-3.46235633e-01 3.71344507e-01 -1.40741721e-01 -4.03248727e-01
-1.10904360e+00 4.95763958e-01 7.00086951e-01 -1.30705103e-01
-8.47339809e-01 9.52769160e-01 6.32127881e-01 -1.77846640e-01
1.46346092e-01 -4.33780134e-01 -3.23176086e-01 -2.26707608e-02
5.33339202e-01 -1.67349339e-01 7.69901946e-02 -1.01872766e+00
-6.88244045e-01 3.93375337e-01 -5.65350801e-02 5.93183003e-03
1.39858353e+00 -1.27255753e-01 -2.74743170e-01 2.12126955e-01
9.64250267e-01 -3.96608710e-01 -1.26560104e+00 -3.87035996e-01
-1.64922595e-01 -3.53458434e-01 2.63183326e-01 -9.49131191e-01
-1.22176921e+00 8.70461762e-01 2.63281286e-01 -1.48322612e-01
1.27169323e+00 5.09232819e-01 6.15669787e-01 4.58797097e-01
7.58705318e-01 -8.69444609e-01 2.78301090e-01 6.58444643e-01
1.36802709e+00 -1.25672817e+00 -2.29654521e-01 -4.53496397e-01
-1.08634162e+00 7.31631875e-01 6.74928367e-01 -1.14487685e-01
4.34428662e-01 1.54679328e-01 -1.10652834e-01 -5.93484163e-01
-7.92732418e-01 -6.38490319e-01 3.00241351e-01 6.12063885e-01
4.42119062e-01 1.23580821e-01 -5.60644502e-03 6.28308177e-01
5.43881953e-02 5.18526696e-02 2.72085339e-01 7.15396702e-01
-3.07056040e-01 -7.23479867e-01 1.99314803e-01 3.46069425e-01
-2.70709217e-01 -4.49029058e-01 -4.13233042e-01 5.12009025e-01
6.61002025e-02 8.81995797e-01 1.48028344e-01 -2.90577173e-01
5.96496999e-01 -1.29677981e-01 5.99439442e-01 -7.79009700e-01
-4.36009556e-01 -2.42859483e-01 1.58355579e-01 -9.85801995e-01
-4.47222203e-01 -7.77567625e-01 -1.64798880e+00 4.00074422e-02
-2.67498661e-02 -1.24969959e-01 2.69929528e-01 8.05526555e-01
5.01942754e-01 7.14288414e-01 3.81169468e-01 -4.78810072e-01
-1.64940193e-01 -6.52473450e-01 -4.48120564e-01 4.48011756e-01
2.37411931e-01 -7.60793865e-01 1.54946297e-01 1.87150285e-01] | [10.375890731811523, 1.4444804191589355] |
378617a0-c1e9-4422-80c2-0bf8c94efd7d | stc-speaker-recognition-systems-for-the | 1904.06093 | null | http://arxiv.org/abs/1904.06093v1 | http://arxiv.org/pdf/1904.06093v1.pdf | STC Speaker Recognition Systems for the VOiCES From a Distance Challenge | This paper presents the Speech Technology Center (STC) speaker recognition
(SR) systems submitted to the VOiCES From a Distance challenge 2019. The
challenge's SR task is focused on the problem of speaker recognition in single
channel distant/far-field audio under noisy conditions. In this work we
investigate different deep neural networks architectures for speaker embedding
extraction to solve the task. We show that deep networks with residual frame
level connections outperform more shallow architectures. Simple energy based
speech activity detector (SAD) and automatic speech recognition (ASR) based SAD
are investigated in this work. We also address the problem of data preparation
for robust embedding extractors training. The reverberation for the data
augmentation was performed using automatic room impulse response generator. In
our systems we used discriminatively trained cosine similarity metric learning
model as embedding backend. Scores normalization procedure was applied for each
individual subsystem we used. Our final submitted systems were based on the
fusion of different subsystems. The results obtained on the VOiCES development
and evaluation sets demonstrate effectiveness and robustness of the proposed
systems when dealing with distant/far-field audio under noisy conditions. | ['Galina Lavrentyeva', 'Vladimir Volokhov', 'Timur Pekhovsky', 'Sergey Novoselov', 'Artem Ivanov', 'Andrey Shulipa', 'Alexandr Kozlov', 'Aleksei Gusev'] | 2019-04-12 | null | null | null | null | ['room-impulse-response'] | ['audio'] | [ 2.01723203e-01 -1.74467206e-01 7.70757079e-01 -5.61416507e-01
-1.10630226e+00 -3.06908935e-01 6.52270079e-01 -2.27367714e-01
-6.55476809e-01 3.33269745e-01 7.40205228e-01 -1.16010882e-01
2.47110482e-02 -1.93254501e-02 -3.57524842e-01 -8.94763708e-01
-2.56333470e-01 -2.17401296e-01 -1.86319813e-01 -5.06758869e-01
1.12844467e-01 9.18118238e-01 -1.57470095e+00 4.15368825e-01
1.51320249e-01 8.56444597e-01 3.72980207e-01 1.52996659e+00
3.71748716e-01 6.21763945e-01 -1.08334863e+00 2.43521228e-01
1.58681065e-01 -3.90690953e-01 -6.31266475e-01 -3.17724615e-01
3.65070075e-01 -1.09351978e-01 -4.74741757e-01 7.80338585e-01
1.41397369e+00 6.07896090e-01 6.32920146e-01 -9.95787799e-01
-3.35591167e-01 8.16107810e-01 1.47343203e-01 8.69310141e-01
3.74449313e-01 5.17333932e-02 7.30305731e-01 -1.36770606e+00
1.87833473e-01 1.33098710e+00 6.52958632e-01 8.44590902e-01
-1.02203870e+00 -4.79235202e-01 -1.15557522e-01 6.53328717e-01
-1.50601530e+00 -1.50237131e+00 9.99364793e-01 -2.15012059e-01
1.81432557e+00 5.88667333e-01 -5.01289479e-02 1.44933236e+00
-1.73926935e-01 4.63189870e-01 7.53039777e-01 -7.30555475e-01
3.37955028e-01 3.57165664e-01 3.31924021e-01 2.22430855e-01
-6.00207865e-01 2.73050666e-01 -7.02698946e-01 7.57953301e-02
1.88574255e-01 -6.14842713e-01 -4.12637800e-01 5.58661997e-01
-1.23509228e+00 6.91429317e-01 2.46331573e-01 1.01546144e+00
-3.88572961e-01 1.60745695e-01 6.85763538e-01 6.87783241e-01
4.43391502e-01 1.87054604e-01 -2.84632742e-01 -1.18197277e-01
-1.20939338e+00 -1.74647465e-01 7.87808418e-01 6.72990263e-01
-1.67090714e-01 9.46775258e-01 -3.14399719e-01 1.26352644e+00
5.72733223e-01 3.96694034e-01 9.24331546e-01 -6.25412166e-01
4.73969758e-01 -3.41809928e-01 -1.48317084e-01 -6.10113442e-01
-4.30965662e-01 -7.27290392e-01 -8.80290806e-01 2.12115273e-01
-1.06031694e-01 -2.83400804e-01 -8.67155075e-01 1.67147744e+00
4.18419167e-02 2.20053062e-01 6.06183589e-01 8.70742381e-01
1.03911197e+00 1.04874790e+00 -3.36698741e-01 -1.85100883e-01
1.20009744e+00 -9.70461249e-01 -1.20380092e+00 1.34945482e-01
3.95330787e-01 -1.03606892e+00 7.75536597e-01 5.71080327e-01
-1.06073296e+00 -1.01274908e+00 -1.50102794e+00 5.64707704e-02
-5.40625930e-01 4.77689236e-01 -4.54389542e-01 1.13892066e+00
-1.61738133e+00 4.02459770e-01 -6.39689982e-01 -2.40755543e-01
-1.14365466e-01 5.28820753e-01 -3.49278182e-01 4.19661582e-01
-1.30504167e+00 9.84944642e-01 -7.04298690e-02 4.28936452e-01
-1.38835573e+00 -2.90829122e-01 -8.01682472e-01 1.74507171e-01
-3.87968034e-01 -6.11283667e-02 1.41829491e+00 -9.59306300e-01
-1.99764109e+00 5.88620424e-01 -1.79243177e-01 -8.58192325e-01
4.51330468e-02 -7.95146525e-02 -1.18527162e+00 -6.12412952e-02
-5.35356760e-01 4.36628044e-01 1.06499517e+00 -8.11569631e-01
-1.53930739e-01 -1.98891431e-01 -6.22312367e-01 1.67140141e-01
-3.49228531e-01 6.51766539e-01 4.32455450e-01 -8.43939543e-01
2.24116594e-02 -5.78798771e-01 2.98212677e-01 -6.93732321e-01
-3.60041022e-01 -2.58590996e-01 7.46550262e-01 -1.21211791e+00
1.11630380e+00 -2.48940086e+00 1.78503633e-01 1.79351456e-02
-3.56037319e-01 3.69230866e-01 -3.53594452e-01 3.93930465e-01
-5.80439448e-01 -1.60244122e-01 -4.94326130e-02 -6.92284882e-01
2.44747311e-01 -3.21895242e-01 -3.73740464e-01 6.63706303e-01
3.50043774e-01 4.18180414e-02 -6.84925914e-02 -2.19563648e-01
2.45030269e-01 1.09250236e+00 -2.95166314e-01 4.05133635e-01
4.73926544e-01 1.72507346e-01 2.46251971e-01 3.01712334e-01
6.61437988e-01 7.76219249e-01 -5.33738434e-01 -2.00820059e-01
-1.94394723e-01 7.05168188e-01 -1.37269711e+00 1.75565135e+00
-8.24290097e-01 1.21368480e+00 7.74737656e-01 -9.81952131e-01
1.17307365e+00 1.15210044e+00 2.41894629e-02 -3.72258991e-01
2.14345336e-01 3.32631767e-01 4.15279359e-01 -5.22289634e-01
4.33014363e-01 -1.95844293e-01 3.76194298e-01 2.75700748e-01
5.96891582e-01 -8.10991749e-02 -4.74739432e-01 -1.13019399e-01
1.07404828e+00 -4.44583207e-01 -1.95564628e-02 -3.44351470e-01
1.23923969e+00 -6.64140284e-01 1.99405983e-01 3.03578734e-01
-5.36935508e-01 7.74832666e-01 -2.21863598e-01 -8.36213306e-03
-1.00734031e+00 -9.11015511e-01 -1.87340155e-01 1.19748342e+00
-7.35579908e-01 -1.20565861e-01 -1.08266687e+00 -8.65980983e-02
-7.03937590e-01 9.61192727e-01 -2.70499438e-01 -9.21038687e-02
-7.93830693e-01 -5.83346367e-01 1.19488561e+00 5.15918493e-01
1.96937516e-01 -1.16502774e+00 -1.69045515e-02 4.57764149e-01
-1.22990139e-01 -1.08091629e+00 -6.71118259e-01 6.10598564e-01
-4.92468417e-01 -3.30656260e-01 -9.32984471e-01 -1.06011784e+00
8.42011869e-02 1.12163164e-01 5.66920280e-01 -5.52825570e-01
-3.90502095e-01 5.34369707e-01 -1.82716027e-01 -5.59364617e-01
-8.37178707e-01 2.38355063e-02 5.61583936e-01 2.74552315e-01
3.87607187e-01 -6.07607305e-01 -4.88251358e-01 3.93426418e-01
-6.29470706e-01 -6.24586821e-01 2.70756394e-01 6.50802910e-01
-1.93104930e-02 5.53082787e-02 1.05913174e+00 3.45725864e-01
9.22102511e-01 -2.24687383e-01 -3.89048189e-01 -6.88866526e-02
-5.83137721e-02 2.33071387e-01 5.24109483e-01 -4.55202729e-01
-1.10340190e+00 1.38348073e-01 -8.03002000e-01 -1.93555459e-01
-4.20300663e-01 -1.12906158e-01 -5.29853225e-01 1.82650328e-01
1.02335870e+00 4.07761872e-01 -1.13071576e-01 -7.38376796e-01
4.47135009e-02 1.50713360e+00 4.55321223e-01 -1.44175455e-01
4.74400550e-01 -3.33533157e-03 -4.88590777e-01 -1.50884676e+00
1.00375514e-03 -5.93979836e-01 -5.10461390e-01 -2.65057802e-01
1.11515927e+00 -1.22050548e+00 -5.13442695e-01 5.35382330e-01
-1.47891128e+00 -6.82320073e-02 -1.98207781e-01 9.85768080e-01
-3.41917634e-01 8.69486928e-02 -5.95106184e-01 -1.32693338e+00
-7.89111912e-01 -1.25582957e+00 9.14252162e-01 -1.36133850e-01
-9.41293463e-02 -7.78700531e-01 5.14667690e-01 3.95647347e-01
9.90285695e-01 -3.04589838e-01 3.53027403e-01 -1.17960036e+00
-1.30681232e-01 -2.66775399e-01 5.49919844e-01 1.26198649e+00
2.67709512e-02 -5.69098927e-02 -2.06458235e+00 -2.86955655e-01
6.32268965e-01 3.80258337e-02 8.26787949e-01 4.13576961e-01
8.90234470e-01 -1.67857096e-01 3.66306126e-01 3.61750931e-01
9.90339994e-01 3.27879727e-01 6.52062476e-01 7.87937120e-02
3.37063134e-01 7.61400402e-01 -1.92470057e-03 1.65698513e-01
-3.14696163e-01 8.04061174e-01 4.23409790e-02 1.75616622e-01
-5.09671390e-01 3.47782046e-01 1.15966964e+00 1.49985182e+00
3.23668048e-02 -4.99496728e-01 -7.16639221e-01 5.30147254e-01
-1.08628905e+00 -1.24302435e+00 -8.08757097e-02 2.05141926e+00
5.63004494e-01 2.13753767e-02 1.17087975e-01 8.75505507e-01
8.55392575e-01 1.44317403e-01 -6.75540343e-02 -1.01501381e+00
-3.42095852e-01 4.45884466e-01 6.87641576e-02 9.40598905e-01
-9.80701804e-01 2.62941390e-01 5.69262123e+00 7.75029302e-01
-1.28185403e+00 5.49814820e-01 2.90973663e-01 -4.98066455e-01
1.90653846e-01 -8.17939222e-01 -8.30454528e-01 1.98638141e-01
1.98218572e+00 2.23405361e-01 5.01630068e-01 7.68327713e-01
5.14386117e-01 4.17869359e-01 -1.24912226e+00 1.34488058e+00
4.15263683e-01 -7.99182773e-01 -5.56325078e-01 -2.90177595e-02
3.68242621e-01 4.93877381e-01 1.74058795e-01 3.49101186e-01
-2.96684474e-01 -1.31343949e+00 8.40017319e-01 4.73479390e-01
5.24582446e-01 -7.98844874e-01 8.14435542e-01 1.70977265e-01
-1.22220814e+00 -1.54796273e-01 -2.24046856e-01 3.06244850e-01
3.54751460e-02 3.94893259e-01 -1.34189498e+00 2.01633796e-01
5.24294555e-01 2.02000991e-01 -2.90457249e-01 7.38771439e-01
2.40099415e-01 9.00776267e-01 -4.28441018e-01 -9.96415839e-02
1.06322415e-01 4.04486001e-01 9.84467328e-01 1.78711057e+00
6.81685925e-01 -3.10416281e-01 -6.87353611e-01 6.09609365e-01
-1.52410433e-01 3.01978905e-02 -6.07535958e-01 1.45087600e-01
3.37091416e-01 1.19677317e+00 -6.88849688e-02 -2.58039385e-01
-3.61900777e-02 9.37205255e-01 -2.92446971e-01 4.00578052e-01
-7.37868547e-01 -8.05086136e-01 6.89173877e-01 -1.33433878e-01
2.88873971e-01 -2.90034443e-01 -1.82715766e-02 -8.67351234e-01
1.12192601e-01 -7.85558701e-01 -7.47412592e-02 -7.90972471e-01
-9.99219954e-01 1.26364458e+00 -2.07249939e-01 -1.10391319e+00
-3.58583659e-01 -7.41006374e-01 -9.65896726e-01 1.33551097e+00
-1.35169554e+00 -5.55630445e-01 -7.27102086e-02 8.07578921e-01
1.02277875e+00 -8.79214466e-01 1.03344810e+00 6.31766796e-01
-6.66128695e-01 7.51374245e-01 3.62803251e-01 -9.16138943e-03
6.74109459e-01 -1.28494513e+00 5.84617138e-01 1.00787163e+00
2.85233915e-01 4.05684054e-01 7.77569234e-01 1.25582799e-01
-1.25827587e+00 -9.36088860e-01 1.08182371e+00 -2.52587646e-01
3.49304110e-01 -7.75749266e-01 -9.12962079e-01 3.14057916e-01
8.10752869e-01 -1.28302366e-01 8.35063994e-01 -3.52665693e-01
-2.72915930e-01 -4.13604766e-01 -1.22373188e+00 -8.27466510e-03
5.07872641e-01 -9.91553247e-01 -8.33189309e-01 3.37463647e-01
7.56938636e-01 1.52753785e-01 -6.36789024e-01 -7.81166553e-02
7.87591860e-02 -6.63624823e-01 1.12231004e+00 -3.76924306e-01
-2.41958529e-01 -4.80312526e-01 -8.38295817e-01 -1.43110013e+00
-2.79817265e-02 -9.21481133e-01 -4.91865352e-02 1.47419238e+00
5.97542167e-01 -4.02146667e-01 1.33071974e-01 -1.05141178e-02
-5.33931434e-01 8.38379636e-02 -1.46774232e+00 -9.80026543e-01
-9.66921374e-02 -7.88147569e-01 3.77763957e-01 5.75822413e-01
-1.81968912e-01 7.25228310e-01 -2.86911041e-01 4.97032791e-01
4.00009185e-01 -9.16091084e-01 2.90642947e-01 -9.39082742e-01
-3.55092913e-01 -2.81870872e-01 -6.49864793e-01 -4.03835237e-01
1.20608762e-01 -9.25341904e-01 3.56951982e-01 -1.19724333e+00
-6.06876969e-01 2.64135420e-01 -5.34901977e-01 -1.85831845e-01
3.70745182e-01 -9.64098647e-02 1.91203319e-02 -3.70641083e-01
-2.20821351e-02 6.96754932e-01 3.52877021e-01 -3.62782449e-01
-2.85209447e-01 3.05358946e-01 -1.34942248e-01 3.58816415e-01
1.04136086e+00 -4.26399231e-01 -3.95783663e-01 -3.53350639e-01
-2.99381256e-01 3.12580764e-02 5.01059532e-01 -1.46572733e+00
3.52259785e-01 6.74906135e-01 2.69327700e-01 -6.37314677e-01
8.78460288e-01 -8.31149399e-01 1.71636179e-01 3.91917318e-01
-7.13801086e-01 1.67936962e-02 3.82837176e-01 2.58439809e-01
-4.59722608e-01 -3.76490355e-01 1.05756295e+00 3.07456553e-01
-2.31441632e-01 -4.02972817e-01 -7.66005218e-01 -4.00445104e-01
5.88472545e-01 -7.28734285e-02 -4.53702500e-03 -4.41652805e-01
-1.17361963e+00 -4.84361738e-01 -4.05429780e-01 4.21774030e-01
8.43865275e-01 -1.39212120e+00 -1.25114548e+00 4.35158074e-01
-9.62086692e-02 -5.21738768e-01 4.74126935e-01 7.54335284e-01
-2.36455515e-01 6.42249882e-01 -2.71866005e-02 -4.72298145e-01
-1.73036182e+00 3.77572626e-01 9.12948728e-01 2.69866228e-01
-2.52791554e-01 1.19635856e+00 -1.20432012e-01 -3.39990079e-01
9.00229812e-01 -6.12813175e-01 -4.13623124e-01 7.00641945e-02
9.53956783e-01 7.48279035e-01 7.20504344e-01 -9.12280381e-01
-6.32189631e-01 1.42058983e-01 -7.36377109e-03 -8.63518953e-01
1.70814991e+00 -1.35165483e-01 2.37582564e-01 6.38663054e-01
1.69512296e+00 3.79048125e-03 -6.23251379e-01 -1.76672921e-01
-1.35474920e-01 1.97208881e-01 6.78499937e-01 -9.38738465e-01
-7.46856272e-01 1.51720536e+00 1.44326615e+00 2.36859739e-01
1.16680968e+00 -4.43970442e-01 5.04999638e-01 6.21004760e-01
-3.25906128e-01 -1.39859343e+00 -4.46056128e-02 5.51934719e-01
1.56586766e+00 -1.03223979e+00 -5.54554760e-01 3.31365407e-01
-5.27909577e-01 1.38281691e+00 1.65526241e-01 1.92162152e-02
8.24587464e-01 4.01248634e-01 9.55527350e-02 1.39356911e-01
-8.27800035e-01 9.22964960e-02 4.05388862e-01 7.28284001e-01
7.30353832e-01 -5.20247519e-02 2.81102598e-01 6.79484069e-01
-4.34393764e-01 -5.90085387e-01 4.64954197e-01 7.31685340e-01
-6.72953486e-01 -7.64523506e-01 -9.63511288e-01 -1.67275026e-01
-6.45414829e-01 -2.94900090e-01 -5.72462797e-01 1.47406399e-01
-1.48156121e-01 1.41718352e+00 -1.40558600e-01 -4.84998286e-01
4.99074757e-01 6.13737404e-01 2.26056576e-01 -4.25988615e-01
-1.17700624e+00 3.70910197e-01 2.22189873e-01 -2.16995239e-01
-3.10033411e-01 -7.24332988e-01 -1.09825790e+00 6.86143264e-02
-3.48192543e-01 2.35028550e-01 1.64653015e+00 4.99452472e-01
3.43238026e-01 1.08336127e+00 1.02970338e+00 -1.02275956e+00
-8.35587323e-01 -1.70196211e+00 -6.42953634e-01 -1.88820511e-01
1.16083837e+00 -9.49171707e-02 -8.08175564e-01 2.88525492e-01] | [14.49189281463623, 6.0145111083984375] |
6a807c07-d32f-4e39-9a5c-54de1644f789 | tensorkrowch-smooth-integration-of-tensor | 2306.08595 | null | https://arxiv.org/abs/2306.08595v1 | https://arxiv.org/pdf/2306.08595v1.pdf | TensorKrowch: Smooth integration of tensor networks in machine learning | Tensor networks are factorizations of high-dimensional tensors into networks of smaller tensors. They have applications in physics and mathematics, and recently have been proposed as promising machine learning architectures. To ease the integration of tensor networks in machine learning pipelines, we introduce TensorKrowch, an open source Python library built on top of PyTorch. Providing a user-friendly interface, TensorKrowch allows users to construct any tensor network, train it, and integrate it as a layer in more intricate deep learning models. In this paper, we describe the main functionality and basic usage of TensorKrowch, and provide technical details on its building blocks and the optimizations performed to achieve efficient operation. | ['Alejandro Pozas-Kerstjens', 'David Pérez-García', 'José Ramón Pareja Monturiol'] | 2023-06-14 | null | null | null | null | ['tensor-networks'] | ['methodology'] | [-8.09014678e-01 -2.41787925e-01 -2.08961815e-01 -4.46215034e-01
6.91397488e-02 -7.56361067e-01 3.00947279e-01 6.98731616e-02
-1.77525043e-01 2.55602360e-01 3.47260356e-01 -8.27613533e-01
-1.41877219e-01 -7.84408331e-01 -3.92994702e-01 -6.01553738e-01
-5.68477273e-01 3.49742323e-01 2.78978497e-01 -1.11799173e-01
-4.20886651e-02 1.02187526e+00 -1.18720233e+00 6.58448219e-01
3.40764791e-01 7.93500960e-01 -2.83670634e-01 1.01167285e+00
-1.78339809e-01 1.10295177e+00 -2.29675755e-01 -7.52229929e-01
2.48201355e-01 3.33463043e-01 -1.09110785e+00 -3.83351743e-01
4.96567905e-01 -5.48049271e-01 -5.75771451e-01 7.30596125e-01
3.25784683e-01 8.45372230e-02 1.79183185e-01 -1.25789917e+00
-4.33082372e-01 9.01175380e-01 -1.48257777e-01 4.35924262e-01
5.52731752e-02 2.19461381e-01 1.31806064e+00 -1.01326299e+00
2.99778402e-01 1.28486574e+00 7.36181200e-01 2.93613523e-01
-1.36143231e+00 -5.59390128e-01 -1.90285832e-01 2.36740381e-01
-1.02826107e+00 -3.41604829e-01 7.44673908e-01 -7.45216727e-01
1.20251358e+00 4.79301780e-01 7.72103488e-01 6.38609648e-01
1.85636014e-01 1.00399399e+00 6.23121560e-01 5.11949547e-02
-4.95981798e-02 -2.56924480e-01 5.84742427e-01 8.92381310e-01
8.35621953e-02 -2.60112137e-01 -2.60856926e-01 -4.03902799e-01
1.07952464e+00 1.34465694e-01 1.95286304e-01 -3.52377921e-01
-1.71967936e+00 7.04356968e-01 8.89830649e-01 5.21317422e-01
-4.41957980e-01 5.18798411e-01 6.52431309e-01 1.44872278e-01
4.63461906e-01 5.55011511e-01 -7.55495012e-01 -2.48699427e-01
-5.80502212e-01 6.57724977e-01 7.97599912e-01 6.42985225e-01
7.62493074e-01 2.36826330e-01 -1.21587574e-01 8.66041064e-01
4.26642410e-02 -6.53082831e-03 2.13344350e-01 -1.00127566e+00
2.67420948e-01 7.83076942e-01 -1.87141120e-01 -1.03900647e+00
-6.86201394e-01 -3.94179970e-01 -9.77607429e-01 -7.80122727e-02
2.33245254e-01 -1.44807711e-01 -6.87005103e-01 8.36009383e-01
5.31731486e-01 2.13984251e-01 -2.62714624e-01 1.04908681e+00
1.11717498e+00 3.94891769e-01 -1.17416397e-01 5.92785537e-01
1.52787840e+00 -1.05248940e+00 -3.87461334e-01 1.87432244e-01
9.70776260e-01 -9.38332498e-01 9.17631030e-01 4.08730209e-01
-8.94392729e-01 -4.12557065e-01 -6.82565272e-01 -6.03242397e-01
-6.06650651e-01 3.24486166e-01 1.51084805e+00 3.69917303e-01
-1.24095333e+00 9.91650760e-01 -1.34803748e+00 -4.65341173e-02
3.09768617e-01 3.94006670e-01 -5.58611274e-01 5.87038435e-02
-8.90387893e-01 6.80109143e-01 6.78419471e-01 3.40458840e-01
-8.70823145e-01 -1.05265236e+00 -8.56696486e-01 1.29493296e-01
9.57872495e-02 -8.06197226e-01 1.47374928e+00 -1.13208983e-02
-1.48354185e+00 5.67616403e-01 -2.47343518e-02 -3.49362552e-01
3.60714272e-02 -2.52571970e-01 -3.00819844e-01 4.76483343e-04
-2.33630329e-01 3.37902784e-01 5.56429029e-01 -5.38732946e-01
-2.17232883e-01 -1.95248738e-01 5.41151106e-01 -2.88114399e-01
-3.61491084e-01 3.42117786e-01 -5.05693913e-01 -6.29657745e-01
1.94586381e-01 -8.78099561e-01 -5.04734874e-01 -1.20991945e-01
-7.33812273e-01 -4.47455317e-01 8.45284402e-01 -6.19689107e-01
1.13354540e+00 -2.16859174e+00 2.59874612e-01 2.16630578e-01
1.02012420e+00 3.66538674e-01 -1.98437706e-01 4.47802544e-01
-5.70842683e-01 1.58584312e-01 2.39122108e-01 -3.35565716e-01
-1.12858545e-02 4.19953197e-01 -3.37454714e-02 3.72208029e-01
2.68727064e-01 9.48965847e-01 -1.09943628e+00 -2.16520950e-01
4.08772558e-01 4.95884359e-01 -7.17955351e-01 2.38619372e-01
-1.70965746e-01 3.70222986e-01 -4.73425627e-01 6.76675737e-01
7.53620267e-01 -3.78628641e-01 1.46290630e-01 -6.88781023e-01
-4.12527770e-01 6.71241939e-01 -1.08250141e+00 1.51176763e+00
-4.06269848e-01 5.61459780e-01 2.28565082e-01 -9.65574682e-01
5.00860274e-01 2.86590070e-01 6.27033770e-01 -1.00190872e-02
1.24817111e-01 7.89274424e-02 9.36631411e-02 -4.88316208e-01
7.41359353e-01 1.57781187e-02 2.28589237e-01 6.22249007e-01
4.45877582e-01 -3.29203397e-01 6.22561991e-01 5.16282320e-01
1.17094016e+00 -8.21408853e-02 3.23375091e-02 -2.00860023e-01
5.25108337e-01 -1.99087560e-02 3.66479129e-01 3.10588032e-01
2.75204211e-01 1.57454893e-01 8.22339952e-01 -1.14862728e+00
-1.34514296e+00 -9.18497264e-01 -1.42992884e-01 1.39764810e+00
-6.74536288e-01 -1.01122391e+00 -2.65931368e-01 -6.05491340e-01
2.14535326e-01 2.86154181e-01 -4.72787410e-01 3.46485168e-01
-5.09726763e-01 -8.39033067e-01 7.01155424e-01 5.97423911e-01
3.25037211e-01 -7.11564839e-01 1.65086374e-01 -1.11016236e-01
4.16664146e-02 -1.28063035e+00 -1.23900987e-01 1.66846868e-02
-1.20040977e+00 -1.03884876e+00 -2.88387924e-01 -4.13247764e-01
5.58134437e-01 4.54518855e-01 1.31313396e+00 2.79492289e-01
-2.73343831e-01 2.26034954e-01 -1.41817898e-01 -1.65376633e-01
-2.14163274e-01 4.29069966e-01 2.32593894e-01 -1.68252021e-01
2.37042278e-01 -6.80041730e-01 -4.62467045e-01 1.05792074e-03
-1.11932564e+00 3.21285397e-01 3.07032019e-01 6.06361091e-01
9.80942026e-02 2.16654301e-01 -4.08607423e-01 -6.57058418e-01
8.59156191e-01 -3.37309092e-01 -8.68480027e-01 -9.01303664e-02
-1.58966556e-01 3.95441025e-01 5.62026024e-01 -5.58596477e-02
-5.20195901e-01 -4.65315301e-03 -4.17307228e-01 -5.24150789e-01
1.11436181e-01 7.87268043e-01 1.78408071e-01 -3.36913377e-01
4.54430699e-01 -1.92690611e-01 -2.79723424e-02 -1.08823276e+00
6.94798529e-01 3.95742267e-01 2.63679594e-01 -9.49417233e-01
8.40871692e-01 2.82277942e-01 1.94621965e-01 -7.41359830e-01
-9.98254538e-01 -6.24251246e-01 -1.07516706e+00 -2.06544638e-01
8.21816087e-01 -8.66086841e-01 -1.14168584e+00 5.68596780e-01
-1.38007569e+00 -1.87776446e-01 1.07000373e-01 6.10673845e-01
-7.39925578e-02 1.72057599e-01 -1.01707983e+00 -2.58718789e-01
-3.48449051e-01 -1.25828254e+00 8.66294265e-01 9.67252627e-02
1.65359974e-02 -1.36020291e+00 1.56498849e-01 3.90566826e-01
4.52990860e-01 7.55484849e-02 9.50435579e-01 -6.31658137e-01
-7.78901041e-01 -3.18448335e-01 -5.65420449e-01 7.08471119e-01
-8.51296261e-02 4.41063643e-01 -9.90602076e-01 -1.17116913e-01
-5.59337735e-01 -1.75237149e-01 6.45569146e-01 -3.94185670e-02
1.54650354e+00 -4.62223142e-01 -1.98419556e-01 8.79637182e-01
1.15103495e+00 -5.06736159e-01 4.52243388e-01 2.03553498e-01
1.28798640e+00 3.08591753e-01 -1.95602044e-01 3.34845901e-01
6.38672292e-01 3.93042326e-01 5.37503779e-01 -3.48040491e-01
1.56169102e-01 3.80308069e-02 1.82987839e-01 1.46819222e+00
-5.08027852e-01 6.15708053e-01 -1.10835385e+00 3.82709682e-01
-1.63840830e+00 -7.39055574e-01 -6.25427127e-01 1.75427699e+00
8.46838236e-01 -7.81161338e-02 2.20070928e-01 8.08057040e-02
2.19089568e-01 2.91517258e-01 -2.33496696e-01 -6.98194027e-01
7.94050694e-02 4.58469689e-01 5.49991906e-01 5.69636047e-01
-1.25579309e+00 1.15286386e+00 7.66310310e+00 2.71738499e-01
-1.34936583e+00 1.43881798e-01 1.59937143e-01 -5.85984252e-02
-2.36098707e-01 1.16007097e-01 -6.18991256e-01 -1.80442017e-02
9.11008358e-01 2.45682560e-02 6.28767550e-01 1.07758987e+00
3.63414735e-01 2.52586335e-01 -1.32446301e+00 7.17979372e-01
-6.70024097e-01 -1.73329198e+00 7.19596669e-02 -5.77258505e-02
3.91008168e-01 6.70315146e-01 -1.88508227e-01 2.92648762e-01
8.09677422e-01 -7.49547005e-01 1.86929286e-01 4.68518525e-01
3.93913150e-01 -4.25411373e-01 7.11299181e-01 1.82222854e-03
-1.13315940e+00 8.32407251e-02 -4.37834829e-01 -4.93573517e-01
-5.21335416e-02 1.31114554e+00 -1.03334594e+00 7.28094518e-01
8.63290071e-01 9.75598514e-01 -6.65238857e-01 1.02286267e+00
-2.57544309e-01 6.94372296e-01 -4.02419478e-01 1.74937740e-01
3.32314730e-01 -2.74049670e-01 4.46297944e-01 1.39710057e+00
-1.05533607e-01 1.80902891e-02 2.33612865e-01 8.10380757e-01
-7.68145248e-02 -6.89461380e-02 -3.92160892e-01 -6.30592287e-01
1.13681413e-01 1.71758056e+00 -5.04405558e-01 -5.53857267e-01
-4.78356868e-01 5.73339701e-01 4.80609924e-01 2.83013701e-01
-6.21151686e-01 -5.38302481e-01 1.48673213e+00 8.57673436e-02
1.19926900e-01 -9.19324934e-01 -4.54692990e-01 -1.69524491e+00
5.92776611e-02 -7.78157175e-01 2.06734076e-01 -8.40846598e-01
-1.23441672e+00 5.78427672e-01 9.68356878e-02 -9.20206606e-01
-1.12231731e-01 -1.17752635e+00 -5.18737555e-01 8.73924434e-01
-1.23544312e+00 -1.04688883e+00 -2.17463359e-01 6.64397418e-01
-1.64473474e-01 -4.43877913e-02 8.59360933e-01 5.56862593e-01
-9.04170454e-01 7.09229335e-02 5.37585802e-02 7.16471672e-01
2.01217577e-01 -1.34167302e+00 7.82323778e-01 8.57240796e-01
-7.61953145e-02 1.07922411e+00 6.43965900e-01 -3.70473355e-01
-1.73878980e+00 -9.61943924e-01 8.67999375e-01 -3.83520365e-01
1.48231041e+00 -5.38279116e-01 -9.40599263e-01 1.02633643e+00
5.25543205e-02 5.20671189e-01 7.45712042e-01 5.93566298e-01
-7.58175254e-01 -2.11073354e-01 -7.38453567e-01 6.00953400e-01
7.31948972e-01 -7.37879872e-01 -3.56680810e-01 6.68927729e-01
8.68331492e-01 -5.57542443e-01 -1.58259487e+00 2.39506692e-01
6.01236999e-01 -9.93255377e-01 1.07703769e+00 -1.04314101e+00
3.52389544e-01 -5.35414755e-01 1.83740124e-01 -1.29211748e+00
-7.13918149e-01 -4.98278201e-01 -3.21373850e-01 7.00303912e-01
3.31605017e-01 -5.33754468e-01 8.38637531e-01 5.96504509e-01
-2.10517049e-01 -7.06776261e-01 -5.54507852e-01 -3.61839145e-01
4.46898537e-03 -8.81907225e-01 1.02693880e+00 1.10307193e+00
2.39962965e-01 3.91187459e-01 -2.00380966e-01 2.72769898e-01
4.96718049e-01 -1.64309099e-01 1.03845656e+00 -1.45077431e+00
-1.79960668e-01 -5.80032587e-01 -6.33631051e-01 -1.06665218e+00
-1.16993561e-01 -1.10377336e+00 -5.55252075e-01 -1.62972951e+00
6.37550354e-02 -5.89112222e-01 -1.34605333e-01 1.02624309e+00
1.35154165e-02 3.61578882e-01 4.29628044e-01 2.68517762e-01
-3.24498773e-01 2.44505376e-01 1.43286669e+00 -2.33506978e-01
-3.47896777e-02 1.03548363e-01 -6.00451529e-01 6.30318820e-01
8.27522159e-01 -2.31395483e-01 -6.59514889e-02 -8.77000690e-01
4.97273535e-01 -4.35348123e-01 3.82071763e-01 -9.46317971e-01
-2.79664379e-02 -1.28340974e-01 4.20863867e-01 -7.35817134e-01
4.26944941e-01 -6.29934907e-01 -1.81518123e-02 1.55949548e-01
-1.75397675e-02 5.27814686e-01 3.21000993e-01 -2.50601053e-01
-1.92758039e-01 -3.44545022e-02 4.57369745e-01 -1.97792172e-01
-4.51676011e-01 5.45243800e-01 -2.52515972e-01 -5.45201302e-01
6.62385583e-01 3.10434967e-01 -3.60809326e-01 3.79987471e-02
-9.41589236e-01 2.33992562e-01 1.04018725e-01 5.31650245e-01
4.27164972e-01 -1.50791287e+00 -5.29474795e-01 2.52167761e-01
-2.12290883e-02 1.92949668e-01 9.33858752e-02 9.68844771e-01
-1.03664756e+00 7.35194325e-01 -1.91621333e-01 -5.20068824e-01
-1.16143215e+00 5.73211372e-01 4.32652980e-01 -4.62490320e-01
-6.32038772e-01 8.57338011e-01 -1.57405317e-01 -1.07022417e+00
1.78450868e-01 -9.38821137e-01 -8.72431844e-02 -2.47302413e-01
7.97711551e-01 5.14147520e-01 4.94021803e-01 -3.66883337e-01
-3.74525219e-01 8.77680555e-02 -1.24880314e-01 2.23380506e-01
1.64345682e+00 4.17762846e-01 -1.15388238e+00 4.61891770e-01
1.25279510e+00 -4.80409041e-02 -6.26456976e-01 -2.71243095e-01
1.02236860e-01 -2.10354134e-01 3.17881227e-01 -4.33619022e-01
-1.48728764e+00 9.22303557e-01 6.39943704e-02 4.61652279e-01
8.36564720e-01 7.35755563e-02 7.42758870e-01 5.81774414e-01
8.92915577e-02 -6.10862792e-01 -1.46367654e-01 1.09355843e+00
7.71906614e-01 -8.53877425e-01 4.00042206e-01 -4.62833107e-01
-2.25505456e-01 1.53463435e+00 3.85750920e-01 -1.80309594e-01
1.27099895e+00 4.00397569e-01 4.13006581e-02 -4.39102948e-01
-8.89151812e-01 -1.92568004e-01 4.94915843e-01 3.36768776e-01
9.69682395e-01 3.76402617e-01 -6.88262507e-02 2.73793548e-01
-6.79041207e-01 2.78367251e-01 5.04722238e-01 6.76961660e-01
-1.61430072e-02 -1.31690884e+00 -5.87773502e-01 5.39050162e-01
-6.04836106e-01 -3.40969056e-01 -1.05158642e-01 2.82832861e-01
1.34008020e-01 4.39983517e-01 2.22811386e-01 -8.33407462e-01
1.73071772e-01 -1.59916058e-01 2.93975234e-01 -6.08433366e-01
-7.16415048e-01 -3.26399088e-01 3.31554979e-01 -9.64794278e-01
-2.20979974e-01 -4.54591781e-01 -1.06705284e+00 -8.89601767e-01
-1.69706773e-02 2.66316772e-01 1.13842154e+00 9.92581606e-01
6.28558874e-01 5.81947505e-01 3.75454098e-01 -1.18801367e+00
-1.66284814e-01 -1.21893013e+00 -5.05870402e-01 -7.61503279e-02
4.62952375e-01 -5.13597071e-01 -8.05542693e-02 3.49555456e-04] | [6.3113789558410645, 5.052757740020752] |
56e8db59-1210-4a41-ace8-74cfce296b74 | fooling-state-of-the-art-deepfake-detection | 2305.05282 | null | https://arxiv.org/abs/2305.05282v2 | https://arxiv.org/pdf/2305.05282v2.pdf | Fooling State-of-the-Art Deepfake Detection with High-Quality Deepfakes | Due to the rising threat of deepfakes to security and privacy, it is most important to develop robust and reliable detectors. In this paper, we examine the need for high-quality samples in the training datasets of such detectors. Accordingly, we show that deepfake detectors proven to generalize well on multiple research datasets still struggle in real-world scenarios with well-crafted fakes. First, we propose a novel autoencoder for face swapping alongside an advanced face blending technique, which we utilize to generate 90 high-quality deepfakes. Second, we feed those fakes to a state-of-the-art detector, causing its performance to decrease drastically. Moreover, we fine-tune the detector on our fakes and demonstrate that they contain useful clues for the detection of manipulations. Overall, our results provide insights into the generalization of deepfake detectors and suggest that their training datasets should be complemented by high-quality fakes since training on mere research data is insufficient. | ['Peter Eisert', 'Anna Hilsmann', 'Arian Beckmann'] | 2023-05-09 | null | null | null | null | ['deepfake-detection', 'face-swapping'] | ['computer-vision', 'computer-vision'] | [-1.35246903e-01 1.75537229e-01 7.85304885e-03 -3.59968901e-01
-5.46974242e-01 -7.23363101e-01 4.84363645e-01 -4.09741342e-01
-2.01377362e-01 5.25006235e-01 -5.95205463e-02 -2.16029912e-01
3.28584909e-01 -5.66927969e-01 -9.93269145e-01 -2.89253622e-01
1.54697552e-01 1.12167679e-01 -5.28673790e-02 -3.75068545e-01
1.46175744e-02 5.16262591e-01 -1.60683894e+00 4.99805510e-01
5.52362502e-01 8.51550102e-01 -5.04053831e-01 3.68151724e-01
4.17480707e-01 2.57486045e-01 -9.82384801e-01 -1.23561645e+00
6.53271079e-01 -4.87847328e-01 -3.60819072e-01 -1.07681334e-01
1.07925642e+00 -1.09563649e+00 -4.66269016e-01 1.13141930e+00
4.38938469e-01 -4.41448599e-01 3.15463066e-01 -1.70281243e+00
-7.90599704e-01 4.33396935e-01 -5.77220738e-01 1.69075131e-01
1.66475996e-01 5.13626218e-01 7.50836074e-01 -8.89616430e-01
7.07697093e-01 1.44255269e+00 9.06849325e-01 1.01965392e+00
-1.13660395e+00 -1.32141471e+00 -2.37512112e-01 -1.06147960e-01
-1.03373837e+00 -1.00302780e+00 5.95923901e-01 -4.26733732e-01
5.94441056e-01 -5.25427647e-02 7.37049580e-01 1.94229019e+00
-2.74853498e-01 7.17333734e-01 8.76499414e-01 -3.00024986e-01
1.00862756e-01 3.61270249e-01 -2.42841899e-01 9.71951187e-01
9.49911177e-01 4.71984118e-01 -6.32393301e-01 -4.06194329e-01
9.30475414e-01 -1.43365040e-01 -1.07934326e-01 -1.78554028e-01
-5.74100673e-01 1.02284169e+00 3.83239865e-01 4.67938811e-01
-2.34181449e-01 2.17262462e-01 2.77090371e-01 4.14382219e-01
3.11929047e-01 9.36552286e-01 -3.72635335e-01 1.54744074e-01
-1.06387806e+00 3.16302210e-01 5.60816705e-01 7.24771857e-01
6.21614814e-01 1.22706041e-01 -1.19024858e-01 4.69606519e-01
6.65182844e-02 3.91320974e-01 3.02428126e-01 -8.05079997e-01
2.74588227e-01 5.38966596e-01 2.68914521e-01 -1.31491852e+00
-1.22853488e-01 -4.06590700e-01 -4.68632936e-01 4.35312875e-02
6.61996901e-01 -1.28131419e-01 -7.48001635e-01 1.79041576e+00
2.15512916e-01 5.90229630e-01 -2.12295532e-01 9.01518285e-01
5.32331586e-01 2.45141983e-02 -1.04476146e-01 1.29848972e-01
1.22397304e+00 -5.76163828e-01 -6.04610026e-01 -1.89588636e-01
8.24120581e-01 -3.85131657e-01 1.19151485e+00 4.81915832e-01
-7.53768861e-01 -3.38683039e-01 -1.17038012e+00 7.47449324e-02
-3.49112779e-01 4.04218704e-01 9.44704711e-01 1.39521050e+00
-1.02063906e+00 7.56342828e-01 -6.26427889e-01 -3.83224189e-01
1.12129319e+00 4.79927391e-01 -7.19767153e-01 1.70977600e-03
-1.11283827e+00 8.15814793e-01 6.13178425e-02 2.26778492e-01
-1.15798914e+00 -4.69686568e-01 -5.06653666e-01 -3.66457715e-03
3.23413044e-01 -2.98607320e-01 1.09428310e+00 -1.04814053e+00
-1.16573632e+00 1.06422150e+00 1.30313382e-01 -3.88970971e-01
6.59742355e-01 -1.51446819e-01 -5.59717715e-01 3.33512366e-01
1.45851195e-01 7.18578339e-01 1.39957428e+00 -1.31126225e+00
-2.84580588e-01 -6.06546581e-01 -7.88737535e-02 -7.17468798e-01
-1.00687170e+00 1.23401202e-01 3.17126811e-02 -5.53049862e-01
-3.50712717e-01 -7.76964784e-01 2.36289412e-01 2.29135141e-01
-6.54690325e-01 -2.47855913e-02 1.09682155e+00 -6.70051754e-01
9.63122308e-01 -2.14921618e+00 -4.23828602e-01 1.12764008e-01
5.13828099e-01 6.81569517e-01 -5.09866357e-01 1.74237229e-02
-1.96841836e-01 4.65752363e-01 2.55122073e-02 -4.50520366e-01
-1.76941827e-02 -9.42753348e-03 -5.51919699e-01 8.01236928e-01
7.50149906e-01 1.04616666e+00 -9.74472284e-01 1.55806690e-02
-9.93548557e-02 3.57843637e-01 -9.34981167e-01 1.88421771e-01
-1.32606238e-01 4.11324322e-01 -4.43070799e-01 1.03710973e+00
1.00176072e+00 -1.74148995e-02 9.86109301e-02 -5.81947677e-02
3.73598874e-01 1.35716856e-01 -7.20276475e-01 1.17890131e+00
-3.91936451e-02 7.11994350e-01 2.43769944e-01 -8.05150211e-01
7.52422631e-01 6.83018267e-02 1.92632630e-01 -6.26879215e-01
4.26466405e-01 4.40008610e-01 1.21382072e-05 -5.90676665e-01
3.67893130e-01 -1.33821502e-01 1.24677233e-01 5.40853024e-01
2.98558772e-01 7.01280162e-02 -2.37796202e-01 8.89390856e-02
1.42736876e+00 -9.73630622e-02 -3.66786897e-01 3.67853530e-02
1.24383375e-01 -3.62743348e-01 4.01277691e-01 9.81509626e-01
-6.77695155e-01 4.28131938e-01 7.45261490e-01 -5.29021919e-01
-1.17926002e+00 -6.04937375e-01 -1.74251795e-02 1.04968870e+00
-2.43222922e-01 -3.75244021e-01 -1.20069325e+00 -1.11158073e+00
4.67141658e-01 3.75297010e-01 -8.48640800e-01 -5.81073344e-01
-4.78808761e-01 -6.99332535e-01 1.38641763e+00 2.06257343e-01
5.43495357e-01 -8.30778360e-01 -4.93548989e-01 -9.36224237e-02
3.51852439e-02 -1.26203954e+00 -9.38479826e-02 -2.29006544e-01
-5.09010732e-01 -1.42010844e+00 -4.29714173e-01 -4.76968974e-01
6.09626591e-01 3.42471421e-01 8.44144523e-01 7.02373803e-01
-3.95868570e-01 2.56779790e-02 -2.35264704e-01 -2.41462186e-01
-6.28247440e-01 -8.11249241e-02 3.58426660e-01 7.98203424e-02
7.03966260e-01 -4.43660855e-01 -3.49333555e-01 2.95624048e-01
-8.08171213e-01 -6.00646138e-01 5.46803534e-01 1.03275692e+00
-2.83681303e-01 3.37988958e-02 7.07033932e-01 -8.06940079e-01
8.23606491e-01 -2.51514256e-01 -8.24625075e-01 1.84428021e-01
-3.93364370e-01 9.31044519e-02 4.74738628e-01 -6.89824760e-01
-5.96930504e-01 -1.18381092e-02 -5.72456978e-02 -7.87533343e-01
-1.43238544e-01 -2.05454186e-01 -2.20910996e-01 -4.89127785e-01
1.10959935e+00 -1.36116415e-01 4.94350912e-04 -3.89331996e-01
5.04391134e-01 7.38355517e-01 4.77735162e-01 -5.55772960e-01
1.09379733e+00 5.62202811e-01 -7.21439943e-02 -7.65114546e-01
-7.12022781e-01 -6.84976652e-02 -3.22396636e-01 -1.22819154e-03
3.88684213e-01 -8.59637797e-01 -1.07004201e+00 7.59115815e-01
-1.34402990e+00 -1.68037325e-01 1.63680092e-01 1.53019518e-01
-1.79073200e-01 4.72521603e-01 -7.19201028e-01 -9.43598628e-01
-3.35896574e-02 -1.24283135e+00 1.31378162e+00 -2.32275836e-02
2.16086879e-02 -4.30671960e-01 -1.91322848e-01 5.55396557e-01
3.24887812e-01 2.46427149e-01 4.16742265e-01 -6.65573955e-01
-6.42906666e-01 -3.86596531e-01 -3.72033805e-01 3.34451735e-01
1.05291322e-01 1.54539406e-01 -1.49427140e+00 -4.43041950e-01
1.11754261e-01 -7.42563784e-01 1.09310627e+00 -2.24683881e-02
1.36802638e+00 -6.20271802e-01 -3.60458106e-01 7.66409039e-01
9.69139993e-01 -2.77604103e-01 5.91850102e-01 1.39939159e-01
7.53796339e-01 7.81391561e-01 8.75187367e-02 3.78497839e-01
4.91271801e-02 6.10520363e-01 5.40758371e-01 7.46234059e-02
1.19211331e-01 -5.72449744e-01 4.21076655e-01 2.25400757e-02
2.27948308e-01 -2.04686314e-01 -4.99698043e-01 3.54808867e-01
-1.54608059e+00 -1.05850899e+00 -1.06239788e-01 2.09160447e+00
7.00145721e-01 1.14290714e-01 4.21094984e-01 -1.26015898e-02
7.46653497e-01 -1.25244632e-01 -6.41569138e-01 -2.57603198e-01
-2.49416515e-01 3.82123142e-01 6.14934325e-01 1.54827699e-01
-1.24575448e+00 1.50523555e+00 6.79547930e+00 4.67732936e-01
-1.27627003e+00 3.24201673e-01 7.00698674e-01 -9.91652627e-03
-2.60953307e-01 -1.99748576e-01 -7.68950880e-01 6.12472951e-01
9.89761710e-01 3.99592876e-01 7.73925304e-01 8.10235083e-01
-1.05981501e-02 1.75696120e-01 -1.11449194e+00 1.10484242e+00
2.16243908e-01 -1.31630719e+00 -1.59534588e-01 3.81981343e-01
5.55889189e-01 -3.01554978e-01 3.34490985e-01 3.08433235e-01
4.36443269e-01 -1.18148983e+00 4.89677697e-01 3.99379618e-02
7.50085711e-01 -6.23754501e-01 6.15224898e-01 1.84817016e-01
-2.78542250e-01 -2.84750879e-01 -5.86132884e-01 2.28359643e-02
-3.57307404e-01 6.18523180e-01 -1.22257805e+00 -3.07306405e-02
5.61836183e-01 5.13565719e-01 -8.12908351e-01 6.74733758e-01
-3.11209381e-01 5.36555469e-01 -5.02553642e-01 1.50281087e-01
-1.03173994e-01 1.84542686e-01 1.20493650e-01 8.78220081e-01
4.41523403e-01 -1.91551074e-01 -3.10389638e-01 1.26166916e+00
-7.22880840e-01 -4.54642922e-01 -9.72568333e-01 -6.75277412e-01
5.80241859e-01 1.24735975e+00 -4.13944870e-01 -1.30604342e-01
-1.48311272e-01 1.16025162e+00 5.42755187e-01 1.41262218e-01
-7.60446191e-01 1.29197240e-01 1.00291395e+00 1.72346145e-01
2.58134544e-01 1.74924880e-01 -2.93001622e-01 -1.53248489e+00
1.37039065e-01 -1.22160435e+00 1.68411732e-01 -5.40129244e-01
-1.53851902e+00 5.61559439e-01 -5.15904784e-01 -8.84898067e-01
-2.14550510e-01 -7.93112457e-01 -3.29235196e-01 5.78029752e-01
-1.39202189e+00 -1.32585120e+00 -3.09992820e-01 4.97703850e-01
-9.72247347e-02 -4.12246853e-01 7.35413074e-01 3.40071291e-01
-8.59072506e-01 1.31232417e+00 -2.61931956e-01 5.07629991e-01
9.18326139e-01 -6.05705142e-01 7.85656273e-01 1.04256749e+00
2.96370864e-01 1.01873553e+00 5.84259570e-01 -7.72900462e-01
-1.60531890e+00 -9.06725347e-01 5.62347591e-01 -8.86316180e-01
7.56474853e-01 -8.75770032e-01 -9.70277607e-01 7.33231843e-01
-3.08758527e-01 3.65202129e-01 4.49000835e-01 7.30507746e-02
-1.00835562e+00 9.85373929e-03 -1.72138357e+00 4.43732888e-01
1.12888563e+00 -9.28904116e-01 -2.85178751e-01 3.87848496e-01
4.56604838e-01 -1.82763383e-01 -2.92671531e-01 1.67813525e-01
8.20438504e-01 -1.30363488e+00 9.27323461e-01 -1.00623953e+00
4.90573674e-01 -1.94876064e-02 1.83746979e-01 -1.18697000e+00
7.52695575e-02 -6.30972862e-01 -2.49215350e-01 1.00015295e+00
1.33525357e-01 -6.85840905e-01 1.29349971e+00 6.01075888e-01
2.81969666e-01 -2.90431082e-01 -1.04582977e+00 -9.00688589e-01
3.90250096e-03 -2.78263479e-01 8.26470315e-01 1.39906895e+00
-2.29910493e-01 -3.78347598e-02 -7.40597904e-01 2.04222560e-01
7.04447806e-01 -4.18862671e-01 1.06807804e+00 -1.35236812e+00
-3.13703328e-01 -4.52287525e-01 -5.63009083e-01 -7.16600239e-01
4.94666666e-01 -5.12488782e-01 -2.38823652e-01 -4.62157071e-01
1.33817896e-01 -2.95912504e-01 4.83034253e-02 7.89864957e-01
-1.05604671e-01 6.07407868e-01 2.84881387e-02 7.00151697e-02
-7.84807205e-02 4.92085963e-01 1.06882906e+00 -9.41877961e-02
-1.59457605e-02 -1.42896563e-01 -1.13690197e+00 5.45298636e-01
5.91515362e-01 -6.95119917e-01 -1.00739613e-01 -3.24384421e-01
2.21636668e-01 -3.37724596e-01 7.97785878e-01 -7.70534933e-01
-1.01344690e-01 -1.54935047e-02 6.15840197e-01 -1.46598652e-01
4.66684878e-01 -6.96536183e-01 -3.02474022e-01 4.61580187e-01
-1.19865343e-01 -2.27152482e-01 1.76420540e-01 4.00590479e-01
2.45590016e-01 -1.34027317e-01 8.09386730e-01 -1.51827827e-01
-3.13407153e-01 2.31809139e-01 1.20488361e-01 -1.36269540e-01
7.42760658e-01 -2.35129640e-01 -7.20193028e-01 -3.87604654e-01
-2.51271963e-01 -1.95434377e-01 7.53338993e-01 4.84975010e-01
6.13807142e-01 -1.16284251e+00 -5.46585441e-01 7.11766899e-01
1.49987057e-01 -4.92940098e-01 -1.18229710e-01 3.94958407e-01
-3.65061402e-01 1.64501280e-01 -4.60809946e-01 -2.73294389e-01
-9.89597499e-01 6.82142973e-01 3.48862529e-01 2.56120563e-01
-3.10357481e-01 1.19325185e+00 1.46791968e-03 -3.80576253e-01
1.58597380e-01 -2.65785754e-01 2.92141080e-01 9.30004045e-02
6.10004425e-01 1.78632755e-02 2.40232810e-01 -6.69630527e-01
-6.03396773e-01 -1.15174517e-01 -3.36021259e-02 5.46635799e-02
1.36241233e+00 3.94526213e-01 -1.34859085e-01 -4.02424246e-01
1.14201689e+00 2.31929153e-01 -1.26564956e+00 2.54702300e-01
1.04884483e-01 -9.43406582e-01 8.37624148e-02 -7.51430333e-01
-1.21671689e+00 8.45658302e-01 6.13796115e-01 2.73125410e-01
9.11530316e-01 -4.94958460e-02 9.23521817e-01 5.63547432e-01
4.30449426e-01 -8.10945809e-01 4.60341275e-01 2.46611461e-01
6.87535763e-01 -1.43369043e+00 -1.16536349e-01 -4.21503901e-01
-2.66234219e-01 9.98002052e-01 7.78927863e-01 -8.98927227e-02
3.35410208e-01 1.24780118e-01 -1.96619015e-02 -3.21504653e-01
-5.12742043e-01 1.08221546e-01 -4.00055051e-02 7.91845620e-01
-4.03518882e-03 1.16186023e-01 1.99975401e-01 6.72899544e-01
-5.49815297e-01 1.33913532e-01 6.85018480e-01 5.29606283e-01
-1.95068583e-01 -1.17330933e+00 -4.51516300e-01 4.70033586e-01
-4.40974087e-01 4.16516736e-02 -8.68855953e-01 6.73735142e-01
2.21026719e-01 1.05991054e+00 -1.58850163e-01 -7.33267367e-01
2.02237204e-01 -2.64842995e-02 7.44756103e-01 -3.98882180e-01
-8.56652796e-01 -2.21778885e-01 7.33743757e-02 -7.33649313e-01
2.87391450e-02 -4.93957281e-01 -4.59059477e-01 -6.46890819e-01
-6.41016901e-01 -2.19816610e-01 7.22814798e-01 9.58849728e-01
6.40824020e-01 -1.15555428e-01 6.65100276e-01 -7.14202940e-01
-8.25360239e-01 -1.00867438e+00 -4.52216238e-01 4.90566373e-01
6.86992764e-01 -8.90564263e-01 -6.04251146e-01 -3.67302895e-01] | [12.655986785888672, 1.0501515865325928] |
deb1913c-437e-4f30-9400-825a1d50ae87 | pre-scaling-and-codebook-design-for-joint | 2111.10527 | null | https://arxiv.org/abs/2111.10527v1 | https://arxiv.org/pdf/2111.10527v1.pdf | Pre-scaling and Codebook Design for Joint Radar and Communication Based on Index Modulation | This paper develops an efficient index modulation (IM) approach for the joint radar-communication (JRC) system based on a multi-carrier multiple-input multiple-output (MIMO) radar. The communication information is embedded into the transmitted radar pulses by selecting the corresponding indices of the carrier frequencies and antenna allocations, providing two degrees of freedom. Our contribution involves the development of a novel codebook based minimum Euclidean distance (MED) maximization and a constellation randomization pre-scaling (CRPS) scheme for efficient IM-JRC transmission. It can be inferred that the IM approach integrating the CRPS scheme followed by the codebook design maximizes the signal-to-noise ratio gain. The numerical results support the effectiveness of the proposed approach and show enhanced bit error rate performance when compared to the existing baseline. | ['Christos Masouros', 'Aryan Kaushik', 'Shengyang Chen'] | 2021-11-20 | null | null | null | null | ['joint-radar-communication'] | ['robots'] | [ 8.13005090e-01 -1.22901775e-01 -1.30886734e-01 -1.99090376e-01
-7.17267215e-01 -4.64333922e-01 9.15462255e-01 -8.17052349e-02
-4.55686986e-01 6.93764150e-01 2.83056777e-02 -6.80326402e-01
-8.95971537e-01 -6.44163311e-01 -1.55762872e-02 -9.54194069e-01
-7.75966406e-01 8.82150978e-02 -2.45649189e-01 -2.06366226e-01
4.98433858e-01 9.59755540e-01 -1.24137080e+00 -4.63524014e-01
6.48857832e-01 1.31206107e+00 2.06665710e-01 1.15670049e+00
5.08995771e-01 2.34152064e-01 -1.09384775e+00 -1.37385145e-01
5.79418540e-01 -5.01747549e-01 2.20162451e-01 -7.62486756e-02
1.49857821e-02 -1.18512437e-01 -4.20790225e-01 9.42456543e-01
6.38405800e-01 -1.69525981e-01 1.08516502e+00 -1.17032576e+00
-3.71759862e-01 2.40920275e-01 -6.49767220e-01 1.81516156e-01
-1.62452549e-01 -3.06783587e-01 6.20485127e-01 -5.75632334e-01
2.18861222e-01 1.10105073e+00 2.28198022e-01 -1.09720923e-01
-1.03199029e+00 -9.46520925e-01 -6.80279791e-01 1.69936806e-01
-1.82868242e+00 -5.20956755e-01 6.46168828e-01 -1.91781104e-01
4.65426832e-01 6.21384144e-01 3.82593900e-01 -8.43853503e-03
7.04664052e-01 -1.14490569e-01 1.26712596e+00 -1.04171073e+00
3.51305783e-01 -5.14729321e-02 -1.79343283e-01 4.46007878e-01
6.27356529e-01 1.00954831e+00 -1.91629827e-01 -3.81479889e-01
7.29949296e-01 -4.35812116e-01 -3.28720599e-01 -2.08537459e-01
-1.04100621e+00 1.13599610e+00 1.70295790e-01 3.66372973e-01
-4.27051902e-01 4.46100980e-01 -3.84415865e-01 6.01933181e-01
-8.32303092e-02 4.41449910e-01 2.35835761e-01 3.49859953e-01
-1.13681459e+00 2.70170093e-01 8.01830053e-01 1.22590363e+00
3.14389706e-01 3.15239787e-01 -5.15656352e-01 3.48034799e-01
6.33867621e-01 1.42393172e+00 -4.83474553e-01 -9.93537545e-01
3.67829949e-01 -2.29887292e-01 1.85625657e-01 -1.10247743e+00
-3.58672947e-01 -1.17636192e+00 -1.04928684e+00 3.02501589e-01
-2.06561103e-01 -5.71059108e-01 -6.96829140e-01 1.44621301e+00
6.18915036e-02 -9.83652920e-02 6.91559315e-01 7.89045691e-01
2.13150933e-01 6.20999455e-01 -4.25699681e-01 -6.96779907e-01
1.49765944e+00 -7.61197135e-02 -1.15184474e+00 1.65266529e-01
1.99171081e-01 -9.56234515e-01 -4.88113582e-01 2.32901752e-01
-6.06539190e-01 -3.69929433e-01 -1.77809656e+00 9.60346699e-01
-3.24001932e-03 3.49617183e-01 4.03673083e-01 1.18731058e+00
-7.32814789e-01 7.37237781e-02 -2.32123826e-02 2.01224253e-01
1.11091010e-01 4.16625410e-01 2.76948512e-01 -9.74712893e-02
-1.20489597e+00 9.15376604e-01 3.18178862e-01 1.60094932e-01
-4.13873106e-01 -4.93764281e-01 -5.69277227e-01 -3.19081903e-01
1.58848211e-01 -3.25595349e-01 8.15290511e-01 -1.03044234e-01
-1.32775617e+00 6.22671582e-02 1.52756557e-01 -9.33790088e-01
-8.35543200e-02 2.21902400e-01 -8.43569875e-01 6.27403617e-01
-2.61115730e-01 4.77816969e-01 8.79047334e-01 -1.31626713e+00
-7.74585485e-01 -5.74888960e-02 -3.13589603e-01 5.65881617e-02
4.99116093e-01 4.35405876e-03 2.58299354e-02 -9.23489928e-01
1.08984873e-01 -7.68800080e-01 -4.82032150e-01 -2.98047781e-01
-3.17476183e-01 3.77438664e-01 8.69415164e-01 -2.84867585e-01
1.42659771e+00 -2.16206932e+00 2.97673792e-02 6.70813084e-01
-1.89558506e-01 2.84502566e-01 -1.09028141e-03 7.81655610e-01
6.45797551e-02 -3.80425900e-01 -2.81505585e-01 1.65300280e-01
-4.62229997e-02 3.96023132e-02 -4.06090945e-01 7.96392381e-01
1.08272232e-01 4.59883481e-01 -3.57731044e-01 -3.87823611e-01
2.05445603e-01 2.41985828e-01 -1.90320373e-01 -9.23183933e-02
1.00797571e-01 2.71437645e-01 -3.03446501e-01 9.02875781e-01
1.18259728e+00 3.86157632e-01 1.08367175e-01 -4.18593675e-01
-4.43125516e-01 -4.62209731e-01 -1.18004560e+00 8.46109152e-01
-1.21696018e-01 7.37102032e-01 3.95811349e-01 -1.16938376e+00
1.24637473e+00 3.69819343e-01 4.69632119e-01 -9.16204035e-01
3.36593568e-01 3.44338089e-01 2.78405964e-01 5.20232767e-02
5.13580978e-01 -3.36347520e-01 -4.32879806e-01 2.90208936e-01
-7.49759227e-02 -1.04507565e-01 2.60002147e-02 2.26197869e-01
8.84286582e-01 -6.11147404e-01 7.40353346e-01 -3.95776510e-01
7.11324513e-01 -3.60656790e-02 2.37216920e-01 8.71894240e-01
-5.27643338e-02 -1.32947266e-01 -1.09700978e-01 5.79712272e-01
-7.54811764e-01 -9.64970648e-01 -7.63423443e-01 5.15244007e-02
4.16565746e-01 6.88216686e-02 -1.74045905e-01 1.17888682e-01
3.54238927e-01 9.13141012e-01 -1.03672901e-02 -1.20766053e-03
-2.17985094e-01 -7.29372859e-01 7.80708909e-01 -2.83047885e-01
5.71186185e-01 -2.04990000e-01 -8.33448708e-01 4.42253828e-01
2.84157902e-01 -1.32117510e+00 -1.54073253e-01 4.94091213e-01
-5.39050400e-01 -7.22624362e-01 -4.69093025e-01 -4.84191805e-01
5.22930861e-01 5.18171132e-01 2.53877163e-01 -4.30457741e-01
-4.74397779e-01 5.95502377e-01 -5.98390698e-01 -4.80702370e-01
-2.19921857e-01 -6.59443974e-01 1.22356296e-01 3.07539217e-02
1.34148449e-01 -4.92182702e-01 -3.92055362e-01 8.94289389e-02
-5.77689588e-01 -2.20500425e-01 1.27765357e+00 4.52347368e-01
1.22100689e-01 4.34189230e-01 8.44287872e-01 -1.54273406e-01
6.21460438e-01 -2.86050141e-01 -1.04380000e+00 1.75345540e-01
-7.41651416e-01 2.53452033e-01 3.63756388e-01 -6.80800006e-02
-5.72461247e-01 -1.25563681e-01 1.77537709e-01 8.62368420e-02
1.59606785e-01 4.43595082e-01 -5.50231934e-02 -8.64905179e-01
1.76120758e-01 4.90522236e-01 3.50847244e-02 -9.91252959e-02
6.07708454e-01 1.23647511e+00 9.01417196e-01 -2.52914101e-01
1.44073713e+00 2.72318244e-01 6.77826703e-01 -1.35131669e+00
-4.07513112e-01 -4.25203204e-01 -4.36099201e-01 -4.22339469e-01
6.08974695e-01 -1.03404462e+00 -6.91990137e-01 -3.28235552e-02
-1.09766531e+00 4.86328602e-01 -6.45450223e-03 1.15888739e+00
-4.55405474e-01 3.54916781e-01 9.40261483e-02 -1.57265210e+00
-3.28729063e-01 -7.26460755e-01 8.94277990e-01 5.62665910e-02
1.69544384e-01 -5.89677989e-01 -1.06690340e-01 1.54367119e-01
7.82185018e-01 5.73284686e-01 9.93181229e-01 -2.19274506e-01
-1.12335360e+00 -4.30921793e-01 -3.39906156e-01 -2.94729657e-02
-1.43443778e-01 -5.98195612e-01 -3.53366047e-01 -5.40106297e-01
-4.08049785e-02 2.35038608e-01 5.71208715e-01 6.51358068e-01
2.72886336e-01 -3.09904426e-01 -4.69308138e-01 7.19510138e-01
1.88990021e+00 7.90285766e-01 6.88398182e-01 4.31192480e-02
-3.50438774e-01 8.03214461e-02 8.52776766e-01 7.98681617e-01
-7.17828721e-02 9.75142002e-01 2.79248565e-01 2.02221155e-01
5.82568087e-02 4.38795179e-01 2.72654921e-01 6.15464330e-01
2.28389695e-01 -3.78120899e-01 -1.36123165e-01 3.12294155e-01
-1.22794271e+00 -1.09923363e+00 -2.02304333e-01 2.36406279e+00
6.60354912e-01 -3.49086709e-02 -4.22677100e-01 3.19003463e-01
7.40863919e-01 1.38157219e-01 -8.98804739e-02 -5.48710048e-01
-2.01235726e-01 5.20148993e-01 1.42985308e+00 7.78791606e-01
-1.17911017e+00 5.13575710e-02 6.52490616e+00 1.14494562e+00
-6.04868948e-01 -2.26019621e-02 -2.55031407e-01 8.26056972e-02
-2.04279989e-01 1.33555427e-01 -1.11413407e+00 3.17639947e-01
1.37110519e+00 -4.79958594e-01 2.47076988e-01 5.52202500e-02
1.71334520e-01 -6.41365767e-01 -4.78840858e-01 8.99909437e-01
2.17708468e-01 -1.33815718e+00 -3.75159532e-02 6.17633104e-01
3.75567824e-01 -5.16065896e-01 9.37239677e-02 2.73290545e-01
3.86315957e-02 -8.42895865e-01 4.96509582e-01 8.38643789e-01
1.05874431e+00 -9.97776508e-01 6.84785962e-01 3.86948735e-01
-1.32525194e+00 -2.84635425e-01 -1.83179259e-01 1.13243364e-01
4.54470217e-01 8.92882168e-01 -6.62273765e-01 1.16702271e+00
-2.16862798e-01 -1.60586447e-01 -2.82290280e-01 1.13654423e+00
1.82150692e-01 5.13184369e-01 -4.70572710e-01 -3.48070294e-01
3.01442653e-01 -4.21314925e-01 1.14760888e+00 1.39512658e+00
7.86067963e-01 5.32007933e-01 1.16503865e-01 4.43379909e-01
4.08827186e-01 -2.69070476e-01 -7.14232147e-01 -1.78748310e-01
1.19189727e+00 9.90729749e-01 -2.81207293e-01 -1.98981073e-03
-2.17306972e-01 3.14337134e-01 -6.77530527e-01 3.08619171e-01
-6.87069654e-01 -1.17506588e+00 4.86510605e-01 -3.98988456e-01
8.41491699e-01 -7.10307598e-01 -2.76064306e-01 -5.07338420e-02
-2.19603479e-01 -5.88540673e-01 -7.99771026e-03 -2.05717504e-01
-7.42690086e-01 1.30091980e-01 4.63217497e-01 -1.79717255e+00
-2.42677674e-01 -1.86560452e-01 -9.69645679e-02 1.05135715e+00
-1.87908649e+00 -7.45640039e-01 2.07159370e-01 1.92348763e-01
-1.82970166e-01 -6.37014449e-01 8.44971538e-01 3.34210932e-01
4.53405268e-02 5.72060108e-01 6.62127495e-01 -3.90983559e-02
5.61785772e-02 -7.29356468e-01 -3.77295554e-01 1.07777202e+00
-2.30747193e-01 5.81555247e-01 8.97319317e-01 -5.78607798e-01
-1.71124613e+00 -1.12425947e+00 7.03878939e-01 3.51214141e-01
4.35536593e-01 -2.32351631e-01 1.39187068e-01 -2.86393128e-02
5.92823565e-01 -5.77631831e-01 1.04040360e+00 -6.06124222e-01
8.81407410e-02 -4.87945862e-02 -1.18337345e+00 4.01598215e-01
4.32564080e-01 -1.63134038e-01 -5.55281997e-01 2.50722677e-01
4.90121096e-01 -6.87174052e-02 -1.04620171e+00 7.01386750e-01
5.95973492e-01 -4.68824089e-01 1.15760171e+00 2.05438465e-01
-2.59332657e-01 -8.16290021e-01 -1.00847924e+00 -1.05069482e+00
-5.15485346e-01 -9.11742985e-01 -2.11257845e-01 1.04396391e+00
3.76558185e-01 -5.51805913e-01 1.71642303e-01 -5.24642825e-01
2.84187924e-02 -6.05862141e-01 -1.37366045e+00 -1.19507968e+00
-3.29427809e-01 -4.80109453e-01 1.13169640e-01 1.37427583e-01
-4.25710268e-02 4.99314398e-01 -6.65220797e-01 8.76830220e-01
1.49219716e+00 3.81048679e-01 6.36045277e-01 -1.38830161e+00
-4.65792567e-01 -1.24267533e-01 -7.58830726e-01 -1.27275884e+00
-4.24834281e-01 -9.38758135e-01 7.59001598e-02 -1.34256554e+00
-4.09722298e-01 -5.89759231e-01 1.66006796e-02 -2.62835801e-01
3.91454667e-01 1.72089845e-01 3.84823710e-01 4.88526262e-02
-2.48901486e-01 8.58871281e-01 1.08293772e+00 3.65289338e-02
1.03328601e-01 4.54777747e-01 -7.47668087e-01 -4.08664644e-02
7.55720913e-01 -4.97391164e-01 -1.38023853e-01 3.31342995e-01
-3.04128826e-01 7.19339848e-01 4.74065632e-01 -1.55607343e+00
4.41507161e-01 -1.51007801e-01 3.22477400e-01 -1.20439708e+00
6.52694404e-01 -1.14669585e+00 3.43953192e-01 1.08390200e+00
-9.72231403e-02 -3.56240183e-01 -4.93643433e-02 1.19662786e+00
-2.90588904e-02 -4.02847201e-01 9.16790068e-01 5.19942462e-01
-4.43407357e-01 -2.40570277e-01 -7.61824131e-01 -2.97118783e-01
1.35944259e+00 -2.77476728e-01 -5.69076657e-01 -5.62344909e-01
-1.99098170e-01 3.83854210e-01 -3.55478853e-01 1.03922524e-01
6.89880788e-01 -1.58731306e+00 -9.82936859e-01 4.36260492e-01
5.29305637e-02 -9.51504707e-01 -2.12683961e-01 1.00076985e+00
-3.11803937e-01 1.30798614e+00 -5.98262511e-02 -5.84427238e-01
-1.33192670e+00 -8.61220658e-02 1.30098477e-01 -5.27437823e-03
-5.00668526e-01 4.29334909e-01 -6.59890890e-01 -4.85474877e-02
3.37538660e-01 1.74653739e-01 -2.96860546e-01 -6.94220066e-02
6.28610551e-01 6.73198849e-02 -1.75196230e-01 -6.97638035e-01
-4.04719234e-01 8.33863676e-01 1.94368765e-01 -8.01692963e-01
8.58265221e-01 -1.68516308e-01 8.56529996e-02 -5.21102175e-02
9.76316333e-01 1.89380258e-01 -6.39484167e-01 -3.24377775e-01
-1.44763663e-02 -5.60894370e-01 4.09981400e-01 -6.62550807e-01
-5.65724909e-01 3.81280065e-01 1.06366980e+00 1.01687416e-01
1.18206644e+00 -4.49520767e-01 5.13491035e-01 6.52077973e-01
9.03221369e-01 -1.11294830e+00 -4.00571436e-01 4.07666773e-01
6.80646360e-01 -3.92496496e-01 5.70430517e-01 -3.58011216e-01
-4.20350462e-01 1.13750148e+00 -4.37971473e-01 -6.31871074e-02
1.02673137e+00 8.65989089e-01 -1.00614361e-01 -3.77969652e-01
-7.44338632e-01 -7.45988786e-01 1.03424422e-01 9.54920530e-01
3.78024280e-02 3.48638892e-01 -9.37940896e-01 1.88793153e-01
-3.61864716e-01 -3.05147529e-01 4.68241185e-01 1.13001513e+00
-1.28385901e+00 -1.08508170e+00 -1.01047707e+00 5.80371916e-01
-1.00735262e-01 -1.34736493e-01 -1.59326404e-01 7.91897595e-01
1.12452008e-01 1.58996570e+00 1.00310870e-01 -5.56765378e-01
3.78679156e-01 -2.51619041e-01 6.26770854e-01 -1.32221162e-01
9.97348651e-02 3.09222788e-01 2.66421914e-01 1.17863148e-01
-5.98340034e-01 -7.01691151e-01 -1.06743157e+00 2.93227918e-02
-5.99318862e-01 6.48704708e-01 7.97081411e-01 9.27173376e-01
3.79582435e-01 7.24974990e-01 1.32892048e+00 -4.69109058e-01
-8.82578552e-01 -7.76759386e-01 -9.17769432e-01 -6.17192686e-01
5.49425900e-01 -8.58526945e-01 -5.62932432e-01 -1.89158395e-01] | [6.3611741065979, 1.2474673986434937] |
e064ce3b-700b-4bfa-927b-3228b8f5d215 | discriminative-deep-feature-visualization-for | 2306.00402 | null | https://arxiv.org/abs/2306.00402v1 | https://arxiv.org/pdf/2306.00402v1.pdf | Discriminative Deep Feature Visualization for Explainable Face Recognition | Despite the huge success of deep convolutional neural networks in face recognition (FR) tasks, current methods lack explainability for their predictions because of their "black-box" nature. In recent years, studies have been carried out to give an interpretation of the decision of a deep FR system. However, the affinity between the input facial image and the extracted deep features has not been explored. This paper contributes to the problem of explainable face recognition by first conceiving a face reconstruction-based explanation module, which reveals the correspondence between the deep feature and the facial regions. To further interpret the decision of an FR model, a novel visual saliency explanation algorithm has been proposed. It provides insightful explanation by producing visual saliency maps that represent similar and dissimilar regions between input faces. A detailed analysis has been presented for the generated visual explanation to show the effectiveness of the proposed method. | ['Touradj Ebrahimi', 'Yuhang Lu', 'Zewei Xu'] | 2023-06-01 | null | null | null | null | ['face-recognition', 'face-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 3.98862243e-01 7.69326925e-01 -8.17617849e-02 -8.39706957e-01
4.25653309e-01 1.42620206e-01 6.20751679e-01 -3.94495040e-01
5.62022388e-01 3.74753892e-01 3.22193027e-01 -1.49202347e-01
-2.61242181e-01 -4.61471677e-01 -7.78086364e-01 -3.87300372e-01
2.05651864e-01 2.21821554e-02 -2.25598827e-01 -1.04548059e-01
7.24413097e-01 7.54785657e-01 -2.10174131e+00 6.40696406e-01
6.27734542e-01 1.02820933e+00 4.59417641e-01 1.94796816e-01
-3.52546483e-01 7.20024288e-01 -4.74707156e-01 -4.71086174e-01
3.70260775e-02 -1.00425947e+00 -9.12857592e-01 2.24673003e-01
3.63666773e-01 -3.80204171e-01 -2.18552515e-01 1.21633816e+00
1.21266417e-01 -1.70580938e-01 5.85904658e-01 -1.65327013e+00
-1.30190361e+00 3.67667556e-01 -4.58229452e-01 3.32861006e-01
5.34265995e-01 -1.19851038e-01 7.82407284e-01 -1.33627272e+00
5.25214553e-01 1.64478517e+00 2.16258913e-01 9.80298638e-01
-8.42628717e-01 -6.60272539e-01 1.64878238e-02 6.32575631e-01
-1.21559620e+00 -4.71994728e-01 9.94081318e-01 -9.21358094e-02
8.99265349e-01 3.27762574e-01 9.03102636e-01 7.89275765e-01
4.36163634e-01 6.73948824e-01 1.05059266e+00 -5.78312159e-01
4.81328405e-02 4.16005701e-01 -1.24615014e-01 8.99446130e-01
2.52348930e-01 1.48965791e-01 -7.52137780e-01 2.69316584e-01
1.02378070e+00 3.09268683e-01 -2.61258632e-01 -2.45091274e-01
-4.46691483e-01 7.84958601e-01 8.82045329e-01 5.26507437e-01
-4.94322509e-01 1.17179237e-01 -1.32788926e-01 -1.28421098e-01
4.10104036e-01 3.61776561e-01 -2.29524329e-01 4.44571823e-01
-8.22056413e-01 -5.39884344e-02 2.42043689e-01 5.31539738e-01
9.01352644e-01 4.35310811e-01 -1.71735406e-01 3.66974473e-01
7.12007642e-01 2.74706125e-01 3.37851584e-01 -8.54336083e-01
-1.31287187e-01 1.00239146e+00 -1.01904631e-01 -1.58662713e+00
-3.08492154e-01 -3.10553521e-01 -6.03458345e-01 4.85228539e-01
-8.08210224e-02 2.10311294e-01 -9.46581244e-01 1.60529792e+00
1.73546687e-01 3.12554151e-01 1.04193106e-01 1.36789525e+00
1.12325990e+00 4.63133097e-01 1.36990339e-01 9.90517288e-02
1.37133610e+00 -9.14362788e-01 -1.04362512e+00 -3.08681399e-01
3.32368538e-02 -5.09672105e-01 9.98238146e-01 -3.69595028e-02
-9.13310349e-01 -8.69882941e-01 -9.93063450e-01 -5.63175716e-02
-2.34206364e-01 2.66046643e-01 8.22675526e-01 5.06849110e-01
-1.17029452e+00 6.15817606e-01 -3.99845958e-01 -6.65250242e-01
8.58151853e-01 4.46650624e-01 -4.27019835e-01 1.34409636e-01
-9.34439838e-01 1.07818782e+00 1.57346457e-01 5.99362969e-01
-1.14188969e+00 -4.80734557e-01 -6.97114348e-01 6.41790211e-01
-1.65859818e-01 -6.46525323e-01 8.62364471e-01 -2.07879519e+00
-1.15206897e+00 8.40725422e-01 -7.44660914e-01 -2.95141637e-01
-1.32841617e-01 -1.62215859e-01 -4.15418237e-01 3.62614065e-01
-4.22361642e-02 1.08313704e+00 1.11172640e+00 -1.43931305e+00
-3.58334869e-01 -6.29281878e-01 3.74296610e-03 5.40144257e-02
-1.75567806e-01 1.05218582e-01 8.43140557e-02 -5.58144033e-01
3.11522812e-01 -4.55013067e-01 5.40197454e-03 3.24373469e-02
-3.64948988e-01 -3.14359128e-01 1.05979335e+00 -4.89023745e-01
8.07108045e-01 -2.13706350e+00 -1.01356350e-01 6.77877590e-02
3.62870842e-01 1.97197348e-01 -1.42892733e-01 1.65934622e-01
-6.79252207e-01 3.38863969e-01 7.42559657e-02 -8.49866942e-02
-2.03373939e-01 1.07583396e-01 -7.15365350e-01 3.75104100e-01
8.41993511e-01 8.73569012e-01 -5.49802780e-01 -2.39917055e-01
2.00454772e-01 7.81463623e-01 -6.01393521e-01 3.24708521e-01
-1.18604772e-01 5.10764480e-01 -5.53772509e-01 7.15763688e-01
8.32811534e-01 -1.48499683e-01 9.85975713e-02 -3.77064943e-01
-1.67147927e-02 3.68890315e-02 -7.23129451e-01 1.32480240e+00
4.31160675e-03 1.08768523e+00 -2.83992708e-01 -8.31525147e-01
1.24452937e+00 3.36568058e-01 5.67827374e-02 -7.49499083e-01
2.44297087e-01 8.62261876e-02 1.98178589e-01 -7.26155579e-01
3.28147262e-01 -4.69984442e-01 7.47509778e-01 3.84078145e-01
6.82419240e-02 4.50169444e-01 -5.27967632e-01 1.29797235e-02
7.21958995e-01 3.68984967e-01 3.20001185e-01 -3.43360513e-01
7.28121817e-01 -8.92222598e-02 3.80967051e-01 2.25366041e-01
-2.89676040e-01 7.97718883e-01 4.77981269e-01 -9.94873583e-01
-8.29475224e-01 -6.41016245e-01 -6.45447671e-02 6.65955245e-01
5.58365464e-01 -1.31183103e-01 -1.10814571e+00 -6.17048621e-01
-1.19666591e-01 8.88427377e-01 -1.09765029e+00 -4.06732947e-01
-2.20470697e-01 -2.94252187e-01 2.83970535e-01 4.81579959e-01
6.94276631e-01 -1.66153872e+00 -1.02434742e+00 -2.86108911e-01
4.34258543e-02 -8.73743892e-01 1.71726644e-01 -1.73970342e-01
-9.72619176e-01 -1.11907029e+00 -3.54444116e-01 -9.24009979e-01
1.36125064e+00 6.91353679e-01 8.48281801e-01 8.48708391e-01
-4.84745890e-01 1.14800900e-01 -2.93181151e-01 -5.64300358e-01
-1.18690372e-01 -3.99940759e-01 9.52782482e-02 2.43118942e-01
7.24512339e-01 -1.58698007e-01 -7.93885589e-01 2.61790097e-01
-8.03015709e-01 4.30071294e-01 7.23579407e-01 6.34017944e-01
2.20974192e-01 -7.19454736e-02 8.31090391e-01 -5.53989708e-01
4.27336335e-01 -5.07396400e-01 -1.66371211e-01 2.85329044e-01
-5.85577011e-01 2.53701985e-01 3.87402236e-01 5.28496057e-02
-1.35803735e+00 1.12518510e-02 -9.43812653e-02 -5.61224818e-01
-4.96344805e-01 4.81812805e-02 -2.86281347e-01 -1.14147380e-01
4.31057543e-01 2.87025303e-01 3.10569465e-01 -2.89550871e-01
2.59487778e-01 5.11967480e-01 3.18568140e-01 -1.86043769e-01
6.49071395e-01 5.27802289e-01 -6.04219548e-03 -5.48713028e-01
-8.88186514e-01 3.07880372e-01 -6.20694220e-01 -6.09043360e-01
9.91299689e-01 -5.20332694e-01 -9.26496148e-01 -7.87664279e-02
-1.58898318e+00 3.24570775e-01 -1.13409676e-01 1.94937482e-01
-5.80802143e-01 1.34567484e-01 -1.15722500e-01 -9.61999476e-01
-1.97392151e-01 -1.24084735e+00 9.65076745e-01 7.86839962e-01
-2.86909521e-01 -6.96828365e-01 -4.36791778e-01 3.91225100e-01
4.51204151e-01 -8.96556396e-03 1.12119305e+00 -5.88561296e-01
-8.53560686e-01 1.50892615e-01 -7.29735553e-01 1.10052703e-02
2.72212356e-01 1.29155040e-01 -1.44094336e+00 8.58254284e-02
1.67949989e-01 1.11785745e-02 6.35130703e-01 4.96715367e-01
1.50625324e+00 -4.68172044e-01 -4.67937171e-01 4.17178392e-01
1.34075117e+00 2.53847361e-01 1.00039554e+00 2.42325142e-01
3.97196710e-01 1.15099049e+00 7.17693746e-01 3.39538515e-01
3.39446962e-02 4.43899661e-01 1.16993475e+00 -5.14529407e-01
-3.18435758e-01 -4.06423151e-01 3.69763047e-01 -1.05735689e-01
-1.38305351e-01 -3.12373620e-02 -6.05717301e-01 4.58652109e-01
-1.76120543e+00 -1.09740472e+00 -1.53189316e-01 1.59223115e+00
8.19332600e-02 -1.19743973e-01 -3.60232532e-01 6.05360977e-02
8.54070008e-01 -4.95595597e-02 -5.59276879e-01 -8.83348942e-01
-9.44853500e-02 1.29864290e-01 -3.16805184e-01 4.70893085e-01
-5.68091452e-01 1.18369114e+00 6.64933825e+00 3.85789335e-01
-1.30079019e+00 -2.13209137e-01 9.12777245e-01 3.48861933e-01
-6.49636805e-01 2.23923430e-01 -6.77811086e-01 7.12739080e-02
6.27523065e-01 -2.79950742e-02 3.45933557e-01 1.11370170e+00
3.65093619e-01 -4.78301197e-02 -1.21982121e+00 9.12217677e-01
4.75893050e-01 -1.49527121e+00 7.69043505e-01 3.27249430e-03
3.93198788e-01 -7.23937571e-01 3.92495692e-01 -9.47250277e-02
-3.89539838e-01 -1.72893393e+00 8.88321340e-01 6.79827750e-01
5.16072869e-01 -7.75058806e-01 8.34322333e-01 1.33735314e-01
-8.88457358e-01 -2.55162001e-01 -7.75242627e-01 -4.69694465e-01
-2.82118261e-01 2.42606014e-01 -1.33054626e+00 3.18205923e-01
7.99322188e-01 5.02339780e-01 -8.04828048e-01 6.15971267e-01
-3.80765289e-01 3.14650595e-01 3.10253859e-01 3.04664969e-02
9.20427963e-03 7.64286984e-03 3.34099352e-01 6.76923931e-01
5.79160690e-01 1.29399940e-01 -5.89083254e-01 1.61650169e+00
7.14072362e-02 7.16060773e-02 -9.54563439e-01 -1.84478797e-02
1.86918810e-01 1.36646092e+00 -8.62128615e-01 -1.00500405e-01
-1.28604874e-01 1.02967191e+00 2.30218381e-01 2.48369589e-01
-9.14253831e-01 -5.61190806e-02 7.00790167e-01 4.08978075e-01
2.76709408e-01 2.14487836e-01 -6.80528820e-01 -6.55781567e-01
-1.53987706e-01 -6.05536163e-01 -1.70877412e-01 -1.34151590e+00
-7.99754441e-01 1.02823317e+00 -1.88714176e-01 -8.18951428e-01
-2.07379878e-01 -6.31537557e-01 -8.99519145e-01 9.65544105e-01
-1.48195529e+00 -1.13335621e+00 -4.16197062e-01 4.57196981e-01
8.51250947e-01 -5.03061473e-01 8.49395216e-01 -1.24336399e-01
-5.97900629e-01 3.46761495e-01 -5.96803725e-01 -1.67793751e-01
1.40073821e-01 -7.94799030e-01 1.71911985e-01 8.34174514e-01
2.60267943e-01 9.38855946e-01 1.07986152e+00 -6.39930367e-01
-1.30722475e+00 -7.86850870e-01 1.11463201e+00 -3.76628101e-01
-2.66052391e-02 -5.17698489e-02 -1.02806604e+00 4.47702616e-01
5.37190855e-01 -6.28114343e-02 6.69743419e-01 -1.54245183e-01
-1.90843895e-01 -1.20654114e-01 -1.38153386e+00 7.05554366e-01
9.04969335e-01 -3.42481315e-01 -8.14826906e-01 -4.46641035e-02
4.21680272e-01 2.17550918e-01 -2.05634981e-01 5.42241573e-01
6.05605006e-01 -1.34924376e+00 8.07531953e-01 -8.75444233e-01
8.03219140e-01 -4.87295747e-01 3.56875337e-03 -1.01215637e+00
-3.96217465e-01 -1.54765965e-02 1.17380612e-01 1.18873596e+00
2.92994589e-01 -3.17690343e-01 9.26885307e-01 5.07793844e-01
-9.54847857e-02 -8.35657060e-01 -7.85232365e-01 -1.55161321e-01
-5.90208471e-01 -9.75965932e-02 7.76805103e-01 7.29705870e-01
-1.61971077e-01 3.28347504e-01 -2.40068153e-01 3.72329265e-01
3.79620373e-01 2.81103104e-01 5.04305303e-01 -1.34238315e+00
2.20351815e-01 -3.95160884e-01 -6.34047091e-01 -5.42188406e-01
5.53339720e-01 -8.63856494e-01 -1.20289288e-01 -1.49573278e+00
5.20012140e-01 7.53914565e-02 -2.07123533e-01 7.13784158e-01
-1.14228591e-01 4.38674927e-01 2.45942220e-01 2.78933525e-01
-2.25889578e-01 8.45135391e-01 1.21278775e+00 9.51089114e-02
2.44916901e-01 -4.00079519e-01 -1.08260322e+00 9.88022029e-01
9.26760972e-01 -4.74783629e-01 -5.18628538e-01 -4.58325654e-01
5.85147133e-03 -5.27441986e-02 7.90433764e-01 -8.55443060e-01
-7.40200803e-02 -2.17406482e-01 1.03762388e+00 -4.37610418e-01
2.35921219e-01 -9.85138535e-01 3.48800540e-01 6.60487890e-01
-4.62020785e-01 1.85366482e-01 2.52391845e-01 4.12723422e-01
-3.64662439e-01 -2.53125608e-01 8.41017365e-01 3.35380959e-04
-9.51093018e-01 3.19868028e-02 -2.95091748e-01 -7.37034678e-01
1.14763522e+00 -7.51561463e-01 -2.44408801e-01 -4.11869168e-01
-6.65037930e-01 -3.14581424e-01 3.13676119e-01 6.61143839e-01
1.32346821e+00 -1.39944899e+00 -4.36097950e-01 6.33859813e-01
-1.44338682e-01 -5.00697196e-01 4.04427528e-01 5.10258317e-01
-4.79139209e-01 6.64762795e-01 -9.81551230e-01 -4.03878033e-01
-1.37990570e+00 6.25447392e-01 4.68974292e-01 4.71869886e-01
-5.00174701e-01 7.26440847e-01 6.27577305e-01 5.85677549e-02
5.66022135e-02 -1.40619781e-02 -7.12229252e-01 -3.22093666e-01
7.04980731e-01 6.52931780e-02 -2.99683094e-01 -8.65006626e-01
-5.67234516e-01 4.85586405e-01 1.56863511e-01 2.40008786e-01
1.33677793e+00 -2.72689521e-01 -3.84363770e-01 -6.31902143e-02
7.79349029e-01 -4.21940207e-01 -1.20869744e+00 3.09127837e-01
1.18283041e-01 -8.28311503e-01 -1.12506822e-01 -7.56420910e-01
-1.37113869e+00 1.29377353e+00 8.44769537e-01 9.25488397e-02
1.34229028e+00 1.19146064e-01 3.69499207e-01 -9.99848917e-02
1.51014134e-01 -4.89945859e-01 3.71004969e-01 1.14059068e-01
1.35902429e+00 -1.30448472e+00 -2.41794273e-01 -5.59880197e-01
-6.79674506e-01 1.50532341e+00 9.63426173e-01 -1.34086832e-01
4.86603975e-01 -1.55284405e-01 -2.93452106e-02 -5.92752159e-01
-7.83434808e-01 -1.47070378e-01 4.78070617e-01 6.80783451e-01
5.82211137e-01 -1.79214239e-01 -2.41141394e-01 8.84231329e-01
-1.56543106e-01 1.51266366e-01 3.35495174e-01 4.60445613e-01
-7.93159246e-01 -7.06892848e-01 -4.72075790e-01 6.99641258e-02
-5.16660452e-01 4.45432439e-02 -7.23752379e-01 6.47221625e-01
2.79480129e-01 9.22768950e-01 8.94981772e-02 -5.13272882e-01
-4.32650074e-02 6.71915933e-02 3.54866147e-01 -5.92439294e-01
-4.14696872e-01 -3.08798611e-01 -4.04090673e-01 -5.91228724e-01
-5.68788052e-01 -2.69396216e-01 -1.77719152e+00 -1.75202966e-01
-1.72713935e-01 2.44901195e-01 8.74253094e-01 1.18308008e+00
6.34960413e-01 5.04189968e-01 6.68209016e-01 -9.08857346e-01
-5.45238471e-03 -7.83249974e-01 -3.22437763e-01 3.37786436e-01
2.25481167e-01 -8.96055043e-01 -3.15035820e-01 -8.77618790e-04] | [10.315152168273926, 2.095500946044922] |
3483eea1-f1f4-406d-a929-a71c7a61e911 | detect-what-you-can-detecting-and | 1406.2031 | null | http://arxiv.org/abs/1406.2031v1 | http://arxiv.org/pdf/1406.2031v1.pdf | Detect What You Can: Detecting and Representing Objects using Holistic Models and Body Parts | Detecting objects becomes difficult when we need to deal with large shape
deformation, occlusion and low resolution. We propose a novel approach to i)
handle large deformations and partial occlusions in animals (as examples of
highly deformable objects), ii) describe them in terms of body parts, and iii)
detect them when their body parts are hard to detect (e.g., animals depicted at
low resolution). We represent the holistic object and body parts separately and
use a fully connected model to arrange templates for the holistic object and
body parts. Our model automatically decouples the holistic object or body parts
from the model when they are hard to detect. This enables us to represent a
large number of holistic object and body part combinations to better deal with
different "detectability" patterns caused by deformations, occlusion and/or low
resolution.
We apply our method to the six animal categories in the PASCAL VOC dataset
and show that our method significantly improves state-of-the-art (by 4.1% AP)
and provides a richer representation for objects. During training we use
annotations for body parts (e.g., head, torso, etc), making use of a new
dataset of fully annotated object parts for PASCAL VOC 2010, which provides a
mask for each part. | ['Sanja Fidler', 'Roozbeh Mottaghi', 'Raquel Urtasun', 'Xianjie Chen', 'Alan Yuille', 'Xiaobai Liu'] | 2014-06-08 | detect-what-you-can-detecting-and-1 | http://openaccess.thecvf.com/content_cvpr_2014/html/Chen_Detect_What_You_2014_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2014/papers/Chen_Detect_What_You_2014_CVPR_paper.pdf | cvpr-2014-6 | ['semantic-part-detection'] | ['computer-vision'] | [ 1.75424606e-01 1.59556657e-01 1.25105172e-01 -2.01363727e-01
-3.79621297e-01 -8.99333179e-01 5.04875660e-01 3.77303585e-02
-3.42775673e-01 2.40830243e-01 -9.69028473e-02 4.36927795e-01
2.90260285e-01 -6.43246770e-01 -8.45669568e-01 -6.38647676e-01
-2.58355021e-01 6.85259879e-01 9.74049985e-01 -1.72265530e-01
-3.41482520e-01 8.64835322e-01 -1.82840908e+00 3.78435761e-01
6.42671764e-01 9.00472343e-01 1.70998558e-01 6.20435715e-01
1.16163075e-01 8.02745074e-02 -7.56663322e-01 -5.82332850e-01
3.87561500e-01 -8.14658254e-02 -7.98184454e-01 4.56427008e-01
1.19745088e+00 -4.17179734e-01 -6.54945299e-02 1.01022530e+00
4.12301719e-01 3.02948733e-03 8.22768629e-01 -9.36632574e-01
-2.75768340e-01 2.34787583e-01 -9.16843057e-01 -4.53244243e-03
3.22144270e-01 2.96291411e-01 5.88398874e-01 -8.66992891e-01
9.92626488e-01 1.61621714e+00 7.02971041e-01 9.81635928e-01
-1.54896474e+00 -4.44537163e-01 3.41510832e-01 -7.35894889e-02
-1.17713904e+00 -5.02211571e-01 6.20021224e-01 -6.53822005e-01
6.46127880e-01 4.71505076e-01 6.47915781e-01 8.29778910e-01
1.30669624e-01 8.48484159e-01 7.30381310e-01 -6.00249544e-02
7.09563419e-02 -7.58092031e-02 2.02438474e-01 7.83317804e-01
4.91087228e-01 -7.10387230e-02 6.86448589e-02 -1.83448374e-01
6.38074577e-01 5.19338772e-02 -2.12459937e-01 -9.02804732e-01
-1.15680146e+00 4.06338215e-01 5.62026262e-01 1.13011532e-01
-2.89228708e-01 2.34731734e-01 2.78414696e-01 -2.27739602e-01
1.83140293e-01 1.83970705e-01 -5.49428880e-01 4.27805156e-01
-8.66269350e-01 4.02150154e-01 7.56645322e-01 1.05756950e+00
6.90498888e-01 -1.44113258e-01 -3.15196395e-01 1.04919648e+00
1.77484229e-01 3.67675573e-01 1.85135201e-01 -9.11754966e-01
3.68919462e-01 7.28600025e-01 2.06169039e-01 -8.49836469e-01
-5.77410638e-01 -3.39296430e-01 -6.96646571e-01 2.86965072e-01
5.99224627e-01 1.88155487e-01 -1.37653768e+00 1.57563174e+00
8.53377998e-01 -3.70071530e-01 -1.49595499e-01 1.00437546e+00
1.35297120e+00 3.40102613e-01 2.74049610e-01 1.72578752e-01
2.03634930e+00 -9.78234470e-01 -4.50687647e-01 -5.81394255e-01
5.65185905e-01 -5.09154797e-01 7.84177601e-01 1.79379448e-01
-1.14947927e+00 -7.71724582e-01 -8.53069246e-01 -2.12615684e-01
-5.28909028e-01 4.93889451e-01 4.80986893e-01 3.28377903e-01
-7.19538987e-01 6.60606325e-01 -9.59251642e-01 -4.72370148e-01
6.57195210e-01 4.62995738e-01 -7.20549941e-01 -3.85390798e-04
-6.10824585e-01 8.78526986e-01 4.99803632e-01 8.49646777e-02
-9.28788066e-01 -4.37134236e-01 -1.14193797e+00 -4.89997119e-02
4.65047598e-01 -7.04732478e-01 8.36225092e-01 -8.37505341e-01
-1.05118716e+00 1.39431930e+00 1.42485678e-01 -2.71872133e-01
7.79399335e-01 -2.98496455e-01 -1.10941947e-01 2.09920481e-01
-1.32762611e-01 1.18120003e+00 9.55494463e-01 -1.40979302e+00
-4.48115945e-01 -5.88317573e-01 5.99279739e-02 -1.09556911e-03
1.53153734e-02 3.43554616e-01 -8.20480406e-01 -7.52643883e-01
2.83605129e-01 -1.08188784e+00 -1.82067841e-01 5.08014977e-01
-3.81588936e-01 -9.60096344e-02 1.24277234e+00 -7.71036327e-01
8.19648385e-01 -2.04467678e+00 3.97203475e-01 -1.90379322e-01
3.11039746e-01 4.01936561e-01 -2.68400431e-01 -3.08660939e-02
-3.30821685e-02 -9.06230602e-03 -5.43084741e-01 -4.76823002e-01
-4.68555912e-02 6.00161374e-01 1.60765722e-01 6.08128846e-01
4.86687422e-01 8.38082612e-01 -6.77821755e-01 -8.37529838e-01
1.59542382e-01 6.37247801e-01 -5.25780916e-01 1.24481879e-01
-2.21030191e-01 3.75254095e-01 -2.49789163e-01 1.05922401e+00
9.47041333e-01 7.29772300e-02 -1.76781937e-02 -5.07132649e-01
8.53978992e-02 -6.13305010e-02 -1.46246827e+00 1.22198427e+00
6.92484900e-02 2.62382746e-01 4.85326499e-01 -6.64827704e-01
7.81069577e-01 4.41841036e-02 2.02223822e-01 -1.77642465e-01
8.07816982e-02 1.52653409e-02 2.46872887e-01 -5.54455340e-01
2.14655474e-01 1.51099131e-01 -4.99557890e-03 -1.36115581e-01
2.46779457e-01 -4.30823684e-01 5.48497379e-01 -6.47828877e-02
9.26162720e-01 4.51218426e-01 3.38124871e-01 -3.39801490e-01
4.52275306e-01 -2.59138286e-01 6.98209047e-01 3.83686900e-01
-2.03787208e-01 9.43845510e-01 4.14017260e-01 -7.80869246e-01
-7.29290187e-01 -8.53797674e-01 -3.53112221e-01 1.25071764e+00
1.87964797e-01 -2.93103427e-01 -7.75750339e-01 -9.20646727e-01
2.88932979e-01 1.92604244e-01 -7.91030884e-01 -4.87960689e-02
-9.14620996e-01 -7.85178423e-01 2.20763743e-01 7.59387791e-01
3.79877418e-01 -1.13551569e+00 -1.05785275e+00 1.39699638e-01
-7.80548677e-02 -1.41818047e+00 -3.37771893e-01 2.15158656e-01
-8.37357879e-01 -1.02753806e+00 -9.57710862e-01 -7.55178809e-01
9.54068124e-01 -4.02566455e-02 1.25766528e+00 1.94807023e-01
-7.51774848e-01 2.42049545e-01 -1.21368632e-01 -3.48341584e-01
-3.94202381e-01 -3.81200135e-01 -4.79285233e-02 6.31353492e-03
-3.05311501e-01 -2.28183165e-01 -6.42194152e-01 7.83644319e-01
-8.63168597e-01 4.89172759e-03 4.47002143e-01 6.25427961e-01
8.57614994e-01 -2.79046595e-01 -1.33482248e-01 -7.95498550e-01
-1.13874055e-01 -1.38925850e-01 -7.47196257e-01 2.70461887e-01
3.45757544e-01 -5.37450910e-02 4.44138944e-01 -8.91649604e-01
-7.04328835e-01 5.82863152e-01 -1.19149931e-01 -4.57829982e-01
-5.80719233e-01 -2.08075672e-01 -4.47783411e-01 -3.56030017e-01
6.95066035e-01 -1.70687094e-01 -2.82164663e-01 -8.44565332e-01
3.11863601e-01 2.45179325e-01 8.27555358e-01 -7.28102803e-01
9.09461737e-01 5.12125790e-01 1.59985036e-01 -6.76386178e-01
-6.34671748e-01 -4.77221102e-01 -1.10977995e+00 -2.82651097e-01
1.03638196e+00 -8.22991133e-01 -4.68093425e-01 3.40924859e-01
-1.26737857e+00 -3.40094060e-01 -3.36466163e-01 2.95693338e-01
-4.05978262e-01 4.84663457e-01 -6.69050455e-01 -6.28123283e-01
-3.38115871e-01 -1.09660518e+00 1.63908315e+00 2.43126437e-01
-2.79774815e-01 -5.22732198e-01 -2.57786423e-01 2.95044065e-01
1.41645342e-01 9.20649767e-01 5.81974208e-01 -4.61864114e-01
-3.48813981e-01 -3.37589145e-01 -1.21006042e-01 2.63646871e-01
1.01034567e-02 4.65579301e-01 -1.03605366e+00 -4.96639907e-01
-2.11207703e-01 -2.90034533e-01 1.05771124e+00 8.58373120e-02
1.24554372e+00 -4.48113769e-01 -6.47375584e-01 6.64387047e-01
1.06978142e+00 -1.15937352e-01 5.50015450e-01 -1.62105888e-01
9.81477797e-01 1.08438969e+00 5.97136199e-01 1.80623770e-01
1.57774702e-01 1.03153527e+00 6.82676435e-01 -2.93510735e-01
-4.45086956e-01 8.74141157e-02 2.92681962e-01 2.57767677e-01
-3.74496579e-01 -1.45409510e-01 -8.22455764e-01 6.06941998e-01
-1.68246996e+00 -7.57917881e-01 -2.97767878e-01 2.10708857e+00
7.29055882e-01 8.01152289e-02 5.44167280e-01 -6.35341629e-02
7.41264880e-01 -4.68190312e-02 -4.26380783e-01 -1.94369510e-01
-1.32257938e-01 1.00239374e-01 4.84594494e-01 9.88660678e-02
-1.49288821e+00 9.10260916e-01 6.42301941e+00 7.29312658e-01
-8.68151486e-01 -1.27531663e-01 4.16580021e-01 1.50370315e-01
1.94031700e-01 -1.69981554e-01 -1.05818892e+00 4.59379286e-01
1.67659372e-01 5.11579633e-01 -3.74876820e-02 1.03304994e+00
-3.04839164e-01 -7.74887204e-02 -1.26422668e+00 8.57052743e-01
5.51108010e-02 -7.73369551e-01 4.86596636e-02 1.67354345e-02
6.74519300e-01 -1.98837608e-01 -2.42618516e-01 1.98548332e-01
3.93528193e-02 -9.52782750e-01 9.97004092e-01 2.83012927e-01
7.04987288e-01 -2.93959498e-01 5.88017523e-01 4.43594247e-01
-1.52144253e+00 6.09303862e-02 -6.70560360e-01 2.41805732e-01
-1.77486554e-01 3.11281294e-01 -4.45676625e-01 3.43029678e-01
9.34640110e-01 4.15954351e-01 -8.41310203e-01 1.26037312e+00
-2.56452292e-01 2.54664660e-01 -6.87521875e-01 2.88000137e-01
-1.56519413e-01 -2.00960655e-02 6.95435643e-01 1.55202162e+00
7.99916778e-03 2.57991731e-01 2.43677124e-01 9.69424427e-01
-1.35190561e-01 8.15382879e-03 -3.71861398e-01 2.66878694e-01
-1.17256530e-02 1.67189395e+00 -1.04677248e+00 -5.32868147e-01
-2.19294578e-01 1.05188727e+00 2.66454488e-01 2.31986165e-01
-9.11541641e-01 -1.27838746e-01 5.13765872e-01 4.11276668e-01
6.82365716e-01 -1.55079752e-01 1.57710180e-01 -1.21871781e+00
2.38223642e-01 -8.17609012e-01 5.52257717e-01 -6.53083920e-01
-9.17352438e-01 5.89230955e-01 3.09969664e-01 -1.22536814e+00
1.99958906e-02 -7.16426849e-01 -5.80413461e-01 5.50947785e-01
-9.92490709e-01 -1.36841702e+00 -5.94321787e-01 3.72498900e-01
4.92624611e-01 5.56498230e-01 7.46055365e-01 3.60938966e-01
-4.62436080e-01 4.68282402e-01 -3.60294104e-01 2.66878605e-01
5.73087096e-01 -1.27150083e+00 4.91873145e-01 7.90280998e-01
9.85848084e-02 4.41259712e-01 7.53448725e-01 -6.20907366e-01
-1.20628524e+00 -1.14716125e+00 4.93117452e-01 -7.29528546e-01
9.36194807e-02 -8.03232908e-01 -1.08046353e+00 5.73139310e-01
-3.20101798e-01 6.93263292e-01 1.66852921e-01 -2.98133016e-01
-4.12620366e-01 -4.12120558e-02 -1.51031399e+00 5.51179409e-01
1.31888688e+00 -6.02917746e-02 -6.95231140e-01 4.08866763e-01
5.33239365e-01 -7.63617516e-01 -7.16549039e-01 9.61968601e-01
7.28779256e-01 -8.19009960e-01 1.24914527e+00 -4.80972141e-01
6.25031590e-02 -7.49415755e-01 9.12918374e-02 -9.67178226e-01
-3.16392511e-01 -3.03075165e-01 -4.77343231e-01 1.07951355e+00
-2.35614926e-02 -3.60660017e-01 5.91479123e-01 5.23014545e-01
-1.53939620e-01 -7.49496222e-01 -8.52021217e-01 -9.80581462e-01
-1.99137211e-01 6.57103732e-02 4.62143540e-01 7.45801747e-01
-4.85971063e-01 -8.85076299e-02 -2.32620597e-01 5.25589734e-02
5.38884103e-01 3.50350171e-01 1.04992545e+00 -1.39981198e+00
-3.98716450e-01 -3.97479028e-01 -7.94170737e-01 -1.04019070e+00
-2.70434946e-01 -5.87415278e-01 1.32772759e-01 -1.54907310e+00
3.16948354e-01 -1.47550628e-01 1.43254444e-01 1.00879490e+00
-2.49764130e-01 6.99902952e-01 4.20014083e-01 1.82829842e-01
-5.85273504e-01 2.29688257e-01 1.45225835e+00 -2.88940758e-01
-3.59321311e-02 4.74447682e-02 -2.95449764e-01 1.15310776e+00
4.19413239e-01 -5.90175569e-01 3.29576641e-01 -3.33218604e-01
-2.76624322e-01 -1.59532189e-01 7.36095250e-01 -1.23496473e+00
-1.69331282e-01 5.79040125e-03 8.13319504e-01 -7.52941966e-01
5.15206933e-01 -9.63882327e-01 3.84563178e-01 7.65569627e-01
5.94345443e-02 -1.60950094e-01 5.25918961e-01 3.36508065e-01
1.27654374e-02 -7.57107213e-02 1.19688082e+00 -2.15283349e-01
-5.49455702e-01 2.88418859e-01 -1.29525699e-02 4.93417354e-03
1.09725440e+00 -3.38485599e-01 -1.74518868e-01 1.35476336e-01
-1.05505681e+00 3.11654031e-01 8.08033407e-01 4.90054816e-01
3.36709529e-01 -1.17077374e+00 -7.57099688e-01 2.65159190e-01
2.92391390e-01 3.76520306e-01 1.54530272e-01 8.06113839e-01
-6.52338445e-01 -1.82220876e-01 -3.21618825e-01 -8.43080103e-01
-1.82117844e+00 6.50611460e-01 5.67367733e-01 -1.18534744e-01
-6.65637314e-01 9.75532174e-01 8.38812113e-01 -2.43053183e-01
3.87537748e-01 -8.83414567e-01 -3.69339198e-01 2.53608525e-01
4.14752424e-01 2.51204461e-01 -8.20948705e-02 -1.04277194e+00
-8.05840373e-01 1.23431635e+00 1.19289972e-01 4.83692080e-01
1.16676569e+00 2.40505189e-01 -2.00489946e-02 -1.02957888e-02
1.08157969e+00 1.32322729e-01 -1.38642085e+00 -1.47123128e-01
-1.96556032e-01 -4.29956347e-01 -6.56887472e-01 -6.80336356e-01
-1.17328179e+00 1.02130175e+00 6.08485699e-01 9.46526304e-02
9.09315765e-01 4.90535200e-01 5.23177385e-01 1.11675411e-01
2.76827127e-01 -9.66954768e-01 2.47139335e-02 3.85710031e-01
1.24267662e+00 -1.09532499e+00 2.93434381e-01 -8.26843500e-01
-4.30238605e-01 1.21962309e+00 7.58765638e-01 -2.05511749e-01
2.31508210e-01 4.45792586e-01 -1.24157801e-01 -3.96269858e-01
-4.38732624e-01 -4.69946295e-01 9.40599382e-01 6.21634007e-01
2.55392641e-01 1.92055017e-01 -4.44493771e-01 4.77899820e-01
-9.04192999e-02 -6.04925513e-01 1.19616553e-01 9.12333786e-01
-4.52379346e-01 -7.97724545e-01 -6.85316503e-01 3.23882461e-01
-6.65148556e-01 2.66327798e-01 -7.92813301e-01 8.98982823e-01
7.12401628e-01 4.30000961e-01 9.86879840e-02 -8.80745500e-02
5.71719468e-01 -4.23551835e-02 6.92663670e-01 -7.30406761e-01
-6.09364271e-01 4.01650310e-01 1.20581202e-01 -7.34678090e-01
-4.50665802e-01 -5.15266299e-01 -1.27002847e+00 1.87013268e-01
-4.40481246e-01 -3.50414097e-01 5.77470183e-01 7.12038577e-01
2.04478487e-01 5.74016988e-01 -1.58958614e-01 -1.34199142e+00
-2.74550438e-01 -8.27030003e-01 -3.24108243e-01 8.11841965e-01
2.31546223e-01 -8.49683344e-01 -2.81605184e-01 3.51139009e-01] | [9.102627754211426, 0.07874105125665665] |
b2a6f3ee-0217-4795-b819-0b51fe8b8f48 | scalable-hybrid-learning-techniques-for | 2212.10733 | null | https://arxiv.org/abs/2212.10733v1 | https://arxiv.org/pdf/2212.10733v1.pdf | Scalable Hybrid Learning Techniques for Scientific Data Compression | Data compression is becoming critical for storing scientific data because many scientific applications need to store large amounts of data and post process this data for scientific discovery. Unlike image and video compression algorithms that limit errors to primary data, scientists require compression techniques that accurately preserve derived quantities of interest (QoIs). This paper presents a physics-informed compression technique implemented as an end-to-end, scalable, GPU-based pipeline for data compression that addresses this requirement. Our hybrid compression technique combines machine learning techniques and standard compression methods. Specifically, we combine an autoencoder, an error-bounded lossy compressor to provide guarantees on raw data error, and a constraint satisfaction post-processing step to preserve the QoIs within a minimal error (generally less than floating point error). The effectiveness of the data compression pipeline is demonstrated by compressing nuclear fusion simulation data generated by a large-scale fusion code, XGC, which produces hundreds of terabytes of data in a single day. Our approach works within the ADIOS framework and results in compression by a factor of more than 150 while requiring only a few percent of the computational resources necessary for generating the data, making the overall approach highly effective for practical scenarios. | ['Sanjay Ranka', 'Anand Rangarajan', 'Scott Klasky', 'Jieyang Chen', 'Qian Gong', 'Jaemoon Lee', 'Jong Choi', 'Tania Banerjee'] | 2022-12-21 | null | null | null | null | ['data-compression'] | ['time-series'] | [ 1.32676482e-01 -1.28029272e-01 4.04147012e-03 -2.85032809e-01
-1.02308381e+00 -5.97535111e-02 7.00307488e-01 1.16280282e+00
-7.28413165e-01 6.33084834e-01 8.41078684e-02 -4.57277507e-01
4.80355462e-03 -1.05848122e+00 -1.10520279e+00 -5.97533941e-01
-9.47818384e-02 9.29878652e-01 6.76989257e-02 9.73167494e-02
5.88962376e-01 7.57219613e-01 -1.69420934e+00 4.04367536e-01
5.93876839e-01 1.25532913e+00 2.65350193e-01 1.01725471e+00
6.71753138e-02 7.98708558e-01 -5.08785605e-01 -1.50611043e-01
4.57402647e-01 -2.57642031e-01 -6.42343998e-01 -4.66590613e-01
1.03409514e-01 -4.09752548e-01 -5.57832956e-01 7.31648505e-01
5.28845251e-01 2.21312582e-01 3.73422563e-01 -9.88746345e-01
2.34744206e-01 3.88485640e-01 -3.04818869e-01 3.32076997e-01
2.28821203e-01 3.30148220e-01 6.69675648e-01 -5.38697422e-01
7.47529805e-01 7.88922191e-01 7.25509107e-01 -1.73028126e-01
-1.18421805e+00 -5.18181443e-01 -1.10627151e+00 3.34790319e-01
-1.37077117e+00 -4.84946728e-01 2.35946253e-01 -2.24601656e-01
1.46579301e+00 4.03851032e-01 1.08630037e+00 2.89148092e-01
9.62160587e-01 6.08831979e-02 7.37942815e-01 -5.89861989e-01
9.91980731e-01 -4.04046386e-01 -1.17591061e-01 1.70419022e-01
5.43230951e-01 3.80194813e-01 -7.99204767e-01 -2.96399415e-01
4.56866294e-01 -1.24003552e-01 8.12550187e-02 5.80988862e-02
-1.27015746e+00 9.66029882e-01 3.48071188e-01 2.00760178e-02
-4.76723999e-01 4.32143718e-01 6.27163589e-01 3.13658088e-01
2.70436436e-01 5.39889216e-01 -3.01452458e-01 -5.42118013e-01
-1.65213048e+00 7.45057285e-01 8.68709862e-01 9.10562694e-01
6.32600009e-01 -3.37186195e-02 -1.39084771e-01 2.06096947e-01
2.49188896e-02 5.23569226e-01 4.49132949e-01 -1.10589647e+00
1.29828319e-01 2.17545673e-01 -9.71738994e-02 -8.75524223e-01
-3.53034794e-01 -4.46065784e-01 -9.19014275e-01 3.97123486e-01
-4.46318239e-02 1.88961625e-01 -7.61511683e-01 1.10362267e+00
6.07033789e-01 1.05650179e-01 1.73367485e-01 7.87485898e-01
4.13485348e-01 1.12535453e+00 1.16793580e-01 -3.73470813e-01
1.38582206e+00 -5.95631063e-01 -6.07144773e-01 1.31855085e-01
8.30941081e-01 -9.26001787e-01 5.95355868e-01 6.78651631e-01
-1.58248818e+00 -3.58995169e-01 -1.44265378e+00 -6.22066379e-01
-8.37215409e-02 -5.15450776e-01 6.49263859e-01 3.22284490e-01
-8.57011199e-01 1.08898747e+00 -1.10736215e+00 6.68649450e-02
3.48357469e-01 4.17122394e-01 -1.44351780e-01 -1.28632858e-01
-7.31694341e-01 6.05120242e-01 7.89857209e-01 -5.92461646e-01
-7.87403226e-01 -1.67143118e+00 -6.58313274e-01 4.77097660e-01
-1.43682852e-01 -7.42124915e-01 1.33074117e+00 -1.56140864e-01
-1.10361612e+00 4.06020552e-01 8.71449336e-03 -1.18815958e+00
3.69704187e-01 -4.01967950e-02 -2.77793050e-01 6.12298787e-01
-3.49670082e-01 6.59485936e-01 5.84531963e-01 -6.22835517e-01
-6.15564644e-01 -9.23411176e-02 -6.16696060e-01 -6.02274276e-02
-4.31764200e-02 -2.67140660e-02 -3.93922657e-01 -5.71512043e-01
6.00801408e-02 -6.48777485e-01 -1.98674500e-01 6.95680231e-02
1.20436035e-01 1.34255677e-01 7.84234822e-01 -9.39578235e-01
8.80735576e-01 -1.82484329e+00 8.33726823e-02 3.40223223e-01
2.60227889e-01 2.42629889e-02 3.54500234e-01 5.13875902e-01
-8.05912092e-02 -2.21076906e-01 -6.46782517e-01 -3.08102936e-01
-2.09504932e-01 8.42388645e-02 -2.26614714e-01 5.75702429e-01
-1.37331113e-01 8.16879332e-01 -7.37147868e-01 -3.53235692e-01
2.99949884e-01 6.02257371e-01 -8.11646521e-01 4.12936240e-01
-5.38987219e-01 2.58741111e-01 -8.28839242e-02 4.99474436e-01
9.17545140e-01 -1.93644673e-01 2.13011988e-02 -4.60170031e-01
-2.13631481e-01 3.15238416e-01 -8.47923934e-01 1.96306491e+00
-3.58075649e-01 4.22868818e-01 3.74959439e-01 -1.12244606e+00
7.88239181e-01 1.26897678e-01 8.50915611e-01 -9.85096872e-01
3.05998236e-01 3.09374183e-01 -1.48865685e-01 8.31755176e-02
8.66790116e-01 -3.52590710e-01 -8.57968349e-03 7.04278886e-01
1.15465196e-02 -8.88604045e-01 4.26444411e-01 5.05966783e-01
1.45051360e+00 -1.90022886e-01 2.47847661e-01 -4.08972830e-01
1.61602855e-01 5.15092850e-01 4.60775644e-01 5.40722251e-01
1.07653722e-01 5.41473389e-01 3.01701665e-01 -4.94571716e-01
-2.19116473e+00 -4.03050125e-01 -2.91105807e-01 5.68959177e-01
-2.00469479e-01 -9.35592890e-01 -7.76585102e-01 1.77331060e-01
1.86133057e-01 1.04611182e+00 -5.02744829e-03 -2.43473157e-01
-5.83674550e-01 -7.39952922e-01 3.33052725e-01 2.82763898e-01
3.83260429e-01 -8.03950131e-01 -1.05149984e+00 1.96000576e-01
2.07170442e-01 -1.05992675e+00 -5.28994091e-02 2.84102827e-01
-1.04742765e+00 -7.60015905e-01 -7.79280141e-02 -1.15740880e-01
4.49434549e-01 -2.80141849e-02 1.29579985e+00 2.40654618e-01
-6.31745040e-01 -1.10843264e-01 -2.50604957e-01 -4.61456180e-01
-7.93726921e-01 -2.07852751e-01 7.00004771e-02 -7.51770675e-01
3.31227154e-01 -8.37139070e-01 -5.65208614e-01 -5.43460429e-01
-1.26034462e+00 3.45271558e-01 6.88389003e-01 6.68074191e-01
1.15531087e+00 3.91375810e-01 -7.56011978e-02 -5.61053574e-01
1.43559560e-01 -5.46090603e-01 -1.15504074e+00 -2.66320884e-01
-9.33512449e-01 3.08940023e-01 8.39760959e-01 9.44846421e-02
-6.53691173e-01 5.10934778e-02 -3.29282820e-01 -6.38273001e-01
2.24879235e-01 7.42317855e-01 1.58123285e-01 -2.11894557e-01
5.64424813e-01 3.19724768e-01 3.52939367e-01 -3.63875270e-01
2.01603845e-01 5.20693362e-01 1.01985037e+00 -6.35766447e-01
6.33355737e-01 4.93027210e-01 5.69311261e-01 -7.81369805e-01
-3.94272566e-01 -3.05502802e-01 -3.92634839e-01 4.19486314e-02
5.59471071e-01 -1.04094219e+00 -8.31820130e-01 1.36747479e-01
-8.81237268e-01 -1.66465819e-01 -7.99739838e-01 7.63376474e-01
-8.28069210e-01 3.42500955e-01 -6.77399814e-01 -4.32594061e-01
-9.36219275e-01 -9.97686207e-01 1.10383308e+00 1.86878757e-03
3.50096226e-02 -4.13868040e-01 7.35674724e-02 3.32319260e-01
7.06432700e-01 3.88248712e-01 9.66002345e-01 -4.84953910e-01
-9.21722054e-01 -3.63213927e-01 -1.57906041e-01 -1.74220502e-02
-6.51769698e-01 -2.75720328e-01 -5.85656941e-01 -6.53958976e-01
3.96702379e-01 -5.01691401e-01 8.02356124e-01 3.08361471e-01
1.61933982e+00 -2.17855945e-01 -1.09690316e-01 1.06543028e+00
1.61042833e+00 3.53680179e-02 8.93071771e-01 2.05140203e-01
3.41266006e-01 7.67865852e-02 4.16821510e-01 9.55810189e-01
-1.23065658e-01 5.00725150e-01 3.59885693e-01 2.54947901e-01
-2.40494341e-01 -1.30838469e-01 -4.92425673e-02 1.26128936e+00
1.96902335e-01 -1.72562003e-01 -8.74713540e-01 2.79184163e-01
-1.35093200e+00 -9.76832390e-01 -5.81947602e-02 2.54192901e+00
1.15419543e+00 1.44298747e-01 -2.98953384e-01 4.71513540e-01
2.65090793e-01 -1.88095689e-01 -6.27042294e-01 -6.71318173e-01
9.41926166e-02 7.92772889e-01 9.83175337e-01 5.33287764e-01
-6.26972139e-01 4.65923846e-01 6.59920311e+00 1.06913340e+00
-1.13480282e+00 1.35326177e-01 7.32756197e-01 -5.59717894e-01
-4.04075265e-01 1.69338018e-01 -5.44865012e-01 6.66935682e-01
2.02134371e+00 -5.78476787e-01 5.70552289e-01 9.15693343e-01
2.59460390e-01 -5.16162932e-01 -1.08867860e+00 1.22043931e+00
-1.49851516e-01 -2.05611801e+00 -7.79040605e-02 2.40505114e-01
4.19336826e-01 1.06907673e-01 -4.14594829e-01 3.79857682e-02
2.50767261e-01 -1.15931928e+00 7.76042402e-01 6.70165360e-01
1.05593872e+00 -1.19578040e+00 6.65716767e-01 4.10537511e-01
-6.64220095e-01 8.45857561e-02 -8.29508960e-01 -2.45186865e-01
3.50438714e-01 1.46190596e+00 -7.88308203e-01 5.51439404e-01
6.85933471e-01 4.21818405e-01 -3.96638989e-01 9.84715104e-01
5.38494706e-01 7.73790181e-01 -8.76416624e-01 3.14016849e-01
-6.33329377e-02 -2.11527050e-01 4.28341657e-01 1.19148612e+00
7.97099829e-01 3.40432316e-01 -8.99783447e-02 6.59357250e-01
-3.63903403e-01 -2.17604741e-01 -2.03810081e-01 -2.45443657e-01
6.09852970e-01 1.06565046e+00 -6.34964228e-01 -5.01254201e-01
-1.73590675e-01 6.85423315e-01 1.16853818e-01 -3.22821856e-01
-6.96758747e-01 -3.66804689e-01 4.43855792e-01 3.26216519e-01
5.32591879e-01 -5.18861830e-01 -7.07499206e-01 -7.62295187e-01
-1.54321253e-01 -8.99525344e-01 2.78237790e-01 -7.65200138e-01
-7.32355952e-01 1.69652089e-01 -1.37264642e-03 -9.41984594e-01
-3.98840755e-01 -3.38846415e-01 -4.32441384e-01 7.75917888e-01
-1.30363262e+00 -7.38768041e-01 -4.69882846e-01 7.82571137e-02
1.90807134e-01 -4.36881483e-01 8.92115712e-01 2.88502693e-01
-1.53619364e-01 3.02776515e-01 6.76275134e-01 -5.99671066e-01
4.28092360e-01 -8.93642843e-01 4.33521390e-01 7.80359328e-01
-2.00189382e-01 3.44614476e-01 1.19998503e+00 -1.06146932e+00
-2.28871894e+00 -1.05640328e+00 8.89664233e-01 1.47501752e-01
3.95726115e-01 -3.41968745e-01 -9.65764165e-01 2.83118367e-01
2.72593379e-01 3.45094472e-01 5.84093869e-01 -3.97929817e-01
-1.36893973e-01 -2.18492642e-01 -1.49677074e+00 -1.10846438e-01
5.79827249e-01 -4.17352319e-01 -3.42027724e-01 8.41778517e-01
7.73162484e-01 -8.21606994e-01 -1.34940314e+00 3.21533859e-01
2.35490367e-01 -9.27374303e-01 1.11114657e+00 -3.44838262e-01
1.08562160e+00 -3.82809341e-01 -2.66004533e-01 -9.35664058e-01
-2.44696051e-01 -7.26727664e-01 -7.50018477e-01 7.30102777e-01
-2.30311211e-02 1.37251362e-01 8.71767640e-01 7.14047611e-01
-2.53738493e-01 -5.17161906e-01 -1.11712921e+00 -5.75001299e-01
2.08767459e-01 -7.29556739e-01 6.88201606e-01 6.52754068e-01
1.55402981e-02 -1.01881959e-01 -2.28539497e-01 -2.93503609e-02
7.89673984e-01 1.12504765e-01 6.57250047e-01 -9.34896708e-01
-6.35685086e-01 -2.03787051e-02 -7.20050514e-01 -6.85935855e-01
-2.11339474e-01 -1.11484408e+00 -1.25356376e-01 -1.06742465e+00
2.64863074e-01 -6.03003085e-01 1.92188453e-02 2.32658610e-01
3.05719942e-01 2.94539392e-01 1.95157841e-01 5.01637876e-01
-2.78081298e-01 6.29351079e-01 6.77523851e-01 -2.34422535e-01
2.38418043e-01 -6.93977237e-01 -4.60386366e-01 1.18288927e-01
6.55125439e-01 -4.94307905e-01 -1.23279653e-01 -4.90779608e-01
3.76807004e-01 2.17274576e-01 2.76989818e-01 -1.63016224e+00
7.36526191e-01 -3.50477435e-02 6.83314145e-01 -1.02683759e+00
2.60244966e-01 -7.26521671e-01 5.65801799e-01 6.56197369e-01
-2.71574050e-01 1.72757819e-01 4.74059373e-01 4.66657609e-01
-2.04872847e-01 -4.18027788e-01 1.02470839e+00 5.04853204e-02
-3.27609956e-01 2.01932162e-01 -1.25460461e-01 -2.20546335e-01
1.09366119e+00 3.21356446e-01 -1.38443634e-02 -2.88285673e-01
-2.61050969e-01 7.55111128e-02 7.17125058e-01 -3.07928026e-01
4.32624608e-01 -1.11371469e+00 -1.02271080e+00 4.61401016e-01
-2.99277306e-01 3.12161893e-01 2.62713939e-01 5.45928836e-01
-1.40769804e+00 6.77865505e-01 -2.52978832e-01 -6.11562490e-01
-9.85747814e-01 7.75696278e-01 -2.99623907e-02 -3.42280060e-01
-1.09678364e+00 6.57874584e-01 -5.66809118e-01 2.11859383e-02
-1.39248595e-01 -3.51393640e-01 5.21784604e-01 -4.32346612e-01
1.05713141e+00 4.33668733e-01 8.48498046e-01 -2.77508885e-01
4.51607481e-02 3.03023644e-02 -1.66046262e-01 -2.41523668e-01
1.64473820e+00 3.54425877e-01 -2.34873965e-01 -8.60723853e-02
1.25369751e+00 -7.48402476e-02 -1.11850011e+00 3.22307870e-02
-3.00271511e-01 -5.01232803e-01 8.63182425e-01 -6.70720875e-01
-1.04382920e+00 6.29841685e-01 5.88605225e-01 -3.59039716e-02
1.25642109e+00 -7.73741528e-02 1.35489774e+00 4.67402488e-01
2.92570502e-01 -1.29216397e+00 -5.29538393e-01 3.36851716e-01
8.09861183e-01 -8.94839942e-01 9.03659225e-01 -2.14125231e-01
-8.31944272e-02 1.22438562e+00 -2.17843391e-02 -4.85266075e-02
6.22947395e-01 8.99128675e-01 -7.11251736e-01 -1.25273645e-01
-9.28194165e-01 6.03665709e-01 -1.62814289e-01 3.63292217e-01
2.20242932e-01 -1.18179351e-01 -6.36331379e-01 1.87560692e-01
-7.32216179e-01 3.89325291e-01 5.21246433e-01 1.43294871e+00
-6.90593779e-01 -1.13206398e+00 -6.05514109e-01 8.58361900e-01
-2.94968724e-01 -2.08559379e-01 4.30658221e-01 2.81924486e-01
-1.64474621e-02 6.09532833e-01 5.28405845e-01 -1.84336945e-01
-1.55052587e-01 -8.44851509e-03 2.80436009e-01 -8.66793245e-02
-5.16134620e-01 -9.98834372e-02 -2.39999250e-01 -8.73518765e-01
-1.64617345e-01 -7.28421211e-01 -1.66035616e+00 -1.17347276e+00
2.30648890e-01 3.64706457e-01 1.26013863e+00 8.61917436e-01
6.97488487e-01 5.52377820e-01 3.63949358e-01 -9.40581322e-01
-5.39325297e-01 -6.99791074e-01 -3.65745187e-01 4.47342128e-01
5.04931509e-02 -2.16269433e-01 -1.71939000e-01 2.73808479e-01] | [8.293669700622559, 3.117011308670044] |
68efa4f4-45be-450c-8fee-6591bb6a5414 | blindharmony-blind-harmonization-for-mr | 2305.10732 | null | https://arxiv.org/abs/2305.10732v1 | https://arxiv.org/pdf/2305.10732v1.pdf | BlindHarmony: "Blind" Harmonization for MR Images via Flow model | In MRI, images of the same contrast (e.g., T1) from the same subject can show noticeable differences when acquired using different hardware, sequences, or scan parameters. These differences in images create a domain gap that needs to be bridged by a step called image harmonization, in order to process the images successfully using conventional or deep learning-based image analysis (e.g., segmentation). Several methods, including deep learning-based approaches, have been proposed to achieve image harmonization. However, they often require datasets of multiple characteristics for deep learning training and may still be unsuccessful when applied to images of an unseen domain. To address this limitation, we propose a novel concept called "Blind Harmonization," which utilizes only target domain data for training but still has the capability of harmonizing unseen domain images. For the implementation of Blind Harmonization, we developed BlindHarmony using an unconditional flow model trained on target domain data. The harmonized image is optimized to have a correlation with the input source domain image while ensuring that the latent vector of the flow model is close to the center of the Gaussian. BlindHarmony was evaluated using simulated and real datasets and compared with conventional methods. BlindHarmony achieved a noticeable performance in both datasets, highlighting its potential for future use in clinical settings. | ['Jongho Lee', 'Dong Un Kang', 'Heejoon Byun', 'Hwihun Jeong'] | 2023-05-18 | null | null | null | null | ['image-harmonization'] | ['computer-vision'] | [ 2.27457173e-02 -1.30146578e-01 1.03366874e-01 -2.77076185e-01
-7.84379959e-01 -3.95907044e-01 3.03366661e-01 2.17566952e-01
-5.01922250e-01 5.10968745e-01 1.61245286e-01 1.51681434e-02
-1.26569107e-01 -3.27993333e-01 -6.17983222e-01 -9.10158396e-01
1.70408875e-01 3.33309203e-01 2.50249177e-01 3.58584970e-02
1.83148876e-01 3.32851559e-01 -1.05199718e+00 1.44779637e-01
1.14417076e+00 7.74981141e-01 5.26217759e-01 3.81377906e-01
-7.03448206e-02 3.25710386e-01 -8.22728455e-01 -2.54345000e-01
4.44880307e-01 -6.47543669e-01 -6.22853577e-01 2.60284394e-02
5.10532677e-01 -2.26102680e-01 -4.47458267e-01 1.32234931e+00
9.04895425e-01 2.34833658e-01 7.44986176e-01 -9.07160223e-01
-5.00015199e-01 2.86922008e-01 -4.99374032e-01 3.72627407e-01
-2.33490746e-02 1.42024964e-01 3.37769926e-01 -5.05084515e-01
5.33004344e-01 8.83629024e-01 6.72074139e-01 5.76185942e-01
-1.54494512e+00 -6.94067657e-01 -4.45944965e-01 3.13991249e-01
-1.20208347e+00 -2.15769157e-01 8.71093392e-01 -7.01257408e-01
3.88667732e-01 7.17962086e-02 6.61065400e-01 8.22151065e-01
7.03168690e-01 5.25476873e-01 1.34077001e+00 -3.37958246e-01
3.40565205e-01 2.44487494e-01 -1.55478686e-01 3.17071229e-01
7.30714202e-02 9.70524400e-02 -3.18913430e-01 -1.67626478e-02
7.42338657e-01 -3.35399002e-01 -7.65729487e-01 -6.37596011e-01
-1.13348663e+00 1.00947130e+00 7.67268538e-01 6.08917713e-01
-4.18350905e-01 -3.90629172e-01 4.01479095e-01 6.60115480e-02
1.57584816e-01 7.93839216e-01 1.72036782e-01 2.50845969e-01
-1.26565516e+00 -1.58575457e-02 4.90276903e-01 4.81321692e-01
4.82800543e-01 1.01539254e-01 -2.47127861e-01 8.48203123e-01
1.15273468e-01 3.71202499e-01 1.14901173e+00 -8.63557518e-01
1.85504153e-01 2.70687431e-01 -7.55138546e-02 -1.06396186e+00
-6.66627228e-01 -6.37554824e-01 -1.09140968e+00 3.52047473e-01
4.60514992e-01 -4.45900522e-02 -1.26602888e+00 1.54833388e+00
2.51584291e-01 2.76020020e-01 1.62210807e-01 1.32168841e+00
8.53043258e-01 4.95144606e-01 -1.07247801e-02 -2.30743065e-01
1.02299500e+00 -7.65556514e-01 -8.73189747e-01 -2.43504778e-01
1.69834092e-01 -9.82703149e-01 1.03675568e+00 4.15191889e-01
-1.13549781e+00 -6.73251033e-01 -1.33008325e+00 3.44533622e-01
-1.68487087e-01 -1.50769860e-01 9.77447480e-02 6.92048550e-01
-1.06705868e+00 4.97199535e-01 -7.83203244e-01 -1.32499754e-01
2.92253643e-01 3.49764585e-01 -4.33591694e-01 -2.13761646e-02
-1.12708461e+00 9.41041172e-01 5.62221706e-01 4.27984484e-02
-8.54490459e-01 -9.58286583e-01 -7.89478779e-01 -1.61194921e-01
-1.01301260e-01 -6.84025764e-01 1.11644220e+00 -9.83201385e-01
-1.27301455e+00 8.74792099e-01 5.53812049e-02 -6.89596593e-01
6.65086329e-01 -5.90024341e-04 -4.97989953e-01 5.51769376e-01
2.00057313e-01 5.73419392e-01 1.35465515e+00 -1.42871845e+00
-3.13312382e-01 -3.84908885e-01 -3.51325095e-01 2.42450759e-01
-2.22166419e-01 -9.48377103e-02 -5.62004626e-01 -8.17353964e-01
2.76075602e-01 -9.74731565e-01 -1.90570310e-01 -1.73489064e-01
-2.90727347e-01 4.27988112e-01 6.05276942e-01 -1.00651836e+00
1.04505563e+00 -2.30923605e+00 9.86607000e-02 4.66852605e-01
5.31964064e-01 5.26742280e-01 -5.94384670e-02 -2.23937765e-01
-3.72801691e-01 -2.26795152e-01 -5.53854704e-01 -2.30241373e-01
-5.15569806e-01 -1.23773292e-02 4.36004400e-02 9.20889974e-01
-1.84559137e-01 7.68929124e-01 -7.63112068e-01 -7.78261125e-01
5.46454489e-01 5.51528931e-01 -4.10299838e-01 4.95697588e-01
2.24128917e-01 1.07537806e+00 -9.46400538e-02 2.09025200e-02
9.19692278e-01 -1.37077346e-01 1.44514829e-01 -5.80101192e-01
1.00811101e-01 -2.60690063e-01 -1.09352279e+00 1.54097748e+00
-5.60586214e-01 7.66591311e-01 1.85770363e-01 -1.29157066e+00
8.71242166e-01 4.60301131e-01 8.48936975e-01 -1.11261606e+00
2.65569508e-01 4.22599882e-01 3.01315129e-01 -5.75854599e-01
1.88787714e-01 -3.54262739e-01 2.94395208e-01 3.06487530e-01
8.92579108e-02 -4.69959497e-01 -7.23756403e-02 -9.35417786e-02
4.96169746e-01 -4.47496414e-01 1.59877300e-01 -2.84503967e-01
5.22821665e-01 4.46844436e-02 4.19891745e-01 7.63380706e-01
-5.43524921e-01 1.02553189e+00 1.63314000e-01 -3.01570594e-01
-1.21127486e+00 -1.11522663e+00 -5.12481987e-01 2.85019130e-01
5.87257206e-01 1.07815474e-01 -1.07951975e+00 -4.66819614e-01
-2.05132589e-01 4.98690844e-01 -4.98869777e-01 -4.48822528e-01
-7.37264514e-01 -8.67549896e-01 3.35294574e-01 1.95198312e-01
7.41830230e-01 -9.53384221e-01 -7.70320296e-01 4.20057714e-01
-3.76829714e-01 -1.20626092e+00 -6.33310020e-01 1.11419380e-01
-8.19226205e-01 -9.64133024e-01 -1.27650547e+00 -8.91381383e-01
6.90524876e-01 1.93527445e-01 9.07698452e-01 -3.11490387e-01
-3.58352572e-01 3.24411899e-01 -2.05059350e-01 -8.83440152e-02
-6.54644430e-01 -1.33751407e-01 -3.05250008e-02 1.91807374e-01
-5.56512028e-02 -4.06394988e-01 -9.44564283e-01 4.33266968e-01
-1.12631702e+00 -6.75113276e-02 5.31704426e-01 1.24441886e+00
6.55883968e-01 2.18214840e-01 3.39103609e-01 -3.08438003e-01
8.39736223e-01 -2.27864906e-01 -5.54212451e-01 2.48833135e-01
-5.10046363e-01 9.96604413e-02 6.52042866e-01 -6.35879457e-01
-9.19298172e-01 -1.04859114e-01 4.18366119e-02 -6.31062150e-01
2.00328119e-02 4.44285035e-01 1.03826812e-02 -3.64234895e-01
9.28090274e-01 3.63311023e-01 5.89697242e-01 -1.91659138e-01
1.55666873e-01 6.59184098e-01 8.80731225e-01 -1.84428781e-01
5.36976814e-01 5.47649741e-01 -1.43004805e-01 -7.55797327e-01
-3.05909216e-01 -4.23967868e-01 -4.99484301e-01 -3.29761863e-01
1.17445922e+00 -5.22070169e-01 -4.18207347e-01 8.16290677e-01
-1.08629942e+00 -2.21614644e-01 -8.75979513e-02 9.17789280e-01
-5.38523614e-01 4.94332939e-01 -3.08404803e-01 -3.42280835e-01
-4.83555436e-01 -1.65986049e+00 7.23544180e-01 5.66192508e-01
-2.50443518e-01 -1.08870518e+00 -3.26235071e-02 3.40976089e-01
4.93933409e-01 1.99009761e-01 1.00485468e+00 -5.55995226e-01
-2.27994815e-01 -1.43800423e-01 -1.53594941e-01 6.41500115e-01
3.82022619e-01 -4.05970603e-01 -7.66504884e-01 -5.29580534e-01
5.56568384e-01 1.75434146e-02 5.02060652e-01 9.93136883e-01
1.02497029e+00 3.80839482e-02 -1.88431323e-01 7.83579290e-01
1.25660002e+00 6.27535582e-01 6.45165503e-01 5.57304144e-01
5.10759354e-01 6.57008946e-01 3.94049257e-01 4.63050120e-02
-4.59148586e-02 1.03301919e+00 2.05602273e-01 -5.36447465e-01
-4.43528533e-01 5.21595515e-02 -1.74143434e-01 5.73134482e-01
3.18455905e-01 6.17602766e-02 -9.94158208e-01 6.10501587e-01
-1.54074728e+00 -6.05289578e-01 5.76883033e-02 2.37649035e+00
7.06217408e-01 -6.77436031e-03 5.77974925e-03 4.75708470e-02
1.07426202e+00 -1.10184088e-01 -7.77843893e-01 -2.55027294e-01
-1.40892178e-01 2.67885596e-01 5.66763461e-01 6.00322247e-01
-1.12845707e+00 6.49903715e-01 6.38679504e+00 6.72030687e-01
-1.71013439e+00 2.76699930e-01 5.42302907e-01 1.12941913e-01
3.65640339e-03 -2.98775911e-01 -1.99728668e-01 6.40270948e-01
5.17027318e-01 -3.39047879e-01 3.55007410e-01 4.78006840e-01
3.75834018e-01 -1.85260385e-01 -7.06557930e-01 1.39489746e+00
2.10441470e-01 -1.24438202e+00 -2.81903654e-01 -2.24055916e-01
7.84824014e-01 -9.75738689e-02 3.96072179e-01 -2.32499003e-01
-2.06114575e-01 -1.10594904e+00 4.79213476e-01 3.11024070e-01
6.61795616e-01 -5.81968486e-01 7.91001499e-01 -2.33450141e-02
-7.66778290e-01 1.19910933e-01 -2.11413950e-01 7.43963659e-01
2.17432693e-01 5.06096840e-01 -1.02726471e+00 5.97156763e-01
8.50984573e-01 4.31810975e-01 -4.92316097e-01 1.44037676e+00
5.57252616e-02 3.16268981e-01 -3.10986973e-02 4.44219440e-01
1.85074374e-01 -2.34762847e-01 7.48255670e-01 9.28086817e-01
2.55842656e-01 -4.57969576e-01 2.28195265e-02 8.79407942e-01
3.07492316e-01 2.01548904e-01 -4.40173030e-01 1.83053151e-01
2.08545312e-01 9.70295966e-01 -7.07490981e-01 -2.86621541e-01
-3.29704314e-01 1.07748377e+00 -2.17705607e-01 5.22734761e-01
-9.00641501e-01 -3.58009607e-01 4.49802816e-01 1.68347791e-01
9.02555808e-02 -2.35351399e-01 -5.18132567e-01 -8.36748898e-01
4.88391519e-03 -1.06771445e+00 2.96207875e-01 -7.12102890e-01
-1.22440004e+00 9.27482486e-01 5.05896807e-02 -1.49469841e+00
-2.88138986e-01 -4.07175392e-01 -5.46089888e-01 1.11753654e+00
-1.35223365e+00 -7.63523936e-01 -5.30790210e-01 9.83820319e-01
2.01264873e-01 -4.29026872e-01 4.90514249e-01 4.42126662e-01
-2.17150852e-01 6.75772309e-01 3.35405737e-01 4.44074236e-02
9.75520074e-01 -1.11146390e+00 -6.32005408e-02 9.26107943e-01
-8.34158063e-02 4.38581079e-01 9.69503224e-01 -6.56036735e-01
-9.38655198e-01 -8.87063265e-01 3.02727044e-01 1.01082973e-01
4.01714504e-01 -1.14177689e-02 -1.04501522e+00 2.17982277e-01
3.19945902e-01 1.20193921e-02 5.84956706e-01 -5.60589969e-01
7.40721375e-02 -1.68822557e-01 -1.39854717e+00 5.10654628e-01
3.55655670e-01 -5.23854673e-01 -7.24269927e-01 1.87271833e-01
5.49085319e-01 -7.11769521e-01 -8.51215363e-01 3.53162050e-01
4.86124098e-01 -1.03706479e+00 9.32626665e-01 -1.76648423e-01
1.73328236e-01 -2.36018091e-01 2.61904616e-02 -1.78543985e+00
-1.46757469e-01 -5.42248130e-01 1.91690937e-01 7.98996747e-01
1.97436959e-01 -6.82337761e-01 6.40100598e-01 6.06862128e-01
-2.24442095e-01 -3.86367917e-01 -1.06404114e+00 -9.30037260e-01
2.71297336e-01 -1.97456494e-01 5.00442386e-01 8.21266353e-01
-2.81736374e-01 -7.19558075e-02 -2.81576842e-01 2.82221317e-01
7.00826883e-01 7.01245368e-02 5.27665257e-01 -9.11634326e-01
-2.14228719e-01 -5.80496609e-01 -6.63179100e-01 -5.56006789e-01
1.68532595e-01 -9.05397594e-01 1.55512065e-01 -1.53815114e+00
3.54056209e-02 -4.43369448e-01 -1.62500218e-01 1.02583118e-01
-1.49651319e-01 3.53136450e-01 2.06747591e-01 4.12172407e-01
1.63922206e-01 4.46686268e-01 1.74444938e+00 -4.25067961e-01
-5.41183591e-01 -2.93310208e-04 -4.99356836e-01 4.42948639e-01
8.95202398e-01 -3.92868668e-01 -4.27677661e-01 -4.10444170e-01
-4.54982489e-01 1.24875456e-01 3.19442809e-01 -1.32382965e+00
4.28498298e-01 3.52631122e-01 3.69940192e-01 -2.29862168e-01
1.86584666e-01 -9.50665653e-01 2.42887154e-01 6.88123286e-01
-1.74991786e-01 -5.26240207e-02 3.28550249e-01 1.40887454e-01
-6.33685052e-01 -3.05651784e-01 1.42510962e+00 1.56671688e-01
-6.55935705e-01 1.82314903e-01 -1.39627755e-01 1.27894118e-01
1.11775136e+00 -3.29237819e-01 -7.10154325e-02 -5.06129146e-01
-8.71696353e-01 -1.14326365e-02 3.45593482e-01 3.77842128e-01
8.78060877e-01 -1.13142312e+00 -7.27588952e-01 4.72081572e-01
3.87049350e-03 -1.75694034e-01 5.54474354e-01 1.10808265e+00
-5.90048492e-01 3.19178551e-01 -4.76078331e-01 -9.24299955e-01
-1.07740474e+00 5.82788587e-01 8.91580105e-01 -1.70137838e-01
-8.44350755e-01 4.85585004e-01 3.94915074e-01 -2.73816288e-01
1.71721920e-01 -1.47014335e-01 -3.38136673e-01 7.76767358e-02
5.39546907e-01 -1.91699155e-02 3.50631386e-01 -7.32023478e-01
-3.64921898e-01 7.20252573e-01 -1.58215433e-01 -2.70257384e-01
9.35790300e-01 -7.00212419e-02 4.48390581e-02 -5.09197414e-02
1.45667851e+00 -1.54485777e-01 -1.18383241e+00 -1.16105244e-01
-2.89829850e-01 -5.91768742e-01 3.54948103e-01 -6.54927194e-01
-1.42157149e+00 8.76861811e-01 1.35759497e+00 4.00217399e-02
1.37597048e+00 9.00521781e-03 8.04881692e-01 -3.89846921e-01
1.57669932e-01 -1.02190757e+00 2.45835438e-01 5.03605939e-02
9.39791501e-01 -1.31391859e+00 -3.68748248e-01 -1.03768356e-01
-8.72294784e-01 1.04572201e+00 5.39791763e-01 -6.13062792e-02
6.20980680e-01 2.19361693e-01 4.29837942e-01 -5.06862067e-02
3.41220766e-01 5.76764457e-02 4.92297232e-01 9.55691874e-01
1.42459303e-01 9.62397736e-03 -4.94301170e-01 3.38133663e-01
-2.28029922e-01 -2.06652936e-02 4.16490316e-01 6.53498590e-01
-6.32003024e-02 -9.41242874e-01 -8.13992858e-01 3.36579055e-01
-4.51004744e-01 -3.49740908e-02 2.01812744e-01 7.11847305e-01
1.00809276e-01 8.60178411e-01 1.98588595e-01 -2.79648870e-01
2.61500210e-01 -3.63355458e-01 5.60226619e-01 -2.63482273e-01
-4.50195014e-01 2.30540439e-01 -5.51940620e-01 -5.45935333e-01
-3.99327010e-01 -3.65112364e-01 -1.27884269e+00 -9.71357599e-02
1.32370204e-01 2.15244696e-01 8.55096042e-01 8.80689263e-01
6.11358657e-02 5.30614495e-01 7.58831263e-01 -7.28031218e-01
-3.38107795e-01 -7.39877403e-01 -7.07942784e-01 6.10223711e-01
5.74598432e-01 -6.59423232e-01 -3.49869341e-01 7.13097230e-02] | [13.69223690032959, -2.299781084060669] |
5b42fc6c-6219-4799-96a8-6e812683b7ce | fine-tuning-deep-learning-models-for-stereo | 2205.14051 | null | https://arxiv.org/abs/2205.14051v1 | https://arxiv.org/pdf/2205.14051v1.pdf | Fine-tuning deep learning models for stereo matching using results from semi-global matching | Deep learning (DL) methods are widely investigated for stereo image matching tasks due to their reported high accuracies. However, their transferability/generalization capabilities are limited by the instances seen in the training data. With satellite images covering large-scale areas with variances in locations, content, land covers, and spatial patterns, we expect their performances to be impacted. Increasing the number and diversity of training data is always an option, but with the ground-truth disparity being limited in remote sensing due to its high cost, it is almost impossible to obtain the ground-truth for all locations. Knowing that classical stereo matching methods such as Census-based semi-global-matching (SGM) are widely adopted to process different types of stereo data, we therefore, propose a finetuning method that takes advantage of disparity maps derived from SGM on target stereo data. Our proposed method adopts a simple scheme that uses the energy map derived from the SGM algorithm to select high confidence disparity measurements, at the same utilizing the images to limit these selected disparity measurements on texture-rich regions. Our approach aims to investigate the possibility of improving the transferability of current DL methods to unseen target data without having their ground truth as a requirement. To perform a comprehensive study, we select 20 study-sites around the world to cover a variety of complexities and densities. We choose well-established DL methods like geometric and context network (GCNet), pyramid stereo matching network (PSMNet), and LEAStereo for evaluation. Our results indicate an improvement in the transferability of the DL methods across different regions visually and numerically. | ['Rongjun Qin', 'Hessah Albanwan'] | 2022-05-27 | null | null | null | null | ['stereo-matching-1'] | ['computer-vision'] | [ 3.40691328e-01 -3.25777024e-01 3.03412378e-02 -4.02228385e-01
-7.43913949e-01 -4.30155605e-01 7.02474773e-01 4.51622531e-03
-5.66038787e-01 9.12849367e-01 -1.49563560e-02 -2.51553476e-01
-3.21157038e-01 -1.19237053e+00 -6.87172592e-01 -7.22099960e-01
-1.49437726e-01 5.28820872e-01 3.22970748e-01 -3.35455865e-01
3.59360158e-01 8.62393141e-01 -2.13069916e+00 1.12476617e-01
1.36094320e+00 9.94969070e-01 7.00815320e-01 1.30246714e-01
-1.08620517e-01 6.85917139e-02 -9.92006436e-02 -1.56617790e-01
6.62654817e-01 -1.49076581e-01 -5.55215299e-01 8.58218744e-02
7.38427162e-01 -4.15043592e-01 -2.13504776e-01 1.28977895e+00
5.59394836e-01 1.39725491e-01 7.65135229e-01 -9.34348106e-01
-2.44787499e-01 -9.78906266e-03 -6.79282725e-01 2.28125844e-02
3.18875819e-01 1.90321997e-01 7.48237669e-01 -7.99907684e-01
5.84156215e-01 1.15507686e+00 8.59403908e-01 9.33052972e-02
-1.22582781e+00 -6.97284877e-01 1.25791254e-02 2.86670566e-01
-1.64545608e+00 -3.00508142e-01 6.64350927e-01 -4.62391913e-01
6.65113270e-01 2.96985120e-01 6.58671379e-01 5.94196081e-01
2.24648446e-01 3.42296153e-01 1.48887134e+00 -5.20718098e-01
2.36408427e-01 2.00573564e-01 -4.12860125e-01 4.09106761e-01
3.12167317e-01 5.18105865e-01 -4.15498137e-01 -2.96508707e-02
7.75572956e-01 2.56395731e-02 -4.70981508e-01 -4.73034829e-01
-9.77669835e-01 7.58222640e-01 8.53057384e-01 4.11817133e-01
-5.48988581e-01 -2.37256050e-01 1.39398370e-02 3.32324892e-01
5.90334117e-01 2.18284369e-01 -2.99537897e-01 4.01862830e-01
-1.26986337e+00 2.55352646e-01 4.52081174e-01 6.62627399e-01
1.31552649e+00 -8.55136812e-02 5.59186712e-02 9.01865840e-01
7.39599913e-02 8.68892789e-01 3.59040141e-01 -7.82895207e-01
6.01327181e-01 8.14151704e-01 9.39267687e-03 -1.35322583e+00
-3.98620218e-01 -4.61254299e-01 -1.11070251e+00 5.71161032e-01
4.20877755e-01 6.25695735e-02 -9.46345747e-01 1.65675962e+00
2.13777378e-01 -2.20571253e-02 -7.04938918e-02 1.01678205e+00
4.74504203e-01 5.41105568e-01 1.25836849e-01 1.87039390e-01
7.84293354e-01 -2.57664114e-01 -2.97948997e-02 -4.00699019e-01
4.55381274e-01 -6.34498417e-01 9.15217936e-01 1.45441279e-01
-5.25074303e-01 -8.29775333e-01 -9.40158904e-01 2.08742410e-01
-8.07663500e-01 6.32354692e-02 6.48985922e-01 4.75196302e-01
-1.23004925e+00 8.15219164e-01 -4.54179972e-01 -8.11045170e-01
3.46147180e-01 3.60045165e-01 -5.52455068e-01 -2.42006242e-01
-1.22807860e+00 1.03093302e+00 4.67110842e-01 4.25265342e-01
-5.82155108e-01 -5.51394045e-01 -9.43621457e-01 9.07141808e-03
-9.01862308e-02 -5.25680423e-01 4.44367588e-01 -1.20623994e+00
-9.57140565e-01 1.19456697e+00 -3.36038321e-02 -3.43826801e-01
7.90820718e-01 1.11797236e-01 -2.59343922e-01 -8.16753954e-02
3.92429024e-01 1.10443473e+00 5.64598680e-01 -1.39406860e+00
-9.49807763e-01 -4.68373388e-01 1.25780061e-01 4.70968097e-01
-9.35587734e-02 -4.86845136e-01 -1.09398007e-01 -3.70105624e-01
5.83252072e-01 -8.07193935e-01 -2.93522418e-01 2.23927170e-01
-8.03895444e-02 3.46691340e-01 5.52656889e-01 -7.18475521e-01
8.62136602e-01 -2.03242421e+00 -3.73518802e-02 4.53355432e-01
-2.37028167e-01 1.96668968e-01 -2.16799617e-01 3.82149190e-01
1.03310034e-01 -1.76116869e-01 -4.98856783e-01 6.81190342e-02
-1.76933125e-01 2.55071640e-01 -1.60012156e-01 5.87613404e-01
8.70757177e-02 6.77231133e-01 -6.99861407e-01 -6.85342729e-01
7.57831037e-01 4.11436945e-01 -4.11280870e-01 1.01356320e-01
-1.16819227e-02 5.66356301e-01 -3.87074739e-01 6.49345696e-01
1.18738198e+00 8.12963024e-02 3.52933444e-02 -2.37825558e-01
-3.88698936e-01 -9.36114863e-02 -1.50082242e+00 1.47695851e+00
-7.64561892e-01 6.88741326e-01 -8.92328694e-02 -1.01161516e+00
1.07633913e+00 -2.37783790e-02 2.68016011e-01 -1.12706125e+00
-2.74204493e-01 6.98854923e-01 -5.06938435e-02 -4.16750073e-01
4.14686114e-01 -1.23761229e-01 1.59334004e-01 -5.27097993e-02
-2.90952742e-01 -3.45230609e-01 -4.70880307e-02 -3.88362527e-01
5.04199624e-01 2.37806156e-01 3.62504661e-01 -5.68016887e-01
7.87058949e-01 3.02922815e-01 4.48829204e-01 8.00970316e-01
1.53384078e-02 8.83774161e-01 -7.56305531e-02 -5.96072495e-01
-1.14701855e+00 -1.06154001e+00 -5.75181544e-01 6.67519391e-01
3.65109891e-01 5.74638903e-01 -4.64500874e-01 -2.32228518e-01
2.39712417e-01 3.39141190e-01 -4.66926783e-01 9.88338068e-02
-4.29786056e-01 -9.60420489e-01 2.85227239e-01 3.05773079e-01
1.23976362e+00 -1.02989352e+00 -6.82370305e-01 1.75304428e-01
-2.37123996e-01 -9.96198535e-01 9.07770097e-02 4.20059934e-02
-1.04261422e+00 -1.03666663e+00 -8.42300653e-01 -6.41492426e-01
6.79893255e-01 3.71293306e-01 1.11816704e+00 7.93918595e-02
-9.26539674e-02 -5.07936850e-02 -7.54736811e-02 -1.31717115e-03
-3.64570320e-01 2.56853193e-01 -1.38914958e-01 2.64157094e-02
3.83193672e-01 -7.48700142e-01 -7.07582533e-01 4.40552503e-01
-9.15447772e-01 1.12460844e-01 7.62920678e-01 7.51147389e-01
5.28255284e-01 1.06437042e-01 3.28257889e-01 -5.67677855e-01
8.53488520e-02 -4.03344303e-01 -9.84845698e-01 2.52274543e-01
-6.43568933e-01 -2.54027899e-02 3.45855653e-01 -1.01244226e-01
-1.08123350e+00 1.01867296e-01 -1.87727094e-01 -3.27369981e-02
-4.36475426e-01 6.17519200e-01 -2.23220572e-01 -5.45111895e-01
7.03224182e-01 4.45662230e-01 -1.21863753e-01 -4.26916927e-01
-4.75592613e-02 7.89919436e-01 5.27931809e-01 -4.61715281e-01
1.00611484e+00 7.64466643e-01 3.27811658e-01 -8.46489012e-01
-5.58868051e-01 -4.99393433e-01 -7.36950636e-01 -3.66621941e-01
6.32177830e-01 -1.07225978e+00 -3.36828113e-01 6.26891136e-01
-8.54802251e-01 -1.98849499e-01 5.92379197e-02 4.66462970e-01
-3.70385468e-01 4.61898863e-01 -4.91275340e-02 -7.12943077e-01
-2.66340792e-01 -1.05652106e+00 1.20613015e+00 3.31922323e-01
8.97197723e-02 -1.06903052e+00 1.13060074e-02 1.91090167e-01
7.42141962e-01 4.19944018e-01 8.18896115e-01 -1.83476418e-01
-8.54659796e-01 -1.87170714e-01 -6.46509290e-01 3.00763488e-01
3.54966372e-01 -3.25068116e-01 -1.16554296e+00 -3.34166735e-01
-2.15463832e-01 -1.37283579e-01 9.13702905e-01 7.30282605e-01
1.03579235e+00 2.07358912e-01 -3.92417043e-01 8.08153093e-01
1.95312917e+00 3.44728157e-02 8.67101729e-01 7.90970504e-01
3.98955077e-01 9.23802137e-01 7.00219989e-01 1.88607335e-01
2.25979447e-01 7.70127058e-01 5.90066850e-01 -3.85125965e-01
-1.91827834e-01 -2.03268230e-01 3.82473618e-02 2.07864642e-01
-4.21640486e-01 -5.58036007e-02 -9.61568654e-01 6.45180523e-01
-1.65887022e+00 -9.80973482e-01 -1.13914674e-02 2.65892386e+00
4.18418854e-01 1.15614533e-01 -3.09175551e-01 8.56903270e-02
8.73466372e-01 1.88837156e-01 -5.27007818e-01 8.03928301e-02
-5.62975109e-01 1.71150610e-01 9.39450264e-01 5.99199295e-01
-1.15567493e+00 6.96819484e-01 5.37672043e+00 9.82532024e-01
-1.30862701e+00 -1.90990791e-01 7.25515723e-01 3.62067491e-01
-3.66592348e-01 -3.88067104e-02 -6.84612453e-01 4.27795082e-01
5.01680493e-01 1.08552434e-01 3.94506425e-01 6.06373250e-01
4.80236143e-01 -6.05230391e-01 -8.17521155e-01 1.04794621e+00
-1.66498289e-01 -1.27225935e+00 1.21300109e-01 1.36511832e-01
9.00393069e-01 2.67595291e-01 -1.32737337e-02 1.08938411e-01
-2.73610055e-02 -9.34122682e-01 5.75392723e-01 4.59312886e-01
9.62545455e-01 -4.97777700e-01 9.80637014e-01 3.75320494e-01
-1.35251927e+00 -3.92020568e-02 -6.12156630e-01 -2.00562850e-01
-9.94623378e-02 6.56841457e-01 -3.66979003e-01 8.55082691e-01
8.40934277e-01 6.45139098e-01 -7.08017468e-01 1.32013667e+00
-1.82376727e-02 1.24050908e-01 -5.93506753e-01 2.94276178e-01
4.62787509e-01 -4.97220725e-01 3.30423534e-01 9.92115438e-01
7.15673685e-01 -2.13512361e-01 4.42571640e-02 8.79961729e-01
2.93949246e-01 1.69467092e-01 -9.18962717e-01 6.08489096e-01
5.32659173e-01 1.00605714e+00 -5.72264791e-01 -1.04966246e-01
-4.23879892e-01 8.31985235e-01 6.82949126e-02 3.61415923e-01
-5.53420901e-01 -5.51628210e-02 5.99274099e-01 3.13554555e-01
6.67076036e-02 -9.03556198e-02 -3.04738581e-01 -1.00137818e+00
1.94529489e-01 -8.09109330e-01 3.42140019e-01 -8.41843367e-01
-1.23031175e+00 5.67324817e-01 1.88459724e-01 -1.66495180e+00
-2.82781750e-01 -5.16689897e-01 -4.21435565e-01 1.31138742e+00
-1.93188214e+00 -1.17871273e+00 -7.46412098e-01 4.95703459e-01
4.37580317e-01 -1.05370373e-01 5.62592983e-01 4.21829283e-01
-9.26044956e-02 1.03028476e-01 3.74796093e-01 -9.09686834e-02
4.72399831e-01 -9.11219895e-01 2.49332249e-01 8.70424092e-01
-1.87737837e-01 1.05259590e-01 6.26114488e-01 -5.18735945e-01
-8.98638666e-01 -1.09815180e+00 9.74476695e-01 3.55068520e-02
3.34222078e-01 -1.41026616e-01 -9.59570348e-01 3.45325649e-01
-1.48213908e-01 -1.34027883e-01 -7.06803277e-02 -1.11168332e-01
-2.01504435e-02 -5.10474920e-01 -1.36911881e+00 5.18910110e-01
1.05621505e+00 -5.65977514e-01 -4.21456307e-01 1.25578269e-01
1.79988407e-02 -3.17167819e-01 -5.57645082e-01 7.60333419e-01
5.11772692e-01 -1.50025415e+00 9.83093262e-01 6.77838549e-02
3.51440042e-01 -4.18648839e-01 -6.05115891e-01 -1.23508847e+00
-2.32715800e-01 1.51411057e-01 8.03549349e-01 1.14322245e+00
3.23256165e-01 -9.97614503e-01 8.60765040e-01 4.59218830e-01
3.73590849e-02 -2.93067902e-01 -1.17701972e+00 -9.16924357e-01
1.04916506e-01 -2.12178186e-01 7.34189332e-01 7.96991587e-01
-5.46224475e-01 -2.60351568e-01 -1.67265683e-01 5.60365021e-01
8.00202847e-01 5.77402234e-01 6.64731324e-01 -1.51580107e+00
-5.56987561e-02 -3.87590200e-01 -7.22946525e-01 -8.30737591e-01
6.62120506e-02 -6.86484635e-01 2.81024016e-02 -1.50561869e+00
2.55042106e-01 -8.31750333e-01 -9.82501507e-02 1.39651984e-01
-2.55132206e-02 4.95958507e-01 -2.80750506e-02 3.81067842e-01
9.69415754e-02 5.26821196e-01 1.04543245e+00 -2.73390412e-01
-2.05926761e-01 5.19974343e-03 -1.08613938e-01 6.42654955e-01
7.18458891e-01 -3.28408778e-01 -2.91684747e-01 -5.22236645e-01
1.89928159e-01 -7.74203166e-02 5.58903098e-01 -1.50005281e+00
5.06974086e-02 -2.42837682e-01 4.86083180e-01 -6.34715855e-01
2.67936140e-01 -1.03215575e+00 6.03782177e-01 5.74473560e-01
4.03381549e-02 -1.55889004e-01 3.05133700e-01 2.29730874e-01
-5.13165832e-01 -2.10110590e-01 9.97584581e-01 -2.20917314e-01
-1.23554361e+00 3.34078610e-01 -1.19145207e-01 -2.42181778e-01
6.97905719e-01 -7.11621642e-01 -1.99647412e-01 -4.48038727e-01
-3.69016767e-01 4.46193516e-02 7.99472690e-01 2.22640052e-01
4.16028082e-01 -1.24335527e+00 -8.30573201e-01 4.07126635e-01
3.19997519e-01 2.67697386e-02 2.77814269e-01 6.30772591e-01
-7.94402599e-01 5.18934011e-01 -5.83329082e-01 -8.23993325e-01
-8.65787446e-01 3.29123497e-01 7.87681043e-01 -1.57180667e-01
-4.24777061e-01 3.41692656e-01 2.56438106e-01 -6.44890487e-01
-3.89083363e-02 -2.81055272e-01 -2.09213823e-01 4.53200340e-02
1.57751255e-02 2.36433148e-01 2.79472888e-01 -7.12884128e-01
-3.51502120e-01 1.10215557e+00 5.24924159e-01 7.54665658e-02
1.25761998e+00 -3.23815614e-01 -9.71304029e-02 1.05440840e-01
1.08411157e+00 -1.92128539e-01 -1.21903348e+00 -1.95750788e-01
4.04169373e-02 -7.21411526e-01 2.56885052e-01 -4.78977352e-01
-1.16336775e+00 7.29986429e-01 1.05626094e+00 -2.41110828e-02
1.36869109e+00 -2.10665435e-01 4.07324582e-01 3.19312304e-01
7.86461234e-01 -9.90953922e-01 -6.20480359e-01 1.72027186e-01
7.47326374e-01 -1.73506963e+00 -5.54190613e-02 -3.10972661e-01
-2.00212553e-01 1.05769062e+00 5.46392024e-01 -4.20653448e-02
6.80657148e-01 -7.84019530e-02 1.78429291e-01 -6.42334819e-02
-5.67558855e-02 -6.35026932e-01 3.24727356e-01 7.75103986e-01
2.06249699e-01 2.02360507e-02 -2.84800529e-01 -3.63131255e-01
3.84091842e-03 1.35121584e-01 2.11365834e-01 6.27485454e-01
-5.90328932e-01 -9.84585881e-01 -6.62722766e-01 4.22590405e-01
7.48541057e-02 -1.08082302e-01 9.54262838e-02 1.05220127e+00
3.90286326e-01 8.04274499e-01 2.48771742e-01 -2.26457939e-01
3.64579976e-01 -2.32806504e-01 3.97053540e-01 -3.10381889e-01
-1.35717899e-01 -3.06895196e-01 -2.23099664e-02 -5.31265974e-01
-7.92611539e-01 -7.89105415e-01 -7.53870845e-01 -4.65526760e-01
-1.10934541e-01 -1.64582536e-01 5.93161941e-01 1.03807485e+00
1.45066127e-01 -8.70436653e-02 7.62296855e-01 -1.16755950e+00
-4.04022604e-01 -9.12223697e-01 -6.79258525e-01 4.54086065e-01
2.63427883e-01 -8.92999113e-01 -5.23050487e-01 -1.69111863e-01] | [8.885817527770996, -2.2731239795684814] |
e5c4771b-0adb-4564-a180-8b758077bd28 | speech-denoising-with-auditory-models | 2011.10706 | null | https://arxiv.org/abs/2011.10706v3 | https://arxiv.org/pdf/2011.10706v3.pdf | Speech Denoising with Auditory Models | Contemporary speech enhancement predominantly relies on audio transforms that are trained to reconstruct a clean speech waveform. The development of high-performing neural network sound recognition systems has raised the possibility of using deep feature representations as 'perceptual' losses with which to train denoising systems. We explored their utility by first training deep neural networks to classify either spoken words or environmental sounds from audio. We then trained an audio transform to map noisy speech to an audio waveform that minimized the difference in the deep feature representations between the output audio and the corresponding clean audio. The resulting transforms removed noise substantially better than baseline methods trained to reconstruct clean waveforms, and also outperformed previous methods using deep feature losses. However, a similar benefit was obtained simply by using losses derived from the filter bank inputs to the deep networks. The results show that deep features can guide speech enhancement, but suggest that they do not yet outperform simple alternatives that do not involve learned features. | ['Josh H. McDermott', 'Yang Zhang', 'Kaizhi Qian', 'Jenelle Feather', 'Andrew Francl', 'Mark R. Saddler'] | 2020-11-21 | null | null | null | null | ['speech-denoising'] | ['speech'] | [ 3.61980170e-01 5.84117174e-02 7.30074525e-01 -5.06412327e-01
-1.10874069e+00 -4.08047080e-01 5.72739005e-01 -7.09380135e-02
-5.45190036e-01 5.44110477e-01 6.23179615e-01 -4.85684425e-02
-1.10477142e-01 -6.99495435e-01 -7.05019772e-01 -7.19980359e-01
-1.79926053e-01 -1.55783147e-01 9.63188782e-02 -4.54401702e-01
-2.66704417e-04 3.83998096e-01 -1.93001306e+00 6.12338066e-01
5.30846000e-01 1.15623868e+00 2.38888726e-01 9.63468790e-01
1.83662083e-02 5.58746934e-01 -1.12080944e+00 -3.50656599e-01
2.51693010e-01 -6.66229725e-01 -3.42089117e-01 -3.77449840e-01
4.96093631e-01 -4.92245376e-01 -5.44208705e-01 1.12205875e+00
9.09458697e-01 5.06075263e-01 3.91415030e-01 -7.21144140e-01
-7.41560161e-01 5.18843114e-01 2.70201284e-02 4.44735765e-01
3.70703369e-01 5.79042695e-02 1.00484467e+00 -1.03817368e+00
3.64872158e-01 1.27378523e+00 1.21927094e+00 6.38371587e-01
-1.42280686e+00 -6.39541030e-01 -1.66152090e-01 2.61604905e-01
-9.24308240e-01 -1.25892425e+00 5.29370248e-01 -5.56817092e-02
1.34213114e+00 3.80729556e-01 5.61100483e-01 9.54848766e-01
8.45783427e-02 3.76889259e-01 9.51613605e-01 -5.59330463e-01
-7.53137050e-03 1.58831879e-01 -4.02385205e-01 2.83206880e-01
-2.80189693e-01 6.59091532e-01 -6.77871108e-01 -1.28913686e-01
3.62495661e-01 -5.26460409e-01 -6.15445137e-01 3.46295029e-01
-7.74515927e-01 6.74952388e-01 4.46424633e-01 6.05104387e-01
-5.60138762e-01 2.23815084e-01 4.44638073e-01 7.65553772e-01
4.97550339e-01 6.84908926e-01 -4.70520347e-01 -2.17951789e-01
-1.10245931e+00 8.93902630e-02 8.07564795e-01 1.22603245e-01
5.73808610e-01 7.60670960e-01 -9.84174907e-02 1.15678382e+00
1.31454453e-01 2.99194843e-01 5.89380622e-01 -1.26970351e+00
2.34203115e-01 -3.89473915e-01 -1.36408299e-01 -9.29048359e-01
-1.61216632e-01 -5.08130848e-01 -6.01166785e-01 7.08062649e-01
2.83751994e-01 -3.32163692e-01 -1.08773398e+00 1.76136875e+00
-8.56743902e-02 1.35383487e-01 1.30541667e-01 9.34495091e-01
5.76317072e-01 8.47763002e-01 -9.20548961e-02 -6.47259457e-03
7.38732815e-01 -4.94593233e-01 -9.65654969e-01 -1.58961147e-01
1.73920751e-01 -1.20120561e+00 1.05954885e+00 7.91674733e-01
-1.39643395e+00 -8.32274199e-01 -1.13301742e+00 -1.09517612e-01
-4.04981822e-01 -1.74359038e-01 2.86113799e-01 9.38138723e-01
-1.60158384e+00 1.14761639e+00 -6.55672848e-01 7.34622553e-02
3.50691736e-01 4.59343553e-01 -4.21395808e-01 1.07663788e-01
-1.37879705e+00 1.00781322e+00 6.20987415e-02 1.55137450e-01
-1.10976887e+00 -8.63058388e-01 -8.85166883e-01 4.04244512e-01
-2.76589990e-01 -4.78908360e-01 1.45200503e+00 -1.35951769e+00
-1.81442273e+00 3.72918218e-01 -4.79427204e-02 -6.45147622e-01
1.36424199e-01 -3.99051696e-01 -7.16226399e-01 3.47282082e-01
-2.25362524e-01 7.63301969e-01 1.32699013e+00 -1.09706926e+00
-6.16060674e-01 1.88393563e-01 -2.56051958e-01 7.06177612e-04
-4.13852781e-01 2.05063194e-01 1.97713599e-01 -8.67910445e-01
-1.58707015e-02 -2.71457106e-01 -3.06096394e-02 1.73968524e-01
-9.85127985e-02 1.51574239e-01 7.16701150e-01 -1.28801537e+00
8.30382824e-01 -2.55772996e+00 -2.09206402e-01 2.32436001e-01
-4.56276610e-02 7.27344930e-01 -5.87332129e-01 3.92051727e-01
-3.86581928e-01 1.56744823e-01 -1.23392493e-01 -5.11630058e-01
1.50042996e-01 -7.21252710e-02 -6.18356824e-01 4.20848668e-01
4.98117656e-01 3.38686317e-01 -7.16416836e-01 7.63157606e-02
2.56954521e-01 1.22814655e+00 -6.81312084e-01 2.46440262e-01
1.01431049e-01 7.93891996e-02 3.40577275e-01 1.29337639e-01
5.21809459e-01 6.82951450e-01 -3.04063976e-01 -2.48455137e-01
-1.63487852e-01 1.03323221e+00 -9.18922484e-01 1.43596745e+00
-6.25539064e-01 1.21784949e+00 7.75424898e-01 -9.72513676e-01
1.08245099e+00 7.76196420e-01 2.62925208e-01 -9.63973522e-01
8.44260752e-02 3.27139884e-01 2.54401565e-01 -4.40196544e-01
2.01871276e-01 -7.48033941e-01 5.21083832e-01 2.74534017e-01
4.42066848e-01 -5.84583819e-01 -2.55314946e-01 -2.62220591e-01
1.24436128e+00 -5.58099709e-02 -3.51462781e-01 5.03387563e-02
2.62734801e-01 -4.23990130e-01 3.27842772e-01 8.81584346e-01
-2.57144749e-01 1.00177097e+00 6.44918382e-02 -1.46833539e-01
-1.02673256e+00 -1.28421152e+00 -2.23038718e-01 1.29435420e+00
-5.02731085e-01 -2.92618454e-01 -8.02227914e-01 -1.46699414e-01
-1.84128135e-01 8.87576401e-01 -3.81897300e-01 -6.22114658e-01
-4.53230649e-01 -1.96019515e-01 1.01468372e+00 4.21567082e-01
8.71136039e-02 -1.19311237e+00 -2.34573871e-01 5.99759996e-01
-1.33716121e-01 -7.30166852e-01 -4.60847020e-01 8.10056925e-01
-7.38963783e-01 -4.74345058e-01 -7.95870602e-01 -9.73370492e-01
1.94245890e-01 6.24557957e-02 8.68001819e-01 1.50556758e-01
-2.77468532e-01 4.06887621e-01 -2.32152075e-01 -5.80742419e-01
-6.74105763e-01 -2.30534121e-01 3.55076700e-01 -1.68269470e-01
2.11978510e-01 -9.90952015e-01 -4.03617889e-01 -9.31557491e-02
-9.43237126e-01 -6.20918512e-01 5.38671553e-01 1.02800512e+00
2.21901521e-01 3.79969537e-01 1.15926373e+00 -1.74257189e-01
1.01670551e+00 -1.18639149e-01 -1.64142981e-01 -2.61105865e-01
-2.07621053e-01 7.85585120e-02 7.81396985e-01 -4.49144453e-01
-1.35802984e+00 -1.33852184e-01 -1.14822102e+00 -3.08158666e-01
-4.43774194e-01 3.79994959e-01 -1.08166367e-01 -2.82216650e-02
9.64234829e-01 3.34687494e-02 1.76954627e-01 -7.94877291e-01
2.35914081e-01 8.11524689e-01 8.91262054e-01 -2.69110709e-01
8.04635465e-01 1.14265397e-01 -4.61443126e-01 -1.08169889e+00
-5.33221781e-01 -7.92248696e-02 -2.50471145e-01 4.06143144e-02
6.18967175e-01 -8.19657445e-01 -3.79120260e-01 2.87123054e-01
-1.24667478e+00 -3.70427698e-01 -7.63364136e-01 7.46412337e-01
-5.20096838e-01 1.86254323e-01 -9.57226336e-01 -8.43026638e-01
-2.97287494e-01 -8.40955973e-01 7.36889958e-01 -4.95093782e-03
-3.09681594e-01 -9.07169044e-01 1.72998622e-01 -2.75677275e-02
9.56268787e-01 -2.10130870e-01 1.02632678e+00 -5.83706141e-01
-3.73423696e-02 -2.19473124e-01 5.60438596e-02 1.13176811e+00
4.61026669e-01 4.65344302e-02 -1.71831191e+00 -2.48145491e-01
4.12821442e-01 -4.31825101e-01 1.08981287e+00 6.12791717e-01
9.04617548e-01 -2.79599011e-01 2.20991805e-01 8.45145464e-01
1.04456365e+00 3.08133215e-01 9.85361874e-01 2.11679503e-01
9.39260200e-02 6.41157687e-01 3.11779846e-02 -3.27203353e-03
-4.53461528e-01 3.56377900e-01 2.22158685e-01 -3.86713117e-01
-7.32246101e-01 -2.90589273e-01 6.26744866e-01 1.05069900e+00
1.97546706e-01 -4.99988114e-03 -5.10937035e-01 6.97099507e-01
-1.01997972e+00 -1.07626235e+00 2.27949291e-01 1.98663294e+00
1.11420298e+00 4.00480747e-01 -1.22432262e-01 6.39226377e-01
5.60436308e-01 7.85481781e-02 -2.57754266e-01 -8.36012483e-01
-2.31053606e-01 1.15956151e+00 -6.65535405e-02 7.89371848e-01
-1.01601684e+00 6.88520074e-01 7.42104578e+00 6.33752525e-01
-1.20261705e+00 6.46602437e-02 3.61654639e-01 -3.00398558e-01
-3.49696577e-01 -4.85575020e-01 -2.43769139e-01 4.61013727e-02
1.49901772e+00 -8.52553826e-03 6.56282783e-01 4.22011673e-01
4.77026761e-01 2.47587487e-01 -1.04561841e+00 6.59870923e-01
-1.01698868e-01 -1.08729959e+00 -8.01783353e-02 -2.30526790e-01
5.55260360e-01 2.37054974e-02 4.51465100e-01 3.12591851e-01
1.92292765e-01 -1.29280496e+00 7.89408028e-01 5.07233739e-01
7.62508392e-01 -8.01870227e-01 6.01626396e-01 1.21401750e-01
-9.02861118e-01 -6.31207824e-02 -6.14019990e-01 -3.41231525e-01
6.63752109e-02 6.02491975e-01 -1.09045863e+00 -2.30896543e-03
9.01355207e-01 3.47430527e-01 -2.24638596e-01 1.29084074e+00
-9.97297764e-02 9.52748775e-01 -3.82294625e-01 2.95564950e-01
1.20768487e-01 1.51957005e-01 6.16772532e-01 1.44208562e+00
5.04003584e-01 -1.77265525e-01 -5.61086178e-01 7.33204842e-01
-1.15015842e-01 -1.78661987e-01 -7.69929707e-01 -1.39673382e-01
2.85859853e-01 8.71576667e-01 -2.82949358e-01 -9.27997157e-02
-3.01077932e-01 7.85025537e-01 -2.77228896e-02 6.07990682e-01
-4.32215750e-01 -9.55965400e-01 1.17421615e+00 5.31117432e-02
6.07881069e-01 1.80635822e-03 -1.96040273e-01 -6.41308486e-01
-4.12753671e-02 -1.08895504e+00 -1.97005123e-01 -1.02913797e+00
-1.32951868e+00 9.84521747e-01 -5.22672296e-01 -8.24353755e-01
-4.33168828e-01 -4.84734535e-01 -7.71329105e-01 1.30503786e+00
-1.51454282e+00 -4.99053448e-01 9.02138203e-02 5.50929606e-01
4.16709065e-01 -1.69715479e-01 1.16941524e+00 6.55860007e-01
-2.89120860e-02 6.09501183e-01 2.81252295e-01 -6.17065206e-02
8.80336881e-01 -1.24984038e+00 6.39477670e-01 8.27886522e-01
4.03207839e-01 4.77785170e-01 9.81039226e-01 -2.61470884e-01
-8.11696231e-01 -8.06904018e-01 7.89620638e-01 -2.42263116e-02
3.95797074e-01 -4.06144202e-01 -1.26233077e+00 3.70309204e-01
6.45348549e-01 -8.98994729e-02 6.03097200e-01 -1.08031631e-01
-3.31048340e-01 -3.78029168e-01 -1.08904088e+00 2.81240821e-01
8.38047981e-01 -9.44104552e-01 -9.54684556e-01 -1.62604734e-01
8.90758932e-01 -8.77571106e-02 -5.25734663e-01 1.05932020e-01
4.49298263e-01 -1.09211385e+00 1.29671931e+00 -7.53545046e-01
2.81716257e-01 -5.04769385e-02 -3.11545879e-01 -1.88088071e+00
-3.06000680e-01 -6.60226822e-01 1.47241652e-01 1.35129702e+00
5.30019999e-01 -6.14123344e-01 5.47517598e-01 3.26617151e-01
-6.91889167e-01 -2.75779426e-01 -1.01055670e+00 -6.17745757e-01
2.57113993e-01 -5.85112214e-01 5.45373201e-01 5.27698517e-01
-2.71478295e-01 2.59039979e-02 -1.49893075e-01 1.27007827e-01
3.23592216e-01 -6.13559961e-01 1.77036867e-01 -1.15126610e+00
-2.84252673e-01 -6.14096880e-01 -4.60157722e-01 -7.99252510e-01
2.35486940e-01 -7.45489299e-01 6.40147567e-01 -1.51731801e+00
-6.69353187e-01 -2.19153538e-01 -4.62001532e-01 4.84714448e-01
-3.52566130e-02 4.43382710e-01 1.45541489e-01 -3.88098121e-01
1.67801410e-01 7.45921254e-01 9.59646940e-01 -2.10531563e-01
-2.91843235e-01 1.63866922e-01 -8.04535568e-01 7.30156124e-01
9.27795887e-01 -6.20642722e-01 -2.49458984e-01 -6.85130656e-01
1.89125314e-02 -1.73523203e-01 4.42196786e-01 -1.34286380e+00
2.72221595e-01 3.79320920e-01 6.81025505e-01 -1.87431484e-01
8.14819813e-01 -8.05882215e-01 9.13604498e-02 3.83766085e-01
-6.42885864e-01 -3.85308474e-01 6.18864536e-01 3.39281797e-01
-6.73023522e-01 -4.91616130e-01 9.69899416e-01 4.54535671e-02
-3.41740906e-01 -3.39219809e-01 -8.95885170e-01 -8.35958645e-02
2.39812136e-01 -1.86772689e-01 -1.13272764e-01 -8.64394724e-01
-1.06043851e+00 -6.33902729e-01 -3.01334318e-02 2.55529404e-01
9.02300298e-01 -1.28916168e+00 -7.33508170e-01 6.22002780e-01
-7.63931096e-01 -4.51041579e-01 7.28881881e-02 4.66213793e-01
-4.29875344e-01 2.33154774e-01 -2.49071345e-01 -2.64288694e-01
-1.14432895e+00 2.41944149e-01 8.73477459e-01 3.93327236e-01
-6.89760506e-01 1.20777798e+00 8.76753330e-02 -4.11695212e-01
6.40927672e-01 -4.23201263e-01 1.00027442e-01 -2.80222129e-02
7.75428951e-01 2.67830521e-01 5.92491627e-01 -4.13735300e-01
-1.60589829e-01 1.32315665e-01 1.65306911e-01 -5.91011524e-01
1.88512397e+00 -1.53291738e-02 2.34913722e-01 2.08293661e-01
1.52774763e+00 1.57547310e-01 -1.43954480e+00 -1.39314145e-01
-1.81515038e-01 -5.25294304e-01 5.90418756e-01 -9.93638158e-01
-1.22757328e+00 1.30851507e+00 8.91455054e-01 4.34652865e-01
1.50840080e+00 -2.95286387e-01 9.02501583e-01 3.55690926e-01
-1.03880815e-01 -1.00287437e+00 2.53731817e-01 5.37822008e-01
1.11556745e+00 -8.45370352e-01 -4.29090142e-01 -5.53977825e-02
-2.45339707e-01 1.34251547e+00 2.70839989e-01 -2.23628640e-01
8.12318087e-01 6.51612997e-01 5.01825452e-01 5.44588044e-02
-6.35562599e-01 -3.02329212e-01 3.08232427e-01 1.14475775e+00
5.43984652e-01 -3.54847401e-01 3.63365471e-01 7.11455524e-01
-6.34747386e-01 -1.41175270e-01 3.25650066e-01 7.11243749e-01
-7.96130240e-01 -1.09447598e+00 -7.36655176e-01 6.54154718e-01
-6.78665757e-01 -3.52073371e-01 -3.63709241e-01 4.57107574e-01
2.11304486e-01 1.24132931e+00 7.99643174e-02 -4.87583727e-01
5.97226083e-01 3.23563635e-01 3.74598593e-01 -5.96627712e-01
-9.71780062e-01 3.58900189e-01 1.66173324e-01 -2.89654911e-01
-1.50699392e-02 -6.11088514e-01 -1.02387929e+00 -1.76750302e-01
-4.75698411e-01 2.15606436e-01 7.74738908e-01 7.41970956e-01
1.31941304e-01 8.39529574e-01 6.76418364e-01 -1.31917155e+00
-8.74650300e-01 -9.68538702e-01 -6.13315046e-01 2.73375124e-01
1.07669902e+00 -2.38324091e-01 -8.17425668e-01 3.49501848e-01] | [15.146655082702637, 5.8433003425598145] |
56032dd0-9a20-4f8e-bb0e-73bff3dd8542 | rethinking-the-design-principles-of-robust | 2105.07926 | null | https://arxiv.org/abs/2105.07926v4 | https://arxiv.org/pdf/2105.07926v4.pdf | Towards Robust Vision Transformer | Recent advances on Vision Transformer (ViT) and its improved variants have shown that self-attention-based networks surpass traditional Convolutional Neural Networks (CNNs) in most vision tasks. However, existing ViTs focus on the standard accuracy and computation cost, lacking the investigation of the intrinsic influence on model robustness and generalization. In this work, we conduct systematic evaluation on components of ViTs in terms of their impact on robustness to adversarial examples, common corruptions and distribution shifts. We find some components can be harmful to robustness. By using and combining robust components as building blocks of ViTs, we propose Robust Vision Transformer (RVT), which is a new vision transformer and has superior performance with strong robustness. We further propose two new plug-and-play techniques called position-aware attention scaling and patch-wise augmentation to augment our RVT, which we abbreviate as RVT*. The experimental results on ImageNet and six robustness benchmarks show the advanced robustness and generalization ability of RVT compared with previous ViTs and state-of-the-art CNNs. Furthermore, RVT-S* also achieves Top-1 rank on multiple robustness leaderboards including ImageNet-C and ImageNet-Sketch. The code will be available at \url{https://github.com/alibaba/easyrobust}. | ['Hui Xue', 'Yuan He', 'Ranjie Duan', 'Shaokai Ye', 'Xiaodan Li', 'Yuefeng Chen', 'Gege Qi', 'Xiaofeng Mao'] | 2021-05-17 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Mao_Towards_Robust_Vision_Transformer_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Mao_Towards_Robust_Vision_Transformer_CVPR_2022_paper.pdf | cvpr-2022-1 | ['robust-design'] | ['miscellaneous'] | [-1.77600518e-01 -3.73398334e-01 1.05592854e-01 -8.42265859e-02
-4.82616425e-01 -5.38788736e-01 6.08300924e-01 -4.89197999e-01
-2.64015228e-01 4.25210625e-01 2.19377279e-01 -2.88459450e-01
-8.95765126e-02 -4.89466816e-01 -9.63599384e-01 -7.32390881e-01
2.10139811e-01 -3.15064877e-01 5.37234783e-01 -5.40133953e-01
1.72495455e-01 4.64476436e-01 -1.20679033e+00 -1.68668795e-02
1.03735709e+00 1.04043305e+00 5.54227270e-02 6.08220935e-01
4.16139632e-01 1.03486907e+00 -6.43091381e-01 -6.78709209e-01
6.41257584e-01 -3.75069454e-02 -5.58813930e-01 -3.26871902e-01
7.48069048e-01 -3.37760061e-01 -1.10407984e+00 1.17681682e+00
8.92374814e-01 1.56688660e-01 4.10024166e-01 -1.58261812e+00
-1.29607141e+00 6.49624228e-01 -6.68858230e-01 4.27811980e-01
4.37088497e-02 7.89284885e-01 6.49508059e-01 -8.42934787e-01
2.23657310e-01 1.39221549e+00 9.65452969e-01 7.07216382e-01
-9.70144272e-01 -1.09739065e+00 4.33977604e-01 5.88700533e-01
-1.29961264e+00 -5.76829255e-01 7.12438762e-01 -1.60551012e-01
1.02302492e+00 2.35298052e-01 2.24973679e-01 1.42404032e+00
8.30137208e-02 8.44344676e-01 1.10766947e+00 1.39715686e-01
-3.14541310e-02 -1.58873260e-01 1.94617316e-01 6.36768997e-01
2.78966784e-01 3.61158550e-01 -2.12472364e-01 3.55654299e-01
8.22397172e-01 4.47850302e-02 -5.43692768e-01 -4.08897810e-02
-1.09137321e+00 6.02111816e-01 1.06392491e+00 -1.13654528e-02
-1.36367902e-01 4.99429494e-01 6.90549374e-01 3.73230308e-01
3.64272743e-01 4.19875443e-01 -4.72901404e-01 1.86140150e-01
-5.77947557e-01 1.16349407e-01 2.66022831e-01 9.09952164e-01
3.90310764e-01 5.58535397e-01 -4.70645875e-01 9.93787706e-01
2.82568652e-02 5.93274653e-01 5.34381986e-01 -7.50376642e-01
3.95188212e-01 6.18066072e-01 -3.10868233e-01 -9.94269788e-01
-3.17575634e-01 -6.37578547e-01 -1.31859064e+00 3.49368125e-01
6.11079410e-02 5.06340712e-02 -1.37567496e+00 1.84928095e+00
1.29374832e-01 5.96194267e-01 1.19395323e-01 7.82429278e-01
1.35916293e+00 4.19537395e-01 -1.21712409e-01 2.50215828e-01
1.15508962e+00 -1.32672918e+00 -5.32855392e-01 -2.44437959e-02
1.21231759e-02 -8.20018113e-01 1.27029419e+00 2.66644567e-01
-1.06717694e+00 -9.21592236e-01 -1.12690747e+00 -6.40524849e-02
-5.29743016e-01 -2.56892806e-03 3.79327059e-01 6.12022161e-01
-1.12641752e+00 7.34614015e-01 -6.88787639e-01 -2.91183144e-01
7.78517067e-01 2.82503009e-01 -5.36364734e-01 -6.58338368e-02
-1.24036884e+00 1.03284812e+00 1.47949994e-01 1.90282166e-01
-1.30427003e+00 -8.52047026e-01 -8.83672833e-01 -1.36338115e-01
4.94801015e-01 -8.52036536e-01 1.18248844e+00 -6.72128916e-01
-1.55544519e+00 3.65618616e-01 1.68817759e-01 -5.83886325e-01
6.46642387e-01 -4.39543605e-01 -3.60056758e-01 4.75723781e-02
-5.81282154e-02 7.52849758e-01 1.29750216e+00 -1.27611625e+00
-2.02459082e-01 -7.31039746e-03 2.32941866e-01 -1.25566468e-01
-5.21050274e-01 1.49881884e-01 -8.38603258e-01 -1.00410974e+00
-1.71846643e-01 -8.51342440e-01 -1.92754716e-01 -7.81500489e-02
-7.28098392e-01 -2.47545972e-01 1.03938305e+00 -6.19986892e-01
1.14287376e+00 -2.04593706e+00 1.09903991e-01 -1.53704211e-02
3.36293608e-01 9.23401117e-01 -5.98516464e-01 1.98838368e-01
-3.58186036e-01 4.33101863e-01 -6.92451745e-02 -3.29161495e-01
8.66167247e-02 1.74708113e-01 -4.36812282e-01 5.80895782e-01
3.81630182e-01 1.16815317e+00 -6.86897814e-01 -3.24398786e-01
3.88190746e-01 6.37206793e-01 -4.30784315e-01 1.11539871e-01
1.68926015e-01 3.88390899e-01 -2.38390163e-01 8.86246264e-01
8.69580448e-01 -3.30810063e-02 -6.28346980e-01 -5.41713953e-01
-5.42249791e-02 -1.12355426e-01 -9.29386735e-01 1.28496635e+00
-2.94565201e-01 6.10521793e-01 -2.35374644e-01 -6.72335625e-01
7.19018042e-01 2.07524866e-01 2.83575859e-02 -8.54300857e-01
2.09129661e-01 -7.34801441e-02 4.26322371e-02 -4.75361049e-01
3.91614050e-01 4.57636148e-01 2.74918884e-01 1.76980551e-02
1.73389554e-01 -3.46760005e-02 2.21249133e-01 4.19683009e-01
1.17977357e+00 1.47596762e-01 -1.07988611e-01 -1.91293970e-01
4.60537881e-01 -5.03472686e-01 7.74237156e-01 1.02133405e+00
-4.46935415e-01 8.62107873e-01 4.73937869e-01 -3.75146776e-01
-9.38977301e-01 -9.89036441e-01 3.59281562e-02 9.39880788e-01
2.53147751e-01 -3.32914531e-01 -8.44136238e-01 -7.44113803e-01
-3.19089182e-02 3.76741499e-01 -8.82764161e-01 -5.29669106e-01
-4.82432187e-01 -8.28939974e-01 9.74747300e-01 8.32093596e-01
1.10293317e+00 -1.08658075e+00 -2.88621068e-01 -2.79784083e-01
-1.86886460e-01 -1.11188281e+00 -7.79968679e-01 -1.29255384e-01
-5.01352906e-01 -1.14923239e+00 -6.05460703e-01 -5.85881293e-01
5.66450536e-01 6.77219689e-01 9.29162741e-01 2.78530836e-01
-2.50404984e-01 3.16112697e-01 -5.25189340e-01 -5.68695724e-01
-1.74302861e-01 8.05942640e-02 3.80443633e-01 -6.70092851e-02
-4.03569490e-01 -6.73543274e-01 -9.04255867e-01 5.72858751e-01
-1.09770858e+00 -1.65363550e-01 7.42116034e-01 8.97594213e-01
3.06098580e-01 6.73983619e-02 5.03468215e-01 -4.65663999e-01
6.52491331e-01 -3.41664433e-01 -4.57266182e-01 2.92205691e-01
-5.46557367e-01 -6.22547232e-02 7.49589264e-01 -7.42910385e-01
-8.30998123e-01 -2.60520190e-01 -2.44168624e-01 -1.13033998e+00
7.59449601e-02 2.51450688e-01 -3.22763622e-01 -5.27563035e-01
7.54853070e-01 2.90226698e-01 -4.01395597e-02 -4.03592497e-01
5.08403182e-01 1.49565294e-01 8.72025907e-01 -4.06403571e-01
1.31277609e+00 2.45222807e-01 -2.10161194e-01 -6.89197540e-01
-6.14745796e-01 -1.64928973e-01 -2.47328714e-01 -1.30678743e-01
4.99663293e-01 -9.09763694e-01 -8.93767118e-01 1.07747781e+00
-1.00155914e+00 -4.72780645e-01 -2.60121524e-01 5.33674564e-03
-3.00483346e-01 5.02240360e-01 -7.23659873e-01 -5.28022110e-01
-6.69244826e-01 -1.52793372e+00 7.05329895e-01 4.42860246e-01
2.74658710e-01 -7.86371112e-01 -1.19333260e-01 4.38225746e-01
7.93107033e-01 2.79176891e-01 5.31087101e-01 -6.04583621e-01
-5.08740306e-01 -1.62431657e-01 -5.22354722e-01 7.98414588e-01
-6.51017874e-02 3.42301607e-01 -1.18339646e+00 -6.23289168e-01
-2.58009076e-01 -4.75117624e-01 1.46278787e+00 4.54918504e-01
1.33945990e+00 -5.35790741e-01 -4.92383838e-02 1.18757498e+00
1.45894396e+00 -1.15868747e-01 1.12526095e+00 4.96917516e-01
9.22238648e-01 1.29700184e-01 1.42923981e-01 3.68497409e-02
2.65971720e-01 5.40437162e-01 9.23673451e-01 -3.64797294e-01
-5.21521926e-01 -2.90299177e-01 6.76158369e-01 6.47199035e-01
-4.41257685e-01 -2.28701040e-01 -6.56050980e-01 5.08575141e-01
-1.82432997e+00 -1.03914106e+00 2.96308491e-02 1.89179420e+00
6.27820075e-01 2.24517852e-01 9.30480435e-02 2.50271738e-01
7.06065834e-01 2.33959153e-01 -8.74430954e-01 -1.24326311e-01
-4.03778404e-01 1.56480268e-01 7.40471303e-01 2.83061206e-01
-1.23789120e+00 1.27231264e+00 6.18306494e+00 1.09065890e+00
-1.17460275e+00 8.44186991e-02 6.11863792e-01 -5.86742461e-02
-3.07357647e-02 -2.38751277e-01 -5.64767122e-01 3.66835207e-01
3.57395887e-01 1.28050849e-01 4.86317247e-01 6.94101810e-01
1.72206670e-01 3.62925321e-01 -6.87839091e-01 8.11883926e-01
6.19985834e-02 -1.29173696e+00 2.33277872e-01 -2.50357002e-01
8.60783517e-01 4.11951751e-01 6.07797205e-01 4.63114530e-01
5.93297482e-01 -1.14670324e+00 8.59464526e-01 5.32381475e-01
8.09965074e-01 -6.99255824e-01 9.68808591e-01 -2.27502808e-01
-1.17355335e+00 -4.17815953e-01 -5.45804977e-01 2.89732605e-01
-3.11088771e-01 3.31844866e-01 -3.15269709e-01 6.27038240e-01
1.25439560e+00 9.16446090e-01 -1.02833939e+00 1.12522256e+00
-5.03565013e-01 7.95188546e-01 -1.49938881e-01 2.51451313e-01
2.27158353e-01 2.01216668e-01 7.93393612e-01 1.15186572e+00
6.19553849e-02 -6.81655481e-02 -6.56242389e-03 8.22254777e-01
-3.11006844e-01 -2.92067766e-01 -4.47105825e-01 2.85968393e-01
5.20136952e-01 1.36346400e+00 -5.68078399e-01 -2.65954971e-01
-2.82385916e-01 9.94098544e-01 3.35025191e-01 6.11249983e-01
-1.16142559e+00 -5.27669132e-01 1.08460414e+00 -3.12394530e-01
5.36728680e-01 2.15067994e-02 -2.26764381e-01 -1.18681002e+00
1.40654027e-01 -1.23787820e+00 1.89047575e-01 -8.42592061e-01
-1.46971440e+00 8.24403048e-01 -1.00138895e-01 -1.13393164e+00
5.23223162e-01 -5.43549120e-01 -1.07803464e+00 6.03273809e-01
-1.68789268e+00 -1.67279911e+00 -4.40793037e-01 1.09094417e+00
5.79612494e-01 -3.93267512e-01 4.31073874e-01 1.69608399e-01
-1.23382306e+00 1.28726435e+00 2.70479396e-02 3.97114933e-01
7.75993049e-01 -1.11233354e+00 9.12699878e-01 1.38031220e+00
-1.56127870e-01 8.08359623e-01 7.19971418e-01 -4.01220471e-01
-1.43686128e+00 -1.42766845e+00 1.67316884e-01 -5.76298714e-01
8.04352701e-01 -1.16536558e-01 -9.59189892e-01 5.66279292e-01
3.81759256e-01 3.05070281e-01 1.34439036e-01 -2.77591139e-01
-1.03291857e+00 -4.07813877e-01 -1.16469252e+00 9.34235156e-01
1.30699944e+00 -3.78956884e-01 -3.45396519e-01 1.44317150e-01
1.17897189e+00 -3.92129660e-01 -6.56567097e-01 6.36610329e-01
4.54279274e-01 -1.13713133e+00 1.35175157e+00 -5.55887401e-01
4.73906904e-01 -4.75414038e-01 -1.19556248e-01 -1.57323718e+00
-7.28504539e-01 -8.88168514e-01 -8.88531581e-02 1.30580640e+00
1.55315772e-01 -9.34843957e-01 3.01151693e-01 2.57747084e-01
-4.78635877e-01 -6.79347575e-01 -8.80009890e-01 -1.05784476e+00
2.40703151e-01 -4.26898628e-01 7.11848617e-01 8.39347422e-01
-4.77540612e-01 7.40482807e-02 -6.37126684e-01 3.62529457e-01
6.77920580e-01 -4.17545795e-01 9.19999897e-01 -7.29665458e-01
-6.49273992e-02 -8.03216100e-01 -6.09041989e-01 -7.20548034e-01
-1.92995034e-02 -7.25512564e-01 -7.01336861e-02 -1.42339909e+00
2.08056957e-01 -1.61279127e-01 -6.80593312e-01 1.00515258e+00
-5.81798077e-01 6.46451712e-01 5.08049965e-01 1.77417651e-01
-6.90806746e-01 6.84815109e-01 1.31841564e+00 -4.98761266e-01
-7.49153942e-02 -1.15340538e-01 -9.87759233e-01 5.99170148e-01
1.18255782e+00 -1.77231282e-01 -4.78139371e-01 -5.96765459e-01
-3.78181809e-03 -5.53212881e-01 7.45615125e-01 -1.20798004e+00
3.59214365e-01 -9.05013531e-02 2.83587962e-01 -2.91606605e-01
1.75726801e-01 -5.64549208e-01 -5.21503724e-02 6.07101083e-01
-2.62864739e-01 2.69976109e-01 5.59317291e-01 5.14424860e-01
3.56694753e-03 2.70375967e-01 1.04509842e+00 7.58413821e-02
-8.15356910e-01 7.47483015e-01 -1.24423675e-01 9.62852761e-02
8.24135542e-01 -1.75738499e-01 -8.39911163e-01 -3.00181895e-01
-1.81078881e-01 3.07905793e-01 3.64840984e-01 8.10033083e-01
9.99386430e-01 -1.37199616e+00 -1.04290950e+00 2.19242394e-01
2.42930055e-01 -8.79802182e-02 5.98788857e-01 8.80778074e-01
-3.76945198e-01 2.03274637e-01 -3.75842571e-01 -3.93839896e-01
-1.36658919e+00 8.35854709e-01 5.47348857e-01 -6.48589507e-02
-6.78141415e-01 1.27644801e+00 3.70069027e-01 -3.99835318e-01
5.32089949e-01 -3.96797240e-01 -1.59892648e-01 -4.03231502e-01
6.19888842e-01 5.31063676e-01 -6.61430657e-02 -6.03723824e-01
-3.43842924e-01 6.12890184e-01 -2.70747900e-01 3.85510981e-01
1.28677976e+00 9.41619352e-02 -2.00981051e-01 -2.29086861e-01
9.91754711e-01 -2.79119045e-01 -1.60285902e+00 -3.67377698e-01
-5.05976439e-01 -4.50590104e-01 2.87454966e-02 -8.95178199e-01
-1.76535046e+00 7.60579824e-01 6.56845570e-01 -3.57355885e-02
1.39828348e+00 -1.78965211e-01 8.58003020e-01 2.14184612e-01
1.73374619e-02 -8.77987146e-01 4.38061088e-01 6.36518061e-01
1.35579121e+00 -1.31388879e+00 -4.21992317e-02 -1.69246361e-01
-6.31067574e-01 6.40644550e-01 9.63797927e-01 -1.83542818e-01
6.41257107e-01 2.58088201e-01 2.10582688e-01 8.57951492e-02
-8.01820159e-01 -2.70231098e-01 3.71970266e-01 8.81864667e-01
-7.56016001e-02 -1.50750428e-01 2.07059696e-01 4.94674176e-01
-2.74007112e-01 -2.10895941e-01 3.11103135e-01 5.68603516e-01
-2.05766395e-01 -7.60058701e-01 -5.23575783e-01 2.15757802e-01
-5.53536594e-01 -3.08205187e-01 -4.62289065e-01 7.10511446e-01
2.83769906e-01 1.10395467e+00 -4.83897597e-01 -9.99046087e-01
6.13389969e-01 -5.26269794e-01 3.33827257e-01 -5.95088191e-02
-8.54807138e-01 -3.02558273e-01 -2.59617031e-01 -7.46560931e-01
-2.15592504e-01 -3.04187238e-01 -7.36372650e-01 -6.08763635e-01
-4.89248157e-01 -3.35815817e-01 5.37110388e-01 6.58715069e-01
3.86005372e-01 8.70710135e-01 7.17881441e-01 -1.08136714e+00
-7.11617649e-01 -1.10200131e+00 -1.02059759e-01 3.98898751e-01
5.84820330e-01 -6.39463246e-01 -3.40253115e-01 -1.37361646e-01] | [5.459530353546143, 7.947899341583252] |
84afafb0-7df8-4e66-8971-0da4276b2745 | frustratingly-easy-label-projection-for-cross | 2211.15613 | null | https://arxiv.org/abs/2211.15613v4 | https://arxiv.org/pdf/2211.15613v4.pdf | Frustratingly Easy Label Projection for Cross-lingual Transfer | Translating training data into many languages has emerged as a practical solution for improving cross-lingual transfer. For tasks that involve span-level annotations, such as information extraction or question answering, an additional label projection step is required to map annotated spans onto the translated texts. Recently, a few efforts have utilized a simple mark-then-translate method to jointly perform translation and projection by inserting special markers around the labeled spans in the original sentence. However, as far as we are aware, no empirical analysis has been conducted on how this approach compares to traditional annotation projection based on word alignment. In this paper, we present an extensive empirical study across 57 languages and three tasks (QA, NER, and Event Extraction) to evaluate the effectiveness and limitations of both methods, filling an important gap in the literature. Experimental results show that our optimized version of mark-then-translate, which we call EasyProject, is easily applied to many languages and works surprisingly well, outperforming the more complex word alignment-based methods. We analyze several key factors that affect the end-task performance, and show EasyProject works well because it can accurately preserve label span boundaries after translation. We will publicly release all our code and data. | ['Wei Xu', 'Alan Ritter', 'Chao Jiang', 'Yang Chen'] | 2022-11-28 | null | null | null | null | ['word-alignment', 'event-extraction', 'cross-lingual-transfer'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 2.49084741e-01 4.39743511e-03 -4.11809027e-01 -4.77804482e-01
-1.66696584e+00 -8.75656188e-01 4.29758698e-01 2.79836088e-01
-8.24254990e-01 9.02514875e-01 3.85838240e-01 -5.76815784e-01
3.91669691e-01 -2.92911798e-01 -7.50478148e-01 -1.50630236e-01
5.45226872e-01 5.87353170e-01 2.84220278e-01 -1.04115807e-01
3.32673104e-03 -5.83373010e-02 -8.39197516e-01 4.77136195e-01
1.26770473e+00 4.57688540e-01 8.86845663e-02 7.10867047e-02
-4.56905484e-01 3.95168781e-01 -5.96872985e-01 -1.05539882e+00
2.06189722e-01 -6.91429853e-01 -1.09326661e+00 -2.86862910e-01
5.73828876e-01 1.52372956e-01 1.52390346e-01 9.64020669e-01
6.95054948e-01 -4.94600646e-02 3.32632840e-01 -7.19559848e-01
-7.19602168e-01 1.13699150e+00 -5.35584033e-01 2.06811979e-01
5.17981708e-01 -2.26067781e-01 1.10024989e+00 -9.95990753e-01
7.80974686e-01 9.12049830e-01 1.00025785e+00 5.34296095e-01
-1.29023051e+00 -5.73832154e-01 7.12472424e-02 1.56894177e-01
-1.46466351e+00 -6.04998291e-01 5.21323204e-01 -3.50340933e-01
1.14589286e+00 2.20035002e-01 1.83155552e-01 1.01586926e+00
1.16293766e-01 6.44354820e-01 1.48036647e+00 -1.14141512e+00
-1.23992890e-01 4.21832502e-01 2.05155134e-01 6.25007391e-01
1.08480705e-02 -2.98222214e-01 -6.68343008e-01 -1.15539022e-01
-7.43089840e-02 -6.02661371e-01 -4.15838838e-01 7.95471743e-02
-1.34227180e+00 7.36600280e-01 -1.02009985e-03 4.84196156e-01
-1.12109967e-01 -1.95055604e-01 6.98852599e-01 3.37930262e-01
8.23322713e-01 5.60255349e-01 -5.63593626e-01 -2.42363945e-01
-1.18261778e+00 2.41629407e-02 8.25990140e-01 9.95924234e-01
7.17642844e-01 -3.04395139e-01 -4.83887970e-01 9.89492118e-01
-1.01426400e-01 4.45133239e-01 6.24435008e-01 -6.94947779e-01
1.00397110e+00 3.89150798e-01 4.58299890e-02 -4.94474232e-01
-3.21499616e-01 -2.42953241e-01 -2.17511103e-01 -3.94682467e-01
6.15175486e-01 -4.88792956e-01 -6.85936928e-01 1.93450165e+00
2.94887453e-01 -2.81802803e-01 4.56571579e-02 6.66379631e-01
6.76342726e-01 5.65667987e-01 3.18694234e-01 -2.84122109e-01
1.64970219e+00 -1.21729350e+00 -1.00950909e+00 -3.81324619e-01
1.06222522e+00 -1.33435130e+00 1.63787234e+00 -4.32551913e-02
-1.07158816e+00 -4.30523455e-01 -9.25595164e-01 -4.45444375e-01
-4.36989874e-01 4.14818078e-01 5.16981304e-01 8.60069752e-01
-8.00974429e-01 4.63982016e-01 -9.08737004e-01 -5.94220817e-01
2.75340136e-02 3.23830992e-02 -4.15323913e-01 -2.03560561e-01
-1.39756823e+00 1.39087808e+00 3.78080398e-01 -2.15050071e-01
-3.25204104e-01 -6.15948796e-01 -9.40738559e-01 -1.67175457e-01
3.18370432e-01 -5.02411485e-01 1.44835925e+00 -7.40091562e-01
-1.52639985e+00 1.15131164e+00 -4.60895985e-01 -3.25134218e-01
4.12654847e-01 -5.84420264e-01 -3.42630357e-01 -1.37337223e-01
2.60671020e-01 6.72077954e-01 1.84893668e-01 -8.63208711e-01
-5.42061627e-01 -2.34890625e-01 -8.26975331e-02 4.31616366e-01
-5.45918882e-01 6.41531825e-01 -6.37624085e-01 -7.19150603e-01
-1.68090314e-01 -1.04158068e+00 -3.40940431e-02 -5.06157637e-01
-3.84702295e-01 -4.87104595e-01 2.92772144e-01 -1.20382798e+00
1.55101788e+00 -2.00329041e+00 -2.48497128e-02 -2.48089418e-01
-3.88935983e-01 3.82151783e-01 -1.12428546e-01 8.08247149e-01
1.12795934e-01 1.89265370e-01 -4.22076732e-01 -6.14972532e-01
1.41767889e-01 1.31905168e-01 -4.71573561e-01 3.54419947e-01
6.78750128e-02 1.06024182e+00 -8.26093435e-01 -7.06445634e-01
-2.19049275e-01 3.47566515e-01 -3.29318643e-01 7.42647797e-02
-1.35765672e-01 3.68432879e-01 -1.38764620e-01 5.21628618e-01
3.73716176e-01 1.29264951e-01 2.20532611e-01 -1.25945881e-01
-2.19122320e-01 1.01660454e+00 -6.62402987e-01 2.06326580e+00
-6.47499621e-01 6.69214129e-01 -3.55531961e-01 -6.77540481e-01
7.86297560e-01 6.14454329e-01 2.10978255e-01 -6.75134361e-01
-1.52912691e-01 5.11905134e-01 -1.59833267e-01 -4.09253269e-01
7.13836610e-01 -3.56864005e-01 -3.93906265e-01 5.16198635e-01
1.04457997e-01 -1.67385891e-01 4.55722898e-01 4.05604914e-02
1.01269329e+00 5.51007867e-01 4.14387196e-01 -2.29492545e-01
4.80533779e-01 5.04848540e-01 5.91024756e-01 3.95024478e-01
-1.65102288e-01 4.86916721e-01 3.04750919e-01 -1.17273197e-01
-9.15206730e-01 -8.28644454e-01 -8.87807980e-02 1.32857025e+00
-1.88769147e-01 -6.71704352e-01 -1.23622060e+00 -1.03522098e+00
-3.40342432e-01 1.05461037e+00 -2.31367797e-01 6.99115172e-02
-8.67011428e-01 -6.59773827e-01 9.77028966e-01 4.69246060e-01
3.13976079e-01 -9.87021744e-01 -2.64115661e-01 3.32739294e-01
-8.06125462e-01 -1.44084013e+00 -9.34557855e-01 1.85952663e-01
-7.62709141e-01 -6.65975571e-01 -5.93401015e-01 -9.71651793e-01
4.98754263e-01 7.40699768e-02 1.28532207e+00 -2.23643094e-01
2.62288958e-01 8.66841301e-02 -5.59740722e-01 -5.08245289e-01
-4.76234704e-01 6.61706686e-01 -6.54374138e-02 -3.67807090e-01
7.66227961e-01 -2.34799013e-01 -1.52130380e-01 3.06561947e-01
-6.68056250e-01 8.15965384e-02 6.14088476e-01 6.49583578e-01
6.79146707e-01 -5.19337058e-01 7.11113930e-01 -1.31814373e+00
8.90796602e-01 -1.68707237e-01 -3.81354690e-01 6.53767765e-01
-6.59617782e-01 9.09778848e-02 6.67159379e-01 -3.17324281e-01
-1.06915510e+00 1.62459224e-01 -4.48482811e-01 1.03170343e-01
-6.31082617e-03 8.56339514e-01 -2.22350225e-01 3.37414406e-02
7.38944113e-01 6.34804787e-03 -2.60799468e-01 -7.97585368e-01
5.76601744e-01 8.84481072e-01 4.85539496e-01 -7.50548184e-01
6.76332653e-01 7.29672164e-02 -5.04247308e-01 -3.76349419e-01
-1.26591063e+00 -4.72888649e-01 -8.02481234e-01 9.12169814e-02
8.52818131e-01 -9.70797539e-01 3.44980806e-02 9.60056484e-02
-1.35766709e+00 -3.72639149e-01 -4.02982831e-01 6.52345657e-01
-4.07312751e-01 5.30515015e-01 -8.58559251e-01 -2.76413530e-01
-6.27412558e-01 -1.08176327e+00 9.87389863e-01 1.54554462e-02
-5.92014015e-01 -1.12392044e+00 3.92967135e-01 5.50524175e-01
3.95117939e-01 2.61353853e-04 1.04848993e+00 -8.33049655e-01
1.32408245e-02 -1.55429140e-01 -8.13813880e-02 3.20771605e-01
1.00787297e-01 -2.86343604e-01 -7.95496643e-01 -2.29760319e-01
-7.90750887e-03 -3.59029114e-01 6.67701483e-01 1.03445001e-01
7.13163018e-01 -2.46574968e-01 -2.42405698e-01 4.22691494e-01
1.29634964e+00 -1.44171566e-01 5.22609651e-01 4.88631159e-01
5.87323964e-01 7.59174109e-01 7.75701046e-01 -2.28889257e-01
6.58632994e-01 1.01839602e+00 -1.84659913e-01 -1.37181997e-01
-4.15705264e-01 -4.52446908e-01 7.71480441e-01 1.40017688e+00
9.19547305e-02 -2.50371993e-01 -8.96446526e-01 6.08050287e-01
-1.59318888e+00 -5.86223066e-01 -8.99905860e-02 2.33489799e+00
1.41750455e+00 3.02058943e-02 -3.70825194e-02 -1.85606658e-01
6.80802584e-01 -1.47197684e-02 -4.77463193e-02 -4.63966638e-01
-1.53433368e-01 4.61066931e-01 5.62755525e-01 6.76840007e-01
-9.61060762e-01 1.45955002e+00 6.70269251e+00 1.02963388e+00
-1.04102325e+00 5.80428541e-01 2.99718857e-01 2.11218044e-01
-4.34056908e-01 2.85696268e-01 -1.10654724e+00 4.72893417e-01
1.11483204e+00 -1.69248641e-01 2.39474133e-01 6.50718153e-01
1.37150526e-01 4.01341207e-02 -1.21740758e+00 6.06738269e-01
2.07665473e-01 -8.90501380e-01 -3.25512849e-02 -1.66320264e-01
7.44693339e-01 8.97806287e-02 -6.45587519e-02 5.21844447e-01
3.80500793e-01 -7.71298885e-01 8.20900440e-01 1.22025192e-01
1.01998365e+00 -5.72112560e-01 8.66183043e-01 3.68507028e-01
-1.06500745e+00 5.07064521e-01 -3.60606492e-01 2.82786060e-02
6.03938103e-01 6.65196717e-01 -8.55879188e-01 6.92411005e-01
4.04837847e-01 2.99102783e-01 -6.29420042e-01 9.45571423e-01
-7.19499409e-01 1.07697296e+00 -1.81645423e-01 1.33094475e-01
2.33570367e-01 -1.36846080e-01 3.97141546e-01 1.66081583e+00
5.86636364e-01 -2.12971032e-01 2.64398336e-01 5.39821684e-01
-3.20525259e-01 7.16402888e-01 -4.59478527e-01 -1.69689015e-01
5.27120650e-01 1.25608432e+00 -7.14020669e-01 -4.10813987e-01
-7.28311360e-01 1.09072828e+00 7.10449457e-01 2.55779117e-01
-8.90887141e-01 -4.32015747e-01 3.14051807e-01 -4.56932746e-02
8.70228335e-02 -3.62724960e-01 -4.07367527e-01 -1.18938267e+00
3.65798891e-01 -1.04028916e+00 6.26057804e-01 -6.06779516e-01
-1.19041538e+00 8.54095817e-01 -2.44423933e-02 -1.09992540e+00
-1.06645450e-01 -3.84973973e-01 -3.73322785e-01 1.08832085e+00
-1.57330811e+00 -1.23501277e+00 9.40568522e-02 3.35983485e-01
6.55901372e-01 7.97977671e-02 1.04660881e+00 5.30245662e-01
-6.24678135e-01 9.25707281e-01 -3.27125676e-02 2.53939956e-01
1.29181767e+00 -1.24040973e+00 4.81521726e-01 1.19850373e+00
7.01397181e-01 6.52386189e-01 5.78531504e-01 -7.16383398e-01
-9.55316365e-01 -9.92987871e-01 1.79306114e+00 -6.82840228e-01
6.51210606e-01 -4.05398250e-01 -1.03842366e+00 8.92609954e-01
7.97521532e-01 -2.97732204e-01 9.83409464e-01 2.94745564e-01
-4.09889281e-01 -7.58550838e-02 -7.60188460e-01 6.11553073e-01
9.38663483e-01 -6.17820680e-01 -9.65354562e-01 5.04497349e-01
9.77735579e-01 -6.77446127e-01 -8.71328950e-01 4.37891632e-01
2.22335473e-01 -5.75223982e-01 5.60531795e-01 -5.10424554e-01
3.34853560e-01 -2.30414867e-01 -1.19982362e-01 -1.50998080e+00
-9.72399488e-02 -7.49425054e-01 1.63260341e-01 1.62764883e+00
9.33838665e-01 -6.98700786e-01 4.72972155e-01 5.19639552e-01
-4.84674096e-01 -6.42053306e-01 -9.81270254e-01 -8.78085554e-01
3.83414388e-01 -3.82943600e-01 4.82980102e-01 1.10807967e+00
2.12992802e-01 6.48215830e-01 -2.99492419e-01 -4.74696383e-02
1.82152689e-01 1.28747910e-01 5.52288711e-01 -6.83960080e-01
-1.96388572e-01 -2.79311925e-01 2.14117676e-01 -1.09001935e+00
3.41463685e-01 -1.21449721e+00 1.63429976e-01 -1.62509549e+00
2.48684317e-01 -4.47071224e-01 -1.39022052e-01 8.54548931e-01
-5.78157902e-01 5.47197282e-01 1.73570365e-01 3.84910315e-01
-5.33369720e-01 3.44065607e-01 9.63919044e-01 9.74027812e-02
-1.55196831e-01 -6.84311315e-02 -7.95158327e-01 6.12911999e-01
8.19640338e-01 -7.53049254e-01 -3.22024152e-02 -9.33524549e-01
3.17412049e-01 -1.71414718e-01 -4.10765588e-01 -7.94081330e-01
1.58946812e-01 4.20233086e-02 -1.10177644e-01 -4.98335212e-01
2.36268863e-02 -6.75935328e-01 -8.32954072e-04 2.42580593e-01
-4.70302820e-01 5.42111993e-01 3.40558082e-01 -1.71107985e-02
-4.40483630e-01 -5.79560816e-01 5.29257834e-01 -8.77903253e-02
-5.60602784e-01 -8.80382657e-02 -2.37883210e-01 4.64389950e-01
8.53186727e-01 6.75460547e-02 -3.22259545e-01 -1.39732033e-01
-3.78861427e-01 5.48410304e-02 4.48658168e-01 4.05646175e-01
1.97930615e-02 -1.39065850e+00 -9.31684971e-01 -2.26203248e-01
1.80515558e-01 -3.07496041e-01 -1.57086730e-01 1.07884252e+00
-4.64606404e-01 5.99492550e-01 8.07638764e-02 -3.07177126e-01
-1.35118091e+00 4.88087326e-01 -9.79419984e-03 -8.04288447e-01
-5.01261234e-01 6.59394681e-01 -2.23844424e-01 -7.75644064e-01
1.26620412e-01 -1.68123007e-01 -1.31685778e-01 1.96876034e-01
3.69276226e-01 2.37159953e-01 4.71445709e-01 -6.88398957e-01
-3.28619659e-01 6.17340207e-01 -1.09167956e-01 -4.69352633e-01
9.55326080e-01 -4.11003500e-01 -3.17649573e-01 5.17682433e-01
9.31653500e-01 5.74063122e-01 -7.41779923e-01 -6.24930918e-01
4.72979188e-01 -2.70642012e-01 -3.78233045e-01 -9.35979307e-01
-6.31966114e-01 9.59818900e-01 1.64721385e-01 2.78346576e-02
1.00325942e+00 1.37163084e-02 1.13921261e+00 2.10561827e-01
4.13815796e-01 -1.30003846e+00 -2.66764730e-01 7.58298397e-01
6.51814520e-01 -1.13129699e+00 -8.85397121e-02 -6.15878284e-01
-7.57566035e-01 7.83865750e-01 6.36097729e-01 3.80259365e-01
1.41630575e-01 2.30214357e-01 4.19337898e-01 1.29713088e-01
-4.54943299e-01 -2.79334158e-01 4.67058510e-01 4.14367348e-01
9.79765177e-01 9.02623907e-02 -9.78077173e-01 5.64141214e-01
-6.29651189e-01 -1.71798468e-01 2.35036835e-01 7.72535563e-01
-3.04505467e-01 -1.60465944e+00 -2.51200080e-01 6.06605150e-02
-9.69463766e-01 -5.21397114e-01 -4.65502590e-01 8.26481819e-01
1.24341182e-01 8.57192814e-01 -2.10522071e-01 -2.05687225e-01
3.94893438e-01 6.82883799e-01 4.98237759e-01 -1.02880561e+00
-7.99986303e-01 -2.70149205e-02 5.01450300e-01 -3.55791718e-01
-5.48358679e-01 -7.30202079e-01 -1.11568749e+00 -4.89552170e-02
-5.08560121e-01 5.00149012e-01 7.32023895e-01 1.16439736e+00
2.49420643e-01 2.75731415e-01 2.59914637e-01 -3.60800773e-01
-6.79092228e-01 -1.22009027e+00 -8.45302045e-02 3.75394404e-01
-2.47065276e-01 -3.25178146e-01 -1.70262083e-01 2.84372747e-01] | [11.213830947875977, 10.044257164001465] |
e48a067b-e1dc-4ab2-959d-eae76b64a938 | grammatical-error-correction-a-survey-of-the | 2211.05166 | null | https://arxiv.org/abs/2211.05166v4 | https://arxiv.org/pdf/2211.05166v4.pdf | Grammatical Error Correction: A Survey of the State of the Art | Grammatical Error Correction (GEC) is the task of automatically detecting and correcting errors in text. The task not only includes the correction of grammatical errors, such as missing prepositions and mismatched subject-verb agreement, but also orthographic and semantic errors, such as misspellings and word choice errors respectively. The field has seen significant progress in the last decade, motivated in part by a series of five shared tasks, which drove the development of rule-based methods, statistical classifiers, statistical machine translation, and finally neural machine translation systems which represent the current dominant state of the art. In this survey paper, we condense the field into a single article and first outline some of the linguistic challenges of the task, introduce the most popular datasets that are available to researchers (for both English and other languages), and summarise the various methods and techniques that have been developed with a particular focus on artificial error generation. We next describe the many different approaches to evaluation as well as concerns surrounding metric reliability, especially in relation to subjective human judgements, before concluding with an overview of recent progress and suggestions for future work and remaining challenges. We hope that this survey will serve as comprehensive resource for researchers who are new to the field or who want to be kept apprised of recent developments. | ['Ted Briscoe', 'Hwee Tou Ng', 'Hannan Cao', 'Muhammad Reza Qorib', 'Zheng Yuan', 'Christopher Bryant'] | 2022-11-09 | null | null | null | null | ['grammatical-error-correction'] | ['natural-language-processing'] | [ 4.17093843e-01 2.71601379e-01 1.98106632e-01 -6.65816307e-01
-1.08393359e+00 -4.53192264e-01 4.61690813e-01 6.61228657e-01
-7.51683235e-01 1.22647023e+00 3.35768402e-01 -3.46919358e-01
1.08681638e-02 -3.55661809e-01 -6.19131982e-01 -1.56599939e-01
1.03859834e-01 7.76696384e-01 -1.64050624e-01 -6.49767578e-01
9.23658252e-01 1.24355957e-01 -1.72663319e+00 4.18058395e-01
1.44563210e+00 5.25445879e-01 2.18787849e-01 5.05953491e-01
-3.74766260e-01 5.29062569e-01 -8.11726630e-01 -1.02262378e+00
-2.59724945e-01 -4.58805531e-01 -1.17004991e+00 -3.06775510e-01
4.38648373e-01 3.84821475e-01 4.66455638e-01 1.35933530e+00
6.48191750e-01 2.74530500e-01 6.10310435e-01 -9.25892055e-01
-1.01448822e+00 6.47496223e-01 8.75617936e-02 3.19823146e-01
7.53541291e-01 -2.22201869e-01 1.04411495e+00 -1.43089235e+00
7.17400372e-01 1.19131219e+00 8.85903955e-01 8.81123483e-01
-6.41198456e-01 -1.73607081e-01 -1.67212095e-02 5.99340677e-01
-1.10905659e+00 -7.47731447e-01 3.24809313e-01 -1.88230619e-01
1.71245241e+00 4.65322018e-01 1.33737847e-01 1.06796706e+00
2.85189331e-01 6.03805661e-01 1.28692889e+00 -1.21011972e+00
9.17097777e-02 2.38383234e-01 9.29441229e-02 5.92699707e-01
3.68155330e-01 7.18536079e-02 -6.34086728e-01 6.25775605e-02
-6.14552610e-02 -7.16458559e-01 -2.14819685e-01 3.76192272e-01
-1.32175958e+00 8.75307739e-01 -2.30314732e-01 7.68605828e-01
-1.59848422e-01 -2.36715645e-01 5.83099365e-01 6.79211736e-01
5.16394615e-01 7.12451279e-01 -6.80422068e-01 -5.14309227e-01
-8.83374691e-01 4.82876211e-01 9.55321074e-01 1.13187802e+00
4.23973471e-01 1.83270305e-01 2.94846762e-02 1.26607871e+00
3.23973924e-01 5.12139678e-01 6.78157210e-01 -8.91112089e-01
8.14310312e-01 4.11348164e-01 3.73674631e-01 -1.01892996e+00
-3.86900574e-01 3.72593515e-02 -5.37806928e-01 -4.11463268e-02
4.31723803e-01 -4.23385575e-02 -6.58365309e-01 1.52112520e+00
-5.25722578e-02 -5.93623281e-01 -9.74832103e-03 7.99235106e-01
1.05312407e+00 4.82929796e-01 2.07296789e-01 -5.31551003e-01
1.39572024e+00 -8.64781201e-01 -1.07131386e+00 -3.68906021e-01
7.76471198e-01 -1.43507493e+00 1.02186704e+00 2.44501010e-01
-1.48739600e+00 -3.51536810e-01 -8.48147690e-01 -4.32820141e-01
-7.94076025e-01 1.67300418e-01 2.89848417e-01 5.89570582e-01
-1.12199271e+00 9.26260173e-01 -6.76269770e-01 -8.43911290e-01
-1.69063330e-01 3.72498184e-01 -4.68303382e-01 -7.17384145e-02
-1.51505148e+00 1.73439550e+00 2.45984063e-01 1.51585817e-01
9.92474556e-02 -3.07146639e-01 -1.02063191e+00 -5.18033803e-01
1.12781376e-01 -5.23423374e-01 1.43445015e+00 -1.04736543e+00
-1.40847826e+00 1.57465279e+00 -5.27071118e-01 -3.84941727e-01
3.79536271e-01 -1.06870718e-01 -8.72848809e-01 -3.52544427e-01
3.47828090e-01 5.18975854e-01 2.93652773e-01 -9.62700248e-01
-9.52838361e-01 -5.54797232e-01 -4.17286724e-01 4.73242670e-01
2.38529995e-01 1.10683298e+00 -5.67116030e-02 -1.02466917e+00
1.72836766e-01 -7.47069895e-01 9.19160917e-02 -4.58732426e-01
-1.87578246e-01 -7.47451901e-01 6.56624883e-02 -1.13452220e+00
1.47817576e+00 -1.67853808e+00 3.58330309e-01 -2.04377383e-01
-4.21961427e-01 4.60086703e-01 -6.48273677e-02 9.10291135e-01
-1.69355959e-01 1.94168523e-01 -7.04732776e-01 -7.27890849e-01
2.01594427e-01 3.59967291e-01 -2.52490789e-01 3.20009083e-01
2.52174914e-01 8.86952639e-01 -1.01804626e+00 -3.83926749e-01
1.36340097e-01 1.75330684e-01 -3.96524630e-02 2.80628242e-02
-7.98284169e-03 3.35985214e-01 1.95337161e-01 7.89242804e-01
4.77475047e-01 2.69466788e-01 3.97751993e-03 3.41209888e-01
-5.77220798e-01 9.63920534e-01 -1.16407037e+00 1.56613517e+00
-5.59666693e-01 4.32717204e-01 -7.54384100e-02 -1.06301904e+00
1.10816467e+00 5.48465073e-01 -1.57889724e-01 -6.91468894e-01
1.69033006e-01 1.19252872e+00 -6.14493340e-02 -5.08522689e-01
9.80005383e-01 -1.22674651e-01 -2.29302675e-01 4.37697709e-01
1.17422007e-01 -1.71547234e-01 4.76974308e-01 -1.82092413e-01
7.63775289e-01 9.76587906e-02 7.95650482e-01 -4.14187044e-01
8.95743430e-01 2.25011006e-01 4.99840766e-01 6.07358456e-01
-4.13708180e-01 6.72239602e-01 -4.01246548e-02 -6.52743638e-01
-1.04098392e+00 -6.65549517e-01 -2.18614593e-01 1.00410402e+00
-2.17159614e-01 -4.99314100e-01 -8.88309836e-01 -6.49627388e-01
-2.40917385e-01 1.03681135e+00 -4.99484450e-01 1.65449649e-01
-9.38364267e-01 -8.84152710e-01 6.15438104e-01 3.76261592e-01
2.35964715e-01 -1.70742655e+00 -3.21703792e-01 4.72990781e-01
-8.04251969e-01 -1.01001596e+00 -3.31411660e-01 1.05741791e-01
-1.01574385e+00 -1.04149818e+00 -5.43442249e-01 -1.20286107e+00
5.72874486e-01 -1.43405572e-01 1.49930787e+00 6.47062957e-01
-8.37653726e-02 1.96409896e-01 -8.27078700e-01 -5.07164299e-01
-7.84100354e-01 -2.45448977e-01 2.85929680e-01 -5.21512330e-01
1.08432150e+00 -3.09397638e-01 -5.95293939e-02 1.07024230e-01
-5.30579627e-01 -2.55071640e-01 2.21868441e-01 9.51048076e-01
3.94193739e-01 -5.18908322e-01 5.76612651e-01 -1.09345579e+00
9.22032595e-01 -2.87247479e-01 -1.98774040e-01 3.55035722e-01
-9.65162337e-01 -1.18518457e-01 4.03116405e-01 3.83763433e-01
-1.11456680e+00 -2.70212173e-01 -6.21378660e-01 6.01247787e-01
-2.35921443e-01 4.78385866e-01 1.16515540e-01 -2.98848569e-01
8.95841539e-01 2.44463161e-01 -6.90461770e-02 -6.54283702e-01
1.17596157e-01 1.10042262e+00 6.50856435e-01 -7.19208896e-01
2.65304953e-01 -7.88737610e-02 -2.98670173e-01 -5.16877651e-01
-7.01704443e-01 -3.89382392e-01 -7.82521069e-01 -8.97909552e-02
5.63528121e-01 -5.13111711e-01 4.24549468e-02 7.24791288e-01
-1.83538234e+00 1.28604725e-01 -1.53679624e-01 3.35077941e-01
-7.10810483e-01 6.67128205e-01 -8.56307864e-01 -5.51628053e-01
-6.93467021e-01 -1.03401947e+00 1.02249575e+00 1.39345631e-01
-7.04590440e-01 -1.26230574e+00 1.31267413e-01 4.64758426e-01
5.16636550e-01 6.64662346e-02 1.04286921e+00 -8.22837114e-01
2.66274493e-02 -1.63929448e-01 1.12349563e-03 6.39916480e-01
-1.17200047e-01 -6.91991597e-02 -4.89763319e-01 -4.44330201e-02
-1.28566800e-02 -2.82481164e-01 5.07034183e-01 4.27758284e-02
6.14205956e-01 -2.85342008e-01 -3.77001800e-02 1.44470528e-01
1.31589091e+00 1.74263969e-01 6.34220064e-01 4.69075471e-01
2.52946168e-01 9.28561032e-01 8.21006238e-01 1.04106009e-01
6.16341412e-01 8.97676766e-01 8.49109069e-02 4.04684067e-01
-3.99028331e-01 -9.15716514e-02 3.24429870e-01 1.61298692e+00
-5.15897691e-01 -2.76773125e-01 -9.49486196e-01 7.01449335e-01
-1.75019062e+00 -6.32046342e-01 -7.21611798e-01 2.36339021e+00
1.03134298e+00 -3.43668640e-01 -2.23203242e-01 3.64215672e-01
1.06756687e+00 -2.37176523e-01 2.06688166e-01 -1.25193024e+00
-3.99435252e-01 5.64454675e-01 1.98290586e-01 9.73442733e-01
-9.90352273e-01 1.33775020e+00 7.18658352e+00 5.15971363e-01
-6.85940683e-01 3.31208348e-01 2.08952889e-01 5.19334197e-01
-2.59313852e-01 -4.95339092e-03 -8.25212419e-01 6.62532747e-01
1.00173485e+00 7.51290238e-03 5.95145702e-01 4.93309766e-01
1.68594792e-01 -2.37254024e-01 -9.35717642e-01 8.92755508e-01
6.36971951e-01 -9.82498050e-01 -2.06085250e-01 -3.55328977e-01
8.63214433e-01 1.87028751e-01 -3.12414497e-01 4.15512025e-01
2.23671302e-01 -9.52412963e-01 7.61265516e-01 3.92572284e-01
7.43990123e-01 -4.85177606e-01 1.07479107e+00 1.58289284e-01
-5.04067123e-01 1.27099976e-01 -6.32162452e-01 -6.28180563e-01
2.89675832e-01 6.65941417e-01 -2.13806584e-01 4.92327273e-01
8.36107194e-01 5.78732193e-01 -5.39679825e-01 1.16867495e+00
-7.18247712e-01 4.99566823e-01 -7.00619966e-02 -3.44737291e-01
2.42865328e-02 -2.06636235e-01 8.11441660e-01 1.49372447e+00
6.12834215e-01 7.71518573e-02 -4.36219007e-01 4.82644647e-01
-3.10304225e-03 8.02280426e-01 -4.43480641e-01 1.03384495e-01
7.04433501e-01 9.67073977e-01 -4.35344309e-01 -3.01272243e-01
-5.18196166e-01 1.42014813e+00 6.97709322e-01 2.03023534e-02
-4.47264582e-01 -8.39212656e-01 3.97070736e-01 -2.33963490e-01
-1.38463080e-01 -1.41210958e-01 -7.35518217e-01 -1.22801578e+00
4.19093698e-01 -1.28034556e+00 4.84205157e-01 -7.54413605e-01
-1.39006841e+00 8.22384357e-01 -2.60718495e-01 -9.58234906e-01
-4.28809941e-01 -8.86848271e-01 -5.13543308e-01 1.28633201e+00
-1.47733426e+00 -7.27311552e-01 -6.71296194e-02 7.35428035e-02
7.16737032e-01 -3.26028198e-01 1.32318592e+00 4.36057240e-01
-3.13751429e-01 7.57338166e-01 2.50693738e-01 -1.09340608e-01
1.28284252e+00 -1.34294045e+00 6.52142823e-01 9.42384303e-01
1.76984981e-01 6.70177460e-01 8.22012722e-01 -6.51801050e-01
-7.94811070e-01 -8.96773160e-01 2.57796955e+00 -8.75098646e-01
4.42880064e-01 -3.90932709e-02 -8.78885865e-01 5.23223519e-01
3.84903431e-01 -2.41045564e-01 5.18499315e-01 1.06492616e-01
-5.08418009e-02 1.89780504e-01 -1.31968546e+00 4.25980866e-01
1.08126545e+00 -3.85962576e-01 -1.06267822e+00 6.32666290e-01
3.84338707e-01 -7.27954030e-01 -5.37862360e-01 3.67719412e-01
1.67592049e-01 -8.97668540e-01 3.87220412e-01 -8.99731398e-01
5.47656596e-01 -2.29040742e-01 -2.60397822e-01 -1.60670233e+00
-3.07767630e-01 -6.12568378e-01 2.12347001e-01 1.28234816e+00
6.72665656e-01 -5.64020157e-01 1.57147586e-01 5.92281818e-01
-7.26309299e-01 -5.60804665e-01 -1.22000420e+00 -6.52355194e-01
4.05623794e-01 -5.36351502e-01 3.91881168e-01 9.85298932e-01
3.82475287e-01 1.98279575e-01 -3.99933189e-01 -3.51176918e-01
2.29741082e-01 -3.08216184e-01 8.72710869e-02 -1.12584472e+00
1.59893408e-01 -4.46430236e-01 -3.84710997e-01 -6.01595938e-01
9.38053429e-02 -1.13755774e+00 3.60303491e-01 -1.55319738e+00
-1.72520742e-01 -1.89613342e-01 9.32628587e-02 3.16383988e-01
-3.46561581e-01 5.83825886e-01 2.93321963e-02 2.01533169e-01
-4.45266366e-01 1.88393444e-01 9.13218379e-01 1.46498889e-01
5.27461581e-02 -1.72998123e-02 -9.10532713e-01 7.10265696e-01
8.10213447e-01 -6.69381618e-01 3.17666918e-01 -9.16678488e-01
7.80523062e-01 -2.81688213e-01 2.60056276e-02 -7.69145012e-01
2.66974807e-01 1.01558581e-01 -4.58457656e-02 -1.37322634e-01
-1.79000199e-01 -4.27662045e-01 -1.49475336e-01 2.35758334e-01
-2.82012880e-01 8.63530040e-01 3.85904200e-02 1.09746322e-01
-5.02012193e-01 -9.98677135e-01 8.55405629e-01 -4.36306447e-01
-9.19647217e-01 -3.05984199e-01 -5.54144919e-01 2.62879997e-01
7.66256392e-01 -4.16501313e-02 -1.49384454e-01 -1.45701215e-01
-6.97824836e-01 6.82187080e-02 3.44315797e-01 7.32150376e-01
4.34597522e-01 -1.32314181e+00 -1.15961730e+00 2.30649463e-03
4.87719923e-01 -3.77799243e-01 -4.97757979e-02 1.00001681e+00
-7.40611613e-01 7.26086974e-01 -2.26596043e-01 -2.15553660e-02
-1.31815398e+00 2.36819550e-01 2.49028325e-01 -1.14809655e-01
-3.42396021e-01 9.45538700e-01 -9.90380228e-01 -9.86635149e-01
2.63540238e-01 -9.42735597e-02 -4.84442264e-01 -1.73253998e-01
6.50143504e-01 6.24764204e-01 8.84202659e-01 -9.90054905e-01
-5.24317980e-01 5.72587907e-01 -5.63754588e-02 -3.41799296e-02
9.63022113e-01 -4.77618307e-01 -6.76926434e-01 4.95166153e-01
7.34151781e-01 4.71308082e-02 -7.16838092e-02 -3.72211605e-01
3.46041739e-01 -3.53898197e-01 -2.69164294e-01 -1.54458237e+00
-3.79545510e-01 1.06971896e+00 3.67911190e-01 -1.23967879e-01
8.41215491e-01 -1.31476462e-01 8.62536371e-01 4.75454777e-01
6.07832968e-01 -1.85930300e+00 -4.35842097e-01 1.11598802e+00
1.26474595e+00 -1.38190305e+00 -3.23596328e-01 -5.40181637e-01
-5.67330837e-01 1.30307746e+00 4.06948149e-01 -5.23345806e-02
2.74581283e-01 -6.03529215e-02 2.45266572e-01 1.33884940e-02
-5.49277842e-01 -7.43882954e-02 2.98518091e-01 7.65903711e-01
1.13999557e+00 1.02320574e-01 -1.47058678e+00 6.64791822e-01
-4.66082215e-01 -1.91563591e-01 5.47953486e-01 7.77398050e-01
-5.23700833e-01 -1.72811770e+00 -4.31805491e-01 3.63599807e-01
-6.18898571e-01 -5.29682159e-01 -5.84250748e-01 6.69796169e-01
4.49503154e-01 1.22286224e+00 -1.44047812e-01 -2.71139573e-02
6.10115290e-01 4.99410838e-01 7.09531009e-01 -8.46453607e-01
-7.75971711e-01 -6.42709494e-01 6.25070512e-01 -2.87334502e-01
-3.89934957e-01 -1.03940725e+00 -1.12243652e+00 -4.89673525e-01
-3.02998990e-01 4.15411204e-01 9.60023403e-01 1.20848560e+00
1.23324007e-01 2.87048426e-02 6.51326105e-02 -3.83224219e-01
-8.49104524e-01 -1.42906988e+00 -3.58773708e-01 4.93605703e-01
-2.24019051e-01 -5.53864419e-01 -3.95271987e-01 1.02128871e-02] | [11.059906959533691, 10.641997337341309] |
6dc46fa7-2579-446c-90b7-aafaab7186b7 | domain-aware-no-reference-image-quality | 1911.00673 | null | https://arxiv.org/abs/1911.00673v3 | https://arxiv.org/pdf/1911.00673v3.pdf | Domain Fingerprints for No-reference Image Quality Assessment | Human fingerprints are detailed and nearly unique markers of human identity. Such a unique and stable fingerprint is also left on each acquired image. It can reveal how an image was degraded during the image acquisition procedure and thus is closely related to the quality of an image. In this work, we propose a new no-reference image quality assessment (NR-IQA) approach called domain-aware IQA (DA-IQA), which for the first time introduces the concept of domain fingerprint to the NR-IQA field. The domain fingerprint of an image is learned from image collections of different degradations and then used as the unique characteristics to identify the degradation sources and assess the quality of the image. To this end, we design a new domain-aware architecture, which enables simultaneous determination of both the distortion sources and the quality of an image. With the distortion in an image better characterized, the image quality can be more accurately assessed, as verified by extensive experiments, which show that the proposed DA-IQA performs better than almost all the compared state-of-the-art NR-IQA methods. | ['Jing-Hao Xue', 'Yujiu Yang', 'Weihao Xia', 'Jing Xiao'] | 2019-11-02 | null | null | null | null | ['no-reference-image-quality-assessment'] | ['computer-vision'] | [ 3.40829074e-01 -5.68616092e-01 -1.41059849e-02 -2.99835950e-01
-6.59503758e-01 -6.73836410e-01 4.62556332e-01 -2.22283915e-01
-1.61984004e-02 3.69213432e-01 9.85936075e-02 1.79630414e-01
-4.28689420e-01 -7.82930195e-01 -5.40410399e-01 -7.78227389e-01
1.82224885e-01 9.90769342e-02 6.49382770e-02 -8.58663488e-03
5.30330658e-01 4.47970957e-01 -1.63113499e+00 8.23470503e-02
1.02892470e+00 1.25134957e+00 2.82442689e-01 5.28708875e-01
1.25373259e-01 5.85698128e-01 -7.60046482e-01 -3.92024457e-01
2.66366243e-01 -4.86425459e-01 -4.84003901e-01 1.11252040e-01
4.82744247e-01 -5.20713866e-01 -2.54559100e-01 1.27128673e+00
4.35377240e-01 -2.15446115e-01 7.32501268e-01 -1.06327760e+00
-6.78007364e-01 -2.10201796e-02 -4.46195722e-01 8.49788710e-02
4.79971021e-01 3.84626426e-02 5.81357002e-01 -7.60162354e-01
7.08240092e-01 1.23275328e+00 5.14365256e-01 2.58814692e-01
-1.17884874e+00 -8.57604980e-01 -2.48935133e-01 5.43771267e-01
-1.40662062e+00 -3.81160438e-01 1.08384585e+00 -6.06567800e-01
-1.30685672e-01 6.53056875e-02 2.59910673e-01 9.98209715e-01
7.13454410e-02 6.02437198e-01 1.47323298e+00 -3.84763807e-01
9.06141996e-02 1.57482803e-01 1.58952296e-01 7.07620919e-01
3.74099582e-01 4.37324435e-01 -5.93865216e-01 1.03351824e-01
7.76708484e-01 -3.56480293e-02 -3.73832613e-01 -4.86801535e-01
-1.13719606e+00 1.66078299e-01 3.13382745e-01 3.15783501e-01
-4.44379449e-01 -2.61338264e-01 8.91140625e-02 1.95496082e-01
-4.44953404e-02 3.44046563e-01 1.80680409e-01 -3.77083942e-02
-9.74957883e-01 4.76336032e-02 2.99764782e-01 6.76470816e-01
1.02610433e+00 -1.05085306e-01 -3.87667596e-01 9.29056764e-01
3.68821919e-02 9.31963384e-01 3.60195577e-01 -1.02993846e+00
4.04629141e-01 7.47920930e-01 6.31613135e-01 -1.41834474e+00
-1.86113507e-01 -6.12610161e-01 -8.58487546e-01 4.75010484e-01
5.78131914e-01 2.78021246e-01 -7.91951418e-01 1.52234435e+00
9.91198570e-02 1.08901329e-01 4.89407517e-02 9.97914255e-01
5.33534229e-01 6.13185823e-01 -4.13801551e-01 -9.67932716e-02
1.40660584e+00 -5.64069152e-01 -7.73962736e-01 -6.41636178e-02
-2.45924249e-01 -8.22695196e-01 1.05591404e+00 7.46022701e-01
-6.56296968e-01 -1.23328185e+00 -1.54212093e+00 3.69532675e-01
-8.77258256e-02 4.58850801e-01 -4.71922243e-03 1.06871474e+00
-7.38245964e-01 5.85109532e-01 -3.32218885e-01 -2.67596185e-01
1.55559763e-01 -2.20345538e-02 -5.07601738e-01 -4.78313297e-01
-1.10838413e+00 7.42098868e-01 2.63155878e-01 -1.07677475e-01
-1.03585815e+00 -6.13445222e-01 -5.83771825e-01 -1.45164505e-01
2.01405063e-01 -3.22596133e-01 8.81636024e-01 -9.91070271e-01
-1.43624890e+00 9.22082305e-01 -3.34602177e-01 -1.47327900e-01
4.73934919e-01 -9.32132229e-02 -9.13852513e-01 3.70627344e-01
2.31649175e-01 -1.54193295e-02 1.23864841e+00 -1.63911963e+00
-4.77789968e-01 -6.39700294e-01 -1.28020808e-01 -7.42607489e-02
-2.43936405e-01 -3.56874168e-02 -7.56501734e-01 -5.37852287e-01
1.05905488e-01 -6.72811627e-01 3.71208012e-01 7.10075945e-02
-2.33816445e-01 2.92632163e-01 7.07839847e-01 -8.88789892e-01
1.34602249e+00 -2.32180691e+00 -1.36543781e-01 5.02021313e-01
2.20312439e-02 5.48352778e-01 -3.57794583e-01 3.52882773e-01
2.09533796e-01 -1.56729043e-01 -2.51650363e-01 6.78136051e-02
-1.18252471e-01 -4.10154685e-02 -1.00099713e-01 4.84564304e-01
2.84467302e-02 4.87995237e-01 -8.56268942e-01 -5.79421639e-01
8.24112296e-02 2.76092738e-01 5.27509227e-02 5.14107347e-01
2.21489325e-01 6.35719061e-01 -4.29086208e-01 8.52137268e-01
1.14392960e+00 -1.48990065e-01 2.43981466e-01 -5.81981659e-01
-1.13651358e-01 -2.92008340e-01 -1.41616333e+00 1.51489091e+00
-3.91924620e-01 4.79793936e-01 1.31707698e-01 -7.79157043e-01
1.36059225e+00 2.17785314e-01 4.32324678e-01 -1.40303540e+00
-1.49002150e-01 2.98805833e-01 -6.99954927e-02 -6.41738296e-01
4.41224784e-01 1.95740640e-01 8.81495997e-02 4.52010721e-01
-9.70926415e-03 3.84739012e-01 1.27883196e-01 -9.25513953e-02
8.41195226e-01 1.23367034e-01 2.11243063e-01 -2.71470547e-01
1.00382662e+00 -3.04093927e-01 8.00355911e-01 8.91773582e-01
-3.90139043e-01 6.22137070e-01 2.53377378e-01 -3.03069055e-01
-1.28823972e+00 -1.26741624e+00 -2.84539878e-01 6.30819559e-01
8.90036643e-01 1.96261164e-02 -9.02955532e-01 -3.43321860e-01
1.30157605e-01 1.55422434e-01 -5.50089300e-01 -1.49615398e-02
-5.05117416e-01 -4.32852983e-01 5.71169257e-01 2.06103370e-01
1.06609595e+00 -7.70298958e-01 -2.72836983e-01 1.72055766e-01
-4.63685989e-01 -1.16063869e+00 -4.08936441e-01 -5.34274995e-01
-6.46036744e-01 -1.21387303e+00 -9.77649748e-01 -5.44632971e-01
6.08403802e-01 3.43426287e-01 8.73004019e-01 -7.92008936e-02
-1.94348678e-01 2.16875255e-01 -2.59416580e-01 -7.88878053e-02
-5.58858633e-01 -3.36788893e-01 2.61353403e-02 8.59542787e-01
-9.17916745e-02 -5.91778278e-01 -1.01343179e+00 7.84591317e-01
-6.80238724e-01 -2.06421897e-01 9.15460944e-01 7.03577399e-01
7.51405299e-01 3.88884485e-01 4.93388623e-01 -8.04572701e-01
6.54939532e-01 -1.88049927e-01 -7.38969743e-01 5.81337571e-01
-9.49368894e-01 2.49313131e-01 6.69240355e-01 -2.51567870e-01
-1.36374843e+00 -1.32306397e-01 1.63174734e-01 -2.82808810e-01
-1.83797017e-01 2.26749063e-01 -6.77097976e-01 -3.73922497e-01
7.82335818e-01 5.90681911e-01 1.08070888e-01 -7.55042195e-01
1.46602884e-01 8.93585384e-01 1.25506628e+00 -6.53157175e-01
9.37534153e-01 5.41757405e-01 2.30750337e-01 -7.58021057e-01
-4.38914239e-01 -7.81080782e-01 -5.53470790e-01 -5.66818833e-01
4.64917541e-01 -7.93275833e-01 -7.23441660e-01 1.12905681e+00
-1.07503819e+00 8.67643133e-02 2.86838114e-01 3.07843953e-01
-4.31436509e-01 7.36842394e-01 -1.67684123e-01 -8.74442041e-01
-2.83347398e-01 -1.12455869e+00 8.37746084e-01 3.84004921e-01
2.04182982e-01 -5.12111247e-01 1.08300857e-01 5.26800573e-01
3.02348733e-01 2.45144308e-01 9.03400302e-01 -6.18075542e-02
-7.16080308e-01 -2.59031504e-01 -6.10983551e-01 6.04461253e-01
3.20825309e-01 -1.37271732e-01 -1.06378329e+00 -1.94917306e-01
3.51830199e-02 2.38434002e-01 4.74162847e-01 1.35460272e-01
9.79632914e-01 -1.65481687e-01 -1.49422586e-01 7.61013150e-01
1.72166240e+00 5.03282368e-01 1.12670982e+00 5.64816535e-01
4.67026442e-01 4.49035227e-01 1.04650664e+00 3.77944440e-01
1.91248972e-02 1.13343728e+00 3.20362598e-01 -1.20832726e-01
-3.27130228e-01 -3.15205157e-01 3.26250643e-01 4.88291621e-01
-1.52255937e-01 -2.04903200e-01 -7.46369898e-01 4.21171576e-01
-1.56920397e+00 -1.07845306e+00 -7.73399994e-02 2.36528444e+00
4.58408624e-01 -7.92971775e-02 2.20087916e-01 4.17893291e-01
1.08660924e+00 1.67771548e-01 -7.35969782e-01 -1.30543575e-01
-2.36388236e-01 -8.92320648e-03 5.26908934e-01 1.75422832e-01
-9.83190179e-01 2.94217676e-01 6.35233688e+00 7.82640398e-01
-1.03087580e+00 -1.09300695e-01 3.35209519e-01 5.23174703e-01
-1.46316290e-01 -1.94581434e-01 -4.08083528e-01 8.23429644e-01
6.20331109e-01 -3.37103248e-01 6.54151559e-01 7.99096227e-01
2.18565658e-01 -1.94031656e-01 -9.05801415e-01 1.29165483e+00
1.73906192e-01 -9.62779820e-01 1.18214218e-02 1.42616227e-01
7.29400814e-01 -6.14247501e-01 4.57966477e-01 -3.68746489e-01
-2.11632907e-01 -6.65988684e-01 8.53014886e-01 9.98416245e-01
1.01717675e+00 -7.25352883e-01 7.48713613e-01 2.83199221e-01
-1.12884951e+00 -2.29724348e-01 -3.26355994e-01 3.37345541e-01
-1.56372055e-01 8.03100169e-01 -5.61049104e-01 9.38270450e-01
6.70314133e-01 6.35758340e-01 -7.62413859e-01 1.11390674e+00
-3.07680994e-01 4.71931428e-01 2.25849822e-01 3.77831697e-01
-2.70079195e-01 -3.14005286e-01 5.63969314e-01 1.06850731e+00
6.33298457e-01 4.43641171e-02 -2.48082891e-01 1.04689789e+00
-1.09015532e-01 6.21216111e-02 -4.27487403e-01 8.32295865e-02
6.60440087e-01 1.00544107e+00 -3.52867365e-01 -1.57386273e-01
-1.78029746e-01 1.03235543e+00 -2.57800668e-01 2.93239385e-01
-5.33177733e-01 -4.17717963e-01 7.42844641e-01 1.11890167e-01
2.92017132e-01 -9.33090523e-02 -1.92090914e-01 -1.06710899e+00
2.97844619e-01 -1.01718485e+00 1.59054652e-01 -9.19238389e-01
-1.29589784e+00 5.22716641e-01 -3.02491665e-01 -1.73404515e+00
-3.07294369e-01 -6.10077977e-01 -4.55305666e-01 1.10282683e+00
-1.60153937e+00 -1.07266915e+00 -8.53393912e-01 7.56954908e-01
-3.22508179e-02 -3.04772973e-01 6.86701000e-01 4.86536562e-01
-3.76888812e-01 8.77214789e-01 5.35471678e-01 2.88000464e-01
9.64917123e-01 -9.42280591e-01 2.85477102e-01 1.31958508e+00
-1.74189080e-02 5.29976964e-01 6.07041001e-01 -6.33185685e-01
-1.50646150e+00 -8.84133458e-01 3.09474736e-01 -2.30409905e-01
2.50747263e-01 -1.59344841e-02 -9.33025658e-01 -1.03962041e-01
-1.47677228e-01 -9.40681770e-02 3.52367431e-01 -1.26894161e-01
-6.82891369e-01 -9.41788077e-01 -1.32098007e+00 -5.92105370e-03
9.52197850e-01 -9.51612830e-01 -2.03752905e-01 -3.42292964e-01
3.17782730e-01 -6.97295964e-02 -1.00247407e+00 3.61036867e-01
9.94853437e-01 -1.41377747e+00 1.12827790e+00 2.39304066e-01
2.44360924e-01 -8.15743327e-01 -1.93912864e-01 -1.02747190e+00
-5.25818467e-01 -1.86971590e-01 8.18298087e-02 1.63130248e+00
8.41581356e-03 -4.63872790e-01 5.76132119e-01 1.67999059e-01
2.88941830e-01 -7.68697336e-02 -6.19315684e-01 -1.30871558e+00
-4.00229901e-01 -1.53391659e-01 1.13224173e+00 6.83115840e-01
-3.42452049e-01 -2.64595538e-01 -7.28843391e-01 5.46383321e-01
1.13416493e+00 4.09197807e-01 8.41152906e-01 -1.21594262e+00
-1.78590581e-01 -2.64354259e-01 -8.64356756e-01 -7.29248226e-01
-3.44567239e-01 -4.46969897e-01 6.52526915e-02 -1.25589526e+00
4.04407382e-01 -3.08080554e-01 -6.90970123e-01 -1.04247026e-01
-2.28715241e-01 3.47948402e-01 1.74592301e-01 5.88333130e-01
-4.73116338e-01 1.14056446e-01 1.23568249e+00 -3.87069374e-01
4.25423533e-02 1.06832035e-01 -6.78402007e-01 2.53587753e-01
5.65672517e-01 -2.38430396e-01 -3.28843981e-01 -2.27220580e-01
-1.35398269e-01 2.25885361e-01 5.00117183e-01 -1.56674111e+00
1.86565995e-01 -8.67887586e-02 4.44025576e-01 -5.41979194e-01
7.06815645e-02 -7.83742368e-01 5.18876493e-01 4.26815540e-01
-1.70152918e-01 -2.37657860e-01 -2.46231794e-01 6.00005984e-01
-6.05941176e-01 -1.86418802e-01 1.04000950e+00 -4.55838181e-02
-1.11741710e+00 1.50264889e-01 1.49025291e-01 -2.97894597e-01
6.67950213e-01 -4.48416710e-01 -5.75406194e-01 -3.51117671e-01
-2.30105549e-01 -2.96439588e-01 8.06842864e-01 1.59008250e-01
6.95834577e-01 -1.46389759e+00 -6.33318007e-01 2.98681021e-01
5.69314480e-01 -6.05414093e-01 6.29805684e-01 3.66210967e-01
-4.01232541e-01 2.15708464e-01 -6.85837090e-01 -6.08538389e-01
-1.33581138e+00 9.14419293e-01 3.09368223e-01 -2.28765383e-01
-2.74031073e-01 2.25325629e-01 3.06127429e-01 -1.04108946e-02
1.53579846e-01 1.05228089e-01 -3.28563601e-01 -4.16951701e-02
7.96773791e-01 7.08341837e-01 1.42002553e-01 -8.56321990e-01
-3.45834762e-01 1.08917606e+00 2.57868141e-01 -1.55280262e-01
9.85055447e-01 -4.89382356e-01 -3.46778445e-02 1.37740001e-01
1.25947630e+00 -8.30668956e-03 -1.31249189e+00 -5.07717550e-01
-3.84639502e-02 -9.64943647e-01 -1.82753250e-01 -1.26527703e+00
-9.57146764e-01 9.44062293e-01 1.25452268e+00 8.52230657e-03
1.64930272e+00 -4.36766624e-01 5.73694587e-01 6.06148969e-03
6.84302390e-01 -1.09933889e+00 3.69005829e-01 -5.89872226e-02
8.93873215e-01 -1.14335215e+00 9.11294203e-03 -1.94709182e-01
-5.09669483e-01 1.11296856e+00 3.01486641e-01 1.79961443e-01
3.17610592e-01 -1.38883159e-01 2.38482654e-01 5.45752868e-02
1.99567080e-01 -1.81508407e-01 4.75357950e-01 9.90828812e-01
-1.32931188e-01 1.03931345e-01 -9.29760113e-02 7.24500418e-01
6.38086870e-02 2.30693966e-01 3.23145151e-01 4.07020062e-01
-4.15024489e-01 -1.36037278e+00 -8.99837136e-01 -7.23230913e-02
-3.18952084e-01 4.07704860e-01 -2.73896903e-01 4.11951423e-01
3.95358741e-01 1.13112497e+00 -3.39810133e-01 -7.96468973e-01
5.31119406e-01 -1.21413387e-01 5.40844023e-01 2.34123971e-02
-8.04570541e-02 -3.34172666e-01 -6.60517141e-02 -7.34543145e-01
-6.01997435e-01 -6.03178501e-01 -7.71077037e-01 -2.96574146e-01
-2.87940968e-02 -3.18483338e-02 9.68912721e-01 6.03699028e-01
4.27013576e-01 2.39448577e-01 1.09879124e+00 -5.75717509e-01
-2.55761743e-01 -6.02450371e-01 -9.45980251e-01 7.70854175e-01
4.35021579e-01 -7.52644479e-01 -9.90371630e-02 2.10354328e-01] | [11.790627479553223, -1.893448829650879] |
b6484b25-bba9-4bac-bc17-684bffc16986 | inspecting-the-geographical | 2305.11080 | null | https://arxiv.org/abs/2305.11080v1 | https://arxiv.org/pdf/2305.11080v1.pdf | Inspecting the Geographical Representativeness of Images from Text-to-Image Models | Recent progress in generative models has resulted in models that produce both realistic as well as relevant images for most textual inputs. These models are being used to generate millions of images everyday, and hold the potential to drastically impact areas such as generative art, digital marketing and data augmentation. Given their outsized impact, it is important to ensure that the generated content reflects the artifacts and surroundings across the globe, rather than over-representing certain parts of the world. In this paper, we measure the geographical representativeness of common nouns (e.g., a house) generated through DALL.E 2 and Stable Diffusion models using a crowdsourced study comprising 540 participants across 27 countries. For deliberately underspecified inputs without country names, the generated images most reflect the surroundings of the United States followed by India, and the top generations rarely reflect surroundings from all other countries (average score less than 3 out of 5). Specifying the country names in the input increases the representativeness by 1.44 points on average for DALL.E 2 and 0.75 for Stable Diffusion, however, the overall scores for many countries still remain low, highlighting the need for future models to be more geographically inclusive. Lastly, we examine the feasibility of quantifying the geographical representativeness of generated images without conducting user studies. | ['Danish Pruthi', 'R. Venkatesh Babu', 'Abhipsa Basu'] | 2023-05-18 | null | null | null | null | ['marketing'] | ['miscellaneous'] | [ 7.61865266e-03 3.04937541e-01 -1.13095399e-02 -2.50784252e-02
-6.93493605e-01 -9.77886140e-01 1.17910838e+00 1.44814536e-01
-4.05355334e-01 6.73733711e-01 8.39555740e-01 -2.15930194e-01
2.60511160e-01 -1.13710105e+00 -8.31672728e-01 -2.53564239e-01
4.20833975e-01 4.22339439e-01 -1.26837477e-01 -1.27354279e-01
6.88573495e-02 2.75330901e-01 -1.35554552e+00 2.14981705e-01
1.10556102e+00 2.67501473e-01 1.67760730e-01 4.76132989e-01
-3.04442972e-01 4.07610148e-01 -1.12165070e+00 -8.55973065e-01
2.40785360e-01 -6.40708625e-01 -3.04816663e-01 2.18859240e-01
7.69894242e-01 -3.95666540e-01 -1.95150882e-01 1.03476787e+00
5.69044709e-01 1.05368115e-01 9.89515603e-01 -1.11606455e+00
-1.61868727e+00 7.90484250e-01 -6.20524168e-01 6.65793717e-02
5.82017720e-01 3.56658906e-01 8.37708354e-01 -9.12327826e-01
1.35404801e+00 1.35736990e+00 5.67485332e-01 9.88237783e-02
-1.19218266e+00 -7.61707246e-01 2.58797586e-01 -6.07115448e-01
-1.58521307e+00 -3.14080477e-01 3.65132838e-01 -6.45450830e-01
6.78465724e-01 3.45624238e-01 9.19668555e-01 1.29373276e+00
2.72252951e-02 2.92710215e-01 1.10630393e+00 -2.14452252e-01
1.67140126e-01 4.22623992e-01 -6.61454320e-01 3.08663309e-01
6.37713730e-01 -3.77755016e-01 -3.44382733e-01 -1.02063701e-01
1.09479690e+00 -1.83106363e-01 -4.57798466e-02 1.13093004e-01
-1.41494787e+00 9.11380947e-01 7.01267421e-01 2.47150064e-01
-5.39357543e-01 5.95692582e-02 -1.79947063e-01 -1.91457912e-01
8.50192010e-01 6.51701927e-01 2.39485398e-01 -1.57577977e-01
-1.07490218e+00 6.67584419e-01 6.41457736e-01 1.23576295e+00
5.65102458e-01 7.56598413e-02 -2.07880691e-01 9.93816972e-01
-2.94251298e-03 9.26428556e-01 3.67102847e-02 -1.01300728e+00
5.31127393e-01 7.46076584e-01 5.29059589e-01 -1.32487571e+00
-2.17187166e-01 -5.91828525e-01 -6.51161969e-01 3.58715355e-02
6.26185417e-01 -4.40251797e-01 -1.15809381e+00 1.71042001e+00
1.74837664e-01 -3.75203401e-01 -2.53515184e-01 9.81482804e-01
8.38424325e-01 9.33478773e-01 4.64903802e-01 3.03806543e-01
1.33735716e+00 -5.32688379e-01 -4.96755153e-01 -5.38784146e-01
3.08807343e-01 -1.17828631e+00 1.40715110e+00 -3.32934633e-02
-1.16221440e+00 -4.69760835e-01 -8.34740937e-01 2.50894520e-02
-6.77403033e-01 -2.90756226e-02 5.70143759e-01 7.97239661e-01
-1.15834939e+00 1.71744511e-01 -1.15162604e-01 -7.89189398e-01
4.86338705e-01 -2.07006827e-01 -3.56178701e-01 -3.14882398e-01
-1.09893680e+00 9.23681617e-01 4.59701046e-02 -4.87924337e-01
-4.51884300e-01 -6.81527436e-01 -6.94313943e-01 -1.74245507e-01
-1.43692970e-01 -7.75788486e-01 9.50559914e-01 -1.09867656e+00
-6.84828997e-01 8.12816620e-01 1.25056570e-02 -6.59499094e-02
7.78291464e-01 -7.55259022e-02 -7.08053946e-01 -6.92278817e-02
6.65920615e-01 1.16994083e+00 4.05876696e-01 -1.49863315e+00
-6.63478076e-01 -1.96334422e-01 2.70136863e-01 5.45117080e-01
-3.74297023e-01 1.41673237e-01 -5.85084677e-01 -1.02887464e+00
-2.86487415e-02 -1.03418171e+00 -3.44172508e-01 -1.50834277e-01
-3.02039713e-01 2.87520707e-01 3.52238119e-01 -1.01781380e+00
1.23359787e+00 -2.01592255e+00 -3.03080320e-01 3.34185392e-01
1.22471966e-01 -3.01668912e-01 -1.95796356e-01 6.89010859e-01
2.93488979e-01 7.26822376e-01 9.82806087e-03 -8.24915022e-02
2.10250653e-02 -1.23542815e-01 -1.42085955e-01 3.70503813e-01
1.48378238e-01 9.17870343e-01 -1.01747990e+00 -3.20598543e-01
8.22480619e-02 7.03421295e-01 -3.88349652e-01 -5.77110529e-01
-2.62968600e-01 2.15978190e-01 -6.13414757e-02 5.57311296e-01
6.77003622e-01 -3.17065299e-01 -1.17961898e-01 1.68404028e-01
-2.48861387e-01 2.62815095e-02 -1.26813221e+00 1.30318093e+00
-4.60152030e-01 8.61430526e-01 -2.34495372e-01 3.71819019e-01
8.43750954e-01 -1.03232153e-02 7.28297308e-02 -1.01922297e+00
-2.09820956e-01 2.15038881e-01 4.19026017e-02 -1.99428976e-01
1.09155464e+00 -1.37642041e-01 -2.69154459e-01 7.53014028e-01
-4.98457253e-01 -3.97953361e-01 2.54613549e-01 4.85865712e-01
6.21543884e-01 1.01099000e-03 1.75908078e-02 -2.55112767e-01
-4.73708570e-01 5.18085241e-01 8.19249004e-02 8.48539948e-01
2.17554688e-01 1.09185827e+00 3.71459305e-01 -2.53501594e-01
-1.52869737e+00 -1.16792095e+00 5.70743456e-02 8.38668823e-01
1.11891448e-01 -4.09615457e-01 -1.02352345e+00 -2.37594724e-01
7.32269287e-02 1.27854717e+00 -6.28798723e-01 8.90585594e-03
-7.24853054e-02 -6.77368522e-01 5.43280840e-01 3.86820734e-01
2.75873512e-01 -8.51617277e-01 -3.06376398e-01 1.81584463e-01
-5.57735801e-01 -9.85077739e-01 -6.54476345e-01 -7.81595707e-01
-4.59819645e-01 -5.25363624e-01 -1.37709689e+00 -5.85768819e-01
7.86666155e-01 3.40546846e-01 1.34037447e+00 -3.24517488e-01
-1.25214383e-02 3.94814849e-01 -2.62719572e-01 -7.11680889e-01
-5.55800915e-01 1.15832135e-01 -1.76732942e-01 -1.87667280e-01
2.80531555e-01 -2.27670565e-01 -6.54692948e-01 2.95132041e-01
-9.45187211e-01 3.05608481e-01 5.43193161e-01 1.15356162e-01
4.24449891e-01 -3.16404402e-02 5.98417997e-01 -1.07795179e+00
1.00528693e+00 -9.03459668e-01 -1.54636770e-01 1.13793172e-01
-4.08523113e-01 -3.74202192e-01 3.59678507e-01 -7.60781467e-01
-1.14126861e+00 -3.84944826e-01 2.96994597e-01 7.18247816e-02
-9.20568965e-03 5.85763991e-01 8.89067575e-02 4.01968122e-01
1.17972732e+00 -3.26560228e-03 -2.01806039e-01 -1.61178321e-01
6.77871406e-01 7.63924003e-01 3.02166611e-01 -3.28955650e-01
8.01200449e-01 6.10550702e-01 -6.40625536e-01 -1.04580569e+00
-2.97894448e-01 -9.04815868e-02 -9.75224972e-02 -5.02087831e-01
7.77101934e-01 -1.27655816e+00 -7.61814713e-02 2.73124814e-01
-1.01964986e+00 -3.39832127e-01 -4.21380877e-01 5.11572063e-01
-1.43448502e-01 -1.43436059e-01 -3.37342352e-01 -7.42210031e-01
-1.24452800e-01 -1.00394404e+00 1.05573225e+00 3.48259747e-01
-8.42115045e-01 -8.81756127e-01 -1.22472286e-01 4.39450592e-01
5.04345655e-01 5.16561925e-01 8.42784345e-01 -2.74088115e-01
-5.81249177e-01 -4.64127421e-01 -4.85889703e-01 -1.70215443e-01
3.99905980e-01 2.99146473e-01 -7.52661049e-01 2.11437885e-02
-5.70844769e-01 8.92296731e-02 3.89307290e-01 4.73195493e-01
3.23712021e-01 -4.78841752e-01 -2.84397393e-01 9.55881700e-02
1.36041200e+00 1.38562322e-01 8.18899333e-01 4.46555376e-01
6.85167551e-01 6.58166230e-01 3.33307356e-01 4.61490661e-01
6.85168326e-01 3.13045770e-01 1.20917790e-01 -2.72521317e-01
-4.65031922e-01 -7.86861062e-01 1.60948545e-01 3.67898017e-01
-2.06609070e-01 -5.75852215e-01 -1.08814633e+00 9.25886869e-01
-1.29790545e+00 -1.12113488e+00 -2.60197490e-01 2.24998403e+00
6.06168628e-01 9.98405814e-02 2.46654376e-01 -5.78654945e-01
9.10988212e-01 2.55592138e-01 -4.65200216e-01 -3.43189687e-01
-6.39791906e-01 -7.39174411e-02 5.80842376e-01 5.40769219e-01
-6.96853399e-01 9.39316154e-01 6.95031643e+00 7.44036436e-01
-9.18363929e-01 -1.00922640e-02 1.17924273e+00 -3.98490667e-01
-1.06902385e+00 1.08542919e-01 -5.77156425e-01 6.80244505e-01
8.29090893e-01 -3.69384497e-01 4.63620901e-01 6.67740822e-01
4.45628911e-01 -4.33888227e-01 -3.81986469e-01 9.56791699e-01
1.78385288e-01 -1.56904697e+00 4.51909184e-01 3.61676604e-01
1.45801425e+00 -1.67008713e-01 3.71911943e-01 -1.15022510e-01
7.30273008e-01 -1.24208844e+00 1.31286001e+00 5.57670891e-01
1.18235791e+00 -7.78239906e-01 3.44867915e-01 1.76194400e-01
-8.28715026e-01 1.12459846e-01 -3.21236134e-01 -2.67478883e-01
3.37490410e-01 6.76627934e-01 -9.65913355e-01 8.91443416e-02
6.92886353e-01 3.37780863e-01 -8.14795852e-01 7.81978607e-01
-2.39957735e-01 4.41496193e-01 -5.20086467e-01 -2.38455072e-01
4.89998341e-01 -3.49095941e-01 3.91294330e-01 1.28740370e+00
8.27737629e-01 -1.05790645e-01 -2.31901735e-01 9.45428371e-01
-4.13920939e-01 2.98266023e-01 -9.85932231e-01 -3.54684234e-01
7.66345501e-01 9.75667894e-01 -1.02569771e+00 -3.14255744e-01
-4.31226909e-01 9.10436988e-01 3.25244777e-02 6.84067726e-01
-9.40196991e-01 -1.39594823e-01 4.92624104e-01 7.89046586e-01
-1.71392679e-01 -1.56580254e-01 -5.89929283e-01 -6.73843324e-01
2.24239733e-02 -9.46322322e-01 3.27525251e-02 -1.32846177e+00
-1.08310711e+00 4.88170832e-01 1.47983339e-02 -1.02584124e+00
-9.18984339e-02 -4.10404196e-03 -3.56466830e-01 1.25098777e+00
-6.63620472e-01 -1.31098926e+00 -3.57231140e-01 1.51515156e-01
4.70176101e-01 1.78190202e-01 5.13093174e-01 2.32506767e-01
-2.83259600e-01 4.47264791e-01 1.40217692e-01 -4.90544029e-02
8.37801516e-01 -1.00699615e+00 1.12804890e+00 9.33092952e-01
2.67847002e-01 9.48160887e-01 7.19671190e-01 -1.16975212e+00
-6.89329207e-01 -1.04892313e+00 1.06447625e+00 -6.87806308e-01
4.67137069e-01 -2.63633072e-01 -3.12794298e-01 7.43544400e-01
4.64972287e-01 -6.98379755e-01 6.92064404e-01 1.44654019e-02
-2.82096565e-01 2.16570035e-01 -1.16226935e+00 1.32266283e+00
1.20622945e+00 -4.04781699e-01 -1.19039463e-02 4.35987264e-01
5.63398361e-01 -2.23947152e-01 -8.34093094e-01 -3.71573657e-01
6.29284382e-01 -8.83801699e-01 9.10308123e-01 -2.42733970e-01
8.09895337e-01 -1.85172722e-01 -6.57760948e-02 -1.62537706e+00
-6.36207581e-01 -5.86259186e-01 3.59738261e-01 1.58665037e+00
9.76132810e-01 -6.32974207e-01 7.32253909e-01 1.04360723e+00
4.08730432e-02 -1.72243968e-01 -5.08444667e-01 -4.44102257e-01
2.04336092e-01 -4.10647959e-01 8.59876990e-01 1.07672226e+00
-4.34506804e-01 2.57069439e-01 -2.13104844e-01 -7.40866661e-02
3.22211653e-01 -1.29860818e-01 8.30484569e-01 -9.03300464e-01
2.12203592e-01 -7.01545119e-01 -1.48479462e-01 -6.36002779e-01
-5.02988577e-01 -6.20369732e-01 -4.25825596e-01 -2.28083873e+00
3.36085588e-01 -7.01879025e-01 4.85804975e-01 9.29822549e-02
-2.46310234e-01 7.12335646e-01 3.19016814e-01 2.75907278e-01
-2.67677754e-02 5.32624349e-02 1.65862298e+00 -1.61244124e-01
-1.83730155e-01 -2.42993221e-01 -1.43027115e+00 6.52160704e-01
7.51944244e-01 -2.96580046e-01 -5.96280515e-01 -7.89458513e-01
6.80492222e-01 -5.36890745e-01 3.61030459e-01 -9.56117213e-01
-1.33542372e-02 -3.85038853e-01 1.03748548e+00 -2.96604723e-01
4.41872597e-01 -6.81322157e-01 9.59544063e-01 1.30575106e-01
-1.30607128e-01 3.95123929e-01 2.35007614e-01 3.66725683e-01
8.94194990e-02 3.30583230e-02 4.78450000e-01 -4.10935462e-01
-5.31309962e-01 -6.27445802e-02 -5.17391324e-01 3.45799387e-01
1.06590831e+00 -7.59660006e-01 -5.86036801e-01 -9.87215757e-01
-3.86760950e-01 -2.27578685e-01 1.16164780e+00 6.58566594e-01
3.77739131e-01 -1.44492888e+00 -9.69561338e-01 1.86294001e-02
2.83032864e-01 -2.40676418e-01 3.38815331e-01 2.13839948e-01
-8.51882875e-01 1.28642946e-01 -1.06011331e-01 -1.59834012e-01
-7.37899899e-01 2.36314818e-01 -7.14415610e-02 2.05169916e-01
-4.75211471e-01 6.62466884e-01 3.83533657e-01 -2.99022466e-01
-2.47128710e-01 -2.09296569e-01 2.66287476e-02 4.84600037e-01
5.64366162e-01 5.66967666e-01 -2.61035562e-01 -1.09904611e+00
-8.55469778e-02 2.59798527e-01 1.63369238e-01 -5.88905871e-01
1.21489060e+00 -6.76006824e-02 2.00627595e-01 3.31166953e-01
8.23779941e-01 5.01782894e-01 -1.21970451e+00 1.68588862e-01
-4.48964179e-01 -6.37879372e-01 -2.79855818e-01 -1.14008725e+00
-8.87647271e-01 3.61391038e-01 2.69015640e-01 2.78971285e-01
7.59954274e-01 1.98088199e-01 6.95170105e-01 -4.25339878e-01
4.21593726e-01 -1.13590109e+00 -2.01621860e-01 3.20569947e-02
1.10458827e+00 -8.34646106e-01 9.45697129e-02 -3.67090255e-01
-1.15746212e+00 5.16978443e-01 4.73923206e-01 8.46130028e-02
2.97595650e-01 7.93210417e-02 9.57829207e-02 -1.02998570e-01
-2.83174038e-01 -1.36143073e-01 3.38536173e-01 8.96985292e-01
5.12498975e-01 4.46504593e-01 -2.83670753e-01 4.03820127e-01
-5.79803526e-01 -2.82649755e-01 7.68730819e-01 5.74398518e-01
-2.18767092e-01 -5.82152247e-01 -7.44409323e-01 7.32003987e-01
-4.20200080e-01 -2.79020011e-01 -9.07079041e-01 1.11138451e+00
1.55554578e-01 9.79962587e-01 4.40863460e-01 -2.15736568e-01
4.81152952e-01 -1.24109782e-01 3.39985460e-01 -5.60629606e-01
-6.87947273e-01 7.80152455e-02 3.42652380e-01 -6.02467768e-02
-1.52298108e-01 -7.40804493e-01 -8.84069324e-01 -7.82723010e-01
-3.89460213e-02 -2.97029942e-01 8.79968226e-01 3.03271383e-01
6.00626290e-01 1.73020810e-01 1.43590733e-01 -7.11717725e-01
1.77784979e-01 -1.17112076e+00 -5.23792267e-01 6.79108024e-01
-2.34382644e-01 -4.69840378e-01 -1.63050458e-01 1.29430354e-01] | [11.404382705688477, 0.6188852190971375] |
2271fdcc-e83e-4760-9a32-0bedce8c2c5d | latent-nerf-for-shape-guided-generation-of-3d | 2211.07600 | null | https://arxiv.org/abs/2211.07600v1 | https://arxiv.org/pdf/2211.07600v1.pdf | Latent-NeRF for Shape-Guided Generation of 3D Shapes and Textures | Text-guided image generation has progressed rapidly in recent years, inspiring major breakthroughs in text-guided shape generation. Recently, it has been shown that using score distillation, one can successfully text-guide a NeRF model to generate a 3D object. We adapt the score distillation to the publicly available, and computationally efficient, Latent Diffusion Models, which apply the entire diffusion process in a compact latent space of a pretrained autoencoder. As NeRFs operate in image space, a naive solution for guiding them with latent score distillation would require encoding to the latent space at each guidance step. Instead, we propose to bring the NeRF to the latent space, resulting in a Latent-NeRF. Analyzing our Latent-NeRF, we show that while Text-to-3D models can generate impressive results, they are inherently unconstrained and may lack the ability to guide or enforce a specific 3D structure. To assist and direct the 3D generation, we propose to guide our Latent-NeRF using a Sketch-Shape: an abstract geometry that defines the coarse structure of the desired object. Then, we present means to integrate such a constraint directly into a Latent-NeRF. This unique combination of text and shape guidance allows for increased control over the generation process. We also show that latent score distillation can be successfully applied directly on 3D meshes. This allows for generating high-quality textures on a given geometry. Our experiments validate the power of our different forms of guidance and the efficiency of using latent rendering. Implementation is available at https://github.com/eladrich/latent-nerf | ['Daniel Cohen-Or', 'Raja Giryes', 'Or Patashnik', 'Elad Richardson', 'Gal Metzer'] | 2022-11-14 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Metzer_Latent-NeRF_for_Shape-Guided_Generation_of_3D_Shapes_and_Textures_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Metzer_Latent-NeRF_for_Shape-Guided_Generation_of_3D_Shapes_and_Textures_CVPR_2023_paper.pdf | cvpr-2023-1 | ['text-to-3d'] | ['computer-vision'] | [ 3.48564148e-01 2.98854500e-01 2.02497005e-01 -1.86629638e-01
-7.66983390e-01 -7.62793243e-01 9.02485967e-01 -1.82700172e-01
5.97513504e-02 3.31807852e-01 4.04085755e-01 -2.41515204e-01
4.31872532e-03 -1.26280355e+00 -7.19430447e-01 -6.39345288e-01
1.22563511e-01 6.06579483e-01 5.11533096e-02 -1.87717512e-01
3.36128145e-01 8.06293786e-01 -1.34389043e+00 3.19341153e-01
7.59717584e-01 6.31021380e-01 3.05323184e-01 9.65403795e-01
-2.50736296e-01 2.52962559e-01 -4.88966763e-01 -2.11928591e-01
4.40815866e-01 -5.57859600e-01 -6.06806934e-01 1.67120382e-01
6.01597250e-01 -6.34048581e-01 -1.71651140e-01 7.19055891e-01
3.77620429e-01 1.12440459e-01 1.02027202e+00 -1.02437830e+00
-8.67172778e-01 1.72174186e-01 -2.49336556e-01 -5.52187979e-01
3.73098522e-01 1.41481329e-02 9.09111559e-01 -1.09506845e+00
9.47723389e-01 1.29371786e+00 5.25081515e-01 7.52458811e-01
-1.62224364e+00 -3.30612957e-01 2.51991972e-02 -6.43982410e-01
-1.22692525e+00 -2.49992669e-01 1.05794764e+00 -7.36788094e-01
7.69172847e-01 2.86843300e-01 9.70746458e-01 9.76910353e-01
2.24354014e-01 8.08665752e-01 1.01946151e+00 -5.49332321e-01
3.76666397e-01 -1.52542308e-01 -2.91373193e-01 9.69005764e-01
-1.69129074e-01 3.06326449e-01 -3.62384230e-01 -4.01707053e-01
1.58905923e+00 -2.30199378e-03 -1.74783483e-01 -6.43638015e-01
-1.05590653e+00 1.15110755e+00 4.46984291e-01 1.00259975e-01
-2.01219290e-01 4.77518618e-01 -3.12625952e-02 1.09942541e-01
7.90356040e-01 6.01429701e-01 -1.67947263e-01 -1.32602975e-01
-1.11169744e+00 5.62253237e-01 8.16834688e-01 8.62567544e-01
7.97283173e-01 1.61245927e-01 -2.31635228e-01 7.62178123e-01
2.96890855e-01 5.31323493e-01 7.95262530e-02 -1.51084864e+00
3.29936631e-02 4.37420338e-01 2.65261769e-01 -1.11699295e+00
-1.94222690e-03 -2.16646925e-01 -8.05636287e-01 8.41106236e-01
2.65207916e-01 -8.46599266e-02 -1.21431530e+00 1.79278362e+00
2.82876253e-01 -5.53771481e-02 -1.81099012e-01 7.52996087e-01
3.73671353e-01 1.07257783e+00 -2.35086754e-01 1.88465625e-01
9.49247241e-01 -8.12552691e-01 -2.93361932e-01 1.19050227e-01
7.20358014e-01 -7.91586936e-01 1.31010985e+00 3.54021847e-01
-1.36704218e+00 -4.54197973e-01 -1.03935146e+00 -2.30926260e-01
-1.02212176e-01 -2.99309939e-02 7.32472837e-01 3.88413668e-01
-1.34482670e+00 8.83665562e-01 -1.01950395e+00 -3.34071964e-01
3.01142842e-01 1.96731791e-01 -2.89974213e-01 1.13476977e-01
-7.29946196e-01 5.30964971e-01 -5.98624647e-02 -2.85305560e-01
-8.72758865e-01 -6.78371370e-01 -7.26559162e-01 -1.44288644e-01
1.39841765e-01 -1.25261390e+00 1.17468536e+00 -7.64902949e-01
-1.81415045e+00 8.14613223e-01 -1.39741004e-01 -8.75245705e-02
6.32865608e-01 -2.26359680e-01 2.84189075e-01 1.70784175e-01
4.47711833e-02 1.03396463e+00 9.71257508e-01 -1.35221481e+00
-9.94450003e-02 7.20877573e-02 3.41606200e-01 2.31535822e-01
-2.75307506e-01 -3.82747561e-01 -5.22282481e-01 -1.11358380e+00
2.60946989e-01 -8.99748325e-01 -3.34703028e-01 5.53210437e-01
-3.10877562e-01 -2.29487240e-01 9.48351204e-01 -3.53675187e-01
8.36756408e-01 -2.09512281e+00 5.03773570e-01 2.97313273e-01
3.17315698e-01 -1.02074489e-01 -3.11320364e-01 4.78023171e-01
4.14896943e-02 4.75481480e-01 -5.16738594e-01 -6.43069983e-01
2.05659464e-01 3.89975965e-01 -4.28275704e-01 1.36400819e-01
2.19614252e-01 9.63043094e-01 -8.20133924e-01 -1.64470509e-01
3.55306178e-01 8.50602388e-01 -1.22964060e+00 2.77366757e-01
-5.44810772e-01 4.56202447e-01 -5.74522913e-01 1.71613380e-01
4.66085523e-01 -2.86150217e-01 -1.95791543e-01 -1.47924110e-01
-2.11609796e-01 2.43506745e-01 -1.10895300e+00 2.08823180e+00
-7.22740948e-01 4.93996382e-01 1.79677278e-01 -5.14884293e-01
1.06084299e+00 1.51898548e-01 4.52776104e-01 -2.42832959e-01
-9.14264098e-02 3.41762602e-01 -4.75914717e-01 1.45734688e-02
6.10817254e-01 -3.44895989e-01 1.32586702e-03 8.43814373e-01
-5.37315086e-02 -8.42545569e-01 -3.64951119e-02 4.73264515e-01
9.12906945e-01 5.55328548e-01 -2.85508692e-01 -2.84373790e-01
1.46932319e-01 -1.33444130e-01 1.13902293e-01 7.09073484e-01
4.99294370e-01 1.07539153e+00 3.71200651e-01 -3.93994093e-01
-1.49387717e+00 -1.39510787e+00 -4.17248532e-02 7.49285758e-01
-2.44655579e-01 -6.00188732e-01 -9.24066901e-01 -5.63060105e-01
-6.80955052e-02 6.31582975e-01 -7.42479324e-01 5.15900552e-02
-6.89344883e-01 -3.40494305e-01 3.61199111e-01 5.08539259e-01
2.17963144e-01 -8.65687072e-01 -4.70580518e-01 3.31894487e-01
6.06959946e-02 -5.71176648e-01 -6.47753477e-01 -2.28536874e-02
-1.12705410e+00 -6.69598639e-01 -1.22639942e+00 -6.37715757e-01
1.04865420e+00 -4.81604338e-02 1.11988389e+00 2.34041438e-01
-1.40017897e-01 5.12417316e-01 -6.40431792e-02 -1.04166241e-02
-6.54098749e-01 1.11899219e-01 -1.34625033e-01 -1.34836197e-01
-2.25545764e-01 -9.86513853e-01 -7.77870476e-01 2.17278928e-01
-9.92067456e-01 6.46103382e-01 2.22758114e-01 8.02670062e-01
6.64771020e-01 -2.01517552e-01 1.21999197e-01 -6.27376437e-01
7.12790906e-01 3.44431115e-04 -6.92355275e-01 -2.32533053e-01
-3.66258383e-01 4.45860684e-01 4.03465986e-01 -5.07917285e-01
-8.50247264e-01 -6.53455406e-02 -5.17590225e-01 -4.75678533e-01
-1.55790105e-01 3.24557841e-01 8.92318338e-02 -8.51876941e-03
6.15326762e-01 5.41733317e-02 -6.35457411e-02 -5.90165496e-01
7.53471494e-01 2.27158576e-01 2.85270423e-01 -1.12244987e+00
9.50607657e-01 6.97339773e-01 -3.92493345e-02 -5.30196965e-01
-6.13306165e-01 2.14442804e-01 -5.53356111e-01 -3.73421498e-02
1.00989854e+00 -4.58836406e-01 -3.49645287e-01 4.15176570e-01
-1.33042431e+00 -9.17401433e-01 -7.30271995e-01 1.09656252e-01
-1.03432775e+00 1.22019291e-01 -7.60786951e-01 -6.49409592e-01
-2.17199281e-01 -1.14042759e+00 1.45584142e+00 -2.04279393e-01
-4.51129019e-01 -1.20690429e+00 2.02774003e-01 -1.15580253e-01
5.41884124e-01 4.86338019e-01 1.11186874e+00 8.84599090e-02
-8.68009865e-01 -1.39880925e-01 -6.89044446e-02 2.91294277e-01
2.55799741e-01 1.50792643e-01 -7.52842546e-01 -2.31868520e-01
-6.13758452e-02 -2.08372384e-01 7.66456246e-01 4.29506123e-01
1.00162792e+00 -3.88106406e-01 -2.73284644e-01 9.16755319e-01
1.36939359e+00 -6.51274249e-03 4.02597159e-01 1.36670962e-04
8.21999490e-01 4.49751794e-01 3.04226987e-02 3.38711888e-01
1.35773614e-01 7.12468088e-01 3.92234400e-02 -3.52194548e-01
-3.71239901e-01 -6.96429014e-01 3.40565264e-01 9.84680355e-01
-3.20626616e-01 -1.18925482e-01 -7.72558033e-01 4.31399196e-01
-1.64477062e+00 -7.19245255e-01 1.27512112e-01 1.97560084e+00
8.83749664e-01 1.15566239e-01 -8.29971209e-02 7.28368107e-03
3.67776155e-01 2.21596599e-01 -4.03824270e-01 -6.22477233e-01
6.38006553e-02 4.64285254e-01 1.55446976e-01 1.08039975e+00
-7.31980443e-01 1.06032002e+00 6.34505367e+00 9.17916894e-01
-1.18714547e+00 1.18147008e-01 6.04100525e-01 -7.42018297e-02
-1.01218474e+00 1.28619149e-01 -6.45384133e-01 -9.05780029e-03
4.74849612e-01 -3.47719908e-01 4.59762663e-01 7.47174740e-01
3.66879940e-01 1.30568057e-01 -1.23633087e+00 7.79647470e-01
-1.79169819e-01 -1.64710963e+00 5.29637575e-01 2.95824766e-01
9.75357592e-01 -2.93727875e-01 2.41885409e-01 5.59813641e-02
5.98796427e-01 -1.03284287e+00 8.73766601e-01 6.33298874e-01
1.21435714e+00 -7.20342457e-01 4.58406284e-03 4.52204019e-01
-1.09826183e+00 3.75780642e-01 -3.57688069e-01 -8.57841000e-02
3.87480378e-01 7.10062444e-01 -4.35959488e-01 2.60278344e-01
1.99417293e-01 5.56367218e-01 -1.36211514e-01 5.79140961e-01
-3.32999557e-01 4.15349334e-01 -5.79486728e-01 2.28125408e-01
2.81135619e-01 -3.42138827e-01 7.15815246e-01 9.92717743e-01
7.32992291e-01 3.25142980e-01 2.38612041e-01 1.45037627e+00
9.13166404e-02 4.00907779e-03 -8.86307240e-01 -1.96944922e-01
1.28021330e-01 1.00514197e+00 -8.41581702e-01 -3.73648971e-01
1.09019950e-02 1.29821730e+00 1.65642738e-01 5.94843924e-01
-5.98795474e-01 -2.92904854e-01 5.57414889e-01 3.74013841e-01
1.73084900e-01 -8.29292655e-01 -4.79947776e-01 -1.10790014e+00
-1.31747678e-01 -5.76654732e-01 -2.00126916e-01 -1.17219484e+00
-1.11874199e+00 6.80623710e-01 3.38977762e-02 -8.78281355e-01
-3.86393994e-01 -5.38370430e-01 -4.84622866e-01 1.20562720e+00
-1.09116066e+00 -1.19387031e+00 -2.13308811e-01 5.74702680e-01
6.10457122e-01 2.93517765e-02 1.09626222e+00 -3.60305943e-02
1.20825380e-01 3.30795765e-01 -1.73446268e-01 6.58582821e-02
4.77971226e-01 -1.24752557e+00 7.56278932e-01 6.39160454e-01
3.04950058e-01 8.81094098e-01 5.21070600e-01 -8.66764247e-01
-1.24877226e+00 -9.68284488e-01 5.94791293e-01 -6.29981935e-01
4.34146792e-01 -6.69982433e-01 -7.42994249e-01 5.67682087e-01
1.27650052e-01 -3.06746185e-01 3.77198994e-01 -2.95426343e-02
-1.85004339e-01 3.18561465e-01 -9.70843732e-01 9.31066036e-01
1.25396836e+00 -7.16499090e-01 -4.09596771e-01 3.12489927e-01
7.77962804e-01 -5.30031919e-01 -7.40160286e-01 2.99694031e-01
5.39433062e-01 -9.48729217e-01 9.68827724e-01 -2.20886961e-01
6.43928349e-01 -3.68411958e-01 -2.25837052e-01 -1.30107093e+00
-5.20716369e-01 -8.39816511e-01 -1.40209362e-01 9.89724994e-01
2.94867784e-01 -4.63639528e-01 1.10892391e+00 6.48376584e-01
-2.60943979e-01 -9.91955161e-01 -5.28233469e-01 -5.70662379e-01
5.32088339e-01 -5.52264512e-01 6.29254937e-01 8.05899501e-01
-5.17077148e-01 -1.01814970e-01 -1.62055090e-01 6.79655895e-02
6.65336370e-01 1.39419422e-01 8.58479083e-01 -1.19317210e+00
-6.75746799e-01 -5.60366690e-01 3.78144085e-02 -1.76587749e+00
-1.80161297e-01 -1.10973692e+00 -1.31101578e-01 -1.94840789e+00
-4.13937010e-02 -8.42730582e-01 1.50288641e-01 6.61581218e-01
1.53070331e-01 4.84014779e-01 3.90447617e-01 3.00952971e-01
2.72385269e-01 7.72658408e-01 1.67449439e+00 -6.09704889e-02
-3.88736367e-01 -3.11088115e-01 -5.46167731e-01 8.70907545e-01
6.52135551e-01 -4.64409143e-01 -5.92841446e-01 -6.77085042e-01
3.97924870e-01 2.09978029e-01 5.58649063e-01 -8.56958270e-01
5.50202746e-03 -1.75061047e-01 3.60402435e-01 -5.88053048e-01
5.94242454e-01 -4.77626354e-01 2.48764858e-01 2.54597127e-01
-3.45134199e-01 -1.50915340e-01 1.78239614e-01 2.48323828e-01
3.66996229e-02 -2.19600707e-01 6.66190684e-01 -2.15670452e-01
-1.91772729e-01 5.34296155e-01 -2.53556758e-01 -2.46998876e-01
5.61890066e-01 -4.49774683e-01 1.03927836e-01 -6.16432846e-01
-1.12280416e+00 -2.50713527e-01 9.47662532e-01 2.89923519e-01
7.63906956e-01 -1.63226283e+00 -6.62254930e-01 6.50816143e-01
-2.82858431e-01 2.68424928e-01 8.42224211e-02 3.01024884e-01
-6.80350959e-01 9.34638083e-02 -4.49462794e-02 -7.77387977e-01
-9.09040272e-01 1.71586007e-01 3.43478769e-01 -1.63377061e-01
-9.91690159e-01 9.13606644e-01 6.79477513e-01 -5.27365506e-01
-7.06750974e-02 -3.47656101e-01 4.53206301e-01 -1.89049125e-01
2.93777674e-01 -7.45642334e-02 -2.02145755e-01 -2.36325666e-01
1.17238604e-01 1.06668413e+00 1.74730524e-01 -7.39368141e-01
1.47063184e+00 1.05981022e-01 4.13479097e-02 3.20253223e-01
1.14250112e+00 4.12139654e-01 -1.73521173e+00 2.54462093e-01
-5.64654052e-01 -4.52693135e-01 1.88342884e-01 -6.46497607e-01
-1.03721070e+00 1.10318661e+00 2.47807592e-01 2.12275505e-01
8.79097879e-01 -8.15396756e-02 9.07534897e-01 1.09845139e-01
5.35056889e-01 -6.41772151e-01 4.00697291e-01 6.45591378e-01
1.24944091e+00 -5.30605018e-01 -2.39774883e-01 -3.89769793e-01
-9.77180600e-02 1.19960809e+00 2.08194867e-01 -4.43371236e-01
7.45104492e-01 5.42992175e-01 5.50590940e-02 -2.73906261e-01
-5.74600518e-01 2.06440762e-01 3.79893482e-01 5.87295175e-01
5.47718465e-01 -2.15533935e-02 8.58336408e-03 1.14278428e-01
-4.91814107e-01 8.10840279e-02 4.99345958e-01 9.79638934e-01
-3.48009437e-01 -1.53412211e+00 -3.19585949e-01 1.85068473e-01
-1.80399008e-02 -1.52120680e-01 -2.54904032e-01 5.07892311e-01
1.08956136e-01 5.52683949e-01 1.33729756e-01 -1.21768691e-01
1.75943792e-01 1.40113533e-01 7.46485293e-01 -9.07588959e-01
-1.91967100e-01 4.21434492e-01 -1.24959581e-01 -6.95177078e-01
-1.88951671e-01 -5.29647708e-01 -1.24673414e+00 -4.38789129e-01
-4.74853888e-02 1.84014201e-01 6.81026757e-01 5.44568777e-01
4.88122791e-01 3.47546250e-01 4.52145427e-01 -1.39431584e+00
-1.95526302e-01 -5.27379453e-01 -3.89518142e-01 3.71185988e-01
2.58522451e-01 -6.28485858e-01 -3.91385406e-01 2.40374833e-01] | [9.139368057250977, -3.4630701541900635] |
b04e0d6b-bdd4-465e-ace4-c2e3303d8e1b | revisiting-estimation-bias-in-policy | 2301.08442 | null | https://arxiv.org/abs/2301.08442v2 | https://arxiv.org/pdf/2301.08442v2.pdf | Revisiting Estimation Bias in Policy Gradients for Deep Reinforcement Learning | We revisit the estimation bias in policy gradients for the discounted episodic Markov decision process (MDP) from Deep Reinforcement Learning (DRL) perspective. The objective is formulated theoretically as the expected returns discounted over the time horizon. One of the major policy gradient biases is the state distribution shift: the state distribution used to estimate the gradients differs from the theoretical formulation in that it does not take into account the discount factor. Existing discussion of the influence of this bias was limited to the tabular and softmax cases in the literature. Therefore, in this paper, we extend it to the DRL setting where the policy is parameterized and demonstrate how this bias can lead to suboptimal policies theoretically. We then discuss why the empirically inaccurate implementations with shifted state distribution can still be effective. We show that, despite such state distribution shift, the policy gradient estimation bias can be reduced in the following three ways: 1) a small learning rate; 2) an adaptive-learning-rate-based optimizer; and 3) KL regularization. Specifically, we show that a smaller learning rate, or, an adaptive learning rate, such as that used by Adam and RSMProp optimizers, makes the policy optimization robust to the bias. We further draw connections between optimizers and the optimization regularization to show that both the KL and the reverse KL regularization can significantly rectify this bias. Moreover, we provide extensive experiments on continuous control tasks to support our analysis. Our paper sheds light on how successful PG algorithms optimize policies in the DRL setting, and contributes insights into the practical issues in DRL. | ['Mingfei Sun', 'Jianping He', 'Wei Yang', 'Qiang Fu', 'Xiaoming Duan', 'Deheng Ye', 'Haoxuan Pan'] | 2023-01-20 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [-1.82361752e-01 1.96174875e-01 -6.66512489e-01 -1.07964359e-01
-5.11469781e-01 -5.14998794e-01 5.28408170e-01 6.89333230e-02
-8.26542437e-01 1.10723221e+00 1.31046981e-01 -6.72374189e-01
-2.23820761e-01 -6.20662570e-01 -8.93123209e-01 -8.78797472e-01
-1.37588874e-01 3.18801969e-01 5.69119751e-02 -1.67687356e-01
3.49116802e-01 4.25447971e-01 -1.20802414e+00 -3.47584814e-01
9.30719197e-01 1.05671537e+00 4.64408956e-02 5.93943000e-01
3.04134265e-02 1.02350950e+00 -6.13119662e-01 -1.40011117e-01
5.28553367e-01 -4.38164979e-01 -5.26911855e-01 -2.67158775e-03
-7.13377595e-02 -7.96948135e-01 -3.17811161e-01 1.07066369e+00
4.91864026e-01 4.28305387e-01 5.06322920e-01 -1.21622443e+00
-3.53055060e-01 6.63929284e-01 -3.82495046e-01 3.80120158e-01
-1.67362154e-01 4.74703580e-01 8.86940777e-01 -3.28692049e-01
3.61719042e-01 1.52329946e+00 4.79739606e-01 6.41214073e-01
-1.04938424e+00 -3.19059968e-01 7.35781312e-01 -2.54279710e-02
-7.86992311e-01 -1.94866940e-01 4.07495022e-01 -3.38794649e-01
1.04135060e+00 -1.02310784e-01 8.13607872e-01 1.13873243e+00
5.16604781e-01 9.12487507e-01 1.45553696e+00 -3.25105101e-01
7.69147515e-01 2.24722162e-01 6.99945784e-04 5.76706290e-01
4.62782115e-01 6.57998145e-01 -1.68189988e-01 -2.10160017e-01
9.23058271e-01 -1.98145792e-01 -2.65772879e-01 -4.45959151e-01
-7.94712365e-01 1.05215025e+00 1.76564515e-01 -9.11070853e-02
-5.36271513e-01 5.56219518e-01 3.23627353e-01 5.17804623e-01
4.81417239e-01 7.76477814e-01 -5.61822116e-01 -4.17349100e-01
-6.82434201e-01 5.84892750e-01 9.89551365e-01 5.67590773e-01
3.78423899e-01 4.27683532e-01 -3.72631073e-01 4.46981966e-01
4.30341035e-01 6.72287524e-01 4.10144448e-01 -1.36859560e+00
5.74956298e-01 -9.73468870e-02 7.20547557e-01 -5.70487201e-01
-4.56903368e-01 -4.47239429e-01 -1.90896600e-01 5.13992906e-01
7.60205865e-01 -8.11022997e-01 -7.56240726e-01 2.17820024e+00
1.82524741e-01 -1.77215785e-02 2.01280281e-01 1.03907323e+00
-2.06164122e-01 6.07499659e-01 1.03490062e-01 -5.06244183e-01
1.06008089e+00 -7.99909830e-01 -8.77813220e-01 -4.04095560e-01
6.83238447e-01 -4.00045127e-01 1.12836921e+00 2.20784113e-01
-1.28606367e+00 6.60024509e-02 -1.11811936e+00 4.23931330e-01
-3.04257404e-02 -2.05661375e-02 6.59848154e-01 5.62674999e-01
-9.12027955e-01 1.14978373e+00 -1.19062221e+00 -3.47151399e-01
1.27410099e-01 1.37962818e-01 6.60100102e-01 4.90668505e-01
-1.35553002e+00 1.30031872e+00 4.06760186e-01 1.91920791e-02
-1.01770115e+00 -5.99997520e-01 -5.86336613e-01 2.51501054e-01
7.41544783e-01 -6.28474951e-01 1.71502626e+00 -1.13132286e+00
-1.98918998e+00 2.60807604e-01 1.68921519e-03 -1.05498493e+00
8.84309113e-01 -2.55009800e-01 5.73064238e-02 1.01104207e-01
-2.70670354e-01 2.87719250e-01 1.04750907e+00 -1.05964029e+00
-7.05135643e-01 -2.38814533e-01 2.93802410e-01 5.06457329e-01
-3.53584476e-02 -2.20842212e-01 1.41451299e-01 -5.40310621e-01
-3.38494658e-01 -1.23017478e+00 -5.43513238e-01 -3.31800133e-01
-3.28317247e-02 1.93679947e-02 3.25246155e-01 -4.72985983e-01
1.20507729e+00 -1.70451283e+00 3.41843665e-02 1.40869081e-01
-2.06198990e-01 1.67205587e-01 1.06284663e-01 4.22540307e-01
1.94048688e-01 1.04266375e-01 -1.87720284e-01 -2.25941196e-01
2.98151940e-01 5.79867005e-01 -8.70635867e-01 5.09787142e-01
1.17896721e-01 8.46202075e-01 -1.04124439e+00 -1.60919204e-01
4.46745045e-02 1.98024437e-01 -6.25298023e-01 -3.72758619e-02
-4.71088976e-01 3.15252364e-01 -6.89250648e-01 3.41291636e-01
3.83523285e-01 -1.22269653e-01 3.72501105e-01 8.86246264e-02
-3.01854938e-01 5.39408803e-01 -1.14254177e+00 9.43560183e-01
-4.04137254e-01 3.59657615e-01 1.97988153e-01 -1.11212909e+00
5.23980439e-01 1.80450246e-01 4.63583380e-01 -5.49286366e-01
1.51201695e-01 3.03071260e-01 5.86363301e-02 -3.81116360e-01
5.71492672e-01 -4.91455793e-01 1.56625256e-01 6.35600269e-01
-1.25664741e-01 -1.50794312e-01 2.02960983e-01 6.82107955e-02
7.82097340e-01 4.15839463e-01 1.62774608e-01 -4.14113611e-01
1.17410151e-02 -7.21379891e-02 8.54738355e-01 1.04430735e+00
-5.89627981e-01 -8.85051414e-02 8.68311346e-01 -6.23298436e-02
-1.02869415e+00 -8.65608752e-01 -4.19969782e-02 1.15595007e+00
1.34963512e-01 -2.02700973e-01 -6.57890260e-01 -7.40256786e-01
4.62253988e-01 1.04013777e+00 -4.76634979e-01 -3.47267658e-01
-6.17201328e-01 -1.10774946e+00 3.43365341e-01 6.15987122e-01
3.12824011e-01 -9.07373488e-01 -9.84766901e-01 3.99232775e-01
9.05656815e-02 -7.90174544e-01 -4.12098795e-01 3.41802388e-01
-1.16698265e+00 -7.86964118e-01 -6.99871480e-01 -8.26221481e-02
4.84897405e-01 -1.21999443e-01 9.23450828e-01 -2.44271457e-01
4.73232239e-01 9.30049062e-01 -9.25853103e-02 -6.24483824e-01
-5.59898794e-01 -5.85420355e-02 4.06445324e-01 -2.20598340e-01
1.09724313e-01 -4.36592281e-01 -7.65792966e-01 2.34189466e-01
-6.04960740e-01 -2.04106614e-01 4.60974365e-01 7.60121584e-01
4.21900034e-01 -1.36720940e-01 7.82688200e-01 -6.94967926e-01
1.07934427e+00 -4.41304296e-01 -1.12749684e+00 2.20258385e-01
-1.23112655e+00 6.84348404e-01 5.10995686e-01 -6.02023780e-01
-1.15450680e+00 -2.61303425e-01 -2.20787693e-02 -4.05589193e-01
4.19688672e-01 4.47750926e-01 3.28004181e-01 -2.15350017e-02
3.68912727e-01 1.33763358e-01 2.05719382e-01 -3.88336122e-01
3.83892655e-01 2.60952443e-01 -1.09258983e-02 -1.02238619e+00
3.74084562e-01 4.41298336e-01 5.72833940e-02 -4.02277172e-01
-9.51230586e-01 2.37062380e-01 4.87649888e-02 -2.86914617e-01
6.23200417e-01 -7.66315639e-01 -9.47034955e-01 1.67109683e-01
-8.08799267e-01 -9.55358565e-01 -6.96520865e-01 8.10250282e-01
-1.08759165e+00 2.03492463e-01 -9.39271748e-01 -1.33226359e+00
-1.34491265e-01 -1.25486207e+00 3.90167296e-01 3.91751736e-01
1.04780301e-01 -1.13500750e+00 1.62528187e-01 -2.46682733e-01
5.90471804e-01 8.11565444e-02 8.42021942e-01 -4.48976398e-01
-3.77724826e-01 3.77785653e-01 1.99075520e-01 4.87251252e-01
-1.47029281e-01 -6.67295828e-02 -6.57998145e-01 -6.61612749e-01
4.18693751e-01 -3.23567420e-01 1.06083977e+00 7.11490929e-01
8.43182385e-01 -5.97356260e-01 -1.12117343e-01 4.61612731e-01
1.41795969e+00 4.27695662e-01 2.08101004e-01 7.11115718e-01
3.14196289e-01 4.72456157e-01 7.68426359e-01 7.78755963e-01
2.97756821e-01 3.77917111e-01 5.21634340e-01 2.47051299e-01
3.79934371e-01 -4.16167736e-01 7.95772791e-01 5.94045758e-01
-2.10552514e-01 -5.35287634e-02 -6.36265039e-01 3.13270003e-01
-2.24535942e+00 -9.38788891e-01 5.07004678e-01 2.52227902e+00
8.86217654e-01 4.13748473e-01 3.37990224e-01 -2.75237024e-01
5.91412842e-01 2.96608627e-01 -1.11020458e+00 -7.18747675e-01
-7.16842338e-02 -7.70304799e-02 1.19552267e+00 8.08603168e-01
-8.28760207e-01 9.18758273e-01 7.04785442e+00 6.11887574e-01
-1.27943456e+00 9.88522395e-02 5.76904595e-01 -4.84267652e-01
-1.36837393e-01 1.96907148e-01 -9.93715107e-01 6.63224518e-01
1.08600700e+00 -4.54349071e-01 8.46087337e-01 9.95373845e-01
5.55791795e-01 -1.38037786e-01 -9.66605783e-01 6.53749764e-01
-4.74869877e-01 -1.08109546e+00 -2.44288012e-01 3.20181876e-01
6.44190371e-01 1.21446252e-01 2.94607460e-01 7.74464309e-01
7.35420763e-01 -7.99703002e-01 9.44961667e-01 4.26838011e-01
1.50808766e-01 -9.79928672e-01 4.65217769e-01 3.49485099e-01
-6.78272426e-01 -5.68826020e-01 -5.69763303e-01 -2.31294960e-01
4.01953638e-01 4.88943309e-01 -6.90889001e-01 8.49919692e-02
3.98270816e-01 5.44349909e-01 -3.33357230e-02 7.68523216e-01
-4.09723610e-01 9.58437920e-01 -5.11502624e-01 -2.55203247e-01
6.58154845e-01 -5.18747330e-01 8.09859812e-01 9.57348347e-01
2.26933017e-01 -6.87278435e-02 1.43863082e-01 9.61488962e-01
2.53757238e-01 -2.23991558e-01 -3.64062935e-01 -2.84281015e-01
4.15568829e-01 8.26773763e-01 -5.06470501e-01 -2.27461427e-01
-1.70706794e-01 6.25128448e-01 4.63437617e-01 7.69054115e-01
-9.96777654e-01 -5.64888567e-02 8.73003364e-01 3.33350301e-02
4.77861285e-01 -3.75505835e-01 -8.68708491e-02 -1.19018340e+00
6.38310332e-03 -7.74172068e-01 2.85173655e-01 -2.19070941e-01
-1.12340844e+00 8.04154277e-02 2.13533923e-01 -8.27325165e-01
-6.16989732e-01 -4.82473731e-01 -3.89932096e-01 5.98307788e-01
-1.78168821e+00 -2.98586041e-01 5.88366270e-01 3.35484505e-01
4.32098776e-01 9.12302285e-02 2.16020331e-01 -8.65509827e-03
-8.66076708e-01 4.08808470e-01 6.00991905e-01 -2.94692099e-01
5.93331635e-01 -1.57039738e+00 1.39099404e-01 6.98857307e-01
-3.92758578e-01 5.18228531e-01 1.07274604e+00 -6.47546828e-01
-1.46349525e+00 -8.78665388e-01 1.86374024e-01 -2.48183042e-01
8.41784716e-01 -2.69355047e-02 -8.63212824e-01 9.22507703e-01
-2.62009315e-02 -3.00907761e-01 1.12267599e-01 -5.05779572e-02
1.25967294e-01 6.92272037e-02 -1.12819135e+00 8.31681311e-01
7.45301723e-01 -1.73890725e-01 -4.69912231e-01 2.21272796e-01
7.38290846e-01 -5.42802274e-01 -8.99064898e-01 2.36156687e-01
5.20652115e-01 -7.67078400e-01 7.46053576e-01 -9.72689748e-01
1.89268187e-01 -8.15945584e-03 -1.40550181e-01 -1.61718500e+00
-7.94581920e-02 -8.94498885e-01 -7.70253956e-01 8.96048486e-01
3.22121173e-01 -1.12451541e+00 5.29245496e-01 9.78134990e-01
8.61324072e-02 -1.18639863e+00 -9.77104127e-01 -1.21775460e+00
6.21428609e-01 -3.81788820e-01 4.73824978e-01 6.31016076e-01
1.53067177e-02 4.25789086e-03 -5.20303428e-01 -1.53741110e-02
5.56239069e-01 1.38380587e-01 3.35098952e-01 -6.62821412e-01
-6.55273736e-01 -6.30078435e-01 2.77014285e-01 -1.43050385e+00
3.15060824e-01 -5.30280173e-01 6.50050193e-02 -1.15382993e+00
-1.33269638e-01 -6.36632979e-01 -4.50728297e-01 2.65416712e-01
-2.07967907e-01 -6.95699215e-01 5.79392374e-01 2.77460963e-01
-4.54805583e-01 8.44011486e-01 1.34624970e+00 1.53642714e-01
-4.68188256e-01 2.06947058e-01 -5.97571075e-01 7.58898199e-01
1.10940838e+00 -6.22404695e-01 -5.58343589e-01 -1.90868348e-01
4.48010951e-01 1.11209571e-01 1.50541767e-01 -5.03120661e-01
-3.48733477e-02 -6.42347872e-01 3.74055691e-02 -1.42221779e-01
2.05949217e-01 -3.99588645e-01 -4.55963045e-01 8.21497977e-01
-5.37504971e-01 -7.77701382e-03 2.24925786e-01 7.58151591e-01
3.04838479e-01 -3.38791281e-01 9.54501569e-01 -2.72427291e-01
-3.74464363e-01 1.59785420e-01 -7.37873018e-01 4.05463994e-01
7.59797037e-01 1.75878301e-01 -2.94986159e-01 -7.91733027e-01
-6.78092718e-01 5.46964705e-01 4.51036513e-01 1.20795265e-01
2.50816166e-01 -1.16357350e+00 -3.86401683e-01 -3.62609416e-01
-5.57131112e-01 -4.68930691e-01 -1.81908861e-01 1.11760497e+00
-1.28038034e-01 5.14980674e-01 4.14815322e-02 -2.36073270e-01
-5.87928534e-01 5.55283487e-01 7.78344512e-01 -5.86128294e-01
-5.27409673e-01 3.12387556e-01 -3.61477733e-02 -8.07281062e-02
4.57887560e-01 -6.73457384e-01 1.25636026e-01 8.00697133e-02
3.98913205e-01 5.27651906e-01 -2.56603152e-01 -1.55880734e-01
-1.88262269e-01 2.99351722e-01 -1.14697486e-01 -6.26500964e-01
1.10074747e+00 -2.84285992e-01 4.05402333e-01 5.42606950e-01
6.67411566e-01 -3.37471306e-01 -1.99413836e+00 -1.29671142e-01
3.35736573e-02 -2.39855886e-01 3.37152511e-01 -7.04693377e-01
-9.86420572e-01 7.35603094e-01 6.16995752e-01 2.82064497e-01
7.73154914e-01 -4.82254356e-01 6.56385362e-01 5.76813042e-01
4.85770822e-01 -1.60954928e+00 -1.48863494e-01 7.64925599e-01
7.40670681e-01 -1.18060505e+00 7.71453679e-02 3.25250030e-01
-1.00293946e+00 1.03469110e+00 4.49700594e-01 -3.36616457e-01
5.30384064e-01 1.62421480e-01 -1.26937106e-01 1.31638855e-01
-1.04180527e+00 -3.39836329e-01 -2.62483180e-01 3.30929130e-01
1.32748485e-01 6.22228086e-02 -6.19713843e-01 4.89870369e-01
-7.80800432e-02 7.91669264e-02 6.05862081e-01 1.24636471e+00
-5.62734127e-01 -8.49992335e-01 -3.80317897e-01 3.47384542e-01
-6.84052110e-01 5.98714575e-02 -7.03766057e-03 8.23039651e-01
-4.70628768e-01 7.78622329e-01 6.40571266e-02 1.76414803e-01
3.46623540e-01 1.29206419e-01 7.23831594e-01 -8.25067014e-02
-5.72812080e-01 -1.79319573e-03 -6.74969405e-02 -7.70113230e-01
-1.70334607e-01 -6.65951371e-01 -1.39278197e+00 -3.19827735e-01
-2.45566815e-01 8.91083181e-02 7.46060908e-01 1.03773296e+00
3.51125777e-01 3.72084886e-01 5.98850012e-01 -7.99750209e-01
-1.47858012e+00 -7.27660537e-01 -6.88075662e-01 1.02111988e-01
5.61055183e-01 -9.51277018e-01 -9.12571490e-01 -5.48149407e-01] | [4.193652629852295, 2.414416551589966] |
d659e1d9-a140-4573-a12f-917106138e7e | monocular-2d-camera-based-proximity | 2305.17931 | null | https://arxiv.org/abs/2305.17931v1 | https://arxiv.org/pdf/2305.17931v1.pdf | Monocular 2D Camera-based Proximity Monitoring for Human-Machine Collision Warning on Construction Sites | Accident of struck-by machines is one of the leading causes of casualties on construction sites. Monitoring workers' proximities to avoid human-machine collisions has aroused great concern in construction safety management. Existing methods are either too laborious and costly to apply extensively, or lacking spatial perception for accurate monitoring. Therefore, this study proposes a novel framework for proximity monitoring using only an ordinary 2D camera to realize real-time human-machine collision warning, which is designed to integrate a monocular 3D object detection model to perceive spatial information from 2D images and a post-processing classification module to identify the proximity as four predefined categories: Dangerous, Potentially Dangerous, Concerned, and Safe. A virtual dataset containing 22000 images with 3D annotations is constructed and publicly released to facilitate the system development and evaluation. Experimental results show that the trained 3D object detection model achieves 75% loose AP within 20 meters. Besides, the implemented system is real-time and camera carrier-independent, achieving an F1 of roughly 0.8 within 50 meters under specified settings for machines of different sizes. This study preliminarily reveals the potential and feasibility of proximity monitoring using only a 2D camera, providing a new promising and economical way for early warning of human-machine collisions. | ['Xiaowei Luo', 'Yuexiong Ding'] | 2023-05-29 | null | null | null | null | ['monocular-3d-object-detection'] | ['computer-vision'] | [ 1.02127045e-01 -7.88668916e-02 2.43572041e-01 -1.09780997e-01
-5.89785695e-01 -2.40791708e-01 3.51029128e-01 3.04379404e-01
-6.37182832e-01 -9.86356754e-03 -1.44582227e-01 -4.77946967e-01
-2.99455166e-01 -8.53819788e-01 -4.15445536e-01 -6.73545301e-01
1.25173226e-01 2.31289685e-01 7.02063322e-01 -3.48379493e-01
2.30533078e-01 9.45148110e-01 -1.80210507e+00 1.26759589e-01
2.33832389e-01 1.21431565e+00 6.60664141e-01 6.24805987e-01
3.51161689e-01 4.16492730e-01 -6.59998298e-01 -2.65189052e-01
4.17766929e-01 1.25112787e-01 -4.74836975e-01 1.59263641e-01
-3.85760218e-02 -2.61592895e-01 -4.23785806e-01 6.32434785e-01
8.88912082e-01 5.52154966e-02 5.56517541e-01 -1.33433092e+00
-4.42750633e-01 -3.82459573e-02 -6.27228379e-01 2.65290856e-01
7.14115560e-01 3.12500060e-01 2.66137481e-01 -7.57344484e-01
-1.21708801e-02 9.40091550e-01 7.29171216e-01 2.42893174e-01
-8.56763124e-01 -5.74352503e-01 -1.50732249e-01 2.61355937e-01
-1.89175093e+00 -1.90836489e-01 8.28717172e-01 -6.15000069e-01
9.31653619e-01 5.49379110e-01 4.91788507e-01 9.53624427e-01
4.25973922e-01 2.38416702e-01 8.30528200e-01 -5.61376214e-01
3.65908295e-02 3.75246823e-01 -5.44425175e-02 6.46266341e-01
3.71810317e-01 2.96079457e-01 -3.02647024e-01 2.68191509e-02
8.50993395e-01 2.02024832e-01 -1.70670927e-01 -3.24664414e-01
-1.26847303e+00 4.47538882e-01 5.99756777e-01 3.61251116e-01
-4.63844895e-01 -7.82378241e-02 3.37225825e-01 -1.28315255e-01
2.72621393e-01 7.18246475e-02 -1.61651056e-02 -1.64472759e-01
-1.69499651e-01 -4.51400643e-03 4.49461266e-02 1.23845053e+00
4.83616382e-01 -5.04167259e-01 -2.15647429e-01 7.66165376e-01
2.14362085e-01 6.66676104e-01 6.09690771e-02 -6.64992988e-01
7.03926027e-01 6.20601654e-01 3.21361750e-01 -1.44538307e+00
-6.97561800e-01 -3.46598446e-01 -7.27524042e-01 5.29342711e-01
2.48741940e-01 2.70959526e-01 -3.78491461e-01 9.47012424e-01
6.30507588e-01 -1.60527512e-01 -3.25671494e-01 1.11316204e+00
6.41117930e-01 3.98807049e-01 8.55637640e-02 -8.73911232e-02
1.64775622e+00 -6.07416809e-01 -5.25634646e-01 -2.62805015e-01
7.14792430e-01 -7.86777616e-01 1.10828710e+00 4.42055881e-01
-7.80254960e-01 -9.53521848e-01 -9.10723388e-01 8.39762241e-02
-2.23268002e-01 3.41685921e-01 3.58208150e-01 8.49590123e-01
-4.55113709e-01 1.87464699e-01 -5.52920699e-01 -3.89111400e-01
2.07727626e-01 1.84810653e-01 -6.51890755e-01 -3.50739285e-02
-1.09010410e+00 1.08581984e+00 4.46567804e-01 4.47117180e-01
-5.65323234e-01 -2.19910145e-01 -7.86020935e-01 -2.64535666e-01
4.59090054e-01 -5.12492120e-01 1.17911577e+00 7.91402608e-02
-9.71884251e-01 9.54458177e-01 2.36270994e-01 -6.21224893e-03
6.03767574e-01 -3.98601830e-01 -6.80482447e-01 1.56458825e-01
4.03604954e-01 -6.05219603e-02 3.09090823e-01 -1.36788619e+00
-1.07029247e+00 -4.41155940e-01 -6.78299647e-03 1.88574672e-01
-1.68402150e-01 1.12161323e-01 -4.93579894e-01 -2.94483334e-01
3.27841252e-01 -8.32981706e-01 -2.84178078e-01 -1.67577597e-03
-3.18536639e-01 -2.45990977e-01 5.38182497e-01 -4.58399504e-01
1.27445209e+00 -2.10874462e+00 -7.04554915e-01 2.67900079e-01
-7.78709799e-02 1.01356387e-01 2.10022420e-01 4.93438900e-01
2.48780604e-02 -5.14937639e-01 -1.91110577e-02 -1.17356122e-01
-1.93185940e-01 -3.28453965e-02 1.28116980e-01 7.46000946e-01
-1.64871067e-01 4.81458783e-01 -1.03238356e+00 -6.43459141e-01
8.01849484e-01 3.41411471e-01 -6.16518520e-02 2.65409321e-01
7.25100636e-01 4.81430590e-01 -3.74934047e-01 7.41833091e-01
7.74213612e-01 2.22708315e-01 -3.20593417e-01 -1.70534819e-01
-4.11352128e-01 -7.05273077e-02 -1.39227617e+00 1.37681007e+00
-5.93670130e-01 2.12305143e-01 1.32550637e-03 -8.24013174e-01
1.06151342e+00 4.39487785e-01 5.27484238e-01 -8.62966239e-01
2.81255335e-01 3.95115726e-02 -7.95925319e-01 -1.04766011e+00
3.25080335e-01 -5.00078723e-02 -4.87446219e-01 -3.69072519e-03
-5.22012413e-01 -2.08707284e-02 -2.87707806e-01 -1.61842614e-01
1.37590170e+00 -1.81452870e-01 1.66489497e-01 1.10841513e-01
5.20327330e-01 1.79832876e-02 2.43729576e-01 7.72604048e-01
-4.01663363e-01 7.39760697e-01 -9.85984653e-02 -5.49800158e-01
-6.95769548e-01 -1.40345168e+00 -5.19498698e-02 7.33783543e-01
7.88920462e-01 -1.13826734e-03 -5.88081121e-01 -8.03937912e-01
5.75760473e-03 7.81866193e-01 -2.67173648e-01 -4.39602375e-01
-4.91766781e-01 -3.46259266e-01 4.65526670e-01 7.60334671e-01
5.33198297e-01 -8.50557923e-01 -1.00489306e+00 1.40029013e-01
-5.01457155e-01 -1.24003828e+00 -2.33673126e-01 1.14580821e-02
-4.25751925e-01 -1.21084583e+00 -5.86411476e-01 -9.34803724e-01
7.52905011e-01 1.04408145e+00 7.10109293e-01 1.93460822e-01
-6.65302455e-01 4.79134351e-01 -3.40083987e-01 -5.30536890e-01
-1.92694917e-01 -1.95624352e-01 2.33095095e-01 -2.62888931e-02
5.04340827e-01 -2.22844973e-01 -8.27555358e-01 9.17618811e-01
-3.09235394e-01 -1.02565646e-01 8.33792508e-01 3.59256238e-01
2.90368021e-01 5.89299858e-01 2.69902200e-01 -4.21994388e-01
4.39663529e-01 -4.75805342e-01 -5.05549788e-01 1.42600164e-01
-5.16531110e-01 -5.54191947e-01 2.16628939e-01 -2.28519425e-01
-1.30651665e+00 5.01830697e-01 -2.41559818e-01 -9.12351310e-02
-8.87817085e-01 -1.62070226e-02 -3.05674642e-01 1.69821456e-01
9.73247111e-01 3.13747348e-03 -2.06298277e-01 -5.57670414e-01
1.22999825e-01 1.17212892e+00 1.09250319e+00 -3.10140699e-01
1.05008519e+00 6.87545896e-01 9.37149823e-02 -7.99605310e-01
-6.90064549e-01 -1.28524804e+00 -9.79758382e-01 -8.05778265e-01
1.05001533e+00 -1.01388896e+00 -1.05351877e+00 3.69807661e-01
-1.37424719e+00 2.09349066e-01 4.00225520e-02 8.15170467e-01
-3.54422420e-01 6.35339081e-01 -3.33108604e-01 -1.19300652e+00
-6.08940609e-02 -8.29841018e-01 1.52617431e+00 -1.58714890e-01
-1.51418999e-01 -3.78021121e-01 -9.81569663e-02 8.96246374e-01
7.74119496e-02 3.45153689e-01 3.76717657e-01 5.05341776e-02
-3.89519185e-01 -8.96929681e-01 -2.06531405e-01 1.70994982e-01
2.91875094e-01 -4.98933911e-01 -1.11373758e+00 4.18871194e-02
1.85606897e-01 2.72653282e-01 4.62196499e-01 1.31319389e-01
1.16601300e+00 5.88405319e-02 -7.53931820e-01 -2.77826656e-02
1.11737990e+00 2.63911992e-01 7.79207826e-01 5.68821430e-01
5.67978203e-01 8.61637771e-01 1.25333369e+00 6.09728396e-01
2.98019558e-01 8.56935263e-01 7.68694580e-01 -3.04266393e-01
1.97365433e-01 -2.31057882e-01 1.30505532e-01 4.74962503e-01
-4.75188017e-01 -2.12673500e-01 -1.01790309e+00 4.96146590e-01
-1.74383640e+00 -1.30398047e+00 -7.11924374e-01 2.34502530e+00
2.49914587e-01 4.85729635e-01 1.22330524e-01 7.41765857e-01
1.07227552e+00 -2.53487974e-01 3.32384050e-01 -1.05498232e-01
4.43482578e-01 -2.23111883e-01 7.89006114e-01 3.08638632e-01
-1.26078069e+00 3.04492593e-01 5.61624050e+00 7.47187495e-01
-7.09747970e-01 2.21816361e-01 1.70260549e-01 8.30114819e-03
1.57053038e-01 -2.49689847e-01 -6.90960824e-01 4.75461155e-01
4.77851123e-01 5.45515716e-01 -2.23008841e-01 1.25514078e+00
5.55738628e-01 -4.47477371e-01 -8.04135799e-01 1.24961472e+00
-1.09101392e-01 -8.06049168e-01 -3.53302419e-01 7.08226487e-02
1.41679943e-01 -5.19722641e-01 -1.05555907e-01 1.18949018e-01
-3.50332141e-01 -5.98871589e-01 1.14805567e+00 5.23397028e-01
6.32567048e-01 -8.94411564e-01 9.55162585e-01 9.00888264e-01
-1.43855309e+00 -4.14976865e-01 -4.73158419e-01 -4.51021165e-01
7.05403984e-01 7.19909072e-01 -1.00840104e+00 6.91864610e-01
1.02020526e+00 5.61053082e-02 -5.37144721e-01 1.07139158e+00
-1.91723928e-01 8.91045853e-02 -1.64185360e-01 -4.48052585e-02
-7.00938655e-03 8.10792521e-02 2.64233589e-01 1.14011836e+00
6.21202290e-01 1.65023834e-01 1.98755711e-01 3.73404682e-01
6.40732527e-01 -1.21785827e-01 -1.06851232e+00 9.26720262e-01
5.47855616e-01 1.16128421e+00 -8.52709055e-01 9.43137035e-02
-4.61051404e-01 9.91268396e-01 -3.42398770e-02 -1.64480470e-02
-1.26921630e+00 -6.79170370e-01 2.77794629e-01 6.74440980e-01
8.22131857e-02 -5.19208312e-01 -4.26001638e-01 -3.23707491e-01
3.06942165e-01 -1.41338155e-01 3.68179440e-01 -9.91256833e-01
-9.72360730e-01 5.65399826e-01 2.96679586e-01 -1.66011906e+00
3.52918208e-01 -8.33655238e-01 -4.96639192e-01 6.62869871e-01
-1.03002810e+00 -1.14802563e+00 -8.82712662e-01 6.07751727e-01
4.23942000e-01 -6.86371233e-03 4.84839529e-01 5.92302382e-01
-5.10303676e-01 4.82425570e-01 -4.28297997e-01 1.63641348e-02
4.75266784e-01 -7.88125694e-01 3.37044895e-01 7.90422976e-01
-6.73642606e-02 3.30061913e-01 6.16260648e-01 -7.21098304e-01
-1.14143252e+00 -1.09702504e+00 1.16857624e+00 -1.02651024e+00
6.08434230e-02 -5.77821374e-01 -4.26230818e-01 2.81311631e-01
-2.90679604e-01 -5.65127991e-02 3.73661548e-01 -2.54665911e-01
1.20553769e-01 -3.21935028e-01 -1.17924511e+00 5.47303677e-01
1.42627144e+00 -4.75111604e-01 -4.73412752e-01 4.00126755e-01
5.21509111e-01 -3.40496510e-01 -7.60487139e-01 5.23684740e-01
4.66655910e-01 -1.17517269e+00 1.36541903e+00 2.42976576e-01
-1.22582853e-01 -5.46068490e-01 -1.19620688e-01 -6.07349396e-01
-6.90469623e-01 -2.44473487e-01 2.93573886e-01 1.10363770e+00
1.51270285e-01 -4.83284175e-01 5.98160326e-01 5.32177985e-01
-6.36264443e-01 -3.87779027e-01 -1.13647461e+00 -1.11811316e+00
-7.27765858e-01 -1.10473776e+00 5.37882030e-01 4.84235853e-01
-1.36007881e-02 2.01765284e-01 -2.03502834e-01 8.65393877e-01
4.98072237e-01 -3.52455884e-01 9.95267451e-01 -1.24259365e+00
6.87353238e-02 -3.06916922e-01 -8.06373894e-01 -1.17011154e+00
-4.17369097e-01 -4.64454055e-01 4.46297705e-01 -1.64631546e+00
7.54857529e-03 -4.56988126e-01 -1.81957781e-01 2.80212283e-01
5.92039824e-02 3.23459774e-01 -3.28012526e-01 1.28106177e-01
-4.50499833e-01 3.23418349e-01 1.09636021e+00 -7.11934417e-02
2.04349548e-01 6.06699347e-01 -3.80562723e-01 7.12386310e-01
6.95438504e-01 -5.39868653e-01 -3.36110860e-01 -4.50894892e-01
1.22739524e-01 -2.00265646e-02 9.43330288e-01 -1.47057152e+00
3.51476222e-01 -2.99625784e-01 2.26576567e-01 -9.17257786e-01
4.35925514e-01 -1.38214612e+00 2.39813104e-01 8.26546013e-01
5.07840812e-02 1.68402255e-01 1.37984054e-02 8.07617188e-01
2.11050659e-01 -2.30547756e-01 4.96312380e-01 -9.19695050e-02
-9.26730931e-01 -1.72736291e-02 -5.19871414e-01 -5.77718973e-01
1.69968104e+00 -6.31493926e-01 -1.36551872e-01 -8.54116306e-02
-4.84418750e-01 -9.15786326e-02 1.87350258e-01 6.19163692e-01
8.12514722e-01 -1.61128938e+00 -4.64809895e-01 2.83619821e-01
5.54905713e-01 1.65037245e-01 5.18814743e-01 8.28000426e-01
-8.37518156e-01 4.00517374e-01 -1.58306230e-02 -9.48462486e-01
-1.41901922e+00 1.02333367e+00 4.34966683e-02 3.20770293e-01
-7.99322188e-01 7.68360496e-01 1.64725021e-01 -3.55046540e-01
4.09194440e-01 -4.12317574e-01 -1.36544853e-01 -3.74728739e-01
3.89204085e-01 8.86169314e-01 3.38188738e-01 -9.62666333e-01
-7.38220453e-01 8.60663176e-01 6.57108605e-01 7.76493773e-02
8.51555824e-01 -5.06347418e-01 2.90101051e-01 2.93217152e-01
1.03242528e+00 3.01619098e-02 -1.04539824e+00 9.22678225e-03
2.92577632e-02 -7.57037938e-01 -7.39346370e-02 -4.33135211e-01
-6.75349593e-01 8.00322473e-01 8.55806589e-01 4.24928576e-01
1.04883575e+00 3.45743567e-01 8.44268084e-01 3.24730784e-01
9.01169837e-01 -1.20841205e+00 1.79710209e-01 1.71329491e-02
1.06903315e+00 -1.35884142e+00 -3.30064297e-02 -7.46261299e-01
-6.80323362e-01 8.84952307e-01 6.81890368e-01 3.17605287e-01
7.32700944e-01 5.37303425e-02 1.89658463e-01 -5.26730895e-01
-1.30988851e-01 -1.91803068e-01 3.16446930e-01 9.62238789e-01
1.57089196e-02 -5.71125001e-02 -7.94339180e-03 6.10577404e-01
-4.69609946e-02 -3.46428782e-01 5.56561798e-02 1.22100282e+00
-7.29634106e-01 -4.48615611e-01 -8.09605479e-01 -5.06171510e-02
-1.40708983e-01 6.70535862e-01 9.98108089e-02 8.18520844e-01
5.39994657e-01 1.46681106e+00 6.23482801e-02 -7.94240117e-01
1.18571186e+00 -4.22801644e-01 2.58170068e-01 -6.56655371e-01
-3.98685932e-01 -1.97486714e-01 6.61109686e-02 -6.25318885e-01
-1.82156011e-01 -3.51746678e-01 -1.10227573e+00 -2.41698354e-01
-7.63991058e-01 8.73848610e-03 6.74044788e-01 7.81277776e-01
2.10627973e-01 4.27959025e-01 1.05448771e+00 -1.04294360e+00
-4.43322569e-01 -6.83096111e-01 -5.97141266e-01 4.11685616e-01
7.88915902e-02 -1.04170334e+00 -1.90453559e-01 -6.48247525e-02] | [7.726075649261475, -1.1352204084396362] |
6c4b9606-e161-4580-9837-2d4151485360 | the-best-of-both-worlds-combining-model-based | 2205.00508 | null | https://arxiv.org/abs/2205.00508v1 | https://arxiv.org/pdf/2205.00508v1.pdf | The Best of Both Worlds: Combining Model-based and Nonparametric Approaches for 3D Human Body Estimation | Nonparametric based methods have recently shown promising results in reconstructing human bodies from monocular images while model-based methods can help correct these estimates and improve prediction. However, estimating model parameters from global image features may lead to noticeable misalignment between the estimated meshes and image evidence. To address this issue and leverage the best of both worlds, we propose a framework of three consecutive modules. A dense map prediction module explicitly establishes the dense UV correspondence between the image evidence and each part of the body model. The inverse kinematics module refines the key point prediction and generates a posed template mesh. Finally, a UV inpainting module relies on the corresponding feature, prediction and the posed template, and completes the predictions of occluded body shape. Our framework leverages the best of non-parametric and model-based methods and is also robust to partial occlusion. Experiments demonstrate that our framework outperforms existing 3D human estimation methods on multiple public benchmarks. | ['Charless Fowlkes', 'Jimei Yang', 'Zhe Wang'] | 2022-05-01 | null | null | null | null | ['3d-absolute-human-pose-estimation'] | ['computer-vision'] | [ 1.01942718e-01 2.68394768e-01 -3.08622658e-01 -1.69890493e-01
-6.11335814e-01 -2.56879181e-01 5.45906186e-01 -3.18159997e-01
1.81976363e-01 7.76979685e-01 3.86083633e-01 6.08686566e-01
1.61707804e-01 -6.69057965e-01 -1.07614732e+00 -2.52048999e-01
3.29633534e-01 1.10594332e+00 4.29737955e-01 1.27270281e-01
6.99209422e-02 3.46807510e-01 -1.55126441e+00 5.32705188e-02
7.06266761e-01 1.21770668e+00 -3.71915773e-02 5.44293702e-01
1.34866104e-01 4.37619686e-01 -1.68141752e-01 -4.99168068e-01
5.66142201e-01 -1.46149129e-01 -5.08493543e-01 3.78577620e-01
1.00057852e+00 -7.63054967e-01 -3.36659551e-01 8.08516443e-01
3.48100245e-01 -1.29377708e-01 6.87666178e-01 -1.14217091e+00
-1.40010849e-01 -3.94363254e-02 -9.14086103e-01 -4.52505916e-01
8.87519896e-01 1.48486808e-01 6.50199831e-01 -1.21354294e+00
1.24419308e+00 1.35658216e+00 1.12853265e+00 2.67980099e-01
-1.36152470e+00 -6.15655661e-01 -1.56339034e-02 -8.72114748e-02
-1.37911069e+00 -5.37837386e-01 8.83586168e-01 -7.84229398e-01
5.19938350e-01 1.18229315e-02 1.18811786e+00 8.27816129e-01
4.12241369e-01 6.43666148e-01 9.88176882e-01 -1.70264259e-01
-1.51221186e-01 -2.82576293e-01 -3.62731040e-01 1.09512758e+00
2.44924977e-01 3.89856935e-01 -9.94874537e-01 -4.99852985e-01
1.33110332e+00 -1.02234297e-01 -2.81915694e-01 -1.13949585e+00
-1.21293056e+00 4.38568890e-01 1.78183585e-01 -4.43271995e-01
-5.77315271e-01 3.31418693e-01 5.66261820e-02 -1.81074083e-01
6.87147141e-01 1.18918747e-01 -5.43949306e-01 -7.76070282e-02
-1.29886961e+00 6.30374789e-01 8.82940531e-01 1.07100749e+00
8.65769923e-01 -1.78471744e-01 -2.69679707e-02 8.63910735e-01
5.54499567e-01 6.47755504e-01 -2.37978041e-01 -1.48164988e+00
2.92430311e-01 6.36913598e-01 3.85074645e-01 -1.16040134e+00
-3.39084864e-01 -2.74819434e-01 -5.70389032e-01 2.45650098e-01
6.08278513e-01 1.70169771e-01 -9.72351432e-01 1.50571609e+00
1.06061673e+00 5.11077464e-01 -7.43354321e-01 1.14711857e+00
7.21630275e-01 2.89373428e-01 -2.60671884e-01 6.55012950e-02
1.23376262e+00 -1.06214523e+00 -4.32565928e-01 -2.41493762e-01
-2.19164453e-02 -8.79112959e-01 4.81743544e-01 3.68706763e-01
-1.62667155e+00 -6.72946095e-01 -1.03271449e+00 -3.51544082e-01
3.53975654e-01 9.94050726e-02 2.37844244e-01 2.60976225e-01
-9.27176118e-01 7.46748984e-01 -9.42954421e-01 -2.62794495e-01
2.89397418e-01 3.93542945e-01 -5.08881450e-01 -2.19284177e-01
-6.45916522e-01 9.71049428e-01 -1.41217485e-01 1.24833979e-01
-7.17514932e-01 -9.74736869e-01 -1.21345484e+00 -4.46913660e-01
2.51886636e-01 -1.50178790e+00 1.09394658e+00 -6.34840012e-01
-1.55804491e+00 8.96439731e-01 -3.39381248e-01 -3.05478096e-01
1.09990799e+00 -4.65906978e-01 3.57765406e-01 3.30138624e-01
2.22841531e-01 1.01250470e+00 1.18229055e+00 -1.42952621e+00
-4.21145707e-01 -3.59017491e-01 -2.95175642e-01 2.08733082e-01
4.50876981e-01 -5.36551178e-01 -8.64114881e-01 -6.32627726e-01
5.39030313e-01 -1.05875266e+00 -6.80714250e-02 7.73651481e-01
-4.24680024e-01 1.76670611e-01 5.74159861e-01 -1.05303454e+00
6.17636979e-01 -1.64026701e+00 5.10778844e-01 5.94845772e-01
2.56022006e-01 -3.91893208e-01 3.11376333e-01 9.38752294e-02
3.38066459e-01 -4.46173996e-01 -1.54029638e-01 -8.43149841e-01
1.42349556e-01 3.12957406e-01 -1.50246203e-01 9.04577136e-01
-6.02933615e-02 9.82064784e-01 -8.07627857e-01 -8.89692903e-01
5.66110492e-01 6.26331389e-01 -8.41618538e-01 2.72106171e-01
-3.10470998e-01 8.20373535e-01 -3.13947201e-01 9.29786205e-01
7.71127164e-01 -2.37660542e-01 3.19636285e-01 -6.18357897e-01
3.15596238e-02 8.09789896e-02 -1.38561940e+00 2.07550335e+00
-1.08952165e-01 8.30868557e-02 2.95374542e-01 -5.36723495e-01
7.10271418e-01 3.67043734e-01 9.04430687e-01 -2.13723496e-01
1.20087020e-01 1.85108840e-01 -5.17179668e-01 -1.65124282e-01
3.40819210e-01 2.71050502e-02 2.86413878e-01 1.58966243e-01
8.02360326e-02 -6.03044093e-01 -1.61194116e-01 -4.96343635e-02
8.28251421e-01 1.16923904e+00 3.65807772e-01 -1.95374966e-01
2.88483351e-01 3.19927186e-02 6.47333324e-01 4.46299434e-01
-1.44548893e-01 1.03334796e+00 2.07729667e-01 -7.12221146e-01
-1.26178408e+00 -1.29070747e+00 -1.97078541e-01 4.56994742e-01
4.50550944e-01 -5.01413643e-01 -7.19435155e-01 -3.73673379e-01
4.92727786e-01 1.50995314e-01 -9.13057089e-01 1.26615286e-01
-6.46435261e-01 -3.55609097e-02 8.64349604e-02 6.49073541e-01
3.88522267e-01 -5.16197860e-01 -6.00165606e-01 2.04723895e-01
-4.14914340e-01 -1.24874914e+00 -7.27055550e-01 -4.48953241e-01
-1.05229616e+00 -1.21552980e+00 -6.92062438e-01 -4.87576544e-01
8.06316435e-01 -1.66804060e-01 1.38806188e+00 3.86368096e-01
-2.98837423e-01 6.84307337e-01 7.04576224e-02 -1.34152770e-01
-4.41767365e-01 -3.17953378e-01 1.36053681e-01 -1.08607337e-01
-3.45100492e-01 -7.87090659e-01 -9.70750690e-01 6.05606556e-01
-3.26219320e-01 5.35702169e-01 3.40655893e-01 7.41758049e-01
1.04609704e+00 -5.89775085e-01 -1.75472423e-01 -4.94280279e-01
-1.87679499e-01 -3.33759069e-01 -5.70420444e-01 1.83364470e-02
-3.31105173e-01 7.07914159e-02 -1.60468388e-02 -3.53803009e-01
-1.04053247e+00 6.79364443e-01 1.39894858e-02 -8.04295421e-01
1.44317821e-01 1.91263512e-01 -8.73359367e-02 -1.84132233e-01
4.45882499e-01 -6.78601637e-02 2.50351042e-01 -7.06449449e-01
3.03669930e-01 -8.97315964e-02 9.05499041e-01 -9.18397486e-01
1.04598129e+00 9.28139091e-01 2.31839091e-01 -5.60250282e-01
-8.29019010e-01 -4.42018569e-01 -1.07997441e+00 -4.62731928e-01
8.64572406e-01 -1.05523622e+00 -4.31894809e-01 3.24107021e-01
-1.49053967e+00 -1.67723998e-01 -3.52807283e-01 4.33198750e-01
-9.42351043e-01 6.40992284e-01 -7.70405293e-01 -5.49813569e-01
-3.65870357e-01 -1.14439058e+00 1.77484524e+00 -1.54370159e-01
-6.25286400e-01 -6.70254529e-01 2.66213298e-01 4.78976190e-01
-3.26281399e-01 6.52466297e-01 4.74710375e-01 2.14173287e-01
-8.17813337e-01 -1.86678812e-01 8.09463486e-03 1.93870459e-02
-9.06555206e-02 1.11355156e-01 -7.40363359e-01 -1.61339551e-01
-1.54926047e-01 -2.54922837e-01 5.11532784e-01 7.77339637e-01
7.83323705e-01 -1.27545118e-01 -5.53928614e-01 7.10186183e-01
1.02740359e+00 -5.34923255e-01 5.92526197e-01 2.48185620e-01
9.03897524e-01 6.49382949e-01 8.63761067e-01 7.82397389e-01
6.55007243e-01 1.04194915e+00 4.84694570e-01 1.44781657e-02
-4.66247350e-01 -6.85643613e-01 1.18468069e-01 7.45701611e-01
-3.82146537e-01 4.79981482e-01 -8.04007649e-01 3.91477853e-01
-1.91865015e+00 -6.40938103e-01 -1.63101181e-01 2.26959467e+00
9.42154408e-01 4.91528064e-02 2.73703575e-01 -1.94788858e-01
5.19749403e-01 -9.90058929e-02 -4.35936540e-01 5.01483791e-02
2.51452565e-01 2.36717895e-01 4.19971019e-01 6.41519666e-01
-9.48243976e-01 9.27267849e-01 6.77126122e+00 4.75053549e-01
-5.69752514e-01 2.13005617e-01 2.25623831e-01 -1.27094179e-01
-1.49695143e-01 1.59386411e-01 -7.08805680e-01 3.17779154e-01
3.15485567e-01 9.53222737e-02 2.24852279e-01 9.31559563e-01
1.34046942e-01 -4.08334345e-01 -1.43281019e+00 1.16492081e+00
1.28720433e-01 -1.42163742e+00 -2.51544006e-02 3.93782675e-01
1.00036728e+00 -1.83699369e-01 -2.35675201e-01 -1.95080251e-01
1.56076685e-01 -8.42426777e-01 1.31267560e+00 1.03724027e+00
7.16048717e-01 -4.71236557e-01 4.22551155e-01 4.35388207e-01
-1.54401505e+00 4.79333162e-01 -2.94718474e-01 -1.89472571e-01
4.16713834e-01 6.79589570e-01 -8.76101017e-01 5.49725831e-01
7.15122044e-01 8.18817556e-01 -3.07295561e-01 1.12197077e+00
-2.91259348e-01 9.56533924e-02 -5.55699170e-01 6.16550624e-01
-4.43742573e-01 -3.34510446e-01 7.60721982e-01 6.33788228e-01
1.69020250e-01 -5.55486185e-03 4.82132703e-01 9.52510476e-01
-5.10203876e-02 6.45742938e-02 -4.26171631e-01 5.15788198e-01
4.27075088e-01 1.07961965e+00 -6.03423417e-01 -4.45388228e-01
-2.44518816e-01 1.05288887e+00 4.01112586e-01 7.30709583e-02
-9.43392098e-01 7.84577310e-01 6.42138243e-01 5.50413370e-01
2.37377644e-01 -4.55506712e-01 -5.91986120e-01 -1.16342425e+00
2.98888296e-01 -8.17788422e-01 1.34947032e-01 -9.74362552e-01
-1.00797772e+00 2.28441935e-02 1.41681984e-01 -1.29928315e+00
-4.23245639e-01 -2.83566147e-01 -7.51589984e-02 6.75622523e-01
-1.11570704e+00 -1.55313897e+00 -4.82820898e-01 4.22211856e-01
5.96495092e-01 5.41004598e-01 6.04502439e-01 8.66969824e-02
-4.70074825e-02 1.70602962e-01 -4.15110767e-01 -8.24996382e-02
7.58954704e-01 -9.98216867e-01 5.13680637e-01 5.83281457e-01
6.11294024e-02 5.10363936e-01 8.51908505e-01 -1.29552925e+00
-1.61199951e+00 -8.58613074e-01 4.95423883e-01 -8.94731402e-01
2.44942978e-01 -2.28735521e-01 -5.97358346e-01 7.60630786e-01
-3.64553273e-01 2.18898118e-01 8.53808448e-02 -1.84296921e-01
-3.60368490e-01 1.61962166e-01 -1.07383585e+00 5.22929668e-01
1.33572340e+00 -2.86112368e-01 -5.45527220e-01 2.17007667e-01
2.76479155e-01 -1.27804875e+00 -1.18490326e+00 6.85090542e-01
1.25412548e+00 -9.95418251e-01 1.41014850e+00 -1.05536021e-01
6.48831427e-01 -3.72050017e-01 -1.54659212e-01 -9.66158986e-01
-1.10039376e-01 -5.93443811e-01 -6.60103977e-01 8.93645763e-01
-4.52062227e-02 -1.80117428e-01 1.12607324e+00 8.94451976e-01
-1.23941973e-01 -8.72370720e-01 -9.93437469e-01 -6.19633555e-01
-2.30344996e-01 -5.07405460e-01 4.44704533e-01 6.12891674e-01
-2.94817954e-01 -2.46358395e-01 -7.94954062e-01 1.27019003e-01
1.02810252e+00 3.25742543e-01 1.46431792e+00 -1.35108185e+00
-4.39650297e-01 5.62653802e-02 -4.66199577e-01 -1.48162317e+00
1.85104638e-01 -5.11742115e-01 3.49820048e-01 -1.59066379e+00
2.32807487e-01 -2.64764279e-01 4.41972911e-01 3.17351609e-01
-3.75015475e-02 6.33269548e-01 -2.82053538e-02 3.91768783e-01
-3.73742312e-01 4.91311967e-01 1.39516389e+00 3.86018679e-02
-7.94151723e-02 -2.04162598e-01 -5.87194897e-02 1.23795593e+00
1.93704814e-01 -4.54703540e-01 -1.54816717e-01 -4.11879331e-01
2.00750202e-01 2.21831903e-01 7.72632420e-01 -1.03948879e+00
2.69333303e-01 -8.16579312e-02 9.40221965e-01 -9.04940546e-01
8.78700197e-01 -8.53790104e-01 7.40673840e-01 5.76846242e-01
1.72861114e-01 -1.40532866e-01 -6.85486645e-02 7.58043528e-01
3.15629125e-01 1.69021204e-01 9.06861424e-01 -1.85146824e-01
-3.72635007e-01 7.07835376e-01 2.99199551e-01 4.28789929e-02
6.94423079e-01 -7.01971591e-01 1.69032678e-01 -4.63383496e-01
-6.94440544e-01 1.52201205e-01 1.08178186e+00 3.34909350e-01
8.10824096e-01 -1.33460128e+00 -6.24492466e-01 2.06413060e-01
-1.28306538e-01 2.49034077e-01 3.37008499e-02 1.23817039e+00
-6.59014940e-01 6.55687898e-02 -3.02256439e-02 -1.07069135e+00
-1.30455482e+00 8.47562253e-02 4.68009114e-01 -1.19836152e-01
-9.14595902e-01 7.48135328e-01 3.47803861e-01 -5.64396501e-01
1.32613137e-01 -4.03665602e-01 3.82269144e-01 -2.98846662e-01
1.88727736e-01 6.28937066e-01 -1.28951699e-01 -1.13958967e+00
-2.95106530e-01 1.12761092e+00 3.63401473e-01 -3.48423988e-01
1.23245859e+00 -2.19482467e-01 -9.83564556e-02 1.85734734e-01
1.01448083e+00 2.47040540e-01 -1.79012012e+00 -1.69795737e-01
-3.13928336e-01 -7.05983818e-01 -1.86997324e-01 -6.00460410e-01
-8.66899908e-01 5.20828605e-01 1.78179383e-01 -5.43554425e-01
6.10099435e-01 6.59139380e-02 1.13388598e+00 -1.99049592e-01
8.29997420e-01 -1.04821217e+00 -2.98714098e-02 1.94512904e-01
1.09481013e+00 -9.83066082e-01 6.23021424e-01 -1.17726541e+00
-1.53577104e-01 1.01493740e+00 6.75022483e-01 -3.71629685e-01
7.26990581e-01 1.91511869e-01 -2.97779649e-01 -3.32464188e-01
-4.25312728e-01 1.45023391e-01 8.62540722e-01 6.44375622e-01
9.88697559e-02 -1.39181122e-01 -1.82089448e-01 3.23916376e-01
-4.71851677e-01 1.40108749e-01 -8.60581473e-02 9.37389255e-01
-2.90732354e-01 -1.13988817e+00 -8.05707812e-01 3.92240465e-01
-2.19993502e-01 1.35259509e-01 -4.33961123e-01 8.18503737e-01
2.16091931e-01 5.45848012e-01 -3.04384390e-03 -2.20547542e-01
3.81978959e-01 -9.53753665e-02 1.00681007e+00 -5.51499546e-01
-2.61788040e-01 4.49447960e-01 3.18095297e-01 -1.10119069e+00
-5.46041846e-01 -8.58754992e-01 -1.08412790e+00 -2.27097988e-01
-3.02647918e-01 -1.98507592e-01 6.38278604e-01 1.06371391e+00
3.14541042e-01 -3.05096619e-02 5.07407598e-02 -1.56840861e+00
-3.52491021e-01 -5.72659373e-01 -4.15253431e-01 5.32980740e-01
-1.91214401e-02 -1.30037653e+00 -1.20212629e-01 3.94473344e-01] | [7.060630798339844, -1.1734353303909302] |
8f458e80-d86f-4507-bfc5-34542390a05f | acenet-anatomical-context-encoding-network | 2002.05773 | null | https://arxiv.org/abs/2002.05773v3 | https://arxiv.org/pdf/2002.05773v3.pdf | ACEnet: Anatomical Context-Encoding Network for Neuroanatomy Segmentation | Segmentation of brain structures from magnetic resonance (MR) scans plays an important role in the quantification of brain morphology. Since 3D deep learning models suffer from high computational cost, 2D deep learning methods are favored for their computational efficiency. However, existing 2D deep learning methods are not equipped to effectively capture 3D spatial contextual information that is needed to achieve accurate brain structure segmentation. In order to overcome this limitation, we develop an Anatomical Context-Encoding Network (ACEnet) to incorporate 3D spatial and anatomical contexts in 2D convolutional neural networks (CNNs) for efficient and accurate segmentation of brain structures from MR scans, consisting of 1) an anatomical context encoding module to incorporate anatomical information in 2D CNNs and 2) a spatial context encoding module to integrate 3D image information in 2D CNNs. In addition, a skull stripping module is adopted to guide the 2D CNNs to attend to the brain. Extensive experiments on three benchmark datasets have demonstrated that our method achieves promising performance compared with state-of-the-art alternative methods for brain structure segmentation in terms of both computational efficiency and segmentation accuracy. | ['Yuemeng Li', 'Yong Fan', 'Hongming Li'] | 2020-02-13 | null | null | null | null | ['skull-stripping'] | ['medical'] | [ 6.07286356e-02 1.03223033e-01 2.18853027e-01 -7.18449593e-01
-5.48678398e-01 -4.73958850e-02 1.19784623e-01 2.89267808e-01
-6.79685831e-01 4.02686208e-01 3.67674255e-03 -3.43697399e-01
-1.01830345e-02 -7.80812383e-01 -3.65804523e-01 -5.75325072e-01
-4.19051856e-01 5.39672673e-01 4.81826276e-01 2.74474639e-03
1.83960095e-01 9.13556635e-01 -9.00422513e-01 -1.33442566e-01
8.78555596e-01 1.08469093e+00 4.78111178e-01 2.04826593e-01
-6.25168920e-01 5.56103945e-01 -2.71912336e-01 2.23980546e-02
2.70859480e-01 -2.61484981e-01 -1.03867877e+00 8.07957277e-02
1.72975257e-01 -5.60866594e-01 -2.17687204e-01 1.19762087e+00
7.71372974e-01 5.31541705e-02 7.57853389e-01 -5.10052562e-01
-3.68436575e-01 2.94411570e-01 -7.11962819e-01 8.13762188e-01
-7.08230287e-02 -2.94212066e-02 4.79239821e-01 -8.16742361e-01
3.19462657e-01 9.18212295e-01 6.94562376e-01 5.05819619e-01
-1.02301872e+00 -5.71179748e-01 1.08765021e-01 -9.89982188e-02
-1.21919644e+00 -5.04195131e-02 1.04782832e+00 -7.09318757e-01
8.68965447e-01 -1.56806841e-01 1.03369474e+00 4.82836425e-01
3.89703870e-01 7.99417675e-01 1.00348568e+00 -8.66755322e-02
2.07208931e-01 -5.96078932e-01 2.49213248e-01 8.89396071e-01
7.77588040e-02 -8.23480487e-02 1.85762584e-01 4.35821600e-02
1.38125086e+00 1.18681699e-01 -5.59373423e-02 -2.36266211e-01
-1.01267529e+00 9.10043895e-01 1.12945735e+00 5.25962293e-01
-6.65115237e-01 7.88279027e-02 3.78325939e-01 -2.17922002e-01
6.85746968e-01 3.25107425e-01 -2.41811126e-01 3.52308303e-01
-1.02009797e+00 1.79113686e-01 1.64355978e-01 4.51356858e-01
4.32012975e-01 -1.98187064e-02 -2.31341675e-01 1.20130610e+00
3.02979201e-01 -2.30802502e-02 5.90295553e-01 -6.06911242e-01
3.59435618e-01 9.43131626e-01 -5.49835324e-01 -7.84440815e-01
-1.02830505e+00 -6.64328933e-01 -1.03577924e+00 1.58646494e-01
2.16636226e-01 -1.92461964e-02 -1.19584823e+00 1.35782444e+00
5.76093674e-01 2.63910238e-02 -4.57687795e-01 1.31150687e+00
1.13504136e+00 1.97473437e-01 1.69760674e-01 -2.09505893e-02
1.42359447e+00 -8.34747195e-01 -4.10693944e-01 -2.87738085e-01
7.18347490e-01 -5.12784481e-01 8.57283115e-01 -2.16530859e-01
-1.23542821e+00 -2.89407134e-01 -1.00062418e+00 -2.06635192e-01
-2.40985364e-01 -9.17203799e-02 6.69339299e-01 5.11965275e-01
-9.16110277e-01 2.81709433e-01 -1.18906772e+00 9.35830250e-02
1.15839326e+00 4.59668875e-01 -4.50563908e-01 -5.19110300e-02
-1.04615116e+00 8.92298877e-01 4.25213695e-01 2.86669254e-01
-8.42716575e-01 -8.28837454e-01 -8.94711375e-01 -1.53007388e-01
1.44161806e-01 -7.56571114e-01 1.12794411e+00 -4.17107373e-01
-1.12870049e+00 1.26079941e+00 1.52844280e-01 -4.21215087e-01
4.36457872e-01 1.26254678e-01 3.58408019e-02 6.30607665e-01
2.57056892e-01 8.03777993e-01 4.53587502e-01 -9.30015504e-01
-4.14374501e-01 -8.32373023e-01 1.34442508e-01 2.91244864e-01
1.13254882e-01 2.29447503e-02 -4.13603783e-01 -8.58105302e-01
7.72173584e-01 -5.02694190e-01 -5.77153862e-01 1.24639675e-01
-3.81838441e-01 -9.07187015e-02 7.29772091e-01 -9.50849712e-01
8.68480802e-01 -1.72954190e+00 7.18918368e-02 1.55146435e-01
6.94583178e-01 3.48941982e-01 7.47770965e-02 -4.16799843e-01
-1.01924837e-01 -2.25638878e-03 -7.21172512e-01 -2.97728807e-01
-3.49314630e-01 1.24538332e-01 4.21645075e-01 5.49236715e-01
2.68549412e-01 9.50258315e-01 -7.46955395e-01 -6.78726256e-01
3.56650233e-01 5.43273211e-01 -6.67061329e-01 3.57503772e-01
1.41232917e-02 9.16597784e-01 -8.00663710e-01 6.26591802e-01
7.79404044e-01 -1.58512920e-01 -1.35827422e-01 -2.60539740e-01
-4.64264341e-02 2.19459012e-01 -6.66587770e-01 1.94550562e+00
-2.99833506e-01 8.22002217e-02 3.32237035e-01 -1.63382494e+00
9.41071391e-01 4.41040874e-01 8.94889653e-01 -1.02225804e+00
4.77472425e-01 3.52886766e-01 1.62384078e-01 -7.19922245e-01
-2.91754007e-01 -4.19029057e-01 1.72058076e-01 4.20572728e-01
1.35315523e-01 -4.53226626e-01 2.64948513e-02 1.46230496e-03
9.67270076e-01 -1.02590002e-01 3.36039439e-02 -4.80805874e-01
6.84312284e-01 -1.63987726e-01 6.10144556e-01 3.86576563e-01
-4.91203040e-01 6.90662324e-01 5.16780078e-01 -9.48531628e-01
-1.06251454e+00 -9.39209819e-01 -3.87690723e-01 5.16030550e-01
1.07520722e-01 1.22100338e-01 -1.21398211e+00 -6.61435425e-01
-3.44735473e-01 2.13597566e-01 -7.46059060e-01 7.82595202e-02
-9.37346041e-01 -1.03505492e+00 4.12089854e-01 7.85851955e-01
9.59424496e-01 -8.71731162e-01 -1.01727629e+00 4.81435567e-01
-2.55451888e-01 -1.17293394e+00 -5.50484061e-01 2.46303454e-01
-1.29344356e+00 -1.30303848e+00 -1.09819567e+00 -9.93430078e-01
9.13249731e-01 6.21581115e-02 7.10312545e-01 4.20229614e-01
-6.59930468e-01 -7.70521089e-02 -2.19120532e-01 -3.56999129e-01
-1.84919797e-02 2.33953938e-01 -3.22546333e-01 -4.08024490e-01
1.84248567e-01 -7.88256705e-01 -8.94807637e-01 2.77185827e-01
-9.01476026e-01 3.52963835e-01 7.64452100e-01 7.61569500e-01
9.54560518e-01 5.00995219e-02 4.40250665e-01 -8.26364815e-01
5.62633634e-01 -3.76887769e-01 -4.67720449e-01 1.12814996e-02
-1.89840868e-01 -1.56227037e-01 3.75015408e-01 -1.61593914e-01
-8.60736549e-01 1.61113247e-01 -7.87808716e-01 -1.33594990e-01
-3.87285829e-01 6.81367338e-01 5.57511151e-02 -1.94344237e-01
2.94101268e-01 3.43841344e-01 2.73644358e-01 -6.59944534e-01
-1.35747358e-01 4.14039463e-01 6.50144339e-01 -6.56328321e-01
3.18814546e-01 4.68305171e-01 1.70145243e-01 -6.12819791e-01
-9.51929331e-01 -2.77509421e-01 -1.31726241e+00 -1.56912908e-01
1.35702276e+00 -6.04505897e-01 -1.97239459e-01 5.28840244e-01
-9.80847359e-01 -3.99036586e-01 1.02342159e-01 7.22453296e-01
-5.78065634e-01 3.00145537e-01 -8.49051535e-01 -3.50026548e-01
-8.26561034e-01 -1.82854009e+00 9.98116016e-01 2.55407661e-01
9.08168256e-02 -1.14773095e+00 -2.49885172e-01 4.68472987e-01
5.59983492e-01 6.60073221e-01 1.48509562e+00 -6.15296245e-01
-2.76532024e-01 -2.42445529e-01 -4.63467956e-01 2.93665737e-01
1.63455620e-01 -6.05456412e-01 -6.08888805e-01 -5.91737181e-02
1.10265799e-01 3.87362693e-03 6.44734859e-01 8.91259253e-01
1.47016358e+00 -2.50057690e-02 -2.48449549e-01 7.56410003e-01
1.20837080e+00 2.98540175e-01 3.54105324e-01 3.07362437e-01
7.49743581e-01 6.26376748e-01 4.69178781e-02 1.92895740e-01
5.21004498e-01 4.72087055e-01 6.44904673e-01 -4.64682102e-01
-3.62497181e-01 9.01394933e-02 -3.84345204e-01 9.24299181e-01
-1.34125233e-01 6.22616410e-01 -1.22406089e+00 5.74308395e-01
-1.39590108e+00 -4.36622709e-01 -1.93064332e-01 1.83821595e+00
8.89551997e-01 2.15611547e-01 2.46671960e-01 8.41449797e-02
7.35188901e-01 2.01713704e-02 -6.42476618e-01 -8.00030306e-02
1.90621004e-01 2.72341639e-01 1.51990235e-01 3.37221265e-01
-1.24598551e+00 7.51777709e-01 6.03565645e+00 6.70512736e-01
-1.23159969e+00 3.89620215e-01 8.14708352e-01 5.25515480e-03
-2.35405974e-02 -3.88822317e-01 -4.87673342e-01 4.19394732e-01
2.80010402e-01 3.57961953e-01 7.97485784e-02 6.56598091e-01
3.27459127e-01 -5.78666702e-02 -8.33553433e-01 1.01258147e+00
-2.75945008e-01 -1.34559155e+00 -4.31847572e-02 1.80824265e-01
5.10010898e-01 2.77298898e-01 -1.91511199e-01 2.22461037e-02
-2.34852776e-01 -1.18991387e+00 6.60682619e-01 4.35217232e-01
6.88350737e-01 -8.71451616e-01 8.82782578e-01 4.49607372e-01
-1.22456670e+00 1.05467841e-01 -4.43869025e-01 1.06061824e-01
2.51590520e-01 7.53348827e-01 -6.92666948e-01 3.97718310e-01
7.87006021e-01 4.01336074e-01 -2.91848987e-01 1.37585306e+00
-2.73710489e-01 3.45742434e-01 -2.73906559e-01 3.00388515e-01
8.00207615e-01 -1.97445288e-01 3.05727452e-01 1.06543660e+00
8.13050419e-02 5.88795960e-01 3.63619983e-01 1.07422507e+00
-1.02755316e-01 3.38726252e-01 -3.34587634e-01 2.72180170e-01
1.57560438e-01 1.22362363e+00 -1.29452622e+00 -2.36123025e-01
-3.86326641e-01 5.77094018e-01 4.45892572e-01 1.08896151e-01
-5.69372237e-01 -3.35213572e-01 5.44987798e-01 3.57700348e-01
7.44611993e-02 -5.34443736e-01 -5.01453817e-01 -8.56458724e-01
-1.56286120e-01 -2.44477123e-01 4.06946301e-01 -6.17327750e-01
-1.25241578e+00 7.14912355e-01 4.74414006e-02 -5.93749642e-01
2.92444289e-01 -5.34626126e-01 -7.25780129e-01 1.01961195e+00
-1.70170307e+00 -9.86309469e-01 -2.82829374e-01 7.06885099e-01
5.55923343e-01 2.65602786e-02 5.83811224e-01 4.27330375e-01
-6.41085804e-01 1.54826120e-01 -4.08141851e-01 6.13312185e-01
1.89913586e-01 -1.26056421e+00 3.04425061e-01 6.44282520e-01
-2.49749586e-01 6.01649642e-01 5.78096434e-02 -6.84535742e-01
-9.34202611e-01 -1.09925580e+00 4.89934534e-01 -2.51909103e-02
3.22174281e-01 -1.41295254e-01 -1.08050346e+00 5.37395418e-01
-2.23512098e-01 5.27777612e-01 7.85875320e-01 -3.20160061e-01
-2.53137257e-02 1.43613592e-01 -1.44402075e+00 3.99443865e-01
1.08430851e+00 -3.77581775e-01 -9.60937500e-01 3.72110963e-01
6.82811260e-01 -7.13760555e-01 -1.25274062e+00 6.56817794e-01
3.93603593e-01 -7.15516627e-01 1.17570722e+00 -3.10910553e-01
5.82830787e-01 -1.08198799e-01 5.35808643e-03 -1.16918671e+00
-2.98286766e-01 1.28036320e-01 1.08267903e-01 6.64005637e-01
4.95133512e-02 -5.22629201e-01 7.02147126e-01 6.90509439e-01
-6.21166945e-01 -1.29037607e+00 -1.06028867e+00 -4.65474546e-01
6.26451194e-01 -4.81109828e-01 7.74387777e-01 8.55280519e-01
-4.76683289e-01 1.97392061e-01 3.37910771e-01 9.56027284e-02
8.01794887e-01 1.49938837e-01 1.91453278e-01 -1.32914209e+00
4.20228988e-01 -9.65530396e-01 -4.78908896e-01 -1.15197980e+00
9.95422974e-02 -1.18123710e+00 -2.30562568e-01 -1.96123743e+00
4.30341393e-01 -6.02245808e-01 -4.13301528e-01 4.12630945e-01
-1.31640866e-01 3.70625496e-01 -1.82560399e-01 3.41114193e-01
-2.82628804e-01 5.75298071e-01 1.85604668e+00 -1.06288150e-01
-8.20137784e-02 -6.44640774e-02 -5.36802769e-01 8.50984156e-01
7.42543578e-01 -2.96315730e-01 -4.18859065e-01 -8.59171569e-01
-4.27368313e-01 1.48689359e-01 4.88345891e-01 -9.58581507e-01
3.22960198e-01 2.19311789e-01 6.97924614e-01 -8.68201494e-01
7.39342272e-02 -7.12613225e-01 -4.24019188e-01 5.17909527e-01
-2.35487238e-01 -7.47442469e-02 1.95004880e-01 7.48546422e-02
-2.85073280e-01 -2.48998553e-01 1.06064367e+00 -6.57156765e-01
-4.41561908e-01 1.03246105e+00 -4.94454175e-01 5.41960374e-02
9.03615296e-01 -4.45610970e-01 2.05049425e-01 2.33327866e-01
-1.04659081e+00 2.52089828e-01 1.75854847e-01 6.91424087e-02
8.16140354e-01 -1.22085667e+00 -5.32485127e-01 2.78388828e-01
-1.51276186e-01 5.78977585e-01 4.81634974e-01 1.32089174e+00
-9.99239147e-01 5.77433646e-01 -5.72999239e-01 -7.66326070e-01
-8.37346196e-01 1.40283689e-01 8.99805188e-01 -3.47110987e-01
-1.03856838e+00 9.87792671e-01 4.72183496e-01 -6.31054223e-01
2.42923200e-01 -6.15155518e-01 -4.52618062e-01 -1.95156053e-01
5.69253325e-01 -1.05961263e-01 3.27581078e-01 -8.01464677e-01
-5.07988453e-01 7.94813871e-01 -1.79988891e-01 2.37387344e-01
1.65239155e+00 -1.56349450e-01 -2.36532331e-01 -6.46909103e-02
1.40710592e+00 -6.46664917e-01 -1.37436080e+00 -4.68691856e-01
5.55202700e-02 -2.40547895e-01 6.01632237e-01 -7.75682688e-01
-1.61065292e+00 1.38031006e+00 6.75623477e-01 -1.32463023e-01
1.16630816e+00 3.11349356e-03 1.24458170e+00 -9.21085328e-02
4.86631542e-01 -9.80265737e-01 -2.78426018e-02 4.77451980e-01
8.31116557e-01 -1.31435263e+00 -1.36759102e-01 -3.34313422e-01
-2.68039912e-01 1.31117880e+00 7.69037724e-01 -2.40645260e-01
9.40074563e-01 3.15490514e-01 2.89704534e-03 -6.97894633e-01
-6.88218996e-02 -1.45392627e-01 3.00240785e-01 5.69050789e-01
5.60694277e-01 -4.79718447e-02 -3.90116811e-01 8.02400708e-01
-1.20585551e-02 -1.68461591e-01 -4.28705588e-02 9.08786356e-01
-5.35984933e-01 -8.25412691e-01 -2.36435339e-01 5.34989536e-01
-6.74083769e-01 -2.63529774e-02 -4.32051532e-02 6.20621383e-01
2.77997732e-01 3.75064403e-01 1.91536620e-01 -9.07259155e-03
2.83460498e-01 -1.32941514e-01 6.88575029e-01 -8.31992924e-01
-6.97667420e-01 3.09957325e-01 -3.58800173e-01 -4.61196631e-01
-3.61969322e-01 -3.98658425e-01 -1.73910856e+00 -4.39572632e-02
-1.23145804e-02 -8.22993964e-02 8.63162756e-01 1.24919605e+00
1.07795574e-01 7.34674990e-01 3.74311924e-01 -1.00544775e+00
-1.09363563e-01 -9.01706755e-01 -6.92252040e-01 2.30470479e-01
1.51754424e-01 -9.47772563e-01 2.59706676e-01 -1.64785877e-01] | [14.365531921386719, -2.393667221069336] |
ff3b0f68-a404-46e9-a65d-5ec57835c5a0 | yolov7-trainable-bag-of-freebies-sets-new | 2207.02696 | null | https://arxiv.org/abs/2207.02696v1 | https://arxiv.org/pdf/2207.02696v1.pdf | YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors | YOLOv7 surpasses all known object detectors in both speed and accuracy in the range from 5 FPS to 160 FPS and has the highest accuracy 56.8% AP among all known real-time object detectors with 30 FPS or higher on GPU V100. YOLOv7-E6 object detector (56 FPS V100, 55.9% AP) outperforms both transformer-based detector SWIN-L Cascade-Mask R-CNN (9.2 FPS A100, 53.9% AP) by 509% in speed and 2% in accuracy, and convolutional-based detector ConvNeXt-XL Cascade-Mask R-CNN (8.6 FPS A100, 55.2% AP) by 551% in speed and 0.7% AP in accuracy, as well as YOLOv7 outperforms: YOLOR, YOLOX, Scaled-YOLOv4, YOLOv5, DETR, Deformable DETR, DINO-5scale-R50, ViT-Adapter-B and many other object detectors in speed and accuracy. Moreover, we train YOLOv7 only on MS COCO dataset from scratch without using any other datasets or pre-trained weights. Source code is released in https://github.com/WongKinYiu/yolov7. | ['Hong-Yuan Mark Liao', 'Alexey Bochkovskiy', 'Chien-Yao Wang'] | 2022-07-06 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wang_YOLOv7_Trainable_Bag-of-Freebies_Sets_New_State-of-the-Art_for_Real-Time_Object_Detectors_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wang_YOLOv7_Trainable_Bag-of-Freebies_Sets_New_State-of-the-Art_for_Real-Time_Object_Detectors_CVPR_2023_paper.pdf | cvpr-2023-1 | ['real-time-object-detection', 'video-object-tracking'] | ['computer-vision', 'computer-vision'] | [-4.90444750e-01 -3.67801309e-01 8.23871866e-02 3.78290296e-01
-5.40796816e-01 -4.57771480e-01 8.04719329e-02 -1.53167218e-01
-6.75660431e-01 2.81821221e-01 -6.82527483e-01 -1.15289599e-01
7.27831364e-01 -8.57992828e-01 -8.21706116e-01 -4.70327973e-01
1.53151199e-01 8.40959400e-02 1.17818677e+00 -1.15036160e-01
-2.27352351e-01 6.14759743e-01 -1.44239509e+00 1.28241390e-01
2.66483516e-01 1.19414222e+00 1.61979496e-01 1.27080262e+00
4.64111626e-01 8.43902946e-01 -6.33897722e-01 -3.87514979e-01
4.09123003e-01 1.46452747e-02 -2.65275002e-01 -3.55721980e-01
5.69096446e-01 -4.99828219e-01 -6.13841116e-01 8.48482013e-01
7.13286221e-01 -2.83947349e-01 4.62493062e-01 -1.28409553e+00
-7.84779727e-01 2.53148228e-01 -9.63269413e-01 7.54964352e-01
6.99615851e-02 7.46951938e-01 3.86595160e-01 -1.26103234e+00
5.67073464e-01 1.15926933e+00 1.03862107e+00 6.99305236e-01
-9.88181055e-01 -1.02305770e+00 -1.62758768e-01 2.39729229e-02
-1.66407490e+00 3.48915383e-02 -1.14032224e-01 -4.53568101e-01
1.42378128e+00 -1.65564239e-01 5.10635853e-01 9.93849874e-01
4.26041335e-01 9.12475705e-01 6.68489277e-01 -1.37331411e-01
-1.35449201e-01 5.50906397e-02 1.73886701e-01 8.79580617e-01
3.53066146e-01 1.61214516e-01 -2.71465212e-01 1.87221184e-01
1.20608795e+00 8.56069326e-02 1.47824079e-01 3.54883969e-01
-1.17237687e+00 5.54556191e-01 6.85321569e-01 2.20422134e-01
-1.40478715e-01 6.81474984e-01 7.25797594e-01 1.42796576e-01
1.26703560e-01 7.19442666e-02 -6.64832950e-01 -1.74426287e-01
-5.79681098e-01 1.98714495e-01 2.49325320e-01 1.42825103e+00
4.51821029e-01 8.09941888e-01 -1.34089693e-01 4.52235937e-01
1.31497845e-01 1.14566004e+00 4.97099340e-01 -7.00795174e-01
3.15614998e-01 6.39370680e-01 3.67406964e-01 -6.05532885e-01
-6.08843207e-01 -5.00047386e-01 -6.55026197e-01 4.53970671e-01
4.86912638e-01 -1.93823472e-01 -1.10209382e+00 1.14275098e+00
2.92666912e-01 2.18118533e-01 1.34959996e-01 9.88847315e-01
1.47683930e+00 8.34869862e-01 2.64806062e-01 3.66285354e-01
1.66287839e+00 -1.13130116e+00 5.44687621e-02 -3.60414237e-02
6.50202990e-01 -1.09867144e+00 8.94020379e-01 4.27194744e-01
-1.12760472e+00 -1.28512812e+00 -1.00557709e+00 -2.09267169e-01
-1.36647344e-01 9.08200681e-01 2.96044379e-01 6.46858215e-01
-9.76124406e-01 3.98222685e-01 -1.05855870e+00 -5.56608140e-01
5.00028253e-01 5.30278623e-01 -1.52698740e-01 2.77421117e-01
-7.49637663e-01 7.52790987e-01 4.48042423e-01 -6.41369149e-02
-1.15525627e+00 -9.21214938e-01 -5.27173579e-01 -1.42723573e-02
2.90227175e-01 -3.94421875e-01 1.46440220e+00 -6.78247035e-01
-1.42292726e+00 9.71025527e-01 2.28860155e-01 -9.36879814e-01
8.80135298e-01 -5.59263766e-01 -6.84345067e-01 2.40518212e-01
9.98875424e-02 1.04764283e+00 6.67678177e-01 -5.77361107e-01
-9.77592111e-01 1.17728766e-02 -2.74991523e-02 -8.48577991e-02
8.36579576e-02 2.28577971e-01 -6.32367551e-01 -2.67812729e-01
-3.79965574e-01 -9.75202143e-01 -1.19037136e-01 3.22825044e-01
-3.57970297e-01 -4.49331522e-01 1.38359857e+00 -2.74395704e-01
8.11878204e-01 -2.23799682e+00 -6.92271471e-01 -4.09576237e-01
4.20484543e-01 1.27226973e+00 -2.37305492e-01 2.63529629e-01
1.58281997e-01 -3.14952314e-01 2.98754573e-01 -1.87989637e-01
-3.69764864e-01 -1.75898179e-01 -1.40131027e-01 7.37894177e-01
3.72223973e-01 1.15589440e+00 -9.93288219e-01 -6.01505041e-01
8.19081128e-01 8.22032332e-01 -4.44588393e-01 6.60599470e-02
1.86556533e-01 2.25871608e-01 -4.20000046e-01 1.26398325e+00
8.52154493e-01 -2.86441803e-01 -5.01439154e-01 -2.94278979e-01
-5.72468042e-01 -3.04021716e-01 -1.14045274e+00 1.16394341e+00
-4.16231900e-01 1.18300748e+00 -2.26523429e-01 -5.94363272e-01
1.10402954e+00 2.11394519e-01 4.50909525e-01 -6.16838515e-01
5.22532880e-01 1.98201507e-01 9.43779573e-03 -3.33298743e-01
5.80294907e-01 3.35685372e-01 2.93866582e-02 -1.30357429e-01
3.15746844e-01 1.21331781e-01 3.28182906e-01 2.92859286e-01
1.14443278e+00 9.15684104e-02 2.82157362e-01 -2.56259799e-01
4.95707721e-01 3.25269364e-02 4.78699923e-01 1.10721004e+00
-5.56582034e-01 8.14633906e-01 2.90668309e-01 -6.20837510e-01
-1.09904230e+00 -1.17018199e+00 -3.99493605e-01 8.90265822e-01
4.63282198e-01 -2.90828824e-01 -5.61573982e-01 -4.22195047e-01
3.10569674e-01 4.27592933e-01 -6.21649921e-01 -3.62164080e-02
-8.70667875e-01 -5.29529810e-01 9.40539956e-01 1.18857014e+00
8.05026174e-01 -1.37078047e+00 -1.18338966e+00 4.08325672e-01
2.57507920e-01 -1.38454366e+00 -3.18150759e-01 9.08667222e-02
-7.50386775e-01 -1.14192843e+00 -7.52575219e-01 -5.59833050e-01
3.58486474e-01 5.32329738e-01 1.18123627e+00 1.14386834e-01
-8.96558642e-01 4.07591373e-01 -4.95307565e-01 -6.78266943e-01
-1.99347571e-01 5.06353825e-02 1.94097415e-01 -3.38053733e-01
5.70667565e-01 1.99667048e-02 -9.74123895e-01 7.15015173e-01
-5.11236012e-01 -1.25523463e-01 3.90696824e-01 5.20283401e-01
6.00250602e-01 -6.96705282e-01 3.15742791e-01 -3.50094467e-01
-3.24404329e-01 -2.74044514e-01 -1.19493353e+00 -1.02094792e-01
-2.90713161e-01 -4.89180356e-01 8.12313080e-01 -9.11274910e-01
-5.49188018e-01 3.73550832e-01 -3.62788111e-01 -1.03307545e+00
6.83643743e-02 -6.85046136e-01 2.70538896e-01 -2.32670203e-01
1.08658624e+00 1.70744121e-01 -3.27510417e-01 -3.40262204e-01
3.61384898e-01 4.06909734e-01 7.69045711e-01 -4.79366966e-02
8.05318058e-01 6.98143363e-01 -2.45321020e-01 -9.42032278e-01
-6.18294835e-01 -8.35414588e-01 -4.80421424e-01 -2.86120802e-01
1.27865863e+00 -1.25070858e+00 -1.16158414e+00 7.05559790e-01
-9.97857988e-01 -3.44241112e-01 -4.51204062e-01 4.93496567e-01
-2.76801467e-01 -6.08907342e-02 -1.00212467e+00 -6.68764889e-01
-8.42671454e-01 -1.25246704e+00 1.14788306e+00 6.14211082e-01
8.51131380e-02 -4.34047699e-01 -4.45645362e-01 -2.74212286e-02
4.28915530e-01 3.74690026e-01 -4.09807324e-01 -2.58437157e-01
-5.77564716e-01 -4.71913517e-01 -9.21944797e-01 4.25269336e-01
-1.75390393e-01 5.62591016e-01 -8.87010992e-01 -4.07205403e-01
-4.34558362e-01 -3.98306251e-01 1.03242564e+00 7.76817381e-01
9.14797783e-01 2.80816555e-01 -5.10789573e-01 7.65518486e-01
1.74342489e+00 5.02444565e-01 5.00776172e-01 4.80538636e-01
8.31855237e-01 -5.59403479e-01 5.47744751e-01 4.51964080e-01
1.37260899e-01 6.27355397e-01 5.95310748e-01 -2.78382421e-01
-4.25707966e-01 -1.05115123e-01 4.63972896e-01 2.32355952e-01
-5.36900103e-01 -3.54573369e-01 -9.10468698e-01 3.76792341e-01
-1.72753298e+00 -1.03486073e+00 -7.53919125e-01 1.84593487e+00
1.53382763e-01 5.36700070e-01 6.63466156e-01 -2.00702146e-01
7.75226653e-01 -1.60832360e-01 -7.97938943e-01 -3.72552991e-01
-1.79130569e-01 1.84134603e-01 1.12330520e+00 1.09974675e-01
-1.32350957e+00 1.30078936e+00 5.33222342e+00 7.62920260e-01
-1.28208399e+00 3.41629505e-01 5.40256381e-01 -3.20350438e-01
9.45666850e-01 -2.60646284e-01 -1.52543652e+00 5.10333002e-01
8.67696404e-01 6.07043914e-02 -2.06456214e-01 1.53831291e+00
-3.41931023e-02 -2.98316628e-01 -9.70220625e-01 1.17191708e+00
-2.16737688e-01 -1.49256825e+00 -3.72722417e-01 -2.22322673e-01
5.54239154e-01 6.61714017e-01 7.09208250e-02 6.62648380e-01
4.46989745e-01 -8.98567319e-01 9.92845595e-01 8.09108466e-02
9.98301506e-01 -7.13728249e-01 8.73839676e-01 1.53476104e-01
-1.82260084e+00 7.48784617e-02 -9.72594082e-01 1.04974754e-01
2.63931286e-02 1.71147466e-01 -8.45936835e-01 1.01552904e-01
1.33497345e+00 8.28634679e-01 -5.08440316e-01 9.76374447e-01
-1.24878526e-01 7.56775737e-01 -6.26507461e-01 -2.43692517e-01
5.77287793e-01 3.37601304e-01 4.03148621e-01 1.78315616e+00
2.25615174e-01 3.59081268e-01 2.21389815e-01 3.83652627e-01
-8.24880823e-02 -1.63366511e-01 -1.49610564e-01 4.03049171e-01
3.04101676e-01 1.43132889e+00 -1.05285907e+00 -6.44665182e-01
-5.41363239e-01 9.88122523e-01 -2.40479177e-03 -1.17592430e-02
-1.47992063e+00 -2.93851048e-01 6.87610388e-01 2.14470387e-01
7.95504034e-01 -1.10546336e-01 2.42441922e-01 -1.18850625e+00
-2.36416608e-01 -4.94582146e-01 4.53808308e-01 -9.46977258e-01
-9.36605871e-01 8.25293720e-01 -2.78002769e-01 -1.68066370e+00
2.73462296e-01 -1.00436234e+00 -6.61420286e-01 4.68283594e-01
-1.26404703e+00 -1.17160833e+00 -5.26998580e-01 5.67569196e-01
8.61219645e-01 -1.55134141e-01 3.36081594e-01 4.70664799e-01
-8.17595720e-01 8.94343555e-01 1.16528295e-01 6.67781055e-01
4.27375555e-01 -1.02991211e+00 9.45724607e-01 7.55180597e-01
-7.76605383e-02 4.36874591e-02 4.20167983e-01 -3.72328550e-01
-1.43447483e+00 -1.41940403e+00 3.69268000e-01 -7.05431521e-01
8.38134170e-01 -3.00941288e-01 -9.18863058e-01 6.97143197e-01
4.52396348e-02 1.02716625e+00 -3.46245989e-03 -7.11844027e-01
-4.42291200e-01 -2.47070223e-01 -1.21363914e+00 2.92981476e-01
9.92866039e-01 -8.90594050e-02 -7.61867985e-02 4.68022555e-01
8.32409620e-01 -9.93708193e-01 -9.32153642e-01 2.06586257e-01
7.08246171e-01 -1.11667585e+00 1.23810554e+00 -1.52902633e-01
1.25491247e-01 -5.67159653e-01 1.02386093e-02 -4.65333939e-01
-4.13379908e-01 -3.19720238e-01 -1.26657352e-01 1.01553941e+00
2.86099166e-01 -8.63705218e-01 9.02680933e-01 -1.23242363e-01
-1.63784117e-01 -6.96707726e-01 -9.62332308e-01 -1.20526135e+00
-3.22784809e-03 -6.14290833e-01 1.31072417e-01 4.10487741e-01
-8.05343866e-01 -5.94361499e-02 -2.69924104e-01 1.70772880e-01
5.14738321e-01 8.13241675e-02 1.05753922e+00 -6.74800754e-01
-2.79477388e-01 -4.23151702e-01 -9.09556746e-01 -1.06378198e+00
-5.77555716e-01 -6.30006611e-01 -3.23270619e-01 -1.17335296e+00
6.73850700e-02 -3.79087806e-01 -3.37655544e-01 5.59674680e-01
-8.30661133e-02 8.93833101e-01 5.91698825e-01 3.89012247e-01
-7.40800321e-01 1.71864197e-01 1.08805108e+00 -3.57277282e-02
-3.89465898e-01 1.21644311e-01 -5.73999621e-02 9.32170570e-01
8.32890928e-01 -7.05864668e-01 5.98328337e-02 -2.54689246e-01
-3.64109874e-01 1.50110155e-01 5.32878995e-01 -1.50561452e+00
7.10934848e-02 2.34878466e-01 8.85080278e-01 -9.82851923e-01
3.84023368e-01 -3.26035619e-01 -6.56972229e-02 1.21745622e+00
3.37589920e-01 5.10746598e-01 6.20498896e-01 1.25293151e-01
3.88001166e-02 6.51721731e-02 1.16703963e+00 -1.18711732e-01
-1.17514598e+00 5.17363667e-01 -4.80979919e-01 2.31262922e-01
1.31448305e+00 -6.16359591e-01 -5.23062885e-01 2.29818806e-01
-6.60871148e-01 1.47260636e-01 2.97237098e-01 5.17670691e-01
5.78463733e-01 -1.07409644e+00 -9.94176924e-01 -3.09934728e-02
2.25487664e-01 -1.27916902e-01 3.75990659e-01 8.48756969e-01
-1.16594887e+00 4.21459377e-01 -2.59563386e-01 -1.08043098e+00
-1.28890002e+00 5.24991214e-01 4.76220369e-01 -6.86824992e-02
-1.21286511e+00 1.08990562e+00 2.43012577e-01 2.38651603e-01
8.85180309e-02 -6.94296837e-01 1.31428748e-01 -2.72540808e-01
6.45448864e-01 6.90704346e-01 -5.76346368e-02 -6.43699110e-01
-8.14706206e-01 6.31538391e-01 -5.70609048e-02 4.61178750e-01
1.05084705e+00 5.46114922e-01 6.56925321e-01 8.27577785e-02
1.13318396e+00 -2.21221313e-01 -1.59712100e+00 -1.63903497e-02
-3.07168037e-01 -3.33899498e-01 -3.11295837e-01 -4.13457870e-01
-1.23144412e+00 7.60827065e-01 1.13995445e+00 5.20954840e-02
1.01243615e+00 3.82473171e-01 8.59631181e-01 2.13287264e-01
1.98939145e-01 -8.44420254e-01 3.18090379e-01 4.60745394e-01
6.40395999e-01 -1.30440009e+00 1.28296226e-01 -2.53249526e-01
-6.57643318e-01 1.27850914e+00 1.31286430e+00 -9.30123091e-01
4.74768370e-01 7.10004032e-01 1.91241995e-01 1.01087339e-01
-6.24804676e-01 -3.99905294e-01 1.99854419e-01 4.82856512e-01
4.40639645e-01 2.62929678e-01 -9.06433538e-02 5.88630922e-02
-1.17364883e-01 -1.06194280e-01 5.16604781e-01 8.24615002e-01
-4.82951701e-01 -5.00400066e-01 -4.50226068e-01 3.37516099e-01
-5.82893372e-01 6.90142363e-02 7.74006322e-02 1.44424760e+00
4.05437112e-01 7.09431648e-01 3.44049662e-01 -2.96905756e-01
5.68720043e-01 -6.23931110e-01 3.01634341e-01 -3.01778167e-01
-1.00342858e+00 2.59434611e-01 -2.35229060e-01 -6.62065983e-01
-1.80890962e-01 -5.73942363e-01 -1.62451017e+00 -5.75400710e-01
-6.77532971e-01 -3.41714472e-01 5.32892287e-01 5.24982810e-01
2.51768917e-01 4.57149327e-01 2.42442444e-01 -1.02147031e+00
-1.65434718e-01 -8.74139547e-01 -1.53151631e-01 -1.44120738e-01
2.68932819e-01 -6.04634225e-01 -1.23826630e-01 6.41410276e-02] | [8.738911628723145, -0.24069388210773468] |
10aec96f-0d16-4de0-8295-79d096fae93c | effect-of-choice-of-probability-distribution | 1910.12383 | null | https://arxiv.org/abs/1910.12383v1 | https://arxiv.org/pdf/1910.12383v1.pdf | Effect of choice of probability distribution, randomness, and search methods for alignment modeling in sequence-to-sequence text-to-speech synthesis using hard alignment | Sequence-to-sequence text-to-speech (TTS) is dominated by soft-attention-based methods. Recently, hard-attention-based methods have been proposed to prevent fatal alignment errors, but their sampling method of discrete alignment is poorly investigated. This research investigates various combinations of sampling methods and probability distributions for alignment transition modeling in a hard-alignment-based sequence-to-sequence TTS method called SSNT-TTS. We clarify the common sampling methods of discrete variables including greedy search, beam search, and random sampling from a Bernoulli distribution in a more general way. Furthermore, we introduce the binary Concrete distribution to model discrete variables more properly. The results of a listening test shows that deterministic search is more preferable than stochastic search, and the binary Concrete distribution is robust with stochastic search for natural alignment transition. | ['Yusuke Yasuda', 'Junichi Yamagishi', 'Xin Wang'] | 2019-10-28 | null | null | null | null | ['hard-attention'] | ['methodology'] | [ 4.00921673e-01 -1.87907636e-01 -2.76183069e-01 -3.86823148e-01
-1.17341137e+00 -2.63101637e-01 3.64428043e-01 -3.68808538e-01
-4.36403632e-01 9.51026917e-01 3.39868546e-01 -6.38484478e-01
1.16901360e-01 -1.90546989e-01 -4.56677496e-01 -8.92397404e-01
4.72584903e-01 1.01285899e+00 2.41079032e-01 -1.97665989e-01
3.82164866e-01 1.63908243e-01 -1.32904601e+00 -6.31375164e-02
1.01382792e+00 5.36567569e-01 9.19202328e-01 1.15443885e+00
-5.28282702e-01 4.43084031e-01 -1.16591895e+00 -4.23493147e-01
-1.67018026e-02 -9.00234222e-01 -7.91821897e-01 -3.26265752e-01
1.93866819e-01 -1.76512703e-01 1.67632192e-01 1.16351950e+00
1.01968503e+00 2.10727960e-01 7.78167903e-01 -1.21581912e+00
-6.55856967e-01 9.38139856e-01 -3.14273655e-01 5.53431213e-01
6.25410795e-01 3.16629916e-01 1.09064412e+00 -5.35129130e-01
2.29443148e-01 1.47014964e+00 7.46934712e-01 5.56725919e-01
-9.61225986e-01 -6.05844736e-01 -9.72493589e-02 5.53096294e-01
-1.39134359e+00 -5.41717410e-01 4.43519145e-01 -1.63967460e-01
1.37463939e+00 6.92228436e-01 6.22893035e-01 1.56052458e+00
3.33090425e-01 8.64556372e-01 9.86760080e-01 -8.42500985e-01
1.72209099e-01 -2.10714221e-01 1.45416617e-01 1.52670652e-01
-3.59272286e-02 6.32630661e-02 -7.20726132e-01 -9.68567729e-02
4.85657096e-01 -4.30050671e-01 -2.01339990e-01 5.08439660e-01
-9.40169215e-01 7.68810153e-01 -5.68029284e-01 3.49220961e-01
-5.24056137e-01 1.49338106e-02 3.56841981e-01 1.51210234e-01
3.98305714e-01 3.54589462e-01 -4.58246857e-01 -9.45139050e-01
-1.16622186e+00 1.84201002e-01 6.65726185e-01 1.20701540e+00
3.38144302e-01 5.36599994e-01 -6.32002175e-01 1.10336828e+00
3.58084530e-01 1.00520027e+00 6.93761051e-01 -6.97037816e-01
5.56330204e-01 -3.27779710e-01 1.22033484e-01 -5.67960083e-01
-3.27137895e-02 -3.05021554e-01 -8.09092641e-01 -1.45407900e-01
3.04126829e-01 -2.29061544e-02 -9.53425825e-01 1.65999293e+00
9.10285339e-02 -2.17337050e-02 -1.81742266e-01 6.89226031e-01
4.69226032e-01 9.60816681e-01 -4.32208516e-02 -7.25250483e-01
1.10514688e+00 -9.49292660e-01 -1.48335791e+00 -1.58798546e-01
3.20538759e-01 -1.13244152e+00 1.54933643e+00 4.73868102e-01
-1.23355377e+00 -6.90111816e-01 -8.24722290e-01 9.77451652e-02
7.46082440e-02 -1.30254224e-01 -2.47992292e-01 1.00699961e+00
-1.14103496e+00 4.74186838e-01 -6.78399205e-01 -2.46053442e-01
-2.36209691e-01 5.72287589e-02 2.39353269e-01 3.05432528e-01
-1.49208391e+00 1.14440286e+00 2.73996025e-01 -5.00684567e-02
-6.87807322e-01 -4.42850888e-02 -6.18584156e-01 1.19098917e-01
2.48898864e-01 -5.63233674e-01 1.58389390e+00 -6.33796513e-01
-2.06651640e+00 5.56849241e-01 -6.89892352e-01 -6.10564530e-01
2.87694514e-01 -5.53102136e-01 -2.80383170e-01 -3.68259707e-03
-9.78756174e-02 2.11571991e-01 9.77070570e-01 -9.38987255e-01
-3.82358104e-01 1.77994221e-01 -6.12279236e-01 4.04123127e-01
-4.54278365e-02 6.08258486e-01 -2.64780611e-01 -7.65700817e-01
-1.48186311e-01 -7.02006817e-01 -1.76232204e-01 -7.11161911e-01
-6.75904036e-01 -3.14712107e-01 3.75087649e-01 -1.06942201e+00
1.89739370e+00 -1.94824541e+00 1.71171144e-01 -1.07037090e-01
-3.70208234e-01 5.67042053e-01 -2.05584392e-01 7.19471991e-01
-2.43185963e-02 2.94797152e-01 -1.48909211e-01 -7.12729514e-01
1.20265096e-01 3.20221722e-01 -2.93523818e-01 8.25641975e-02
-2.31163114e-01 6.36487782e-01 -8.13030958e-01 -8.14590275e-01
2.76104569e-01 5.01921698e-02 -3.38358372e-01 7.27019012e-01
-3.56314152e-01 4.40951854e-01 -9.11256820e-02 6.54075623e-01
5.14768898e-01 1.36055261e-01 2.43707951e-02 1.29326999e-01
-2.22969100e-01 5.48225522e-01 -8.33802342e-01 1.22538769e+00
-3.88920933e-01 6.91437006e-01 -1.52072877e-01 -7.09986031e-01
1.15481853e+00 4.80976820e-01 8.86006728e-02 -4.58432615e-01
2.26828322e-01 2.33198851e-01 3.44761580e-01 -8.19883227e-01
8.93082678e-01 -1.07834861e-01 2.10659117e-01 1.08144723e-01
-2.50669628e-01 -4.26875830e-01 -2.10171249e-02 -8.36710930e-02
8.96883667e-01 1.69097587e-01 5.01311183e-01 -6.37992471e-02
2.37294599e-01 -2.08255321e-01 7.60388851e-01 1.19702590e+00
-2.60986298e-01 9.33783412e-01 2.43829548e-01 5.58811761e-02
-1.30184460e+00 -9.91676688e-01 3.25855911e-02 9.42099452e-01
-9.61308330e-02 -5.04010439e-01 -9.94198918e-01 -3.71556222e-01
-5.13821483e-01 1.26107347e+00 -6.57534748e-02 -7.99103230e-02
-6.03520811e-01 -3.99410933e-01 7.40421593e-01 2.67293155e-01
1.75243303e-01 -1.38356256e+00 -2.70794392e-01 4.17127520e-01
-8.01586151e-01 -9.80444789e-01 -8.26242149e-01 2.54755884e-01
-5.43518066e-01 -5.26299596e-01 -1.02335691e+00 -5.72584987e-01
-2.72774585e-02 -8.67860690e-02 9.72969234e-01 -6.04915284e-02
-2.08075140e-02 3.71796452e-02 -8.39867711e-01 -5.32657743e-01
-9.71597373e-01 1.40641764e-01 2.27049783e-01 -1.73969552e-01
4.41920877e-01 -3.16520959e-01 -2.36237749e-01 3.66733581e-01
-5.54917634e-01 -1.18799232e-01 6.37362361e-01 1.09025395e+00
4.15799141e-01 -3.37840289e-01 2.00930372e-01 -3.77117574e-01
8.93517435e-01 -2.34691218e-01 -2.85971105e-01 4.24308866e-01
-5.08933544e-01 4.42707986e-02 5.94330549e-01 -7.90704429e-01
-8.68435264e-01 -4.43873227e-01 -7.83881426e-01 -5.90886235e-01
-2.68028408e-01 4.99099135e-01 -4.16908324e-01 3.70014340e-01
4.70202774e-01 8.93108547e-01 1.79153696e-01 -6.07128084e-01
-1.07761487e-01 1.28509259e+00 3.83203864e-01 -5.51921546e-01
3.32617640e-01 -2.60778189e-01 -4.01699126e-01 -1.05921113e+00
-3.84337693e-01 -4.06660259e-01 -2.09278241e-01 -2.48542219e-01
9.92315173e-01 -2.85712510e-01 -5.20860493e-01 8.97770286e-01
-1.42773151e+00 -5.56398094e-01 -6.30359799e-02 6.83608115e-01
-8.55061471e-01 9.21307981e-01 -7.07562149e-01 -1.38393533e+00
-5.59563935e-01 -1.42779374e+00 1.23467112e+00 8.83971155e-02
-6.82092190e-01 -5.51677585e-01 1.98034033e-01 3.47839862e-01
5.48582256e-01 -3.88361156e-01 7.86110282e-01 -8.29568207e-01
-2.86671311e-01 2.41067801e-02 4.14096564e-01 5.82877219e-01
7.75798634e-02 3.69015187e-01 -5.95075727e-01 -1.36894673e-01
1.53785110e-01 1.12220488e-01 3.03477108e-01 7.97488630e-01
1.15809429e+00 -4.50010031e-01 -1.72514781e-01 4.34055239e-01
9.58199441e-01 7.94685721e-01 1.05374908e+00 3.87311578e-01
4.16556895e-01 1.78291857e-01 1.03880799e+00 5.46452641e-01
1.22878574e-01 9.26332951e-01 2.78619856e-01 1.07013173e-02
-1.29417986e-01 -1.49445400e-01 5.62099040e-01 1.38083613e+00
6.79196641e-02 -1.00514853e+00 -8.86356890e-01 5.20291805e-01
-1.60356176e+00 -1.36450434e+00 -2.57064372e-01 2.27630973e+00
1.23096514e+00 3.35607260e-01 2.71311760e-01 2.89873511e-01
1.36298394e+00 2.62428284e-01 -1.54140843e-02 -9.28337455e-01
-2.74220943e-01 1.97713256e-01 4.10992801e-01 8.16998184e-01
-4.87738460e-01 1.20579743e+00 7.37742043e+00 1.55854201e+00
-8.73316705e-01 1.31932169e-01 4.07904029e-01 7.55070662e-03
-4.83650267e-01 5.93623780e-02 -1.27360821e+00 1.09127176e+00
1.25831425e+00 1.83421835e-01 2.52959967e-01 5.50560117e-01
5.24533689e-01 -5.22796884e-02 -7.24355578e-01 1.07778811e+00
-9.04077068e-02 -9.88251925e-01 1.82154343e-01 -1.86027549e-02
3.36511165e-01 -2.12180838e-01 -6.00512810e-02 1.24065146e-01
3.75415742e-01 -9.24992859e-01 1.02437186e+00 6.12396061e-01
8.24343026e-01 -4.85428542e-01 8.81712437e-01 4.36209917e-01
-8.65792036e-01 2.69613564e-01 -3.32132876e-01 -5.28663918e-02
7.91542053e-01 6.51479602e-01 -9.57520425e-01 5.40683091e-01
5.08694351e-01 1.50808126e-01 1.84269026e-01 1.27918136e+00
-1.84743911e-01 1.38485050e+00 -4.63288724e-01 -7.21101880e-01
2.12375730e-01 -3.03473324e-01 1.07604945e+00 1.44351256e+00
6.43860459e-01 -1.81372151e-01 -2.68389583e-02 5.00317931e-01
7.09599376e-01 2.07882479e-01 -2.26863593e-01 -2.62673181e-02
1.03668153e+00 4.00562644e-01 -3.72417778e-01 -3.44766170e-01
-8.82908627e-02 1.00650227e+00 1.04797184e-01 4.20811743e-01
-1.16526592e+00 -5.50339460e-01 6.11076891e-01 -1.04871407e-01
4.97194469e-01 -4.59420055e-01 -2.69869149e-01 -5.99741638e-01
-3.99573296e-02 -1.26142347e+00 9.58650261e-02 -1.08695853e+00
-1.08479440e+00 8.66603315e-01 2.37301916e-01 -1.08291304e+00
-4.79591817e-01 -1.74115494e-01 -7.38121331e-01 1.37275636e+00
-1.01112616e+00 -5.09308755e-01 8.56252983e-02 2.19548866e-01
1.15205586e+00 -3.45240951e-01 6.95357680e-01 2.10615113e-01
-6.51871324e-01 9.82879639e-01 2.51960516e-01 -1.43184558e-01
6.81980014e-01 -1.11307561e+00 8.18890512e-01 9.16146934e-01
9.73942056e-02 5.90111971e-01 1.27640641e+00 -9.53148425e-01
-1.04853892e+00 -4.83007371e-01 1.32809722e+00 -4.46755663e-02
1.75384685e-01 -1.50001824e-01 -9.35284674e-01 3.68629545e-01
5.40328503e-01 -8.29167306e-01 5.85753024e-01 8.89726505e-02
9.02307555e-02 1.31657541e-01 -9.24405396e-01 9.84418750e-01
1.02769220e+00 -4.09412026e-01 -8.99890900e-01 2.01284409e-01
9.22018945e-01 -5.94499528e-01 -4.80023891e-01 3.29780430e-01
4.23521131e-01 -9.73623574e-01 7.56120861e-01 -3.38164836e-01
2.95550767e-02 -3.20353508e-01 -3.82899880e-01 -1.53531337e+00
-5.97796321e-01 -1.28359139e+00 2.31696456e-03 1.42257726e+00
4.07783598e-01 -6.89439416e-01 5.72444022e-01 7.00374320e-02
-6.46244884e-01 -6.67312920e-01 -1.35411072e+00 -1.27114022e+00
-1.28233433e-01 -5.43212593e-01 7.23280728e-01 5.75275898e-01
-1.52747259e-01 2.40870997e-01 -7.19227910e-01 3.81533504e-02
2.38266736e-01 -2.81526238e-01 6.83886111e-01 -4.79622900e-01
-7.26079106e-01 -6.51775062e-01 -2.36957282e-01 -1.52892852e+00
-6.80946326e-03 -2.85115361e-01 5.31904459e-01 -1.46285117e+00
2.00444181e-03 -3.20875764e-01 7.21238479e-02 1.30931348e-01
-6.09819531e-01 -3.05627495e-01 1.07157141e-01 1.70902133e-01
-2.10826352e-01 9.03325319e-01 1.02338922e+00 -3.69490571e-02
-8.51300582e-02 3.49077284e-01 -5.86529747e-02 2.57571399e-01
9.56901848e-01 -7.36610770e-01 -3.02625716e-01 -2.19677195e-01
-2.05112591e-01 3.94596100e-01 -3.52867842e-01 -7.93069005e-01
1.47832081e-01 -4.47381884e-01 -2.90846348e-01 -1.05827999e+00
4.32897210e-01 -4.11078870e-01 3.47083718e-01 4.50574636e-01
-4.34950501e-01 3.61449063e-01 1.03885472e-01 3.35042059e-01
-1.91326141e-01 -8.23517561e-01 5.68248689e-01 6.69953355e-04
-2.62841046e-01 -3.43890525e-02 -9.36284542e-01 4.22069460e-01
6.46872044e-01 -7.03398228e-01 1.57326814e-02 -8.53619635e-01
-7.02812910e-01 -7.17983842e-02 1.77138805e-01 1.81401283e-01
4.85101432e-01 -1.07355762e+00 -1.00035226e+00 7.38609433e-02
-3.03496093e-01 -2.03847587e-01 9.94975343e-02 8.44804287e-01
-5.57121992e-01 2.27678344e-01 2.90650167e-02 -9.20380831e-01
-1.82046366e+00 3.59240472e-01 1.21073097e-01 -7.60265142e-02
-3.10365379e-01 1.18152165e+00 -2.37260208e-01 -1.06139809e-01
6.30834758e-01 -2.56899714e-01 9.42870006e-02 -3.36140364e-01
3.64875615e-01 5.31504214e-01 -1.36837214e-01 -4.71985966e-01
-4.39344466e-01 5.11218548e-01 -1.04563579e-01 -5.43727398e-01
6.66698277e-01 -5.02354562e-01 6.93300366e-02 6.09734654e-01
8.31384957e-01 2.12026775e-01 -9.57976401e-01 -2.22214565e-01
-3.32854539e-02 -6.37077808e-01 -3.49639654e-01 -7.30925679e-01
-4.25145149e-01 9.75098133e-01 1.92262203e-01 5.94750702e-01
8.60435188e-01 -2.20027789e-01 1.02711141e+00 -9.52001289e-02
2.38271028e-01 -1.22352850e+00 -1.18689604e-01 9.85442400e-01
1.00455022e+00 -8.63337398e-01 -2.78188765e-01 -3.24573576e-01
-8.30947876e-01 1.08763254e+00 5.63969672e-01 5.07130623e-01
2.63705969e-01 5.39373100e-01 3.01025480e-01 5.10972679e-01
-5.53635299e-01 -3.37427914e-01 1.20011240e-01 8.97683382e-01
5.06090641e-01 7.41296634e-02 -4.89193320e-01 5.37460506e-01
-6.03340805e-01 -2.78179049e-01 3.70060384e-01 6.59158528e-01
-6.25088394e-01 -1.30331910e+00 -7.48065650e-01 2.19322413e-01
-5.23311913e-01 -8.05764139e-01 -3.51699620e-01 3.17349374e-01
-2.71089584e-01 1.16259277e+00 1.75085917e-01 -5.58884621e-01
-5.11915758e-02 3.10897470e-01 4.29886192e-01 -5.05571663e-01
-6.23532593e-01 4.18376982e-01 3.79900783e-01 -1.12976417e-01
-8.54538009e-03 -9.13378239e-01 -9.95470524e-01 -4.12450701e-01
-7.19011605e-01 4.47191387e-01 6.63086474e-01 1.03809476e+00
2.22609043e-01 5.54085851e-01 5.80986500e-01 -8.47623467e-01
-1.04956234e+00 -1.45091450e+00 -4.58396286e-01 5.36698736e-02
3.63494903e-01 -3.91893804e-01 -5.34853816e-01 -2.67053455e-01] | [14.636889457702637, 6.828321933746338] |
6ee04a20-8b29-43e3-84ed-a265ded4a6b7 | eiseg-an-efficient-interactive-segmentation | 2210.08788 | null | https://arxiv.org/abs/2210.08788v2 | https://arxiv.org/pdf/2210.08788v2.pdf | EISeg: An Efficient Interactive Segmentation Tool based on PaddlePaddle | In recent years, the rapid development of deep learning has brought great advancements to image and video segmentation methods based on neural networks. However, to unleash the full potential of such models, large numbers of high-quality annotated images are necessary for model training. Currently, many widely used open-source image segmentation software relies heavily on manual annotation which is tedious and time-consuming. In this work, we introduce EISeg, an Efficient Interactive SEGmentation annotation tool that can drastically improve image segmentation annotation efficiency, generating highly accurate segmentation masks with only a few clicks. We also provide various domain-specific models for remote sensing, medical imaging, industrial quality inspections, human segmentation, and temporal aware models for video segmentation. The source code for our algorithm and user interface are available at: https://github.com/PaddlePaddle/PaddleSeg. | ['Baohua Lai', 'Zeyu Chen', 'Zewu Wu', 'Guowei Chen', 'Shiyu Tang', 'Juncai Peng', 'Lin Han', 'Yizhou Chen', 'Yi Liu', 'Yuying Hao'] | 2022-10-17 | null | null | null | null | ['interactive-segmentation'] | ['computer-vision'] | [ 3.36922795e-01 -6.73617274e-02 2.99871862e-02 -3.90856177e-01
-8.99879575e-01 -5.85447371e-01 -7.39179850e-02 5.12689129e-02
-3.78218979e-01 4.80625302e-01 -5.90437055e-01 -7.13323236e-01
1.97421983e-01 -9.32702243e-01 -3.99116307e-01 -5.39846301e-01
2.02615023e-01 3.44655603e-01 5.04322529e-01 1.43129021e-01
4.11154181e-02 4.43102539e-01 -1.23212969e+00 4.25649015e-03
1.25528836e+00 8.40317369e-01 6.02872729e-01 8.11522245e-01
-2.40828380e-01 1.84240177e-01 -4.21118766e-01 -1.30669445e-01
1.66984856e-01 -3.77636611e-01 -9.15194571e-01 1.62749067e-01
-1.61489155e-02 -4.63735044e-01 4.53764535e-02 1.14215684e+00
4.69147921e-01 -4.78465459e-04 3.33122432e-01 -9.64699149e-01
-4.71256822e-01 4.00669754e-01 -6.24266028e-01 1.48573056e-01
-6.34191185e-02 3.98910552e-01 6.64107084e-01 -5.79037666e-01
4.71233517e-01 6.78617954e-01 6.80449188e-01 4.12012488e-01
-9.26565826e-01 -8.15875828e-01 5.93956374e-03 4.01984788e-02
-1.40660000e+00 -1.75468788e-01 5.88670313e-01 -6.19226456e-01
7.54078805e-01 3.40698272e-01 9.19511676e-01 5.72702229e-01
-4.39714491e-02 7.88780689e-01 8.14398229e-01 -2.60437787e-01
7.76116326e-02 -9.47955996e-02 1.85707659e-02 8.16546619e-01
4.59716357e-02 -2.83949316e-01 6.38623163e-02 1.46632001e-01
1.23333311e+00 6.68393075e-02 -2.47122765e-01 -7.11998567e-02
-1.03291929e+00 7.60596931e-01 4.25167143e-01 3.61433715e-01
-4.00573939e-01 6.06950372e-02 2.63495386e-01 -2.25264475e-01
7.00059950e-01 2.01520190e-01 -7.27671862e-01 -1.21347569e-01
-1.19368577e+00 -3.69973890e-02 5.07737458e-01 7.97882915e-01
1.09831953e+00 -6.33518100e-02 1.38152167e-01 8.60769272e-01
2.55237103e-01 4.13367599e-01 2.17539907e-01 -1.17789221e+00
8.26021805e-02 6.03660822e-01 1.53486341e-01 -8.88215125e-01
-4.94671524e-01 -1.68633342e-01 -8.29730988e-01 1.55317277e-01
4.55039054e-01 -2.54257470e-01 -1.20869374e+00 1.01876521e+00
5.30722022e-01 8.88941959e-02 -4.37235951e-01 8.98647785e-01
8.39497149e-01 7.71253347e-01 3.95868540e-01 3.49452123e-02
1.20424616e+00 -1.12367940e+00 -5.94173551e-01 -2.49209955e-01
6.46257818e-01 -7.42494106e-01 1.40480125e+00 3.43341351e-01
-9.29255903e-01 -5.16623199e-01 -8.72427821e-01 -1.54313013e-01
-5.71333885e-01 2.53783971e-01 8.15782249e-01 6.11141682e-01
-7.32376695e-01 6.93274796e-01 -1.30875969e+00 -4.31259334e-01
9.05038655e-01 3.43789935e-01 -1.01311244e-01 2.98557598e-02
-9.30140078e-01 3.75891447e-01 3.87327105e-01 2.96779960e-01
-7.58515418e-01 -4.99893457e-01 -8.10934484e-01 -1.26876011e-01
6.75859392e-01 -3.34456146e-01 1.49867678e+00 -9.51763988e-01
-1.37065589e+00 9.49357450e-01 2.56027076e-02 -7.18858466e-02
5.65585554e-01 -3.07877660e-01 -1.19759813e-01 2.59610772e-01
1.76711991e-01 9.27276969e-01 3.90625119e-01 -1.11968076e+00
-5.65484524e-01 -2.25588918e-01 -5.56020141e-02 -8.19004327e-02
-1.81486294e-01 1.67923883e-01 -8.74856293e-01 -5.04475534e-01
6.79236352e-02 -7.01195002e-01 -6.74768090e-01 1.92206010e-01
-4.69515383e-01 -3.60731930e-02 1.04221594e+00 -1.13788259e+00
1.14104557e+00 -2.11404634e+00 -3.53254855e-01 -3.63852046e-02
2.01153252e-02 6.03566527e-01 6.10224232e-02 1.19189620e-01
1.25265017e-01 4.83659208e-01 -8.25463295e-01 2.88972910e-02
-2.06906855e-01 1.96662143e-01 2.35545173e-01 3.34030509e-01
1.65456012e-02 9.54865336e-01 -7.88296819e-01 -7.04973161e-01
6.23773456e-01 4.85265940e-01 -3.45063388e-01 2.25743145e-01
-4.85473841e-01 1.01089883e+00 -5.37841439e-01 8.97361696e-01
6.88534856e-01 -5.38726866e-01 8.92749727e-02 -1.84147790e-01
-2.50169218e-01 -9.56137776e-02 -1.08830643e+00 1.85782444e+00
-3.12407732e-01 4.41686869e-01 3.69921207e-01 -9.70448256e-01
6.16730869e-01 2.52648175e-01 7.36413598e-01 -4.60535824e-01
2.56121278e-01 2.48690382e-01 -4.86834735e-01 -7.28028893e-01
3.30472350e-01 2.60946304e-01 1.57067522e-01 4.54984605e-01
-1.82920367e-01 -3.20662826e-01 3.75532776e-01 -6.37660772e-02
6.43155515e-01 5.50097048e-01 2.84428179e-01 2.07821839e-02
3.47782880e-01 3.71393114e-01 7.38234043e-01 4.60989833e-01
-2.83245116e-01 7.79943943e-01 1.29728079e-01 -5.00785589e-01
-1.02106297e+00 -7.01245964e-01 -7.14578182e-02 1.02253950e+00
1.23858534e-01 -1.97425351e-01 -1.33910942e+00 -4.78606492e-01
-4.26581472e-01 4.63492632e-01 -1.61904618e-01 4.78640050e-01
-5.17416656e-01 -8.69430959e-01 5.76628387e-01 5.96032917e-01
8.91076565e-01 -1.28640819e+00 -8.79476309e-01 1.73344374e-01
-4.58884031e-01 -1.14647257e+00 -3.96492809e-01 1.71053056e-02
-1.00601351e+00 -1.18664420e+00 -8.75819027e-01 -7.26768792e-01
7.77630270e-01 2.47956559e-01 9.95844483e-01 4.91963506e-01
-7.08093226e-01 1.78264156e-01 -3.25852931e-01 -5.14186502e-01
-2.08398655e-01 2.86418736e-01 -6.58475459e-01 -4.47262347e-01
3.01881373e-01 -3.75501156e-01 -7.18871713e-01 3.58204424e-01
-1.17337155e+00 5.46227276e-01 4.06496882e-01 3.60156775e-01
9.01293337e-01 1.50016934e-01 2.82766640e-01 -1.04289663e+00
2.08025932e-01 -1.53587475e-01 -1.02974224e+00 2.17139885e-01
-4.65873539e-01 -5.74330151e-01 2.91439176e-01 -1.10706508e-01
-1.10585439e+00 3.68721396e-01 -6.00836813e-01 -2.27145866e-01
-7.82294095e-01 7.32186019e-01 -1.77882522e-01 1.00991704e-01
4.58044052e-01 4.58345041e-02 -6.76852316e-02 -6.49298370e-01
3.77263665e-01 9.10446584e-01 4.61911082e-01 -3.59336883e-01
5.26914358e-01 4.52738076e-01 -4.86836106e-01 -9.56966043e-01
-8.09171796e-01 -4.34849560e-01 -9.77273285e-01 -3.58773172e-01
1.30601895e+00 -6.80245697e-01 -3.48161191e-01 8.59416783e-01
-9.72391129e-01 -9.40223038e-01 -8.56438931e-03 1.74683407e-01
-2.46190995e-01 4.78468478e-01 -6.92894638e-01 -7.13485956e-01
-5.38384378e-01 -1.27284884e+00 1.04444861e+00 5.65806150e-01
-2.52271742e-02 -9.81572866e-01 -3.52365196e-01 6.93148077e-01
3.16255361e-01 4.19741511e-01 7.23553896e-01 -3.96234006e-01
-8.56753290e-01 -2.60523140e-01 -3.70363325e-01 5.19331694e-01
2.14516714e-01 3.48082751e-01 -9.58586693e-01 6.73594549e-02
-3.73511642e-01 -1.41389385e-01 5.37159562e-01 7.39895701e-01
1.74347043e+00 8.00108016e-02 -5.42766392e-01 7.71897137e-01
1.38180459e+00 4.82335001e-01 5.93426049e-01 3.45955908e-01
1.00319219e+00 5.85323036e-01 7.23100901e-01 3.97475034e-01
3.30991417e-01 3.98520708e-01 3.66073877e-01 -6.94853246e-01
2.35730857e-01 2.94883084e-02 -2.44048342e-01 7.35238671e-01
-2.77203083e-01 -2.54497051e-01 -1.26711607e+00 5.64109743e-01
-1.67931426e+00 -6.75678194e-01 -4.47859615e-01 1.94233680e+00
8.97257209e-01 -1.36451513e-01 -8.29482544e-03 1.10467754e-01
8.07192862e-01 -3.15037817e-02 -6.95539773e-01 -1.08782658e-02
3.07556182e-01 3.00550878e-01 5.11616707e-01 4.49301511e-01
-1.49737799e+00 1.24446273e+00 6.25974369e+00 7.91501880e-01
-1.19701791e+00 1.97349086e-01 9.81163919e-01 2.37321869e-01
1.33280205e-02 -1.43615678e-01 -5.62390983e-01 2.96302289e-01
8.08988094e-01 1.26462445e-01 2.56541938e-01 9.44741726e-01
6.14363968e-01 -4.69473124e-01 -4.32941705e-01 7.84867764e-01
-4.00051951e-01 -1.34416449e+00 -4.06283677e-01 5.68743125e-02
6.52744174e-01 1.95056006e-01 -1.91561922e-01 -1.04981877e-01
1.90076917e-01 -9.11414206e-01 3.61022353e-01 1.81546986e-01
8.64965677e-01 -5.77852786e-01 7.00740278e-01 3.40298414e-01
-1.21929491e+00 2.13978305e-01 -2.80538380e-01 9.17895883e-02
5.82619667e-01 7.82076359e-01 -5.43026984e-01 5.02934694e-01
9.13189769e-01 5.95397472e-01 -5.42148769e-01 1.13328147e+00
-3.60443205e-01 8.39503706e-01 -2.39632159e-01 4.38980699e-01
2.30695888e-01 -5.83897114e-01 3.94787900e-02 1.19907415e+00
3.02965462e-01 2.64591992e-01 5.25116444e-01 9.02235150e-01
1.20343708e-01 1.47964492e-01 -3.06991965e-01 -2.94241607e-01
2.91058213e-01 1.61488914e+00 -1.50938284e+00 -3.97524714e-01
-4.30358410e-01 1.02633500e+00 -2.26415217e-01 3.72442931e-01
-1.24779809e+00 -5.90024173e-01 4.03938413e-01 1.24356404e-01
1.65905640e-01 -4.50031608e-01 -4.37974215e-01 -8.81957829e-01
-2.19569728e-01 -7.29037344e-01 2.87856638e-01 -8.72233808e-01
-7.98514187e-01 4.87797797e-01 1.69799868e-02 -8.62512946e-01
1.06790394e-01 -4.97454077e-01 -6.17768884e-01 7.87052631e-01
-1.30350828e+00 -1.21694517e+00 -7.44839132e-01 3.33685070e-01
7.54432201e-01 2.96484590e-01 7.83456802e-01 6.45026505e-01
-9.80203629e-01 2.61311531e-02 2.70318072e-02 4.90677088e-01
3.08404535e-01 -1.02335691e+00 6.69086456e-01 1.00671637e+00
4.11902331e-02 3.45363587e-01 5.27771771e-01 -6.86544597e-01
-7.39090919e-01 -1.30018628e+00 4.78085041e-01 -1.51997700e-01
4.37439352e-01 -2.56329834e-01 -1.09682930e+00 7.37827063e-01
2.23282963e-01 -4.16520089e-02 7.44601965e-01 -2.80055732e-01
3.57799500e-01 9.79086906e-02 -1.06613994e+00 4.44318324e-01
8.13072085e-01 -2.69584209e-01 1.04755186e-01 6.03577077e-01
6.00181699e-01 -6.43288791e-01 -8.94966841e-01 3.15812200e-01
4.88585144e-01 -1.08616376e+00 7.95576811e-01 6.02849238e-02
3.43806177e-01 -4.70788211e-01 2.73602098e-01 -7.10789561e-01
-1.55996094e-02 -4.07384098e-01 3.85558695e-01 1.11378539e+00
4.28244472e-01 -5.24005055e-01 9.11287844e-01 8.35282803e-01
-3.78491402e-01 -7.79427648e-01 -4.69792962e-01 -5.66623807e-01
-4.02461439e-02 -5.36783934e-01 5.77580571e-01 9.37855482e-01
-4.15736288e-01 -1.28686473e-01 -1.22200258e-01 2.61531472e-01
5.05445957e-01 2.22998150e-02 6.22688234e-01 -1.23932886e+00
-1.41769215e-01 -3.67335051e-01 9.70519707e-02 -9.65878308e-01
-6.33263737e-02 -5.42346478e-01 2.52925754e-01 -2.04320526e+00
1.05434932e-01 -4.98074532e-01 5.14164940e-02 8.76050889e-01
-2.49435425e-01 6.86933160e-01 4.84736338e-02 2.05043480e-01
-5.53229451e-01 3.14917900e-02 1.36206555e+00 -1.08714931e-01
-2.89980292e-01 7.27610886e-02 -5.37665665e-01 9.92603302e-01
1.19462538e+00 -3.90983552e-01 -3.26459885e-01 -6.73328578e-01
-4.74061966e-02 -1.46026209e-01 4.36006039e-01 -1.00320399e+00
4.77757603e-02 -3.56045216e-01 3.35042685e-01 -6.42189443e-01
4.36276607e-02 -7.97155261e-01 4.13034439e-01 4.74524438e-01
-1.41248386e-02 -1.46166205e-01 3.62171531e-01 1.20471485e-01
-2.35011563e-01 -3.72237295e-01 9.63347137e-01 -6.45446658e-01
-8.62537980e-01 6.35439336e-01 -5.63444495e-01 -1.74754798e-01
1.10572529e+00 -4.85897481e-01 2.67262687e-03 -1.76401749e-01
-9.19408441e-01 1.87565893e-01 6.25001848e-01 1.10039100e-01
3.56375068e-01 -7.37436473e-01 -2.80325472e-01 1.63926005e-01
-2.10307524e-01 5.98751366e-01 5.24677992e-01 9.19917464e-01
-1.14055419e+00 4.06214356e-01 -1.40834719e-01 -5.59611499e-01
-1.31683815e+00 1.95534334e-01 3.94668490e-01 -1.69159085e-01
-7.45051205e-01 8.51715088e-01 1.41285151e-01 -5.54304898e-01
-9.44053568e-03 -4.09083188e-01 -4.84624580e-02 -1.55879825e-01
3.83429796e-01 5.61106741e-01 -7.40258908e-03 -5.09362578e-01
-2.32546031e-01 6.11723125e-01 2.30084360e-01 1.70251951e-02
1.31007671e+00 -2.54110783e-01 -1.91357613e-01 2.61998147e-01
6.86308742e-01 -3.35201979e-01 -1.46925128e+00 1.83325648e-01
4.93334373e-03 -3.79164755e-01 2.20117211e-01 -8.33483398e-01
-1.38764966e+00 1.11015630e+00 6.62496924e-01 3.43183666e-01
1.31329346e+00 -1.39303818e-01 1.03765583e+00 1.51910871e-01
3.52711052e-01 -1.18228042e+00 -2.39048213e-01 4.23034579e-01
3.58113199e-01 -1.42351234e+00 6.53061047e-02 -6.91039026e-01
-5.65010309e-01 9.86853421e-01 7.23612368e-01 2.64753759e-01
6.66750789e-01 3.05171847e-01 5.60388803e-01 -1.67798981e-01
-5.60570089e-03 -1.61148310e-01 -2.94436282e-03 7.37005472e-01
7.12113380e-01 -1.04632005e-02 -2.69696563e-01 3.80349398e-01
1.55101016e-01 2.27915227e-01 3.54998022e-01 9.85864758e-01
-5.48449814e-01 -1.20743155e+00 -3.29887956e-01 4.56104398e-01
-7.06425250e-01 -1.37616858e-01 -5.64519055e-02 6.80333674e-01
1.85120463e-01 8.80591214e-01 -9.30623896e-03 1.77655630e-02
6.28342852e-03 5.33317812e-02 1.04696117e-01 -7.18740940e-01
-4.73033726e-01 4.11869377e-01 -5.26230671e-02 -6.15290821e-01
-6.33877099e-01 -4.01502341e-01 -1.61835063e+00 -1.10021882e-01
-4.07477766e-01 5.67138083e-02 7.61052847e-01 8.79373729e-01
3.60562891e-01 5.34116268e-01 7.77079314e-02 -1.19845748e+00
1.37631744e-01 -1.00449717e+00 -5.98871946e-01 2.18201876e-02
-1.94458514e-01 -3.09916019e-01 -5.44087775e-02 6.37758017e-01] | [9.561454772949219, 0.016164978966116905] |
4347f32d-40b9-4f6e-b446-61c7d803de54 | processing-energy-modeling-for-neural-network | 2306.16755 | null | https://arxiv.org/abs/2306.16755v1 | https://arxiv.org/pdf/2306.16755v1.pdf | Processing Energy Modeling for Neural Network Based Image Compression | Nowadays, the compression performance of neural-networkbased image compression algorithms outperforms state-of-the-art compression approaches such as JPEG or HEIC-based image compression. Unfortunately, most neural-network based compression methods are executed on GPUs and consume a high amount of energy during execution. Therefore, this paper performs an in-depth analysis on the energy consumption of state-of-the-art neural-network based compression methods on a GPU and show that the energy consumption of compression networks can be estimated using the image size with mean estimation errors of less than 7%. Finally, using a correlation analysis, we find that the number of operations per pixel is the main driving force for energy consumption and deduce that the network layers up to the second downsampling step are consuming most energy. | ['André Kaup', 'Felix Rievel', 'Andy Regensky', 'Fabian Brand', 'Christian Herglotz'] | 2023-06-29 | null | null | null | null | ['image-compression'] | ['computer-vision'] | [ 4.41988975e-01 -1.50050253e-01 -4.02010471e-01 -2.41974235e-01
5.58024049e-02 2.12282002e-01 2.87874579e-01 3.15735847e-01
-9.58692729e-01 3.61231059e-01 -1.20825738e-01 -4.87198740e-01
6.14169799e-02 -1.12993741e+00 -9.41119313e-01 -6.32111073e-01
-6.09700717e-02 1.97034836e-01 1.65069312e-01 -7.35020638e-02
3.98365378e-01 4.92267102e-01 -1.74694109e+00 2.32943267e-01
3.30200970e-01 1.26688147e+00 1.81500584e-01 7.68538654e-01
5.23071513e-02 9.59185898e-01 -6.99893475e-01 -4.97513115e-01
1.44631267e-01 -5.37474215e-01 -3.69890124e-01 -5.57218015e-01
2.54125834e-01 -8.94842744e-01 -1.04476869e+00 1.16028154e+00
5.38694203e-01 -5.67462901e-03 4.03694302e-01 -8.37723374e-01
-7.74756521e-02 7.90715516e-01 -3.13139558e-01 3.25474292e-01
-2.85599917e-01 -1.14544369e-01 4.18977737e-01 -5.05267262e-01
5.73559642e-01 8.06132972e-01 8.52367401e-01 3.66748273e-01
-7.81134605e-01 -7.08140612e-01 -5.20577013e-01 7.67890036e-01
-1.44427168e+00 -5.44646204e-01 8.21312308e-01 2.65758723e-01
1.42721045e+00 3.30397785e-01 1.22449958e+00 7.59647012e-01
4.15448159e-01 4.57080781e-01 6.62252486e-01 -6.80427134e-01
5.13228774e-01 -3.13434243e-01 -1.34440213e-01 6.88611388e-01
7.43225336e-01 1.41434744e-01 -7.49183536e-01 1.30566627e-01
9.37858522e-01 5.38232848e-02 -2.17852175e-01 2.96915889e-01
-6.42442167e-01 9.55733180e-01 4.66882586e-01 5.80788136e-01
-5.64501584e-01 8.67959857e-01 8.36069763e-01 2.08707899e-01
3.28211993e-01 7.33314455e-02 -4.14943665e-01 -6.41505957e-01
-1.76482844e+00 3.98858488e-02 8.27773929e-01 8.44244421e-01
4.13335800e-01 3.83291274e-01 3.14844728e-01 7.24978089e-01
1.17296830e-01 2.29677543e-01 8.72881055e-01 -1.13974869e+00
2.56603599e-01 3.52760464e-01 -7.42813766e-01 -1.22129655e+00
-1.15955219e-01 -5.46447814e-01 -1.51271296e+00 1.57195896e-01
1.09774053e-01 2.13876497e-02 -7.12953150e-01 1.31742537e+00
-2.15050176e-01 -3.25132906e-02 3.90334576e-02 6.71741962e-01
6.18559957e-01 7.63140082e-01 -2.06095859e-01 -3.00385714e-01
1.29476357e+00 -1.17586184e+00 -7.87148893e-01 -2.27258891e-01
3.77270222e-01 -5.32738864e-01 5.10047376e-01 4.89441693e-01
-1.65887034e+00 -5.78935087e-01 -1.63199234e+00 -2.32773304e-01
-2.44432315e-01 1.62012681e-01 6.62225902e-01 7.86085904e-01
-1.03100479e+00 1.33228230e+00 -1.24196994e+00 -1.40269846e-01
6.19468629e-01 5.79143524e-01 9.33960974e-02 -7.11579621e-02
-8.10147285e-01 8.37501943e-01 8.25937510e-01 -1.12262219e-01
-7.43551016e-01 -6.75071180e-01 -4.32800919e-01 6.11949086e-01
-9.80975255e-02 -4.19475138e-01 1.12875879e+00 -9.54368174e-01
-1.45656979e+00 3.48343074e-01 -1.44029111e-01 -1.23204708e+00
3.69549096e-01 -1.70957968e-01 -1.81751266e-01 6.80047333e-01
-9.51661885e-01 7.74520218e-01 8.20226014e-01 -6.71408355e-01
-3.96632075e-01 -1.14178866e-01 -2.70116061e-01 -7.24116117e-02
-8.08267474e-01 -1.82247132e-01 -5.12255371e-01 -5.95118999e-01
6.56562522e-02 -7.05163479e-01 -1.82430193e-01 2.80068636e-01
-9.74789262e-02 -6.00284711e-02 1.04579020e+00 -7.25537062e-01
1.53816986e+00 -2.01882148e+00 -1.53088450e-01 1.62221342e-01
2.78866708e-01 5.17743170e-01 2.00936407e-01 2.63553679e-01
8.66166949e-02 7.88893998e-02 -1.60360709e-01 -2.46995986e-01
-2.83398181e-01 2.99314857e-01 -7.45183975e-02 4.50044841e-01
-3.83034319e-01 7.09844828e-01 -2.72221088e-01 -6.32321239e-01
3.15922141e-01 1.14313781e+00 -5.78276217e-01 1.77617997e-01
9.95128080e-02 -4.28743184e-01 7.97833037e-03 3.96904469e-01
7.11549461e-01 -1.65568471e-01 3.15290481e-01 -6.64140821e-01
-9.19980812e-04 3.95437598e-01 -4.92634982e-01 1.81276953e+00
-4.34222341e-01 1.23424327e+00 1.62507370e-02 -1.15796340e+00
6.11066163e-01 3.61909598e-01 4.65151906e-01 -7.81174779e-01
7.76979864e-01 4.11295235e-01 1.86218426e-01 -8.85967463e-02
8.58581364e-01 2.38741577e-01 4.78160262e-01 3.72640967e-01
2.20465109e-01 -2.09683906e-02 5.04162550e-01 -5.39760888e-02
1.09282088e+00 -2.01574653e-01 2.67790705e-01 -1.93786353e-01
1.70685649e-01 -5.22830784e-02 1.45747826e-01 6.12365961e-01
-5.73151857e-02 3.19446146e-01 2.05215529e-01 -5.81121564e-01
-1.63100529e+00 -3.54272097e-01 -1.50033176e-01 7.39199162e-01
-1.95742175e-01 -7.20696867e-01 -1.44503951e+00 1.27030984e-01
-5.81840336e-01 6.56930208e-01 -2.93293506e-01 -3.49191785e-01
-7.57676899e-01 -7.60220885e-01 1.04270768e+00 5.36770701e-01
1.07872260e+00 -9.07651603e-01 -1.81428647e+00 3.03867459e-01
-4.74474877e-02 -1.19428897e+00 4.36152853e-02 5.33389270e-01
-1.60005677e+00 -8.41464639e-01 -5.06813943e-01 -5.85434556e-01
6.91423535e-01 6.33287337e-03 1.20967197e+00 5.12477040e-01
-2.55768567e-01 -2.75221139e-01 -1.33282974e-01 -3.55761170e-01
-5.31689703e-01 3.19746733e-01 -2.57274240e-01 -8.04975212e-01
4.01757538e-01 -1.12040150e+00 -9.68387604e-01 -2.81864852e-01
-9.37608540e-01 4.83863145e-01 7.45868564e-01 5.47809958e-01
7.49152780e-01 7.95326948e-01 -3.24348837e-01 -5.93348444e-01
5.91134787e-01 1.23204859e-02 -6.60466611e-01 -1.56979442e-01
-1.03878939e+00 7.01418966e-02 9.23545003e-01 -5.38193583e-01
-5.47622859e-01 5.96182905e-02 -3.14199060e-01 -6.33619964e-01
5.66498972e-02 6.02521479e-01 4.49400902e-01 -2.45132729e-01
4.49406505e-01 5.49372137e-01 -6.92232628e-04 -3.14117432e-01
-6.45038337e-02 3.87622416e-01 8.38747978e-01 -5.52322939e-02
3.89648199e-01 3.42589706e-01 5.66325665e-01 -8.95028830e-01
-2.03436121e-01 1.80385187e-01 -2.45866016e-01 -3.45617175e-01
7.01804280e-01 -9.43785071e-01 -8.30602884e-01 5.30108392e-01
-1.32147288e+00 -2.50029266e-01 -3.50081354e-01 5.24884522e-01
-4.87914741e-01 1.55440763e-01 -1.11371887e+00 -7.54469514e-01
-9.75068867e-01 -1.10330307e+00 4.70459461e-01 2.86988378e-01
7.16520101e-02 -7.51016736e-01 -2.78591752e-01 -7.55005376e-03
8.54331374e-01 1.60903081e-01 9.42107975e-01 -3.36782992e-01
-5.31151116e-01 -3.29891264e-01 -3.10508430e-01 2.65149504e-01
-6.03831232e-01 1.31928436e-02 -7.59593487e-01 -3.50315362e-01
5.17011821e-01 -1.32540241e-01 9.95918870e-01 7.08709121e-01
1.71807516e+00 -6.42349243e-01 -1.11922279e-01 9.61381435e-01
2.05325556e+00 3.26788694e-01 1.04740429e+00 2.39939064e-01
4.39481586e-01 1.11473493e-01 7.74109154e-04 4.75236237e-01
-4.29123826e-02 4.51398462e-01 7.77572632e-01 1.07639328e-01
-1.22554645e-01 -1.90132037e-01 1.05193637e-01 1.51708472e+00
-5.01105130e-01 -4.30117995e-01 -5.85553050e-01 2.15920478e-01
-1.42005670e+00 -9.02923942e-01 -6.60590678e-02 2.22778320e+00
9.17480946e-01 2.88655490e-01 -2.51187772e-01 8.71943653e-01
4.82002020e-01 2.61473924e-01 -5.57043135e-01 -7.70286560e-01
3.19292396e-02 6.64177418e-01 1.19539893e+00 7.09832367e-03
-7.47955263e-01 5.32652020e-01 7.23599863e+00 1.23470342e+00
-1.25556529e+00 2.08603725e-01 7.17655957e-01 -4.91431415e-01
2.80026048e-01 -2.59780824e-01 -4.82616305e-01 5.06946683e-01
1.89732587e+00 -4.14673537e-02 7.13706255e-01 1.12865508e+00
1.22124357e-02 -4.13776517e-01 -9.98435378e-01 1.24261069e+00
4.34936471e-02 -1.57560730e+00 -3.18004042e-02 3.57389450e-01
4.62069094e-01 3.05500716e-01 -1.81370214e-01 -2.32284516e-01
-2.11511761e-01 -1.11311555e+00 7.09948659e-01 1.46927297e-01
1.07273972e+00 -1.19087970e+00 9.54198122e-01 3.42583239e-01
-1.16246140e+00 -4.79262620e-02 -9.64358747e-01 -1.44624576e-01
2.14718461e-01 1.00788164e+00 -4.24894504e-02 1.29612878e-01
8.12118232e-01 4.39965397e-01 -3.82663041e-01 7.83163846e-01
-5.66182658e-03 8.51370633e-01 -4.38861877e-01 -6.38752729e-02
9.64094847e-02 -1.11980967e-01 -2.17851121e-02 1.31988490e+00
7.87238002e-01 1.06489316e-01 -6.85935676e-01 5.17905653e-01
-4.70103800e-01 -1.68981984e-01 -2.83243477e-01 -2.17141569e-01
4.64966059e-01 8.68287265e-01 -8.10023189e-01 -6.27075553e-01
-2.35342979e-01 1.07720494e+00 -1.89458560e-02 -1.42900810e-01
-7.99875438e-01 -6.38666868e-01 3.81764859e-01 5.68723306e-02
5.11396229e-01 -3.67687792e-01 -4.06130344e-01 -7.86091268e-01
-6.92746490e-02 -6.24253809e-01 -2.72575706e-01 -7.29191303e-01
-1.87640801e-01 5.75881720e-01 -1.11937039e-01 -9.71828222e-01
-3.06795239e-01 -5.25995433e-01 -3.07789981e-01 3.66701931e-01
-1.51663589e+00 -4.58407521e-01 -4.60941911e-01 4.08199579e-01
5.22787154e-01 -2.68782467e-01 9.98239100e-01 6.54801548e-01
-5.31624973e-01 7.07663894e-01 2.09104672e-01 1.58565387e-01
-2.48540476e-01 -2.29729161e-01 4.15686637e-01 8.76454115e-01
8.45294744e-02 4.15732712e-01 9.17855918e-01 -3.34384888e-01
-1.89303899e+00 -9.38672721e-01 8.05638731e-01 7.22316265e-01
1.59924820e-01 -6.55967817e-02 -7.37413228e-01 1.46586746e-01
5.89242637e-01 8.68619978e-02 4.58103299e-01 -5.95065773e-01
-2.67768860e-01 -1.04682945e-01 -1.19378197e+00 3.58490050e-01
9.23435152e-01 -5.79404414e-01 9.40628797e-06 2.35065937e-01
4.71274555e-01 -4.54007655e-01 -9.66582298e-01 3.09347659e-01
8.62657666e-01 -1.37568676e+00 1.01009178e+00 1.27810165e-01
1.23923349e+00 2.61101186e-01 -3.32021207e-01 -7.57851779e-01
-1.19385896e-02 -1.15956046e-01 -8.39528799e-01 6.36196613e-01
-4.98805866e-02 -1.60262674e-01 1.18383586e+00 1.93896189e-01
1.22031324e-01 -1.01557851e+00 -1.19390213e+00 -7.19197989e-01
-2.21038580e-01 -5.79763949e-01 4.29940701e-01 5.30480146e-01
-1.52749613e-01 -7.10346401e-02 -3.55733424e-01 -5.14633656e-01
6.14277065e-01 -4.01644856e-01 2.82511503e-01 -1.18304968e+00
-3.36674154e-01 -6.43828571e-01 -6.38432026e-01 -9.45018470e-01
-1.73853934e-01 -4.20411319e-01 -1.32310092e-01 -1.26669431e+00
2.14637369e-01 1.24132134e-01 -8.79951417e-02 1.59842730e-01
6.22202992e-01 6.12879217e-01 2.57653922e-01 4.29983616e-01
-1.35237649e-01 4.35982853e-01 7.94979513e-01 -1.42168611e-01
1.32003739e-01 -5.71386874e-01 -2.30833709e-01 9.70147192e-01
9.30008769e-01 -6.41920388e-01 -4.98152643e-01 -6.85397863e-01
3.20359141e-01 1.75514162e-01 7.78560191e-02 -1.73968506e+00
6.20171726e-01 1.90560207e-01 3.93557936e-01 -7.76219964e-01
5.75967073e-01 -1.25122166e+00 3.31001282e-01 1.18498397e+00
-3.22022438e-01 3.17146838e-01 1.02965854e-01 1.64188325e-01
-4.19322073e-01 -6.01428807e-01 8.66154134e-01 -1.61456704e-01
-2.29299635e-01 1.53374881e-01 -6.06702328e-01 -3.29940945e-01
6.41019523e-01 -5.09635985e-01 -3.05383742e-01 -3.88255566e-01
-9.18390825e-02 -8.11031818e-01 3.80885035e-01 -2.14384019e-01
7.59562254e-01 -1.03059018e+00 -4.72659379e-01 2.50886172e-01
-6.31382465e-01 -9.01306272e-02 1.68320566e-01 6.27206266e-01
-1.35742974e+00 6.79177344e-01 -4.40230310e-01 -3.40350181e-01
-1.53706467e+00 2.80957282e-01 2.60129958e-01 -4.01148230e-01
-6.72202826e-01 7.82085419e-01 -7.70123661e-01 5.72541833e-01
4.37478930e-01 -2.70896912e-01 7.75791192e-03 -2.65555233e-01
5.73568583e-01 9.13195908e-01 2.23137856e-01 -5.19993603e-01
-6.54424727e-02 5.50225437e-01 1.96103215e-01 -7.16798753e-02
1.40865529e+00 2.73842812e-01 -3.69749665e-01 4.65197936e-02
1.51481950e+00 -6.81250632e-01 -9.55063939e-01 1.16893291e-01
-5.07840157e-01 -2.74324596e-01 7.10151613e-01 -3.37325275e-01
-1.83446360e+00 1.03228259e+00 9.99559104e-01 8.65322128e-02
1.85004067e+00 -5.97398162e-01 1.13287950e+00 6.69176519e-01
3.60004753e-01 -1.58856463e+00 -3.49063456e-01 3.34760368e-01
3.46177429e-01 -5.69367707e-01 6.53048098e-01 -3.55118304e-01
2.30472639e-01 1.42657959e+00 2.57046789e-01 -4.44414914e-01
8.10862601e-01 8.18547547e-01 -5.48077643e-01 5.09388149e-02
-8.31294417e-01 5.69750667e-01 -1.27994880e-01 2.57397413e-01
4.98510778e-01 -5.82681634e-02 -7.82507539e-01 6.54909238e-02
-7.64217436e-01 4.07305837e-01 5.30310214e-01 1.07046783e+00
-4.43634033e-01 -9.20582891e-01 -6.12691604e-02 7.35527217e-01
-8.69343877e-01 -4.10658926e-01 1.96831331e-01 5.60965955e-01
8.11445341e-02 8.20722938e-01 4.87747878e-01 -6.26187801e-01
-1.12495951e-01 -1.38610706e-01 6.43128812e-01 2.47703582e-01
-8.87435615e-01 -9.36142430e-02 -6.17984086e-02 -7.32477725e-01
-6.65194392e-01 -1.19655199e-01 -1.19390309e+00 -1.04038739e+00
-2.38900065e-01 -5.59865423e-02 1.59257972e+00 7.15489149e-01
2.66019106e-01 8.56260598e-01 1.82037756e-01 -1.08526540e+00
-1.79465666e-01 -9.59207773e-01 -4.42465633e-01 -4.61049937e-02
-4.30926196e-02 1.15343846e-01 -4.08758968e-01 1.42317891e-01] | [8.474725723266602, 2.9804797172546387] |
a75e15ad-711d-436c-9d10-99775900ef32 | radio-slam-for-6g-systems-at-thz-frequencies | 2212.12388 | null | https://arxiv.org/abs/2212.12388v1 | https://arxiv.org/pdf/2212.12388v1.pdf | Radio SLAM for 6G Systems at THz Frequencies: Design and Experimental Validation | Next-generation wireless networks will see the convergence of communication and sensing, also exploiting the availability of large bandwidths in the Terahertz (THz) spectrum and electrically large antenna arrays on handheld devices. In particular, it is envisaged that user devices will be able to automatically scan their surroundings by steering a very narrow antenna beam and collecting echoes reflected by objects and walls to derive a map of indoors and infer users' trajectories using simultaneous localization and mapping (SLAM) techniques. In this paper, we address this scenario by proposing original radioSLAM (R-SLAM) algorithms, derived from image processing techniques, to map the environment and pinpoint the device position in the map starting from measurements sensed by a mobile THz radar. Initially, to fully understand the THz backscattering phenomenon, we provide an experimental characterization of the THz backscattering channel in indoor environments. Then, the performance of the proposed algorithms is assessed using real-world THz radar measurements and is compared with state-of-the-art SLAM techniques, demonstrating the superiority of the proposed approaches. | ['Davide Dardari', "Raffaele D'Errico", 'Francesco Guidi', 'Anna Guerra', 'Gianni Pasolini', 'Marina Lotti'] | 2022-12-23 | null | null | null | null | ['simultaneous-localization-and-mapping'] | ['computer-vision'] | [ 6.62795722e-01 -1.28002539e-02 3.42790931e-01 -6.30873799e-01
-4.56244409e-01 -5.11961162e-01 4.72784281e-01 -2.00589955e-01
-4.44277763e-01 8.69300902e-01 -3.26123297e-01 -3.67578268e-01
-4.90596801e-01 -1.26281238e+00 -4.07003045e-01 -8.64686251e-01
-4.42720920e-01 1.05475473e+00 1.53593659e-01 -1.44288223e-02
2.23718777e-01 6.82823718e-01 -1.45090294e+00 -2.35953420e-01
9.66551006e-01 1.21305525e+00 7.28026390e-01 4.54972982e-01
-7.29620010e-02 -1.45654842e-01 -3.67578566e-01 1.38685688e-01
8.15369636e-02 6.58088177e-02 -1.77838765e-02 -5.68903208e-01
2.69810498e-01 -4.02800262e-01 -1.67539582e-01 9.45237458e-01
2.35386133e-01 -2.76594192e-01 5.21778822e-01 -7.65050650e-01
5.17744869e-02 3.57775986e-01 -2.31200576e-01 -3.53959203e-01
5.66986024e-01 -4.84006017e-01 3.13341171e-01 -6.13173902e-01
9.52283740e-01 5.19301057e-01 7.45767653e-01 -3.52577902e-02
-6.16268218e-01 -9.83561516e-01 -5.01637638e-01 3.24487686e-02
-1.57857335e+00 -4.50825959e-01 6.74109101e-01 -1.22944862e-01
4.64296877e-01 5.03280759e-01 7.42614090e-01 1.18155801e+00
4.82881695e-01 -1.65795594e-01 1.21745074e+00 -7.49254286e-01
5.92098713e-01 2.71172583e-01 -1.89239770e-01 9.02069569e-01
8.34367812e-01 3.14406514e-01 -7.31654286e-01 -1.14293940e-01
4.76480335e-01 -6.72288612e-02 -4.35154825e-01 -6.33625686e-01
-1.21087396e+00 2.93144017e-01 7.05332637e-01 8.41099322e-01
-7.96533704e-01 2.56551445e-01 -5.96550286e-01 7.11760297e-02
3.66047233e-01 5.17019153e-01 9.81249139e-02 1.12849176e-02
-1.19340980e+00 4.29707430e-02 6.84925914e-01 1.15085852e+00
8.92970026e-01 -3.87772948e-01 1.60010591e-01 1.10289946e-01
6.16657495e-01 1.27897942e+00 1.14399239e-01 -5.70485890e-01
4.68546271e-01 -1.72132868e-02 6.91735089e-01 -8.39693666e-01
-5.01225710e-01 -8.67466331e-01 -7.53312111e-01 -2.14930966e-01
2.82732546e-01 -2.45072648e-01 -9.24340487e-01 1.46395087e+00
3.81844908e-01 4.95808661e-01 1.25811234e-01 6.40017569e-01
5.36973298e-01 5.07467985e-01 -3.58155131e-01 -4.29880619e-01
1.26931953e+00 -1.06595762e-01 -6.69698060e-01 -3.17075461e-01
3.74451369e-01 -4.56988454e-01 2.79332846e-01 5.55273294e-01
-5.46087682e-01 -1.56168520e-01 -1.25434482e+00 5.51584244e-01
-4.65246886e-01 2.04485655e-01 8.37339580e-01 1.24261975e+00
-7.97790647e-01 3.83209080e-01 -9.90994453e-01 -8.28051209e-01
3.29167455e-01 3.69126499e-01 2.04635710e-01 -3.85078162e-01
-1.01600111e+00 4.31431055e-01 -1.77387968e-01 2.77960867e-01
-2.76233613e-01 -9.35392320e-01 -2.33686075e-01 -2.96566963e-01
-2.00925767e-01 -7.36998439e-01 8.46650839e-01 -2.33197995e-02
-1.49387705e+00 4.09986854e-01 -4.23119843e-01 -5.53406656e-01
4.62215036e-01 -3.14841568e-02 -6.94617987e-01 2.74720222e-01
2.17317596e-01 -4.34824713e-02 4.75589871e-01 -1.29699433e+00
-4.41612303e-01 -8.90924275e-01 -6.97086990e-01 -2.26885974e-01
-2.45788440e-01 -5.80734730e-01 1.61161691e-01 4.17908370e-01
1.08147264e+00 -1.02175307e+00 -1.98778495e-01 -3.15526456e-01
-4.86629814e-01 5.43231726e-01 1.18623018e+00 -7.27218017e-02
5.67803502e-01 -1.86758077e+00 -5.53470373e-01 7.94983208e-01
6.38171285e-02 -1.41728669e-01 1.24622129e-01 8.31744969e-01
5.51210046e-01 -2.83564895e-01 -4.41768356e-02 -2.54100144e-01
-2.49836147e-01 -1.39469564e-01 -4.45695132e-01 6.69864118e-01
-9.78273153e-01 6.95595682e-01 -9.91631746e-01 1.11775942e-01
2.53686756e-01 3.99601638e-01 1.14092976e-02 -1.11015022e-01
-4.39380296e-02 8.35859358e-01 -1.17551041e+00 8.62513661e-01
1.39674461e+00 1.22218737e-02 4.08501804e-01 -2.71481305e-01
-9.35372591e-01 1.83882251e-01 -8.63460660e-01 1.85757911e+00
-6.31201029e-01 5.87960780e-01 3.45520675e-01 -7.94523895e-01
1.22279263e+00 9.48607251e-02 7.18722701e-01 -1.18101108e+00
-1.54039031e-02 6.62057340e-01 -5.10688424e-01 -5.69276094e-01
5.57543039e-01 -1.69443920e-01 -6.21601976e-02 5.44585466e-01
-1.78976625e-01 -3.86538804e-01 -5.55556536e-01 -2.23296490e-02
1.51520014e+00 5.79831712e-02 1.93012238e-01 -4.25516158e-01
2.31836647e-01 6.12032861e-02 -2.71186769e-01 1.15546596e+00
2.91453689e-01 1.57220252e-02 -4.58257407e-01 -3.31983924e-01
-6.59602940e-01 -1.40273106e+00 -4.86989856e-01 1.76201254e-01
5.66709876e-01 -1.78223640e-01 -4.70635384e-01 -1.37193590e-01
3.54982652e-02 7.72197902e-01 -2.42790535e-01 3.82616162e-01
-2.60772288e-01 -9.22581732e-01 2.65836686e-01 -3.21924627e-01
9.84469056e-01 -3.71797800e-01 -1.11688614e+00 2.36043990e-01
-2.77270496e-01 -1.59032488e+00 7.24903166e-01 1.38458416e-01
-6.42980039e-01 -6.30629778e-01 -3.75270605e-01 -2.68014222e-01
5.78627646e-01 7.67074466e-01 6.57132566e-01 -3.68163407e-01
-6.04177237e-01 7.34886527e-01 -4.54751134e-01 -6.77465439e-01
9.30859819e-02 7.45456740e-02 3.94531816e-01 -5.41042313e-02
5.23614958e-02 -1.16897213e+00 -7.33761728e-01 3.20503563e-01
-6.21345416e-02 1.22925051e-01 7.46667445e-01 -1.50305882e-01
3.27988684e-01 2.52373219e-01 3.90466034e-01 -8.56830299e-01
1.09696344e-01 -5.35671890e-01 -1.09591103e+00 1.83920935e-01
-5.76994717e-01 -2.03536958e-01 2.65435666e-01 1.53517067e-01
-9.82063353e-01 1.39054269e-01 -7.78393000e-02 4.75203902e-01
-2.18072698e-01 2.62654275e-01 -1.96784541e-01 -9.89596605e-01
5.28333902e-01 4.10081536e-01 -6.44688010e-01 -2.89037488e-02
2.52855331e-01 8.50180745e-01 3.56653094e-01 -3.58797193e-01
1.27598274e+00 1.07080007e+00 4.42835271e-01 -1.66115773e+00
-6.27721667e-01 -4.87648070e-01 -5.43539286e-01 -5.34636259e-01
6.83902264e-01 -7.14966953e-01 -8.92789006e-01 6.34665489e-02
-1.05802512e+00 -7.73663819e-02 1.37827784e-01 8.81890535e-01
-3.34375024e-01 1.25415817e-01 3.12217563e-01 -1.28163469e+00
-3.12026560e-01 -4.68975306e-01 1.14105082e+00 2.15343013e-01
2.03508092e-03 -6.70301497e-01 3.19681108e-01 2.61638671e-01
9.27380979e-01 3.36659849e-01 7.09322989e-01 -7.33369738e-02
-1.26014709e+00 -4.42562938e-01 -2.26359829e-01 -8.98371398e-01
3.10739446e-02 -7.67934740e-01 -1.18245316e+00 -1.69130743e-01
2.27714613e-01 3.04594547e-01 5.21707535e-01 6.57472610e-01
9.76543367e-01 2.19756871e-01 -1.22877610e+00 9.75235283e-01
1.72995412e+00 2.06145972e-01 8.00497115e-01 1.67026684e-01
2.30178028e-01 3.50813448e-01 7.64588118e-01 4.88126338e-01
2.79180616e-01 7.26999044e-01 8.14676821e-01 2.73405880e-01
2.60135770e-01 -1.38457911e-02 -1.44825235e-01 2.43007720e-01
-3.29143554e-01 -6.36677921e-01 -7.09987521e-01 4.33411188e-02
-1.52934647e+00 -8.07233632e-01 -5.57254791e-01 2.59249663e+00
-4.16728646e-01 -1.70325235e-01 -6.20684683e-01 -3.00779164e-01
3.67965758e-01 2.41698653e-01 -1.20884739e-01 3.99995536e-01
2.23622978e-01 6.56649888e-01 1.03967416e+00 7.61912346e-01
-6.09379649e-01 6.90055013e-01 5.84584427e+00 1.47177175e-01
-1.22071290e+00 2.06248716e-01 -1.58761114e-01 7.31217414e-02
-6.89544141e-01 4.99798656e-02 -6.53733194e-01 2.42469996e-01
1.02500975e+00 3.61699224e-01 4.44865942e-01 2.39385515e-01
4.68528807e-01 -6.18386924e-01 -5.83067775e-01 1.00184703e+00
-2.96573251e-01 -1.34573555e+00 -3.43432844e-01 5.77529788e-01
5.06165981e-01 2.73402691e-01 4.18746797e-03 -7.36200511e-02
-1.33269399e-01 -5.24986327e-01 4.38650489e-01 7.62753546e-01
7.05257893e-01 -4.87895280e-01 4.20302927e-01 6.83616281e-01
-1.24052036e+00 4.96629402e-02 -2.88530767e-01 5.46947867e-02
2.50419050e-01 1.31352568e+00 -1.55889809e+00 8.72532547e-01
6.29612803e-01 1.84935480e-01 -9.34428070e-03 8.00400138e-01
-1.14876159e-01 5.03327250e-01 -7.01246798e-01 -4.86363709e-01
1.05534703e-01 -7.32373595e-01 5.70440531e-01 1.00601292e+00
1.54420745e+00 2.71119267e-01 -3.23935717e-01 1.14715946e+00
-6.54434040e-02 -4.18840498e-01 -8.95971537e-01 3.14533040e-02
8.03798437e-01 1.52421832e+00 -8.07556629e-01 1.46171018e-01
1.39771283e-01 9.48568225e-01 -4.42752182e-01 6.57191515e-01
-6.35812581e-01 -4.78660911e-01 4.79937047e-01 5.75967312e-01
2.07198173e-01 -1.02706587e+00 -1.46180734e-01 -7.88911760e-01
3.48612554e-02 3.38239193e-01 -4.80254203e-01 -1.00800502e+00
-4.68085408e-01 2.92655587e-01 -2.66566843e-01 -1.02644122e+00
-1.50016412e-01 -3.12843591e-01 -4.17712629e-01 7.32460141e-01
-1.72124112e+00 -1.20819366e+00 -1.08802485e+00 3.49785209e-01
-2.02898696e-01 2.79134005e-01 1.08376920e+00 1.96872845e-01
2.68836439e-01 -1.02968767e-01 8.07727993e-01 -4.47949857e-01
5.02002001e-01 -8.01098466e-01 3.84727329e-01 5.91500640e-01
2.82994151e-01 6.94809318e-01 9.72367704e-01 -9.84745383e-01
-1.99657726e+00 -1.10181713e+00 7.71010399e-01 -1.79483101e-01
4.74469036e-01 -1.13957918e+00 -1.47024214e-01 3.84721100e-01
-1.30564809e-01 1.77151747e-02 4.65053290e-01 1.68341354e-01
8.24988633e-02 -4.72122401e-01 -1.68242931e+00 2.40864515e-01
1.46483064e+00 -2.91468322e-01 2.75902480e-01 7.63425708e-01
1.41629100e-01 -4.15132850e-01 -4.19079393e-01 4.54784721e-01
1.08301318e+00 -1.21634054e+00 1.17108214e+00 3.60473067e-01
-4.38665152e-01 7.28252484e-03 -5.93003809e-01 -1.17202771e+00
-6.21695966e-02 -6.87401474e-01 2.96350330e-01 1.06879282e+00
1.75378397e-01 -1.10767543e+00 1.34517205e+00 -1.19117886e-01
3.96605209e-02 -3.33706662e-02 -1.34734690e+00 -8.15626740e-01
-7.36788154e-01 -7.15789258e-01 6.35217845e-01 7.08161294e-01
-2.10430607e-01 2.39557736e-02 -1.97745711e-01 1.15651309e+00
1.41145277e+00 2.92325228e-01 6.38785243e-01 -1.67448044e+00
9.55337882e-02 4.27612931e-01 -4.25846100e-01 -9.43840027e-01
-1.56328842e-01 -7.20246851e-01 2.87416965e-01 -1.63625515e+00
-3.27268898e-01 -1.16651773e+00 1.11662865e-01 -3.20657194e-01
9.70886648e-01 3.99623215e-01 -3.67290825e-01 1.26958519e-01
-4.49851722e-01 3.84875029e-01 7.30812848e-01 1.54200643e-01
-2.83333212e-01 5.07449627e-01 -2.27625266e-01 4.50913727e-01
5.14183939e-01 -4.67253178e-01 -4.03021932e-01 -4.04980719e-01
3.20110768e-01 3.35990340e-01 6.21933520e-01 -1.84702599e+00
5.48986256e-01 -1.85790414e-03 5.50393343e-01 -7.79003322e-01
8.87164116e-01 -1.04733622e+00 4.79600489e-01 6.81800604e-01
3.44556391e-01 -8.82927775e-01 -3.13614428e-01 7.78137565e-01
4.83306199e-01 -2.52417237e-01 4.77534086e-01 -5.65664992e-02
-4.18829054e-01 3.17843109e-01 -4.98456776e-01 -7.76508451e-01
9.11750317e-01 -2.19602957e-01 -4.74954516e-01 -5.06590962e-01
-2.12508947e-01 -3.20939034e-01 3.35237861e-01 -1.96981445e-01
6.92348540e-01 -1.00956798e+00 -3.20228994e-01 4.08086717e-01
1.57709762e-01 -6.00048661e-01 3.40906680e-01 7.51852036e-01
-5.24514914e-01 7.80754387e-01 -5.96757978e-03 -6.91789567e-01
-1.24276590e+00 -1.60351321e-01 2.94388503e-01 3.86386067e-02
-4.93439078e-01 5.80016851e-01 -3.03568780e-01 -4.42553729e-01
-2.51762003e-01 -5.41752815e-01 3.26397389e-01 -3.74277055e-01
3.47249925e-01 1.45154715e-01 5.15178144e-01 -1.30474359e-01
-7.37909973e-01 1.12886953e+00 4.18509424e-01 -4.52667713e-01
1.57738853e+00 -3.05106491e-01 -4.37985174e-02 4.67773154e-02
7.40411997e-01 5.92703044e-01 -3.29649210e-01 -1.24606967e-01
-1.48279369e-01 -4.80572104e-01 1.92494214e-01 -8.22507918e-01
-3.37147832e-01 6.74397469e-01 1.16941166e+00 4.78006840e-01
9.67235446e-01 1.23669719e-02 6.11540496e-01 9.56946015e-01
1.55722976e+00 -4.95648235e-01 -4.70896751e-01 1.79608792e-01
2.36923158e-01 -8.49649012e-01 3.63507539e-01 -7.68601239e-01
3.69639874e-01 1.17792296e+00 -1.93069369e-01 3.05977575e-02
8.15032542e-01 5.53698063e-01 -2.60403842e-01 -3.62516761e-01
6.72423691e-02 1.36844907e-02 -3.24181676e-01 9.54623938e-01
1.76707342e-01 2.34852955e-01 1.08680777e-01 1.85164884e-01
-4.64281410e-01 7.06600696e-02 5.00399113e-01 9.44759429e-01
-1.12664127e+00 -1.14595246e+00 -6.02740765e-01 4.31713849e-01
2.78932005e-01 7.62390867e-02 -2.27842569e-01 8.50828886e-01
1.80186599e-01 9.51212049e-01 -2.14493811e-01 -2.79614776e-01
2.00929046e-01 -5.73049784e-02 9.75074530e-01 -2.97447354e-01
2.63511837e-01 -2.39941373e-01 4.98346463e-02 -6.46411598e-01
-3.18517834e-01 -6.18135810e-01 -1.20275044e+00 1.86089680e-01
-2.13858694e-01 3.94375265e-01 1.69225883e+00 1.06381559e+00
4.41305429e-01 2.37805620e-01 5.98828077e-01 -1.00773025e+00
-8.58966485e-02 -6.80809736e-01 -1.00608897e+00 -6.65471673e-01
2.78252244e-01 -5.72227359e-01 -1.79192707e-01 -7.87311018e-01] | [6.289161205291748, 1.0856685638427734] |
6c99f3bc-7736-453e-8e18-464000942f2c | pre-trained-language-models-for-keyphrase | 2212.10233 | null | https://arxiv.org/abs/2212.10233v1 | https://arxiv.org/pdf/2212.10233v1.pdf | Pre-trained Language Models for Keyphrase Generation: A Thorough Empirical Study | Neural models that do not rely on pre-training have excelled in the keyphrase generation task with large annotated datasets. Meanwhile, new approaches have incorporated pre-trained language models (PLMs) for their data efficiency. However, there lacks a systematic study of how the two types of approaches compare and how different design choices can affect the performance of PLM-based models. To fill in this knowledge gap and facilitate a more informed use of PLMs for keyphrase extraction and keyphrase generation, we present an in-depth empirical study. Formulating keyphrase extraction as sequence labeling and keyphrase generation as sequence-to-sequence generation, we perform extensive experiments in three domains. After showing that PLMs have competitive high-resource performance and state-of-the-art low-resource performance, we investigate important design choices including in-domain PLMs, PLMs with different pre-training objectives, using PLMs with a parameter budget, and different formulations for present keyphrases. Further results show that (1) in-domain BERT-like PLMs can be used to build strong and data-efficient keyphrase generation models; (2) with a fixed parameter budget, prioritizing model depth over width and allocating more layers in the encoder leads to better encoder-decoder models; and (3) introducing four in-domain PLMs, we achieve a competitive performance in the news domain and the state-of-the-art performance in the scientific domain. | ['Kai-Wei Chang', 'Wasi Uddin Ahmad', 'Di wu'] | 2022-12-20 | null | null | null | null | ['keyphrase-generation', 'keyphrase-extraction'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.72223085e-01 6.71735592e-03 -5.43629229e-01 1.15824617e-01
-1.02474272e+00 -6.75485134e-01 1.04420626e+00 2.75730699e-01
-8.17374825e-01 9.27577853e-01 5.98208010e-01 -5.55252850e-01
3.50150727e-02 -8.78259897e-01 -8.04627657e-01 -2.75184184e-01
7.93372467e-02 3.99954826e-01 2.48268723e-01 -3.41058701e-01
4.48836982e-01 4.10584182e-01 -1.31286907e+00 4.80627447e-01
7.21786618e-01 7.86284029e-01 5.75612724e-01 8.57120037e-01
-4.52380508e-01 9.25515234e-01 -8.28983009e-01 -4.74288553e-01
2.99542844e-01 -4.93619949e-01 -9.40606594e-01 -2.08414435e-01
3.02209169e-01 -3.66291344e-01 -3.80472034e-01 7.64064610e-01
7.87097812e-01 -9.10164937e-02 6.75188243e-01 -1.05766678e+00
-7.41038799e-01 1.11233091e+00 -3.47654194e-01 2.39537314e-01
3.44933748e-01 8.23973641e-02 1.27063537e+00 -8.20381284e-01
8.23551178e-01 9.59114730e-01 3.67681593e-01 5.73079169e-01
-1.19704378e+00 -7.69088864e-01 2.33581178e-02 6.90688193e-02
-1.39192152e+00 -5.01882792e-01 4.22910035e-01 -1.62235856e-01
1.25963497e+00 1.61864281e-01 5.62027395e-01 1.19139349e+00
3.67892444e-01 1.21895182e+00 9.42564905e-01 -5.70007622e-01
-1.14166643e-02 1.48244694e-01 -1.16982110e-01 5.59878230e-01
4.25942749e-01 5.72063066e-02 -7.98752844e-01 -1.74554393e-01
7.56150007e-01 -4.37365323e-01 -2.90635794e-01 3.01566541e-01
-1.56461775e+00 8.56542468e-01 -1.57001644e-01 4.72895265e-01
-5.00392854e-01 1.10690825e-01 5.74601054e-01 4.33483690e-01
5.01962364e-01 1.19603562e+00 -7.25356758e-01 -3.13742667e-01
-1.62602854e+00 6.27151668e-01 9.78463650e-01 1.20410812e+00
5.13166428e-01 1.71198517e-01 -5.69077551e-01 8.43915403e-01
-1.25580102e-01 3.57918561e-01 7.40997732e-01 -5.00129521e-01
7.30030000e-01 4.38546747e-01 7.90567137e-03 -6.76089406e-01
-3.39316249e-01 -3.72751445e-01 -8.97691786e-01 -5.00322938e-01
1.75810546e-01 -3.93688560e-01 -1.13358402e+00 1.63232732e+00
-2.28358015e-01 7.77219608e-02 1.76648468e-01 5.19321978e-01
9.58973944e-01 1.12086618e+00 8.73548239e-02 -1.70770541e-01
1.71244705e+00 -9.06103134e-01 -6.85775280e-01 -4.50926691e-01
7.75731623e-01 -1.01518214e+00 9.46969986e-01 3.99690360e-01
-1.19965589e+00 -5.83373606e-01 -1.04506326e+00 -3.56619656e-01
-5.37696600e-01 4.36702609e-01 6.76718950e-01 4.29402053e-01
-1.05956721e+00 5.74226856e-01 -5.59508145e-01 -2.80609071e-01
1.98274434e-01 7.87581597e-03 -3.01425576e-01 -1.20005421e-02
-1.57335103e+00 1.04155624e+00 9.45217669e-01 -3.97586435e-01
-1.04435480e+00 -1.02005208e+00 -7.29241073e-01 -1.07893515e-02
5.58046281e-01 -8.52849245e-01 1.31006789e+00 -4.20089424e-01
-1.48756301e+00 7.79966652e-01 -5.95450662e-02 -7.50138283e-01
2.03681186e-01 -3.76741886e-01 -5.58928668e-01 1.80138171e-01
3.43186669e-02 9.47916746e-01 7.38558531e-01 -8.48301828e-01
-9.45822656e-01 2.53568918e-01 1.69275571e-02 1.74436435e-01
-5.11581123e-01 2.75030375e-01 -6.22696996e-01 -1.14914310e+00
-3.78084153e-01 -8.40793312e-01 -3.22108746e-01 -4.67701584e-01
-5.59280515e-01 -1.36752129e-01 3.82854134e-01 -7.86133707e-01
1.79190874e+00 -1.70012379e+00 5.01677319e-02 -1.67665973e-01
7.01429546e-02 4.90906537e-01 -4.63506788e-01 8.49590659e-01
5.46926074e-02 4.19240952e-01 -4.10163552e-02 -1.52717605e-01
1.08951062e-01 1.09958917e-01 -7.37544596e-01 -1.39045820e-01
3.87697637e-01 1.09562874e+00 -8.45322728e-01 -6.22677922e-01
-1.96563423e-01 2.10373491e-01 -5.68586409e-01 4.24475104e-01
-5.51150739e-01 1.40609033e-02 -4.37766403e-01 4.08055872e-01
2.74272144e-01 -3.66229951e-01 -2.55259108e-02 -1.56243682e-01
-1.54752597e-01 7.54747033e-01 -1.11919332e+00 1.64190578e+00
-6.27947092e-01 7.39040673e-01 -9.85457525e-02 -6.76554918e-01
8.00309300e-01 5.59007704e-01 4.13524419e-01 -7.11035252e-01
-1.71514124e-01 3.40656102e-01 -8.77759233e-02 -6.22214712e-02
1.33587468e+00 1.25454934e-02 -9.42886025e-02 6.85966313e-01
4.35211778e-01 -3.59360486e-01 7.43631303e-01 5.41893125e-01
1.09745657e+00 1.39188152e-02 4.26987469e-01 -3.13665152e-01
3.75248730e-01 -2.22150749e-03 1.56075284e-01 8.96420896e-01
4.73288745e-01 5.09221494e-01 4.85474020e-01 -2.08794430e-01
-1.26968586e+00 -4.97655571e-01 3.73487920e-02 1.18306339e+00
-2.50737309e-01 -1.01563072e+00 -5.95330000e-01 -5.48301578e-01
-1.66246876e-01 9.93560016e-01 -2.66242594e-01 -1.28755182e-01
-7.36544192e-01 -1.08045781e+00 1.12490475e+00 4.06627089e-01
5.01260817e-01 -1.13202786e+00 -5.12662113e-01 4.56120878e-01
-4.76121962e-01 -1.32660568e+00 -5.77369034e-01 4.05033588e-01
-7.32394576e-01 -5.67268193e-01 -1.23346376e+00 -6.16670549e-01
5.48754275e-01 6.68843985e-02 1.50761819e+00 -1.49728462e-01
-6.05118908e-02 2.00566083e-01 -5.83473623e-01 -6.07043684e-01
-7.40176201e-01 6.98984563e-01 -1.19244814e-01 -4.93597865e-01
2.16530874e-01 -3.56895506e-01 -3.91127348e-01 3.46523477e-03
-1.27243865e+00 4.72537488e-01 9.97677028e-01 7.34551668e-01
4.14438307e-01 1.06918633e-01 6.01963341e-01 -8.59492242e-01
1.12209654e+00 -3.74367386e-01 -6.17082596e-01 3.30710620e-01
-9.84304130e-01 5.13428867e-01 7.49053836e-01 -5.79245567e-01
-8.99095893e-01 -3.80005658e-01 -2.49949545e-01 5.84943704e-02
1.58819169e-01 7.66680241e-01 3.84487659e-02 3.34425300e-01
7.80426800e-01 7.70194948e-01 -3.22872102e-01 -6.35673940e-01
6.65245771e-01 5.46179116e-01 2.01587051e-01 -8.85331690e-01
7.28224516e-01 1.61029410e-03 -1.71199083e-01 -8.64720702e-01
-7.66581774e-01 -3.11745405e-01 -3.83607298e-01 2.71816432e-01
6.33258462e-01 -1.23810410e+00 -1.75119713e-01 3.72478157e-01
-1.21582580e+00 -4.32624042e-01 -2.80036986e-01 3.43189687e-01
-4.39016223e-01 5.03529787e-01 -8.70311141e-01 -6.28336191e-01
-8.15334916e-01 -1.17057657e+00 1.23421347e+00 1.18980519e-01
-5.77826142e-01 -9.14763033e-01 -1.21042162e-01 3.88260670e-02
4.69859451e-01 -2.64270872e-01 1.03770697e+00 -8.19211185e-01
-6.26736522e-01 -1.36526868e-01 -2.73823380e-01 2.90058911e-01
4.09536511e-02 -1.93189874e-01 -6.67250216e-01 -1.57852918e-01
-4.70598668e-01 -3.56261015e-01 1.02930713e+00 2.29130700e-01
1.08256221e+00 -6.34534895e-01 -3.65742534e-01 5.74164212e-01
1.24667990e+00 2.62135565e-02 7.14104533e-01 4.04593647e-01
6.30810916e-01 5.05036175e-01 5.24411261e-01 4.47581917e-01
4.70437437e-01 5.82914293e-01 -2.28233144e-01 -1.03973515e-01
-3.03445399e-01 -6.57456696e-01 7.27581143e-01 9.54359770e-01
-2.63070222e-02 -5.81218243e-01 -7.18261242e-01 7.35450268e-01
-1.66688287e+00 -8.84627223e-01 1.42315924e-01 2.06218076e+00
1.66816902e+00 3.12905461e-01 1.21385798e-01 2.40078457e-02
1.67548984e-01 5.22224128e-01 -3.37145999e-02 -3.23464572e-01
-2.61956185e-01 6.03822768e-01 9.31562245e-01 3.37089241e-01
-9.79195058e-01 1.32010043e+00 6.57878733e+00 1.50373304e+00
-1.04920435e+00 -1.76506504e-01 5.35715640e-01 4.66958940e-04
-4.80037659e-01 -1.83688675e-03 -1.44167578e+00 4.79478896e-01
1.29616737e+00 -3.96922410e-01 3.77446324e-01 7.36359954e-01
-2.16811504e-02 -2.53163930e-02 -1.21105480e+00 8.15411866e-01
-6.97494075e-02 -1.82511771e+00 4.94019240e-01 7.81783909e-02
6.83358908e-01 -3.60024907e-02 -2.96651334e-01 4.80808467e-01
6.60533667e-01 -1.09120011e+00 8.32151771e-01 3.88707459e-01
8.83482873e-01 -6.52318597e-01 5.58209419e-01 4.49660182e-01
-1.17761838e+00 2.80941814e-01 -4.01249200e-01 1.74158409e-01
2.98684865e-01 8.32969785e-01 -1.00260448e+00 7.54189372e-01
2.37681881e-01 3.92815530e-01 -5.41978657e-01 7.09934711e-01
-4.58020061e-01 7.65839517e-01 -2.46891201e-01 -2.91450530e-01
4.50520068e-01 1.74972266e-01 4.09788638e-01 1.86367202e+00
3.55010897e-01 -9.28841531e-02 1.92371964e-01 8.83791029e-01
-2.03159645e-01 1.21113047e-01 -4.04660523e-01 -7.15509236e-01
5.35597205e-01 1.23046982e+00 -7.39060998e-01 -6.99326992e-01
-3.48814845e-01 9.39738929e-01 -5.17059006e-02 3.84307235e-01
-6.09128118e-01 -4.90162522e-01 6.63467467e-01 2.33930126e-01
2.94508040e-01 -4.51708823e-01 -2.59411663e-01 -1.19965422e+00
-2.58238018e-01 -1.30762506e+00 4.75669026e-01 -6.56309962e-01
-1.11963105e+00 7.01783478e-01 3.43921483e-01 -1.04088628e+00
-5.79191148e-01 -5.13093472e-01 -8.15476924e-02 1.10630810e+00
-1.80900598e+00 -1.23043418e+00 3.18000942e-01 2.32563481e-01
6.93285525e-01 -3.38384241e-01 9.27334309e-01 3.93396318e-01
-2.91637301e-01 6.62748873e-01 -1.70086116e-01 3.57643604e-01
7.36880064e-01 -1.18415046e+00 9.86958563e-01 9.36937511e-01
4.43071216e-01 9.39458609e-01 5.26319623e-01 -8.29841256e-01
-1.52419078e+00 -1.01359308e+00 1.25264323e+00 -5.37046492e-01
6.77013397e-01 -4.87826675e-01 -6.26618922e-01 4.97725815e-01
3.97275686e-01 -4.07828212e-01 6.19886220e-01 -2.13100333e-02
-3.17069948e-01 2.13357463e-01 -6.20258868e-01 9.85439777e-01
8.52767050e-01 -5.54136217e-01 -5.34079313e-01 2.63831198e-01
1.05382657e+00 -5.63753784e-01 -9.13443685e-01 3.06811631e-01
5.67229986e-01 -2.87750691e-01 1.03306329e+00 -7.27367878e-01
8.71186972e-01 -1.54426426e-01 -7.55324438e-02 -1.46520042e+00
-4.34033096e-01 -8.95760417e-01 -4.22708273e-01 1.18197715e+00
7.90425360e-01 -3.12868178e-01 6.60291731e-01 3.02523226e-01
-6.42677173e-02 -8.81342173e-01 -4.53936875e-01 -8.29627037e-01
1.41770184e-01 -5.67125797e-01 5.68004012e-01 7.91554630e-01
-1.17938772e-01 7.56607652e-01 -6.37172282e-01 -1.39053077e-01
4.25086580e-02 1.85795203e-02 8.88436615e-01 -8.58198225e-01
-4.27293837e-01 -6.55582726e-01 2.18813255e-01 -1.57653928e+00
2.88319066e-02 -9.48455036e-01 -6.47408441e-02 -1.68674695e+00
1.21344440e-01 -3.47891659e-01 -9.49275419e-02 6.95617259e-01
-5.15082061e-01 -5.18731289e-02 3.19581002e-01 1.40369326e-01
-2.79831082e-01 3.57081205e-01 1.24209619e+00 -6.51520789e-02
-1.12147607e-01 -1.46068841e-01 -1.01940274e+00 3.09120536e-01
5.67037463e-01 -5.47273755e-01 -5.04536867e-01 -3.31631720e-01
5.27579546e-01 -8.85427836e-03 -6.93795532e-02 -7.47950077e-01
2.47097164e-01 -3.67529631e-01 2.77724087e-01 -6.03497088e-01
1.96710676e-01 -4.35068429e-01 -9.62283313e-02 3.66231382e-01
-5.70365965e-01 1.83443964e-01 5.87482572e-01 1.82775468e-01
-1.92575887e-01 -4.38100696e-01 4.78149205e-01 -5.25701702e-01
-8.01513493e-01 3.15864891e-01 -5.85130155e-01 3.91811579e-01
3.77373785e-01 -2.68827137e-02 -2.29854971e-01 -5.65422654e-01
6.78752363e-02 6.54871389e-02 1.79869056e-01 6.65490210e-01
4.70579714e-01 -1.16133010e+00 -1.06334007e+00 1.15160212e-01
1.65274471e-01 -9.32182074e-02 -1.09782338e-01 4.80238289e-01
-4.92966890e-01 9.22630072e-01 2.10780874e-02 -6.10634089e-02
-9.80112493e-01 3.99474323e-01 -1.92205325e-01 -9.67879772e-01
-4.32115227e-01 9.55974579e-01 1.50830418e-01 -4.42841917e-01
1.23797461e-01 -5.12198985e-01 -1.06646530e-01 3.29624832e-01
7.88621843e-01 1.46098420e-01 2.26803079e-01 -2.75403500e-01
-4.00908440e-02 3.68457586e-02 -4.34837699e-01 -3.53629023e-01
1.24158859e+00 3.02461326e-01 -1.66700669e-02 2.25166887e-01
1.01907372e+00 1.91589102e-01 -8.23247910e-01 -4.87103999e-01
1.99017227e-01 -3.55858952e-02 2.33277380e-01 -1.03686547e+00
-8.17911386e-01 7.14752018e-01 -7.55607933e-02 2.10521463e-03
1.20284128e+00 -2.30874538e-01 1.27530396e+00 3.84496182e-01
3.66840988e-01 -1.19255102e+00 1.76825941e-01 8.59606862e-01
8.04169953e-01 -8.52576494e-01 4.05332834e-01 -1.55489936e-01
-7.63293743e-01 1.08046424e+00 3.59120905e-01 2.72129685e-01
3.71674120e-01 5.76677263e-01 -2.51962095e-01 1.67668965e-02
-1.01002872e+00 -2.78384745e-01 6.95359170e-01 2.30460480e-01
6.82330132e-01 -1.71797499e-02 -4.90537971e-01 7.17487991e-01
-6.16124213e-01 1.41689405e-01 4.46504205e-01 9.95281994e-01
-3.67249608e-01 -1.40265143e+00 -1.48323014e-01 6.22050703e-01
-8.72717261e-01 -8.31294119e-01 -4.58470166e-01 7.13329673e-01
-5.31883240e-02 7.24274099e-01 -1.89586937e-01 -4.37377423e-01
1.74682990e-01 3.23182344e-01 4.39683765e-01 -9.40897107e-01
-7.16735840e-01 7.12453499e-02 5.03549516e-01 -2.48694092e-01
-1.40628070e-01 -1.72533944e-01 -9.46337223e-01 -4.16670978e-01
-3.94965351e-01 2.81990737e-01 5.05542576e-01 8.27062488e-01
6.58888161e-01 6.27442002e-01 1.26814842e-01 -5.52531898e-01
-5.40904462e-01 -1.29918647e+00 -3.27427387e-01 3.41354869e-03
1.25529021e-01 -6.45409152e-02 1.15875863e-01 3.11669469e-01] | [12.205060958862305, 8.945391654968262] |
986cc64c-b9bc-4d08-95ca-0f691b661f88 | an-equivalent-circuit-approach-to-distributed | 2305.14607 | null | https://arxiv.org/abs/2305.14607v1 | https://arxiv.org/pdf/2305.14607v1.pdf | An Equivalent Circuit Approach to Distributed Optimization | Distributed optimization is an essential paradigm to solve large-scale optimization problems in modern applications where big-data and high-dimensionality creates a computational bottleneck. Distributed optimization algorithms that exhibit fast convergence allow us to fully utilize computing resources and effectively scale to larger optimization problems in a myriad of areas ranging from machine learning to power systems. In this work, we introduce a new centralized distributed optimization algorithm (ECADO) inspired by an equivalent circuit model of the distributed problem. The equivalent circuit (EC) model provides a physical analogy to derive new insights to develop a fast-convergent algorithm. The main contributions of this approach are: 1) a weighting scheme based on a circuit-inspired aggregate sensitivity analysis, and 2) an adaptive step-sizing derived from a stable, Backward-Euler numerical integration. We demonstrate that ECADO exhibits faster convergence compared to state-of-the art distributed optimization methods and provably converges for nonconvex problems. We leverage the ECADO features to solve convex and nonconvex optimization problems with large datasets such as: distributing data for logistic regression, training a deep neural network model for classification, and solving a high-dimensional problem security-constrained optimal power flow problem. Compared to state-of-the-art centralized methods, including ADMM, centralized gradient descent, and DANE, this new ECADO approach is shown to converge in fewer iterations. | ['Larry Pileggi', 'Aayushya Agarwal'] | 2023-05-24 | null | null | null | null | ['distributed-optimization', 'numerical-integration'] | ['methodology', 'miscellaneous'] | [-6.02351189e-01 -4.47057158e-01 -2.73183495e-01 -1.62704661e-01
-8.32179368e-01 -5.84271729e-01 1.56680904e-02 9.94147956e-02
-2.94998169e-01 1.08533382e+00 6.33784905e-02 -3.56485903e-01
-6.62093699e-01 -7.21000373e-01 -6.88263953e-01 -9.68026161e-01
-2.89946079e-01 4.95699137e-01 -5.02261162e-01 -2.32696041e-01
5.61772846e-02 7.69784272e-01 -8.51397872e-01 -2.77268678e-01
1.20472884e+00 1.35165799e+00 -9.31558460e-02 3.35788101e-01
2.52718359e-01 6.64536297e-01 -5.39243281e-01 -1.02708917e-02
6.63173079e-01 -4.92021114e-01 -5.33236086e-01 -2.37284690e-01
1.78075418e-01 -5.32706976e-01 -2.01655805e-01 9.70322073e-01
9.22393024e-01 4.32855934e-01 4.28493828e-01 -1.46444678e+00
-5.27071238e-01 4.12826538e-01 -7.12183058e-01 1.55438423e-01
-3.67925018e-01 -2.92714150e-03 1.07423615e+00 -8.16954136e-01
3.01102787e-01 7.79108524e-01 9.20359135e-01 1.68834656e-01
-1.34229791e+00 -5.14006853e-01 6.59161955e-02 1.51281834e-01
-1.61574066e+00 -8.81271735e-02 8.65612745e-01 -1.56528428e-01
1.00840342e+00 2.52390772e-01 7.74236441e-01 2.84091026e-01
3.13903362e-01 6.73718154e-01 8.17303538e-01 -2.17926726e-01
7.41379142e-01 -1.16175562e-01 -2.10714579e-01 5.86360633e-01
4.14137632e-01 1.16208326e-02 -4.61406916e-01 -5.83300650e-01
5.16718090e-01 1.36219844e-01 -2.36158833e-01 -4.09154475e-01
-7.21187592e-01 1.16293275e+00 6.95326984e-01 2.10173443e-01
-5.57870150e-01 4.00060117e-01 4.02527034e-01 1.66304708e-01
8.70054364e-01 5.22702754e-01 -6.55955434e-01 -2.30050921e-01
-1.03110027e+00 6.05978668e-01 1.16794825e+00 5.23797989e-01
6.98025048e-01 3.76941711e-01 -1.25220686e-01 8.54940295e-01
1.51744112e-01 8.30528080e-01 1.72245413e-01 -1.09997272e+00
5.64463794e-01 4.54115599e-01 3.05476580e-02 -1.11382127e+00
-6.94985986e-01 -7.04925776e-01 -1.15348160e+00 2.88057476e-01
3.39396179e-01 -6.83895946e-01 1.23308025e-01 1.53247261e+00
5.98424673e-01 -2.17703044e-01 -2.70020694e-01 1.25857615e+00
-1.51067421e-01 1.01184618e+00 -4.35828716e-01 -3.34637493e-01
8.52597892e-01 -1.21700048e+00 -5.18775284e-01 2.14696929e-01
7.50430584e-01 -2.40890443e-01 5.49994409e-01 3.83906662e-01
-1.20008314e+00 8.00097957e-02 -1.06474650e+00 -1.65050372e-01
-1.94252610e-01 1.45303801e-01 8.07222605e-01 6.83744133e-01
-1.14986253e+00 9.05281007e-01 -7.47164786e-01 1.00880191e-01
8.10079098e-01 5.32326460e-01 1.32178351e-01 5.81674241e-02
-6.63179755e-01 8.00646901e-01 -1.24083973e-01 4.25899744e-01
-7.48784125e-01 -1.29741943e+00 -5.55516720e-01 3.80234599e-01
4.07781065e-01 -1.00585353e+00 1.09028947e+00 -1.00706971e+00
-1.90343165e+00 -1.20262709e-02 -4.21560258e-02 -4.18796957e-01
3.76497149e-01 -7.35805780e-02 1.15774147e-01 1.00325041e-01
-1.26290679e-01 -1.11725211e-01 7.23920286e-01 -7.96092331e-01
-3.69168550e-01 -4.74194616e-01 -2.80952394e-01 3.33285838e-01
-9.22775269e-01 -3.80487368e-02 8.69426280e-02 -6.55982494e-01
-4.88361984e-01 -6.36694551e-01 -6.20743811e-01 2.34706789e-01
-2.53948241e-01 -1.42724514e-02 9.20726538e-01 -6.88283682e-01
1.31926537e+00 -1.80231786e+00 5.35717666e-01 5.96447468e-01
6.57601953e-01 9.66180116e-02 1.89592708e-02 5.14568031e-01
2.28974968e-01 6.76719248e-02 -2.12164000e-01 -3.86052489e-01
4.52216625e-01 7.99018964e-02 -1.42333105e-01 9.86230433e-01
-2.62576044e-01 9.67623770e-01 -6.59574628e-01 -4.41605709e-02
4.61210087e-02 4.10598934e-01 -8.35506201e-01 -3.22538540e-02
2.52401046e-02 3.02715272e-01 -6.39993012e-01 4.23896998e-01
6.54769778e-01 -5.73519945e-01 3.21196169e-01 -3.48968059e-01
-2.43622497e-01 -2.00422585e-01 -1.23072696e+00 1.84484994e+00
-9.49448764e-01 6.98301494e-01 8.82019877e-01 -1.56345940e+00
5.60917437e-01 1.59680977e-01 9.37963128e-01 -8.49431694e-01
3.15670431e-01 5.07470191e-01 -1.86636627e-01 -7.68035650e-02
4.72958088e-01 -2.12866589e-02 -8.27071220e-02 7.79163361e-01
-9.10407454e-02 -1.62937194e-01 1.51467741e-01 3.10908705e-01
1.26521134e+00 -4.98971760e-01 -5.21226227e-02 -9.86080527e-01
3.31184328e-01 1.12291882e-02 7.08264351e-01 5.93721688e-01
-3.36160026e-02 2.61544347e-01 6.20287597e-01 -6.49369776e-01
-1.27951622e+00 -7.67145872e-01 -7.04892129e-02 9.52089846e-01
1.25475049e-01 -3.44153970e-01 -7.40484416e-01 -4.14173096e-01
4.91925031e-01 2.55724221e-01 -1.64798960e-01 -3.36002605e-03
-6.17877603e-01 -1.02412260e+00 2.36400500e-01 6.46406770e-01
4.57418412e-01 -2.21588701e-01 -3.74428004e-01 2.82763362e-01
1.71308339e-01 -9.82360303e-01 -7.55189419e-01 1.95918009e-01
-9.40287113e-01 -7.06575394e-01 -8.79012644e-01 -4.68961060e-01
9.54308748e-01 -8.91731083e-02 1.10525537e+00 1.24305725e-01
-7.93089509e-01 3.97567183e-01 2.03460991e-01 -2.04470441e-01
9.85982195e-02 1.57915652e-01 3.04827332e-01 1.36197522e-01
-3.64175230e-01 -6.66037083e-01 -7.72904277e-01 3.51273626e-01
-6.94596827e-01 -1.20947845e-01 3.44407082e-01 8.63173902e-01
4.72566873e-01 2.37093374e-01 7.69870639e-01 -5.80638051e-01
1.02996182e+00 -7.29031205e-01 -1.35974860e+00 1.52902111e-01
-8.28983665e-01 2.71567076e-01 1.46777260e+00 -2.80730635e-01
-7.64398694e-01 -1.31581128e-01 3.95476073e-02 -6.36427939e-01
6.61906898e-01 5.59690595e-01 1.07772334e-03 -4.81639236e-01
4.90169436e-01 -9.98809114e-02 1.25328094e-01 -5.07192671e-01
5.35436451e-01 5.28699875e-01 1.15065537e-01 -1.08113003e+00
7.94295251e-01 6.22025371e-01 4.86253560e-01 -6.38600469e-01
-4.35933411e-01 -3.52414250e-01 -2.22070199e-02 -7.18515515e-02
4.98690873e-01 -7.57229745e-01 -1.21845591e+00 4.84788537e-01
-1.07916570e+00 -7.71887362e-01 -6.27914429e-01 2.88401991e-01
-3.44087243e-01 -5.98678328e-02 -6.95591807e-01 -7.42480516e-01
-7.59426713e-01 -9.64171946e-01 8.95425141e-01 4.71343040e-01
2.43426576e-01 -1.53050196e+00 1.52298525e-01 1.85449913e-01
1.22740650e+00 3.61974210e-01 6.99863434e-01 -2.77684271e-01
-5.10334015e-01 -9.61758569e-02 -2.40461886e-01 3.91691893e-01
-2.29971692e-01 -8.61252397e-02 -4.13532883e-01 -6.28445029e-01
8.07364434e-02 -3.78494263e-01 2.66066343e-01 6.02857590e-01
1.47809267e+00 -6.84263885e-01 -3.57192606e-01 1.16831470e+00
1.94248927e+00 -2.24510878e-01 1.93131641e-01 -6.63767681e-02
6.00042880e-01 1.09305261e-02 -3.58541124e-02 9.61296141e-01
3.61916959e-01 7.65594840e-01 2.95844406e-01 -3.38995248e-01
2.79322773e-01 1.51015231e-02 1.43551439e-01 1.01405489e+00
-4.12580408e-02 -9.57502052e-02 -7.32579708e-01 5.34383118e-01
-1.94429541e+00 -7.73439646e-01 -1.24723576e-01 2.08858919e+00
6.66041553e-01 -5.71883738e-01 1.88864723e-01 -1.11729592e-01
4.97902304e-01 -2.71228943e-02 -9.32883382e-01 -6.49400651e-01
-6.72235116e-02 5.57910800e-01 9.81819093e-01 3.99642318e-01
-7.45832920e-01 1.92835480e-01 6.23364496e+00 1.20894957e+00
-1.28457153e+00 6.55723333e-01 8.53140533e-01 -6.54594004e-01
-3.28991622e-01 -6.75535277e-02 -6.47138238e-01 5.08588314e-01
9.19099152e-01 -6.83588564e-01 1.11741936e+00 9.52250659e-01
5.10903060e-01 -1.08714879e-01 -9.26450133e-01 1.20505607e+00
-4.89395054e-04 -1.65043426e+00 -6.60033166e-01 3.39186221e-01
1.32340074e+00 4.23974156e-01 -9.38435346e-02 9.59878266e-02
3.01997453e-01 -1.10741961e+00 4.92107630e-01 2.37406433e-01
5.70982695e-01 -6.94342852e-01 4.00422364e-01 2.69018561e-01
-1.23484755e+00 -6.45615160e-01 -2.98615843e-01 -4.11717892e-01
4.00494665e-01 1.26269853e+00 3.56581695e-02 7.06023753e-01
6.80535197e-01 6.99953079e-01 9.50709954e-02 1.19755757e+00
2.91747935e-02 5.79619586e-01 -9.33358729e-01 -3.76775473e-01
1.64784804e-01 -6.29959762e-01 5.81313074e-01 6.24243677e-01
3.48309040e-01 6.25522286e-02 1.17337545e-02 1.37061226e+00
-4.76698369e-01 2.13876516e-01 -3.02667499e-01 9.59815308e-02
5.20585775e-01 1.74299383e+00 -4.69067663e-01 -4.65026312e-03
-4.19788152e-01 7.88674772e-01 3.60208303e-01 5.69702685e-01
-9.49386954e-01 -6.61791265e-01 6.95950925e-01 -2.38142535e-01
3.27246577e-01 -5.45986414e-01 -6.18971765e-01 -1.18815100e+00
3.93994093e-01 -6.43400073e-01 2.63035238e-01 -3.68863910e-01
-1.52733421e+00 2.19721049e-01 -1.08609259e-01 -9.57819343e-01
-1.36007547e-01 -6.38394594e-01 -9.35641229e-01 1.07023096e+00
-1.57284629e+00 -8.54490340e-01 -2.74767190e-01 6.55283332e-01
9.27299112e-02 -1.99665055e-01 5.94531238e-01 8.32386553e-01
-9.33581054e-01 5.53829312e-01 9.68048751e-01 -4.94807027e-02
1.83227554e-01 -1.17224967e+00 -1.92487866e-01 7.26682425e-01
-2.14620575e-01 1.92520589e-01 3.03319395e-01 -2.38389209e-01
-2.35231590e+00 -8.76512766e-01 4.00769025e-01 3.98698673e-02
9.05443907e-01 -5.11786520e-01 -4.74582434e-01 1.63027704e-01
2.03351319e-01 7.81343460e-01 5.61804235e-01 2.30325043e-01
1.55144304e-01 -8.47479582e-01 -1.20708978e+00 7.10459650e-02
7.24969268e-01 -2.13880226e-01 2.14948073e-01 6.38515174e-01
1.75310075e-01 -5.03907263e-01 -1.09737015e+00 -9.64436382e-02
4.78837729e-01 -5.62550247e-01 8.40517521e-01 -4.06947881e-01
4.34747785e-01 -1.00188859e-01 -4.95584719e-02 -1.61514974e+00
-1.44822765e-02 -1.30480552e+00 -4.85394388e-01 8.72266114e-01
4.34423894e-01 -1.02882087e+00 6.83581829e-01 9.08967555e-01
-2.00108588e-01 -1.27465105e+00 -1.26975858e+00 -9.79062915e-01
5.44302166e-01 -1.61157340e-01 5.41552007e-01 8.46704364e-01
2.85353303e-01 7.52675831e-02 -1.92440584e-01 1.13642678e-01
7.34100699e-01 2.22594962e-01 7.14871287e-01 -7.58103073e-01
-5.90502083e-01 -7.16445148e-01 -1.63716614e-01 -1.18397951e+00
1.53295606e-01 -9.57175910e-01 -2.81897932e-01 -1.39550340e+00
3.84828001e-02 -8.59440684e-01 -6.19663410e-02 5.89018106e-01
2.01353028e-01 1.34720847e-01 2.43625998e-01 1.89446583e-01
-6.36316299e-01 9.34272051e-01 8.76173377e-01 -2.60028630e-01
-2.96223044e-01 -2.19428599e-01 -6.78867519e-01 3.10066611e-01
6.13020480e-01 -5.41938782e-01 -2.52190769e-01 -8.77828181e-01
7.47599423e-01 8.99679810e-02 2.83109546e-01 -7.81230569e-01
7.09089875e-01 -1.51971713e-01 1.27669156e-01 -1.50805665e-02
1.44642042e-02 -1.04355550e+00 -1.56752840e-01 4.74796921e-01
1.14249296e-01 2.04978436e-01 3.31293941e-01 1.51661277e-01
-1.64698437e-01 6.73661940e-03 5.89836776e-01 4.34006691e-01
-2.05605000e-01 4.59471017e-01 -2.17292726e-01 3.72447044e-01
1.16333294e+00 3.78178537e-01 -5.84122360e-01 -4.97740805e-01
-1.36464193e-01 6.44302368e-01 2.00742409e-01 5.10464422e-02
2.85332292e-01 -1.45830858e+00 -7.65207410e-01 7.81804994e-02
-7.71395028e-01 1.27383128e-01 2.70271450e-01 1.39774227e+00
-8.12192321e-01 2.19075829e-01 1.84837088e-01 -5.88351309e-01
-4.56326038e-01 1.28451303e-01 6.74336255e-01 -5.15837848e-01
-5.19242764e-01 5.43959081e-01 -3.45361263e-01 -4.60849524e-01
-1.51677907e-01 -2.84590513e-01 6.81907773e-01 -2.78756410e-01
3.54161233e-01 9.32610273e-01 3.99324834e-01 1.02864824e-01
-3.31245869e-01 6.09815061e-01 4.68650818e-01 -1.12228775e-02
1.65171278e+00 -7.92888366e-03 -5.72980285e-01 -7.62510598e-02
1.75387728e+00 1.55504599e-01 -1.21868217e+00 7.58230314e-02
-6.29011929e-01 -3.59228849e-01 5.48490763e-01 -7.50554025e-01
-1.74070740e+00 5.60464859e-01 4.18961912e-01 1.87481031e-01
1.36119139e+00 -2.93856889e-01 1.07654631e+00 5.21271765e-01
4.45958376e-01 -1.39403737e+00 -2.59434819e-01 4.04680341e-01
6.83568895e-01 -9.26232219e-01 5.57266653e-01 3.56048085e-02
-1.31262198e-01 1.23202431e+00 2.74471283e-01 -3.19030195e-01
9.09693837e-01 7.65448153e-01 -3.86511475e-01 -1.26579599e-02
-6.28883660e-01 4.87699091e-01 1.10671729e-01 -3.58623862e-02
-1.25410408e-01 -1.20679088e-01 -5.18812239e-01 8.74175191e-01
7.53617361e-02 1.15284719e-01 2.73272425e-01 9.81488287e-01
5.34376837e-02 -1.01235807e+00 -3.24798971e-01 6.01078928e-01
-3.98774207e-01 2.74880528e-02 1.38385683e-01 4.18896765e-01
-2.84053773e-01 7.43914425e-01 2.21724406e-01 1.33991182e-01
2.83272434e-02 -3.75613749e-01 4.02478337e-01 -1.05807327e-01
-7.80309200e-01 -1.75809205e-01 -2.79657662e-01 -8.29500973e-01
-7.77613372e-02 -2.81823218e-01 -1.23212063e+00 -7.59221196e-01
-6.00232482e-01 4.01695162e-01 1.28870070e+00 1.12742603e+00
8.87172163e-01 5.05323887e-01 1.04533553e+00 -1.20156777e+00
-1.11298120e+00 -3.55242193e-01 -7.88861573e-01 1.04228929e-01
1.66858181e-01 -3.08490008e-01 -5.31334400e-01 -5.11967480e-01] | [6.256715297698975, 4.975368022918701] |
8d3cd312-7175-4f35-8f65-c9775fadaeff | emphcmsalgan-rgb-d-salient-object-detection | 1912.10280 | null | https://arxiv.org/abs/1912.10280v2 | https://arxiv.org/pdf/1912.10280v2.pdf | \emph{cm}SalGAN: RGB-D Salient Object Detection with Cross-View Generative Adversarial Networks | Image salient object detection (SOD) is an active research topic in computer vision and multimedia area. Fusing complementary information of RGB and depth has been demonstrated to be effective for image salient object detection which is known as RGB-D salient object detection problem. The main challenge for RGB-D salient object detection is how to exploit the salient cues of both intra-modality (RGB, depth) and cross-modality simultaneously which is known as cross-modality detection problem. In this paper, we tackle this challenge by designing a novel cross-modality Saliency Generative Adversarial Network (\emph{cm}SalGAN). \emph{cm}SalGAN aims to learn an optimal view-invariant and consistent pixel-level representation for RGB and depth images via a novel adversarial learning framework, which thus incorporates both information of intra-view and correlation information of cross-view images simultaneously for RGB-D saliency detection problem. To further improve the detection results, the attention mechanism and edge detection module are also incorporated into \emph{cm}SalGAN. The entire \emph{cm}SalGAN can be trained in an end-to-end manner by using the standard deep neural network framework. Experimental results show that \emph{cm}SalGAN achieves the new state-of-the-art RGB-D saliency detection performance on several benchmark datasets. | ['Xiao Wang', 'Zitai Zhou', 'Jin Tang', 'Bin Luo', 'Bo Jiang'] | 2019-12-21 | null | null | null | null | ['rgb-d-salient-object-detection'] | ['computer-vision'] | [ 5.36273777e-01 5.07871024e-02 3.06109469e-02 2.72972567e-04
-8.55477512e-01 -8.55935588e-02 2.57511109e-01 -7.12157115e-02
-2.77574599e-01 4.50228006e-01 1.79718897e-01 -2.94036535e-03
1.67454511e-01 -6.16782904e-01 -8.17548335e-01 -8.38212907e-01
3.24552119e-01 -4.99892741e-01 8.03643882e-01 -5.63653409e-01
2.74942398e-01 3.42339039e-01 -1.69621360e+00 2.80956268e-01
7.01389015e-01 1.17310810e+00 6.74695671e-01 6.31468296e-01
1.93838775e-01 8.73784661e-01 -2.33543843e-01 -1.18262626e-01
4.55571413e-01 -5.09516180e-01 -5.42286575e-01 2.31574148e-01
4.18898761e-01 -3.77024770e-01 -3.84993255e-01 1.28808343e+00
8.19369972e-01 1.24587029e-01 2.01339409e-01 -1.43944800e+00
-8.16034138e-01 3.09621602e-01 -1.13868940e+00 8.21078181e-01
2.41616696e-01 1.66310236e-01 5.94836354e-01 -1.14761496e+00
4.24255133e-01 1.19927394e+00 3.62251997e-01 4.68999296e-01
-8.37863505e-01 -6.08498693e-01 2.88706392e-01 2.76857376e-01
-1.17288399e+00 -4.90016565e-02 1.62129331e+00 5.48293106e-02
4.63152468e-01 3.66202772e-01 6.79195166e-01 7.43779838e-01
1.19875960e-01 1.42211616e+00 1.20007849e+00 -4.31398243e-01
1.30643591e-01 9.79072899e-02 -2.93306410e-01 8.30697179e-01
1.77437216e-01 1.22595504e-01 -7.28851020e-01 2.44197100e-01
1.11818159e+00 3.90238851e-01 -2.89395332e-01 -7.52407312e-01
-1.15601933e+00 7.81289876e-01 1.13262939e+00 2.15444446e-01
-4.39462930e-01 7.67433941e-02 3.02161872e-01 -3.92458253e-02
4.86620337e-01 4.49599884e-02 -3.47910464e-01 5.25025725e-01
-7.20219254e-01 1.12546638e-01 -9.11619067e-02 9.92084920e-01
7.95791805e-01 4.20681387e-01 -2.04776809e-01 5.41653693e-01
4.07443494e-01 7.75996327e-01 6.38298333e-01 -6.59618795e-01
4.16022927e-01 8.10638189e-01 1.01905584e-01 -1.22538817e+00
-4.42287147e-01 -6.18454218e-01 -8.24198842e-01 4.97810453e-01
2.83298083e-02 -8.29085633e-02 -1.06771135e+00 1.64487696e+00
6.99744165e-01 3.23829710e-01 2.35942915e-01 1.45273757e+00
1.35244262e+00 4.80232328e-01 1.61488384e-01 -1.23596273e-01
1.31997073e+00 -9.72758055e-01 -3.61120671e-01 -7.40678310e-01
-8.88846070e-02 -9.21970963e-01 1.00930214e+00 -2.69358963e-01
-1.30297995e+00 -6.61410034e-01 -1.09795964e+00 -4.62913424e-01
-3.58090729e-01 -2.46074591e-02 6.60275280e-01 3.37687016e-01
-8.94094348e-01 -2.04875305e-01 -7.31289625e-01 -8.22532922e-02
6.56622469e-01 3.13560694e-01 -2.49153927e-01 -1.56590402e-01
-1.22829247e+00 6.57289743e-01 5.56200325e-01 1.68880835e-01
-9.83526707e-01 -6.50768101e-01 -1.19199789e+00 -1.16842248e-01
4.83718485e-01 -8.02352369e-01 9.46191072e-01 -1.25078404e+00
-1.00449419e+00 1.07687700e+00 -6.11761250e-02 -1.32438660e-01
2.34789386e-01 3.10934242e-02 -3.55684012e-01 4.62882280e-01
3.86774391e-01 6.85784698e-01 1.11641800e+00 -1.58020294e+00
-9.19459701e-01 -5.44096351e-01 4.29325476e-02 5.41311502e-01
-1.08312674e-01 1.84267983e-01 -5.15839279e-01 -9.56947684e-01
5.18921554e-01 -6.61625803e-01 -9.38968286e-02 1.92946440e-03
-5.07144690e-01 2.64755432e-02 1.18766487e+00 -7.24509716e-01
6.89553380e-01 -1.90060866e+00 2.97186702e-01 -1.65571943e-01
2.16521516e-01 4.65183884e-01 1.19451784e-01 -3.10157277e-02
-8.64597484e-02 -3.97943825e-01 -4.90509868e-01 -3.07004362e-01
-3.54128003e-01 -1.60237476e-01 -1.36994898e-01 5.91896772e-01
4.59116846e-01 1.25626123e+00 -1.19320273e+00 -6.88593686e-01
6.96431458e-01 6.23258710e-01 -2.64831692e-01 2.56258905e-01
-8.52255970e-02 4.15841222e-01 -7.11684406e-01 1.19158149e+00
9.56502676e-01 -1.40192866e-01 -4.49549049e-01 -3.07149321e-01
-8.67403448e-02 -3.07446003e-01 -1.15407300e+00 1.86423802e+00
-2.83450544e-01 5.63976347e-01 -4.76157814e-02 -8.84301186e-01
8.82131457e-01 -1.20892473e-01 3.08268487e-01 -9.28110003e-01
2.00180143e-01 2.44212836e-01 -4.68825161e-01 -3.73158932e-01
4.92248774e-01 -2.23025978e-01 -1.93992630e-01 1.97279140e-01
-1.55971706e-01 -2.21065253e-01 -3.64477277e-01 1.65614560e-01
5.80051005e-01 -3.81956249e-03 3.05328578e-01 -2.80117076e-02
8.28726113e-01 -1.31312430e-01 5.41119099e-01 5.22918165e-01
-4.60484743e-01 9.49558973e-01 5.77756576e-02 -1.75837502e-01
-8.57602000e-01 -1.21199870e+00 1.78249717e-01 1.01469052e+00
9.75788891e-01 2.87405193e-01 -5.54155767e-01 -7.58222640e-01
2.90157460e-02 4.25340325e-01 -8.84921193e-01 -4.58432138e-01
-5.07681668e-01 -4.66803908e-01 -4.36634338e-03 6.81143165e-01
9.92312729e-01 -1.35737896e+00 -1.08962178e+00 4.12047394e-02
-1.91372231e-01 -1.02507639e+00 -5.99900305e-01 2.13198766e-01
-6.78853512e-01 -9.46027696e-01 -1.19309247e+00 -1.17109001e+00
6.55212700e-01 1.08480000e+00 8.70240510e-01 -1.15916617e-01
-4.24844205e-01 3.33394021e-01 -3.41278315e-01 -7.19798207e-01
-1.26079833e-02 -1.77267149e-01 -2.89922684e-01 9.52716023e-02
2.17631459e-01 -4.06164795e-01 -1.19218707e+00 2.61320382e-01
-1.30576229e+00 2.98899740e-01 1.01462352e+00 8.70291054e-01
8.24986219e-01 -1.29222989e-01 5.02393186e-01 -4.60677981e-01
1.28184929e-01 -3.39044511e-01 -3.21032822e-01 8.09393898e-02
-2.44418100e-01 -2.73079097e-01 2.46453285e-01 -2.44498588e-02
-1.10435283e+00 2.55655289e-01 -1.10954151e-01 -7.24810362e-01
-9.52238217e-02 2.11780220e-01 -4.04255867e-01 -2.40341052e-01
2.17485011e-01 7.83865869e-01 -8.36502835e-02 -1.97886065e-01
3.12582195e-01 6.61787629e-01 8.39817464e-01 1.75471067e-01
8.77816617e-01 7.45898187e-01 1.82973538e-02 -7.02663600e-01
-1.16798913e+00 -7.78748095e-01 -3.76849830e-01 -3.13093811e-01
9.66185033e-01 -1.27725017e+00 -4.59331691e-01 7.15370238e-01
-9.25622821e-01 -5.29719479e-02 -2.04030439e-01 1.02640204e-01
-5.53326368e-01 3.79510701e-01 -2.31787533e-01 -6.63119018e-01
-6.70641065e-01 -1.15763855e+00 1.67998123e+00 8.94561410e-01
4.40498590e-01 -8.98552597e-01 -3.47056091e-01 4.69708592e-01
2.99664885e-01 6.86228693e-01 4.38869476e-01 -6.43063262e-02
-8.19726050e-01 -1.89031512e-01 -6.87742889e-01 3.93560708e-01
1.49837062e-01 -6.72292352e-01 -8.45644951e-01 -2.97157586e-01
1.31493822e-01 -1.13437414e-01 9.64093328e-01 4.37872559e-01
1.00984740e+00 -2.44110059e-02 -2.38781556e-01 5.67865372e-01
1.66489291e+00 -1.03954457e-01 5.82011402e-01 4.95876193e-01
1.15575469e+00 2.13349611e-01 9.45237041e-01 3.38244230e-01
5.51642478e-01 7.09185958e-01 8.69165599e-01 -6.90808177e-01
-3.13360512e-01 -3.18567157e-01 2.02205271e-01 2.29440570e-01
1.91836074e-01 1.03439167e-01 -7.12961316e-01 7.85495818e-01
-1.78868079e+00 -1.02944076e+00 -5.64309210e-02 1.85492218e+00
8.40784848e-01 1.30951792e-01 2.01891258e-01 2.99782902e-01
8.76555383e-01 2.52136856e-01 -8.24734747e-01 -2.69496050e-02
-4.47628438e-01 1.32855356e-01 6.57860756e-01 3.06587219e-02
-1.33927870e+00 7.68713534e-01 4.22360802e+00 8.76517951e-01
-1.38627839e+00 2.57948637e-01 6.69599473e-01 3.21453693e-03
-2.34209657e-01 -2.52690732e-01 -6.43946290e-01 7.69310415e-01
3.74367670e-03 4.64797728e-02 -7.77641535e-02 9.03644383e-01
1.38352528e-01 -4.61805701e-01 -5.22717595e-01 1.20226848e+00
5.44308960e-01 -1.20440602e+00 -1.85841545e-01 -3.43151599e-01
1.18364310e+00 -7.31778657e-03 3.96401227e-01 2.29825545e-03
-6.51771277e-02 -6.18100643e-01 8.97944987e-01 3.25453460e-01
5.78416467e-01 -1.02495885e+00 8.79399121e-01 1.46541953e-01
-1.29016268e+00 -2.96623200e-01 -3.37665707e-01 2.24691793e-01
1.96113676e-01 4.49353695e-01 -4.57195014e-01 6.69675350e-01
1.01341331e+00 9.36229110e-01 -7.08926380e-01 9.75158453e-01
-3.50791335e-01 6.85966313e-02 -4.28185165e-02 4.49286848e-02
4.36097443e-01 2.77496755e-01 8.89918029e-01 9.89425600e-01
8.68583694e-02 1.57632187e-01 7.40412921e-02 8.52073908e-01
-2.38427576e-02 -2.36622915e-01 -3.71703327e-01 5.96987844e-01
1.52974874e-01 1.22124970e+00 -9.15957928e-01 -1.20321207e-01
-4.00599748e-01 1.34742022e+00 -3.30330469e-02 1.63357690e-01
-8.98527145e-01 -3.99295717e-01 4.31233883e-01 1.30429015e-01
8.15877080e-01 1.57685399e-01 -3.06370705e-01 -9.54368353e-01
3.83643769e-02 -7.13836730e-01 4.06974435e-01 -1.26457238e+00
-9.42542076e-01 2.87008107e-01 -1.78443998e-01 -1.43715942e+00
2.63183534e-01 -4.42662030e-01 -6.47319734e-01 9.49065626e-01
-2.25195336e+00 -1.70120323e+00 -8.15827370e-01 9.95240808e-01
6.44189835e-01 -2.97773294e-02 2.69965827e-01 6.82633445e-02
-4.72496003e-01 5.29581428e-01 7.20668361e-02 8.08197111e-02
5.21732330e-01 -1.25035393e+00 1.55874833e-01 1.25141442e+00
-2.33491763e-01 1.55208826e-01 7.55688667e-01 -5.19115388e-01
-1.51775014e+00 -1.52235913e+00 3.57836604e-01 -2.18012661e-01
2.08518088e-01 -5.75945936e-02 -6.31401598e-01 3.95207405e-01
2.07378745e-01 4.37685102e-01 2.53406882e-01 -7.68788218e-01
5.10235950e-02 -1.78484365e-01 -1.23036289e+00 4.85658526e-01
8.95111620e-01 -5.70466220e-01 -7.48635650e-01 1.13870047e-01
1.06011176e+00 -8.39869440e-01 -5.05247891e-01 5.47602773e-01
1.81601837e-01 -1.30385983e+00 1.49975026e+00 -1.35327533e-01
7.66972423e-01 -6.80517673e-01 -1.15074001e-01 -1.07645798e+00
1.12418719e-01 -4.04157281e-01 -3.31174612e-01 1.13966978e+00
-1.58271313e-01 -3.92235339e-01 7.63485610e-01 3.10201496e-01
-4.55945104e-01 -9.17624831e-01 -8.02710950e-01 -2.79044271e-01
-3.95917267e-01 -1.56471640e-01 4.04201061e-01 7.55391955e-01
-4.02537793e-01 1.29197344e-01 -3.87559563e-01 3.93073291e-01
9.19948459e-01 6.86760366e-01 6.27014935e-01 -8.18882465e-01
-3.57926153e-02 -3.90381366e-01 -7.10816145e-01 -9.27325010e-01
-2.26797476e-01 -7.88318932e-01 1.20909460e-01 -1.65278018e+00
3.16781163e-01 -3.22454780e-01 -7.33002782e-01 2.98935831e-01
-7.79326916e-01 8.69205356e-01 2.21699864e-01 -3.53545733e-02
-8.63023162e-01 7.70711780e-01 1.54146135e+00 -2.36871243e-01
-2.71620810e-01 -9.43783447e-02 -1.07040226e+00 6.50400937e-01
6.67860031e-01 -2.75578529e-01 -3.26271862e-01 -2.04720005e-01
-5.03136963e-02 1.60473466e-01 1.03946137e+00 -1.07090700e+00
3.49664778e-01 -6.83254302e-02 8.83880973e-01 -9.46457982e-01
3.99357170e-01 -6.64076090e-01 -3.69974256e-01 3.76045704e-01
-1.04562305e-01 -9.77530852e-02 4.46488202e-01 8.21896076e-01
-3.93164814e-01 2.26381510e-01 1.03658748e+00 -2.35273182e-01
-1.34256148e+00 2.45638654e-01 2.62604535e-01 2.22355321e-01
1.38159430e+00 -4.77709621e-01 -2.46529043e-01 -1.25850096e-01
-3.75335336e-01 2.47394711e-01 4.66993302e-01 6.39854550e-01
1.10448861e+00 -1.33179581e+00 -6.24697864e-01 3.58081996e-01
3.76243502e-01 2.62959719e-01 6.79012358e-01 9.48774636e-01
-3.50785732e-01 2.70752281e-01 -5.00844836e-01 -7.89507449e-01
-1.24487293e+00 7.91457593e-01 2.45493472e-01 9.44461077e-02
-4.16040331e-01 1.21409285e+00 2.77253002e-01 -3.11949337e-03
1.27811711e-02 -2.04843044e-01 -1.68852732e-01 -1.28934860e-01
4.92756814e-01 1.65894762e-01 -5.64409681e-02 -9.87070799e-01
-5.40978432e-01 6.58179700e-01 -5.25950221e-03 3.77175987e-01
1.36044073e+00 -5.93121946e-01 5.23822047e-02 1.95403308e-01
1.29781866e+00 -1.48268774e-01 -1.45099688e+00 -4.60941941e-01
-5.86184680e-01 -7.09800899e-01 4.94633853e-01 -6.78716719e-01
-1.34730530e+00 7.13280618e-01 1.08763361e+00 -6.69050664e-02
1.54574955e+00 6.99218214e-02 1.08638322e+00 -3.44101965e-01
1.95666462e-01 -9.62517977e-01 5.49142957e-01 -3.38393971e-02
9.39247727e-01 -1.93381989e+00 1.97512314e-01 -4.55528408e-01
-8.12245429e-01 6.72596931e-01 6.67745471e-01 -2.46149778e-01
5.35738051e-01 -3.12341869e-01 6.28612861e-02 -2.71763831e-01
-6.22155028e-04 -7.37000048e-01 5.55983245e-01 5.16452789e-01
-7.71657303e-02 -7.47132301e-02 2.21869033e-02 5.77079654e-01
2.91954011e-01 -1.72934026e-01 3.65477324e-01 1.19976437e+00
-4.97787714e-01 -6.67019665e-01 -4.45237905e-01 1.45569012e-01
-4.27143157e-01 -1.32480919e-01 -2.71925420e-01 7.77921975e-01
3.46204519e-01 9.36644137e-01 -1.63933516e-01 -2.34031320e-01
3.30869287e-01 -4.09902185e-01 2.74124831e-01 -3.53839666e-01
-4.68782425e-01 6.04347438e-02 -5.97817004e-01 -4.43565488e-01
-7.76752889e-01 -7.69148171e-01 -1.35928595e+00 8.55981186e-02
-3.45704585e-01 -2.89510697e-01 5.26394010e-01 7.91233361e-01
3.32769424e-01 7.77153730e-01 9.44740415e-01 -1.15659893e+00
-2.10206568e-01 -6.32513642e-01 -4.82853740e-01 4.56231147e-01
7.48460948e-01 -8.11665893e-01 -2.39738435e-01 1.97865069e-02] | [9.709718704223633, -0.751375138759613] |
8c712e67-cf9b-4193-88d2-fea3739921cb | delta-training-simple-semi-supervised-text | 1901.07651 | null | https://arxiv.org/abs/1901.07651v3 | https://arxiv.org/pdf/1901.07651v3.pdf | Delta-training: Simple Semi-Supervised Text Classification using Pretrained Word Embeddings | We propose a novel and simple method for semi-supervised text classification. The method stems from the hypothesis that a classifier with pretrained word embeddings always outperforms the same classifier with randomly initialized word embeddings, as empirically observed in NLP tasks. Our method first builds two sets of classifiers as a form of model ensemble, and then initializes their word embeddings differently: one using random, the other using pretrained word embeddings. We focus on different predictions between the two classifiers on unlabeled data while following the self-training framework. We also use early-stopping in meta-epoch to improve the performance of our method. Our method, Delta-training, outperforms the self-training and the co-training framework in 4 different text classification datasets, showing robustness against error accumulation. | ['Ceyda Cinarel', 'Hwiyeol Jo'] | 2019-01-22 | delta-training-simple-semi-supervised-text-1 | https://aclanthology.org/D19-1347 | https://aclanthology.org/D19-1347.pdf | ijcnlp-2019-11 | ['semi-supervised-text-classification-1'] | ['natural-language-processing'] | [ 1.77418321e-01 1.41018599e-01 -4.43121433e-01 -5.52008390e-01
-3.49548161e-01 -5.99948823e-01 8.62832129e-01 6.15622759e-01
-9.07164156e-01 5.09431243e-01 3.39310795e-01 -5.16716421e-01
1.62829265e-01 -7.76136458e-01 -1.16294339e-01 -6.77349567e-01
2.01153174e-01 6.99531972e-01 3.05212826e-01 -1.20511189e-01
2.09838375e-01 4.08761464e-02 -1.61276841e+00 2.05071256e-01
9.26834524e-01 8.02808762e-01 -7.75992498e-02 7.74061561e-01
-5.65973461e-01 5.89454353e-01 -4.98577207e-01 -5.90021729e-01
1.55288696e-01 -1.66059092e-01 -8.88920188e-01 9.29215699e-02
1.92636833e-01 -6.48002625e-02 -1.64294273e-01 7.89314151e-01
4.35460180e-01 2.67304242e-01 8.94997954e-01 -1.19589138e+00
-7.88336933e-01 9.49787021e-01 -9.55092460e-02 -7.08684027e-02
1.96508229e-01 -1.56745136e-01 1.37867975e+00 -9.86702681e-01
4.35017020e-01 1.05539560e+00 8.33360970e-01 6.37464583e-01
-1.46712959e+00 -6.60947323e-01 3.17437351e-01 -1.92300335e-03
-1.32507050e+00 -4.30461228e-01 5.58330119e-01 -5.65283716e-01
9.85536933e-01 -1.52695710e-02 3.18552822e-01 1.47815621e+00
1.33742228e-01 6.57555223e-01 1.01460516e+00 -8.45196307e-01
4.64153051e-01 6.64640367e-01 1.05660915e+00 4.49767917e-01
5.16811430e-01 8.71944726e-02 -3.38739485e-01 -6.35064065e-01
9.00831306e-04 7.57101998e-02 -1.75503626e-01 -4.25172597e-01
-1.04780233e+00 1.05980134e+00 -5.59868552e-02 7.01010883e-01
-9.22918841e-02 -1.57723173e-01 4.70255762e-01 4.85034883e-01
8.87661934e-01 5.64983606e-01 -8.79246771e-01 -1.56317025e-01
-8.93294275e-01 2.66942084e-02 9.22419369e-01 5.99925578e-01
7.58692443e-01 -1.38359472e-01 -8.66055042e-02 1.11731279e+00
3.28479290e-01 -6.81993440e-02 1.42165673e+00 9.63626057e-02
2.75523335e-01 7.79997766e-01 5.72190210e-02 -5.19917667e-01
-3.99707586e-01 -4.39966381e-01 -5.25367022e-01 -4.85109491e-03
4.47147757e-01 -4.42383707e-01 -1.03867424e+00 1.43243682e+00
2.74776816e-01 1.92624778e-01 3.33526671e-01 4.10151869e-01
8.56193960e-01 6.24225497e-01 2.52062589e-01 -1.92980558e-01
1.21984124e+00 -1.23826671e+00 -8.12166512e-01 -2.96966046e-01
1.32858264e+00 -4.68642682e-01 8.30760181e-01 3.15804303e-01
-3.54898721e-01 -7.49632120e-01 -1.26342213e+00 1.11152850e-01
-8.45010936e-01 1.71714887e-01 3.82235974e-01 9.89875674e-01
-7.28771329e-01 6.92842841e-01 -6.23212814e-01 -3.94473642e-01
3.40827167e-01 3.56937349e-01 -5.69923520e-01 3.51204053e-02
-1.31537771e+00 1.15287817e+00 8.29795837e-01 -4.35976744e-01
-2.77590573e-01 -6.31417632e-01 -8.25293481e-01 4.92429398e-02
-3.54309194e-02 -2.22755924e-01 1.16284859e+00 -1.20793498e+00
-1.52310181e+00 9.82948422e-01 -3.34666550e-01 -4.74719703e-01
5.23415446e-01 -2.93055654e-01 -3.96816164e-01 -3.40844035e-01
-2.07424760e-01 5.32636762e-01 8.75531793e-01 -1.36697245e+00
-5.50146759e-01 -3.76545936e-01 -3.06513846e-01 1.04846507e-02
-1.03136647e+00 -1.98456675e-01 -1.09065481e-01 -6.18144929e-01
-7.22912028e-02 -7.50394106e-01 -3.28288943e-01 -2.52065301e-01
-3.41835946e-01 -7.64420748e-01 8.93987477e-01 -1.55337989e-01
1.52043021e+00 -2.01985788e+00 -1.36235747e-02 1.65861204e-01
1.99619889e-01 5.71706772e-01 -3.64909858e-01 5.92524350e-01
-4.42250788e-01 3.90352130e-01 -1.54255062e-01 -7.41458476e-01
-5.21948077e-02 4.55253094e-01 -4.52221602e-01 2.98249334e-01
2.62042254e-01 7.90687203e-01 -1.12305832e+00 -5.53126276e-01
2.47754641e-02 2.00546905e-01 -4.04768795e-01 4.04000968e-01
-1.50226936e-01 -1.31942406e-01 -3.27320546e-01 3.02942783e-01
5.49245656e-01 -1.64683625e-01 5.04821599e-01 6.26862794e-02
1.65924355e-02 2.36421958e-01 -1.27610540e+00 1.29400671e+00
-5.87086201e-01 5.76852381e-01 -5.84059179e-01 -1.25840008e+00
1.04922664e+00 4.01680678e-01 3.40658128e-01 -1.23330504e-01
4.24714983e-01 1.72324687e-01 1.62365437e-01 -4.56203103e-01
5.03339350e-01 -3.95077169e-02 2.87265271e-01 9.26061332e-01
7.14666784e-01 8.02271813e-02 2.16692224e-01 6.88365251e-02
1.11205459e+00 4.08228487e-02 5.57761312e-01 -4.23215657e-01
4.36180621e-01 -1.27102003e-01 1.71188056e-01 8.59422982e-01
-1.36139020e-01 4.00351673e-01 4.66822475e-01 -6.44204974e-01
-8.43737900e-01 -7.36314774e-01 -6.49064422e-01 1.60004234e+00
-7.93337002e-02 -7.15231478e-01 -4.84881222e-01 -1.46930051e+00
2.70743370e-01 1.01538646e+00 -1.11664426e+00 -2.23906979e-01
-1.45332813e-01 -1.03410375e+00 3.86169285e-01 4.83429015e-01
3.88693474e-02 -7.55322039e-01 -2.52285540e-01 3.21388721e-01
3.05843711e-01 -9.19587374e-01 -2.36541748e-01 9.35932398e-01
-8.90328169e-01 -1.10734355e+00 -4.70172077e-01 -9.36403155e-01
8.87736619e-01 1.63520768e-01 1.14018369e+00 3.73110473e-01
1.11136228e-01 3.22742581e-01 -9.28793669e-01 -5.02558053e-01
-6.79931104e-01 4.60141480e-01 3.40527117e-01 2.07598656e-01
7.67683148e-01 -2.86087155e-01 -1.00601263e-01 3.31314743e-01
-1.01124036e+00 -2.17193931e-01 2.00978577e-01 1.39942873e+00
1.83223054e-01 4.58100140e-02 4.85277683e-01 -1.27935874e+00
9.17648494e-01 -7.38045037e-01 -2.58278549e-01 3.61828506e-01
-1.20949292e+00 4.71419007e-01 7.88415611e-01 -9.33306038e-01
-6.72233880e-01 1.50744975e-01 -3.28397930e-01 -1.12153843e-01
-2.58593976e-01 4.32863802e-01 2.65966821e-02 1.67919308e-01
1.03275454e+00 1.97631583e-01 -5.00194430e-02 -4.09514725e-01
4.88619506e-01 1.24111152e+00 -7.89986253e-02 -5.83955109e-01
9.45588350e-01 2.02416629e-01 -6.20725036e-01 -8.71285975e-01
-1.06939912e+00 -6.15094841e-01 -1.02394259e+00 9.48786363e-02
5.73749781e-01 -5.95039785e-01 -4.10698913e-02 3.21856111e-01
-1.14738917e+00 -4.27685142e-01 -3.08658957e-01 3.20671469e-01
4.35901470e-02 4.52001601e-01 -1.78216740e-01 -8.56991231e-01
-2.73949146e-01 -9.69121575e-01 7.23714411e-01 6.40212651e-03
-5.68127990e-01 -1.55434930e+00 5.48079729e-01 -6.62771389e-02
2.70809412e-01 -2.43554130e-01 7.50763416e-01 -1.86691964e+00
5.52500367e-01 -5.60195148e-01 3.73568721e-02 6.53614163e-01
5.04156053e-01 1.79426864e-01 -1.28211558e+00 -2.63608873e-01
-2.55373240e-01 -5.84319592e-01 1.11648381e+00 -3.43345776e-02
1.12752509e+00 -1.57011881e-01 -4.92138743e-01 2.95594007e-01
1.33147752e+00 4.79860455e-02 4.36964363e-01 3.65675926e-01
5.03846407e-01 6.41533554e-01 2.22219437e-01 2.74517000e-01
1.91145837e-01 4.49744284e-01 9.29457024e-02 7.58089200e-02
4.11417931e-01 -3.25241446e-01 3.90364349e-01 7.59980798e-01
2.25859478e-01 -6.54122174e-01 -1.05957472e+00 5.35952091e-01
-1.89647532e+00 -7.69214034e-01 -1.57345667e-01 2.17885160e+00
1.00621068e+00 3.02513272e-01 8.42188150e-02 5.59800744e-01
5.94396830e-01 2.25575387e-01 -2.67721355e-01 -6.29780352e-01
3.22795771e-02 7.75746286e-01 4.75549489e-01 5.43105602e-01
-1.29303062e+00 1.02379203e+00 6.92516232e+00 6.89895213e-01
-9.19756591e-01 1.93848208e-01 4.78718698e-01 2.51484305e-01
-1.99747309e-01 -7.11735934e-02 -1.01677060e+00 5.01573563e-01
1.14674330e+00 -2.91219771e-01 -9.21948552e-02 9.10169899e-01
-1.97519675e-01 8.01174194e-02 -1.34989548e+00 6.68949068e-01
1.75284550e-01 -1.24507248e+00 2.33679458e-01 -9.78301018e-02
8.59939158e-01 1.19048454e-01 -2.13886902e-01 6.07429624e-01
7.82941222e-01 -1.03396499e+00 4.29639876e-01 -2.00538826e-03
4.96696770e-01 -5.13699889e-01 1.05143893e+00 4.50256646e-01
-9.06145573e-01 -2.13188663e-01 -2.68717021e-01 -1.17685392e-01
-2.24788174e-01 6.91849291e-01 -6.39262855e-01 3.98551077e-01
2.10158288e-01 7.20178604e-01 -8.72453749e-01 6.73837304e-01
-2.16781050e-01 9.98054385e-01 -1.66013822e-01 -4.42312419e-01
2.22042561e-01 -7.90922865e-02 1.08402416e-01 1.57847035e+00
-1.32231012e-01 -1.23203076e-01 2.55059540e-01 2.38414645e-01
-7.63014928e-02 3.63173515e-01 -5.71693599e-01 -2.15530843e-01
4.58805829e-01 1.25081551e+00 -6.79031491e-01 -6.15628600e-01
-4.63741153e-01 9.24371719e-01 5.80362022e-01 2.52903670e-01
-5.27858675e-01 -5.02260208e-01 5.58213651e-01 -7.60762915e-02
3.99070084e-01 -1.76655367e-01 -3.99235010e-01 -1.34253824e+00
-2.32003450e-01 -6.16812348e-01 4.43328708e-01 -2.49746397e-01
-1.54200029e+00 8.69434118e-01 -6.29147589e-02 -1.19503665e+00
-2.95412570e-01 -1.10076642e+00 -7.68694639e-01 6.11936390e-01
-1.41470492e+00 -8.44357669e-01 -2.05057099e-01 1.45863712e-01
6.90199733e-01 -3.30200136e-01 1.34545326e+00 -1.85661897e-01
-7.81906247e-01 9.49076295e-01 5.22084534e-01 4.79952335e-01
9.81953025e-01 -1.50022507e+00 3.43647301e-01 4.68923301e-01
5.34103096e-01 6.25610590e-01 6.48719907e-01 -3.19526136e-01
-9.11142051e-01 -1.19139016e+00 1.17360544e+00 -6.76650941e-01
1.00244403e+00 -6.14046574e-01 -1.20874655e+00 6.54348910e-01
3.34740490e-01 2.04067007e-01 1.22765672e+00 4.75823551e-01
-6.67775631e-01 1.69449911e-01 -1.04496837e+00 5.15951157e-01
6.75028920e-01 -3.03492993e-01 -1.04199910e+00 5.95735550e-01
6.31475627e-01 -1.26221001e-01 -6.54070854e-01 1.25896269e-02
6.93550825e-01 -6.81066930e-01 5.66859841e-01 -1.08868110e+00
5.46751320e-01 1.80776447e-01 -8.27565417e-02 -1.85196483e+00
-2.33126208e-01 -3.41632694e-01 -2.13345960e-01 1.26241934e+00
6.69002473e-01 -1.11611295e+00 6.04225874e-01 3.37891072e-01
2.75268495e-01 -8.28583896e-01 -8.21017206e-01 -9.55774486e-01
5.75150728e-01 -6.07294977e-01 3.22196364e-01 1.28016186e+00
2.78351784e-01 4.74628806e-01 8.36267173e-02 -6.06287606e-02
3.77922326e-01 -2.60011464e-01 7.79463232e-01 -1.67907393e+00
-8.06360245e-02 -5.26468933e-01 -4.86141413e-01 -1.01351774e+00
7.35837042e-01 -1.08599496e+00 4.53300551e-02 -1.22772372e+00
-4.42035310e-03 -4.20012474e-01 -5.61542869e-01 6.97067261e-01
-5.36256373e-01 1.80727482e-01 -2.05477744e-01 8.04950576e-03
-4.02618349e-01 4.88103598e-01 5.82929671e-01 -2.11607948e-01
-3.28625977e-01 -1.29715070e-01 -6.08505011e-01 5.57989120e-01
8.93115461e-01 -7.84499347e-01 -2.93581635e-01 -2.11149141e-01
5.71319237e-02 -7.62325525e-01 -8.14373717e-02 -8.22592497e-01
6.46765903e-02 -4.06681411e-02 3.54232937e-01 -1.51056066e-01
-1.64567396e-01 -7.32278407e-01 -5.23844540e-01 4.34643537e-01
-7.00546086e-01 1.86972581e-02 1.38749659e-01 4.79756892e-01
-1.01516552e-01 -8.91393900e-01 7.97863662e-01 1.88634321e-01
-5.26571155e-01 1.43706650e-01 -5.33442199e-01 1.19957663e-01
9.52909768e-01 -2.89940029e-01 -9.10691097e-02 6.98857009e-02
-7.63883889e-01 3.90300184e-01 3.56990576e-01 7.39045143e-01
3.88852060e-01 -1.38559234e+00 -7.18647122e-01 3.69174212e-01
5.75144887e-01 -2.55032390e-01 -3.22491825e-01 4.09505099e-01
-8.77876803e-02 1.43840149e-01 3.31929505e-01 -5.57374418e-01
-1.20244944e+00 4.25179213e-01 3.02412391e-01 -6.17793083e-01
-2.82737195e-01 6.99316323e-01 -2.71623194e-01 -8.42361867e-01
2.85915822e-01 -3.55850428e-01 -4.48544055e-01 4.40814018e-01
5.66601694e-01 -1.28765358e-02 2.38366678e-01 -4.32659298e-01
-3.01150918e-01 3.14268440e-01 -4.62021261e-01 -1.03064649e-01
1.21504831e+00 2.71185130e-01 2.41792157e-01 9.54212904e-01
1.46106279e+00 -2.94580832e-02 -7.21685290e-01 -4.65573311e-01
2.23963901e-01 -1.97930366e-01 4.36502278e-01 -6.27752602e-01
-6.08690441e-01 7.04454482e-01 5.97185016e-01 5.24016082e-01
5.16994119e-01 -2.81052232e-01 4.01467770e-01 6.37024343e-01
-1.83485635e-02 -1.26341450e+00 5.89998923e-02 7.37005055e-01
4.99983042e-01 -1.43932498e+00 1.05606645e-01 6.77945390e-02
-6.87521577e-01 1.55734146e+00 5.69974840e-01 -2.77298063e-01
1.07081509e+00 1.78925768e-01 2.51795679e-01 2.25833923e-01
-1.08514726e+00 -3.81811440e-01 3.57782155e-01 6.60057366e-01
7.74472117e-01 -1.53209940e-01 -3.81710142e-01 7.58868217e-01
-1.00420900e-01 -2.52711549e-02 3.27621549e-01 1.00473762e+00
-3.87065798e-01 -1.44997227e+00 -1.45568147e-01 7.21004605e-01
-1.97729230e-01 -1.60180390e-01 -5.94759583e-01 8.67849171e-01
2.32749000e-01 1.05997503e+00 3.34563881e-01 -7.58878648e-01
2.09491089e-01 7.55642235e-01 1.40448675e-01 -9.46640611e-01
-7.78188705e-01 -4.50930417e-01 7.00220913e-02 4.00765166e-02
-3.10462266e-01 -4.32367146e-01 -9.57665563e-01 1.36901001e-02
-8.63898516e-01 4.92848784e-01 5.21483958e-01 1.19700599e+00
1.09953597e-01 3.50570172e-01 1.01483536e+00 -7.30623186e-01
-8.11819077e-01 -1.45410788e+00 -3.59223455e-01 5.35119414e-01
3.65210325e-01 -8.29440296e-01 -8.91705811e-01 -1.89591423e-02] | [10.548768997192383, 7.865195274353027] |
a55adbd3-d6f9-479b-873d-a4b7d1817c53 | person-search-in-videos-with-one-portrait | 1807.10510 | null | http://arxiv.org/abs/1807.10510v1 | http://arxiv.org/pdf/1807.10510v1.pdf | Person Search in Videos with One Portrait Through Visual and Temporal Links | In real-world applications, e.g. law enforcement and video retrieval, one
often needs to search a certain person in long videos with just one portrait.
This is much more challenging than the conventional settings for person
re-identification, as the search may need to be carried out in the environments
different from where the portrait was taken. In this paper, we aim to tackle
this challenge and propose a novel framework, which takes into account the
identity invariance along a tracklet, thus allowing person identities to be
propagated via both the visual and the temporal links. We also develop a novel
scheme called Progressive Propagation via Competitive Consensus, which
significantly improves the reliability of the propagation process. To promote
the study of person search, we construct a large-scale benchmark, which
contains 127K manually annotated tracklets from 192 movies. Experiments show
that our approach remarkably outperforms mainstream person re-id methods,
raising the mAP from 42.16% to 62.27%. | ['Wentao Liu', 'Qingqiu Huang', 'Dahua Lin'] | 2018-07-27 | person-search-in-videos-with-one-portrait-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Qingqiu_Huang_Person_Search_in_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Qingqiu_Huang_Person_Search_in_ECCV_2018_paper.pdf | eccv-2018-9 | ['person-search'] | ['computer-vision'] | [ 1.87976778e-01 -5.73646188e-01 -2.43482143e-02 -1.95162266e-01
-4.90625352e-01 -7.92198122e-01 7.43618548e-01 1.78949401e-01
-5.86643040e-01 6.47914708e-01 1.96640790e-01 2.84178734e-01
-1.96436703e-01 -5.82286119e-01 -4.80541229e-01 -7.12884247e-01
5.12218587e-02 5.95368326e-01 3.24159771e-01 -7.15016052e-02
1.36949494e-01 2.88455993e-01 -1.45143139e+00 -6.32693097e-02
6.73805177e-01 7.93414772e-01 -4.30815034e-02 3.29795867e-01
2.92205244e-01 4.10325617e-01 -6.66932940e-01 -8.57575238e-01
2.51922518e-01 -2.90023327e-01 -8.21389198e-01 2.11697906e-01
7.57529736e-01 -3.22348177e-01 -4.05059218e-01 1.33774972e+00
6.16299510e-01 4.16996092e-01 3.78947973e-01 -1.21740282e+00
-5.31380296e-01 4.44067419e-01 -8.85958612e-01 4.13427472e-01
7.97034979e-01 -1.08706154e-01 8.90696883e-01 -6.02156758e-01
7.98401356e-01 1.15409815e+00 7.67725170e-01 6.39208257e-01
-1.07200289e+00 -8.42908025e-01 4.14586276e-01 5.46600521e-01
-1.77414405e+00 -4.71572846e-01 5.79319775e-01 -4.47375476e-01
1.85856685e-01 3.72074217e-01 7.01233208e-01 1.20614052e+00
-3.58552903e-01 5.85451663e-01 9.52328920e-01 -2.05594257e-01
-1.83754772e-01 1.67555869e-01 2.31342822e-01 4.27867025e-01
3.58506292e-01 -1.90356240e-01 -6.54722691e-01 -2.95580059e-01
6.91239357e-01 9.93706957e-02 -4.98913080e-01 -2.78520495e-01
-1.44421709e+00 3.92019629e-01 3.80287260e-01 3.77661496e-01
-2.71383435e-01 -1.01455986e-01 4.69650358e-01 1.87439367e-01
3.30326617e-01 2.68043846e-01 1.71068400e-01 -5.52582070e-02
-1.00775385e+00 4.45278645e-01 5.82418621e-01 9.37635064e-01
4.69726712e-01 -6.66090727e-01 -3.84386271e-01 8.72561574e-01
4.17928882e-02 4.22262937e-01 2.18349352e-01 -7.89207041e-01
3.97076249e-01 4.02750283e-01 4.37119007e-01 -1.47320235e+00
-1.84546679e-01 -5.07580876e-01 -1.07537997e+00 -3.14089358e-01
7.28439033e-01 7.13868365e-02 -5.52968979e-01 1.77329898e+00
6.52782738e-01 5.14707983e-01 -2.55774289e-01 1.23393333e+00
7.10237622e-01 4.85974014e-01 -2.35648956e-02 -2.42669046e-01
1.57255030e+00 -1.01050675e+00 -6.74620390e-01 8.28879252e-02
2.56340921e-01 -7.18738735e-01 5.54720342e-01 2.86469311e-01
-1.02022922e+00 -6.19879723e-01 -7.14213848e-01 1.79973200e-01
-1.53329223e-01 6.74994066e-02 1.94931433e-01 4.15557712e-01
-1.00732684e+00 4.96863395e-01 -3.98728520e-01 -7.55761385e-01
2.07606420e-01 3.70715559e-01 -6.51779413e-01 -2.39115581e-01
-1.21822858e+00 5.63208580e-01 3.57645452e-01 3.29073489e-01
-5.02477050e-01 -5.04691422e-01 -5.57087123e-01 4.66082897e-03
6.06646538e-01 -7.35512853e-01 9.92466211e-01 -9.88707423e-01
-1.02577734e+00 9.62573290e-01 -3.32410842e-01 -2.93479145e-01
8.52668822e-01 -2.65048981e-01 -5.48294783e-01 2.11435556e-01
4.26156729e-01 3.00564349e-01 7.18330801e-01 -1.21674669e+00
-8.29509556e-01 -4.28351492e-01 2.77131915e-01 3.18541616e-01
-5.87925851e-01 3.26699227e-01 -1.17039835e+00 -7.73154676e-01
-3.92529219e-02 -1.26549125e+00 -1.45256426e-02 5.52977286e-02
-3.56605768e-01 -4.30862188e-01 5.36782146e-01 -8.60074878e-01
1.37632334e+00 -2.09867430e+00 3.96904707e-01 5.04935205e-01
3.39299679e-01 3.68225813e-01 1.64321750e-01 5.39133370e-01
2.17587575e-01 -1.52149588e-01 1.08053774e-01 -6.12804949e-01
-1.24332137e-01 -1.33194551e-01 -5.82105145e-02 6.41017973e-01
-3.47503752e-01 6.46111071e-01 -1.10707450e+00 -6.77564025e-01
-2.10995421e-01 4.31991369e-01 -2.41617382e-01 -1.18918747e-01
2.90819377e-01 6.36093318e-01 -5.18260241e-01 5.50946593e-01
6.67086363e-01 -5.02177179e-01 2.49273986e-01 -2.93928355e-01
-9.92187113e-02 -2.38403827e-01 -1.39018285e+00 1.81902480e+00
1.64185584e-01 5.17506123e-01 -3.95713970e-02 -6.77024007e-01
5.78037977e-01 2.44576558e-01 5.89175820e-01 -5.94403148e-01
-1.25374705e-01 1.24813337e-03 -3.00968707e-01 -3.01357806e-01
6.63716555e-01 2.49043956e-01 -1.43660396e-01 4.30618584e-01
-3.00376177e-01 7.77620256e-01 4.61913317e-01 3.29766959e-01
9.42696929e-01 -3.73725872e-03 4.52967063e-02 -1.50924727e-01
8.62746358e-01 -8.27771425e-03 6.85798943e-01 8.10562313e-01
-3.74206543e-01 5.83359957e-01 9.33153033e-02 -5.32322884e-01
-8.28359485e-01 -8.37873161e-01 3.83364968e-02 8.79515648e-01
8.00248504e-01 -6.38876021e-01 -8.32004964e-01 -6.04390502e-01
-9.36451927e-02 -5.85436933e-02 -6.34342015e-01 -7.70305842e-02
-7.37342238e-01 -5.88094115e-01 5.91030717e-01 2.23456740e-01
9.32389081e-01 -6.64356947e-01 -6.17385544e-02 8.15181211e-02
-6.89940453e-01 -1.30324876e+00 -1.10563564e+00 -9.58602726e-01
-3.19370538e-01 -1.10792482e+00 -1.23474622e+00 -8.35327983e-01
7.13669419e-01 6.07205153e-01 8.09866130e-01 3.84648085e-01
-1.38601512e-01 4.99321580e-01 -4.98187006e-01 1.42310858e-01
5.30035831e-02 2.76950032e-01 4.04477835e-01 6.08257771e-01
3.97381216e-01 -3.91922802e-01 -7.17056751e-01 6.92234755e-01
-5.80684245e-01 -5.57627622e-03 2.19030008e-01 8.06265712e-01
4.60912555e-01 3.78860384e-01 4.07578111e-01 -6.91061914e-01
5.04497588e-01 -4.19475287e-01 -4.03864413e-01 6.55410528e-01
-5.21497548e-01 -3.90149415e-01 3.50177169e-01 -6.49779916e-01
-1.01622653e+00 6.10805079e-02 1.27803952e-01 -3.56956095e-01
-7.80135468e-02 1.61718667e-01 4.03400362e-02 -3.93379599e-01
4.01743680e-01 3.69292349e-01 -2.26423979e-01 -5.95039785e-01
2.64618155e-02 5.85786760e-01 7.92488098e-01 -3.91607076e-01
1.13335371e+00 6.21215045e-01 -2.06951633e-01 -6.85047388e-01
-6.81161702e-01 -8.73277962e-01 -6.68590486e-01 -5.73844314e-01
7.42695093e-01 -9.76802588e-01 -1.04728568e+00 6.04821742e-01
-1.19420433e+00 1.31434724e-01 2.40185246e-01 3.72987092e-01
-7.74993300e-02 7.45292485e-01 -6.62720680e-01 -6.46043479e-01
-3.62960726e-01 -8.25295925e-01 9.61277485e-01 4.60765064e-01
-3.06232095e-01 -7.79599190e-01 1.96967393e-01 4.57819462e-01
1.94057822e-01 2.23187685e-01 2.56065607e-01 -5.70669472e-01
-5.17400920e-01 -2.92904407e-01 -3.73616040e-01 -1.86900318e-01
2.32460648e-01 -2.76799947e-01 -8.28943670e-01 -6.57559514e-01
-4.39285427e-01 4.75296602e-02 7.09864259e-01 -7.52086937e-02
8.37526441e-01 -2.16576591e-01 -6.93903148e-01 4.65080410e-01
1.05702806e+00 3.26204598e-02 4.42147702e-01 4.87799078e-01
8.64102423e-01 7.89464831e-01 5.76700568e-01 4.95263457e-01
5.93891680e-01 1.27270508e+00 -4.47647870e-02 7.90058225e-02
-8.52805749e-02 -3.27002347e-01 5.42347468e-02 6.81054592e-01
-5.51979601e-01 -3.19133461e-01 -7.35291123e-01 5.07689953e-01
-2.18675327e+00 -1.31465304e+00 9.45690181e-03 2.31867862e+00
6.83635175e-01 -1.32861257e-01 4.49475944e-01 1.17005564e-01
1.22837615e+00 4.94667664e-02 -3.04963917e-01 5.72138071e-01
-1.27402425e-01 -4.11588371e-01 3.80288959e-01 3.86727303e-01
-1.16771090e+00 7.92634130e-01 5.97913980e+00 8.84708345e-01
-8.23829174e-01 9.75122228e-02 2.95487821e-01 -8.39938447e-02
1.25029296e-01 -2.21440434e-01 -1.03570497e+00 7.43670642e-01
4.23272431e-01 -2.07895935e-01 4.57222909e-01 3.98566723e-01
8.39724466e-02 -1.05847262e-01 -1.04959190e+00 1.41421187e+00
3.12414080e-01 -1.09923542e+00 -5.81952296e-02 1.12562880e-01
6.30356133e-01 -5.49470901e-01 -1.03298880e-01 -4.42515202e-02
-1.12189762e-02 -7.05596924e-01 6.51280403e-01 6.70119464e-01
7.00972736e-01 -7.06725121e-01 5.40305793e-01 2.61588573e-01
-1.46339977e+00 2.91636772e-02 -1.19840294e-01 1.42372504e-01
4.77533370e-01 3.02572131e-01 -4.77033496e-01 7.77745485e-01
1.07505512e+00 9.97082293e-01 -5.89452088e-01 1.44084382e+00
1.01290710e-01 1.69569522e-01 -2.83118010e-01 1.35268867e-01
-2.97429562e-02 -2.02307716e-01 6.13747597e-01 1.18589902e+00
3.66027594e-01 1.48313984e-01 2.88570851e-01 4.19888020e-01
-6.26529083e-02 1.45430520e-01 -4.15626228e-01 2.56226510e-01
4.94329870e-01 1.30238450e+00 -7.43434608e-01 -3.56380492e-01
-2.16564059e-01 1.51337576e+00 3.14342231e-01 4.49023962e-01
-9.46977258e-01 -1.97221130e-01 6.61539137e-01 2.17288271e-01
7.00617880e-02 -2.07126975e-01 4.40463781e-01 -1.29896569e+00
3.41899574e-01 -8.16383541e-01 6.55478239e-01 -6.41203821e-01
-1.40202188e+00 7.78410137e-01 1.28564760e-01 -1.30376422e+00
-1.43524036e-01 -1.57939658e-01 -3.55553389e-01 7.68524587e-01
-1.36129296e+00 -1.25377369e+00 -6.07252479e-01 8.32794368e-01
4.10511643e-01 -1.19805217e-01 5.02022326e-01 8.85987759e-01
-8.37619066e-01 8.95178497e-01 6.14329353e-02 3.26808006e-01
1.12104785e+00 -7.88052619e-01 3.78568292e-01 9.93046224e-01
1.73782364e-01 8.17262709e-01 7.21201122e-01 -7.18109548e-01
-1.24529052e+00 -9.06092584e-01 1.08347464e+00 -4.59846258e-01
5.90649247e-01 -4.07261401e-01 -8.63621175e-01 5.10302842e-01
7.56718591e-02 -1.62168190e-01 5.78661680e-01 6.98431656e-02
-3.39675039e-01 -2.97338188e-01 -9.25540686e-01 6.93136930e-01
1.51596785e+00 -4.85501438e-01 -1.56161964e-01 4.15077686e-01
4.03538913e-01 -3.00999194e-01 -8.40965986e-01 2.31528550e-01
7.73194194e-01 -7.33951092e-01 1.19841218e+00 -3.15095156e-01
-2.08421528e-01 -6.37472272e-01 1.27340809e-01 -9.97260094e-01
-6.59296513e-01 -8.65686059e-01 3.45542878e-02 1.65083122e+00
1.01280035e-02 -6.22509718e-01 7.09206104e-01 6.75923586e-01
3.71640742e-01 -2.58012444e-01 -8.85679185e-01 -8.31833720e-01
-5.34794569e-01 7.04596192e-02 7.13373840e-01 9.06835735e-01
-8.98154899e-02 1.63739026e-01 -9.22679186e-01 3.71800274e-01
8.03239524e-01 4.99367826e-02 8.02671909e-01 -1.40600646e+00
-2.86110282e-01 -3.22496533e-01 -5.02396703e-01 -1.23990822e+00
1.21532418e-01 -5.54004014e-01 -6.78738803e-02 -1.25507641e+00
6.50365353e-01 -5.88637292e-01 -2.49891430e-01 2.10927069e-01
-2.57778138e-01 6.07625604e-01 2.58639157e-01 7.56843507e-01
-1.03811896e+00 3.03854406e-01 1.14598441e+00 -2.11956471e-01
-1.12745836e-02 2.43721634e-01 -6.27019286e-01 5.31801701e-01
5.50131798e-01 -2.93885529e-01 -2.61206031e-01 -5.49067378e-01
2.17769668e-01 -1.81126416e-01 6.24373078e-01 -9.99468148e-01
8.08106065e-01 -3.02832872e-02 3.83486778e-01 -3.96266907e-01
5.66632569e-01 -8.60910475e-01 6.16487801e-01 2.90815651e-01
-2.39386618e-01 3.75166088e-01 -1.34095490e-01 9.25636590e-01
-2.41782397e-01 -4.59692217e-02 4.04065162e-01 -6.90737516e-02
-8.42292964e-01 5.33465326e-01 3.41335908e-02 -1.61753640e-01
1.15340996e+00 -3.18243057e-01 -3.66092175e-01 -4.77705538e-01
-7.03711033e-01 4.93186653e-01 6.60241127e-01 4.11661297e-01
4.07149851e-01 -1.64108312e+00 -7.46782899e-01 -6.60112649e-02
3.00415546e-01 -3.73866171e-01 5.32193005e-01 9.47106898e-01
-2.51967132e-01 2.24917516e-01 -9.17903334e-02 -4.85603839e-01
-1.62337339e+00 6.76241755e-01 2.39182353e-01 -1.38299629e-01
-7.44327784e-01 7.03276634e-01 2.00544506e-01 3.19396034e-02
4.34701800e-01 3.67916524e-01 -5.47250032e-01 2.45265886e-01
9.56991851e-01 4.65096444e-01 -2.36678615e-01 -1.17842865e+00
-4.72580045e-01 9.74911928e-01 -2.27295771e-01 -7.75623918e-02
8.53035748e-01 -4.57564443e-01 -1.67816788e-01 1.82311580e-01
9.45565104e-01 1.22291543e-01 -1.03653800e+00 -6.48368955e-01
5.08897007e-02 -7.49381244e-01 -4.55242693e-01 -4.72324669e-01
-1.04198909e+00 2.98122823e-01 5.18588543e-01 2.47879356e-01
1.08864784e+00 -1.35844052e-01 1.03129578e+00 3.20747584e-01
7.46340990e-01 -1.07962966e+00 -8.68403018e-02 1.69772208e-01
6.00111067e-01 -1.24117136e+00 4.80539128e-02 -5.69808185e-01
-4.72918600e-01 8.71864855e-01 4.06077534e-01 -3.91968600e-02
4.36629474e-01 -3.33613843e-01 -1.81595787e-01 -1.27753079e-01
-3.30447227e-01 -1.54911876e-01 5.12961864e-01 3.76386762e-01
4.71763313e-02 -7.83976689e-02 -3.05783629e-01 4.45138544e-01
2.54696291e-02 2.82245502e-02 1.65099859e-01 7.01426625e-01
-1.44864053e-01 -1.21778989e+00 -5.18468916e-01 9.69656929e-02
-5.23614824e-01 4.56588753e-02 -4.46383446e-01 5.09368658e-01
2.21732572e-01 9.33047175e-01 -7.15685412e-02 -4.23084646e-01
2.90435106e-01 -2.07227275e-01 4.05605227e-01 -3.27411413e-01
-6.45440996e-01 -3.52105759e-02 1.31204009e-01 -4.90427136e-01
-8.20112169e-01 -9.18269694e-01 -6.86144292e-01 -6.68446779e-01
-3.19659829e-01 2.82629609e-01 1.76210463e-01 8.32775950e-01
3.43537003e-01 1.00558572e-01 4.94647056e-01 -7.42217839e-01
-2.04545617e-01 -6.10019386e-01 -5.10712147e-01 7.13552058e-01
3.29173446e-01 -7.73227632e-01 -2.11597875e-01 2.76368886e-01] | [14.768012046813965, 1.0342285633087158] |
569f42b8-dff7-4beb-bb60-4a648584749c | baam-monocular-3d-pose-and-shape | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Lee_BAAM_Monocular_3D_Pose_and_Shape_Reconstruction_With_Bi-Contextual_Attention_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Lee_BAAM_Monocular_3D_Pose_and_Shape_Reconstruction_With_Bi-Contextual_Attention_CVPR_2023_paper.pdf | BAAM: Monocular 3D Pose and Shape Reconstruction With Bi-Contextual Attention Module and Attention-Guided Modeling | 3D traffic scene comprises various 3D information about car objects, including their pose and shape. However, most recent studies pay relatively less attention to reconstructing detailed shapes. Furthermore, most of them treat each 3D object as an independent one, resulting in losses of relative context inter-objects and scene context reflecting road circumstances. A novel monocular 3D pose and shape reconstruction algorithm, based on bi-contextual attention and attention-guided modeling (BAAM), is proposed in this work. First, given 2D primitives, we reconstruct 3D object shape based on attention-guided modeling that considers the relevance between detected objects and vehicle shape priors. Next, we estimate 3D object pose through bi-contextual attention, which leverages relation-context inter objects and scene-context between an object and road environment. Finally, we propose a 3D non maximum suppression algorithm to eliminate spurious objects based on their Bird-Eye-View distance. Extensive experiments demonstrate that the proposed BAAM yields state-of-the-art performance on ApolloCar3D. Also, they show that the proposed BAAM can be plugged into any mature monocular 3D object detector on KITTI and significantly boost their performance. | ['Yeong Jun Koh', 'Seong-Gyun Jeong', 'Su-Min Choi', 'HanUl Kim', 'Hyo-Jun Lee'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['3d-car-instance-understanding'] | ['computer-vision'] | [-5.90676628e-02 -6.43501803e-02 1.21285655e-01 -4.33785200e-01
-7.22987354e-01 -3.26798856e-01 7.32741058e-01 -2.93174863e-01
-7.57654458e-02 -8.61650240e-03 7.88426921e-02 -2.28651732e-01
1.38454944e-01 -6.22218549e-01 -9.35194790e-01 -5.15608132e-01
6.65418208e-01 7.51549006e-01 8.23776186e-01 -9.43515673e-02
2.30255827e-01 9.87744868e-01 -1.73656225e+00 1.87616237e-02
5.83898544e-01 9.08498108e-01 7.09077060e-01 5.14664233e-01
4.51295897e-02 5.04530251e-01 -2.96124011e-01 -3.82859409e-01
4.60976541e-01 9.46609899e-02 -2.11699367e-01 5.49308956e-01
8.48959684e-01 -5.54188013e-01 -6.15051329e-01 9.50819790e-01
2.54950851e-01 1.77103370e-01 8.78476202e-01 -1.16467619e+00
-6.53970420e-01 -3.62796009e-01 -8.56904805e-01 2.22808510e-01
6.10459633e-02 3.89377624e-01 7.75147915e-01 -1.55458188e+00
5.03786922e-01 1.73541439e+00 4.40013230e-01 2.96962023e-01
-1.09881783e+00 -7.17784524e-01 5.81830442e-01 6.44598782e-01
-1.37753797e+00 -4.60648417e-01 1.20580781e+00 -4.70131457e-01
1.01198041e+00 2.61976898e-01 5.63201666e-01 6.91930771e-01
8.08978453e-02 8.62595141e-01 7.79919326e-01 -3.23680520e-01
-8.36566240e-02 2.72683769e-01 1.87164724e-01 4.75452363e-01
4.41542268e-01 4.43258166e-01 -3.94747704e-01 1.13254838e-01
5.56443214e-01 6.35089502e-02 1.98149160e-01 -7.37030089e-01
-8.90652597e-01 5.21020293e-01 4.16733593e-01 -4.15886343e-01
-2.91532189e-01 -1.00038955e-02 -6.25605062e-02 -3.49889874e-01
7.47765362e-01 -2.38196954e-01 -2.06530049e-01 4.93153423e-01
-2.80754328e-01 5.65869749e-01 1.26181170e-01 1.43252158e+00
9.25012350e-01 -4.81262105e-03 -1.69444427e-01 7.86930919e-01
7.67979562e-01 1.06962788e+00 -4.80947584e-01 -9.57906842e-01
4.55415994e-01 6.64725065e-01 1.28969982e-01 -1.10862708e+00
-3.61687809e-01 -3.21599662e-01 -3.75376552e-01 4.79691446e-01
1.47459209e-01 3.42255831e-01 -9.31629419e-01 1.43863773e+00
9.39434290e-01 2.99504966e-01 -2.70571649e-01 1.24461818e+00
1.03634381e+00 4.47690248e-01 -1.04633890e-01 8.23102221e-02
1.38342464e+00 -8.78169954e-01 -5.09069145e-01 -6.50576949e-01
7.80951977e-02 -1.05955935e+00 6.33761227e-01 6.60915151e-02
-1.04787719e+00 -8.47029328e-01 -8.37294698e-01 -4.85768378e-01
-1.73568502e-01 1.91977665e-01 2.20306113e-01 6.32585108e-01
-5.62891006e-01 -2.44166657e-01 -5.10576904e-01 -1.88194826e-01
5.90335608e-01 9.97827873e-02 -2.04001322e-01 -5.08749723e-01
-6.07878685e-01 1.24872649e+00 2.37420037e-01 2.10521698e-01
-1.08315992e+00 -8.27146292e-01 -1.00633657e+00 -1.95115119e-01
7.84834921e-01 -6.84951246e-01 1.14212084e+00 -1.24488726e-01
-1.26286006e+00 1.05874276e+00 -5.89405060e-01 -3.32932949e-01
2.37781093e-01 -4.21927094e-01 -3.66036028e-01 1.61010753e-02
1.88462779e-01 7.52384722e-01 9.08887565e-01 -1.73950231e+00
-8.83953512e-01 -6.80505276e-01 -1.52631640e-01 4.57479239e-01
4.99525785e-01 -1.46192303e-02 -9.62488830e-01 -3.80813569e-01
4.73979294e-01 -8.68786871e-01 -2.37246066e-01 3.01120311e-01
-3.66426587e-01 -2.47267157e-01 1.33310425e+00 -5.53346694e-01
6.11583114e-01 -2.16524243e+00 -1.16020061e-01 -5.84849827e-02
2.37572342e-01 3.37881446e-01 -9.29328129e-02 -1.02579175e-02
2.23706082e-01 -3.66997868e-01 -1.50603652e-01 -6.44205749e-01
9.50961187e-02 3.49187732e-01 -4.93787169e-01 7.28888869e-01
6.26572311e-01 1.07001531e+00 -6.59802377e-01 -3.78750890e-01
8.59295785e-01 7.32087255e-01 -7.60162175e-01 1.29439592e-01
-2.60442793e-01 3.70600998e-01 -5.62426209e-01 7.73519933e-01
1.41339076e+00 1.16381273e-01 -2.33268574e-01 -3.63342375e-01
-2.91062862e-01 1.11620344e-01 -9.75640416e-01 1.22369075e+00
-3.62689316e-01 6.40984774e-01 1.94219172e-01 -7.88620353e-01
1.07186007e+00 -1.26033753e-01 2.23040611e-01 -8.01261187e-01
1.57577857e-01 -3.83284129e-03 -2.07357451e-01 -3.33585143e-01
7.48050570e-01 3.53696421e-02 1.10013582e-01 2.69782156e-01
-2.57509679e-01 -6.82910979e-01 -2.24450573e-01 1.65361866e-01
5.08556068e-01 2.70749122e-01 1.90749466e-01 -6.30471706e-02
6.37676239e-01 -2.73402911e-02 5.09094536e-01 7.22466707e-01
-3.20314109e-01 7.63855457e-01 4.83841635e-02 -3.84813577e-01
-1.20312154e+00 -1.14102340e+00 -3.20729941e-01 7.69031763e-01
8.22315812e-01 9.09160674e-02 -3.57605487e-01 -6.43989503e-01
2.74866730e-01 1.05714464e+00 -4.50598955e-01 -1.85371801e-01
-8.06262970e-01 -4.94476497e-01 -1.38530508e-01 4.77505505e-01
3.30254078e-01 -7.60268748e-01 -6.97535634e-01 3.92819308e-02
-4.54166830e-02 -1.49086022e+00 -6.80357873e-01 -2.50426918e-01
-6.83921993e-01 -1.15693831e+00 -2.94858575e-01 -6.65223181e-01
7.18205690e-01 1.23392308e+00 1.11818397e+00 -1.53657153e-01
-3.00537318e-01 5.90877712e-01 -2.21288472e-01 -6.85125470e-01
-2.84377098e-01 -5.23124456e-01 2.22867616e-02 3.82518947e-01
6.71566844e-01 -3.42824697e-01 -5.26190579e-01 7.11143374e-01
-2.31408998e-01 2.72398412e-01 8.08393657e-01 4.66963589e-01
9.35535967e-01 -1.48344532e-01 3.18673164e-01 -5.56436241e-01
-2.55214363e-01 -4.41311032e-01 -8.61736476e-01 -1.41310096e-01
-1.54711589e-01 -2.37606972e-01 5.21776490e-02 -2.20349178e-01
-1.55900204e+00 2.73047119e-01 -2.24686065e-03 -8.27063918e-01
-3.80599648e-01 -3.72303784e-01 -6.76484883e-01 -5.43777868e-02
3.71294796e-01 3.12245399e-01 -4.53388840e-02 -6.52034581e-01
4.83169734e-01 5.23788750e-01 6.16436839e-01 -4.60728258e-01
1.03994310e+00 1.00426102e+00 1.87807277e-01 -1.03009558e+00
-1.10203433e+00 -7.03832567e-01 -6.51611269e-01 -4.39583778e-01
9.20930207e-01 -1.06124020e+00 -7.09768355e-01 4.29003596e-01
-1.39111626e+00 6.47187009e-02 -1.49989381e-01 5.87505579e-01
-6.01093531e-01 4.26002532e-01 -1.71871223e-02 -1.09523582e+00
7.04005808e-02 -1.23018503e+00 1.51926160e+00 -9.69185755e-02
3.36255014e-01 -4.52938735e-01 -4.15418923e-01 7.76687264e-01
1.34154946e-01 -8.59334171e-02 7.20424354e-01 -2.32369214e-01
-1.32779038e+00 -1.96264759e-01 -6.75200522e-01 2.52430916e-01
-7.03364015e-02 -3.35386962e-01 -1.11404121e+00 2.43954062e-02
1.02453940e-01 1.17958575e-01 8.75787020e-01 6.52260542e-01
9.24961507e-01 1.75466508e-01 -5.61082602e-01 4.22577530e-01
9.40662146e-01 1.99889228e-01 3.35812688e-01 -1.34428635e-01
1.01327789e+00 8.08574200e-01 9.93526578e-01 3.07641476e-01
6.23841882e-01 1.13286114e+00 8.46111715e-01 7.11712614e-02
-5.97586095e-01 -3.12733710e-01 2.01244250e-01 6.10020459e-01
6.72047064e-02 -2.59763096e-02 -7.90120304e-01 5.30084372e-01
-1.67470133e+00 -9.36808825e-01 -5.19418299e-01 2.11240029e+00
1.19228415e-01 4.09486830e-01 8.20615515e-02 -2.25592539e-01
8.21409941e-01 -5.49606886e-03 -9.06582117e-01 1.42025232e-01
-2.74574250e-01 -3.08932543e-01 5.80056190e-01 7.30812252e-01
-9.30007517e-01 9.61545527e-01 5.69233370e+00 9.52816844e-01
-6.32968307e-01 1.77405566e-01 4.66393262e-01 -2.65039671e-02
-4.79438633e-01 2.94804536e-02 -1.40728712e+00 1.53575182e-01
2.99380273e-01 -3.55104031e-03 1.16985962e-01 8.94859374e-01
3.75465900e-01 -2.42804438e-01 -9.56743717e-01 1.03707266e+00
3.71535003e-01 -1.08880019e+00 1.54860646e-01 2.57995844e-01
6.18003428e-01 1.24874681e-01 1.72536179e-01 3.07007492e-01
1.56273723e-01 -6.29360318e-01 1.33705032e+00 5.71295798e-01
5.24470627e-01 -9.19000030e-01 3.20999622e-01 6.31496310e-01
-1.44121337e+00 -3.16406712e-02 -4.82752144e-01 5.12384586e-02
4.31763798e-01 5.83265185e-01 -9.66072023e-01 6.02591693e-01
6.51144207e-01 7.00922430e-01 -5.06639957e-01 1.01311219e+00
-9.69319716e-02 3.31229419e-01 -3.21337193e-01 1.73360839e-01
3.84397320e-02 -1.43190607e-01 1.14979994e+00 9.11154747e-01
1.71768188e-01 2.69116640e-01 3.31667900e-01 1.17241991e+00
2.88281381e-01 -2.04646364e-01 -6.84295058e-01 7.35661030e-01
5.61879456e-01 9.30189610e-01 -5.47846496e-01 -3.01070064e-01
-7.41356909e-01 3.74413937e-01 1.83840066e-01 4.46943372e-01
-1.00657594e+00 1.49549276e-01 9.00615811e-01 4.29704249e-01
7.31477797e-01 -3.69090199e-01 -5.22597849e-01 -8.57394338e-01
2.61339724e-01 -3.64118487e-01 -8.08412433e-02 -1.02760911e+00
-1.34426773e+00 3.11692595e-01 3.32297504e-01 -1.39815700e+00
2.62610376e-01 -7.35668659e-01 -2.12382719e-01 8.33799124e-01
-1.74166512e+00 -1.42539787e+00 -2.77431309e-01 3.91538084e-01
1.01244736e+00 -5.31248655e-03 1.09439164e-01 3.72543961e-01
-2.98468828e-01 1.90744027e-01 -3.44814867e-01 -3.33151251e-01
3.65706414e-01 -7.46106625e-01 6.84583724e-01 8.66958201e-01
1.29616156e-01 1.46574900e-01 6.23205304e-01 -8.46497059e-01
-1.53214812e+00 -1.44976735e+00 8.69614840e-01 -9.85490382e-01
1.37684718e-01 -6.47943974e-01 -8.37462187e-01 6.94833636e-01
-2.25068480e-01 3.26668561e-01 2.44471207e-02 -1.38393760e-01
-3.58121276e-01 -1.54686257e-01 -9.29862618e-01 6.78552866e-01
1.46751225e+00 -4.70168680e-01 -7.10308850e-01 3.33954096e-01
8.76581132e-01 -7.25122213e-01 -1.36694327e-01 6.55458689e-01
4.10907775e-01 -9.76491392e-01 1.41622889e+00 -2.97679275e-01
-3.02099194e-02 -7.61027634e-01 -4.74063367e-01 -7.89227009e-01
-5.13692260e-01 -2.99841940e-01 -4.51343089e-01 8.94523025e-01
1.89935844e-02 -3.67670089e-01 7.03400552e-01 3.90954673e-01
-7.74072468e-01 -4.97066200e-01 -1.12053847e+00 -7.24621236e-01
-1.55802026e-01 -1.00338376e+00 6.63390696e-01 4.35216159e-01
-8.19020092e-01 5.49454629e-01 -4.88009512e-01 6.24939859e-01
1.01694596e+00 4.28252727e-01 1.18747389e+00 -1.31540716e+00
-2.57269051e-02 -3.66675943e-01 -5.06590247e-01 -1.67857444e+00
2.51892209e-01 -8.12958717e-01 2.34981298e-01 -1.14025235e+00
3.29283088e-01 -4.30145055e-01 1.72628760e-01 -9.32667851e-02
-2.43607327e-01 3.62553746e-01 2.61162043e-01 5.37503026e-02
-6.59625709e-01 8.63672137e-01 1.52910137e+00 -7.19111413e-02
5.38924290e-03 1.29842266e-01 -6.03697896e-01 8.81671369e-01
3.30748379e-01 -4.52031821e-01 -3.57065767e-01 -5.64641893e-01
-1.87422797e-01 -1.14345022e-01 9.47975755e-01 -8.28822732e-01
2.08094507e-01 -2.67924964e-01 3.60501587e-01 -1.79700017e+00
7.81233370e-01 -1.10662341e+00 3.76997739e-02 9.92450044e-02
2.35989124e-01 -3.12628359e-01 3.18627417e-01 1.04799652e+00
9.20006111e-02 1.35006100e-01 8.55505466e-01 1.04179876e-02
-9.09580529e-01 4.81273949e-01 -2.70779252e-01 1.31711081e-01
1.25307691e+00 -5.04476488e-01 -1.73042212e-02 -3.43279429e-02
-5.71953177e-01 3.41538996e-01 2.69636601e-01 6.26871586e-01
8.47077787e-01 -1.57389939e+00 -9.29563642e-01 5.91917276e-01
3.38088185e-01 4.39291507e-01 6.43369317e-01 6.98189378e-01
-1.63391888e-01 7.29630589e-01 8.05172697e-02 -1.19414485e+00
-1.36873162e+00 7.86960900e-01 2.28234768e-01 3.54824513e-01
-9.36811864e-01 7.20705986e-01 9.41065669e-01 -5.87238312e-01
2.30695624e-02 -3.71037513e-01 -5.45310415e-02 -2.29733691e-01
4.63022828e-01 3.44192386e-01 7.06709400e-02 -1.23176169e+00
-4.30115193e-01 1.07102275e+00 -1.93896875e-01 1.45674855e-01
1.18748689e+00 -5.62565804e-01 3.94091040e-01 2.20306367e-01
8.22215259e-01 3.08726076e-03 -1.86301899e+00 -6.30839467e-01
-4.01988626e-01 -9.41515446e-01 1.57357916e-01 -3.38454366e-01
-8.76567006e-01 9.31431115e-01 4.39445734e-01 -3.33980322e-01
8.75750244e-01 4.16245878e-01 6.28724158e-01 2.44566783e-01
3.44517291e-01 -7.28345513e-01 1.07049860e-01 7.83776224e-01
1.02419722e+00 -1.38628352e+00 1.37316108e-01 -9.09415007e-01
-5.96464694e-01 7.10573912e-01 8.34277570e-01 -1.74378216e-01
7.47548640e-01 -5.28978370e-02 -1.97881952e-01 -3.56303781e-01
-7.42415488e-01 -5.42642832e-01 7.19542921e-01 8.22045922e-01
-3.95501465e-01 -2.22820919e-02 3.33604395e-01 5.48916161e-01
1.15013085e-01 -4.64580923e-01 2.31203675e-01 5.76638222e-01
-6.30812049e-01 -5.81411600e-01 -7.24151790e-01 1.72098130e-01
1.79516599e-01 7.55862817e-02 -2.10332692e-01 9.08622622e-01
3.60370427e-01 9.36491311e-01 3.01502138e-01 -2.29425609e-01
7.37339914e-01 -3.10750782e-01 6.22297585e-01 -7.31049061e-01
5.91183193e-02 2.07309112e-01 -3.00052483e-02 -4.51435715e-01
-2.42544368e-01 -8.84416342e-01 -9.91177499e-01 -1.69929460e-01
-5.85712671e-01 -4.39993113e-01 5.60390949e-01 9.45747554e-01
5.45726001e-01 2.79504031e-01 5.93317151e-01 -1.44914150e+00
-2.93617904e-01 -6.28781378e-01 -3.00373495e-01 3.14311504e-01
3.88323486e-01 -1.24019122e+00 -2.07334325e-01 -3.54277231e-02] | [7.770467758178711, -2.558281421661377] |
969aa3e3-ca62-4770-9eeb-7916813020f8 | learning-video-independent-eye-contact | 2210.02033 | null | https://arxiv.org/abs/2210.02033v1 | https://arxiv.org/pdf/2210.02033v1.pdf | Learning Video-independent Eye Contact Segmentation from In-the-Wild Videos | Human eye contact is a form of non-verbal communication and can have a great influence on social behavior. Since the location and size of the eye contact targets vary across different videos, learning a generic video-independent eye contact detector is still a challenging task. In this work, we address the task of one-way eye contact detection for videos in the wild. Our goal is to build a unified model that can identify when a person is looking at his gaze targets in an arbitrary input video. Considering that this requires time-series relative eye movement information, we propose to formulate the task as a temporal segmentation. Due to the scarcity of labeled training data, we further propose a gaze target discovery method to generate pseudo-labels for unlabeled videos, which allows us to train a generic eye contact segmentation model in an unsupervised way using in-the-wild videos. To evaluate our proposed approach, we manually annotated a test dataset consisting of 52 videos of human conversations. Experimental results show that our eye contact segmentation model outperforms the previous video-dependent eye contact detector and can achieve 71.88% framewise accuracy on our annotated test set. Our code and evaluation dataset are available at https://github.com/ut-vision/Video-Independent-ECS. | ['Yusuke Sugano', 'Tianyi Wu'] | 2022-10-05 | null | null | null | null | ['contact-detection'] | ['robots'] | [ 1.79335430e-01 -2.84468591e-01 -2.37573698e-01 -4.07773852e-01
-5.58621526e-01 -6.01373196e-01 4.08465236e-01 -4.26454484e-01
-5.79806268e-01 3.54964525e-01 1.07860630e-02 -4.89815101e-02
3.45528305e-01 9.43886414e-02 -7.46264160e-01 -6.91410840e-01
1.91248238e-01 1.01700082e-01 5.02166629e-01 1.83072031e-01
3.55372816e-01 2.31400684e-01 -1.87028515e+00 6.62467554e-02
7.90116668e-01 9.62812066e-01 1.77849680e-01 8.05357277e-01
3.24392438e-01 6.58509851e-01 -4.69800115e-01 -1.74296886e-01
2.25985855e-01 -7.15538561e-01 -7.99473226e-01 5.86305916e-01
8.28374207e-01 -4.71309632e-01 -1.46935776e-01 1.06066477e+00
3.68841857e-01 3.83739978e-01 5.70985734e-01 -1.75356233e+00
-1.48541495e-01 -1.03115804e-01 -7.66675532e-01 6.13397896e-01
7.69670725e-01 3.19413543e-01 7.46145010e-01 -7.48850942e-01
6.26343846e-01 1.05188417e+00 3.78002673e-01 8.74600053e-01
-9.55555856e-01 -7.57156789e-01 2.17179045e-01 5.04886210e-01
-1.50172770e+00 -1.02502358e+00 7.77797520e-01 -6.45674348e-01
3.68866324e-01 3.52340847e-01 6.96366310e-01 1.20352864e+00
-2.27497071e-01 1.05137181e+00 1.06871045e+00 -4.52526003e-01
-1.39044419e-01 1.32411942e-01 2.05394566e-01 6.35728955e-01
1.26067298e-02 -4.23476733e-02 -6.58741236e-01 9.77479666e-02
5.21422923e-01 1.12072051e-01 -7.02511549e-01 -1.98771805e-01
-1.02761328e+00 5.70445955e-01 3.31849694e-01 3.23127270e-01
-6.39393255e-02 1.54080353e-05 6.66807666e-02 2.20827833e-01
6.29129946e-01 -2.48375107e-02 -2.70010144e-01 -3.80973876e-01
-1.03073740e+00 6.16644062e-02 8.45525444e-01 1.05009401e+00
6.24185860e-01 -5.86563349e-01 -3.30532908e-01 6.26500368e-01
3.77727360e-01 4.41534251e-01 3.36330831e-01 -8.12929690e-01
3.64887506e-01 6.15764201e-01 3.59339863e-01 -8.16962540e-01
-3.90528351e-01 1.23266324e-01 -8.03748369e-02 6.10153079e-02
8.78708661e-01 -5.30531168e-01 -6.95955634e-01 1.62933099e+00
7.68434048e-01 5.29177427e-01 -2.71965086e-01 1.18763912e+00
8.65455210e-01 4.02691513e-01 -1.10727206e-01 -5.34098804e-01
1.46571183e+00 -1.20599055e+00 -9.05795634e-01 -9.68203321e-02
7.23593891e-01 -8.72320294e-01 9.42583084e-01 5.45228645e-02
-8.27022135e-01 -3.94513637e-01 -5.49307466e-01 -4.97557707e-02
-5.19827083e-02 4.37590718e-01 2.51468748e-01 5.93003035e-01
-9.93833601e-01 3.40045355e-02 -8.21183741e-01 -7.23765314e-01
3.28346372e-01 3.06188494e-01 -3.99300784e-01 6.20951280e-02
-7.06573904e-01 4.66573387e-01 1.15761600e-01 3.29597175e-01
-7.19771504e-01 -2.73236722e-01 -9.32961941e-01 -2.32885987e-01
8.64265621e-01 -2.70110428e-01 1.47748995e+00 -1.46906555e+00
-1.44305968e+00 1.14791024e+00 -9.22416031e-01 -6.20866418e-02
6.42743409e-01 -2.33269393e-01 -4.15838152e-01 5.29383361e-01
9.20533314e-02 6.78423405e-01 1.25036645e+00 -1.25226176e+00
-9.88166869e-01 -4.26947951e-01 1.69109702e-01 3.13182503e-01
-2.02065945e-01 5.85270584e-01 -9.50001657e-01 -2.54177302e-01
-3.73575240e-01 -1.33323967e+00 4.54409629e-01 2.50067804e-02
-5.24955094e-01 -5.51171184e-01 9.53440845e-01 -6.83465660e-01
1.31187105e+00 -2.15950894e+00 6.28843382e-02 -1.64959863e-01
5.19841254e-01 2.92043000e-01 1.08147919e-01 7.57625885e-03
6.06623106e-02 -1.19454466e-01 6.06415123e-02 -6.67028248e-01
-2.87179708e-01 -2.99256206e-01 3.50012243e-01 7.54213214e-01
9.33555067e-02 8.21660578e-01 -8.33015025e-01 -7.92859077e-01
2.91875191e-02 4.28612381e-01 -4.35624480e-01 5.41281760e-01
-7.63986856e-02 7.95096636e-01 -3.62393528e-01 7.17926204e-01
6.27327621e-01 -3.21446031e-01 -1.93129882e-01 -1.86669037e-01
-2.19685689e-01 -2.76116341e-01 -1.02757311e+00 1.34964883e+00
-1.48067653e-01 1.09604776e+00 7.03258589e-02 -6.47971809e-01
5.02291560e-01 2.66464055e-01 4.16680455e-01 -4.16782826e-01
6.24219298e-01 7.68394023e-02 2.48636097e-01 -1.02411664e+00
1.96570978e-01 3.76561314e-01 1.56794906e-01 6.01696432e-01
1.66931301e-01 5.10662198e-01 5.10367453e-01 -9.97597948e-02
7.92380750e-01 1.83129027e-01 -4.30270359e-02 -1.35521535e-02
9.12111759e-01 -2.39682302e-01 5.13569176e-01 2.66858667e-01
-6.95598662e-01 6.99821949e-01 4.51346695e-01 -1.48549184e-01
-5.57111382e-01 -4.89294827e-01 -5.63402027e-02 1.33773112e+00
3.82438898e-01 -1.92529097e-01 -1.31743252e+00 -7.24856555e-01
-3.96782041e-01 2.99587041e-01 -8.60528827e-01 2.15277478e-01
-4.96126831e-01 -3.70714694e-01 1.65112749e-01 9.40447971e-02
5.45249820e-01 -1.07707405e+00 -6.69338584e-01 -3.23651969e-01
-4.99273598e-01 -1.53613925e+00 -1.30668330e+00 -5.56879222e-01
-3.31608802e-01 -1.62878561e+00 -9.47613895e-01 -8.79965603e-01
9.53885198e-01 7.39638925e-01 7.09782362e-01 2.98383921e-01
-2.52174079e-01 7.38640964e-01 -4.53660071e-01 -3.47929031e-01
-1.73466027e-01 7.21846595e-02 1.01941615e-01 9.14582014e-01
9.19441044e-01 -2.46629864e-02 -9.00937080e-01 7.67082870e-01
-5.77434361e-01 -6.23811558e-02 1.44145265e-01 5.37347376e-01
2.80838579e-01 -4.06665981e-01 1.23655148e-01 -7.26190031e-01
6.12093031e-01 -3.70467246e-01 -6.19878531e-01 3.17211658e-01
-1.44354515e-02 -3.71661544e-01 8.05849507e-02 -7.71582067e-01
-9.62705374e-01 2.88672298e-01 2.51569003e-01 -7.86525965e-01
-4.72533494e-01 -2.91182399e-02 -1.09906688e-01 -2.11520731e-01
4.62023705e-01 1.97008312e-01 2.70788699e-01 -3.67413968e-01
-4.87954207e-02 1.04620695e+00 2.38273591e-01 -1.75378457e-01
6.69617832e-01 4.70921904e-01 -3.14241916e-01 -1.17609608e+00
-1.02397072e+00 -1.03974485e+00 -8.76721740e-01 -5.90748608e-01
1.17030323e+00 -9.68045235e-01 -1.17763543e+00 7.60147810e-01
-1.05173910e+00 -3.85882765e-01 4.06365246e-01 5.24894178e-01
-5.59039712e-01 2.95580417e-01 -1.82113677e-01 -8.65503073e-01
-2.18835011e-01 -1.20774972e+00 1.15620518e+00 5.94709098e-01
-1.69906378e-01 -8.86483014e-01 2.07810309e-02 7.05601096e-01
-1.39977047e-02 1.90534979e-01 -2.13887557e-01 -7.21286952e-01
-5.97526252e-01 -1.79232493e-01 -1.20084599e-01 2.34074324e-01
2.32341617e-01 1.32885352e-01 -1.02640271e+00 -2.65568554e-01
3.38669792e-02 -2.06926614e-01 5.38394451e-01 4.18876380e-01
9.00792837e-01 -1.94174096e-01 -5.91673195e-01 6.20564222e-01
8.40205967e-01 8.25928599e-02 3.57085049e-01 1.34302946e-02
9.28914905e-01 9.41213310e-01 1.00403190e+00 2.77478337e-01
4.11271811e-01 8.95000160e-01 1.91896886e-01 -4.47784401e-02
-6.11484274e-02 2.60666255e-02 3.55380505e-01 3.65182638e-01
-3.13876659e-01 -4.44301188e-01 -8.75745773e-01 6.39880478e-01
-1.88890433e+00 -1.14913011e+00 -2.99299359e-01 2.17971897e+00
7.73899198e-01 -1.67202115e-01 7.67750382e-01 -6.45433664e-02
1.17245555e+00 -2.33322039e-01 -6.45049930e-01 1.71029851e-01
3.58330339e-01 -3.98264319e-01 2.99210846e-01 5.31010687e-01
-1.27163994e+00 1.00290823e+00 4.85865402e+00 4.64533180e-01
-1.36880863e+00 2.70428389e-01 5.45763373e-01 -3.98971915e-01
3.81232619e-01 -4.40459549e-01 -1.15669918e+00 8.70232701e-01
9.56857502e-01 -1.35372207e-01 3.40634733e-01 3.82083684e-01
6.44166589e-01 -3.91585827e-01 -1.28335857e+00 1.39728940e+00
5.81468940e-01 -8.05762231e-01 -4.59441900e-01 1.89029783e-01
4.63064909e-01 9.46396310e-03 -1.46236166e-01 -1.05591029e-01
-4.63513166e-01 -8.49012375e-01 5.33892632e-01 6.57508194e-01
8.77320290e-01 -4.62692946e-01 4.24415380e-01 3.98375541e-01
-1.17230725e+00 -7.64595717e-03 2.70815063e-02 1.25946760e-01
1.53348491e-01 -2.43304238e-01 -7.55517244e-01 -6.74451664e-02
8.46850812e-01 9.61290002e-01 -7.93100238e-01 1.52904844e+00
-1.64955825e-01 5.50838232e-01 -4.29582238e-01 -1.24142267e-01
-8.47308263e-02 -1.23480402e-01 6.31881475e-01 9.40772533e-01
1.33046493e-01 1.39927089e-01 3.72046530e-02 6.95489049e-01
-1.87956408e-01 8.20066184e-02 -6.03706300e-01 9.03712586e-02
4.20416415e-01 1.40429020e+00 -7.04401791e-01 5.17492220e-02
-6.27331197e-01 1.11085546e+00 1.16140142e-01 5.69090009e-01
-9.48299468e-01 -2.52352148e-01 7.84707129e-01 4.87268835e-01
3.10066372e-01 -3.88733633e-02 5.78622282e-01 -1.36121833e+00
2.04923049e-01 -8.52636397e-01 1.90373972e-01 -6.31759882e-01
-8.60672891e-01 6.71639502e-01 1.43083915e-01 -1.62168300e+00
-4.00524259e-01 -4.49860573e-01 -8.49411488e-01 6.45554304e-01
-1.41329467e+00 -1.20016432e+00 -9.16824579e-01 1.07674623e+00
7.21679449e-01 -6.74495101e-02 2.86525548e-01 2.46230200e-01
-9.92664516e-01 9.16361153e-01 -1.40007883e-01 3.84683430e-01
9.76174057e-01 -8.62325609e-01 8.42501000e-02 9.63118076e-01
9.53149721e-02 5.01599133e-01 7.43769467e-01 -3.07110757e-01
-1.18717980e+00 -9.24823105e-01 7.91222394e-01 -6.43325686e-01
6.06849253e-01 -5.29156208e-01 -7.57265747e-01 8.12510073e-01
2.77047992e-01 2.38545746e-01 8.34078431e-01 -2.88449615e-01
2.22425506e-01 1.76111236e-02 -1.05985892e+00 7.08315492e-01
1.22721946e+00 -5.09408772e-01 -4.97634053e-01 4.02237058e-01
5.52228093e-01 -4.68078792e-01 -5.65809011e-01 -3.75817381e-02
5.04889786e-01 -9.92711961e-01 5.66784382e-01 -2.83035636e-01
-7.24690482e-02 -3.83083433e-01 4.29676443e-01 -8.92740965e-01
3.43550146e-01 -1.05952370e+00 -1.19119942e-01 1.32657790e+00
8.31932575e-02 -6.77173853e-01 6.86884344e-01 8.10928226e-01
2.14783549e-01 -4.89045620e-01 -7.71412969e-01 -4.29280967e-01
-5.39873958e-01 -1.05359010e-01 2.23035887e-01 7.19784558e-01
-8.69809538e-02 4.48652089e-01 -4.03255492e-01 1.01917416e-01
7.95098364e-01 1.94781758e-02 1.16981792e+00 -1.37639129e+00
5.29155955e-02 -1.95289224e-01 -4.10548270e-01 -1.31609273e+00
3.94632012e-01 -3.27727795e-01 9.65700373e-02 -9.47217047e-01
2.89671540e-01 1.80555899e-02 7.86830764e-03 2.08367318e-01
-2.57678479e-01 5.48351347e-01 1.78076342e-01 4.34083730e-01
-9.47992146e-01 4.08370495e-01 1.33446312e+00 4.37767915e-02
-3.56562644e-01 3.72423857e-01 -3.75938892e-01 7.20429480e-01
8.37347209e-01 -3.63964766e-01 -4.13229734e-01 -9.91245508e-02
-2.44739711e-01 4.55259196e-02 4.77227420e-01 -9.15803313e-01
5.11385262e-01 -1.16592288e-01 1.06076263e-01 -5.15917778e-01
3.88073117e-01 -8.76532197e-01 -2.72060424e-01 1.82693183e-01
-1.30754575e-01 -2.49462068e-01 1.17807552e-01 5.56870461e-01
-2.26037189e-01 -4.17969584e-01 6.91825032e-01 1.72205225e-01
-7.44380772e-01 5.83492100e-01 -2.60785788e-01 3.62091869e-01
1.49537122e+00 -4.26690698e-01 -3.23061883e-01 -4.12122697e-01
-8.25021207e-01 4.91596073e-01 7.21098840e-01 7.13147283e-01
4.94950205e-01 -9.60048139e-01 -5.57249308e-01 2.32794717e-01
3.82752597e-01 -2.40076765e-01 1.70552716e-01 1.21756136e+00
-2.87617117e-01 3.73419315e-01 -2.41236866e-01 -9.34284270e-01
-1.88716936e+00 6.19929075e-01 5.92782378e-01 5.33234894e-01
-3.54496151e-01 8.21987748e-01 3.41324985e-01 8.06713626e-02
4.25017118e-01 -2.12298140e-01 -7.07192123e-01 2.86416322e-01
7.69422650e-01 4.07727391e-01 -3.19991916e-01 -1.30565691e+00
-2.95583516e-01 7.84481168e-01 -1.36950761e-01 2.25844711e-01
1.02282572e+00 -7.31692851e-01 1.11507457e-02 5.35389721e-01
1.42863774e+00 -8.08793306e-02 -1.54207563e+00 -2.62584478e-01
-1.95405439e-01 -7.02832580e-01 -1.54468715e-01 -1.41961440e-01
-1.04755187e+00 8.23199034e-01 7.75404334e-01 1.62862614e-01
1.11074865e+00 1.78885490e-01 7.13207603e-01 2.14562237e-01
1.42658740e-01 -9.84490335e-01 2.01532170e-01 2.15720803e-01
7.56986797e-01 -1.84593272e+00 -3.88290256e-01 -5.99859238e-01
-6.54612005e-01 9.45739448e-01 8.31796110e-01 1.04309574e-01
7.56528556e-01 -2.41802678e-01 2.15705782e-01 -2.32683524e-01
-5.10582685e-01 -7.63670743e-01 6.41695261e-01 5.64491451e-01
3.99251908e-01 -2.08135858e-01 -1.86954156e-01 1.43454537e-01
-4.54902276e-03 1.69167042e-01 5.94407201e-01 6.79908395e-01
-2.72338212e-01 -9.84329164e-01 -3.57242733e-01 2.57322252e-01
-7.07925797e-01 1.38451770e-01 -3.72989863e-01 7.24924147e-01
-4.63097878e-02 1.38808608e+00 2.67398894e-01 -2.53089190e-01
5.33588789e-02 7.81230778e-02 4.82384413e-01 -6.06774271e-01
-2.60624915e-01 2.73844749e-01 -2.34751636e-03 -7.07771778e-01
-9.41524982e-01 -9.18768644e-01 -8.72075796e-01 -3.00690174e-01
-3.86249751e-01 9.26863328e-02 3.03410858e-01 8.90818655e-01
3.44625533e-01 1.68733671e-01 6.77470267e-01 -1.21839249e+00
-8.03035200e-02 -1.18928683e+00 -4.17436898e-01 7.63249338e-01
8.08454156e-01 -1.01099074e+00 -5.93137145e-01 4.71124709e-01] | [14.064658164978027, 0.10031753778457642] |
7eadd19d-b717-482a-977a-f3e884d5b52c | unter-a-unified-knowledge-interface-for | 2305.01624 | null | https://arxiv.org/abs/2305.01624v2 | https://arxiv.org/pdf/2305.01624v2.pdf | UNTER: A Unified Knowledge Interface for Enhancing Pre-trained Language Models | Recent research demonstrates that external knowledge injection can advance pre-trained language models (PLMs) in a variety of downstream NLP tasks. However, existing knowledge injection methods are either applicable to structured knowledge or unstructured knowledge, lacking a unified usage. In this paper, we propose a UNified knowledge inTERface, UNTER, to provide a unified perspective to exploit both structured knowledge and unstructured knowledge. In UNTER, we adopt the decoder as a unified knowledge interface, aligning span representations obtained from the encoder with their corresponding knowledge. This approach enables the encoder to uniformly invoke span-related knowledge from its parameters for downstream applications. Experimental results show that, with both forms of knowledge injected, UNTER gains continuous improvements on a series of knowledge-driven NLP tasks, including entity typing, named entity recognition and relation extraction, especially in low-resource scenarios. | ['Maosong Sun', 'Zhengyan Zhang', 'Yankai Lin', 'Deming Ye'] | 2023-05-02 | null | null | null | null | ['entity-typing'] | ['natural-language-processing'] | [-1.60570875e-01 3.63715798e-01 -8.27431798e-01 -1.97541654e-01
-6.83954000e-01 -1.04912758e+00 3.66566956e-01 -6.07203022e-02
-5.22545576e-01 1.02486932e+00 4.09659892e-01 -4.84341294e-01
1.16636261e-01 -8.30255747e-01 -9.34125364e-01 -1.49256721e-01
2.47921079e-01 4.64534104e-01 2.13654727e-01 -2.80594565e-02
-2.39605382e-01 3.55116010e-01 -7.84642756e-01 3.58167529e-01
1.15757573e+00 6.92458391e-01 1.92407757e-01 2.61594296e-01
-5.51450253e-01 9.32370305e-01 -4.46220726e-01 -1.07990932e+00
1.16696404e-02 2.00058799e-02 -1.31331825e+00 -5.76569796e-01
-4.19524573e-02 -2.83365995e-01 -4.90256399e-01 6.86284602e-01
3.86502922e-01 1.55791417e-01 7.30043352e-01 -9.43490386e-01
-1.19283855e+00 1.29891622e+00 -2.26050779e-01 1.61841497e-01
2.84082115e-01 2.81308711e-01 1.07868457e+00 -8.95121753e-01
8.68909895e-01 9.53108013e-01 7.21529007e-01 5.06702244e-01
-1.04388320e+00 -6.80812895e-01 2.71571726e-01 3.21511418e-01
-1.50303626e+00 -6.33358598e-01 3.61919284e-01 -2.08211109e-01
1.77262008e+00 -3.17493856e-01 2.82986939e-01 1.20091701e+00
-3.08853388e-01 1.21019423e+00 5.38603842e-01 -4.34602797e-01
-1.22941665e-01 4.12401587e-01 2.51297325e-01 5.85411072e-01
3.91350836e-01 6.33789077e-02 -8.00610125e-01 -6.25720695e-02
6.43213809e-01 -5.58311045e-01 -3.74998569e-01 -1.94940716e-01
-1.06561446e+00 4.96734440e-01 1.46917954e-01 3.06054771e-01
-4.37644094e-01 8.52179304e-02 5.62567174e-01 2.55221248e-01
4.23386753e-01 7.05930054e-01 -1.28498960e+00 -3.89636457e-01
-5.76235533e-01 -2.02934835e-02 1.47910678e+00 1.57434928e+00
8.21996272e-01 -2.09988374e-02 -5.27486205e-01 7.82381892e-01
1.77765369e-01 6.90986693e-01 4.75641251e-01 -5.38466752e-01
9.37236488e-01 7.27605700e-01 1.36155888e-01 -5.21500051e-01
-1.91691503e-01 -4.35217768e-01 -4.76627022e-01 -7.19303250e-01
3.21658194e-01 -5.36719143e-01 -9.86293018e-01 1.91483939e+00
2.64648438e-01 4.45459515e-01 7.52317846e-01 5.08386731e-01
9.49090123e-01 5.93310356e-01 5.54161727e-01 -1.04391009e-01
1.46163154e+00 -1.15163887e+00 -7.95809805e-01 -2.98487544e-01
1.03139377e+00 -3.75816703e-01 9.14393961e-01 -4.73792739e-02
-1.05378699e+00 -3.33948225e-01 -6.79260910e-01 -3.84409696e-01
-7.85553515e-01 2.65023321e-01 6.33712292e-01 4.45035845e-01
-9.03661847e-01 3.35860968e-01 -7.69492328e-01 -1.40776336e-01
4.19420987e-01 1.39463484e-01 -4.01923537e-01 -7.30154756e-03
-1.77226686e+00 1.17011821e+00 1.03968382e+00 8.41502994e-02
-7.73530245e-01 -1.19072521e+00 -1.16011441e+00 3.61975253e-01
5.89364231e-01 -1.09418666e+00 1.44819582e+00 -6.97298050e-01
-1.77005863e+00 6.00845635e-01 -3.84110421e-01 -5.52681386e-01
1.59007534e-01 -7.95888960e-01 -4.48395699e-01 -6.39965460e-02
3.27840522e-02 4.17782992e-01 4.73236769e-01 -1.06357086e+00
-4.88011539e-01 -3.88328061e-02 1.18903883e-01 3.34244192e-01
-5.36871433e-01 8.37085024e-02 -9.08387899e-01 -5.54619908e-01
-6.36953950e-01 -3.53683203e-01 -7.03422502e-02 -7.65858293e-01
-6.70051515e-01 -4.00383949e-01 4.26185608e-01 -8.86512160e-01
1.68265522e+00 -2.02283049e+00 2.70439982e-01 1.60398349e-01
9.70075652e-02 7.78270781e-01 -1.30202189e-01 6.63007796e-01
2.34512374e-01 3.84202689e-01 -1.60145849e-01 -2.09565759e-01
2.47650743e-01 4.51446354e-01 -6.37773335e-01 -2.74740010e-01
6.12128139e-01 1.73116279e+00 -1.11120903e+00 -6.04938984e-01
-2.52418995e-01 4.07799214e-01 -3.07656974e-01 2.92539895e-01
-6.47652328e-01 2.84307599e-01 -6.98707819e-01 5.37525356e-01
4.68153477e-01 -4.41531420e-01 3.70177090e-01 -1.20345816e-01
9.41214710e-02 5.29406846e-01 -7.76608765e-01 1.89959884e+00
-7.33664513e-01 2.51972347e-01 -1.51021510e-01 -5.64376473e-01
6.63403749e-01 6.66701138e-01 -9.84283816e-03 -3.61041874e-01
-1.92485332e-01 1.58320963e-01 -2.49830872e-01 -7.00450301e-01
5.69179595e-01 -1.59316108e-01 -1.29952446e-01 3.28795224e-01
4.81437862e-01 3.59725475e-01 2.71748424e-01 3.27992290e-01
1.33941483e+00 4.09406364e-01 5.90106606e-01 3.74382794e-01
4.36252803e-01 1.08722381e-01 6.50313973e-01 8.21164787e-01
-1.35595584e-02 -7.76547194e-02 3.03721160e-01 1.36112258e-01
-6.36372209e-01 -1.34712839e+00 -3.84384654e-02 1.18315148e+00
-8.04005787e-02 -5.95570982e-01 -4.98331547e-01 -1.19321525e+00
3.05617541e-01 8.66166234e-01 -1.73098519e-01 -4.08579171e-01
-7.28495955e-01 -3.33171964e-01 1.21848524e+00 9.60938692e-01
5.56970537e-01 -1.27393687e+00 -5.83202355e-02 4.00090694e-01
-2.25497559e-01 -1.61415970e+00 -4.09521312e-01 2.66922563e-01
-6.81789577e-01 -9.49890256e-01 -6.06457651e-01 -9.03626382e-01
4.09018368e-01 -1.31233141e-01 1.28641558e+00 -3.69951993e-01
2.15054125e-01 5.26080549e-01 -6.10199332e-01 -1.72280639e-01
-3.90907735e-01 7.84654558e-01 -9.20735523e-02 -2.12896287e-01
7.52478361e-01 -5.31776011e-01 -2.02234760e-01 -6.50912151e-02
-7.55096793e-01 7.15643093e-02 9.02110398e-01 7.70284414e-01
3.56336176e-01 -2.00209662e-01 9.92086053e-01 -1.04898369e+00
6.62650466e-01 -7.86519945e-01 -2.89365858e-01 7.16433525e-01
-4.70243454e-01 4.80101317e-01 9.66947377e-01 -3.42014283e-01
-1.68590319e+00 -8.79119337e-03 -5.14846966e-02 -4.88340467e-01
-3.03190827e-01 1.10654926e+00 -5.75809896e-01 2.66580015e-01
5.13554394e-01 4.85197514e-01 -6.04436100e-01 -8.21391463e-01
1.14241457e+00 6.66795075e-01 6.06351554e-01 -9.89457011e-01
7.84588754e-01 -1.16514927e-02 -6.66438282e-01 -4.57534164e-01
-1.11555648e+00 -5.73769391e-01 -7.89491117e-01 4.41432118e-01
7.17357934e-01 -1.17369807e+00 -6.22329593e-01 3.46952707e-01
-1.38797915e+00 -5.23633778e-01 -3.82367879e-01 4.56745356e-01
-4.51626390e-01 3.45425129e-01 -8.20011735e-01 -4.26793307e-01
-4.69572067e-01 -7.00551271e-01 9.51967061e-01 3.90504032e-01
-1.58355832e-01 -1.33739388e+00 7.11613968e-02 4.04450238e-01
3.71341795e-01 -4.13646936e-01 1.31457114e+00 -1.08662117e+00
-5.95558286e-01 6.75682127e-02 -4.06042039e-01 4.46905941e-01
1.26531675e-01 -3.84157181e-01 -9.37039316e-01 1.45903483e-01
-5.13400674e-01 -5.35177469e-01 8.24172318e-01 -2.16276154e-01
9.97207642e-01 -5.49054205e-01 -7.23768294e-01 7.31003761e-01
1.30307317e+00 4.47619185e-02 5.07463217e-01 2.09754556e-01
8.19960535e-01 3.01785618e-01 3.22773129e-01 2.01072097e-01
6.85642064e-01 5.85546613e-01 -2.75191218e-01 2.21797138e-01
-2.78973103e-01 -6.51131630e-01 5.24064600e-01 9.42154408e-01
2.35099588e-02 -3.34541082e-01 -1.16028202e+00 8.52216542e-01
-1.90382957e+00 -6.89993739e-01 2.70415813e-01 1.79951656e+00
1.84989154e+00 -1.64391622e-01 -3.88865441e-01 -6.29961550e-01
4.86783773e-01 -1.24509871e-01 -7.92709470e-01 -3.51433933e-01
-5.78983910e-02 4.53181207e-01 5.64910352e-01 3.47296864e-01
-9.47489321e-01 1.71576202e+00 6.15448093e+00 1.02401626e+00
-8.64987016e-01 2.34743446e-01 6.16432680e-03 1.00775987e-01
-4.64177191e-01 1.35405689e-01 -1.35030997e+00 4.21806931e-01
1.15292752e+00 -6.82220936e-01 1.49246648e-01 8.42447400e-01
-1.46377772e-01 2.94133067e-01 -1.23018050e+00 5.14936686e-01
-3.09763283e-01 -1.29894710e+00 3.94390583e-01 -3.94248635e-01
6.66638494e-01 8.36926997e-02 -2.01881871e-01 9.47372139e-01
7.29029775e-01 -1.01486659e+00 2.49334574e-01 6.30972564e-01
9.87362623e-01 -7.17871249e-01 6.76345348e-01 4.62950826e-01
-1.19117033e+00 1.76342666e-01 -3.96487713e-02 1.13324925e-01
6.08463109e-01 6.25936747e-01 -1.12342417e+00 1.07144022e+00
1.88814357e-01 7.44022965e-01 -2.51379013e-01 7.62000084e-01
-7.91494966e-01 8.86194766e-01 -2.29675293e-01 1.30977020e-01
5.04031554e-02 2.13531539e-01 3.08075219e-01 1.67342591e+00
-6.58590272e-02 5.59155308e-02 1.07628532e-01 9.56197500e-01
-4.93284494e-01 1.84763148e-01 -6.94266617e-01 -4.22366768e-01
8.66196871e-01 9.28238392e-01 -1.23158894e-01 -6.47544205e-01
-6.88472986e-01 1.10588074e+00 1.01755512e+00 8.25278282e-01
-7.88329363e-01 -6.87838495e-01 7.73634613e-01 -5.23494542e-01
6.43720567e-01 -2.65608728e-01 -1.88963905e-01 -1.45183957e+00
1.81395784e-01 -6.40976787e-01 4.47002023e-01 -5.69625139e-01
-1.34768379e+00 3.11369240e-01 2.21173242e-01 -6.06202662e-01
-3.07559937e-01 -5.58177173e-01 -2.07351342e-01 1.06159472e+00
-1.98841727e+00 -1.36251152e+00 3.07043195e-01 4.37592834e-01
2.32659385e-01 -9.71649066e-02 9.82436836e-01 3.80249351e-01
-7.04316735e-01 1.02996588e+00 9.99895185e-02 6.81480825e-01
7.84971774e-01 -1.31454051e+00 5.05153477e-01 7.71002471e-01
2.92241365e-01 1.04827631e+00 6.80671856e-02 -9.36754525e-01
-1.69053674e+00 -1.31323719e+00 1.30922616e+00 -7.70986438e-01
8.25345635e-01 -1.50484696e-01 -1.08708918e+00 1.21312320e+00
2.06623510e-01 -3.51591036e-02 1.09233975e+00 5.19443810e-01
-7.96514452e-01 1.90319821e-01 -7.47586191e-01 5.99946916e-01
1.35410559e+00 -7.45553553e-01 -1.03629470e+00 3.73720974e-02
1.25660193e+00 -5.39267361e-01 -1.32113945e+00 5.47976553e-01
3.17924410e-01 -1.00039497e-01 9.40574765e-01 -1.16050732e+00
5.16019344e-01 -1.99348494e-01 1.86768711e-01 -1.50746691e+00
-1.99852809e-01 -6.21822715e-01 -1.05551553e+00 1.68208289e+00
9.20800149e-01 -7.18568146e-01 6.46055877e-01 5.82841635e-01
-2.98675686e-01 -8.07974815e-01 -7.56488085e-01 -1.13955653e+00
2.20046103e-01 -4.93516833e-01 6.42375350e-01 1.13714612e+00
4.69027936e-01 7.90746808e-01 -1.08759761e-01 3.35022092e-01
6.08254746e-02 1.33928671e-01 6.49059415e-01 -7.23018050e-01
-4.43517357e-01 -3.61021817e-01 -4.67411950e-02 -1.48093653e+00
7.42135227e-01 -1.40961802e+00 -8.60142857e-02 -1.71915519e+00
1.55499771e-01 -5.37677288e-01 -3.56859773e-01 1.12232637e+00
-5.03493369e-01 -4.25071031e-01 6.23196028e-02 -2.25233659e-02
-7.28635490e-01 6.63147092e-01 1.13586593e+00 -3.23343230e-03
-3.38159770e-01 -3.48677307e-01 -9.87035334e-01 4.85081732e-01
7.08091676e-01 -3.04949880e-01 -5.32394469e-01 -8.56174290e-01
5.31793177e-01 -4.49573398e-02 6.46540001e-02 -5.54452538e-01
4.24474746e-01 -1.85612649e-01 1.37409642e-01 -2.24365160e-01
4.99578305e-02 -7.31645823e-01 -1.72428548e-01 2.34237500e-02
-3.07736427e-01 -3.58372033e-01 5.32350242e-01 5.24814248e-01
-4.00350779e-01 -2.30259925e-01 1.70282632e-01 -1.52293727e-01
-1.13479686e+00 3.26258004e-01 -6.10272698e-02 5.55347025e-01
9.16546106e-01 1.15104027e-01 -6.45157158e-01 -8.42221454e-02
-7.65597224e-01 6.00157857e-01 1.61770180e-01 4.99518603e-01
5.11093497e-01 -1.23238719e+00 -5.97562730e-01 7.65401796e-02
3.43801111e-01 1.92466468e-01 8.41008276e-02 8.48231137e-01
-1.07814901e-01 7.71824062e-01 2.43801385e-01 1.66217815e-02
-6.95555270e-01 7.33853936e-01 1.11771546e-01 -7.83836305e-01
-5.33732951e-01 8.85210752e-01 1.80359066e-01 -6.58458054e-01
2.76935816e-01 -4.72386241e-01 -2.24777728e-01 5.28369751e-03
5.52194715e-01 7.92894289e-02 -4.31546606e-02 -2.32911617e-01
-3.65279078e-01 3.12095582e-01 -3.67313355e-01 7.37566054e-02
9.89160061e-01 -8.75580236e-02 -5.14969137e-03 2.68034518e-01
1.10086811e+00 1.85189635e-01 -1.05749989e+00 -7.54662216e-01
5.05125940e-01 2.43979879e-02 -3.59482080e-01 -1.26076210e+00
-8.49332154e-01 6.32564068e-01 -3.23078275e-01 -2.86906898e-01
1.00264454e+00 2.10905924e-01 1.13742435e+00 8.40794623e-01
4.40278918e-01 -9.37665939e-01 -3.59120369e-01 1.19170833e+00
5.44394612e-01 -9.28672373e-01 -4.43285078e-01 -6.40918195e-01
-8.06192517e-01 8.11968684e-01 8.03219080e-01 5.52053750e-01
6.28609002e-01 4.39227909e-01 1.56404063e-01 -7.19057992e-02
-1.01304162e+00 -6.02719069e-01 2.37693623e-01 7.58496761e-01
4.96719003e-01 1.03237946e-02 -2.19121277e-01 1.30131876e+00
-6.57234564e-02 5.40648878e-01 -6.87372610e-02 9.59525466e-01
-2.38294110e-01 -1.26819420e+00 2.68565655e-01 3.88447464e-01
-5.06837606e-01 -6.88631594e-01 -3.63093317e-01 6.57144606e-01
6.58470839e-02 6.40358031e-01 -3.35700333e-01 -2.76407242e-01
4.50901240e-01 7.24736392e-01 4.22224998e-01 -1.04014575e+00
-8.27818930e-01 -6.39577746e-01 6.86828732e-01 -4.29723680e-01
-1.72098652e-01 -2.18072191e-01 -1.47954202e+00 4.58754078e-02
-4.32870746e-01 3.70889127e-01 2.55216449e-01 1.11092091e+00
8.59595299e-01 4.33390439e-01 7.10551292e-02 -7.07098097e-02
-6.31352842e-01 -9.98100519e-01 -3.80614817e-01 2.90423095e-01
-4.61728089e-02 -5.70562899e-01 3.89401652e-02 3.77643287e-01] | [10.407671928405762, 8.2855863571167] |
d5fb4aea-dec5-4590-98cc-439bb760b04f | deepfake-detection-with-inconsistent-head | 2108.12715 | null | https://arxiv.org/abs/2108.12715v1 | https://arxiv.org/pdf/2108.12715v1.pdf | DeepFake Detection with Inconsistent Head Poses: Reproducibility and Analysis | Applications of deep learning to synthetic media generation allow the creation of convincing forgeries, called DeepFakes, with limited technical expertise. DeepFake detection is an increasingly active research area. In this paper, we analyze an existing DeepFake detection technique based on head pose estimation, which can be applied when fake images are generated with an autoencoder-based face swap. Existing literature suggests that this method is an effective DeepFake detector, and its motivating principles are attractively simple. With an eye towards using these principles to develop new DeepFake detectors, we conduct a reproducibility study of the existing method. We conclude that its merits are dramatically overstated, despite its celebrated status. By investigating this discrepancy we uncover a number of important and generalizable insights related to facial landmark detection, identity-agnostic head pose estimation, and algorithmic bias in DeepFake detectors. Our results correct the current literature's perception of state of the art performance for DeepFake detection. | ['Robert Bassett', 'Kevin Lutz'] | 2021-08-28 | null | null | null | null | ['head-pose-estimation'] | ['computer-vision'] | [-1.66650429e-01 4.27931339e-01 -7.03443289e-02 -2.62072146e-01
-5.84556222e-01 -4.42033857e-01 7.52461672e-01 -8.71523440e-01
-1.57074600e-01 5.65541685e-01 3.67957771e-01 -1.18246786e-01
1.48981929e-01 -5.09658813e-01 -7.54968762e-01 -6.41090930e-01
1.22237898e-01 4.20654975e-02 -2.10807279e-01 -3.60443562e-01
2.58726776e-01 7.81473756e-01 -1.70112455e+00 1.87978670e-01
2.46420950e-01 8.99175882e-01 -5.94058812e-01 5.00274599e-01
3.25243950e-01 7.23153353e-01 -1.06991923e+00 -1.19204664e+00
4.64816809e-01 -4.21501547e-01 -5.96410692e-01 2.94427108e-02
1.34164536e+00 -1.14114285e+00 -4.97178584e-01 1.02769196e+00
1.02002025e+00 -4.66123819e-01 6.43832922e-01 -1.63306963e+00
-8.68466079e-01 2.00813413e-01 -5.44291139e-01 1.34630635e-01
1.93779007e-01 3.68230224e-01 6.11090124e-01 -1.25284851e+00
3.60920370e-01 1.49679923e+00 1.23368168e+00 1.08925450e+00
-8.65833104e-01 -1.20832753e+00 -3.61197203e-01 7.86895230e-02
-1.61234021e+00 -1.07692420e+00 8.01835597e-01 -5.30043304e-01
4.75969762e-01 4.85647023e-02 6.53997540e-01 1.57875073e+00
-3.40055376e-02 9.40363348e-01 1.09990716e+00 -4.85090286e-01
-8.29780325e-02 1.95745945e-01 -2.14780509e-01 1.09794164e+00
6.63209617e-01 6.65365100e-01 -8.13301563e-01 -3.22806150e-01
9.66354668e-01 -7.19853699e-01 -1.83604494e-01 -6.09738380e-02
-5.97115636e-01 1.28287256e+00 3.56544495e-01 9.60011780e-02
-5.97661547e-02 4.25476164e-01 3.39490980e-01 1.70473307e-01
4.97388482e-01 4.76779789e-01 1.00345388e-01 1.77699909e-01
-1.22633708e+00 5.17207146e-01 6.15576625e-01 5.81562698e-01
5.44095039e-01 5.84083259e-01 -6.84510320e-02 5.68468690e-01
3.52959722e-01 4.12998885e-01 6.36604667e-01 -9.74057138e-01
-8.64991918e-02 -1.08556770e-01 1.62092909e-01 -1.57779956e+00
-3.00593615e-01 -2.17873380e-01 -4.80547756e-01 5.85866928e-01
6.48239315e-01 -2.95439512e-01 -6.74561203e-01 1.63914955e+00
4.13759977e-01 3.87201220e-01 -2.75590092e-01 9.43424344e-01
1.04318476e+00 -1.28297895e-01 -1.35734066e-01 2.94462323e-01
1.39924490e+00 -7.21944511e-01 -6.94569528e-01 -1.71538860e-01
6.67012751e-01 -8.92803371e-01 8.46813083e-01 3.88733000e-01
-9.22388434e-01 -2.73835778e-01 -1.14153755e+00 -9.81727242e-02
-3.00946742e-01 4.30209249e-01 1.02371371e+00 1.59035289e+00
-1.29061377e+00 6.37005031e-01 -1.63929895e-01 -3.86561394e-01
9.35419261e-01 4.12457228e-01 -3.99355888e-01 3.41234028e-01
-1.09297168e+00 1.02959239e+00 -1.15504954e-02 1.63990021e-01
-1.03553331e+00 -4.28838223e-01 -5.75342894e-01 -3.84201169e-01
1.39727086e-01 -6.90467656e-01 1.54450703e+00 -1.27282095e+00
-1.57276285e+00 1.49737871e+00 -7.02526271e-02 -5.38744569e-01
8.57277453e-01 -1.77664235e-01 -6.42535448e-01 2.41784498e-01
8.28647688e-02 7.33536422e-01 1.68495023e+00 -9.98145819e-01
-1.52615860e-01 -6.08048975e-01 -3.42553914e-01 -2.33186811e-01
-4.43571448e-01 2.25049987e-01 1.54426694e-01 -9.01030719e-01
-1.59020752e-01 -9.92306471e-01 5.17055333e-01 4.22407031e-01
-6.46949470e-01 -2.64725626e-01 1.02424240e+00 -7.25899696e-01
1.04701006e+00 -1.97538412e+00 -6.14685118e-01 2.04943232e-02
8.31533670e-01 4.81776237e-01 -1.12210371e-01 8.04463029e-02
-1.22779496e-01 2.44835705e-01 2.03671828e-01 -4.35494870e-01
-5.29262284e-03 -2.20500380e-01 -3.29151183e-01 1.00438750e+00
1.84463412e-01 1.18428433e+00 -6.76846504e-01 -1.92581549e-01
6.25412315e-02 3.99983823e-01 -6.35199666e-01 1.32513093e-02
3.02345783e-01 2.04864427e-01 -1.86532617e-01 8.35953236e-01
1.17743766e+00 -5.59144793e-03 -1.87436372e-01 -3.95007193e-01
4.70821140e-03 6.91569597e-02 -8.08524787e-01 1.04433787e+00
2.22438388e-02 1.19942081e+00 2.17885464e-01 -6.98814034e-01
7.57695735e-01 1.62448198e-01 8.91026631e-02 -5.10549128e-01
5.29633701e-01 4.23226237e-01 -1.13633731e-02 -4.76490051e-01
6.41105711e-01 -4.59494293e-01 3.56775790e-01 5.91366887e-01
2.42266059e-01 -9.56929475e-02 -6.31617606e-01 1.16363488e-01
8.82064342e-01 -9.94502157e-02 3.15714031e-02 -4.23349172e-01
1.33817062e-01 -3.04333836e-01 1.70442369e-02 9.36969995e-01
-8.12522054e-01 7.21283853e-01 2.28832528e-01 -6.01945937e-01
-1.05934834e+00 -8.61874044e-01 -1.24059856e-01 9.78986204e-01
-2.64169186e-01 -2.03577623e-01 -1.20410550e+00 -8.15589845e-01
4.04500633e-01 2.69513041e-01 -1.05049658e+00 -3.21373284e-01
-5.04795611e-01 -7.83651531e-01 1.55820334e+00 3.80406976e-01
6.34754419e-01 -7.76622176e-01 -4.66061592e-01 -2.22920597e-01
-8.49437267e-02 -1.07437015e+00 -3.50306630e-01 -7.14150131e-01
-3.68724346e-01 -1.11965215e+00 -1.18005979e+00 -7.64795601e-01
4.00393158e-01 3.84324521e-01 1.06847525e+00 3.27413261e-01
-4.15086299e-01 3.90138745e-01 2.86848750e-02 -6.71166003e-01
-6.54576361e-01 -2.13169143e-01 4.55757260e-01 2.08289310e-01
7.65202522e-01 -2.49614224e-01 -8.89430404e-01 3.29687744e-01
-6.54766262e-01 -3.30592006e-01 4.88665581e-01 9.06230271e-01
-3.19718271e-01 -4.30884510e-01 8.53015840e-01 -6.92387879e-01
9.35722709e-01 -2.54771858e-01 -4.59756613e-01 -2.06049800e-01
-5.40916443e-01 -1.28826812e-01 1.12911966e-02 -3.93581659e-01
-1.11803639e+00 -8.95367637e-02 -4.07125413e-01 -4.97322649e-01
-6.99223205e-02 -7.20095560e-02 4.81382497e-02 -8.25643182e-01
1.16348588e+00 2.06729591e-01 3.50889117e-01 -2.35976934e-01
5.07670164e-01 7.37062037e-01 7.04093993e-01 -3.41612101e-01
7.74118781e-01 8.82129014e-01 -1.43114433e-01 -1.03384280e+00
-5.57245910e-01 -2.12301742e-02 -3.74928385e-01 -4.20767307e-01
4.45910245e-01 -9.76926386e-01 -1.04930079e+00 1.02483869e+00
-1.36147237e+00 -4.30874666e-03 5.88282831e-02 2.00496763e-01
-3.93597811e-01 6.93137527e-01 -6.82967365e-01 -9.42089856e-01
-2.79003382e-01 -8.60790730e-01 1.41162622e+00 -5.87570257e-02
-4.88373041e-01 -7.39631891e-01 6.54647723e-02 5.18223882e-01
4.53912497e-01 2.32763901e-01 1.98625132e-01 -4.42040890e-01
-5.24325252e-01 -3.25124234e-01 -3.58825773e-01 4.64852065e-01
-1.77268777e-02 1.04890637e-01 -1.72958541e+00 -3.06470633e-01
7.15138689e-02 -5.27029812e-01 8.10894132e-01 4.75922614e-01
1.07735968e+00 -7.65884638e-01 -1.72626063e-01 6.63226306e-01
9.66872334e-01 -3.59491467e-01 8.69078696e-01 2.68581420e-01
6.34678006e-01 7.68340707e-01 -8.38714242e-02 5.54958403e-01
2.93701261e-01 7.17371583e-01 4.45066214e-01 -6.33406490e-02
-4.59531158e-01 -5.36071181e-01 3.34232599e-01 1.48693666e-01
1.06900195e-02 -1.09490819e-01 -5.79553783e-01 4.25939649e-01
-1.25807953e+00 -1.13501358e+00 -1.37574822e-01 2.08029509e+00
4.59210157e-01 -2.40075022e-01 5.10248363e-01 1.80153951e-01
8.30882788e-01 2.28537679e-01 -3.21223944e-01 -3.40522170e-01
-4.01952863e-01 2.07373783e-01 6.57320201e-01 2.21891761e-01
-1.12399673e+00 1.38359308e+00 7.54733562e+00 7.82258511e-01
-1.35595751e+00 2.01853305e-01 6.88416898e-01 1.23340525e-01
7.63101503e-03 -5.99431396e-01 -8.01559806e-01 4.64877605e-01
6.81850553e-01 -2.59007514e-02 2.70350158e-01 9.78057683e-01
1.43802255e-01 1.18896261e-01 -1.07920563e+00 1.39575040e+00
6.28329813e-01 -1.66809916e+00 -5.73780648e-02 2.67709345e-01
5.27160466e-01 -1.00631930e-01 6.73350513e-01 3.83966565e-02
2.48281673e-01 -1.36783016e+00 8.14293981e-01 7.70736486e-02
9.33623791e-01 -5.82052708e-01 4.58672792e-01 -7.62758106e-02
-4.77523655e-01 -5.56977792e-03 -2.90379107e-01 -2.17144668e-01
-1.74649462e-01 3.95505816e-01 -1.26765549e+00 -1.98203489e-01
7.01759398e-01 3.61143440e-01 -4.76981670e-01 7.82441616e-01
-2.95048833e-01 5.17411172e-01 -2.90396333e-01 -2.39769761e-02
-3.06305271e-02 4.57372308e-01 4.52963352e-01 1.14061570e+00
4.12201375e-01 -2.21829817e-01 -5.27581215e-01 1.02364922e+00
-3.06194425e-01 -9.84175652e-02 -1.00962651e+00 -9.76171121e-02
4.63829756e-01 1.08268344e+00 -3.60805780e-01 -1.74920693e-01
-3.55751753e-01 1.29087031e+00 8.87594223e-02 1.01075068e-01
-7.35782385e-01 -2.28381269e-02 1.06750441e+00 3.52679402e-01
4.28628922e-02 2.37849578e-02 -3.79778445e-01 -1.24486184e+00
-6.75662532e-02 -1.04106092e+00 2.60760844e-01 -7.64170289e-01
-1.43491280e+00 4.36470628e-01 -3.64329189e-01 -1.01928830e+00
-2.85561800e-01 -9.15003419e-01 -5.40091693e-01 5.47124803e-01
-1.44991148e+00 -1.44321561e+00 -4.06812459e-01 7.25203097e-01
2.11134925e-01 -4.19220269e-01 8.05350959e-01 3.29602599e-01
-5.63799679e-01 1.42094612e+00 -6.49450049e-02 4.67634827e-01
7.63998568e-01 -5.91639102e-01 7.51403809e-01 7.97848403e-01
1.01824597e-01 6.02991223e-01 7.61208534e-01 -5.74704051e-01
-1.27685392e+00 -8.46173882e-01 6.81352079e-01 -8.98001969e-01
6.26732945e-01 -3.63772959e-01 -5.86630940e-01 6.47039711e-01
-1.26227006e-01 -6.26844764e-02 6.66566432e-01 -1.04546964e-01
-6.66212380e-01 1.65529460e-01 -1.62571764e+00 6.86939061e-01
1.04830277e+00 -7.05115318e-01 -3.96126837e-01 3.76910001e-01
5.95874991e-03 -3.64625484e-01 -2.72077471e-01 2.21399829e-01
1.07388830e+00 -1.40700555e+00 1.10477352e+00 -5.71959972e-01
3.42445672e-01 1.44773111e-01 1.48167357e-01 -1.24090958e+00
-2.77153581e-01 -9.74067748e-01 -2.28366360e-01 8.82307649e-01
-4.67179753e-02 -7.89477944e-01 1.45583940e+00 5.23593247e-01
-2.56047677e-03 -2.87253857e-01 -1.16271877e+00 -8.88350368e-01
3.34312260e-01 -4.29232329e-01 6.69484913e-01 1.22367072e+00
-3.33301574e-01 1.86220318e-01 -1.05430686e+00 1.59427628e-01
8.44912708e-01 -5.22783995e-01 1.17353821e+00 -1.32498014e+00
-7.83024728e-02 -6.04735255e-01 -6.95062459e-01 -1.06639647e+00
4.36288357e-01 -5.63626409e-01 -2.40125105e-01 -5.47399640e-01
-5.06237708e-02 6.76863864e-02 2.31754914e-01 3.01946253e-01
2.36635447e-01 7.24323392e-01 1.16498739e-01 2.21339956e-01
1.80843577e-01 4.04236883e-01 1.14933658e+00 -1.87254306e-02
3.43687028e-01 -3.49320173e-02 -9.72140014e-01 1.10054183e+00
7.25183666e-01 -2.64615923e-01 1.63881015e-02 -4.15794522e-01
1.61665305e-02 -4.21566576e-01 9.23912108e-01 -8.09818089e-01
5.59405722e-02 2.22522184e-01 7.40115166e-01 -7.79074803e-02
5.52576125e-01 -3.84368092e-01 -2.86611736e-01 4.49848443e-01
-1.15022631e-02 -1.15238369e-01 2.93653637e-01 3.41300964e-01
1.18192822e-01 -9.33782384e-02 9.86613989e-01 -2.57443488e-01
-9.03225660e-01 2.57437289e-01 -4.54703748e-01 -2.09514033e-02
7.14920640e-01 -6.17942989e-01 -5.47866881e-01 -9.08133626e-01
-4.75936711e-01 -6.12180829e-01 3.59600633e-01 4.98979241e-01
8.90364408e-01 -1.43383467e+00 -8.96751940e-01 5.67238748e-01
1.61691681e-01 -9.41226065e-01 2.39786386e-01 5.25661409e-01
-5.68852842e-01 2.49342024e-01 -2.69070417e-01 -2.83780098e-01
-1.32425928e+00 3.39472711e-01 6.17265701e-01 6.75477564e-01
-4.10979241e-01 1.35804248e+00 3.91771525e-01 -1.52641043e-01
1.56439260e-01 1.49849206e-01 8.81484002e-02 2.71410053e-03
8.06103945e-01 4.97274786e-01 1.95140719e-01 -1.12761223e+00
-3.97325605e-01 2.75148839e-01 3.12592052e-02 4.06291597e-02
8.46144021e-01 -3.27708423e-02 1.43395677e-01 -3.90703589e-01
1.21947825e+00 3.94811720e-01 -1.12472606e+00 1.18160918e-01
-2.49690592e-01 -8.31046820e-01 1.32838160e-01 -6.73555911e-01
-1.18461716e+00 8.99190545e-01 8.96112204e-01 6.76071597e-03
8.20232451e-01 1.31605402e-01 8.05485547e-01 1.52169928e-01
3.20170403e-01 -1.01209569e+00 5.52152097e-01 3.08556378e-01
1.02035534e+00 -1.39500618e+00 -1.23658545e-01 -4.81141657e-01
-4.33070689e-01 1.00416851e+00 6.71799004e-01 -1.37825981e-01
5.81125975e-01 1.19930878e-01 1.72401205e-01 -3.99550110e-01
3.59781571e-02 -3.46804038e-02 2.07092807e-01 9.51445997e-01
3.43516648e-01 4.71075661e-02 -9.58931223e-02 4.72628534e-01
-7.77712405e-01 6.55502528e-02 5.91252029e-01 4.56220031e-01
-2.82753229e-01 -8.00782025e-01 -5.70540965e-01 3.64844739e-01
-7.97447741e-01 -1.65914133e-01 -9.64665353e-01 9.20692503e-01
6.53022975e-02 8.02432954e-01 -1.74011141e-01 -5.23712814e-01
3.60337310e-02 -1.37515999e-02 8.31189096e-01 -2.89114803e-01
-3.80684584e-01 -2.42779851e-01 8.01110566e-02 -4.66022789e-01
-1.73911005e-01 -4.77377325e-01 -3.67653877e-01 -8.83427918e-01
-5.21360457e-01 -3.06075960e-01 5.83911002e-01 9.07725275e-01
4.73823398e-01 -3.41811746e-01 4.97700065e-01 -1.05850053e+00
-6.20873451e-01 -7.88028061e-01 -5.93597710e-01 3.41345876e-01
6.59661472e-01 -7.47990906e-01 -5.18064618e-01 -1.83339208e-01] | [12.631976127624512, 1.0536731481552124] |
c736b56e-4e5d-4845-9fff-72c37ec9b686 | uwb-at-semeval-2016-task-11-exploring | null | null | https://aclanthology.org/S16-1162 | https://aclanthology.org/S16-1162.pdf | UWB at SemEval-2016 Task 11: Exploring Features for Complex Word Identification | null | ['Michal Konkol'] | 2016-06-01 | null | null | null | semeval-2016-6 | ['complex-word-identification'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.2383294105529785, 3.5662012100219727] |
0ba280e2-f31f-4e5b-961c-9bfd5be1fd87 | self-reinforcement-attention-mechanism-for | 2305.11684 | null | https://arxiv.org/abs/2305.11684v1 | https://arxiv.org/pdf/2305.11684v1.pdf | Self-Reinforcement Attention Mechanism For Tabular Learning | Apart from the high accuracy of machine learning models, what interests many researchers in real-life problems (e.g., fraud detection, credit scoring) is to find hidden patterns in data; particularly when dealing with their challenging imbalanced characteristics. Interpretability is also a key requirement that needs to accompany the used machine learning model. In this concern, often, intrinsically interpretable models are preferred to complex ones, which are in most cases black-box models. Also, linear models are used in some high-risk fields to handle tabular data, even if performance must be sacrificed. In this paper, we introduce Self-Reinforcement Attention (SRA), a novel attention mechanism that provides a relevance of features as a weight vector which is used to learn an intelligible representation. This weight is then used to reinforce or reduce some components of the raw input through element-wise vector multiplication. Our results on synthetic and real-world imbalanced data show that our proposed SRA block is effective in end-to-end combination with baseline models. | ['Gregoire Jaffre', 'Zaineb Chelly Dagdia', 'Mustapha Lebbah', 'Hanene Azzag', 'Mohamed Djallel Dilmi', 'Kodjo Mawuena Amekoe'] | 2023-05-19 | null | null | null | null | ['fraud-detection'] | ['miscellaneous'] | [ 2.74100512e-01 7.85168186e-02 -1.29765615e-01 -7.79012084e-01
-3.55095267e-01 -2.88826749e-02 2.60247111e-01 6.66514933e-01
-3.68431419e-01 7.06149995e-01 1.96266100e-01 -3.24487627e-01
-1.55314103e-01 -6.98315322e-01 -6.22009397e-01 -5.45328915e-01
2.90042423e-02 4.29385424e-01 -2.41836205e-01 -3.73776555e-01
5.12290299e-01 2.95122415e-01 -1.56813931e+00 4.35786575e-01
1.16410112e+00 9.25619006e-01 1.05305061e-01 2.34952569e-01
-3.39834869e-01 1.04070210e+00 -5.58013260e-01 -7.29094028e-01
-1.14907455e-02 -3.10268909e-01 -7.50667572e-01 1.54219210e-01
5.48473932e-02 -1.86948091e-01 1.50115997e-01 9.30603445e-01
1.74806088e-01 5.70921004e-02 7.19943523e-01 -1.13704896e+00
-5.77213526e-01 6.60778522e-01 -9.07931805e-01 3.37823987e-01
1.95814312e-01 1.40402868e-01 1.22376871e+00 -8.62032890e-01
2.61064798e-01 1.37158895e+00 3.51013541e-01 4.77872282e-01
-1.23073828e+00 -5.43467939e-01 4.16846514e-01 5.51997721e-01
-8.39082956e-01 -2.74441183e-01 1.04239345e+00 -2.60900468e-01
8.36990476e-01 4.26098734e-01 4.80566561e-01 8.99320781e-01
2.28973389e-01 9.39475596e-01 9.48664784e-01 -5.33269763e-01
1.95796326e-01 4.00600612e-01 4.91439462e-01 2.88613230e-01
4.74649400e-01 -4.03929234e-01 -4.81888741e-01 -6.73968643e-02
2.77661562e-01 2.95630336e-01 -2.56633610e-01 -2.58672774e-01
-8.62585962e-01 8.72932315e-01 6.48533463e-01 1.93014428e-01
-6.23322129e-01 -2.39063911e-02 3.51081192e-01 2.97723234e-01
5.21234691e-01 5.27257681e-01 -3.18833590e-01 -2.59383917e-01
-5.60716212e-01 3.56778562e-01 3.56369823e-01 3.84106904e-01
5.66777945e-01 1.20689288e-01 -1.12943381e-01 9.09060657e-01
3.34615380e-01 1.06888898e-01 6.32119060e-01 -3.81926745e-01
7.12351263e-01 1.15945339e+00 2.10948542e-01 -1.22292113e+00
-5.14557481e-01 -9.17884111e-01 -1.18047535e+00 4.04859006e-01
3.29547733e-01 2.35035121e-01 -7.01421082e-01 1.66809773e+00
2.86626816e-01 -2.23496884e-01 1.20369658e-01 1.09004998e+00
2.96603501e-01 7.23215938e-01 2.27374315e-01 -1.99550524e-01
1.54091597e+00 -8.45254719e-01 -8.92756641e-01 -6.09587789e-01
4.47877824e-01 -4.88554686e-01 1.35721302e+00 5.83490133e-01
-9.08907771e-01 -5.05217671e-01 -1.02426982e+00 -2.83667445e-01
-3.11411023e-01 4.66949344e-02 5.66075921e-01 4.01410669e-01
-4.00114596e-01 6.62548184e-01 -7.07489371e-01 -1.39541283e-01
5.28954744e-01 3.40367258e-01 -2.81804055e-01 2.39877310e-02
-1.20687008e+00 1.10886848e+00 3.81285369e-01 4.14876014e-01
-2.28073582e-01 -4.13053155e-01 -7.67624021e-01 4.46320862e-01
3.08652908e-01 -5.04165173e-01 1.04886234e+00 -1.38546336e+00
-1.11312008e+00 6.84628367e-01 -9.45073962e-02 -7.76880503e-01
6.39379978e-01 -6.16835237e-01 -3.52402449e-01 -2.83841252e-01
-2.39840850e-01 3.71392280e-01 8.77150059e-01 -1.22708678e+00
-4.54608977e-01 -4.78281051e-01 1.01984203e-01 2.73704380e-01
-5.50392270e-01 -7.06493556e-02 6.07615598e-02 -8.69646311e-01
2.13258803e-01 -6.47693336e-01 -3.20893824e-01 -1.47880435e-01
-4.08627391e-01 -4.02633160e-01 7.49233723e-01 -8.64840329e-01
1.43882120e+00 -1.99642658e+00 1.98461741e-01 2.52126217e-01
1.77219421e-01 2.52767473e-01 1.32677659e-01 2.97181159e-01
-4.43180382e-01 1.25025228e-01 -4.88146722e-01 -3.02188337e-01
-2.69680750e-02 -8.40232447e-02 -4.22111124e-01 2.44434670e-01
7.21875250e-01 7.60702312e-01 -6.40659273e-01 -2.65331179e-01
4.09331396e-02 3.04928482e-01 -6.84431255e-01 3.95243615e-01
-1.70701921e-01 3.01124632e-01 -3.85945380e-01 4.46085334e-01
6.15029335e-01 -3.15922886e-01 6.36680499e-02 -2.10270137e-01
1.98720425e-01 1.84037715e-01 -1.16551149e+00 1.26861346e+00
-5.41064262e-01 3.62611592e-01 -3.54316503e-01 -1.22686791e+00
1.09588051e+00 -8.52098390e-02 -8.82158577e-02 -8.53682637e-01
1.56396508e-01 1.72621205e-01 3.30924362e-01 -6.34615123e-01
5.81484020e-01 -1.67395651e-01 4.17916849e-03 4.24421340e-01
-4.60505426e-01 2.96474844e-01 1.57441199e-01 1.24389820e-01
7.86072075e-01 -1.32504501e-03 5.23508966e-01 -1.78817198e-01
6.94593012e-01 -1.79489478e-01 8.41286361e-01 3.06558728e-01
4.60945927e-02 6.96208894e-01 8.47709656e-01 -7.85003245e-01
-9.80702817e-01 -4.83444870e-01 -8.50102603e-02 7.93236792e-01
7.17897043e-02 -1.58426613e-01 -6.22566938e-01 -6.87055349e-01
-5.55863008e-02 9.03870046e-01 -7.85293579e-01 -5.08479476e-01
-5.57497859e-01 -9.54129517e-01 -3.21434252e-02 5.79583168e-01
3.57003510e-01 -1.39911902e+00 -9.22769666e-01 4.80968714e-01
-6.14888892e-02 -6.34283960e-01 -1.72375008e-01 5.05353510e-01
-1.04729795e+00 -8.84501815e-01 -4.55274403e-01 -4.99201208e-01
9.11583543e-01 1.51215538e-01 1.20418811e+00 3.90290380e-01
-1.98929295e-01 -4.40332174e-01 -4.46534216e-01 -4.99778181e-01
-2.39375457e-01 2.07143679e-01 -6.53402209e-02 5.13131678e-01
5.47886729e-01 -4.40414488e-01 -6.00758255e-01 6.66504055e-02
-1.16440570e+00 3.30936074e-01 7.35454619e-01 1.23226428e+00
3.54359597e-01 -8.39966685e-02 9.05770838e-01 -1.16412032e+00
8.67434680e-01 -5.46812594e-01 -4.41929847e-01 2.82562792e-01
-6.51638508e-01 5.00396907e-01 9.03278589e-01 -3.65675539e-01
-9.30915058e-01 -1.59315348e-01 -2.15267643e-01 -2.26682022e-01
-5.34556918e-02 7.68646598e-01 -4.19340789e-01 4.83780324e-01
6.01360917e-01 8.05809125e-02 5.38795777e-02 -5.81283748e-01
-5.77067432e-04 9.05772805e-01 1.97858095e-01 -3.84361565e-01
5.87959945e-01 1.47393778e-01 -9.63090882e-02 -3.46799821e-01
-8.35211396e-01 -1.78934440e-01 -4.06893939e-01 9.94057059e-02
4.74824876e-01 -5.93816817e-01 -7.50558913e-01 5.29053986e-01
-1.09190869e+00 4.02821414e-02 -2.20606908e-01 2.59843886e-01
-3.59302372e-01 1.37040347e-01 -4.54223216e-01 -1.14861953e+00
-5.09943068e-01 -1.15924585e+00 7.72449255e-01 5.15767813e-01
-5.18721461e-01 -7.63991952e-01 -2.03858256e-01 5.05607128e-01
4.51940328e-01 2.71052271e-01 1.46899140e+00 -8.31891119e-01
-2.98875332e-01 -3.25139016e-01 -2.08670124e-01 4.28333431e-01
1.37260646e-01 -5.98215237e-02 -1.05685520e+00 -1.56462297e-01
7.51788095e-02 -4.71526772e-01 8.95160735e-01 1.60914049e-01
1.69716346e+00 -5.48408389e-01 6.89227134e-02 3.59997511e-01
1.20741594e+00 2.11640209e-01 6.97537482e-01 3.80938470e-01
6.20754719e-01 9.03441966e-01 7.39989519e-01 4.66622323e-01
3.85885537e-01 6.49511218e-01 6.53921664e-01 -3.28994334e-01
3.35459232e-01 -2.11331442e-01 1.93758547e-01 7.88239658e-01
-1.49830980e-02 -2.02576384e-01 -8.04441988e-01 4.01250094e-01
-1.90051627e+00 -8.94971788e-01 -1.31721005e-01 2.25681424e+00
8.74696434e-01 5.02742350e-01 -1.64633058e-02 7.21848369e-01
6.80901647e-01 8.48992392e-02 -7.74677217e-01 -6.30077422e-01
-1.26177788e-01 1.17778908e-02 3.45292166e-02 2.49993384e-01
-1.02407765e+00 5.68715513e-01 4.71564293e+00 8.24226320e-01
-1.23449683e+00 -2.05819935e-01 1.27049851e+00 1.28371110e-02
-4.88580912e-01 -2.94808596e-01 -3.63938898e-01 6.34433150e-01
7.33441830e-01 -4.66368794e-02 2.79182553e-01 8.03550005e-01
2.93036789e-01 -1.65185407e-01 -1.17192781e+00 9.44653213e-01
-1.45443857e-01 -9.34722185e-01 1.36841357e-01 -9.61752757e-02
3.85879666e-01 -7.21006215e-01 2.05102503e-01 4.14458662e-01
-1.47672147e-01 -1.27473056e+00 6.85579717e-01 3.05603713e-01
4.71140921e-01 -1.04012311e+00 1.07809663e+00 3.45854223e-01
-5.49791098e-01 -2.89972365e-01 -4.59106237e-01 -1.72372818e-01
-4.95844781e-02 7.37880766e-01 -7.64608681e-01 4.22703445e-01
5.65428913e-01 6.05499685e-01 -5.54183722e-01 1.02907884e+00
-2.44428456e-01 6.54542625e-01 -8.82211253e-02 -1.79245889e-01
7.04773739e-02 -1.57190159e-01 2.68578321e-01 1.06207681e+00
2.01848090e-01 1.26596302e-01 -3.35322350e-01 8.17933500e-01
-2.85832971e-01 3.95065665e-01 -3.97631794e-01 6.31090552e-02
9.58278924e-02 1.37303793e+00 -6.33197427e-01 -2.32908323e-01
-1.55344889e-01 8.70277703e-01 5.60884774e-01 1.89215824e-01
-7.63090611e-01 -5.07162631e-01 6.46937847e-01 3.61025244e-01
-2.58462653e-02 2.83361197e-01 -6.03429377e-01 -1.14417958e+00
3.68715584e-01 -1.26141763e+00 3.81511807e-01 -6.13464832e-01
-1.36934674e+00 8.83045137e-01 -2.71144032e-01 -1.52960432e+00
-2.67679423e-01 -4.60615456e-01 -6.81761980e-01 9.71893966e-01
-1.71154153e+00 -8.48185539e-01 -4.10839498e-01 2.04095840e-01
7.34110832e-01 -3.99045758e-02 6.86344922e-01 3.08027744e-01
-6.79665625e-01 6.86127305e-01 -2.68006548e-02 2.66918074e-02
5.72245479e-01 -1.34288430e+00 4.94061857e-01 6.77696645e-01
7.32102841e-02 6.21342182e-01 9.16335464e-01 -4.92475092e-01
-1.22237349e+00 -9.89950061e-01 1.03370130e+00 -1.29204065e-01
2.29275376e-01 -4.45480675e-01 -1.47036624e+00 4.13355559e-01
2.02725753e-01 7.64961764e-02 5.08333266e-01 1.37306482e-01
-3.03623348e-01 -4.67464656e-01 -1.21976662e+00 5.17415941e-01
5.95558226e-01 -5.76538444e-02 -6.82567120e-01 2.57303398e-02
5.56252599e-01 -4.28159595e-01 -3.56463462e-01 4.15757626e-01
5.97968102e-01 -9.84302878e-01 6.81858778e-01 -1.02630877e+00
7.39578843e-01 -3.13318670e-01 2.33277246e-01 -1.54365075e+00
-2.90537298e-01 -2.05947474e-01 -1.98088884e-01 1.22420943e+00
4.80260104e-01 -5.27282834e-01 7.22771287e-01 7.28921592e-01
1.78047389e-01 -1.00461566e+00 -8.28912437e-01 -4.07555014e-01
-1.14343613e-01 -3.23510736e-01 7.40405977e-01 1.02969408e+00
6.49039298e-02 3.67442220e-01 -6.34334624e-01 -8.59660953e-02
5.95629692e-01 1.76776350e-01 5.65018654e-01 -1.24586892e+00
-2.63942182e-01 -4.77296442e-01 -5.09122431e-01 -7.12208033e-01
3.26595432e-03 -6.52747869e-01 -2.30861560e-01 -1.23118258e+00
2.67557889e-01 -4.92522776e-01 -5.53991199e-01 4.89624172e-01
-6.23005033e-01 7.46818334e-02 3.04407150e-01 8.07884149e-03
-3.96505594e-01 7.97981620e-01 9.38577056e-01 -3.40496689e-01
-2.33279109e-01 2.05069616e-01 -1.05514681e+00 6.52162731e-01
9.33643281e-01 -5.78626215e-01 -3.08349818e-01 -4.42120820e-01
4.94962454e-01 -1.07468544e-02 2.03539610e-01 -8.59279275e-01
-5.16123772e-02 -3.38684648e-01 5.05419850e-01 -4.21687365e-01
1.37251779e-01 -1.04195952e+00 -5.88470772e-02 5.78230560e-01
-5.39320409e-01 4.14230138e-01 1.17316628e-02 3.44933361e-01
-4.94438976e-01 -3.36400807e-01 8.00932527e-01 7.38019571e-02
-3.94200951e-01 -2.48747915e-02 1.15709368e-03 -2.38096844e-02
9.30255651e-01 -1.47496387e-01 -2.23272681e-01 -4.68549788e-01
-4.30624694e-01 4.65009183e-01 1.96177855e-01 6.66450500e-01
6.93254590e-01 -1.34754789e+00 -9.68924463e-01 3.47910821e-01
2.63544828e-01 1.83874846e-01 4.60255891e-01 6.49239004e-01
-5.13592303e-01 2.69980580e-01 -3.34910333e-01 -4.29895580e-01
-1.15215135e+00 5.19828022e-01 1.07377589e-01 -5.65881550e-01
-4.23248619e-01 6.05810046e-01 1.76495284e-01 -9.10963044e-02
4.71785635e-01 -5.54044783e-01 -5.83478749e-01 1.80162311e-01
7.52672851e-01 1.51120886e-01 3.68114620e-01 -3.43044937e-01
-3.58649909e-01 3.02938581e-01 -3.31438065e-01 2.83465087e-01
1.57617021e+00 1.17406800e-01 5.40683046e-03 5.81594408e-01
8.65989447e-01 -2.21462682e-01 -1.06878376e+00 -1.95339099e-01
3.23544711e-01 -5.61415970e-01 -5.13948798e-02 -8.86998713e-01
-1.05841947e+00 1.27884471e+00 6.42942309e-01 3.65645498e-01
1.30727386e+00 -3.73467326e-01 5.28427482e-01 3.18948627e-01
1.67160615e-01 -9.71508741e-01 -4.38336283e-02 1.69525743e-01
1.17945445e+00 -1.49393201e+00 -1.00259572e-01 -1.00259222e-01
-9.71817374e-01 1.19749665e+00 8.37529302e-01 -7.75837302e-02
1.82293087e-01 2.89275981e-02 7.76455700e-02 1.05216816e-01
-9.13284898e-01 1.42558590e-01 2.28821546e-01 2.41276741e-01
6.38052225e-01 -1.59780428e-01 -5.28016150e-01 1.08566201e+00
-1.51216000e-01 -9.31746662e-02 4.96215284e-01 7.79495120e-01
-3.57724607e-01 -1.05534554e+00 -5.50433278e-01 5.95534027e-01
-7.53666282e-01 3.56196240e-03 -5.15575446e-02 4.50125992e-01
3.39805037e-02 7.63856769e-01 1.55785978e-01 -2.97644109e-01
4.16170835e-01 1.23663776e-01 5.72656989e-02 -3.24403644e-01
-8.55882704e-01 -1.50565848e-01 -1.35889262e-01 -2.37431467e-01
-1.82858184e-01 -5.68284869e-01 -1.16084456e+00 -1.80325136e-01
-2.77222961e-01 2.29568496e-01 6.32903755e-01 8.61103475e-01
3.13082457e-01 7.22430348e-01 8.65960777e-01 -4.53615338e-01
-8.59437823e-01 -1.17344236e+00 -4.15723264e-01 9.45546150e-01
4.56447542e-01 -6.24125302e-01 -2.52974272e-01 -5.37538715e-02] | [9.299490928649902, 5.244102478027344] |
68692de7-b84f-4453-b9f6-66944987e43b | learning-spatial-attention-for-face-super | 2012.01211 | null | https://arxiv.org/abs/2012.01211v2 | https://arxiv.org/pdf/2012.01211v2.pdf | Learning Spatial Attention for Face Super-Resolution | General image super-resolution techniques have difficulties in recovering detailed face structures when applying to low resolution face images. Recent deep learning based methods tailored for face images have achieved improved performance by jointly trained with additional task such as face parsing and landmark prediction. However, multi-task learning requires extra manually labeled data. Besides, most of the existing works can only generate relatively low resolution face images (e.g., $128\times128$), and their applications are therefore limited. In this paper, we introduce a novel SPatial Attention Residual Network (SPARNet) built on our newly proposed Face Attention Units (FAUs) for face super-resolution. Specifically, we introduce a spatial attention mechanism to the vanilla residual blocks. This enables the convolutional layers to adaptively bootstrap features related to the key face structures and pay less attention to those less feature-rich regions. This makes the training more effective and efficient as the key face structures only account for a very small portion of the face image. Visualization of the attention maps shows that our spatial attention network can capture the key face structures well even for very low resolution faces (e.g., $16\times16$). Quantitative comparisons on various kinds of metrics (including PSNR, SSIM, identity similarity, and landmark detection) demonstrate the superiority of our method over current state-of-the-arts. We further extend SPARNet with multi-scale discriminators, named as SPARNetHD, to produce high resolution results (i.e., $512\times512$). We show that SPARNetHD trained with synthetic data cannot only produce high quality and high resolution outputs for synthetically degraded face images, but also show good generalization ability to real world low quality face images. | ['Kwan-Yee K. Wong', 'Zhifeng Li', 'Hao Wang', 'Dihong Gong', 'Chaofeng Chen'] | 2020-12-02 | null | null | null | null | ['face-parsing'] | ['computer-vision'] | [ 1.79733053e-01 7.41562396e-02 9.50302184e-02 -4.17375475e-01
-7.72869468e-01 -1.88432917e-01 2.32886493e-01 -7.25917101e-01
-2.74496786e-02 6.35007143e-01 2.56242841e-01 2.35563502e-01
-7.96037316e-02 -1.03165627e+00 -8.43036294e-01 -5.14822423e-01
-1.00928932e-01 1.42788410e-01 1.44097209e-03 -3.98238182e-01
-9.35370252e-02 8.76162112e-01 -1.80631673e+00 4.34673131e-01
6.14453733e-01 9.66378391e-01 2.93878078e-01 2.81265169e-01
3.70382108e-02 5.08954048e-01 -5.39231539e-01 -5.14578879e-01
4.31473553e-01 -4.71972257e-01 -6.57666981e-01 -1.36537969e-01
9.06463623e-01 -8.58333170e-01 -4.06199098e-01 1.07900655e+00
7.49532759e-01 -1.22974487e-02 2.86525071e-01 -9.19730783e-01
-1.07058549e+00 3.73422027e-01 -1.02591050e+00 3.45692515e-01
4.53939646e-01 3.48541290e-02 7.06274271e-01 -1.34624124e+00
6.18098021e-01 1.69616044e+00 6.54795468e-01 1.03284502e+00
-1.20180774e+00 -1.17720592e+00 -4.53506261e-02 2.06044897e-01
-1.65640342e+00 -8.75148237e-01 6.88009501e-01 -8.17088112e-02
7.13405728e-01 1.56892985e-01 -1.16744600e-02 1.03179455e+00
-1.99166965e-03 2.25252673e-01 1.15258670e+00 4.23988849e-02
-2.12812245e-01 2.75127068e-02 -4.88310695e-01 9.28520739e-01
1.96215093e-01 1.75037771e-01 -4.41637695e-01 2.55746394e-01
1.53068721e+00 7.61358663e-02 -4.76732552e-01 5.45577146e-02
-7.30305254e-01 7.09767342e-01 7.58551180e-01 3.60752255e-01
-2.70394295e-01 -1.56324208e-02 9.80105400e-02 1.13232084e-01
4.65215981e-01 1.60810336e-01 -3.46499950e-01 5.36384344e-01
-9.76393104e-01 -3.70082818e-02 1.00017257e-01 9.11397159e-01
9.40048039e-01 4.18242961e-01 -3.65762889e-01 1.22529459e+00
1.61902651e-01 3.77771348e-01 2.47703984e-01 -1.08043873e+00
3.98341358e-01 4.36150372e-01 3.53692770e-02 -9.63255167e-01
-3.12646776e-01 -3.55976552e-01 -1.23802173e+00 4.93365288e-01
1.92019165e-01 -1.47431433e-01 -1.09600496e+00 1.91878247e+00
2.77376622e-01 3.92595738e-01 8.28478113e-02 1.02316678e+00
1.38975894e+00 5.42448580e-01 5.08742221e-03 -3.82269502e-01
1.46041250e+00 -8.13573837e-01 -7.08856404e-01 -2.22364858e-01
-1.67692125e-01 -6.64463937e-01 1.05536664e+00 -5.58832511e-02
-1.40371835e+00 -1.01332045e+00 -8.99029911e-01 -2.28797257e-01
-1.68432295e-01 1.57061651e-01 4.70061660e-01 6.29388869e-01
-1.45323932e+00 6.18432641e-01 -4.41866785e-01 -1.26423180e-01
9.95649695e-01 6.26433969e-01 -6.91231310e-01 -3.74412358e-01
-1.14782906e+00 6.30819380e-01 -1.61740452e-01 1.28794894e-01
-1.12829220e+00 -8.27284098e-01 -9.84729230e-01 3.37374240e-01
2.88810223e-01 -4.24946517e-01 7.37416685e-01 -8.93417716e-01
-1.38384640e+00 8.25893402e-01 -3.18825811e-01 -3.68702635e-02
1.93938315e-01 -2.60552950e-02 -6.28003657e-01 4.52409148e-01
2.03430742e-01 9.63609517e-01 1.07713008e+00 -1.47020650e+00
-5.30891538e-01 -5.48832536e-01 1.27708897e-01 -3.01656401e-04
-4.05900031e-01 4.30986315e-01 -6.77028596e-01 -7.69921660e-01
4.25973684e-02 -4.17415798e-01 1.47173889e-02 9.18284655e-02
-4.64910120e-02 3.50496769e-02 9.02125597e-01 -9.07150984e-01
9.57026899e-01 -2.02842259e+00 5.85313663e-02 -1.34564817e-01
2.04314083e-01 3.53569001e-01 -4.66339469e-01 -3.18900108e-01
-3.31934005e-01 4.12925512e-01 -1.69541687e-01 -3.28849018e-01
-3.95551801e-01 9.37756058e-03 -2.02807426e-01 2.87304044e-01
4.80928302e-01 9.08501804e-01 -4.26654994e-01 -4.37572181e-01
1.10475235e-01 1.05617595e+00 -7.71767259e-01 2.06232026e-01
4.76556569e-02 6.32911265e-01 -2.77126372e-01 8.96327257e-01
1.14639962e+00 -2.55284250e-01 5.97911589e-02 -5.46676457e-01
8.29808265e-02 -2.18462080e-01 -1.12253451e+00 1.51147985e+00
-6.46646023e-01 4.04329896e-01 5.16931534e-01 -5.77082157e-01
9.59639490e-01 3.54619056e-01 4.13275063e-01 -1.11596596e+00
-9.66964737e-02 -6.57578409e-02 -1.36602312e-01 -2.82582790e-01
3.40027779e-01 -2.22009897e-01 2.94809699e-01 2.93611318e-01
2.32362255e-01 3.61924648e-01 -4.62144986e-02 -3.17875370e-02
7.54570723e-01 1.00685805e-01 3.67214270e-02 -5.23764007e-02
6.75470114e-01 -7.30612516e-01 7.08018839e-01 2.63971299e-01
-1.03771687e-01 1.02755702e+00 3.76000106e-01 -4.19320613e-01
-1.11718798e+00 -9.58740234e-01 -3.55538309e-01 1.28038943e+00
2.81341523e-01 -2.16524854e-01 -9.31791127e-01 -5.98653555e-01
-2.09726334e-01 3.74041460e-02 -7.89639890e-01 7.09214509e-02
-8.73608768e-01 -7.76319504e-01 4.69886273e-01 7.08051801e-01
9.77882743e-01 -1.45497620e+00 -2.88444251e-01 -3.32360156e-02
-5.95865697e-02 -1.34891593e+00 -4.58819479e-01 -5.34546375e-01
-6.83610916e-01 -9.45117354e-01 -8.92812192e-01 -9.82721508e-01
7.68524349e-01 3.80653411e-01 1.08756316e+00 2.87798762e-01
-5.50283670e-01 6.58778846e-02 -7.14569762e-02 1.94347963e-01
-5.34162074e-02 -3.29356164e-01 1.71422437e-01 1.61404714e-01
7.02992603e-02 -6.03361487e-01 -8.54989827e-01 5.12726247e-01
-7.82185197e-01 -8.08932558e-02 8.15179110e-01 9.58147049e-01
8.09542894e-01 2.28105158e-01 8.00753355e-01 -7.49993265e-01
2.32858911e-01 -3.02815735e-01 -4.59992409e-01 2.21016020e-01
-3.48263621e-01 1.17071429e-02 8.08470964e-01 -2.83980936e-01
-1.38957584e+00 3.01829036e-02 -3.12851071e-01 -6.86127424e-01
-7.95772225e-02 -1.77116930e-01 -5.96861422e-01 -3.78604203e-01
5.52740097e-01 1.41638175e-01 -2.14392431e-02 -6.38950408e-01
1.86316252e-01 4.01746720e-01 5.86670637e-01 -3.52575511e-01
9.17995632e-01 5.84218144e-01 3.40130590e-02 -6.55310154e-01
-6.09267116e-01 2.69647539e-01 -4.73100960e-01 3.22377682e-02
7.60905027e-01 -1.15042901e+00 -9.09125388e-01 1.76072776e-01
-8.69628668e-01 -2.12920472e-01 -3.39317098e-02 1.73288986e-01
-4.45899099e-01 8.10237899e-02 -8.59507799e-01 -6.34939373e-01
-4.99730468e-01 -1.25079334e+00 1.21792209e+00 5.45334995e-01
3.38639289e-01 -5.32690048e-01 -4.98479366e-01 4.58632827e-01
6.22480333e-01 2.44127184e-01 8.38059068e-01 3.70260216e-02
-7.22585917e-01 3.14760089e-01 -8.11535060e-01 2.44778156e-01
2.10022777e-01 -1.15136147e-01 -1.24648583e+00 -6.07372046e-01
-1.31832227e-01 -2.65261620e-01 9.55239713e-01 3.67313206e-01
1.50981438e+00 -4.91895944e-01 -3.19531471e-01 1.04275715e+00
1.58603096e+00 1.11225039e-01 1.02231348e+00 -1.28253356e-01
8.86441827e-01 7.56715953e-01 4.57035810e-01 2.14102179e-01
1.55997768e-01 8.83317232e-01 5.52976072e-01 -5.02632856e-01
-6.09534323e-01 -2.43129507e-01 3.35945010e-01 1.48039833e-01
-4.09663469e-01 1.85086489e-01 -4.72201347e-01 3.97871524e-01
-1.17281878e+00 -1.07916021e+00 3.51697147e-01 2.05797648e+00
8.02384138e-01 -2.63452023e-01 -3.74033302e-02 -6.86390400e-02
1.00256813e+00 1.80238396e-01 -6.43180132e-01 -1.51052043e-01
-1.94673300e-01 8.69668901e-01 2.39186004e-01 5.02015173e-01
-9.46624994e-01 1.17243695e+00 5.84401751e+00 9.30396795e-01
-1.09903789e+00 3.15705806e-01 9.75104690e-01 -2.23041534e-01
-2.17782259e-01 -5.79113245e-01 -1.01824844e+00 3.57676387e-01
7.26019919e-01 1.08374089e-01 6.20078325e-01 7.95829952e-01
-6.28992319e-02 2.90051907e-01 -9.25633013e-01 1.31412709e+00
3.14473540e-01 -1.27773714e+00 2.87145823e-01 3.56555432e-02
8.53087366e-01 -4.10195261e-01 5.16010463e-01 2.85388738e-01
1.07919194e-01 -1.81237805e+00 2.57278800e-01 3.26739639e-01
1.68531525e+00 -1.06195140e+00 5.49833715e-01 -1.97700545e-01
-1.58018565e+00 -2.82604128e-01 -7.38767684e-01 4.86992925e-01
-8.09862465e-02 1.06020860e-01 -4.59228039e-01 4.74061877e-01
1.13510811e+00 5.46891570e-01 -7.07375288e-01 4.96595562e-01
-1.19849987e-01 -2.76231226e-02 -1.63915176e-02 7.30078101e-01
-2.45635465e-01 7.90513605e-02 2.01916441e-01 8.25493753e-01
6.48954809e-01 4.22725469e-01 -2.91205436e-01 1.01279044e+00
-5.77616096e-01 5.24665117e-02 -5.87930858e-01 3.98058295e-01
4.48241025e-01 1.48772013e+00 -5.75448811e-01 -1.06390923e-01
-6.07695699e-01 1.10956299e+00 4.99441326e-01 4.27745819e-01
-8.56791556e-01 -2.17230067e-01 1.07696950e+00 3.66085410e-01
6.44199133e-01 1.58772886e-01 -1.42667547e-01 -9.17215526e-01
-6.64040912e-03 -9.59498227e-01 3.10815245e-01 -7.47007966e-01
-1.10110581e+00 1.15737164e+00 -2.35198200e-01 -9.62537408e-01
-2.00694606e-01 -4.84684289e-01 -4.10357654e-01 1.10938418e+00
-1.59191620e+00 -1.25727463e+00 -5.92229545e-01 9.94913876e-01
5.87869585e-01 -3.60771954e-01 8.73870134e-01 5.59181452e-01
-7.97503710e-01 1.19957721e+00 -2.14081332e-01 3.37081492e-01
7.50109076e-01 -7.68289566e-01 4.03875262e-01 7.11928725e-01
3.38954479e-02 6.64424837e-01 3.32415789e-01 -4.19181049e-01
-1.30294979e+00 -1.36193883e+00 3.15886050e-01 -2.95201868e-01
-8.53288844e-02 -4.47121918e-01 -1.13588631e+00 6.08393192e-01
-7.64583573e-02 5.23577750e-01 3.13900024e-01 -1.45235017e-01
-5.50021410e-01 -5.06860197e-01 -1.76964080e+00 5.86412251e-01
1.41224694e+00 -5.69021821e-01 -7.84911141e-02 6.99287355e-02
5.59224725e-01 -3.52896243e-01 -9.70794320e-01 7.90055215e-01
5.88808954e-01 -1.17847407e+00 1.43505049e+00 -3.02852571e-01
6.65606737e-01 -2.98924834e-01 -1.28142059e-01 -9.26911592e-01
-5.57710886e-01 -3.27136844e-01 -1.80661678e-01 1.39404011e+00
2.00052351e-01 -4.96208459e-01 7.42308021e-01 2.97423035e-01
1.03618048e-01 -7.33355761e-01 -1.03798592e+00 -4.10603970e-01
-7.17771053e-03 2.57079571e-01 9.90035594e-01 8.60449433e-01
-5.85128188e-01 2.41783947e-01 -5.47006011e-01 4.68149006e-01
8.48019540e-01 2.12798357e-01 4.23759639e-01 -1.15918207e+00
6.65801764e-02 -3.28774095e-01 -3.34288239e-01 -7.99105227e-01
3.27116936e-01 -6.37844145e-01 -2.42119610e-01 -1.35172880e+00
3.20509821e-01 -4.24574643e-01 -1.59591973e-01 6.23834968e-01
-2.04860821e-01 1.04231966e+00 -7.72077292e-02 1.09939270e-01
-2.74037391e-01 4.79954034e-01 1.59080708e+00 3.20154093e-02
-9.07228962e-02 -3.11568081e-01 -1.06669605e+00 6.32296145e-01
7.10709035e-01 3.68296728e-02 -2.69931942e-01 -5.52104115e-01
-2.69277275e-01 2.35949859e-01 4.26594615e-01 -1.06595576e+00
2.88523547e-02 -3.37917954e-02 1.18155468e+00 -2.31253102e-01
5.53609073e-01 -6.43497765e-01 3.16755891e-01 1.27289116e-01
-6.43925443e-02 -1.38842493e-01 4.22169954e-01 2.57406741e-01
-2.05859274e-01 2.44619995e-01 1.27785611e+00 -2.34203622e-01
-8.12917292e-01 8.35352659e-01 3.35882157e-01 -2.29563564e-01
1.03624737e+00 -2.73713380e-01 -3.96839112e-01 -3.07906508e-01
-7.66885698e-01 -1.79566965e-02 5.20137310e-01 6.16146207e-01
9.91389334e-01 -1.52062213e+00 -9.65846658e-01 6.34347558e-01
-2.11062655e-01 8.68323201e-04 7.63600409e-01 2.47316346e-01
-3.83925527e-01 3.39435846e-01 -7.61514127e-01 -3.22903633e-01
-1.55201828e+00 7.96620369e-01 4.54597831e-01 1.79846399e-02
-6.75942838e-01 9.26145136e-01 7.54804790e-01 -6.53463677e-02
7.65896663e-02 1.23130605e-01 -6.69889390e-01 3.02290078e-02
1.05571187e+00 2.41798267e-01 -7.89152831e-02 -1.13153636e+00
-4.61619377e-01 1.05549872e+00 -2.26359546e-01 1.56411305e-01
1.48747241e+00 -1.50551096e-01 -1.41624868e-01 -5.20953178e-01
1.19050169e+00 1.12306483e-01 -1.64677036e+00 -2.44619459e-01
-5.84557176e-01 -8.28244269e-01 -1.56810402e-03 -6.91281140e-01
-1.76694906e+00 9.19241488e-01 9.82252896e-01 -2.72223979e-01
1.56847179e+00 1.34785920e-01 5.91173589e-01 -3.00710648e-01
6.39544725e-01 -6.94360673e-01 2.87498564e-01 1.93250999e-01
1.26124525e+00 -1.34868324e+00 -1.86959992e-03 -5.96575975e-01
-4.44924384e-01 8.90881777e-01 1.21652222e+00 3.41110229e-02
3.86743635e-01 2.83201665e-01 -1.65304273e-01 -1.95138410e-01
-5.56569040e-01 -3.62039119e-01 2.29120731e-01 7.79423952e-01
4.41715628e-01 -1.75422594e-01 1.21695518e-01 6.93161368e-01
-1.67002961e-01 -2.73441970e-01 2.57612795e-01 2.90706813e-01
-3.08688134e-01 -9.06749904e-01 -5.22422075e-01 3.49406153e-01
-8.64256620e-01 -1.93180785e-01 2.25167591e-02 6.82130039e-01
3.79131049e-01 8.17758918e-01 1.70881778e-01 -3.97747129e-01
3.00901771e-01 -2.91171968e-01 7.08012640e-01 -6.98237062e-01
-4.25606281e-01 -5.89042902e-02 -3.63128781e-01 -7.72650778e-01
-2.84421742e-01 -1.51768595e-01 -1.16653490e+00 -5.89484751e-01
7.60242790e-02 -1.95494428e-01 3.57553214e-01 3.93089741e-01
4.85458076e-01 7.21056998e-01 8.47573698e-01 -1.05597448e+00
-1.47120543e-02 -1.07849777e+00 -5.49691856e-01 3.65699261e-01
4.26842481e-01 -7.77675807e-01 -1.70378745e-01 -9.32284296e-02] | [12.800127029418945, -0.07248575985431671] |
dd8e83b5-8f16-4f56-9bfa-87f2c6d5bb9b | a-fuzzy-expert-system-for-earthquake | 1610.04028 | null | http://arxiv.org/abs/1610.04028v2 | http://arxiv.org/pdf/1610.04028v2.pdf | A fuzzy expert system for earthquake prediction, case study: the Zagros range | A methodology for the development of a fuzzy expert system (FES) with
application to earthquake prediction is presented. The idea is to reproduce the
performance of a human expert in earthquake prediction. To do this, at the
first step, rules provided by the human expert are used to generate a fuzzy
rule base. These rules are then fed into an inference engine to produce a fuzzy
inference system (FIS) and to infer the results. In this paper, we have used a
Sugeno type fuzzy inference system to build the FES. At the next step, the
adaptive network-based fuzzy inference system (ANFIS) is used to refine the FES
parameters and improve its performance. The proposed framework is then employed
to attain the performance of a human expert used to predict earthquakes in the
Zagros area based on the idea of coupled earthquakes. While the prediction
results are promising in parts of the testing set, the general performance
indicates that prediction methodology based on coupled earthquakes needs more
investigation and more complicated reasoning procedure to yield satisfactory
predictions. | ['Farid Atry', 'Mehdi Zare', 'Arash Andalib'] | 2016-10-13 | null | null | null | null | ['earthquake-prediction'] | ['computer-vision'] | [-2.68483534e-02 1.55227274e-01 6.46415532e-01 -2.88086504e-01
2.24456385e-01 -4.33109477e-02 1.69173390e-01 2.73544520e-01
-2.98174083e-01 7.83014834e-01 -2.34382600e-01 -4.95845616e-01
-5.63058197e-01 -1.14388204e+00 -1.82245985e-01 -4.82376963e-01
1.65751114e-01 6.65827453e-01 5.88201344e-01 -9.33960974e-01
7.25890458e-01 8.63808692e-01 -1.69991267e+00 3.98291528e-01
9.62652087e-01 6.67557418e-01 5.19127965e-01 4.60936695e-01
2.08697826e-01 9.78318691e-01 -6.16074383e-01 7.31502473e-02
-9.39574987e-02 -2.59716511e-01 -7.83605814e-01 4.11244389e-03
-7.11895287e-01 -3.26405674e-01 1.94605872e-01 8.17707360e-01
2.67148167e-01 5.66603184e-01 1.10489535e+00 -7.62104332e-01
-3.16147923e-01 5.45112491e-01 2.69570321e-01 1.42755419e-01
4.70340997e-01 -1.99190766e-01 3.46070737e-01 -8.87191415e-01
2.46810973e-01 1.01780140e+00 7.05166101e-01 -1.02380894e-01
-6.72466815e-01 -1.87513590e-01 -4.83679622e-01 5.12113035e-01
-1.72545946e+00 -4.00150344e-02 7.34021664e-01 -5.59703946e-01
9.92119491e-01 -2.26758863e-03 7.88939595e-01 6.77708760e-02
3.56255740e-01 -1.85924217e-01 9.79433656e-01 -7.98471391e-01
5.57325661e-01 4.77313221e-01 9.76432785e-02 3.81093472e-01
4.80462044e-01 1.87040374e-01 3.31309110e-01 1.05839223e-01
6.04891360e-01 -1.90136433e-02 3.68743055e-02 5.86639643e-01
-2.96312302e-01 9.15277481e-01 3.88647467e-01 1.18418849e+00
-9.46163416e-01 -5.64291954e-01 2.79808849e-01 1.16660476e-01
5.10633327e-02 6.22641742e-01 -2.45325521e-01 2.33483329e-01
-1.27513099e+00 3.57570827e-01 1.16173947e+00 1.40087247e-01
6.14251018e-01 4.65699136e-01 3.70798945e-01 5.75770855e-01
5.77363193e-01 4.59951818e-01 5.22583961e-01 -7.18202591e-01
6.89632148e-02 8.81434798e-01 3.45267326e-01 -1.54757881e+00
-3.69499087e-01 -1.59187168e-01 -4.76328641e-01 6.84513450e-01
2.24257603e-01 -4.52065408e-01 -6.65974855e-01 8.95697594e-01
2.64273155e-02 1.88730527e-02 4.88814265e-01 7.02230453e-01
4.97661620e-01 9.50465798e-01 -1.94795638e-01 -2.08143935e-01
1.19276333e+00 -2.63696194e-01 -7.30716586e-01 3.16738605e-01
4.65518475e-01 -6.18180752e-01 5.57953835e-01 7.73595989e-01
-9.00635779e-01 -9.00390029e-01 -1.05731142e+00 6.39021277e-01
-6.55771017e-01 3.47192764e-01 5.01853526e-02 4.31323022e-01
-1.03288257e+00 5.48500717e-01 -7.21916854e-01 -4.00042713e-01
-3.11465800e-01 5.25383353e-01 8.76208246e-02 2.84693569e-01
-1.64077342e+00 1.50171983e+00 1.00210559e+00 5.02197862e-01
-3.10577065e-01 -1.05413318e-01 -8.13479841e-01 3.70504737e-01
1.31795108e-01 -4.12352324e-01 8.42915654e-01 -7.92145789e-01
-1.40049708e+00 2.22393461e-02 2.02021345e-01 -6.23910248e-01
1.86514586e-01 3.11019868e-01 -7.63259828e-01 7.03042448e-01
2.28683986e-02 2.06403255e-01 2.46347770e-01 -1.30424750e+00
-7.33783424e-01 -6.33983836e-02 -5.11661246e-02 1.05994619e-01
-1.35948658e-01 -1.54456124e-01 1.56717405e-01 -4.17172998e-01
-4.01984006e-02 -6.41896307e-01 -3.24080378e-01 -9.26168919e-01
1.25239521e-01 -3.70088816e-01 6.33150041e-01 -9.28796172e-01
1.61702657e+00 -1.80831325e+00 -3.67332041e-01 1.14773989e+00
-4.23160315e-01 6.33995533e-01 6.26845002e-01 6.56119168e-01
-8.11369345e-02 -8.85242894e-02 -2.85539865e-01 7.03062057e-01
-3.62551212e-01 4.34633791e-01 -3.19515467e-02 -3.40750307e-01
2.07423434e-01 -5.98312952e-02 -6.95426822e-01 -7.30187535e-01
6.01286232e-01 3.82071137e-01 -5.55479646e-01 1.10491745e-01
8.18768293e-02 1.28577679e-01 -8.63943279e-01 3.16692740e-01
3.41401249e-01 2.58839220e-01 3.50399852e-01 -3.74239117e-01
-3.55720520e-01 -1.30980089e-01 -1.52135563e+00 5.75346351e-01
-3.57489735e-01 2.96176635e-02 -1.44040361e-01 -1.60649419e+00
1.72628498e+00 1.03195560e+00 4.77947176e-01 -3.05409938e-01
3.87086779e-01 5.60338438e-01 1.55999333e-01 -1.11146641e+00
4.81342971e-01 -3.44269305e-01 3.09114188e-01 8.72687250e-02
-8.59004855e-02 -5.38683414e-01 7.57070184e-01 -3.30726504e-01
4.55358475e-01 -1.92747135e-02 4.41680551e-01 -3.93036485e-01
1.10709596e+00 3.81586313e-01 2.40420088e-01 4.32827830e-01
3.25102657e-01 -3.43817472e-02 5.96159399e-02 -6.67860210e-01
-9.93892908e-01 -4.96842474e-01 -1.93392962e-01 3.52402776e-01
-2.19347998e-01 1.09302305e-01 -8.72676730e-01 8.95608515e-02
-4.48164999e-01 9.86231863e-01 -1.69267029e-01 4.73366715e-02
-3.48054618e-01 -6.29057288e-01 3.61045659e-01 2.09844828e-01
5.10791659e-01 -1.28590977e+00 -9.05364990e-01 7.81489968e-01
-1.73957318e-01 -4.94339705e-01 5.61251819e-01 8.69760960e-02
-9.23442423e-01 -8.43999803e-01 -2.92649508e-01 -8.90289843e-01
6.10770822e-01 -3.43436390e-01 5.69629550e-01 5.70791006e-01
2.41364568e-01 8.12985897e-02 -6.68408573e-01 -5.63556671e-01
-9.18775320e-01 -4.22989838e-02 -1.31752476e-01 -2.10836351e-01
2.07679823e-01 -4.79724705e-01 -1.82857975e-01 2.32566178e-01
-1.20270026e+00 -3.29525173e-01 5.30983806e-01 5.33289909e-01
1.47200987e-01 1.22381163e+00 1.18445957e+00 -5.34590244e-01
8.46594036e-01 -5.09826660e-01 -7.35026300e-01 4.05624837e-01
-7.79259145e-01 -7.92048406e-03 9.59942162e-01 -1.17286876e-01
-1.55840826e+00 -1.16097309e-01 -5.83194852e-01 3.88986804e-02
-4.72899318e-01 1.21620703e+00 9.88557115e-02 -1.00112088e-01
8.72666895e-01 6.15239106e-02 1.34800062e-01 -2.82471508e-01
-5.50948262e-01 9.78542030e-01 4.77241993e-01 -4.15423661e-01
3.82841319e-01 -1.16851397e-01 2.28377372e-01 -6.67534351e-01
-2.44750351e-01 -6.74960539e-02 -6.24655902e-01 -6.76338136e-01
8.94250572e-01 -4.51926172e-01 -7.97393203e-01 2.10930765e-01
-1.01303899e+00 4.64618169e-02 1.30593315e-01 8.49822402e-01
-4.57025677e-01 2.15046093e-01 -5.51309705e-01 -1.19342351e+00
-2.59039968e-01 -1.07137954e+00 7.67138675e-02 4.30943042e-01
-3.25496733e-01 -1.27212989e+00 -5.31131737e-02 2.81283706e-01
1.55248225e-01 1.96572021e-01 8.63482237e-01 -9.02312934e-01
3.44743393e-02 -4.44032401e-01 3.56140792e-01 6.34801507e-01
1.31399959e-01 4.97817248e-01 -2.89901674e-01 2.44463876e-01
3.38908046e-01 1.23789996e-01 2.79433489e-01 3.19145173e-01
6.55201554e-01 -2.16977641e-01 -1.21889964e-01 -1.27132520e-01
1.69887841e+00 1.08227265e+00 8.92868102e-01 4.72327858e-01
3.67489979e-02 6.04707420e-01 1.10233700e+00 6.79269791e-01
4.44284618e-01 4.11113471e-01 8.37783068e-02 1.45518273e-01
4.70369995e-01 3.29342514e-01 1.78347126e-01 6.64519429e-01
-7.67635465e-01 -1.18274689e-01 -1.08802676e+00 4.33535904e-01
-1.62101877e+00 -1.33793366e+00 -2.90359825e-01 1.93370080e+00
6.12059295e-01 3.27571929e-01 1.23080134e-01 1.09242737e+00
8.79401982e-01 -6.14872873e-01 4.47890431e-01 -1.14058030e+00
1.80048078e-01 4.17390853e-01 4.77850363e-02 7.98463404e-01
-6.66572094e-01 6.04441285e-01 5.36380386e+00 7.43292391e-01
-1.35753739e+00 -5.20737529e-01 2.35773668e-01 6.96094811e-01
-5.36032133e-02 1.47327827e-02 -5.10823071e-01 6.07609987e-01
1.04446685e+00 -1.30339980e-01 6.76717639e-01 5.26300311e-01
8.30980837e-01 -5.78087926e-01 -2.01891273e-01 2.43531570e-01
-5.48450649e-02 -1.06973994e+00 2.36800626e-01 -2.02239692e-01
7.23720372e-01 -6.01671696e-01 -6.18281603e-01 8.41902196e-02
3.75994630e-02 -7.31159389e-01 4.93686765e-01 1.42062986e+00
-5.57768047e-02 -1.15022779e+00 1.44383752e+00 4.90742147e-01
-1.01709723e+00 -3.94511759e-01 -4.50755656e-01 -5.46391428e-01
2.78455794e-01 5.18816411e-01 -1.44658983e+00 1.03183532e+00
4.84693021e-01 2.00090017e-02 -3.83155644e-01 9.13129032e-01
1.03166275e-01 7.53537118e-01 -5.15494406e-01 -2.22309247e-01
3.32878530e-01 -4.37877327e-01 2.41666451e-01 9.95755970e-01
8.81690145e-01 5.19322276e-01 -1.16759511e-02 7.75279880e-01
8.62477839e-01 3.08310717e-01 -4.36092943e-01 7.79624954e-02
7.05785334e-01 9.59874690e-01 -9.34219718e-01 -7.78580964e-01
-1.29566610e-01 3.56366187e-01 -1.41479433e-01 2.74974614e-01
-5.97486079e-01 -6.86276913e-01 -5.02790749e-01 3.62129331e-01
4.91835654e-01 1.76093668e-01 -5.73414981e-01 -5.82110167e-01
-3.69139910e-01 -6.82362974e-01 4.55942214e-01 -1.17249620e+00
-9.18560266e-01 7.74019837e-01 4.70318466e-01 -1.06281137e+00
-9.12553608e-01 -6.79294884e-01 -9.81550872e-01 1.19128978e+00
-6.47235692e-01 -7.88005710e-01 -1.52312024e-02 6.10500634e-01
1.43434823e-01 -3.04990172e-01 8.36824894e-01 2.52553344e-01
-2.24615633e-01 -1.20144352e-01 -1.13293409e-01 3.49531113e-03
2.08716616e-01 -9.84028876e-01 -7.85740197e-01 1.01804352e+00
-6.49128616e-01 3.57769281e-01 9.81292307e-01 -1.06640041e+00
-3.46135795e-01 -6.08845413e-01 1.35698271e+00 2.84595460e-01
5.15307963e-01 7.50620008e-01 -1.02755201e+00 3.02419633e-01
-1.31891994e-03 -4.52589482e-01 3.40087533e-01 -5.36603749e-01
6.14624321e-01 -2.80599445e-01 -1.47802246e+00 4.12581325e-01
-1.82100683e-01 -2.68998086e-01 -1.30603528e+00 -2.81084210e-01
-4.10363004e-02 1.49402656e-02 -1.52314436e+00 5.80299616e-01
4.88878161e-01 -9.99524057e-01 8.14472854e-01 -5.51032573e-02
2.97859371e-01 -7.86340833e-01 6.87264800e-02 -1.12419176e+00
-3.68968278e-01 8.95693153e-02 6.32087827e-01 1.06502354e+00
6.09511197e-01 -7.58287907e-01 3.31611902e-01 5.43215215e-01
-4.72536683e-01 -6.64238095e-01 -5.56171179e-01 -2.43524373e-01
-3.22538435e-01 -2.03886986e-01 5.76583207e-01 8.18367183e-01
3.29346120e-01 -1.53428060e-03 -1.08410634e-01 3.87672842e-01
1.15572825e-01 -2.36070067e-01 1.67177379e-01 -1.38653243e+00
-1.69386268e-01 -3.07211161e-01 -4.40995753e-01 8.31016526e-02
-2.32319683e-01 -3.89973938e-01 7.91181773e-02 -1.66641963e+00
-6.52271032e-01 -2.60596871e-01 -2.78453320e-01 3.98240387e-01
-1.41313180e-01 2.34651968e-01 2.56588995e-01 7.87048414e-02
2.19825715e-01 -4.44368348e-02 1.04930663e+00 3.59883904e-01
-4.20057714e-01 5.29527605e-01 -4.58574057e-01 8.42352092e-01
9.10516202e-01 -3.14550191e-01 -5.52483380e-01 2.87574947e-01
4.75928545e-01 5.99833250e-01 2.58246303e-01 -1.35792458e+00
2.11574271e-01 -2.48323500e-01 4.82444644e-01 -8.49412799e-01
-4.08035703e-02 -1.33498561e+00 7.71010160e-01 9.15084720e-01
6.42725676e-02 5.73433489e-02 4.31241095e-03 -1.72537923e-01
-5.76069176e-01 -9.95484412e-01 4.91480082e-01 6.81717768e-02
-6.13916636e-01 -4.55605567e-01 -1.12797499e+00 -6.65668905e-01
9.95150805e-01 -7.35506475e-01 3.90990138e-01 -2.37051129e-01
-1.29344308e+00 -2.62817945e-02 1.40873820e-01 -3.70616376e-01
4.97101426e-01 -1.04811323e+00 -6.69286370e-01 7.05151781e-02
-2.73351640e-01 -2.15157896e-01 4.54833657e-01 8.97871614e-01
-1.18258929e+00 4.16183114e-01 -6.60075843e-01 -1.01603836e-01
-9.35775518e-01 4.86978173e-01 4.49644208e-01 -3.40331674e-01
-2.82275546e-02 2.67812610e-02 -6.15701675e-01 -1.17227770e-01
-3.84109348e-01 -4.46650654e-01 -1.19427812e+00 1.04097292e-01
3.64610463e-01 7.54396319e-01 2.64646173e-01 -7.71677315e-01
-2.26371869e-01 4.05834377e-01 7.11649239e-01 -3.67650807e-01
1.41324854e+00 -1.03893533e-01 -1.30799189e-01 2.32217744e-01
4.03227866e-01 4.87626018e-03 -8.29979897e-01 2.87457645e-01
1.07835405e-01 3.85634340e-02 2.78324503e-02 -1.01561105e+00
-7.14341044e-01 4.39711154e-01 2.95532286e-01 6.36357009e-01
1.53200924e+00 -5.66154063e-01 3.47305238e-01 7.55013347e-01
1.79031074e-01 -1.56632507e+00 -5.31663656e-01 5.24310768e-01
7.94867635e-01 -7.09279895e-01 -1.41999304e-01 -3.45531970e-01
-7.77375281e-01 1.80604064e+00 2.08439857e-01 -4.73228902e-01
1.02108479e+00 4.35076714e-01 -3.06877136e-01 -1.28212303e-01
-3.36372346e-01 -2.00328007e-01 2.83350736e-01 4.12438661e-01
3.24653357e-01 1.68335289e-01 -9.63314533e-01 1.03532088e+00
-2.72847831e-01 8.81810486e-01 6.49274707e-01 7.46258795e-01
-1.00548315e+00 -8.24943900e-01 -1.00260115e+00 4.58976716e-01
-4.90867138e-01 5.46533726e-02 -4.68469821e-02 8.52040887e-01
7.06160367e-01 1.42795503e+00 5.00425957e-02 -4.39983010e-01
2.76756883e-01 -3.95988859e-03 1.97419882e-01 -3.51303458e-01
-8.23395073e-01 5.38750328e-02 3.27508807e-01 2.68416911e-01
-5.13021350e-01 -5.11752665e-01 -1.73259842e+00 -3.40632975e-01
-3.58570516e-01 9.88278866e-01 5.83396494e-01 1.43525374e+00
-1.63355798e-01 6.14036024e-01 7.82431006e-01 -7.32621789e-01
-3.01585972e-01 -1.23833525e+00 -6.12140357e-01 3.30931604e-01
-4.57253814e-01 -6.58657670e-01 -2.90592611e-01 2.42587239e-01] | [6.0330810546875, 3.415147066116333] |
30d007a9-be45-4ab6-b628-9321ad0267e5 | an-efficient-cnn-for-spectral-reconstruction | 1804.04647 | null | http://arxiv.org/abs/1804.04647v1 | http://arxiv.org/pdf/1804.04647v1.pdf | An efficient CNN for spectral reconstruction from RGB images | Recently, the example-based single image spectral reconstruction from RGB
images task, aka, spectral super-resolution was approached by means of deep
learning by Galliani et al. The proposed very deep convolutional neural network
(CNN) achieved superior performance on recent large benchmarks. However,
Aeschbacher et al showed that comparable performance can be achieved by shallow
learning method based on A+, a method introduced for image super-resolution by
Timofte et al. In this paper, we propose a moderately deep CNN model and
substantially improve the reported performance on three spectral reconstruction
standard benchmarks: ICVL, CAVE, and NUS. | ['Radu Timofte', 'Yigit Baran Can'] | 2018-04-12 | null | null | null | null | ['spectral-reconstruction', 'spectral-super-resolution'] | ['computer-vision', 'computer-vision'] | [ 4.91291434e-01 -2.22859815e-01 -8.96161946e-04 -1.33546308e-01
-9.19139564e-01 -2.59581417e-01 5.76382160e-01 -5.52383065e-01
-3.89326125e-01 1.09759367e+00 2.23941773e-01 1.50661409e-01
-2.14171521e-02 -1.03630257e+00 -7.97467470e-01 -7.38215625e-01
1.86160624e-01 8.22687000e-02 1.83202669e-01 -4.07052487e-01
-1.52947843e-01 6.99682176e-01 -1.64791965e+00 5.78596532e-01
6.60929620e-01 1.19354701e+00 3.98153692e-01 7.12153077e-01
-6.70600533e-02 9.56799746e-01 -2.00633973e-01 4.31810021e-02
7.55608737e-01 -2.83545196e-01 -9.46590364e-01 -3.71547230e-02
9.00667846e-01 -4.70416516e-01 -3.48002166e-01 1.12343669e+00
7.72016883e-01 1.02782629e-01 2.79120505e-01 -4.12755221e-01
-8.99588704e-01 8.19130719e-01 -7.55307317e-01 1.89398285e-02
6.50660619e-02 2.48450369e-01 7.62112439e-01 -1.01547170e+00
4.55209404e-01 8.28166902e-01 1.38105714e+00 4.36578274e-01
-1.38457382e+00 -7.34646082e-01 -5.46089232e-01 4.58122373e-01
-1.59763849e+00 -2.13077381e-01 5.43214977e-01 -1.29572600e-01
1.25610065e+00 2.15664357e-01 7.11209774e-01 1.02979445e+00
-3.37059438e-01 3.37833554e-01 1.55766654e+00 -2.33381107e-01
-7.02816853e-03 -2.29086354e-01 -3.32388997e-01 5.39113939e-01
1.18320301e-01 4.39770073e-01 -7.05923617e-01 1.08355641e-01
1.15608966e+00 -3.45685571e-01 -4.58072454e-01 -1.56075228e-02
-1.08318853e+00 7.19766557e-01 1.11301219e+00 4.33723256e-02
-5.11996627e-01 2.33786762e-01 3.62056196e-01 1.09702922e-01
6.65869772e-01 3.42235297e-01 -5.81490815e-01 2.04786122e-01
-1.18346155e+00 2.34736711e-01 1.63048401e-01 7.02759326e-01
8.29038918e-01 6.74448907e-01 2.12460011e-01 8.49593639e-01
-1.16995731e-02 3.98708969e-01 3.34711343e-01 -1.34758925e+00
4.96210754e-02 2.20322981e-01 3.57795149e-01 -1.47466168e-01
-7.60191679e-01 -5.93987286e-01 -1.22419441e+00 5.71944594e-01
2.09838226e-01 -4.64547090e-02 -9.41820502e-01 1.27263033e+00
6.88421400e-03 3.05427074e-01 5.69437742e-01 1.32758009e+00
1.44594610e+00 7.26237297e-01 4.41994425e-03 -2.29386210e-01
9.93081272e-01 -1.02226913e+00 -6.39285088e-01 -8.32583979e-02
2.03281030e-01 -7.78467596e-01 1.07298315e+00 4.80770111e-01
-1.20230675e+00 -8.42887402e-01 -1.09559071e+00 -2.67311573e-01
-3.79655600e-01 1.69676527e-01 1.04436839e+00 6.82357430e-01
-1.27583897e+00 9.93380725e-01 -3.71372491e-01 -5.83032608e-01
7.19140232e-01 8.94269198e-02 -3.52624357e-01 1.80481821e-01
-1.12821460e+00 7.18822181e-01 8.34593832e-01 1.48129046e-01
-9.62702692e-01 -9.79477823e-01 -4.55642939e-01 -3.46040167e-02
2.47311905e-01 -7.70973861e-01 1.07842445e+00 -1.11068714e+00
-1.83838606e+00 9.85108554e-01 2.85121292e-01 -9.43313777e-01
3.34308296e-01 -1.00663736e-01 -5.26290178e-01 1.47371516e-01
-3.07091951e-01 5.90329051e-01 6.71034276e-01 -1.43039715e+00
-4.18968111e-01 -3.30737382e-01 -8.80990848e-02 3.09934676e-01
1.20456927e-01 1.25767767e-01 1.26363039e-01 -3.89541656e-01
2.09604934e-01 -6.43468797e-01 -2.35730767e-01 1.15446828e-01
-1.11463852e-01 2.77221620e-01 6.13694429e-01 -8.53810132e-01
4.52587575e-01 -1.95609689e+00 2.18855172e-01 -2.95863241e-01
2.06498370e-01 5.00962615e-01 -8.55678990e-02 1.04106992e-01
-3.99612606e-01 2.30713129e-01 -6.01859510e-01 -1.68113381e-01
-2.86704928e-01 1.55248865e-01 -1.34539664e-01 5.10098577e-01
2.80809961e-03 7.24523664e-01 -6.57057643e-01 -3.40629667e-02
5.35640419e-01 8.83008718e-01 -1.60645828e-01 1.91827103e-01
-1.81707755e-01 4.75112349e-01 4.11163270e-01 6.74296558e-01
1.25929224e+00 -2.40734830e-01 -2.11711321e-02 -7.85122871e-01
-6.35706007e-01 -1.90518647e-02 -1.26283109e+00 2.01449823e+00
-6.95559144e-01 6.97543144e-01 2.79466301e-01 -6.65651441e-01
8.38018119e-01 2.03431651e-01 4.82113540e-01 -9.31426823e-01
-2.07963333e-01 2.89381802e-01 -3.79373848e-01 -1.78967178e-01
6.70246661e-01 -3.40852320e-01 4.92672861e-01 -1.78188495e-02
1.78494781e-01 -4.25425500e-01 -4.01009619e-01 -1.43170133e-01
7.29284704e-01 7.10788310e-01 3.22188288e-01 -4.13170427e-01
5.07308722e-01 2.26072475e-01 3.69327098e-01 8.33842993e-01
-2.93220818e-01 1.21921337e+00 -6.76607415e-02 -8.85553360e-01
-1.59372818e+00 -1.18307781e+00 -4.59066182e-01 8.45318377e-01
-9.69759449e-02 -3.33802104e-01 -5.55489242e-01 -4.72552190e-03
-2.45537043e-01 3.77129376e-01 -7.20742106e-01 1.11438729e-01
-2.95790136e-01 -1.25086749e+00 7.25667775e-01 6.19442821e-01
1.40473843e+00 -1.09234953e+00 -6.39064431e-01 1.18350856e-01
-1.50984168e-01 -1.46890163e+00 4.30829763e-01 3.53341639e-01
-8.77013922e-01 -8.77564907e-01 -6.80186689e-01 -3.76780182e-01
-3.96520317e-01 2.68076003e-01 1.03732014e+00 -3.18493664e-01
-6.23576045e-01 -2.18932927e-01 -3.77247751e-01 -1.20182365e-01
-1.19294211e-01 6.67872000e-03 -2.23106053e-02 -1.27941847e-01
-3.04772835e-02 -6.94585085e-01 -8.50953281e-01 -7.26993456e-02
-7.33008265e-01 5.15892029e-01 4.45370406e-01 8.06089759e-01
1.02685237e+00 1.13928169e-01 4.42917109e-01 -7.69420326e-01
-5.86992353e-02 -8.53489339e-02 -6.12801909e-01 -3.40783484e-02
-7.46344388e-01 1.84203703e-02 7.02134728e-01 -1.56534575e-02
-1.56603456e+00 4.77904797e-01 -6.35486186e-01 -2.90102959e-01
-3.72886866e-01 2.59837419e-01 1.26004010e-01 -8.51698101e-01
1.12197828e+00 5.64120531e-01 -5.01731455e-01 -6.17349565e-01
4.50737476e-01 7.82806635e-01 9.59978878e-01 -1.06359057e-01
8.75795007e-01 8.57520282e-01 3.76655042e-01 -1.00244045e+00
-1.00891328e+00 -3.49075407e-01 -6.80516779e-01 -1.47339076e-01
1.11044109e+00 -1.49203670e+00 -4.40060347e-01 6.07083738e-01
-9.02150333e-01 -7.57449269e-01 -4.30268914e-01 3.56335670e-01
-7.68005133e-01 2.91636795e-01 -5.33200383e-01 -6.61271155e-01
-5.54365337e-01 -9.32000399e-01 1.20743930e+00 3.73015821e-01
4.81526256e-01 -7.38124073e-01 9.37538296e-02 3.63989234e-01
9.58382666e-01 6.52588010e-01 3.57684493e-01 2.81505078e-01
-6.49463952e-01 2.18910724e-01 -1.08528554e+00 4.96938556e-01
2.17759356e-01 -8.25637579e-02 -1.58543181e+00 -1.74780995e-01
-1.68542489e-01 -6.51712179e-01 1.31552589e+00 5.22123396e-01
1.59565377e+00 7.87322596e-02 4.02514726e-01 1.64425540e+00
2.22025156e+00 -4.96818304e-01 1.06283045e+00 5.94785273e-01
1.01052046e+00 6.26992285e-02 2.38848016e-01 5.68399847e-01
3.98171805e-02 5.77021778e-01 1.07234395e+00 -3.87803793e-01
-8.38575363e-01 3.34809840e-01 -1.09661490e-01 2.20905572e-01
-1.01304817e+00 2.91615188e-01 -5.80935001e-01 4.17096943e-01
-1.85770524e+00 -1.09002137e+00 -6.36974156e-01 2.00654221e+00
7.68159866e-01 -1.81588277e-01 -7.76932836e-02 3.30357216e-02
3.36827666e-01 5.56365371e-01 -5.13410568e-01 -8.38787779e-02
-8.69106054e-01 6.75357580e-01 1.20360053e+00 4.97990668e-01
-1.05601323e+00 1.33281898e+00 7.11430645e+00 8.09519708e-01
-1.00709009e+00 5.00446081e-01 2.03916281e-01 -2.19333187e-01
-1.65484160e-01 -2.46013686e-01 -4.68886256e-01 -4.62902486e-02
1.20559669e+00 2.11673230e-01 9.66474533e-01 4.84530777e-01
1.81904584e-01 -2.41157338e-01 -2.71609396e-01 1.43496132e+00
-6.31777197e-02 -1.62445700e+00 4.27437201e-02 -2.87403405e-01
1.06890178e+00 6.02679372e-01 1.83187947e-01 1.69164196e-01
2.72483289e-01 -1.37482941e+00 3.61536533e-01 6.70559704e-01
1.15804207e+00 -1.13252127e+00 7.88112819e-01 -8.64252150e-02
-1.50806379e+00 -1.49275675e-01 -1.13224614e+00 -1.29465416e-01
-2.16619849e-01 5.30149043e-01 -4.85201746e-01 9.83070076e-01
1.18388832e+00 6.88173175e-01 -7.26570427e-01 9.22428310e-01
-1.03766799e-01 7.09334135e-01 -8.05529207e-02 6.45034611e-01
3.81577253e-01 -1.87573075e-01 2.16673359e-01 1.36968172e+00
2.83574730e-01 2.22271502e-01 -1.92306578e-01 1.13397861e+00
-1.60835773e-01 -6.77387370e-03 -5.34149766e-01 1.05943590e-01
-4.12346460e-02 1.69312990e+00 -4.63830113e-01 -1.49351627e-01
-6.13192558e-01 1.17023492e+00 2.83706784e-01 4.01366234e-01
-7.79843211e-01 -7.84833804e-02 6.51362360e-01 -1.55318705e-02
2.08559945e-01 -1.26815736e-01 -5.42135954e-01 -9.41927195e-01
-4.37582374e-01 -5.25767624e-01 2.76326418e-01 -1.55308950e+00
-8.97680819e-01 6.60325050e-01 -3.21985662e-01 -1.05009556e+00
1.32276073e-01 -9.35706913e-01 -1.64276481e-01 1.37525570e+00
-1.94206166e+00 -1.55057108e+00 -1.05646598e+00 9.24734414e-01
5.96051812e-01 -2.02043056e-01 9.07391787e-01 1.00165710e-01
-3.75465631e-01 -2.18002517e-02 3.15829158e-01 -2.26440310e-01
7.19914854e-01 -1.62575269e+00 4.00811613e-01 6.98685884e-01
-2.45343432e-01 -1.04235508e-01 7.58197963e-01 -3.65789086e-01
-1.33548951e+00 -1.50245285e+00 8.30766782e-02 -2.46111955e-02
5.86508453e-01 -7.12292567e-02 -1.10308182e+00 6.54876947e-01
6.22260213e-01 3.47676277e-01 4.22506988e-01 -2.03604743e-01
-6.18032396e-01 -3.30929846e-01 -1.49139571e+00 2.37408236e-01
1.23934805e+00 -8.98235977e-01 -5.37482426e-02 6.86312735e-01
8.83567989e-01 -5.03158450e-01 -1.24464440e+00 7.67669618e-01
5.23220539e-01 -1.66983843e+00 1.29716218e+00 -2.04312786e-01
5.06697714e-01 -3.25425148e-01 -5.46832681e-01 -1.34273911e+00
-5.54302633e-01 -3.44390184e-01 -8.14881027e-02 7.02485502e-01
9.51481611e-02 -2.33342573e-01 6.53196752e-01 -2.38960207e-01
-4.44954336e-01 -1.80605292e-01 -9.47512269e-01 -7.37389624e-01
1.41307190e-01 -3.28440577e-01 8.67599130e-01 9.40160692e-01
-7.86831856e-01 8.48326366e-03 -6.50786400e-01 4.91512805e-01
1.06443989e+00 -6.64939079e-03 4.94543582e-01 -1.35942352e+00
-1.33136109e-01 -3.52958024e-01 -3.14091519e-02 -2.29723915e-01
6.71945438e-02 -9.04635310e-01 -3.63522023e-02 -1.82663953e+00
4.05392051e-01 1.67808443e-01 -2.79949933e-01 3.51498157e-01
1.33065194e-01 9.59742129e-01 -4.04960178e-02 2.09539309e-01
-3.75138462e-01 4.55811799e-01 1.21590972e+00 -1.29884288e-01
5.45933358e-02 -3.62234175e-01 -3.53900105e-01 7.23741770e-01
1.09743941e+00 4.83347587e-02 1.10762194e-01 -4.53237563e-01
6.82531595e-01 9.53355357e-02 6.94286585e-01 -1.38310778e+00
-7.93532282e-02 2.01749988e-02 7.46622086e-01 -7.24721074e-01
4.74255562e-01 -7.25931227e-01 5.37366986e-01 2.37091601e-01
-1.44341692e-01 -4.15490687e-01 4.72161323e-01 2.03208551e-01
1.08898930e-01 -8.63883048e-02 1.42686701e+00 -5.64929187e-01
-1.13512933e+00 3.72135311e-01 -2.69391648e-02 -5.24008214e-01
3.69567066e-01 -2.08123744e-01 -4.69731480e-01 -1.95389882e-01
-8.88231754e-01 -4.43586498e-01 5.26026011e-01 -1.98974222e-01
8.10937405e-01 -1.61064398e+00 -1.02323711e+00 -1.29233479e-01
1.33732468e-01 1.63234353e-01 5.88324189e-01 7.06167519e-01
-8.62830639e-01 7.05723986e-02 -6.89774275e-01 -5.22355735e-01
-1.00669193e+00 4.66839731e-01 8.83032680e-01 6.05052412e-02
-9.13540840e-01 7.95623422e-01 -1.05664596e-01 -6.39619827e-01
-1.23195305e-01 -1.07719511e-01 -1.31660342e-01 -2.20087767e-01
6.91368043e-01 4.29092020e-01 3.09354305e-01 -6.78106368e-01
-1.11185692e-01 6.65211499e-01 3.83172274e-01 8.66732895e-02
1.67628777e+00 -2.67541200e-01 -2.72539377e-01 4.40356582e-01
9.05378938e-01 -3.58471841e-01 -1.62055862e+00 -5.36078572e-01
-4.38198268e-01 -4.78531539e-01 7.75472105e-01 -1.12297511e+00
-1.42333210e+00 7.15238512e-01 1.31359422e+00 1.21979959e-01
1.64788175e+00 -2.26433843e-01 6.50582373e-01 4.39277411e-01
3.75219524e-01 -1.04127252e+00 -1.13779135e-01 4.92342263e-01
9.94584143e-01 -1.62743890e+00 2.51496553e-01 -2.70861387e-01
-3.93346101e-01 1.40652966e+00 6.59602106e-01 -2.19650552e-01
3.26564521e-01 3.46230328e-01 -1.61132902e-01 -1.63502648e-01
-3.56148005e-01 -8.41501117e-01 1.38650015e-01 9.36387002e-01
4.86381859e-01 1.47921786e-01 5.77899106e-02 4.12762612e-01
-1.40444621e-01 1.96945459e-01 7.92039752e-01 3.66292924e-01
-7.71580696e-01 -3.86170089e-01 -5.16415715e-01 1.62809983e-01
-1.43170491e-01 -5.13766050e-01 -5.11725068e-01 8.06679964e-01
2.63756454e-01 4.89842772e-01 3.01549956e-02 -3.95255834e-01
2.59386659e-01 1.26730492e-02 8.17915022e-01 -4.68849242e-01
-6.38780057e-01 -2.24354658e-02 1.20199591e-01 -9.05659854e-01
-8.83261621e-01 -4.21595544e-01 -9.94965911e-01 -3.98558885e-01
1.53331868e-02 -3.99170160e-01 6.54107630e-01 6.51555836e-01
-1.57212228e-01 8.76188099e-01 4.89279687e-01 -1.20578206e+00
-1.30756170e-01 -1.22093976e+00 -1.06400549e+00 5.35287336e-02
4.75512654e-01 -3.92249674e-01 -3.38598549e-01 -1.60206810e-01] | [10.304314613342285, -1.999470829963684] |
2bef5eed-2d00-4b82-a341-1731a0767e3f | blended-multi-modal-deep-convnet-features-for | 2006.00197 | null | https://arxiv.org/abs/2006.00197v1 | https://arxiv.org/pdf/2006.00197v1.pdf | Blended Multi-Modal Deep ConvNet Features for Diabetic Retinopathy Severity Prediction | Diabetic Retinopathy (DR) is one of the major causes of visual impairment and blindness across the world. It is usually found in patients who suffer from diabetes for a long period. The major focus of this work is to derive optimal representation of retinal images that further helps to improve the performance of DR recognition models. To extract optimal representation, features extracted from multiple pre-trained ConvNet models are blended using proposed multi-modal fusion module. These final representations are used to train a Deep Neural Network (DNN) used for DR identification and severity level prediction. As each ConvNet extracts different features, fusing them using 1D pooling and cross pooling leads to better representation than using features extracted from a single ConvNet. Experimental studies on benchmark Kaggle APTOS 2019 contest dataset reveals that the model trained on proposed blended feature representations is superior to the existing methods. In addition, we notice that cross average pooling based fusion of features from Xception and VGG16 is the most appropriate for DR recognition. With the proposed model, we achieve an accuracy of 97.41%, and a kappa statistic of 94.82 for DR identification and an accuracy of 81.7% and a kappa statistic of 71.1% for severity level prediction. Another interesting observation is that DNN with dropout at input layer converges more quickly when trained using blended features, compared to the same model trained using uni-modal deep features. | ['S. N. Shareef', 'M. Bilal', 'J. D. Bodapati', 'O. Jo', 'P. K. R. Maddikunta', 'S. Hakak', 'N. Veeranjaneyulu'] | 2020-05-30 | null | null | null | null | ['severity-prediction'] | ['computer-vision'] | [-1.91323590e-02 -1.46804780e-01 2.11296254e-03 -4.98944312e-01
-6.07520401e-01 -1.83063775e-01 3.55761021e-01 -7.69621432e-02
-5.01578271e-01 9.69262242e-01 4.32535470e-01 -1.70807764e-02
-2.52676368e-01 -7.15885818e-01 -1.61292285e-01 -8.93207729e-01
1.60145223e-01 -2.00982317e-02 4.92930375e-02 1.31932190e-02
3.57614905e-01 9.62818325e-01 -1.86185026e+00 6.79683745e-01
1.17035139e+00 1.18869805e+00 -8.73510446e-03 6.75877273e-01
3.42056490e-02 9.89122450e-01 -6.02047622e-01 -2.95715362e-01
5.42874634e-01 -1.63709000e-01 -5.74145377e-01 -1.03251740e-01
9.00509357e-01 -5.95467687e-01 -5.18021643e-01 9.78838384e-01
1.09334028e+00 -3.34210061e-02 8.19462180e-01 -4.96120572e-01
-6.95514321e-01 2.12822109e-01 -7.36458242e-01 6.57521784e-01
-1.56984195e-01 3.32567006e-01 5.26603937e-01 -5.56850374e-01
4.40879196e-01 1.12290001e+00 3.68631899e-01 4.09520417e-01
-1.13124096e+00 -4.51514274e-01 -3.63633871e-01 3.72574687e-01
-1.38519621e+00 -3.92625242e-01 3.04335624e-01 -6.45870924e-01
9.31124747e-01 8.37998018e-02 5.03868580e-01 5.86219847e-01
4.30702090e-01 6.19049370e-01 1.60501361e+00 -1.04024000e-01
-2.56643891e-01 1.17361128e-01 4.70542073e-01 4.67348665e-01
4.99964088e-01 1.52301073e-01 -1.09180063e-02 3.53492163e-02
6.95215940e-01 -1.85657162e-02 -4.49792057e-01 1.60083354e-01
-8.68167818e-01 7.10809350e-01 6.37233913e-01 2.99047619e-01
-7.90942132e-01 -5.61235324e-02 3.25508416e-01 3.49921942e-01
-7.41528869e-02 2.43363097e-01 -2.17572525e-01 2.85237104e-01
-7.98818290e-01 2.08268538e-01 1.85905993e-01 3.41015846e-01
5.72411239e-01 2.41653979e-01 -7.14092612e-01 1.11414993e+00
2.38560036e-01 4.28834140e-01 5.54492533e-01 -7.32324123e-01
2.54483912e-02 1.06478155e+00 -1.30058125e-01 -6.48234069e-01
-4.24076557e-01 -8.90397370e-01 -1.07331586e+00 8.76474380e-01
4.04886127e-01 -3.62757295e-01 -1.55367708e+00 1.03854489e+00
-2.94866741e-01 -1.23772472e-01 6.20419621e-01 1.01286268e+00
1.13902056e+00 4.33650285e-01 1.33542523e-01 5.99704012e-02
1.46219110e+00 -4.30460244e-01 -4.20761228e-01 9.34234113e-02
3.93761098e-01 -9.54648435e-01 2.59225219e-01 6.63586140e-01
-9.59445655e-01 -7.88173795e-01 -1.02127314e+00 -2.31876016e-01
-3.06465179e-01 9.44317758e-01 4.50313091e-01 4.97204155e-01
-1.29555368e+00 9.96291339e-02 -1.49640381e-01 -5.50642610e-01
7.85066962e-01 7.52062082e-01 -5.06320834e-01 -4.73627955e-01
-7.56930292e-01 1.21779013e+00 3.40322167e-01 3.78587276e-01
-5.84790349e-01 -3.98207337e-01 -3.66399676e-01 -2.43044674e-01
-3.43881935e-01 -9.12116528e-01 6.59870267e-01 -1.00301790e+00
-9.90697682e-01 8.48872244e-01 -2.42629096e-01 -7.86885321e-01
4.33426678e-01 -2.37884089e-01 -6.81759298e-01 3.89279813e-01
-3.27823251e-01 7.46434391e-01 5.39823830e-01 -9.93644416e-01
-9.67693329e-01 -6.81207836e-01 -3.62655781e-02 1.90894052e-01
1.63026795e-01 1.49504542e-01 8.57525170e-02 -2.02143908e-01
1.81993768e-02 -4.77990121e-01 -1.16016485e-01 -1.92537308e-01
-3.30979139e-01 -4.40005332e-01 7.27904499e-01 -9.95248258e-01
9.49308038e-01 -1.94596887e+00 -1.04245901e-01 2.63492286e-01
5.40748894e-01 9.38377321e-01 -1.01499312e-01 -9.59258154e-02
-1.94688365e-01 -2.18104254e-02 1.72284946e-01 1.46755785e-01
-5.58042705e-01 -7.04240873e-02 9.00981873e-02 3.98881257e-01
4.78737801e-01 5.95893383e-01 -1.33476406e-01 -2.17701688e-01
3.76352698e-01 9.93367434e-01 -1.20912090e-01 4.55289222e-02
3.43462586e-01 3.10011894e-01 -4.73080724e-01 7.70586431e-01
9.62076843e-01 1.47753581e-02 -2.48564497e-01 -8.46769750e-01
-3.06249410e-01 -3.62876654e-01 -1.04759753e+00 9.12409067e-01
-1.89589381e-01 8.38722527e-01 -5.97627044e-01 -1.01064754e+00
1.41221833e+00 2.98182428e-01 1.47331759e-01 -7.33070672e-01
5.51103413e-01 2.66202211e-01 3.87194008e-01 -8.48524272e-01
1.21021725e-01 2.45388448e-01 6.50827944e-01 -2.86949039e-01
1.74839303e-01 8.21926594e-01 1.56578586e-01 -2.24263757e-01
9.17027950e-01 -2.85554111e-01 3.50585848e-01 4.47269566e-02
9.13603365e-01 -2.47564986e-01 5.25443971e-01 5.77700675e-01
-2.41455808e-01 7.44742990e-01 6.59666181e-01 -6.45415127e-01
-8.72229278e-01 -8.60346913e-01 -5.83787739e-01 -4.71377708e-02
-2.94162422e-01 4.09661800e-01 -1.56538397e-01 -5.00364423e-01
1.01522140e-01 3.53480875e-01 -7.65188396e-01 1.15364447e-01
-4.22902226e-01 -1.05147982e+00 5.86976349e-01 5.45666158e-01
1.07329798e+00 -7.51145005e-01 -4.77667451e-01 -8.47515929e-03
3.49318475e-01 -8.10752928e-01 2.23529994e-01 -2.68117070e-01
-1.00756574e+00 -1.24286962e+00 -1.23617637e+00 -8.80049229e-01
6.26502573e-01 6.46063462e-02 6.20069563e-01 -6.53417110e-02
-6.60655320e-01 -6.28084689e-02 -1.90666839e-01 -2.08105281e-01
-5.46573177e-02 -2.30814189e-01 -3.70937049e-01 5.07805765e-01
8.20442855e-01 -3.64944547e-01 -1.07828867e+00 -1.85228392e-01
-5.64839602e-01 -2.82315731e-01 1.35590279e+00 6.59294426e-01
5.40065169e-01 -9.07102078e-02 8.11838031e-01 -5.89643002e-01
5.51982045e-01 -2.29526564e-01 -6.70380950e-01 2.30593681e-01
-5.39666176e-01 -5.34892567e-02 2.65189618e-01 -1.42475009e-01
-1.02781594e+00 1.15013979e-01 1.66298121e-01 -1.20339565e-01
-5.12171924e-01 2.65930057e-01 5.60146570e-02 -3.69115561e-01
7.56584287e-01 2.20257431e-01 2.69101173e-01 -7.02775359e-01
-2.61547440e-03 1.15871966e+00 4.92787987e-01 6.30668476e-02
1.73444241e-01 2.10130289e-01 1.96204528e-01 -8.95287991e-01
-5.76786220e-01 -3.69086713e-01 -4.22194749e-01 -1.59613043e-01
1.33795786e+00 -1.19672632e+00 -9.28009808e-01 8.77947032e-01
-1.07873094e+00 2.61996090e-01 2.89267182e-01 9.71358895e-01
-6.64243847e-02 9.40813199e-02 -2.53937960e-01 -8.18020761e-01
-5.45967400e-01 -1.16994154e+00 4.99103755e-01 8.58231246e-01
2.97079265e-01 -6.59009278e-01 -1.43416926e-01 6.49656296e-01
6.54187739e-01 5.09071589e-01 1.06653726e+00 -5.92480302e-01
-6.18187547e-01 -3.94598991e-01 -9.70132291e-01 1.00350964e+00
1.58981726e-01 2.00643629e-01 -1.05619705e+00 -9.49065089e-02
-3.93860608e-01 -7.85934851e-02 1.47810292e+00 7.80002952e-01
8.94050598e-01 7.64938518e-02 -1.14221618e-01 5.08752942e-01
2.14720893e+00 6.08026981e-01 1.18568981e+00 2.42002249e-01
5.06546080e-01 3.95731747e-01 1.67511463e-01 2.03682721e-01
3.34716767e-01 3.36114943e-01 4.53484684e-01 -3.18874896e-01
-6.63787305e-01 5.92268229e-01 1.04211465e-01 4.61895391e-02
-6.02445781e-01 -1.73252910e-01 -9.09330785e-01 7.32068002e-01
-1.51336837e+00 -8.86199713e-01 -5.04044175e-01 2.25849128e+00
5.17855644e-01 -1.34650707e-01 1.27807751e-01 5.11661023e-02
8.52575362e-01 -1.77937374e-01 -3.58196527e-01 -5.14191449e-01
-6.41683996e-01 6.22449636e-01 7.95133471e-01 3.71835649e-01
-9.79587913e-01 5.61572611e-01 5.41156435e+00 2.89619416e-01
-1.38946700e+00 -3.35435778e-01 7.21926212e-01 -1.22310869e-01
3.07285219e-01 -2.28986830e-01 -9.92147267e-01 2.92650074e-01
6.94341063e-01 5.02943322e-02 1.48011133e-01 2.67758816e-01
4.99776423e-01 -3.91916364e-01 -6.90013826e-01 1.01570272e+00
1.69368625e-01 -1.27828634e+00 4.88734365e-01 1.99089080e-01
8.70183885e-01 3.30728382e-01 2.02546105e-01 -1.59602407e-02
1.83234736e-01 -1.65854454e+00 -2.67749310e-01 1.32059729e+00
5.76028943e-01 -1.09825611e+00 1.54630101e+00 -3.51997167e-01
-7.16965973e-01 -3.03811163e-01 -5.72432220e-01 3.20306987e-01
-1.97880447e-01 5.60541749e-01 -7.02446878e-01 6.07518733e-01
6.05991185e-01 7.82661974e-01 -8.22522640e-01 1.79839253e+00
-1.46943275e-02 4.09815341e-01 9.84309707e-03 2.37039447e-01
1.10861003e-01 -6.02012686e-02 5.28106570e-01 1.00369442e+00
4.84196663e-01 1.30742565e-01 -3.71464252e-01 6.97627485e-01
1.05002716e-01 2.33351752e-01 -5.71287870e-01 1.13274761e-01
-1.62779894e-02 1.21458876e+00 1.81288701e-02 -2.99563199e-01
-6.30680501e-01 5.25186419e-01 4.76182951e-03 7.70468950e-01
-3.68091315e-01 -7.28420675e-01 5.67819238e-01 6.83406889e-02
4.72578824e-01 1.17302999e-01 -3.43119532e-01 -7.67346919e-01
-4.95124869e-02 -8.15560222e-01 3.38179886e-01 -9.11125541e-01
-1.29078603e+00 9.97162402e-01 -4.18016583e-01 -1.24008894e+00
1.88328594e-01 -8.22446585e-01 -5.34033954e-01 1.67518568e+00
-1.92586863e+00 -1.33248806e+00 -7.12910533e-01 7.06866562e-01
9.36496332e-02 -7.58278847e-01 5.05535245e-01 3.35835278e-01
-7.69921243e-01 4.65443283e-01 -6.39206320e-02 2.58508295e-01
7.93780923e-01 -1.18308175e+00 -5.53234518e-01 9.06165302e-01
-6.53690159e-01 5.74937582e-01 3.61825198e-01 -5.79942942e-01
-9.12792861e-01 -1.31725860e+00 7.53431916e-01 1.78260133e-01
4.41263616e-02 7.13962853e-01 -8.19278538e-01 3.78723800e-01
3.33043128e-01 2.29160354e-01 5.54838181e-01 -2.81652898e-01
-2.29519084e-01 -6.66796803e-01 -1.46329212e+00 2.76815861e-01
3.18855911e-01 -1.77225247e-01 -5.36513925e-01 -4.73683625e-02
9.73271206e-03 -1.53008312e-01 -1.20313311e+00 6.01911068e-01
6.71750546e-01 -1.19813013e+00 7.75023997e-01 -6.25462174e-01
3.54891092e-01 -5.90983987e-01 -3.50223005e-01 -9.54103053e-01
-1.12054698e-01 -1.06150419e-01 3.06831747e-01 1.09785044e+00
4.58521247e-01 -9.84483123e-01 4.95303869e-01 2.69307941e-01
-4.23558839e-02 -6.85361624e-01 -5.71754456e-01 -3.93132210e-01
3.68325636e-02 3.01828086e-01 2.36449495e-01 7.06874430e-01
-1.00450683e+00 1.17503867e-01 -1.94189325e-01 5.47911465e-01
7.56754220e-01 -2.35031664e-01 4.84107405e-01 -1.58358920e+00
1.30299270e-01 -5.00320077e-01 -1.08243382e+00 -5.07202387e-01
-3.55358094e-01 -7.96899855e-01 -6.19448602e-01 -2.10918736e+00
2.34651878e-01 -2.08838135e-01 -7.57969022e-01 6.30771756e-01
2.06618413e-01 6.27052546e-01 -5.90808801e-02 -3.19937505e-02
1.31436929e-01 1.41636983e-01 1.46324682e+00 -6.95061162e-02
-4.04915541e-01 1.12824045e-01 -9.42003667e-01 4.90853012e-01
1.04530156e+00 1.31280459e-02 -3.63971651e-01 -4.88589734e-01
-4.16858792e-02 -6.67394325e-02 7.70483077e-01 -1.26971281e+00
1.39734954e-01 1.07614025e-01 8.75390708e-01 -7.49443769e-01
2.85766453e-01 -4.59304869e-01 4.81914841e-02 3.05136591e-01
-1.98707685e-01 -3.89034241e-01 2.68385351e-01 2.49213234e-01
-4.62486088e-01 3.62844504e-02 1.13649905e+00 1.06754424e-02
-8.03252220e-01 1.34230316e-01 -3.60408247e-01 -3.59558731e-01
9.78422344e-01 -6.74264848e-01 -7.73220956e-01 1.04476959e-02
-9.18723643e-01 8.92928839e-02 2.12262534e-02 2.78316498e-01
8.15584660e-01 -1.10768831e+00 -1.20647848e+00 7.25048035e-02
1.00145482e-01 -1.86325535e-01 5.44460058e-01 1.19480968e+00
-6.91332400e-01 4.57383871e-01 -8.81758392e-01 -5.83932698e-01
-1.65082908e+00 -1.43053770e-01 7.88478494e-01 7.83650279e-02
-6.42720222e-01 5.77837825e-01 -1.45385161e-01 3.94337803e-01
1.51398107e-01 -4.46973920e-01 -1.12362874e+00 2.25793615e-01
8.03302586e-01 6.43740535e-01 2.14366838e-01 -6.63157880e-01
-2.41271645e-01 1.05916357e+00 -5.56944966e-01 2.89208353e-01
1.38555849e+00 1.12104192e-01 -3.56938094e-01 -3.35450843e-02
1.08611119e+00 -4.61365320e-02 -8.45779657e-01 -2.15803146e-01
-3.16983849e-01 -3.89107585e-01 6.29742920e-01 -1.41036117e+00
-1.47732961e+00 8.52373004e-01 1.43172181e+00 -1.80879205e-01
1.38140821e+00 -3.10079247e-01 5.28767109e-01 1.93226427e-01
-1.51451692e-01 -5.73259652e-01 -2.98526764e-01 2.11671561e-01
9.88983929e-01 -1.11633801e+00 -3.81645337e-02 1.88403931e-02
-7.11605787e-01 1.46527207e+00 6.52564526e-01 -5.60672343e-01
5.16098082e-01 -2.79222280e-01 3.93358290e-01 -2.12771371e-01
-5.06285489e-01 -5.92387676e-01 7.12080121e-01 8.47203195e-01
5.19266963e-01 -6.67479783e-02 -7.08633184e-01 4.12202984e-01
1.51496381e-01 4.03789669e-01 8.92242968e-01 4.96271729e-01
-7.67790675e-01 -1.07370567e+00 -4.34801131e-02 1.02434719e+00
-6.84088945e-01 -3.13303433e-02 -2.95401543e-01 7.40781426e-01
5.49601018e-01 9.21160460e-01 9.62024108e-02 -2.45416343e-01
3.19143951e-01 -2.20543481e-02 5.21936774e-01 -3.66392344e-01
-6.08994484e-01 3.69229652e-02 3.14879447e-01 -4.08793867e-01
-6.91827059e-01 -4.49684232e-01 -1.07254708e+00 -1.18054815e-01
1.19142205e-01 -4.45890456e-01 5.76144695e-01 7.93300688e-01
4.79570627e-01 6.92855299e-01 4.67128396e-01 -6.99818060e-02
-3.85215491e-01 -1.05362618e+00 -6.45523667e-01 -1.62023500e-01
7.52636313e-01 -6.70407951e-01 -2.43520424e-01 -5.71793281e-02] | [15.836979866027832, -3.9881322383880615] |
51558238-c960-4eea-a03a-57ffa4c09954 | efficientface-an-efficient-deep-network-with | 2302.11816 | null | https://arxiv.org/abs/2302.11816v1 | https://arxiv.org/pdf/2302.11816v1.pdf | EfficientFace: An Efficient Deep Network with Feature Enhancement for Accurate Face Detection | In recent years, deep convolutional neural networks (CNN) have significantly advanced face detection. In particular, lightweight CNNbased architectures have achieved great success due to their lowcomplexity structure facilitating real-time detection tasks. However, current lightweight CNN-based face detectors trading accuracy for efficiency have inadequate capability in handling insufficient feature representation, faces with unbalanced aspect ratios and occlusion. Consequently, they exhibit deteriorated performance far lagging behind the deep heavy detectors. To achieve efficient face detection without sacrificing accuracy, we design an efficient deep face detector termed EfficientFace in this study, which contains three modules for feature enhancement. To begin with, we design a novel cross-scale feature fusion strategy to facilitate bottom-up information propagation, such that fusing low-level and highlevel features is further strengthened. Besides, this is conducive to estimating the locations of faces and enhancing the descriptive power of face features. Secondly, we introduce a Receptive Field Enhancement module to consider faces with various aspect ratios. Thirdly, we add an Attention Mechanism module for improving the representational capability of occluded faces. We have evaluated EfficientFace on four public benchmarks and experimental results demonstrate the appealing performance of our method. In particular, our model respectively achieves 95.1% (Easy), 94.0% (Medium) and 90.1% (Hard) on validation set of WIDER Face dataset, which is competitive with heavyweight models with only 1/15 computational costs of the state-of-the-art MogFace detector. | ['Wankou Yang', 'Jifeng Shen', 'Jianhua Xu', 'Zhijian Wu', 'Jun Li', 'Guangtao Wang'] | 2023-02-23 | null | null | null | null | ['face-detection'] | ['computer-vision'] | [-1.96375132e-01 -1.06267825e-01 5.48730753e-02 -4.55472618e-01
-2.26115689e-01 -1.19089372e-01 4.36402529e-01 -4.24279392e-01
-4.44352776e-01 3.37697715e-01 3.95387933e-02 8.85751843e-02
8.01598430e-02 -7.68250704e-01 -5.55570602e-01 -7.54285097e-01
-2.32210487e-01 -1.27239823e-01 3.23796347e-02 -1.69911057e-01
-1.05192084e-02 7.59560645e-01 -1.89769173e+00 2.92785436e-01
6.28861308e-01 1.39116406e+00 1.81597583e-02 2.42332563e-01
-2.82111242e-02 3.11359733e-01 -4.52246785e-01 -6.32411420e-01
3.10898840e-01 1.06363744e-01 -2.24236265e-01 4.24668826e-02
5.68037689e-01 -7.42326498e-01 -4.78507757e-01 1.16128182e+00
7.69615650e-01 -2.22503811e-01 5.12853861e-01 -1.30567372e+00
-6.42014146e-01 3.70055556e-01 -1.16718829e+00 3.17642361e-01
2.72180170e-01 1.28284335e-01 6.04377747e-01 -1.59588695e+00
1.57729879e-01 1.68540144e+00 7.82307982e-01 6.37167037e-01
-6.30452216e-01 -1.12964761e+00 3.12916666e-01 1.05677225e-01
-1.66955996e+00 -8.60354006e-01 7.05195904e-01 -1.05824269e-01
1.04699540e+00 5.53297140e-02 4.16948199e-01 8.50918293e-01
9.45547819e-02 6.22602105e-01 6.98341310e-01 -2.01233238e-01
-2.47778192e-01 1.95169315e-01 -1.44704461e-01 1.23963094e+00
4.81472224e-01 2.37035200e-01 -4.73829806e-01 6.20040633e-02
7.86838889e-01 3.89903426e-01 -2.03174397e-01 2.92661805e-02
-5.07074177e-01 7.59596348e-01 9.05775785e-01 1.53391600e-01
-3.62532347e-01 8.12176522e-03 3.82460475e-01 4.23652604e-02
4.20027435e-01 -2.51011848e-01 -3.21334094e-01 4.68656361e-01
-7.46038496e-01 1.03612691e-01 3.63519937e-01 8.88163447e-01
5.48629344e-01 2.82032222e-01 -3.72359186e-01 8.85231614e-01
6.60943568e-01 5.49467802e-01 2.33179599e-01 -5.05061030e-01
5.05316079e-01 7.99621701e-01 -2.34680980e-01 -1.24118173e+00
-5.97666562e-01 -7.45650470e-01 -1.15525901e+00 2.74513602e-01
3.08265234e-03 -5.42976260e-02 -1.01809895e+00 1.62103927e+00
4.21632469e-01 9.21738669e-02 -2.05527529e-01 8.43196392e-01
1.18438578e+00 5.56866288e-01 3.36976737e-01 -2.17735469e-01
1.88352036e+00 -7.64830947e-01 -4.87340569e-01 -1.37764618e-01
2.41422176e-01 -8.26349497e-01 6.19037271e-01 2.00206786e-01
-9.65864718e-01 -9.60398316e-01 -1.09053576e+00 1.20300725e-01
-3.77671719e-01 5.53910971e-01 8.14334273e-01 9.77673709e-01
-1.15403986e+00 2.01331481e-01 -5.11028171e-01 -1.30938187e-01
1.10796869e+00 7.92733312e-01 -5.77717960e-01 -1.50094092e-01
-9.25736010e-01 5.87402940e-01 2.16994017e-01 4.19149786e-01
-9.25134778e-01 -7.06745684e-01 -9.94161844e-01 3.30304056e-01
3.71029586e-01 -5.48738122e-01 9.40202653e-01 -8.19686532e-01
-1.12539327e+00 7.62582541e-01 -1.21631607e-01 -1.28964797e-01
4.06549305e-01 -2.23285004e-01 -6.46064579e-01 1.25745907e-01
6.74367994e-02 1.01769936e+00 1.16106403e+00 -9.80265558e-01
-8.80558789e-01 -7.06153631e-01 -8.49250611e-03 4.21811938e-02
-8.53572190e-01 4.45503712e-01 -5.54233611e-01 -5.71785450e-01
5.57654314e-02 -3.98308337e-01 1.13659322e-01 4.46533650e-01
-2.58270860e-01 -6.46588385e-01 1.18102252e+00 -4.64470237e-01
1.11399686e+00 -2.24945450e+00 -3.88290554e-01 5.38506992e-02
4.17654544e-01 7.43151009e-01 -1.89285398e-01 -1.14247367e-01
-9.74483714e-02 -1.22446585e-02 2.45292354e-02 -4.55455393e-01
-6.53812885e-02 -2.31583551e-01 -1.06490530e-01 6.36673808e-01
8.15781772e-01 8.15510988e-01 -5.02613783e-01 -6.67613804e-01
2.54864722e-01 9.66392457e-01 -8.53531718e-01 1.17176764e-01
4.07912463e-01 -1.58767343e-01 -4.85869497e-01 1.19553387e+00
1.21533167e+00 1.38062891e-02 -9.73165482e-02 -5.20485401e-01
-1.22246183e-01 -2.99198478e-01 -1.35652637e+00 1.10380375e+00
-4.15220231e-01 2.00435147e-01 4.41791117e-01 -7.52449811e-01
8.60407531e-01 2.36892238e-01 1.57072261e-01 -6.93477750e-01
4.51872110e-01 1.02055550e-01 2.07609963e-02 -3.20227534e-01
1.02270260e-01 9.95924547e-02 3.55294675e-01 -1.16674304e-01
1.87118545e-01 6.59653366e-01 -5.45627158e-03 -1.65418629e-02
7.14150846e-01 2.00465061e-02 2.72418767e-01 -3.51012379e-01
8.66399109e-01 -7.93379664e-01 6.11956418e-01 3.83864582e-01
-5.02159834e-01 4.09149617e-01 1.84524640e-01 -6.39163017e-01
-5.00062704e-01 -8.55697453e-01 -4.10167307e-01 1.25843251e+00
5.03691882e-02 -3.34954083e-01 -7.45271266e-01 -7.07703292e-01
1.79128110e-01 -2.65520990e-01 -7.69961715e-01 -2.35227570e-01
-6.62742615e-01 -1.01637316e+00 6.75468624e-01 8.50596905e-01
1.16862988e+00 -1.07306421e+00 -5.60861349e-01 8.41882750e-02
2.32554182e-01 -1.05511367e+00 -4.83297765e-01 -1.29299030e-01
-5.35526514e-01 -9.89623487e-01 -6.48196638e-01 -1.13257194e+00
8.49009454e-01 6.47375405e-01 8.59565794e-01 4.89731342e-01
-7.51157165e-01 -1.30043060e-01 -1.38142416e-02 -5.53325057e-01
2.62788475e-01 -1.80129707e-01 2.99284965e-01 1.58651844e-01
4.66473103e-01 -2.06410378e-01 -9.67302620e-01 2.76902169e-01
-6.54403150e-01 -3.45200956e-01 8.87152910e-01 8.42612326e-01
2.14019269e-01 1.76693007e-01 7.72218704e-01 -5.09282231e-01
3.37803721e-01 -3.36886138e-01 -5.72700143e-01 8.12851414e-02
-4.01371270e-01 -2.54203856e-01 6.60072863e-01 -2.16652095e-01
-1.37481713e+00 2.20633671e-01 -4.19446856e-01 -3.08709025e-01
-7.31666908e-02 -8.87418613e-02 -5.50726473e-01 -4.78597134e-01
3.95165384e-01 1.65389925e-01 3.92549150e-02 -4.65375990e-01
8.93620178e-02 7.82019854e-01 4.81522769e-01 -3.95684272e-01
7.61783779e-01 5.39474666e-01 1.00901954e-01 -7.12511420e-01
-4.89165336e-01 -3.32578003e-01 -3.12079102e-01 -1.78927884e-01
4.75755751e-01 -1.29631495e+00 -1.26586354e+00 6.90857768e-01
-1.05501044e+00 6.06033802e-01 2.71381438e-01 1.79110870e-01
3.87620814e-02 3.33618313e-01 -6.48239195e-01 -1.04071999e+00
-7.07838595e-01 -1.08007431e+00 1.31422579e+00 6.38221025e-01
2.90917873e-01 -4.36970860e-01 -5.85718453e-01 7.31542856e-02
6.27555013e-01 1.33574530e-01 4.78077531e-01 -2.73951143e-01
-3.92595500e-01 -3.12351912e-01 -8.87079060e-01 2.55362868e-01
1.58579960e-01 1.22779615e-01 -1.47559512e+00 -6.04154468e-01
-1.65442213e-01 -2.09614441e-01 1.21015668e+00 3.14994186e-01
1.37280667e+00 -2.35059768e-01 -5.68250954e-01 8.30769718e-01
1.40448952e+00 1.48116663e-01 6.00449979e-01 6.25586603e-04
5.94439507e-01 5.41250408e-01 4.32779759e-01 6.33435547e-01
2.47119129e-01 6.07107043e-01 5.75086474e-01 -3.75979364e-01
-4.22236979e-01 -1.52339160e-01 3.23336750e-01 2.13583350e-01
-2.12850630e-01 -6.47785366e-02 -5.21256685e-01 3.93196881e-01
-1.34939945e+00 -8.50672543e-01 1.72108948e-01 1.83974767e+00
5.14217138e-01 1.72940999e-01 1.82630613e-01 1.65831327e-01
7.92840421e-01 5.35872951e-02 -3.83168936e-01 -6.82551563e-02
-8.73538405e-02 2.93008089e-01 2.52826124e-01 -2.88616568e-02
-1.52843547e+00 9.13473964e-01 5.64919996e+00 9.65211153e-01
-1.06675386e+00 1.28217950e-01 8.05093169e-01 3.92065383e-03
3.96146804e-01 -6.51196361e-01 -1.42350972e+00 5.26585281e-01
5.12656331e-01 2.81183720e-01 9.62872580e-02 1.05260372e+00
-1.02945097e-01 2.24889129e-01 -7.60655403e-01 1.27403128e+00
1.88872501e-01 -1.05795586e+00 1.65034905e-01 9.14848968e-02
4.54904765e-01 -2.66862571e-01 3.41992110e-01 5.89956522e-01
-2.21719682e-01 -1.26121938e+00 4.96259302e-01 1.08654760e-01
9.64519799e-01 -1.07653606e+00 1.00269568e+00 -7.98904151e-02
-1.68127763e+00 -5.21917403e-01 -6.44469202e-01 -5.43115381e-03
-2.32496113e-01 3.66471857e-01 -5.49348295e-01 4.20816302e-01
9.53780055e-01 4.27421689e-01 -6.25668824e-01 1.00897741e+00
6.25592098e-02 1.68971792e-01 -4.33974862e-01 1.69293731e-01
1.59923434e-01 3.80591512e-01 6.62521645e-02 1.27136493e+00
3.08368206e-01 1.99879795e-01 1.97363988e-01 6.72984779e-01
-5.07246852e-01 1.19603030e-01 -5.66372514e-01 2.94472665e-01
5.49002707e-01 1.73993707e+00 -7.25988865e-01 -1.70332715e-01
-6.52726948e-01 7.07128406e-01 3.51078272e-01 7.85029754e-02
-9.01554346e-01 -5.31582415e-01 8.85159612e-01 9.37110186e-02
4.63564456e-01 2.04207897e-01 -4.52441536e-02 -7.93045580e-01
1.79608092e-01 -7.16285467e-01 4.65044290e-01 -1.65825441e-01
-1.12211776e+00 9.04120743e-01 -2.46608168e-01 -9.52447712e-01
1.86861783e-01 -1.05127597e+00 -6.85817778e-01 7.43248820e-01
-1.80942285e+00 -1.39114952e+00 -6.40470266e-01 6.67299449e-01
6.18576765e-01 -4.49397296e-01 5.63158870e-01 7.75783837e-01
-1.02054846e+00 1.15185952e+00 -4.19387221e-01 3.98672163e-01
5.29918373e-01 -6.64959788e-01 4.10155386e-01 8.89890134e-01
-1.08227462e-01 7.91385889e-01 2.27562889e-01 -4.16578889e-01
-1.58510733e+00 -1.34918594e+00 5.39194226e-01 -9.96477306e-02
4.40163687e-02 -4.20550257e-01 -7.76014030e-01 2.34196812e-01
-1.55544192e-01 4.97484416e-01 3.05527717e-01 -4.01233286e-02
-6.20101869e-01 -4.81444865e-01 -1.46741962e+00 4.58530754e-01
1.30236125e+00 -2.97389150e-01 -1.63300067e-01 1.21478677e-01
3.74359429e-01 -9.90568399e-02 -5.77869296e-01 7.83429086e-01
7.92978466e-01 -1.05317259e+00 1.25325310e+00 -4.05926883e-01
1.74374580e-01 -3.27850610e-01 -1.09557599e-01 -6.94845617e-01
-6.54809415e-01 -2.79547513e-01 -3.53045911e-01 1.38457870e+00
-2.61650998e-02 -7.82435834e-01 8.40226769e-01 2.07653493e-01
-9.95188504e-02 -9.83313799e-01 -8.93243015e-01 -6.82812452e-01
-3.42987239e-01 1.55082177e-02 8.64969134e-01 4.68565166e-01
-3.90111536e-01 1.16351716e-01 -3.03919345e-01 3.53481412e-01
7.04701304e-01 -7.12861419e-02 4.33440953e-01 -1.31846952e+00
1.97279245e-01 -6.79142118e-01 -7.07493126e-01 -7.97739863e-01
1.06626660e-01 -5.20009398e-01 -1.04301222e-01 -1.05883563e+00
5.23845613e-01 -2.31754497e-01 -4.50290859e-01 7.12428927e-01
-3.84023666e-01 8.80280435e-01 9.65783149e-02 -1.34078428e-01
-5.94644189e-01 6.50969684e-01 1.12136734e+00 -7.03777075e-02
1.49222925e-01 -1.82289362e-01 -9.66512561e-01 8.28251362e-01
6.29652083e-01 -1.77150682e-01 -1.92464888e-01 -3.71498734e-01
-2.18543366e-01 -4.99357462e-01 5.30933380e-01 -1.14939547e+00
2.72186518e-01 3.81345868e-01 1.17076385e+00 -5.01256406e-01
4.10124689e-01 -6.57024801e-01 -2.07285583e-01 7.50026584e-01
3.37540120e-01 -7.62765761e-03 4.06739354e-01 4.39001948e-01
-2.03390613e-01 1.44433036e-01 1.01562250e+00 -5.66328950e-02
-1.01027250e+00 7.16166437e-01 9.71795395e-02 -2.24090904e-01
1.06174576e+00 -3.67876530e-01 -3.45131129e-01 9.05135274e-02
-3.00024390e-01 1.82760656e-01 -4.74155322e-03 7.24466741e-01
9.81763899e-01 -1.35878885e+00 -8.77540946e-01 7.57538021e-01
4.78130057e-02 -2.81156003e-01 4.16534573e-01 6.08110189e-01
-3.93154353e-01 5.16788602e-01 -4.26871479e-01 -4.50355202e-01
-1.46013498e+00 5.14051259e-01 3.18200260e-01 1.83255672e-01
-4.75866467e-01 1.29396367e+00 6.76102519e-01 -7.16580898e-02
4.26058948e-01 -7.17758536e-02 -4.71240580e-01 1.68099537e-01
1.10958219e+00 3.47801805e-01 3.25303137e-01 -8.89751673e-01
-8.66277754e-01 6.74415588e-01 -3.92381877e-01 8.13818514e-01
1.20421445e+00 8.42782110e-02 -3.69633846e-02 -7.05229521e-01
1.18362534e+00 -1.46542341e-01 -1.42256236e+00 -1.11231349e-01
-3.45670789e-01 -6.19292915e-01 2.65965194e-01 -5.26906133e-01
-1.49957013e+00 9.43213165e-01 1.05205929e+00 -8.36947784e-02
1.32937789e+00 -1.47997588e-01 6.92805469e-01 1.72338262e-01
4.75528538e-01 -7.50876307e-01 1.88107535e-01 2.12159440e-01
9.09900367e-01 -1.54401731e+00 4.50394340e-02 -7.81372190e-01
-1.06790967e-01 1.12821221e+00 1.06324553e+00 5.68844704e-03
7.60966837e-01 4.62074667e-01 -1.90599188e-01 -3.35926235e-01
-5.56447804e-01 -3.38204205e-01 3.57420087e-01 5.55953801e-01
3.99354994e-01 1.90928450e-03 2.70890389e-02 7.80493259e-01
5.59541062e-02 -1.51102006e-01 -2.14391559e-01 7.74485111e-01
-7.43743539e-01 -5.83235145e-01 -4.41933602e-01 4.65809047e-01
-7.43022501e-01 -1.08120419e-01 -1.01416864e-01 8.31676006e-01
5.38186610e-01 1.03981340e+00 1.26493767e-01 -3.18919808e-01
3.01108122e-01 -2.81394988e-01 4.65375990e-01 -4.63117182e-01
-5.40030360e-01 -1.06238034e-02 -2.08493426e-01 -5.62040389e-01
-8.83831233e-02 -3.07789177e-01 -1.06976056e+00 -4.12321836e-01
-4.80382532e-01 -1.14593141e-01 4.84662890e-01 6.61226571e-01
5.01658678e-01 5.26290178e-01 7.01857090e-01 -9.86155152e-01
-7.03129768e-01 -9.63322341e-01 -4.99204427e-01 1.02102004e-01
2.92467296e-01 -1.08177197e+00 -1.10001639e-01 -2.74005175e-01] | [13.333688735961914, 0.7139497399330139] |
325c8cf5-904f-4576-acb7-2fe228d34bf8 | syntactically-informed-unsupervised | null | null | https://aclanthology.org/2021.emnlp-main.203 | https://aclanthology.org/2021.emnlp-main.203.pdf | Syntactically-Informed Unsupervised Paraphrasing with Non-Parallel Data | Previous works on syntactically controlled paraphrase generation heavily rely on large-scale parallel paraphrase data that is not easily available for many languages and domains. In this paper, we take this research direction to the extreme and investigate whether it is possible to learn syntactically controlled paraphrase generation with nonparallel data. We propose a syntactically-informed unsupervised paraphrasing model based on conditional variational auto-encoder (VAE) which can generate texts in a specified syntactic structure. Particularly, we design a two-stage learning method to effectively train the model using non-parallel data. The conditional VAE is trained to reconstruct the input sentence according to the given input and its syntactic structure. Furthermore, to improve the syntactic controllability and semantic consistency of the pre-trained conditional VAE, we fine-tune it using syntax controlling and cycle reconstruction learning objectives, and employ Gumbel-Softmax to combine these new learning objectives. Experiment results demonstrate that the proposed model trained only on non-parallel data is capable of generating diverse paraphrases with specified syntactic structure. Additionally, we validate the effectiveness of our method for generating syntactically adversarial examples on the sentiment analysis task. | ['Yufeng Chen', 'Jinan Xu', 'Changjian Hu', 'Yao Meng', 'Yujie Zhang', 'Deyi Xiong', 'Mingtong Liu', 'Erguang Yang'] | null | null | null | null | emnlp-2021-11 | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 3.69549930e-01 1.25801802e-01 -1.52505860e-01 -4.45414811e-01
-8.00123334e-01 -7.83176899e-01 6.19058967e-01 -3.38604778e-01
-2.28628919e-01 8.39966834e-01 5.02673090e-01 -5.01111269e-01
3.49340171e-01 -1.08513403e+00 -1.09333527e+00 -4.51246649e-01
8.20477188e-01 3.68606836e-01 -2.61077940e-01 -4.17561084e-01
1.45701453e-01 2.84382910e-01 -1.18560064e+00 5.68465710e-01
1.19121003e+00 5.18468499e-01 4.33765054e-01 4.57334429e-01
-2.57659882e-01 8.30380082e-01 -7.52240121e-01 -6.32938504e-01
1.23562828e-01 -9.27613616e-01 -9.77579057e-01 5.13941161e-02
6.76777139e-02 -2.72350423e-02 -1.10454842e-01 1.08708525e+00
2.77273536e-01 -4.70594428e-02 8.00594568e-01 -9.38037157e-01
-1.07169843e+00 8.65077794e-01 -2.83416927e-01 5.74833266e-02
3.46491426e-01 2.11530864e-01 1.30347872e+00 -9.01324451e-01
7.08680570e-01 1.11893189e+00 4.39510822e-01 8.59150469e-01
-1.37437737e+00 -6.54415846e-01 -6.77135214e-02 2.43945383e-02
-9.93144393e-01 -1.96876332e-01 1.35650480e+00 -2.99293518e-01
8.00170302e-01 1.82902604e-01 5.67291796e-01 1.73704541e+00
3.86764854e-01 7.86766291e-01 1.17775083e+00 -5.98038614e-01
2.95197666e-01 2.95688957e-01 2.09372733e-02 5.78447163e-01
-1.12593155e-02 2.13227704e-01 -3.24603111e-01 -1.06486054e-02
5.02037704e-01 -2.10793689e-01 -3.55472714e-01 -4.18071210e-01
-1.05201542e+00 1.23151195e+00 4.18532193e-01 3.53694171e-01
-8.39213356e-02 -5.25467694e-02 5.59852898e-01 6.03143156e-01
4.66827214e-01 8.28503072e-01 -3.36834699e-01 1.85503885e-01
-1.03533435e+00 2.94145733e-01 7.44674325e-01 1.08921683e+00
5.21706760e-01 3.53489041e-01 -5.40398359e-01 9.16312218e-01
5.41684441e-02 6.88257039e-01 8.52462888e-01 -5.53223252e-01
8.06199133e-01 4.51535791e-01 -3.00317574e-02 -6.23099983e-01
-1.31556308e-02 -3.82443279e-01 -9.23110187e-01 -4.22704965e-02
1.27260000e-01 -3.30791771e-01 -7.91063786e-01 1.92238152e+00
-4.95427474e-03 -2.07330242e-01 5.45704782e-01 8.09697747e-01
7.44663775e-01 8.72563601e-01 2.36940198e-02 -2.42683426e-01
1.17404723e+00 -1.10928011e+00 -6.35682046e-01 -2.30604887e-01
5.93883634e-01 -6.47336245e-01 1.74780750e+00 -1.65593848e-02
-1.24062848e+00 -7.64297426e-01 -1.19999063e+00 -2.44260937e-01
-4.62519154e-02 2.89979368e-01 2.32664317e-01 4.95853186e-01
-6.07568920e-01 5.47293067e-01 -6.30025148e-01 4.37614304e-04
2.40952894e-01 1.86250433e-02 -1.53123096e-01 6.87034428e-02
-1.53586209e+00 9.66137469e-01 5.24393022e-01 3.15329321e-02
-6.38410568e-01 -6.27133787e-01 -1.15747893e+00 2.85291851e-01
7.34323785e-02 -1.13889503e+00 1.15268052e+00 -1.40199685e+00
-2.00553036e+00 7.42316484e-01 -2.71676511e-01 -4.44159120e-01
4.45220649e-01 1.91273138e-01 -2.22867690e-02 -4.22586985e-02
1.30539194e-01 4.63465661e-01 1.04444444e+00 -1.18545985e+00
-1.37403756e-01 -9.64483544e-02 1.37509584e-01 2.44869426e-01
-4.78924215e-01 -1.14415511e-02 7.35540092e-02 -9.90134537e-01
-3.15310657e-01 -9.33606386e-01 -1.21093012e-01 -5.21824062e-01
-5.52636743e-01 -7.35197067e-02 5.03500342e-01 -8.12110245e-01
1.13767958e+00 -1.87117350e+00 7.68405974e-01 9.19251665e-02
-3.13066304e-01 2.37096339e-01 -2.04199776e-01 4.40733850e-01
-1.21870779e-01 1.69589654e-01 -5.91553867e-01 -4.52806324e-01
7.75140375e-02 3.09465557e-01 -8.58301759e-01 -3.18683051e-02
5.04782200e-01 1.25031877e+00 -7.96274781e-01 -4.29469377e-01
6.68677986e-02 1.32018641e-01 -8.53685439e-01 5.20954609e-01
-3.84935170e-01 6.21699333e-01 -5.46219647e-01 1.79152474e-01
5.24130404e-01 -1.05023764e-01 2.40181699e-01 -1.29942089e-01
1.66756868e-01 3.97400022e-01 -5.23767292e-01 1.89326036e+00
-1.13812816e+00 2.88071454e-01 -2.34577373e-01 -1.20913339e+00
9.84067559e-01 2.10288495e-01 -8.20063576e-02 -5.67522943e-01
8.51378813e-02 2.17611656e-01 -1.57039642e-01 -5.12059271e-01
4.16148663e-01 -6.65168464e-01 -4.09479916e-01 4.55922455e-01
1.02261201e-01 -5.28214991e-01 7.66635835e-02 6.86770678e-02
7.84547448e-01 2.68958151e-01 1.04555435e-01 -2.01701656e-01
9.40261066e-01 -2.25378517e-02 4.77468282e-01 6.64125025e-01
2.18219265e-01 8.33436728e-01 6.81421757e-01 -5.43433949e-02
-1.48851657e+00 -1.12540424e+00 -5.43418638e-02 8.47319424e-01
-3.08037791e-02 -3.27278703e-01 -1.00710177e+00 -8.79408002e-01
-2.66491920e-01 1.24652481e+00 -5.43082654e-01 -4.27066147e-01
-6.71917200e-01 -6.33580625e-01 5.93139291e-01 5.74205160e-01
5.70250034e-01 -1.43555713e+00 -2.25517645e-01 1.25138506e-01
-5.37868083e-01 -1.01697874e+00 -6.37143433e-01 -8.07007998e-02
-6.41620815e-01 -7.49706566e-01 -7.37547398e-01 -1.13060033e+00
6.75668836e-01 -1.26156434e-01 1.05408478e+00 -3.69592309e-01
1.88395366e-01 -4.93756309e-02 -3.64680976e-01 -1.93981990e-01
-1.16244900e+00 4.38838542e-01 -2.14597404e-01 1.03549823e-01
8.62909853e-02 -7.02710450e-01 -3.07198465e-01 -2.25885138e-02
-1.02660060e+00 3.99144024e-01 5.35877049e-01 1.22901905e+00
5.02520561e-01 -2.43318677e-01 9.71127808e-01 -1.16293788e+00
1.00610292e+00 -5.43903232e-01 -5.25981843e-01 3.75442177e-01
-4.20041323e-01 5.06006360e-01 1.34176993e+00 -4.17903423e-01
-1.29995561e+00 9.98413041e-02 -4.14009660e-01 -6.56289637e-01
5.90387918e-02 6.17068112e-01 -3.41290951e-01 4.38188434e-01
7.63681591e-01 6.51593029e-01 -3.46669741e-02 -1.87819213e-01
6.23078883e-01 7.99778521e-01 4.09965068e-01 -7.64797449e-01
1.02717876e+00 1.76809072e-01 -1.60601601e-01 -5.37796974e-01
-9.98009205e-01 1.43932641e-01 -4.90867108e-01 2.53378540e-01
1.05727851e+00 -8.99854779e-01 -1.72907948e-01 4.58700210e-01
-1.37230957e+00 -3.36832434e-01 -4.21149433e-01 1.93840191e-01
-7.13602662e-01 4.56078082e-01 -6.77767813e-01 -2.69805014e-01
-8.79498720e-01 -1.13526964e+00 1.00270987e+00 -1.04838520e-01
-2.16900900e-01 -1.04063010e+00 2.35506445e-01 5.44380367e-01
4.01392460e-01 1.58860445e-01 1.13328707e+00 -6.13938332e-01
-3.21174353e-01 7.92878401e-03 6.45500768e-05 7.87168860e-01
2.21949697e-01 -1.65276363e-01 -7.62395084e-01 -3.53356957e-01
4.09150839e-01 -6.59784198e-01 7.35316873e-01 9.03612673e-02
1.15480769e+00 -5.41008949e-01 7.48226931e-03 6.61360145e-01
1.14594936e+00 -1.82702795e-01 5.29736578e-01 1.85728759e-01
7.70970821e-01 7.03114688e-01 2.60899693e-01 6.89716935e-02
1.56871289e-01 5.94248474e-01 3.99874523e-03 1.45115674e-01
-1.06939554e-01 -8.31294656e-01 6.08579934e-01 1.09538031e+00
3.77313644e-01 -2.35870227e-01 -4.83927071e-01 4.41488892e-01
-1.70603549e+00 -1.06803644e+00 1.73719421e-01 1.92032707e+00
1.18131959e+00 2.69673586e-01 -3.33565980e-01 -6.40040264e-02
7.02029765e-01 3.21135163e-01 -4.74504411e-01 -6.82461977e-01
-1.36688069e-01 7.16450334e-01 1.12334907e-01 5.83529413e-01
-8.69699955e-01 1.12786877e+00 5.26728773e+00 8.64098668e-01
-1.29208529e+00 1.55271098e-01 3.75838637e-01 -7.51435757e-02
-9.67193663e-01 1.33270830e-01 -5.21656573e-01 7.80663371e-01
1.00168550e+00 -4.64706391e-01 4.70210552e-01 9.27727461e-01
4.10253197e-01 4.97726321e-01 -1.21727991e+00 6.33148730e-01
1.24892130e-01 -1.43559718e+00 4.61083025e-01 -3.08371484e-01
8.97123814e-01 -3.59864950e-01 9.74507630e-03 6.91206098e-01
2.26563975e-01 -1.08119678e+00 8.06339920e-01 2.24266976e-01
8.30068529e-01 -8.47792566e-01 7.05670774e-01 6.17007077e-01
-8.73149514e-01 -5.22598773e-02 -4.87926006e-01 -6.98659988e-03
3.31954628e-01 3.18592578e-01 -6.94507301e-01 6.51911497e-01
1.77230477e-01 6.70420587e-01 -4.54370260e-01 2.18058199e-01
-8.08105350e-01 7.59456158e-01 3.34467813e-02 -2.58786768e-01
1.32477373e-01 -3.89614075e-01 5.36119878e-01 1.10208595e+00
2.86859989e-01 -4.71188165e-02 -4.34807353e-02 1.42605615e+00
-3.25075805e-01 2.12481067e-01 -7.48587966e-01 1.77518222e-02
3.19052368e-01 9.59796369e-01 -1.34775087e-01 -2.50321567e-01
-3.62158954e-01 1.34205472e+00 6.98944867e-01 3.98103923e-01
-1.07431567e+00 -5.53989053e-01 2.22358674e-01 -1.59185022e-01
3.41848701e-01 -1.81511752e-02 -5.07920265e-01 -1.76484549e+00
3.16125512e-01 -9.86767232e-01 7.47219399e-02 -1.00924778e+00
-1.49630582e+00 7.26293921e-01 -9.68764275e-02 -1.12627447e+00
-5.66105068e-01 -3.39223802e-01 -1.04888046e+00 1.19304478e+00
-1.47605491e+00 -1.49831259e+00 2.79896725e-02 5.76935828e-01
8.36241245e-01 -2.77259767e-01 8.78582239e-01 -7.95520395e-02
-6.36374056e-01 9.42216277e-01 1.66608572e-01 1.87432453e-01
5.41380942e-01 -1.20178723e+00 6.36510670e-01 9.89637494e-01
1.71075612e-01 6.92084312e-01 7.23591268e-01 -4.90747988e-01
-1.22605288e+00 -1.28494334e+00 1.04914868e+00 -4.56513315e-01
7.30547130e-01 -6.60858274e-01 -9.66970623e-01 7.11312830e-01
3.92297179e-01 -4.27130371e-01 4.16696429e-01 -2.00556219e-01
-4.56397086e-01 8.87656678e-03 -1.06164539e+00 7.47165442e-01
9.06357169e-01 -7.29696751e-01 -1.15908885e+00 5.28460264e-01
1.01711118e+00 -3.70983034e-01 -6.64843202e-01 4.70602155e-01
1.98519334e-01 -6.29608631e-01 7.79687822e-01 -9.58086669e-01
1.20607519e+00 3.94265614e-02 3.91087160e-02 -1.59454811e+00
-1.96868002e-01 -4.46549326e-01 2.16556489e-01 1.37783945e+00
6.01741910e-01 -5.33450067e-01 8.75531495e-01 3.80158633e-01
-3.38595837e-01 -7.18546271e-01 -8.35066020e-01 -6.80236757e-01
7.01912105e-01 -7.55388439e-02 4.88963604e-01 8.21672738e-01
2.47336924e-01 9.41786706e-01 -4.73495871e-01 -1.23435836e-02
3.27180624e-01 6.33419394e-01 8.47043157e-01 -5.03748298e-01
-6.87330127e-01 -2.93217301e-01 2.05466568e-01 -1.03113198e+00
9.70596313e-01 -1.40816545e+00 -6.80380240e-02 -1.29858494e+00
2.42347568e-01 -1.35484204e-01 -2.04646140e-02 3.01005512e-01
-3.45473349e-01 -1.77851647e-01 1.83599740e-01 2.32755139e-01
9.87549424e-02 1.12070262e+00 1.40781355e+00 -2.94019550e-01
-2.21419469e-01 2.26194799e-01 -6.59694493e-01 3.83984476e-01
8.67460668e-01 -4.11745936e-01 -7.66273439e-01 -4.67972189e-01
1.16945863e-01 2.32050061e-01 3.43359053e-01 -5.89561820e-01
-1.60204232e-01 -3.23880553e-01 7.47310221e-02 -1.22111507e-01
1.57770097e-01 -5.38039923e-01 -2.04228554e-02 4.34810102e-01
-7.45682120e-01 1.22988727e-02 4.23526950e-02 5.41939020e-01
-4.21008289e-01 -5.63252747e-01 9.00397122e-01 -2.14354321e-01
-1.50354862e-01 -1.98688675e-02 -3.48620653e-01 4.38622832e-01
8.35900843e-01 1.41130626e-01 -1.70799941e-01 -3.72100115e-01
-5.11971951e-01 1.46871194e-01 5.42373002e-01 5.79806268e-01
5.36434829e-01 -1.41907215e+00 -8.04500401e-01 4.87799942e-01
1.18389547e-01 -7.41472840e-02 2.31038183e-01 1.93905402e-02
-5.24291098e-01 3.71955365e-01 -2.11675525e-01 -2.52122015e-01
-9.05281544e-01 8.53768885e-01 3.09351623e-01 -5.88526249e-01
-4.60459232e-01 5.65762818e-01 2.24494770e-01 -8.26303542e-01
-1.84735566e-01 -3.81202281e-01 -1.61869392e-01 -2.69039512e-01
8.33650604e-02 -2.32743084e-01 -1.41933933e-01 -4.18546289e-01
5.77736422e-02 3.79205972e-01 -9.58858803e-02 -3.55574399e-01
1.06811488e+00 1.09008774e-01 -2.58106459e-02 2.18624949e-01
1.36466074e+00 2.30279073e-01 -1.04847407e+00 -1.21369675e-01
-3.07549089e-01 -8.77493843e-02 -2.50666082e-01 -4.99275267e-01
-9.16923583e-01 9.35907006e-01 -3.50701623e-02 1.17951650e-02
8.52782428e-01 -1.54310465e-01 1.13645494e+00 4.08867538e-01
2.03308076e-01 -7.98912644e-01 2.95755416e-01 5.55965006e-01
1.09368360e+00 -9.73075569e-01 -4.28234011e-01 -3.30973655e-01
-9.04107571e-01 9.54843879e-01 6.24612868e-01 -3.44133586e-01
3.09981436e-01 -1.70837324e-02 -1.85356200e-01 1.84696704e-01
-7.17305362e-01 3.20172280e-01 1.96459278e-01 2.95438141e-01
2.52305090e-01 -9.04531479e-02 -5.93412936e-01 7.84159541e-01
-7.51552284e-01 -1.79530233e-02 7.65998900e-01 7.16076672e-01
-1.48401186e-01 -1.38607395e+00 -4.20502089e-02 2.09411815e-01
-4.10133481e-01 -3.59765440e-01 -4.00009632e-01 5.10433376e-01
-2.23749653e-01 7.88830817e-01 -1.03609964e-01 -2.98016578e-01
3.97291481e-01 3.80563110e-01 4.49952066e-01 -8.89552951e-01
-6.99990451e-01 -4.19589549e-01 5.41974492e-02 -1.27380580e-01
-1.38142198e-01 -4.75860029e-01 -1.10238886e+00 -1.73457399e-01
-2.38185704e-01 3.69088948e-01 5.39503694e-01 1.00740063e+00
3.22900474e-01 4.84728426e-01 9.94604826e-01 -5.72463632e-01
-1.16652358e+00 -1.02954841e+00 -1.98946074e-01 8.22624385e-01
1.58636451e-01 -2.08790630e-01 -4.68582153e-01 2.32953757e-01] | [11.67210578918457, 9.313986778259277] |
ab801610-bfb6-4322-9a27-d90544ac6633 | invariances-and-data-augmentation-for | 1711.04845 | null | http://arxiv.org/abs/1711.04845v1 | http://arxiv.org/pdf/1711.04845v1.pdf | Invariances and Data Augmentation for Supervised Music Transcription | This paper explores a variety of models for frame-based music transcription,
with an emphasis on the methods needed to reach state-of-the-art on human
recordings. The translation-invariant network discussed in this paper, which
combines a traditional filterbank with a convolutional neural network, was the
top-performing model in the 2017 MIREX Multiple Fundamental Frequency
Estimation evaluation. This class of models shares parameters in the
log-frequency domain, which exploits the frequency invariance of music to
reduce the number of model parameters and avoid overfitting to the training
data. All models in this paper were trained with supervision by labeled data
from the MusicNet dataset, augmented by random label-preserving pitch-shift
transformations. | ['Sham M. Kakade', 'Dean Foster', 'Zaid Harchaoui', 'John Thickstun'] | 2017-11-13 | null | null | null | null | ['music-transcription'] | ['music'] | [ 4.43982095e-01 -1.63290814e-01 -2.51703352e-01 -2.44058464e-02
-8.80254328e-01 -7.83101022e-01 3.84878963e-01 -4.62993950e-01
-4.15913016e-01 5.26091337e-01 6.72027111e-01 1.54603779e-01
-3.72588009e-01 -2.87120342e-01 -5.54493785e-01 -4.80430275e-01
-1.11879073e-01 7.15052783e-02 -5.11731863e-01 -2.39743665e-01
2.16874313e-02 1.81800440e-01 -1.34906113e+00 5.98438442e-01
2.54208356e-01 9.51515853e-01 -3.54919508e-02 7.39238262e-01
3.80264908e-01 7.92551935e-01 -9.81919229e-01 -3.30673009e-01
4.00645435e-01 -7.79905975e-01 -7.65521586e-01 -3.59280348e-01
8.46476078e-01 -1.34840295e-01 -5.86808801e-01 6.40119731e-01
9.07990813e-01 4.42309886e-01 4.43540096e-01 -7.02827156e-01
-4.73934948e-01 1.22636914e+00 3.62175070e-02 3.19207311e-01
3.33178699e-01 -1.30378336e-01 1.27534544e+00 -8.54303718e-01
5.31459987e-01 9.46250260e-01 1.13332939e+00 3.45528334e-01
-1.36583090e+00 -8.28123033e-01 -4.46377337e-01 2.60000795e-01
-1.54650009e+00 -6.49054825e-01 1.06701875e+00 -4.98455852e-01
1.27581322e+00 4.09705669e-01 9.73480761e-01 1.44172060e+00
-3.03699020e-02 5.86131394e-01 5.79526305e-01 -6.99952483e-01
-1.21642292e-01 -6.90561593e-01 -2.72563040e-01 1.83444977e-01
-3.55076820e-01 3.34555149e-01 -1.26044750e+00 -2.70943642e-01
9.22793448e-01 -4.26263779e-01 -5.12146652e-01 -8.64931494e-02
-1.68264771e+00 6.08690858e-01 2.20240697e-01 6.59606397e-01
-2.27856293e-01 4.92282331e-01 6.97412550e-01 4.18574601e-01
3.93911958e-01 1.05178785e+00 -6.40108824e-01 -4.55660343e-01
-1.46979201e+00 5.24877667e-01 7.44149208e-01 6.35425806e-01
1.86756909e-01 6.97158396e-01 -4.79014426e-01 8.37904572e-01
3.56905386e-02 4.26314443e-01 7.14797795e-01 -1.39238429e+00
3.89444023e-01 -1.19635612e-01 -1.31303538e-02 -6.57587349e-01
-6.18694544e-01 -1.26232421e+00 -7.19629586e-01 -2.78214633e-01
4.74180341e-01 -1.96092963e-01 -7.01569021e-01 1.98969078e+00
-3.38675290e-01 5.88020623e-01 -2.01588407e-01 9.34038877e-01
6.23529434e-01 4.68625873e-01 -1.48199931e-01 -1.71001986e-01
1.07839417e+00 -1.10449886e+00 -1.00790238e+00 4.27466929e-02
2.65780926e-01 -1.22011328e+00 1.03164864e+00 6.56842828e-01
-1.22968948e+00 -9.95182335e-01 -1.25472605e+00 -1.52406365e-01
2.03920919e-02 5.48525035e-01 5.13451815e-01 4.55348879e-01
-9.39465523e-01 1.17786515e+00 -5.38603127e-01 -1.21334895e-01
7.30793551e-02 3.45934808e-01 -1.11899428e-01 4.37203258e-01
-1.57926393e+00 7.66473413e-01 7.34901130e-01 -2.51360741e-02
-7.84167826e-01 -9.07652855e-01 -5.23716331e-01 2.07223132e-01
6.67494312e-02 -8.20208251e-01 1.73154187e+00 -1.16551387e+00
-1.81101882e+00 6.12793446e-01 7.46198893e-02 -8.76220167e-01
2.91630328e-01 -6.12974584e-01 -7.09380805e-01 1.04534678e-01
-1.54725596e-01 6.89231873e-01 1.20303345e+00 -6.33075058e-01
-2.91822910e-01 2.55617350e-01 -2.63295807e-02 2.51889080e-01
-2.85933793e-01 -5.89076318e-02 4.31384481e-02 -1.41645241e+00
-1.13644246e-02 -1.15768147e+00 2.44878903e-01 -6.75347447e-01
-2.77090073e-01 1.61245484e-02 5.09934306e-01 -7.73973525e-01
1.60566556e+00 -2.49698877e+00 5.08156061e-01 -4.96514253e-02
-1.18577987e-01 2.32165799e-01 -2.65698075e-01 6.48232758e-01
-4.78410661e-01 -3.28893587e-02 -1.06258266e-01 -5.18633842e-01
2.44187877e-01 -1.21150836e-01 -7.24722385e-01 3.92171741e-01
-4.70739510e-03 8.22947025e-01 -6.92235172e-01 8.14137608e-02
2.07890451e-01 7.28314281e-01 -6.75370872e-01 -1.02006324e-01
-1.22974612e-01 7.39454567e-01 2.61449039e-01 3.29231620e-01
2.22292900e-01 1.68867648e-01 2.34798640e-02 -5.23477077e-01
-6.21029846e-02 9.87705529e-01 -1.13143790e+00 2.47901058e+00
-4.44736540e-01 7.55913079e-01 -2.84082294e-01 -5.88233948e-01
7.35890269e-01 8.88882577e-01 7.71467388e-01 -2.35346854e-01
1.27254739e-01 4.06398296e-01 4.49376315e-01 -1.08694457e-01
4.90741342e-01 -2.28719220e-01 -2.46244092e-02 4.33138311e-01
7.22413898e-01 -1.93863183e-01 3.44688088e-01 -3.47627759e-01
9.23422039e-01 5.68692267e-01 2.34551832e-01 -2.40355954e-01
3.42894197e-01 -2.88774222e-01 6.22842729e-01 6.94725037e-01
9.05666314e-03 1.00490391e+00 1.69271268e-02 -6.20784104e-01
-9.25122738e-01 -1.02454710e+00 -1.00098729e-01 1.19050443e+00
-7.26103127e-01 -1.01790404e+00 -8.77044916e-01 -9.99726653e-02
-2.69464165e-01 6.40671074e-01 -4.89845484e-01 -1.48751557e-01
-8.59575570e-01 -2.44000033e-01 1.24001491e+00 3.45760822e-01
5.52092433e-01 -1.29553151e+00 -4.69144225e-01 5.05927742e-01
-7.75644004e-01 -9.00490046e-01 -6.50148928e-01 4.19397980e-01
-8.35519791e-01 -8.53537440e-01 -8.15231085e-01 -5.71340382e-01
-3.56364906e-01 -1.88518524e-01 1.23251498e+00 -3.25721264e-01
1.45591795e-01 8.93968195e-02 -4.47167367e-01 -6.90904200e-01
-1.56249121e-01 7.74213314e-01 2.47261047e-01 -1.96603909e-02
1.60233766e-01 -8.14572215e-01 -4.43774462e-01 6.33346513e-02
-7.50713587e-01 -2.62510926e-02 3.63044173e-01 8.97534609e-01
6.51381612e-01 -4.65358458e-02 6.22575402e-01 -3.04993421e-01
5.64235687e-01 1.00684449e-01 -5.10358401e-02 -2.52944499e-01
-6.61961734e-02 -3.58756870e-01 6.36618495e-01 -7.66047657e-01
-5.31529546e-01 8.01447481e-02 -2.60702282e-01 -8.15867424e-01
-9.73810814e-03 7.06740618e-01 1.91656664e-01 1.10796750e-01
9.57066536e-01 -5.51072657e-02 -5.49649775e-01 -7.18537569e-01
4.10633147e-01 4.31526870e-01 9.65629816e-01 -4.54416126e-01
6.81717694e-01 1.47788510e-01 1.56323880e-01 -7.42887139e-01
-1.38111162e+00 -3.84090930e-01 -8.63823652e-01 -1.81601822e-01
8.06351006e-01 -1.01934898e+00 -3.49794269e-01 5.09683430e-01
-1.11067903e+00 -3.32343698e-01 -8.45217586e-01 1.00237525e+00
-9.93152857e-01 -7.06190541e-02 -9.03977394e-01 -5.03414929e-01
-4.84033704e-01 -8.11440408e-01 9.03153002e-01 -2.47514457e-01
-7.85713732e-01 -7.08579361e-01 3.72535586e-01 1.04100689e-01
6.73848391e-01 2.10903928e-01 7.72596002e-01 -4.41400081e-01
-1.71515852e-01 -9.80940312e-02 5.62894225e-01 5.17456949e-01
5.99213392e-02 -1.94182485e-01 -1.51138878e+00 -3.44507188e-01
2.56106466e-01 -2.03917667e-01 1.05939198e+00 6.80073261e-01
1.08144343e+00 -1.59310102e-01 2.47203887e-01 1.05907261e+00
7.51648605e-01 -1.47272676e-01 4.98910308e-01 3.48964334e-01
6.03862822e-01 9.05759037e-02 2.06177056e-01 2.59696841e-01
-2.13135481e-01 8.92751336e-01 1.94239214e-01 -9.79380403e-03
-4.73647803e-01 -5.20867467e-01 4.32305306e-01 1.37331212e+00
-4.66958672e-01 -1.43629581e-01 -7.24155188e-01 2.88820595e-01
-1.77702940e+00 -1.20211160e+00 2.01648235e-01 2.17766953e+00
8.18565130e-01 5.88142052e-02 2.48008594e-01 7.14500129e-01
4.74291563e-01 3.50778490e-01 -2.34530374e-01 -2.34622404e-01
-3.77868056e-01 8.20460796e-01 2.76156157e-01 1.44605443e-01
-1.58373988e+00 8.36614788e-01 7.45792580e+00 1.02526665e+00
-1.20538688e+00 2.68367946e-01 -8.43363255e-02 -5.48678398e-01
1.86867952e-01 -7.00613856e-02 -1.81647331e-01 4.75638807e-02
1.42514729e+00 -1.68439057e-02 1.02933502e+00 4.50704843e-01
2.31719255e-01 5.77187777e-01 -1.16580760e+00 1.17810607e+00
1.74143419e-01 -1.40992475e+00 -1.11690268e-01 -8.44053403e-02
8.88689399e-01 2.98621327e-01 2.44967356e-01 4.83333915e-01
-2.51564145e-01 -1.12128186e+00 1.20641232e+00 7.81942248e-01
1.08763635e+00 -1.01776254e+00 6.30814791e-01 1.45492658e-01
-1.34150696e+00 -1.58778280e-01 -8.33991542e-02 -4.04733539e-01
2.21411839e-01 4.61547583e-01 -5.23506165e-01 7.02621579e-01
7.47201562e-01 1.05392587e+00 -2.34975487e-01 1.12828565e+00
-2.32104272e-01 1.06873167e+00 -2.89913505e-01 7.57633388e-01
-3.29780928e-03 2.02587303e-02 8.11437428e-01 1.31238961e+00
4.99209166e-01 -3.38109046e-01 1.46199942e-01 6.76576614e-01
-3.61787289e-01 4.01800908e-02 -3.17746997e-01 -3.30043525e-01
3.90160084e-01 1.05879653e+00 -3.46380413e-01 -6.30017594e-02
-8.52090940e-02 6.96682870e-01 6.71584308e-02 4.47343260e-01
-8.60785484e-01 -4.74157333e-01 5.06834626e-01 4.55526412e-02
2.38300711e-01 -5.06078005e-01 -1.51168466e-01 -1.09434640e+00
-3.23185444e-01 -1.09134090e+00 1.56940639e-01 -7.58651495e-01
-1.05057585e+00 5.70556879e-01 -1.08511798e-01 -1.53857720e+00
-5.76074421e-01 -3.00450504e-01 -4.09602314e-01 1.05189121e+00
-1.11136901e+00 -1.10282266e+00 1.38889521e-01 5.72601557e-01
4.51814711e-01 -3.45422626e-01 1.44075406e+00 6.23828232e-01
-1.03349328e-01 6.52881324e-01 -7.79586211e-02 2.73753315e-01
1.10163987e+00 -1.08576524e+00 6.93930447e-01 5.19356310e-01
8.63534510e-01 8.52211893e-01 5.82638204e-01 -2.83323139e-01
-9.98095691e-01 -1.03301299e+00 1.03023601e+00 -4.37018871e-01
5.78188062e-01 -3.36487293e-01 -6.60734951e-01 8.27068031e-01
4.13990736e-01 -1.40331164e-01 9.66032207e-01 2.52363980e-01
-4.88793224e-01 4.80471849e-02 -6.26174688e-01 5.19931793e-01
1.18458116e+00 -1.10206687e+00 -7.15237737e-01 2.18664378e-01
8.55695963e-01 -7.96690583e-01 -1.05595362e+00 6.31720424e-01
9.34487939e-01 -8.29174221e-01 9.63565052e-01 -4.96768504e-01
2.16606095e-01 -3.72328460e-01 -2.95664817e-01 -1.46624649e+00
-7.57285357e-01 -1.19990313e+00 -4.36954588e-01 1.16253662e+00
3.47156286e-01 -9.75253657e-02 3.38574111e-01 -5.27598143e-01
-4.50271457e-01 -3.44158053e-01 -1.23915231e+00 -8.75122011e-01
3.91831156e-03 -7.17970192e-01 6.48989081e-01 1.04140580e+00
-9.71640646e-02 4.78874415e-01 -6.53535903e-01 -4.23881590e-01
1.94138780e-01 -1.28672957e-01 6.87963903e-01 -1.25023532e+00
-6.36266649e-01 -5.53284943e-01 -2.35872701e-01 -6.79986417e-01
2.25677922e-01 -1.14062893e+00 -2.19275877e-01 -1.14887083e+00
-2.41042212e-01 2.12993562e-01 -7.14430034e-01 4.67800617e-01
2.82857239e-01 6.76012874e-01 6.65066004e-01 4.44487184e-01
-2.71491140e-01 4.98780400e-01 1.15054631e+00 -1.83304653e-01
-2.80042320e-01 1.04132555e-01 -3.02358508e-01 1.09965432e+00
8.17878723e-01 -5.02402842e-01 -4.17409450e-01 -4.50625151e-01
3.57143939e-01 5.37841208e-02 2.04412386e-01 -1.60299289e+00
-1.25705048e-01 2.78963596e-01 5.80130994e-01 -5.58632195e-01
7.34265387e-01 -5.68232596e-01 6.28976882e-01 2.07233161e-01
-7.72562325e-01 2.67972089e-02 4.73218650e-01 2.33156040e-01
-3.89764279e-01 -1.47668168e-01 6.24626040e-01 -9.07877535e-02
-1.22507855e-01 7.39706168e-03 -2.55493879e-01 1.76600888e-01
3.46572474e-02 1.81035757e-01 -1.85436279e-01 -5.13017178e-01
-1.12756121e+00 -7.20891118e-01 9.19394288e-03 5.59822917e-01
-2.16759136e-03 -1.71784914e+00 -8.37713003e-01 2.25875035e-01
-4.55698138e-03 -5.03976524e-01 2.79016197e-01 7.32811511e-01
-3.28551590e-01 6.58600032e-01 -2.65010059e-01 -4.90612239e-01
-8.22502136e-01 1.79733440e-01 6.79114997e-01 -3.59598935e-01
-6.72510386e-01 6.89904869e-01 -2.65791416e-01 -5.72527111e-01
3.75065058e-01 -5.17817676e-01 -1.19737655e-01 1.03670999e-01
4.81596708e-01 3.73721272e-01 3.55932057e-01 -9.25190628e-01
-1.47825792e-01 5.46593964e-01 3.68769825e-01 -6.12748981e-01
1.10713279e+00 1.05265908e-01 1.03225589e-01 9.91117001e-01
8.87340307e-01 1.44704357e-01 -1.01083744e+00 -2.70213872e-01
-4.74399589e-02 -1.89380407e-01 3.09109449e-01 -1.12849045e+00
-8.90243948e-01 8.76153648e-01 4.98043656e-01 1.12230740e-01
1.15072393e+00 -5.86712778e-01 7.66777813e-01 4.65666682e-01
3.02206755e-01 -1.21145523e+00 -1.29656360e-01 9.72930491e-01
1.15203404e+00 -4.33642805e-01 -5.06012291e-02 1.34465024e-01
-2.50741661e-01 1.36970520e+00 5.06495535e-02 -3.90948445e-01
7.13965416e-01 2.16919795e-01 1.73505351e-01 1.50543854e-01
-5.58735192e-01 -6.28192946e-02 8.11990798e-01 3.83508891e-01
9.08654988e-01 5.02282977e-02 -1.88452497e-01 8.69129181e-01
-1.09061277e+00 2.19312102e-01 1.68048054e-01 5.95339954e-01
-1.57270223e-01 -1.10933828e+00 -3.20107788e-01 1.87823266e-01
-1.01844764e+00 -3.44681323e-01 -4.88536447e-01 6.55453503e-01
5.63666940e-01 9.91704226e-01 -7.32870400e-02 -5.41460395e-01
4.86846149e-01 5.66783309e-01 8.03799272e-01 -5.37419021e-01
-1.13229167e+00 7.17217565e-01 1.53805614e-01 -4.70422536e-01
-7.45099425e-01 -5.43935418e-01 -1.04631388e+00 -1.32072093e-02
-3.01785290e-01 -9.11241993e-02 6.23443127e-01 9.28203881e-01
2.09245399e-01 1.21785295e+00 2.61994421e-01 -1.34921050e+00
-4.83486503e-01 -1.57163846e+00 -6.79654717e-01 2.06503242e-01
5.12544811e-01 -4.76155430e-01 -3.01123202e-01 2.12592170e-01] | [15.83660888671875, 5.351710319519043] |
0ddf1fad-a7fe-48c6-b89c-6893666a013b | class-guided-image-to-image-diffusion-cell | 2303.08863 | null | https://arxiv.org/abs/2303.08863v2 | https://arxiv.org/pdf/2303.08863v2.pdf | Class-Guided Image-to-Image Diffusion: Cell Painting from Brightfield Images with Class Labels | Image-to-image reconstruction problems with free or inexpensive metadata in the form of class labels appear often in biological and medical image domains. Existing text-guided or style-transfer image-to-image approaches do not translate to datasets where additional information is provided as discrete classes. We introduce and implement a model which combines image-to-image and class-guided denoising diffusion probabilistic models. We train our model on a real-world dataset of microscopy images used for drug discovery, with and without incorporating metadata labels. By exploring the properties of image-to-image diffusion with relevant labels, we show that class-guided image-to-image diffusion can improve the meaningful content of the reconstructed images and outperform the unguided model in useful downstream tasks. | ['Carola-Bibiane Schönlieb', 'Yinhai Wang', 'Elizabeth Mouchet', 'Guy Williams', 'Praveen Anand', 'Jan Oscar Cross-Zamirski'] | 2023-03-15 | null | null | null | null | ['drug-discovery'] | ['medical'] | [ 8.58188450e-01 -1.81711745e-02 -2.11044997e-01 -7.22306669e-01
-1.11513388e+00 -7.10850358e-01 8.51620913e-01 1.76267534e-01
-6.00465596e-01 8.99775624e-01 2.86722898e-01 -1.65646330e-01
-2.74792969e-01 -4.57905948e-01 -8.29897285e-01 -9.43952978e-01
3.05708379e-01 7.91764200e-01 2.96299547e-01 4.28420514e-01
3.67325902e-01 4.01688635e-01 -9.87525582e-01 7.02145636e-01
4.94501382e-01 6.70675099e-01 5.52388012e-01 8.04626346e-01
-2.18126431e-01 1.07881796e+00 -3.27920973e-01 -1.15849271e-01
2.50202864e-01 -6.87824607e-01 -9.34814274e-01 3.53124946e-01
6.27673507e-01 -5.02472162e-01 -3.27515304e-01 1.25861263e+00
6.57975793e-01 -1.58706293e-01 1.11508870e+00 -9.32121813e-01
-1.08745909e+00 2.48000085e-01 -6.16273463e-01 4.61286634e-01
1.91449076e-02 3.41604829e-01 4.42068636e-01 -8.39525342e-01
1.58760440e+00 1.43956852e+00 3.95417184e-01 8.18679333e-01
-2.00184584e+00 -4.02087539e-01 -2.30057165e-01 -6.43720925e-02
-1.09313476e+00 -6.88337803e-01 4.12777781e-01 -1.01212072e+00
8.34815323e-01 -5.60429171e-02 2.48490140e-01 1.07554388e+00
2.56147534e-01 6.08889341e-01 1.45712042e+00 -3.31322759e-01
2.38051564e-01 -2.17043832e-02 -4.29704711e-02 8.63544881e-01
9.58016366e-02 -9.88090858e-02 -6.02480829e-01 -2.11825699e-01
1.03591049e+00 1.62808985e-01 -1.29964769e-01 -1.98155776e-01
-1.69028127e+00 6.76561594e-01 1.62573129e-01 2.28349015e-01
-1.93490312e-01 3.62749487e-01 2.14756876e-01 2.08630592e-01
9.14300501e-01 2.64981359e-01 -5.57472229e-01 1.76905453e-01
-1.40497196e+00 1.94824725e-01 4.96314764e-01 9.54659581e-01
8.66315007e-01 -2.32303038e-01 -4.03253824e-01 5.67586005e-01
3.04982271e-02 6.60500705e-01 4.93101716e-01 -1.68343508e+00
-1.83973297e-01 2.53983270e-02 9.94740278e-02 -6.19690061e-01
-3.45078826e-01 -4.38962542e-02 -9.42681313e-01 2.19183758e-01
6.64416194e-01 2.39831284e-01 -1.28037536e+00 1.68563390e+00
2.44448915e-01 2.23049387e-01 5.38895233e-03 6.61276937e-01
8.70142043e-01 4.98530298e-01 4.65418488e-01 -3.11138093e-01
1.18509197e+00 -7.43161082e-01 -8.04971099e-01 1.55946866e-01
8.01609397e-01 -7.05769598e-01 9.74689841e-01 2.93147117e-01
-1.21140254e+00 -1.78803712e-01 -3.80387664e-01 -5.97602189e-01
-4.40526605e-01 4.04885225e-02 1.32848248e-01 3.31569910e-01
-1.42026293e+00 8.56821537e-01 -7.68248975e-01 -5.06575525e-01
1.08664465e+00 3.30876976e-01 -7.17069387e-01 -5.53487062e-01
-4.64968950e-01 7.55377173e-01 1.11222148e-01 -7.62143850e-01
-1.31373954e+00 -1.08150077e+00 -4.94699389e-01 -4.13662732e-01
-2.82224327e-01 -1.04824936e+00 9.11107123e-01 -1.01353359e+00
-1.18767667e+00 1.51854908e+00 -3.94157261e-01 -3.62445712e-01
5.17297506e-01 3.36982697e-01 3.11527282e-01 6.22027993e-01
4.52934027e-01 1.41457963e+00 9.60334241e-01 -1.27235711e+00
-5.09696305e-01 -3.77007723e-01 -2.48654112e-01 -8.03610496e-03
-2.68976688e-01 4.33598571e-02 -4.15479541e-01 -9.08866823e-01
1.83688439e-02 -6.54762030e-01 -1.84731483e-01 8.82381797e-01
-5.21011770e-01 6.55727759e-02 9.27123725e-01 -7.00566947e-01
4.09907341e-01 -2.09712481e+00 3.05771410e-01 -1.69282869e-01
5.81437588e-01 -1.24287337e-01 -4.61711258e-01 7.35971406e-02
-1.07959304e-02 2.81124383e-01 -7.54072845e-01 -4.35966939e-01
-2.95560032e-01 4.28067178e-01 6.77960739e-02 8.57531011e-01
1.67653978e-01 1.04943144e+00 -9.68238592e-01 -9.85800624e-01
6.48621768e-02 7.35445917e-01 -7.74346352e-01 1.50318235e-01
-3.99350166e-01 1.15043557e+00 -4.00746077e-01 5.95219076e-01
7.19022870e-01 -6.67857468e-01 2.96159238e-02 -2.65021205e-01
2.46034428e-01 -1.73336536e-01 -3.27084422e-01 2.04164457e+00
-4.41305228e-02 7.08467185e-01 3.01298678e-01 -8.17848325e-01
2.85488456e-01 1.14642046e-01 9.70124900e-01 -6.14181995e-01
-2.93676883e-01 1.68008775e-01 -1.42631352e-01 -5.33217847e-01
3.61920185e-02 -5.15638292e-01 4.14229423e-01 6.16279244e-01
2.12179780e-01 -4.50326890e-01 2.19957009e-01 6.10102654e-01
1.32081556e+00 2.69211680e-01 -2.87987947e-01 -6.53368354e-01
2.48261958e-01 8.97287503e-02 3.92457485e-01 8.15333486e-01
-3.50647479e-01 9.00704980e-01 2.76012033e-01 -2.34194174e-01
-1.50286353e+00 -1.02910614e+00 -5.40194929e-01 9.25917268e-01
-4.06875126e-02 -3.33315618e-02 -9.93378103e-01 -7.33352840e-01
-1.38444841e-01 3.23474586e-01 -6.33395135e-01 1.33496106e-01
-2.49291454e-02 -1.06759596e+00 5.83249092e-01 -5.36803380e-02
1.26856893e-01 -9.28043187e-01 -5.73989488e-02 2.16383263e-01
-1.68787569e-01 -1.19596493e+00 -6.87683821e-01 2.07625180e-01
-1.06121135e+00 -9.93853629e-01 -1.30362594e+00 -9.15061474e-01
1.08332396e+00 2.49203816e-01 1.14223862e+00 -5.86345494e-02
-7.21517444e-01 7.11007833e-01 -2.92584188e-02 -1.40540138e-01
-5.65158665e-01 -3.09741080e-01 -1.34912759e-01 -7.05758259e-02
-4.87386696e-02 -4.96608496e-01 -1.07091928e+00 6.36930689e-02
-1.44328988e+00 1.79879948e-01 4.58200425e-01 9.88701701e-01
1.12233078e+00 -1.06471412e-01 4.04222697e-01 -1.38016498e+00
3.76813203e-01 -4.77416068e-01 -2.17623875e-01 1.02864221e-01
-6.16562247e-01 1.88257784e-01 1.06984690e-01 -6.62257433e-01
-9.58862901e-01 4.23347831e-01 -2.27598965e-01 -4.77348834e-01
-1.69911548e-01 2.20863014e-01 -8.69290531e-02 -2.97761679e-01
7.56858468e-01 4.13500309e-01 2.43500561e-01 -4.14621800e-01
7.90671647e-01 5.23807228e-01 6.31244898e-01 -7.28553236e-01
3.70833248e-01 1.15327287e+00 2.05902100e-01 -8.32855582e-01
-8.73207867e-01 -6.51407599e-01 -7.71356523e-01 9.25231427e-02
1.15611196e+00 -1.05504143e+00 -2.12558925e-01 5.33491313e-01
-1.22544932e+00 -8.45369816e-01 -6.43008828e-01 2.35550180e-01
-1.09778357e+00 4.74102944e-01 -1.11173356e+00 -3.07962179e-01
-1.98106647e-01 -1.48993015e+00 1.49142432e+00 -1.80648252e-01
7.12513477e-02 -1.34065890e+00 -2.34953925e-01 1.61217943e-01
4.79482651e-01 1.13826446e-01 1.20710194e+00 -4.06061381e-01
-6.77140117e-01 2.57676959e-01 -6.76576734e-01 1.76447660e-01
2.79039651e-01 -3.84859294e-01 -1.06150055e+00 -1.97608814e-01
-4.91932034e-03 -5.16891360e-01 1.21434784e+00 9.71195221e-01
1.38011885e+00 -2.18079567e-01 -5.03606915e-01 7.50147045e-01
1.34771764e+00 -3.03391010e-01 8.68846536e-01 3.47493701e-02
7.91719735e-01 6.35772109e-01 1.32571667e-01 2.45611534e-01
3.00646991e-01 4.29537028e-01 1.19594522e-01 -5.50210834e-01
-8.02766860e-01 -8.29917565e-02 1.77439258e-01 5.58451593e-01
4.14383531e-01 -3.91633630e-01 -3.83436710e-01 6.89023972e-01
-1.47658467e+00 -9.14435387e-01 -3.10494483e-01 1.76331127e+00
1.47531593e+00 -3.05939466e-01 -5.44619039e-02 -3.70716304e-01
6.93261147e-01 -2.65970975e-01 -7.76123345e-01 3.44789833e-01
-6.10881805e-01 -8.35022181e-02 7.33219445e-01 6.49317861e-01
-1.07328033e+00 9.85017061e-01 7.29776096e+00 1.30863881e+00
-8.86355042e-01 6.70126319e-01 1.35349929e+00 4.56019901e-02
-5.34381986e-01 9.28899739e-03 -5.39663553e-01 4.27430362e-01
9.78632629e-01 6.34861365e-02 4.58278835e-01 2.69899875e-01
4.20776874e-01 -4.90040556e-02 -1.34194982e+00 1.10086906e+00
-4.04005982e-02 -1.70285034e+00 4.20116514e-01 3.52885574e-01
1.02284861e+00 2.36581594e-01 3.14121813e-01 -6.24439001e-01
6.61340356e-01 -9.26562071e-01 7.48871803e-01 1.12716448e+00
1.32031775e+00 -7.20894784e-02 2.91384399e-01 2.98487991e-01
-2.80772388e-01 3.01103413e-01 -6.02696478e-01 5.37723064e-01
9.94298514e-03 1.05431080e+00 -4.74495798e-01 3.90068442e-02
6.80263638e-01 1.28881288e+00 -5.20735085e-01 8.59018922e-01
2.54029185e-01 5.34885168e-01 -1.50296718e-01 7.08409786e-01
9.89525020e-02 -1.75863475e-01 3.44810992e-01 1.43198621e+00
5.29100955e-01 1.49246424e-01 2.29009345e-01 1.06695092e+00
-4.03902203e-01 6.48304168e-03 -7.77807355e-01 -1.40677720e-01
3.37017477e-02 1.09857559e+00 -1.12519276e+00 -7.53451645e-01
-4.02990669e-01 1.40455794e+00 1.27775386e-01 5.59550524e-01
-2.03515723e-01 9.10686180e-02 4.43615884e-01 5.98623157e-01
2.19335645e-01 -1.11722544e-01 -1.54846564e-01 -9.87461746e-01
-5.64132512e-01 -7.31392741e-01 2.03577101e-01 -1.10509300e+00
-1.85708296e+00 2.58929789e-01 -5.81718981e-02 -7.74404287e-01
2.87921131e-01 -6.02746189e-01 -1.73102003e-02 7.97109425e-01
-1.75339329e+00 -1.37148535e+00 1.31037965e-01 6.81118369e-01
5.73991776e-01 -1.96696911e-02 7.04453647e-01 5.25005996e-01
-1.17011964e-01 2.37294301e-01 8.61374021e-01 -4.21311408e-02
1.17925131e+00 -1.24243689e+00 5.65314256e-02 4.57937568e-01
4.48381416e-02 6.17420316e-01 5.69662154e-01 -7.71318436e-01
-1.10940421e+00 -1.49671924e+00 6.53353333e-01 -9.63220477e-01
6.00392759e-01 -3.08071580e-02 -6.82573318e-01 5.22853434e-01
3.63374770e-01 4.22722518e-01 5.81989706e-01 -6.98097885e-01
-4.16404217e-01 1.88083082e-01 -1.56506479e+00 2.78620631e-01
1.05291224e+00 -8.42316031e-01 9.07675475e-02 1.04877794e+00
6.74963951e-01 -1.10861637e-01 -1.15745652e+00 -1.75219506e-01
2.23418683e-01 -4.74665761e-01 1.27169180e+00 -4.74933803e-01
7.32172132e-01 -2.86931068e-01 -3.76462966e-01 -1.09742963e+00
-4.16629225e-01 -5.67980528e-01 3.46699357e-01 1.12468851e+00
3.09921533e-01 -2.46016547e-01 6.82602108e-01 5.23600101e-01
5.84450476e-02 -1.80525780e-01 -9.60534930e-01 -5.49157441e-01
6.69738948e-01 -1.86242489e-03 1.27685979e-01 9.01409030e-01
-3.08580905e-01 -3.00325081e-02 -2.31112227e-01 -3.33745629e-01
9.88200009e-01 -1.19391739e-01 2.56821394e-01 -1.12202322e+00
-1.75619796e-01 -5.07982016e-01 -3.62327605e-01 -1.06767392e+00
3.02341729e-01 -1.45401883e+00 2.15002149e-01 -1.68187916e+00
1.00422299e+00 -5.48269391e-01 -1.75926417e-01 5.33094108e-01
-1.00815669e-01 9.76874173e-01 -3.33769284e-02 9.01847303e-01
-9.28823292e-01 1.78664431e-01 1.71683812e+00 -7.86231160e-01
3.07864875e-01 -7.09866822e-01 -6.14155948e-01 5.72195113e-01
4.66474324e-01 -1.07401359e+00 -3.57848167e-01 -6.02511346e-01
-1.06890142e-01 -8.46946612e-03 4.55283165e-01 -4.65499669e-01
2.34558031e-01 -2.58147448e-01 4.05875415e-01 -1.71607792e-01
3.07903945e-01 -6.39161229e-01 1.23128526e-01 4.75167066e-01
-8.33902001e-01 -3.32852036e-01 -9.91248563e-02 9.97600913e-01
9.86628681e-02 -2.77567387e-01 1.04804957e+00 -4.44366157e-01
-4.41231847e-01 6.26762688e-01 -5.72433293e-01 2.19449550e-01
8.00971568e-01 4.67025749e-02 -5.65463483e-01 -5.35368383e-01
-1.09080803e+00 -2.80000061e-01 8.73349011e-01 -1.08834729e-01
6.48115575e-01 -1.08723903e+00 -9.08649564e-01 -8.93520564e-02
1.81466252e-01 -2.08231062e-01 1.91682428e-01 8.61867964e-01
-6.76795304e-01 3.60460669e-01 -1.44878969e-01 -9.81987357e-01
-1.13436818e+00 6.34405017e-01 3.68442982e-01 -1.01109184e-01
-6.98777616e-01 8.60780537e-01 8.78905416e-01 -5.92342675e-01
-4.39111702e-02 -1.26874551e-01 3.33233267e-01 -3.35962296e-01
4.85561132e-01 -3.22534703e-02 -2.14999780e-01 -8.73176396e-01
-1.15507185e-01 7.78794885e-01 -1.50911823e-01 -3.08347791e-01
1.46398902e+00 -5.06376266e-01 -3.67465556e-01 2.15238303e-01
1.18563187e+00 -3.96446586e-01 -1.62727773e+00 -3.93986791e-01
-2.36264989e-02 -4.42744285e-01 4.01758879e-01 -8.80584896e-01
-1.19928956e+00 1.22507477e+00 1.03506732e+00 -3.39160979e-01
9.07502651e-01 1.65236518e-01 6.59679294e-01 2.53101259e-01
2.06877887e-01 -8.59108806e-01 3.28840524e-01 1.17254205e-01
6.27177298e-01 -1.46661425e+00 2.22988889e-01 -3.47151726e-01
-5.15903354e-01 6.08693600e-01 -1.62480265e-01 2.12543234e-01
1.02185464e+00 4.28119630e-01 2.17287585e-01 -2.71303982e-01
-6.86097383e-01 -2.39895597e-01 -7.77472556e-03 1.22196019e+00
4.43631113e-01 -2.56051511e-01 -1.67455211e-01 1.17511116e-01
6.64374471e-01 3.82863104e-01 5.64250708e-01 8.93070638e-01
-3.98113132e-01 -1.29838789e+00 -2.61575073e-01 7.36789823e-01
-7.58128703e-01 -4.99316692e-01 -2.55645663e-01 1.43958211e-01
3.43644917e-01 8.88169944e-01 6.91733733e-02 2.40241602e-01
-3.10230821e-01 -3.27902101e-02 7.27359474e-01 -6.84553742e-01
-2.75000066e-01 4.93061304e-01 -3.24610025e-01 -3.96386802e-01
-1.03987026e+00 -6.57788575e-01 -1.35165966e+00 -3.89575779e-01
-1.09489419e-01 -1.05025880e-01 6.79122031e-01 8.76446247e-01
6.99589193e-01 4.58082020e-01 8.27430412e-02 -1.02898383e+00
1.08568251e-01 -8.40920866e-01 -8.88273716e-01 7.92659342e-01
4.02125627e-01 -3.88087213e-01 -7.47274533e-02 1.14508140e+00] | [11.282906532287598, -0.3488941490650177] |
349cc2bf-b579-4924-8134-4ce1e0b2782c | decoupled-multimodal-distilling-for-emotion | 2303.13802 | null | https://arxiv.org/abs/2303.13802v1 | https://arxiv.org/pdf/2303.13802v1.pdf | Decoupled Multimodal Distilling for Emotion Recognition | Human multimodal emotion recognition (MER) aims to perceive human emotions via language, visual and acoustic modalities. Despite the impressive performance of previous MER approaches, the inherent multimodal heterogeneities still haunt and the contribution of different modalities varies significantly. In this work, we mitigate this issue by proposing a decoupled multimodal distillation (DMD) approach that facilitates flexible and adaptive crossmodal knowledge distillation, aiming to enhance the discriminative features of each modality. Specially, the representation of each modality is decoupled into two parts, i.e., modality-irrelevant/-exclusive spaces, in a self-regression manner. DMD utilizes a graph distillation unit (GD-Unit) for each decoupled part so that each GD can be performed in a more specialized and effective manner. A GD-Unit consists of a dynamic graph where each vertice represents a modality and each edge indicates a dynamic knowledge distillation. Such GD paradigm provides a flexible knowledge transfer manner where the distillation weights can be automatically learned, thus enabling diverse crossmodal knowledge transfer patterns. Experimental results show DMD consistently obtains superior performance than state-of-the-art MER methods. Visualization results show the graph edges in DMD exhibit meaningful distributional patterns w.r.t. the modality-irrelevant/-exclusive feature spaces. Codes are released at \url{https://github.com/mdswyz/DMD}. | ['Zhen Cui', 'Yuanzhi Wang', 'Yong Li'] | 2023-03-24 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Li_Decoupled_Multimodal_Distilling_for_Emotion_Recognition_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_Decoupled_Multimodal_Distilling_for_Emotion_Recognition_CVPR_2023_paper.pdf | cvpr-2023-1 | ['multimodal-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'speech'] | [-9.39119309e-02 -1.70392677e-01 -3.25560793e-02 -2.37571850e-01
-4.83652830e-01 -5.70398808e-01 5.53591311e-01 1.05399586e-01
-5.69420815e-01 4.68144923e-01 2.69325286e-01 5.27470894e-02
-1.23155922e-01 -6.17477894e-01 -4.53588426e-01 -9.83379483e-01
9.72055644e-02 2.58740902e-01 -1.87434599e-01 -3.98853779e-01
-2.71664094e-02 3.66786540e-01 -1.55437565e+00 1.82427526e-01
1.05615163e+00 1.00778818e+00 2.43032396e-01 2.69043684e-01
-5.09570301e-01 2.19882473e-01 -2.93157041e-01 -4.88310158e-01
-2.09445450e-02 -4.90929067e-01 -2.99693823e-01 -7.84820914e-02
1.42188460e-01 1.90020397e-01 -3.61382902e-01 1.16934633e+00
7.31505632e-01 3.89484078e-01 8.22629273e-01 -1.41124809e+00
-8.47495556e-01 5.78802407e-01 -8.40135872e-01 -1.12036718e-02
3.61953467e-01 1.28416762e-01 9.14213479e-01 -9.93593872e-01
4.79395986e-01 1.31759274e+00 2.57869333e-01 5.91128945e-01
-1.32552254e+00 -8.70777547e-01 3.47633511e-01 2.44604111e-01
-1.49035490e+00 -1.91916615e-01 1.22283351e+00 -5.48031092e-01
6.07828498e-01 2.96331406e-01 6.77412570e-01 1.17599428e+00
-1.20526947e-01 8.62286925e-01 9.52187240e-01 -3.31561029e-01
8.97188634e-02 3.94262969e-01 1.27647594e-01 6.37731016e-01
-6.06036112e-02 -2.08451211e-01 -6.60120785e-01 6.02995569e-04
6.23387396e-01 9.19849128e-02 -4.00348425e-01 -6.69198036e-01
-1.32513905e+00 7.30623722e-01 6.15222752e-01 4.58370090e-01
-3.19116563e-01 -2.48362482e-01 6.70513272e-01 2.89235443e-01
1.36907538e-02 2.10169405e-01 -2.60883242e-01 -1.73456520e-01
-4.67407942e-01 -2.36509934e-01 4.91555691e-01 8.53148520e-01
1.02644503e+00 1.30192876e-01 -1.88052714e-01 1.19766510e+00
3.42989981e-01 5.25682211e-01 6.74924791e-01 -4.61736828e-01
5.59954107e-01 9.29078877e-01 -2.91739792e-01 -1.24912238e+00
-5.10071278e-01 -2.50091285e-01 -1.16032279e+00 -3.33456993e-02
1.00950524e-01 -2.44620353e-01 -6.81217790e-01 2.17845249e+00
5.06267667e-01 -1.35812223e-01 1.52881905e-01 9.38469768e-01
1.06802654e+00 6.27418756e-01 4.54592913e-01 -2.03483686e-01
1.53081191e+00 -7.58251548e-01 -8.30750108e-01 -8.74768347e-02
4.33326572e-01 -5.94261169e-01 1.24829352e+00 3.58399123e-01
-8.32364798e-01 -5.59142709e-01 -8.15460742e-01 -4.18784395e-02
-7.03866363e-01 1.92610770e-01 4.02185291e-01 3.70705962e-01
-7.48196363e-01 2.56999061e-02 -5.35716891e-01 -4.15673137e-01
1.78697452e-01 3.15620661e-01 -7.29450047e-01 1.23715729e-01
-1.46710718e+00 6.94714308e-01 8.13003898e-01 3.28185141e-01
-3.80926937e-01 -3.95801216e-01 -1.20456493e+00 -3.14930305e-02
4.48201805e-01 -7.38341272e-01 7.15362847e-01 -1.14528728e+00
-1.51052177e+00 7.26376891e-01 -2.58392356e-02 1.94289818e-01
3.07274252e-01 -1.50579721e-01 -8.20304751e-01 3.33472997e-01
-2.24063724e-01 5.80876112e-01 1.00369477e+00 -1.47621870e+00
-5.04019320e-01 -5.19798636e-01 -1.31135374e-01 6.33788228e-01
-8.88353407e-01 -1.09968640e-01 -7.17517793e-01 -7.73258805e-01
-1.17670752e-01 -8.52702975e-01 1.73267320e-01 -1.24034017e-01
-3.81700546e-01 -2.47231901e-01 8.28558505e-01 -5.05570292e-01
1.67797887e+00 -2.41224051e+00 8.05216551e-01 4.15855169e-01
3.31076741e-01 2.13320628e-01 -2.59016424e-01 6.65014088e-01
-1.32060677e-01 1.74838454e-02 -3.47694039e-01 -4.06339020e-01
3.47023517e-01 2.88903266e-01 1.11381337e-01 2.45607674e-01
1.58061132e-01 9.81029689e-01 -7.53478706e-01 -5.73515415e-01
4.64742392e-01 7.72535443e-01 -3.35325032e-01 2.74374366e-01
1.39672652e-01 4.86370921e-01 -5.03625214e-01 7.11833537e-01
7.44023144e-01 -1.61895752e-01 3.10956031e-01 -6.06599331e-01
-1.01627052e-01 -4.16629404e-01 -1.30414164e+00 1.94893014e+00
-4.47679937e-01 2.50560224e-01 1.87522128e-01 -1.02622271e+00
1.14262235e+00 8.02998915e-02 2.79296160e-01 -7.98288763e-01
3.27666968e-01 1.11928947e-01 -1.63043499e-01 -6.32939577e-01
5.50959110e-01 -3.16795945e-01 -3.14438254e-01 6.42637536e-02
2.65322208e-01 1.88974142e-01 6.74102679e-02 2.62783945e-01
4.19581056e-01 -4.18990925e-02 3.14907908e-01 -1.72456503e-01
6.90948606e-01 -4.36420768e-01 6.30817831e-01 2.90197432e-01
-1.52595848e-01 3.52385342e-01 4.17955041e-01 8.63512829e-02
-5.10925949e-01 -1.25068426e+00 -7.95785189e-02 1.20606399e+00
5.85689962e-01 -2.82536060e-01 -3.84364039e-01 -5.51733434e-01
1.66871056e-01 6.68464482e-01 -8.84768248e-01 -4.28808570e-01
-3.23534906e-01 -4.33257997e-01 4.43756670e-01 4.85059708e-01
5.13344586e-01 -1.14011848e+00 -4.36599016e-01 -8.80529508e-02
-1.75182119e-01 -7.95305908e-01 -4.72499251e-01 1.99756280e-01
-5.32150865e-01 -6.30876839e-01 -7.98592448e-01 -8.29969227e-01
6.42713904e-01 1.54803663e-01 7.67776132e-01 -3.83229703e-01
-1.66249096e-01 7.15543747e-01 -4.44946110e-01 -2.56693363e-02
9.14616287e-02 -6.95134103e-02 7.13191330e-02 4.19220954e-01
5.71804762e-01 -7.12182939e-01 -7.64340162e-01 1.95013791e-01
-1.11538577e+00 7.54369572e-02 7.57236362e-01 1.02592921e+00
7.27017999e-01 -1.49213582e-01 5.60370684e-01 -5.47089398e-01
8.31705987e-01 -8.77007067e-01 -1.14200011e-01 4.95087862e-01
-2.66137779e-01 2.27925088e-02 5.85638463e-01 -7.44041502e-01
-1.17986214e+00 -1.27571926e-01 9.22848880e-02 -6.69377089e-01
-4.32956994e-01 7.52145410e-01 -5.82058012e-01 1.35486975e-01
3.37889850e-01 4.19339299e-01 7.77927041e-02 -4.61910933e-01
7.94299006e-01 6.12679303e-01 4.83897179e-01 -7.71454453e-01
6.66562259e-01 1.94121182e-01 -3.48533541e-01 -8.66145372e-01
-1.28112525e-01 -4.45932597e-01 -4.98731107e-01 -4.45930928e-01
9.55514014e-01 -8.12140584e-01 -7.46899903e-01 3.59272271e-01
-7.70414948e-01 3.38947983e-03 -2.38227308e-01 5.91017425e-01
-2.01638430e-01 4.48342651e-01 -4.83832151e-01 -6.70449197e-01
-2.46319503e-01 -1.08579886e+00 1.02315021e+00 7.42212296e-01
-2.04930291e-01 -1.14070809e+00 5.46717420e-02 2.33231321e-01
2.63504326e-01 3.98569167e-01 9.45234656e-01 -6.18741572e-01
1.07161872e-01 -7.07410425e-02 -3.62089217e-01 3.55482459e-01
1.77881598e-01 -2.80863009e-02 -8.63974571e-01 -3.04423273e-01
-4.26158130e-01 -3.92232448e-01 8.35885942e-01 4.48523648e-03
1.03066945e+00 -1.73027098e-01 -2.55586058e-01 4.37955171e-01
1.26321244e+00 2.57571578e-01 5.11722445e-01 2.65925050e-01
9.09991920e-01 5.65951943e-01 4.23022389e-01 5.80198526e-01
5.70827425e-01 5.89936256e-01 3.19410175e-01 -1.73932672e-01
1.40968964e-01 -2.75861174e-01 3.33775401e-01 1.09433293e+00
-1.03488386e-01 -2.08024099e-01 -8.23022068e-01 3.86987895e-01
-1.99178159e+00 -8.16103220e-01 2.12346673e-01 2.16257477e+00
9.41306531e-01 -3.18728834e-01 2.00899094e-01 -1.27724826e-01
8.11013222e-01 1.49858534e-01 -5.70309043e-01 -5.49574792e-01
-5.03916502e-01 -1.54731959e-01 -2.06407472e-01 3.28437388e-01
-9.24022019e-01 8.09437096e-01 4.16294718e+00 1.09820724e+00
-1.23426330e+00 -4.17537577e-02 1.70726001e-01 -1.27478410e-03
-7.08450556e-01 -3.23169887e-01 -3.88584733e-01 4.79909718e-01
3.37690860e-01 -2.34719381e-01 3.19352388e-01 5.24702668e-01
-1.17887035e-01 -8.05704594e-02 -8.04192364e-01 1.52262509e+00
1.68632716e-01 -7.23940372e-01 3.48393798e-01 -1.53458148e-01
4.24649715e-01 -4.01618570e-01 3.04662913e-01 5.81938863e-01
-1.18944243e-01 -7.97284245e-01 5.06940126e-01 9.15571153e-01
7.67796278e-01 -9.58343208e-01 5.09145379e-01 1.47636652e-01
-1.48856163e+00 -2.26506982e-02 -3.78311127e-02 3.41421276e-01
1.52412519e-01 4.74554747e-01 -1.22710079e-01 1.02499449e+00
7.66558886e-01 6.87583983e-01 -4.91548240e-01 7.07412243e-01
-8.84062797e-02 1.31754085e-01 -1.99657932e-01 -3.46586742e-02
1.12059072e-01 -5.82998514e-01 7.06958234e-01 1.55349076e+00
3.98791105e-01 1.73765361e-01 1.50626376e-01 7.30598986e-01
-9.20295343e-02 4.97837901e-01 -6.36486888e-01 -2.47039035e-01
4.62212890e-01 1.53169000e+00 -3.45472187e-01 -1.98849022e-01
-4.71926212e-01 1.10616696e+00 5.58979750e-01 6.77116036e-01
-8.53580832e-01 -6.54110074e-01 6.50610030e-01 -4.28898722e-01
3.55248153e-01 -2.84645170e-01 -9.88007784e-02 -1.46675920e+00
-6.47863979e-03 -8.79978597e-01 7.56882608e-01 -6.84134960e-01
-1.73312986e+00 7.73648858e-01 2.94778515e-02 -1.31220174e+00
3.63157578e-02 -5.06832719e-01 -3.94757390e-01 1.00520968e+00
-1.40027821e+00 -1.35149086e+00 -4.75255638e-01 1.11024737e+00
1.67990997e-01 -2.32252300e-01 8.99698079e-01 3.63697559e-01
-8.07393551e-01 9.49180126e-01 5.38073704e-02 -5.71940131e-02
9.24623311e-01 -1.15285945e+00 -6.70056164e-01 5.52880704e-01
-4.80156168e-02 8.45675707e-01 4.35942411e-01 -4.83711660e-01
-1.60550570e+00 -6.91144347e-01 4.26573932e-01 1.18180551e-02
9.50580597e-01 -2.89199501e-01 -1.10744786e+00 3.14643562e-01
4.73610282e-01 -2.43602201e-01 1.14843321e+00 2.67392516e-01
-6.82232082e-01 -3.26045871e-01 -9.15987790e-01 7.88298726e-01
7.71106720e-01 -7.06430137e-01 -6.27532184e-01 -2.64138430e-01
5.07028401e-01 -1.04880452e-01 -1.02148306e+00 5.39935350e-01
6.40679538e-01 -8.68640244e-01 7.42523193e-01 -4.45721418e-01
2.28688866e-01 -5.37492514e-01 -1.87833920e-01 -1.37903047e+00
-2.66057163e-01 -5.46926618e-01 -3.36082518e-01 1.50759363e+00
3.55737358e-01 -8.89700949e-01 4.88542877e-02 6.23920321e-01
-8.08164030e-02 -7.19033957e-01 -9.79569852e-01 -6.93127275e-01
2.66815424e-02 -2.76499003e-01 5.71786344e-01 1.38854027e+00
5.02880752e-01 4.89374280e-01 -4.05286312e-01 5.06571941e-02
4.43559736e-01 2.39676863e-01 5.55240333e-01 -1.01175749e+00
-3.82933855e-01 -8.91951084e-01 -3.64916146e-01 -9.52773333e-01
3.86015289e-02 -9.83033419e-01 -2.85955876e-01 -1.43175268e+00
2.76578337e-01 -2.63872236e-01 -8.13226998e-01 7.76674926e-01
-2.47569695e-01 1.23175845e-01 3.78926307e-01 8.48965198e-02
-6.47857249e-01 1.08678007e+00 1.36423385e+00 -1.80754066e-01
-4.12220865e-01 -5.72987020e-01 -7.49211907e-01 4.59501028e-01
1.05117393e+00 3.06536630e-02 -5.73164463e-01 -3.96838516e-01
1.23347551e-01 -2.54482422e-02 3.41047138e-01 -5.41105688e-01
2.88863659e-01 -1.45260647e-01 3.31163824e-01 -4.07275468e-01
4.18832093e-01 -9.60992038e-01 2.90076017e-01 8.45432729e-02
-1.93249688e-01 2.89305151e-02 6.28486574e-01 7.03870237e-01
-5.82165003e-01 2.78381795e-01 6.34038508e-01 2.43227959e-01
-1.02938449e+00 2.42455199e-01 -6.38862774e-02 -4.46436517e-02
1.09385908e+00 -3.17966133e-01 -2.82010883e-01 -1.98020384e-01
-9.29077029e-01 4.30561453e-01 2.02400327e-01 7.50478625e-01
7.32688725e-01 -1.63383090e+00 -4.15095657e-01 1.58264011e-01
5.36951602e-01 -2.55410820e-01 9.30113971e-01 1.06891239e+00
6.45536855e-02 -9.01408941e-02 -5.05928636e-01 -5.24147451e-01
-1.20570242e+00 6.45374715e-01 2.24747092e-01 -6.73498772e-03
-4.23735172e-01 7.68220782e-01 6.42219365e-01 -4.65789646e-01
2.27483749e-01 5.26635423e-02 -5.73558390e-01 5.02791345e-01
3.04866135e-01 2.69024938e-01 -2.64618069e-01 -9.52688575e-01
-4.36054230e-01 8.07693303e-01 8.05705562e-02 -1.27795786e-01
1.09449017e+00 -3.69921148e-01 -2.30624199e-01 8.16063225e-01
1.27689862e+00 4.51322086e-02 -9.48451817e-01 -4.10773337e-01
-3.72805864e-01 -3.48576516e-01 -1.02216102e-01 -9.57369328e-01
-1.17949855e+00 1.05499160e+00 6.41612589e-01 1.50736079e-01
1.60828114e+00 3.88409719e-02 5.66534996e-01 6.70583025e-02
8.43809023e-02 -1.11805081e+00 5.68982251e-02 4.27035004e-01
1.09632146e+00 -1.26038361e+00 -4.16716725e-01 -1.33972272e-01
-1.18384635e+00 1.22673810e+00 7.17950046e-01 2.50981212e-01
6.45417929e-01 1.23179536e-02 7.60585293e-02 -2.67582446e-01
-2.79383153e-01 -2.17819005e-01 7.46124029e-01 4.14807379e-01
3.08808386e-01 3.77236843e-01 -4.60545421e-01 8.99979889e-01
9.52089652e-02 -3.05911481e-01 -2.41919234e-01 7.68183827e-01
-2.08051264e-01 -1.11830628e+00 -2.28158310e-01 5.57267070e-02
2.89639145e-01 -7.25961179e-02 -3.62144381e-01 9.71669793e-01
2.80616343e-01 7.84431636e-01 -1.59706876e-01 -7.22088158e-01
4.22737688e-01 1.66654274e-01 3.59539211e-01 -1.83539450e-01
-4.74160999e-01 3.05876404e-01 -6.64444715e-02 -4.42330778e-01
-4.24480200e-01 -5.58529377e-01 -1.45336699e+00 -1.47240490e-01
-1.92766219e-01 2.26943403e-01 3.98656249e-01 6.27976835e-01
6.05611503e-01 5.48091412e-01 5.85480034e-01 -7.67240822e-01
-1.04315616e-01 -7.68693805e-01 -8.45812023e-01 5.39858103e-01
1.00240037e-01 -9.47148740e-01 -3.82511884e-01 -2.58823335e-02] | [13.085381507873535, 4.99919319152832] |
f2bea63f-b4db-4a88-8c35-352516fb29a9 | conic-colon-nuclei-identification-and | 2111.14485 | null | https://arxiv.org/abs/2111.14485v1 | https://arxiv.org/pdf/2111.14485v1.pdf | CoNIC: Colon Nuclei Identification and Counting Challenge 2022 | Nuclear segmentation, classification and quantification within Haematoxylin & Eosin stained histology images enables the extraction of interpretable cell-based features that can be used in downstream explainable models in computational pathology (CPath). However, automatic recognition of different nuclei is faced with a major challenge in that there are several different types of nuclei, some of them exhibiting large intra-class variability. To help drive forward research and innovation for automatic nuclei recognition in CPath, we organise the Colon Nuclei Identification and Counting (CoNIC) Challenge. The challenge encourages researchers to develop algorithms that perform segmentation, classification and counting of nuclei within the current largest known publicly available nuclei-level dataset in CPath, containing around half a million labelled nuclei. Therefore, the CoNIC challenge utilises over 10 times the number of nuclei as the previous largest challenge dataset for nuclei recognition. It is important for algorithms to be robust to input variation if we wish to deploy them in a clinical setting. Therefore, as part of this challenge we will also test the sensitivity of each submitted algorithm to certain input variations. | ['Nasir Rajpoot', 'Fayyaz Minhas', 'Shan E Ahmed Raza', 'David Snead', 'Thomas Leech', 'Giorgos Hadjigeorghiou', 'Quoc Dang Vu', 'Mostafa Jahanifar', 'Simon Graham'] | 2021-11-29 | null | null | null | null | ['explainable-models', 'nuclear-segmentation'] | ['computer-vision', 'medical'] | [ 4.84747767e-01 -4.55097295e-02 4.31038029e-02 -1.55122206e-01
-9.30357575e-01 -9.07474220e-01 5.32338083e-01 8.86349142e-01
-8.03693235e-01 7.04774082e-01 1.97904423e-01 -3.91173840e-01
-4.26134616e-02 -4.97899622e-01 -1.23181619e-01 -1.13054121e+00
5.07658496e-02 9.07329500e-01 3.58015329e-01 -2.00483575e-02
2.60034770e-01 8.61479521e-01 -1.21120918e+00 5.87497473e-01
2.39415318e-01 3.91606897e-01 6.22257998e-04 1.25519788e+00
2.31586229e-02 5.64348221e-01 -3.32221836e-01 5.54462522e-02
-1.09646291e-01 -5.13600290e-01 -1.20883179e+00 -6.12415150e-02
2.86062825e-02 1.99834377e-01 3.03449869e-01 7.15026140e-01
7.19863236e-01 -2.47971237e-01 1.10646582e+00 -8.78976107e-01
2.74136752e-01 4.82116163e-01 -2.86983728e-01 7.67953694e-01
-3.22933681e-02 2.23606318e-01 7.62005389e-01 -5.17772019e-01
1.00954437e+00 6.40082240e-01 7.75784433e-01 5.57783723e-01
-1.49396181e+00 -2.25691497e-01 -6.32689595e-01 -1.38035119e-01
-1.35725474e+00 -3.29423994e-01 -1.71742588e-01 -6.49447560e-01
1.04400992e+00 7.62273490e-01 9.59261954e-01 5.07773995e-01
1.08915552e-01 4.37750131e-01 1.39964366e+00 -4.57079023e-01
3.47566545e-01 1.14670426e-01 -1.22488938e-01 4.03321803e-01
2.84250021e-01 -3.52819026e-01 -2.86947973e-02 -1.93104684e-01
5.68439722e-01 -2.11116076e-01 -2.21416295e-01 -6.67279139e-02
-1.67586279e+00 6.65518939e-01 2.04823777e-01 8.22551191e-01
4.66622077e-02 1.58361524e-01 7.26069510e-01 -1.03799053e-01
1.84156597e-01 5.48368633e-01 -4.40418988e-01 -1.53854266e-01
-1.11372733e+00 1.04391418e-01 7.67767191e-01 2.90423959e-01
3.05481493e-01 -5.38563728e-01 4.89102975e-02 7.23030746e-01
3.54809731e-01 9.17600840e-02 7.35871851e-01 -7.68992901e-01
-3.41747761e-01 8.33079100e-01 -2.09036991e-01 -8.11603069e-01
-9.49146569e-01 -2.41633877e-01 -1.07864320e+00 1.48830056e-01
8.41098964e-01 2.58165479e-01 -9.91690516e-01 1.18215668e+00
6.84405863e-01 -1.75659269e-01 -7.66374394e-02 7.53764331e-01
7.81172276e-01 2.43186519e-01 3.24871957e-01 -9.99132246e-02
1.59104574e+00 -3.45588148e-01 -2.64937699e-01 3.73738855e-02
1.37578094e+00 -1.03797138e+00 5.65932274e-01 2.78184742e-01
-6.94584787e-01 1.31502330e-01 -8.69895577e-01 -1.64861590e-01
-7.45608628e-01 -6.00541644e-02 6.26794994e-01 6.22977078e-01
-1.29357696e+00 3.45514894e-01 -1.04828608e+00 -8.86920691e-01
6.49156570e-01 8.65826964e-01 -7.68452168e-01 5.37202433e-02
-6.03962839e-01 1.02165782e+00 6.16761565e-01 1.76285446e-01
-3.33703578e-01 -8.76179218e-01 -4.91140217e-01 -3.59275669e-01
-1.72785535e-01 -7.35049546e-01 1.01759255e+00 -3.58824104e-01
-8.30458164e-01 1.56554246e+00 -1.11611657e-01 -2.26732552e-01
4.77187723e-01 8.51670325e-01 -4.30210074e-03 2.78498292e-01
6.81027770e-02 9.27563548e-01 8.35355446e-02 -9.68601525e-01
-9.23476875e-01 -2.31956422e-01 -4.21330810e-01 -7.96087310e-02
1.55962646e-01 8.47074762e-02 -2.08479121e-01 -3.76560807e-01
7.97750950e-02 -1.06588817e+00 -5.10195792e-01 1.05003767e-01
-3.73073548e-01 -1.54323041e-01 5.38400888e-01 -4.96851414e-01
7.50108182e-01 -2.02184987e+00 5.59409745e-02 4.17379647e-01
6.45091414e-01 2.48497978e-01 2.16225132e-01 3.10620874e-01
-4.59006019e-02 6.78401530e-01 -1.35235369e-01 -7.17181563e-02
-7.33699277e-02 1.90404534e-01 3.69186431e-01 1.00256443e+00
3.35998893e-01 8.03499460e-01 -9.23929095e-01 -9.18443084e-01
1.56552345e-01 5.90129077e-01 -3.61455977e-01 -1.87372029e-01
5.95007576e-02 5.43620408e-01 -2.00015187e-01 8.74453485e-01
4.11600441e-01 -3.53346944e-01 3.92451137e-01 -4.36212495e-02
-2.73069218e-02 -1.90971226e-01 -8.80915582e-01 8.98747325e-01
3.62981893e-02 6.01750970e-01 1.67307273e-01 -8.44197333e-01
5.38648009e-01 4.69105691e-01 6.49262905e-01 -2.05300346e-01
4.78684038e-01 7.12255180e-01 4.27154899e-01 -6.54572621e-02
1.06509022e-01 -7.80031621e-01 8.60401019e-02 4.15886462e-01
5.19262254e-02 -2.88190901e-01 6.98456824e-01 7.37430677e-02
1.49865782e+00 -4.54237998e-01 5.76572001e-01 -5.30806005e-01
7.42109895e-01 3.39704365e-01 4.03701931e-01 3.35747480e-01
-7.14827657e-01 8.83107841e-01 8.27859044e-01 -4.64567751e-01
-1.30661070e+00 -5.95943809e-01 -5.11024177e-01 7.44272053e-01
-3.73748034e-01 -1.43915683e-01 -6.40349925e-01 -4.81951803e-01
-1.94290921e-01 -9.38177109e-03 -9.72391367e-01 3.04685384e-01
-4.20836121e-01 -1.34287179e+00 9.36892509e-01 3.86900812e-01
-1.00162782e-01 -1.03439522e+00 -6.25042915e-01 2.87255585e-01
-6.17097951e-02 -9.62576985e-01 -1.12139761e-01 6.65530622e-01
-7.47093141e-01 -1.43126309e+00 -8.65899682e-01 -1.01766217e+00
9.73956883e-01 -1.03026703e-01 1.05697215e+00 6.57571852e-01
-9.18098032e-01 3.04842561e-01 -2.96550065e-01 -5.31987786e-01
-8.50650489e-01 3.09886545e-01 -1.54774964e-01 -5.02239406e-01
4.14184511e-01 -3.17717999e-01 -7.34163344e-01 2.03492478e-01
-1.13961959e+00 -2.10756343e-02 6.99139059e-01 9.02395964e-01
1.15661490e+00 1.07209295e-01 1.72094330e-01 -1.24592316e+00
2.33200073e-01 -5.18050551e-01 -2.81054765e-01 8.52424428e-02
-2.79700607e-01 -2.72391200e-01 6.04241192e-01 -6.58425689e-02
-3.77173305e-01 2.02804416e-01 -1.98104754e-01 3.73200685e-01
-5.20020127e-01 5.47745824e-01 3.26986969e-01 -3.25338483e-01
6.93352759e-01 7.53730312e-02 2.05069244e-01 2.08459832e-02
-1.08335195e-02 5.15008152e-01 3.87012094e-01 3.68742347e-02
5.57998657e-01 7.85399556e-01 6.67854249e-01 -8.41891110e-01
-5.06912529e-01 -1.16431522e+00 -9.02857244e-01 -7.54051059e-02
8.73263359e-01 -5.36189854e-01 -7.03173995e-01 4.20489728e-01
-6.23570383e-01 -4.99759108e-01 -1.90480202e-01 3.10362846e-01
-5.83875418e-01 2.01885775e-01 -7.41243780e-01 -5.23687363e-01
-5.18251657e-01 -1.14449501e+00 1.08340979e+00 2.34592155e-01
-7.05958843e-01 -1.32952142e+00 6.23838425e-01 5.96004546e-01
2.44244903e-01 6.46211207e-01 1.25781536e+00 -9.14307654e-01
-1.15049168e-01 -4.23283577e-01 -1.63973942e-01 -1.71787441e-01
1.04151489e-02 6.27394259e-01 -8.35382819e-01 -3.30069721e-01
-3.17787886e-01 -4.57860529e-01 9.49006617e-01 4.07560438e-01
8.40608120e-01 -1.80085242e-01 -6.50122643e-01 5.78177631e-01
1.53607976e+00 -1.03729300e-01 6.69701815e-01 6.63531423e-01
2.72589892e-01 8.38488877e-01 4.71745133e-01 -1.18322581e-01
8.66713598e-02 1.00408934e-01 1.39851004e-01 -5.03408611e-01
-2.25153752e-02 4.00313765e-01 -3.35589111e-01 7.45357335e-01
-2.37839550e-01 -2.99713820e-01 -1.54405951e+00 9.17098224e-01
-1.31890786e+00 -9.05010343e-01 -4.01014954e-01 1.77540493e+00
1.07796872e+00 1.00325376e-01 3.76290202e-01 6.14042342e-01
5.87448776e-01 -3.44413877e-01 -1.45800188e-01 -4.14479971e-01
-5.14408275e-02 3.71888995e-01 5.06953359e-01 2.71386564e-01
-1.14131582e+00 4.41847980e-01 6.87620354e+00 8.33644986e-01
-1.09423947e+00 -2.44659156e-01 1.24119723e+00 1.26045570e-01
1.14685692e-01 -1.33842900e-01 -8.52464736e-01 2.52902120e-01
1.07507336e+00 -1.03653491e-01 6.99831322e-02 2.89852679e-01
3.20720792e-01 -5.65517843e-01 -1.02583253e+00 5.48141479e-01
-1.93267033e-01 -1.42403603e+00 -3.90976489e-01 3.72199237e-01
4.51335400e-01 2.55310744e-01 -1.31860271e-01 1.73483491e-02
1.11930557e-01 -1.54121101e+00 5.82001917e-02 3.82994264e-01
9.51605380e-01 -7.30417550e-01 1.39774644e+00 2.92168528e-01
-1.02025986e+00 3.63333911e-01 -3.11166495e-01 2.89275974e-01
-1.58138767e-01 3.47637475e-01 -1.61677170e+00 1.27115464e-02
3.73581648e-01 3.05093199e-01 -8.54187727e-01 1.47056377e+00
4.29855376e-01 6.89041853e-01 -5.87045312e-01 -2.24582240e-01
2.39332050e-01 1.65892944e-01 1.68742165e-01 1.58806086e+00
1.87989071e-01 1.81811899e-01 -1.50559753e-01 4.10297275e-01
1.77223966e-01 3.61527592e-01 -4.13758717e-02 -5.48294663e-01
2.79127777e-01 1.86511827e+00 -1.87709939e+00 -5.85596226e-02
5.00587747e-02 5.21409571e-01 2.52708495e-01 -1.98471949e-01
-3.74906093e-01 -6.16094097e-02 2.50221372e-01 4.71970379e-01
1.42905027e-01 2.16650113e-01 -2.55204648e-01 -4.37972486e-01
-6.11756146e-01 -8.53621602e-01 6.22494400e-01 -3.67575943e-01
-1.33870828e+00 3.03811252e-01 -2.48482600e-01 -9.09116387e-01
-1.16078757e-01 -8.19330037e-01 -5.52599847e-01 7.18284130e-01
-1.25498176e+00 -1.35965991e+00 -1.16465971e-01 -4.09385338e-02
-5.48582450e-02 2.67767340e-01 1.18801165e+00 -5.07384501e-02
-3.50143492e-01 3.41713905e-01 1.92654774e-01 1.67332232e-01
5.75319469e-01 -1.78884578e+00 -7.27540776e-02 1.69864982e-01
2.33216099e-02 5.20534754e-01 7.85158873e-01 -3.98028284e-01
-9.87300038e-01 -1.18735099e+00 1.10779297e+00 -7.06270456e-01
6.83639944e-01 -1.50894314e-01 -7.47681499e-01 5.24423897e-01
-1.46052256e-01 3.38245571e-01 1.54752731e+00 -1.75466001e-01
1.19842157e-01 3.66193533e-01 -1.36194646e+00 5.42330563e-01
2.39936858e-01 -3.22809249e-01 -1.47860542e-01 5.24093211e-01
-6.38607442e-02 -6.08478308e-01 -1.20744479e+00 1.83649614e-01
4.76429313e-01 -8.28171253e-01 7.38858283e-01 -3.58002275e-01
3.31296980e-01 -4.72806126e-01 2.64732808e-01 -8.76913309e-01
-3.48184913e-01 -1.61738068e-01 6.09770954e-01 1.04869878e+00
4.92688686e-01 -4.95071411e-01 1.17443216e+00 5.24507344e-01
1.12662941e-01 -1.07921183e+00 -1.12460721e+00 -1.80883512e-01
4.20581222e-01 -1.96052156e-02 2.30099395e-01 8.47333193e-01
3.54300231e-01 3.12691554e-02 2.75211722e-01 -2.51229912e-01
3.95591944e-01 -3.56479943e-01 6.03135586e-01 -1.14841557e+00
-7.54201636e-02 -6.86114609e-01 -1.15106702e+00 1.27025262e-01
-2.86273181e-01 -1.24559259e+00 -6.81486577e-02 -1.53972948e+00
7.03320920e-01 -4.30463910e-01 -2.34331146e-01 6.01881385e-01
-1.87145755e-01 9.48883176e-01 -4.12690453e-02 3.11815739e-01
-6.31218672e-01 -2.59829611e-01 1.19465029e+00 -1.39454901e-01
1.27578735e-01 -2.28508607e-01 -6.97205901e-01 7.27350831e-01
8.10051322e-01 -8.72830272e-01 7.86775425e-02 4.31485474e-01
4.42420036e-01 -3.14781964e-01 4.53618735e-01 -1.06803954e+00
3.23008090e-01 -2.45422930e-01 7.40591824e-01 -8.19736898e-01
7.05699772e-02 -4.60149378e-01 4.01490390e-01 7.19178379e-01
-1.75486624e-01 -2.33282283e-01 2.60083109e-01 4.71000671e-01
-1.44939259e-01 -3.13942909e-01 9.25419271e-01 -4.56057250e-01
-2.15725854e-01 2.51119286e-02 -1.00248909e+00 3.17816660e-02
1.12006629e+00 -7.48657882e-01 -3.85689765e-01 9.02259797e-02
-8.41082275e-01 3.48781556e-01 8.22428524e-01 -5.06731510e-01
1.41242594e-01 -9.91702974e-01 -7.89930701e-01 -1.83632180e-01
2.08542258e-01 4.34717059e-01 4.37221676e-01 1.37054443e+00
-1.34696066e+00 6.98779166e-01 -1.78929046e-01 -7.52608120e-01
-1.63532794e+00 1.61536351e-01 5.88602126e-01 -7.63483763e-01
-2.32616365e-01 1.18445885e+00 -1.49690971e-01 -6.02147460e-01
-3.17611128e-01 -2.36890227e-01 -5.21762550e-01 1.52046859e-01
4.12863076e-01 2.66034991e-01 1.83416486e-01 -8.78278792e-01
-4.36352819e-01 1.87917292e-01 -3.25183392e-01 1.34789512e-01
1.41761279e+00 2.68042356e-01 -5.35753608e-01 5.82064211e-01
1.08908010e+00 -1.32487506e-01 -6.92858875e-01 4.71661687e-01
2.37102807e-01 -2.11949244e-01 8.36064741e-02 -7.94377983e-01
-7.05390990e-01 5.75271487e-01 5.81473053e-01 2.15021282e-01
9.34519231e-01 -7.09538534e-03 4.89439279e-01 3.19584012e-02
1.73625037e-01 -9.75941718e-01 -4.11565810e-01 2.79370844e-01
4.54386443e-01 -1.10776627e+00 3.61123711e-01 -5.28879344e-01
-1.96032539e-01 1.29629588e+00 1.38286203e-01 1.85297951e-01
4.78016078e-01 6.33982837e-01 3.23787004e-01 -5.31546175e-01
-9.30625439e-01 -8.10479745e-02 1.88758463e-01 7.50338912e-01
8.71356308e-01 2.18478823e-03 -5.63091576e-01 4.60912079e-01
-3.51917505e-01 9.70486253e-02 7.36911833e-01 1.04404438e+00
-3.65897745e-01 -1.25155592e+00 -3.56649995e-01 8.76194894e-01
-8.91538680e-01 1.33717939e-01 -7.31761456e-01 1.11196804e+00
2.00129390e-01 4.98607993e-01 -3.84072177e-02 6.53311834e-02
-1.20232515e-02 -1.49557963e-01 3.46253514e-01 -7.79286027e-01
-9.07494187e-01 2.95359403e-01 4.00110856e-02 -1.22805409e-01
-6.91747904e-01 -8.02523732e-01 -1.72511220e+00 -4.87669498e-01
-4.44462776e-01 9.42767784e-03 3.66652012e-01 9.97637570e-01
-2.26090640e-01 3.84335399e-01 1.30053714e-01 -7.19588578e-01
-1.20690808e-01 -8.51879895e-01 -9.92993236e-01 2.33303532e-01
1.32793814e-01 -3.26957285e-01 -8.07815671e-01 4.47064936e-01] | [15.070718765258789, -3.1342720985412598] |
89803d67-fed6-4862-82db-27bea6792b13 | text-preprocessing-and-its-implications-in-a | null | null | https://aclanthology.org/2021.ranlp-srw.13 | https://aclanthology.org/2021.ranlp-srw.13.pdf | Text Preprocessing and its Implications in a Digital Humanities Project | This paper focuses on data cleaning as part of a preprocessing procedure applied to text data retrieved from the web. Although the importance of this early stage in a project using NLP methods is often highlighted by researchers, the details, general principles and techniques are usually left out due to consideration of space. At best, they are dismissed with a comment “The usual data cleaning and preprocessing procedures were applied”. More coverage is usually given to automatic text annotation such as lemmatisation, part-of-speech tagging and parsing, which is often included in preprocessing. In the literature, the term ‘preprocessing’ is used to refer to a wide range of procedures, from filtering and cleaning to data transformation such as stemming and numeric representation, which might create confusion. We argue that text preprocessing might skew original data distribution with regard to the metadata, such as types, locations and times of registered datapoints. In this paper we describe a systematic approach to cleaning text data mined by a data-providing company for a Digital Humanities (DH) project focused on cultural analytics. We reveal the types and amount of noise in the data coming from various web sources and estimate the changes in the size of the data associated with preprocessing. We also compare the results of a text classification experiment run on the raw and preprocessed data. We hope that our experience and approaches will help the DH community to diagnose the quality of textual data collected from the web and prepare it for further natural language processing. | ['Alistair Plum', 'Maria Kunilovskaya'] | null | null | null | null | ranlp-2021-9 | ['text-annotation'] | ['natural-language-processing'] | [ 2.05627978e-01 8.12471099e-03 1.43720523e-01 -4.75663006e-01
-7.99193084e-01 -9.87555385e-01 6.75437033e-01 9.54559505e-01
-7.71351695e-01 5.33802330e-01 8.22997630e-01 -3.48700613e-01
-2.98826724e-01 -7.04037666e-01 -5.42806506e-01 -5.93339741e-01
3.10299754e-01 3.41556579e-01 -1.18218340e-01 -1.88350558e-01
5.94562769e-01 3.32487106e-01 -1.81648314e+00 4.51616704e-01
6.60278082e-01 4.91650313e-01 1.66177243e-01 4.03147846e-01
-9.82272267e-01 5.31470597e-01 -8.42504680e-01 -6.82909012e-01
2.31900349e-01 -1.35264188e-01 -1.13351119e+00 1.60095021e-01
1.45218879e-01 2.29067966e-01 2.17999548e-01 1.17995977e+00
3.81221771e-01 -2.71116972e-01 3.00149024e-01 -9.26840603e-01
-3.86320174e-01 9.07407641e-01 -3.96335274e-01 1.85645521e-01
6.65302396e-01 -2.85325885e-01 8.58480155e-01 -6.20035827e-01
9.66539383e-01 8.87888372e-01 8.41719091e-01 1.87907331e-02
-1.02038527e+00 -3.16252708e-01 -9.64563265e-02 -2.00546592e-01
-1.19597137e+00 -5.02600670e-01 3.69197279e-01 -6.28647029e-01
8.53394926e-01 5.68688989e-01 4.89291966e-01 1.10547876e+00
-9.04644132e-02 6.13236666e-01 1.15240872e+00 -9.92917657e-01
1.22483328e-01 4.62642223e-01 3.12255502e-01 1.79926604e-01
6.14123464e-01 -6.05558217e-01 -5.13531804e-01 -3.58549803e-01
1.31757095e-01 -8.63531828e-02 3.55523191e-02 -2.17979968e-01
-1.07056999e+00 7.31957734e-01 -2.85080612e-01 7.04645932e-01
-4.24413502e-01 -5.35269976e-01 9.04017568e-01 3.10173273e-01
6.55374587e-01 5.80028474e-01 -9.02121186e-01 -6.00253642e-01
-8.90814900e-01 2.48905301e-01 1.15239942e+00 1.07990837e+00
4.62121218e-01 -5.45972407e-01 2.54177064e-01 1.08236337e+00
3.30160111e-01 3.19258928e-01 5.82764149e-01 -7.36591876e-01
6.92319751e-01 9.28475142e-01 1.53499484e-01 -1.01650274e+00
-4.36756700e-01 2.84866601e-01 -3.63835871e-01 -8.90649110e-02
5.97777545e-01 -2.85021663e-01 -9.59704161e-01 1.11454988e+00
2.78093934e-01 -1.05953252e+00 -1.17562853e-01 4.19747233e-01
8.01182091e-01 3.31851035e-01 3.17132831e-01 -3.27628136e-01
1.64202392e+00 -9.65961367e-02 -1.17177641e+00 -5.07098176e-02
8.84928167e-01 -1.29686046e+00 1.15342569e+00 6.26134455e-01
-9.01135683e-01 4.13300097e-02 -5.76378405e-01 -2.50077605e-01
-1.16651392e+00 -3.98472190e-01 6.55079663e-01 8.29774201e-01
-6.26724780e-01 5.73181927e-01 -6.68691039e-01 -1.25908196e+00
2.45699897e-01 -1.24170939e-02 -4.93652135e-01 1.77135766e-02
-9.69280303e-01 8.16595793e-01 3.24397683e-01 -7.46777579e-02
2.40664110e-01 -5.07746935e-01 -6.73379481e-01 -2.50394017e-01
7.18332827e-01 -7.81394914e-02 1.09297585e+00 -8.65813076e-01
-8.06464314e-01 1.30271125e+00 -2.06169426e-01 -1.63457558e-01
4.17102635e-01 -2.80569792e-01 -7.19800234e-01 -3.03793460e-01
4.08755332e-01 -5.60185015e-02 4.13346291e-01 -1.27850711e+00
-1.01647687e+00 -5.36368549e-01 -4.58167791e-01 -1.99013963e-01
-2.32562393e-01 7.32960820e-01 -3.55464011e-01 -9.98915732e-01
1.40960321e-01 -7.19220579e-01 2.79435702e-02 -6.56741142e-01
-2.03022242e-01 -1.35721743e-01 5.34593403e-01 -1.32464635e+00
1.83004856e+00 -2.11794853e+00 -2.36001402e-01 3.74222457e-01
-1.26921788e-01 1.00584961e-01 2.37241000e-01 1.11463380e+00
-2.28134260e-01 8.23629141e-01 -2.25645766e-01 -7.50774369e-02
-1.97140817e-02 4.89591032e-01 -4.43259180e-01 4.55907792e-01
-1.48841515e-01 4.89256263e-01 -8.40557992e-01 -6.13501549e-01
-4.41585742e-02 4.66916472e-01 -2.05143899e-01 -3.04065138e-01
9.04911160e-02 -4.78224754e-02 -2.02563301e-01 8.73701513e-01
4.96142775e-01 4.13937539e-01 1.95850000e-01 -1.64454460e-01
-6.78347826e-01 7.48338401e-01 -1.29703689e+00 1.63774240e+00
3.31497379e-02 8.47131193e-01 3.76588553e-01 -6.13289595e-01
7.29507804e-01 3.27618480e-01 6.60742760e-01 -5.84542155e-01
1.83064654e-01 2.80822039e-01 -2.01110885e-01 -9.11964417e-01
9.77905869e-01 2.79736280e-01 -3.59320045e-01 5.21904707e-01
-8.80935267e-02 -1.53996825e-01 4.36264902e-01 8.87933597e-02
1.01902080e+00 3.01399291e-01 4.01975662e-01 -3.56141955e-01
3.85005683e-01 6.66237235e-01 3.76537323e-01 4.63971496e-01
-8.11068490e-02 5.65451264e-01 5.97053051e-01 -4.63634849e-01
-1.53245997e+00 -2.65317619e-01 -4.57008839e-01 1.07812440e+00
-6.75028920e-01 -9.23434675e-01 -9.09394562e-01 -7.17153549e-01
-1.33629724e-01 8.55323255e-01 -7.05033123e-01 3.00572693e-01
-5.24917483e-01 -9.31769669e-01 4.71139103e-01 -7.42762908e-02
-4.81338613e-02 -1.12409234e+00 -7.30178297e-01 2.61281461e-01
-4.28572118e-01 -7.97190487e-01 -1.21357128e-01 6.36315048e-01
-5.96196949e-01 -1.13949847e+00 -8.69236514e-02 -6.31532073e-01
6.54561579e-01 7.96417966e-02 1.02900231e+00 2.30403677e-01
-3.02171379e-01 3.31262469e-01 -1.01747692e+00 -8.11573505e-01
-6.37257278e-01 3.17735195e-01 -2.27712914e-01 -3.88078004e-01
1.04522097e+00 -2.44897455e-01 -2.22286433e-02 1.24292597e-02
-1.50718629e+00 -4.35668230e-01 4.83625531e-01 4.43027139e-01
3.62943947e-01 3.17336529e-01 -5.47186425e-03 -1.40191996e+00
8.93285036e-01 -5.32965004e-01 -1.72416314e-01 2.26722822e-01
-7.92743921e-01 -5.36039285e-02 2.93356389e-01 -2.00787261e-01
-8.93310726e-01 -2.27070022e-02 -9.79720280e-02 4.22142863e-01
-6.66814506e-01 7.62742043e-01 -2.62744159e-01 2.40709558e-01
6.10038161e-01 -9.78459418e-02 2.39080247e-02 -9.02586043e-01
1.54070869e-01 1.28590977e+00 4.45592403e-01 -4.71694082e-01
6.97789848e-01 4.98839200e-01 -5.65191746e-01 -9.15821016e-01
-6.88486099e-01 -7.14376390e-01 -9.23915446e-01 2.38532834e-02
5.72441995e-01 -6.69033170e-01 -2.10908383e-01 4.19969201e-01
-8.12559128e-01 -1.87774897e-02 -5.24755776e-01 2.14792937e-01
-1.54022828e-01 5.08615017e-01 -3.11713815e-01 -8.88583302e-01
-2.56273866e-01 -5.66873550e-01 7.78701246e-01 -2.00407803e-01
-9.07771587e-01 -8.65118861e-01 2.01989504e-04 6.41567230e-01
2.17639238e-01 5.16496956e-01 1.01535225e+00 -1.02403021e+00
1.92018121e-01 -5.06663918e-01 1.25900224e-01 1.66337714e-01
4.58245575e-01 5.86641073e-01 -7.99226522e-01 7.87213296e-02
1.40930504e-01 1.22865491e-01 3.14850509e-01 1.35671705e-01
9.44455564e-01 -7.08394468e-01 -1.66328892e-01 2.30290547e-01
1.48311234e+00 2.41848469e-01 7.35973001e-01 1.00446534e+00
6.09703302e-01 1.18172669e+00 7.36754417e-01 5.07637382e-01
2.93879896e-01 3.95603895e-01 1.81155458e-01 1.89458802e-01
1.42624214e-01 -2.08924860e-01 2.89225549e-01 8.94453764e-01
3.46268341e-02 -1.04110785e-01 -1.15207720e+00 8.32952678e-01
-1.62574124e+00 -9.63499963e-01 -7.18001842e-01 2.25431108e+00
9.69237864e-01 1.59441113e-01 2.19888240e-01 4.84735399e-01
6.14243090e-01 -2.01268584e-01 1.81665465e-01 -7.16652751e-01
-2.56688476e-01 1.32789150e-01 1.00329220e+00 3.51240933e-01
-9.83938515e-01 6.70458734e-01 6.63053656e+00 6.25615835e-01
-7.46375561e-01 -8.31339061e-02 1.86661705e-01 -6.99989274e-02
-2.29643539e-01 3.65004651e-02 -8.26470137e-01 7.80844152e-01
1.04603899e+00 1.68603938e-02 4.91177142e-01 6.34272397e-01
4.68021899e-01 -5.41542470e-01 -8.27042937e-01 6.54080093e-01
2.03042522e-01 -9.38697875e-01 1.94340516e-02 3.69874120e-01
3.63784313e-01 8.28414038e-03 -5.63488066e-01 8.09638798e-02
4.11317348e-01 -8.17792177e-01 1.04507840e+00 4.06311870e-01
3.78702849e-01 -7.24174500e-01 1.01968646e+00 1.92298159e-01
-7.08282888e-01 -3.81049328e-02 -1.31096497e-01 -1.32292628e-01
2.64786221e-02 7.56756186e-01 -6.55695736e-01 5.04126966e-01
1.32704246e+00 3.46080393e-01 -9.55761015e-01 1.10653615e+00
6.25003725e-02 4.96685386e-01 -4.11727548e-01 -1.28518716e-01
-9.64752957e-03 -3.94333243e-01 4.61682975e-01 1.46074915e+00
3.27231526e-01 -7.22911432e-02 -3.01106334e-01 2.16565937e-01
1.39702633e-01 5.71394622e-01 -4.35771912e-01 -3.48454207e-01
5.44447899e-01 1.37910295e+00 -1.01428711e+00 -2.22057760e-01
-5.54454982e-01 6.45916760e-01 5.05233034e-02 1.21968217e-01
-3.66448939e-01 -6.75924897e-01 5.95854521e-01 5.96931636e-01
4.01414156e-01 -1.59535468e-01 -7.72554219e-01 -9.87694919e-01
2.12254167e-01 -1.18356872e+00 6.54011607e-01 -7.58548200e-01
-1.20862246e+00 3.56764436e-01 2.12906659e-01 -8.22660089e-01
-8.58886689e-02 -4.16100651e-01 4.46288362e-02 8.68947983e-01
-9.28533792e-01 -7.45174646e-01 -3.20913941e-01 3.24058831e-01
5.05513608e-01 1.68749318e-01 8.60173702e-01 4.45463866e-01
-4.97696787e-01 9.90726948e-02 2.50151247e-01 5.13152957e-01
1.12590921e+00 -1.25981998e+00 4.22088087e-01 9.07869399e-01
7.47997239e-02 9.13669348e-01 1.06903267e+00 -9.97628689e-01
-1.21634948e+00 -7.38108337e-01 1.57788289e+00 -8.19924176e-01
1.00499249e+00 -4.44854736e-01 -1.14047778e+00 6.64658487e-01
6.45567119e-01 -8.26106191e-01 1.04523182e+00 1.81854442e-01
-1.77489236e-01 9.62196887e-02 -1.26011050e+00 5.21226048e-01
8.08333874e-01 -3.24449748e-01 -1.10665917e+00 3.52576077e-01
3.50545645e-01 -1.76038116e-01 -1.02693009e+00 -6.82531446e-02
5.01156807e-01 -5.84517479e-01 4.76314664e-01 -5.81197917e-01
1.97966293e-01 -1.95458874e-01 -4.68133949e-02 -1.11469185e+00
-1.40536860e-01 -7.09026098e-01 4.10941482e-01 2.04014111e+00
4.25326467e-01 -4.88730073e-01 4.51615870e-01 1.01153934e+00
8.80272761e-02 4.31601703e-02 -7.20496774e-01 -2.11474225e-01
-1.03389136e-02 -7.40717471e-01 6.27295971e-01 1.39103973e+00
2.22448528e-01 1.02000974e-01 -7.85710216e-02 -2.05404878e-01
3.12506944e-01 -4.45541233e-01 7.56621063e-01 -1.58963883e+00
5.27780652e-01 -4.22342807e-01 -3.89709626e-03 8.48660991e-03
-3.81297112e-01 -5.77943325e-01 -8.42809305e-02 -1.72681653e+00
-3.24694812e-02 -3.30476522e-01 3.71080041e-01 4.44695532e-01
1.25067592e-01 -1.53178066e-01 1.25230383e-02 4.97708708e-01
-8.20541754e-02 -3.52519333e-01 5.30969143e-01 2.39396870e-01
-4.11810309e-01 -3.07475299e-01 -1.09156120e+00 8.35466385e-01
7.56365538e-01 -9.37296450e-01 1.28021166e-01 -4.45596933e-01
8.17683578e-01 -8.51028800e-01 1.22823015e-01 -6.72939122e-01
9.19909626e-02 -2.56287754e-01 3.77384931e-01 -6.88726842e-01
-3.30508649e-01 -1.35783696e+00 5.67005277e-01 1.03305340e-01
-2.94971734e-01 1.92441687e-01 1.63115352e-01 1.15423337e-01
-1.14443734e-01 -6.48093820e-01 2.83366203e-01 -4.91548270e-01
-5.24993479e-01 -3.13937962e-01 -7.05163181e-01 5.40137887e-02
8.80609393e-01 -4.96183455e-01 -2.88767844e-01 -1.89804569e-01
-6.15735590e-01 3.88708338e-02 9.61808443e-01 5.06313145e-01
-2.26247877e-01 -1.00995159e+00 -5.28610826e-01 1.31824851e-01
2.33296230e-01 -1.84890314e-03 -1.80581957e-01 7.44169533e-01
-7.67993569e-01 3.82579684e-01 -8.40279683e-02 -7.34375045e-02
-1.41567361e+00 7.56053507e-01 -2.20960885e-01 -2.97677368e-02
-6.72588527e-01 1.28260404e-01 -7.77854621e-01 -4.14073795e-01
4.00138766e-01 -6.21110320e-01 -4.95355934e-01 8.94210279e-01
4.73183572e-01 4.58443075e-01 5.38270652e-01 -7.76909709e-01
-2.96538085e-01 3.30423713e-01 1.71638690e-02 -1.71431661e-01
1.76949060e+00 -6.95417285e-01 -5.28712869e-01 8.31462681e-01
9.27871943e-01 6.03164613e-01 -5.01934946e-01 -1.45169556e-01
6.62446976e-01 -4.86089438e-01 -2.20367640e-01 -8.92078876e-01
-6.77456915e-01 2.62315124e-01 2.68332511e-01 9.46927965e-01
1.03310299e+00 4.28757221e-02 4.39406514e-01 2.44115576e-01
6.87659681e-02 -1.82157719e+00 -8.60718727e-01 4.64446455e-01
8.57354045e-01 -1.15501416e+00 4.55342829e-01 -5.19264817e-01
-6.08888268e-01 1.18711853e+00 -4.00324073e-03 3.38022441e-01
7.60375261e-01 6.43564343e-01 3.09893638e-01 -3.62140805e-01
-4.56613153e-01 -3.02357107e-01 -1.41078785e-01 6.57569945e-01
7.34332979e-01 -2.69775957e-01 -9.92562532e-01 5.23593962e-01
-6.83572948e-01 4.25143987e-02 8.90955269e-01 1.26877475e+00
-3.39234233e-01 -1.44267964e+00 -8.10655117e-01 6.40262902e-01
-1.12964630e+00 -2.07518995e-01 -9.34849918e-01 1.14702642e+00
3.49924475e-01 1.03473616e+00 1.50325090e-01 -4.67856266e-02
5.49279630e-01 5.33501565e-01 5.15788794e-02 -5.66631079e-01
-1.17714417e+00 3.13078046e-01 5.13838351e-01 -2.83679932e-01
-7.62598813e-01 -1.40410638e+00 -1.01055539e+00 -4.99170065e-01
-5.27328067e-02 4.19295341e-01 1.19572151e+00 8.16042542e-01
2.84135640e-01 2.39475444e-01 5.73894149e-03 -3.80623877e-01
-9.78776515e-02 -1.08042276e+00 -4.97877836e-01 7.40697384e-01
1.46390289e-01 -1.47689730e-01 -5.35200596e-01 5.88470995e-01] | [10.017542839050293, 9.840523719787598] |
aa7ab65a-8873-417f-b268-aaebe32d4926 | a-hybrid-deep-learning-framework-for-covid-19 | 2107.03904 | null | https://arxiv.org/abs/2107.03904v2 | https://arxiv.org/pdf/2107.03904v2.pdf | A hybrid deep learning framework for Covid-19 detection via 3D Chest CT Images | In this paper, we present a hybrid deep learning framework named CTNet which combines convolutional neural network and transformer together for the detection of COVID-19 via 3D chest CT images. It consists of a CNN feature extractor module with SE attention to extract sufficient features from CT scans, together with a transformer model to model the discriminative features of the 3D CT scans. Compared to previous works, CTNet provides an effective and efficient method to perform COVID-19 diagnosis via 3D CT scans with data resampling strategy. Advanced results on a large and public benchmarks, COV19-CT-DB database was achieved by the proposed CTNet, over the state-of-the-art baseline approachproposed together with the dataset. | ['Shuang Liang'] | 2021-07-08 | null | null | null | null | ['covid-19-detection'] | ['medical'] | [-1.03637375e-01 -3.72906984e-03 -1.16965033e-01 -2.74342060e-01
-1.10885942e+00 -1.35207534e-01 2.92582452e-01 -8.58241245e-02
-6.04661226e-01 3.70132238e-01 2.26137146e-01 -4.02185410e-01
-1.25710219e-01 -8.59433591e-01 -5.79532921e-01 -6.48522615e-01
-3.08439136e-01 9.31254983e-01 4.03895736e-01 -5.30880876e-03
-4.19638962e-01 8.38499665e-01 -1.01605773e+00 4.69770342e-01
1.26944169e-01 1.45908415e+00 1.95119902e-01 8.22200239e-01
2.26205245e-01 7.79723287e-01 -3.23143065e-01 -1.02731630e-01
1.74061194e-01 -7.72165135e-02 -7.57556200e-01 -2.89616082e-02
3.14597070e-01 -9.08717811e-01 -5.65746129e-01 3.86578202e-01
8.35994363e-01 -3.66894275e-01 9.43941355e-01 -8.99922132e-01
-7.91064501e-02 2.98751205e-01 -6.34986699e-01 8.33495200e-01
1.51196390e-01 2.79426843e-01 8.36306095e-01 -1.16849375e+00
7.79720724e-01 9.01307881e-01 1.11338067e+00 3.18233162e-01
-5.70468605e-01 -8.94097269e-01 -5.78432381e-01 3.17705333e-01
-1.23944545e+00 4.60758597e-01 3.82159591e-01 -3.66250753e-01
1.15357161e+00 3.15150768e-01 1.05919433e+00 1.48207331e+00
7.68127561e-01 9.11718249e-01 8.31914961e-01 6.78437799e-02
-2.28277430e-01 -2.54457861e-01 2.54122168e-01 1.11605811e+00
1.52293518e-01 3.87212902e-01 -1.70103073e-01 -3.30087245e-01
7.94512093e-01 4.42026734e-01 -5.61063066e-02 -3.90825838e-01
-9.68146384e-01 8.62485409e-01 8.23690236e-01 3.06673795e-01
-6.38207734e-01 1.40814662e-01 9.16195452e-01 1.26427328e-02
5.20350158e-01 -7.51507804e-02 -3.89226556e-01 4.18829113e-01
-7.35896170e-01 3.59011650e-01 2.63337374e-01 8.01640570e-01
1.96693316e-02 -1.19950548e-01 -4.78839487e-01 6.27893448e-01
3.57500672e-01 5.13970315e-01 8.69460762e-01 -9.53712240e-02
4.85465080e-01 5.31977892e-01 -5.08210361e-01 -4.47454304e-01
-9.47273731e-01 -7.41902471e-01 -1.15350306e+00 -8.79822299e-02
-1.32863134e-01 -1.36898428e-01 -1.52377856e+00 1.17328143e+00
4.46965098e-01 2.90848076e-01 -2.88042217e-01 8.71834576e-01
1.50403225e+00 9.27147344e-02 3.40008847e-02 1.99305236e-01
1.71159792e+00 -8.08949411e-01 -3.21161062e-01 3.81474167e-01
4.71353531e-01 -6.74951315e-01 6.48896992e-01 4.74874318e-01
-1.00281048e+00 -3.85673165e-01 -1.17034698e+00 -5.93911000e-02
-3.55534941e-01 2.83372343e-01 4.40394193e-01 5.95320761e-01
-1.01478660e+00 4.29273933e-01 -1.15476298e+00 -3.18860501e-01
8.67888808e-01 4.77932811e-01 -4.02211249e-01 -1.48530617e-01
-1.18955708e+00 1.03190029e+00 3.87057304e-01 -5.50200045e-02
-1.75897670e+00 -9.96280730e-01 -7.13301897e-01 -6.08947724e-02
3.96739870e-01 -1.12005734e+00 1.45143735e+00 -2.16491953e-01
-9.49083984e-01 1.14589715e+00 3.71592462e-01 -6.37043715e-01
8.96753788e-01 -2.61546612e-01 -1.91108704e-01 4.58596140e-01
1.52111620e-01 3.22821438e-01 9.59719956e-01 -7.42297769e-01
-6.62155330e-01 -3.97025734e-01 -3.03575307e-01 -9.46751088e-02
1.42510876e-01 2.15290919e-01 -5.99855185e-01 -8.43415916e-01
1.13250345e-01 -8.05467784e-01 -5.62084734e-01 1.06022514e-01
-6.85238957e-01 -3.94457906e-01 1.09599054e+00 -6.69631064e-01
6.10501468e-01 -2.00819182e+00 -2.09927514e-01 2.73488492e-01
1.00722826e+00 3.13088030e-01 2.86632150e-01 -2.67046504e-02
-5.92149258e-01 4.33021039e-02 -1.64104909e-01 -3.43983024e-01
-3.26032788e-01 9.10681784e-02 1.47201464e-01 7.26218700e-01
3.77920866e-01 1.06264448e+00 -6.62756205e-01 -9.07815099e-01
4.71645892e-01 5.04389286e-01 -4.46430296e-01 4.37631786e-01
2.05504701e-01 2.80220866e-01 -6.51257694e-01 9.69028771e-01
8.96360874e-01 -5.79560339e-01 -1.75717443e-01 -2.43537188e-01
3.73300016e-01 5.99698052e-02 -6.63074255e-01 1.50074351e+00
-4.22507018e-01 2.79459953e-01 7.34511539e-02 -9.26366508e-01
4.37254012e-01 6.67743802e-01 8.13099444e-01 -4.33649868e-01
6.99046016e-01 2.52075672e-01 -1.13315992e-01 -8.79424453e-01
1.51410744e-01 -3.56099486e-01 1.26722117e-03 5.08238018e-01
3.42868358e-01 -2.77602881e-01 -1.62852392e-01 8.97606239e-02
1.39697111e+00 -2.96612293e-01 5.11568487e-01 -5.65143414e-02
5.17437279e-01 1.01789542e-01 2.92673498e-01 7.13126242e-01
-3.18390518e-01 8.36507201e-01 1.82598189e-01 -7.67491639e-01
-1.02301741e+00 -1.37202454e+00 -6.26782894e-01 4.94567066e-01
-2.42653176e-01 -2.85426408e-01 -4.30979222e-01 -1.25535452e+00
-3.91531996e-02 8.23976174e-02 -1.06334257e+00 4.91579026e-02
-8.76785278e-01 -9.35543776e-01 8.33715975e-01 1.06908691e+00
6.91087961e-01 -8.56401265e-01 -8.83004427e-01 1.90000921e-01
-1.61031559e-01 -1.09719360e+00 -2.09188074e-01 5.49552858e-01
-9.07256663e-01 -1.51007533e+00 -8.34175050e-01 -6.82505667e-01
3.89159173e-01 6.01288378e-02 1.23948419e+00 2.75614411e-01
-1.03527427e+00 3.85940485e-02 -3.18704963e-01 -8.00059974e-01
-2.97125578e-01 2.71736205e-01 -3.32835048e-01 -4.87519652e-01
2.38503009e-01 -4.74724472e-01 -7.11942017e-01 1.32805839e-01
-8.56368124e-01 3.38676050e-02 8.85346293e-01 1.18569171e+00
8.68587494e-01 -1.53865308e-01 1.58632964e-01 -9.18028891e-01
2.41389662e-01 -8.40597510e-01 -1.89181343e-01 3.00085247e-02
-3.07650268e-01 -3.54392231e-01 5.69792807e-01 1.15994923e-01
-7.70192981e-01 6.54115379e-02 -7.30932117e-01 -7.79838204e-01
-2.05509216e-01 3.18215996e-01 3.53805006e-01 1.70833454e-01
7.10713208e-01 2.82676041e-01 -1.18898816e-01 -5.13252079e-01
-1.57705113e-01 6.79066718e-01 4.79056448e-01 -1.35075077e-01
6.50828838e-01 7.46468127e-01 4.23515975e-01 -3.95016253e-01
-8.84335518e-01 -6.69147134e-01 -8.46352637e-01 -5.74065857e-02
1.22339189e+00 -1.06292176e+00 -3.42632204e-01 5.35452425e-01
-9.08055842e-01 2.36573860e-01 -3.37032825e-01 6.81552887e-01
-5.12242973e-01 1.51872769e-01 -9.72430468e-01 -1.46601021e-01
-1.02595675e+00 -1.58208728e+00 1.40475869e+00 -1.66679367e-01
4.63963341e-04 -7.25989521e-01 1.99186534e-01 2.52272934e-01
5.39768159e-01 6.84262395e-01 9.97591972e-01 -1.00192010e+00
-4.93866026e-01 -4.32889432e-01 -4.45375264e-01 3.29239249e-01
-1.12195186e-01 -1.54508576e-01 -1.07873869e+00 -2.83525944e-01
7.76663646e-02 -6.43959343e-01 1.07954443e+00 4.96872038e-01
1.68381703e+00 2.41736144e-01 -6.57490075e-01 1.21073508e+00
1.60829878e+00 7.35817775e-02 2.23513275e-01 1.06174506e-01
6.95794404e-01 -2.41398245e-01 3.76236618e-01 7.28438795e-01
2.11690471e-01 3.06997359e-01 9.05611575e-01 -5.91955781e-01
-2.22235665e-01 4.57253344e-02 -4.25346196e-01 5.90254605e-01
-2.85468131e-01 -1.21763706e-01 -1.21896541e+00 5.99430442e-01
-1.12344038e+00 -4.45852906e-01 -2.58123040e-01 1.34973502e+00
4.01823819e-01 1.83275953e-01 1.16785318e-01 2.27509156e-01
5.36552668e-01 6.10448718e-02 -5.43829858e-01 -8.73590186e-02
2.43147761e-01 1.00413728e+00 5.61388314e-01 -3.40275168e-01
-1.34050930e+00 2.78325945e-01 7.35537291e+00 9.50848043e-01
-1.39925480e+00 6.72950387e-01 6.40848994e-01 -1.69720277e-01
2.01824442e-01 -8.33428502e-01 -7.12785363e-01 1.20964214e-01
7.38902271e-01 1.64879069e-01 -4.75276351e-01 1.08009863e+00
-3.45626473e-02 1.57385647e-01 -9.48986053e-01 8.53768885e-01
1.16066672e-01 -1.41135299e+00 -6.71074688e-02 3.42935584e-02
2.57652313e-01 7.66472220e-01 1.36806265e-01 4.40714478e-01
4.51006681e-01 -1.20298183e+00 4.32018012e-01 8.74830782e-02
1.21983349e+00 -7.93882489e-01 1.29232728e+00 6.57448769e-02
-1.37709534e+00 -3.61706130e-02 -1.72057614e-01 6.46805346e-01
-3.27242091e-02 4.58793283e-01 -1.66536880e+00 9.94241178e-01
1.20327163e+00 7.80835629e-01 -6.54589772e-01 1.17006528e+00
-1.98777676e-01 7.64972448e-01 -3.32335174e-01 8.91424641e-02
3.48826736e-01 6.39549077e-01 5.06437778e-01 1.42206383e+00
3.89768928e-01 1.57679543e-01 1.60938665e-01 7.26547778e-01
-1.23367384e-01 1.80546954e-01 -5.36952972e-01 4.74252850e-01
9.57644954e-02 1.43896437e+00 -6.68540418e-01 -6.69896960e-01
-2.37421125e-01 5.15249014e-01 1.81306712e-02 -3.08697343e-01
-1.21984148e+00 -1.70201734e-01 1.27004385e-01 1.44400626e-01
6.31242752e-01 2.69814849e-01 -1.47385567e-01 -9.65564191e-01
-3.37851584e-01 -7.14084208e-01 1.15167785e+00 -8.97718608e-01
-1.38249445e+00 9.73402321e-01 2.62658626e-01 -1.49614787e+00
-3.58473092e-01 -8.97442520e-01 -7.33992457e-01 6.65375531e-01
-1.64993465e+00 -1.39097095e+00 -6.33529723e-01 1.03193569e+00
6.45612776e-01 -1.10871285e-01 8.62330377e-01 2.47831047e-01
-4.69797224e-01 7.20314026e-01 -5.42205460e-02 3.56470615e-01
5.70794940e-01 -1.39394355e+00 4.81204599e-01 3.56161118e-01
-5.67811310e-01 1.88333526e-01 1.40032858e-01 -4.73419398e-01
-1.20836508e+00 -1.40848136e+00 4.20892835e-01 -4.10860002e-01
4.46193010e-01 -2.10714430e-01 -4.93271589e-01 7.72964299e-01
3.13642621e-01 5.66177249e-01 7.37915933e-01 -4.26888078e-01
-2.55505055e-01 2.23534808e-01 -1.62070191e+00 -9.44176093e-02
9.51025665e-01 -2.17528075e-01 -9.15710032e-01 4.92034405e-01
7.80248940e-01 -1.01359892e+00 -1.24610198e+00 8.70856285e-01
4.61231112e-01 -7.93557167e-01 1.33528161e+00 -5.85494161e-01
6.31291509e-01 1.69038966e-01 -6.20920248e-02 -1.18060660e+00
-3.94478291e-01 1.16604246e-01 2.06458420e-02 3.48645121e-01
1.64433062e-01 -3.17638636e-01 8.88196886e-01 -2.71105558e-01
-4.70900863e-01 -1.21253669e+00 -1.12525535e+00 -5.22345245e-01
3.34241539e-01 -6.62515461e-01 8.39744329e-01 7.14903355e-01
-5.76467931e-01 2.42368616e-02 9.95237567e-03 -8.56656488e-03
5.47757328e-01 -1.86707541e-01 4.81403232e-01 -9.98594642e-01
-4.75687742e-01 -2.70883083e-01 -6.38315201e-01 -5.94788730e-01
-2.56914496e-01 -1.01580846e+00 -2.10331753e-01 -1.42087245e+00
7.24799931e-01 -1.49243653e-01 -5.67772150e-01 3.25181425e-01
1.70720574e-02 6.27668202e-01 2.00973395e-02 -3.34415026e-03
-3.30518425e-01 2.91077703e-01 1.61526144e+00 -2.40885228e-01
3.43711406e-01 1.39964089e-01 -8.34119692e-02 7.13778496e-01
5.59614778e-01 -6.72745645e-01 -1.99421614e-01 -5.97532392e-02
-2.24214926e-01 4.55308437e-01 5.58211267e-01 -1.00921929e+00
-7.14126527e-02 2.46255919e-01 9.52378869e-01 -1.78814960e+00
3.46902728e-01 -8.63523841e-01 -2.95148864e-02 9.41025913e-01
-4.79808450e-02 3.68142277e-01 1.71831265e-01 3.75921965e-01
-2.47897461e-01 -9.90640670e-02 9.29583848e-01 -6.05047464e-01
-3.35655212e-01 8.37177575e-01 -7.85184920e-01 1.46401152e-01
1.20665085e+00 1.41130900e-02 -1.44129038e-01 1.04426727e-01
-9.28632259e-01 1.67062521e-01 -2.12667152e-01 3.32106948e-02
9.80608046e-01 -1.37188840e+00 -9.60063159e-01 4.47450787e-01
2.24636823e-01 5.45404613e-01 4.16496545e-01 1.17648208e+00
-8.48749638e-01 5.08016288e-01 -3.64011616e-01 -1.17924118e+00
-1.44060588e+00 5.67195892e-01 9.34334338e-01 -9.99402225e-01
-9.61115181e-01 1.03311014e+00 3.26470554e-01 -4.83696282e-01
2.22056344e-01 -8.48338187e-01 -2.53151655e-01 -6.14318922e-02
4.62987185e-01 8.91851038e-02 6.41629100e-01 -4.77916449e-01
-6.54854596e-01 4.86711025e-01 -4.30692732e-01 2.22618386e-01
1.75275743e+00 2.42390826e-01 2.05746889e-01 -9.48289484e-02
1.47559321e+00 -5.58289111e-01 -6.42549574e-01 -4.07263547e-01
-4.54062343e-01 -2.78682172e-01 2.11023092e-01 -7.93607295e-01
-1.48226285e+00 8.54949594e-01 1.15047836e+00 -2.18009278e-01
1.15944231e+00 2.01049522e-01 1.10990894e+00 1.25197411e-01
9.95441377e-02 -6.59904659e-01 4.29823279e-01 4.63393301e-01
7.62481511e-01 -1.42579103e+00 1.22981399e-01 -4.03920621e-01
-3.98352563e-01 1.18507993e+00 6.74432874e-01 -5.97532153e-01
1.27398658e+00 4.39813972e-01 -1.88378594e-03 -9.25823212e-01
-8.30151796e-01 -1.51938960e-01 2.77304858e-01 5.42644143e-01
3.47496480e-01 2.54102826e-01 3.78995351e-02 8.11597466e-01
-3.39067966e-01 2.23763898e-01 2.25157693e-01 1.00372255e+00
-1.77292421e-01 -6.85938835e-01 -4.25089061e-01 8.96353781e-01
-1.01915991e+00 -4.00713906e-02 -2.82694846e-01 1.50836003e+00
4.53677028e-01 2.70480096e-01 4.94636223e-02 -6.21224165e-01
5.13773441e-01 -1.98204607e-01 3.71887773e-01 -6.41194403e-01
-1.18006492e+00 2.72982776e-01 2.55865715e-02 -6.69955552e-01
-2.35846013e-01 -7.07152843e-01 -1.15659988e+00 2.02737749e-02
-5.27998507e-01 -4.37499285e-01 4.89275426e-01 9.51780915e-01
-1.39352515e-01 1.00455534e+00 7.53062308e-01 -5.71809173e-01
-7.59436965e-01 -1.10826659e+00 -6.06359184e-01 3.11764389e-01
6.71106696e-01 -8.51718009e-01 -6.19356297e-02 -3.47093403e-01] | [15.263042449951172, -1.9276984930038452] |
4b710391-84ba-4f62-ab11-4a9b79d45466 | punctuation-restoration-in-swedish-through | 2202.06769 | null | https://arxiv.org/abs/2202.06769v1 | https://arxiv.org/pdf/2202.06769v1.pdf | Punctuation restoration in Swedish through fine-tuned KB-BERT | Presented here is a method for automatic punctuation restoration in Swedish using a BERT model. The method is based on KB-BERT, a publicly available, neural network language model pre-trained on a Swedish corpus by National Library of Sweden. This model has then been fine-tuned for this specific task using a corpus of government texts. With a lower-case and unpunctuated Swedish text as input, the model is supposed to return a grammatically correct punctuated copy of the text as output. A successful solution to this problem brings benefits for an array of NLP domains, such as speech-to-text and automated text. Only the punctuation marks period, comma and question marks were considered for the project, due to a lack of data for more rare marks such as semicolon. Additionally, some marks are somewhat interchangeable with the more common, such as exclamation points and periods. Thus, the data set had all exclamation points replaced with periods. The fine-tuned Swedish BERT model, dubbed prestoBERT, achieved an overall F1-score of 78.9. The proposed model scored similarly to international counterparts, with Hungarian and Chinese models obtaining F1-scores of 82.2 and 75.6 respectively. As further comparison, a human evaluation case study was carried out. The human test group achieved an overall F1-score of 81.7, but scored substantially worse than prestoBERT on both period and comma. Inspecting output sentences from the model and humans show satisfactory results, despite the difference in F1-score. The disconnect seems to stem from an unnecessary focus on replicating the exact same punctuation used in the test set, rather than providing any of the number of correct interpretations. If the loss function could be rewritten to reward all grammatically correct outputs, rather than only the one original example, the performance could improve significantly for both prestoBERT and the human group. | ['John Björkman Nilsson'] | 2022-02-14 | null | null | null | null | ['punctuation-restoration'] | ['natural-language-processing'] | [ 1.38311654e-01 5.48160851e-01 2.85932600e-01 -4.03091311e-01
-8.81921053e-01 -6.35964334e-01 4.72672433e-01 2.77184993e-01
-8.06539893e-01 1.05074298e+00 2.16556266e-01 -6.47674263e-01
-1.47809401e-01 -5.32135308e-01 -6.25537932e-01 -5.09236693e-01
2.97951221e-01 5.06469548e-01 1.69766501e-01 -4.47764635e-01
1.41598955e-01 1.80498838e-01 -1.44487023e+00 6.85225546e-01
9.91899788e-01 3.01892430e-01 5.68647563e-01 5.09354472e-01
-3.18464458e-01 3.91016901e-01 -1.19091320e+00 -6.35801435e-01
1.22072041e-01 -3.97478700e-01 -1.01569343e+00 -1.25831753e-01
3.64806890e-01 1.43629476e-01 2.73521133e-02 1.01586819e+00
4.20440406e-01 8.86988267e-02 5.16189575e-01 -6.96191728e-01
-5.45331359e-01 1.15169382e+00 -4.17134427e-02 -8.65574107e-02
4.53916103e-01 8.43367074e-03 1.03817153e+00 -5.76059282e-01
7.46477664e-01 1.25403619e+00 6.33469760e-01 6.34995401e-01
-1.07956886e+00 -3.56134295e-01 -8.39059353e-02 -1.47671029e-01
-1.21412718e+00 -1.63368374e-01 3.21866155e-01 -1.31673157e-01
1.35538805e+00 5.09135664e-01 4.55473393e-01 9.25422788e-01
1.54301733e-01 6.74558103e-01 1.23533297e+00 -9.51758206e-01
-8.30864348e-03 5.67283213e-01 7.40092397e-02 1.84634656e-01
2.43218690e-01 1.34260207e-01 -2.18499705e-01 9.34176072e-02
4.24451798e-01 -5.62732100e-01 -3.08188915e-01 3.02859068e-01
-1.13823879e+00 4.80568916e-01 1.34056553e-01 9.03794944e-01
-2.74748236e-01 -2.97728926e-01 5.12625456e-01 6.27109528e-01
2.60931760e-01 5.91135979e-01 -9.07196403e-01 -3.08388323e-01
-1.09349871e+00 3.05433065e-01 9.25180495e-01 7.12631702e-01
5.55921733e-01 2.32486621e-01 -1.43754989e-01 1.08009911e+00
1.22229420e-01 2.07822621e-01 8.81103337e-01 -6.66090012e-01
7.79786527e-01 5.39849639e-01 2.78621614e-01 -6.00534022e-01
-3.88790578e-01 -4.39572185e-01 -5.03424048e-01 2.94109523e-01
9.60837066e-01 -2.69832760e-01 -1.05914187e+00 1.75720489e+00
-1.27637118e-01 -6.47562385e-01 4.25676167e-01 6.86049044e-01
6.95277572e-01 9.08243120e-01 6.87096044e-02 -1.86969310e-01
1.28255677e+00 -8.28363597e-01 -8.17715406e-01 -3.04627985e-01
8.92678797e-01 -9.50028896e-01 1.44590974e+00 5.46275377e-01
-1.25634515e+00 -5.91650248e-01 -1.05444741e+00 6.50196970e-02
-5.95496476e-01 2.41551340e-01 7.05045462e-02 8.70995641e-01
-1.25354075e+00 8.53492618e-01 -5.63946664e-01 -5.09335697e-01
-4.28065389e-01 3.52999270e-01 -3.57283890e-01 2.07940131e-01
-1.46557403e+00 1.04477644e+00 8.18540156e-01 -2.37746276e-02
-1.71029747e-01 -1.76094949e-01 -9.26360071e-01 1.56980306e-01
8.32843035e-02 -1.18096158e-01 1.52182341e+00 -1.25765169e+00
-1.43675554e+00 8.88080597e-01 -2.48529743e-02 -6.17234766e-01
7.29966044e-01 -5.92139214e-02 -6.17721736e-01 -1.86056629e-01
2.30541646e-01 5.30743241e-01 3.55290681e-01 -1.09387898e+00
-8.27981234e-01 -1.30467236e-01 -2.16931999e-01 1.47672161e-01
-2.08913777e-02 3.86795402e-01 -1.50372550e-01 -7.50042975e-01
1.17470790e-02 -7.33101249e-01 -1.08598605e-01 -8.56736243e-01
-3.16986769e-01 -4.21455801e-01 3.98478746e-01 -1.16256356e+00
1.48442495e+00 -2.03865385e+00 -2.85667986e-01 1.47511229e-01
-4.02927220e-01 5.17244577e-01 -1.45412236e-01 5.28854489e-01
-5.24718642e-01 4.61358994e-01 -3.81197959e-01 -2.38331154e-01
6.69300631e-02 2.29952887e-01 -1.80827510e-02 7.95304254e-02
3.98543745e-01 6.59028053e-01 -8.62934053e-01 -3.54507715e-02
-4.35445085e-02 1.84419885e-01 -2.01296464e-01 -2.14512035e-01
-1.20223381e-01 -2.66718958e-02 -2.48301420e-02 3.84958744e-01
4.54224706e-01 3.68508935e-01 2.40695804e-01 4.29094851e-01
-3.97175044e-01 7.03308880e-01 -1.23510313e+00 1.34617257e+00
-4.68351841e-01 5.97737551e-01 3.00347507e-02 -7.21707880e-01
1.08583534e+00 7.48589575e-01 -1.54187381e-01 -7.67392099e-01
4.92313355e-02 7.46784449e-01 3.94449830e-01 -4.69910622e-01
1.03212595e+00 -2.89861262e-01 -4.23055708e-01 2.48030975e-01
3.29649239e-03 -2.57097453e-01 4.10327464e-01 9.65281725e-02
9.06751394e-01 2.72824913e-01 3.32428336e-01 -2.81643838e-01
5.10355175e-01 2.24539861e-01 5.33665895e-01 7.01162279e-01
-1.17079355e-01 8.47662747e-01 6.05286241e-01 -1.98431611e-01
-1.15762687e+00 -8.02685618e-01 -1.36455685e-01 7.34906375e-01
-4.89644200e-01 -5.37513494e-01 -1.23510253e+00 -7.23634362e-01
-3.06938142e-01 1.27024376e+00 -2.76802331e-01 1.23857342e-01
-7.96483576e-01 -4.00842875e-01 8.50457013e-01 7.56905675e-02
2.76586205e-01 -1.82997024e+00 -3.91467392e-01 5.66538692e-01
-3.40220630e-01 -8.20857167e-01 -1.46617740e-01 4.80882406e-01
-6.85978889e-01 -7.18129396e-01 -7.91999221e-01 -1.10190749e+00
4.46819842e-01 -1.97597668e-01 9.93939519e-01 2.44063661e-01
2.09506765e-01 -8.53799954e-02 -4.65423197e-01 -5.25558949e-01
-1.15910995e+00 1.66899502e-01 -9.38264504e-02 -4.78259116e-01
4.40593719e-01 -1.48191720e-01 1.88742369e-01 2.56126970e-01
-1.09233785e+00 -2.30367273e-01 5.41080654e-01 1.01590610e+00
2.03026727e-01 3.32613848e-02 8.86328518e-01 -9.14667606e-01
7.92285144e-01 -1.06395923e-01 -3.69696736e-01 1.16600357e-01
-5.57184756e-01 -4.70590517e-02 8.00753772e-01 -4.06747639e-01
-1.06132972e+00 -8.55226740e-02 -4.96906787e-01 1.86762080e-01
-5.75985730e-01 6.63154542e-01 -4.80802894e-01 4.49400455e-01
8.39501619e-01 1.45045131e-01 -1.09806515e-01 -5.61425269e-01
-3.32576707e-02 1.11943293e+00 5.83771050e-01 -5.18323541e-01
5.15171885e-01 -3.53603393e-01 -7.35718846e-01 -9.04373825e-01
-3.72929275e-01 -3.05058241e-01 -4.80516195e-01 -1.90370083e-02
5.52488506e-01 -5.63179374e-01 -2.49329999e-01 6.44330978e-01
-1.40264249e+00 -4.66403842e-01 -5.00652969e-01 3.99510682e-01
-4.65964019e-01 6.03244543e-01 -5.35940886e-01 -8.20827186e-01
-2.86550820e-01 -1.15565825e+00 6.97854042e-01 1.58841655e-01
-7.09490716e-01 -8.83276761e-01 -1.98743165e-01 2.99843252e-01
1.94092661e-01 1.11641679e-02 1.11678195e+00 -1.09621119e+00
4.33444642e-02 -3.88423443e-01 1.15776911e-01 8.60150397e-01
1.76549718e-01 -1.53714612e-01 -8.92225385e-01 -1.26089290e-01
4.06832770e-02 -9.63395908e-02 6.37008011e-01 3.50168273e-02
3.69357347e-01 -4.77216035e-01 2.73600277e-02 -6.84602186e-02
1.11846733e+00 4.92756933e-01 9.02506173e-01 8.86524141e-01
1.69353172e-01 9.52592075e-01 6.29164696e-01 -1.53145734e-02
1.63455188e-01 5.68757057e-01 1.39503717e-01 1.57621652e-02
-2.30856627e-01 -2.39883602e-01 7.72565067e-01 8.74464273e-01
1.96732134e-01 -5.06630659e-01 -1.03718150e+00 9.54654813e-01
-1.64312232e+00 -9.00132895e-01 -5.32741249e-01 2.45217490e+00
1.10023785e+00 5.37330210e-01 -6.50584772e-02 6.10571384e-01
9.47551727e-01 -1.47529423e-01 1.81733742e-01 -1.13593554e+00
-3.77166718e-01 4.03174460e-01 2.10030586e-01 7.05606878e-01
-8.69506836e-01 1.32540810e+00 5.98997498e+00 9.60836530e-01
-1.06566429e+00 -4.37320471e-02 4.80832249e-01 1.31251924e-02
-1.51203141e-01 4.41904105e-02 -8.35449457e-01 6.19835317e-01
1.57295358e+00 -1.95887268e-01 4.13758188e-01 4.96079773e-01
6.00292325e-01 -2.25204974e-01 -8.04612339e-01 5.32800257e-01
-4.73373458e-02 -9.56323564e-01 2.48481538e-02 -5.40047474e-02
6.56021953e-01 -1.23527929e-01 -2.94653624e-01 6.17495835e-01
1.83221191e-01 -9.81977761e-01 1.21351731e+00 1.56102031e-01
6.12002611e-01 -8.65180314e-01 1.11727989e+00 6.49773717e-01
-5.49037516e-01 9.75181460e-02 -3.06838214e-01 -3.35694969e-01
7.82023147e-02 4.30618674e-01 -1.11444294e+00 7.47422278e-01
6.36220515e-01 6.99162781e-02 -5.14315188e-01 1.08408356e+00
-6.05816722e-01 8.88025045e-01 -4.86590892e-01 -3.35439950e-01
4.50802565e-01 -1.44312501e-01 7.12240279e-01 1.56570506e+00
4.72545117e-01 -2.34285861e-01 -1.95106834e-01 5.83390296e-01
1.93064705e-01 4.53423738e-01 -4.46246564e-01 1.07082287e-02
4.04136717e-01 1.09999084e+00 -6.89647675e-01 -2.81084359e-01
-1.64523125e-01 9.67177451e-01 2.34881982e-01 3.57648641e-01
-5.57837367e-01 -9.24692810e-01 2.71203876e-01 2.73562551e-01
2.99115092e-01 -5.42056784e-02 -4.85139281e-01 -6.20686591e-01
3.31242591e-01 -1.07384050e+00 2.02269658e-01 -8.65348697e-01
-8.85769665e-01 9.62994337e-01 1.84261687e-02 -1.06376505e+00
-5.46745598e-01 -7.52373755e-01 -5.65861285e-01 1.40378451e+00
-1.28030694e+00 -8.37295711e-01 2.81929553e-01 1.03459045e-01
5.78890622e-01 -2.27501616e-01 9.04395461e-01 2.64669389e-01
-3.87580663e-01 7.98638284e-01 2.22626656e-01 2.45666564e-01
7.96347558e-01 -1.41676784e+00 5.66631138e-01 9.39165592e-01
9.52616036e-02 6.98658645e-01 9.36230779e-01 -6.80867314e-01
-3.76207143e-01 -1.01137710e+00 1.79693758e+00 -2.59596795e-01
5.91629505e-01 -1.61122695e-01 -1.03866899e+00 5.15286684e-01
5.30421138e-01 -7.79132843e-01 2.93695539e-01 -7.69062266e-02
1.79608256e-01 2.33085304e-01 -1.23749554e+00 7.49201000e-01
5.57051420e-01 -1.71984598e-01 -1.07926214e+00 2.23977000e-01
7.89679945e-01 -4.36343342e-01 -5.41375101e-01 3.88934016e-02
2.70935267e-01 -9.34891582e-01 3.61977041e-01 -5.29201925e-01
3.80907923e-01 -3.39092702e-01 3.23195942e-02 -1.53220427e+00
-1.29513979e-01 -7.88610220e-01 5.97747505e-01 1.57225537e+00
9.56152976e-01 -7.07440972e-01 6.43369019e-01 5.49468160e-01
-6.54483795e-01 -3.53602976e-01 -1.26019585e+00 -9.62036908e-01
3.54922444e-01 -5.23462594e-01 4.33430880e-01 7.65835285e-01
2.08547428e-01 3.27688396e-01 -7.13786036e-02 -9.84558985e-02
-5.44320084e-02 -4.30327296e-01 4.05579299e-01 -1.26738024e+00
-2.27828205e-01 -5.08940518e-01 -1.84182510e-01 -5.04679322e-01
2.07516328e-01 -1.09254313e+00 3.30936372e-01 -1.60418332e+00
-3.91339421e-01 -3.03506136e-01 -1.13081085e-02 8.62500846e-01
-5.62359281e-02 1.24120563e-01 3.17928851e-01 -4.97294553e-02
1.13308616e-01 1.19318351e-01 1.04148936e+00 -1.10969674e-02
-6.35863006e-01 1.53212637e-01 -7.63515413e-01 7.65385032e-01
1.14955187e+00 -6.57590628e-01 5.99594228e-02 -4.32759970e-01
2.09590882e-01 -5.38810045e-02 1.72922894e-01 -9.89004314e-01
-4.84297946e-02 8.16054270e-03 2.98154235e-01 -4.69079524e-01
1.58432767e-01 -6.33746445e-01 1.91379398e-01 4.33555275e-01
-2.44560152e-01 2.31760547e-01 5.61080337e-01 -2.52170078e-02
-2.99127400e-01 -8.82254004e-01 6.32950485e-01 -3.03154171e-01
-4.88318115e-01 -6.61611974e-01 -9.15989518e-01 -1.46178767e-01
6.85370743e-01 -7.89565742e-01 -1.73799813e-01 -3.59954536e-01
-1.05512345e+00 -9.13230330e-02 4.29525286e-01 4.56754744e-01
2.13574290e-01 -9.30218101e-01 -8.77731979e-01 2.42999673e-01
-1.55087605e-01 -2.91570872e-02 -1.78082399e-02 5.71043253e-01
-5.34920156e-01 5.70668757e-01 -6.82603121e-02 -9.93324742e-02
-1.26101756e+00 1.27874479e-01 3.46972376e-01 -5.53870022e-01
-4.57429677e-01 4.45390642e-01 -2.46398777e-01 -9.21607375e-01
2.43714482e-01 -5.95648408e-01 -2.56571710e-01 1.75360665e-02
4.51732010e-01 1.58268347e-01 6.53035223e-01 -6.92283273e-01
-1.71183661e-01 1.04783528e-01 -1.64175063e-01 -4.32788581e-01
1.18508673e+00 6.40934333e-02 -5.18271625e-02 4.27050948e-01
7.43283093e-01 4.11066443e-01 -6.68440640e-01 8.42533037e-02
4.39738601e-01 -2.74799895e-02 -2.61741012e-01 -1.43286228e+00
-4.72434908e-01 6.55989170e-01 2.66571134e-01 4.23791349e-01
7.71558642e-01 -2.30527252e-01 5.86409509e-01 5.27503133e-01
3.04809988e-01 -1.28343880e+00 -4.58286256e-01 1.13926041e+00
9.64922309e-01 -6.48232579e-01 -3.77821356e-01 -2.64330685e-01
-6.98127508e-01 1.18389285e+00 4.98580039e-01 3.07692029e-02
2.01652180e-02 1.14099786e-01 3.37211281e-01 2.82856166e-01
-4.69427556e-01 -6.79670041e-03 1.23145040e-02 6.58331990e-01
7.80958593e-01 1.40697926e-01 -9.80055392e-01 1.01737356e+00
-8.68764997e-01 -1.82217285e-01 9.44181740e-01 7.87885606e-01
-4.82681185e-01 -1.43845582e+00 -7.08539128e-01 2.75489926e-01
-7.15973437e-01 -3.04655939e-01 -6.55526280e-01 1.13555455e+00
1.56696692e-01 1.03701770e+00 9.10432637e-02 -2.19883591e-01
6.18357003e-01 5.27087033e-01 3.73565033e-02 -8.04023087e-01
-1.40827715e+00 1.99304044e-01 5.47823131e-01 4.17318530e-02
1.49185229e-02 -8.29181433e-01 -1.50653553e+00 -4.72247563e-02
-3.39157134e-01 5.84183812e-01 8.52788329e-01 9.29836452e-01
9.23542026e-03 4.71621096e-01 2.29211152e-01 -6.66674018e-01
-6.80476010e-01 -1.44026899e+00 -6.43498361e-01 3.11870098e-01
5.52931465e-02 -8.72047022e-02 -4.96495396e-01 -1.80002376e-02] | [11.053295135498047, 10.152790069580078] |
9f1388e3-9184-4582-8c5c-3b288f2d687a | diffss-diffusion-model-for-few-shot-semantic | 2307.00773 | null | https://arxiv.org/abs/2307.00773v1 | https://arxiv.org/pdf/2307.00773v1.pdf | DifFSS: Diffusion Model for Few-Shot Semantic Segmentation | Diffusion models have demonstrated excellent performance in image generation. Although various few-shot semantic segmentation (FSS) models with different network structures have been proposed, performance improvement has reached a bottleneck. This paper presents the first work to leverage the diffusion model for FSS task, called DifFSS. DifFSS, a novel FSS paradigm, can further improve the performance of the state-of-the-art FSS models by a large margin without modifying their network structure. Specifically, we utilize the powerful generation ability of diffusion models to generate diverse auxiliary support images by using the semantic mask, scribble or soft HED boundary of the support image as control conditions. This generation process simulates the variety within the class of the query image, such as color, texture variation, lighting, $etc$. As a result, FSS models can refer to more diverse support images, yielding more robust representations, thereby achieving a consistent improvement in segmentation performance. Extensive experiments on three publicly available datasets based on existing advanced FSS models demonstrate the effectiveness of the diffusion model for FSS task. Furthermore, we explore in detail the impact of different input settings of the diffusion model on segmentation performance. Hopefully, this completely new paradigm will bring inspiration to the study of FSS task integrated with AI-generated content. | ['Bo Yan', 'Siyuan Chen', 'Weimin Tan'] | 2023-07-03 | null | null | null | null | ['few-shot-image-segmentation', 'image-generation'] | ['computer-vision', 'computer-vision'] | [ 6.27300620e-01 1.40968561e-01 -1.00649081e-01 -4.54026401e-01
-5.08835256e-01 -5.21365583e-01 6.24975741e-01 -3.25634152e-01
-2.28465125e-01 5.08607566e-01 3.62599641e-02 5.93827143e-02
1.01946943e-01 -8.99432182e-01 -6.28624737e-01 -6.46092474e-01
4.21738505e-01 3.44022840e-01 8.67058456e-01 -4.75757629e-01
3.93495589e-01 3.81646574e-01 -1.73218012e+00 2.59294361e-01
1.17450440e+00 9.88234818e-01 5.48044384e-01 4.35135663e-01
-4.69035447e-01 5.47598362e-01 -9.97719884e-01 -1.88730180e-01
4.68485713e-01 -6.32415712e-01 -7.04177797e-01 1.96175352e-01
4.71172482e-01 -1.46811947e-01 -3.27401489e-01 1.13354969e+00
5.85426688e-01 3.99992377e-01 7.19472945e-01 -1.21655166e+00
-1.04998970e+00 8.73768687e-01 -5.97384512e-01 2.17673987e-01
2.37359986e-01 4.89386201e-01 8.42779696e-01 -6.50394142e-01
1.11770058e+00 1.35209513e+00 5.62053204e-01 8.44901383e-01
-1.24324167e+00 -6.06740713e-01 2.58921742e-01 3.75185102e-01
-1.22612083e+00 -3.73828322e-01 9.31471407e-01 -1.24132663e-01
7.14438200e-01 2.72402525e-01 8.00382614e-01 1.36337590e+00
-1.30231902e-01 1.42776144e+00 1.36733317e+00 -3.93695027e-01
4.13953692e-01 1.51535302e-01 1.76946700e-01 8.36027980e-01
8.46796408e-02 -4.69344705e-02 -5.48294485e-01 2.53239453e-01
7.52631843e-01 -1.91574872e-01 -2.92856425e-01 -3.14538360e-01
-9.14470375e-01 8.77480388e-01 5.81115961e-01 4.08876926e-01
-8.66533071e-02 1.20346926e-01 2.30911463e-01 8.16367939e-02
6.35899484e-01 6.04630589e-01 -2.77705073e-01 1.51047379e-01
-1.36204243e+00 3.41428339e-01 8.00412893e-01 1.14680862e+00
5.39518952e-01 4.37859416e-01 -8.77802670e-01 1.10576785e+00
1.29674345e-01 4.31239277e-01 6.03851914e-01 -1.13683629e+00
1.55235693e-01 7.39146233e-01 -1.65960386e-01 -9.99309301e-01
-2.18557388e-01 -2.88745701e-01 -6.89897418e-01 1.39517084e-01
1.51411012e-01 -1.34836465e-01 -1.72914159e+00 1.85982060e+00
3.47976536e-01 4.09010202e-01 1.25401467e-01 8.62804115e-01
1.02316809e+00 7.08839774e-01 1.41296118e-01 1.49272487e-01
1.02845633e+00 -1.45265114e+00 -7.08932996e-01 -3.47385377e-01
2.47012034e-01 -7.46239960e-01 1.12318897e+00 5.07446118e-02
-1.11952102e+00 -6.02436543e-01 -1.14337003e+00 1.70561552e-01
-5.96276462e-01 -1.05707608e-01 8.53218555e-01 6.89782500e-01
-1.42211747e+00 7.81338990e-01 -7.07862854e-01 -4.54474121e-01
7.86346436e-01 1.01039194e-01 2.32730344e-01 -3.31467360e-01
-1.36918128e+00 6.42945230e-01 4.37823772e-01 -5.58607653e-02
-8.82591665e-01 -7.28558838e-01 -1.04258931e+00 4.27382765e-03
7.71388113e-01 -1.02062368e+00 1.10862446e+00 -1.14547896e+00
-1.62510276e+00 8.03497076e-01 -1.91718452e-02 -4.81679142e-01
6.43081665e-01 9.89704952e-02 -1.93748593e-01 2.84576893e-01
3.81320953e-01 1.25501382e+00 1.25699878e+00 -1.34256697e+00
-3.08396965e-01 -1.14114940e-01 -2.70157810e-02 2.65831143e-01
-1.67100057e-01 -1.00468740e-01 -8.50013018e-01 -1.02951109e+00
-6.56587332e-02 -9.61010575e-01 -6.59590364e-01 -5.12459092e-02
-6.79237008e-01 -1.72239751e-01 9.47117209e-01 -1.78462207e-01
1.11449909e+00 -2.12799287e+00 2.27256477e-01 1.94759995e-01
-7.23098516e-02 6.91742301e-01 -5.82337976e-01 2.60582656e-01
2.78127611e-01 3.36605817e-01 -7.68701136e-01 -3.99489492e-01
-1.09116975e-02 2.77849257e-01 -1.73969492e-01 -1.37383893e-01
3.38742644e-01 1.23753047e+00 -8.07980359e-01 -6.11873865e-01
1.89718008e-01 3.06684405e-01 -4.42396790e-01 1.30763754e-01
-5.39148450e-01 2.55058736e-01 -5.28901815e-01 7.31600583e-01
5.37022233e-01 -2.79194266e-01 -3.95199805e-01 -4.61411439e-02
2.14321390e-01 -3.25817049e-01 -1.28868484e+00 2.03527164e+00
-7.56305531e-02 4.30933774e-01 -7.37650171e-02 -8.76441538e-01
1.00052524e+00 -1.12358458e-01 3.25039417e-01 -8.42789710e-01
7.37309307e-02 1.50328830e-01 -5.41886054e-02 -3.85551423e-01
6.48938477e-01 -4.00914587e-02 -5.51836975e-02 3.82683665e-01
1.47478133e-01 -6.12829447e-01 6.32818997e-01 5.61326385e-01
8.81002903e-01 2.67859817e-01 -1.68751404e-01 -3.57135892e-01
3.05945903e-01 1.16379961e-01 5.38324416e-01 1.19129288e+00
-4.25563693e-01 9.75814521e-01 2.54906744e-01 -1.45117193e-02
-7.99309313e-01 -1.11684740e+00 8.61539990e-02 9.01887417e-01
5.03361881e-01 -1.96284279e-01 -1.20524716e+00 -8.15557480e-01
-2.99700741e-02 8.06426942e-01 -7.22512484e-01 -4.49836701e-01
-3.60768676e-01 -9.86968100e-01 7.31181145e-01 6.40677810e-01
9.00427818e-01 -1.49963272e+00 -5.36156356e-01 1.90765619e-01
-8.54965821e-02 -1.08132052e+00 -5.36324501e-01 -7.14218989e-02
-8.02455902e-01 -9.41607237e-01 -1.22680664e+00 -7.75676727e-01
5.75661361e-01 4.28016067e-01 1.03209519e+00 1.65433764e-01
-7.62681007e-01 3.98997486e-01 -5.37189543e-01 -3.81263286e-01
-3.94319534e-01 1.56344458e-01 -4.14728165e-01 1.24239000e-02
-2.93001495e-02 -2.68825531e-01 -9.35795665e-01 2.11797759e-01
-1.34686160e+00 2.36737013e-01 3.98943633e-01 6.56071603e-01
6.30180538e-01 8.15731734e-02 7.49547422e-01 -1.25216949e+00
8.02010417e-01 -4.48925018e-01 -2.44326502e-01 2.95421809e-01
-7.42654324e-01 4.29432169e-02 4.21540707e-01 -3.79446626e-01
-1.43210983e+00 -1.40464216e-01 -1.75300807e-01 -4.48400289e-01
-2.61519641e-01 3.59692812e-01 1.11273982e-01 -1.12024829e-01
6.57803416e-01 3.34204197e-01 -1.46259293e-01 -2.81807572e-01
7.10920632e-01 3.88028949e-01 2.77479112e-01 -5.02144337e-01
6.02227032e-01 5.86742043e-01 -1.80499569e-01 -8.42204690e-01
-9.87211108e-01 -4.21141356e-01 -4.18191105e-01 -1.31945357e-01
7.80411661e-01 -5.39479733e-01 2.02046588e-01 1.05672622e+00
-9.73249018e-01 -6.58941329e-01 -7.29951501e-01 -1.81755632e-01
-6.12237513e-01 3.68781269e-01 -7.98175454e-01 -5.61760664e-01
-3.20839554e-01 -1.27809107e+00 1.11715949e+00 6.09368801e-01
-1.38889745e-01 -9.93375003e-01 -2.55782455e-01 5.24667978e-01
6.92643285e-01 4.14271772e-01 8.13092709e-01 -5.62961578e-01
-7.24014580e-01 1.34590507e-01 -3.73070806e-01 4.91156042e-01
1.69327065e-01 1.45160735e-01 -7.88014770e-01 8.25242605e-03
-7.13175461e-02 -3.44125032e-01 1.37259936e+00 6.29114687e-01
1.33509195e+00 5.24231456e-02 -3.65210980e-01 7.88754821e-01
1.29709470e+00 3.35391045e-01 7.71385610e-01 3.09724063e-01
8.19716513e-01 5.49003303e-01 6.97473943e-01 1.75846592e-01
2.61412442e-01 4.74469513e-01 1.60961807e-01 -4.44365114e-01
-8.11910689e-01 -1.38913184e-01 2.14173600e-01 5.58921933e-01
4.82796617e-02 -5.43630719e-01 -8.41174364e-01 5.26901305e-01
-1.91186929e+00 -8.62312496e-01 -7.46236816e-02 1.60114992e+00
7.86474824e-01 1.31860793e-01 -6.37013465e-02 -3.54174785e-02
8.29795718e-01 5.74868917e-01 -9.02461708e-01 -3.12775224e-01
-3.88200432e-01 4.16137189e-01 3.29154164e-01 2.56961375e-01
-9.61162329e-01 1.78037333e+00 6.92840624e+00 1.26422238e+00
-1.04914904e+00 -9.84589010e-02 8.85714352e-01 -4.48478200e-02
-5.12335837e-01 -2.57998496e-01 -9.41194892e-01 5.85251033e-01
4.46735203e-01 -8.90198052e-02 3.34697902e-01 7.89127529e-01
8.42397586e-02 -3.55069876e-01 -7.03376651e-01 7.69472599e-01
4.35909033e-01 -1.46869838e+00 4.18880939e-01 -3.98099273e-01
1.22796381e+00 -1.65082347e-02 3.47415537e-01 3.80007267e-01
5.20557046e-01 -8.89145195e-01 8.12229693e-01 4.24131840e-01
8.61268163e-01 -4.51078296e-01 4.98642683e-01 1.54724002e-01
-1.02742386e+00 -2.62042787e-02 -2.15767071e-01 3.72790068e-01
4.91525590e-01 4.85939771e-01 -6.93589151e-01 4.41282719e-01
6.70783758e-01 7.41629004e-01 -9.34787035e-01 9.55486476e-01
-3.26896161e-01 6.90864801e-01 -2.07961768e-01 -2.11055949e-03
6.46775901e-01 -3.38301361e-01 6.11862898e-01 1.37885129e+00
2.35736102e-01 3.19025740e-02 1.63256645e-01 1.20950818e+00
-9.47401114e-03 6.54516183e-03 -5.82585871e-01 -1.93890139e-01
3.38700444e-01 1.04938579e+00 -1.36915207e+00 -6.33441031e-01
-1.27661958e-01 1.35101652e+00 1.42064571e-01 6.04073346e-01
-8.71443152e-01 -3.81757230e-01 4.13815051e-01 5.54094352e-02
4.47458565e-01 2.86241360e-02 -4.24520463e-01 -9.44295168e-01
-2.50947326e-01 -8.55794489e-01 3.36936504e-01 -7.25184858e-01
-1.39680231e+00 5.99434972e-01 8.57364386e-02 -6.26952887e-01
-4.60768305e-02 -2.94053823e-01 -6.60209775e-01 6.66850150e-01
-1.65073812e+00 -1.31275427e+00 -2.73170322e-01 7.19190717e-01
1.11428940e+00 -1.83742166e-01 5.08028269e-01 1.27137676e-02
-7.22185135e-01 3.75058383e-01 -3.50103639e-02 -1.18913077e-01
5.53635001e-01 -1.35172856e+00 7.48802364e-01 9.19804811e-01
2.87811875e-01 4.80249465e-01 4.64291006e-01 -8.97948384e-01
-8.55832338e-01 -1.23687482e+00 1.93177119e-01 -1.74745485e-01
4.85470682e-01 -2.37057999e-01 -9.18473780e-01 3.16491604e-01
2.55137384e-01 -1.28640637e-01 5.02875507e-01 -1.67145088e-01
-1.73010170e-01 7.90766701e-02 -1.16987681e+00 9.47735608e-01
1.48645771e+00 -1.50606185e-01 -5.35118163e-01 1.35967940e-01
9.42672014e-01 -4.18721706e-01 -3.36237729e-01 5.81867099e-01
1.20612450e-01 -1.10921681e+00 1.19079173e+00 -4.76781905e-01
5.70367813e-01 -1.64552838e-01 5.32507990e-03 -1.50344908e+00
-3.50277245e-01 -5.74141920e-01 -8.07785019e-02 1.42376375e+00
4.35517550e-01 -3.85908157e-01 9.06452775e-01 7.83937931e-01
-3.60047489e-01 -9.33241665e-01 -7.26843417e-01 -9.13639665e-01
-8.84778276e-02 -2.67453521e-01 6.70071363e-01 8.55816722e-01
-6.13282800e-01 1.58084169e-01 -1.64078251e-01 -3.47623855e-01
4.93384928e-01 2.30124578e-01 5.17893314e-01 -9.88618493e-01
-3.22629929e-01 -7.12144077e-01 -1.58960983e-01 -1.11489344e+00
1.99212223e-01 -9.72071648e-01 2.42931232e-01 -1.85240173e+00
1.86978936e-01 -7.54115760e-01 -3.37030292e-01 3.07595342e-01
-6.62822723e-01 4.89721119e-01 5.18907189e-01 6.19648956e-02
-7.02963650e-01 7.49989867e-01 1.74702036e+00 -2.71905243e-01
-3.81504416e-01 -1.55891836e-01 -8.92820239e-01 8.78270090e-01
9.29838896e-01 -3.69437009e-01 -8.02723467e-01 -4.91333961e-01
-2.31131807e-01 -2.99439400e-01 1.73471093e-01 -9.44727123e-01
1.24897167e-01 -3.32352310e-01 3.07203948e-01 -2.20487446e-01
3.67698163e-01 -4.36224312e-01 -8.88822600e-02 3.49406719e-01
-2.77861029e-01 -4.15514380e-01 2.41741210e-01 8.16492260e-01
-1.08185098e-01 -4.37577814e-01 8.81044507e-01 -4.97968405e-01
-1.22829366e+00 4.85602796e-01 -2.41445914e-01 3.89001459e-01
1.28152263e+00 -6.03341341e-01 -2.39091113e-01 -1.37788430e-01
-5.69813609e-01 1.91536456e-01 5.13272107e-01 6.90802336e-01
8.28569114e-01 -1.00416946e+00 -5.23797452e-01 2.04633281e-01
7.11925467e-03 2.88578928e-01 4.69303221e-01 3.94000977e-01
-3.08288366e-01 -2.33374909e-02 -1.13463812e-01 -5.99276364e-01
-1.07855999e+00 4.32498068e-01 8.04895759e-02 -1.35374010e-01
-7.05501378e-01 1.26914048e+00 3.73836279e-01 -2.12975726e-01
6.83216825e-02 -5.75442202e-02 -2.01889157e-01 1.66378483e-01
2.98494607e-01 3.60544473e-01 -3.21022421e-01 -3.74959558e-01
-7.45287463e-02 5.19047976e-01 -3.09186727e-01 -2.29445755e-01
1.23550558e+00 -1.27879515e-01 2.84418255e-01 3.63526464e-01
7.61919320e-01 -2.92336881e-01 -1.53266954e+00 -1.63074628e-01
1.09229973e-02 -5.35426557e-01 -4.18173149e-02 -1.11419332e+00
-1.41295457e+00 6.03440106e-01 4.12015200e-01 8.19057897e-02
1.09990990e+00 1.61598444e-01 1.14055371e+00 -3.50243486e-02
3.38350862e-01 -1.47823524e+00 5.35571933e-01 4.05203909e-01
8.43565047e-01 -1.37375462e+00 -3.79352301e-01 -7.43902981e-01
-1.03565776e+00 6.97100222e-01 6.93475008e-01 -2.16485709e-01
3.60267669e-01 2.03451067e-01 1.69074133e-01 -3.78393888e-01
-3.29774469e-01 -4.54674393e-01 2.58008033e-01 8.41035545e-01
1.68743581e-02 -6.60632849e-02 -2.23621637e-01 5.19910097e-01
-6.84404224e-02 -3.50733884e-02 4.64006186e-01 8.85277927e-01
-4.09140140e-01 -1.05284357e+00 -1.41281962e-01 5.24884105e-01
-2.24638447e-01 -3.19456637e-01 -6.63379848e-01 6.09500647e-01
2.18262210e-01 9.10252333e-01 -2.37060487e-01 -6.58147931e-02
3.26796591e-01 5.74297234e-02 5.06515920e-01 -9.03183162e-01
-5.14162660e-01 -1.19650997e-01 -1.24898136e-01 -6.84890270e-01
-5.12943685e-01 -6.63167715e-01 -1.25478935e+00 -1.09333783e-01
-3.52208555e-01 -1.92091614e-01 5.10682940e-01 8.13433349e-01
5.17514586e-01 7.98659980e-01 4.11612064e-01 -7.93624878e-01
-3.70486140e-01 -7.47460186e-01 -6.52508438e-01 7.73820817e-01
-1.33633003e-01 -7.18541324e-01 -2.48845816e-01 1.29634634e-01] | [9.688645362854004, 0.6339647173881531] |
6c37feb4-3fdc-4f76-ac88-82da84e48600 | zhijun-wu-chinese-semantic-dependency-parsing | null | null | https://aclanthology.org/S12-1058 | https://aclanthology.org/S12-1058.pdf | Zhijun Wu: Chinese Semantic Dependency Parsing with Third-Order Features | null | ['Xinxin Li', 'Zhijun Wu', 'Xuan Wang'] | 2012-07-01 | null | null | null | semeval-2012-7 | ['transition-based-dependency-parsing', 'semantic-dependency-parsing'] | ['natural-language-processing', 'natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.346733570098877, 3.732409954071045] |
146349d4-f20a-4887-afde-b2076d3b2388 | barriers-for-the-performance-of-graph-neural | 2306.02555 | null | https://arxiv.org/abs/2306.02555v1 | https://arxiv.org/pdf/2306.02555v1.pdf | Barriers for the performance of graph neural networks (GNN) in discrete random structures. A comment on~\cite{schuetz2022combinatorial},\cite{angelini2023modern},\cite{schuetz2023reply} | Recently graph neural network (GNN) based algorithms were proposed to solve a variety of combinatorial optimization problems, including Maximum Cut problem, Maximum Independent Set problem and similar other problems~\cite{schuetz2022combinatorial},\cite{schuetz2022graph}. The publication~\cite{schuetz2022combinatorial} stirred a debate whether GNN based method was adequately benchmarked against best prior methods. In particular, critical commentaries~\cite{angelini2023modern} and~\cite{boettcher2023inability} point out that simple greedy algorithm performs better than GNN in the setting of random graphs, and in fact stronger algorithmic performance can be reached with more sophisticated methods. A response from the authors~\cite{schuetz2023reply} pointed out that GNN performance can be improved further by tuning up the parameters better. We do not intend to discuss the merits of arguments and counter-arguments in~\cite{schuetz2022combinatorial},\cite{angelini2023modern},\cite{boettcher2023inability},\cite{schuetz2023reply}. Rather in this note we establish a fundamental limitation for running GNN on random graphs considered in these references, for a broad range of choices of GNN architecture. These limitations arise from the presence of the Overlap Gap Property (OGP) phase transition, which is a barrier for many algorithms, both classical and quantum. As we demonstrate in this paper, it is also a barrier to GNN due to its local structure. We note that at the same time known algorithms ranging from simple greedy algorithms to more sophisticated algorithms based on message passing, provide best results for these problems \emph{up to} the OGP phase transition. This leaves very little space for GNN to outperform the known algorithms, and based on this we side with the conclusions made in~\cite{angelini2023modern} and~\cite{boettcher2023inability}. | ['David Gamarnik'] | 2023-06-05 | null | null | null | null | ['combinatorial-optimization'] | ['methodology'] | [ 3.74174029e-01 4.18344229e-01 -1.95520520e-01 5.12458235e-02
-4.51137215e-01 -6.85978711e-01 4.87155855e-01 3.22756767e-01
-4.96261597e-01 1.24799752e+00 -3.16375852e-01 -8.27784896e-01
-1.08062589e+00 -1.09103727e+00 -7.20238924e-01 -7.84806311e-01
-6.89227045e-01 6.38641953e-01 2.09793281e-02 -4.94943589e-01
4.18781698e-01 5.65030634e-01 -1.48817980e+00 -2.77016252e-01
5.15355349e-01 7.69133568e-01 -6.34293398e-03 9.00421381e-01
1.05516829e-01 2.22306147e-01 -3.70477766e-01 -3.16317916e-01
4.90292698e-01 -6.22590125e-01 -1.01154423e+00 -4.14102525e-01
-3.13623920e-02 3.92510891e-01 -4.06891763e-01 1.44733059e+00
4.95144755e-01 1.10918134e-01 7.25736618e-01 -1.31807816e+00
-2.65730441e-01 9.07828629e-01 -6.29279137e-01 4.66517746e-01
2.51488298e-01 -1.79629922e-02 1.23358214e+00 -2.89680511e-01
6.49061263e-01 7.65099406e-01 6.40447438e-01 5.26296377e-01
-1.01752996e+00 -7.99405158e-01 -7.44093582e-03 2.87784010e-01
-1.43577695e+00 -1.95211425e-01 6.24089241e-01 -5.33559136e-02
7.87081540e-01 8.00645351e-01 4.75460559e-01 7.76553690e-01
6.57531768e-02 4.37023908e-01 1.44743609e+00 -6.50114954e-01
1.88792974e-01 -3.04264039e-01 2.09476769e-01 5.97092807e-01
7.88944900e-01 2.20467880e-01 -1.75267369e-01 -6.74550086e-02
7.70189106e-01 -4.92500633e-01 -3.17852139e-01 5.99278174e-02
-9.34460104e-01 1.03985405e+00 3.62399846e-01 5.53703725e-01
-2.51155142e-02 4.56957668e-01 2.12830797e-01 6.38458133e-01
1.58802673e-01 6.25796854e-01 -3.07998300e-01 -2.05544502e-01
-7.83599019e-01 3.95802945e-01 1.16178560e+00 1.06363773e+00
7.00991213e-01 -5.56622632e-03 3.22247475e-01 3.21250737e-01
3.10709417e-01 3.84360701e-01 -1.71172529e-01 -1.03392291e+00
5.37522435e-01 2.27506757e-01 -6.90386593e-02 -9.21044052e-01
-9.03218627e-01 -4.78668869e-01 -1.36585057e+00 4.72151220e-01
7.50146747e-01 -4.00976866e-01 -5.86001754e-01 1.67202818e+00
-4.52502929e-02 -2.72655249e-01 -2.03034967e-01 7.68866420e-01
9.65653896e-01 6.23594463e-01 -3.14683884e-01 -6.30343497e-01
1.06954229e+00 -5.34510791e-01 -4.15489018e-01 2.37804756e-01
7.72198260e-01 -8.28076661e-01 3.14948827e-01 6.40057921e-01
-1.25549710e+00 -2.87484583e-02 -1.23397982e+00 5.17964959e-01
-3.57800990e-01 -5.77391267e-01 1.07081747e+00 1.00252557e+00
-1.38633561e+00 9.86106217e-01 -6.22068107e-01 -5.65855742e-01
1.30826652e-01 9.41790640e-01 -1.81087241e-01 -7.65274987e-02
-1.22764444e+00 7.45595515e-01 6.88365340e-01 3.35009784e-01
-3.53661150e-01 -4.43090908e-02 -6.60588205e-01 -1.06643504e-02
6.69003904e-01 -4.40739810e-01 8.75599384e-01 -6.75565898e-01
-1.23637176e+00 7.93164313e-01 3.21407437e-01 -5.01283228e-01
5.92010021e-01 5.23692787e-01 -4.97959673e-01 1.09707393e-01
-1.97970152e-01 2.58460313e-01 -4.10951413e-02 -1.00455678e+00
-4.51871037e-01 -9.73067060e-02 4.07321960e-01 9.78447720e-02
1.15419336e-01 2.54208952e-01 -2.03859314e-01 -3.73187542e-01
4.00852948e-01 -1.05177808e+00 -3.92331064e-01 -6.76612914e-01
-7.88899839e-01 -6.13981009e-01 4.21746641e-01 -6.08476922e-02
1.19929767e+00 -1.60215569e+00 2.11253405e-01 7.09280968e-01
4.63107377e-01 3.20649832e-01 -4.69737984e-02 9.29850340e-01
-3.28352630e-01 4.02535349e-01 -1.30615532e-01 3.64385813e-01
2.34786972e-01 3.23461086e-01 3.12475890e-01 7.69797325e-01
-1.38552606e-01 8.14834714e-01 -7.59742081e-01 -1.61733642e-01
4.21949029e-02 -2.73256525e-02 -4.42700148e-01 -5.41142344e-01
-3.20555985e-01 1.84164494e-01 -4.94399160e-01 6.30337358e-01
6.95977986e-01 -5.07368624e-01 3.77983600e-01 1.14514634e-01
-3.74560922e-01 2.14571446e-01 -1.40721977e+00 1.14453578e+00
1.68448478e-01 6.12984300e-01 3.75282615e-01 -1.54429090e+00
7.09567785e-01 3.05141836e-01 6.08479083e-01 -5.86861730e-01
6.39987707e-01 3.39101762e-01 2.88349122e-01 -2.39817575e-01
4.55929041e-01 -2.16056749e-01 -1.06513783e-01 5.93566656e-01
-9.64426696e-02 -1.76842913e-01 5.29628158e-01 2.45077059e-01
1.35153425e+00 3.29867704e-03 2.69960463e-01 -7.84141302e-01
3.93468887e-01 -1.16436169e-01 6.00377731e-02 1.29030728e+00
-2.58444786e-01 6.42484128e-01 7.21123874e-01 -4.05669183e-01
-9.61923778e-01 -7.50118792e-01 -1.74338862e-01 9.13313150e-01
2.42993101e-01 -5.01973152e-01 -6.20426297e-01 -5.15949488e-01
-5.65265238e-01 4.92085725e-01 -6.54156685e-01 1.87997997e-01
-6.01205349e-01 -1.32775688e+00 7.55368650e-01 1.29559860e-01
5.18739164e-01 -1.12726247e+00 -3.21402401e-01 2.28108451e-01
9.79169384e-02 -7.17292666e-01 2.77467668e-01 7.13673115e-01
-6.13603115e-01 -1.22582710e+00 -5.43769002e-01 -6.05648577e-01
4.64059234e-01 4.99806777e-02 1.03496015e+00 5.04221082e-01
8.66637006e-03 1.80361345e-01 -3.72223049e-01 -4.36214268e-01
-4.46764797e-01 3.11016291e-01 5.90142049e-02 -7.21069813e-01
2.71558642e-01 -9.04954851e-01 -5.16805708e-01 1.91567108e-01
-9.17927086e-01 5.58935814e-02 5.25726259e-01 6.99173093e-01
2.59967327e-01 3.63672644e-01 4.37608302e-01 -9.87880945e-01
5.24240673e-01 -3.91330749e-01 -7.92680562e-01 -2.07150057e-02
-8.63924563e-01 4.45165597e-02 7.41502941e-01 6.04560263e-02
-2.38189057e-01 -5.87773860e-01 -3.86752695e-01 1.65540487e-01
-2.22552374e-01 5.99674106e-01 1.58777751e-03 -2.31810674e-01
7.23303378e-01 6.93620518e-02 -2.80234903e-01 -2.23515272e-01
1.21449925e-01 3.36896509e-01 3.05996925e-01 -5.85772157e-01
1.04164541e+00 3.74946594e-01 5.84008813e-01 -7.87360489e-01
-5.78795433e-01 -2.43699655e-01 -2.81661212e-01 -5.86184822e-02
7.21636891e-01 -2.40019143e-01 -1.28181756e+00 3.22489232e-01
-8.85382593e-01 -7.82542378e-02 -2.59387732e-01 6.69146717e-01
-6.41892910e-01 3.95513237e-01 -5.87952018e-01 -1.01776004e+00
-7.78318271e-02 -8.32313120e-01 3.52961540e-01 4.43328977e-01
-1.40223533e-01 -1.22966301e+00 -1.46964699e-01 3.06708425e-01
3.92051876e-01 4.50519890e-01 1.00177443e+00 -9.24650967e-01
-7.65571833e-01 -3.37182939e-01 -6.62878036e-01 3.73339728e-02
-2.18504682e-01 -7.92460591e-02 -6.64211810e-01 -6.29304826e-01
-2.32399926e-01 -8.58008862e-02 6.68446183e-01 3.52784485e-01
9.50891674e-01 -4.28487748e-01 -2.68102080e-01 5.16207278e-01
1.78926694e+00 3.20379913e-01 8.30938041e-01 6.53046966e-01
4.65302914e-01 3.92290980e-01 1.81535613e-02 1.59865156e-01
2.85134584e-01 4.55864280e-01 7.04989254e-01 8.33972916e-02
3.46616097e-02 2.30940685e-01 7.49479979e-02 9.05953407e-01
-6.65232360e-01 -7.29455471e-01 -8.17868233e-01 2.97379851e-01
-1.60318136e+00 -1.05784333e+00 -5.93336582e-01 2.18116403e+00
5.72506785e-01 4.38931942e-01 2.16010138e-01 3.01835209e-01
9.20294762e-01 1.63916007e-01 -1.81416661e-01 -7.92374074e-01
-2.01517865e-01 6.37469590e-01 8.90200317e-01 4.62112784e-01
-8.41580570e-01 5.34582973e-01 5.65143347e+00 1.04964066e+00
-8.05971920e-01 -1.27066925e-01 4.43391860e-01 -7.72351250e-02
-4.07703012e-01 5.15931010e-01 -6.88336968e-01 3.89479607e-01
1.11040354e+00 -4.97977495e-01 6.35222435e-01 1.04878969e-01
-8.71697441e-02 -3.18173200e-01 -9.74268138e-01 8.64532888e-01
4.27278280e-02 -1.42430162e+00 -2.93397367e-01 3.94318938e-01
6.80756748e-01 2.98018426e-01 -2.18620151e-01 3.61283064e-01
7.12590039e-01 -1.37760830e+00 2.62536883e-01 1.77022278e-01
7.72926986e-01 -8.93133998e-01 6.86072707e-01 2.97735482e-01
-9.44279015e-01 1.87412724e-01 -3.23653311e-01 -4.40027028e-01
5.30159473e-03 5.38269639e-01 -5.05932331e-01 1.26245892e+00
5.83009541e-01 5.22087179e-02 -2.81843781e-01 1.31822479e+00
9.20893811e-03 7.43852854e-01 -5.93146801e-01 -6.05219305e-01
4.86686230e-01 -5.08376420e-01 8.75301063e-01 1.05582690e+00
2.17163801e-01 3.89939964e-01 4.63953649e-04 7.13243544e-01
-1.11878976e-01 1.06920153e-02 -6.17504060e-01 -1.87950805e-01
5.47142401e-02 1.04596043e+00 -1.20522213e+00 7.77439796e-04
-2.62298882e-01 4.80982423e-01 5.00263944e-02 3.86573285e-01
-6.52479529e-01 -7.88075089e-01 -3.65009494e-02 1.89216718e-01
3.12609732e-01 -3.03703874e-01 -2.86293089e-01 -1.01927102e+00
-8.67225081e-02 -7.94379950e-01 5.56773841e-01 -4.25903618e-01
-1.07823777e+00 5.27582109e-01 4.57643032e-01 -7.98442006e-01
-8.57225880e-02 -7.93318272e-01 -4.88840342e-01 7.18703508e-01
-1.16481864e+00 -6.84748113e-01 2.10311443e-01 2.75625795e-01
-3.52593809e-02 2.62778960e-02 8.18908215e-01 1.13632306e-01
-3.20839494e-01 4.08220083e-01 5.59030354e-01 7.53294900e-02
4.24506031e-02 -1.22609723e+00 1.51955932e-01 7.74672806e-01
1.65521726e-01 4.85984862e-01 1.06282151e+00 -3.61203760e-01
-1.49228632e+00 -6.93977416e-01 8.37342381e-01 -1.34496525e-01
8.13065588e-01 -3.71916145e-01 -5.05853176e-01 5.05494118e-01
4.47735995e-01 -2.93847054e-01 4.07824188e-01 2.79350549e-01
1.36404097e-01 1.97474465e-01 -1.05576718e+00 6.12375557e-01
1.19179916e+00 -1.30482972e-01 -2.58752584e-01 7.89599538e-01
3.06015730e-01 -3.96173775e-01 -7.04636812e-01 4.61867303e-01
3.29641432e-01 -1.31783628e+00 5.27084827e-01 -3.53957087e-01
1.85157388e-01 -7.61459172e-02 -1.42005548e-01 -7.94940710e-01
-3.59861672e-01 -1.13163149e+00 2.21636266e-01 8.72511387e-01
4.98432994e-01 -9.61478770e-01 8.91325355e-01 3.10061276e-01
-4.07233427e-04 -7.92007089e-01 -1.46958649e+00 -8.98090124e-01
6.19993746e-01 -5.16369700e-01 1.44926161e-01 9.14763987e-01
3.53105390e-03 1.15598008e-01 -2.26585001e-01 -2.76065096e-02
7.52125382e-01 3.16950940e-02 4.74929392e-01 -1.23164952e+00
-5.09919047e-01 -7.64033973e-01 -6.00169182e-01 -8.48641932e-01
-2.70003974e-01 -1.06429279e+00 -3.36504638e-01 -1.64716434e+00
2.36036718e-01 -7.57066727e-01 -4.12403762e-01 3.69655073e-01
4.35113043e-01 4.32504177e-01 6.00851104e-02 2.67735776e-02
-8.05456579e-01 -6.32190406e-02 1.18324590e+00 1.12473838e-01
9.93245021e-02 1.56229734e-01 -9.06478286e-01 7.38853812e-01
7.61408627e-01 -4.43210989e-01 -2.06714734e-01 -6.47178665e-02
8.06448758e-01 2.70655245e-01 3.60255241e-01 -1.04837608e+00
3.57306451e-01 -2.04628244e-01 1.91134155e-01 -4.87001300e-01
2.11374145e-02 -4.96859759e-01 2.95029074e-01 5.20149529e-01
-6.85231462e-02 1.73700482e-01 1.93667829e-01 4.65134889e-01
2.37003699e-01 -8.10693979e-01 5.10474384e-01 -3.34748447e-01
-2.18026295e-01 2.79561669e-01 -4.96862113e-01 2.13477939e-01
9.88065302e-01 -5.10105848e-01 -4.60182220e-01 -6.00134671e-01
-7.74804294e-01 3.37345988e-01 2.96726257e-01 1.29996836e-01
1.09579191e-01 -1.09504867e+00 -6.21393800e-01 -7.34430831e-03
-3.87267709e-01 1.17599860e-01 9.69510376e-02 1.29200387e+00
-7.15870798e-01 5.47855198e-01 8.81440565e-02 -3.19950998e-01
-1.13727689e+00 5.40473938e-01 2.17609510e-01 -3.81503820e-01
-4.39891785e-01 1.00566471e+00 -2.40447745e-01 -3.45117509e-01
1.02323547e-01 -1.11380570e-01 3.03557497e-02 -1.86970249e-01
8.85080174e-02 6.24160171e-01 1.66982412e-01 -3.84856761e-01
-3.99309903e-01 2.80194432e-01 -1.05850615e-01 -1.17083108e-02
1.47684741e+00 -4.22595739e-02 -4.46375281e-01 2.71663725e-01
1.11592698e+00 8.55617821e-02 -4.84981358e-01 1.63043305e-01
-8.51783827e-02 3.71962748e-02 -2.74271041e-01 -8.08432162e-01
-8.07952046e-01 4.38345730e-01 3.60833675e-01 8.85411620e-01
1.12658250e+00 -9.07771215e-02 4.09366399e-01 5.81861973e-01
8.78158212e-01 -1.03640819e+00 -6.66171193e-01 5.88460267e-01
5.40995538e-01 -8.67755413e-01 4.29348826e-01 -3.99990052e-01
3.06677502e-02 1.39120173e+00 2.43124768e-01 -3.71699214e-01
5.76230407e-01 1.78831011e-01 -4.28082675e-01 -4.76966709e-01
-5.56120694e-01 -2.35554740e-01 -3.63282338e-02 2.75653183e-01
1.46550894e-01 -2.79775951e-02 -1.02594054e+00 2.00074896e-01
-4.18223381e-01 -6.77384064e-02 8.11792314e-01 9.54903722e-01
-3.64136517e-01 -1.53100562e+00 -4.29537475e-01 5.89567006e-01
-3.96054208e-01 -2.63829708e-01 -4.81377155e-01 1.31520760e+00
3.14194798e-01 1.16151798e+00 -3.47062677e-01 -2.48067737e-01
-6.14737310e-02 -2.45431840e-01 9.15142179e-01 -2.27711901e-01
-6.37222946e-01 9.03285667e-02 5.11457801e-01 -1.07494891e-01
-6.27837181e-01 -5.28136551e-01 -1.25725496e+00 -8.02534044e-01
-7.71741927e-01 7.75799453e-01 5.36091685e-01 1.21274841e+00
-1.37703326e-02 3.86774212e-01 2.95322210e-01 -6.85012162e-01
-6.08370602e-01 -7.60068536e-01 -8.42343330e-01 -2.52956506e-02
-3.59282978e-02 -3.42082143e-01 -5.77715516e-01 -6.66934431e-01] | [6.556680202484131, 5.3163838386535645] |
577d8b18-248d-4a3b-aef8-79d522a20819 | gridtopix-training-embodied-agents-with | 2105.00931 | null | https://arxiv.org/abs/2105.00931v2 | https://arxiv.org/pdf/2105.00931v2.pdf | GridToPix: Training Embodied Agents with Minimal Supervision | While deep reinforcement learning (RL) promises freedom from hand-labeled data, great successes, especially for Embodied AI, require significant work to create supervision via carefully shaped rewards. Indeed, without shaped rewards, i.e., with only terminal rewards, present-day Embodied AI results degrade significantly across Embodied AI problems from single-agent Habitat-based PointGoal Navigation (SPL drops from 55 to 0) and two-agent AI2-THOR-based Furniture Moving (success drops from 58% to 1%) to three-agent Google Football-based 3 vs. 1 with Keeper (game score drops from 0.6 to 0.1). As training from shaped rewards doesn't scale to more realistic tasks, the community needs to improve the success of training with terminal rewards. For this we propose GridToPix: 1) train agents with terminal rewards in gridworlds that generically mirror Embodied AI environments, i.e., they are independent of the task; 2) distill the learned policy into agents that reside in complex visual worlds. Despite learning from only terminal rewards with identical models and RL algorithms, GridToPix significantly improves results across tasks: from PointGoal Navigation (SPL improves from 0 to 64) and Furniture Moving (success improves from 1% to 25%) to football gameplay (game score improves from 0.1 to 0.6). GridToPix even helps to improve the results of shaped reward training. | ['Alexander Schwing', 'Luca Weihs', 'Aniruddha Kembhavi', 'Svetlana Lazebnik', 'Iou-Jen Liu', 'Unnat Jain'] | 2021-04-14 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Jain_GridToPix_Training_Embodied_Agents_With_Minimal_Supervision_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Jain_GridToPix_Training_Embodied_Agents_With_Minimal_Supervision_ICCV_2021_paper.pdf | iccv-2021-1 | ['pointgoal-navigation'] | ['robots'] | [-1.46272779e-01 3.71582329e-01 2.09245086e-02 2.11192518e-01
-7.64388382e-01 -8.04303586e-01 8.22279572e-01 -1.95383668e-01
-7.70612538e-01 1.07818711e+00 7.74257183e-02 -2.86246598e-01
-1.52299613e-01 -7.20699489e-01 -9.19069409e-01 -7.03667223e-01
-6.08063340e-01 6.09512746e-01 1.18883051e-01 -9.06800151e-01
1.14673972e-01 2.68049777e-01 -1.57738328e+00 6.19158894e-02
7.79752910e-01 7.10772812e-01 3.21104735e-01 9.95120525e-01
2.10964367e-01 1.23505962e+00 -6.16556644e-01 2.56842235e-03
5.04868746e-01 -3.55496287e-01 -7.30298698e-01 -3.38191986e-01
7.42558166e-02 -4.20380086e-01 -2.47875854e-01 7.56955147e-01
5.07521033e-01 3.29642832e-01 5.54345787e-01 -1.67687643e+00
-9.67176795e-01 5.60443044e-01 -4.37659085e-01 -2.79015332e-01
2.96131670e-01 6.25415206e-01 9.04460132e-01 -4.46455121e-01
6.69531345e-01 1.24080575e+00 6.38724267e-01 8.92080128e-01
-9.94085491e-01 -4.60030228e-01 4.02359605e-01 -1.01445250e-01
-9.47334766e-01 -4.87155735e-01 2.72670954e-01 -1.98990405e-01
1.42926466e+00 1.70695364e-01 8.61221850e-01 1.39076602e+00
1.28386825e-01 8.79322827e-01 1.12469101e+00 -2.05854341e-01
4.99040991e-01 -1.91454917e-01 -4.91872579e-01 8.28216016e-01
3.57174017e-02 5.98515630e-01 -4.50686991e-01 1.41500682e-01
1.10426748e+00 -2.47055814e-01 4.70510311e-02 -6.33861482e-01
-1.53383684e+00 5.51576793e-01 7.25237370e-01 2.40219548e-01
-4.17647153e-01 8.95567656e-01 1.96026549e-01 5.18118918e-01
6.16347231e-02 1.04803038e+00 -3.63484323e-01 -4.47761118e-01
-4.90776807e-01 7.81010330e-01 6.44180536e-01 1.17849839e+00
7.15700150e-01 5.96736550e-01 6.47304803e-02 6.29785657e-01
5.23318276e-02 8.64941657e-01 3.28106225e-01 -1.50246751e+00
3.01012933e-01 3.26241225e-01 6.90750420e-01 -6.37860000e-01
-8.65324497e-01 -4.32044953e-01 -4.78406370e-01 1.09123361e+00
7.37537265e-01 -5.89826822e-01 -1.16021502e+00 1.91729999e+00
-3.05443280e-03 -2.62340039e-01 4.26824838e-01 1.14174473e+00
7.11489260e-01 8.32250893e-01 2.25923032e-01 3.45484078e-01
1.11233127e+00 -1.24126160e+00 -1.53045252e-01 -6.14672363e-01
8.52207303e-01 -1.32781923e-01 1.49139988e+00 4.92919326e-01
-1.26221204e+00 -3.54109794e-01 -1.04938459e+00 8.40611756e-02
-3.55396867e-01 -3.09987605e-01 7.62238204e-01 4.30388987e-01
-1.39710593e+00 6.04503930e-01 -9.94837046e-01 -4.39583868e-01
3.51783127e-01 4.85863268e-01 -4.38545406e-01 -1.52972579e-01
-8.98567796e-01 1.30777740e+00 3.57596070e-01 -3.32258523e-01
-1.61352742e+00 -4.88068730e-01 -7.66396105e-01 -2.47512702e-02
7.01751590e-01 -9.26577747e-01 1.47632265e+00 -1.16573119e+00
-1.72100484e+00 6.31461620e-01 4.01008785e-01 -5.92944860e-01
7.28978932e-01 -4.06227469e-01 -9.14684907e-02 -4.39555272e-02
2.41951212e-01 1.10146928e+00 4.64705467e-01 -1.61397004e+00
-7.93050110e-01 -8.95761847e-02 7.77457833e-01 7.28553057e-01
-7.06103668e-02 -1.98328868e-01 -2.03808174e-01 -3.42326462e-01
-3.38464469e-01 -1.25117671e+00 -3.86576086e-01 1.58175364e-01
1.29635455e-02 -1.28893167e-01 4.37281460e-01 -4.17390019e-01
2.42588565e-01 -2.04134035e+00 4.52693611e-01 -2.11474299e-01
3.56706291e-01 -2.38027245e-01 -5.91892421e-01 4.45469856e-01
2.91807652e-01 8.43058750e-02 -7.45735765e-02 -3.09382498e-01
4.28246379e-01 3.73169541e-01 7.19506368e-02 1.74059212e-01
1.28699720e-01 1.10038698e+00 -1.45982707e+00 -8.32510069e-02
2.03717247e-01 2.18575895e-01 -7.56270051e-01 1.70357265e-02
-4.33498085e-01 4.86257255e-01 -3.77564847e-01 5.85369647e-01
1.48369163e-01 4.54856968e-03 -5.48264123e-02 5.61606169e-01
-4.67927963e-01 8.08900818e-02 -8.87320161e-01 2.03457785e+00
-6.04373693e-01 5.07499397e-01 2.53081352e-01 -6.61375940e-01
8.20446491e-01 2.69658536e-01 6.41573787e-01 -1.20280540e+00
-1.04219645e-01 1.23439141e-01 2.34477431e-01 -3.54124546e-01
8.32557738e-01 -3.70666206e-01 -3.13626349e-01 3.31015617e-01
-7.86067620e-02 -5.24304688e-01 8.87422115e-02 1.48867667e-01
1.46738434e+00 7.06296682e-01 -1.89603895e-01 -3.31305683e-01
-2.79407173e-01 5.84474266e-01 3.33961427e-01 9.78155911e-01
-1.87642321e-01 5.13020277e-01 4.98117536e-01 -5.53687096e-01
-1.32882524e+00 -1.25439215e+00 4.64884818e-01 1.64379597e+00
3.99779975e-01 -8.60343874e-02 -7.54058897e-01 -5.26564300e-01
2.52072848e-02 9.06570196e-01 -7.59693503e-01 -3.33100051e-01
-6.69429004e-01 -4.75554615e-01 6.30347788e-01 6.30754411e-01
7.21357346e-01 -1.53598416e+00 -1.18390524e+00 2.67608762e-01
-1.20652944e-01 -7.90469110e-01 3.29921171e-02 5.13804555e-01
-4.92176563e-01 -6.30974770e-01 -1.06815290e+00 -5.89992285e-01
3.49200726e-01 2.82471567e-01 1.27842271e+00 1.06532633e-01
5.93239777e-02 5.47526240e-01 -5.82661271e-01 -4.69060034e-01
-2.77890205e-01 8.35307986e-02 2.40331858e-01 -8.20097983e-01
-3.79660964e-01 -5.22509813e-01 -6.32788718e-01 3.05584013e-01
-4.94946867e-01 2.92266697e-01 7.19070017e-01 8.80957425e-01
1.01131119e-01 -3.31636876e-01 6.52646184e-01 -2.46979982e-01
6.23926401e-01 -5.63384891e-01 -5.32823980e-01 8.03533867e-02
-4.46067840e-01 1.41214371e-01 7.51034975e-01 -6.81519985e-01
-8.32547486e-01 -2.76197553e-01 6.20469116e-02 -2.70029485e-01
-8.00106898e-02 2.38218844e-01 2.70643711e-01 -1.64210442e-02
1.24965322e+00 1.57930106e-02 -1.23605885e-01 5.12283109e-02
6.92210078e-01 1.70031399e-01 7.33795047e-01 -9.84149516e-01
6.74250782e-01 4.10733163e-01 -3.67744833e-01 -3.95177811e-01
-3.17656189e-01 2.48503879e-01 7.46286064e-02 -2.68829048e-01
8.96392703e-01 -8.81908834e-01 -1.06792200e+00 2.66446710e-01
-6.46206677e-01 -1.34095216e+00 -6.35932922e-01 4.60934520e-01
-1.14724517e+00 -1.93490565e-01 -6.04941308e-01 -9.91514504e-01
-9.24150199e-02 -1.08092523e+00 1.03489614e+00 2.22386003e-01
-3.82889241e-01 -6.05762124e-01 1.94242820e-01 7.44073018e-02
5.06484568e-01 5.48578382e-01 7.02796459e-01 -1.92755237e-01
-6.74495578e-01 2.05334157e-01 -1.79991961e-01 -6.79484382e-02
8.62791464e-02 -3.81959289e-01 -8.01722348e-01 -4.57130015e-01
-5.22130311e-01 -8.43627989e-01 4.12778318e-01 1.90464139e-01
5.32100737e-01 2.55453940e-02 -2.32128009e-01 3.96239877e-01
1.05914593e+00 6.62190795e-01 6.49882078e-01 9.87855017e-01
5.10682940e-01 4.98908460e-01 6.16052747e-01 4.64505941e-01
5.60113847e-01 5.28698027e-01 1.02276123e+00 -1.45982146e-01
-2.10934237e-01 -3.40110213e-01 7.90847600e-01 1.07805908e-01
-6.19721055e-01 -3.86816949e-01 -1.01033366e+00 5.32173932e-01
-1.99363375e+00 -9.83697116e-01 3.12546432e-01 2.01331902e+00
5.68731785e-01 3.41698349e-01 4.61916894e-01 -2.69590914e-01
2.01483235e-01 8.50381479e-02 -1.05613959e+00 -4.71061528e-01
-1.60852354e-02 3.87211218e-02 3.50261599e-01 3.46128047e-01
-7.87859619e-01 1.40857983e+00 6.22124863e+00 5.19523084e-01
-1.04975414e+00 1.81652643e-02 3.19618613e-01 -5.07161975e-01
-1.94177374e-01 -8.03725421e-02 -1.93875939e-01 3.19117308e-01
7.28512764e-01 -7.28620365e-02 1.14799380e+00 1.05415511e+00
-6.21133223e-02 -2.53882259e-01 -1.08726811e+00 1.01178753e+00
-2.92490512e-01 -1.12652361e+00 -3.91348690e-01 1.89940482e-01
7.22326756e-01 3.65425587e-01 2.99492627e-01 1.01459026e+00
1.36549652e+00 -1.49465466e+00 1.30861199e+00 3.96038890e-01
7.95544088e-01 -9.27789450e-01 3.54737937e-01 6.01547897e-01
-9.09151375e-01 -1.82686359e-01 -3.52847010e-01 -5.30828655e-01
8.69869590e-02 -2.63003469e-01 -8.66624713e-01 4.41573888e-01
8.85138452e-01 4.39662993e-01 6.98544905e-02 8.04006040e-01
-1.96302161e-01 3.81833781e-03 -3.89881670e-01 -3.60845834e-01
7.31038094e-01 -9.13514942e-02 4.57214564e-01 7.39640474e-01
3.72757912e-01 1.86923012e-01 3.72940838e-01 7.88130641e-01
-8.21131805e-04 -2.71705598e-01 -8.80771399e-01 -1.29235052e-02
3.11138570e-01 1.04944956e+00 -6.17848873e-01 -1.36742756e-01
-5.79749569e-02 1.09829867e+00 6.65606081e-01 5.07841527e-01
-1.05079603e+00 -2.71581560e-01 8.81982565e-01 4.39137369e-02
1.50219843e-01 -4.99286473e-01 7.25688552e-03 -8.59644473e-01
-3.53883117e-01 -1.12973154e+00 -7.50966594e-02 -1.21656251e+00
-9.60014880e-01 7.77201176e-01 -1.76317975e-01 -1.02669811e+00
-3.83526683e-01 -5.79032540e-01 -3.95614088e-01 6.12519860e-01
-1.30789793e+00 -1.23057866e+00 -3.60996246e-01 5.36289155e-01
4.73412991e-01 -1.34800091e-01 9.35621500e-01 -2.01224566e-01
-2.27990419e-01 4.31400537e-01 8.20748135e-02 5.91296609e-03
5.60076356e-01 -1.40034723e+00 8.01563442e-01 3.44988674e-01
1.55907393e-01 2.81261533e-01 7.72019029e-01 -4.41362053e-01
-1.34874237e+00 -5.21315694e-01 -1.95553266e-02 -8.01758349e-01
6.63712919e-01 -3.24505359e-01 -3.62758428e-01 6.67869508e-01
4.30734783e-01 -6.10376447e-02 1.15164749e-01 3.53860646e-01
-3.83432597e-01 1.34186894e-01 -1.21418571e+00 1.22056925e+00
1.52530265e+00 -2.55198538e-01 -5.01889646e-01 2.43287355e-01
7.16225505e-01 -5.43513298e-01 -7.56334484e-01 2.69887894e-01
7.91513741e-01 -9.66492236e-01 1.04695261e+00 -7.71088243e-01
5.43200493e-01 -4.06450659e-01 -2.67852575e-01 -1.77262092e+00
-6.24729574e-01 -5.94315946e-01 4.18338589e-02 4.69247103e-01
3.80264968e-01 -7.22594082e-01 9.25988972e-01 5.97254336e-01
-4.81137544e-01 -6.05490148e-01 -7.51480639e-01 -1.02171028e+00
4.20393050e-01 -3.15146446e-01 4.51697111e-01 9.43331718e-01
2.31305048e-01 2.10198581e-01 -4.49004024e-01 3.67872268e-02
2.83688664e-01 -1.95027485e-01 9.58109915e-01 -8.46834600e-01
-5.57977736e-01 -7.13371992e-01 -9.31376666e-02 -1.04225540e+00
-5.16953841e-02 -7.50320256e-01 3.50819677e-01 -1.86702991e+00
-1.80273503e-02 -1.00129414e+00 -2.31635183e-01 9.65913594e-01
9.65608880e-02 1.67961910e-01 7.51339495e-01 9.16939452e-02
-8.06516469e-01 7.31511652e-01 1.37120903e+00 -2.14934081e-01
-2.28969589e-01 -5.47225118e-01 -8.82185519e-01 6.76829338e-01
9.40932572e-01 -4.27968323e-01 -6.51046991e-01 -8.57698143e-01
4.49276745e-01 2.16567993e-01 5.65871000e-01 -1.29594135e+00
3.94701548e-02 -4.36966509e-01 3.34136873e-01 4.78844158e-02
6.47461295e-01 -6.57132030e-01 5.74541502e-02 5.74082494e-01
-2.22254857e-01 2.57284224e-01 4.74934131e-01 4.12886113e-01
3.70727688e-01 -1.42809138e-01 3.12640399e-01 -4.89694804e-01
-9.14909065e-01 -4.30306718e-02 -6.21309996e-01 2.89395094e-01
1.16607511e+00 -4.91158575e-01 -6.61226809e-01 -7.07686901e-01
-9.27161753e-01 5.48580468e-01 8.33971858e-01 4.12760645e-01
4.63443577e-01 -1.17802072e+00 -6.37990355e-01 -2.14426383e-01
5.03020287e-02 2.29460582e-01 5.53274043e-02 4.21439230e-01
-5.79336464e-01 -4.45317775e-02 -6.92868114e-01 -3.91699612e-01
-8.86991084e-01 4.55759704e-01 4.63301659e-01 -2.28746206e-01
-5.77237248e-01 9.26334500e-01 4.23429787e-01 -6.54237151e-01
1.91768080e-01 -4.28505331e-01 1.08958296e-01 -3.24486107e-01
2.00434059e-01 4.20686752e-01 -3.12565655e-01 -1.07539862e-01
-2.78322428e-01 4.95891124e-01 1.11281157e-01 -5.86365998e-01
1.59900522e+00 3.00584257e-01 4.68677461e-01 3.76111329e-01
6.12382770e-01 -1.15674891e-01 -2.00519371e+00 3.17632705e-01
-1.84705734e-01 -2.19157487e-02 -1.79006919e-01 -1.21489859e+00
-7.63631403e-01 6.13863945e-01 4.70170796e-01 1.42923087e-01
7.68405676e-01 -4.01118770e-02 5.54729402e-01 8.47512603e-01
1.14263105e+00 -9.85061705e-01 6.98978662e-01 6.96882367e-01
1.10183585e+00 -1.20648396e+00 -2.77425289e-01 3.74849558e-01
-1.10789156e+00 6.62381589e-01 8.50936532e-01 -4.44334328e-01
-2.25190684e-01 5.34766734e-01 2.88074855e-02 -1.64769322e-01
-6.97831273e-01 -5.65539122e-01 -2.15948537e-01 1.00024438e+00
-5.57492021e-03 1.25947565e-01 5.74860036e-01 2.96700954e-01
-4.66629893e-01 -1.54216573e-01 2.81539410e-01 1.16643071e+00
-7.27698386e-01 -7.99517035e-01 -3.89786214e-01 3.21772061e-02
1.61533490e-01 -5.73959351e-02 -3.87555510e-01 1.24833560e+00
1.57214507e-01 9.26041186e-01 9.27666575e-02 -3.69000137e-01
3.35899442e-01 -2.04920113e-01 7.79006004e-01 -5.26270270e-01
-7.94779837e-01 -4.52891886e-02 3.76367092e-01 -6.88477159e-01
-2.98191875e-01 -4.71139342e-01 -1.90686679e+00 -6.42961979e-01
9.54752192e-02 1.15446754e-01 5.44642270e-01 6.26305342e-01
2.98106253e-01 7.18267977e-01 1.14294782e-01 -1.21983743e+00
-3.78177017e-01 -7.07037270e-01 -3.73164833e-01 1.24984592e-01
4.63798434e-01 -8.14105690e-01 -1.38877183e-01 -2.68715113e-01] | [4.291147232055664, 0.9834454655647278] |
2eb4b42d-8d40-4090-b50d-549a02ff966b | longitudinal-quantitative-assessment-of-covid | 2103.07240 | null | https://arxiv.org/abs/2103.07240v2 | https://arxiv.org/pdf/2103.07240v2.pdf | Longitudinal Quantitative Assessment of COVID-19 Infection Progression from Chest CTs | Chest computed tomography (CT) has played an essential diagnostic role in assessing patients with COVID-19 by showing disease-specific image features such as ground-glass opacity and consolidation. Image segmentation methods have proven to help quantify the disease burden and even help predict the outcome. The availability of longitudinal CT series may also result in an efficient and effective method to reliably assess the progression of COVID-19, monitor the healing process and the response to different therapeutic strategies. In this paper, we propose a new framework to identify infection at a voxel level (identification of healthy lung, consolidation, and ground-glass opacity) and visualize the progression of COVID-19 using sequential low-dose non-contrast CT scans. In particular, we devise a longitudinal segmentation network that utilizes the reference scan information to improve the performance of disease identification. Experimental results on a clinical longitudinal dataset collected in our institution show the effectiveness of the proposed method compared to the static deep neural networks for disease quantification. | ['Thomas Wendler', 'Nassir Navab', 'Rickmer Braren', 'Egon Burian', 'Tobias Czempiel', 'Matthias Keicher', 'Ashkan Khakzar', 'Magdalini Paschali', 'Leili Goli', 'Seong Tae Kim'] | 2021-03-12 | null | null | null | null | ['covid-19-image-segmentation'] | ['computer-vision'] | [ 8.98738205e-02 -6.43937111e-01 -6.08103760e-02 -5.21027185e-02
-3.74888569e-01 -3.09666872e-01 2.02315435e-01 3.72499883e-01
-2.99793750e-01 4.98849779e-01 1.07939102e-01 -4.04171526e-01
-2.19696805e-01 -7.76322126e-01 -9.23163295e-02 -6.74485743e-01
-4.53423023e-01 1.20070672e+00 3.53338331e-01 3.39370340e-01
-2.12150067e-01 7.02996790e-01 -7.27729738e-01 1.38876021e-01
5.61732292e-01 8.54682088e-01 5.36016524e-01 7.25821376e-01
1.14745922e-01 8.07703018e-01 -4.53564376e-01 2.34621271e-01
1.26679748e-01 -4.45828110e-01 -7.15169549e-01 3.16909067e-02
2.20351070e-01 -9.43867505e-01 -7.17000440e-02 5.66367269e-01
4.93170768e-01 -1.75664052e-01 9.01774049e-01 -8.84200335e-01
9.51919928e-02 1.76027656e-01 -6.24037623e-01 8.53022933e-01
-5.16199879e-02 2.26377517e-01 4.58862543e-01 -6.05341971e-01
7.03321874e-01 7.82854080e-01 8.73592615e-01 3.86550725e-01
-6.84487879e-01 -5.38546741e-01 -4.09831464e-01 1.16546527e-01
-1.00812423e+00 5.44737577e-01 1.44680619e-01 -9.76710260e-01
6.58796132e-01 3.29255402e-01 1.00660527e+00 8.68598998e-01
4.20384288e-01 4.29274201e-01 1.06799507e+00 -5.41411303e-02
-7.28665292e-02 -4.30599511e-01 3.21784019e-01 1.06394923e+00
5.50939381e-01 1.25258937e-01 1.39495209e-01 -2.13171631e-01
1.00669873e+00 6.08830929e-01 -5.46817601e-01 -3.56611371e-01
-1.03513396e+00 8.11854303e-01 5.26524663e-01 5.28896570e-01
-5.93364954e-01 2.02957988e-01 6.15064681e-01 -5.06796837e-01
3.47037524e-01 2.03824043e-01 -1.01554073e-01 1.88427255e-01
-1.07767153e+00 5.08468151e-02 7.98768625e-02 -1.91303883e-02
-5.24930805e-02 7.12646767e-02 -4.39325213e-01 6.32166207e-01
3.46911460e-01 8.14848363e-01 6.86530888e-01 -6.64701521e-01
3.41588438e-01 5.71843624e-01 4.46319617e-02 -6.20623231e-01
-8.49017680e-01 -3.93634379e-01 -1.12020934e+00 1.48524120e-02
2.76012629e-01 -8.65012035e-02 -1.22561145e+00 1.24096739e+00
2.73077339e-01 3.17123085e-01 -4.01997447e-01 1.06665838e+00
5.36627114e-01 5.31032324e-01 2.84028798e-01 -1.89217150e-01
1.45692313e+00 -6.79712594e-01 -4.84719515e-01 2.73614615e-01
5.08109093e-01 -2.51449168e-01 8.48816872e-01 1.37008488e-01
-8.52714181e-01 -1.36087507e-01 -9.90092218e-01 5.45330465e-01
-1.35580286e-01 5.34253009e-02 2.17824325e-01 6.27880394e-01
-7.03375638e-01 5.22513449e-01 -1.30378234e+00 -4.31203395e-01
4.39859390e-01 3.08951646e-01 -4.15907763e-02 9.96382092e-04
-1.17636847e+00 9.67471957e-01 3.09179932e-01 -1.05295470e-02
-1.19858241e+00 -7.71098077e-01 -4.12960470e-01 6.88107386e-02
1.68727279e-01 -1.13094389e+00 1.02851784e+00 -1.69887841e-02
-6.79182112e-01 7.41021395e-01 2.79903002e-02 -5.51192105e-01
7.45777547e-01 -1.21025229e-02 6.67645484e-02 6.65868282e-01
3.93807963e-02 3.12500685e-01 5.99346995e-01 -1.22778845e+00
-6.04061365e-01 -6.38045907e-01 -2.11315677e-01 4.88953367e-02
-9.29062366e-02 1.49702623e-01 -2.60683775e-01 -5.96416950e-01
3.33921276e-02 -8.40459824e-01 -2.24829078e-01 -2.41523311e-01
-2.53737926e-01 9.26261395e-02 1.16874683e+00 -1.05142212e+00
1.02238500e+00 -1.51659238e+00 -3.37381095e-01 2.62774885e-01
8.06704283e-01 4.80615288e-01 3.73891562e-01 -5.25726564e-02
9.53864157e-02 3.86285722e-01 -6.25594854e-01 7.64940754e-02
-4.95253503e-01 9.79332253e-02 4.28854674e-02 5.39616525e-01
1.85529608e-02 9.63001549e-01 -5.73454440e-01 -7.45896757e-01
3.57171685e-01 6.60512686e-01 -2.26338729e-01 3.90773714e-01
-1.23178847e-01 9.05732810e-01 -6.01643801e-01 4.88660693e-01
4.62162107e-01 -8.90503228e-01 2.09066764e-01 1.90977439e-01
7.29745552e-02 -1.41304299e-01 -5.36004126e-01 7.30832577e-01
-5.07435858e-01 5.39881229e-01 1.09897755e-01 -6.27678096e-01
3.05419058e-01 6.15492284e-01 9.45556343e-01 -6.54462934e-01
5.08822381e-01 1.20685976e-02 5.26423641e-02 -7.86316454e-01
-6.59583881e-02 -4.10423845e-01 3.45241010e-01 6.96861625e-01
-5.84384739e-01 -1.49516001e-01 1.37714133e-01 -6.87738284e-02
1.02548730e+00 -3.18199188e-01 2.49212623e-01 5.29073514e-02
4.81300175e-01 2.72259027e-01 1.67791605e-01 7.14222729e-01
-3.24536353e-01 7.48735070e-01 3.14294487e-01 -4.63702440e-01
-1.09171593e+00 -1.25461328e+00 -3.47056925e-01 4.93743896e-01
3.51200625e-02 3.81846428e-01 -8.21384430e-01 -5.40135622e-01
-2.06106618e-01 2.96505183e-01 -5.80730319e-01 3.15757304e-01
-1.05097580e+00 -1.11089838e+00 5.23496389e-01 8.07427406e-01
5.51905036e-01 -9.99333322e-01 -1.23581934e+00 2.01217264e-01
-3.82755488e-01 -9.62075889e-01 -1.69228911e-01 7.77209029e-02
-1.20652497e+00 -1.34655046e+00 -1.11285496e+00 -7.32865155e-01
5.10332763e-01 1.18646011e-01 8.76519680e-01 4.83043075e-01
-7.86240220e-01 2.87289917e-01 1.29689081e-02 -2.14532226e-01
-5.77618420e-01 -1.97042763e-01 -1.31835476e-01 -5.43396473e-01
-2.66031891e-01 -2.37818256e-01 -8.69307458e-01 1.74788192e-01
-8.23437989e-01 5.68995401e-02 3.03878099e-01 6.61612272e-01
8.32326353e-01 -1.97927933e-02 1.93802193e-01 -9.36804116e-01
8.87513816e-01 -6.55720353e-01 -5.19349396e-01 2.40484402e-01
-8.63544822e-01 -4.14216638e-01 5.96796274e-01 1.78793430e-01
-8.98592114e-01 -3.97156686e-01 7.78791029e-03 -6.62954330e-01
-1.19464196e-01 4.54582781e-01 7.13478506e-01 1.95217043e-01
3.35087895e-01 3.06077898e-01 2.60286033e-02 -3.34327191e-01
-1.33005396e-01 5.20123482e-01 5.40941417e-01 -3.77749681e-01
5.08419454e-01 7.97168255e-01 4.03020650e-01 -6.54489577e-01
-5.14250338e-01 -7.86354005e-01 -7.54348040e-01 -3.02363932e-01
1.40502822e+00 -5.86756647e-01 -7.66251683e-01 4.69506741e-01
-1.10844398e+00 -2.92515039e-01 -6.96613491e-02 5.10562599e-01
-3.75023067e-01 3.78027529e-01 -8.85945439e-01 -2.82474071e-01
-8.88609767e-01 -1.41303122e+00 7.68834293e-01 4.38009351e-02
-2.01600790e-01 -1.15153146e+00 4.10364985e-01 4.97055531e-01
5.88594913e-01 5.68664074e-01 1.38288903e+00 -5.97029865e-01
-7.56591320e-01 -2.17677355e-01 -5.26905835e-01 2.92166531e-01
-1.32867359e-02 -7.77927414e-02 -3.77340883e-01 -1.38560101e-01
2.52806991e-01 -5.46881780e-02 7.38538325e-01 8.23262155e-01
1.17924380e+00 1.53337330e-01 -4.80821103e-01 7.04844534e-01
1.52959275e+00 6.01351142e-01 2.05650721e-02 1.92757815e-01
8.48018050e-01 2.46349707e-01 3.17293882e-01 5.22503614e-01
3.18675563e-02 4.29475337e-01 5.39608419e-01 -3.26162815e-01
-2.84275085e-01 1.80760413e-01 -3.76843959e-01 8.40592742e-01
-5.16185403e-01 -4.29373652e-01 -1.54609704e+00 4.34644520e-01
-1.22628808e+00 -8.29639077e-01 -6.10174656e-01 1.71496379e+00
2.93623745e-01 -1.19437516e-01 2.19161361e-01 -1.00595906e-01
8.64915848e-01 -5.66397049e-02 -2.92482466e-01 -2.40551040e-01
2.16362506e-01 4.49232936e-01 4.19709116e-01 2.40806311e-01
-1.02274990e+00 2.19020009e-01 6.92079163e+00 3.87168318e-01
-1.62531340e+00 5.33213377e-01 6.73070192e-01 2.74681151e-02
1.04253059e-02 -6.82557523e-01 -4.46834534e-01 4.58870798e-01
6.19227231e-01 1.32001668e-01 -8.73099081e-03 5.08160591e-01
5.52605927e-01 -2.80232698e-01 -6.94844365e-01 5.34915984e-01
-8.79135802e-02 -1.44913232e+00 8.06701481e-02 -2.19594985e-02
6.24858737e-01 3.20958167e-01 2.15039462e-01 -5.74048311e-02
-2.95158383e-02 -9.91091430e-01 2.84692049e-01 1.98100135e-01
1.06181324e+00 -4.17865515e-01 9.94521260e-01 1.99897766e-01
-1.28690743e+00 1.16497397e-01 1.68392509e-01 4.79351133e-01
3.00403625e-01 8.59452877e-03 -1.68084526e+00 2.60981590e-01
5.87070823e-01 1.23668611e-01 -3.53644311e-01 1.19719946e+00
-8.02579969e-02 8.33494604e-01 -2.87559807e-01 5.79090901e-02
3.50921780e-01 -2.19193935e-01 5.20384252e-01 1.29639745e+00
3.30100775e-01 3.60724211e-01 1.76054478e-01 9.03750718e-01
9.47520956e-02 2.14380831e-01 -4.52152848e-01 8.65881667e-02
-6.62944466e-02 9.14426088e-01 -1.28135943e+00 -4.88466054e-01
-2.42961928e-01 5.13589382e-01 -2.37027615e-01 2.86669075e-01
-1.01412892e+00 3.36723626e-01 -3.58612873e-02 7.05832720e-01
1.98697746e-01 -1.75419047e-01 -5.24951339e-01 -5.99710584e-01
-3.23780119e-01 -5.43259025e-01 5.46771348e-01 -7.45358944e-01
-8.54956925e-01 7.53977954e-01 2.43217424e-01 -9.36498940e-01
-3.40599328e-01 -4.96051311e-01 -8.40141654e-01 8.55882347e-01
-1.23769891e+00 -8.63872528e-01 -7.11045265e-01 3.98946166e-01
3.13961983e-01 4.55131158e-02 7.12889254e-01 2.73864567e-01
-6.08602822e-01 -4.09625918e-02 1.80895939e-01 1.34202167e-01
4.89140153e-02 -1.26383984e+00 -5.04664220e-02 6.84360564e-01
-4.67324108e-01 6.28717661e-01 2.76144385e-01 -1.01601338e+00
-5.63685358e-01 -1.17745757e+00 2.54619896e-01 -2.68285036e-01
5.44622958e-01 2.85394222e-01 -7.51138687e-01 5.40227413e-01
4.57088947e-02 3.88455838e-02 5.02861738e-01 -5.25894761e-01
1.33439466e-01 2.36014396e-01 -1.23109603e+00 2.30045885e-01
4.92582589e-01 -2.55455822e-01 -5.59054017e-01 3.05557489e-01
4.07008290e-01 -4.29517448e-01 -6.61234617e-01 8.73034239e-01
4.91399348e-01 -9.59724545e-01 1.08824146e+00 -4.62355524e-01
5.38951218e-01 -4.34427857e-02 1.16222553e-01 -8.02414715e-01
-2.44667560e-01 2.15410605e-01 6.26635328e-02 4.00442690e-01
-6.49747029e-02 -4.13996845e-01 1.04850161e+00 3.44935320e-02
5.70588373e-02 -1.17301023e+00 -6.47220612e-01 -4.83273655e-01
-5.81368245e-02 -2.25425914e-01 2.84909695e-01 8.04423153e-01
-6.91111743e-01 -1.34953320e-01 1.08809909e-02 1.64573297e-01
5.52468061e-01 2.86579639e-01 8.23691711e-02 -1.24796605e+00
-1.75246105e-01 -5.14084995e-01 -2.44420588e-01 -4.64478463e-01
-3.15172762e-01 -1.08176506e+00 -2.52771884e-01 -2.07983208e+00
8.20689797e-01 -5.06553471e-01 -5.27957976e-01 7.71674141e-02
-2.18623906e-01 4.50934500e-01 1.82878450e-01 5.66783071e-01
-1.67137131e-01 -5.38910702e-02 1.70175636e+00 -1.43110901e-01
-9.14501492e-03 1.49018735e-01 2.36982241e-01 9.11698937e-01
8.66429508e-01 -6.69564128e-01 -3.69359583e-01 -1.25304058e-01
-1.76828697e-01 6.39354765e-01 6.51892126e-01 -1.01101792e+00
-2.54777253e-01 -5.78301698e-02 5.17382801e-01 -1.00321031e+00
1.58407941e-01 -8.07698190e-01 1.99758768e-01 1.32026184e+00
-4.64518592e-02 2.51009643e-01 1.63412526e-01 4.05420631e-01
-1.71741813e-01 -3.32780242e-01 1.05882299e+00 -3.62368524e-01
-4.51727122e-01 5.20765781e-01 -7.20769286e-01 2.62672156e-01
1.12639189e+00 -1.62360549e-01 -2.88779527e-01 -1.57070547e-01
-7.20573068e-01 2.46678423e-02 2.73028761e-01 -6.23827055e-03
6.62400663e-01 -7.30166852e-01 -6.85504138e-01 -8.10759217e-02
-2.37470359e-01 5.80405705e-02 3.55055630e-01 1.29809296e+00
-1.55082178e+00 7.06945121e-01 -2.47885093e-01 -9.63088691e-01
-1.54148006e+00 3.81645381e-01 6.82699621e-01 -9.84977841e-01
-6.81453466e-01 8.52675259e-01 5.02482355e-01 1.51427463e-01
1.19390577e-01 -3.26251805e-01 -3.36748004e-01 -1.34649372e-03
1.63749322e-01 4.65357989e-01 8.01214725e-02 -6.05358481e-01
-3.59721452e-01 6.48881435e-01 7.94095360e-03 1.80138528e-01
1.14592195e+00 -1.16172440e-01 -1.93690211e-01 1.98526174e-01
1.07112098e+00 -2.28397250e-01 -8.00477147e-01 6.87993914e-02
-2.10410923e-01 -1.00321136e-01 1.35754094e-01 -8.74309242e-01
-1.10295069e+00 1.14558768e+00 1.27606595e+00 1.54399201e-01
9.56697285e-01 5.63131422e-02 1.12837911e+00 -1.14892021e-01
1.86725602e-01 -3.57579380e-01 2.93804020e-01 1.61096931e-01
4.59508240e-01 -1.10961735e+00 5.26015088e-02 -3.32160264e-01
-4.60464865e-01 1.11691558e+00 4.62393671e-01 7.78749771e-03
4.21483696e-01 3.85234773e-01 2.91442186e-01 -6.98389351e-01
-1.85813054e-01 -2.25527994e-02 1.26839608e-01 5.22114694e-01
4.29458231e-01 4.21396136e-01 -1.62379131e-01 2.65135914e-01
2.10481301e-01 -5.85164828e-03 3.27752352e-01 7.42745161e-01
-6.26976848e-01 -7.73631275e-01 -6.05892181e-01 8.47394168e-01
-8.49502921e-01 -9.11390111e-02 -1.18377008e-01 1.11133230e+00
4.03782487e-01 4.14930046e-01 1.18581899e-01 3.03801745e-02
3.93942520e-02 1.78153679e-01 6.31660104e-01 -6.00305140e-01
-6.51568830e-01 1.74450368e-01 -2.34933749e-01 -2.19648957e-01
-3.62503737e-01 -5.35064876e-01 -1.73940408e+00 -4.79773507e-02
-1.82905998e-02 -1.51417702e-01 7.01781154e-01 8.18009555e-01
-1.47799775e-01 1.01871502e+00 3.97629976e-01 -2.89177686e-01
-4.25807744e-01 -7.78857827e-01 -5.48537612e-01 4.43345755e-01
4.23350185e-01 -5.95214128e-01 -1.39492884e-01 1.43460006e-01] | [15.446248054504395, -1.8366994857788086] |
8e3a2c77-2930-41ae-9dba-5ed6dfeaa92c | jsi-gan-gan-based-joint-super-resolution-and | 1909.04391 | null | https://arxiv.org/abs/1909.04391v2 | https://arxiv.org/pdf/1909.04391v2.pdf | JSI-GAN: GAN-Based Joint Super-Resolution and Inverse Tone-Mapping with Pixel-Wise Task-Specific Filters for UHD HDR Video | Joint learning of super-resolution (SR) and inverse tone-mapping (ITM) has been explored recently, to convert legacy low resolution (LR) standard dynamic range (SDR) videos to high resolution (HR) high dynamic range (HDR) videos for the growing need of UHD HDR TV/broadcasting applications. However, previous CNN-based methods directly reconstruct the HR HDR frames from LR SDR frames, and are only trained with a simple L2 loss. In this paper, we take a divide-and-conquer approach in designing a novel GAN-based joint SR-ITM network, called JSI-GAN, which is composed of three task-specific subnets: an image reconstruction subnet, a detail restoration (DR) subnet and a local contrast enhancement (LCE) subnet. We delicately design these subnets so that they are appropriately trained for the intended purpose, learning a pair of pixel-wise 1D separable filters via the DR subnet for detail restoration and a pixel-wise 2D local filter by the LCE subnet for contrast enhancement. Moreover, to train the JSI-GAN effectively, we propose a novel detail GAN loss alongside the conventional GAN loss, which helps enhancing both local details and contrasts to reconstruct high quality HR HDR results. When all subnets are jointly trained well, the predicted HR HDR results of higher quality are obtained with at least 0.41 dB gain in PSNR over those generated by the previous methods. | ['Munchurl Kim', 'Soo Ye Kim', 'Jihyong Oh'] | 2019-09-10 | null | null | null | null | ['tone-mapping'] | ['computer-vision'] | [ 6.12143159e-01 1.19098425e-01 -7.80732110e-02 -3.57321411e-01
-1.03202617e+00 -8.70826393e-02 4.61776942e-01 -7.53222525e-01
-6.14678711e-02 9.12201703e-01 4.38818127e-01 -1.53312489e-01
1.98252678e-01 -9.89585578e-01 -8.13483596e-01 -7.62307107e-01
2.34384194e-01 -2.46577948e-01 1.90613419e-01 -4.38843429e-01
-1.89459309e-01 4.55816299e-01 -1.38975406e+00 3.86787325e-01
9.18695807e-01 1.23967874e+00 3.32961082e-01 7.98244715e-01
2.64606774e-01 1.34194899e+00 -4.19581413e-01 -1.42623723e-01
5.40400505e-01 -8.83769691e-01 -5.66025078e-01 -2.08049007e-02
4.41087961e-01 -9.11977172e-01 -8.22102010e-01 8.52972567e-01
5.88015795e-01 2.53425926e-01 2.59952217e-01 -6.30492091e-01
-9.12594974e-01 4.36097980e-01 -8.82085323e-01 1.90918148e-01
3.55034709e-01 2.76082397e-01 7.77928352e-01 -5.74409306e-01
6.16656005e-01 1.27642131e+00 5.82235932e-01 7.10776508e-01
-1.40805018e+00 -8.42056096e-01 -7.38998204e-02 7.94068873e-02
-1.14433146e+00 -3.61291856e-01 9.63648677e-01 8.04168284e-02
6.74556613e-01 4.11884308e-01 6.32923782e-01 9.74761844e-01
1.10858858e-01 4.75334853e-01 1.44426060e+00 -5.79073206e-02
-1.62674949e-01 -1.66753143e-01 -5.66381693e-01 3.76415640e-01
-1.79639965e-01 3.39421332e-01 -2.95536488e-01 3.94099772e-01
1.61247146e+00 1.21498428e-01 -6.89115107e-01 2.40754262e-01
-9.61741567e-01 6.00849867e-01 7.12207556e-01 3.26145440e-01
-4.27816480e-01 2.81504244e-01 1.10829107e-01 4.79461014e-01
6.11186743e-01 3.02163273e-01 -2.26462200e-01 2.59888679e-01
-1.01979756e+00 1.01674013e-02 2.15129137e-01 8.06669593e-01
7.86271811e-01 2.97764093e-01 -3.46982896e-01 1.19969952e+00
9.47202817e-02 3.93101007e-01 2.61347592e-01 -1.23801363e+00
4.38016504e-01 1.27149373e-01 1.10359654e-01 -7.59092510e-01
8.62504356e-03 -7.12819576e-01 -1.32029736e+00 3.21195960e-01
-1.73107665e-02 1.33036986e-01 -1.02273011e+00 1.81798637e+00
3.06076407e-01 2.52907932e-01 1.12282880e-01 1.26465523e+00
8.58406663e-01 1.00602186e+00 2.60482915e-03 -3.00436407e-01
1.19107091e+00 -1.13596237e+00 -8.82790625e-01 -8.37166756e-02
-3.72090898e-02 -8.57809782e-01 1.05485630e+00 2.84081936e-01
-1.75269115e+00 -1.04978240e+00 -1.27288270e+00 -7.11226046e-01
2.65395761e-01 -1.89130078e-03 3.33738953e-01 1.62472382e-01
-1.27090490e+00 5.71881890e-01 -4.30530667e-01 1.30577058e-01
3.91128629e-01 5.94538711e-02 -1.29270718e-01 -5.48155546e-01
-1.42172921e+00 6.51697338e-01 -3.55310999e-02 2.62175173e-01
-1.18337941e+00 -9.16305363e-01 -7.94487000e-01 5.72870560e-02
2.02332214e-01 -8.42910111e-01 8.70752871e-01 -1.10727322e+00
-1.81052005e+00 7.81108320e-01 1.34721577e-01 -3.56151283e-01
6.11702383e-01 -4.53577116e-02 -5.93665719e-01 3.47902417e-01
-9.77932662e-02 7.41493523e-01 1.00926447e+00 -1.58232605e+00
-7.79070258e-01 -5.69418743e-02 1.65501237e-01 3.23554516e-01
8.97751600e-02 -7.44028436e-03 -5.15630543e-01 -1.02894056e+00
4.09814343e-02 -3.70689213e-01 -1.27314344e-01 1.60978541e-01
-2.62888551e-01 3.82947654e-01 1.03033257e+00 -1.52252185e+00
1.16378140e+00 -2.11415315e+00 1.53607264e-01 -6.05652146e-02
3.69139344e-01 3.24422061e-01 -4.37684238e-01 -1.25807211e-01
-1.93386048e-01 -1.01572588e-01 -4.12907809e-01 -2.40364790e-01
-3.62415522e-01 6.42537773e-02 -3.80271494e-01 3.42571646e-01
2.65537262e-01 9.79050517e-01 -7.12987423e-01 -2.17145190e-01
3.53379190e-01 1.13822758e+00 -4.61947352e-01 6.12541735e-01
-6.23591617e-02 9.27813530e-01 -2.76253283e-01 4.94309813e-01
9.83669043e-01 -1.12764612e-01 1.32483140e-01 -6.79796278e-01
-1.96824059e-01 1.77462727e-01 -1.01066720e+00 1.48904991e+00
-9.68118191e-01 5.35733819e-01 4.03524429e-01 -7.31735766e-01
1.19582522e+00 2.52834707e-01 4.80809599e-01 -1.35416460e+00
-7.41552189e-02 3.23950142e-01 -3.58661652e-01 -1.33662090e-01
5.49727857e-01 -4.51321483e-01 3.17590982e-01 8.97123441e-02
-1.09417878e-01 -3.67947742e-02 -2.10245013e-01 -1.35190040e-01
1.02285969e+00 2.47809693e-01 4.29415666e-02 1.13131598e-01
7.54073918e-01 -5.43439806e-01 6.33786380e-01 3.03105950e-01
1.55654117e-01 1.19097662e+00 2.89145142e-01 -4.30114537e-01
-1.57008481e+00 -1.05901456e+00 -4.56541181e-02 6.99991703e-01
2.45380268e-01 5.93947992e-02 -5.84813118e-01 -2.66611010e-01
-5.36108553e-01 5.36542594e-01 -3.50483984e-01 -1.99564785e-01
-8.75668406e-01 -4.03562486e-01 2.19598830e-01 5.65567255e-01
1.25204110e+00 -1.01285553e+00 -2.40202770e-01 3.09490383e-01
-4.78174627e-01 -1.26710093e+00 -7.48203874e-01 1.19819738e-01
-6.95858657e-01 -7.59985983e-01 -1.18460059e+00 -7.21546590e-01
4.14658934e-01 2.64324963e-01 1.18011928e+00 1.23253562e-01
-1.79140508e-01 -2.09196601e-02 -3.73320788e-01 4.20344710e-01
-6.03481054e-01 -2.37697735e-01 -5.47342777e-01 1.90213978e-01
-4.78257477e-01 -8.25414360e-01 -1.14061272e+00 4.81738448e-01
-1.23716295e+00 4.86992806e-01 8.75194311e-01 9.95452464e-01
9.74863231e-01 4.16180104e-01 5.36338508e-01 -7.95291841e-01
1.76579237e-01 -1.25365645e-01 -5.23126483e-01 1.07213274e-01
-4.49781775e-01 -3.18312258e-01 9.43665326e-01 -3.04146737e-01
-1.52108157e+00 -2.11290732e-01 -5.44277191e-01 -6.69263482e-01
1.58765875e-02 -1.25272378e-01 -4.81451482e-01 -2.39425629e-01
3.55726898e-01 5.31153202e-01 -1.09197289e-01 -4.48528528e-01
3.52374703e-01 6.56967759e-01 7.10322261e-01 -2.42313892e-01
9.13990140e-01 5.18244267e-01 1.39864413e-02 -5.84414601e-01
-9.06884789e-01 -6.11925535e-02 4.27247304e-03 -2.19818756e-01
1.24721181e+00 -1.38400650e+00 -4.26673502e-01 5.38855433e-01
-8.01769137e-01 -8.80931318e-01 -6.41745329e-01 2.86396474e-01
-6.78281605e-01 2.00369596e-01 -1.03020871e+00 -5.42873800e-01
-5.51320374e-01 -1.21990156e+00 1.18825305e+00 3.45891744e-01
3.84085029e-01 -7.42590785e-01 -2.70707577e-01 6.40852690e-01
9.61779952e-01 4.47288990e-01 8.74890029e-01 3.48028809e-01
-8.36992383e-01 2.14703292e-01 -7.00713992e-01 9.28155780e-01
1.28878832e-01 -4.69335556e-01 -9.76087928e-01 -4.96872753e-01
3.71275663e-01 -2.10675523e-01 1.03926682e+00 6.73290014e-01
1.38800919e+00 -4.11672771e-01 2.36308783e-01 1.15825343e+00
1.77957058e+00 2.33602598e-01 1.39917886e+00 1.93968266e-01
9.74203825e-01 3.05324227e-01 5.48091888e-01 1.81968778e-01
3.27045798e-01 8.41984332e-01 3.27985883e-01 -8.37817192e-01
-1.00607574e+00 -4.59037870e-01 4.57089901e-01 4.91270274e-01
-2.51809627e-01 -1.71925604e-01 -8.75012130e-02 2.47870579e-01
-1.33435214e+00 -9.07241881e-01 -8.88676499e-04 2.14829731e+00
1.04422390e+00 -1.66056350e-01 -4.22176197e-02 2.30791755e-02
8.15793931e-01 5.02864063e-01 -7.11712241e-01 -1.63503841e-01
-4.06626463e-01 3.92199785e-01 4.77884054e-01 6.03079319e-01
-7.76672304e-01 6.84503615e-01 5.02048874e+00 1.22325373e+00
-1.28599966e+00 3.20508331e-01 1.28170228e+00 9.07830447e-02
-6.50669754e-01 -3.60284559e-02 -5.00341535e-01 4.38579053e-01
7.32269943e-01 5.65366924e-01 8.92086387e-01 5.03305614e-01
4.06539619e-01 6.40206039e-02 -5.91574609e-01 1.12386036e+00
-1.33725658e-01 -1.22473562e+00 2.04302788e-01 7.58360475e-02
9.63629007e-01 -4.64797318e-01 2.48081923e-01 1.55206800e-01
2.73490809e-02 -1.09006095e+00 6.57441080e-01 5.52302599e-01
1.47286296e+00 -9.50633466e-01 4.34548199e-01 -1.64018691e-01
-1.35537159e+00 -2.05928028e-01 -4.70810980e-01 4.27865207e-01
4.06060159e-01 8.63319993e-01 1.27926514e-01 6.73454821e-01
9.52683032e-01 8.39195013e-01 -1.70012206e-01 5.07037580e-01
-3.87287468e-01 4.34275985e-01 1.13644712e-01 9.63590026e-01
1.96989570e-02 -3.96251321e-01 6.29601955e-01 8.47236156e-01
2.99137026e-01 3.21573287e-01 -1.60296440e-01 1.18329930e+00
-3.28828305e-01 -2.40361795e-01 -2.36488864e-01 3.13500434e-01
1.27063766e-01 1.40451288e+00 -4.75710183e-01 -2.96496868e-01
-4.31221038e-01 1.29704940e+00 -6.38977438e-02 5.80266595e-01
-9.85564291e-01 -3.47206295e-01 5.86140335e-01 4.11883026e-01
4.91852999e-01 4.52872366e-02 -1.07700385e-01 -1.12444794e+00
-2.70626247e-02 -1.13879156e+00 1.35640815e-01 -1.09746981e+00
-1.28447211e+00 8.64865661e-01 -3.85735393e-01 -1.36187434e+00
7.86639377e-02 -9.33481008e-02 -5.08026183e-01 1.13415599e+00
-2.14892197e+00 -1.36401021e+00 -5.85652590e-01 6.74656987e-01
5.45456886e-01 1.76305041e-01 1.30659953e-01 7.45476961e-01
-2.81406850e-01 5.06804943e-01 2.03097262e-03 -3.43453651e-03
7.16822147e-01 -1.03839803e+00 1.90778330e-01 9.52894747e-01
-5.34173012e-01 2.23483309e-01 6.31892085e-01 -6.10735774e-01
-1.26162350e+00 -1.51830363e+00 4.11930978e-01 3.06750357e-01
1.12896323e-01 -1.12085849e-01 -9.30339217e-01 5.29972017e-01
1.02406830e-01 1.92167252e-01 1.14628375e-01 -6.00194335e-01
-2.98826098e-01 -5.35472274e-01 -1.64192998e+00 4.58864331e-01
1.06290925e+00 -5.77852428e-01 -7.86678307e-03 4.22120802e-02
1.20127702e+00 -5.93790174e-01 -1.27774632e+00 5.62987864e-01
4.13959622e-01 -1.22525477e+00 1.21899390e+00 2.23189414e-01
1.16776705e+00 -6.48456931e-01 -4.07259852e-01 -1.02448380e+00
-3.82850498e-01 -6.02651417e-01 -1.04832456e-01 1.43974984e+00
8.93434957e-02 -5.61953485e-01 3.62897038e-01 2.22185910e-01
-3.74283373e-01 -8.22254121e-01 -6.57039762e-01 -5.64293206e-01
-2.63186812e-01 3.61562669e-02 6.75930381e-01 9.00181890e-01
-9.10143137e-01 3.57108802e-01 -9.07051384e-01 9.66850519e-02
7.26151407e-01 9.17558819e-02 5.05980194e-01 -5.50718904e-01
-5.77727318e-01 -2.39060596e-01 -2.64659040e-02 -1.42226708e+00
-1.80339754e-01 -5.63233733e-01 7.19183162e-02 -1.62564087e+00
3.40638936e-01 -4.38645840e-01 -2.10870326e-01 1.22638702e-01
-1.47235438e-01 6.88320518e-01 1.28859699e-01 1.91556379e-01
-3.29562455e-01 7.08723366e-01 1.93448305e+00 4.79496829e-02
-3.17121774e-01 -2.01920137e-01 -7.78836548e-01 2.41963193e-01
4.12662864e-01 -7.82258213e-02 -3.72203827e-01 -4.10536289e-01
6.29089773e-02 6.61169171e-01 6.10883236e-01 -1.06287730e+00
-1.44182533e-01 8.60806182e-02 7.62530327e-01 -3.50043565e-01
4.48360324e-01 -7.90561438e-01 5.69146097e-01 2.73887336e-01
-5.56263685e-01 -3.46393049e-01 -1.56386152e-01 4.95128185e-01
-5.28748274e-01 4.49071825e-01 1.41422331e+00 -1.47906944e-01
-5.72172999e-01 5.95138013e-01 1.52957678e-01 -1.02652229e-01
6.14382684e-01 -2.45883152e-01 -3.74087483e-01 -7.04742193e-01
-4.61589813e-01 -1.12981290e-01 6.01833403e-01 1.99730337e-01
8.52056146e-01 -1.33129227e+00 -8.24621856e-01 2.38273427e-01
-3.79023165e-01 2.69883245e-01 9.49173331e-01 8.28046024e-01
-5.80701172e-01 -3.21674794e-02 -5.16044915e-01 -2.32324153e-01
-8.46287727e-01 5.17241538e-01 3.96954566e-01 -5.82150519e-01
-1.06341219e+00 6.43667102e-01 5.78864276e-01 -2.06722021e-01
3.34149972e-02 -1.48789689e-01 -1.56260148e-01 -4.28210527e-01
7.37568915e-01 4.82297927e-01 -1.82257697e-01 -6.04236662e-01
2.24060103e-01 6.81347966e-01 7.56809711e-02 4.70854789e-02
1.61822069e+00 -5.04526138e-01 -1.61520228e-01 -8.05275515e-02
1.37105656e+00 -9.30807590e-02 -1.80175459e+00 -2.65100449e-01
-7.73433030e-01 -6.78186655e-01 4.49004292e-01 -8.75074923e-01
-1.69641495e+00 6.04720175e-01 8.70718002e-01 2.93770805e-02
1.90492189e+00 -1.22764260e-02 1.43042481e+00 -5.80387235e-01
1.78729221e-01 -7.59607255e-01 2.58080572e-01 1.64402723e-01
1.04363751e+00 -9.89430070e-01 -1.55526072e-01 -4.51001883e-01
-6.37441218e-01 9.59294200e-01 5.28971016e-01 -2.98571110e-01
1.80100635e-01 2.01547325e-01 -1.18402988e-01 8.70524943e-02
-5.51041245e-01 -1.64033979e-01 3.39801073e-01 6.04360104e-01
4.35050339e-01 -1.89167246e-01 -2.96512604e-01 3.92152011e-01
-1.75729375e-02 6.20179921e-02 4.24191147e-01 3.81843746e-01
-2.79757231e-01 -8.15285325e-01 -2.20961735e-01 3.83067340e-01
-5.59787691e-01 -1.69630572e-01 3.03452075e-01 7.48365223e-01
2.37903550e-01 8.08310211e-01 8.48023891e-02 -4.56469297e-01
3.36235881e-01 -5.93767107e-01 4.73928928e-01 -1.57003477e-01
-4.48614955e-01 3.38304818e-01 -1.02689795e-01 -8.09862375e-01
-4.48166728e-01 -8.16680118e-02 -1.04801822e+00 -4.61055994e-01
5.40755652e-02 -3.34804386e-01 4.83326167e-01 6.34942889e-01
8.78448114e-02 9.27752793e-01 1.02387643e+00 -9.12680864e-01
-1.68604270e-01 -6.51471257e-01 -8.95535827e-01 3.80385518e-01
6.86722398e-01 -1.49452806e-01 -4.89444345e-01 5.50654791e-02] | [11.057978630065918, -1.9655081033706665] |
4694dac5-94a7-4cb1-ba28-e37f72eff4cc | regular-splitting-graph-network-for-3d-human | 2305.05785 | null | https://arxiv.org/abs/2305.05785v1 | https://arxiv.org/pdf/2305.05785v1.pdf | Regular Splitting Graph Network for 3D Human Pose Estimation | In human pose estimation methods based on graph convolutional architectures, the human skeleton is usually modeled as an undirected graph whose nodes are body joints and edges are connections between neighboring joints. However, most of these methods tend to focus on learning relationships between body joints of the skeleton using first-order neighbors, ignoring higher-order neighbors and hence limiting their ability to exploit relationships between distant joints. In this paper, we introduce a higher-order regular splitting graph network (RS-Net) for 2D-to-3D human pose estimation using matrix splitting in conjunction with weight and adjacency modulation. The core idea is to capture long-range dependencies between body joints using multi-hop neighborhoods and also to learn different modulation vectors for different body joints as well as a modulation matrix added to the adjacency matrix associated to the skeleton. This learnable modulation matrix helps adjust the graph structure by adding extra graph edges in an effort to learn additional connections between body joints. Instead of using a shared weight matrix for all neighboring body joints, the proposed RS-Net model applies weight unsharing before aggregating the feature vectors associated to the joints in order to capture the different relations between them. Experiments and ablations studies performed on two benchmark datasets demonstrate the effectiveness of our model, achieving superior performance over recent state-of-the-art methods for 3D human pose estimation. | ['A. Ben Hamza', 'Tanvir Hassan'] | 2023-05-09 | null | null | null | null | ['3d-human-pose-estimation'] | ['computer-vision'] | [-6.44057691e-02 4.98909622e-01 -1.70499489e-01 -1.24038823e-01
2.24393696e-01 -3.25862348e-01 3.20708424e-01 1.53383791e-01
-4.28243190e-01 2.41987213e-01 5.04722774e-01 4.21779841e-01
-2.95369416e-01 -7.28377819e-01 -6.28711939e-01 -4.28675473e-01
-7.60633230e-01 7.37002075e-01 4.91240650e-01 -5.48559487e-01
-2.18923792e-01 5.57116449e-01 -1.00813484e+00 -1.32390916e-01
4.11731750e-01 5.50887704e-01 -3.26638371e-01 4.54726368e-01
3.33560318e-01 4.05065924e-01 -5.60129821e-01 -3.06626409e-01
5.16152620e-01 -3.69747192e-01 -6.61570430e-01 3.06848496e-01
5.34327149e-01 -1.27859518e-01 -7.59144306e-01 6.44003987e-01
6.07898116e-01 3.16380441e-01 6.31777287e-01 -1.13828957e+00
-2.46719867e-01 5.69830954e-01 -7.71625757e-01 -1.99961942e-02
4.21465993e-01 2.40926728e-01 1.30950212e+00 -4.00446028e-01
8.33495975e-01 1.42434990e+00 8.97612393e-01 5.25514305e-01
-1.42826343e+00 -3.87229949e-01 2.68919647e-01 2.39097878e-01
-1.35756624e+00 3.53127792e-02 1.09158003e+00 -3.24555874e-01
9.80001032e-01 4.27071825e-02 1.22061265e+00 9.82653141e-01
4.99373496e-01 4.98657614e-01 3.76642346e-01 -1.93845287e-01
-1.59019440e-01 -7.86150336e-01 1.85073316e-01 1.08842051e+00
3.69957358e-01 -1.99067131e-01 -7.32946217e-01 -1.21354461e-01
1.00798643e+00 -9.45141241e-02 -1.88945353e-01 -1.06641948e+00
-1.36569321e+00 8.31272066e-01 1.08718967e+00 1.87949970e-01
-4.12416190e-01 4.33602571e-01 4.32921827e-01 1.41626373e-01
2.91379362e-01 3.77268016e-01 -3.07292968e-01 3.03913385e-01
-5.67841291e-01 4.06703830e-01 7.99570799e-01 4.80679780e-01
8.17256451e-01 -1.31455958e-01 -3.79308134e-01 7.78006017e-01
5.58136225e-01 1.26250342e-01 2.54884779e-01 -8.26616108e-01
7.19506383e-01 8.88798356e-01 -2.57922530e-01 -1.42641532e+00
-1.03976166e+00 -6.77654564e-01 -1.02723992e+00 1.21376731e-01
5.04528522e-01 -1.97417945e-01 -1.19005620e+00 1.87705958e+00
6.11715138e-01 2.46317014e-01 -5.66269100e-01 1.06953669e+00
6.17275476e-01 1.46871641e-01 -1.32334679e-01 4.19659287e-01
1.35389471e+00 -9.58405197e-01 -4.01011139e-01 -4.92837191e-01
5.53753495e-01 -4.79541004e-01 5.77500463e-01 1.12149324e-02
-9.36749220e-01 -6.41200423e-01 -1.19384885e+00 -2.23005250e-01
-1.23379432e-01 1.58473663e-02 5.23065686e-01 2.79885799e-01
-8.81668389e-01 8.05747688e-01 -1.14156842e+00 -2.91530401e-01
4.08704281e-02 7.26464152e-01 -7.35681772e-01 1.53504491e-01
-1.46457469e+00 8.60502481e-01 1.99555427e-01 4.74092335e-01
-3.81575644e-01 -4.14945483e-01 -1.24008083e+00 -1.84221193e-01
4.80637163e-01 -1.20494080e+00 5.02289593e-01 -5.53313196e-01
-1.59703076e+00 6.44199967e-01 2.46439978e-01 -5.43272257e-01
4.84421909e-01 -4.13344026e-01 -1.66381761e-01 3.01624030e-01
-8.16021562e-02 8.72752905e-01 1.04982400e+00 -8.97216201e-01
-1.30776942e-01 -6.61856055e-01 1.30478323e-01 5.25265276e-01
-3.03551346e-01 -6.83119059e-01 -8.89735937e-01 -8.91387880e-01
5.60804486e-01 -1.48810041e+00 -4.66256857e-01 2.89245341e-02
-7.21382499e-01 -1.22366704e-01 6.80594087e-01 -7.66367614e-01
1.38946509e+00 -1.85544956e+00 9.45694625e-01 8.25574934e-01
5.83544254e-01 3.89769040e-02 -3.98861319e-01 4.28274572e-01
-4.39828560e-02 -4.42207813e-01 -2.02969238e-01 -2.10442811e-01
-8.88416842e-02 4.52122599e-01 5.61509132e-01 7.56225824e-01
2.73037612e-01 8.79649520e-01 -8.34311604e-01 -3.15296412e-01
4.06422138e-01 9.36545253e-01 -7.37267852e-01 -7.66526163e-02
6.34461120e-02 4.89405215e-01 -4.62389290e-01 1.58808395e-01
4.08931553e-01 -2.33507663e-01 4.77918625e-01 -6.67941630e-01
4.17279005e-01 3.59323084e-01 -1.56248319e+00 2.04617500e+00
-2.79145658e-01 5.51480353e-02 6.92305639e-02 -1.03675723e+00
7.80099928e-01 1.04661606e-01 8.36372554e-01 -1.94202736e-01
2.37240344e-01 -1.21932700e-01 4.81644511e-01 -7.22965673e-02
1.19438201e-01 3.14737231e-01 -1.78382471e-01 1.61389276e-01
2.63248473e-01 4.04830985e-02 3.81351113e-01 2.30846599e-01
1.24010122e+00 2.67394871e-01 1.28457829e-01 -2.50502467e-01
6.85420156e-01 -4.34441775e-01 5.51527143e-01 2.21904665e-01
-2.67114211e-02 6.05324745e-01 5.92642546e-01 -6.17129982e-01
-7.71844029e-01 -1.27428436e+00 2.92203933e-01 8.66018236e-01
1.65559083e-01 -7.73101032e-01 -7.87092447e-01 -6.71943784e-01
2.90378362e-01 -9.57115442e-02 -9.39920366e-01 -5.24718046e-01
-1.01304710e+00 -5.02423227e-01 4.27950531e-01 4.61463660e-01
5.53204119e-01 -6.58285022e-01 -7.22372532e-01 2.95634508e-01
-2.72566944e-01 -9.61906135e-01 -7.37169266e-01 3.91607374e-01
-9.68856931e-01 -1.19666612e+00 -7.70672917e-01 -6.13211393e-01
7.33539462e-01 -8.06632265e-02 1.00769448e+00 1.73416823e-01
-5.14515996e-01 4.21402961e-01 -1.20468505e-01 2.27796853e-01
6.08952716e-02 4.72479820e-01 -4.26369254e-03 4.69172932e-02
-1.97698995e-02 -8.08586121e-01 -7.24890232e-01 3.33253086e-01
-7.93365300e-01 -9.53596234e-02 6.33811951e-01 9.93365467e-01
3.97897810e-01 -1.98663831e-01 -1.73953474e-02 -6.64761961e-01
3.73007536e-01 -2.27992103e-01 -1.63990140e-01 5.76453432e-02
-1.21185683e-01 4.79582489e-01 3.28832895e-01 -4.74045634e-01
-4.17103022e-01 4.16584104e-01 -8.81612599e-02 -3.24492604e-01
2.27336794e-01 4.79193002e-01 -2.26426050e-01 -1.82006493e-01
5.40683746e-01 -2.44114742e-01 3.38654757e-01 -3.90856624e-01
5.12394428e-01 -1.56865463e-01 3.78833979e-01 -4.20834661e-01
1.03462529e+00 3.54135334e-01 7.06132174e-01 -8.16900611e-01
-6.31155312e-01 -4.64243174e-01 -1.23528826e+00 -2.93097377e-01
1.04302049e+00 -7.87491739e-01 -6.04576707e-01 4.03052002e-01
-8.43133748e-01 -3.28489512e-01 -2.29729772e-01 6.11456215e-01
-3.77380461e-01 6.18703306e-01 -8.65064859e-01 -2.40331069e-01
-2.18310446e-01 -9.96440768e-01 1.23677933e+00 -2.45398849e-01
-6.68529510e-01 -1.08337843e+00 3.28871936e-01 1.95048839e-01
-1.46098603e-02 4.62388009e-01 8.45271528e-01 -1.97612196e-01
-1.45181760e-01 -3.12445313e-01 1.98270649e-01 2.08696440e-01
2.36200079e-01 -1.57822028e-01 -2.60962129e-01 -4.26685095e-01
-6.09738290e-01 -7.78134912e-02 1.09897685e+00 4.56506521e-01
5.99606216e-01 -1.10681787e-01 -4.15386975e-01 5.72812319e-01
8.28376949e-01 -6.25419259e-01 5.50457001e-01 2.13960320e-01
1.32394731e+00 7.03309715e-01 7.67596737e-02 3.98550123e-01
3.64173114e-01 1.00089383e+00 6.29264474e-01 -9.04300883e-02
-4.27736908e-01 -2.69431502e-01 4.02340144e-01 8.43230546e-01
-5.69014370e-01 -3.97743098e-02 -8.38349938e-01 3.20045650e-01
-1.94326627e+00 -5.68377256e-01 -1.98440939e-01 2.44986534e+00
5.77958643e-01 4.03576881e-01 4.20920670e-01 2.26488248e-01
6.73322439e-01 5.93033969e-01 -4.25623327e-01 1.07257500e-01
9.06059369e-02 3.80750597e-01 4.75272864e-01 5.49424648e-01
-1.23438120e+00 8.92174602e-01 5.53061962e+00 4.23633248e-01
-8.92902374e-01 -2.16254070e-01 -8.97033792e-03 -1.25992954e-01
-6.67383671e-02 -9.84886512e-02 -5.95778644e-01 3.03882845e-02
3.31804305e-01 4.73335534e-01 2.88856238e-01 3.62366796e-01
-2.01799944e-01 1.44417539e-01 -1.04269505e+00 7.25279033e-01
-1.81355998e-02 -7.80941308e-01 2.76012719e-01 1.21464677e-01
5.25025249e-01 3.00793024e-03 -7.76535794e-02 -6.86320215e-02
2.00801596e-01 -7.80569017e-01 5.39988339e-01 3.57287407e-01
2.21805379e-01 -8.68902504e-01 6.19576037e-01 1.65608693e-02
-1.60357666e+00 1.52983829e-01 -3.03565919e-01 -1.76707879e-01
1.36059210e-01 6.57785356e-01 -6.14576280e-01 6.75452650e-01
6.78988636e-01 9.28765953e-01 -6.28855228e-01 7.94237554e-01
-4.85667139e-01 2.43579328e-01 -6.20350718e-01 1.60685450e-01
2.78642774e-01 -2.66606927e-01 7.90771425e-01 8.99951100e-01
4.16394975e-03 -2.28076279e-01 3.73969704e-01 4.65605736e-01
-6.82865456e-02 4.07259539e-02 -4.03263897e-01 1.60262048e-01
-4.50434200e-02 1.34886932e+00 -8.81241679e-01 -5.92665635e-02
-2.77822495e-01 1.29289460e+00 5.55661142e-01 2.87466824e-01
-8.21971059e-01 -3.92760903e-01 8.95900726e-01 5.17845571e-01
2.93970764e-01 -7.57808983e-01 1.43958762e-01 -1.01272881e+00
1.45622402e-01 -6.85292065e-01 6.16900027e-01 -1.87798917e-01
-1.23590851e+00 3.15373272e-01 5.05498759e-02 -1.17740822e+00
-2.58376598e-01 -5.38370490e-01 -3.03220630e-01 7.50133395e-01
-9.53929782e-01 -1.39349270e+00 -1.53784722e-01 8.81831288e-01
1.42530009e-01 2.76388496e-01 6.62936389e-01 1.44741684e-01
-4.25670922e-01 5.83419025e-01 -5.34302294e-01 4.76649076e-01
7.30897903e-01 -1.18476391e+00 6.26163483e-01 5.11273682e-01
3.50565702e-01 8.22198153e-01 6.52707696e-01 -9.28055704e-01
-1.34183574e+00 -9.76266265e-01 7.36510754e-01 -1.80575997e-01
6.72010064e-01 -5.54377675e-01 -7.79455245e-01 7.70894945e-01
-3.99950035e-02 3.31973284e-01 4.32836771e-01 3.67905259e-01
-5.11010826e-01 -9.81273800e-02 -6.75179780e-01 7.29895413e-01
1.46187830e+00 -3.04761618e-01 -5.43354511e-01 2.47321963e-01
3.52454841e-01 -5.32913029e-01 -9.25658643e-01 6.52622283e-01
8.91491234e-01 -7.97721148e-01 1.39755297e+00 -5.25364161e-01
2.46100307e-01 -4.64327008e-01 2.84509301e-01 -1.45275545e+00
-6.31257415e-01 -4.10433799e-01 -2.95257151e-01 5.68454981e-01
3.84481877e-01 -3.94609630e-01 1.09039712e+00 2.32049450e-01
-1.31598599e-02 -6.95322871e-01 -1.10450590e+00 -7.54277885e-01
-2.20700294e-01 -1.63141847e-01 2.41408899e-01 7.97296584e-01
-1.05932184e-01 7.65645444e-01 -7.54821241e-01 3.26330483e-01
8.10784161e-01 -1.93440855e-01 9.38756287e-01 -1.35710084e+00
-6.00111246e-01 -3.51828396e-01 -9.82325792e-01 -1.21270943e+00
1.31556377e-01 -8.92033696e-01 -4.08292711e-02 -1.55921149e+00
-5.79568744e-02 -1.16580151e-01 -2.74264812e-01 5.35278797e-01
-1.91475853e-01 5.60803950e-01 2.46106952e-01 -1.05453521e-01
-4.26178604e-01 5.71612179e-01 1.43106151e+00 -3.01518291e-01
-3.90834957e-01 3.01939286e-02 6.62552612e-03 9.20739651e-01
5.37616551e-01 -3.76487285e-01 -4.34723169e-01 -5.05592823e-01
3.00459415e-01 -1.07950605e-01 6.49988353e-01 -1.33051312e+00
2.11360335e-01 3.65427494e-01 6.45271778e-01 -4.66615230e-01
5.33600807e-01 -7.90464282e-01 3.70305479e-01 8.90983880e-01
-2.07591876e-01 -1.15822461e-02 -1.16337940e-01 6.83922887e-01
-1.97773680e-01 1.90860778e-01 5.84861517e-01 -2.56382108e-01
-5.67850173e-01 4.32985693e-01 -2.18455538e-01 -5.52711077e-02
7.03805983e-01 -2.53567398e-01 2.58064777e-01 -3.61260146e-01
-1.30111217e+00 2.24936843e-01 3.82640928e-01 6.34789050e-01
5.38745701e-01 -1.47583961e+00 -5.30740559e-01 2.21199617e-01
9.03367996e-02 3.18851322e-02 3.13757151e-01 9.16460633e-01
-5.28073430e-01 1.74301520e-01 -5.94013810e-01 -7.35548913e-01
-1.35259247e+00 1.41539678e-01 4.37749594e-01 -7.14597940e-01
-8.07292938e-01 1.07800484e+00 1.17009655e-01 -6.58340573e-01
2.14403152e-01 -4.04145569e-01 -3.14164668e-01 6.92803711e-02
-1.27023414e-01 5.00824809e-01 5.01321182e-02 -9.81330812e-01
-5.30879855e-01 9.40022469e-01 2.05799729e-01 2.07579639e-02
1.17913640e+00 7.07551837e-03 -1.48392633e-01 3.39479178e-01
1.28170228e+00 -5.09882122e-02 -1.19314742e+00 -2.78807610e-01
-1.57210995e-02 -2.84020483e-01 -7.07408190e-02 -3.29604179e-01
-1.40095067e+00 7.78953075e-01 4.74796146e-01 -1.38014242e-01
8.80150020e-01 5.67622948e-03 8.56186271e-01 5.08961797e-01
3.71227741e-01 -1.20431674e+00 4.58210468e-01 5.69458008e-01
9.36149836e-01 -9.34294224e-01 3.82213324e-01 -7.05597579e-01
-3.13073754e-01 1.21001267e+00 5.62762678e-01 -6.47108257e-01
8.44089329e-01 -1.08942538e-01 -7.10798427e-02 -3.73964339e-01
-1.95950866e-01 -3.34170133e-01 9.04388547e-01 5.00537097e-01
6.05471909e-01 -3.32738012e-02 -4.43858653e-01 1.62714019e-01
-1.77472353e-01 -4.51302648e-01 3.99022084e-03 7.94009089e-01
-2.61592478e-01 -1.50952649e+00 -3.00002664e-01 4.60358500e-01
-3.19237709e-01 1.59489945e-01 -5.88280797e-01 1.11633790e+00
2.32035905e-01 6.03173614e-01 8.58846456e-02 -5.76037049e-01
6.28889561e-01 -6.42557964e-02 8.49367142e-01 -7.48326957e-01
-6.89012587e-01 3.17095399e-01 1.87945589e-01 -8.39426577e-01
-6.03662550e-01 -6.22082710e-01 -1.45292294e+00 -1.69810466e-02
-1.83016196e-01 -6.89001232e-02 2.60290205e-01 7.60313928e-01
3.96642894e-01 7.53989100e-01 1.19804129e-01 -1.24321961e+00
-3.13676894e-01 -8.92838418e-01 -7.73202300e-01 7.05447018e-01
2.52316892e-01 -1.10726368e+00 -1.43333435e-01 -2.45973736e-01] | [7.093308925628662, -0.5873224139213562] |
50c754b9-9964-4bae-910e-f731bc68cf79 | contrastive-transformer-based-multiple | 2203.12121 | null | https://arxiv.org/abs/2203.12121v2 | https://arxiv.org/pdf/2203.12121v2.pdf | Contrastive Transformer-based Multiple Instance Learning for Weakly Supervised Polyp Frame Detection | Current polyp detection methods from colonoscopy videos use exclusively normal (i.e., healthy) training images, which i) ignore the importance of temporal information in consecutive video frames, and ii) lack knowledge about the polyps. Consequently, they often have high detection errors, especially on challenging polyp cases (e.g., small, flat, or partially visible polyps). In this work, we formulate polyp detection as a weakly-supervised anomaly detection task that uses video-level labelled training data to detect frame-level polyps. In particular, we propose a novel convolutional transformer-based multiple instance learning method designed to identify abnormal frames (i.e., frames with polyps) from anomalous videos (i.e., videos containing at least one frame with polyp). In our method, local and global temporal dependencies are seamlessly captured while we simultaneously optimise video and snippet-level anomaly scores. A contrastive snippet mining method is also proposed to enable an effective modelling of the challenging polyp cases. The resulting method achieves a detection accuracy that is substantially better than current state-of-the-art approaches on a new large-scale colonoscopy video dataset introduced in this work. | ['Gustavo Carneiro', 'Johan W Verjans', 'Yuanhong Chen', 'Chong Wang', 'Yuyuan Liu', 'Fengbei Liu', 'Guansong Pang', 'Yu Tian'] | 2022-03-23 | null | null | null | null | ['supervised-anomaly-detection'] | ['computer-vision'] | [ 6.09911501e-01 7.12532997e-02 -1.93281978e-01 -5.50949499e-02
-7.36654937e-01 -4.34254557e-01 1.95758775e-01 7.87956536e-01
-3.73353064e-01 1.91621587e-01 -3.88763566e-03 -2.24058568e-01
-1.75402254e-01 -6.29338801e-01 -9.83952165e-01 -7.74370074e-01
-4.51248616e-01 -1.09901861e-03 5.13919175e-01 1.74179614e-01
1.22324012e-01 9.03866626e-03 -1.59209061e+00 7.35172331e-01
1.00346291e+00 1.13759673e+00 1.31624892e-01 1.06963372e+00
2.75369138e-01 1.03583717e+00 -3.13448429e-01 -2.23142341e-01
2.50960678e-01 -4.57000792e-01 -4.66786176e-01 3.72784257e-01
7.15950727e-01 -4.04469460e-01 -3.71816218e-01 1.13961804e+00
1.30334944e-01 -1.16346158e-01 2.81286925e-01 -7.34942675e-01
1.79594204e-01 1.69170499e-01 -5.10147214e-01 9.66561139e-01
3.86787832e-01 1.88692659e-01 7.52584517e-01 -6.37743413e-01
4.86510903e-01 6.38296723e-01 7.18814194e-01 2.67277688e-01
-1.09215915e+00 -8.64715725e-02 5.51603794e-01 2.56893098e-01
-9.15094197e-01 -1.98833317e-01 5.19433498e-01 -6.21089220e-01
8.60985577e-01 5.28807402e-01 1.12773979e+00 1.19416487e+00
4.22911197e-01 9.45624232e-01 3.96570742e-01 -3.33945185e-01
-2.41979584e-02 -2.78635234e-01 -1.99681878e-01 8.97433341e-01
4.34040636e-01 9.89467502e-02 -3.35211545e-01 -2.99352944e-01
5.63517153e-01 4.76180822e-01 -2.54013866e-01 -5.05639791e-01
-1.70027435e+00 7.19282985e-01 2.24164143e-01 3.29044580e-01
-5.34353793e-01 -7.60276392e-02 6.76554978e-01 3.57923925e-01
2.49575779e-01 3.18381250e-01 -3.17433029e-01 -1.01644762e-01
-8.87854874e-01 9.78408903e-02 7.46138930e-01 6.13797247e-01
2.64343828e-01 -8.16304684e-02 -2.75626063e-01 4.97907102e-01
2.75854379e-01 -6.65497705e-02 8.87726665e-01 -5.47419310e-01
3.91259670e-01 8.56113553e-01 1.55160069e-01 -1.14357650e+00
-4.42669809e-01 -5.53457439e-01 -8.63945186e-01 -1.98615804e-01
4.82770562e-01 9.15218219e-02 -8.37447166e-01 1.29693222e+00
3.12477708e-01 7.42176533e-01 1.05860017e-01 7.44479775e-01
6.75773144e-01 2.26056054e-01 -1.36426777e-01 -2.96120226e-01
1.50186837e+00 -1.04657102e+00 -5.69490969e-01 -4.49771017e-01
8.15239191e-01 -6.21729195e-01 7.11246490e-01 6.45908475e-01
-1.12205124e+00 -3.99604142e-01 -1.12204063e+00 3.79145712e-01
-6.54913560e-02 4.09978151e-01 3.67254674e-01 4.15517777e-01
-6.23199642e-01 4.65472341e-01 -1.49657416e+00 -1.56208798e-01
2.97183037e-01 4.62715358e-01 -3.46397489e-01 -2.63014406e-01
-7.66058087e-01 4.72261280e-01 5.80961823e-01 7.13602751e-02
-9.00099099e-01 -7.28008747e-01 -1.17924571e+00 -2.26404414e-01
6.58080757e-01 -6.75911009e-01 1.39613187e+00 -1.40677989e+00
-8.15194249e-01 8.83325100e-01 -8.14760625e-02 -9.16399300e-01
7.23162830e-01 -3.57036054e-01 -6.09885454e-01 6.25501037e-01
-7.44212419e-02 2.90997505e-01 1.13328969e+00 -7.26511717e-01
-1.01976383e+00 -1.51330277e-01 2.84413189e-01 1.34269908e-01
-4.39874560e-01 -2.14185789e-01 -4.54677284e-01 -9.23782647e-01
1.68727338e-01 -9.22882020e-01 -4.48218763e-01 2.59284735e-01
-2.48643532e-01 2.63287455e-01 7.82522202e-01 -8.31908762e-01
1.61766529e+00 -2.41324449e+00 9.11371782e-02 2.54630506e-01
4.99474049e-01 4.38044250e-01 9.82935429e-02 -9.10124704e-02
-1.39257118e-01 -8.15344080e-02 -2.41959527e-01 3.15582901e-02
-5.84050238e-01 4.32240725e-01 4.82363731e-01 6.81936800e-01
5.09914041e-01 5.45016170e-01 -1.33201039e+00 -6.07847750e-01
5.36276639e-01 9.68818963e-02 -9.43293273e-01 1.57967821e-01
-2.14954138e-01 8.20644081e-01 -3.31317872e-01 8.81917357e-01
2.52958208e-01 -4.71541911e-01 8.13802928e-02 -3.17899019e-01
-2.48952836e-01 -9.71637368e-02 -9.91903961e-01 1.97550189e+00
-3.85167658e-01 4.08013701e-01 -5.34901135e-02 -1.15967870e+00
3.22851777e-01 5.86043358e-01 8.49328339e-01 -4.53311801e-01
-3.03465705e-02 4.98123169e-01 5.74880898e-01 -9.71622467e-01
7.63214678e-02 2.90699691e-01 2.16291398e-01 -1.59032330e-01
-1.06653430e-01 4.75159943e-01 6.03351414e-01 -2.79613473e-02
1.64024627e+00 1.87001415e-02 8.10424149e-01 -1.62946180e-01
8.91958773e-01 2.33236942e-02 7.96501458e-01 8.44545245e-01
-4.82959688e-01 7.87433505e-01 5.93156934e-01 -8.36018801e-01
-8.90089750e-01 -7.97155023e-01 -1.35346338e-01 5.14933169e-01
3.55433136e-01 -4.56570864e-01 -3.05640459e-01 -8.48280430e-01
-1.13526657e-01 -2.62506902e-02 -6.55622423e-01 -3.04882646e-01
-9.17564690e-01 -9.41000640e-01 1.90674543e-01 3.55483562e-01
1.82325885e-01 -8.28556418e-01 -1.07792377e+00 5.50911784e-01
-3.42789590e-01 -1.06618071e+00 -4.44605589e-01 -2.00179685e-02
-9.39410388e-01 -1.60289288e+00 -6.59694016e-01 -9.08163249e-01
8.12856615e-01 2.46182352e-01 1.18179464e+00 4.11273986e-01
-7.92230904e-01 4.63552654e-01 -5.49366295e-01 -3.04453939e-01
-4.79282141e-01 -2.75541723e-01 -1.92256346e-01 3.88060898e-01
2.84916669e-01 -3.34199101e-01 -1.16888821e+00 4.40532774e-01
-1.11797845e+00 -6.50914013e-02 8.83724689e-01 1.24738634e+00
1.05705822e+00 -7.17232227e-02 2.09755570e-01 -6.89918876e-01
7.41455778e-02 -7.23071873e-01 -5.10013759e-01 2.12928548e-01
-4.83868159e-02 -2.94180900e-01 6.33171737e-01 -7.05376744e-01
-6.09369278e-01 2.94155419e-01 2.17327606e-02 -6.68680906e-01
-3.19945395e-01 5.65331399e-01 5.02396405e-01 -1.49091899e-01
7.57148266e-01 2.57112265e-01 4.43532728e-02 -8.78508165e-02
-4.39176351e-01 2.22656563e-01 6.45014107e-01 -1.15991317e-01
3.39253545e-01 6.60606027e-01 -1.30335331e-01 -8.20906341e-01
-7.80611217e-01 -1.11392248e+00 -5.71395695e-01 -2.82061875e-01
8.19553852e-01 -1.12441444e+00 -2.03686342e-01 4.85025942e-01
-6.46171391e-01 3.14539149e-02 -3.93035829e-01 8.97614360e-01
-4.94509041e-01 6.30436659e-01 -6.23608351e-01 -5.96049964e-01
-1.77496597e-01 -1.21397734e+00 1.16557276e+00 1.64554715e-02
-2.05817387e-01 -9.24911320e-01 3.09968963e-02 -4.64859791e-02
1.31925389e-01 7.47890651e-01 6.17938578e-01 -9.69924092e-01
-7.30453014e-01 -4.76853937e-01 -5.43906586e-03 3.13435346e-01
2.06214577e-01 1.46255538e-01 -5.97055912e-01 -4.61930156e-01
3.09131742e-02 6.12552762e-02 9.17403758e-01 6.84454918e-01
1.37623632e+00 -4.37440097e-01 -4.82331872e-01 6.44335806e-01
1.15327883e+00 2.58347750e-01 4.38865691e-01 4.51796532e-01
4.59888697e-01 2.60018319e-01 8.74934912e-01 5.84079027e-01
-1.97916105e-03 5.62950015e-01 8.67781401e-01 -1.50461003e-01
4.30947468e-02 1.54544115e-01 3.38640988e-01 6.92988992e-01
-7.88340196e-02 -2.00866297e-01 -8.91582191e-01 8.35646093e-01
-2.15463614e+00 -8.47938120e-01 -1.29300788e-01 2.25210786e+00
3.72373343e-01 2.35632271e-01 1.38274804e-01 2.69550502e-01
6.81023300e-01 -7.38418847e-03 -6.56753838e-01 8.81522596e-02
3.27029899e-02 -1.22794613e-01 2.54819930e-01 5.11071049e-02
-1.78273964e+00 1.12506613e-01 4.82298470e+00 3.90345663e-01
-9.19311106e-01 8.80388357e-03 5.54807067e-01 -3.51471245e-01
4.13713604e-02 -3.62626880e-01 -3.27585995e-01 6.60994709e-01
6.62186861e-01 2.88240518e-02 -2.45905779e-02 7.68896699e-01
6.49338737e-02 -7.61599019e-02 -1.18838549e+00 9.39026535e-01
2.25762412e-01 -1.11928618e+00 -1.42846450e-01 -3.67788263e-02
6.21540308e-01 1.66946296e-02 6.40713796e-02 5.78139350e-02
-4.06631470e-01 -7.14565575e-01 5.73741972e-01 4.67581093e-01
5.49937189e-01 -5.08700371e-01 1.04999757e+00 2.65108198e-01
-1.32368970e+00 -4.42544132e-01 -4.48059151e-03 2.02864334e-01
-1.84019972e-02 7.39688039e-01 -6.08069301e-01 8.36501360e-01
8.31548154e-01 1.30824876e+00 -6.27595007e-01 1.68160367e+00
1.46071285e-01 3.66258115e-01 -4.46973860e-01 4.31367725e-01
3.85935038e-01 8.20369273e-02 9.57566500e-01 1.24323010e+00
7.59368122e-01 -1.07719213e-01 5.39958894e-01 2.09255517e-01
3.55186820e-01 1.46058977e-01 -7.02403307e-01 1.53771490e-01
-1.24509580e-01 1.21541095e+00 -7.63977289e-01 -3.46838355e-01
-1.04590821e+00 8.41131508e-01 -2.11212844e-01 -5.65210879e-02
-7.79474437e-01 1.45848259e-01 5.37055075e-01 9.26872641e-02
5.20958602e-01 2.20921680e-01 3.55326533e-01 -1.57271385e+00
3.82016629e-01 -9.22776580e-01 8.69871378e-01 -2.88907229e-03
-1.26794863e+00 4.51799750e-01 -3.74887496e-01 -2.04137516e+00
-5.13869107e-01 -6.63037419e-01 -7.69353926e-01 -9.87838395e-03
-1.63853204e+00 -8.79541934e-01 -5.89592755e-01 7.17960715e-01
6.94058836e-01 -3.50881666e-02 7.57700622e-01 5.05703032e-01
-4.26012814e-01 5.36496758e-01 -1.41371474e-01 1.91155151e-01
5.27720451e-01 -1.37534952e+00 1.15504555e-01 1.25276935e+00
5.88783175e-02 1.26973659e-01 4.78099406e-01 -5.65698266e-01
-1.25158942e+00 -1.26436758e+00 4.04867262e-01 -2.60761142e-01
6.99099362e-01 -2.94441674e-02 -8.29952061e-01 6.94635391e-01
-2.49263361e-01 6.56748056e-01 7.53277898e-01 -4.29558098e-01
6.08818084e-02 1.62275076e-01 -1.11557198e+00 4.47188735e-01
9.47706401e-01 -2.29754094e-02 -4.89145964e-01 5.60264707e-01
5.02510905e-01 -9.44104195e-01 -8.54459465e-01 8.11138988e-01
6.25705957e-01 -1.18743396e+00 1.18428946e+00 -4.45026636e-01
3.87098193e-01 -2.63048708e-01 9.59573612e-02 -1.02814186e+00
-9.69302058e-02 -4.65726048e-01 -8.10638189e-01 1.75684690e-01
3.97718638e-01 -3.73634368e-01 7.34132349e-01 -6.79164603e-02
-5.97099543e-01 -8.55461895e-01 -1.00272632e+00 -5.05866289e-01
-6.34645224e-01 -1.08924299e-01 -1.80816259e-02 8.03482413e-01
8.21979567e-02 -5.44797003e-01 -2.44328618e-01 4.58145559e-01
3.57465208e-01 2.39243731e-02 2.98470259e-01 -1.18337047e+00
-4.70161527e-01 -3.17695141e-01 -9.80150163e-01 -7.04353809e-01
-3.68879318e-01 -7.51665473e-01 -2.37685964e-02 -1.19832194e+00
1.23910867e-01 -5.06792730e-03 -6.25342071e-01 2.02160373e-01
-4.13107663e-01 1.56297490e-01 -2.83590734e-01 1.74976096e-01
-7.71313548e-01 -8.11450109e-02 1.12470841e+00 -2.55292207e-01
-2.31703788e-01 5.85249782e-01 -2.47620642e-01 1.04644132e+00
6.91887259e-01 -3.32983881e-01 -2.38087669e-01 -2.01682776e-01
2.75030673e-01 1.69452146e-01 7.34390318e-01 -1.25910592e+00
1.44612610e-01 2.26700306e-01 4.21930790e-01 -4.52620089e-01
1.10941038e-01 -8.82318020e-01 -1.09783001e-01 1.13154876e+00
-2.39592060e-01 2.57561237e-01 2.00457871e-01 1.12250459e+00
-7.02946961e-01 -3.16433549e-01 6.74820006e-01 -6.35441303e-01
-7.93594360e-01 5.21451950e-01 -4.44335550e-01 5.29026100e-03
1.31407428e+00 -2.17487589e-01 -2.40411870e-02 -1.88022964e-02
-1.13785982e+00 2.14482561e-01 3.06001008e-01 4.23968881e-01
6.62745595e-01 -1.12884390e+00 -8.37055504e-01 4.95306045e-01
6.37002647e-01 3.60105753e-01 5.70955157e-01 1.48847544e+00
-6.30458951e-01 2.71938920e-01 1.41815230e-01 -1.21367252e+00
-1.46033704e+00 5.82544148e-01 5.49120247e-01 -7.51440048e-01
-1.09427285e+00 7.17546999e-01 2.78648496e-01 7.62633085e-02
2.04930514e-01 -8.17853570e-01 -3.49337131e-01 8.21608528e-02
7.93241024e-01 9.02669784e-03 4.82078731e-01 -2.73052365e-01
-2.14164957e-01 2.66476363e-01 -4.56756353e-01 5.32814920e-01
9.97225642e-01 1.32149413e-01 1.48813710e-01 3.19588453e-01
9.04737175e-01 -3.24841328e-02 -1.33113694e+00 -3.48185927e-01
4.65388745e-02 -6.86803818e-01 -9.30607840e-02 -4.78866398e-01
-1.05729914e+00 6.18278146e-01 8.54953766e-01 2.60416240e-01
1.14473999e+00 -2.02629790e-01 8.05295765e-01 1.95758715e-01
4.73217927e-02 -8.35801005e-01 3.57508689e-01 1.62645727e-01
5.49760938e-01 -1.50356662e+00 -8.95127952e-02 -3.52853030e-01
-4.27465796e-01 1.17853642e+00 6.31372631e-01 -2.13106290e-01
4.60573912e-01 1.40960276e-01 -1.59245089e-01 -2.09956050e-01
-7.01881647e-01 -5.53456545e-02 4.30512220e-01 4.26906109e-01
2.81087965e-01 -8.78049955e-02 -2.17772052e-02 4.60504293e-01
3.39784116e-01 1.88670978e-02 5.09962976e-01 1.20779967e+00
-4.15023863e-01 -7.33102381e-01 -2.35996172e-01 1.06494522e+00
-9.46941018e-01 -1.47052169e-01 2.33920366e-01 8.66327465e-01
3.06307137e-01 7.56639481e-01 2.35153824e-01 -8.18739459e-02
4.59692448e-01 -1.92124426e-01 4.75011915e-01 -5.44600248e-01
-6.73830152e-01 3.99227858e-01 6.64025620e-02 -8.46620977e-01
-7.69260764e-01 -9.39200819e-01 -8.19054067e-01 3.43090653e-01
-1.68522298e-01 -1.72947243e-01 1.35381609e-01 8.23682606e-01
2.16333479e-01 9.08600330e-01 3.46740514e-01 -6.68170869e-01
-1.95854768e-01 -5.84830940e-01 -2.31673986e-01 6.52914882e-01
9.97729838e-01 -5.39713562e-01 -4.99578148e-01 1.68471277e-01] | [14.013113021850586, -3.193646192550659] |
a05ac5ce-bdf2-4a7e-8aca-e6cbd3e2db0e | bait-barometer-for-information | 2206.07535 | null | https://arxiv.org/abs/2206.07535v2 | https://arxiv.org/pdf/2206.07535v2.pdf | BaIT: Barometer for Information Trustworthiness | This paper presents a new approach to the FNC-1 fake news classification task which involves employing pre-trained encoder models from similar NLP tasks, namely sentence similarity and natural language inference, and two neural network architectures using this approach are proposed. Methods in data augmentation are explored as a means of tackling class imbalance in the dataset, employing common pre-existing methods and proposing a method for sample generation in the under-represented class using a novel sentence negation algorithm. Comparable overall performance with existing baselines is achieved, while significantly increasing accuracy on an under-represented but nonetheless important class for FNC-1. | ['Callum Rhys Tilbury', 'Jeroen van Mourik', 'Oisín Nolan'] | 2022-06-15 | null | null | null | null | ['news-classification'] | ['natural-language-processing'] | [ 6.13324940e-01 7.00160325e-01 -6.36490464e-01 -8.32340360e-01
-1.09055269e+00 -1.37372613e-01 7.77187943e-01 6.29202306e-01
-6.61370039e-01 1.14676046e+00 5.94407439e-01 -4.15175021e-01
1.49294376e-01 -8.28288138e-01 -8.66055846e-01 -2.38523677e-01
2.09641472e-01 7.59161472e-01 2.77211983e-02 -8.71906400e-01
7.90628672e-01 2.31948450e-01 -1.33905327e+00 1.18361938e+00
7.29848862e-01 1.09383690e+00 -7.32959628e-01 7.83208609e-01
3.52172460e-03 1.46981144e+00 -1.27962458e+00 -9.80493665e-01
3.79154906e-02 -6.85596585e-01 -1.33634686e+00 -1.76312804e-01
9.35881734e-01 -3.44492257e-01 -2.54706830e-01 1.12286401e+00
2.10370734e-01 2.45510377e-02 7.46007681e-01 -9.95252132e-01
-7.65953779e-01 7.51878321e-01 -4.47297283e-02 5.11643827e-01
6.69292033e-01 -1.94300815e-01 9.74829495e-01 -6.77357793e-01
8.86847317e-01 1.40731680e+00 1.11096752e+00 8.55831444e-01
-1.08203065e+00 -3.04663926e-01 -2.98330247e-01 1.52700067e-01
-7.48869538e-01 -5.58603644e-01 6.49658978e-01 -1.75310209e-01
1.48855162e+00 2.86225885e-01 4.64030236e-01 1.52809453e+00
3.93825620e-01 9.27645147e-01 9.91209209e-01 -7.33213484e-01
2.36125037e-01 5.38819313e-01 1.97034419e-01 4.51715559e-01
3.16615969e-01 4.57174070e-02 -7.02400804e-01 -3.68820399e-01
4.43247147e-03 -3.87016416e-01 -3.14056844e-01 3.48818183e-01
-8.74311745e-01 1.39053488e+00 4.79501754e-01 4.70790178e-01
-2.33283207e-01 5.15996218e-02 1.05479658e+00 6.90867126e-01
1.25002348e+00 1.10262382e+00 -7.25600064e-01 -1.09476745e-01
-1.20368850e+00 8.67774785e-01 1.16771066e+00 1.02453732e+00
8.46571773e-02 6.89839795e-02 -4.86724526e-01 4.67285454e-01
-2.15914827e-02 1.95002332e-01 8.92441452e-01 -7.92917430e-01
8.71654272e-01 6.34832621e-01 1.10658124e-01 -1.27255189e+00
-2.19850048e-01 -4.40433711e-01 -6.01445496e-01 -1.64957613e-01
1.33931711e-01 2.15909615e-01 -8.88065040e-01 1.33765841e+00
-4.49341759e-02 -2.19584808e-01 7.10054576e-01 6.35063827e-01
8.97964478e-01 5.46821177e-01 -1.94550291e-01 -2.09701329e-01
1.27187073e+00 -1.06570041e+00 -9.00434077e-01 -3.64751756e-01
1.04890084e+00 -7.08065629e-01 8.31928492e-01 3.75777662e-01
-1.09597611e+00 -2.24527285e-01 -1.30666208e+00 -5.15667737e-01
-8.23442638e-01 1.32545188e-01 4.79872555e-01 7.07963884e-01
-9.23802316e-01 1.01191032e+00 -2.22038969e-01 -1.94931865e-01
8.15645814e-01 2.00989291e-01 -4.16640013e-01 -9.90782380e-02
-1.55391848e+00 1.48019016e+00 8.52024436e-01 -1.68596774e-01
-5.59366345e-01 -5.25480688e-01 -9.91461039e-01 -3.84203158e-02
1.65275838e-02 -3.93412590e-01 1.55448949e+00 -1.25065458e+00
-1.31771505e+00 9.22728658e-01 -6.54377565e-02 -1.26582897e+00
4.55167919e-01 -2.88316458e-01 -7.18557715e-01 3.65809828e-01
5.37721694e-01 5.97907126e-01 8.37278843e-01 -1.00848413e+00
-4.92923409e-01 -1.00561738e-01 2.70579532e-02 2.02890173e-01
-3.05004954e-01 1.50183082e-01 5.48325419e-01 -8.31209719e-01
-4.87494282e-02 -2.32637674e-01 -6.29771948e-02 -3.61329883e-01
-5.90087771e-01 -3.60249400e-01 1.03840971e+00 -7.12610722e-01
1.04181480e+00 -1.54883599e+00 -3.98450017e-01 -4.25682552e-02
9.30831954e-02 5.18282235e-01 -4.76503670e-02 4.02584851e-01
-2.06476510e-01 3.49686623e-01 -2.91096181e-01 -3.83415788e-01
-3.49029034e-01 2.16136962e-01 -6.36713505e-01 3.64228070e-01
8.83698463e-01 5.99179387e-01 -1.19681466e+00 -2.73217231e-01
-2.53846109e-01 1.64033830e-01 -6.82451189e-01 -9.47550386e-02
-4.43767101e-01 -5.89153133e-02 1.59432337e-01 6.49340928e-01
5.96947193e-01 -5.20024495e-03 -4.47220579e-02 -4.81278710e-02
2.75874376e-01 1.15034807e+00 -4.44130480e-01 1.22904313e+00
-1.65211365e-01 8.74184549e-01 -2.77921647e-01 -1.72314334e+00
1.06288838e+00 4.61903453e-01 -2.85501093e-01 -7.86594093e-01
4.56301779e-01 7.42403865e-01 -3.29891928e-02 -6.50322735e-01
1.02772129e+00 -6.66563094e-01 -1.84584692e-01 3.69562119e-01
4.87042695e-01 -3.31890523e-01 3.51946205e-01 4.10121262e-01
9.88599241e-01 -1.33903921e-01 4.64847028e-01 -4.36434746e-01
4.78400141e-01 6.01052523e-01 3.94307196e-01 1.02897978e+00
-3.74134600e-01 5.47629416e-01 7.93761134e-01 -8.38961601e-01
-1.06261718e+00 -3.14045727e-01 -1.67028576e-01 7.22713530e-01
-2.65621871e-01 -2.48984709e-01 -6.20456159e-01 -1.22606325e+00
5.23211062e-02 1.43220973e+00 -9.17488694e-01 -5.99934697e-01
-5.48450828e-01 -7.60773420e-01 1.06917870e+00 4.77885246e-01
4.69913930e-01 -1.10811496e+00 -3.89767259e-01 2.73025811e-01
-4.67728645e-01 -1.11421466e+00 -3.63065414e-02 4.89483774e-01
-1.02665162e+00 -9.55484807e-01 -2.35933751e-01 -9.50457394e-01
5.70133328e-01 -3.94119024e-02 1.60117650e+00 2.58080751e-01
-9.00683329e-02 -2.22314849e-01 -4.77671415e-01 -6.57808483e-01
-1.09328032e+00 -5.29851951e-02 2.83622928e-02 -4.03609633e-01
7.80313492e-01 1.71138763e-01 -1.13748349e-01 -3.99217159e-01
-8.03088367e-01 -1.47654312e-02 3.27226996e-01 1.30332041e+00
4.49852832e-02 -4.70232926e-02 1.04674077e+00 -1.48066509e+00
9.65660274e-01 -7.16014504e-01 -7.28590926e-03 6.78875446e-02
-7.16192245e-01 -1.19071387e-01 8.32852304e-01 -1.57432452e-01
-1.06457460e+00 -3.97002816e-01 -3.91943455e-01 7.23673478e-02
-1.34548873e-01 5.68611324e-01 2.90768087e-01 9.29410681e-02
1.20413256e+00 1.99041709e-01 3.62148970e-01 -1.58299401e-01
1.94063455e-01 1.07216418e+00 4.26469773e-01 -1.86221376e-01
3.37856084e-01 3.87944043e-01 -4.87186015e-01 -5.38084328e-01
-1.50184357e+00 -3.56293827e-01 -4.36465949e-01 1.07686333e-01
4.85242724e-01 -8.53036404e-01 -2.20994815e-01 3.53772908e-01
-1.54207110e+00 1.95008978e-01 -4.92758453e-01 4.23515230e-01
-4.56479698e-01 9.79909822e-02 -1.02042353e+00 -7.75757790e-01
-6.01517320e-01 -7.32431352e-01 1.16724670e+00 -2.48855036e-02
-3.67149651e-01 -1.01361561e+00 1.23553704e-02 7.02215314e-01
4.38285232e-01 1.97710901e-01 9.53807712e-01 -1.58220851e+00
4.53183874e-02 -8.07432950e-01 -1.80214509e-01 8.24028373e-01
-2.63831496e-01 -3.83637846e-01 -8.99596691e-01 -2.98052549e-01
4.40017521e-01 -9.96094167e-01 1.03727341e+00 3.87150794e-02
9.18573678e-01 -8.57261002e-01 -2.33513176e-01 -5.38673289e-02
1.28990495e+00 1.31794468e-01 6.28726780e-01 5.54004788e-01
2.10721761e-01 6.62864506e-01 6.14802599e-01 -4.69065569e-02
1.78239226e-01 2.44965911e-01 2.90408254e-01 -6.45145262e-03
8.54464844e-02 -7.38979951e-02 2.75237054e-01 5.59002280e-01
3.54985774e-01 -5.79301536e-01 -6.89668894e-01 6.78390682e-01
-1.58098459e+00 -1.29172778e+00 -1.97072297e-01 1.67446077e+00
1.17563760e+00 5.71794748e-01 -2.53992289e-01 4.50071037e-01
5.49311221e-01 -7.47065712e-03 -1.63790926e-01 -9.94095802e-01
-5.94024539e-01 2.46404722e-01 3.63214552e-01 5.73167384e-01
-1.44876897e+00 1.01726854e+00 7.48160172e+00 7.74862647e-01
-9.49592829e-01 1.18394338e-01 1.03131318e+00 -1.89554498e-01
-2.65112489e-01 3.02354656e-02 -7.94849873e-01 4.99549627e-01
1.18329906e+00 1.17107801e-01 -1.90288238e-02 8.18233550e-01
-1.10748699e-02 -1.67708695e-01 -1.13690281e+00 4.80979055e-01
9.30524290e-01 -1.93946314e+00 3.80804032e-01 -3.82865965e-01
8.40716541e-01 -1.79336742e-02 -2.40502700e-01 7.86432266e-01
5.85301779e-02 -9.78189588e-01 7.38553107e-01 4.65382695e-01
5.80035746e-01 -8.92711520e-01 1.62454176e+00 4.21169847e-01
3.24245207e-02 -1.02955982e-01 -3.88741702e-01 -4.62110728e-01
-1.31274953e-01 6.88358366e-01 -1.16140783e+00 5.51877022e-01
6.41036987e-01 6.53671026e-01 -7.48095036e-01 4.69062597e-01
-1.12908706e-01 6.20876670e-01 -6.31693378e-02 -3.65786999e-01
2.43526429e-01 4.01086003e-01 4.70863581e-01 1.39482570e+00
4.51064855e-02 -1.41452730e-01 6.23717904e-02 5.55127203e-01
-5.05180955e-01 -4.74802554e-02 -8.22755516e-01 -1.72905117e-01
3.38257164e-01 6.91107929e-01 -4.95511293e-01 -1.01729620e+00
-2.89564669e-01 1.11032617e+00 5.29516995e-01 -2.51875184e-02
-6.01159096e-01 -8.49642336e-01 4.27763760e-02 -2.74046101e-02
-6.00301027e-02 6.12560570e-01 -5.30544817e-01 -1.26398492e+00
1.27980143e-01 -1.29308844e+00 4.73402977e-01 -5.47939718e-01
-1.58877146e+00 7.04344273e-01 -2.74830937e-01 -1.29404831e+00
-5.61366558e-01 -7.72309422e-01 -5.75410664e-01 6.24646485e-01
-1.37718391e+00 -9.66575921e-01 3.01054507e-01 9.52471867e-02
6.42842948e-01 -6.19109988e-01 1.12857175e+00 1.52663738e-01
-1.46460593e-01 7.59823620e-01 1.31955102e-01 1.00421675e-01
7.73367167e-01 -1.27475631e+00 4.03485298e-01 8.17906380e-01
-9.58252102e-02 5.58668911e-01 8.04353535e-01 -9.57868874e-01
-6.30734324e-01 -1.10328019e+00 1.75922704e+00 -5.09140432e-01
6.09644234e-01 -4.15134549e-01 -9.37744677e-01 5.72779179e-01
4.20848519e-01 -1.98707983e-01 7.27306306e-01 -1.92040671e-02
-5.66542387e-01 2.31519446e-01 -1.71433616e+00 2.42035925e-01
4.84166086e-01 -9.10731316e-01 -1.05388165e+00 1.02013600e+00
8.22167218e-01 -6.90242827e-01 -4.50955421e-01 3.78576756e-01
8.68970877e-04 -7.91740477e-01 6.53842747e-01 -1.40926516e+00
1.22193480e+00 1.33337528e-01 -7.19005913e-02 -1.40993845e+00
-4.45517413e-02 -2.99553096e-01 -3.40853333e-01 9.37047899e-01
7.56353259e-01 -5.03927290e-01 1.02063656e+00 -1.24887817e-01
-1.95839837e-01 -9.82835650e-01 -1.08646822e+00 -6.62009537e-01
3.12094480e-01 -3.43806446e-01 1.19098708e-01 1.25516331e+00
3.47904772e-01 7.29472697e-01 -5.15001357e-01 -1.49839386e-01
2.78923005e-01 -2.31453970e-01 1.86019167e-01 -1.12264514e+00
8.96610226e-03 -3.53166819e-01 -5.99428356e-01 -8.37911963e-01
3.20745200e-01 -1.11784697e+00 1.75436452e-01 -1.34727585e+00
1.38244897e-01 1.19865842e-01 2.69945621e-01 2.79025823e-01
-9.65663344e-02 4.59662288e-01 -1.38871253e-01 9.50332358e-02
-4.56049055e-01 6.19057000e-01 8.29294503e-01 -3.67680490e-01
2.51472682e-01 -5.11279143e-02 -9.64958727e-01 7.05083251e-01
9.79828358e-01 -8.25175345e-01 -1.00232735e-01 -4.23834652e-01
2.38223568e-01 -5.75396419e-02 5.93370736e-01 -9.45290208e-01
1.42381668e-01 4.49802488e-01 6.65151596e-01 -7.56838143e-01
2.06009552e-01 -4.01078552e-01 -8.84558499e-01 8.54855478e-01
-1.00842166e+00 1.06937125e-01 2.78215617e-01 7.00988173e-01
-5.26091576e-01 -8.24839830e-01 8.57908070e-01 -4.00204211e-01
-1.46934301e-01 -6.23575509e-01 -8.67341459e-01 4.44049299e-01
7.43088245e-01 1.29276872e-01 -8.32068324e-01 -5.50703228e-01
-6.46094143e-01 -2.98896451e-02 2.12765974e-03 4.57145780e-01
7.36456871e-01 -1.27770960e+00 -9.95386600e-01 3.13280314e-01
2.17653409e-01 -2.94264227e-01 -1.14573680e-01 6.73702836e-01
-7.80235350e-01 6.10846281e-01 -7.94812222e-04 -2.28413090e-01
-1.03485382e+00 3.88367027e-01 3.53467077e-01 -5.16433835e-01
-2.80345917e-01 1.14023411e+00 -8.56781602e-01 -9.80735600e-01
8.39603618e-02 -1.60064921e-01 -3.01162750e-01 6.94006160e-02
6.96011543e-01 1.93087548e-01 6.27963722e-01 -6.33555591e-01
-3.15074325e-01 -4.89553064e-01 -6.53272808e-01 1.39658824e-01
1.18409157e+00 2.30267361e-01 -2.65217483e-01 3.86462778e-01
1.50975096e+00 -3.83081406e-01 -2.91817933e-01 -2.14254960e-01
4.43529397e-01 -2.89936781e-01 1.38173059e-01 -1.24815059e+00
-4.85946745e-01 7.76716173e-01 3.92125100e-01 2.72219896e-01
8.62207055e-01 -1.75746471e-01 7.47768819e-01 8.75237286e-01
-1.35176238e-02 -1.53445232e+00 1.80710137e-01 1.01264250e+00
1.06346917e+00 -1.53615534e+00 1.62365302e-01 -3.20633799e-01
-6.34780049e-01 1.49280810e+00 7.11440682e-01 -4.08294201e-01
2.29449093e-01 1.84482142e-01 1.22732125e-01 -4.95355278e-01
-8.06391776e-01 4.11939025e-01 4.90879826e-02 4.23253000e-01
7.75190771e-01 -1.74186915e-01 -6.08295262e-01 4.33382004e-01
-4.29714411e-01 -2.07283169e-01 9.16454613e-01 1.15361154e+00
-2.54289597e-01 -6.21824324e-01 -1.85930416e-01 9.66078639e-01
-9.71933186e-01 -4.27493840e-01 -7.32079387e-01 8.98871601e-01
1.42322689e-01 9.10243094e-01 3.18800867e-01 -3.93687487e-01
1.96661472e-01 2.17764348e-01 1.09096758e-01 -7.18893826e-01
-1.05937207e+00 -4.54318911e-01 8.76775861e-01 -5.16036570e-01
-5.36443591e-01 -4.10103023e-01 -8.59521627e-01 -1.80746794e-01
-6.95089698e-01 3.07265133e-01 3.94992352e-01 1.06358540e+00
3.13536108e-01 3.93418044e-01 4.36552316e-01 -6.48272336e-01
-1.05911744e+00 -1.36563921e+00 -6.44951016e-02 6.81371450e-01
5.48516452e-01 -2.80179858e-01 -6.54635251e-01 -1.02862827e-01] | [8.26508617401123, 10.181344985961914] |
9490c708-7266-48c3-ba12-da712143bad8 | ungeneralizable-contextual-logistic-bandit-in | 2212.07632 | null | https://arxiv.org/abs/2212.07632v1 | https://arxiv.org/pdf/2212.07632v1.pdf | Ungeneralizable Contextual Logistic Bandit in Credit Scoring | The application of reinforcement learning in credit scoring has created a unique setting for contextual logistic bandit that does not conform to the usual exploration-exploitation tradeoff but rather favors exploration-free algorithms. Through sufficient randomness in a pool of observable contexts, the reinforcement learning agent can simultaneously exploit an action with the highest reward while still learning more about the structure governing that environment. Thus, it is the case that greedy algorithms consistently outperform algorithms with efficient exploration, such as Thompson sampling. However, in a more pragmatic scenario in credit scoring, lenders can, to a degree, classify each borrower as a separate group, and learning about the characteristics of each group does not infer any information to another group. Through extensive simulations, we show that Thompson sampling dominates over greedy algorithms given enough timesteps which increase with the complexity of underlying features. | ['Seksan Kiatsupaibul', 'Kantapong Visantavarakul', 'Pojtanut Manopanjasiri'] | 2022-12-15 | null | null | null | null | ['thompson-sampling'] | ['methodology'] | [-1.11752778e-01 2.27059752e-01 -8.70205760e-01 -1.33403614e-01
-8.49422932e-01 -7.14105666e-01 4.88076955e-01 1.54305249e-01
-7.14043558e-01 1.11263895e+00 2.89169550e-01 -4.29576159e-01
-4.46134716e-01 -8.97373021e-01 -4.36347902e-01 -8.24205041e-01
-1.17129214e-01 9.38443542e-01 -2.68391687e-02 1.65390387e-01
4.65676218e-01 3.63011628e-01 -1.16819656e+00 6.11704104e-02
5.10226667e-01 7.48777986e-01 8.55366364e-02 6.52344048e-01
-8.55414569e-02 1.15550494e+00 -4.34434086e-01 -1.20497666e-01
3.19221199e-01 -7.17986166e-01 -7.33123600e-01 8.77195075e-02
-1.99799135e-01 -6.58785939e-01 -4.04658556e-01 8.96368384e-01
7.99329802e-02 4.28844750e-01 7.12294638e-01 -8.61789048e-01
-4.30580705e-01 1.20310259e+00 -6.14164114e-01 2.81091839e-01
3.30994129e-01 3.13065439e-01 1.53472149e+00 -3.14614087e-01
5.68957508e-01 1.03077221e+00 2.75950283e-01 3.39840472e-01
-1.37987113e+00 -7.22567081e-01 6.06670976e-01 6.43173084e-02
-5.08746684e-01 -9.77008268e-02 6.33838058e-01 -1.55414075e-01
8.16391468e-01 2.51440942e-01 1.15954518e+00 9.28139031e-01
-1.38132125e-01 1.13837218e+00 1.45570016e+00 -4.70028579e-01
9.44175601e-01 8.33472833e-02 2.43304566e-01 3.89770180e-01
5.11077702e-01 7.73122549e-01 -8.34340334e-01 -6.39778137e-01
7.29845524e-01 2.86558121e-01 -3.86443697e-02 -6.95266604e-01
-1.01381803e+00 1.23552728e+00 2.82810181e-01 8.48543458e-03
-6.37733221e-01 5.84226012e-01 1.48508966e-01 6.85328186e-01
2.02494115e-01 6.99708879e-01 -3.48991394e-01 -5.02673328e-01
-1.09260237e+00 7.97511339e-01 9.28126037e-01 4.71057236e-01
1.00324750e+00 6.04701377e-02 -4.96074893e-02 3.88257027e-01
4.18295950e-01 3.93824726e-01 6.00850105e-01 -1.21477902e+00
6.87643349e-01 2.22711235e-01 6.77833974e-01 -2.22329780e-01
-3.03919196e-01 -3.82916898e-01 -4.37618010e-02 7.07749009e-01
6.94844723e-01 -6.00041628e-01 -6.69847310e-01 1.80179513e+00
1.98801637e-01 -1.34127468e-01 -5.15306629e-02 8.28376353e-01
-4.48899329e-01 4.87481415e-01 1.02101572e-01 -5.11110961e-01
7.57799804e-01 -8.38609934e-01 -2.82471210e-01 -5.65780640e-01
5.67997813e-01 -1.63415596e-01 7.99343407e-01 7.35296905e-01
-1.23840892e+00 2.72566471e-02 -1.08492911e+00 6.08799040e-01
1.13298774e-01 -6.15986645e-01 1.26818216e+00 8.95947456e-01
-7.35664129e-01 8.52503300e-01 -9.13561285e-01 -5.19695394e-02
6.84933662e-01 2.83505619e-01 1.84329912e-01 1.61046386e-01
-9.70623970e-01 7.41807759e-01 2.87086427e-01 -3.39468569e-01
-1.19564927e+00 -3.70114774e-01 -2.19446868e-01 3.78833413e-01
5.69990396e-01 -4.72288221e-01 1.65488803e+00 -1.18912494e+00
-1.55628252e+00 2.75070876e-01 1.15713477e-01 -7.95249820e-01
8.12741756e-01 -3.37609768e-01 2.55387008e-01 1.63151398e-02
1.62975654e-01 3.00816983e-01 7.30751932e-01 -9.41200793e-01
-9.45307136e-01 -4.68158334e-01 7.17483312e-02 5.10603607e-01
-6.47423491e-02 8.92219394e-02 4.97023880e-01 -5.14535546e-01
3.19809139e-01 -1.10389757e+00 -5.75165153e-01 -2.06498817e-01
-2.65873820e-02 -1.12705551e-01 8.04722235e-02 -1.95075348e-02
1.00819814e+00 -1.92223382e+00 -2.15403885e-01 4.75003839e-01
1.15909278e-01 -4.64065075e-01 2.58569628e-01 4.94990110e-01
1.52346641e-01 1.24829235e-02 -9.73777995e-02 -2.15313514e-03
2.26480350e-01 1.69977367e-01 -7.03606427e-01 4.98809159e-01
-3.80511582e-01 7.35677242e-01 -1.04242802e+00 -2.05358610e-01
-8.92666355e-02 -6.24818265e-01 -8.46417248e-01 1.27704963e-01
-3.59796822e-01 6.78054616e-03 -1.01141012e+00 5.49634755e-01
3.07101458e-01 -2.56018460e-01 5.39408743e-01 1.12438965e+00
-8.96069482e-02 5.82554817e-01 -1.48484647e+00 1.17850769e+00
-3.19527864e-01 3.35739732e-01 2.71357503e-02 -1.17552578e+00
5.72607636e-01 4.51071620e-01 5.10489881e-01 -3.41330975e-01
-1.55432269e-01 4.26460505e-01 -1.48721203e-01 -2.17697904e-01
4.13496226e-01 -5.14374912e-01 -9.78363752e-02 1.43014622e+00
-1.30409405e-01 5.14802942e-03 -8.64435732e-02 2.99333949e-02
1.17142594e+00 2.24367365e-01 4.77310359e-01 -1.07045837e-01
-4.65316594e-01 2.64852524e-01 7.35107541e-01 1.59207880e+00
-2.40893945e-01 2.46709004e-01 6.61082387e-01 -6.03846908e-01
-9.06992316e-01 -1.16616499e+00 -6.83696195e-02 1.34842074e+00
3.28886919e-02 1.85606480e-01 -4.08184350e-01 -8.92699361e-01
4.43172902e-01 7.38805056e-01 -8.47545266e-01 1.42331094e-01
-5.00215530e-01 -1.16008544e+00 1.16272688e-01 4.95941639e-01
2.85928845e-01 -1.44807279e+00 -1.17632830e+00 7.23856807e-01
4.00557965e-01 -2.94532418e-01 -2.38237187e-01 9.63668704e-01
-1.26577497e+00 -9.24033582e-01 -5.50369143e-01 -2.83619016e-01
3.31777960e-01 9.69109461e-02 9.65405464e-01 -7.19429255e-02
2.04392269e-01 1.06735505e-01 -2.97174066e-01 -4.85724479e-01
-3.08228314e-01 1.23576611e-01 6.08559651e-03 -2.79460073e-01
6.00589335e-01 -7.21497059e-01 -6.57978952e-01 2.35758405e-02
-3.79714698e-01 -2.90492237e-01 6.32155955e-01 1.06337166e+00
9.28696394e-02 -2.02271179e-03 8.38087261e-01 -1.25271130e+00
6.72516525e-01 -7.33136237e-01 -6.84091210e-01 1.27552107e-01
-9.19977903e-01 2.93544918e-01 3.72763395e-01 -8.20187092e-01
-1.12779796e+00 7.68269598e-02 5.09740472e-01 9.46993604e-02
-5.18815406e-02 5.53257048e-01 7.42767975e-02 2.85692930e-01
8.44818473e-01 1.53063297e-01 -1.01526633e-01 -6.99366271e-01
2.50621259e-01 5.52713275e-01 7.10714310e-02 -9.22283173e-01
4.38743919e-01 4.55224782e-01 -2.15410978e-01 -3.00768673e-01
-7.57600009e-01 -2.49826282e-01 -1.64764658e-01 -3.82006802e-02
3.54757130e-01 -6.73337400e-01 -7.21185327e-01 2.29062900e-01
-5.10546803e-01 -6.76561415e-01 -7.53839970e-01 8.05830359e-01
-1.05504870e+00 9.88163520e-03 -6.02785766e-01 -1.28484082e+00
2.90517777e-01 -8.66579533e-01 3.14936489e-01 3.77325356e-01
-3.65019649e-01 -7.32724249e-01 4.19458926e-01 3.24234933e-01
2.80789882e-01 -1.74871340e-01 8.68137360e-01 -9.92630959e-01
-9.87346828e-01 -1.33493558e-01 2.86364645e-01 -1.82203874e-01
-6.95924386e-02 -3.58557045e-01 -9.56972778e-01 -3.81056964e-01
1.60460174e-01 -6.22225523e-01 9.17588651e-01 5.27528703e-01
7.80954838e-01 -5.13304949e-01 -2.66226143e-01 3.70220453e-01
1.16141319e+00 4.58520263e-01 1.02988020e-01 9.18381870e-01
1.09053753e-01 4.39264208e-01 7.20027864e-01 8.61718655e-01
5.64938635e-02 5.46290100e-01 4.09039110e-01 3.05985242e-01
5.70780337e-01 -4.93019968e-01 4.12188917e-01 -1.20996259e-01
-4.54106890e-02 8.23747888e-02 -5.66324234e-01 6.21922910e-01
-2.07405567e+00 -1.25165606e+00 6.59485817e-01 2.28266406e+00
1.03147054e+00 4.07976866e-01 5.97793043e-01 6.94565028e-02
4.75721747e-01 2.38464072e-01 -1.03640378e+00 -5.86474121e-01
1.94492191e-02 1.61517486e-01 8.88885677e-01 6.49910510e-01
-8.26053321e-01 7.73702919e-01 7.25493622e+00 7.24049985e-01
-7.03727484e-01 -5.23076020e-02 8.27096999e-01 -6.49661839e-01
-5.94896555e-01 6.13913417e-01 -7.91336417e-01 5.19143403e-01
8.56810331e-01 -3.71147603e-01 8.77948761e-01 1.11749911e+00
1.54113516e-01 -4.46072400e-01 -1.21366203e+00 2.71333277e-01
-5.88783205e-01 -1.33585525e+00 -2.55303472e-01 4.03212398e-01
9.00683045e-01 1.67522550e-01 -2.24290006e-02 2.88889468e-01
1.34116161e+00 -1.02754211e+00 1.07926166e+00 2.31874660e-01
3.01437110e-01 -8.06366146e-01 3.80020946e-01 7.83289671e-01
-4.36220884e-01 -7.51138091e-01 -3.75817925e-01 -5.43337345e-01
-9.39525105e-03 2.57769585e-01 -7.80726314e-01 -1.41128108e-01
4.05498445e-01 2.66148418e-01 -8.18565637e-02 1.05797911e+00
-4.13466334e-01 1.02839410e+00 -3.92801285e-01 -5.07518113e-01
6.74303710e-01 -3.48781288e-01 3.78139138e-01 8.60063672e-01
1.87249288e-01 1.95032343e-01 2.97994524e-01 6.44328475e-01
1.07382506e-01 8.99448693e-02 -3.65893751e-01 2.43194163e-01
7.75751173e-01 6.65386260e-01 -6.59353852e-01 -4.00293291e-01
-3.77767235e-01 2.90646374e-01 5.13041556e-01 5.49297631e-01
-3.40057194e-01 1.27467483e-01 5.89657605e-01 7.95162544e-02
5.61516583e-01 -6.22546412e-02 -5.55014193e-01 -1.00703001e+00
-3.93269807e-02 -9.78591084e-01 8.37826729e-01 -4.36370432e-01
-1.04206610e+00 1.69180706e-02 6.06489591e-02 -8.36263418e-01
-7.74178982e-01 -2.25204706e-01 -5.41003585e-01 7.00609505e-01
-1.45955145e+00 -4.99969423e-01 4.61037457e-01 2.93164998e-01
5.87306738e-01 -3.49267542e-01 6.77752018e-01 -7.33483255e-01
-1.73117504e-01 2.43763879e-01 7.11586893e-01 -1.37396142e-01
3.67616832e-01 -1.60651541e+00 1.56093925e-01 4.45885032e-01
5.83074450e-01 4.96027082e-01 8.26565444e-01 -5.38320005e-01
-1.05442595e+00 -4.44565326e-01 4.78857696e-01 -2.98998475e-01
9.57888722e-01 -4.95890938e-02 -8.04259777e-01 7.41928101e-01
-4.74847779e-02 -2.45671362e-01 5.61500788e-01 4.71104711e-01
-2.69142926e-01 -1.81620214e-02 -1.11882007e+00 5.37630737e-01
7.32449055e-01 -4.02548134e-01 -7.18378305e-01 2.61357367e-01
2.54302055e-01 -1.34725749e-01 -5.13841569e-01 -2.72261173e-01
6.84572875e-01 -1.17525339e+00 6.56920135e-01 -9.92076874e-01
1.74982324e-01 9.26899686e-02 -1.25187978e-01 -1.35584676e+00
-4.15753394e-01 -1.09772456e+00 -2.41426617e-01 6.71591818e-01
5.34089863e-01 -8.75208795e-01 1.52104783e+00 8.88242066e-01
4.31451976e-01 -8.21326077e-01 -1.24143934e+00 -8.09330583e-01
6.50300920e-01 -2.64775127e-01 9.55068529e-01 7.68718779e-01
5.56904495e-01 -1.93371832e-01 -2.47264832e-01 -2.08062887e-01
8.81937087e-01 7.73424268e-01 5.37150323e-01 -1.03156638e+00
-9.35700774e-01 -5.85112989e-01 6.67762160e-02 -1.27355993e+00
-2.34889463e-02 -4.87924576e-01 3.22994702e-02 -1.01261747e+00
5.80895960e-01 -9.56759930e-01 -7.43866861e-01 4.39642519e-01
-2.88019866e-01 -2.68090665e-01 2.70902783e-01 5.37681997e-01
-5.16949356e-01 1.98414028e-01 9.86729741e-01 2.99929380e-02
-4.25635874e-01 5.37503839e-01 -1.09495640e+00 7.67448962e-01
8.67017508e-01 -8.17623258e-01 -4.14747655e-01 -1.06652021e-01
6.21991634e-01 7.56835282e-01 1.57975823e-01 -4.76662427e-01
1.81444556e-01 -8.39263737e-01 3.06272626e-01 -1.86232314e-01
8.10530558e-02 -4.76984292e-01 -5.67907421e-03 5.45476079e-01
-9.79519308e-01 -1.97890028e-01 -5.58736384e-01 7.11665273e-01
1.12707458e-01 -7.81679630e-01 7.21486628e-01 -4.60694939e-01
-1.67888850e-01 1.98494419e-01 -7.38954961e-01 2.77796328e-01
7.66083360e-01 -4.59130436e-01 -1.60930037e-01 -8.28778982e-01
-8.15376997e-01 1.50060713e-01 7.28541195e-01 -1.61418885e-01
2.47263357e-01 -1.12603867e+00 -6.55828595e-01 -3.00769159e-03
-1.61949649e-01 -1.93756044e-01 -3.24014798e-02 2.76452631e-01
-1.34397119e-01 3.68736982e-01 -1.55648872e-01 -4.31194492e-02
-7.21296549e-01 4.91306335e-01 3.73799086e-01 -6.33456349e-01
-6.84318662e-01 6.68696642e-01 3.54513407e-01 7.43735656e-02
2.31781632e-01 -1.85142979e-01 -1.00894786e-01 1.03875063e-01
5.38800299e-01 3.45681399e-01 -2.38529444e-01 1.60543233e-01
6.95516840e-02 -2.73975544e-02 -3.33407164e-01 -6.59909070e-01
1.65621650e+00 3.50002646e-02 3.91443670e-01 4.52396661e-01
5.88508308e-01 -6.24592304e-02 -2.00162721e+00 -3.90566379e-01
3.97019088e-01 -8.65734935e-01 1.64837763e-01 -7.24557042e-01
-7.99158275e-01 5.26812077e-01 1.72654003e-01 5.50110221e-01
7.08659291e-01 -4.14785594e-02 3.22713733e-01 6.69975102e-01
6.70056343e-01 -1.49782562e+00 2.35122647e-02 1.16683252e-01
3.59578162e-01 -1.03772593e+00 3.52464676e-01 4.80665684e-01
-8.81110191e-01 9.50711429e-01 1.87687457e-01 -5.65853536e-01
5.64819515e-01 2.36821130e-01 3.04357037e-02 -1.11170612e-01
-1.20382881e+00 7.64342919e-02 -6.03851020e-01 5.61968923e-01
2.92472001e-02 3.82103235e-01 -5.28052822e-02 5.02161205e-01
-4.10299122e-01 1.42988950e-01 7.66373217e-01 1.00846827e+00
-1.01694155e+00 -1.17217505e+00 -5.27238667e-01 7.43509293e-01
-7.38194644e-01 9.34007093e-02 -8.93793032e-02 8.54664922e-01
-3.73195589e-01 7.10800529e-01 1.27937347e-01 2.65982985e-01
-7.78138936e-02 6.45364970e-02 4.75431681e-01 -6.76272213e-01
-6.78038418e-01 2.38154978e-01 1.19678125e-01 -5.12418270e-01
-1.37022913e-01 -1.40944839e+00 -1.00750220e+00 -2.06573531e-01
-3.87427449e-01 5.75245023e-01 4.16420221e-01 1.03659761e+00
-3.21589351e-01 -1.74537957e-01 1.07555723e+00 -5.44714749e-01
-1.49689567e+00 -7.22072601e-01 -1.20840919e+00 2.10078061e-01
6.32794201e-01 -7.63239443e-01 -6.81584001e-01 -3.81357789e-01] | [4.402288913726807, 3.0330331325531006] |
7582e4bc-bca2-45bd-9f05-1feef466f783 | musclemap-towards-video-based-activated | 2303.00952 | null | https://arxiv.org/abs/2303.00952v2 | https://arxiv.org/pdf/2303.00952v2.pdf | MuscleMap: Towards Video-based Activated Muscle Group Estimation | In this paper, we tackle the new task of video-based Activated Muscle Group Estimation (AMGE) aiming at identifying active muscle regions during physical activity. To this intent, we provide the MuscleMap136 dataset featuring >15K video clips with 136 different activities and 20 labeled muscle groups. This dataset opens the vistas to multiple video-based applications in sports and rehabilitation medicine. We further complement the main MuscleMap136 dataset, which specifically targets physical exercise, with Muscle-UCF90 and Muscle-HMDB41, which are new variants of the well-known activity recognition benchmarks extended with AMGE annotations. To make the AMGE model applicable in real-life situations, it is crucial to ensure that the model can generalize well to types of physical activities not present during training and involving new combinations of activated muscles. To achieve this, our benchmark also covers an evaluation setting where the model is exposed to activity types excluded from the training set. Our experiments reveal that generalizability of existing architectures adapted for the AMGE task remains a challenge. Therefore, we also propose a new approach, TransM3E, which employs a transformer-based model with cross-modal multi-label knowledge distillation and surpasses all popular video classification models when dealing with both, previously seen and new types of physical activities. The datasets and code will be publicly available at https://github.com/KPeng9510/MuscleMap. | ['Rainer Stiefelhagen', 'M. Saquib Sarfraz', 'Jiaming Zhang', 'Kailun Yang', 'Alina Roitberg', 'David Schneider', 'Kunyu Peng'] | 2023-03-02 | null | null | null | null | ['video-classification', 'human-activity-recognition', 'human-activity-recognition'] | ['computer-vision', 'computer-vision', 'time-series'] | [ 5.08608699e-01 -2.35999569e-01 -7.46030927e-01 1.82710811e-01
-6.99697495e-01 -4.69308317e-01 3.94448280e-01 -3.15048367e-01
-6.17576182e-01 7.59534538e-01 3.23553532e-01 1.31252229e-01
-2.28967458e-01 -4.58337456e-01 -8.41207922e-01 -7.44278133e-01
-4.39110130e-01 2.51409948e-01 3.71351272e-01 -1.44642349e-02
-2.19876878e-03 1.83563545e-01 -1.74315071e+00 8.39787602e-01
3.85372549e-01 9.06050265e-01 1.89290978e-02 8.64564121e-01
3.49391311e-01 1.08573103e+00 -4.94153649e-01 -6.31125197e-02
4.35862243e-01 -4.25621748e-01 -1.05512536e+00 2.66063333e-01
8.49998116e-01 -7.56679103e-02 -6.08923793e-01 5.63257515e-01
5.79794168e-01 3.96599501e-01 3.47419620e-01 -1.10629988e+00
-1.19671598e-01 4.62126851e-01 -3.92427355e-01 7.06256807e-01
4.16818559e-01 5.04867017e-01 7.21761882e-01 -5.76186299e-01
8.43043506e-01 7.07143366e-01 7.44573474e-01 6.94424748e-01
-1.11315942e+00 -4.96732205e-01 3.27954739e-01 6.69313192e-01
-1.29914427e+00 -3.67324978e-01 7.66241670e-01 -6.44433916e-01
7.60112107e-01 5.48333406e-01 9.96873498e-01 1.69424272e+00
-5.58368526e-02 1.17174768e+00 1.29081345e+00 -1.37322277e-01
6.61176518e-02 -4.86278385e-01 3.94070297e-02 4.10502195e-01
-1.10805318e-01 -4.22919635e-03 -9.10595953e-01 1.95038006e-01
6.54922307e-01 -9.43385214e-02 -5.84316492e-01 -5.60539246e-01
-1.71361613e+00 3.43008161e-01 9.36182365e-02 4.74213690e-01
-5.58115840e-01 4.28997606e-01 8.86207283e-01 3.01370502e-01
3.91284049e-01 2.74595290e-01 -4.37572628e-01 -6.75575495e-01
-9.73316014e-01 4.64990169e-01 6.20196700e-01 7.23292768e-01
4.38631982e-01 -7.04496577e-02 -3.83220345e-01 9.14610982e-01
-1.58593893e-01 -2.77394522e-03 6.68642581e-01 -1.23347950e+00
5.35650194e-01 5.55321097e-01 -2.95313418e-01 -5.99164903e-01
-4.56546366e-01 -6.38987422e-01 -6.61772490e-01 8.81915912e-02
5.70920110e-01 1.58917069e-01 -7.90297389e-01 1.63467479e+00
4.81062174e-01 7.49732733e-01 -6.24602139e-02 1.11645198e+00
7.93784142e-01 2.22870573e-01 1.86861336e-01 -1.07636467e-01
1.44580376e+00 -1.33175147e+00 -5.81658244e-01 -2.80687988e-01
9.26020384e-01 -2.16981441e-01 1.06971490e+00 7.92162359e-01
-9.74570394e-01 -7.35650897e-01 -9.35641170e-01 9.77769867e-02
-3.35335195e-01 1.90153778e-01 5.89490831e-01 4.15950000e-01
-7.80851543e-01 7.29083836e-01 -8.53651583e-01 -3.33694786e-01
4.96250033e-01 3.53333890e-01 -7.43410110e-01 -1.34250268e-01
-1.30895603e+00 8.23394060e-01 6.55541003e-01 2.69137502e-01
-9.52223480e-01 -7.80820787e-01 -7.25544930e-01 -3.16563874e-01
9.27299619e-01 -6.42137945e-01 1.15647531e+00 -1.16757846e+00
-1.44144404e+00 1.09299898e+00 3.92689466e-01 -3.58088076e-01
6.52560532e-01 -3.90132725e-01 -5.11102855e-01 4.95522320e-01
-1.55245796e-01 4.80786681e-01 5.25983930e-01 -7.36764908e-01
-3.70801330e-01 -1.88780114e-01 3.96204919e-01 3.86235327e-01
-2.86596388e-01 -7.31109232e-02 -7.21422672e-01 -8.52674067e-01
-5.31169236e-01 -1.18789673e+00 -4.72717807e-02 -8.55578948e-03
-2.51604319e-01 -1.34516150e-01 4.88294423e-01 -7.65435755e-01
1.35436988e+00 -2.10643744e+00 7.53368497e-01 2.67169140e-02
1.75732240e-01 4.00476187e-01 -2.62313694e-01 2.99768269e-01
-3.67706716e-01 -1.71346635e-01 -2.44036168e-01 -1.01266891e-01
-9.73637700e-02 4.30999041e-01 5.15562057e-01 4.37846988e-01
1.55741155e-01 8.73620570e-01 -9.18503881e-01 -7.11579084e-01
3.42780471e-01 1.57217056e-01 -6.11346602e-01 -6.74974769e-02
-1.03205264e-01 6.96019948e-01 -3.15637320e-01 8.42154443e-01
2.80585140e-01 -5.89798167e-02 2.97403604e-01 -3.85338217e-01
1.25197664e-01 -1.35480955e-01 -1.30311441e+00 2.41575265e+00
-2.63621747e-01 3.06801587e-01 -1.07828416e-02 -1.34859335e+00
3.60515475e-01 3.63562703e-01 1.30401170e+00 -3.62438977e-01
9.29638296e-02 -5.45972288e-02 4.98003997e-02 -7.85393953e-01
1.89533889e-01 -4.12158296e-02 -2.52082825e-01 1.14710771e-01
4.37967420e-01 5.34212291e-01 8.07484448e-01 -2.35319138e-02
1.56306469e+00 6.89657927e-01 3.01459581e-01 -2.03620628e-01
6.85242057e-01 4.55470346e-02 7.75960565e-01 5.39797127e-01
-5.01601934e-01 5.84539413e-01 3.04547608e-01 -4.03353691e-01
-6.10987842e-01 -1.06252348e+00 1.10929318e-01 1.15809703e+00
2.71804370e-02 -6.22114062e-01 -6.81470811e-01 -9.80398417e-01
-7.87845179e-02 4.37069759e-02 -8.85773361e-01 -2.73728520e-01
-9.02864039e-01 -7.24690437e-01 6.39387369e-01 6.74425542e-01
4.66229886e-01 -1.18507719e+00 -6.03029966e-01 1.67435631e-01
-6.08551264e-01 -1.34059262e+00 -5.91256142e-01 8.20694864e-02
-7.58216858e-01 -1.49220395e+00 -9.07032371e-01 -5.67839026e-01
9.81761366e-02 -4.91724685e-02 1.18071699e+00 2.51948577e-03
-4.22049910e-01 6.78646564e-01 -6.13596141e-01 2.10716715e-03
-5.00633061e-01 3.22010398e-01 4.40437496e-02 2.60236561e-01
4.05426025e-01 -6.32725775e-01 -8.29265952e-01 4.12650317e-01
-1.02488351e+00 2.56997973e-01 7.07666993e-01 7.30138719e-01
8.00184667e-01 -3.06555390e-01 4.49894577e-01 -7.34395623e-01
2.66522299e-02 -6.78514242e-01 9.59769115e-02 2.46127188e-01
-1.01179346e-01 -2.50783205e-01 3.02073270e-01 -9.65307653e-01
-7.17197001e-01 2.53191024e-01 -2.26467296e-01 -8.09520543e-01
-3.44951361e-01 5.98550439e-01 -1.22842960e-01 5.81928082e-02
8.28615487e-01 3.40352237e-01 1.70774430e-01 -6.66885614e-01
1.92092612e-01 4.30359989e-01 8.17214549e-01 -6.10347569e-01
3.93094212e-01 3.84685278e-01 2.00886041e-01 -7.14206696e-01
-7.68311441e-01 -8.25064659e-01 -8.22544336e-01 -7.74545312e-01
9.23131645e-01 -1.09728968e+00 -5.45152128e-01 7.34577715e-01
-5.01565754e-01 -7.98767745e-01 -5.18470228e-01 7.58858144e-01
-1.04797375e+00 7.99779415e-01 -7.42969751e-01 -4.23430979e-01
-2.29446277e-01 -9.84455466e-01 1.04296458e+00 -2.35308915e-01
-1.98129714e-01 -9.03320551e-01 3.52363467e-01 8.31406653e-01
6.36453778e-02 6.99320376e-01 3.41607839e-01 -4.74564135e-01
-3.18717182e-01 -7.42756054e-02 4.64688092e-01 5.94136000e-01
9.15364400e-02 -3.82861972e-01 -7.26549149e-01 -3.01666170e-01
-4.06511188e-01 -5.90302229e-01 1.14014304e+00 2.83890158e-01
1.35110950e+00 -7.25180358e-02 -1.93094671e-01 5.69784820e-01
1.04960406e+00 -2.13696420e-01 9.13424969e-01 5.72641969e-01
7.25103498e-01 6.20445788e-01 8.48977149e-01 2.04503343e-01
2.17957690e-01 1.17238677e+00 4.47662979e-01 -2.45875031e-01
-4.64910001e-01 7.54902661e-02 5.59551418e-01 8.94570947e-01
-7.61991084e-01 -1.57153353e-01 -6.99712813e-01 6.38608515e-01
-1.89003682e+00 -1.34801388e+00 -1.28015518e-01 2.05371952e+00
1.06110883e+00 -2.32620500e-02 6.60020411e-01 3.88653934e-01
4.46433604e-01 2.46628180e-01 -5.07395625e-01 -1.06490746e-01
-5.22800870e-02 4.13005263e-01 4.26664501e-01 -1.23434357e-01
-1.49158871e+00 5.53652167e-01 5.69270802e+00 1.13613200e+00
-8.10448945e-01 4.98926759e-01 1.55342788e-01 -6.07390583e-01
3.90207201e-01 -2.78861612e-01 -4.03626978e-01 6.40662909e-01
1.04380107e+00 2.84373343e-01 3.72431964e-01 5.84415495e-01
2.24482313e-01 -1.43018663e-01 -9.78548944e-01 8.53269517e-01
1.72870934e-01 -1.11021149e+00 -3.03950220e-01 6.17366768e-02
3.63569170e-01 2.22031653e-01 -3.43446642e-01 6.49785161e-01
-1.73525020e-01 -6.88319802e-01 6.21456265e-01 7.30057180e-01
7.47848630e-01 -2.74503738e-01 5.29098272e-01 2.95118600e-01
-1.30818617e+00 -1.82803273e-01 2.13046297e-01 -1.61365897e-03
3.04069400e-01 2.49362141e-01 -1.91825897e-01 9.19915497e-01
7.67039895e-01 1.10379231e+00 -5.90666950e-01 1.05749595e+00
6.49800524e-02 6.70095146e-01 -2.76294768e-01 5.01670122e-01
1.33409798e-01 4.75015156e-02 5.73682666e-01 1.26561463e+00
2.07122669e-01 -3.21690559e-01 3.52561444e-01 3.68407160e-01
1.23193584e-01 2.18756616e-01 -5.22706151e-01 -8.67182836e-02
-1.60702497e-01 1.32250512e+00 -4.68448699e-01 -3.83746952e-01
-5.99021137e-01 1.11029041e+00 2.54783213e-01 2.15752110e-01
-1.11776030e+00 1.84505284e-01 7.64862955e-01 3.02673727e-01
1.87216982e-01 -9.41083059e-02 5.24309576e-01 -1.43283558e+00
1.98207065e-01 -1.23030090e+00 8.02504003e-01 -5.95717132e-01
-1.10006773e+00 3.04210961e-01 3.92534077e-01 -1.51834965e+00
-2.03737527e-01 -7.18794286e-01 -2.23124862e-01 2.91860908e-01
-1.26335478e+00 -1.48444188e+00 -5.05005121e-01 6.69051707e-01
8.62541139e-01 1.10135756e-01 6.68898046e-01 8.55029464e-01
-7.85834610e-01 4.66938078e-01 -3.13041300e-01 1.32927746e-01
7.29101598e-01 -1.10929835e+00 -6.14544600e-02 7.38392353e-01
2.63327926e-01 1.34337261e-01 5.21811068e-01 -5.26558340e-01
-1.37125266e+00 -1.28504872e+00 3.46632153e-01 -5.62250137e-01
5.54431438e-01 -3.36017340e-01 -9.20559585e-01 7.59705961e-01
-4.76359129e-02 2.81348646e-01 7.94907749e-01 -3.46179418e-02
1.16587500e-04 -2.08044257e-02 -7.25523412e-01 5.09958267e-01
1.77454615e+00 -2.13293850e-01 -3.90862912e-01 4.94655818e-01
2.78389305e-01 -6.54641688e-01 -1.51401687e+00 7.40881383e-01
6.32618785e-01 -7.46700644e-01 1.09679365e+00 -8.82236421e-01
5.14215887e-01 -2.56977856e-01 1.26589769e-02 -1.06705344e+00
-2.96126276e-01 -3.36645663e-01 -7.70702243e-01 8.39496791e-01
3.75031978e-02 -1.29597321e-01 1.00225174e+00 5.00416905e-02
-6.05463028e-01 -9.47126210e-01 -1.22533464e+00 -1.10367775e+00
4.04595118e-03 -5.11973262e-01 2.05297738e-01 1.09058654e+00
1.99699014e-01 -1.01492666e-01 -9.19720352e-01 -3.04724187e-01
3.11863095e-01 -3.18711549e-02 9.21832979e-01 -8.88557732e-01
-7.93481886e-01 -3.71814847e-01 -8.08265388e-01 -9.44205940e-01
1.44653365e-01 -1.13536382e+00 -2.49916911e-01 -1.29361236e+00
4.17702347e-01 4.28418070e-02 -6.50978923e-01 7.38099396e-01
-5.14285155e-02 7.55744994e-01 2.27171645e-01 1.62474826e-01
-8.69003057e-01 3.98539424e-01 1.32932436e+00 -2.77212650e-01
-2.34673843e-02 1.24697492e-01 -3.45791548e-01 5.37433863e-01
8.57487381e-01 -2.71096677e-01 -4.14247811e-01 -7.89667293e-02
4.60159816e-02 -3.34558375e-02 6.25012636e-01 -1.34204495e+00
-1.12783797e-01 -2.28823751e-01 1.27270808e-02 -4.25728917e-01
6.06395125e-01 -6.45518899e-01 6.94747686e-01 6.37854397e-01
-4.08294052e-01 -1.32767186e-01 1.82182521e-01 7.32315779e-01
-2.07353622e-01 -1.17916418e-02 4.23528820e-01 -2.64507651e-01
-1.22667933e+00 4.02403265e-01 -6.79784179e-01 1.45897582e-01
1.33236969e+00 -6.38943255e-01 -2.99194187e-01 -4.65680333e-03
-1.32904494e+00 2.62166351e-01 2.66602725e-01 7.21291721e-01
2.80997485e-01 -1.51718521e+00 -7.87914932e-01 -1.97643027e-01
6.12745106e-01 -3.85526001e-01 8.22175145e-01 1.62411857e+00
-3.52330089e-01 1.55407965e-01 -4.00854766e-01 -7.22374380e-01
-1.44312942e+00 7.20634282e-01 5.12009323e-01 -5.67166805e-01
-8.24749291e-01 3.67137611e-01 -2.96298191e-02 -3.15809101e-01
1.37133688e-01 -3.43145311e-01 -4.03894544e-01 -1.24005005e-01
3.38606745e-01 5.05471647e-01 1.29684657e-01 -7.62030721e-01
-3.36177260e-01 4.45659697e-01 2.61542320e-01 3.37115347e-01
1.09292555e+00 1.21284230e-02 3.08069825e-01 6.24674976e-01
9.89230931e-01 -2.08468258e-01 -1.35801339e+00 -2.89329171e-01
2.20009629e-02 -3.83934915e-01 -4.69871573e-02 -8.46854091e-01
-1.35544586e+00 6.69313312e-01 9.37546492e-01 -3.52215469e-01
1.37419927e+00 4.63195443e-02 7.38585353e-01 -9.56067443e-03
3.75005186e-01 -1.27011037e+00 2.62509555e-01 -5.80740124e-02
8.08739305e-01 -1.00979400e+00 9.06986892e-02 -4.00467813e-01
-7.33202875e-01 8.39577198e-01 8.67823482e-01 1.15617283e-01
3.50252807e-01 1.55335799e-01 3.76175129e-04 -3.75145316e-01
-7.18094885e-01 -4.79493439e-01 4.72960383e-01 5.73843122e-01
2.78128505e-01 -3.76764201e-02 -6.22419477e-01 6.34568334e-01
2.95787960e-01 6.04385555e-01 3.84982318e-01 1.15950775e+00
-9.05280281e-03 -1.18299890e+00 -8.09893534e-02 5.65313339e-01
-6.35556281e-01 2.39170074e-01 -2.76093692e-01 1.00865030e+00
5.48835516e-01 4.82561022e-01 -2.15262637e-01 -4.80127424e-01
5.90075791e-01 2.45518729e-01 6.89850152e-01 -6.08904600e-01
-7.67346978e-01 -1.18675731e-01 5.36110520e-01 -9.15801764e-01
-1.07623315e+00 -8.02231491e-01 -8.30469906e-01 -1.01030491e-01
-2.02665791e-01 -8.27837288e-02 2.00419232e-01 7.51964092e-01
3.87108952e-01 8.61253202e-01 1.12013139e-01 -9.53135788e-01
-4.74568039e-01 -1.18759716e+00 -7.18007028e-01 8.41675103e-01
-1.14629559e-01 -1.08733642e+00 -1.12029523e-01 2.49362886e-01] | [8.368563652038574, 0.6097179651260376] |
330cac37-3f9f-4f6d-9a12-357119567ee8 | cross-speaker-style-transfer-with-prosody | 2107.12562 | null | https://arxiv.org/abs/2107.12562v1 | https://arxiv.org/pdf/2107.12562v1.pdf | Cross-speaker Style Transfer with Prosody Bottleneck in Neural Speech Synthesis | Cross-speaker style transfer is crucial to the applications of multi-style and expressive speech synthesis at scale. It does not require the target speakers to be experts in expressing all styles and to collect corresponding recordings for model training. However, the performances of existing style transfer methods are still far behind real application needs. The root causes are mainly twofold. Firstly, the style embedding extracted from single reference speech can hardly provide fine-grained and appropriate prosody information for arbitrary text to synthesize. Secondly, in these models the content/text, prosody, and speaker timbre are usually highly entangled, it's therefore not realistic to expect a satisfied result when freely combining these components, such as to transfer speaking style between speakers. In this paper, we propose a cross-speaker style transfer text-to-speech (TTS) model with explicit prosody bottleneck. The prosody bottleneck builds up the kernels accounting for speaking style robustly, and disentangles the prosody from content and speaker timbre, therefore guarantees high quality cross-speaker style transfer. Evaluation result shows the proposed method even achieves on-par performance with source speaker's speaker-dependent (SD) model in objective measurement of prosody, and significantly outperforms the cycle consistency and GMVAE-based baselines in objective and subjective evaluations. | ['Lei He', 'Shifeng Pan'] | 2021-07-27 | null | null | null | null | ['expressive-speech-synthesis'] | ['speech'] | [-5.74422348e-03 -2.65212327e-01 -7.15330094e-02 -4.18971837e-01
-1.06673837e+00 -7.51068056e-01 4.57076997e-01 -4.17694718e-01
-1.22891851e-01 6.82233453e-01 4.60443288e-01 -5.33206500e-02
4.00920719e-01 -3.27594072e-01 -3.92142534e-01 -8.76159966e-01
3.48940462e-01 2.57495433e-01 -7.94054344e-02 -5.67469418e-01
-3.40853930e-01 2.73650438e-01 -1.21259785e+00 2.71982610e-01
8.18155885e-01 1.00728595e+00 3.41786444e-01 9.32587564e-01
-4.07692492e-01 5.50723076e-01 -9.07308638e-01 -4.18290585e-01
-1.89397093e-02 -9.60762203e-01 -4.34585989e-01 3.85065610e-03
1.71276718e-01 -2.04068094e-01 -5.29305786e-02 1.13064790e+00
7.72415817e-01 -1.05130430e-02 7.00615287e-01 -9.86726582e-01
-5.45658767e-01 6.30009592e-01 -3.11653852e-01 1.60079934e-02
2.33006433e-01 2.03859121e-01 1.06244445e+00 -7.70990908e-01
2.47895613e-01 1.50544822e+00 4.41144437e-01 1.03171408e+00
-1.43554139e+00 -7.69024789e-01 6.69372454e-02 -1.58043325e-01
-1.25165403e+00 -9.96912062e-01 1.21251655e+00 -2.64596850e-01
5.58351994e-01 5.55029392e-01 4.71411169e-01 1.71475375e+00
-2.31235027e-02 8.10227633e-01 1.16062033e+00 -3.68081093e-01
1.05790332e-01 6.92918956e-01 -1.12543687e-01 4.82162759e-02
-5.53687811e-01 2.24677816e-01 -8.52244198e-01 3.08477227e-02
7.47272193e-01 -6.87751830e-01 -5.55157959e-01 5.14238961e-02
-1.16050363e+00 6.36548996e-01 4.12666686e-02 4.52423692e-01
-2.00104296e-01 -6.15526177e-02 7.70012140e-01 7.38237023e-01
6.17953360e-01 7.03756213e-02 -4.22952384e-01 -3.92801821e-01
-1.09148955e+00 1.37143940e-01 8.17911446e-01 1.14762747e+00
3.17252457e-01 7.67006099e-01 -9.91985202e-02 1.27410400e+00
3.42890531e-01 8.75175476e-01 5.95451832e-01 -6.81780875e-01
4.19733644e-01 -4.17327322e-02 3.06223314e-02 -5.57630301e-01
-3.50225419e-02 -4.00621027e-01 -1.03357136e+00 1.66457087e-01
2.76191741e-01 -3.41374964e-01 -3.77016753e-01 2.11262417e+00
1.28484011e-01 -2.55341120e-02 1.78494990e-01 9.74182367e-01
7.61553407e-01 1.03925073e+00 1.05458431e-01 -5.29451728e-01
1.34927773e+00 -9.26393449e-01 -1.09864700e+00 -2.89827406e-01
5.12754768e-02 -1.16723692e+00 1.48827004e+00 3.51321042e-01
-1.41458607e+00 -1.07479250e+00 -1.12938511e+00 -5.25623448e-02
6.02710582e-02 2.43413672e-01 7.08228722e-02 8.60621989e-01
-9.67423916e-01 4.46197718e-01 -5.11175752e-01 -6.04812913e-02
-1.67891368e-01 1.93405151e-01 -1.42676592e-01 4.76315320e-01
-1.36801827e+00 8.03204834e-01 2.08741337e-01 4.39951047e-02
-7.45227993e-01 -8.02976012e-01 -7.31107354e-01 5.26330769e-02
-1.12483740e-01 -6.33643031e-01 1.51030850e+00 -1.41151524e+00
-2.26372671e+00 6.07047856e-01 -3.72090906e-01 -1.51369944e-01
4.56599236e-01 -3.28086406e-01 -8.79825950e-01 -7.78338313e-02
-1.86716497e-01 5.61943889e-01 1.24093902e+00 -1.28892052e+00
-4.84620184e-01 -1.08238436e-01 -4.74282831e-01 5.14399827e-01
-5.11738420e-01 4.25162822e-01 -3.21164519e-01 -8.95424902e-01
-2.72485286e-01 -8.00618052e-01 4.04949725e-01 -1.30180240e-01
-1.95593059e-01 -1.04618393e-01 9.83606637e-01 -8.92712295e-01
1.15141797e+00 -2.50120759e+00 4.53038692e-01 -3.57212007e-01
-2.12721348e-01 3.22789043e-01 -1.04702801e-01 3.24030489e-01
-5.82837150e-04 -1.55944481e-01 4.05943170e-02 -7.30333924e-01
1.92773908e-01 2.22113967e-01 -7.04338133e-01 2.06584290e-01
2.88312286e-01 6.76181018e-01 -7.25968361e-01 -4.09177184e-01
2.15519667e-01 7.26935267e-01 -5.06804526e-01 5.24244010e-01
-1.80215418e-01 8.32948387e-01 -2.33678281e-01 2.52648056e-01
4.64159667e-01 3.43053102e-01 7.58160129e-02 -3.73387188e-01
-2.58922223e-02 6.98109865e-01 -1.17767239e+00 1.83187306e+00
-7.64848113e-01 6.09208584e-01 5.87807775e-01 -5.51473260e-01
1.05315769e+00 9.17384863e-01 1.44080967e-01 -4.70769674e-01
1.89600140e-01 3.27469617e-01 1.88848928e-01 -8.39302540e-02
2.62386531e-01 -8.07935536e-01 -2.50900179e-01 1.63945153e-01
4.75979358e-01 -6.14526510e-01 -2.54471064e-01 -2.32893214e-01
3.60729039e-01 1.28482223e-01 1.76980153e-01 -5.45589685e-01
8.48758519e-01 -6.75109148e-01 6.79581821e-01 8.14225972e-02
-1.93343043e-01 6.73833728e-01 2.79948264e-01 9.33355242e-02
-1.01888907e+00 -1.38505673e+00 1.59171686e-01 1.31661654e+00
-1.53477624e-01 -1.77267969e-01 -9.82418656e-01 -2.95830548e-01
-4.31564957e-01 9.84846234e-01 -2.07476735e-01 -1.31434590e-01
-6.83997035e-01 -2.01085702e-01 8.43697011e-01 4.35557634e-01
3.30811322e-01 -1.19462037e+00 6.26984537e-02 4.55539644e-01
-4.14082378e-01 -1.14428127e+00 -1.24835992e+00 1.30503252e-01
-6.47985637e-01 -3.40171486e-01 -8.26361179e-01 -7.55586743e-01
-8.24688450e-02 -6.75594434e-02 9.91513371e-01 -6.10877633e-01
3.06418538e-01 1.11914106e-01 -1.41296625e-01 -5.12286127e-01
-1.00356555e+00 1.08482815e-01 3.76837969e-01 3.95231426e-01
1.01524442e-01 -9.05637801e-01 -3.69553357e-01 3.26854020e-01
-6.74144506e-01 1.31995931e-01 3.27913970e-01 8.80134761e-01
2.63169408e-01 -5.17499782e-02 1.07889998e+00 -4.29450423e-01
7.38458872e-01 -1.51325598e-01 -3.88141870e-01 2.59177666e-02
-2.14606255e-01 -2.14558411e-02 9.36511219e-01 -6.97770059e-01
-1.43418622e+00 -2.80281574e-01 -5.59882224e-01 -5.42520285e-01
-3.25247705e-01 -6.99402988e-02 -8.02996576e-01 3.35403711e-01
6.22884929e-01 4.58574802e-01 -1.56470519e-02 -6.09823644e-01
3.93469125e-01 1.10308611e+00 5.89894652e-01 -7.29938030e-01
7.42090404e-01 4.83045839e-02 -4.80602920e-01 -1.32934129e+00
-6.93984032e-01 -3.99612010e-01 -4.64168280e-01 -1.24318767e-02
8.96778822e-01 -1.18438864e+00 -5.47186494e-01 6.44668639e-01
-1.42431867e+00 -4.27736402e-01 -4.04114872e-01 5.79049408e-01
-8.95849586e-01 2.79063344e-01 -8.94402683e-01 -9.35897648e-01
-5.59277415e-01 -1.23463178e+00 1.11575806e+00 7.29616582e-02
-4.54528302e-01 -1.21151137e+00 -3.71765941e-02 2.33723477e-01
7.72186160e-01 -8.03310424e-02 8.12296510e-01 -4.37770158e-01
-9.13878232e-02 2.85963237e-01 6.81534559e-02 9.73014951e-01
5.82741320e-01 -2.17686668e-01 -1.59592271e+00 -3.63622993e-01
6.06540322e-01 -3.27503443e-01 2.11775869e-01 3.07934672e-01
6.10511303e-01 -3.75593156e-01 2.63169140e-01 5.49114287e-01
9.45565820e-01 2.12908596e-01 4.41703409e-01 -4.88308012e-01
7.07904518e-01 6.56931996e-01 2.08353326e-01 1.68426529e-01
8.12388510e-02 9.00710642e-01 -2.12748155e-01 -9.33814794e-02
-6.12088084e-01 -3.21668506e-01 9.63650823e-01 1.70102787e+00
1.15398370e-01 -3.08661073e-01 -2.96554595e-01 4.35744852e-01
-1.22942615e+00 -9.68526125e-01 6.42081350e-02 2.20428395e+00
1.38323426e+00 1.76417574e-01 2.68402249e-01 2.66190648e-01
6.26656890e-01 2.56081522e-01 -4.63828474e-01 -7.13998556e-01
-2.21196353e-01 2.51247138e-01 -3.85874435e-02 6.73909605e-01
-7.85746396e-01 9.67076957e-01 5.69033480e+00 1.24893296e+00
-1.46415854e+00 3.39647114e-01 3.73454303e-01 -2.72640228e-01
-4.03724283e-01 -2.98141420e-01 -9.13660824e-01 4.90702033e-01
1.42916954e+00 -3.34730834e-01 7.99811959e-01 6.59759760e-01
4.31599170e-01 5.24879336e-01 -1.39602828e+00 1.09839010e+00
-2.34037321e-02 -7.26884961e-01 -1.44416066e-02 -1.24050878e-01
4.62674350e-01 -2.92798281e-02 1.78268775e-01 5.31104088e-01
1.52105722e-03 -9.45696115e-01 1.13479269e+00 1.90380454e-01
1.13512123e+00 -7.11515307e-01 2.85069913e-01 4.69553083e-01
-1.28735638e+00 2.83763021e-01 -1.22347288e-01 2.46777982e-01
3.42481226e-01 3.37457031e-01 -6.18930578e-01 6.26512289e-01
2.64120609e-01 4.25674319e-01 3.08078621e-02 3.68869811e-01
-1.23496175e-01 1.01853001e+00 -1.24758363e-01 1.29038274e-01
-5.77636296e-03 -1.61985114e-01 7.95259297e-01 1.50585496e+00
3.95772487e-01 -1.83989972e-01 -5.63739873e-02 9.41684008e-01
-1.07535617e-02 2.84775317e-01 -4.40815598e-01 -2.45348081e-01
3.67374688e-01 1.10390759e+00 -1.20427094e-01 -2.16231123e-01
-5.05924463e-01 1.37849760e+00 -4.59377393e-02 4.83773053e-01
-8.15381467e-01 -1.75072983e-01 1.08155417e+00 -1.38954923e-01
7.12066144e-02 -2.36881182e-01 -3.95902455e-01 -1.16875815e+00
-2.27706656e-02 -1.25644791e+00 1.30822202e-02 -6.82224751e-01
-1.44444954e+00 8.98387611e-01 -2.87696630e-01 -1.20299375e+00
-5.24594545e-01 -5.13441086e-01 -7.14160621e-01 1.60319221e+00
-1.42907143e+00 -1.37247610e+00 8.88206661e-02 6.89866841e-01
1.10546446e+00 -2.36630514e-01 1.14771283e+00 4.48709339e-01
-3.43935788e-01 8.89475346e-01 -4.28572781e-02 -6.07931521e-03
1.04123163e+00 -1.30280185e+00 4.31739539e-01 5.63786745e-01
1.47861481e-01 4.23058093e-01 8.98710608e-01 -1.91975743e-01
-1.30643117e+00 -1.02358925e+00 8.75734985e-01 -2.96649098e-01
6.83453500e-01 -6.65369630e-01 -1.15862250e+00 4.11184639e-01
6.37230873e-01 -1.27974674e-01 7.98838139e-01 7.39854714e-03
-3.72005463e-01 -4.20947850e-01 -8.46786797e-01 6.84397340e-01
6.59073114e-01 -8.96546900e-01 -6.49876177e-01 -4.28686440e-02
9.35243309e-01 -2.18346134e-01 -8.40267658e-01 1.08259834e-01
4.65889424e-01 -1.00357389e+00 8.54442716e-01 -3.27932447e-01
1.16205536e-01 -2.05753863e-01 -4.83214587e-01 -1.70989692e+00
-1.16914250e-01 -9.96433377e-01 1.43063262e-01 1.92960441e+00
3.68414223e-01 -3.83791029e-01 1.21219262e-01 2.99197435e-01
-3.26850474e-01 -9.88529995e-02 -8.73246908e-01 -1.02384484e+00
4.01546896e-01 -4.80104566e-01 6.56560957e-01 8.70944440e-01
-5.92918396e-02 8.58758807e-01 -6.01487160e-01 2.14232296e-01
5.44003785e-01 -1.20351473e-02 6.90480173e-01 -9.35285926e-01
-7.32931912e-01 -7.73199618e-01 1.68881923e-01 -1.03237283e+00
2.60445982e-01 -7.07420290e-01 1.41142905e-01 -9.85889614e-01
-3.08931559e-01 -4.23638672e-01 -3.10530514e-01 -2.13792417e-02
-1.91172719e-01 -1.08489484e-01 3.43191385e-01 -1.48440292e-02
-3.60616657e-04 9.91551578e-01 1.39306283e+00 -5.74116968e-02
-4.44841266e-01 3.02805603e-01 -6.40468895e-01 7.53367424e-01
8.52057219e-01 -1.70677394e-01 -6.24650538e-01 -2.62595206e-01
-3.86287987e-01 6.62794173e-01 7.14258477e-02 -8.56675506e-01
-7.93692246e-02 -2.09620759e-01 6.26180246e-02 -1.94220155e-01
6.72717929e-01 -7.82242954e-01 2.21901119e-01 1.08962290e-01
-4.04071569e-01 -3.57516706e-01 4.90923494e-01 2.58734941e-01
-4.97613072e-01 -6.55390397e-02 1.11000836e+00 7.29739517e-02
-1.98755547e-01 1.09772548e-01 -2.46077105e-01 1.75338179e-01
4.18646246e-01 1.10677116e-01 1.20798446e-01 -4.79936242e-01
-7.18547285e-01 -2.84123421e-01 2.54180491e-01 6.77281857e-01
2.51320630e-01 -1.66623831e+00 -9.96871293e-01 4.67256069e-01
-3.74277718e-02 -2.57352978e-01 4.61903781e-01 5.66484332e-01
9.16522071e-02 3.55166107e-01 -9.34093148e-02 -5.25012314e-01
-1.24781883e+00 5.31970084e-01 3.71861160e-01 2.70595513e-02
-3.21687698e-01 8.02708924e-01 7.27612972e-01 -3.47878933e-01
2.70200640e-01 -2.77733505e-01 1.00394167e-01 1.25832692e-01
6.84309602e-01 1.94275588e-01 6.55859113e-02 -8.19532514e-01
-5.46374135e-02 4.16139632e-01 1.25247777e-01 -5.15436888e-01
1.02527595e+00 -4.37711775e-01 1.26781151e-01 1.14487147e+00
1.33859861e+00 4.73945409e-01 -1.33927321e+00 -2.90456742e-01
-3.37873012e-01 -4.48417142e-02 1.31976962e-01 -7.77911007e-01
-8.20806086e-01 1.33707857e+00 4.65249807e-01 2.53270209e-01
1.28723145e+00 -3.27035069e-01 9.59992766e-01 -1.53208554e-01
2.39876226e-01 -1.11872351e+00 6.81759492e-02 4.79314178e-01
1.30806613e+00 -9.89317238e-01 -7.64207721e-01 -3.13202649e-01
-9.64314759e-01 1.04386818e+00 4.27694917e-01 1.51526406e-01
6.49006844e-01 4.93987471e-01 4.93218541e-01 4.26357806e-01
-8.26984167e-01 9.04854238e-02 6.81644082e-01 5.21379530e-01
7.17274249e-01 2.27864742e-01 1.43483639e-01 7.70956457e-01
-5.23371518e-01 -3.44581455e-01 1.18393384e-01 1.05512939e-01
-3.85754287e-01 -1.40838325e+00 -4.53465492e-01 -1.62174255e-01
-5.41921437e-01 -1.78844422e-01 -2.37422585e-01 3.45405370e-01
-9.07331482e-02 1.08826888e+00 -2.02334911e-01 -2.96732455e-01
4.39039230e-01 5.54386497e-01 4.50693786e-01 -5.25140107e-01
-7.04768479e-01 7.58139312e-01 1.24438159e-01 -2.72285342e-01
-1.45940661e-01 -5.86690307e-01 -9.89288688e-01 -2.93151975e-01
-2.37167031e-01 1.32504910e-01 6.73842192e-01 8.22111428e-01
1.05266608e-01 7.38065064e-01 9.11046505e-01 -8.66625845e-01
-8.46923769e-01 -1.28614497e+00 -8.26296985e-01 4.59802628e-01
5.68091750e-01 -2.43669733e-01 -5.13802230e-01 3.23822886e-01] | [14.939105987548828, 6.554594039916992] |
78cbcd19-c7fd-44ad-bdd4-db2edff8f52c | characterization-and-learning-of-causal-1 | 2301.09028 | null | https://arxiv.org/abs/2301.09028v1 | https://arxiv.org/pdf/2301.09028v1.pdf | Characterization and Learning of Causal Graphs with Small Conditioning Sets | Constraint-based causal discovery algorithms learn part of the causal graph structure by systematically testing conditional independences observed in the data. These algorithms, such as the PC algorithm and its variants, rely on graphical characterizations of the so-called equivalence class of causal graphs proposed by Pearl. However, constraint-based causal discovery algorithms struggle when data is limited since conditional independence tests quickly lose their statistical power, especially when the conditioning set is large. To address this, we propose using conditional independence tests where the size of the conditioning set is upper bounded by some integer $k$ for robust causal discovery. The existing graphical characterizations of the equivalence classes of causal graphs are not applicable when we cannot leverage all the conditional independence statements. We first define the notion of $k$-Markov equivalence: Two causal graphs are $k$-Markov equivalent if they entail the same conditional independence constraints where the conditioning set size is upper bounded by $k$. We propose a novel representation that allows us to graphically characterize $k$-Markov equivalence between two causal graphs. We propose a sound constraint-based algorithm called the $k$-PC algorithm for learning this equivalence class. Finally, we conduct synthetic, and semi-synthetic experiments to demonstrate that the $k$-PC algorithm enables more robust causal discovery in the small sample regime compared to the baseline PC algorithm. | ['Murat Kocaoglu'] | 2023-01-22 | null | null | null | null | ['causal-discovery'] | ['knowledge-base'] | [ 2.48419389e-01 1.42310292e-01 -6.95566237e-01 -2.97883779e-01
-4.32865053e-01 -7.59673238e-01 4.37462986e-01 2.97459632e-01
3.78816389e-02 9.40501392e-01 3.13995272e-01 -8.10525239e-01
-9.16535318e-01 -1.04597795e+00 -8.92656267e-01 -4.72141922e-01
-9.04432118e-01 5.47962129e-01 1.88807547e-01 3.26706529e-01
1.38079703e-01 2.72042871e-01 -1.21456110e+00 -8.04689005e-02
9.97022927e-01 2.44136289e-01 -2.01165929e-01 6.32053375e-01
2.66891092e-01 5.38719237e-01 -1.96001008e-01 -4.11353298e-02
1.36485651e-01 -8.05381417e-01 -9.07364964e-01 -3.03873539e-01
2.58425534e-01 -1.80811137e-01 -2.64235228e-01 1.02986646e+00
3.31559815e-02 -3.25894058e-02 7.89955556e-01 -1.81102252e+00
-2.75738031e-01 1.20096493e+00 -7.42865086e-01 3.21490586e-01
6.79831862e-01 -1.38101831e-01 1.46804118e+00 -3.94043803e-01
6.31604731e-01 1.59472430e+00 3.71744871e-01 2.47982100e-01
-1.65613830e+00 -1.06171608e+00 3.21391881e-01 3.64384890e-01
-1.45382071e+00 1.07895173e-02 6.71999216e-01 -5.45538187e-01
6.25071526e-01 3.86562169e-01 5.32172143e-01 8.40818405e-01
-1.79143809e-02 5.74138701e-01 1.29466426e+00 -4.92165387e-01
5.12879550e-01 -5.02756774e-01 4.88692790e-01 8.89629602e-01
7.29689837e-01 4.17684793e-01 -6.13292634e-01 -6.21298671e-01
8.96612585e-01 -2.49585539e-01 -3.24870884e-01 -5.54358542e-01
-1.02021098e+00 1.07325053e+00 2.42014557e-01 7.66827241e-02
1.11911878e-01 5.01656950e-01 9.07611474e-02 4.09908772e-01
-1.08876191e-01 3.06019694e-01 -4.45926964e-01 2.29502141e-01
-7.49138057e-01 2.41509661e-01 8.54889393e-01 1.01098597e+00
6.34429038e-01 -2.93491185e-01 -8.27959105e-02 1.38178572e-01
5.29188573e-01 5.48447847e-01 -3.28137934e-01 -7.91798413e-01
2.87984014e-01 7.00677276e-01 2.00317100e-01 -8.59886467e-01
-3.04058254e-01 2.35934909e-02 -8.31738234e-01 -5.07151186e-02
6.05141819e-01 -1.44822851e-01 -8.80340099e-01 2.25982881e+00
3.80756557e-01 5.37264347e-01 -1.86518267e-01 6.10645115e-01
3.04264367e-01 3.14237714e-01 2.24848002e-01 -6.09216452e-01
1.03326690e+00 -3.29605266e-02 -5.34099519e-01 9.05712470e-02
4.37456042e-01 -3.16998869e-01 1.15520597e+00 -5.60013726e-02
-6.28576338e-01 3.59760178e-03 -1.11223185e+00 5.84394634e-01
2.12098900e-02 -5.76341152e-01 1.01296628e+00 8.92755330e-01
-6.65840864e-01 4.34997052e-01 -9.51559722e-01 -2.39793167e-01
2.70423740e-01 3.51342201e-01 -3.75295609e-01 -4.88082647e-01
-1.27025735e+00 3.47044736e-01 5.38964212e-01 -2.43433878e-01
-1.21353161e+00 -8.68690491e-01 -7.88719952e-01 1.57989025e-01
9.25823390e-01 -6.48018241e-01 8.72749925e-01 -4.70550776e-01
-7.67247736e-01 4.57607836e-01 -2.30228826e-01 -1.27372503e-01
3.30399454e-01 -1.82619896e-02 -4.78691906e-01 6.82048127e-02
1.98388085e-01 -6.17375337e-02 6.54457211e-01 -1.30434191e+00
-7.01196551e-01 -4.38967228e-01 2.99095720e-01 -2.66698509e-01
1.54902548e-01 7.66342506e-02 -3.10759395e-01 -3.70021313e-01
2.06979960e-01 -8.40974033e-01 -3.73083740e-01 -3.55791867e-01
-8.75614345e-01 -3.28916818e-01 6.49354160e-01 -9.86040011e-02
1.51583827e+00 -1.98053455e+00 -3.33875418e-03 7.58766174e-01
3.58041495e-01 -4.57698315e-01 -1.13320775e-01 4.29708064e-01
-5.34055531e-01 5.92402279e-01 -4.97062206e-01 3.95045131e-01
-7.86454044e-03 3.65637809e-01 -2.90716141e-01 5.76247573e-01
2.22483978e-01 6.41326129e-01 -1.14667892e+00 -4.93674189e-01
-1.13920309e-01 -1.83620661e-01 -7.80956209e-01 2.44827464e-01
-4.00812805e-01 3.40143293e-01 -5.37223935e-01 5.60652614e-01
5.14558434e-01 -3.29220861e-01 8.83689702e-01 1.17001787e-01
-4.61693108e-02 2.49191880e-01 -1.64453232e+00 1.19420326e+00
1.21994331e-01 2.68262774e-01 -2.37468317e-01 -1.15096128e+00
5.13068020e-01 3.62984985e-01 4.55520123e-01 -3.72415870e-01
-1.66122347e-01 3.36210169e-02 2.68940479e-01 -2.79470682e-01
-2.53275245e-01 -4.75231558e-01 -3.82384390e-01 6.86297953e-01
-2.74156690e-01 1.57878011e-01 6.01863682e-01 5.77408671e-01
1.71229541e+00 -1.08825907e-01 6.07262433e-01 -6.12505496e-01
9.09877345e-02 3.80940400e-02 1.02899134e+00 1.14699471e+00
-3.37168295e-03 1.84506103e-01 1.16885471e+00 -5.08825332e-02
-8.07556450e-01 -1.63150442e+00 -2.00112879e-01 7.85208821e-01
8.77975523e-02 -7.97940493e-01 -3.07923108e-01 -9.76124883e-01
2.04956204e-01 7.28444695e-01 -8.42155874e-01 -2.54988551e-01
-4.84012067e-01 -9.73160386e-01 4.44016278e-01 6.18594229e-01
2.11977586e-01 -5.54306686e-01 -4.95132327e-01 4.93310280e-02
-7.86001906e-02 -8.07690501e-01 -2.67097682e-01 3.89559865e-01
-8.65999341e-01 -1.86719787e+00 1.80816818e-02 -2.71705598e-01
7.54772246e-01 -3.78679633e-02 1.05898690e+00 -1.05481550e-01
-1.87367424e-01 4.81738716e-01 -3.10173124e-01 -1.77630693e-01
-3.02341312e-01 -3.91235739e-01 1.44017383e-01 -3.35965365e-01
1.81952029e-01 -9.22459245e-01 -3.17219943e-01 3.42110604e-01
-8.81808698e-01 -1.92159131e-01 5.88066876e-01 7.19786048e-01
6.65873408e-01 5.65333784e-01 4.09327477e-01 -1.04378772e+00
5.17747819e-01 -6.74163163e-01 -8.00991833e-01 3.77345353e-01
-7.16649055e-01 5.34455597e-01 4.28929448e-01 -3.74619305e-01
-7.65941322e-01 2.88385735e-03 4.69286263e-01 -2.37416431e-01
-6.01347834e-02 1.18925273e+00 -5.53508580e-01 4.63276327e-01
5.44305503e-01 -4.61433679e-01 -6.79684699e-01 -2.89356470e-01
5.59764445e-01 -5.06838299e-02 5.59781253e-01 -8.40976596e-01
9.65327084e-01 5.36964774e-01 4.39071149e-01 -4.33591992e-01
-6.95871532e-01 -3.56627464e-01 -6.08734667e-01 2.23235856e-03
5.81492126e-01 -4.90988702e-01 -1.10227609e+00 -2.85196066e-01
-8.03959191e-01 -4.89479035e-01 -2.18649104e-01 5.36980689e-01
-5.61099589e-01 3.82143259e-01 -1.50249183e-01 -9.59906876e-01
2.34896436e-01 -7.37758517e-01 5.33431828e-01 -1.20420352e-01
-4.77884501e-01 -9.14809763e-01 5.60530186e-01 -2.06279337e-01
-3.63224834e-01 5.37375927e-01 1.80030000e+00 -4.85510409e-01
-4.03670728e-01 -1.87165797e-01 -4.29498047e-01 -3.83351117e-01
1.78683147e-01 2.58715659e-01 -3.08006942e-01 -2.62682706e-01
-5.97172737e-01 -7.70569721e-04 8.34652662e-01 5.67549527e-01
9.09027338e-01 -4.43776697e-01 -8.20964932e-01 1.65757328e-01
1.41609824e+00 2.12370574e-01 4.61291760e-01 -1.48744330e-01
6.22346520e-01 4.21990037e-01 3.70437205e-01 4.36181664e-01
2.46419251e-01 4.36603099e-01 4.65952933e-01 8.51747617e-02
1.42235056e-01 -6.08998120e-01 2.42017210e-01 3.76114398e-01
-2.18900442e-01 -2.63464212e-01 -1.17908621e+00 5.02389491e-01
-1.95868576e+00 -1.01779544e+00 -5.20413101e-01 2.56858468e+00
1.16311693e+00 3.00182372e-01 2.38315344e-01 3.67829263e-01
8.00549746e-01 -2.48068690e-01 -4.47212994e-01 -1.54759392e-01
-2.94556078e-02 5.19129455e-01 5.18817544e-01 7.13436127e-01
-8.85165334e-01 5.76196432e-01 7.16396236e+00 5.60703754e-01
-5.56368053e-01 1.39995022e-02 2.39377096e-01 -2.54137833e-02
-5.62837183e-01 6.80878580e-01 -6.18045211e-01 2.82836229e-01
1.05731130e+00 -5.08497655e-01 7.25354105e-02 6.61991060e-01
3.60353231e-01 -3.44008058e-01 -1.63459420e+00 6.19083762e-01
-5.01830995e-01 -1.19231462e+00 5.64456619e-02 1.34337753e-01
7.97729492e-01 -3.26784760e-01 -2.25215748e-01 5.36619686e-03
1.40708160e+00 -1.16710401e+00 4.34811741e-01 2.31648803e-01
1.06435776e+00 -7.78719723e-01 4.02699739e-01 1.00859754e-01
-1.34219813e+00 -2.73443639e-01 -1.11762069e-01 -2.57396072e-01
1.61316365e-01 1.05467582e+00 -7.75851309e-01 7.82858968e-01
7.75714934e-01 5.35828233e-01 -2.48781890e-01 9.93439496e-01
-7.02487230e-01 1.08002293e+00 -4.25902426e-01 2.16231108e-01
-7.69503117e-02 -5.56861199e-02 5.35505593e-01 1.10754812e+00
1.07647061e-01 4.05509502e-01 2.30806023e-01 1.01392066e+00
3.53729054e-02 -2.92490393e-01 -6.10572517e-01 -2.45175689e-01
6.77088678e-01 5.89752734e-01 -8.24263930e-01 -1.36122286e-01
-3.45419586e-01 4.50110406e-01 2.23635420e-01 3.75800252e-01
-8.05870295e-01 -5.17534837e-03 5.17027020e-01 1.23476572e-01
4.80356487e-03 -3.65195334e-01 -1.73531383e-01 -9.78633165e-01
-3.65442932e-01 -7.64873207e-01 1.20295119e+00 -4.27000165e-01
-1.43212330e+00 -1.42396197e-01 5.78764260e-01 -5.83106875e-01
-2.51946211e-01 -4.14424062e-01 -6.91677511e-01 6.34942889e-01
-1.10474217e+00 -8.16517353e-01 -2.90490012e-03 9.85727906e-01
-1.45266965e-01 5.39633334e-01 8.60930860e-01 1.22746860e-04
-8.55280042e-01 3.75857472e-01 -8.21372271e-02 1.40048638e-01
5.27110636e-01 -1.50448501e+00 -1.57670692e-01 1.16056502e+00
7.50436336e-02 9.09668684e-01 9.31723058e-01 -1.09335566e+00
-1.39906907e+00 -9.16142046e-01 6.66688621e-01 -2.51798272e-01
9.64054704e-01 -3.92693549e-01 -7.02919126e-01 1.04837763e+00
-1.76087976e-01 -2.14700103e-02 9.13306832e-01 8.71354580e-01
-8.57273936e-01 -3.72606739e-02 -7.53593683e-01 8.20633292e-01
1.24742270e+00 -4.04074937e-01 -6.82248056e-01 1.53286189e-01
6.58747852e-01 2.65147597e-01 -8.50797772e-01 7.84511447e-01
6.25054598e-01 -7.41406560e-01 7.02770174e-01 -1.05717254e+00
4.79054064e-01 -5.16957760e-01 -2.99404413e-01 -1.03330588e+00
-5.29939711e-01 -5.49276114e-01 -8.47895965e-02 9.41304564e-01
6.35881722e-01 -4.17661697e-01 6.26061082e-01 5.66906214e-01
3.24738503e-01 -1.93357930e-01 -1.09589708e+00 -9.65350270e-01
4.63447422e-02 -7.47429729e-01 4.67326820e-01 1.42064738e+00
5.43615937e-01 4.29285705e-01 -1.72488302e-01 5.15185356e-01
9.33003306e-01 5.02528191e-01 7.05041051e-01 -1.49893153e+00
-5.90025604e-01 -3.89299631e-01 -1.45470873e-01 -3.77614528e-01
1.71584666e-01 -8.37585747e-01 -1.15650013e-01 -1.21212542e+00
9.19706881e-01 -7.86892056e-01 -2.63060629e-01 7.18072176e-01
-3.40520084e-01 -1.46908671e-01 -1.67958379e-01 1.18474802e-02
-3.31542909e-01 2.36646250e-01 9.25208032e-01 -1.22830771e-01
-3.04793298e-01 -1.49620742e-01 -5.83055913e-01 7.33188272e-01
5.44535160e-01 -6.95580482e-01 -6.75511301e-01 1.57356605e-01
6.84673965e-01 5.00535786e-01 5.48892081e-01 -3.87099147e-01
9.18052047e-02 -7.98768997e-01 1.41835976e-02 -4.32446033e-01
-4.69804615e-01 -7.42077470e-01 7.07860768e-01 6.82049453e-01
-5.31763434e-01 -1.87939554e-01 1.21584721e-01 8.91455948e-01
7.46994466e-02 1.05783246e-01 6.01835847e-01 1.05610318e-01
-4.87780064e-01 2.43941367e-01 -2.69235820e-01 1.06699973e-01
9.47173834e-01 1.04893915e-01 -3.89773309e-01 -3.55546176e-01
-8.34991932e-01 3.41082603e-01 8.70741680e-02 1.51721522e-01
4.76418704e-01 -1.25803888e+00 -7.99932480e-01 -4.64753099e-02
2.38313347e-01 -2.02694461e-01 -9.16133448e-02 1.01849318e+00
5.06742075e-02 3.77044767e-01 1.80911757e-02 -4.06033605e-01
-1.19710171e+00 8.71269941e-01 4.07742448e-02 -5.10780811e-01
-4.65815753e-01 6.07385933e-01 4.99186993e-01 -6.12745956e-02
-6.76077157e-02 -3.12748075e-01 1.72728732e-01 -1.85109749e-01
3.92574519e-01 4.26473379e-01 -4.42503035e-01 2.92444825e-02
-6.46302819e-01 2.36245751e-01 1.61681250e-01 -3.46381664e-01
1.13192439e+00 5.49546964e-02 -3.40311021e-01 4.28384304e-01
7.33648896e-01 3.44189733e-01 -9.96483386e-01 -1.34162664e-01
2.79157281e-01 -4.18288380e-01 -2.50362337e-01 -9.91339743e-01
-8.07614326e-01 4.39178586e-01 1.10764988e-01 2.93053359e-01
1.14727199e+00 4.46437448e-01 -2.00245962e-01 1.16220154e-01
3.60663414e-01 -5.28698027e-01 -1.00912161e-01 2.75155336e-01
6.71379626e-01 -8.60847950e-01 2.34933242e-01 -9.01226640e-01
3.52756456e-02 8.15177321e-01 4.01271313e-01 -3.70075218e-02
7.53035784e-01 5.42162001e-01 -6.19527757e-01 -4.74233598e-01
-8.10941458e-01 -3.11547190e-01 2.28410468e-01 5.64893067e-01
3.17404300e-01 5.03909290e-01 -5.59112310e-01 8.28854382e-01
-2.05410957e-01 -2.29702175e-01 5.07697284e-01 9.15672600e-01
-3.11922222e-01 -1.02487624e+00 -5.62154949e-01 4.36065316e-01
-2.37772956e-01 -2.60990322e-01 -6.05263054e-01 1.09535074e+00
-9.34478119e-02 1.20971835e+00 8.09988305e-02 -2.47455388e-01
1.94966763e-01 1.54182054e-02 7.09722996e-01 -5.85084021e-01
2.29063645e-01 1.53623730e-01 1.05390579e-01 -7.78806031e-01
-6.03463829e-01 -7.40942478e-01 -1.43053401e+00 -6.10532045e-01
-4.03336465e-01 3.56640190e-01 2.28841063e-02 1.16277874e+00
-2.86906287e-02 3.94527853e-01 5.09929359e-01 -1.25136049e-02
-2.03279898e-01 -6.75429761e-01 -9.38660264e-01 2.03799114e-01
2.05169003e-02 -9.15468395e-01 -4.82769519e-01 2.52500474e-01] | [7.8135528564453125, 5.339077949523926] |
2f1a8463-f374-4a1c-9ea4-fa93f348a3ea | statik-structure-and-text-for-inductive | null | null | https://aclanthology.org/2022.findings-naacl.46 | https://aclanthology.org/2022.findings-naacl.46.pdf | StATIK: Structure and Text for Inductive Knowledge Graph Completion | Knowledge graphs (KGs) often represent knowledge bases that are incomplete. Machine learning models can alleviate this by helping automate graph completion. Recently, there has been growing interest in completing knowledge bases that are dynamic, where previously unseen entities may be added to the KG with many missing links. In this paper, we present StATIK–Structure And Text for Inductive Knowledge Completion. StATIK uses Language Models to extract the semantic information from text descriptions, while using Message Passing Neural Networks to capture the structural information. StATIK achieves state of the art results on three challenging inductive baselines. We further analyze our hybrid model through detailed ablation studies. | ['Greg Ver Steeg', 'Aram Galstyan', 'Murali Annavaram', 'Mehrnoosh Mirtaheri', 'Keshav Balasubramanian', 'Elan Markowitz'] | null | null | null | null | findings-naacl-2022-7 | ['inductive-knowledge-graph-completion'] | ['knowledge-base'] | [-2.84500457e-02 9.32417870e-01 -6.06726587e-01 -1.15929730e-01
-7.14878500e-01 -7.69166708e-01 5.67910671e-01 6.85867429e-01
-3.47391337e-01 1.02067149e+00 6.84975564e-01 -3.05672348e-01
-1.77752420e-01 -1.15400541e+00 -9.24282551e-01 1.14127301e-01
-2.77295977e-01 8.21124613e-01 8.37687999e-02 -1.98796898e-01
-3.11116666e-01 5.77964485e-02 -9.21949625e-01 3.94589633e-01
8.16677690e-01 1.61369503e-01 -3.79304215e-02 5.75829804e-01
-5.57469070e-01 1.51285052e+00 -2.50142425e-01 -9.45698678e-01
-1.33820936e-01 2.29482844e-01 -1.52301526e+00 -3.83695275e-01
4.04326975e-01 -2.38509655e-01 -9.28440511e-01 8.39392364e-01
1.21931799e-01 9.50053334e-02 6.29367590e-01 -1.39862943e+00
-7.40926147e-01 1.50073361e+00 -2.98446298e-01 -7.84335881e-02
7.07347810e-01 -1.39920339e-01 1.50572860e+00 -1.03151131e+00
1.07585776e+00 1.30380201e+00 7.99346924e-01 3.25055510e-01
-1.23067784e+00 -5.25988519e-01 3.75277251e-01 5.71067631e-01
-1.61326098e+00 -4.50330555e-01 6.40166938e-01 -3.07799548e-01
1.45627236e+00 2.64108647e-02 6.01169169e-01 1.01733828e+00
-3.29209089e-01 1.16694868e+00 3.17681998e-01 -6.23905241e-01
-1.53682441e-01 4.65269834e-02 5.47051966e-01 1.13683641e+00
8.24964285e-01 -3.42176437e-01 -6.29174888e-01 -2.72548974e-01
3.08849186e-01 -3.04692417e-01 -1.29047185e-01 -4.53251660e-01
-9.65161443e-01 6.82512999e-01 4.62214857e-01 -9.57510173e-02
-2.29152203e-01 4.51611459e-01 4.37506914e-01 2.86128372e-01
3.42368037e-01 6.78260207e-01 -7.02524006e-01 5.87621070e-02
-5.04369915e-01 2.43353039e-01 1.39542198e+00 1.51312554e+00
1.10547209e+00 -1.78476363e-01 -1.34497836e-01 6.35186374e-01
3.38597149e-01 3.40159893e-01 -9.60357636e-02 -6.42455459e-01
7.60300159e-01 8.95889938e-01 6.56690225e-02 -8.33262026e-01
-5.22929132e-01 -4.41608399e-01 -6.18121505e-01 -5.36050975e-01
7.64403567e-02 -2.58407116e-01 -1.10469353e+00 1.68616688e+00
2.24263251e-01 2.33413264e-01 4.24161166e-01 2.83276558e-01
1.19407284e+00 4.10839200e-01 5.37656307e-01 1.25129238e-01
1.16934133e+00 -1.04890180e+00 -7.31414557e-01 -5.28646290e-01
1.22739124e+00 -2.37185776e-01 8.45040262e-01 5.45019992e-02
-9.04113114e-01 1.23732552e-01 -8.52069676e-01 -3.92724752e-01
-7.73622811e-01 7.57361390e-03 9.81314182e-01 3.42398018e-01
-1.27539206e+00 4.04967517e-01 -8.78014028e-01 -4.56932127e-01
4.16831970e-01 3.30582231e-01 -5.73135018e-01 -5.11941969e-01
-1.77818787e+00 9.45716321e-01 1.23722494e+00 6.64601028e-02
-9.59750414e-01 -8.19919586e-01 -1.37765050e+00 1.43770084e-01
8.64090443e-01 -1.22205389e+00 1.15487695e+00 -3.83285016e-01
-8.56185555e-01 6.06187344e-01 -2.38721356e-01 -4.61068898e-01
-7.29731172e-02 -2.87016422e-01 -4.71715748e-01 -8.00283179e-02
8.44918787e-02 6.83977365e-01 5.09445012e-01 -1.29362750e+00
-5.24916768e-01 -2.67641455e-01 5.83812535e-01 2.50143737e-01
-3.96902949e-01 -2.78301030e-01 -9.11899090e-01 -4.06548977e-01
-2.86813341e-02 -8.02241147e-01 -1.39019251e-01 -6.95250988e-01
-1.03483570e+00 -3.66166741e-01 5.82929969e-01 -8.46425056e-01
1.51666701e+00 -1.63913298e+00 3.34633589e-01 4.78471220e-01
7.46248066e-01 1.01175606e-01 -4.15125608e-01 8.34017992e-01
1.01148926e-01 4.35081214e-01 -4.92611751e-02 -2.93148637e-01
8.72683525e-02 5.13787031e-01 -3.18759233e-01 -1.93179950e-01
1.58019483e-01 1.53148293e+00 -1.21576524e+00 -6.54580116e-01
-2.50732780e-01 8.64297897e-02 -5.54583251e-01 -3.56460772e-02
-6.73998356e-01 -2.61572123e-01 -6.09747529e-01 8.51800084e-01
3.99195850e-01 -5.90262473e-01 5.89274526e-01 -4.38175440e-01
6.01669431e-01 4.20444369e-01 -1.08423054e+00 1.77999592e+00
-3.17784280e-01 3.81190747e-01 6.84799533e-03 -7.56264627e-01
2.62858570e-01 1.95660859e-01 1.91199839e-01 -1.32356614e-01
-4.21208650e-01 -1.69127241e-01 -3.90002400e-01 -4.93808180e-01
8.28797996e-01 -5.20326085e-02 -1.08035065e-01 4.06616092e-01
3.81421477e-01 2.65185237e-02 5.57142198e-01 1.22107601e+00
1.58817565e+00 9.69914719e-02 3.95998687e-01 2.62789339e-01
2.61719912e-01 4.07627493e-01 4.12640393e-01 9.58372951e-01
3.61818194e-01 -8.08059797e-02 6.28724575e-01 -1.78118870e-01
-8.02384257e-01 -1.17808163e+00 5.16147912e-01 1.01947045e+00
3.67901139e-02 -1.21611178e+00 -2.53127038e-01 -1.13132453e+00
4.05530363e-01 8.29578221e-01 -5.28493583e-01 -4.37636435e-01
-3.57998043e-01 -4.47731465e-01 1.10270786e+00 7.58663058e-01
6.46395609e-02 -9.89254594e-01 6.67017996e-01 2.91137844e-01
-5.60351610e-01 -1.50701642e+00 -1.34171084e-01 -1.03169084e-02
-7.89525986e-01 -1.44630659e+00 2.62083020e-03 -8.67714942e-01
9.47835445e-01 1.14818931e-01 1.60415912e+00 1.24734640e-01
-1.54320717e-01 8.82123828e-01 -4.83554870e-01 -3.73548359e-01
-4.69677240e-01 8.46039414e-01 -5.55292927e-02 -5.88969707e-01
4.31953549e-01 -5.69373012e-01 -3.55177671e-02 -2.55449474e-01
-1.01592445e+00 3.38874519e-01 7.12255180e-01 5.32636821e-01
4.46667314e-01 2.67009854e-01 6.38065696e-01 -1.66810942e+00
7.21313477e-01 -7.85063505e-01 -2.57299215e-01 6.38772905e-01
-7.58661509e-01 6.20349109e-01 4.53687310e-01 2.34083571e-02
-1.28663623e+00 8.20922181e-02 -6.31787479e-02 -5.76125272e-02
2.21210659e-01 1.37083960e+00 -1.94269553e-01 -5.23931459e-02
6.94341540e-01 4.02902849e-02 -5.17679632e-01 -4.94800419e-01
9.97744679e-01 2.31344372e-01 4.64142978e-01 -7.99306452e-01
1.10894918e+00 3.59210402e-01 -1.26371935e-01 -6.64813578e-01
-1.38972998e+00 -6.30579472e-01 -6.67075217e-01 4.67995480e-02
3.27790499e-01 -1.25726998e+00 -4.95359212e-01 1.44663811e-01
-1.13154900e+00 -3.86033475e-01 -3.58794332e-01 3.57686400e-01
-1.27550676e-01 5.21126807e-01 -8.71156335e-01 -4.27366465e-01
-5.50698221e-01 -3.14414203e-01 8.19808006e-01 -1.03766702e-01
-2.87476301e-01 -1.43852711e+00 1.95419073e-01 5.08802116e-01
5.30466102e-02 -6.39528409e-02 1.25434983e+00 -8.64377677e-01
-8.77458692e-01 -2.18354374e-01 -3.14608246e-01 4.55335202e-03
9.15582851e-02 -1.90297231e-01 -7.49433577e-01 -3.49279344e-01
-1.07784021e+00 -6.64385796e-01 1.20806253e+00 -7.29098618e-02
8.84971559e-01 -6.56841874e-01 -8.33486319e-01 5.62609017e-01
1.41256905e+00 -4.52202082e-01 6.62282169e-01 2.05228254e-01
1.20687044e+00 3.37788314e-01 2.54869729e-01 2.07813859e-01
1.10306728e+00 1.53161585e-01 2.93200046e-01 -9.56976116e-02
-2.64422178e-01 -9.89900291e-01 3.01718682e-01 9.53087866e-01
2.04105079e-01 -4.20319021e-01 -1.16737378e+00 9.02437568e-01
-1.97696781e+00 -7.12972701e-01 -1.96045369e-01 1.72234142e+00
1.19375372e+00 1.91606894e-01 -3.27549309e-01 -3.04249287e-01
5.37641406e-01 1.25458032e-01 -4.74484116e-01 2.03069612e-01
-3.13683420e-01 4.88134585e-02 7.17567325e-01 9.69274879e-01
-1.03096616e+00 1.52679765e+00 6.70103025e+00 7.61415005e-01
-2.12962896e-01 -7.38935545e-02 -6.54061213e-02 6.00404330e-02
-7.70417333e-01 6.22874916e-01 -1.01843190e+00 -2.20112279e-01
7.26886392e-01 -6.05539262e-01 4.93662775e-01 7.77100265e-01
-2.66279787e-01 -1.28921360e-01 -1.21186197e+00 6.53385162e-01
3.10089607e-02 -1.52831185e+00 6.40150726e-01 -1.57747164e-01
7.71521747e-01 2.88834631e-01 -5.41392744e-01 8.41136873e-01
1.27411830e+00 -8.80427539e-01 1.30815223e-01 6.70502067e-01
8.83184791e-01 -5.80359697e-01 7.22700596e-01 3.03803891e-01
-1.45149529e+00 1.73746929e-01 -1.97605878e-01 2.09189281e-02
1.62511304e-01 7.21412957e-01 -1.44879472e+00 1.07367754e+00
2.37142146e-01 9.75319326e-01 -8.79214346e-01 7.46407330e-01
-8.24885190e-01 7.68105984e-01 -3.90248328e-01 4.04279530e-02
1.18010364e-01 1.62615195e-01 5.77612400e-01 1.37818229e+00
-2.92972535e-01 1.14905588e-01 3.26299846e-01 8.04470897e-01
-7.80793667e-01 2.19975300e-02 -9.07873154e-01 -5.22131264e-01
6.31964028e-01 1.32771373e+00 -3.26865196e-01 -5.94832182e-01
-5.35854936e-01 7.98020005e-01 9.99174356e-01 7.11600244e-01
-3.55663687e-01 -5.10935128e-01 4.31656778e-01 6.74163476e-02
-4.86224331e-02 -3.87890697e-01 6.93704858e-02 -1.49193585e+00
-2.05507572e-03 -4.05947119e-01 1.10068202e+00 -9.87954140e-01
-1.23424649e+00 1.20381275e-02 1.80314913e-01 -3.34061950e-01
-3.23653579e-01 -2.59795368e-01 -2.39785418e-01 6.15791559e-01
-1.63079333e+00 -1.56047404e+00 -2.36454457e-01 5.82838893e-01
1.20971404e-01 1.25124291e-01 8.58938873e-01 2.25621447e-01
-3.44729602e-01 4.80250984e-01 -1.41171589e-01 6.79667413e-01
6.62414074e-01 -1.50218332e+00 7.08514214e-01 8.81073594e-01
3.08471888e-01 8.74196589e-01 4.87210691e-01 -1.30409598e+00
-1.70908558e+00 -1.53847718e+00 1.21529627e+00 -7.94096947e-01
9.14081573e-01 -3.56628090e-01 -9.99730170e-01 1.53149343e+00
-9.99754667e-02 -1.35860801e-01 6.53008640e-01 9.08604205e-01
-6.85488760e-01 1.07381947e-01 -6.70284212e-01 6.32353783e-01
1.45315671e+00 -8.59278500e-01 -7.52417862e-01 3.91079158e-01
1.05019855e+00 -4.12844509e-01 -9.99160647e-01 3.06444019e-01
2.98032999e-01 4.02863435e-02 9.39764798e-01 -1.12670147e+00
2.23450512e-01 -3.52195501e-01 3.14672649e-01 -1.54682314e+00
-4.84727740e-01 -6.46709204e-01 -8.66040647e-01 1.12091982e+00
9.13970947e-01 -3.09484541e-01 1.18038297e+00 8.40065002e-01
-1.38353519e-02 -2.37822190e-01 -3.88392597e-01 -7.49222875e-01
-1.82019308e-01 -6.04014874e-01 3.76280665e-01 1.25378036e+00
7.71111071e-01 9.94746387e-01 -3.14151019e-01 2.79041320e-01
7.56335139e-01 -1.65933117e-01 7.41762161e-01 -1.36313534e+00
6.65059406e-03 1.30309537e-01 -2.43986443e-01 -8.86338413e-01
7.02047646e-01 -1.51564312e+00 -3.26916665e-01 -2.39045858e+00
6.76692724e-01 -3.63467604e-01 -9.73858386e-02 1.29152107e+00
-3.84493560e-01 -3.24220806e-01 -2.22514346e-01 8.53119697e-03
-1.22661042e+00 6.75059021e-01 9.99028862e-01 -5.63909709e-01
-1.53926492e-01 -4.06115323e-01 -1.04895246e+00 6.48763597e-01
7.78608441e-01 -5.39470673e-01 -7.62664795e-01 -7.66764462e-01
1.08519387e+00 -7.71999359e-02 2.21147969e-01 -5.47884643e-01
8.21909547e-01 5.31955017e-03 5.83056509e-02 -5.61387658e-01
1.46029398e-01 -6.32583618e-01 2.04856694e-01 2.41927356e-01
-4.77497309e-01 -2.17372298e-01 3.40313971e-01 9.64049160e-01
-3.04723382e-01 -1.90840721e-01 3.85770574e-02 -2.52553254e-01
-1.07502997e+00 5.64518631e-01 -2.18654603e-01 5.91858208e-01
5.68071544e-01 1.52814463e-01 -6.02254570e-01 -6.37640059e-01
-1.05802214e+00 8.52041781e-01 2.10839763e-01 4.48217392e-01
7.96772420e-01 -1.28862476e+00 -7.99606442e-01 -2.81828374e-01
6.43062532e-01 3.45574081e-01 1.13025799e-01 4.77323174e-01
-2.96914071e-01 5.76367021e-01 4.31306124e-01 7.75342062e-02
-1.15694904e+00 6.24329984e-01 1.79556325e-01 -7.53633738e-01
-6.45362794e-01 6.76253617e-01 4.79504652e-03 -7.37442732e-01
2.90513307e-01 -2.07886308e-01 -3.14398885e-01 -3.29358242e-02
3.52175683e-01 4.48193103e-01 6.12486303e-02 -1.64165907e-02
-3.44931841e-01 -1.44366935e-01 -6.29536450e-01 6.12685643e-02
1.21985853e+00 -2.33852938e-01 -3.98859501e-01 6.89502507e-02
9.02252555e-01 1.42610148e-01 -5.79495192e-01 -8.86583686e-01
6.75042212e-01 4.21111956e-02 -2.45249402e-02 -1.10991657e+00
-7.48868585e-01 2.35063851e-01 -5.60352504e-01 -6.02578819e-02
7.09937990e-01 4.23284233e-01 7.92250991e-01 1.13258541e+00
3.37020785e-01 -1.12594151e+00 -2.43841276e-01 9.77820337e-01
7.17231095e-01 -1.14094090e+00 2.31224805e-01 -9.50078666e-01
-5.02197266e-01 8.13479722e-01 7.56869912e-01 4.49230760e-01
6.40120745e-01 2.44797379e-01 -5.59663951e-01 -5.85600853e-01
-1.12050402e+00 -5.18149018e-01 2.59043932e-01 8.89938176e-01
1.56020969e-01 1.57371163e-01 1.00569189e-01 8.58554840e-01
-1.28991148e-02 7.59425759e-02 5.66605091e-01 9.60197866e-01
-3.90621275e-01 -1.23148239e+00 1.30636454e-01 7.17166483e-01
-3.13821286e-01 -6.84308887e-01 -8.98320079e-01 9.09182250e-01
-3.12781274e-01 9.12574172e-01 -4.40870255e-01 -4.33418691e-01
4.37608838e-01 3.41753572e-01 5.50644934e-01 -9.40179765e-01
-1.68488577e-01 -6.16187155e-01 9.78555441e-01 -4.13771033e-01
-8.36051181e-02 -1.65680945e-01 -1.59966922e+00 -4.03082669e-01
-3.50610077e-01 3.90686870e-01 2.39171684e-01 9.29009140e-01
4.94278550e-01 4.89085525e-01 5.11406176e-03 -2.83606686e-02
-1.63661614e-01 -8.75320315e-01 -6.49819314e-01 3.96734595e-01
1.46750184e-02 -3.94564778e-01 -1.19417056e-01 2.39510566e-01] | [9.168331146240234, 8.09712028503418] |
50401ecc-c431-4e64-b0fd-cb38df4577b0 | weakly-supervised-3d-human-pose-learning-via | 2003.07581 | null | https://arxiv.org/abs/2003.07581v1 | https://arxiv.org/pdf/2003.07581v1.pdf | Weakly-Supervised 3D Human Pose Learning via Multi-view Images in the Wild | One major challenge for monocular 3D human pose estimation in-the-wild is the acquisition of training data that contains unconstrained images annotated with accurate 3D poses. In this paper, we address this challenge by proposing a weakly-supervised approach that does not require 3D annotations and learns to estimate 3D poses from unlabeled multi-view data, which can be acquired easily in in-the-wild environments. We propose a novel end-to-end learning framework that enables weakly-supervised training using multi-view consistency. Since multi-view consistency is prone to degenerated solutions, we adopt a 2.5D pose representation and propose a novel objective function that can only be minimized when the predictions of the trained model are consistent and plausible across all camera views. We evaluate our proposed approach on two large scale datasets (Human3.6M and MPII-INF-3DHP) where it achieves state-of-the-art performance among semi-/weakly-supervised methods. | ['Pavlo Molchanov', 'Umar Iqbal', 'Jan Kautz'] | 2020-03-17 | weakly-supervised-3d-human-pose-learning-via-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Iqbal_Weakly-Supervised_3D_Human_Pose_Learning_via_Multi-View_Images_in_the_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Iqbal_Weakly-Supervised_3D_Human_Pose_Learning_via_Multi-View_Images_in_the_CVPR_2020_paper.pdf | cvpr-2020-6 | ['monocular-3d-human-pose-estimation', 'weakly-supervised-3d-human-pose-estimation'] | ['computer-vision', 'computer-vision'] | [-2.17227012e-01 2.19288692e-01 -1.11503318e-01 -5.39052427e-01
-9.37750041e-01 -6.14623964e-01 2.76798159e-01 -3.80602032e-01
-6.13595307e-01 6.35893047e-01 1.43301412e-01 3.23778689e-01
2.67502755e-01 -1.14732392e-01 -1.02593410e+00 -2.09935814e-01
2.65742630e-01 1.05472243e+00 2.92935580e-01 -9.98916328e-02
-2.82662809e-01 2.80037940e-01 -1.40640032e+00 1.21939965e-01
3.63935262e-01 7.62244165e-01 2.01619461e-01 7.30906308e-01
6.55452430e-01 4.97341245e-01 -1.29090577e-01 -4.53857809e-01
6.55494213e-01 -3.01695615e-01 -8.15677941e-01 6.04105413e-01
9.17296827e-01 -5.92293024e-01 -2.12745324e-01 8.02435577e-01
7.26353288e-01 -1.19991571e-01 4.63275671e-01 -1.32726896e+00
-2.99944804e-05 -3.44550669e-01 -6.08582854e-01 -2.18137756e-01
8.48190248e-01 1.04950428e-01 7.99473405e-01 -1.05170667e+00
9.70783234e-01 1.20139515e+00 8.14258456e-01 6.00566447e-01
-1.13549078e+00 -2.15664417e-01 3.08357060e-01 -5.01912683e-02
-1.48399198e+00 -2.76992887e-01 6.77518785e-01 -5.70905387e-01
9.00282979e-01 6.36601970e-02 8.37630391e-01 1.33740354e+00
1.51797518e-01 1.01989508e+00 1.16556382e+00 -4.19088036e-01
4.24848162e-02 3.48953195e-02 -3.52557540e-01 9.22458887e-01
2.84792185e-01 8.69384333e-02 -8.11236203e-01 -3.76667157e-02
8.78056228e-01 1.02933168e-01 -1.55105725e-01 -1.19200432e+00
-1.34557378e+00 4.18738216e-01 3.72642606e-01 -2.73695469e-01
-2.69087493e-01 7.32988119e-02 2.07595035e-01 7.81957153e-03
5.66576064e-01 7.57537261e-02 -8.20471108e-01 -7.35504776e-02
-7.93572545e-01 5.33962607e-01 4.81550783e-01 1.16735184e+00
5.16990066e-01 -4.49897081e-01 3.14525634e-01 5.84837437e-01
4.21277165e-01 6.65523529e-01 1.34863377e-01 -9.87235427e-01
7.56903350e-01 6.04135752e-01 4.95403171e-01 -6.43150210e-01
-6.32788241e-01 -3.70290220e-01 -4.51813579e-01 1.87253922e-01
5.30680358e-01 -5.50225712e-02 -8.32236290e-01 1.64727235e+00
8.20697129e-01 -1.18569918e-01 -5.41507602e-02 1.32190061e+00
6.85107172e-01 1.68308452e-01 -2.63090014e-01 4.54240479e-02
1.07949817e+00 -1.24021447e+00 -3.84516656e-01 -6.66400194e-01
5.53445816e-01 -6.56705916e-01 9.71500278e-01 3.62859666e-01
-1.04886353e+00 -7.35936880e-01 -9.30535138e-01 -1.87464356e-01
4.30426821e-02 3.84626120e-01 1.51142523e-01 3.23828548e-01
-7.18371034e-01 3.09737831e-01 -1.06819785e+00 -6.10568345e-01
-7.03950599e-02 2.34880745e-01 -1.08776248e+00 -2.56621808e-01
-6.90116942e-01 1.11298454e+00 4.28240418e-01 1.79092288e-01
-9.42459345e-01 -2.84790546e-01 -1.11685884e+00 -5.52445889e-01
8.39910865e-01 -9.86142755e-01 1.22101164e+00 -5.44162095e-01
-1.32521605e+00 1.37861562e+00 -1.12258777e-01 -1.89049825e-01
1.02876937e+00 -9.84092891e-01 1.91523045e-01 1.87385648e-01
2.64385402e-01 7.94514000e-01 6.19870186e-01 -1.63878775e+00
-3.68704259e-01 -9.41236675e-01 6.57698140e-02 6.71995699e-01
8.13572332e-02 -3.19356978e-01 -9.42741871e-01 -4.39331740e-01
5.31696916e-01 -1.47812331e+00 -2.94583321e-01 3.02506804e-01
-5.20148873e-01 1.97770953e-01 5.58778346e-01 -7.32434273e-01
4.79763657e-01 -1.63010967e+00 7.18745470e-01 -3.40095721e-02
-9.54673998e-03 -1.15593160e-02 1.66819751e-01 2.57226020e-01
2.07869470e-01 -5.60980678e-01 -1.51129633e-01 -9.55710769e-01
2.68498156e-03 3.68948102e-01 2.27830932e-01 8.61166894e-01
-2.01696027e-02 9.26791489e-01 -8.22783589e-01 -6.13004267e-01
4.40055519e-01 4.17594910e-01 -6.99776173e-01 7.18477011e-01
-1.74326941e-01 9.55695748e-01 -1.99875757e-01 7.16693938e-01
5.47276139e-01 -4.36794221e-01 3.52705985e-01 -2.99056619e-01
1.12295501e-01 -1.86956942e-01 -1.43937230e+00 2.34615612e+00
-3.88930708e-01 4.29788046e-02 -6.31839633e-02 -7.30318487e-01
5.15899658e-01 3.15422535e-01 5.09427547e-01 -2.01039940e-01
1.21622093e-01 2.48345852e-01 -5.13483047e-01 -4.10738111e-01
2.93227762e-01 -4.53869179e-02 -2.89163589e-01 1.66957662e-01
3.98378432e-01 -3.06988776e-01 -2.39358723e-01 -3.41128670e-02
7.04318702e-01 9.02025223e-01 5.81898391e-01 -2.43540425e-02
4.56362218e-01 -1.05783314e-01 6.26555383e-01 3.61146003e-01
-2.08808109e-01 1.08387601e+00 1.68574184e-01 -6.04022324e-01
-1.15388060e+00 -1.30424511e+00 2.06539139e-01 7.98453033e-01
2.83410251e-01 -4.99820828e-01 -6.59878731e-01 -1.10402703e+00
3.61178070e-02 7.84886703e-02 -5.62154889e-01 2.15134934e-01
-4.38815862e-01 -2.90258199e-01 2.35701874e-01 7.48388290e-01
4.42212701e-01 -4.62404668e-01 -9.79897261e-01 -1.18113734e-01
-4.89572257e-01 -1.74480939e+00 -6.03329957e-01 1.34708077e-01
-8.47294271e-01 -1.24576688e+00 -1.00049925e+00 -7.31803298e-01
9.86040473e-01 2.04894647e-01 1.15727806e+00 -2.99391329e-01
-2.37008482e-01 6.77265286e-01 -3.21117789e-01 -7.72938058e-02
4.47748862e-02 -1.03249006e-01 5.41249096e-01 -5.02527803e-02
1.14629023e-01 -3.49520624e-01 -5.14651954e-01 7.24433422e-01
-2.76611358e-01 1.45263419e-01 5.37670135e-01 9.25558448e-01
1.14075661e+00 -5.05305171e-01 5.06030135e-02 -8.32038760e-01
-3.01303059e-01 4.25780937e-02 -5.83114147e-01 3.15850973e-01
-2.59827882e-01 3.06150373e-02 3.41771215e-01 -2.01792449e-01
-1.10381544e+00 9.21315908e-01 -9.78830978e-02 -6.90382183e-01
-3.33885491e-01 2.21363440e-01 -3.83755088e-01 -1.61586612e-01
6.14546299e-01 8.79752729e-03 -1.00217201e-01 -6.64850473e-01
3.77833664e-01 4.27502036e-01 8.85232329e-01 -5.85576832e-01
9.45562363e-01 7.15945423e-01 1.48628175e-01 -6.84016883e-01
-1.27567458e+00 -7.87205875e-01 -1.36913514e+00 -4.24861550e-01
1.02535272e+00 -1.60041296e+00 -4.45937812e-01 5.60708404e-01
-1.12622583e+00 -2.51292139e-01 7.30239321e-03 7.36040413e-01
-9.69454467e-01 5.91780603e-01 -3.90966088e-01 -7.93476820e-01
-1.15672894e-01 -1.08634055e+00 1.68827236e+00 -1.88566670e-01
-4.07156795e-01 -6.85555100e-01 3.02515298e-01 7.58889854e-01
-3.58157247e-01 6.14725947e-01 8.61181170e-02 -3.40017647e-01
-3.62797529e-01 -3.61727029e-01 1.40359014e-01 3.59759510e-01
-1.33733228e-01 -6.50363743e-01 -9.40544307e-01 -5.49146533e-01
-9.95235667e-02 -9.88202155e-01 4.77359384e-01 3.31026673e-01
7.84595549e-01 -1.34052653e-02 -1.84888363e-01 6.61699653e-01
1.16346371e+00 -6.07446194e-01 2.15552568e-01 2.53290355e-01
1.04508495e+00 7.30794549e-01 1.04340053e+00 6.07953787e-01
7.35490799e-01 1.16934478e+00 4.32346165e-01 -5.75821474e-02
1.31796271e-01 -6.80236816e-01 2.94606894e-01 5.86745560e-01
-2.29439601e-01 -1.86330527e-02 -8.59803557e-01 3.81533504e-01
-2.08613753e+00 -6.82855129e-01 6.27751350e-02 2.29265380e+00
6.42183959e-01 3.48342061e-01 3.91139776e-01 -2.56922618e-02
4.22538489e-01 -1.38820792e-02 -5.75496972e-01 3.65226775e-01
1.40779957e-01 -1.59951553e-01 4.68338042e-01 6.11361384e-01
-1.36262774e+00 8.87456775e-01 5.76055050e+00 8.39288160e-02
-7.88727641e-01 2.12008759e-01 4.13141280e-01 -3.81596804e-01
1.74660102e-01 -1.38097882e-01 -8.97659242e-01 1.04024149e-01
3.92516106e-01 5.64018667e-01 -5.57573140e-02 1.10800695e+00
1.16297975e-01 -2.99717307e-01 -1.54828703e+00 1.36094368e+00
4.30323511e-01 -7.83449590e-01 -3.21409702e-01 9.10091102e-02
9.56010222e-01 5.92812989e-03 -1.70192704e-01 -2.71600951e-02
-9.92473215e-02 -7.83216476e-01 1.06399238e+00 3.48955423e-01
8.14965367e-01 -6.86228991e-01 8.08378458e-01 9.15892959e-01
-1.23606360e+00 1.89614743e-01 -1.48348466e-01 -1.57446593e-01
4.18734163e-01 4.06284958e-01 -6.23172581e-01 7.00171471e-01
9.65533197e-01 8.79102945e-01 -7.20101178e-01 7.53028989e-01
-3.39838356e-01 -1.11266807e-01 -4.16104734e-01 3.69893909e-01
-1.09156065e-01 7.29146004e-02 3.37292254e-01 8.83809268e-01
2.10074425e-01 4.40486334e-02 7.00732708e-01 4.32216853e-01
8.73561874e-02 -2.13296324e-01 -6.44132197e-01 5.30810773e-01
2.10919619e-01 1.01663804e+00 -6.44800603e-01 -1.13634951e-01
-3.97415549e-01 1.48570466e+00 5.58757126e-01 -3.37312669e-02
-7.85528362e-01 3.76665413e-01 3.11278164e-01 2.54440457e-01
3.52773726e-01 -4.90134120e-01 4.14561145e-02 -1.48313153e+00
5.92582941e-01 -8.12279224e-01 4.58579630e-01 -9.23314214e-01
-1.27682114e+00 6.17025137e-01 4.46143776e-01 -1.49654496e+00
-6.29892111e-01 -7.94211745e-01 7.72251561e-02 5.06750762e-01
-1.21934056e+00 -1.56635129e+00 -5.33890188e-01 6.30302608e-01
5.43169856e-01 -8.01636837e-03 6.89417303e-01 1.43895835e-01
-1.23031721e-01 6.09553993e-01 -2.55190045e-01 -1.20319277e-02
1.10613906e+00 -1.33131456e+00 5.28964341e-01 7.38920927e-01
3.12177062e-01 3.29519004e-01 6.99601352e-01 -6.52771652e-01
-1.48873353e+00 -1.00136626e+00 9.29373801e-01 -1.07131135e+00
-1.70377176e-02 -7.12446392e-01 -3.11371297e-01 9.63568389e-01
-2.69437462e-01 5.44481397e-01 4.90224898e-01 9.20742825e-02
-3.44950885e-01 7.23204166e-02 -1.18780732e+00 2.89138466e-01
1.56779802e+00 -5.65885603e-01 -6.63668633e-01 4.39606011e-01
4.43376720e-01 -1.03778923e+00 -7.46118784e-01 4.54979539e-01
6.88828647e-01 -1.04401088e+00 1.17473853e+00 -5.51403284e-01
1.71439901e-01 -4.27282661e-01 -4.67317343e-01 -1.15791392e+00
1.15899459e-01 -4.90910530e-01 -2.38196567e-01 6.02746129e-01
1.22019023e-01 -1.55207559e-01 1.19563448e+00 4.36585546e-01
8.59493986e-02 -6.58171535e-01 -1.04915333e+00 -1.00204027e+00
-2.96856582e-01 -2.98268288e-01 -1.02703094e-01 5.30778825e-01
-1.41402304e-01 3.11890900e-01 -8.36936235e-01 3.46478730e-01
1.10158408e+00 5.24233617e-02 1.36928821e+00 -1.21486270e+00
-6.38092399e-01 5.65953434e-01 -6.35047078e-01 -1.40893161e+00
2.50037730e-01 -4.72955912e-01 3.34556580e-01 -1.22168446e+00
3.87173891e-01 8.90409574e-02 2.81880170e-01 3.50432068e-01
-1.78872064e-01 4.93062794e-01 2.24123031e-01 2.58124232e-01
-1.02690661e+00 5.91519713e-01 1.08986497e+00 3.48353803e-01
8.98428187e-02 2.88081374e-02 -1.03536502e-01 1.04946280e+00
2.84775227e-01 -4.01045918e-01 -5.20596683e-01 -4.77393836e-01
6.83885291e-02 1.36067554e-01 8.01540256e-01 -1.17563164e+00
7.50218108e-02 -3.46492864e-02 7.57977605e-01 -1.01764119e+00
7.50023246e-01 -9.87571776e-01 3.10984015e-01 4.64506239e-01
-1.94376290e-01 3.39520454e-01 -1.30025715e-01 7.68761933e-01
4.42081783e-03 1.74547762e-01 7.97052622e-01 -5.92040122e-01
-7.15752304e-01 3.77655208e-01 2.99020588e-01 4.48229522e-01
1.09178483e+00 -1.09280087e-01 2.70907253e-01 -4.66338038e-01
-7.98388422e-01 3.05265486e-01 8.74225855e-01 5.58813334e-01
7.35160112e-01 -1.41219592e+00 -6.80421829e-01 1.41041994e-01
5.21044433e-01 4.45195049e-01 2.08789036e-01 5.14536440e-01
-6.93614900e-01 5.01446307e-01 -2.12188497e-01 -1.12136292e+00
-1.41767597e+00 4.02801126e-01 3.96383852e-01 -3.59457821e-01
-5.67029297e-01 7.56179631e-01 1.18125319e-01 -1.01765907e+00
3.43277007e-01 -1.45036578e-01 3.82412285e-01 -3.64102036e-01
2.04294100e-01 2.35890701e-01 1.58471823e-01 -1.19714141e+00
-6.60200596e-01 9.20339704e-01 1.35112613e-01 -3.29245031e-01
1.37418413e+00 -2.85396695e-01 2.74703920e-01 4.72513378e-01
1.39130986e+00 -2.21802533e-01 -1.73814201e+00 -3.71389449e-01
-2.54706115e-01 -7.38910735e-01 -2.49193728e-01 -7.55275190e-01
-7.53064692e-01 6.75047219e-01 7.33295500e-01 -7.76927948e-01
6.18872881e-01 3.29931140e-01 7.44789362e-01 6.30804598e-01
8.49591911e-01 -1.34719074e+00 5.72506249e-01 4.85304683e-01
9.78522480e-01 -1.69870067e+00 3.30781996e-01 -5.49642503e-01
-8.67889345e-01 8.27245474e-01 1.07845676e+00 -1.26561195e-01
5.30375898e-01 6.60778210e-02 1.23708755e-01 -1.59560442e-01
-4.49436307e-01 -2.35639468e-01 5.91340899e-01 5.95209956e-01
3.64298433e-01 -9.61518660e-03 -1.77283883e-02 4.13364530e-01
-1.03138596e-01 -7.29648620e-02 1.41318873e-01 1.25617075e+00
-1.77294225e-01 -1.06360793e+00 -5.65171301e-01 -1.03343524e-01
-1.73567817e-01 4.46681589e-01 -5.43693900e-01 8.13555956e-01
1.96652010e-01 7.34507024e-01 -2.73582369e-01 -5.04970968e-01
7.00846970e-01 1.32926255e-01 9.97257829e-01 -7.76181638e-01
-1.31495446e-01 3.19725692e-01 1.31808817e-01 -8.55753303e-01
-7.18292236e-01 -8.76900733e-01 -1.10033643e+00 7.61662126e-02
-2.66307741e-01 -1.93140611e-01 3.49407166e-01 1.20084000e+00
2.61438787e-01 1.69740487e-02 4.02908981e-01 -1.31177533e+00
-7.21026242e-01 -6.92583382e-01 -4.27723587e-01 6.94730520e-01
3.38101774e-01 -9.94936109e-01 -6.75085187e-02 1.68705657e-01] | [6.984831809997559, -0.9112699031829834] |
feeec8ce-7e41-46ae-9255-3c0072130bea | model-of-the-weak-reset-process-in-hfox | 2107.06064 | null | https://arxiv.org/abs/2107.06064v2 | https://arxiv.org/pdf/2107.06064v2.pdf | Model of the Weak Reset Process in HfOx Resistive Memory for Deep Learning Frameworks | The implementation of current deep learning training algorithms is power-hungry, owing to data transfer between memory and logic units. Oxide-based RRAMs are outstanding candidates to implement in-memory computing, which is less power-intensive. Their weak RESET regime, is particularly attractive for learning, as it allows tuning the resistance of the devices with remarkable endurance. However, the resistive change behavior in this regime suffers many fluctuations and is particularly challenging to model, especially in a way compatible with tools used for simulating deep learning. In this work, we present a model of the weak RESET process in hafnium oxide RRAM and integrate this model within the PyTorch deep learning framework. Validated on experiments on a hybrid CMOS/RRAM technology, our model reproduces both the noisy progressive behavior and the device-to-device (D2D) variability. We use this tool to train Binarized Neural Networks for the MNIST handwritten digit recognition task and the CIFAR-10 object classification task. We simulate our model with and without various aspects of device imperfections to understand their impact on the training process and identify that the D2D variability is the most detrimental aspect. The framework can be used in the same manner for other types of memories to identify the device imperfections that cause the most degradation, which can, in turn, be used to optimize the devices to reduce the impact of these imperfections. | ['Damien Querlioz', 'Jean-Michel Portal', 'Elisa Vianello', 'Etienne Nowak', 'Jacques-Olivier Klein', 'Axel Laborieux', 'Tifenn Hirtzlin', 'Marc Bocquet', 'Atreya Majumdar'] | 2021-07-02 | null | null | null | null | ['handwritten-digit-recognition'] | ['computer-vision'] | [ 1.62136167e-01 -3.79411936e-01 -1.59302175e-01 -3.63763385e-02
-1.73068643e-02 -1.98587313e-01 3.14215541e-01 3.49957049e-01
-6.61683023e-01 8.46285403e-01 -4.85413074e-01 -3.67280662e-01
-1.06404752e-01 -1.02950966e+00 -8.10154855e-01 -1.15557337e+00
2.10948333e-01 4.10577625e-01 6.26055539e-01 -4.54165608e-01
3.92056197e-01 7.94996321e-01 -1.64810801e+00 3.23939532e-01
5.66756785e-01 1.03510141e+00 8.85884687e-02 3.13824207e-01
5.89359850e-02 7.67721474e-01 -8.67393374e-01 2.29831524e-02
-1.61728188e-01 -2.92698979e-01 -2.84995645e-01 -7.59538114e-01
1.88390225e-01 -9.92934257e-02 -5.47637701e-01 6.92351401e-01
7.99090981e-01 -1.68787893e-02 8.43331218e-01 -4.84305918e-01
-1.29031658e-01 1.00466955e+00 -1.52593866e-01 7.39106596e-01
-2.71361113e-01 2.90156573e-01 2.59375572e-01 -5.24720132e-01
4.82917577e-01 8.66088808e-01 5.78795314e-01 8.50196183e-01
-1.28733373e+00 -8.19648027e-01 -3.79522860e-01 6.06681742e-02
-1.45604396e+00 -4.35211182e-01 6.10062957e-01 -1.37257874e-01
1.65886962e+00 -1.25340417e-01 7.26031423e-01 1.28308344e+00
1.01394844e+00 2.09040210e-01 1.32698178e+00 -4.60892946e-01
4.58659112e-01 3.01976323e-01 5.00847042e-01 5.39260864e-01
5.78423500e-01 1.31714776e-01 -7.81045854e-01 3.74431275e-02
6.49895191e-01 -5.61718084e-02 -1.31013393e-01 2.41503134e-01
-3.06095481e-01 3.41592371e-01 8.12506497e-01 8.05945396e-01
-9.02045965e-02 7.48218834e-01 3.77800405e-01 3.50271165e-01
-7.38326088e-02 6.05823338e-01 -1.74527407e-01 -6.45960942e-02
-9.84488249e-01 2.88437661e-02 5.76993585e-01 2.27382928e-01
4.43235308e-01 5.78613162e-01 -1.04773492e-01 7.50930011e-01
2.98109055e-01 7.02722251e-01 7.26430357e-01 -3.01087976e-01
1.27328813e-01 5.52734792e-01 -2.64285713e-01 -3.39899749e-01
-6.90118790e-01 -5.12050629e-01 -1.03843749e+00 4.07956183e-01
2.53065944e-01 -1.27051584e-02 -1.26857793e+00 1.34411788e+00
-3.64748418e-01 -1.08038500e-01 9.60627869e-02 6.66164696e-01
8.04822266e-01 8.37788939e-01 7.93214515e-02 -1.39096946e-01
1.09765565e+00 -7.31176734e-02 -6.97604597e-01 -1.73736185e-01
5.43346584e-01 -9.95213985e-02 9.01296496e-01 3.77163231e-01
-1.09156799e+00 -7.25886583e-01 -1.76102531e+00 7.25141987e-02
-6.49958789e-01 -1.09332018e-01 1.74846858e-01 1.29407835e+00
-1.00030923e+00 1.22404885e+00 -1.26388216e+00 -6.29443154e-02
4.90806848e-01 1.02855504e+00 4.00969386e-01 1.77153587e-01
-1.36463904e+00 1.02384460e+00 3.87681901e-01 1.63123876e-01
-9.74265933e-01 -6.50093138e-01 -3.05325776e-01 -5.80527401e-03
-1.14449382e-01 -4.67545837e-01 8.21606159e-01 -7.52972364e-01
-1.72872841e+00 7.78903663e-01 -5.73936589e-02 -8.59475017e-01
1.71590269e-01 -7.97797553e-03 -4.35648054e-01 -3.44965488e-01
-7.56470263e-01 4.63138431e-01 8.15777123e-01 -1.04360914e+00
2.92747110e-01 -5.01014113e-01 -3.65202188e-01 -4.31726336e-01
-7.64355600e-01 -3.27828318e-01 2.16488726e-03 -3.37130755e-01
3.32682393e-02 -9.23950851e-01 -1.99430287e-01 -4.78208989e-01
-1.98850289e-01 1.24061517e-01 8.72997344e-01 2.13265922e-02
1.18099356e+00 -1.99681735e+00 -1.03902303e-01 5.85744381e-01
-3.67333479e-02 5.13913155e-01 2.65581787e-01 3.08191001e-01
2.49938160e-01 1.48283079e-01 -2.63895452e-01 -2.34565437e-02
-5.63928068e-01 1.90764174e-01 -3.98238838e-01 4.00629103e-01
3.57493550e-01 8.26246798e-01 -2.18334675e-01 1.73275799e-01
7.86253065e-02 7.58314073e-01 -2.08137169e-01 -8.02712813e-02
-1.38313845e-01 3.79342139e-01 -3.20022166e-01 4.89489704e-01
5.69394410e-01 6.93053603e-02 4.81575072e-01 -2.88098425e-01
-2.80579031e-01 4.11585689e-01 -7.23989248e-01 1.18496323e+00
-5.18969119e-01 7.82140017e-01 -2.16144711e-01 -8.91560853e-01
1.56556118e+00 -3.76192555e-02 1.63347825e-01 -1.48025239e+00
4.84631926e-01 6.00841463e-01 3.95125926e-01 -2.40294069e-01
5.54225624e-01 -2.87525028e-01 -5.14278281e-03 2.16762528e-01
3.34636085e-02 -2.37849012e-01 -1.76468596e-01 -2.31072247e-01
1.13550293e+00 -2.44327754e-01 -3.14529240e-01 -8.23639631e-01
3.22628349e-01 -3.18061680e-01 2.02056274e-01 8.24327826e-01
1.58610672e-01 2.78107911e-01 5.60715795e-01 -4.45324481e-01
-1.13336635e+00 -1.09673846e+00 -5.84460735e-01 5.21838307e-01
2.30380043e-01 -6.05062954e-02 -7.48564243e-01 1.41566500e-01
-2.72864252e-01 6.52996540e-01 -4.31029409e-01 -8.15165102e-01
-8.04128826e-01 -1.34161639e+00 1.10811281e+00 6.40861154e-01
4.44004923e-01 -1.02847481e+00 -1.13226426e+00 4.70176816e-01
7.37880528e-01 -9.79824901e-01 6.27672493e-01 1.10123932e+00
-1.45289552e+00 -4.54418987e-01 -2.73161113e-01 -5.62671244e-01
4.59780306e-01 -6.78633034e-01 1.15660441e+00 3.88041109e-01
-5.69928944e-01 -5.94478920e-02 8.74526575e-02 -5.63412011e-01
-6.56840384e-01 3.74925286e-01 3.44881505e-01 -4.20297205e-01
2.65014499e-01 -6.55568540e-01 -6.89312696e-01 1.06782556e-01
-9.93641615e-01 -4.34779584e-01 5.23847759e-01 6.57155454e-01
6.57308459e-01 4.63692725e-01 6.95425451e-01 -1.07368958e+00
3.03177625e-01 -2.00047478e-01 -6.88075066e-01 -1.68499500e-01
-7.27772176e-01 4.09951478e-01 8.52079690e-01 -5.30826271e-01
-8.83368969e-01 4.09591980e-02 -6.43884480e-01 -1.87610283e-01
1.11103395e-03 2.77149051e-01 -3.01779360e-01 -1.70207873e-01
9.87768233e-01 -7.08289561e-04 -3.88150126e-01 -2.42085949e-01
-4.44651872e-01 5.66261590e-01 1.35659099e-01 -3.52179259e-01
4.33752835e-01 4.61541414e-01 3.82960677e-01 -1.05934143e+00
-4.81137037e-02 3.07273805e-01 -3.92727107e-01 -2.10388467e-01
5.95262468e-01 -8.17247033e-01 -7.94134974e-01 8.06382418e-01
-8.90209496e-01 -9.06801701e-01 -1.65994987e-01 7.84748942e-02
-7.63669461e-02 -6.42461538e-01 -1.00751245e+00 -9.84431505e-01
-6.85959697e-01 -1.31248081e+00 5.84885836e-01 7.49411404e-01
1.69270057e-02 -1.07083285e+00 -1.19029015e-01 7.93711469e-02
8.72120380e-01 9.49635878e-02 1.52257073e+00 -2.79996693e-01
-5.40322244e-01 7.19682947e-02 6.55637160e-02 3.35304052e-01
-2.83930540e-01 -3.12694395e-03 -1.48569870e+00 -4.45968121e-01
1.73040316e-01 -2.56973952e-01 1.48414779e+00 5.15469849e-01
1.28938651e+00 5.07218838e-01 -5.12883663e-01 2.91448891e-01
1.86811233e+00 2.38778129e-01 1.11602247e+00 -1.06475074e-02
7.11905360e-01 -3.76015007e-02 4.32453342e-02 2.49074161e-01
-1.74066082e-01 8.54835272e-01 4.27527964e-01 -1.52301043e-01
-2.27113031e-02 -7.22000655e-03 6.85851276e-01 7.22847700e-01
-1.49319991e-01 -3.36349070e-01 -1.19603288e+00 1.20340861e-01
-1.24641979e+00 -5.18280983e-01 -4.54996347e-01 2.38708234e+00
9.17652488e-01 8.61543179e-01 -3.05986375e-01 4.75135684e-01
2.88271934e-01 -4.76023182e-02 -7.32621372e-01 -9.35176373e-01
-3.14538538e-01 1.12690842e+00 8.68900836e-01 2.36412957e-01
-5.50403535e-01 7.48141408e-01 5.89694881e+00 6.55844986e-01
-1.90294719e+00 2.18047857e-01 7.78523207e-01 -3.26987028e-01
-2.18639523e-01 -4.47738707e-01 -1.36582017e+00 4.52654600e-01
1.77277088e+00 5.85700154e-01 1.83862075e-01 3.20820302e-01
-4.14887816e-02 -4.18376118e-01 -1.23898041e+00 8.57139051e-01
-1.61955521e-01 -1.53467286e+00 -1.79863691e-01 -6.38442412e-02
7.63709068e-01 1.88213050e-01 4.00296718e-01 2.32789591e-01
-4.50591654e-01 -1.44040513e+00 5.98735869e-01 6.95034862e-01
9.04977262e-01 -9.03425395e-01 9.07555580e-01 7.18329251e-02
-7.46087611e-01 -3.77461553e-01 -6.27101660e-01 -7.41750896e-02
-2.09215239e-01 1.11195457e+00 -4.38746452e-01 7.60640055e-02
9.21133578e-01 3.81338060e-01 -6.98218524e-01 7.70902932e-01
2.59580642e-01 7.98003435e-01 -4.38698649e-01 -4.42539006e-01
-1.22176960e-01 1.05913281e-01 1.77858561e-01 1.23382020e+00
3.52759302e-01 -2.76322901e-01 -4.82168168e-01 1.22050953e+00
-5.14720082e-01 -3.81079555e-01 -5.37117779e-01 -5.50808907e-02
6.07243061e-01 9.62232769e-01 -1.20669079e+00 -2.59333886e-02
3.53688776e-01 7.86994636e-01 7.47009590e-02 1.66026041e-01
-6.59976661e-01 -4.17643219e-01 3.71451914e-01 6.94057763e-01
3.25657308e-01 -4.61143821e-01 -7.42609322e-01 -4.86102462e-01
-2.01934669e-02 -4.86360699e-01 -2.01710388e-01 -2.36769393e-01
-6.57174885e-01 4.95190352e-01 -3.13416123e-01 -5.31773984e-01
1.07660010e-01 -1.05788457e+00 -8.68734121e-01 8.56017232e-01
-1.55025756e+00 -5.19235969e-01 -2.83531159e-01 2.32119605e-01
-9.40561667e-02 2.92762779e-02 8.57995391e-01 5.13805270e-01
-7.93336987e-01 8.25374365e-01 3.16934675e-01 -1.53148085e-01
3.93592447e-01 -8.59332800e-01 4.55649912e-01 4.22422796e-01
-1.22610494e-01 5.92052937e-01 7.48380661e-01 -4.90002066e-01
-1.76140487e+00 -1.09535766e+00 2.46238232e-01 -3.77588570e-01
3.10206354e-01 -8.57171297e-01 -1.38639915e+00 -4.89653721e-02
-2.89341118e-02 -5.83935864e-02 3.80712003e-01 -2.26335898e-01
-1.73511967e-01 -5.99421620e-01 -1.13650024e+00 3.88719976e-01
7.55262792e-01 -5.38167477e-01 -1.06741332e-01 6.12713993e-02
3.39140624e-01 -4.05735165e-01 -1.05370498e+00 6.23056114e-01
6.73238575e-01 -1.14616048e+00 6.63189828e-01 4.59652357e-02
2.59007066e-01 -7.23310113e-02 -2.06316262e-02 -1.08670795e+00
2.02142835e-01 -1.54753774e-01 -1.63872898e-01 1.30011237e+00
7.36840725e-01 -6.90052748e-01 8.76630843e-01 2.04360753e-01
3.70618738e-02 -9.58155036e-01 -1.12431848e+00 -7.57606626e-01
4.72844630e-01 -1.44079253e-01 2.30861336e-01 1.68920442e-01
-5.23939133e-01 3.32100451e-01 -3.06651294e-02 -1.38296094e-02
1.49735242e-01 -4.71311539e-01 1.60828754e-02 -1.25594246e+00
-3.12204003e-01 -4.27311540e-01 -4.20378596e-01 -5.43335497e-01
1.99472502e-01 -4.73912835e-01 2.23268807e-01 -9.08968210e-01
-2.68648386e-01 -1.06047130e+00 -5.98569930e-01 8.39210525e-02
1.55723512e-01 4.92745519e-01 -6.73673302e-02 1.39474392e-01
-1.73765346e-01 2.91150421e-01 5.15974998e-01 -1.47745073e-01
-5.78362286e-01 1.03747556e-02 -1.99517477e-02 3.32466900e-01
8.18874180e-01 -7.30898440e-01 -1.99343547e-01 -3.77464443e-01
4.97943878e-01 -2.04142690e-01 2.01842099e-01 -1.81571472e+00
1.94069251e-01 4.53189999e-01 8.57721686e-01 -9.30244923e-02
5.87895453e-01 -7.91579366e-01 2.36344948e-01 9.47302580e-01
-2.41585582e-01 -9.64169353e-02 7.79442787e-01 2.56868124e-01
7.17777684e-02 -6.33224726e-01 1.18472660e+00 7.91025385e-02
-3.58692378e-01 -6.40135109e-02 -7.80939996e-01 -1.54181033e-01
5.22119164e-01 -1.91945910e-01 -4.44728225e-01 1.95293516e-01
-3.11415046e-01 -3.91089529e-01 4.90316570e-01 1.70505226e-01
6.79859400e-01 -8.86517704e-01 -2.43369788e-01 5.03731310e-01
-2.89827645e-01 -2.25662768e-01 4.38079506e-01 7.06936061e-01
-5.19156873e-01 2.65628040e-01 -5.72362125e-01 -7.39126980e-01
-9.74147201e-01 2.18779176e-01 8.82143021e-01 -4.61036801e-01
-4.03430402e-01 7.34497190e-01 -5.28765023e-01 9.60008278e-02
1.02052301e-01 -4.72638994e-01 -2.86900491e-01 3.17577161e-02
5.05626976e-01 1.99153364e-01 8.00548196e-01 -9.49146077e-02
-4.63108808e-01 7.50353932e-01 -9.16761309e-02 1.12332627e-01
1.25799501e+00 2.68656820e-01 -2.14873374e-01 1.02058911e+00
9.34238255e-01 -3.48795891e-01 -9.23723876e-01 1.01733813e-02
2.14690119e-01 4.11769181e-01 4.50781018e-01 -6.23260200e-01
-1.20058239e+00 1.15289378e+00 1.44599569e+00 1.48456767e-02
1.21804690e+00 -3.31875443e-01 7.32366681e-01 5.26501417e-01
3.10707510e-01 -1.35803735e+00 1.95635617e-01 6.16754532e-01
3.04581642e-01 -6.39796257e-01 2.53569707e-02 5.47113875e-03
-1.37968630e-01 1.34118676e+00 7.69558072e-01 -3.06877792e-01
8.60811591e-01 1.05000174e+00 6.82452843e-02 -1.61574200e-01
-8.08817565e-01 1.74273774e-01 -5.87766357e-02 5.25700629e-01
6.44793034e-01 -2.41741654e-03 -6.52186945e-02 4.19530153e-01
9.82707143e-02 1.78575426e-01 6.57158494e-01 1.03762066e+00
-5.15392184e-01 -1.14711118e+00 -2.47416317e-01 7.37353504e-01
-5.14718592e-01 -9.11857411e-02 -1.40069559e-01 3.35345685e-01
2.38888726e-01 5.74157357e-01 4.32898223e-01 -4.75948274e-01
4.49865222e-01 2.66987532e-01 8.07574153e-01 -3.78742814e-01
-1.09317231e+00 -4.21865702e-01 -2.02096701e-01 -1.28391907e-01
-8.98304954e-02 -1.34072110e-01 -1.82211864e+00 -3.48691314e-01
-4.13610011e-01 -2.03972772e-01 1.11198854e+00 9.34616208e-01
9.02988911e-02 1.08583021e+00 2.93916941e-01 -9.62477863e-01
-3.06597173e-01 -7.15775907e-01 -6.65168643e-01 -1.19381368e-01
2.08492652e-01 -7.43334889e-01 -8.43942091e-02 -5.79282105e-01] | [8.232172966003418, 2.5553107261657715] |
116a6529-ae5d-48cd-95a9-50964cae7fd2 | fairvis-visual-analytics-for-discovering | 1904.05419 | null | https://arxiv.org/abs/1904.05419v4 | https://arxiv.org/pdf/1904.05419v4.pdf | FairVis: Visual Analytics for Discovering Intersectional Bias in Machine Learning | The growing capability and accessibility of machine learning has led to its application to many real-world domains and data about people. Despite the benefits algorithmic systems may bring, models can reflect, inject, or exacerbate implicit and explicit societal biases into their outputs, disadvantaging certain demographic subgroups. Discovering which biases a machine learning model has introduced is a great challenge, due to the numerous definitions of fairness and the large number of potentially impacted subgroups. We present FairVis, a mixed-initiative visual analytics system that integrates a novel subgroup discovery technique for users to audit the fairness of machine learning models. Through FairVis, users can apply domain knowledge to generate and investigate known subgroups, and explore suggested and similar subgroups. FairVis' coordinated views enable users to explore a high-level overview of subgroup performance and subsequently drill down into detailed investigation of specific subgroups. We show how FairVis helps to discover biases in two real datasets used in predicting income and recidivism. As a visual analytics system devoted to discovering bias in machine learning, FairVis demonstrates how interactive visualization may help data scientists and the general public understand and create more equitable algorithmic systems. | ['Jamie Morgenstern', 'Minsuk Kahng', 'Ángel Alexander Cabrera', 'Fred Hohman', 'Will Epperson', 'Duen Horng Chau'] | 2019-04-10 | null | null | null | null | ['subgroup-discovery'] | ['methodology'] | [-9.78620425e-02 4.16165411e-01 -6.12265766e-01 -7.14017749e-01
-4.77371961e-02 -4.26041454e-01 7.85238683e-01 8.03468287e-01
-3.16990823e-01 6.32713079e-01 6.22178614e-01 -6.96173668e-01
-4.07036999e-03 -8.71615112e-01 -1.92747518e-01 -2.51394004e-01
-2.79812783e-01 4.38211799e-01 -3.80287647e-01 -1.82414845e-01
5.58306038e-01 4.73170966e-01 -1.73722291e+00 5.16989708e-01
1.26431584e+00 3.51335078e-01 -5.35498977e-01 2.53788918e-01
-3.28756094e-01 7.83581316e-01 -5.91275156e-01 -6.54258132e-01
1.28104776e-01 -1.88508496e-01 -5.92795908e-01 -5.72369099e-01
7.73873508e-01 -4.67376225e-02 4.25728410e-01 9.84586477e-01
5.33435881e-01 -1.55039027e-01 7.48536110e-01 -1.79825044e+00
-7.30967462e-01 8.01216006e-01 -7.13576019e-01 1.88896179e-01
3.74349684e-01 4.51032460e-01 7.38395810e-01 -6.26760423e-01
6.62277460e-01 1.72805226e+00 8.92880201e-01 4.16596115e-01
-1.64211512e+00 -1.20012951e+00 2.87857831e-01 1.42617911e-01
-1.09883547e+00 -4.68141288e-01 7.37416506e-01 -1.01687872e+00
6.74564421e-01 9.15169835e-01 8.49798620e-01 9.39458013e-01
-2.28852451e-01 4.58103985e-01 1.64206898e+00 -1.56550765e-01
2.36232698e-01 7.53987908e-01 6.58286273e-01 7.96538770e-01
6.89677179e-01 2.16829345e-01 -6.44839883e-01 -7.78279543e-01
6.70775920e-02 2.21821383e-01 1.38472229e-01 -3.18858296e-01
-9.33936596e-01 9.13557351e-01 6.98169768e-01 -1.43143281e-01
-4.08950478e-01 -2.28585184e-01 4.38706785e-01 1.20239727e-01
8.21410596e-01 6.70079887e-01 -9.35345963e-02 1.04251809e-01
-1.06218731e+00 5.97488225e-01 6.46026492e-01 2.76447952e-01
8.93252313e-01 -2.04097807e-01 -1.00388207e-01 6.92942441e-01
2.34416500e-01 4.23529983e-01 1.91820040e-01 -1.00202715e+00
1.49933770e-01 1.38402033e+00 2.66323447e-01 -1.42683446e+00
-7.17960835e-01 -3.16703439e-01 -9.83959675e-01 7.21079350e-01
6.87164247e-01 -1.09036767e-03 -6.04708374e-01 1.59290814e+00
6.10138476e-01 -4.99390364e-01 -4.02745575e-01 1.03510380e+00
8.90896738e-01 1.45093471e-01 7.61493266e-01 1.36454955e-01
1.36659884e+00 -3.05576492e-02 -5.40760279e-01 -3.60869020e-01
8.44532549e-01 -3.01399499e-01 1.26891017e+00 2.82778323e-01
-8.07229936e-01 -2.44680688e-01 -8.21756721e-01 -2.99237430e-01
-8.32367182e-01 -2.12687537e-01 7.51041234e-01 9.34106052e-01
-6.50309622e-01 7.59663522e-01 -3.94111305e-01 -4.99053299e-01
1.20194173e+00 2.37103790e-01 -1.03118733e-01 1.54689059e-01
-1.14698792e+00 9.42036033e-01 8.28096643e-02 -4.08630252e-01
-1.50630042e-01 -1.46376896e+00 -5.98461151e-01 -2.81867813e-02
4.07586172e-02 -8.17457795e-01 6.94452226e-01 -1.35335875e+00
-3.63019645e-01 1.30118263e+00 -2.88795173e-01 -2.76361495e-01
9.52983677e-01 8.62491727e-02 -3.84007901e-01 -5.67464769e-01
2.26774171e-01 5.36340654e-01 3.69298398e-01 -1.16034842e+00
-6.51028335e-01 -8.10795009e-01 -3.19286585e-01 -7.48154521e-02
-4.67408568e-01 4.86425966e-01 7.03280807e-01 -5.24155557e-01
-3.08257192e-01 -4.45206434e-01 -4.27498519e-01 1.84645474e-01
-5.50493062e-01 -2.29318112e-01 7.88863599e-01 -8.08486402e-01
1.39546847e+00 -1.85825026e+00 -2.79158235e-01 5.36817491e-01
7.34141052e-01 2.70012200e-01 4.04232800e-01 2.35265270e-01
-1.62171215e-01 7.21863866e-01 -1.09970933e-02 -1.09211773e-01
2.11846262e-01 -2.47546151e-01 -2.22036511e-01 4.42112356e-01
-9.14181117e-03 6.85648441e-01 -9.41906154e-01 -5.04736960e-01
1.57293901e-01 3.87135297e-01 -6.18281603e-01 6.18321672e-02
7.43118078e-02 1.19970821e-01 -1.24758691e-01 5.12184381e-01
9.59197223e-01 -1.50320247e-01 3.96436036e-01 4.40710634e-02
-4.89483804e-01 3.65896583e-01 -1.01578736e+00 6.54876173e-01
4.90552820e-02 6.68820977e-01 9.76745188e-02 -5.43267846e-01
1.10512519e+00 -3.88131022e-01 -2.29002554e-02 -6.14152133e-01
-9.01790261e-02 7.29830712e-02 -4.10926454e-02 -2.76369750e-01
5.90595067e-01 -2.02411920e-01 -1.30305532e-02 9.75534379e-01
-5.65854430e-01 3.28039408e-01 -1.59865960e-01 3.84023547e-01
6.14131808e-01 -4.38354731e-01 6.99574530e-01 -7.18708754e-01
1.11367993e-01 3.38189065e-01 5.28679907e-01 6.74041808e-01
-3.16584945e-01 1.27807513e-01 8.69244337e-01 -1.05171120e+00
-9.63920116e-01 -9.24120367e-01 -1.50229678e-01 1.59442437e+00
-2.94061750e-01 -3.80023956e-01 -3.70351076e-01 -6.42016888e-01
7.72062361e-01 9.65689301e-01 -1.09245956e+00 -1.67932481e-01
-8.11046436e-02 -1.06168926e+00 3.67285669e-01 1.24425881e-01
3.74003440e-01 -7.93639064e-01 -9.14283633e-01 -3.53356391e-01
-4.27823626e-02 -1.79180786e-01 -3.75394709e-02 -2.99720764e-01
-8.73556316e-01 -1.38595593e+00 -3.85905504e-01 -2.73633122e-01
7.77505219e-01 6.57751784e-02 1.45792747e+00 3.46837342e-01
-5.56448281e-01 -7.96604007e-02 1.95672408e-01 -1.18193889e+00
-6.83012664e-01 -6.09762706e-02 6.77586719e-02 -2.28453487e-01
1.03066123e+00 -5.10449529e-01 -7.33622849e-01 2.79671103e-01
-4.43742990e-01 5.71498692e-01 -8.37601423e-02 3.50812197e-01
-5.37429005e-02 -5.70107698e-01 7.09307313e-01 -1.56836808e+00
1.02026820e+00 -9.44631636e-01 -4.16936338e-01 1.36168361e-01
-1.22498691e+00 -9.45372730e-02 3.08575541e-01 -2.50963122e-01
-9.88298595e-01 -3.82943869e-01 5.68001747e-01 2.58167565e-01
-2.71113694e-01 3.44999760e-01 -4.78581041e-02 3.07779253e-01
1.33921969e+00 -5.19447803e-01 2.69988388e-01 -5.45141578e-01
5.47312558e-01 8.40141952e-01 1.99948192e-01 -2.89878368e-01
5.58935285e-01 5.55961967e-01 -2.69066483e-01 -5.88215888e-01
-4.92598772e-01 -2.62484640e-01 -6.32097065e-01 -5.23042023e-01
3.55020851e-01 -7.76631117e-01 -1.07121670e+00 2.49772593e-01
-6.78765655e-01 -4.99125421e-01 -2.20958740e-01 -2.90574450e-02
6.82690293e-02 -1.31121278e-01 -1.53430328e-02 -1.04883063e+00
-4.65389192e-01 -5.34738421e-01 2.39933029e-01 3.97608310e-01
-1.20208633e+00 -9.82520640e-01 8.84530246e-02 3.32890302e-01
6.59073353e-01 6.40916109e-01 1.25698745e+00 -8.26279640e-01
-1.43244192e-01 8.01974982e-02 -7.44245827e-01 -3.02542210e-01
2.00389802e-01 2.05690667e-01 -1.27388144e+00 -4.66685295e-02
-8.75742435e-01 -2.31820956e-01 7.54041314e-01 1.88120857e-01
1.28196907e+00 -7.72096455e-01 -5.71096539e-01 3.53273153e-01
9.78040874e-01 -1.67455524e-01 2.88568407e-01 4.48050439e-01
7.34019816e-01 1.24744701e+00 3.96354288e-01 6.58538818e-01
7.03576922e-01 2.97590077e-01 3.38292181e-01 -7.81325519e-01
1.62379026e-01 -4.23607796e-01 -7.11547732e-02 -1.24810882e-01
-3.24277788e-01 6.13160670e-01 -1.32979298e+00 3.05421829e-01
-1.87439954e+00 -1.13864052e+00 -4.13983852e-01 2.29306459e+00
7.92422473e-01 1.56038003e-02 7.42421329e-01 5.26660169e-03
8.00703108e-01 2.57170796e-01 -7.79087007e-01 -1.13366008e+00
-9.41882730e-02 -1.31366208e-01 4.40965831e-01 4.06911403e-01
-7.95714200e-01 5.22830963e-01 6.77544689e+00 3.64186406e-01
-1.12734783e+00 -1.56946346e-01 1.17860436e+00 -3.61182898e-01
-7.81583130e-01 -1.67673811e-01 -2.90732741e-01 4.07478690e-01
8.39649022e-01 -8.42267513e-01 2.85883546e-01 9.24661279e-01
6.75644696e-01 -2.06771180e-01 -1.16736591e+00 7.73576260e-01
-1.48011491e-01 -1.56342435e+00 5.80538362e-02 2.48845458e-01
6.49027586e-01 -7.69200549e-02 1.73151508e-01 1.80405766e-01
8.00145864e-01 -1.36381340e+00 6.68229640e-01 8.20202291e-01
9.00847435e-01 -1.05261362e+00 3.06992978e-01 2.99318820e-01
-4.25921053e-01 -3.00008863e-01 -2.36760661e-01 -7.89507687e-01
-2.93628484e-01 7.61954784e-01 -1.09080589e+00 -8.94155204e-02
8.62926722e-01 6.69392765e-01 -9.33597565e-01 8.43317866e-01
1.75989255e-01 6.01403236e-01 -1.02203220e-01 -1.60088956e-01
-3.88344139e-01 -1.08838759e-01 2.58177221e-01 1.59814405e+00
2.64768191e-02 1.14901260e-01 -1.31591156e-01 1.30086446e+00
-8.56372491e-02 3.03485125e-01 -8.31605613e-01 -1.32199541e-01
4.78790164e-01 1.25543165e+00 -3.93035144e-01 -4.89246458e-01
-1.35609612e-01 2.47933611e-01 4.56888914e-01 3.39478821e-01
-1.97772712e-01 -1.49711639e-01 1.23967648e+00 7.63584852e-01
-6.70210242e-01 1.64030775e-01 -9.97665405e-01 -6.56657994e-01
-5.45049250e-01 -1.35934365e+00 7.52070248e-01 -6.51347697e-01
-1.32961690e+00 -8.86139125e-02 4.88523394e-02 -6.55232608e-01
-5.89163899e-02 -3.78030568e-01 -7.92090714e-01 1.26318574e+00
-1.09212589e+00 -1.10512757e+00 -5.34599900e-01 2.62846202e-01
-1.63626939e-01 -1.83246717e-01 7.93178737e-01 1.09942369e-02
-6.19421840e-01 4.96510923e-01 -1.40307218e-01 -2.37056743e-02
1.01487303e+00 -1.30612421e+00 5.08961916e-01 3.19813043e-01
-3.62275571e-01 7.29346573e-01 9.57149923e-01 -8.87967169e-01
-5.90103507e-01 -1.02106154e+00 1.06243575e+00 -7.05944121e-01
4.96317178e-01 -5.00368774e-01 -1.18684733e+00 6.21996820e-01
2.04121750e-02 -3.24477911e-01 1.31656218e+00 7.69319355e-01
-6.89386308e-01 -8.04297775e-02 -1.48343325e+00 8.52014124e-01
1.03150749e+00 -4.51283842e-01 -4.42874253e-01 7.93275237e-02
2.17749700e-01 9.06132311e-02 -4.98556882e-01 1.74869508e-01
8.86624157e-01 -1.51839161e+00 8.50212276e-01 -1.17894435e+00
6.00550711e-01 1.24059126e-01 4.24745679e-01 -1.48268175e+00
-4.26134020e-01 -4.28509086e-01 2.69782305e-01 1.37361729e+00
5.49691200e-01 -8.55730951e-01 8.23581576e-01 1.54655135e+00
5.70209324e-01 -4.00444955e-01 -4.05090272e-01 -1.29545887e-03
3.10749799e-01 -5.34819424e-01 1.20040560e+00 1.60236287e+00
4.92314845e-01 -2.26415381e-01 -2.45330751e-01 -5.20662256e-02
9.90893722e-01 3.77705187e-01 1.06342041e+00 -1.65516078e+00
3.50158602e-01 -8.40759039e-01 -2.29746968e-01 1.95883483e-01
-8.88985395e-02 -1.17580128e+00 -7.80113399e-01 -1.34618497e+00
6.96793258e-01 -9.43477809e-01 -2.87640274e-01 6.29430711e-01
-4.41555738e-01 2.23445296e-01 2.78543651e-01 2.40601927e-01
-2.57910371e-01 -8.99681170e-03 7.50041008e-01 -1.45247012e-01
-3.81013364e-01 -1.15929753e-01 -1.51002073e+00 9.94374573e-01
9.08963561e-01 -5.30787706e-01 -3.95812504e-02 1.62889846e-02
7.04354286e-01 -7.58702993e-01 8.44887793e-01 -5.75581491e-01
-7.58262426e-02 -6.39971972e-01 9.00459409e-01 -3.45712632e-01
-2.41797060e-01 -3.76220137e-01 2.71537602e-01 7.21159041e-01
-6.51581585e-01 4.15347889e-02 3.05697232e-01 1.03848271e-01
3.53202075e-01 3.50335270e-01 6.58867836e-01 -1.62427738e-01
-2.84428358e-01 -8.31702054e-02 -3.15327287e-01 1.82750732e-01
8.86327624e-01 -6.86997175e-02 -8.68605375e-01 -4.19986963e-01
-6.01737797e-01 6.16650760e-01 8.02119792e-01 4.56544310e-01
2.50821859e-01 -1.21182001e+00 -8.20613801e-01 3.96612704e-01
2.39611074e-01 -5.40353954e-01 1.57351106e-01 5.77439547e-01
-2.84751356e-01 -4.67111841e-02 -6.78876758e-01 -3.90739769e-01
-1.71223533e+00 5.49758017e-01 1.64575621e-01 1.17353655e-01
-3.70599747e-01 6.34659290e-01 2.16788933e-01 -6.50622845e-01
1.23048365e-01 -5.80605455e-02 -4.11914617e-01 5.93280733e-01
8.00436080e-01 1.05850136e+00 -3.02400649e-01 -4.83029038e-01
-5.26592135e-01 1.24653764e-02 5.51152602e-02 2.08334357e-01
1.27864552e+00 -7.38887042e-02 -2.52860278e-01 6.17758751e-01
7.32116044e-01 8.70020762e-02 -9.33783710e-01 -7.48041570e-02
1.70930311e-01 -8.18479061e-01 -1.99875563e-01 -1.25968266e+00
-6.03995323e-01 1.01551843e+00 7.33921766e-01 4.33598310e-01
8.07822526e-01 -1.37617275e-01 -2.11407822e-02 -1.25807732e-01
4.88844663e-02 -1.10096037e+00 -4.80763912e-01 -1.23612791e-01
9.64299142e-01 -1.66545761e+00 5.68752348e-01 -1.81419879e-01
-6.91941023e-01 8.45634997e-01 5.51809430e-01 2.30116457e-01
6.04219079e-01 1.26973510e-01 4.68368202e-01 -3.24899614e-01
-7.81958759e-01 1.22968122e-01 4.42789167e-01 8.04936647e-01
6.10282421e-01 6.55149221e-01 -4.81686771e-01 7.20120072e-01
-5.50094664e-01 8.82640034e-02 3.77730280e-01 3.52645516e-01
-4.18602854e-01 -9.30359244e-01 -7.39033222e-01 9.67777848e-01
-3.87554109e-01 1.35218680e-01 -9.59969521e-01 8.01933527e-01
4.59382743e-01 6.52587891e-01 4.50551361e-01 -3.10458660e-01
2.55905837e-01 1.93265006e-01 -1.47161365e-01 -3.37713540e-01
-1.00571561e+00 -7.61160612e-01 4.39663410e-01 -7.07040906e-01
-1.38644308e-01 -8.05360198e-01 -1.01839006e+00 -9.49996293e-01
4.31282252e-01 1.48879170e-01 6.88507199e-01 4.71769333e-01
5.99922299e-01 3.49773429e-02 3.69056642e-01 -6.38878942e-01
-1.64733648e-01 -8.15923452e-01 -3.72577459e-01 8.49891841e-01
3.22691947e-01 -4.20468390e-01 -4.33979958e-01 -2.96893179e-01] | [8.915249824523926, 5.4866557121276855] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.